From owner-freebsd-scsi Sat Jul 10 23: 3: 4 1999 Delivered-To: freebsd-scsi@freebsd.org Received: from metriclient-2.uoregon.edu (metriclient-2.uoregon.edu [128.223.172.2]) by hub.freebsd.org (Postfix) with ESMTP id 1237314E3D for ; Sat, 10 Jul 1999 23:02:55 -0700 (PDT) (envelope-from gurney_j@efn.org) Received: (from jmg@localhost) by metriclient-2.uoregon.edu (8.9.1/8.8.7) id XAA00669; Sat, 10 Jul 1999 23:02:37 -0700 (PDT) Message-ID: <19990710230236.35387@hydrogen.fircrest.net> Date: Sat, 10 Jul 1999 23:02:36 -0700 From: John-Mark Gurney To: "Kenneth D. Merry" Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: microp 4421-07 fails on tags... References: <19990710214320.56085@hydrogen.fircrest.net> <199907110522.XAA19783@panzer.kdm.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.69 In-Reply-To: <199907110522.XAA19783@panzer.kdm.org>; from Kenneth D. Merry on Sat, Jul 10, 1999 at 11:22:36PM -0600 Reply-To: John-Mark Gurney Organization: Cu Networking X-Operating-System: FreeBSD 3.0-RELEASE i386 X-PGP-Fingerprint: B7 EC EF F8 AE ED A7 31 96 7A 22 B3 D8 56 36 F4 X-Files: The truth is out there X-URL: http://resnet.uoregon.edu/~gurney_j/ Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Kenneth D. Merry scribbled this message on Jul 10: > John-Mark Gurney wrote... > > Kenneth D. Merry scribbled this message on Jul 10: > > > John-Mark Gurney wrote... > > > > well, after a long time testing, the microp 4421-07: > > > > da0: Fixed Direct Access SCSI2 device > > > > > > > > has serious problems with tag queuing... the only way I was able to > > > > get this drive to work reliably was to turn of tags completely... does > > > > anyone have a drive with a different firmware that does do tags? > > > > > > > > if no one does, I'll just commit the attached patch.... > > > > > > You need to be a little more specific about how the drive is broken, and > > > how you tested it. > > > > the drive will not return ccb's (the ccb's will timeout) if you have 2 > > or more tags defined in the quirk structure... I tried various levels > > of tags from the default limit and 2, 8, 16, and 32 if I remeber > > correctly... all of those would cause a ccb to not be returned and end > > up timing out, and now that I have this entry in, I can run with the > > fs's on it no longer mounted sync (which helped prevent to many commands > > from hitting the drive)... > > Hmm, sounds like buggy firmware all right. Don't you mean "no longer > mounted _async_"? In any case, softupdates would have had a similar > effect, and isn't dangerous to run with. nope, I did mean sync, it seemed to help and actually let me do some installs on the disk, while w/o it, it would fail... > > since I didn't have the nice camcontrol ability to set tags, I had to > > build kernels to test this... > > That's unfortunate. You would have been able to do it with -current or > -stable from early May on. yeh, I noticed, but I did most of the work before the tree was integrated, and I had problems with 3.1-R being worse than 3.0-R as far as the vm system was concerned... > > of course this still doesn't address the fact that Justin hasn't gotten > > back to me on the fact that the adv driver is broken wrt chinon cdrom > > drives... > > Not surprising, I suppose. You might want to ping him again. > > Anyway, your drive sounds broken enough. Feel free to commit the patch. > Make sure you mention the name of the drive in the commit message, and make > sure you commit it to -stable as well as -current. I'm sure someone will > squawk if they have a 4421 that works properly with tagged queueing. yeh, the Microp don't actually have a name for their drives, just their model number... and if they do, we can figure out the fireware rev part of it... -- John-Mark Gurney Voice: +1 541 684 8449 Cu Networking P.O. Box 5693, 97405 "The soul contains in itself the event that shall presently befall it. The event is only the actualizing of its thought." -- Ralph Waldo Emerson To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message