From owner-freebsd-stable Tue Feb 18 20:59:52 2003 Delivered-To: freebsd-stable@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1C62437B401 for ; Tue, 18 Feb 2003 20:59:51 -0800 (PST) Received: from olmec.nighttide.net (jasper.nighttide.net [207.5.141.146]) by mx1.FreeBSD.org (Postfix) with ESMTP id 38C7343F3F for ; Tue, 18 Feb 2003 20:59:49 -0800 (PST) (envelope-from darren@nighttide.net) Received: from olmec.nighttide.net (localhost [127.0.0.1]) by olmec.nighttide.net (8.12.6/8.12.6) with ESMTP id h1J4xmC9071655 for ; Tue, 18 Feb 2003 23:59:48 -0500 (EST) (envelope-from darren@nighttide.net) Received: from localhost (darren@localhost) by olmec.nighttide.net (8.12.6/8.12.6/Submit) with ESMTP id h1J4xmYD071652 for ; Tue, 18 Feb 2003 23:59:48 -0500 (EST) (envelope-from darren@nighttide.net) X-Authentication-Warning: olmec.nighttide.net: darren owned process doing -bs Date: Tue, 18 Feb 2003 23:59:48 -0500 (EST) From: Darren Henderson To: stable list Subject: spontaneous reboot & gcc In-Reply-To: <20030219033811.GA25426@luke.immure.com> Message-ID: References: <20030218161705.GA4876@luke.immure.com> <20030219011808.GK50962@rot13.obsecurity.org> <20030219033811.GA25426@luke.immure.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII 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 I'm seeing a spontaneous reboot in one very specific circumstance that appears to be attributable to the use of gcc with -O2 optimization. For some time now I've been seeing spontaneous reboots when trying to compile pike (http://pike.ida.liu.se/) on a 1.8Ghz P4 system running 4.7-STABLE (though its been a problem for a few versions of both Pike and FreeBSD). This does not occur on a 700Mhz P3 or on a 133Mhz Pentium Pro system. All the systems have recent source. The 1.8Ghz system has 256MB of ram and appears to be rock solid in all other respects - the reboot only occurs while attempting to compile Pike, and always in the same location (peep.c). Pike defaults to using -O2 optimizations. Forcing it not to do that stops the spontaneous reboot from occurring. There isn't any spore left behind after the reboot though I have softupdates turned on and fsck has a bit of work to do when the box comes back up. Turning softupdates off might reveal a bit more. It doesn't seem reasonable that gcc could bring the box down but the fact that the -O2 elimination keeps it from happening gives me pause to wonder about gcc on P4 systems. I suppose not using -O2 would mean that gcc is doing a bit less work, maybe using a bit less memory etc but the other systems have less then half the memory of the system in question and they don't see a problem. Any thoughts appreciated. ______________________________________________________________________ Darren Henderson darren@nighttide.net Help fight junk e-mail, visit http://www.cauce.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message