From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: [DOC PATCH] T10/04-262r8 Date: Thu, 10 Nov 2005 10:58:39 -0500 Message-ID: <1131638319.3336.9.camel@mulgrave> References: <4371BB82.3090208@gmx.de> <4371C557.80807@gmx.de> <4371F7C3.5080702@rtr.ca> <43720232.6090609@gmx.de> <4372172A.50003@gmx.de> <43721C3C.1010103@pobox.com> <1131553363.3271.1.camel@mulgrave> <4372E66F.5070802@gmx.de> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <4372E66F.5070802@gmx.de> Sender: linux-scsi-owner@vger.kernel.org To: t.schorpp@gmx.de Cc: Linux SCSI list , linux-ide@vger.kernel.org List-Id: linux-ide@vger.kernel.org On Thu, 2005-11-10 at 07:19 +0100, thomas schorpp wrote: > how about full definitions of the pasthru cdb's in kernel structs and approbiate ioctls? Why would we want to do that? The whole point about pass through is that it is designed to use the existing SCSI infrastructure transparently (hence use SG_IO to send an ATA command to a bridge that understands it---but the key point here is that the device so addressed must understand the opcodes). > the passthru cdb is also not clear enough to me: > according to sat-r06 spec byte 0-2 is scsi, 3-9 are ata, 11 is scsi, yes? Actually, no; byte 0 is the opcode, so you could call that SCSI and byte 11 (or 15) is the control as defined by SAM; the rest all contain SAT specific fields. James