From owner-freebsd-hackers Mon Jul 27 11:59:43 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA14940 for freebsd-hackers-outgoing; Mon, 27 Jul 1998 11:59:43 -0700 (PDT) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from alpo.whistle.com (alpo.whistle.com [207.76.204.38]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA14746 for ; Mon, 27 Jul 1998 11:58:40 -0700 (PDT) (envelope-from ambrisko@whistle.com) Received: (from daemon@localhost) by alpo.whistle.com (8.8.5/8.8.5) id LAA22592; Mon, 27 Jul 1998 11:46:41 -0700 (PDT) Received: from crab.whistle.com(207.76.205.112), claiming to be "whistle.com" via SMTP by alpo.whistle.com, id smtpdk22585; Mon Jul 27 18:46:34 1998 Received: (from ambrisko@localhost) by whistle.com (8.8.8/8.8.7) id SAA04478; Mon, 27 Jul 1998 18:35:01 GMT (envelope-from ambrisko) From: Doug Ambrisko Message-Id: <199807271835.SAA04478@whistle.com> Subject: Re: AMD PCNet/FAST cards; suppliers? In-Reply-To: <199807270826.BAA22455@usr08.primenet.com> from Terry Lambert at "Jul 27, 98 08:26:10 am" To: tlambert@primenet.com (Terry Lambert) Date: Mon, 27 Jul 1998 18:34:59 +0000 (GMT) Cc: mike@smith.net.au, tlambert@primenet.com, rjs@fdy2.demon.co.uk, hackers@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL29 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Terry Lambert writes: | > > You should talk to Doug Ambrisko about this. | > | > I would hope he'll drop into the discussion. | | Unfortunately, he has a life. 8-). Atleast on weekends. | > > I believe there were also problems with the hardware | > > autodetection, unlike the 82558 part. | > | > Detection of the PCI parts is, not surprisingly, trivial. | | Autodetection of the 100 vs. 10 rate, not autodetection of the device. One of the issues is that at the time you couldn't set the card in 10 or 100 mode since the driver support wasn't there. Connecting multiple autosense cards via a cross cable sometimes causes both to bounce between 10 and 100 when trying to auto-detect. Using a hub got rid of this problem. Routing was a big issue. At 100BaseT speeds and one AMD card things were okay. I forget the numbers but "acceptable" not as good as DEC or Intel. The problem came as soon as you tried routing through two cards. performance would fall through the floor <.03Mbit/s. I also tried them with SCO Openserver (Personal edition). Routing was again terrible with SCO non-routing was okay. FYI SCO was <1/2 the routing performance of FreeBSD with good Intel cards. FreeBSD with our version of natd routing faster then SCO. Since AMD wasn't willing to put any effort into trying to reproduce the SCO case we dropped that idea (they blamed FreeBSD and SCO even though they supplied the SCO driver). I'm not sure if we still have the eval cards. BTW testing either card separately was okay, only when routing between cards or running both cards at the same time was there a problem. This problem was similar to the ISA AMD cards that bus master. One card is fine, start routing and things blow up due to bus master issues. This also means an AMD ISA card with a Adaptec 1542 is a bad idea (which I had at home). Doug A. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message