From owner-freebsd-stable Sat Dec 22 9:14:41 2001 Delivered-To: freebsd-stable@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [216.240.41.2]) by hub.freebsd.org (Postfix) with ESMTP id B68C037B419 for ; Sat, 22 Dec 2001 09:14:35 -0800 (PST) Received: (from dillon@localhost) by apollo.backplane.com (8.11.6/8.9.1) id fBMHEX195032; Sat, 22 Dec 2001 09:14:33 -0800 (PST) (envelope-from dillon) Date: Sat, 22 Dec 2001 09:14:33 -0800 (PST) From: Matthew Dillon Message-Id: <200112221714.fBMHEX195032@apollo.backplane.com> To: "Kristian K. Nielsen" Cc: , , , , , Subject: Re: RE: 4.4-STABLE crashes - suspects new ata-driver over wd-drivers References: 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 did in the bug report at: :http://www.freebsd.org/cgi/query-pr.cgi?pr=31233 : :Is that enough information? : :/Kristian Yah. I've been analyzing the thread and focusing on Brady's kernel cores to try to track this down, but it's tough going. This is what I see so far: * Cpu does not appear to matter (P3, P4, and AMD have been indicated) * All machines are doing heavy IDE I/O * All machines are running the new ATA drivers * DMA vs PIO does not seem to matter * All or most machines appear to be running softupdates ?? * None of the disk activity is over SCSI * VM system is not necessarily being stressed, just the disks. I would like you to turn off softupdates if you have it on and see if that makes a difference. I do not know whether the problem is the ATA driver or whether it is simply a side effect - for example, heavy disk I/O creating backlogs and delays that are causing another bug to rear its ugly head. So far I haven't been able to track down the cause - the problem appears to be random corruption. Each of Brady's three kernel cores are corrupted in different places - Most typically different portions of the vm_page_array[] which is the absolute *last* thing I would every expect to get corrupted considering their static, fixed ram. The only bug I know about that could have an effect is the BUF_TIMELOCK bug which I fixed two days go in -current (not yet in stable), but a number of people have already reported -current working prior to the fix. Of course, the timing is very different in -current so that might not matter. Longer IDE delays could cause this bug to rear its ugly head but it's a long shot. I am pursuing this with Brady since he seems to be able to reliably reproduce the crash. -Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message