From owner-freebsd-scsi Tue Apr 11 02:29:11 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id CAA17506 for freebsd-scsi-outgoing; Tue, 11 Apr 1995 02:29:11 -0700 Received: from tuminfo2.informatik.tu-muenchen.de (root@tuminfo2.informatik.tu-muenchen.de [131.159.0.81]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id CAA17500 for ; Tue, 11 Apr 1995 02:28:55 -0700 Received: from hpsystem1.informatik.tu-muenchen.de ([131.159.0.176]) by tuminfo2.informatik.tu-muenchen.de with ESMTP id <26507-3>; Tue, 11 Apr 1995 11:27:22 +0200 Received: by hpsystem1.informatik.tu-muenchen.de id <233074>; Tue, 11 Apr 1995 11:27:11 +0100 To: freebsd-scsi@freefall.cdrom.com Path: news From: hafner@suncog2.forwiss.tu-muenchen.de (Walter 'madhouse' Hafner) Newsgroups: comp.periphs.scsi,comp.unix.bsd.freebsd.misc,muc.lists.freebsd.scsi,muc.lists.freebsd.questions Subject: Two hostadapters in one PC? Followup-To: comp.unix.bsd.freebsd.misc Date: 11 Apr 1995 09:25:12 GMT Organization: Technische Universitaet Muenchen, Germany Lines: 82 Distribution: world Message-ID: NNTP-Posting-Host: suncog2.forwiss.tu-muenchen.de Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk Hello! I've got a problem with a PC and two SCSI-hostadapters. My setup: - Mainboard: ASUS SP3 (486/DX2-66) (SP3, not 3G !!) - NCR 53C810 SCSI-controller on board (rom v. 2.00) - Adaptec 1542B in an ISA slot. - NEC 31200 (1 GB, HostID 0) - DEC 5200 (2 GB, HostID 1) - CD-ROM (HostID 6) - DAT-Tape (HostID 5) Currently I have DOS and FreeBSD on the NEC HD and BSD/OS 2.0 on the DEC. (BSD/OS is the only reason I still need the Adaptec ... And I can't dump BSD/OS because I still run Mail/News/UUCP/SLIP on it.) Up to now I simply disabled the NCR chip in the setup and connected all devices to the Adaptec. No problems. Last week I finally was fed up with BSD/OS 2.0: the way they handle their shared libs, the small but annoying bugs in the SW installations, the lack of support for the NCR ... and I decided to give FreeBSD a try. Since friday I have it up and running and want to keep it. Of course I wanted to use my on-board NCR SCSI controller for both DOS and FreeBSD. And the problems began: General: The external devices remained at the Adaptec all the time. When I didn't use the NCR, I disabled it in the CMOS-Setup. When I used the NCR I disabled the BIOS-check of the Adaptec via jumper. When I connected the drives to the Adaptec (NCR disabled), all went well (This was my setup for about a year now). When I connected the drives to the NCR (Adaptec BIOS disabled), all went well: I could reach both drives from DOS, but didn't check the external devices. I could reach and use all devices under FreeBSD, including the external ones.. BSD/OS didn't work because of the lack of NCR support. Then I connected the NEC drive (HostID 0) to the NCR and the DEC drive (HostID 1) to the Adaptec. When I booted DOS, the NEC at the NCR was found and the Adaptec wasn't tested. DOS fdisk only found the drive at the NCR and there was no way to reach the drive at the Adaptec. When I booted FreeBSD (from the boot-floppy), first the Adaptec was probed and the DEC (HostID 1, BSD/OS) was correctly identified (sd0). Aswell as the external devices. Then the NCR was probed and the NEC (HostID 0, DOS, FreeBSD) was correctly identified (sd1). Up to now my setup was NEC: sd0 and DEC sd1. So I switched my configuration again and compiled a new FreeBSD kernel with root and swap on sd1, switched back again and booted from floppy (My bootloader still pointed to sd0 ...). I stopped before the kernel was loaded and tried to load one manually: hd(1,a)/kernel ... Strange Error-messages and sudden death. Some one-liner about disk error (sorry, I don't have my machine at hand). hd(0,a)/kernel ... found the kernel ... strange enough on sd1 ... found all devices as outlined above and then died with "couldn't mount root device". (I compiled the kernel with root on sd1). A newly compiled kernel with sd0 as root device gave the same result. I never tried to boot BSD/OS: It was just too late yesterday. When I come home today I'll try it. I still can think of two or three things to try ... but on the whole I'm quite clueless what I do wrong. I would be very glad if someone out there could enlighten me. Thanks in advance. -Walter -- Walter Hafner_____________________________ hafner@forwiss.tu-muenchen.de FORWISS Muenchen, FG Kognitive Systeme, Raum O-134, Tel: 089/48095-220 So here I am, not being entertained! (Calvin) *CLICK* From owner-freebsd-scsi Tue Apr 11 05:07:11 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id FAA18806 for freebsd-scsi-outgoing; Tue, 11 Apr 1995 05:07:11 -0700 Received: from Root.COM (implode.Root.COM [198.145.90.1]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id FAA18800 for ; Tue, 11 Apr 1995 05:07:07 -0700 Received: from corbin.Root.COM (corbin.Root.COM [198.145.90.18]) by Root.COM (8.6.8/8.6.5) with ESMTP id FAA08885; Tue, 11 Apr 1995 05:06:44 -0700 Received: from localhost (localhost [127.0.0.1]) by corbin.Root.COM (8.6.11/8.6.5) with SMTP id FAA00394; Tue, 11 Apr 1995 05:06:44 -0700 Message-Id: <199504111206.FAA00394@corbin.Root.COM> X-Authentication-Warning: corbin.Root.COM: Host localhost didn't use HELO protocol To: hafner@suncog2.forwiss.tu-muenchen.de (Walter 'madhouse' Hafner) cc: freebsd-scsi@freefall.cdrom.com Subject: Re: Two hostadapters in one PC? In-reply-to: Your message of "11 Apr 95 09:25:12 GMT." From: David Greenman Reply-To: davidg@Root.COM Date: Tue, 11 Apr 1995 05:06:37 -0700 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk >I still can think of two or three things to try ... but on the whole I'm >quite clueless what I do wrong. I would be very glad if someone out >there could enlighten me. One thing you didn't mention was if you changed your /etc/fstab to specify sd1 be mounted on /. It sounds as though it is failing before it gets to this, but I think it is worth mentioning nonetheless. BTW, -questions would probably be a better forum. -DG From owner-freebsd-scsi Tue Apr 11 05:30:08 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id FAA18991 for freebsd-scsi-outgoing; Tue, 11 Apr 1995 05:30:08 -0700 Received: from hda.com (hda.com [199.232.40.182]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id FAA18985 for ; Tue, 11 Apr 1995 05:30:04 -0700 Received: (dufault@localhost) by hda.com (8.6.9/8.3) id