From mboxrd@z Thu Jan 1 00:00:00 1970 From: Willem Riede Subject: Re: Badness in scsi_single_lun_run at /root/scsi/scsi_lib.c:344 Date: Wed, 28 Jan 2004 13:33:23 -0500 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20040128183323.GG23308@serve.riede.org> References: <20040127003244.GM23308@serve.riede.org> <20040128033041.GY23308@serve.riede.org> <20040128084142.A4835@beaverton.ibm.com> <20040128173609.GE23308@serve.riede.org> <20040128095341.A5242@beaverton.ibm.com> Reply-To: wrlk@riede.org Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Return-path: Received: from rwcrmhc11.comcast.net ([204.127.198.35]:53914 "EHLO rwcrmhc11.comcast.net") by vger.kernel.org with ESMTP id S266131AbUA1Sd0 (ORCPT ); Wed, 28 Jan 2004 13:33:26 -0500 Content-Disposition: inline In-Reply-To: <20040128095341.A5242@beaverton.ibm.com> (from patmans@us.ibm.com on Wed, Jan 28, 2004 at 12:53:41 -0500) List-Id: linux-scsi@vger.kernel.org To: Patrick Mansfield Cc: linux-scsi@vger.kernel.org On 2004.01.28 12:53, Patrick Mansfield wrote: > On Wed, Jan 28, 2004 at 12:36:09PM -0500, Willem Riede wrote: > > On 2004.01.28 11:41, Patrick Mansfield wrote: > > > > Do you know if the single lun code is for performance, or because of > > > hardware limitations - that is we don't want a disc change between IO, or > > > does the device just fail? > > > > In the case of the NEC PD/CD, it is a hardware limitation. You can only > > have one disc inserted at a time, either CD or PD. Mixing IO will fail > > for the absent media type. > > BTW what is PD? PD stands for PowerDrive, it uses 650MB (CD size) cartridges that Panasonic invented before there were writable CDs. I assume they're a type of magneto- optical device. > So there is one physical drive that can hold one disc, and it can only use > the LUN associated with the disc that is currently in the drive; and reading > or accessing the wrong LUN will always fail? > > We don't need single_lun code for that, we can just let the IO fail, right? That would be no problem. All the single-lun code does is delay the inevitable: if I address the /dev/scdX which corresponds to the /dev/sdY that's the PD disk when it's loaded, it _will_ fail (after all IO to the PD's lun finishes). > I assume there are other single_lun devices that automatically change discs. Yes, there would be, like CD changers. (That reminds me, I've got one of those lying around too - I just don't remember how broken it was when I removed it...) Regards, Willem Riede.