From owner-freebsd-scsi Sun Jun 17 14:24:38 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.dis.org (mass.dis.org [216.240.45.41]) by hub.freebsd.org (Postfix) with ESMTP id 0941D37B407 for ; Sun, 17 Jun 2001 14:24:25 -0700 (PDT) (envelope-from msmith@mass.dis.org) Received: from mass.dis.org (localhost [127.0.0.1]) by mass.dis.org (8.11.3/8.11.3) with ESMTP id f5HLZXp01329 for ; Sun, 17 Jun 2001 14:35:33 -0700 (PDT) (envelope-from msmith@mass.dis.org) Message-Id: <200106172135.f5HLZXp01329@mass.dis.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: scsi@freebsd.org Subject: 'mly' driver update/bugfix snapshot available Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 17 Jun 2001 14:35:33 -0700 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org There have been a number of problems with the 'mly' driver reported recently, mostly related to the default-on SCSI passthrough and conflicts with enclosure management hardware. There have also been some other subtle problems which have not been reported, but shown up in my own testing. In addition, work on other drivers has brought to light some shortcomings in the design and implementation of the original 'mly' driver. In order to remedy all of these problems, I have reworked the 'mly' driver substantially, and I'm now looking for some testers. The new code has only been tested on 4.3-STABLE, but should work on most or all 4.x systems (and probably on 5.x as well). You can fetch it at: http://people.freebsd.org/~msmith/RAID/mylex/mly-20010617.tar.gz If you do try it out, please let me know how you go. Regards, Mike -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Jun 17 17: 0: 6 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mailout02.sul.t-online.de (mailout02.sul.t-online.com [194.25.134.17]) by hub.freebsd.org (Postfix) with ESMTP id 724A337B408 for ; Sun, 17 Jun 2001 17:00:04 -0700 (PDT) (envelope-from haribeau@gmx.de) Received: from fwd00.sul.t-online.de by mailout02.sul.t-online.de with smtp id 15BmSZ-0004Xv-00; Mon, 18 Jun 2001 02:00:03 +0200 Received: from asterix.local (320080844193-0001@[217.80.84.1]) by fmrl00.sul.t-online.com with smtp id 15BmSP-1F5k6iC; Mon, 18 Jun 2001 01:59:53 +0200 Received: (qmail 2253 invoked from network); 17 Jun 2001 23:59:52 -0000 Received: from homer.local (HELO homer.local.nlocal) (192.168.1.50) by 0 with SMTP; 17 Jun 2001 23:59:52 -0000 Received: (nullmailer pid 447 invoked by uid 1100); Sun, 17 Jun 2001 23:59:52 -0000 Date: Mon, 18 Jun 2001 01:59:52 +0200 From: Clemens Hermann To: Mike Smith Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available Message-ID: <20010618015951.A405@homer.local> References: <200106172135.f5HLZXp01329@mass.dis.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200106172135.f5HLZXp01329@mass.dis.org> von Mike Smith am 17.Jun.2001 um 14:35:33 (-0700) X-Mailer: Mutt 1.2.5i (FreeBSD 4.3-RELEASE i386) X-Sender: 320080844193-0001@t-dialin.net Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Hi, > http://people.freebsd.org/~msmith/RAID/mylex/mly-20010617.tar.gz > > If you do try it out, please let me know how you go. sure :). I installed the new driver as described in the README and disabled the old one. The boot went through perfect (no error messages concerning the enclosure any more). Then I removed one drive and the spare was rebuilt. Later I re-inserted the first drive and removed the spare. The rebuild also went through. I got many messages during the rebuilt, sorry for not being able to reproduce them exactly but they did not appear in syslog. There were several concerning IO-Problems and things like that, I could no longer access the machine. there seems to be big trouble when accessing the array after the rebuild. Finally I rebooted the machine and the syncing before the halt failed. While rebooting the mylex-bios did not report any problems, the two system drives and the spare disk were all online. Freebsd made an fsck (yes, I booted a different kernel) but did not report problems. It might be possible that the error-messages were the result from the driver trying to write to syslog - I have no Idea why else all the messages that appered on the screen were not in syslog. /ch To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Sun Jun 17 17:25:41 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.dis.org (mass.dis.org [216.240.45.41]) by hub.freebsd.org (Postfix) with ESMTP id DEBD337B408 for ; Sun, 17 Jun 2001 17:25:27 -0700 (PDT) (envelope-from msmith@mass.dis.org) Received: from mass.dis.org (localhost [127.0.0.1]) by mass.dis.org (8.11.3/8.11.3) with ESMTP id f5I0aXp03022; Sun, 17 Jun 2001 17:36:33 -0700 (PDT) (envelope-from msmith@mass.dis.org) Message-Id: <200106180036.f5I0aXp03022@mass.dis.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: Clemens Hermann Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available In-reply-to: Your message of "Mon, 18 Jun 2001 01:59:52 +0200." <20010618015951.A405@homer.local> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 17 Jun 2001 17:36:33 -0700 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > Hi, > > > http://people.freebsd.org/~msmith/RAID/mylex/mly-20010617.tar.gz > > > > If you do try it out, please let me know how you go. > > sure :). I installed the new driver as described in the README and > disabled the old one. The boot went through perfect (no error messages > concerning the enclosure any more). That's a good start. > Then I removed one drive and the spare was rebuilt. Later I re-inserted > the first drive and removed the spare. The rebuild also went through. You say that it "went through", but it sounds like there were problems before it completed. What really happened? > I got many messages during the rebuilt, sorry for not being able to > reproduce them exactly but they did not appear in syslog. There were > several concerning IO-Problems and things like that, I could no longer > access the machine. I really need these messages to be able to help, here. 8( > there seems to be big trouble when accessing the array after the > rebuild. > Finally I rebooted the machine and the syncing before the halt failed. > While rebooting the mylex-bios did not report any problems, the two > system drives and the spare disk were all online. Freebsd made an fsck (yes, > I booted a different kernel) but did not report problems. I would have liked to know what happened if you rebooted with the modified driver at this point. > It might be possible that the error-messages were the result from the > driver trying to write to syslog - I have no Idea why else all the > messages that appered on the screen were not in syslog. Unlikely; the driver just uses the normal kernel printf mechanisms. If /var was on the array, it's possible that writes were failing and thus nothing ended up in the log. A transcript from a serial console would be helpful here. -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Jun 18 4: 9:12 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mailout00.sul.t-online.de (mailout00.sul.t-online.com [194.25.134.16]) by hub.freebsd.org (Postfix) with ESMTP id 5960D37B405 for ; Mon, 18 Jun 2001 04:09:00 -0700 (PDT) (envelope-from haribeau@gmx.de) Received: from fwd01.sul.t-online.de by mailout00.sul.t-online.de with smtp id 15Bwtv-0005fN-00; Mon, 18 Jun 2001 13:08:59 +0200 Received: from asterix.local (320080844193-0001@[62.225.210.223]) by fmrl01.sul.t-online.com with smtp id 15Bwtp-1I3hpIC; Mon, 18 Jun 2001 13:08:53 +0200 Received: (qmail 504 invoked from network); 18 Jun 2001 11:08:51 -0000 Received: from homer.local (HELO homer.local.nlocal) (192.168.1.50) by 0 with SMTP; 18 Jun 2001 11:08:51 -0000 Received: (nullmailer pid 1666 invoked by uid 1100); Mon, 18 Jun 2001 11:08:51 -0000 Date: Mon, 18 Jun 2001 13:08:51 +0200 From: Clemens Hermann To: Mike Smith Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available Message-ID: <20010618130851.C1070@homer.local> References: <20010618015951.A405@homer.local> <200106180036.f5I0aXp03022@mass.dis.org> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="82I3+IH0IqGh5yIs" Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200106180036.f5I0aXp03022@mass.dis.org> von Mike Smith am 17.Jun.2001 um 17:36:33 (-0700) X-Mailer: Mutt 1.2.5i (FreeBSD 4.3-RELEASE i386) Organization: Linuxlupe InternetSolutions X-Sender: 320080844193-0001@t-dialin.net Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org --82I3+IH0IqGh5yIs Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Hi Mike, > You say that it "went through", but it sounds like there were problems > before it completed. What really happened? The test setup was as follows: Drive0 and Drive1: 9 GB, System Mirror Drive2 and Drive3: 36 GB, UserSpace Mirror Drive4: 36 GB: Spare. I tested as follows: - Boot with every drive online - removed drive0 - after the rebuild went through I re-inserted the drive - removed drive4 - after the rebuild went through I re-inserted the drive Each time when I removed a drive the spare drive was synchronized (according to the lights on the backplane and the messages to syslog). Afther the above procedure I could no longer access the system. I could not read files or write files, the according error-messages are attached. The system was yet unusable after the first synchronisation. even a "halt" was not possible and after ctrl-alt-del the synchronization of the array