From owner-freebsd-bugs@FreeBSD.ORG Tue Apr 24 03:01:27 2007 Return-Path: X-Original-To: freebsd-bugs@freebsd.org Delivered-To: freebsd-bugs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 1DF3D16A406 for ; Tue, 24 Apr 2007 03:01:27 +0000 (UTC) (envelope-from brett@unix.org.au) Received: from ipmail01.adl2.internode.on.net (ipmail01.adl2.internode.on.net [203.16.214.140]) by mx1.freebsd.org (Postfix) with ESMTP id 96A4813C459 for ; Tue, 24 Apr 2007 03:01:25 +0000 (UTC) (envelope-from brett@unix.org.au) X-IronPort-AV: E=Sophos;i="4.14,444,1170595800"; d="scan'208";a="118704715" Received: from home.unix.org.au (HELO [127.0.0.1]) ([59.167.215.117]) by ipmail01.adl2.internode.on.net with ESMTP; 24 Apr 2007 12:15:44 +0930 Message-ID: <462D6F65.8020500@unix.org.au> Date: Tue, 24 Apr 2007 12:45:57 +1000 From: Brett Johnson User-Agent: Thunderbird 1.5.0.10 (Windows/20070221) MIME-Version: 1.0 To: freebsd-bugs@freebsd.org Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Broadcom BCM5704 Driver Problem X-BeenThere: freebsd-bugs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Bug reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 24 Apr 2007 03:01:27 -0000 Hello All, I've just deployed several IBM (x3650) servers all running 6.2-Release. Each of these systems has two on board gigabit ethernet cards using the BCM5708 chipset which work fine. (bce driver) Four of these machines have an added dual port gigabit pci-x card with the BCM5704 chipset, these cards experience packet loss. (bge drivers) I have tried setting the speed/duplex to gig/full, gig/half, 100/full, 100/half and I still get packet loss of upto 30% across the interfaces. At present I'm running the cards at 10base/utp and that seems to have fixed the packet loss problem. Testing was done using 2 cisco 2960's and a dlink des-1024r+ switch, same problem using any of the three switches. bge0: mem 0xc7ef0000-0xc7efffff irq 24 at device 3.0 on pci8 bge1: mem 0xc7ee0000-0xc7eeffff irq 25 at device 3.1 on pci8 bce0: mem 0xce000000-0xcfffffff irq 16 at device 0.0 on pci3 bce1: mem 0xca000000-0xcbffffff irq 17 at device 0.0 on pci6 The machines in question are load balancing several web servers and the 10mbit solution is causing nearly as many problems as the initial packet loss problem, any suggestions/ideas would be greatly appreciated. Regards, Brett Johnson brett@unix.org.au