From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mike Christie Subject: Re: [PATCH 2/2] blk request timeout handler: Date: Mon, 09 Jul 2007 10:56:48 -0500 Message-ID: <46925AC0.4020102@cs.wisc.edu> References: Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from sabe.cs.wisc.edu ([128.105.6.20]:33632 "EHLO sabe.cs.wisc.edu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754077AbXGIP5D (ORCPT ); Mon, 9 Jul 2007 11:57:03 -0400 In-Reply-To: Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: "Salyzyn, Mark" Cc: linux-scsi@vger.kernel.org, jens.axboe@oracle.com I dropped Jen's old suse address and added his oracle one. Salyzyn, Mark wrote: > What ever happened to the following (admittedly snipped up) patch from a > year ago, was there a reason this initiative was dropped? If so, please > state what I missed? > I have not worked on it because I got busy with iscsi again. I was trying to rework the patches, so that the scsi-ml would not have to do so much digging in the timers. In the end I wanted it so the block layer started the timer when the command either was queued up or when it was sent to the LLD (I was not sure which was better), and the timer was stopped when the command was completed or when the LLD requeued it. The starting of the timer is easy, but with the stopping of it, scsi-ml needed to do some things like stop the timer manually then call a block layer requeue function and I wanted the block layer requeue function to stop the timer for scsi-ml so the LLD would not have to worry about it at all. I also got stuck working out the details for situations like when we get into aborts and error handling and support for subsystems like DM or MD doing timers and stacking them on top of SCSI timers or having MD/DM timers calling into the SCSI timeout handlers.