failed. During reboot the mylex-bios did not mention problems so all the drives must have been set online. > I really need these messages to be able to help, here. 8( O.k.. I just had to get another (IDE) drive because the problem really was that syslog could no longer write to /var/log/messages because it was on the array. I attached the messages from the boot on till the shutdown. In the middle the process I described above took place and produced quite ugly messages :). The Kernel I used is the 4.3-Release Generic with the only difference that the mly-driver has been deactivated and your module has been used. > > I booted a different kernel) but did not report problems. > > I would have liked to know what happened if you rebooted with the > modified driver at this point. The system boots, checks the drives and everything seems to work fine as long as I do not remove a drive. > > It might be possible that the error-messages were the result from the > > driver trying to write to syslog > > Unlikely; the driver just uses the normal kernel printf mechanisms. If > /var was on the array, it was > it's possible that writes were failing and thus > nothing ended up in the log. A transcript from a serial console would be > helpful here. I used another drive but anyway I would be interested how I could send the logs to another machine. Could you give me a pointer how this could be done via serial or parallel cable? /ch --82I3+IH0IqGh5yIs Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename=messages Jun 18 13:46:25 hiob /kernel: Copyright (c) 1992-2001 The FreeBSD Project. Jun 18 13:46:25 hiob /kernel: Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 Jun 18 13:46:25 hiob /kernel: The Regents of the University of California. All rights reserved. Jun 18 13:46:25 hiob /kernel: FreeBSD 4.3-RELEASE #0: Mon Jun 18 02:57:03 CEST 2001 Jun 18 13:46:25 hiob /kernel: root@hiob.linuxlupe.de:/usr/src/sys/compile/TEST Jun 18 13:46:25 hiob /kernel: Timecounter "i8254" frequency 1193182 Hz Jun 18 13:46:25 hiob /kernel: CPU: Pentium III/Pentium III Xeon/Celeron (999.72-MHz 686-class CPU) Jun 18 13:46:25 hiob /kernel: Origin = "GenuineIntel" Id = 0x686 Stepping = 6 Jun 18 13:46:25 hiob /kernel: Features=0x387fbff Jun 18 13:46:25 hiob /kernel: real memory = 1073676288 (1048512K bytes) Jun 18 13:46:25 hiob /kernel: avail memory = 1041354752 (1016948K bytes) Jun 18 13:46:25 hiob /kernel: Preloaded elf kernel "kernel" at 0xc0451000. Jun 18 13:46:25 hiob /kernel: Preloaded elf module "mly.ko" at 0xc045109c. Jun 18 13:46:25 hiob /kernel: Pentium Pro MTRR support enabled Jun 18 13:46:25 hiob /kernel: md0: Malloc disk Jun 18 13:46:25 hiob /kernel: npx0: on motherboard Jun 18 13:46:25 hiob /kernel: npx0: INT 16 interface Jun 18 13:46:25 hiob /kernel: pcib0: on motherboard Jun 18 13:46:25 hiob /kernel: pci0: on pcib0 Jun 18 13:46:25 hiob /kernel: pci0: at 2.0 irq 11 Jun 18 13:46:25 hiob /kernel: fxp0: port 0x5400-0x543f mem 0xfb000000-0xfb0fffff,0xfb201000-0xfb201fff irq 10 at device 3.0 on pci0 Jun 18 13:46:25 hiob /kernel: fxp0: Ethernet address 00:d0:b7:b6:c3:06 Jun 18 13:46:25 hiob /kernel: pcib1: at device 7.0 on pci0 Jun 18 13:46:25 hiob /kernel: pci1: on pcib1 Jun 18 13:46:25 hiob /kernel: mly0: mem 0xfb500000-0xfb501fff irq 9 at device 7.1 on pci0 Jun 18 13:46:25 hiob /kernel: mly0: AcceleRAID 170 , 1 channel, firmware 6.00-7-00 (20001214), 64MB RAM Jun 18 13:46:25 hiob /kernel: fxp1: port 0x5440-0x547f mem 0xfb100000-0xfb1fffff,0xfb202000-0xfb202fff irq 5 at device 9.0 on pci0 Jun 18 13:46:25 hiob /kernel: fxp1: Ethernet address 00:02:b3:1a:e0:84 Jun 18 13:46:25 hiob /kernel: isab0: at device 15.0 on pci0 Jun 18 13:46:25 hiob /kernel: isa0: on isab0 Jun 18 13:46:25 hiob /kernel: atapci0: port 0x5480-0x548f,0x374-0x377,0x170-0x177 at device 15.1 on pci0 Jun 18 13:46:25 hiob /kernel: ata0: at 0x1f0 irq 14 on atapci0 Jun 18 13:46:25 hiob /kernel: ata1: at 0x170 irq 15 on atapci0 Jun 18 13:46:25 hiob /kernel: ohci0: mem 0xfb203000-0xfb203fff irq 9 at device 15.2 on pci0 Jun 18 13:46:25 hiob /kernel: usb0: OHCI version 1.0, legacy support Jun 18 13:46:25 hiob /kernel: usb0: on ohci0 Jun 18 13:46:25 hiob /kernel: usb0: USB revision 1.0 Jun 18 13:46:25 hiob /kernel: uhub0: (unknown) OHCI root hub, class 9/0, rev 1.00/1.00, addr 1 Jun 18 13:46:25 hiob /kernel: uhub0: 4 ports with 4 removable, self powered Jun 18 13:46:25 hiob /kernel: pcib2: on motherboard Jun 18 13:46:25 hiob /kernel: pci2: on pcib2 Jun 18 13:46:25 hiob /kernel: ahc0: port 0x5800-0x58ff mem 0xfd000000-0xfd000fff irq 9 at device 4.0 on pci2 Jun 18 13:46:25 hiob /kernel: aic7899: Wide Channel A, SCSI Id=7, 32/255 SCBs Jun 18 13:46:25 hiob /kernel: ahc1: port 0x6000-0x60ff mem 0xfd001000-0xfd001fff irq 11 at device 4.1 on pci2 Jun 18 13:46:25 hiob /kernel: aic7899: Wide Channel B, SCSI Id=7, 32/255 SCBs Jun 18 13:46:25 hiob /kernel: fdc0: at port 0x3f0-0x3f5,0x3f7 irq 6 drq 2 on isa0 Jun 18 13:46:25 hiob /kernel: fdc0: FIFO enabled, 8 bytes threshold Jun 18 13:46:25 hiob /kernel: fd0: <1440-KB 3.5" drive> on fdc0 drive 0 Jun 18 13:46:25 hiob /kernel: atkbdc0: at port 0x60,0x64 on isa0 Jun 18 13:46:25 hiob /kernel: atkbd0: flags 0x1 irq 1 on atkbdc0 Jun 18 13:46:25 hiob /kernel: kbd0 at atkbd0 Jun 18 13:46:25 hiob /kernel: vga0: at port 0x3c0-0x3df iomem 0xa0000-0xbffff on isa0 Jun 18 13:46:25 hiob /kernel: sc0: at flags 0x100 on isa0 Jun 18 13:46:25 hiob /kernel: sc0: VGA <16 virtual consoles, flags=0x300> Jun 18 13:46:25 hiob /kernel: sio0 at port 0x3f8-0x3ff irq 4 flags 0x10 on isa0 Jun 18 13:46:25 hiob /kernel: sio0: type 16550A Jun 18 13:46:25 hiob /kernel: sio1 at port 0x2f8-0x2ff irq 3 on isa0 Jun 18 13:46:25 hiob /kernel: sio1: type 16550A Jun 18 13:46:25 hiob /kernel: ppc0: at port 0x378-0x37f irq 7 on isa0 Jun 18 13:46:25 hiob /kernel: ppc0: SMC-like chipset (ECP/EPP/PS2/NIBBLE) in COMPATIBLE mode Jun 18 13:46:25 hiob /kernel: ppc0: FIFO with 16/16/8 bytes threshold Jun 18 13:46:25 hiob /kernel: plip0: on ppbus0 Jun 18 13:46:25 hiob /kernel: lpt0: on ppbus0 Jun 18 13:46:25 hiob /kernel: lpt0: Interrupt-driven port Jun 18 13:46:25 hiob /kernel: ppi0: on ppbus0 Jun 18 13:46:25 hiob /kernel: ata0-master: DMA limited to UDMA33, non-ATA66 compliant cable Jun 18 13:46:25 hiob /kernel: ad0: 19569MB [39761/16/63] at ata0-master UDMA33 Jun 18 13:46:25 hiob /kernel: acd0: CDROM at ata0-slave using PIO4 Jun 18 13:46:25 hiob /kernel: Waiting 15 seconds for SCSI devices to settle Jun 18 13:46:25 hiob /kernel: sa0 at ahc0 bus 0 target 3 lun 0 Jun 18 13:46:25 hiob /kernel: sa0: Removable Sequential Access SCSI-2 device Jun 18 13:46:25 hiob /kernel: sa0: 80.000MB/s transfers (40.000MHz, offset 32, 16bit) Jun 18 13:46:25 hiob /kernel: da0 at mly0 bus 0 target 0 lun 0 Jun 18 13:46:25 hiob /kernel: da0: Fixed Direct Access SCSI-3 device Jun 18 13:46:25 hiob /kernel: da0: 135.168MB/s transfers Jun 18 13:46:25 hiob /kernel: da0: 8392MB (17186816 512 byte sectors: 255H 63S/T 1069C) Jun 18 13:46:25 hiob /kernel: da1 at mly0 bus 0 target 1 lun 0 Jun 18 13:46:25 hiob /kernel: da1: Fixed Direct Access SCSI-3 device Jun 18 13:46:25 hiob /kernel: da1: 135.168MB/s transfers Jun 18 13:46:25 hiob /kernel: da1: 34712MB (71090176 512 byte sectors: 255H 63S/T 4425C) Jun 18 13:46:25 hiob /kernel: Mounting root from ufs:/dev/da0s1a Jun 18 13:48:07 hiob login: ROOT LOGIN (root) ON ttyv0 Jun 18 13:49:25 hiob /kernel: mly0: physical device 0:0 device cannot be accessed Jun 18 13:49:25 hiob /kernel: mly0: logical device 0 (da0) critical Jun 18 13:49:25 hiob /kernel: mly0: unknown event code - 61 Jun 18 13:49:25 hiob /kernel: (da0:mly0:0:0:0): lost device Jun 18 13:49:25 hiob /kernel: mly0: unknown event code - 162 Jun 18 13:50:01 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 13:50:01 hiob /kernel: size: 8192, resid: 8192, a_count: 8192, valid: 0x0 Jun 18 13:50:01 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 2 Jun 18 13:51:31 hiob sendmail[251]: f5IBpVX00251: SYSERR(root): collect: Cannot write ./dff5IBpVX00251 (bffsync, uid=0): Device not configured Jun 18 13:55:25 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 13:55:25 hiob /kernel: size: 8192, resid: 8192, a_count: 8192, valid: 0x0 Jun 18 13:55:25 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 2 Jun 18 13:55:55 hiob /kernel: mly0: logical device 0 (da0) rebuild completed Jun 18 13:55:55 hiob /kernel: mly0: physical device 0:4 rebuild completed Jun 18 13:55:55 hiob /kernel: cam_periph_alloc: attempt to re-allocate valid device da0 rejected Jun 18 13:55:55 hiob /kernel: daasync: Unable to attach to new device due to status 0x6 Jun 18 13:55:55 hiob /kernel: mly0: physical device 0:4 online Jun 18 13:56:25 hiob /kernel: mly0: logical device 0 (da0) online Jun 18 13:56:30 hiob sendmail[256]: f5IBuUq00256: SYSERR(root): collect: Cannot write ./dff5IBuUq00256 (bffsync, uid=0): Device not configured Jun 18 13:58:25 hiob /kernel: mly0: physical device 0:0 found Jun 18 13:58:25 hiob /kernel: mly0: physical device 0:0 standby Jun 18 13:58:55 hiob /kernel: mly0: physical device 0:4 device cannot be accessed Jun 18 13:58:55 hiob /kernel: mly0: logical device 0 (da0) critical Jun 18 13:58:55 hiob /kernel: mly0: unknown event code - 61 Jun 18 13:58:55 hiob /kernel: cam_periph_alloc: attempt to re-allocate valid device da0 rejected Jun 18 13:58:55 hiob /kernel: daasync: Unable to attach to new device due to status 0x6 Jun 18 13:58:55 hiob /kernel: mly0: