From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [140.186.70.92] (port=36735 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1OLbq3-00013V-6K for qemu-devel@nongnu.org; Mon, 07 Jun 2010 08:54:03 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.69) (envelope-from ) id 1OLbpy-00080M-1Y for qemu-devel@nongnu.org; Mon, 07 Jun 2010 08:53:59 -0400 Received: from mx1.redhat.com ([209.132.183.28]:28326) by eggs.gnu.org with esmtp (Exim 4.69) (envelope-from ) id 1OLbpx-00080G-NV for qemu-devel@nongnu.org; Mon, 07 Jun 2010 08:53:53 -0400 Received: from int-mx01.intmail.prod.int.phx2.redhat.com (int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11]) by mx1.redhat.com (8.13.8/8.13.8) with ESMTP id o57Crq8N015965 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Mon, 7 Jun 2010 08:53:52 -0400 Message-ID: <4C0CEBD1.2050809@redhat.com> Date: Mon, 07 Jun 2010 14:53:37 +0200 From: Kevin Wolf MIME-Version: 1.0 Subject: Re: [Qemu-devel] [PATCH] monitor: allow device to be ejected if no disk is inserted References: <20100601221219.GB13961@blackpad.lan.raisama.net> <4C0CD574.2020404@redhat.com> <20100607124345.GV14354@blackpad.lan.raisama.net> In-Reply-To: <20100607124345.GV14354@blackpad.lan.raisama.net> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Markus Armbruster , qemu-devel@nongnu.org, Luiz Capitulino Am 07.06.2010 14:43, schrieb Eduardo Habkost: > On Mon, Jun 07, 2010 at 02:19:28PM +0200, Markus Armbruster wrote: >> Kevin Wolf writes: >>> Am 02.06.2010 00:12, schrieb Eduardo Habkost: >>>> The first eject command didn't work because the is_inserted() check >>>> failed. >>> >>> But does it really make a difference? The guest should not see a medium >>> before and it should not see one afterwards. > > It does How that? Even if the host device is still connected, but no there's no medium in it, the guest shouldn't see a medium (I mean, which medium should it see if there is none?) > as the whole purpose of the "eject" command is to disconnect > the block device from the host backing file. > > Awful naming, I agree, but that's the expected semantics of the command. If it's just meant to say "disconnect the image" it's a really bad name. Luiz, can we please get rid of it before QMP becomes stable? > If we want to solve the naming confusion, this could be implemented as a > special case of the "change" command instead, and then the "eject" > command could be deprecated. Sounds much better, though it was suggested to deprecate "change" itself, too. ;-) Kevin