From owner-freebsd-hackers Thu Jan 11 10:40:45 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id KAA09375 for hackers-outgoing; Thu, 11 Jan 1996 10:40:45 -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 KAA09369 for ; Thu, 11 Jan 1996 10:40:38 -0800 (PST) Received: (from nate@localhost) by rocky.sri.MT.net (8.6.12/8.6.12) id LAA21741; Thu, 11 Jan 1996 11:43:10 -0700 Date: Thu, 11 Jan 1996 11:43:10 -0700 From: Nate Williams Message-Id: <199601111843.LAA21741@rocky.sri.MT.net> To: dennis@etinc.com (dennis) Cc: hackers@freebsd.org Subject: Re: pppd vs ijppp In-Reply-To: <199601111824.NAA00402@etinc.com> References: <199601111824.NAA00402@etinc.com> Sender: owner-hackers@freebsd.org Precedence: bulk [ My last post on the subject, I promise! > >> >Or, we could all of those features in the kernel, increase your memory > >> >use by a couple 100K (always, even if you don't use it), and it would > >> >take us 6 months to get it working. :)' > >> > >> (*) Fine print. Per connection. Perhaps less than 5%. > > Your premise that you can't have your cake and eat it too is much more of > a matter of a lack of design foresight than fact. There are clearly solutions > which address both the functional and performance issues. The benefits > of pppij can be integrated into a kernel ppp with gains rather than losses > of overall appeal. If you look above, I admitted that those solutions *could* be integrated into the kernel, but it would take time and memory. Maybe the memory usage seems a bit excessive, but although it *could* be done in less generally speaking the first version is always a pig. Since no-one around here has time to do the integration, it's not going to get done unless you want to do given your well publicized experience in doing such things. Then you can show us how easily and compactly it can be done, and shut me up. :) Nate