From owner-freebsd-hackers Thu Jan 11 21:19:39 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id VAA28886 for hackers-outgoing; Thu, 11 Jan 1996 21:19:39 -0800 (PST) Received: from rocky.sri.MT.net (rocky.sri.MT.net [204.182.243.10]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id VAA28880 for ; Thu, 11 Jan 1996 21:19:35 -0800 (PST) Received: (from nate@localhost) by rocky.sri.MT.net (8.6.12/8.6.12) id WAA23433; Thu, 11 Jan 1996 22:22:08 -0700 Date: Thu, 11 Jan 1996 22:22:08 -0700 From: Nate Williams Message-Id: <199601120522.WAA23433@rocky.sri.MT.net> To: "Michael C. Newell" Cc: hackers@FreeBSD.org Subject: Re: pppd vs ijppp In-Reply-To: References: Sender: owner-hackers@FreeBSD.org Precedence: bulk Michael C. Newell writes: > On Thu, 11 Jan 1996, Jaye Mathisen wrote: > 1. Someone (Nate?) was working on installing predictor-1 compression into > kernel ppp. Is this available for alpha/beta testing? Nope. I got a predictor-1 module from the Linux author which is *extremely* Linux-centric. It also has some patent problems with the CRC algorithm it uses (which is why it's not distributed publically), so I hope to use it as a basis for bringing into our pppd, using the predictor-1 module out of ijppp. Based on my work-load and my lack of free time, I don't expect to have anything for a couple weeks minimum. My first priority for FreeBSD is to look at the PCCARD patches which are supposed to be release RSN. > 2. Has there been any progress in making ijppp a tad more stable? Doug Rabson posted that he was working on the code, and Poul just fixed a couple minor bugs. [ kernel ppp more robust than user ppp ] > But... while [ user-ppp ] WAS up I got REALLY REALLY REALLY good > throughput; due, I understand, to predictor-1. So... Yep. > I'd REALLY like to see predictor-1 added to the kernel mode PPP. I don't > need dialers etc. (although the "connect" command on kernel ppp coupled > with chat does a nice job for me...) Then I could have a nice solid ppp > WITH compression. Happy happy joy joy! That's my goal as well. > If the person doing the predictor-1 stuff needs help/testers, please > lemme know!! :-) If you think you could do the work, I could send you the module. However, I'm not allowed to post it publically. > BTW- One other nice thing would be to add to the kernel ppp man page the > hint that one should NEVER use the proxyarp command [other hint: never > say never... :-)]. I had all kinds of bizzarre problems (mostly with > pppd's not going away) 'till I removed it per the suggestion of someone > from the list; instantly things started working well! :-)! It's a bug in FreeBSD handling of external ARP, and not a bug in PPP. Unfortunately, I'm now kernel hacker and none of the network gurus are interested in helping me fix the problem. :( Nate