From owner-freebsd-stable Sun May 6 5:41: 3 2001 Delivered-To: freebsd-stable@freebsd.org Received: from hamilton.th.physik.uni-bonn.de (hamilton.th.physik.uni-bonn.de [131.220.162.85]) by hub.freebsd.org (Postfix) with ESMTP id 36E8837B424 for ; Sun, 6 May 2001 05:40:58 -0700 (PDT) (envelope-from conrad@th.physik.uni-bonn.de) Received: from merlin.th.physik.uni-bonn.de (merlin.th.physik.uni-bonn.de [131.220.161.121]) by hamilton.th.physik.uni-bonn.de (Postfix) with ESMTP id 106125D1B; Sun, 6 May 2001 14:40:57 +0200 (CEST) Received: by merlin.th.physik.uni-bonn.de (Postfix, from userid 145) id 1F6E93640E; Sun, 6 May 2001 14:40:55 +0200 (CEST) Received: from localhost (localhost [127.0.0.1]) by merlin.th.physik.uni-bonn.de (Postfix) with ESMTP id 0CCC73260E; Sun, 6 May 2001 14:40:55 +0200 (CEST) Date: Sun, 6 May 2001 14:40:55 +0200 (CEST) From: Jan Conrad To: Jan Mikkelsen Cc: Doug Russell , Matt Dillon , freebsd-stable Subject: Tagged Queueing and ATA driver (was soft update should be default) In-Reply-To: <00b601c0d5bd$42ba1340$0901a8c0@haym.transactionsite.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Well, if I read the sources correctly, enabling tagged queueuing in the ata driver *IMPLIES* write caching!!! So it's pretty clear why tagged queueing is as good as write cacheing (or as bad :-) sys/dev/ata/ata-disk.c of 4.3R says /* enable write cacheing if allowed and not default on device */ if (ata_wc || ata_tags) { if (ata_command(adp->controller, adp->unit, ATA_C_SETFEATURES, 0, 0, 0, 0, ATA_C_F_ENAB_WCACHE, ATA_WAIT_INTR)) ata_printf(scp, device, "enabling write cache failed\n"); } What I really do not understand here - maybe it's trivial, as I am not an expert here - is why does tags imply wc for ATA but not for SCSI? The SCSI driver says 'Tagged Queueing Enabled' even on my SCSI disk (IBM DDYS-T18350N) where WCE is off! -Jan On Sun, 6 May 2001, Jan Mikkelsen wrote: > Doug Russell wrote: > > >On Sat, 5 May 2001, Matt Dillon wrote: > > > >> Not only will the hard drive not be able to write the write-cached > >> data to the media, but IDE hard drives will not guarentee write > >> ordering either. Someone did a test a while back and found that > >> under heavy disk loads an IDE drive could hold some of the dirty data > >> in its cache for an indefinite period of time without writing it out. > >> i.e. it would write out some of the dirty data but also hold some of > it > >> indefinitely, unwritten. > > > >Blech. IDE. Need I say more? :) > >(Yes, I do have some IDE disks, but I prefer SCSI. Don't we all? :) ) > >I don't have WCE on any IDE disk i can think of, but I don't normally run > >it on my SCSIs, either, unless I specifically need the write speed. > > > Good write speed is possible without using write cache by using tagged > command queueing. Have you measured tagged command queueing vs. write cache > for write speed on your SCSI drives? I get about the same (~23MB/sec) on an > IBM DLTA-307030 Ultra ATA drive (tags/no WC vs. no tags/WC). With neither > option, it is terrible, of course. > > The question is how much performance the write ordering constraints in the > softupdates protocol removes. I have no idea, and I'm sure it would depend > on all sorts of stuff. Hopefully, aggregate throughput would be high, even > if individual operations were slow. > > Jan Mikkelsen > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-stable" in the body of the message > -- Physikalisches Institut der Universitaet Bonn Nussallee 12 D-53115 Bonn GERMANY To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message