unknown event code - 162 Jun 18 14:00:25 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 14:00:25 hiob /kernel: size: 16384, resid: 16384, a_count: 15976, valid: 0x0 Jun 18 14:00:25 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 4 Jun 18 14:00:25 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 14:00:25 hiob /kernel: size: 8192, resid: 8192, a_count: 8192, valid: 0x0 Jun 18 14:00:25 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 2 Jun 18 14:01:20 hiob sendmail[264]: f5IC1K200264: SYSERR(root): collect: Cannot write ./dff5IC1K200264 (bffsync, uid=0): Device not configured Jun 18 14:01:30 hiob sendmail[263]: f5IC1U100263: SYSERR(root): collect: Cannot write ./dff5IC1U100263 (bffsync, uid=0): Device not configured Jun 18 14:05:25 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 14:05:25 hiob /kernel: size: 8192, resid: 8192, a_count: 8192, valid: 0x0 Jun 18 14:05:25 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 2 Jun 18 14:05:55 hiob /kernel: mly0: logical device 0 (da0) rebuild completed Jun 18 14:05:55 hiob /kernel: mly0: physical device 0:0 rebuild completed Jun 18 14:05:55 hiob /kernel: cam_periph_alloc: attempt to re-allocate valid device da0 rejected Jun 18 14:05:55 hiob /kernel: daasync: Unable to attach to new device due to status 0x6 Jun 18 14:05:55 hiob /kernel: mly0: physical device 0:0 online Jun 18 14:05:55 hiob /kernel: mly0: logical device 0 (probe0) online Jun 18 14:06:30 hiob sendmail[268]: f5IC6U100268: SYSERR(root): collect: Cannot write ./dff5IC6U100268 (bffsync, uid=0): Device not configured Jun 18 14:10:25 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 14:10:25 hiob /kernel: size: 8192, resid: 8192, a_count: 8192, valid: 0x0 Jun 18 14:10:25 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 2 Jun 18 14:11:25 hiob /kernel: mly0: physical device 0:4 found Jun 18 14:11:25 hiob /kernel: mly0: physical device 0:4 standby Jun 18 14:11:30 hiob sendmail[272]: f5ICBUR00272: SYSERR(root): collect: Cannot write ./dff5ICBUR00272 (bffsync, uid=0): Device not configured Jun 18 14:13:56 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 14:13:56 hiob /kernel: size: 7168, resid: 7168, a_count: 7156, valid: 0x0 Jun 18 14:13:56 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 2 Jun 18 14:14:26 hiob /kernel: spec_getpages:(#da/0x20004) I/O read failure: (error=6) bp 0xceb71088 vp 0xd9a28840 Jun 18 14:14:26 hiob /kernel: size: 65536, resid: 65536, a_count: 65536, valid: 0x0 Jun 18 14:14:26 hiob /kernel: nread: 0, reqpage: 0, pindex: 0, pcount: 16 Jun 18 14:14:52 hiob syslogd: exiting on signal 15 --82I3+IH0IqGh5yIs-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Jun 18 19: 0:21 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from a.mx.everquick.net (a.mx.everquick.net [216.89.137.3]) by hub.freebsd.org (Postfix) with ESMTP id 68D8D37B403 for ; Mon, 18 Jun 2001 19:00:18 -0700 (PDT) (envelope-from eddy+public+spam@noc.everquick.net) Received: from localhost (eddy@localhost) by a.mx.everquick.net (8.10.2/8.10.2) with ESMTP id f5J1l0004925; Tue, 19 Jun 2001 01:47:04 GMT X-EverQuick-No-Abuse: Report any e-mail abuse to Date: Tue, 19 Jun 2001 01:47:00 +0000 (GMT) From: "E.B. Dreger" To: mylexsup@us.ibm.com Cc: freebsd-scsi@freebsd.org Subject: Mylex RAID drivers Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Greetings, Where would one find the information necessary to write a device driver for the DAC960 family of controllers? I ask because many FreeBSD users have DAC960 controllers (I alone have one rev 2.xx and two rev 3.xx cards) that we wish to use with FreeBSD. However, it seems that the device driver needs work. We understand that Linux has a working driver for firmware rev 3.xx and newer cards. However, due to licensing differences between Linux and BSD, we [the FreeBSD developer and user community] wish to continue working with the more flexible BSD license. Please direct us to a developer's resource so that will allow us to further develop BSD drivers, and to continue enjoying Mylex RAID controllers. Thanks in advance, Eddy --------------------------------------------------------------------------- Brotsman & Dreger, Inc. EverQuick Internet Division Phone: +1 (316) 794-8922 Wichita/(Inter)national Phone: +1 (785) 865-5885 Lawrence --------------------------------------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Jun 18 19: 1:14 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.dis.org (mass.dis.org [216.240.45.41]) by hub.freebsd.org (Postfix) with ESMTP id D958637B401 for ; Mon, 18 Jun 2001 19:01:12 -0700 (PDT) (envelope-from msmith@mass.dis.org) Received: from mass.dis.org (localhost [127.0.0.1]) by mass.dis.org (8.11.3/8.11.3) with ESMTP id f5J2CTP03735 for ; Mon, 18 Jun 2001 19:12:29 -0700 (PDT) (envelope-from msmith@mass.dis.org) Message-Id: <200106190212.f5J2CTP03735@mass.dis.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: scsi@freebsd.org Subject: device arrival/departure/possible state change? Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 18 Jun 2001 19:12:29 -0700 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Can someone that understands CAM explain what (more or less) one needs to do to handle the possible arrival/departure of a device? It appears that if I send a rescan cdb to rescan an already existing device, and the device is still present, something gets really upset. Specifically, if I know that a particular bus/target may have undergone a status change, and I want CAM to work out whether it's still there and talking, what do I have to do in the case where a) we have a periph attached to the b/t, and b) where there isn't a periph attached to it? Examples would also be welcome. Regards, Mike -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Jun 18 19: 3:19 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.dis.org (mass.dis.org [216.240.45.41]) by hub.freebsd.org (Postfix) with ESMTP id 6F9A437B401 for ; Mon, 18 Jun 2001 19:03:16 -0700 (PDT) (envelope-from msmith@mass.dis.org) Received: from mass.dis.org (localhost [127.0.0.1]) by mass.dis.org (8.11.3/8.11.3) with ESMTP id f5J2E0P03750; Mon, 18 Jun 2001 19:14:00 -0700 (PDT) (envelope-from msmith@mass.dis.org) Message-Id: <200106190214.f5J2E0P03750@mass.dis.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: "E.B. Dreger" Cc: mylexsup@us.ibm.com, freebsd-scsi@freebsd.org Subject: Re: Mylex RAID drivers In-reply-to: Your message of "Tue, 19 Jun 2001 01:47:00 -0000." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 18 Jun 2001 19:14:00 -0700 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > Greetings, > > Where would one find the information necessary to write a device driver > for the DAC960 family of controllers? You ask me. 8) > I ask because many FreeBSD users have DAC960 controllers (I alone have one > rev 2.xx and two rev 3.xx cards) that we wish to use with FreeBSD. > However, it seems that the device driver needs work. That's correct. > We understand that Linux has a working driver for firmware rev 3.xx and > newer cards. However, due to licensing differences between Linux and BSD, > we [the FreeBSD developer and user community] wish to continue working > with the more flexible BSD license. > > Please direct us to a developer's resource so that will allow us to > further develop BSD drivers, and to continue enjoying Mylex RAID > controllers. If you're seriously interested in fixing the problems with these adapters, please contact me off the list and we can talk about it. The issues are relatively straightforward, they just need someone with a suitable adapter and some time to work on it. Regards, Mike -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Mon Jun 18 21:57:21 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.dis.org (mass.dis.org [216.240.45.41]) by hub.freebsd.org (Postfix) with ESMTP id AD97F37B403 for ; Mon, 18 Jun 2001 21:57:19 -0700 (PDT) (envelope-from msmith@mass.dis.org) Received: from mass.dis.org (localhost [127.0.0.1]) by mass.dis.org (8.11.3/8.11.3) with ESMTP id f5J58XP05458; Mon, 18 Jun 2001 22:08:33 -0700 (PDT) (envelope-from msmith@mass.dis.org) Message-Id: <200106190508.f5J58XP05458@mass.dis.