From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pat LaVarre Subject: Re: [PATCH/RFT] check non-scsi part of status in scsi_status_is_good Date: 30 Oct 2003 09:35:31 -0700 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <1067531731.4250.11.camel@pathost1.iomegacorp.com> References: Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from email-out1.iomega.com ([147.178.1.82]:40649 "EHLO email.iomega.com") by vger.kernel.org with ESMTP id S262617AbTJ3Qfs (ORCPT ); Thu, 30 Oct 2003 11:35:48 -0500 In-Reply-To: List-Id: linux-scsi@vger.kernel.org To: stern@rowland.harvard.edu Cc: patmans@us.ibm.com, ronald@kuetemeier.com, linux-scsi@vger.kernel.org, usb-storage@one-eyed-alien.net, james.bottomley@steeleye.com > > With the patch, the MODE SENSE page 3f len 4 fails, so we try a MODE SENSE > > page 0 len 4 and it hangs ... > > If the first MODE SENSE were 10 bytes your device would probably work OK, > > but per comments in sd.c, we should not increase this. > > > > But, the 2.4 code used only a 255 byte MODE SENSE request. ... > > IMO, this is exactly the sort of thing that has been causing us so much > pain. Some device doesn't accept the standard commands, dies when it gets > them, and will only work with some idiosyncratic combination of > parameters. > > I can't think of any way to deal with this other than special-purpose > per-device code, the simplest of which is blacklisting. Not idiosyncratic, if in fact -y "5A 00 08:00:00:00 00 00:1C 00" -i x1C does work. If that test works, then instead here we have a quiet unwitting insistence that all hosts fit within the narrow, massively-distributed, binary-code-only Windows redefinition of SCSI, as opposed to the dreamy fiction published at t10.org and elsewhere. I feel sympathy for the device because from 1994 to 1999 for pay I wrote device firmware. I never wrote anything this broken only because I wrote my own tests to step outside of what Windows tries. Because I was writing my own tests in confidence, I did ship misinterpretations where my initial understanding of the merely English specifications differed from what turned out to be the more broadly accepted understanding. As far as I know, all the companies who have tried to write such tests for pay have gone out of business after a few years. Pat LaVarre