From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from sunset.davemloft.net (unknown [74.93.104.97]) by ozlabs.org (Postfix) with ESMTP id E4B62DDF7E for ; Sat, 16 Jun 2007 07:17:13 +1000 (EST) Date: Fri, 15 Jun 2007 14:17:20 -0700 (PDT) Message-Id: <20070615.141720.112626963.davem@davemloft.net> To: Geert.Uytterhoeven@sonycom.com Subject: Re: [patch 4/6] ps3: Disk Storage Driver From: David Miller In-Reply-To: References: <20070615120848.352093000@pademelon.sonytel.be> <1181918159.25228.543.camel@pmac.infradead.org> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Cc: axboe@kernel.dk, James.Bottomley@SteelEye.com, linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org, linuxppc-dev@ozlabs.org, paulus@samba.org, dwmw2@infradead.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , From: Geert Uytterhoeven Date: Fri, 15 Jun 2007 16:43:05 +0200 (CEST) > On Fri, 15 Jun 2007, David Woodhouse wrote: > > On Fri, 2007-06-15 at 13:39 +0200, Geert Uytterhoeven wrote: > > > Add a Disk Storage Driver for the PS3: > > > - Implemented as a block device driver with a dynamic major > > > - Disk names (and partitions) are of the format ps3d%c(%u) > > > - Uses software scatter-gather with a 64 KiB bounce buffer as the hypervisor > > > doesn't support scatter-gather > > > > Any particular reason why this is done as a separate block device driver > > rather than as SCSI? > > Because no new fake SCSI drivers are accepted anymore. I'm strongly divided on this issue as I'm about to hit the same exact thing for Sun Logical Domains on sparc64 Niagara systems. In fact the interface I get to use allows SCSI commands to be sent pass-through to the device, even though the basic virtual I/O API is purely block I/O based. It's senseless to make people build new major/minor numbers for all these new quirky storage drivers. People have to add support for the new major number to installers and all kinds of other tools. If the SCSI guys were smart, there would be a totally generic helper layer that allows anyone to hook into the SCSI layer as a virtual SCSI disk provider in like 10 lines of code. :-)