Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 6 Nov 2001 15:08:33 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 6 Nov 2001 15:08:23 -0500 Received: from web12301.mail.yahoo.com ([216.136.173.99]:56735 "HELO web12301.mail.yahoo.com") by vger.kernel.org with SMTP id ; Tue, 6 Nov 2001 15:08:16 -0500 Message-ID: <20011106200815.50754.qmail@web12301.mail.yahoo.com> Date: Tue, 6 Nov 2001 12:08:15 -0800 (PST) From: Stephen Cameron Subject: Re: Mylex/Compaq RAID controller placement in config To: linux-kernel@vger.kernel.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Venkatesh Ramamurthy wrote: > > I know it might seem silly, but as to make things clearer for most > > users/admins, wouldn't it be better to just call them SCSI controllers, as > > they all indeed connect SCSI drives to the host? > > Eventhough they connect SCSI Drives, the fact is that they are totally > different beast from the OS/driver perspective. Actually, at least in Alan Cox's tree, the cciss driver is BOTH a SCSI driver and a block driver simultaneously, or can be configured that way anyhow. That change was intruduced for the sole purpose of accessing SCSI tape drives via array controllers. (nevermind why you'd want to do that, suffice it to say, some folks want this ability.) I would consider it mostly a block driver, even when SCSI tape drive support is compiled in. In any case, the "drives" (logical volumes) which are presented by the hardware/firmware and eventually by the driver to the OS do not in general correspond 1-to-1 with physical SCSI drives. Even the SCSI busses presented by the cciss driver with SCSI tape support do not necessarily correspond with physical SCSI busses, so considering these controllers "SCSI controllers" is not really accurate, though typically they _contain_ what you can think of as one or more SCSI controllers as part of the back-end, which the firmware uses to control SCSI devices,. Though, in some cases it could be a fibre-channel back-end instead of SCSI. In any case, the nature of the back-end is irrelevant to the driver for purposes of doing i/o to logical volumes. Incidentally, last time I checked, Linus' tree contained the documentation portion of the patch which allowed a cciss controller to access SCSI tape drives (Documentation/cciss.txt) but none of the actual code that implemented that functionality... :-/ oh well. -- steve (aka steve.cameron@compaq.com) __________________________________________________ Do You Yahoo!? Find a job, post your resume. http://careers.yahoo.com - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/