Date: Sat, 17 Jan 2004 22:49:06 -0500 From: Mike Tancsa <mike@sentex.net> To: Andre Oppermann <andre@freebsd.org> Cc: freebsd-net@freebsd.org Subject: Re: PPPoE problem: "Too many LQR packets lost" Message-ID: <h20k00hj1q413tlqdsirk9p4q6kp7bbvi0@4ax.com> In-Reply-To: <4009CF11.42C481E5@freebsd.org> References: <4009C0D5.2070209@cse.unsw.edu.au> <4009CF11.42C481E5@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
I tried emailing the owner of ppp but got no response. http://www.freebsd.org/cgi/query-pr.cgi?pr=3Dkern/11293 Seems to indicate others have run into this as well. I can confirm that here in Canada, FreeBSD's ppp is indeed broken when connecting to Juniper's ERX as part of a PPPoE session. I worked around the issue locally by patching ppp to force to use just LCP echo. Ideally, the easiest solution would be a way to force LCP echo in ppp.conf. WIthout it, there is no built in way to check to see if the link is up or down when connected to an ERX. --- lcp.c =20 +++ lcp.c.patch=20 @@ -575,6 +575,11 @@ log_Printf(LogLCP, "%s: LayerUp\n", fp->link->name); physical_SetAsyncParams(p, lcp->want_accmap, lcp->his_accmap); lqr_Start(lcp); + +=20 + p->hdlc.lqm.method =3D LQM_ECHO; + + hdlc_StartTimer(&p->hdlc); fp->more.reqs =3D fp->more.naks =3D fp->more.rejs =3D = lcp->cfg.fsm.maxreq * 3; fixes it for me. Does it actually work with anyone's implementation ? If not, perhaps it would be best to just disable it until someone fixes it. ---Mike On Sun, 18 Jan 2004 01:10:57 +0100, in sentex.lists.freebsd.net you wrote: >nil000@cse.unsw.edu.au wrote: >>=20 >> G'day, >>=20 >> Since my ADSL connection was moved at my ISP's end from one set of >> equipment (norte shasta?) to another (unisphere?), I've been having >> problems with my PPPoE sessions. >>=20 >> With "enable lqr" in my ppp.conf, the connection terminates after a = few >> mintues with "Too many LQR packets lost". I've included log output >> below. According to these posts, > >I remember one or two open PRs on LQR in ppp. Try to do a search for >owner brian in the PR database. I don't know what happend to Brian, >but he hasn't done much in past few month to a year and I haven't heard >anything from him since. > >If any of these PRs apply to your problem I could have a look into it >and commit a working fix. I have PPPoE ADSL stuff around here because >at my ISP we are a wholesale partner of our beloved local telco. But >we have only a Cisco 7200 as endpoint. No shasta (dead?) or unisphere >(called Juniper E-Series these days).
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?h20k00hj1q413tlqdsirk9p4q6kp7bbvi0>