From owner-freebsd-scsi Sun Apr 2 5:16:37 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from bluebottle.calcaphon.com (calcaphon.demon.co.uk [193.237.19.5]) by hub.freebsd.org (Postfix) with ESMTP id 57A9837B5BD for ; Sun, 2 Apr 2000 05:16:31 -0700 (PDT) (envelope-from n_hibma@calcaphon.com) Received: from henny.calcaphon.com (henny.calcaphon.com [10.0.0.36]) by bluebottle.calcaphon.com (8.9.3/8.9.1) with ESMTP id NAA98088 for ; Sun, 2 Apr 2000 13:22:05 +0100 (BST) (envelope-from n_hibma@calcaphon.com) Date: Sun, 2 Apr 2000 13:13:43 +0100 (BST) From: Nick Hibma X-Sender: n_hibma@localhost Reply-To: Nick Hibma To: FreeBSD SCSI Mailing List Subject: peripheral Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org How do I send an action to a SIM (XPT_SCAN_BUS)? A SIM does not seem to have a peripheral (cam_find_find_periph), yet periph needs to be initialised in the path passed to xpt_setup_ccb/xpt_action, to avoid a NULL dereference in xpt_done. Nick -- n_hibma@webweaving.org n_hibma@freebsd.org USB project http://www.etla.net/~n_hibma/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Apr 2 14:50:25 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by hub.freebsd.org (Postfix) with ESMTP id 6991937B965 for ; Sun, 2 Apr 2000 14:50:23 -0700 (PDT) (envelope-from gibbs@narnia.plutotech.com) Received: (from gibbs@localhost) by narnia.plutotech.com (8.9.3/8.7.3) id PAA27619; Sun, 2 Apr 2000 15:35:36 -0600 (MDT) Date: Sun, 2 Apr 2000 15:35:36 -0600 (MDT) From: "Justin T. Gibbs" Message-Id: <200004022135.PAA27619@narnia.plutotech.com> To: Nick Hibma Cc: scsi@FreeBSD.org Subject: Re: peripheral X-Newsgroups: pluto.freebsd.scsi In-Reply-To: User-Agent: tin/pre-1.4-980818 ("Laura") (UNIX) (FreeBSD/4.0-CURRENT (i386)) Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In article you wrote: > > How do I send an action to a SIM (XPT_SCAN_BUS)? A SIM does not seem to > have a peripheral (cam_find_find_periph), yet periph needs to be > initialised in the path passed to xpt_setup_ccb/xpt_action, to avoid a > NULL dereference in xpt_done. A SIM is not supposed to send CCBs to itself. To truely implement auto-discovery (which I'm guessing is the root of your question), requires adding in the auto-discovery hooks from CAM3. Until then, either add your hack to get at the xpt_periph, or allocate a peripheral (cam_periph_alloc) for your SIM and use it to generate the SCAN_BUS request. -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Apr 2 16:33:53 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from hotmail.com (law2-f300.hotmail.com [216.32.180.154]) by hub.freebsd.org (Postfix) with SMTP id 0EE1B37B65F for ; Sun, 2 Apr 2000 16:33:49 -0700 (PDT) (envelope-from mut3x@hotmail.com) Received: (qmail 20869 invoked by uid 0); 2 Apr 2000 23:33:47 -0000 Message-ID: <20000402233347.20868.qmail@hotmail.com> Received: from 206.165.200.116 by www.hotmail.com with HTTP; Sun, 02 Apr 2000 16:33:47 PDT X-Originating-IP: [206.165.200.116] From: "Mutex Records USA" To: freebsd-scsi@freebsd.org Subject: AMI Raid Controller Date: Sun, 02 Apr 2000 16:33:47 PDT Mime-Version: 1.0 Content-Type: text/plain; format=flowed Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I'm considering getting some AMI 1400 raid controllers for a few servers. I understand that they don't support booting, that's fine, I want to have a disk for system/swap anyway. My question: I understand that the AMI card supports hotswapping drives. Is this supported by the FreeBSD driver? If so I am planning on buying some removable Kingston drive shells (unless someone can reccommend a better one!) Also, how would I be notified that a drive has failed? console messages? /var/log/messages, or wherever kernel error messages are set to go by /etc/syslog.conf? The boxes are going to be in colocation about 800 miles away, so I need to know if a drive has croaked... My hearing is good, but not *that* good. :) How does the AMI card perform compared to say, the DPT cards? Thanks! adam ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Apr 2 16:33:57 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from hotmail.com (law2-f32.hotmail.com [216.32.181.32]) by hub.freebsd.org (Postfix) with SMTP id C8E2C37BC00 for ; Sun, 2 Apr 2000 16:33:54 -0700 (PDT) (envelope-from mut3x@hotmail.com) Received: (qmail 67631 invoked by uid 0); 2 Apr 2000 23:33:53 -0000 Message-ID: <20000402233353.67630.qmail@hotmail.com> Received: from 206.165.200.116 by www.hotmail.com with HTTP; Sun, 02 Apr 2000 16:33:53 PDT X-Originating-IP: [206.165.200.116] From: "Mutex Records USA" To: freebsd-scsi@freebsd.org Subject: AMI Raid Controller Date: Sun, 02 Apr 2000 16:33:53 PDT Mime-Version: 1.0 Content-Type: text/plain; format=flowed Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I'm considering getting some AMI 1400 raid controllers for a few servers. I understand that they don't support booting, that's fine, I want to have a disk for system/swap anyway. My question: I understand that the AMI card supports hotswapping drives. Is this supported by the FreeBSD driver? If so I am planning on buying some removable Kingston drive shells (unless someone can reccommend a better one!) Also, how would I be notified that a drive has failed? console messages? /var/log/messages, or wherever kernel error messages are set to go by /etc/syslog.conf? The boxes are going to be in colocation about 800 miles away, so I need to know if a drive has croaked... My hearing is good, but not *that* good. :) How does the AMI card perform compared to say, the DPT cards? Thanks! adam ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Apr 2 16:34:55 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from hotmail.com (law2-f236.hotmail.com [216.32.181.236]) by hub.freebsd.org (Postfix) with SMTP id B8A6D37BBCC for ; Sun, 2 Apr 2000 16:34:53 -0700 (PDT) (envelope-from mut3x@hotmail.com) Received: (qmail 70974 invoked by uid 0); 2 Apr 2000 23:34:52 -0000 Message-ID: <20000402233451.70973.qmail@hotmail.com> Received: from 206.165.200.116 by www.hotmail.com with HTTP; Sun, 02 Apr 2000 16:34:51 PDT X-Originating-IP: [206.165.200.116] From: "Mutex Records USA" To: freebsd-scsi@freebsd.org Subject: AMI Raid Controller Date: Sun, 02 Apr 2000 16:34:51 PDT Mime-Version: 1.0 Content-Type: text/plain; format=flowed Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I'm considering getting some AMI 1400 raid controllers for a few servers. I understand that they don't support booting, that's fine, I want to have a disk for system/swap anyway. My question: I understand that the AMI card supports hotswapping drives. Is this supported by the FreeBSD driver? If so I am planning on buying some removable Kingston drive shells (unless someone can reccommend a better one!) Also, how would I be notified that a drive has failed? console messages? /var/log/messages, or wherever kernel error messages are set to go by /etc/syslog.conf? The boxes are going to be in colocation about 800 miles away, so I need to know if a drive has croaked... My hearing is good, but not *that* good. :) How does the AMI card perform compared to say, the DPT cards? Thanks! adam ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Apr 2 16:52:45 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from hotmail.com (law2-f47.hotmail.com [216.32.181.47]) by hub.freebsd.org (Postfix) with SMTP id 8D25537BC29 for ; Sun, 2 Apr 2000 16:52:44 -0700 (PDT) (envelope-from mut3x@hotmail.com) Received: (qmail 4107 invoked by uid 0); 2 Apr 2000 23:52:34 -0000 Message-ID: <20000402235234.4106.qmail@hotmail.com> Received: from 206.165.200.116 by www.hotmail.com with HTTP; Sun, 02 Apr 2000 16:52:34 PDT X-Originating-IP: [206.165.200.116] From: "Mutex Records USA" To: freebsd-scsi@freebsd.org Subject: Hotmail sucks :( Date: Sun, 02 Apr 2000 16:52:34 PDT Mime-Version: 1.0 Content-Type: text/plain; format=flowed Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org They must be trying to port it to NT again today, or something :( my apology! Adam ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Apr 2 20:10:58 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from mail.wolves.k12.mo.us (mail.wolves.k12.mo.us [207.160.214.1]) by hub.freebsd.org (Postfix) with ESMTP id B761B37BB3D for ; Sun, 2 Apr 2000 20:10:54 -0700 (PDT) (envelope-from cdillon@wolves.k12.mo.us) Received: from mail.wolves.k12.mo.us (cdillon@mail.wolves.k12.mo.us [207.160.214.1]) by mail.wolves.k12.mo.us (8.9.3/8.9.3) with ESMTP id WAA59710; Sun, 2 Apr 2000 22:10:32 -0500 (CDT) (envelope-from cdillon@wolves.k12.mo.us) Date: Sun, 2 Apr 2000 22:10:32 -0500 (CDT) From: Chris Dillon To: Mutex Records USA Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: AMI Raid Controller In-Reply-To: <20000402233347.20868.qmail@hotmail.