From owner-cvs-src Mon Mar 17 19:26:54 2003 Delivered-To: cvs-src@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1AD2E37B404; Mon, 17 Mar 2003 19:26:51 -0800 (PST) Received: from rusunix.org (rusunix.org [195.162.58.254]) by mx1.FreeBSD.org (Postfix) with ESMTP id A4B1643F85; Mon, 17 Mar 2003 19:26:49 -0800 (PST) (envelope-from aquatique@rusunix.org) Received: from rusunix.org (aquatique.rsb.local [192.168.1.111]) (using TLSv1 with cipher RC4-MD5 (128/128 bits)) (Client CN "Aquatique", Issuer "El Vampiro" (verified OK)) by rusunix.org (CommuniGay Pro) with ESMTP id 8E0261CE018; Tue, 18 Mar 2003 09:26:46 +0600 (OMST) Message-ID: <3E7691F5.9070902@rusunix.org> Date: Tue, 18 Mar 2003 09:26:45 +0600 From: "Evgueni V. Gavrilov" User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.2.1) Gecko/20030306 X-Accept-Language: ru, en MIME-Version: 1.0 To: Jeff Roberson Cc: Poul-Henning Kamp , Kirk McKusick , dwmalone@FreeBSD.org, Mike Makonnen , src-committers@FreeBSD.org, cvs-src@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: kern/42277 References: <20030317162051.R66343-100000@mail.chesapeake.net> In-Reply-To: <20030317162051.R66343-100000@mail.chesapeake.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-cvs-src@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Jeff Roberson wrote: >>The simple way to nail the culprint would be to move things out of >>the 128 bucket by adding to their malloc byte counts, but for such >>a strategy to work, we obviously need to be able to reproduce it >>fairly consistently. > > UMA has a facility to track the previous owner of the memory. Assuming > that is the problem I could give you a way to cache this information so we > can print it out when we detect corruption. I can work on this tonight. > What is the repro? As I told before the repro is funny: hard jobs such as world/release building runs fine. Panic occurs upon minimal activity. Maximal uptime achieved about 3 days. But usually panic comes faster - in 10-12 hours. In order to track down the issue before tarballing 4.8 I can apply any overbold patches/ideas. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-src" in the body of the message