From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: [PATCH] SCSI and FibreChannel Hotswap for linux 2.5.44-bk2 Date: Wed, 30 Oct 2002 15:42:21 -0500 Sender: linux-kernel-owner@vger.kernel.org Message-ID: <200210302042.g9UKgLS02419@localhost.localdomain> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: In-Reply-To: Message from Steven Dake of "Wed, 30 Oct 2002 11:54:47 MST." <3DC02AF7.6020209@mvista.com> To: Steven Dake Cc: torvalds@transmeta.com, linux-kernel@vger.kernel.org, linux-scsi@vger.kernel.org List-Id: linux-scsi@vger.kernel.org sdake@mvista.com said: > This patch has been reviewed by Alan Cox, Greg KH, Christoph Hellwig, > Patrick Mansfield, Rob Landly, Jeff Garzik, Scott Murray, James > Bottomley, Mike Anderson, Randy Dunlap and Patrick Mochel. Well, I reviewed it but my though was that it should be built on the emerging hotplug infrastructure. I'm currently trying to move things like this into the user layer, so from a design principle I don't want to have 90% using the user space stuff and 10% using its in-kernel equivalent because that means we have two mechanisms to maintain, thus doubling the work. The problem you had with this was the 10ms requirement from removal notification to removal completion. Several people have already suggested that if that really is a hard and fast requirement, then you could simply treat the removal as a surprise removal rather than a planned one and work on fixing our surprise removal problems instead. The bottom line is that I'm not convinced this can't be done using the existing infrastructure or a generic enhancement to it. James