From owner-freebsd-hackers Sun Sep 21 19:01:00 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id TAA02086 for hackers-outgoing; Sun, 21 Sep 1997 19:01:00 -0700 (PDT) Received: from awfulhak.demon.co.uk (awfulhak.demon.co.uk [158.152.17.1]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id TAA02068; Sun, 21 Sep 1997 19:00:42 -0700 (PDT) Received: from gate.lan.awfulhak.org (localhost [127.0.0.1]) by awfulhak.demon.co.uk (8.8.5/8.8.5) with ESMTP id CAA05366; Mon, 22 Sep 1997 02:09:08 +0100 (BST) Message-Id: <199709220109.CAA05366@awfulhak.demon.co.uk> X-Mailer: exmh version 2.0zeta 7/24/97 To: Eivind Eklund cc: Brian Somers , hackers@FreeBSD.ORG Subject: Re: cvs commit: src/usr.sbin/ppp lcp.c In-reply-to: Your message of "Mon, 22 Sep 1997 02:34:10 +0200." <199709220034.CAA24672@bitbox.follo.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 22 Sep 1997 02:09:08 +0100 From: Brian Somers Sender: owner-freebsd-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > > > > brian 1997/09/21 16:02:38 PDT > > > > Modified files: (Branch: RELENG_2_2) > > usr.sbin/ppp lcp.c > > Log: > > MFC: Sleep for a second before sending the first LCP > > config request. This stops us from squirting stuff > > down a line that still has ECHO turned on because the > > peer hasn't had a chance to start yet. > > Lead to the cause by: Greg Lehey > > Isn't detecting this what the magic number more-or-less was introduced > for? You can catch looping LCP-negotiation by looking for the same > magic number, and it really seems a pity to waste a second on > connection, as this could be running on really fast links (like TCP or > ISDN links). Besides, if the other end is slow, 1 second might not be > enough. When we connect to a machine that is running a bit slow, the client-side ppp manages to send 6 LCPs, receive the same 6 back (due to echo being on) and give up - all before the server-side ppp gets to turn ECHO off. Maybe I should incrementally "back out" when receiving config requests w/ the same magic - ie, wait .2 seconds after the first dup magic, .4 after the second etc. > Eivind. > > Sleep for a second before sending the first LCP > > config request. This stops us from squirting stuff > > Don't forget that some boxes, if you wait a second or more, will fall > back into ``terminal server'' mode, and then the user gets hosed. The > Shiva LanRover Access Switch is one such box in wide circulation. Hmm, looks like the incremental backout may be a better idea. > -GAWollman -- Brian , , Don't _EVER_ lose your sense of humour....