From owner-freebsd-hardware Tue Sep 7 12:24:31 1999 Delivered-To: freebsd-hardware@freebsd.org Received: from thoth.mch.sni.de (thoth.mch.sni.de [192.35.17.2]) by hub.freebsd.org (Postfix) with ESMTP id F1C7915B56 for ; Tue, 7 Sep 1999 12:24:22 -0700 (PDT) (envelope-from andre.albsmeier@mchp.siemens.de) X-Envelope-Sender-Is: andre.albsmeier@mchp.siemens.de (at relayer thoth.mch.sni.de) Received: from mail2.siemens.de (mail2.siemens.de [139.25.208.14]) by thoth.mch.sni.de (8.9.3/8.9.3) with ESMTP id VAA05679 for ; Tue, 7 Sep 1999 21:23:45 +0200 (MET DST) Received: from curry.mchp.siemens.de (curry.mchp.siemens.de [139.25.42.7]) by mail2.siemens.de (8.9.3/8.9.3) with ESMTP id VAA29249 for ; Tue, 7 Sep 1999 21:23:45 +0200 (MET DST) Received: (from daemon@localhost) by curry.mchp.siemens.de (8.9.3/8.9.3) id VAA76193 for ; Tue, 7 Sep 1999 21:23:45 +0200 (CEST) Date: Tue, 7 Sep 1999 21:23:43 +0200 From: Andre Albsmeier To: Peter Jeremy Cc: hardware@FreeBSD.ORG Subject: Re: Decent 100baseTX PCMCIA card wanted Message-ID: <19990907212343.A99954@internal> References: <99Sep6.105734est.40333@border.alcanet.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0pre2i In-Reply-To: <99Sep6.105734est.40333@border.alcanet.com.au> Sender: owner-freebsd-hardware@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Mon, 06-Sep-1999 at 10:59:26 +1000, Peter Jeremy wrote: > Does anyone have any suggestions for a PCMCIA card that can snoop a > heavily loaded 100baseTX LAN using a Celeron/300. > > I've tried a 3Com 3C574BT, but it's far too slow (basically, it can > just keep up with a 10baseT network). I just asked the same question on -mobile. I have got an "Intel" "EtherExpress(TM) PRO/100 PC C" but I can't get more than about 1.3 MBytes/sec through it. It appears to be a driver/hardware issue: /* * Card interrupt handler: should return true if the interrupt was for us, in * case we are sharing our IRQ line with other devices (this will probably be * the case for multifunction cards). * * This function is probably more complicated than it needs to be, as it * attempts to deal with the case where multiple packets get sent between * interrupts. This is especially annoying when working out the collision * stats. Not sure whether this case ever really happens or not (maybe on a * slow/heavily loaded machine?) so it's probably best to leave this like it * is. * * Note that the crappy PIO used to get packets on and off the card means that * you will spend a lot of time in this routine -- I can get my P150 to spend * 90% of its time servicing interrupts if I really hammer the network. Could * fix this, but then you'd start dropping/losing packets. The moral of this * story? If you want good network performance _and_ some cycles left over to * get your work done, don't buy a Xircom card. Or convince them to tell me * how to do memory-mapped I/O :) */ and /* * Now get the packet, including the Ethernet header and trailer (?) * We use programmed I/O, because we don't know how to do shared * memory with these cards. So yes, it's real slow, and heavy on * the interrupts (CPU on my P150 maxed out at ~950KBps incoming). */ Since it is named EtherExpress I thought it would be a good deal :-( -Andre To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message