From mboxrd@z Thu Jan 1 00:00:00 1970 From: Patrick Mansfield Subject: Re: [patch] 2.6 -- add IOI Media Bay to SCSI quirk list Date: Thu, 12 Aug 2004 17:27:56 -0700 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20040813002756.GA21763@beaverton.ibm.com> References: <200408122137.i7CLbGU13688@ra.tuxdriver.com> <20040812225118.GA20904@beaverton.ibm.com> <411BF6A5.2030306@tuxdriver.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from e35.co.us.ibm.com ([32.97.110.133]:53500 "EHLO e35.co.us.ibm.com") by vger.kernel.org with ESMTP id S268913AbUHMA2H (ORCPT ); Thu, 12 Aug 2004 20:28:07 -0400 Content-Disposition: inline In-Reply-To: <411BF6A5.2030306@tuxdriver.com> List-Id: linux-scsi@vger.kernel.org To: "John W. Linville" Cc: linux-kernel@vger.kernel.org, linux-scsi@vger.kernel.org, James.Bottomley@SteelEye.com On Thu, Aug 12, 2004 at 07:00:53PM -0400, John W. Linville wrote: > Patrick Mansfield wrote: > > >We seem to be getting quite a few of these. In theory we could add a line > >like this for every multi-lun SCSI device. > > > > > > Isn't that what the quirk list is for? We should not add to the list for devices that behave as expected. I would hope the number of borken BLIST_NOLUN devices is much smaller than the number of good BLIST_FORCELUN devices. The list is backwards compatible, so we have flags in there that are not required, like BLIST_FORCELUN. > >Can you instead try booting with scsi_mod.max_luns=8 (or such) or build > >with SCSI_MULTI_LUN enabled? > > > > > > That works for my box, but what about for others? Like those who may > have both a multi-lun device and a single-lun device that hangs on a > non-zero lun? What about the average luser who can't be bothered to > hack-up his startup scripts or *gasp* rebuild his kernel? Users should first set max_luns (or use a kernel built with SCSI_MULTI_LUN), and then if they find a device that breaks have it added to the quirks as BLIST_NOLUN. In the meantime they can boot using the devinfo flag with BLIST_NOLUN. I do not recall any reports (on linux-scsi) of devices that can't handle LUN 0 requests, there is no change in the number of SCSI_NOLUN devices from 2.4.21 to 2.6 (but I didn't check older kernels). I had thought it was best to build without SCSI_MULTI_LUN, but given the lack of SCSI_NOLUN additions, most users are better off with it on. > It seems like the quirk list is there for a reason. If we start > rejecting certain devices, then what is the criteria for a device to > actually make it on the list? Only add borken devices to the list, so it is a list of bad devices rather than a list of good ones. -- Patrick Mansfield