From owner-freebsd-bugs Sun Sep 1 5:33:46 2002 Delivered-To: freebsd-bugs@freebsd.org Received: from mx1.FreeBSD.org (mx1.FreeBSD.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5BCED37B400 for ; Sun, 1 Sep 2002 05:33:39 -0700 (PDT) Received: from mx1.lublin.pl (mx1.lublin.pl [212.182.63.90]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2D66143EA9 for ; Sun, 1 Sep 2002 05:33:38 -0700 (PDT) (envelope-from pawmal@unia.3lo.lublin.pl) Received: from proxy.zin.lublin.pl ([212.182.126.66]:15740 "EHLO towah-xp") by mx1.lublin.pl with ESMTP id ; Sun, 1 Sep 2002 14:33:16 +0200 From: "=?ISO-8859-2?Q?Pawe=B3_Ma=B3achowski?=" Organization: unidentified flying modems To: David Malone , freebsd-bugs@FreeBSD.org Date: Sun, 1 Sep 2002 14:33:17 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=ISO-8859-2 Content-transfer-encoding: 8BIT Comments: Sender has elected to use 8-bit data in this message. If problems arise, refer to postmaster at sender's site. Subject: Re: kern/42260: 4.6-STABLE kernel panic, fxp related Reply-To: pawmal@unia.3lo.lublin.pl In-reply-to: <200208312000.g7VK096s002453@freefall.freebsd.org> X-mailer: Pegasus Mail for Win32 (v3.12b) Message-Id: <20020901123326Z1510737-2696+4@mx1.lublin.pl> Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On 31 Aug 02, at 13:00, David Malone wrote: > > #16 0xc0224a77 in m_getcl (how=1, type=1, flags=2) > > at /usr/src/sys/kern/uipc_mbuf.c:589 > > This looks like it might be the mcl_pool problem. Can you check what > version of /usr/src/sys/kern/uipc_mbuf.c you have? You might also > like to check and see what "sysctl -a | fgrep mcl_pool_max" outputs. I use cvsup ~every day. It was: $FreeBSD: src/sys/kern/uipc_mbuf.c,v 1.51.2.20 2002/08/12 22:09:12 luigi Exp kern.ipc.mcl_pool_max: 0 Problem looks like a kind of magic: switch---switch | | | A B C Host A is pantagruel, Athlon XP on VIA KT333 Soltek MB. It is connected using Compaq/Intel NIC. Runs fresh GENERIC 4.6-STABLE. It crashes. Subnet a. Host B is gargantua, Celeron on VIA694T Asus MB. It is connected using 3Com 905B NIC. Runs 4.6.2-RELEASE. Subnets a, c. Acts as a router beetween A and C. Host C is Athlon on SiS735 ECS MB. It is connected using Compaq/Intel NIC. Runs fresh GENERIC 4.6-STABLE. Subnet c. When I try to fetch something big from A to B, A crashes. When I try to fetch something big from B to A, A crashes. When I try to fetch something big from A to C (via B), A works fine. When I try to fetch something big from C to A (via B), A works fine. That's interesting. When I try to fetch something big from B to C, C works fine. When I try to fetch something big from C to B, C works fine. So, I have machine C. It runs 4.6-STABLE and has no problems with fxp interface! How-To-Repeat sections in kern/42260 seems to be wrong, I can't reproduce this problem on other machine with fxp interface. I have machine A, it runs 4.6-STABLE and crashes when there is heavy traffic on fxp interface AND traffic goes directly to B. This fxp interface was replaced with identical Compaq/Intel NIC, nothing helps. When traffic from A goes to B and then to C (meens this is A-C connection and B acts as a router beetween them), machine A is not crashing. (?!) There are no compile optimizations. I would suspect some hardware fault, but there is one thing: When I downgrade A from 4.6-STABLE to 4.6.2-RELEASE, A has no problems and everyting works fine. -- Paweł Małachowski, mailto:pawmal#freebsd,lublin,pl To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message