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: Clemens Hermann Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available In-reply-to: Your message of "Mon, 18 Jun 2001 13:08:51 +0200." <20010618130851.C1070@homer.local> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 18 Jun 2001 22:08:33 -0700 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > Jun 18 13:49:25 hiob /kernel: mly0: physical device 0:0 device cannot be accessed > Jun 18 13:49:25 hiob /kernel: mly0: logical device 0 (da0) critical > Jun 18 13:49:25 hiob /kernel: mly0: unknown event code - 61 > Jun 18 13:49:25 hiob /kernel: (da0:mly0:0:0:0): lost device Ok. Here is where we come unstuck; the event code has confused physical device 0:0 with logical device 0. Can you try the new version at: http://people.freebsd.org/~msmith/RAID/mylex/mly-20010618.tar.gz I've undone the madness I had before wrt. switching 'virtual' bus numbers around, and it should survive the above. I discovered the other day that I can't actually find the key for my hotswap bays, so I can't test drive pulls, which is cramping my testing a bit. I still don't think that it'll properly handle a drive going totally offline and then coming back, but I could be wrong. Your test case here should be fixed though. Please let me know how you go; I really appreciate the feedback. Regards, Mike -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Tue Jun 19 2:20:15 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mailout01.sul.t-online.de (mailout01.sul.t-online.com [194.25.134.80]) by hub.freebsd.org (Postfix) with ESMTP id A656D37B401 for ; Tue, 19 Jun 2001 02:20:11 -0700 (PDT) (envelope-from haribeau@gmx.de) Received: from fwd02.sul.t-online.de by mailout01.sul.t-online.de with smtp id 15CHgA-0002tP-0C; Tue, 19 Jun 2001 11:20:10 +0200 Received: from asterix.local (320080844193-0001@[217.80.84.120]) by fmrl02.sul.t-online.com with smtp id 15CHfz-0zfM0WC; Tue, 19 Jun 2001 11:19:59 +0200 Received: (qmail 537 invoked from network); 19 Jun 2001 09:19:52 -0000 Received: from homer.local (HELO homer.local.nlocal) (192.168.1.50) by 0 with SMTP; 19 Jun 2001 09:19:52 -0000 Received: (nullmailer pid 850 invoked by uid 1100); Tue, 19 Jun 2001 09:19:52 -0000 Date: Tue, 19 Jun 2001 11:19:52 +0200 From: Clemens Hermann To: Mike Smith Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available Message-ID: <20010619111952.A823@homer.local> References: <20010618130851.C1070@homer.local> <200106190508.f5J58XP05458@mass.dis.org> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="T4sUOijqQbZv57TR" Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200106190508.f5J58XP05458@mass.dis.org> von Mike Smith am 18.Jun.2001 um 22:08:33 (-0700) X-Mailer: Mutt 1.2.5i (FreeBSD 4.3-RELEASE i386) X-Sender: 320080844193-0001@t-dialin.net Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org --T4sUOijqQbZv57TR Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Hi, > Can you try the new version at: > > http://people.freebsd.org/~msmith/RAID/mylex/mly-20010618.tar.gz sure. It worked :). I attached the according syslog. the behaviour is the same as the old driver with Markus' patch. > I still don't think that it'll properly handle a drive going totally > offline and then coming back, but I could be wrong. I think you are :). As you might have noticed I am no expert but to me it looks like things work as they should. /ch --T4sUOijqQbZv57TR Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename=messages Jun 19 13:02:03 hiob /kernel: mly0: physical device 0:4 device cannot be accessed Jun 19 13:02:03 hiob /kernel: mly0: logical device 0 (da0) critical Jun 19 13:02:03 hiob /kernel: mly0: unknown event code - 61 Jun 19 13:02:03 hiob /kernel: mly0: unknown event code - 162 Jun 19 13:08:33 hiob /kernel: mly0: logical device 0 (da0) rebuild completed Jun 19 13:08:33 hiob /kernel: mly0: physical device 0:0 rebuild completed Jun 19 13:08:33 hiob /kernel: mly0: physical device 0:0 online Jun 19 13:08:33 hiob /kernel: mly0: logical device 0 (da0) online Jun 19 13:10:00 hiob /kernel: mly0: physical device 0:4 found Jun 19 13:10:00 hiob /kernel: mly0: physical device 0:4 standby --T4sUOijqQbZv57TR-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Tue Jun 19 2:43: 1 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.dis.org (mass.dis.org [216.240.45.41]) by hub.freebsd.org (Postfix) with ESMTP id 8A99537B403 for ; Tue, 19 Jun 2001 02:42:59 -0700 (PDT) (envelope-from msmith@mass.dis.org) Received: from mass.dis.org (localhost [127.0.0.1]) by mass.dis.org (8.11.3/8.11.3) with ESMTP id f5J9s4P09110; Tue, 19 Jun 2001 02:54:04 -0700 (PDT) (envelope-from msmith@mass.dis.org) Message-Id: <200106190954.f5J9s4P09110@mass.dis.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: Clemens Hermann Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available In-reply-to: Your message of "Tue, 19 Jun 2001 11:19:52 +0200." <20010619111952.A823@homer.local> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 19 Jun 2001 02:54:04 -0700 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > Hi, > > > Can you try the new version at: > > > > http://people.freebsd.org/~msmith/RAID/mylex/mly-20010618.tar.gz > > sure. It worked :). I attached the according syslog. the behaviour is > the same as the old driver with Markus' patch. > > > I still don't think that it'll properly handle a drive going totally > > offline and then coming back, but I could be wrong. > > I think you are :). As you might have noticed I am no expert but to me > it looks like things work as they should. Well, mostly. Can you pull another disk while it's rebuilding, so that the array goes completely offline? I want to be sure that CAM correctly kills da0. Aside from that though, yes, this looks better. 8) -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Tue Jun 19 3:16:22 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mailout01.sul.t-online.de (mailout01.sul.t-online.com [194.25.134.80]) by hub.freebsd.org (Postfix) with ESMTP id 1E7C737B40A for ; Tue, 19 Jun 2001 03:16:18 -0700 (PDT) (envelope-from haribeau@gmx.de) Received: from fwd02.sul.t-online.de by mailout01.sul.t-online.de with smtp id 15CIYT-0003RQ-01; Tue, 19 Jun 2001 12:16:17 +0200 Received: from asterix.local (320080844193-0001@[217.80.84.69]) by fmrl02.sul.t-online.com with smtp id 15CIYI-1TXdRoC; Tue, 19 Jun 2001 12:16:06 +0200 Received: (qmail 1139 invoked from network); 19 Jun 2001 10:15:59 -0000 Received: from homer.local (HELO homer.local.nlocal) (192.168.1.50) by 0 with SMTP; 19 Jun 2001 10:15:59 -0000 Received: (nullmailer pid 1573 invoked by uid 1100); Tue, 19 Jun 2001 10:15:59 -0000 Date: Tue, 19 Jun 2001 12:15:59 +0200 From: Clemens Hermann To: Mike Smith Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available Message-ID: <20010619121559.A1493@homer.local> References: <20010619111952.A823@homer.local> <200106190954.f5J9s4P09110@mass.dis.org> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="9amGYk9869ThD9tj" Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200106190954.f5J9s4P09110@mass.dis.org> von Mike Smith am 19.Jun.2001 um 02:54:04 (-0700) X-Mailer: Mutt 1.2.5i (FreeBSD 4.3-RELEASE i386) X-Sender: 320080844193-0001@t-dialin.net Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org --9amGYk9869ThD9tj Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Hi again, > Well, mostly. Can you pull another disk while it's rebuilding, so that > the array goes completely offline? I want to be sure that CAM correctly > kills da0. can it be done with da1? da0 is the system and I did not configure the backup yet :) Drive 2 and 3 are the mirror (da1), drive 4 the spare. I removed drive 2 first and while the rebuilt took place I removed drive 3. Later on I re-inserted both of them. The logs are attached. > Aside from that though, yes, this looks better. 8) Would you recommend using this driver instead of the old one with Markus' Patch on a production server? /ch --9amGYk9869ThD9tj Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename=messages Jun 19 13:57:03 hiob /kernel: mly0: physical device 0:2 device cannot be accessed Jun 19 13:57:03 hiob /kernel: mly0: logical device 1 (da1) critical Jun 19 13:57:03 hiob /kernel: mly0: unknown event code - 61 Jun 19 13:57:03 hiob /kernel: mly0: unknown event code - 162 Jun 19 13:57:33 hiob /kernel: mly0: physical device 0:3 scsi bus phase error Jun 19 13:57:33 hiob /kernel: mly0: logical device 1 (da1) offline Jun 19 13:57:33 hiob /kernel: mly0: physical device 0:4 rebuild failed due to logical drive failure Jun 19 13:57:33 hiob /kernel: mly0: logical device 1 (da1) rebuild failed due to logical drive failure Jun 19 13:59:03 hiob /kernel: mly0: physical device 0:3 found Jun 19 13:59:33 hiob /kernel: mly0: physical device 0:2 found Jun 19 13:59:33 hiob /kernel: mly0: physical device 0:2 standby --9amGYk9869ThD9tj-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Tue Jun 19 3:55:24 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.dis.org (mass.dis.org [216.240.45.41]) by hub.freebsd.org (Postfix) with ESMTP id 3FDAF37B401 for ; Tue, 19 Jun 2001 03:55:22 -0700 (PDT) (envelope-from msmith@mass.dis.org) Received: from mass.dis.org (localhost [127.0.0.1]) by mass.dis.org (8.11.3/8.11.3) with ESMTP id f5JB6ZP10309; Tue, 19 Jun 2001 04:06:35 -0700 (PDT) (envelope-from msmith@mass.dis.org) Message-Id: <200106191106.f5JB6ZP10309@mass.dis.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: Clemens Hermann Cc: scsi@freebsd.org Subject: Re: 'mly' driver update/bugfix snapshot available In-reply-to: Your message of "Tue, 19 Jun 2001 12:15:59 +0200." <20010619121559.A1493@homer.local> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 19 Jun 2001 04:06:00 -0700 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > > Well, mostly. Can you pull another disk while it's rebuilding, so that > > the array goes completely offline? I want to be sure that CAM correctly > > kills da0. > > can it be done with da1? da0 is the system and I did not configure the > backup yet :) Heh. Yes, da1 would be fine for the test. > Later on I re-inserted both of them. > The logs are attached. Ok. I note that CAM didn't kill da1; that's bad. I'll have to work that out at some point. > > Aside from that though, yes, this looks better. 