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Sun, 2 Apr 2000, Mutex Records USA wrote: > I'm considering getting some AMI 1400 raid controllers for a few servers. > I understand that they don't support booting, that's fine, I want to have a > disk for system/swap anyway. Mike Smith (the author of the amr driver for FreeBSD) is going to be a much better authority on AMI MegaRAID controllers than I am, but I just recently bought an Enterprise 1200 to play with. I can boot from this thing just fine, by the way. > My question: I understand that the AMI card supports hotswapping > drives. Is this supported by the FreeBSD driver? If so I am > planning on buying some removable Kingston drive shells (unless > someone can reccommend a better one!) Also, how would I be > notified that a drive has failed? console messages? > /var/log/messages, or wherever kernel error messages are set to go > by /etc/syslog.conf? The boxes are going to be in colocation > about 800 miles away, so I need to know if a drive has croaked... > My hearing is good, but not *that* good. :) If I understand correctly, as long as the logical drive doesn't enter the Failed or Offline states, then the driver won't care or need to know and you can just swap a physical drive out and the controller will begin to automatically begin to rebuild the logical drive, assuming you have allowed it to do automatic rebuild. You can also set up a hot spare to have it happen without your intervention. Obviously if a logical drive enters the Failed or Offline state, the driver is going to have to know about it. I have no idea if Mike put anything in the driver to warn you about logical drives entering Degraded mode, but from a quick look through the amr sources, it doesn't look like it. The controller will definately let out a loud and relentless ear-piercing beep when a logical disk is in the Degraded/Failed/Offline states. :-) The controllers support communication with disk cabinets through either a fault-bus connector or serial port, so if you used one of those you could at least get a visual indication of which disk has failed -- great if you want to hire a monkey to look for little red lights once in a while and pull and replace a drive if one lights up. :-) -- Chris Dillon - cdillon@wolves.k12.mo.us - cdillon@inter-linc.net FreeBSD: The fastest and most stable server OS on the planet. For Intel x86 and Alpha architectures. ( http://www.freebsd.org ) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Apr 3 5:45:18 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from springbank.linguistics.ruhr-uni-bochum.de (springbank.linguistics.ruhr-uni-bochum.de [134.147.14.9]) by hub.freebsd.org (Postfix) with ESMTP id B25CE37BDA1 for ; Mon, 3 Apr 2000 05:45:13 -0700 (PDT) (envelope-from wegmann@linguistics.ruhr-uni-bochum.de) Received: from dialppp-7-132.rz.ruhr-uni-bochum.de ([134.147.7.132] helo=linguistics.ruhr-uni-bochum.de) by springbank.linguistics.ruhr-uni-bochum.de with esmtp (Exim 3.12 #1) id 12c6Bb-0002DE-00; Mon, 03 Apr 2000 14:42:31 +0200 Message-ID: <38E89296.78012581@linguistics.ruhr-uni-bochum.de> Date: Mon, 03 Apr 2000 14:46:14 +0200 From: Frank Wegmann X-Mailer: Mozilla 4.72 [en] (WinNT; I) X-Accept-Language: en MIME-Version: 1.0 To: freebsd-scsi@freebsd.org Subject: Can't use 4GB 1024b/sec SCSI disk Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Maybe someone on this list can be of help in this tricky issue. PROBLEM SUMMARY: A Seagate ST34371N (4.3GB) at AHA 1542C on a 486 LocalBus host (64MB), formatted with 1024 bytes per sector can be recognized but not used under FreeBSD at all--on a (true) NeXTstation it can be used without problem. Formatted with 512 bytes per sector yields half capacity on both systems. DETAILS: Originally I formatted it low-level with 1024 bytes per sector (for getting out some more MB) and used it in a NeXT machine as boot disk for years. Although it should--in theory--be possible to use it directly under FreeBSD, this doesn't work at all. Probing the devices (when invoking sysinstall) leads to: /kernel: dscheck(rda2): b_bcount 512 is not on a sector boundary (ssize 1024) At startup, the disk's transliterated geometry plus capacity had been identified correctly: da2 at aha0 bus 0 target 4 lun 0 da2: Fixed Direct Access SCSI-2 device da2: 3.333MB/s transfers (3.333MHz, offset 8) da2: 4341MB (4445468 1024 byte sectors: 64H 32S/T 2170C) This is the physical layout of the drive (by querying it with the help of the commands READ CAPACITY and MODE SENSE page 3 (format device page) and page 4 (rigid disk layout page) -- done with a tool on the NeXT): 1024 bytes per sector 86 sectors per track 10 tracks per cylinder 5168 cylinder per volume (including spare cylinders) 62 spare sectors per cylinder 0 alternate tracks per volume 4445467 usable sectors for the disk If formatting via sysinstall doesn't work, why not doing the old-fashioned way (according to 8.2.2 of the Handbook)? This is what I've got: # dd if=/dev/zero of=/dev/rda2 bs=1k count=1 1+0 records in 1+0 records out 1024 bytes transferred in 0.004346 secs (235618 bytes/sec) # disklabel -Brw da2 auto disklabel: /boot/boot2 too large <-- doh, try the other way: # disklabel /dev/rda2 | disklabel -BrR da2 /dev/stdin Put it in a nutshell, I currently cannot use the disk at all, if it is formatted with 1024 bytes. I tried to reformat it low-level under FreeBSD with camcontrol, but that doesn't help either. However, with 512 bytes per sector I can use *half* of the disk. Strangely enough I then have the same number of physical sectors (4445468), but with 512 bytes per sector I only have 2.1 GB available. At least, it would be useable under FreeBSD, but surely this is not what I want. Finally, using the disk with 512 bytes at my NeXT also gives me 2.1 GB (probably, because that number of sectors is returned on querying the device at the sense mode level). With the SCSI knowledge that I have, I'm at my wits end. It seems to me that FreeBSD refuses to work with the 1024 byte/sec disk because of its uncommon physical geometry data (nearly all disks in the PC world have 512 bytes/sec). Might be some tweaks at the driver level could help here. As far as that halved capacity with 512 bytes/sec is concerned, this could be the drives fault and perhaps it could be solved by somehow (?) writing the correct number of sectors into the mode pages. Could anybody help in this case? TIA, Frank To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Apr 3 5:53:33 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from morpheus.skynet.be (morpheus.skynet.be [195.238.2.39]) by hub.freebsd.org (Postfix) with ESMTP id 0E60637BE72; Mon, 3 Apr 2000 05:53:29 -0700 (PDT) (envelope-from blk@skynet.be) Received: from [195.238.1.121] (brad.techos.skynet.be [195.238.1.121]) by morpheus.skynet.be (Postfix) with ESMTP id 3B111DCBD; Mon, 3 Apr 2000 14:53:03 +0200 (MET DST) Mime-Version: 1.0 X-Sender: blk@pop.skynet.be Message-Id: Date: Mon, 3 Apr 2000 14:52:31 +0200 To: FreeBSD-CURRENT Mailing List From: Brad Knowles Subject: Sysctl for CAM SCSI quirks? Cc: freebsd-scsi@freebsd.org Content-Type: text/plain; charset="us-ascii" ; format="flowed" Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Folks, I just did a bit of searching on the website, and garnered enough information from to be able to add a new "quirk" in /usr/src/sys/cam/cam_xpt.c for a piece of hardware I'm having some difficulty with. Assuming this works, I will file a proper PR, etc.... However, shouldn't there be an easier way to do this? For example, shouldn't there be a sysctl for changing these sorts of things, perhaps something like: kern.cam.cd.changer.min_busy_seconds: 5 kern.cam.cd.changer.max_busy_seconds: 15 kern.cam.quirks.direct.fixed.quantum.XP39100*.*.mintags: 24 kern.cam.quirks.direct.fixed.quantum.XP39100*.*.maxtags: 32 kern.cam.quirks.direct.fixed.quantum.XP34550*.