Date: Fri, 08 Dec 1995 09:25:36 -0800 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: "Michael C. Newell" <mnewell@lupine.nsi.nasa.gov> Cc: Joe Greco <jgreco@brasil.moneng.mei.com>, Angelo Turetta <turetta@trust.stylo.it>, freebsd-hackers@freebsd.org Subject: Re: User space PPP & leased line Message-ID: <28835.818443536@time.cdrom.com> In-Reply-To: Your message of "Fri, 08 Dec 1995 11:34:00 EST." <Pine.SUN.3.91.951208112808.17619E-100000@lupine.nsi.nasa.gov>
next in thread | previous in thread | raw e-mail | index | archive | help
> We tried using the user mode PPP on our systems for a while, but under > very heavy load after 2-3 minutes the line ALWAYS dropped. No indication > of what caused it. We use the identical setup with the kernel PPP (pppd) > and run for days with no line drops. We sent mail to the list a couple > of times but never got responses, so we just stuck with the kernel mode code. I'm afraid that the ijppp stuff is kind of in a special category given that most of it's in Japanese and nobody really pretends to understand it anyway. :-) I have had similar problems with it, I'm afraid, and feel this to be a subtle bug somewhere because sometimes the line drops and ppp just sort of sits there and sometimes it'll drop followed by ppp dumping core. I talked with Atsushi about it a few times but he's been pretty busy with work and my switching over to SLIP (for other reasons) kind of put an end to our testing. The best I can suggest is compiling it with debugging information and then using gdb's `attach' command to try and see what its internal state is when the line drops. It could be that the problem will leap right out at you (or then again, maybe not :-). The debugging facilities are also pretty advanced in ijppp, so it may be that you can modify it to tell you itself what's going wrong if you add more debugging to whatever section deals with line drops (or, more likely, drops it itself in response to some erroneous condition). > One thing we DID notice was that the transfer speeds using the user mode > code were SIGNIFICANTLY higher than those using the kernel mode code - > for example, transferring a medium size binary file (one that could be > fully transferred in < 2 minutes :-) using pppd takes an average of > 2.3KB/s whereas with ijppp it's more like 3.3KB/s (using a V.fc modem > connection at 26.4Kb/s). I'm not sure why this should be, but given the That's also worth investigating. I can't see any inefficiency that should result in *that* much slowdown, and my own tests showed virtually no difference between the two. Perhaps tcpdump might not be a bad idea - you may find that there's something about your setup that ppp just doesn't like. Jordan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?28835.818443536>