From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1MF8Fx-0000Xc-O7 for qemu-devel@nongnu.org; Fri, 12 Jun 2009 11:01:25 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1MF8Fr-0000Wj-BX for qemu-devel@nongnu.org; Fri, 12 Jun 2009 11:01:25 -0400 Received: from [199.232.76.173] (port=37691 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1MF8Fr-0000Wg-4u for qemu-devel@nongnu.org; Fri, 12 Jun 2009 11:01:19 -0400 Received: from mx2.redhat.com ([66.187.237.31]:52487) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1MF8Fp-00024o-VQ for qemu-devel@nongnu.org; Fri, 12 Jun 2009 11:01:19 -0400 Message-ID: <4A326D23.4070401@redhat.com> Date: Fri, 12 Jun 2009 16:58:43 +0200 From: Gerd Hoffmann MIME-Version: 1.0 Subject: Re: [Qemu-devel] [PATCH 3/4] qdev: very first cut of scsi bus support. References: <1244798921-23262-1-git-send-email-kraxel@redhat.com> <1244798921-23262-4-git-send-email-kraxel@redhat.com> <200906121222.46360.paul@codesourcery.com> In-Reply-To: <200906121222.46360.paul@codesourcery.com> Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Paul Brook Cc: qemu-devel@nongnu.org On 06/12/09 13:22, Paul Brook wrote: > On Friday 12 June 2009, Gerd Hoffmann wrote: >> + >> + if (bus) { >> + d = scsi_create_simple(bus, "scsi-disk"); >> + } else { >> + /* temporary until usb is qdev-ified */ >> + d = (SCSIDevice *)qemu_mallocz(sizeof(SCSIDevice)); > >> +static void scsi_disk_initfn(SCSIDevice *dev) >> +{ >> + /* TODO */ >> +} > > These are both fairly good indicators that this patch is nowhere near ready > for integration. The whole point of the qdev API is to abstract device > creation from individual device implementation details. I know. It is tagged "very first cut" for a reason ;) Just skip it for now if you prefer to wait for something more complete. > Your implementation > has the abstraction layers completely backwards. Yes, but this is how it works in qemu today. Changing that without breaking stuff isn't that easy though. Guess I'll first have a look at usb anyway, so I can make sure usb-storage is covered too. cheers, Gerd