*.mintags: 24 kern.cam.quirks.direct.fixed.quantum.XP34550*.*.maxtags: 32 kern.cam.quirks.direct.fixed.quantum.XP32275*.*.mintags: 24 kern.cam.quirks.direct.fixed.quantum.XP32275*.*.maxtags: 32 This would make it very easy to add new quirks as they come along (just update your sysctl.conf), and without having to always file a PR so that /usr/src/sys/cam/cam_xpt.c gets modified so that you can have your quirk and track -CURRENT too (or -STABLE, for that matter). -- These are my opinions -- not to be taken as official Skynet policy ====================================================================== Brad Knowles, || Belgacom Skynet SA/NV Systems Architect, Mail/News/FTP/Proxy Admin || Rue Colonel Bourg, 124 Phone/Fax: +32-2-706.13.11/12.49 || B-1140 Brussels http://www.skynet.be || Belgium To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Apr 3 6:53: 4 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from caladan.tdx.co.uk (caladan.tdx.co.uk [195.188.177.4]) by hub.freebsd.org (Postfix) with ESMTP id 1BA3C37B640 for ; Mon, 3 Apr 2000 06:52:54 -0700 (PDT) (envelope-from kpielorz@tdx.co.uk) Received: from tdx.co.uk (lorca-tx.tdx.co.uk [195.188.177.242]) by caladan.tdx.co.uk (8.10.0/8.10.0/Kp) with ESMTP id e32F0Rx00372 for ; Sun, 2 Apr 2000 16:00:27 +0100 (BST) Message-ID: <38E76085.9E5BE928@tdx.co.uk> Date: Sun, 02 Apr 2000 16:00:21 +0100 From: Karl Pielorz Organization: TDX - The Digital eXchange X-Mailer: Mozilla 4.7 [en-gb] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: scsi@freebsd.org Subject: Problems with 4.0 - CAM when 3.4 was OK? (long post) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi All, I've recently upgraded my 3.4-STABLE box to 4.0-STABLE. The upgrade went fine, except something seems to have gone awry with my SCSI set-up... The machine has 3 x 2940's in it, and one 1542 (not actually used for anything yet). Under 3.4 the system would boot, and run fine... I've got data on most the SCSI drives, I used to backup my NT workstation to them across a 100Mb/FDX link and get good performance (~4-5Mb/sec), and no errors... I'm very sure all the drives etc. are terminated OK - after having spent a lot of money getting decent cables, active termination etc. If I boot a 4.0-STABLE kernel, I get heaps of: " Waiting 4 seconds for SCSI devices to settle de0: enabling 10baseT port (probe15:ahc1:0:0:0): SCB 0xe - timed out in Message-in phase, SEQADDR == 0x157 (probe15:ahc1:0:0:0): Other SCB Timeout (probe16:ahc1:0:1:0): SCB 0xd - timed out in Message-in phase, SEQADDR == 0x157 (probe16:ahc1:0:1:0): BDR message in message buffer (probe16:ahc1:0:1:0): SCB 0xd - timed out in Message-in phase, SEQADDR == 0x157 (probe16:ahc1:0:1:0): no longer in timeout, status = 34b ahc1: Issued Channel A Bus Reset. 15 SCBs aborted Timedout SCB handled by another timeout (probe18:ahc1:0:3:0): SCB 0xd - timed out while idle, SEQADDR == 0xa (probe18:ahc1:0:3:0): Queuing a BDR SCB (probe18:ahc1:0:3:0): no longer in timeout, status = 35b Timedout SCB handled by another timeout (probe15:ahc1:0:0:0): SCB 0xc - timed out while idle, SEQADDR == 0x4a (probe15:ahc1:0:0:0): SCB 12: Immediate reset. Flags = 0x4040 (probe15:ahc1:0:0:0): no longer in timeout, status = 34b ahc1: Issued Channel A Bus Reset. 13 SCBs aborted Timedout SCB handled by another timeout (probe21:ahc1:0:6:0): SCB 0xc - timed out while idle, SEQADDR == 0x9 (probe21:ahc1:0:6:0): Queuing a BDR SCB (probe21:ahc1:0:6:0): no longer in timeout, status = 35b " The box doesn't get any further than this... I've left it for 1/2 an hour before "on the offchance" it would get past all the timeouts, and at least continue with no SCSI drives / missing drives, but it doesn't. I use a 'wired-down' SCSI config (included below), dmesg under 4.0-STABLE shows all the Adaptecs being found OK, " ahc0: port 0xe400-0xe4ff mem 0xffafd000-0xffafdfff irq 17 at device 18.0 on pci0 ahc0: aic7880 Wide Channel A, SCSI Id=7, 16/255 SCBs ahc1: port 0xe800-0xe8ff mem 0xffafe000-0xffafefff irq 16 at device 19.0 on pci0 ahc1: aic7880 Wide Channel A, SCSI Id=7, 16/255 SCBs ahc2: port 0xec00-0xecff mem 0xffaff000-0xffafffff irq 19 at device 20.0 on pci0 ahc2: aic7880 Wide Channel A, SCSI Id=7, 16/255 SCBs " This happens in SMP or non-SMP mode... The previous 3.4-STABLE config would instead of spewing out the above errors find: " Mar 31 09:26:38 caladan /kernel: Waiting 2 seconds for SCSI devices to settle Mar 31 09:26:38 caladan /kernel: SMP: AP CPU #1 Launched! Mar 31 09:26:38 caladan /kernel: de0: enabling 10baseT port Mar 31 09:26:39 caladan /kernel: da0 at ahc0 bus 0 target 0 lun 0 Mar 31 09:26:39 caladan /kernel: da0: Fixed Direct Access SCSI-2 device Mar 31 09:26:39 caladan /kernel: da0: 10.000MB/s transfers (10.000MHz, offset 8) Mar 31 09:26:39 caladan /kernel: da0: 1042MB (2134305 512 byte sectors: 64H 32S/T 1042C) Mar 31 09:26:39 caladan /kernel: da1 at ahc0 bus 0 target 1 lun 0 Mar 31 09:26:39 caladan /kernel: da1: Fixed Direct Access SCSI-2 device Mar 31 09:26:39 caladan /kernel: da1: 10.000MB/s transfers (10.000MHz, offset 8) Mar 31 09:26:39 caladan /kernel: da1: 1042MB (2134305 512 byte sectors: 64H 32S/T 1042C) Mar 31 09:26:39 caladan /kernel: da2 at ahc0 bus 0 target 2 lun 0 Mar 31 09:26:39 caladan /kernel: da2: Fixed Direct Access SCSI-2 device Mar 31 09:26:39 caladan /kernel: da2: 10.000MB/s transfers (10.000MHz, offset 15), Tagged Queueing Enabled Mar 31 09:26:39 caladan /kernel: da2: 1003MB (2055096 512 byte sectors: 64H 32S/T 1003C) Mar 31 09:26:39 caladan /kernel: da8 at ahc2 bus 0 target 0 lun 0 Mar 31 09:26:39 caladan /kernel: da8: Fixed Direct Access SCSI-2 device Mar 31 09:26:39 caladan /kernel: da8: 40.000MB/s transfers (20.000MHz, offset 8, 16bit), Tagged Queueing Enabled Mar 31 09:26:39 caladan /kernel: da8: 4148MB (8496960 512 byte sectors: 255H 63S/T 528C) Mar 31 09:26:39 caladan /kernel: da9 at ahc2 bus 0 target 1 lun 0 Mar 31 09:26:39 caladan /kernel: da9: Fixed Direct Access SCSI-2 device Mar 31 09:26:39 caladan /kernel: da9: 40.000MB/s transfers (20.000MHz, offset 8, 16bit), Tagged Queueing Enabled Mar 31 09:26:39 caladan /kernel: da9: 4148MB (8496960 512 byte sectors: 255H 63S/T 528C) Mar 31 09:26:39 caladan /kernel: da3 at ahc0 bus 0 target 3 lun 0 Mar 31 09:26:39 caladan /kernel: da3: Fixed Direct Access SCSI-2 device Mar 31 09:26:39 caladan /kernel: da3: 10.000MB/s transfers (10.000MHz, offset 15), Tagged Queueing Enabled Mar 31 09:26:39 caladan /kernel: da3: 1003MB (2055096 512 byte sectors: 64H 32S/T 1003C) Mar 31 09:26:39 caladan /kernel: da4 at ahc0 bus 0 target 5 lun 0 Mar 31 09:26:40 caladan /kernel: da4: Fixed Direct Access SCSI-2 device Mar 31 09:26:40 caladan /kernel: da4: 10.000MB/s transfers (10.000MHz, offset 15), Tagged Queueing Enabled Mar 31 09:26:40 caladan /kernel: da4: 1021MB (2091144 512 byte sectors: 64H 32S/T 1021C) Mar 31 09:26:40 caladan /kernel: da5 at ahc0 bus 0 target 6 lun 0 Mar 31 09:26:40 caladan /kernel: da5: Fixed Direct Access SCSI-2 device Mar 31 09:26:40 caladan /kernel: da5: 10.000MB/s transfers (10.000MHz, offset 15), Tagged Queueing Enabled Mar 31 09:26:40 caladan /kernel: da5: 1021MB (2091144 512 byte sectors: 64H 32S/T 1021C) Mar 31 09:26:40 caladan /kernel: da6 at ahc1 bus 0 target 0 lun 0 Mar 31 09:26:40 caladan /kernel: da6: Fixed Direct Access SCSI-2 device Mar 31 09:26:40 caladan /kernel: da6: 40.000MB/s transfers (20.000MHz, offset 8, 16bit), Tagged Queueing Enabled Mar 31 09:26:40 caladan /kernel: da6: 4148MB (8496960 512 byte sectors: 64H 32S/T 4148C) " Before anyone jumps and says "Hmmm.. Seagate ST34371W with firmware 0202 isn't tested / is known 'rogue'" - I know... I've argued with Seagate who refuse to do anything for me with these drives, so I'm stuck with them... They worked fine under 3.4, I'd expect them to under 4.0 as well :) The kernel config I use to wire down the above is: " device ahc device ahc0 device ahc1 device ahc2 device scbus0 at ahc0 device scbus1 at ahc1 device scbus2 at ahc2 device da0 at scbus0 target 0 device da1 at scbus0 target 1 device da2 at scbus0 target 2 device da3 at scbus0 target 3 device da4 at scbus0 target 5 device da5 at scbus0 target 6 device da6 at scbus1 target 0 device da7 at scbus1 target 1 device da8 at scbus2 target 0 device da9 at scbus2 target 1 " Any help / info or suggestions gratefully received, as it is - I can only use the machine with the IDE's (under the new 'ad' drivers) - but these work fine, as did the SCSI once :) Regards, Karl Pielorz To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Apr 3 7:59:29 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from feral.com (feral.com [192.67.166.1]) by hub.freebsd.org (Postfix) with ESMTP id 692CE37BE1A for ; Mon, 3 Apr 2000 07:59:25 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from beppo.feral.com (beppo [192.67.166.79]) by feral.com (8.9.3/8.9.3) with ESMTP id HAA28995; Mon, 3 Apr 2000 07:58:02 -0700 Date: Mon, 3 Apr 2000 07:58:51 -0700 (PDT) From: Matthew Jacob Reply-To: mjacob@feral.com To: Frank Wegmann Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: Can't use 4GB 1024b/sec SCSI disk In-Reply-To: <38E89296.78012581@linguistics.ruhr-uni-bochum.