From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Subject: Re: TCP as a module Date: Sun, 27 Jul 2003 10:19:19 -0500 Sender: linux-newbie-owner@vger.kernel.org Message-ID: <200307271019.19111.eric@cisu.net> References: <000501c3538a$2cd22e30$6400a8c0@bombay> Reply-To: eric@cisu.net Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <000501c3538a$2cd22e30$6400a8c0@bombay> Content-Disposition: inline List-Id: Content-Type: text/plain; charset="us-ascii" To: sks@alum.mit.edu, linux-newbie@vger.kernel.org Have you tried CC'ing linux-kernel mailing list? I could be wrong but linux-newbie seems a bit out of place for a post dealing with kernel code development. the address is linux-kernel@vger.kernel.org and you can subscribe at www.kernel.org Just a suggestion, I hope you find the answers you're looking for there. On Saturday 26 July 2003 10:25 am, Shan Sinha wrote: > Hi everyone- > > My web searches on this seem to turn up very little. > > Many people have suggested that TCP as a loadable module would be nice. > I am actually in a position where I think it might be necessary. > > I am working on a research project on an ad-hoc 802.11 network spanning > through a local neighborhood. There are multiple people developing on > the network with dozens of nodes all running 2.4. Various kernels > across the network are updated at various times, but upgrades do happen > frequently. Occassionally, kernels are uniformally updated across the > entire network. > > My work specifically requires some TCP hacking. As a result, I would > like to insulate my work from other kernel upgrades, so as not to have > to redo work. Creating an isolated sandbox, while possible, is not > optimal, b/c the kernel upgrades are usually necessary and worth > incorporating. > > As a result, I would like to implement a new transport layer module > based on the TCP code, but with its own address family. I realize I > might have to re-compile a few tools, but I use few enough of those, > that it should be no problem. > > Looking at the code, I assume that a starting step would be to figure > out how to implement TCP as a module. > > + Has anyone actually done this or know of someone who has? Any > advice/links would be useful. > + Or does anyone know for certain that doing this is just not possible? > If so, why not? > > I'm willing to live with idiosyncracies in how such a TCP module would > behave for making my development life easier, I think this is necessary! > The idea would be that once some of my work actually demonstrates > positive results, to re-implement back into the core kernel code. > > If you have a better solution, I would love to hear it! > > Cheers- > Shan > > Shan Sinha > Networks and Mobile Systems > Laboratory for Computer Science and AI > MIT > > For information on the research project, visit > http://pdos.lcs.mit.edu/roofnet > > - > To unsubscribe from this list: send the line "unsubscribe linux-newbie" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.linux-learn.org/faqs -- ---------------------- Eric Bambach Eric (at) CISU (dot) net ---------------------- - To unsubscribe from this list: send the line "unsubscribe linux-newbie" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.linux-learn.org/faqs