IAA02925; Tue, 11 Apr 1995 08:28:31 -0400 From: Peter Dufault Message-Id: <199504111228.IAA02925@hda.com> Subject: Re: Two hostadapters in one PC? To: hafner@suncog2.forwiss.tu-muenchen.de (Walter 'madhouse' Hafner) Date: Tue, 11 Apr 1995 08:28:30 -0400 (EDT) Cc: freebsd-scsi@freefall.cdrom.com In-Reply-To: from "Walter 'madhouse' Hafner" at Apr 11, 95 09:25:12 am X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 680 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk Walter 'madhouse' Hafner writes: > > Hello! > > I've got a problem with a PC and two SCSI-hostadapters. Do you have the two host adapters on different SCSI IDs? Usually they come set to seven and you'll have to set them to either 2, 3 or 4 (since you have 0, 1, 5, 6 and 7 in use). Also, some of the host adapter code dies if the scsi bus is reset (the AIC code, for example) and those adapters won't play well with multiple adapters on a bus since the bus is reset at start up. It could be that. Peter -- Peter Dufault Real Time Machine Control and Simulation HD Associates, Inc. Voice: 508 433 6936 dufault@hda.com Fax: 508 433 5267 From owner-freebsd-scsi Tue Apr 11 09:59:48 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id JAA23146 for freebsd-scsi-outgoing; Tue, 11 Apr 1995 09:59:48 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id JAA23140 for ; Tue, 11 Apr 1995 09:59:44 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.8/8.6.6) id JAA03932; Tue, 11 Apr 1995 09:58:24 -0700 From: "Rodney W. Grimes" Message-Id: <199504111658.JAA03932@gndrsh.aac.dev.com> Subject: Re: Two hostadapters in one PC? To: hafner@suncog2.forwiss.tu-muenchen.de (Walter 'madhouse' Hafner) Date: Tue, 11 Apr 1995 09:58:24 -0700 (PDT) Cc: freebsd-scsi@freefall.cdrom.com In-Reply-To: from "Walter 'madhouse' Hafner" at Apr 11, 95 09:25:12 am X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 3190 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > > Hello! > > I've got a problem with a PC and two SCSI-hostadapters. > > My setup: > > - Mainboard: ASUS SP3 (486/DX2-66) (SP3, not 3G !!) > - NCR 53C810 SCSI-controller on board (rom v. 2.00) > - Adaptec 1542B in an ISA slot. > - NEC 31200 (1 GB, HostID 0) > - DEC 5200 (2 GB, HostID 1) > - CD-ROM (HostID 6) > - DAT-Tape (HostID 5) ... > > Then I connected the NEC drive (HostID 0) to the NCR and the DEC drive > (HostID 1) to the Adaptec. When you do this, you may want to set both disk drives to SCSI ID 0, and enable the adaptec BIOS. This should make both drives show up under DOS and FreeBSD. > When I booted DOS, the NEC at the NCR was found and the Adaptec wasn't > tested. DOS fdisk only found the drive at the NCR and there was no way > to reach the drive at the Adaptec. See above about setting the drive on the Adaptec to 0, it will only install a DOS compatible drive for a scsi ID of 1 if it also installed a DOS compatible drive for a scsi ID of 0. Since the adaptec did not see a drive on it's scsi bus at ID 0 it will not install any DOS compatible drive. > When I booted FreeBSD (from the boot-floppy), first the Adaptec was > probed and the DEC (HostID 1, BSD/OS) was correctly identified (sd0). > Aswell as the external devices. Then the NCR was probed and the NEC > (HostID 0, DOS, FreeBSD) was correctly identified (sd1). Okay, this tells me that the Adaptec BIOS is probed first by the system BIOS, this is the normal situation. > Up to now my setup was NEC: sd0 and DEC sd1. So I switched my > configuration again and compiled a new FreeBSD kernel with root and swap > on sd1, switched back again and booted from floppy (My bootloader > still pointed to sd0 ...). I stopped before the kernel was loaded and > tried to load one manually: > > hd(1,a)/kernel ... Strange Error-messages and sudden death. Some > one-liner about disk error (sorry, I don't have my machine at hand). Ahh the hd device name is a compatibility hack for when you have both IDE disk drives and SCSI disk drives. Since you do *NOT* have any IDE drives (or at leaset you did not tell us about any) you should be using sd(1,a)/kernel. > hd(0,a)/kernel ... found the kernel ... strange enough on sd1 ... found > all devices as outlined above and then died with "couldn't mount root > device". (I compiled the kernel with root on sd1). A newly compiled > kernel with sd0 as root device gave the same result. Thats because hd(0,a) says to use the device wd0 as root, you have no wd0 drive. > I never tried to boot BSD/OS: It was just too late yesterday. When I > come home today I'll try it. Fix the scsi ID's first and start using sd(0,a)/kernel and sd(1,a)/kernel and things should be much better for you. > I still can think of two or three things to try ... but on the whole I'm > quite clueless what I do wrong. I would be very glad if someone out > there could enlighten me. If you continue to have problems send me some more email, I'll see what I can figure out for you. > Thanks in advance. Your welcome. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Custom computers for FreeBSD From owner-freebsd-scsi Fri Apr 14 08:39:30 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id IAA21348 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 08:39:30 -0700 Received: from hda.com (hda.com [199.232.40.182]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id IAA21340 ; Fri, 14 Apr 1995 08:39:26 -0700 Received: (dufault@localhost) by hda.com (8.6.9/8.3) id LAA00581; Fri, 14 Apr 1995 11:38:55 -0400 From: Peter Dufault Message-Id: <199504141538.LAA00581@hda.com> Subject: SCSI target To: freebsd-hackers@FreeBSD.org Date: Fri, 14 Apr 1995 11:38:55 -0400 (EDT) Cc: freebsd-scsi@FreeBSD.org X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 1806 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk I've added a SCSI target device. Using this you can act as a target and another system acting as an initiator can send data to you over the SCSI bus. However, the only host adapter that supports this is the Adaptec 1542B. The 1542C will hang the bus big time if you put it into target mode, so I've locked it out. I have little hope that most of the commercial firmware will support this properly, so if some of our firmware crankers want to start cranking firmware they can look at the isa/aha1542c to get a feel for how it should act. I'll have a Acculogic board with the NCR 825 in house for a while. Basically there is an escape added