From owner-freebsd-stable Thu Jun 6 6:27:22 2002 Delivered-To: freebsd-stable@freebsd.org Received: from alogis.com (firewall.solit-ag.de [212.184.102.1]) by hub.freebsd.org (Postfix) with ESMTP id BDB5537B406 for ; Thu, 6 Jun 2002 06:27:13 -0700 (PDT) Received: from alogis.com (kipp@clausthal.int1.b.intern [10.1.1.30]) by alogis.com (8.11.1/8.9.3) with ESMTP id g56DQbl29840; Thu, 6 Jun 2002 15:26:37 +0200 (CEST) (envelope-from holger.kipp@alogis.com) Message-ID: <3CFF5EEF.731DC0A6@alogis.com> Date: Thu, 06 Jun 2002 15:09:03 +0200 From: Holger Kipp X-Mailer: Mozilla 4.7 [en] (X11; U; Linux 2.2.13 i686) X-Accept-Language: en MIME-Version: 1.0 To: Greg Rumple Cc: FreeBSD-STABLE Mailing List Subject: Re: fxp0 timeout errors References: <200205271231.29155.paul@nerdlabs.com> <002401c20d1f$a3931d70$8000000a@Bryce> <20020606125825.GD55256@zaphon.llamas.net> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Greg Rumple wrote: > > I have a stack of dual port intel 10/100 cards that this happens on time > and time again. I after spending weeks trying to solve this problem (it > does not happen under Linux btw) gave up finally, and purchased a > different card for the machines that run FreeBSD. The circumstances to > make this happen purely involve doing heavy volumes of ethernet traffic > (a ping -f or a nice large rsync suffice). It will immediately happen, > and sometimes it will recover and work again after 2-3 minutes, > sometimes it'll stay dead forever. It purely seems limited to this > stack of cards, as I have other cards that do not suffer from the same > problem (single port cards). If anyone has any clues, I would be glad > to assist in helping track this down as well. Sounds like the problem I have - with a single on-board fxp0 and sym scsi- controller (SMP only) Problem went away after enabling extended IO-APIC on mobo which resultet in different IRQs being assigned to fxp0, sym0 and sym1. Presumably due to different IRQ-Handling of ncr, no system hangs with ncr instead of sym. (Without extended IP-APIC, fxp0, sym0 and sym1 are assigned to the same IRQ). According to Gérard Roudier this could be an IRQ stall (in my case at least). Presumably similar problems (other users): - System hangs (sym) with heavy disk io - System hangs with heavy network io I'm still waiting for him to finish some testing patches/routines so I can go and analyze the IRQ behaviour on my system (I'm not in IRQ programming...). -- Holger Kipp, Dipl.-Math., Systemadministrator | alogis AG Fon: +49 (0)30 / 43 65 8 - 114 | Berliner Strasse 26 Fax: +49 (0)30 / 43 65 8 - 214 | D-13507 Berlin Tegel email: holger.kipp@alogis.com | http://www.alogis.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message