From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932503AbWCXML2 (ORCPT ); Fri, 24 Mar 2006 07:11:28 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932597AbWCXML2 (ORCPT ); Fri, 24 Mar 2006 07:11:28 -0500 Received: from outpipe-village-512-1.bc.nu ([81.2.110.250]:32423 "EHLO lxorguk.ukuu.org.uk") by vger.kernel.org with ESMTP id S932503AbWCXML1 (ORCPT ); Fri, 24 Mar 2006 07:11:27 -0500 Subject: Re: [RFC PATCH 35/35] Add Xen virtual block device driver. From: Alan Cox To: Jeff Garzik Cc: Ian Pratt , Anthony Liguori , Chris Wright , virtualization@lists.osdl.org, xen-devel@lists.xensource.com, linux-kernel@vger.kernel.org, Ian Pratt , ian.pratt@cl.cam.ac.uk, SCSI Mailing List In-Reply-To: <4421D943.1090804@garzik.org> References: <4421D943.1090804@garzik.org> Content-Type: text/plain Content-Transfer-Encoding: 7bit Date: Fri, 24 Mar 2006 12:17:53 +0000 Message-Id: <1143202673.18986.5.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.2.3 (2.2.3-2.fc4) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mer, 2006-03-22 at 18:09 -0500, Jeff Garzik wrote: > An IBM hypervisor on ppc64 communicates uses SCSI RPC messages. I think > this would be quite nice for Xen, because SCSI (a) is a message-based > model, and (b) implementing block using SCSI has a very high Just > Works(tm) value which cannot be ignored. And perhaps (c) SCSI target > code already exists, so implementing the server side doesn't require > starting from scratch, but rather simply connecting the Legos. A pure SCSI abstraction doesn't allow for shared head scheduling which you will need to scale Xen sanely on typical PC boxes. SCSI emulations are also always full of bits people got wrong, often critical bits like tagged queues and error sequences - things that break your journalled file system.