From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:36908) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1STrGr-0003lx-MV for qemu-devel@nongnu.org; Mon, 14 May 2012 05:08:53 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1STrGm-0004BB-7I for qemu-devel@nongnu.org; Mon, 14 May 2012 05:08:49 -0400 Received: from mx1.redhat.com ([209.132.183.28]:2462) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1STrGl-0004B1-Vt for qemu-devel@nongnu.org; Mon, 14 May 2012 05:08:44 -0400 Message-ID: <4FB0C6A7.2040308@redhat.com> Date: Mon, 14 May 2012 10:47:35 +0200 From: Kevin Wolf MIME-Version: 1.0 References: <1336749740-18474-1-git-send-email-armbru@redhat.com> <1336749740-18474-3-git-send-email-armbru@redhat.com> In-Reply-To: <1336749740-18474-3-git-send-email-armbru@redhat.com> Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [RFC PATCH 2/2] Split fdd devices off the floppy controller List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Markus Armbruster Cc: aliguori@us.ibm.com, hpoussin@reactos.org, qemu-devel@nongnu.org Am 11.05.2012 17:22, schrieb Markus Armbruster: > For historical reasons, and unlike other block devices, our floppy > devices isa-fdc, sysbus-fdc and SUNW,fdtwo integrate the controller > and the drive(s) in a single qdev. This makes them weird: we need > -global to set up floppy drives, unlike every other optional device. I like the idea of splitting the drives from the controller. In fact, I think we could even try to split them into a separate hw/fdd.c > Unfortunately, eliding the qbus means I can't make the floppy disk a > qdev (sub-class of TYPE_DEVICE), because qdevs can only connect to a > qbus. Anthony tells me that restriction is gone in his latest QOM > series. > > Since it's not a qdev, -device fdd does not work. Pity, because it > defeats the stated purpose of making floppy disk drives work like > other existing optional devices. As long as this is true, committing a patch like this doesn't help a lot, so I hope Anthony's patches will go in before this is ready. > Note: I *break* -global isa-fdc.driveA=... The properties are simply > gone. Fixable if we need backwards compatibility there. We might need it, I seem to remember that libvirt uses it. Kevin