8) > > Would you recommend using this driver instead of the old one with > Markus' Patch on a production server? I'm not certain, yet. Neither of them are interfacing very well to CAM when it comes to drive failures. The newer driver will perform a little better, and IMO is better-behaved, but it has had less testing. I will be continuing to support this driver and eventually commit it within a few weeks; I just need to get my keys. 8) Regards, Mike -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Tue Jun 19 19:36:18 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from mail.datausa.com (mail.datausa.com [207.174.131.1]) by hub.freebsd.org (Postfix) with ESMTP id CDA7337B406 for ; Tue, 19 Jun 2001 19:36:15 -0700 (PDT) (envelope-from brad@wcubed.net) Received: from wcubed.net ([209.180.250.253]) by mail.datausa.com (8.9.3/8.9.1) with ESMTP id UAA20162 for ; Tue, 19 Jun 2001 20:28:55 -0600 (MDT) Message-ID: <3B300C4A.4D771C68@wcubed.net> Date: Tue, 19 Jun 2001 20:36:58 -0600 From: Brad Waite X-Mailer: Mozilla 4.76 [en] (Win98; U) X-Accept-Language: en,ja,zh,zh-CN,en-GB MIME-Version: 1.0 To: freebsd-scsi@freebsd.org Subject: data overruns on 'tar x' from DAT Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org I'm trying to retrieve data from a tar archive created from a Sun box on DAT, but I keep getting these errors: Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): data overrun detected in Data-in phase. Tag == 0x6. Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): data overrun detected in Data-in phase. Tag == 0x6. Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): Have seen Data Phase. Length = 20480. NumSGs = 5. Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): Have seen Data Phase. Length = 20480. NumSGs = 5. After which tar says there's a read error and pukes. What's causing this? Thanks, Brad Waite To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Tue Jun 19 20:53:57 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 7E0CD37B403 for ; Tue, 19 Jun 2001 20:53:54 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from beppo (mjacob@beppo [192.67.166.79]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5K3rZg67092; Tue, 19 Jun 2001 20:53:36 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Tue, 19 Jun 2001 20:53:35 -0700 (PDT) From: Matthew Jacob Reply-To: mjacob@feral.com To: Brad Waite Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: data overruns on 'tar x' from DAT In-Reply-To: <3B300C4A.4D771C68@wcubed.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org The record being read from tape is larger than the buffer you've allocated in tar. It may be you're in trouble. Default maxphys on Sun's is 124KB. That's substantially larger than that which is allowed by default in FreeBSD. Check -questions as to how to cope. On Tue, 19 Jun 2001, Brad Waite wrote: > I'm trying to retrieve data from a tar archive created from a Sun box on DAT, > but I keep getting these errors: > > Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): data overrun detected in Data-in > phase. Tag == 0x6. > Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): data overrun detected in Data-in > phase. Tag == 0x6. > Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): Have seen Data Phase. Length = > 20480. NumSGs = 5. > Jun 19 19:26:31 host /kernel: (sa0:ahc0:0:3:0): Have seen Data Phase. Length = > 20480. NumSGs = 5. > > After which tar says there's a read error and pukes. What's causing this? > > Thanks, > > Brad Waite > > 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 Jun 20 14:47: 3 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from juneau.dogwood.com (juneau.dogwood.com [63.96.228.133]) by hub.freebsd.org (Postfix) with ESMTP id D63C337B401 for ; Wed, 20 Jun 2001 14:47:01 -0700 (PDT) (envelope-from dave@juneau.dogwood.com) Received: (from dave@localhost) by juneau.dogwood.com (8.11.4/8.11.4) id f5KLWp434598 for freebsd-scsi@freebsd.org; Wed, 20 Jun 2001 14:32:51 -0700 (PDT) (envelope-from dave) From: dave Message-Id: <200106202132.f5KLWp434598@juneau.dogwood.com> Subject: CAM/AHC debugging To: freebsd-scsi@freebsd.org Date: Wed, 20 Jun 2001 14:32:51 -0700 (PDT) X-Mailer: ELM [version 2.4ME+ PL92 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Is there a document that explains debugging and/or the architecture of the CAM and ahc drivers somewhere? thanks, dave c To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Wed Jun 20 14:50:27 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 37B1E37B401 for ; Wed, 20 Jun 2001 14:50:25 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from beppo (mjacob@beppo [192.67.166.79]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5KLoMg69946; Wed, 20 Jun 2001 14:50:22 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Wed, 20 Jun 2001 14:50:21 -0700 (PDT) From: Matthew Jacob Reply-To: mjacob@feral.com To: dave Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: CAM/AHC debugging In-Reply-To: <200106202132.f5KLWp434598@juneau.dogwood.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org The formal architecture of CAM is available from www.t10.org (CAM is derivative of SCSI-2). Justin used to have a web page talking about the implementation for FreeBSD, but it wasn't all that complete. On Wed, 20 Jun 2001, dave wrote: > Is there a document that explains debugging and/or the architecture of > the CAM and ahc drivers somewhere? > > thanks, > dave c > > 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 Thu Jun 21 9:25:42 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 4967A37B401 for ; Thu, 21 Jun 2001 09:25:40 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5LGPdg71947 for ; Thu, 21 Jun 2001 09:25:39 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Thu, 21 Jun 2001 09:25:37 -0700 (PDT) From: Matthew Jacob Reply-To: To: Subject: best place for options Message-ID: <20010621092318.A18589-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org There is an option (CAM_NEW_TRAN_CODE) which I need to begin exercising- and then start getting folks to conver their HBAs on over to using. It seems to me that the best pace for this would be in opt_global.h so I don't have to run around figuring out who includes what. Or should I make this an option for opt_cam.h and make sure HBAs include it? -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Thu Jun 21 9:47: 2 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from crufty.research.bell-labs.com (crufty.research.bell-labs.com [204.178.16.49]) by hub.freebsd.org (Postfix) with SMTP id 85DDD37B406 for ; Thu, 21 Jun 2001 09:46:56 -0700 (PDT) (envelope-from jkf@zydeco.research.bell-labs.com) Received: from scummy.research.bell-labs.com ([135.104.2.10]) by crufty; Thu Jun 21 12:42:22 EDT 2001 Received: from zydeco.research.bell-labs.com ([135.104.120.150]) by scummy; Thu Jun 21 12:45:51 EDT 2001 Received: (from jkf@localhost) by zydeco.research.bell-labs.com (8.9.1/8.9.1) id MAA07853; Thu, 21 Jun 2001 12:45:15 -0400 (EDT) Date: Thu, 21 Jun 2001 12:45:15 -0400 (EDT) From: Jeff Fellin Message-Id: <200106211645.MAA07853@zydeco.research.bell-labs.com> To: scsi@FreeBSD.ORG, mjacob@feral.com Subject: Re: best place for options X-Sun-Charset: US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Matt, From looking at what the new option does I think the option should be in opt_cam.h, and make sure all HBAs include it. Jeff Fellin > From owner-freebsd-scsi@FreeBSD.ORG Thu Jun 21 12:26 EDT 2001 > Delivered-To: freebsd-scsi@freebsd.org > Date: Thu, 21 Jun 2001 09:25:37 -0700 (PDT) > From: Matthew Jacob > Reply-To: > To: > Subject: best place for options > Sender: owner-freebsd-scsi@FreeBSD.ORG > List-ID: > List-Archive: (Web Archive) > List-Help: (List Instructions) > List-Subscribe: > List-Unsubscribe: > Precedence: bulk > X-Lines: 14 > > > There is an option (CAM_NEW_TRAN_CODE) which I need to begin exercising- and > then start getting folks to conver their HBAs on over to using. > > It seems to me that the best pace for this would be in opt_global.h so I > don't have to run around figuring out who includes what. Or should I make > this an option for opt_cam.h and make sure HBAs include it? > > -matt > > > > 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 Thu Jun 21 15:20:10 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from munin.odin-corporation.com (munin.odin-corporation.com [216.233.173.18]) by hub.freebsd.org (Postfix) with ESMTP id E914B37B401 for ; Thu, 21 Jun 2001 15:20:03 -0700 (PDT) (envelope-from lars@odin-corporation.com) Received: from odin-corporation.com (localhost [127.0.0.1]) by munin.odin-corporation.com (8.11.3/8.11.1) with ESMTP id f5LMJmS03614 for ; Thu, 21 Jun 2001 17:19:48 -0500 (CDT) (envelope-from lars@odin-corporation.com) Message-ID: <3B327303.62AB5B2E@odin-corporation.com> Date: Thu, 21 Jun 2001 17:19:47 -0500 From: Lars Fredriksen Organization: Odin Corporation X-Mailer: Mozilla 4.76 [en] (X11; U; FreeBSD 5.0-CURRENT i386) X-Accept-Language: no, en MIME-Version: 1.0 To: freebsd-scsi@freebsd.org Subject: upping the writefilemarks timeout Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Hi, Would anyone object if I commit a change that basically doubles the timeout value used when writing filemarks? Perhaps we should make this a sysctl variable? Currently scsi_write_filemarks() is called from sawritefilemarks() with a timeout of 60 seconds. For some older tape drives this is not enough. I have been testing my Exabyte 8500 with this fix, and it now reliably is able to write the filemark without timing out. Comments? Thanks, Lars To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 10:47:25 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from serenity.mcc.ac.uk (serenity.mcc.ac.uk [130.88.200.93]) by hub.freebsd.org (Postfix) with ESMTP id 5A37937B403 for ; Fri, 22 Jun 2001 10:47:22 -0700 (PDT) (envelope-from jcm@freebsd-uk.eu.org) Received: from dogma.freebsd-uk.eu.org ([130.88.200.97] ident=root) by serenity.mcc.ac.uk with esmtp (Exim 2.05 #6) id 15DV1d-000Ajb-00 for freebsd-scsi@freebsd.org; Fri, 22 Jun 2001 18:47:21 +0100 Received: (from jcm@localhost) by dogma.freebsd-uk.eu.org (8.11.3/8.11.1) id f5MHlKA60563 for freebsd-scsi@freebsd.org; Fri, 22 Jun 2001 18:47:20 +0100 (BST) (envelope-from jcm) Date: Fri, 22 Jun 2001 18:47:20 +0100 From: j mckitrick To: freebsd-scsi@freebsd.org Subject: how to invalidate scsi connection to driver module Message-ID: <20010622184720.B60256@dogma.freebsd-uk.eu.