to put the host adapter in target mode. Once it is in target mode it must respond to the mandatory commands: INQUIRY, REQUEST SENSE, SEND DIAGNOSTIC and TEST UNIT READY as well as the two data transfer commands SEND DATA and RECEIVE DATA. The 1542 model is good, and I suggest we just use that as much as possible. You can set up a data transfer in advance, with information that indicates which initiator it is for, and then when that initiator connects the board does the transfer. If an initiator connects to the 1542 when it doesn't have a transfer set up for that initiator, the 1542 will disconnect, generate a device interrupt, and report the ID and LUN of the initiator that connected as well as the amount of data the initiator wants to transfer. Eventually I'd like to add support for reserving units on the SCSI bus, sharing read only disk partitions, and running TCP/IP over SCSI using target mode capable host adapters. I think that could make for a nice little cluster. -- Peter Dufault Real Time Machine Control and Simulation HD Associates, Inc. Voice: 508 433 6936 dufault@hda.com Fax: 508 433 5267 From owner-freebsd-scsi Fri Apr 14 10:50:20 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id KAA26993 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 10:50:20 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id KAA26981 ; Fri, 14 Apr 1995 10:50:13 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id KAA01193; Fri, 14 Apr 1995 10:47:27 -0700 From: "Rodney W. Grimes" Message-Id: <199504141747.KAA01193@gndrsh.aac.dev.com> Subject: Re: SCSI target To: dufault@hda.com (Peter Dufault) Date: Fri, 14 Apr 1995 10:47:27 -0700 (PDT) Cc: freebsd-hackers@FreeBSD.org, freebsd-scsi@FreeBSD.org In-Reply-To: <199504141538.LAA00581@hda.com> from "Peter Dufault" at Apr 14, 95 11:38:55 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2236 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > > I've added a SCSI target device. Using this you can act as a target > and another system acting as an initiator can send data to you over > the SCSI bus. > > However, the only host adapter that supports this is the Adaptec > 1542B. The 1542C will hang the bus big time if you put it into > target mode, so I've locked it out. Great! But one of my 1542B's bit the dust about a month ago so now I only have one of them :-( :-(. > I have little hope that most of the commercial firmware will support > this properly, so if some of our firmware crankers want to start > cranking firmware they can look at the isa/aha1542c to get a feel > for how it should act. I'll have a Acculogic board with the NCR > 825 in house for a while. If you loose that card and still need one to work with I can arrange to put an NCR 53C810 controller in your house on a permanent loan basis. > Basically there is an escape added to put the host adapter in target > mode. Once it is in target mode it must respond to the mandatory > commands: INQUIRY, REQUEST SENSE, SEND DIAGNOSTIC and TEST UNIT > READY as well as the two data transfer commands SEND DATA and > RECEIVE DATA. > > The 1542 model is good, and I suggest we just use that as much as > possible. > > You can set up a data transfer in advance, with information that > indicates which initiator it is for, and then when that initiator > connects the board does the transfer. > > If an initiator connects to the 1542 when it doesn't have a transfer > set up for that initiator, the 1542 will disconnect, generate a > device interrupt, and report the ID and LUN of the initiator that > connected as well as the amount of data the initiator wants to > transfer. > > Eventually I'd like to add support for reserving units on the SCSI > bus, sharing read only disk partitions, and running TCP/IP over > SCSI using target mode capable host adapters. I think that could > make for a nice little cluster. Well, with 100MB/S ethernet support now being a reality TCP/IP over SCSI only has an advantage for *wide* scsi controllers. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Custom computers for FreeBSD From owner-freebsd-scsi Fri Apr 14 11:10:00 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id LAA27770 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 11:10:00 -0700 Received: from hda.com (hda.com [199.232.40.182]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id LAA27764 ; Fri, 14 Apr 1995 11:09:57 -0700 Received: (dufault@localhost) by hda.com (8.6.9/8.3) id OAA01326; Fri, 14 Apr 1995 14:09:15 -0400 From: Peter Dufault Message-Id: <199504141809.OAA01326@hda.com> Subject: Re: SCSI target To: rgrimes@gndrsh.aac.dev.com (Rodney W. Grimes) Date: Fri, 14 Apr 1995 14:09:14 -0400 (EDT) Cc: freebsd-hackers@FreeBSD.org, freebsd-scsi@FreeBSD.org In-Reply-To: <199504141747.KAA01193@gndrsh.aac.dev.com> from "Rodney W. Grimes" at Apr 14, 95 10:47:27 am X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 1223 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk Rodney W. Grimes writes: > > > > > I've added a SCSI target device. Using this you can act as a target > > and another system acting as an initiator can send data to you over > > the SCSI bus. > > > > However, the only host adapter that supports this is the Adaptec > > 1542B. The 1542C will hang the bus big time if you put it into > > target mode, so I've locked it out. > > Great! But one of my 1542B's bit the dust about a month ago so now > I only have one of them :-( :-(. I only have one also. (...) > If you loose that card and still need one to work with I can arrange > to put an NCR 53C810 controller in your house on a permanent loan > basis. I'm not planning on adding the support to the firmware - initially, I could probably use the 1542B more than the 810 for test. If anyone has a 1542B sitting in a closet (as I did until recently) feel free to send it. (...) > Well, with 100MB/S ethernet support now being a reality TCP/IP over > SCSI only has an advantage for *wide* scsi controllers. Yes, but it would still be cool. -- Peter Dufault Real Time Machine Control and Simulation HD Associates, Inc. Voice: 508 433 6936 dufault@hda.com Fax: 508 433 5267 From owner-freebsd-scsi Fri Apr 14 12:29:24 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id MAA02509 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 12:29:24 -0700 Received: from tfs.com (mailhub.tfs.com [140.145.250.1]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id