From owner-freebsd-bugs@FreeBSD.ORG Mon Oct 20 17:00:11 2008 Return-Path: Delivered-To: freebsd-bugs@hub.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D00181065679 for ; Mon, 20 Oct 2008 17:00:11 +0000 (UTC) (envelope-from gnats@FreeBSD.org) Received: from freefall.freebsd.org (freefall.freebsd.org [IPv6:2001:4f8:fff6::28]) by mx1.freebsd.org (Postfix) with ESMTP id BCF568FC1A for ; Mon, 20 Oct 2008 17:00:11 +0000 (UTC) (envelope-from gnats@FreeBSD.org) Received: from freefall.freebsd.org (gnats@localhost [127.0.0.1]) by freefall.freebsd.org (8.14.3/8.14.3) with ESMTP id m9KH0BWp015631 for ; Mon, 20 Oct 2008 17:00:11 GMT (envelope-from gnats@freefall.freebsd.org) Received: (from gnats@localhost) by freefall.freebsd.org (8.14.3/8.14.3/Submit) id m9KH0BkY015630; Mon, 20 Oct 2008 17:00:11 GMT (envelope-from gnats) Date: Mon, 20 Oct 2008 17:00:11 GMT Message-Id: <200810201700.m9KH0BkY015630@freefall.freebsd.org> To: freebsd-bugs@FreeBSD.org From: "Bjoern A. Zeeb" Cc: Subject: Re: kern/128247: [panic] Fatal Trap 12 in ip6_forward (/usr/src/sys/netinet6/ip6_forward.c:420) X-BeenThere: freebsd-bugs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: "Bjoern A. Zeeb" List-Id: Bug reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 20 Oct 2008 17:00:11 -0000 The following reply was made to PR kern/128247; it has been noted by GNATS. From: "Bjoern A. Zeeb" To: bug-followup@FreeBSD.org, m.atkinson@F5.com Cc: Subject: Re: kern/128247: [panic] Fatal Trap 12 in ip6_forward (/usr/src/sys/netinet6/ip6_forward.c:420) Date: Mon, 20 Oct 2008 16:31:17 +0000 (UTC) On Mon, 20 Oct 2008, Mark Atkinson wrote: > #0 doadump () at pcpu.h:221 > 221 __asm __volatile("movl %%fs:0,%0" : "=3Dr" (td)); > (kgdb) l *0xc09af288 > 0xc09af288 is in ip6_forward (/usr/src/sys/netinet6/ip6_forward.c:420). > 415 * address). We use a local copy of ip6_src, since = > in6_setscope() > 416 * will possibly modify its first argument. > 417 * [draft-ietf-ipngwg-icmp-v3-04.txt, Section 3.1] > 418 */ > 419 src_in6 =3D ip6->ip6_src; > 420 if (in6_setscope(&src_in6, rt->rt_ifp, &outzone)) { > 421 /* XXX: this should not happen */ > 422 V_ip6stat.ip6s_cantforward++; > 423 V_ip6stat.ip6s_badscope++; > 424 m_freem(m); > > (kgdb) frame 10 > #10 0xc09af288 in ip6_forward (m=3D0xc5ed8300, srcrt=3D0) > at /usr/src/sys/netinet6/ip6_forward.c:420 > 420 if (in6_setscope(&src_in6, rt->rt_ifp, &outzone)) { > (kgdb) p rt > $3 =3D (struct rtentry *) 0x0 and rt comes from rt = V_ip6_forward_rt.ro_rt; which is an unprotected (no lock) single global cache in the ipv6 stack. So I guess another packet changed it while this one was processed. The problem is well known and will hopefully be eliminated one day along with other caches left. /bz -- Bjoern A. Zeeb Stop bit received. Insert coin for new game.