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Here is the code for a scsi removable media drive. If this is to become a module, the cam/scsi attachment must be removed. I have tried calling cam_sim_free() and xpt_bus_deregister() but when the module is reloaded, the cam system assigns the next higher minor device number, and then crashes when accessed. Any ideas here how to reset cam to a state where it knows nothing about the device? /* ** Now tell the generic SCSI layer ** about our bus. */ devq = cam_simq_alloc(/*maxopenings*/1); /* XXX What about low-level detach on error? */ if (devq == NULL) return (ENXIO); vpo->sim = cam_sim_alloc(vpo_action, vpo_poll, "vpo", vpo, device_get_unit(dev), /*untagged*/1, /*tagged*/0, devq); if (vpo->sim == NULL) { cam_simq_free(devq); return (ENXIO); } if (xpt_bus_register(vpo->sim, /*bus*/0) != CAM_SUCCESS) { cam_sim_free(vpo->sim, /*free_devq*/TRUE); return (ENXIO); } /* all went ok */ vpo_cam_rescan(vpo); /* have CAM rescan the bus */ return (0); Jonathon -- Microsoft complaining about the source license used by Linux is like the event horizon calling the kettle black. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 12: 2:15 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 1060F37B403; Fri, 22 Jun 2001 12:02:10 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5MJ29g75003; Fri, 22 Jun 2001 12:02:09 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 12:02:04 -0700 (PDT) From: Matthew Jacob Reply-To: To: Mike Smith Cc: Subject: Re: device arrival/departure/possible state change? Message-ID: <20010622115317.Q20923-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org I won't claim to answer helpfully except in generalities: >Can someone that understands CAM explain what (more or less) one needs to >do to handle the possible arrival/departure of a device? > >It appears that if I send a rescan cdb to rescan an already existing >device, and the device is still present, something gets really upset. > >Specifically, if I know that a particular bus/target may have undergone a >status change, and I want CAM to work out whether it's still there and >talking, what do I have to do in the case where a) we have a periph >attached to the b/t, and b) where there isn't a periph attached to it? Basically, you have to say a device is gone before you can ask for it to be found again. -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 12: 9:52 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 00C8937B403 for ; Fri, 22 Jun 2001 12:09:49 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5MJ7Fg75013; Fri, 22 Jun 2001 12:07:15 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 12:07:09 -0700 (PDT) From: Matthew Jacob Reply-To: To: j mckitrick Cc: Subject: Re: how to invalidate scsi connection to driver module In-Reply-To: <20010622183909.A60256@dogma.freebsd-uk.eu.org> Message-ID: <20010622120232.B20923-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Fri, 22 Jun 2001, j mckitrick wrote: > > Here is the code for a scsi removable media drive. If this is to become a > module, the cam/scsi attachment must be removed. I have tried calling > cam_sim_free() and xpt_bus_deregister() but when the module is reloaded, the > cam system assigns the next higher minor device number, and then crashes > when accessed. Any ideas here how to reset cam to a state where it knows > nothing about the device? The xpt_bus_deregister is supposed to do an AC_LOST_DEVICE on all attached devices. I don't remember what is supposed to happen to recycle unit numbers.... Take a look at camperiphunit- it really should recycle unit numbers if the old ones have truly gone away. > > > > /* > ** Now tell the generic SCSI layer > ** about our bus. > */ > devq = cam_simq_alloc(/*maxopenings*/1); > /* XXX What about low-level detach on error? */ > if (devq == NULL) > return (ENXIO); > > vpo->sim = cam_sim_alloc(vpo_action, vpo_poll, "vpo", vpo, > device_get_unit(dev), > /*untagged*/1, /*tagged*/0, devq); > if (vpo->sim == NULL) { > cam_simq_free(devq); > return (ENXIO); > } > > if (xpt_bus_register(vpo->sim, /*bus*/0) != CAM_SUCCESS) { > cam_sim_free(vpo->sim, /*free_devq*/TRUE); > return (ENXIO); > } > > /* all went ok */ > > vpo_cam_rescan(vpo); /* have CAM rescan the bus */ > > return (0); > > Jonathon > -- > Microsoft complaining about the source license used by > Linux is like the event horizon calling the kettle black. > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-hackers" 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 Fri Jun 22 12:35:45 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from serenity.mcc.ac.uk (serenity.mcc.ac.uk [130.88.200.93]) by hub.freebsd.org (Postfix) with ESMTP id A0B3237B403 for ; Fri, 22 Jun 2001 12:35:42 -0700 (PDT) (envelope-from jcm@freebsd-uk.eu.org) Received: from dogma.freebsd-uk.eu.org ([130.88.200.97] ident=root) by serenity.mcc.ac.uk with esmtp (Exim 2.05 #6) id 15DWiT-000JW5-00; Fri, 22 Jun 2001 20:35:41 +0100 Received: (from jcm@localhost) by dogma.freebsd-uk.eu.org (8.11.3/8.11.1) id f5MJZeg62840; Fri, 22 Jun 2001 20:35:41 +0100 (BST) (envelope-from jcm) Date: Fri, 22 Jun 2001 20:35:40 +0100 From: j mckitrick To: Matthew Jacob Cc: scsi@freebsd.org Subject: Re: how to invalidate scsi connection to driver module Message-ID: <20010622203540.A62806@dogma.freebsd-uk.eu.org> References: <20010622183909.A60256@dogma.freebsd-uk.eu.org> <20010622120232.B20923-100000@wonky.feral.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <20010622120232.B20923-100000@wonky.feral.com>; from mjacob@feral.com on Fri, Jun 22, 2001 at 12:07:09PM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Fri, Jun 22, 2001 at 12:07:09PM -0700, Matthew Jacob wrote: | | | On Fri, 22 Jun 2001, j mckitrick wrote: | | > | > Here is the code for a scsi removable media drive. If this is to become a | > module, the cam/scsi attachment must be removed. I have tried calling | > cam_sim_free() and xpt_bus_deregister() but when the module is reloaded, the | > cam system assigns the next higher minor device number, and then crashes | > when accessed. Any ideas here how to reset cam to a state where it knows | > nothing about the device? | | The xpt_bus_deregister is supposed to do an AC_LOST_DEVICE on all attached | devices. I don't remember what is supposed to happen to recycle unit | numbers.... | | Take a look at camperiphunit- it really should recycle unit numbers if the old | ones have truly gone away. Hmmm. I'm new to cam/scsi code, so i'm not sure what i need to do here. Apparently, the 2 functions i call *aren't* enough to make the unit number go away. jcm To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 12:40:54 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id BC8B437B406 for ; Fri, 22 Jun 2001 12:40:51 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5MJcIg75072; Fri, 22 Jun 2001 12:38:18 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 12:38:13 -0700 (PDT) From: Matthew Jacob Reply-To: To: j mckitrick Cc: Subject: Re: how to invalidate scsi connection to driver module In-Reply-To: <20010622203540.A62806@dogma.freebsd-uk.eu.org> Message-ID: <20010622123714.R20923-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > | > | > | On Fri, 22 Jun 2001, j mckitrick wrote: > | > | > > | > Here is the code for a scsi removable media drive. If this is to become a > | > module, the cam/scsi attachment must be removed. I have tried calling > | > cam_sim_free() and xpt_bus_deregister() but when the module is reloaded, the > | > cam system assigns the next higher minor device number, and then crashes > | > when accessed. Any ideas here how to reset cam to a state where it knows > | > nothing about the device? > | > | The xpt_bus_deregister is supposed to do an AC_LOST_DEVICE on all attached > | devices. I don't remember what is supposed to happen to recycle unit > | numbers.... > | > | Take a look at camperiphunit- it really should recycle unit numbers if the old > | ones have truly gone away. > > Hmmm. I'm new to cam/scsi code, so i'm not sure what i need to do here. > Apparently, the 2 functions i call *aren't* enough to make the unit number > go away. > Well, xpt_bus_deregister should invalidate all attached peripherals. But it's not clear to me whether it's your periph driver that's fielding (and swallowing) the AC_LOST_DEVICE, thus leaving things around. -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 13:55:44 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 530FF37B401 for ; Fri, 22 Jun 2001 13:55:41 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5MKr3g75178; Fri, 22 Jun 2001 13:53:03 