Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1422767AbWCXMi5 (ORCPT ); Fri, 24 Mar 2006 07:38:57 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1422762AbWCXMi5 (ORCPT ); Fri, 24 Mar 2006 07:38:57 -0500 Received: from srv5.dvmed.net ([207.36.208.214]:29669 "EHLO mail.dvmed.net") by vger.kernel.org with ESMTP id S1422741AbWCXMi4 (ORCPT ); Fri, 24 Mar 2006 07:38:56 -0500 Message-ID: <4423E853.1040707@garzik.org> Date: Fri, 24 Mar 2006 07:38:43 -0500 From: Jeff Garzik User-Agent: Thunderbird 1.5 (X11/20060313) MIME-Version: 1.0 To: Alan Cox 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 Subject: Re: [RFC PATCH 35/35] Add Xen virtual block device driver. References: <4421D943.1090804@garzik.org> <1143202673.18986.5.camel@localhost.localdomain> In-Reply-To: <1143202673.18986.5.camel@localhost.localdomain> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Score: -2.5 (--) X-Spam-Report: SpamAssassin version 3.0.5 on srv5.dvmed.net summary: Content analysis details: (-2.5 points, 5.0 required) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1422 Lines: 37 Alan Cox wrote: > 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. Not true at all. If you can do it with a block device, you can do it with a SCSI block device. In fact, SCSI should make a few things easier, because the notion of host+bus topology is already present, and notion of messaging is already present, so you don't have to recreate that in a Xen block device infrastructure. > 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. This I'll grant you. Jeff - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/