MAA02502 ; Fri, 14 Apr 1995 12:29:23 -0700 Received: by tfs.com (smail3.1.28.1) Message-Id: From: julian@TFS.COM (Julian Elischer) Subject: Re: SCSI target To: dufault@hda.com (Peter Dufault) Date: Fri, 14 Apr 1995 12:28:51 -0700 (PDT) Cc: freebsd-hackers@FreeBSD.org, freebsd-scsi@FreeBSD.org In-Reply-To: <199504141538.LAA00581@hda.com> from "Peter Dufault" at Apr 14, 95 11:38:55 am X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 1269 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > > However, the only host adapter that supports this is the Adaptec > 1542B. The 1542C will hang the bus big time if you put it into > target mode, so I've locked it out. does the buslogic work? JORDAN? you've the busloginc manual...... > > I have little hope that most of the commercial firmware will support > this properly, so if some of our firmware crankers want to start > cranking firmware they can look at the isa/aha1542c to get a feel > for how it should act. I'll have a Acculogic board with the NCR > 825 in house for a while. I guess the NCR and 77x0 based boards would eb our best bet then.. > > The 1542 model is good, and I suggest we just use that as much as > possible. I believe the bt cards follow this model do you know if it works... (I believe you talked to them about it once) > > > Eventually I'd like to add support for reserving units on the SCSI > bus, sharing read only disk partitions, and running TCP/IP over > SCSI using target mode capable host adapters. I think that could > make for a nice little cluster. yeah.... with care, and an MTU > 4096, we should be able to shift whole pages around as well, (though that may take some work in the protocol stacks) (I was thinking of a special page-shifting protocol :) julian From owner-freebsd-scsi Fri Apr 14 12:40:46 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id MAA02858 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 12:40:46 -0700 Received: from mail.barrnet.net (mail.BARRNET.NET [131.119.246.7]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id MAA02852 for ; Fri, 14 Apr 1995 12:40:45 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by mail.barrnet.net (8.6.10/MAIL-RELAY-LEN) with ESMTP id LAA04452 for ; Fri, 14 Apr 1995 11:44:14 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id LAA01339; Fri, 14 Apr 1995 11:42:48 -0700 From: "Rodney W. Grimes" Message-Id: <199504141842.LAA01339@gndrsh.aac.dev.com> Subject: Re: SCSI target To: nate@trout.sri.MT.net (Nate Williams) Date: Fri, 14 Apr 1995 11:42:47 -0700 (PDT) Cc: freebsd-scsi@FreeBSD.org In-Reply-To: <199504141840.MAA29401@trout.sri.MT.net> from "Nate Williams" at Apr 14, 95 12:40:20 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1692 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk [cc: trimmed to scsi list] > > > > However, the only host adapter that supports this is the Adaptec > > > > 1542B. The 1542C will hang the bus big time if you put it into > > > > target mode, so I've locked it out. > > > > > > Great! But one of my 1542B's bit the dust about a month ago so now > > > I only have one of them :-( :-(. > > > > I only have one also. > > > > (...) > > > If you loose that card and still need one to work with I can arrange > > > to put an NCR 53C810 controller in your house on a permanent loan > > > basis. > > > > I'm not planning on adding the support to the firmware - initially, > > I could probably use the 1542B more than the 810 for test. If anyone > > has a 1542B sitting in a closet (as I did until recently) feel free to > > send it. > > If anyone has a controller which is better than a 1542B for a straight > ISA box (No VLB frills), I'd be willing to trade it for my 1542B. I'm > not real pleased with the performance out of the one I have, and would > certainly be willing to trade up to get better disk performance. (I'm > getting about 60% of what I could with my PD2100S) Your not going to get much more out of *any* ISA controller. The 1542CF will do a little better, but basically your problem is the speed of the ISA bus bus mastering (typical max is 5MB/sec *burst*, sustainted goes down to about 3.7-4MB/sec). You should be getting about 3MB/sec out of your 1542B and I have not seen anthing better on ISA. You'll need to go VLB or PCI to get the numbers up there. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Custom computers for FreeBSD From owner-freebsd-scsi Fri Apr 14 12:47:26 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id MAA02981 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 12:47:26 -0700 Received: from trout.sri.MT.net (trout.sri.MT.net [204.182.243.12]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id MAA02975 for ; Fri, 14 Apr 1995 12:47:22 -0700 Received: (from nate@localhost) by trout.sri.MT.net (8.6.11/8.6.11) id NAA29549; Fri, 14 Apr 1995 13:51:19 -0600 Date: Fri, 14 Apr 1995 13:51:19 -0600 From: Nate Williams Message-Id: <199504141951.NAA29549@trout.sri.MT.net> In-Reply-To: "Rodney W. Grimes" "Re: SCSI target" (Apr 14, 11:42am) X-Mailer: Mail User's Shell (7.2.5 10/14/92) To: "Rodney W. Grimes" Subject: Re: SCSI target Cc: freebsd-scsi@FreeBSD.org Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > Your not going to get much more out of *any* ISA controller. The 1542CF > will do a little better, but basically your problem is the speed of > the ISA bus bus mastering (typical max is 5MB/sec *burst*, sustainted > goes down to about 3.7-4MB/sec). You should be getting about 3MB/sec > out of your 1542B and I have not seen anthing better on ISA. I *should* be seeing 3MB/sec, but I've never got #'s even close. This is with 2.0R install created FS's. IOZONE: Performance Test of Sequential File I/O -- V1.16 (10/28/92) By Bill Norcott Operating System: POSIX 1003.1-1988 -- using fsync() Send comments to: norcott_bill@tandem.com IOZONE writes a 32 Megabyte sequential file consisting of 65536 records which are each 512 bytes in length. It then reads the file. It prints the bytes-per-second rate at which the computer can read and write files. Writing the 32 Megabyte file, 'iozone.tmp'...31.257812 seconds Reading the file...24.953125 seconds IOZONE performance measurements: 1073473 bytes/second for writing the file 1344698 bytes/second for reading the file Nate From owner-freebsd-scsi Fri Apr 14 13:26:57 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id NAA03825 