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org This isn't a SCSI problem. This is a disklabel problem. On Mon, 3 Apr 2000, Frank Wegmann wrote: > Maybe someone on this list can be of help in this tricky issue. > > PROBLEM SUMMARY: > A Seagate ST34371N (4.3GB) at AHA 1542C on a 486 LocalBus host (64MB), > formatted with 1024 bytes per sector can be recognized but not used > under FreeBSD at all--on a (true) NeXTstation it can be used without > problem. Formatted with 512 bytes per sector yields half capacity on > both systems. > > DETAILS: > Originally I formatted it low-level with 1024 bytes per sector (for > getting out some more MB) and used it in a NeXT machine as boot disk > for years. Although it should--in theory--be possible to use it > directly under FreeBSD, this doesn't work at all. Probing the devices > (when invoking sysinstall) leads to: > > /kernel: dscheck(rda2): b_bcount 512 is not on a sector boundary (ssize > 1024) > > At startup, the disk's transliterated geometry plus capacity had been > identified correctly: > > da2 at aha0 bus 0 target 4 lun 0 > da2: Fixed Direct Access SCSI-2 device > da2: 3.333MB/s transfers (3.333MHz, offset 8) > da2: 4341MB (4445468 1024 byte sectors: 64H 32S/T 2170C) > > This is the physical layout of the drive (by querying it with the help of > the commands READ CAPACITY and MODE SENSE page 3 (format device page) and > page 4 (rigid disk layout page) -- done with a tool on the NeXT): > > 1024 bytes per sector > 86 sectors per track > 10 tracks per cylinder > 5168 cylinder per volume (including spare cylinders) > 62 spare sectors per cylinder > 0 alternate tracks per volume > 4445467 usable sectors for the disk > > If formatting via sysinstall doesn't work, why not doing the old-fashioned > way (according to 8.2.2 of the Handbook)? This is what I've got: > > # dd if=/dev/zero of=/dev/rda2 bs=1k count=1 > 1+0 records in > 1+0 records out > 1024 bytes transferred in 0.004346 secs (235618 bytes/sec) > # disklabel -Brw da2 auto > disklabel: /boot/boot2 too large <-- doh, try the other way: > # disklabel /dev/rda2 | disklabel -BrR da2 /dev/stdin > > Put it in a nutshell, I currently cannot use the disk at all, if it is > formatted with 1024 bytes. I tried to reformat it low-level under FreeBSD > with camcontrol, but that doesn't help either. > > However, with 512 bytes per sector I can use *half* of the disk. Strangely > enough I then have the same number of physical sectors (4445468), but with > 512 bytes per sector I only have 2.1 GB available. At least, it would be > useable under FreeBSD, but surely this is not what I want. Finally, using > the disk with 512 bytes at my NeXT also gives me 2.1 GB (probably, because > that number of sectors is returned on querying the device at the sense mode > level). > > With the SCSI knowledge that I have, I'm at my wits end. It seems to me that > FreeBSD refuses to work with the 1024 byte/sec disk because of its uncommon > physical geometry data (nearly all disks in the PC world have 512 bytes/sec). > Might be some tweaks at the driver level could help here. As far as that > halved capacity with 512 bytes/sec is concerned, this could be the drives > fault and perhaps it could be solved by somehow (?) writing the correct > number of sectors into the mode pages. > > Could anybody help in this case? > > TIA, > Frank > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-scsi" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Apr 3 8: 4:59 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from feral.com (feral.com [192.67.166.1]) by hub.freebsd.org (Postfix) with ESMTP id 8B64A37B9D7; Mon, 3 Apr 2000 08:04:52 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from beppo.feral.com (beppo [192.67.166.79]) by feral.com (8.9.3/8.9.3) with ESMTP id IAA29053; Mon, 3 Apr 2000 08:03:06 -0700 Date: Mon, 3 Apr 2000 08:03:55 -0700 (PDT) From: Matthew Jacob Reply-To: mjacob@feral.com To: Brad Knowles Cc: FreeBSD-CURRENT Mailing List , freebsd-scsi@FreeBSD.ORG Subject: Re: Sysctl for CAM SCSI quirks? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org These aren't available at boot time when you need them, are they? On Mon, 3 Apr 2000, Brad Knowles wrote: > Folks, > > I just did a bit of searching on the website, and garnered enough > information from > b/text/1998/freebsd-questions/19981115.freebsd-questions> to be able > to add a new "quirk" in /usr/src/sys/cam/cam_xpt.c for a piece of > hardware I'm having some difficulty with. > > Assuming this works, I will file a proper PR, etc.... > > > However, shouldn't there be an easier way to do this? For > example, shouldn't there be a sysctl for changing these sorts of > things, perhaps something like: > > kern.cam.cd.changer.min_busy_seconds: 5 > kern.cam.cd.changer.max_busy_seconds: 15 > kern.cam.quirks.direct.fixed.quantum.XP39100*.*.mintags: 24 > kern.cam.quirks.direct.fixed.quantum.XP39100*.*.maxtags: 32 > kern.cam.quirks.direct.fixed.quantum.XP34550*.*.mintags: 24 > kern.cam.quirks.direct.fixed.quantum.XP34550*.*.maxtags: 32 > kern.cam.quirks.direct.fixed.quantum.XP32275*.*.mintags: 24 > kern.cam.quirks.direct.fixed.quantum.XP32275*.*.maxtags: 32 > > This would make it very easy to add new quirks as they come along > (just update your sysctl.conf), and without having to always file a > PR so that /usr/src/sys/cam/cam_xpt.c gets modified so that you can > have your quirk and track -CURRENT too (or -STABLE, for that matter). > > -- > These are my opinions -- not to be taken as official Skynet policy > ====================================================================== > Brad Knowles, || Belgacom Skynet SA/NV > Systems Architect, Mail/News/FTP/Proxy Admin || Rue Colonel Bourg, 124 > Phone/Fax: +32-2-706.13.11/12.49 || B-1140 Brussels > http://www.skynet.be || Belgium > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-scsi" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Apr 3 12:19:54 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from mailman.zeta.org.au (mailman.zeta.org.au [203.26.10.16]) by hub.freebsd.org (Postfix) with ESMTP id 5DF1D37B630 for ; Mon, 3 Apr 2000 12:19:48 -0700 (PDT) (envelope-from bde@zeta.org.au) Received: from bde.zeta.org.au (bde.zeta.org.au [203.2.228.102]) by mailman.zeta.org.au (8.8.7/8.8.7) with ESMTP id FAA23950; Tue, 4 Apr 2000 05:19:08 +1000 Date: Tue, 4 Apr 2000 05:18:50 +1000 (EST) From: Bruce Evans X-Sender: bde@alphplex.bde.org To: Frank Wegmann Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: Can't use 4GB 1024b/sec SCSI disk In-Reply-To: <38E89296.78012581@linguistics.ruhr-uni-bochum.