From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benny Halevy Subject: Re: [PATCH 4/4] aic7xxx: Enable 16-bit CDBs Date: Mon, 22 Oct 2007 23:24:49 +0200 Message-ID: <471D1521.3090405@panasas.com> References: <20071019083239.D88FE18C61E@pentland.suse.de> <1192983563.3339.4.camel@localhost.localdomain> <471CF5B9.2090800@panasas.com> <20071022193551.GF27248@parisc-linux.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from sa14.bezeqint.net ([192.115.104.29]:47977 "EHLO sa14.bezeqint.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750828AbXJVVYw (ORCPT ); Mon, 22 Oct 2007 17:24:52 -0400 In-Reply-To: <20071022193551.GF27248@parisc-linux.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Matthew Wilcox Cc: Boaz Harrosh , James Bottomley , Hannes Reinecke , linux-scsi@vger.kernel.org On Oct. 22, 2007, 21:35 +0200, Matthew Wilcox wrote: > On Mon, Oct 22, 2007 at 09:10:49PM +0200, Boaz Harrosh wrote: >> I'm about to finish an RFC patchset for the extended commands. >> I have implemented a more aggressive approach than the one >> I've been sending for the last year. >> (Matthew I have an extra 8-bytes save to scsi_cmnd on >> 64bit and 12 bytes for 32bit. Guess how? ;)) > > Well ... the command has to be stored somewhere. If it's an additional > kmalloc, that's a loss. If it's in the request, that's a loss too ... > let's see where you're keeping it ;-) > I love spoiling Boaz's surprise :) The gist of it is scsi_cmnd->cmnd pointing at req->cmd or at req->varlen_cmd as appropriate. Most users can't tell the difference except if they tried using sizeof(scsi_cmnd->cmnd) in which case they can use MAX_COMMAND_SIZE instead. Benny