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 13:26:57 -0700 Received: from netcom18.netcom.com (bakul@netcom18.netcom.com [192.100.81.131]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id NAA03817 for ; Fri, 14 Apr 1995 13:26:53 -0700 Received: from localhost by netcom18.netcom.com (8.6.12/Netcom) id NAA25300; Fri, 14 Apr 1995 13:26:32 -0700 Message-Id: <199504142026.NAA25300@netcom18.netcom.com> To: "Rodney W. Grimes" cc: freebsd-scsi@FreeBSD.org Subject: Re: SCSI target In-reply-to: Your message of "Fri, 14 Apr 95 11:42:47 PDT." <199504141842.LAA01339@gndrsh.aac.dev.com> Date: Fri, 14 Apr 95 13:26:26 -0700 From: Bakul Shah Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk On a somewhat related topic, does anyone know command processing overhead of modern SCSI devices? It used to be in the 50 ~ 100 us range a few years ago. From owner-freebsd-scsi Fri Apr 14 13:45:16 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id NAA04900 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 13:45:16 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id NAA04886 for ; Fri, 14 Apr 1995 13:45:11 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id NAA01599; Fri, 14 Apr 1995 13:42:17 -0700 From: "Rodney W. Grimes" Message-Id: <199504142042.NAA01599@gndrsh.aac.dev.com> Subject: Re: SCSI target To: bakul@netcom.com (Bakul Shah) Date: Fri, 14 Apr 1995 13:42:17 -0700 (PDT) Cc: freebsd-scsi@FreeBSD.org In-Reply-To: <199504142026.NAA25300@netcom18.netcom.com> from "Bakul Shah" at Apr 14, 95 01:26:26 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 509 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > > On a somewhat related topic, does anyone know command > processing overhead of modern SCSI devices? It used to be > in the 50 ~ 100 us range a few years ago. Not sure what it is, but most of my Quantum disk drives are using a 25Mhz 68020 and a Qlogic scsi bus controller chip, should be much much faster than the old stuff that was Z80 based :-) -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Custom computers for FreeBSD From owner-freebsd-scsi Fri Apr 14 13:56:25 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id NAA05589 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 13:56:25 -0700 Received: from husky.cs.vt.edu (jaitken@cslab.cs.vt.edu [128.173.41.87]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id NAA05579 for ; Fri, 14 Apr 1995 13:56:22 -0700 Received: (jaitken@localhost) by husky.cs.vt.edu (8.6.10/8.6.4) id QAA31915; Fri, 14 Apr 1995 16:54:53 -0400 From: Jeff Aitken Message-Id: <199504142054.QAA31915@husky.cs.vt.edu> Subject: Re: SCSI target To: nate@trout.sri.MT.net (Nate Williams) Date: Fri, 14 Apr 1995 16:54:52 -0400 (EDT) Cc: rgrimes@gndrsh.aac.dev.com, freebsd-scsi@FreeBSD.org In-Reply-To: <199504141951.NAA29549@trout.sri.MT.net> from "Nate Williams" at Apr 14, 95 01:51:19 pm X-Mailer: ELM [version 2.4 PL23] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2422 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > > > Your not going to get much more out of *any* ISA controller. The 1542CF > > will do a little better, but basically your problem is the speed of > > the ISA bus bus mastering (typical max is 5MB/sec *burst*, sustainted > > goes down to about 3.7-4MB/sec). You should be getting about 3MB/sec > > out of your 1542B and I have not seen anthing better on ISA. > > I *should* be seeing 3MB/sec, but I've never got #'s even close. > > This is with 2.0R install created FS's. > > Writing the 32 Megabyte file, 'iozone.tmp'...31.257812 seconds > Reading the file...24.953125 seconds > > IOZONE performance measurements: > 1073473 bytes/second for writing the file > 1344698 bytes/second for reading the file Maybe I'm just not understanding something here, but shouldn't I be getting better than 3MB/s with both the NCR Fast-SCSI2 controller + a Micropolis Fast-SCSI2 drive? (486DX2/66, 16MB, PCI/ISA, 256K cache) % iozone 32 IOZONE: Performance Test of Sequential File I/O -- V1.16 (10/28/92) By Bill Norcott Operating System: POSIX 1003.1-1988 -- using fsync() Send comments to: norcott_bill@tandem.com IOZONE writes a 32 Megabyte sequential file consisting of 65536 records which are each 512 bytes in length. It then reads the file. It prints the bytes-per-second rate at which the computer can read and write files. Writing the 32 Megabyte file, 'iozone.tmp'...16.140625 seconds Reading the file...10.984375 seconds IOZONE performance measurements: 2078880 bytes/second for writing the file 3054742 bytes/second for reading the file % iozone 128 IOZONE: Performance Test of Sequential File I/O -- V1.16 (10/28/92) By Bill Norcott Operating System: POSIX 1003.1-1988 -- using fsync() Send comments to: norcott_bill@tandem.com IOZONE writes a 128 Megabyte sequential file consisting of 262144 records which are each 512 bytes in length. It then reads the file. It prints the bytes-per-second rate at which the computer can read and write files. Writing the 128 Megabyte file, 'iozone.tmp'...59.031250 seconds Reading the file...42.664062 seconds IOZONE performance measurements: 2273672 bytes/second for writing the file 3145920 bytes/second for reading the file -- Jeff Aitken jaitken@vt.edu From owner-freebsd-scsi Fri Apr 14 14:03:46 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id OAA06104 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 14:03:46 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id OAA06088 for ; Fri, 14 Apr 1995 14:03:41 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id NAA01671; Fri, 14 Apr 1995 13:59:43 -0700 From: "Rodney W. Grimes" Message-Id: <199504142059.NAA01671@gndrsh.aac.dev.com> Subject: Re: SCSI target To: jaitken@cslab.cs.vt.edu (Jeff Aitken) Date: Fri, 14 Apr 1995 13:59:42 -0700 (PDT) Cc: nate@trout.sri.MT.net, freebsd-scsi@FreeBSD.org In-Reply-To: <199504142054.QAA31915@husky.cs.vt.edu> from "Jeff Aitken" at Apr 14, 95 04:54:52 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1610 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > > > > > > Your not going to get much more out of *any* ISA controller. The 1542CF > > > will do a little better, but basically your problem is the speed of > > > the ISA bus bus mastering (typical max is 5MB/sec *burst*, sustainted > > > goes down to about 3.7-4MB/sec). You should be getting about 3MB/sec > > > out of your 1542B and I have not seen anthing better