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 13:52:57 -0700 (PDT) From: Matthew Jacob Reply-To: To: j mckitrick Cc: Subject: Re: how to invalidate scsi connection to driver module In-Reply-To: <20010622123714.R20923-100000@wonky.feral.com> Message-ID: <20010622135159.B20923-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org To follow up on this- Since I've been fooling around with the CAM_NEW_TRAN_CODE and Fibre Channel, I have the SIM call xpt_async with AC_LOST_DEVICE when a device goes away, so I had: isp0: Retaining Loop ID 0x20 for Target 32 (Port 0xb2) isp0: Retaining Loop ID 0x21 for Target 33 (Port 0xb1) isp0: Retaining Loop ID 0x22 for Target 34 (Port 0xae) isp0: Retaining Loop ID 0x23 for Target 35 (Port 0xad) isp0: Retaining Loop ID 0x24 for Target 36 (Port 0xac) isp0: Retaining Loop ID 0x25 for Target 37 (Port 0xab) isp0: Retaining Loop ID 0x2d for Target 45 (Port 0x9e) isp0: Retaining Loop ID 0x3d for Target 61 (Port 0x75) isp0: Retaining Loop ID 0x6d for Target 109 (Port 0x29) isp0: Retaining Loop ID 0x7d for Target 125 (Port 0x1) isp0: Target 38 (Loop 0x26) Port ID 0xaa (role Target) Departed Port WWN 0x2100002037082546 Node WWN 0x2000002037082546 (pass7:isp0:0:38:0): lost device (pass7:isp0:0:38:0): removing device entry (da7:isp0:0:38:0): lost device (da7:isp0:0:38:0): removing device entry That seems to work well... note it removes all periphs that referred to that device. -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 14: 7:28 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from probity.mcc.ac.uk (probity.mcc.ac.uk [130.88.200.94]) by hub.freebsd.org (Postfix) with ESMTP id DE88E37B401 for ; Fri, 22 Jun 2001 14:07:24 -0700 (PDT) (envelope-from jcm@freebsd-uk.eu.org) Received: from dogma.freebsd-uk.eu.org ([130.88.200.97] ident=root) by probity.mcc.ac.uk with esmtp (Exim 2.05 #7) id 15DY9B-000KPh-00; Fri, 22 Jun 2001 22:07:21 +0100 Received: (from jcm@localhost) by dogma.freebsd-uk.eu.org (8.11.3/8.11.1) id f5ML7Kh64543; Fri, 22 Jun 2001 22:07:20 +0100 (BST) (envelope-from jcm) Date: Fri, 22 Jun 2001 22:07:19 +0100 From: j mckitrick To: Matthew Jacob Cc: scsi@freebsd.org Subject: Re: how to invalidate scsi connection to driver module Message-ID: <20010622220719.A64495@dogma.freebsd-uk.eu.org> References: <20010622123714.R20923-100000@wonky.feral.com> <20010622135159.B20923-100000@wonky.feral.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <20010622135159.B20923-100000@wonky.feral.com>; from mjacob@feral.com on Fri, Jun 22, 2001 at 01:52:57PM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org So the function you call is xpt_bus_deregister() ? On Fri, Jun 22, 2001 at 01:52:57PM -0700, Matthew Jacob wrote: | | To follow up on this- Since I've been fooling around with the | CAM_NEW_TRAN_CODE and Fibre Channel, I have the SIM call xpt_async | with AC_LOST_DEVICE when a device goes away, so I had: | | isp0: Retaining Loop ID 0x20 for Target 32 (Port 0xb2) | isp0: Retaining Loop ID 0x21 for Target 33 (Port 0xb1) | isp0: Retaining Loop ID 0x22 for Target 34 (Port 0xae) | isp0: Retaining Loop ID 0x23 for Target 35 (Port 0xad) | isp0: Retaining Loop ID 0x24 for Target 36 (Port 0xac) | isp0: Retaining Loop ID 0x25 for Target 37 (Port 0xab) | isp0: Retaining Loop ID 0x2d for Target 45 (Port 0x9e) | isp0: Retaining Loop ID 0x3d for Target 61 (Port 0x75) | isp0: Retaining Loop ID 0x6d for Target 109 (Port 0x29) | isp0: Retaining Loop ID 0x7d for Target 125 (Port 0x1) | isp0: Target 38 (Loop 0x26) Port ID 0xaa (role Target) Departed | Port WWN 0x2100002037082546 | Node WWN 0x2000002037082546 | (pass7:isp0:0:38:0): lost device | (pass7:isp0:0:38:0): removing device entry | (da7:isp0:0:38:0): lost device | (da7:isp0:0:38:0): removing device entry | | That seems to work well... note it removes all periphs that referred to that | device. | | -matt | | Jonathon -- Microsoft complaining about the source license used by Linux is like the event horizon calling the kettle black. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 14: 7:37 2001 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 69AA237B406 for ; Fri, 22 Jun 2001 14:07:33 -0700 (PDT) (envelope-from ken@panzer.kdm.org) Received: (from ken@localhost) by panzer.kdm.org (8.9.3/8.9.1) id PAA57806; Fri, 22 Jun 2001 15:06:07 -0600 (MDT) (envelope-from ken) Date: Fri, 22 Jun 2001 15:06:07 -0600 From: "Kenneth D. Merry" To: Matthew Jacob Cc: j mckitrick , scsi@FreeBSD.ORG Subject: Re: how to invalidate scsi connection to driver module Message-ID: <20010622150607.A57782@panzer.kdm.org> References: <20010622123714.R20923-100000@wonky.feral.com> <20010622135159.B20923-100000@wonky.feral.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2i In-Reply-To: <20010622135159.B20923-100000@wonky.feral.com>; from mjacob@feral.com on Fri, Jun 22, 2001 at 01:52:57PM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Fri, Jun 22, 2001 at 13:52:57 -0700, Matthew Jacob wrote: > > To follow up on this- Since I've been fooling around with the > CAM_NEW_TRAN_CODE and Fibre Channel, I have the SIM call xpt_async > with AC_LOST_DEVICE when a device goes away, so I had: I believe that is the right way to do it. > isp0: Retaining Loop ID 0x20 for Target 32 (Port 0xb2) > isp0: Retaining Loop ID 0x21 for Target 33 (Port 0xb1) > isp0: Retaining Loop ID 0x22 for Target 34 (Port 0xae) > isp0: Retaining Loop ID 0x23 for Target 35 (Port 0xad) > isp0: Retaining Loop ID 0x24 for Target 36 (Port 0xac) > isp0: Retaining Loop ID 0x25 for Target 37 (Port 0xab) > isp0: Retaining Loop ID 0x2d for Target 45 (Port 0x9e) > isp0: Retaining Loop ID 0x3d for Target 61 (Port 0x75) > isp0: Retaining Loop ID 0x6d for Target 109 (Port 0x29) > isp0: Retaining Loop ID 0x7d for Target 125 (Port 0x1) > isp0: Target 38 (Loop 0x26) Port ID 0xaa (role Target) Departed > Port WWN 0x2100002037082546 > Node WWN 0x2000002037082546 > (pass7:isp0:0:38:0): lost device > (pass7:isp0:0:38:0): removing device entry > (da7:isp0:0:38:0): lost device > (da7:isp0:0:38:0): removing device entry > > That seems to work well... note it removes all periphs that referred to that > device. Yep. 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 Jun 22 14:10:34 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 4178C37B401 for ; Fri, 22 Jun 2001 14:10:32 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5ML7rg75210; Fri, 22 Jun 2001 14:07:54 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 14:07:47 -0700 (PDT) From: Matthew Jacob Reply-To: To: "Kenneth D. Merry" Cc: j mckitrick , Subject: Re: how to invalidate scsi connection to driver module In-Reply-To: <20010622150607.A57782@panzer.kdm.org> Message-ID: <20010622140658.P20923-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > > That seems to work well... note it removes all periphs that referred to that > > device. > > Yep. I decided to do things more riskily, and the AC_FOUND_DEVICE even works w/o tremendous confusion during probe time- this means that while we're booting and probing, the wad of device arrivals get sorted out as well. Not bad. -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 14:13: 3 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 9900837B401 for ; Fri, 22 Jun 2001 14:12:59 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5MLAQg75223; Fri, 22 Jun 2001 14:10:26 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 14:10:20 -0700 (PDT) From: Matthew Jacob Reply-To: To: j mckitrick Cc: Subject: Re: how to invalidate scsi connection to driver module In-Reply-To: <20010622220719.A64495@dogma.freebsd-uk.eu.org> Message-ID: <20010622140800.F20923-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Fri, 22 Jun 2001, j mckitrick wrote: > > So the function you call is xpt_bus_deregister() ? I did this: if (xpt_create_path(&tmppath, NULL, cam_sim_path(isp->isp_sim), (target_id_t)tgt, CAM_LUN_WILDCARD) != CAM_REQ_CMP) { break; } if (lp->valid && (lp->roles & (SVC3_INI_ROLE >> SVC3_ROLE_SHIFT))) { ISPLOCK_2_CAMLOCK(isp); xpt_async(AC_FOUND_DEVICE, tmppath, NULL); } else { ISPLOCK_2_CAMLOCK(isp); xpt_async(AC_LOST_DEVICE, tmppath, NULL); } CAMLOCK_2_ISPLOCK(isp); xpt_free_path(tmppath); This does a LOST or FOUND device async for that PATH (bus), that Target, and all LUNS. Note that xpt_bus_deregister has: status = xpt_compile_path(&bus_path, NULL, pathid, CAM_TARGET_WILDCARD, CAM_LUN_WILDCARD); ... xpt_async(AC_LOST_DEVICE, &bus_path, NULL); which then does lost devices on all targets and luns on that bus. -matt > > > > On Fri, Jun 22, 2001 at 01:52:57PM -0700, Matthew Jacob wrote: > | > | To follow up on this- Since I've been fooling around with the > | CAM_NEW_TRAN_CODE and Fibre Channel, I have the SIM call xpt_async > | with AC_LOST_DEVICE when a device goes away, so I had: > | > | isp0: Retaining Loop ID 0x20 for Target 32 (Port 0xb2) > | isp0: Retaining Loop ID 0x21 for Target 33 (Port 0xb1) > | isp0: Retaining Loop ID 0x22 for Target 34 (Port 0xae) > | isp0: Retaining Loop ID 0x23 for Target 35 (Port 0xad) > | isp0: Retaining Loop ID 0x24 for Target 36 (Port 0xac) > | isp0: Retaining Loop ID 0x25 for Target 37 (Port 0xab) > | isp0: Retaining Loop ID 0x2d for Target 45 (Port 0x9e) > | isp0: Retaining Loop ID 0x3d for Target 61 (Port 0x75) > | isp0: Retaining Loop ID 0x6d for Target 109 (Port 0x29) > | isp0: Retaining Loop ID 0x7d for Target 125 (Port 0x1) > | isp0: Target 38 (Loop 0x26) Port ID 0xaa (role Target) Departed > | Port