From owner-freebsd-hackers Mon Aug 31 09:56:30 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id JAA17865 for freebsd-hackers-outgoing; Mon, 31 Aug 1998 09:56:30 -0700 (PDT) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from word.smith.net.au (castles346.castles.com [208.214.167.46]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id JAA17841 for ; Mon, 31 Aug 1998 09:56:24 -0700 (PDT) (envelope-from mike@word.smith.net.au) Received: from word.smith.net.au (LOCALHOST [127.0.0.1]) by word.smith.net.au (8.9.1/8.8.8) with ESMTP id JAA20742; Mon, 31 Aug 1998 09:53:01 GMT (envelope-from mike@word.smith.net.au) Message-Id: <199808310953.JAA20742@word.smith.net.au> X-Mailer: exmh version 2.0.2 2/24/98 To: Darren Reed cc: hackers@FreeBSD.ORG Subject: Re: TCP performance. In-reply-to: Your message of "Mon, 31 Aug 1998 20:15:23 +1000." <199808311016.DAA24028@hub.freebsd.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 31 Aug 1998 09:53:00 +0000 From: Mike Smith Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > > Well, FreeBSD TCP performance sucks or it does some things which can > make it suck badly. It might be more fair to say "FreeBSD 2.2.6 using 3c589d and 'zp' driver suffers slow FTP send? receive? performance with Sun LX running SunOS 4.1.x over with and " > >From simple observations (hub leds, ftp's meter), ftp'ing from a laptop > (3c589d) running 2.2.6 to SunOS4 (LX) is considerably worse than in the > other direction. > > Although I do suspect the 3c589d (or its driver - zp ?) > do not handle ethernet collisions at all in a good way to be a possible > cause. The '589 handles collisions correctly; there's no other way to handle them and work. The driver isn't involved in handling collisions; it gives a packet to the card, and waits for the card to report results. Having said that, bear in mind that the 'zp' driver is a bandaid that only exists to aid installation, and you may well be better off using the 'ep' driver inside the pccard framework. If it counts for anything, I regularly move data in and out of both a '589 and an NE2000 clone pccard at around the 1MB/sec rate, so I have a direct counter-example if you feel that's the way to go. Please, if you want to complain about FreeBSD's performance, consider investigating the problem a little to give some substance to your complaint. Without it, all you're doing is abusing the product; with it, you may help us improve, and hopefully reduce your need to complain in the future. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message