on ISA. > > > > I *should* be seeing 3MB/sec, but I've never got #'s even close. > > > > This is with 2.0R install created FS's. > > > > Writing the 32 Megabyte file, 'iozone.tmp'...31.257812 seconds > > Reading the file...24.953125 seconds > > > > IOZONE performance measurements: > > 1073473 bytes/second for writing the file > > 1344698 bytes/second for reading the file > > > Maybe I'm just not understanding something here, but shouldn't I be > getting better than 3MB/s with both the NCR Fast-SCSI2 controller + a > Micropolis Fast-SCSI2 drive? (486DX2/66, 16MB, PCI/ISA, 256K cache) Depends on the model of the drive, I need to know how fast it spends and how many heads and cylinders there are. If you have the sectors/zone info that helps too. Basically max sustained transfer rate is: MB/sec == (#sectors per track) * (RPM / 60) * 512 For Zone Bit Recorded drives (any modern scsi drive is ZBR) you will get higher transfer rates on the lower numbered cylinders. > % iozone 32 Can't say much with the model drive it was on, and the parameters of said drive. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Custom computers for FreeBSD From owner-freebsd-scsi Fri Apr 14 14:40:52 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id OAA08298 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 14:40:52 -0700 Received: from kitten.mcs.com (Kitten.mcs.com [192.160.127.90]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id OAA08290 for ; Fri, 14 Apr 1995 14:40:50 -0700 Received: from mailbox.mcs.com (Mailbox.mcs.com [192.160.127.87]) by kitten.mcs.com (8.6.10/8.6.9) with SMTP id QAA00297; Fri, 14 Apr 1995 16:39:16 -0500 Received: by mailbox.mcs.com (/\==/\ Smail3.1.28.1 #28.5) id ; Fri, 14 Apr 95 16:39 CDT Received: by mercury.mcs.com (/\==/\ Smail3.1.28.1 #28.5) id ; Fri, 14 Apr 95 16:38 CDT Message-Id: Subject: Re: SCSI target To: nate@trout.sri.MT.net (Nate Williams) Date: Fri, 14 Apr 1995 16:38:35 -0500 (CDT) From: "Lars Fredriksen" Cc: rgrimes@gndrsh.aac.dev.com, freebsd-scsi@FreeBSD.org In-Reply-To: <199504141951.NAA29549@trout.sri.MT.net> from "Nate Williams" at Apr 14, 95 01:51:19 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2079 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk Nate Williams writes: > > > Your not going to get much more out of *any* ISA controller. The 1542CF > > will do a little better, but basically your problem is the speed of > > the ISA bus bus mastering (typical max is 5MB/sec *burst*, sustainted > > goes down to about 3.7-4MB/sec). You should be getting about 3MB/sec > > out of your 1542B and I have not seen anthing better on ISA. > > I *should* be seeing 3MB/sec, but I've never got #'s even close. > > This is with 2.0R install created FS's. > > > IOZONE: Performance Test of Sequential File I/O -- V1.16 (10/28/92) > By Bill Norcott > > Operating System: POSIX 1003.1-1988 -- using fsync() > > Send comments to: norcott_bill@tandem.com > > IOZONE writes a 32 Megabyte sequential file consisting of > 65536 records which are each 512 bytes in length. > It then reads the file. It prints the bytes-per-second > rate at which the computer can read and write files. > > > Writing the 32 Megabyte file, 'iozone.tmp'...31.257812 seconds > Reading the file...24.953125 seconds > > IOZONE performance measurements: > 1073473 bytes/second for writing the file > 1344698 bytes/second for reading the file > And you are complaining! :-) On my Mylex 486/66 EISA MB with a 1742 I get 400K/600K W/R. lmbench indicates that my memory subsystem is only capable of 10Mb/s and that the cache is write through. Mylex swears that it is write back, and have no explanation as to the slowness. They said something about EISA memory subsystem being more complex than an ISA one, and therefore slower. John Dyson is seeing 5-6 times the memory bandwidth on an AIR and Micronics ISA MBs, and needless to say he is getting a lot better disk performance too. There might be a snag in that I only have 16 MB of ram, which is all in one bank. If the MB is smart enough to do interleaving this could cut my performance in ~2. Lars -- ------------------------------------------------------------------- Lars Fredriksen fredriks@mcs.com (home) lars@fredriks.pr.mcs.net (home-home) fredriks@asiago.cs.wisc.edu From owner-freebsd-scsi Fri Apr 14 15:25:35 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id PAA10709 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 15:25:35 -0700 Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id PAA10677 for ; Fri, 14 Apr 1995 15:24:46 -0700 Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id IAA23515; Sat, 15 Apr 1995 08:16:59 +1000 Date: Sat, 15 Apr 1995 08:16:59 +1000 From: Bruce Evans Message-Id: <199504142216.IAA23515@godzilla.zeta.org.au> To: bakul@netcom.com, rgrimes@gndrsh.aac.dev.com Subject: Re: SCSI target Cc: freebsd-scsi@FreeBSD.org Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk >> On a somewhat related topic, does anyone know command >> processing overhead of modern SCSI devices? It used to be >> in the 50 ~ 100 us range a few years ago. >Not sure what it is, but most of my Quantum disk drives are using >a 25Mhz 68020 and a Qlogic scsi bus controller chip, should be much >much faster than the old stuff that was Z80 based :-) It seems to be much larger than 50 ~ 100 us :-(. According to the following test program, it is about 3500 us for a Buslogic BT445C controller / Toshiba MK537FB drive and about 2800 us for a BT445C controller / Quantum XP34301 drive. Bruce --- #include #include #include #include #include #define ITERATIONS 1000 static int syserror(const char *where); static long timeit(int fd, char *buf, unsigned blocksize); int main(int argc, char **argv) { char buf[2 * 4096]; int fd; long time_4096; long time_8192; if (argc != 2) { fprintf(stderr, "usage: %s device\n", argv[0]); exit(1); } fd = open(argv[1], O_RDONLY); if (fd == -1) syserror("open"); time_4096 = timeit(fd, buf, 4096); time_8192 = timeit(fd, buf, 8192); printf("Command overhead is %ld usec (time_4096 = %ld, time_8192 = %ld)\n", (time_4096 - (time_8192 - time_4096)) / ITERATIONS, time_4096 / ITERATIONS, time_8192 / ITERATIONS); exit(0); } static int syserror(const char *where) { perror(where); exit(1); } static long timeit(int fd, char *buf, unsigned blocksize) { struct timeval finish; int i; struct timeval start; if (read(fd, buf, blocksize) != blocksize) syserror("read"); if (gettimeofday(&start, (struct timezone *)NULL) != 0) syserror("gettimeofday(start)"); for (i = 0; i < ITERATIONS; ++i) { if (lseek(fd, (off_t)0, SEEK_SET) == -1) syserror("lseek"); if (read(fd, buf, blocksize) != blocksize) syserror("read"); } if (gettimeofday(&finish, (struct timezone *)NULL) != 0) syserror("gettimeofday(finish)"); return (finish.tv_sec - start.tv_sec) * 1000000 + finish.tv_usec - start.tv_usec; } --- From owner-freebsd-scsi Fri Apr 14 16:30:20 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id QAA14007 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 16:30:20 -0700 Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id QAA13921 for ; Fri, 14 Apr 1995 16:29:47 -0700 Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id JAA26487; Sat, 15 Apr 1995 09:23:22 +1000 Date: Sat, 15 Apr 1995 09:23:22 +1000 From: Bruce Evans Message-Id: <199504142323.JAA26487@godzilla.zeta.org.au> To: bakul@netcom.com, bde@zeta.org.au, rgrimes@gndrsh.aac.dev.com Subject: Re: SCSI target Cc: freebsd-scsi@FreeBSD.org Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk >>> On a somewhat related topic, does anyone know command >>> processing overhead of modern SCSI devices? It used to be >>> in the 50 ~ 100 us range a few years ago. >>Not sure what it is, but most of my Quantum disk drives are using >>a 25Mhz 68020 and a Qlogic scsi bus controller chip, should be much >>much faster than the old stuff that was Z80 based :-) >It seems to be much larger than 50 ~ 100 us :-(. According to the >following test program, it is about 3500 us for a Buslogic BT445C >controller / Toshiba MK537FB drive and about 2800 us for a BT445C >controller / Quantum XP34301 drive. PS: I modified the program to print the transfer rate (it was as expected) and tested an IDE drive (the overhead was much better, as expected :-), but still much larger than 50 ~ 100 usec. Maybe the SCSI overhead used to be 50 ~ 100 ms? :-). Summary: Controller Drive Cmd Overhead Transfer Rate BT445C MK537FB 3500 usec 10-MB/sec BT445C XP34301 2800 usec 10+MB/sec No-name ISA-IDE SHD-1312A 654 usec 2.0MB/sec The transfer rate for the SCSI setups are known to be 10MB/sec so the differences in the measured rate give an idea of the inaccuracy of the measurements. The high command overheads destroy the performance of file systems like ufs that do a lot of small i/o's to access metadata. Bruce From owner-freebsd-scsi Fri Apr 14 16:51:21 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id QAA16475 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 16:51:21 -0700 Received: from hda.com (hda.com [199.232.40.182]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id QAA16459 for ; Fri, 14 Apr 1995 16:51:17 -0700 Received: (dufault@localhost) by hda.com (8.6.9/8.3) id TAA03880; Fri, 14 Apr 1995 19:48:40 -0400 From: Peter Dufault Message-Id: <199504142348.TAA03880@hda.com> Subject: SCSI command overhead To: bde@zeta.org.au (Bruce Evans) Date: Fri, 14 Apr 1995 19:48:39 -0400 (EDT) Cc: bakul@netcom.com, rgrimes@gndrsh.aac.dev.com, freebsd-scsi@FreeBSD.org In-Reply-To: <199504142216.IAA23515@godzilla.zeta.org.au> from "Bruce Evans" at Apr 15, 95 08:16:59 am X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 1784 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk Bruce Evans writes: > > >> On a somewhat related topic, does anyone know command > >> processing overhead of modern SCSI devices? It used to be > >> in the 50 ~ 100 us range a few years ago. > > >Not sure what it is, but most of my Quantum disk drives are using > >a 25Mhz 68020 and a Qlogic scsi bus controller chip, should be much > >much faster than the old stuff that was Z80 based :-) > > It seems to be much larger than 50 ~ 100 us :-(. According to the > following test program, it is about 3500 us for a Buslogic BT445C > controller / Toshiba MK537FB drive and about 2800 us for a BT445C > controller / Quantum XP34301 drive. I'd like to work out a careful breakdown of where the overhead is. Bruce is in the right ballpark. Using libscsi I just put together a loop of 1000 "test unit readies" and sent it to the three disks I have here when the system was idle (verified by seeing that the disks were always ready). Then I stepped all over the CDB so that the driver would reject it to get a feel for the overhead of the SCSI code down to the point it was about to submit the CDB to the host adapter. And it is awful (This is with a 1542C): hda# ./tur /dev/rsd0c CDC 94171-9: 2535 microseconds. 92 microseconds. hda# ./tur /dev/rsd1c FUJITSU M2266S-5: 2324 microseconds. 85 microseconds. hda# ./tur /dev/rsd2c FUJITSU M2654S-5: 2280 microseconds. 87 microseconds. So although we're seeing almost a 300 microsecond improvement between the oldest and the newest device, but the overhead is still more than 2 milliseconds per transaction. -- Peter Dufault Real Time Machine Control and Simulation HD Associates, Inc. Voice: 508 433 6936 dufault@hda.com Fax: 508 433 5267 From owner-freebsd-scsi Fri Apr 14 18:27:06 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id SAA21052 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 18:27:06 -0700 Received: from netcom10.netcom.com (root@netcom10.netcom.com [192.100.81.120]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id SAA21044 for ; Fri, 14 Apr 1995 18:27:04 -0700 Received: from localhost by netcom10.netcom.com (8.6.12/Netcom) id SAA21458; Fri, 14 Apr 1995 18:20:49 -0700 Message-Id: <199504150120.SAA21458@netcom10.netcom.com> To: Peter Dufault , bde@zeta.org.au (Bruce Evans) To: rgrimes@gndrsh.aac.dev.com, freebsd-scsi@FreeBSD.org Subject: Re: SCSI command overhead In-reply-to: Your message of "Fri, 14 Apr 95 19:48:39 EDT." <199504142348.TAA03880@hda.com> Date: Fri, 14 Apr 95 18:20:44 -0700 From: Bakul Shah Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk Folks, thanks for the data. The 50 ~ 100 us I saw was entirely the SCSI chip overhead (WD33C93A), measured with a logic analyzer and confirmed by looking at cycle counts data wormed out of Western Digital. No disk or memory access was involved in this measurement. This was for an embedded system where SCSI was to be used for some `real time' communication (don't ask why). I was hoping that some of the new chips used simpler state machines and got out of the way of data transfer ASAP. Part of the problem *is* those more complicated state