Date: Sun, 20 Aug 2000 13:36:04 -0700 (PDT) From: Archie Cobbs <archie@whistle.com> To: Brian Somers <brian@Awfulhak.org> Cc: cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/usr.sbin/ppp README.changes mp.c Message-ID: <200008202036.NAA51971@bubba.whistle.com> In-Reply-To: <200008190941.e7J9fSF94932@hak.lan.Awfulhak.org> from Brian Somers at "Aug 19, 2000 10:41:28 am"
next in thread | previous in thread | raw e-mail | index | archive | help
Brian Somers writes: > > > As ppp currently only does IP as an NCP, and IP packets are allowed > > > to be delivered out-of-order, I think this is ok. > > > > If you're doing Van Jacobson header compression, for example, > > with connection ID compression, out of order packets can really > > screw things up. > > True, but as ppp will only ever generate one out-of-order packet > under normal circumstances, this just means that if the two packets > in question are from the same tcp stream, the first will get dropped. > > Perhaps it would be better if PROTO_IP packets were only sent when > there's only one link open *AND* no other links are in any sort of > trying-to-open state ? I'm sure it's fine the way it is.. in practice a small packet reordering window is probably not going to lead to big problems or anything. It just falls into that cateogory of things that have the potential to lead to really mysterious, hard-to-debug behavior later on, and it's good to at least be aware of that possibility. For example, if in the future someone implements a new network protocol that is particularly sensitive to it. -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
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200008202036.NAA51971>