From mboxrd@z Thu Jan 1 00:00:00 1970 From: Luben Tuikov Subject: Re: Patch: change the serial_number for error-handler commands Date: Wed, 21 May 2003 15:24:25 -0400 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <3ECBD269.7050404@rogers.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from fep03-mail.bloor.is.net.cable.rogers.com ([66.185.86.73]:13214 "EHLO fep03-mail.bloor.is.net.cable.rogers.com") by vger.kernel.org with ESMTP id S262254AbTEUTL0 (ORCPT ); Wed, 21 May 2003 15:11:26 -0400 In-Reply-To: List-Id: linux-scsi@vger.kernel.org To: Alan Stern Cc: Mike Anderson , linux-scsi@vger.kernel.org Alan Stern wrote: >>Alan, could you please explain this patch? > > > It's just to insure that when the error handler sends a TEST-UNIT-READY, > the serial number is different from that of the immediately-prior failed > command (or from that of any command issued recently or likely to be > issued in the near future). This is a _trivial_ bug in SCSI Core (error handler, ...) which I hope will be fixed in the near future (Mike or Patrick will send a patch). That is, as per my other email, two successive commands (CDBs) should never be sent with duplicate cmdsn's. What we're talking about here is a _consistent_ path of a command through SCSI Core, before it reaches a LLDD. This is currently NOT the case, thus the duplicate cmdsn... -- Luben