From owner-freebsd-current Wed Aug 5 07:04:05 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA19648 for freebsd-current-outgoing; Wed, 5 Aug 1998 07:04:05 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from bright.ny.otec.com (bright.ny.otec.com [209.3.16.125]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA19589; Wed, 5 Aug 1998 07:03:58 -0700 (PDT) (envelope-from bright@hotjobs.com) Received: from localhost (bright@localhost) by bright.ny.otec.com (8.8.8/8.8.8) with SMTP id KAA19895; Wed, 5 Aug 1998 10:04:27 -0500 (EST) (envelope-from bright@hotjobs.com) X-Authentication-Warning: bright.ny.otec.com: bright owned process doing -bs Date: Wed, 5 Aug 1998 10:04:27 -0500 (EST) From: Alfred Perlstein X-Sender: bright@bright.ny.otec.com To: Bill Paul cc: current@FreeBSD.ORG Subject: Re: Etherlink XL driver In-Reply-To: <199808050548.WAA16103@hub.freebsd.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG using the first version of the driver posted on 100mb segement seemed to have not great performance. The admin here said he was able to get ~8megs/sec over the segment using Linux and 11megs/sec via Solaris, the FreeBSD driver was never over 4.5megs and averaged around 3-3.5. This is ftp transfers. It seems to grab a LOT of data, then "stall" then grab, stall over and over. I know this is an older driver, i'll compile him a new kernel with the latest snapshot to see if there is an improvement. Just thought to let you know. This is on a -current box with a PCI card 905b on 100/TX. The driver was acting in full duplex mode, perhaps it should have been set to half-duplex and it was tripping on itself? The driver seems quite stable though, playing mp3s over NFS, while not a true "benchmark" works fine. My personal workstation is on a 10mbit segment and has had an uptime of 2days so far, no problems i can see. On a 10mbit segment though the driver (perhaps it's the card as i've not heard great reviews about the 905b) seems not be as fast as the de driver on a tulip chipset. Again, sorry this is a older version of the driver. Both boxes will be setup with the newer snapshots and if you're interested i'll post the performance. Thank you, Alfred Perlstein -- Programmer @ HotJobs Inc. [- http://www.hotjobs.com/ -] |-- There are operating systems, and then there's BSD. \-- http://www.freebsd.org/ > (Incidentally, since this is going to -hackers: if you're having a > problem with the XL driver generating an error message that says > 'MII without any PHY!' and you have a 3c905B card, please grab the > latest version of the driver from www.freebsd.org/~wpaul/3Com and > let me know if the problem goes away.) > > The -current branch may be bleeding edge, but I don't feel that means > that I shouldn't try my best to work out all the bugs before I commit > new code to it, which is what I'm doing, at least when I'm not being > interrupted by Real Work (tm). > > That said, my plan is to commit the 3.0 version to -current sometime > next week. After a couple more weeks there, I'll commit the 2.2 > version. > > -Bill > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-hackers" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message