From owner-freebsd-hackers Mon Sep 14 00:23:41 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id AAA20131 for freebsd-hackers-outgoing; Mon, 14 Sep 1998 00:23:41 -0700 (PDT) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from whistle.com (s205m131.whistle.com [207.76.205.131]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id AAA20110; Mon, 14 Sep 1998 00:23:39 -0700 (PDT) (envelope-from archie@whistle.com) Received: (from smap@localhost) by whistle.com (8.7.5/8.6.12) id AAA18054; Mon, 14 Sep 1998 00:23:24 -0700 (PDT) Received: from bubba.whistle.com(207.76.205.7) by whistle.com via smap (V1.3) id sma018050; Mon Sep 14 00:23:13 1998 Received: (from archie@localhost) by bubba.whistle.com (8.8.7/8.6.12) id AAA16550; Mon, 14 Sep 1998 00:23:09 -0700 (PDT) From: Archie Cobbs Message-Id: <199809140723.AAA16550@bubba.whistle.com> Subject: Re: Still! what appears to be VM problems. In-Reply-To: <13818.5746.872010.922783@lupo.thebarn.com> from Russell Cattelan at "Sep 12, 98 01:45:51 am" To: cattelan@thebarn.com (Russell Cattelan) Date: Mon, 14 Sep 1998 00:23:09 -0700 (PDT) Cc: freebsd-current@FreeBSD.ORG, freebsd-hackers@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Russell Cattelan writes: > Sep 11 17:34:45 lupo /kernel: pid 11035 (sendmail), uid 0: exited on signal 11 > Sep 11 17:49:45 lupo /kernel: pid 11049 (sendmail), uid 0: exited on signal 11 > > perl in free(): warning: junk pointer, too high to make sense. > perl in free(): warning: junk pointer, too high to make sense. > > This is the latest CAM kernel... so it isn't the "most" to date. We've been seeing a similar problem (or problems) literally for years ! Ie, in 2.2, 3.0, ... Could be one bug, could be more than one bug... but the following things seems to crop up all the time (not necessarily at the same time): - some file is being mmap'd - some process is forking - panic because a process on sleep and run queues at the same time - random segfaults in processes (inetd, sendmail -> because they fork alot?) When we try to look at kernel core dumps, it always seems either completely mysterious (eg, impossible CPU register contents) or almost completely mysterious (because no one here understands the VM code). I guess I'm hoping some VM guru will see this list and think "aha!" .. VM gurus, any ideas? Maybe there's some small race condition in the VM system when a process forks and there's a context switch involving a process using mmap and/or shared memory?? Or maybe just a simple mmap/shared memory bug? -Archie References: [DR97] The Dave Rivers memorial panic [INN98] People with news servers (which use mmap) getting random pages written into files. [SEGV97] Numerous inetd/sendmail random segfault reports to this list ___________________________________________________________________________ Archie Cobbs * Whistle Communications, Inc. * http://www.whistle.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message