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Mon, 3 Apr 2000, Frank Wegmann wrote: > ... > DETAILS: > Originally I formatted it low-level with 1024 bytes per sector (for > getting out some more MB) and used it in a NeXT machine as boot disk > for years. Although it should--in theory--be possible to use it > directly under FreeBSD, this doesn't work at all. Probing the devices > (when invoking sysinstall) leads to: > > /kernel: dscheck(rda2): b_bcount 512 is not on a sector boundary (ssize > 1024) This seems to be a bug in sysinstall. More precisely, in libdisk. libdisk uses lots of hard-coded 512's. It doesn't even use DEV_BSIZE. It only uses d_secsize in the label in order to initialize it to 512. > If formatting via sysinstall doesn't work, why not doing the old-fashioned > way (according to 8.2.2 of the Handbook)? This is what I've got: > > # dd if=/dev/zero of=/dev/rda2 bs=1k count=1 > 1+0 records in > 1+0 records out > 1024 bytes transferred in 0.004346 secs (235618 bytes/sec) The label is in the second sector, so you need a count of 2 to overwrite it. > # disklabel -Brw da2 auto "auto" is new-fashioned. > disklabel: /boot/boot2 too large <-- doh, try the other way: This means that the number of sectors specified in the label (either sectors/unit or the offset of the end of a partition) is too large for the disk. The number must be given in sector-sized units, not in DEV_BSIZE units. "disklabel -auto" is guaranteed to get this wrong unless the device described by da2 (i.e., /dev/da2c) has the same size as the whole disk, since it produces a label suitable for applying to the whole disk. However, it should get the units right. > # disklabel /dev/rda2 | disklabel -BrR da2 /dev/stdin This (when uncommented) gives the same result as "auto". The output of "disklabel /dev/rda2" must be edited if the device being labeled (/dev/rda2c) is smaller than /dev/rda2. > Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Apr 3 14:58:30 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from springbank.linguistics.ruhr-uni-bochum.de (springbank.linguistics.ruhr-uni-bochum.de [134.147.14.9]) by hub.freebsd.org (Postfix) with ESMTP id 2A05837BA6C for ; Mon, 3 Apr 2000 14:58:23 -0700 (PDT) (envelope-from wegmann@acm.org) Received: from dialppp-3-117.rz.ruhr-uni-bochum.de ([134.147.3.117] helo=acm.org) by springbank.linguistics.ruhr-uni-bochum.de with esmtp (Exim 3.12 #1) id 12cEot-0004FA-00; Mon, 03 Apr 2000 23:55:39 +0200 Message-ID: <38E913F1.87479CF@acm.org> Date: Mon, 03 Apr 2000 23:58:09 +0200 From: Frank Wegmann Reply-To: wegmann@linguistics.ruhr-uni-bochum.de X-Mailer: Mozilla 4.72 [en] (X11; I; Linux 2.2.12 i386) X-Accept-Language: en MIME-Version: 1.0 To: bde@zeta.org.au Cc: freebsd-scsi@freebsd.org Subject: Re: Can't use 4GB 1024b/sec SCSI disk Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > This seems to be a bug in sysinstall. More precisely, in libdisk. libdisk > uses lots of hard-coded 512's. It doesn't even use DEV_BSIZE. It only > uses d_secsize in the label in order to initialize it to 512. Hopefully someone is going to fix this. Frankly I can't believe that I'm the first stumbling over this problem. > > # dd if=/dev/zero of=/dev/rda2 bs=1k count=1 > > 1+0 records in > > 1+0 records out > > 1024 bytes transferred in 0.004346 secs (235618 bytes/sec) > > The label is in the second sector, so you need a count of 2 to overwrite it. Did that: # dd if=/dev/zero of=/dev/rda2 bs=1k count=1 2+0 records in 2+0 records out 2048 bytes transferred in 0.006840 secs (299416 bytes/sec) > This means that the number of sectors specified in the label (either > sectors/unit or the offset of the end of a partition) is too large > for the disk. The number must be given in sector-sized units, not in > DEV_BSIZE units. "disklabel -auto" is guaranteed to get this wrong > unless the device described by da2 (i.e., /dev/da2c) has the same size > as the whole disk, since it produces a label suitable for applying to > the whole disk. However, it should get the units right. > > > # disklabel /dev/rda2 | disklabel -BrR da2 /dev/stdin > > This (when uncommented) gives the same result as "auto". The output of > "disklabel /dev/rda2" must be edited if the device being labeled > (/dev/rda2c) is smaller than /dev/rda2. Hmm here I'm getting lost, sorry. Here is the output of disklabel: tanqueray<35># disklabel /dev/rda2 # /dev/rda2: type: SCSI disk: SEAGATE label: ST34371N flags: bytes/sector: 1024 sectors/track: 32 tracks/cylinder: 64 sectors/cylinder: 2048 cylinders: 2170 sectors/unit: 4445468 rpm: 3600 interleave: 1 trackskew: 0 cylinderskew: 0 headswitch: 0 # milliseconds track-to-track-seek: 0 # milliseconds drivedata: 0 8 partitions: # size offset fstype [fsize bsize bps/cpg] c: 4445468 0 unused 0 0 # # disklabel /dev/rda2c disklabel: ioctl DIOCGDINFO: Invalid argument One of the things I don't quite understand here is what information must be used for disklabel: data of the physical driver or of the transliterated one. I'm still puzzled and my disk still doesn't want to cooperate... Frank To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Tue Apr 4 12: 4: 1 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by hub.freebsd.org (Postfix) with ESMTP id A4B2A37B6F8 for ; Tue, 4 Apr 2000 12:03:58 -0700 (PDT) (envelope-from gibbs@narnia.plutotech.com) Received: (from gibbs@localhost) by narnia.plutotech.com (8.9.3/8.7.3) id MAA00285; Tue, 4 Apr 2000 12:49:02 -0600 (MDT) Date: Tue, 4 Apr 2000 12:49:02 -0600 (MDT) From: "Justin T. Gibbs" Message-Id: <200004041849.MAA00285@narnia.plutotech.com> To: Karl Pielorz Cc: scsi@FreeBSD.org Subject: Re: Problems with 4.0 - CAM when 3.4 was OK? (long post) X-Newsgroups: pluto.freebsd.scsi In-Reply-To: <38E76085.9E5BE928@tdx.co.uk> User-Agent: tin/pre-1.4-980818 ("Laura") (UNIX) (FreeBSD/4.0-CURRENT (i386)) Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In article <38E76085.9E5BE928@tdx.co.uk> you wrote: > Hi All, > > I've recently upgraded my 3.4-STABLE box to 4.0-STABLE. The upgrade went fine, > except something seems to have gone awry with my SCSI set-up... ... > If I boot a 4.0-STABLE kernel, I get heaps of: > > " > Waiting 4 seconds for SCSI devices to settle > de0: enabling 10baseT port > (probe15:ahc1:0:0:0): SCB 0xe - timed out in Message-in phase, SEQADDR == > 0x157 > (probe15:ahc1:0:0:0): Other SCB Timeout > (probe16:ahc1:0:1:0): SCB 0xd - timed out in Message-in phase, SEQADDR == > 0x157 This really looks like the system is not seeing interrupts from this card. I don't know why (perhaps an mptable problem), but the errors are consistent with that. -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Wed Apr 5 1: 9:25 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from caladan.tdx.co.uk (caladan.tdx.co.uk [195.188.177.4]) by hub.freebsd.org (Postfix) with ESMTP id 50CCC37B8BE for ; Wed, 5 Apr 2000 01:09:12 -0700 (PDT) (envelope-from kpielorz@tdx.co.uk) Received: from tdx.co.uk (lorca-tx.tdx.co.uk [195.188.177.242]) by caladan.tdx.co.uk (8.10.0/8.10.0/Kp) with ESMTP id e3589XG11608; Wed, 5 Apr 2000 09:09:33 +0100 (BST) Message-ID: <38EAF491.93C92B8E@tdx.co.uk> Date: Wed, 05 Apr 2000 09:08:49 +0100 From: Karl Pielorz Organization: TDX - The Digital eXchange X-Mailer: Mozilla 4.7 [en-gb] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: "Justin T. Gibbs" Cc: scsi@FreeBSD.ORG Subject: Re: Problems with 4.0 - CAM when 3.4 was OK? (long post) References: <200004041849.MAA00285@narnia.plutotech.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org "Justin T. Gibbs" wrote: > > If I boot a 4.0-STABLE kernel, I get heaps of: > > > > " > > Waiting 4 seconds for SCSI devices to settle > > de0: enabling 10baseT port > > (probe15:ahc1:0:0:0): SCB 0xe - timed out in Message-in phase, SEQADDR == > > 0x157 > > (probe15:ahc1:0:0:0): Other SCB Timeout > > (probe16:ahc1:0:1:0): SCB 0xd - timed out in Message-in phase, SEQADDR == > > 0x157 > > This really looks like the system is not seeing interrupts from this > card. I don't know why (perhaps an mptable problem), but the > errors are consistent with that. Hi Justin, Thanks for the pointer - I think you may have hit this straight on the head (and thanks for not even mentioning cables :-) Under 3.X, in the dmesg I had lurking some: /kernel: bogus MP table, 2 IO APIC pins connected to the same PCI device or ISA/EISA interrupt /kernel: Registered extra interrupt handler for int 18 (in addition to int 16) Under 4.0, in the dmesg I don't appear to see them... So the question has to be - where's my buggy APIC workaround gone from 3.X to 4.0? :) I'm going to double check on this, then fire a post to -SMP, Thanks -Karl To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Wed Apr 5 8:15:30 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from overcee.netplex.com.au (peter1.yahoo.com [208.48.107.4]) by hub.freebsd.org (Postfix) with ESMTP id F149037BE87 for ; Wed, 5 Apr 2000 08:15:25 -0700 (PDT) (envelope-from peter@netplex.com.au) Received: from netplex.com.au (localhost [127.0.0.1]) by overcee.netplex.com.au (Postfix) with ESMTP id CCB231CE0; Wed, 5 Apr 2000 08:15:23 -0700 (PDT) (envelope-from peter@netplex.com.au) X-Mailer: exmh version 2.1.1 10/15/1999 To: Karl Pielorz Cc: "Justin T. Gibbs" , scsi@FreeBSD.ORG Subject: Re: Problems with 4.0 - CAM when 3.4 was OK? (long post) In-Reply-To: Message from Karl