Date: Wed, 30 Dec 1998 00:39:12 +0900 From: Kenjiro Cho <kjc@csl.sony.co.jp> To: Poul-Henning Kamp <phk@critter.freebsd.dk> Cc: atm@FreeBSD.ORG Subject: Re: ATM in FreeBSD, status requested! Message-ID: <199812291539.AAA13414@hotaka.csl.sony.co.jp> In-Reply-To: Your message of "Tue, 29 Dec 1998 14:24:24 %2B0100." <78560.914937864@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp <phk@critter.freebsd.dk> said: >> Now for a second round. >> Currently the only piece of hardware supported by Cranors code but >> not by the HARP code is the Adaptec if I have understood this right. Yes, but the issue is not only the supported hardware. >> The reson for me asking about removing Cranors stuff, is that I don't >> think we are serviced well with two competing implementations in the >> tree, and of the two HARP is much more complete and usable, and most >> importantly: it is actively maintained. What makes you think Cranor's driver is poorly maintained? As a maintainer of the driver, I never heard of complaints. Many minor problems have been fixed by Chuck and me before other people even notice. >> I am aware that many reseach projects run on Cranors stuff, but if >> they are not even considering upgrading from FreeBSD 2.1 for >> instance, they do not really factor into the equation of FreeBSD >> 3.1 and further down the road, do they ? :-) They do. FreeBSD has been the best platform in network research and we should be proud of the stability of FreeBSD 2.1 :-) Let me put this way. Cranor's driver vs HARP is somewhat similar to slip vs ppp. Even when ppp seems to become a super set of what slip provides for normal use, we still want to keep slip. It is difficult to touch the hairy ppp code or to evaluate a single component of ppp. New ideas will be first implemented into simpler slip, and then, if proved useful, it will be brought into production ppp. Cranor's driver has been playing this role and it will continue to do so for at least another couple of years. For instance, IPv6 people needed the point-to-point interface model for ATM because multicast is essential to IPv6 but multicast in the Classical IP over ATM model is unnecessarily complex for them. >> I think therefore, that unless circumstances change in the meantime, >> that you should expect the Cranor driver to disapper from >> FreeBSD-current sometime after Q299, and prepare to either maintain >> it yourself (some of you already do) or switch to HARP. I still don't see why you are so eager to remove Cranor's driver at this stage. Can we just wait and see how people use ATM? As for ALTQ, I'm not eligible to evaluate it :-) but the issue will be raised along with the IPv6 integration. Both INRIA and KAME already merged ALTQ in their release. --Kenjiro To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-atm" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199812291539.AAA13414>