From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mike Anderson Subject: Re: 2.4.19 SCSI error handing, door locking, etc Date: Fri, 4 Oct 2002 12:41:36 -0700 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20021004194136.GC9544@beaverton.ibm.com> References: <20020926012357.B8733@flint.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20020926012357.B8733@flint.arm.linux.org.uk> List-Id: linux-scsi@vger.kernel.org To: Russell King Cc: linux-scsi@vger.kernel.org Russell, Sorry it took so long to port forward and test these changes. I will be posting the set of patches to the list right after this mail. Here is an overview the forward port in relation to your patches. 01-scsi-cmd-retry-1.diff - Brought forward with minor arg/ variable renaming. 01-scsi-cmd-retry-2.diff - Brought forward though my original implementation of how retries are controlled is different than your code. 01-scsi-cmd-retry-3.diff - I need to think on this. I currently have 1 retry for TUR and non for REQ sense. If a device is not responding the way we want the probability of sending multiple commands in a short time period especially TURs is most likely not going to result in what we want. Though your delay mentioned below is something I currently do not have. 02-scsi-cmd-report.diff - I brought this forward, but it bothers me that the data in a cmd is incorrect as the command should be restored after it has been touched by error handling. 03-scsi-restart-ops.diff - Brought forward. In 2.5 we should probably consider removing all these checks and just call the request_fn so that we do not have the checks spread all over which leads to them becoming out of sync or implemented incorrectly. 04-scsi-door-lock-1.diff - Brought forward. I did need to increase the bus reset settle time from 5 seconds to 10 to get the door lock to re-lock the door. 04-scsi-door-lock-2.diff - Brought forward. In 2.5 we do not have the old error handler so I removed the door lock completely from the the scsi_request_fn function. 05-scsi-unjam-clean-1.diff 05-scsi-unjam-clean-2.diff 05-scsi-unjam-clean-3.diff 05-scsi-unjam-clean-4.diff 05-scsi-unjam-clean-5.diff 05-scsi-unjam-clean-6.diff 05-scsi-unjam-clean-8.diff - The previous patch I created for 2.5 scsi_error cleanup I believe covers similar changes as yours for 2.4. Patch is now part of 2.5.40. 05-scsi-unjam-clean-7.diff - I have not covered all these yet. One change that made a difference is how the tur retry is done and I also increased the bus_reset settle delay from 5 to 10. I will try to ensure that the error policy you want is covered in 2.5. 2.) Added scsi_set_medium_removal to scsi_syms.c and scsi_ioctl.h -andmike -- Michael Anderson andmike@us.ibm.com