From owner-freebsd-hackers Fri Jun 28 08:39:38 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id IAA08733 for hackers-outgoing; Fri, 28 Jun 1996 08:39:38 -0700 (PDT) Received: from gateway.cybernet.com (gateway.cybernet.com [192.245.33.1]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id IAA08724 for ; Fri, 28 Jun 1996 08:39:34 -0700 (PDT) Received: from spiffy.cybernet.com (spiffy.cybernet.com [192.245.33.55]) by gateway.cybernet.com (8.6.12/8.6.12) with SMTP id LAA28049; Fri, 28 Jun 1996 11:40:55 -0400 Message-ID: X-Mailer: XFMail 0.4-prerelease [p0] on FreeBSD Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: Date: Fri, 28 Jun 1996 11:37:35 -0400 (EDT) Organization: Cybernet Systems Corporation From: (Mark J. Taylor) To: Jaye Mathisen Subject: Re: de0: Transmission timeout? Cc: Kurt Olsen , hackers@freebsd.org Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk On 19:09:20 Jaye Mathisen wrote: >>Hmmm, I'm a bit skeptical of this explanation, for the following reason. >The same kernel and source tree is on 4 identical boxes (4 P120's), and >the 1 P6, and the P6 only has this problem. Swapping cards and slots >doesn't fix it either, it's only on the P6. > >So I'm thinking a hardware problem of somekind, but I can't imagine what. >3com cards work fine, the adaptec works fine, just the darn network card. > >On Thu, 27 Jun 1996, Kurt Olsen wrote: > >> I've seen this same behavior and a knowledgable friend tells me it's a >> common bug. Claims that it expires the arp entry for the default router, >> so you can't talk to it from anywhere outside the subnet. A work-around >> is to have either a cron job that pings out of your subnet every few minutes, >> or just do what I do and: >> >> % ping -i 300 >& /dev/null & >> >> I haven't look into the kernel to see if this is the case though, but the >> ping does the job (as well as logging in from the console, then telneting >> out.) >> >> Kurt >> I too have only experienced this on our one P6-200. We have several P5 machines with the SMC Ethernet (de0), and they do not exhibit this message. Sounds like a loop-counter in the de driver code instead of a sleep(). -Mark Taylor mtaylor@cybernet.com