From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:36330)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <marcel@redhat.com>) id 1YdDfN-0001bG-QV
	for qemu-devel@nongnu.org; Wed, 01 Apr 2015 04:06:26 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <marcel@redhat.com>) id 1YdDfK-0006Rm-Jx
	for qemu-devel@nongnu.org; Wed, 01 Apr 2015 04:06:25 -0400
Received: from mx1.redhat.com ([209.132.183.28]:42731)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <marcel@redhat.com>) id 1YdDfK-0006RS-F4
	for qemu-devel@nongnu.org; Wed, 01 Apr 2015 04:06:22 -0400
Message-ID: <551BA6F9.2070202@redhat.com>
Date: Wed, 01 Apr 2015 11:06:17 +0300
From: Marcel Apfelbaum <marcel@redhat.com>
MIME-Version: 1.0
References: <551AAD75.8090909@linux.vnet.ibm.com> <551B963E.9060800@gmail.com>
	<551BA5D1.8050002@redhat.com>
In-Reply-To: <551BA5D1.8050002@redhat.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [qemu devel] disable shared memory is not
 available with this QEMU binary
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Paolo Bonzini <pbonzini@redhat.com>, Tony Krowiak <akrowiak@linux.vnet.ibm.com>, qemu-devel@nongnu.org
Cc: =?UTF-8?B?QW5kcmVhcyBGw6RyYmVy?= <afaerber@suse.de>

On 04/01/2015 11:01 AM, Paolo Bonzini wrote:
>
>
> On 01/04/2015 08:54, Marcel Apfelbaum wrote:
>> This is the first object for which QemuOps are defined per
>> sub-type and are not global (if you don't take "object" under
>> consideration).
>
> We can return the same QemuOpts that were included before.
> Per-machine-type options are new and need not be covered by
> query-command-line-options.

OK, we have them under hw/core/machine.c as "base" machine properties.
We still need a way to fill them back into QemuOpts right?
Maybe return them to the static global list? It seems like a step
back, but if there is no better way...

Thanks,
Marcel

>
> Paolo
>
>> I saw others as well, like netdev, but I am not sure what happens there.
>>
>> Once the QemuOpts are parsed, the only place we can find those options
>> is the machine object itself (as QOM properties).
>>
>> I see a few options here:
>> 1. Add a feature to QemuOpts: "Look for options in QOM properties of
>> this obj"
>> 2. Add a callback to QEMU opts that supplies the options (have machine
>> supply the callback)
>> 3. Have the machine object fill in the corresponding QemuOpts on init.