Date: Sun, 20 Jan 2008 02:53:28 +0200 From: Alexander Motin <mav@FreeBSD.org> To: Michael MacLeod <mikemacleod@gmail.com> Cc: freebsd-net@freebsd.org, Julian Elischer <julian@elischer.org>, Nikos Vassiliadis <nvass@teledomenet.gr> Subject: Re: Multilink PPP Download Speeds With Round-Robin Packets Message-ID: <47929B88.3050303@FreeBSD.org> In-Reply-To: <e8f0b580801191512nee559ady6620adf64a9d148b@mail.gmail.com> References: <1200479046.00010232.1200466203@10.7.7.3> <1200479105.00010249.1200468002@10.7.7.3> <1200482587.00010258.1200469801@10.7.7.3> <1200489793.00010290.1200478201@10.7.7.3> <1200518607.00010486.1200507002@10.7.7.3> <478E834E.4080302@FreeBSD.org> <e8f0b580801191347y42b3a041y821dc4d8805a168b@mail.gmail.com> <47927858.4050702@FreeBSD.org> <e8f0b580801191512nee559ady6620adf64a9d148b@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Michael MacLeod wrote: >> After peer rejected multilink once, mpd is not trying to negotiate it >> any more. I am not sure it is a correct behaviour, but Cisco manuals >> recommend to configure LAC in a way that avoids rejections. >> >> User-land ppp seems to have specific workaround for this case while mpd >> does not. I don't very like this idea, but probably I could add this if >> you like to test it. > > If that's what it takes, I'm happy to test it. Also, the engineers at > my ISP are remarkably approachable, and if you tell me what to ask > them I can probably get you a decent answer pretty quick. After rereading RFC I have found that your provider operation possibly not perfect but it is correct. It is possible to hint peer to use multilink after rejecting by sending NAK. I have made some changes to handle that in mpd5 CVS and merged them into mpd4 branch. You can get new mpd version to test from mpd CVS repository: https://sourceforge.net/cvs/?group_id=14145 -- Alexander Motin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?47929B88.3050303>