Pielorz of "Wed, 05 Apr 2000 09:08:49 BST." <38EAF491.93C92B8E@tdx.co.uk> Date: Wed, 05 Apr 2000 08:15:23 -0700 From: Peter Wemm Message-Id: <20000405151523.CCB231CE0@overcee.netplex.com.au> Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Karl Pielorz wrote: > "Justin T. Gibbs" wrote: > > > > If I boot a 4.0-STABLE kernel, I get heaps of: > > > > > > " > > > Waiting 4 seconds for SCSI devices to settle > > > de0: enabling 10baseT port > > > (probe15:ahc1:0:0:0): SCB 0xe - timed out in Message-in phase, SEQADDR == > > > 0x157 > > > (probe15:ahc1:0:0:0): Other SCB Timeout > > > (probe16:ahc1:0:1:0): SCB 0xd - timed out in Message-in phase, SEQADDR == > > > 0x157 > > > > This really looks like the system is not seeing interrupts from this > > card. I don't know why (perhaps an mptable problem), but the > > errors are consistent with that. > > Hi Justin, > > Thanks for the pointer - I think you may have hit this straight on the head > (and thanks for not even mentioning cables :-) > > Under 3.X, in the dmesg I had lurking some: > /kernel: bogus MP table, 2 IO APIC pins connected to the same PCI device or > ISA/EISA interrupt > /kernel: Registered extra interrupt handler for int 18 (in addition to int > 16) > > Under 4.0, in the dmesg I don't appear to see them... > > So the question has to be - where's my buggy APIC workaround gone from 3.X to > 4.0? :) > > I'm going to double check on this, then fire a post to -SMP, > > Thanks > > -Karl This is a known problem. The workaround for the mptable and/or hardware configuration problem was implemented in the wrong place and hasn't been repaired and made to work for all drivers yet. Cheers, -Peter -- Peter Wemm - peter@FreeBSD.org; peter@yahoo-inc.com; peter@netplex.com.au "All of this is for nothing if we don't go to the stars" - JMS/B5 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Wed Apr 5 17:23:54 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from arjun.niksun.com (gwnew.niksun.com [206.20.52.130]) by hub.freebsd.org (Postfix) with ESMTP id EEA4D37B996 for ; Wed, 5 Apr 2000 17:23:47 -0700 (PDT) (envelope-from ath@niksun.com) Received: from stiegl.niksun.com (stiegl.niksun.com [10.0.0.44]) by arjun.niksun.com (8.9.3/8.9.3) with ESMTP id IAA92261 for ; Wed, 5 Apr 2000 08:18:33 -0400 (EDT) (envelope-from ath@stiegl.niksun.com) Received: from stiegl.niksun.com (localhost.niksun.com [127.0.0.1]) by stiegl.niksun.com (8.9.2/8.8.7) with ESMTP id IAA09904 for ; Wed, 5 Apr 2000 08:18:33 -0400 (EDT) (envelope-from ath@stiegl.niksun.com) Message-Id: <200004051218.IAA09904@stiegl.niksun.com> From: Andrew Heybey To: freebsd-scsi@freebsd.org Subject: newfs on IBM disks slower than Seagate disks? Mime-Version: 1.0 (generated by tm-edit 7.108) Content-Type: text/plain; charset=US-ASCII Date: Wed, 05 Apr 2000 08:18:33 -0400 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Newfs of a ~16GB partition (as performed by sysinstall, so the newfs arguments are the same) is *much* slower on IBM 18GB 10K RPM LVD disks versus similar Seagates. Systems are otherwise identical (same controller (onboard Adaptec AIC7896), same motherboard, same amount of RAM). Once newfs'd, bonnie and iozone give similar performance for the two disks. Rawio also gives similar numbers for the two. Running 3.2-RELEASE. IBM disks are DMVS18V. Seagates are Cheetah ST318203LW. Why would this be the case? andrew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Wed Apr 5 17:28:56 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from panzer.kdm.org (panzer.kdm.org [216.160.178.169]) by hub.freebsd.org (Postfix) with ESMTP id 9E04E37BE41 for ; Wed, 5 Apr 2000 17:28:53 -0700 (PDT) (envelope-from ken@panzer.kdm.org) Received: (from ken@localhost) by panzer.kdm.org (8.9.3/8.9.1) id SAA06981; Wed, 5 Apr 2000 18:28:27 -0600 (MDT) (envelope-from ken) Date: Wed, 5 Apr 2000 18:28:27 -0600 From: "Kenneth D. Merry" To: Andrew Heybey Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: newfs on IBM disks slower than Seagate disks? Message-ID: <20000405182827.A6960@panzer.kdm.org> References: <200004051218.IAA09904@stiegl.niksun.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0i In-Reply-To: <200004051218.IAA09904@stiegl.niksun.com>; from ath@niksun.com on Wed, Apr 05, 2000 at 08:18:33AM -0400 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Wed, Apr 05, 2000 at 08:18:33 -0400, Andrew Heybey wrote: > Newfs of a ~16GB partition (as performed by sysinstall, so the newfs > arguments are the same) is *much* slower on IBM 18GB 10K RPM LVD disks > versus similar Seagates. Systems are otherwise identical (same > controller (onboard Adaptec AIC7896), same motherboard, same amount of > RAM). Once newfs'd, bonnie and iozone give similar performance for > the two disks. Rawio also gives similar numbers for the two. > > Running 3.2-RELEASE. > > IBM disks are DMVS18V. > Seagates are Cheetah ST318203LW. > > Why would this be the case? See if write caching is enabled (WCE == 1) in mode page 8. (See the camcontrol(8) man page for how to display mode pages.) Ken -- Kenneth Merry ken@kdm.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Wed Apr 5 18:59:47 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from feral.com (feral.com [192.67.166.1]) by hub.freebsd.org (Postfix) with ESMTP id ED2CB37B5F0; Wed, 5 Apr 2000 18:59:43 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from beppo.feral.com (beppo [192.67.166.79]) by feral.com (8.9.3/8.9.3) with ESMTP id SAA06338; Wed, 5 Apr 2000 18:59:13 -0700 Date: Wed, 5 Apr 2000 18:59:28 -0700 (PDT) From: Matthew Jacob Reply-To: mjacob@feral.com To: Andrew Heybey Cc: freebsd-bugs@freebsd.org, freebsd-scsi@freebsd.org Subject: Re: kern/17153 (was: newfs on IBM disks slower than Seagate disks?) In-Reply-To: <200004051218.IAA09904@stiegl.niksun.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Very interesting. See kern/17153. This carries over, btw, to NetBSD as well. I have a theory about the problem for the Qlogic controller, but I don't know if the same theory would apply to the Adaptec- maybe not because the bonnie numbers are different. On Wed, 5 Apr 2000, Andrew Heybey wrote: > Newfs of a ~16GB partition (as performed by sysinstall, so the newfs > arguments are the same) is *much* slower on IBM 18GB 10K RPM LVD disks > versus similar Seagates. Systems are otherwise identical (same > controller (onboard Adaptec AIC7896), same motherboard, same amount of > RAM). Once newfs'd, bonnie and iozone give similar performance for > the two disks. Rawio also gives similar numbers for the two. > > Running 3.2-RELEASE. > > IBM disks are DMVS18V. > Seagates are Cheetah ST318203LW. > > Why would this be the case? > > andrew > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-scsi" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Wed Apr 5 19: 6:48 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from caspian.plutotech.com (caspian.plutotech.com [206.168.67.80]) by hub.freebsd.org (Postfix) with ESMTP id 4B4D637B80C; Wed, 5 Apr 2000 19:06:44 -0700 (PDT) (envelope-from gibbs@caspian.plutotech.com) Received: from caspian.plutotech.com (localhost [127.0.0.1]) by caspian.plutotech.com (8.9.3/8.9.1) with ESMTP id UAA24407; Wed, 5 Apr 2000 20:07:26 -0600 (MDT) (envelope-from gibbs@caspian.plutotech.com) Message-Id: <200004060207.UAA24407@caspian.plutotech.com> X-Mailer: exmh version 2.1.1 10/15/1999 To: mjacob@feral.com Cc: Andrew Heybey , freebsd-bugs@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG Subject: Re: kern/17153 (was: newfs on IBM disks slower than Seagate disks?) In-Reply-To: Your message of "Wed, 05 Apr 2000 18:59:28 PDT." