From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christof Schmitt Subject: Re: [patch 02/10] zfcp: Remove SCSI device when removing unit Date: Wed, 28 Jul 2010 10:35:54 +0200 Message-ID: <20100728083553.GA3668@schmichrtp.mainz.de.ibm.com> References: <20100716133733.621672000@de.ibm.com> <20100716134300.006377000@de.ibm.com> <1280263075.2833.404.camel@mulgrave.site> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mtagate2.uk.ibm.com ([194.196.100.162]:44092 "EHLO mtagate2.uk.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751459Ab0G1If4 (ORCPT ); Wed, 28 Jul 2010 04:35:56 -0400 Content-Disposition: inline In-Reply-To: <1280263075.2833.404.camel@mulgrave.site> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: James Bottomley Cc: linux-scsi@vger.kernel.org, linux-s390@vger.kernel.org, schwidefsky@de.ibm.com, heiko.carstens@de.ibm.com On Tue, Jul 27, 2010 at 03:37:55PM -0500, James Bottomley wrote: > On Fri, 2010-07-16 at 15:37 +0200, Christof Schmitt wrote: > > plain text document attachment (705-zfcp-unit-removal.diff) > > From: Christof Schmitt > > > > Configuring a LUN in zfcp, also creates a SCSI device. For > > consistency, it makes sense to remove the SCSI device when the LUN is > > deconfigured. Replace the flush_work with the call to > > scsi_remove_device: scsi_remove_device also takes the scan_mutex that > > synchronizes itself with any long running device discovery. > > This one's rejecting: > > jejb@mulgrave:~/git/scsi-misc-2.6> patch -p1 < ~/tmp/tmp/02 > patching file drivers/s390/scsi/zfcp_def.h > patching file drivers/s390/scsi/zfcp_scsi.c > Hunk #1 FAILED at 564. > 1 out of 1 hunk FAILED -- saving rejects to file > drivers/s390/scsi/zfcp_scsi.c.rej > > It looks like a missing patch somewhere. > > Could you check the current state of scsi-misc-2.6 (it includes > scsi-rc-fixes-2.6) and see if I've lost something? scsi-misc and scsi-rc-fixes have the three fixes i sent for 2.6.35. I sent them again with six additional fixes here: http://marc.info/?l=linux-scsi&m=127857607617662&w=2 The six additional patches are missing and this seems to cause the reject. On top of scsi-misc, these patches apply cleanly in this order: >>From http://marc.info/?l=linux-scsi&m=127857607617662&w=2 zfcp: Do not unblock rport from REOPEN_PORT_FORCED zfcp: Do not try "forced close" when port is already closed zfcp: Register SCSI devices after successful fc_remote_port_add zfcp: Use forced_reopen in terminate_rport_io callback zfcp: Fail erp after timeout zfcp: Fix retry after failed "open port" erp action >>From http://marc.info/?l=linux-scsi&m=127928780100336&w=2 zfcp: Use memdup_user and kstrdup zfcp: Remove SCSI device when removing unit zfcp: Use correct width for timer_interval field zfcp: Cleanup function parameters for sbal value. zfcp: Cleanup QDIO attachment and improve processing. zfcp: Post events through FC transport class zfcp: Prevent access on uninitialized memory. zfcp: Enable data division support for FCP devices zfcp: Introduce experimental support for DIF/DIX zfcp: Trigger logging in the FCP channel on qdio error conditions I can also resend them in one series if this makes things easier. Christof