From owner-cvs-all Tue Dec 7 8:24:14 1999 Delivered-To: cvs-all@freebsd.org Received: from feral.com (feral.com [192.67.166.1]) by hub.freebsd.org (Postfix) with ESMTP id 35E7614CBF; Tue, 7 Dec 1999 08:24:09 -0800 (PST) (envelope-from mjacob@feral.com) Received: from beppo.feral.com (beppo [192.67.166.79]) by feral.com (8.8.7/8.8.7) with ESMTP id IAA21843; Tue, 7 Dec 1999 08:25:29 -0800 Date: Tue, 7 Dec 1999 08:25:29 -0800 (PST) From: Matthew Jacob Reply-To: mjacob@feral.com To: Matthew Dillon Cc: Julian Elischer , Mike Smith , cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/kern kern_shutdown.c In-Reply-To: <199912070829.AAA75674@apollo.backplane.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk On Tue, 7 Dec 1999, Matthew Dillon wrote: > :Softupdates is pointless without this because Soft Updates cannot > :guarantee the state of the filesystem to be sane unless the disk does not > :lie about whether it has written to the disk. Soft updates also makes all > :the writes async, so who cares.. :-) > : > :More precisely, the order of the writes must be absolutly maintained > :and disks that buffer writes usually reserve the right to re-order > :adjacent writes. > : > :This is basically what we decided was needed in discussion with Kirk to > :make Soft Updates safe. (The whole point). > : > :Julian > > The disk can lie about whether it has written data to the disk without We are using a shutdown hook (dashutdown) as well as daclose/dadump to issue a SYNCHRONIZE CACHE to the drives. > messing softupdates up, but it cannot reorder writes that the system has > told it must be ordered. I would note that Solaris/Sun have made a determination (from several years ago- maybe stale) that they do not believe disks reliable in f/w enough to truly respect ORDERED tags. This is why they never really got around to using B_ORDER. Before people go blaming softupdates code, they should quirk tags for a drive *off*. > SCSI has a tag to guarentee ordering (or not), > but I don't know if IDE has the same sort of feature. With ATAPI, yes, I believe so. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message