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 05 Apr 2000 20:07:26 -0600 From: "Justin T. Gibbs" Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> Newfs of a ~16GB partition (as performed by sysinstall, so the newfs >> arguments are the same) is *much* slower on IBM 18GB 10K RPM LVD disks >> versus similar Seagates. Systems are otherwise identical (same >> controller (onboard Adaptec AIC7896), same motherboard, same amount of >> RAM). Once newfs'd, bonnie and iozone give similar performance for >> the two disks. Rawio also gives similar numbers for the two. >> >> Running 3.2-RELEASE. >> >> IBM disks are DMVS18V. >> Seagates are Cheetah ST318203LW. >> >> Why would this be the case? IBM usually ships with the write cache disabled. Seagate almost always has it enabled. As newfs is a "single blocking I/O at a time" kind of application, the additional write latency causes a degradation in performance. When going through the filesystem, this latency is hidden by the buffer cache. You can view mode page 8 with camcontrol to determine the settings for your drives. Man camcontrol for details. -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Thu Apr 6 7:51:49 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from ns1.networkiowa.com (ns1.networkiowa.com [209.234.64.192]) by hub.freebsd.org (Postfix) with ESMTP id 04A2E37B5A8 for ; Thu, 6 Apr 2000 07:51:47 -0700 (PDT) (envelope-from johnl@raccoon.com) Received: from raccoon.com (dsl.72.145.networkiowa.com [209.234.72.145]) by ns1.networkiowa.com (8.9.3/8.9.3) with ESMTP id JAA23823; Thu, 6 Apr 2000 09:54:09 -0500 Message-ID: <38ECA464.97DFEDB3@raccoon.com> Date: Thu, 06 Apr 2000 09:51:16 -0500 From: John Lengeling X-Mailer: Mozilla 4.7 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 To: "Kenneth D. Merry" Cc: Andrew Heybey , freebsd-scsi@FreeBSD.ORG Subject: Re: newfs on IBM disks slower than Seagate disks? References: <200004051218.IAA09904@stiegl.niksun.com> <20000405182827.A6960@panzer.kdm.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > See if write caching is enabled (WCE == 1) in mode page 8. (See the > camcontrol(8) man page for how to display mode pages.) > > Ken > Should it be on or off? What does WCE do? johnl To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Thu Apr 6 8:20: 4 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from panzer.kdm.org (panzer.kdm.org [216.160.178.169]) by hub.freebsd.org (Postfix) with ESMTP id 08B4037B505 for ; Thu, 6 Apr 2000 08:20:01 -0700 (PDT) (envelope-from ken@panzer.kdm.org) Received: (from ken@localhost) by panzer.kdm.org (8.9.3/8.9.1) id JAA11641; Thu, 6 Apr 2000 09:19:53 -0600 (MDT) (envelope-from ken) Date: Thu, 6 Apr 2000 09:19:52 -0600 From: "Kenneth D. Merry" To: John Lengeling Cc: Andrew Heybey , freebsd-scsi@FreeBSD.ORG Subject: Re: newfs on IBM disks slower than Seagate disks? Message-ID: <20000406091952.A11611@panzer.kdm.org> References: <200004051218.IAA09904@stiegl.niksun.com> <20000405182827.A6960@panzer.kdm.org> <38ECA464.97DFEDB3@raccoon.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0i In-Reply-To: <38ECA464.97DFEDB3@raccoon.com>; from johnl@raccoon.com on Thu, Apr 06, 2000 at 09:51:16AM -0500 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, Apr 06, 2000 at 09:51:16 -0500, John Lengeling wrote: > > > > > See if write caching is enabled (WCE == 1) in mode page 8. (See the > > camcontrol(8) man page for how to display mode pages.) > > > > Ken > > > > Should it be on or off? What does WCE do? WCE means Write Cache Enabled. I generally have write caching turned on with my disks, but it can be a rather religious issue. You can probably find lots of discussion about it if you search various FreeBSD list archives. (I'm not going to recommend one way or the other.) Ken -- Kenneth Merry ken@kdm.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Thu Apr 6 13:25:48 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from front6m.grolier.fr (front6m.grolier.fr [195.36.216.56]) by hub.freebsd.org (Postfix) with ESMTP id 5C66737BAF4 for ; Thu, 6 Apr 2000 13:25:38 -0700 (PDT) (envelope-from groudier@club-internet.fr) Received: from ppp-109-30.villette.club-internet.fr (ppp-109-30.villette.club-internet.fr [194.158.109.30]) by front6m.grolier.fr (8.9.3/No_Relay+No_Spam_MGC990224) with ESMTP id WAA22781; Thu, 6 Apr 2000 22:24:58 +0200 (MET DST) Date: Thu, 6 Apr 2000 21:58:11 +0200 (CEST) From: =?ISO-8859-1?Q?G=E9rard_Roudier?= X-Sender: groudier@linux.local To: "Kenneth D. Merry" Cc: John Lengeling , Andrew Heybey , freebsd-scsi@FreeBSD.ORG Subject: Re: newfs on IBM disks slower than Seagate disks? In-Reply-To: <20000406091952.A11611@panzer.kdm.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, 6 Apr 2000, Kenneth D. Merry wrote: > On Thu, Apr 06, 2000 at 09:51:16 -0500, John Lengeling wrote: > >=20 > > >=20 > > > See if write caching is enabled (WCE =3D=3D 1) in mode page 8. (See = the > > > camcontrol(8) man page for how to display mode pages.) > > >=20 > > > Ken > > > > >=20 > > Should it be on or off? What does WCE do? >=20 > WCE means Write Cache Enabled. >=20 > I generally have write caching turned on with my disks, but it can be a > rather religious issue. You can probably find lots of discussion about i= t > if you search various FreeBSD list archives. (I'm not going to recommend > one way or the other.) I will do slightly different. :-) When WCE bit =3D 0, the device shall return GOOD SCSI status after DATA has= =20 been actually written to the medium (on WRITE commands obviously). So, even when WCE=3D0, the device can use its cache memory to optimize=20 actual write operations by sorting and coalescing commands, provided that= =20 it is given several commands simultaneously. When WCE bit =3D 1, the device is allowed to return GOOD status as soon as= =20 it has buffered the data of a WRITE operation. In result, when WCE bit =3D 1, the cache memory of the device can be populated with numerous pending data for writing, reducing the memory available for prefetching and decreasing READ performances. In my opinion, given realistic IO patterns and not too mistaken IO subsystems, WCE =3D 1 is mostly good with reducing IO performances in numerous (most?) situations (of not using a silly benchmark or a silly IO pattern). However, in the situation of a system performing bunches of small WRITES it wants (or believes) to be synchronous, WCE =3D 1 can magically increase performances if the device is just not told about these IOs having to be performed synchronously (FUA bit should be used). I generally have write caching enabled on my testing machine, not because I like this option but for the reason I know people generally use it and that this option is a good stresser for disk firmwares. But I donnot recommend to blindly use this option on production systems without having checked its actual effect on performances. If you are unsure, you should probably disable WCE, since the risk of losing both performances and reliability does exist, in my opinion, when WCE bit is set. Obviously this depends on so many factors, that a single recommendation is not possible. G=E9rard. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Thu Apr 6 19: 4: 3 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from henry.cs.adfa.edu.au (henry.cs.adfa.edu.au [131.236.21.158]) by hub.freebsd.org (Postfix) with ESMTP id E19B337BF01 for ; Thu, 6 Apr 2000 19:03:59 -0700 (PDT) (envelope-from wkt@henry.cs.adfa.edu.au) Received: (from wkt@localhost) by henry.cs.adfa.edu.au (8.9.2/8.9.3) id MAA19720 for freebsd-scsi@freebsd.org; Fri, 7 Apr 2000 12:03:57 +1000 (EST) From: Warren Toomey Message-Id: <200004070203.MAA19720@henry.cs.adfa.edu.au> Subject: Adaptec 2906, any users? To: freebsd-scsi@freebsd.org Date: Fri, 7 Apr 2000 12:03:57 +1000 (EST) Reply-To: wkt@cs.adfa.edu.au X-Mailer: ELM [version 2.4ME+ PL43 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Back in December 1999, Justin Gibbs wrote: > The driver does not have an entry for [the Adaptec 2906] card, > but it should be caught by one of the generic entries. If it doesn't > work for you, I should be able to get you a patch for it as soon as > you can supply the PCI vendor/device/subvendor/subdevice ids > from a boot -v. > > -- > Justin I'm just about to buy this card. Has anybody get this working in 3.4-RELEASE or 4.0-RELEASE since then? Justin, did the person supply you with the PCI device ids? I bought a Ricoh MP7060S CD burner, which comes with an ACARD SCSI controller that isn't recognised by 4.0-RELEASE. Anybody heard of that, by any chance? I don't have the card here with me, so I can't quote chip details. Cheers, Warren To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Apr 7 6:18:28 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from arjun.niksun.com (gwnew.niksun.com [206.20.52.130]) by hub.freebsd.org (Postfix) with ESMTP id 250C237BB71; Fri, 7 Apr 2000 06:18:20 -0700 (PDT) (envelope-from ath@niksun.com) Received: from celis.niksun.com (celis.niksun.com [10.1.0.48]) by arjun.niksun.com (8.9.3/8.9.3) with ESMTP id JAA15229; Fri, 7 