From mboxrd@z Thu Jan 1 00:00:00 1970 From: bugzilla-daemon@bugzilla.kernel.org Subject: [Bug 16547] mptscsih: ioc0: attempting task abort, raid array LUNs not detected properly on some boots Date: Tue, 21 Sep 2010 08:12:10 GMT Message-ID: <201009210812.o8L8CA7s002044@demeter1.kernel.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Return-path: Received: from demeter.kernel.org ([140.211.167.39]:41127 "EHLO demeter1.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755350Ab0IUIML (ORCPT ); Tue, 21 Sep 2010 04:12:11 -0400 Received: from demeter1.kernel.org (localhost.localdomain [127.0.0.1]) by demeter1.kernel.org (8.14.4/8.14.3) with ESMTP id o8L8CAXP002045 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Tue, 21 Sep 2010 08:12:10 GMT In-Reply-To: Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: linux-scsi@vger.kernel.org https://bugzilla.kernel.org/show_bug.cgi?id=16547 --- Comment #7 from Martin Steigerwald 2010-09-21 08:12:01 --- (In reply to comment #5) > (In reply to comment #3) > > Some additional information on the MPT driver version and controller: > > > > backend01:~# grep -r "" /proc/mpt/* > > /proc/mpt/ioc0/summary:ioc0: LSIFC949E A1, FwRev=01030e00h, Ports=1, MaxQ=1023, > > LanAddr=00:06:[...], IRQ=33 > > /proc/mpt/ioc0/info:ioc0: > > /proc/mpt/ioc0/info: ProductID = 0x1005 (LSIFC949E A1) > > /proc/mpt/ioc0/info: FWVersion = 0x01030e00 (fw_size=190556) [...] > Your bug is completely different issue. Whatever you are point to redhat > bugzilla is with respect to SAS controller. I thought it might be related nevertheless. I don't know the inner structure of the MPT driver. It also sounded similar, cause in that bug report there is also the mention that it worked with 2.6.26, but I AFAIR not with 2.6.27. Maybe its a general change in the SCSI layer that triggers the issue. > In your case it is FC controller. Yes, I know. > You have mentioned that > "Latest kernel known to work: 2.6.26 from Debian Backports" > > Can you provide me driver version where things are working fine. Here is the version from a 2.6.26 lenny kernel, which should be the one that has been backported to Etch: pasta:~# modinfo /lib/modules/2.6.26-2-amd64/kernel/drivers/message/fusion/mptfc.ko filename: /lib/modules/2.6.26-2-amd64/kernel/drivers/message/fusion/mptfc.ko version: 3.04.06 license: GPL description: Fusion MPT FC Host driver author: LSI Corporation srcversion: F3D99FE0544BDDD1455BAAA alias: pci:v00001657d00000646sv*sd*bc*sc*i* alias: pci:v00001000d00000646sv*sd*bc*sc*i* alias: pci:v00001000d00000640sv*sd*bc*sc*i* alias: pci:v00001000d00000642sv*sd*bc*sc*i* alias: pci:v00001000d00000626sv*sd*bc*sc*i* alias: pci:v00001000d00000628sv*sd*bc*sc*i* alias: pci:v00001000d00000622sv*sd*bc*sc*i* alias: pci:v00001000d00000624sv*sd*bc*sc*i* alias: pci:v00001000d00000621sv*sd*bc*sc*i* depends: mptscsih,scsi_transport_fc,scsi_mod,mptbase vermagic: 2.6.26-2-amd64 SMP mod_unload modversions parm: mptfc_dev_loss_tmo: Initial time the driver programs the transport to wait for an rport to return following a device loss event. Default=60. (int) parm: max_lun: max lun, default=16895 (int) The 2.6.32 kernel, where we see described issues has: backend01:~# modinfo mptfc filename: /lib/modules/2.6.32-bpo.5-amd64/kernel/drivers/message/fusion/mptfc.ko version: 3.04.12 license: GPL description: Fusion MPT FC Host driver author: LSI Corporation srcversion: 92E350C096B75A9714B8B0E alias: pci:v00001657d00000646sv*sd*bc*sc*i* alias: pci:v00001000d00000646sv*sd*bc*sc*i* alias: pci:v00001000d00000640sv*sd*bc*sc*i* alias: pci:v00001000d00000642sv*sd*bc*sc*i* alias: pci:v00001000d00000626sv*sd*bc*sc*i* alias: pci:v00001000d00000628sv*sd*bc*sc*i* alias: pci:v00001000d00000622sv*sd*bc*sc*i* alias: pci:v00001000d00000624sv*sd*bc*sc*i* alias: pci:v00001000d00000621sv*sd*bc*sc*i* depends: mptscsih,mptbase,scsi_transport_fc,scsi_mod vermagic: 2.6.32-bpo.5-amd64 SMP mod_unload modversions parm: mptfc_dev_loss_tmo: Initial time the driver programs the transport to wait for an rport to return following a device loss event. Default=60. (int) parm: max_lun: max lun, default=16895 (int) backend01:~# > In case of > some working kernel is there, I would like to simply upgrade MPTFUSION driver > (do not upgrade a whole kernel). This way I would like to change only one > component of the system at a time... Well the old 2.6.26 kernel worked. But actually it does not boot on the new servers, cause the old version ata_piix does not talk to the newer onboard SATA controller. Thus it would be required to use a newer ata_piix and a newer MPT FUSION FC driver with 2.6.26 kernel. I don't know whether thats feasible. Its a production machine and I need to be careful with testing. I can only test with agreement of the customer. But for a defined test case it might be workable. Would it be as easy as to replace the directories with the driver source with a newer version? From 2.6.26 to 2.6.32 is quite a step. > This will help to understand where things are broken. I understand. > FYI, > MPTFC drive is highly in mentionation mode. There are very very minimal changes > happened to MPTFC driver since 2008. > > Last change went to upstream for MPTFC is > > http://git.kernel.org/?p=linux/kernel/git/jejb/scsi-misc-2.6.git;a=commit;h=03cb3829e0e5650518ce37e2b4420a35e034dc9e I don't think that commit has landed in 2.6.32, since Linus released it on 3rd december 2009. It also does not seem to be in one of the stable patches: ms@mango:~/Linux/Kernel/Mainline> ls ChangeLog-2.6.32* ChangeLog-2.6.32 ChangeLog-2.6.32.16 ChangeLog-2.6.32.3 ChangeLog-2.6.32.1 ChangeLog-2.6.32.17 ChangeLog-2.6.32.4 ChangeLog-2.6.32.10 ChangeLog-2.6.32.18 ChangeLog-2.6.32.5 ChangeLog-2.6.32.11 ChangeLog-2.6.32.19 ChangeLog-2.6.32.6 ChangeLog-2.6.32.12 ChangeLog-2.6.32.2 ChangeLog-2.6.32.7 ChangeLog-2.6.32.13 ChangeLog-2.6.32.20 ChangeLog-2.6.32.8 ChangeLog-2.6.32.14 ChangeLog-2.6.32.21 ChangeLog-2.6.32.9 ChangeLog-2.6.32.15 ChangeLog-2.6.32.22 ms@mango:~/Linux/Kernel/Mainline> grep 03cb3829e0e5650518ce37e2b4420a35e034dc9e ChangeLog-2.6.32* ms@mango:~/Linux/Kernel/Mainline#1> Thanks, Martin -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. You are watching the assignee of the bug.