transitions in the target mode. Add to that time spent by the drive processor to map a block number to hd/sec/trk, look up any block forwarding, deal with variable number of sectors/trk etc. From owner-freebsd-scsi Fri Apr 14 19:15:57 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id TAA22227 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 19:15:57 -0700 Received: from tfs.com (mailhub.tfs.com [140.145.250.1]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id TAA22217 ; Fri, 14 Apr 1995 19:15:43 -0700 Received: by tfs.com (smail3.1.28.1) Message-Id: From: julian@TFS.COM (Julian Elischer) Subject: Re: SCSI target To: rgrimes@gndrsh.aac.dev.com (Rodney W. Grimes) Date: Fri, 14 Apr 1995 19:15:02 -0700 (PDT) Cc: dufault@hda.com, freebsd-hackers@FreeBSD.org, freebsd-scsi@FreeBSD.org In-Reply-To: <199504141747.KAA01193@gndrsh.aac.dev.com> from "Rodney W. Grimes" at Apr 14, 95 10:47:27 am X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 445 Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > > > > Well, with 100MB/S ethernet support now being a reality TCP/IP over > SCSI only has an advantage for *wide* scsi controllers. weeeellll, no, there are advandages in being able to transfer 128KB of scatter-gather data in one hit with NO cpu intervention..... :) > > > > -- > Rod Grimes rgrimes@gndrsh.aac.dev.com > Accurate Automation Company Custom computers for FreeBSD > From owner-freebsd-scsi Fri Apr 14 20:46:42 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id UAA25252 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 20:46:42 -0700 Received: from estienne.cs.berkeley.edu (estienne.CS.Berkeley.EDU [128.32.42.147]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id UAA25238 ; Fri, 14 Apr 1995 20:46:35 -0700 Received: from localhost (localhost [127.0.0.1]) by estienne.cs.berkeley.edu (8.6.11/8.6.9) with SMTP id UAA01018; Fri, 14 Apr 1995 20:46:19 -0700 Message-Id: <199504150346.UAA01018@estienne.cs.berkeley.edu> X-Authentication-Warning: estienne.cs.berkeley.edu: Host localhost didn't use HELO protocol To: julian@TFS.COM (Julian Elischer) cc: rgrimes@gndrsh.aac.dev.com (Rodney W. Grimes), dufault@hda.com, freebsd-hackers@FreeBSD.org, freebsd-scsi@FreeBSD.org Subject: Re: SCSI target In-reply-to: Your message of "Fri, 14 Apr 1995 19:15:02 PDT." Date: Fri, 14 Apr 1995 20:46:19 -0700 From: "Justin T. Gibbs" Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk >> >> > >> Well, with 100MB/S ethernet support now being a reality TCP/IP over >> SCSI only has an advantage for *wide* scsi controllers. >weeeellll, no, there are advandages in being able to transfer >128KB of scatter-gather data in one hit with NO >cpu intervention..... :) The max is one megabyte (assuming one page per SG) on any of the aic7xxx cards. >> -- >> Rod Grimes rgrimes@gndrsh.aac.dev.com >> Accurate Automation Company Custom computers for FreeBSD -- Justin T. Gibbs ============================================== TCS Instructional Group - Programmer/Analyst 1 Cory | Po | Danube | Volga | Parker | Torus ============================================== From owner-freebsd-scsi Fri Apr 14 21:52:07 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id VAA26685 for freebsd-scsi-outgoing; Fri, 14 Apr 1995 21:52:07 -0700 Received: from localhost (localhost [127.0.0.1]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id VAA26592 ; Fri, 14 Apr 1995 21:47:40 -0700 X-Authentication-Warning: freefall.cdrom.com: Host localhost didn't use HELO protocol To: julian@tfs.com (Julian Elischer) cc: dufault@hda.com (Peter Dufault), freebsd-hackers@FreeBSD.org, freebsd-scsi@FreeBSD.org Subject: Re: SCSI target In-reply-to: Your message of "Fri, 14 Apr 95 12:28:51 PDT." Date: Fri, 14 Apr 1995 21:47:40 -0700 Message-ID: <26591.797921260@freefall.cdrom.com> From: "Jordan K. Hubbard" Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > does the buslogic work? > JORDAN? you've the busloginc manual...... Only for the new WIDE controller series.. Jordan From owner-freebsd-scsi Sat Apr 15 08:22:19 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id IAA15096 for freebsd-scsi-outgoing; Sat, 15 Apr 1995 08:22:19 -0700 Received: from mail.eskimo.com (root@mail.eskimo.com [204.122.16.4]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id IAA15081 ; Sat, 15 Apr 1995 08:21:58 -0700 Received: from eskimo.com by mail.eskimo.com (5.65c/1.35) id AA25349; Sat, 15 Apr 1995 08:21:54 -0700 Date: Sat, 15 Apr 1995 08:21:52 -0700 (PDT) From: Dave Haas To: freebsd-scsi@FreeBSD.org Cc: freebsd-hardware@FreeBSD.org Subject: QLogic Fast!SCSI VESA Local host adapter Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk Hi there. I've been running FreeBSD snapshots in their 2/10 and 3/22 incarnations. I have yet to hear if there is any support planned, in the works, or otherwise on the drawing board to support the QLogic SCSI controller card. A search of the mail archives didn't turn up any definitive answers to this question. And the closest thing to hardware support that I've found has been a Linux port. I thought I would give my queries a try here, as QLogic claimed by phone to have no plans to support any BSD in the near future. Anybody know? Dave Haas daveh@eskimo.com From owner-freebsd-scsi Sat Apr 15 14:40:43 1995 Return-Path: freebsd-scsi-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id OAA02664 for freebsd-scsi-outgoing; Sat, 15 Apr 1995 14:40:43 -0700 Received: from localhost (localhost [127.0.0.1]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id OAA02657 ; Sat, 15 Apr 1995 14:40:42 -0700 X-Authentication-Warning: freefall.cdrom.com: Host localhost didn't use HELO protocol To: Dave Haas cc: freebsd-scsi@FreeBSD.org, freebsd-hardware@FreeBSD.org Subject: Re: QLogic Fast!SCSI VESA Local host adapter In-reply-to: Your message of "Sat, 15 Apr 95 08:21:52 PDT." Date: Sat, 15 Apr 1995 14:40:41 -0700 Message-ID: <2656.797982041@freefall.cdrom.com> From: "Jordan K. Hubbard" Sender: freebsd-scsi-owner@FreeBSD.org Precedence: bulk > Hi there. I've been running FreeBSD snapshots in their 2/10 and 3/22 > incarnations. I have yet to hear if there is any support planned, in the > works, or otherwise on the drawing board to support the QLogic SCSI controlle r AFAIK, nothing is planned. None of us have QLOGIC cards and since no one with the requisite skill and a controller has stepped forward, it's not happened.. Jordan