Apr 2000 09:18:16 -0400 (EDT) (envelope-from ath@celis.niksun.com) Received: from celis (localhost [127.0.0.1]) by celis.niksun.com (8.9.3/8.9.2) with ESMTP id JAA15202; Fri, 7 Apr 2000 09:18:02 -0400 (EDT) (envelope-from ath@celis.niksun.com) Message-Id: <200004071318.JAA15202@celis.niksun.com> From: Andrew Heybey To: "Justin T. Gibbs" Cc: mjacob@feral.com, freebsd-bugs@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG Subject: Re: kern/17153 (was: newfs on IBM disks slower than Seagate disks?) In-reply-to: Your message of Wed, 05 Apr 2000 20:07:26 -0600. <200004060207.UAA24407@caspian.plutotech.com> Mime-Version: 1.0 (generated by tm-edit 7.108) Content-Type: text/plain; charset=US-ASCII Date: Fri, 07 Apr 2000 09:18:02 -0400 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org It would seem that the WCE bit is the answer. The IBM has it set to zero while the Seagates have it set to 1. Unlike Matt Jacob, I get substantially the same performance for normal IO: IBM: install8 1# dd if=/dev/rda0s1c of=/dev/null bs=1024k count=200 200+0 records in 200+0 records out 209715200 bytes transferred in 7.561781 secs (27733572 bytes/sec) Seagate: su-2.03# dd if=/dev/rda0s1c of=/dev/null bs=1024k count=200 200+0 records in 200+0 records out 209715200 bytes transferred in 7.520681 secs (27885135 bytes/sec) I don't have any partitions lying around that I can newfs to test the WCE hypothesis, but it seems the most logical answer. andrew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Apr 7 6:39:15 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from worf.qntm.com (worf.qntm.com [146.174.250.100]) by hub.freebsd.org (Postfix) with ESMTP id 95ECB37BB55 for ; Fri, 7 Apr 2000 06:39:12 -0700 (PDT) (envelope-from Stephen.Byan@quantum.com) Received: from mail3.qntm.com by worf.qntm.com with ESMTP (1.40.112.12/16.2) id AA074444752; Fri, 7 Apr 2000 06:39:12 -0700 Received: from milcmimb.qntm.com (milcmimb.qntm.com [146.174.18.77]) by mail3.qntm.com (8.8.6/8.8.6) with ESMTP id GAA25348 for ; Fri, 7 Apr 2000 06:39:15 -0700 (PDT) Received: by milcmimb.qntm.com with Internet Mail Service (5.5.2650.10) id ; Fri, 7 Apr 2000 06:39:07 -0700 Message-Id: <8133266FE373D11190CD00805FA768BF02EE9F64@shrcmsg1.tdh.qntm.com> From: Stephen Byan To: freebsd-scsi@FreeBSD.ORG Subject: RE: newfs on IBM disks slower than Seagate disks? Date: Fri, 7 Apr 2000 06:39:01 -0700 Mime-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.10) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Does the FreeBSD SCSI subsystem set the FUA bit in the CDB for UFS metadata writes? If so, then data integrity with WCE=1 is probably no worse than for WCE=0, since the filesystem is caching non-metadata writes anyway. If UFS and CAM haven't made arrangements to hint which disk writes are precious, then I think you're best off setting WCE=0, unless your system and your disks are on a UPS. Regards, -Steve To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Apr 7 8:30:50 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from panzer.kdm.org (panzer.kdm.org [216.160.178.169]) by hub.freebsd.org (Postfix) with ESMTP id 9F54937BEDD for ; Fri, 7 Apr 2000 08:30:44 -0700 (PDT) (envelope-from ken@panzer.kdm.org) Received: (from ken@localhost) by panzer.kdm.org (8.9.3/8.9.1) id JAA19624; Fri, 7 Apr 2000 09:29:55 -0600 (MDT) (envelope-from ken) Date: Fri, 7 Apr 2000 09:29:55 -0600 From: "Kenneth D. Merry" To: Stephen Byan Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: newfs on IBM disks slower than Seagate disks? Message-ID: <20000407092955.A19590@panzer.kdm.org> References: <8133266FE373D11190CD00805FA768BF02EE9F64@shrcmsg1.tdh.qntm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0i In-Reply-To: <8133266FE373D11190CD00805FA768BF02EE9F64@shrcmsg1.tdh.qntm.com>; from Stephen.Byan@quantum.com on Fri, Apr 07, 2000 at 06:39:01AM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Fri, Apr 07, 2000 at 06:39:01 -0700, Stephen Byan wrote: > Does the FreeBSD SCSI subsystem set the FUA bit in the CDB for UFS metadata > writes? If so, then data integrity with WCE=1 is probably no worse than for > WCE=0, since the filesystem is caching non-metadata writes anyway. > > If UFS and CAM haven't made arrangements to hint which disk writes are > precious, then I think you're best off setting WCE=0, unless your system and > your disks are on a UPS. CAM doesn't set the FUA bit on metadata writes because it doesn't currently have a way to distinguish between metadata and normal data. Ken -- Kenneth Merry ken@kdm.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Apr 7 8:56:57 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from worf.qntm.com (worf.qntm.com [146.174.250.100]) by hub.freebsd.org (Postfix) with ESMTP id 684AC37BDB0 for ; Fri, 7 Apr 2000 08:56:55 -0700 (PDT) (envelope-from Stephen.Byan@quantum.com) Received: from mail3.qntm.com by worf.qntm.com with ESMTP (1.40.112.12/16.2) id AA167763010; Fri, 7 Apr 2000 08:56:50 -0700 Received: from milcmimb.qntm.com (milcmimb.qntm.com [146.174.18.77]) by mail3.qntm.com (8.8.6/8.8.6) with ESMTP id IAA21117; Fri, 7 Apr 2000 08:56:53 -0700 (PDT) Received: by milcmimb.qntm.com with Internet Mail Service (5.5.2650.10) id ; Fri, 7 Apr 2000 08:56:45 -0700 Message-Id: <8133266FE373D11190CD00805FA768BF02EE9F66@shrcmsg1.tdh.qntm.com> From: Stephen Byan To: "'Kenneth D. Merry'" , Stephen Byan Cc: freebsd-scsi@FreeBSD.ORG Subject: RE: newfs on IBM disks slower than Seagate disks? Date: Fri, 7 Apr 2000 08:56:13 -0700 Mime-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.10) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Too bad. It'd be a useful thing for the I/O subsystem to know, both from a data integrity and from a performance viewpoint. In one of my former lives (at Hitachi), we hacked OSF/1's flavor of UFS to pass a "this is metadata" flag down to the disk device driver in the buf header. I arranged for the big honking mainframe disk controllers to put the unflagged writes in their copious volatile cache, and the flagged writes in their more-limited non-volatile cache. Regards, -Steve -----Original Message----- From: Kenneth D. Merry [mailto:ken@kdm.org] Sent: Friday, April 07, 2000 11:30 AM To: Stephen Byan Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: newfs on IBM disks slower than Seagate disks? On Fri, Apr 07, 2000 at 06:39:01 -0700, Stephen Byan wrote: > Does the FreeBSD SCSI subsystem set the FUA bit in the CDB for UFS metadata > writes? If so, then data integrity with WCE=1 is probably no worse than for > WCE=0, since the filesystem is caching non-metadata writes anyway. > > If UFS and CAM haven't made arrangements to hint which disk writes are > precious, then I think you're best off setting WCE=0, unless your system and > your disks are on a UPS. CAM doesn't set the FUA bit on metadata writes because it doesn't currently have a way to distinguish between metadata and normal data. Ken -- Kenneth Merry ken@kdm.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sat Apr 8 21:26: 2 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from gidora.zeta.org.au (gidora.zeta.org.au [203.26.10.25]) by hub.freebsd.org (Postfix) with SMTP id 1425B37B70D for ; Sat, 8 Apr 2000 21:25:56 -0700 (PDT) (envelope-from bde@zeta.org.au) Received: (qmail 15753 invoked from network); 9 Apr 2000 04:25:52 -0000 Received: from bde.zeta.org.au (203.2.228.102) by gidora.zeta.org.au with SMTP; 9 Apr 2000 04:25:52 -0000 Date: Sun, 9 Apr 2000 14:25:31 +1000 (EST) From: Bruce Evans X-Sender: bde@alphplex.bde.org To: "Kenneth D. Merry" Cc: Stephen Byan , freebsd-scsi@FreeBSD.ORG Subject: Re: newfs on IBM disks slower than Seagate disks? In-Reply-To: <20000407092955.A19590@panzer.kdm.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Fri, 7 Apr 2000, Kenneth D. Merry wrote: > On Fri, Apr 07, 2000 at 06:39:01 -0700, Stephen Byan wrote: > > Does the FreeBSD SCSI subsystem set the FUA bit in the CDB for UFS metadata > > writes? If so, then data integrity with WCE=1 is probably no worse than for > > WCE=0, since the filesystem is caching non-metadata writes anyway. > > ... > > CAM doesn't set the FUA bit on metadata writes because it doesn't currently > have a way to distinguish between metadata and normal data. The B_ASYNC flag gives an even more relevant distinction. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message