WWN 0x2100002037082546 > | Node WWN 0x2000002037082546 > | (pass7:isp0:0:38:0): lost device > | (pass7:isp0:0:38:0): removing device entry > | (da7:isp0:0:38:0): lost device > | (da7:isp0:0:38:0): removing device entry > | > | That seems to work well... note it removes all periphs that referred to that > | device. > | > | -matt > | > | > > > Jonathon > -- > Microsoft complaining about the source license used by > Linux is like the event horizon calling the kettle black. > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 14:35:28 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from serenity.mcc.ac.uk (serenity.mcc.ac.uk [130.88.200.93]) by hub.freebsd.org (Postfix) with ESMTP id BCD1B37B401 for ; Fri, 22 Jun 2001 14:35:25 -0700 (PDT) (envelope-from jcm@freebsd-uk.eu.org) Received: from dogma.freebsd-uk.eu.org ([130.88.200.97] ident=root) by serenity.mcc.ac.uk with esmtp (Exim 2.05 #6) id 15DYaK-0003Ig-00; Fri, 22 Jun 2001 22:35:25 +0100 Received: (from jcm@localhost) by dogma.freebsd-uk.eu.org (8.11.3/8.11.1) id f5MLZOa65215; Fri, 22 Jun 2001 22:35:24 +0100 (BST) (envelope-from jcm) Date: Fri, 22 Jun 2001 22:35:24 +0100 From: j mckitrick To: Matthew Jacob Cc: scsi@freebsd.org Subject: Re: how to invalidate scsi connection to driver module Message-ID: <20010622223523.D64495@dogma.freebsd-uk.eu.org> References: <20010622220719.A64495@dogma.freebsd-uk.eu.org> <20010622140800.F20923-100000@wonky.feral.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <20010622140800.F20923-100000@wonky.feral.com>; from mjacob@feral.com on Fri, Jun 22, 2001 at 02:10:20PM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org | xpt_async(AC_FOUND_DEVICE, tmppath, NULL); Is there any way i can get a cam_path from a cam_sim? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 14:40:47 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id 1827237B401 for ; Fri, 22 Jun 2001 14:40:46 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5MLcDg75299; Fri, 22 Jun 2001 14:38:13 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 14:38:08 -0700 (PDT) From: Matthew Jacob Reply-To: To: j mckitrick Cc: Subject: Re: how to invalidate scsi connection to driver module In-Reply-To: <20010622223523.D64495@dogma.freebsd-uk.eu.org> Message-ID: <20010622143758.D20923-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org cam_sim_path? On Fri, 22 Jun 2001, j mckitrick wrote: > | xpt_async(AC_FOUND_DEVICE, tmppath, NULL); > > Is there any way i can get a cam_path from a cam_sim? > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 14:42:36 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from probity.mcc.ac.uk (probity.mcc.ac.uk [130.88.200.94]) by hub.freebsd.org (Postfix) with ESMTP id 5193537B407 for ; Fri, 22 Jun 2001 14:42:34 -0700 (PDT) (envelope-from jcm@freebsd-uk.eu.org) Received: from dogma.freebsd-uk.eu.org ([130.88.200.97] ident=root) by probity.mcc.ac.uk with esmtp (Exim 2.05 #7) id 15DYhE-000Krs-00; Fri, 22 Jun 2001 22:42:33 +0100 Received: (from jcm@localhost) by dogma.freebsd-uk.eu.org (8.11.3/8.11.1) id f5MLgWo65368; Fri, 22 Jun 2001 22:42:32 +0100 (BST) (envelope-from jcm) Date: Fri, 22 Jun 2001 22:42:32 +0100 From: j mckitrick To: Matthew Jacob Cc: scsi@freebsd.org Subject: Re: how to invalidate scsi connection to driver module Message-ID: <20010622224232.E64495@dogma.freebsd-uk.eu.org> References: <20010622223523.D64495@dogma.freebsd-uk.eu.org> <20010622143758.D20923-100000@wonky.feral.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <20010622143758.D20923-100000@wonky.feral.com>; from mjacob@feral.com on Fri, Jun 22, 2001 at 02:38:08PM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Fri, Jun 22, 2001 at 02:38:08PM -0700, Matthew Jacob wrote: | | cam_sim_path? thanks.... "All too easy...." :-) Jonathon -- Microsoft complaining about the source license used by Linux is like the event horizon calling the kettle black. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 16:10:41 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from probity.mcc.ac.uk (probity.mcc.ac.uk [130.88.200.94]) by hub.freebsd.org (Postfix) with ESMTP id 3CDA737B406 for ; Fri, 22 Jun 2001 16:10:39 -0700 (PDT) (envelope-from jcm@freebsd-uk.eu.org) Received: from dogma.freebsd-uk.eu.org ([130.88.200.97] ident=root) by probity.mcc.ac.uk with esmtp (Exim 2.05 #7) id 15Da4T-000LiF-00; Sat, 23 Jun 2001 00:10:37 +0100 Received: (from jcm@localhost) by dogma.freebsd-uk.eu.org (8.11.3/8.11.1) id f5MNAa967270; Sat, 23 Jun 2001 00:10:36 +0100 (BST) (envelope-from jcm) Date: Sat, 23 Jun 2001 00:10:36 +0100 From: j mckitrick To: Matthew Jacob Cc: scsi@freebsd.org Subject: Re: how to invalidate scsi connection to driver module Message-ID: <20010623001035.A67153@dogma.freebsd-uk.eu.org> References: <20010622220719.A64495@dogma.freebsd-uk.eu.org> <20010622140800.F20923-100000@wonky.feral.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <20010622140800.F20923-100000@wonky.feral.com>; from mjacob@feral.com on Fri, Jun 22, 2001 at 02:10:20PM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org here is what i did so far: xpt_async(AC_LOST_DEVICE, vpo->path, NULL); xpt_free_path(vpo->path); xpt_bus_deregister(cam_sim_path(vpo->sim)); cam_sim_free(vpo->sim, TRUE); It pretty much works, meaning the module unloads and can be reloaded without crashing. However, something odd happens. When you first attach, as expected it says "vpo0... da0..." which is fine. But on each subsequent reattach, it allocates da0 and da1... then da0, da1, and da2. And so on. When i detach, it detaches the entire list. Everything still works, but it should *not* attach /n/ devices for every /nth/ attach. How do i force only one instance of the drive for each attach? Jonathon -- Microsoft complaining about the source license used by Linux is like the event horizon calling the kettle black. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 16:15: 9 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id A01F537B401 for ; Fri, 22 Jun 2001 16:15:05 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from wonky.feral.com (wonky.feral.com [192.67.166.7]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f5MNCWg75497; Fri, 22 Jun 2001 16:12:32 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 22 Jun 2001 16:12:26 -0700 (PDT) From: Matthew Jacob Reply-To: To: j mckitrick Cc: Subject: Re: how to invalidate scsi connection to driver module In-Reply-To: <20010623001035.A67153@dogma.freebsd-uk.eu.org> Message-ID: <20010622161135.A21527-100000@wonky.feral.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Sat, 23 Jun 2001, j mckitrick wrote: > > here is what i did so far: > > xpt_async(AC_LOST_DEVICE, vpo->path, NULL); > xpt_free_path(vpo->path); > xpt_bus_deregister(cam_sim_path(vpo->sim)); > cam_sim_free(vpo->sim, TRUE); > > It pretty much works, meaning the module unloads and can be reloaded without > crashing. However, something odd happens. When you first attach, as > expected it says "vpo0... da0..." which is fine. But on each subsequent > reattach, it allocates da0 and da1... then da0, da1, and da2. And so on. > When i detach, it detaches the entire list. Everything still works, but it > should *not* attach /n/ devices for every /nth/ attach. > > How do i force only one instance of the drive for each attach? I'm not quite following what happens, but if you are deregistering a bus, all of the devices for that bus go away. I need a few more details. -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message From owner-freebsd-scsi Fri Jun 22 16:19:26 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from probity.mcc.ac.uk (probity.mcc.ac.uk [130.88.200.94]) by hub.freebsd.org (Postfix) with ESMTP id 26F2537B401 for ; Fri, 22 Jun 2001 16:19:24 -0700 (PDT) (envelope-from jcm@freebsd-uk.eu.org) Received: from dogma.freebsd-uk.eu.org ([130.88.200.97] ident=root) by probity.mcc.ac.uk with esmtp (Exim 2.05 #7) id 15DaCx-000LuC-00; Sat, 23 Jun 2001 00:19:23 +0100 Received: (from jcm@localhost) by dogma.freebsd-uk.eu.org (8.11.3/8.11.1) id f5MNJNv67484; Sat, 23 Jun 2001 00:19:23 +0100 (BST) (envelope-from jcm) Date: Sat, 23 Jun 2001 00:19:22 +0100 From: j mckitrick To: Matthew Jacob Cc: scsi@freebsd.org Subject: Re: how to invalidate scsi connection to driver module Message-ID: <20010623001922.B67153@dogma.freebsd-uk.eu.org> References: <20010623001035.A67153@dogma.freebsd-uk.eu.org> <20010622161135.A21527-100000@wonky.feral.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <20010622161135.A21527-100000@wonky.feral.com>; from mjacob@feral.com on Fri, Jun 22, 2001 at 04:12:26PM -0700 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org | I'm not quite following what happens, but if you are deregistering a bus, all | of the devices for that bus go away. I need a few more details. The device is vpo, the parallel port zip drive. It is already a loadable module, i am making it unloadable. But the device unit number (or rather, the number of units created) is increasing by one every time i load. When i detach, they all go away. This is happening in probe/attach/identify or even earlier, because after each unit is connected, it just keeps finding one more than last time, and the registering it with cam. Jonathon -- Microsoft complaining about the source license used by Linux is like the event horizon calling the kettle black. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message