From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:52077)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1WL8Aw-0001EZ-Kl
	for qemu-devel@nongnu.org; Wed, 05 Mar 2014 04:31:48 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1WL8An-0002Vo-SV
	for qemu-devel@nongnu.org; Wed, 05 Mar 2014 04:31:42 -0500
Received: from mx1.redhat.com ([209.132.183.28]:38986)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1WL8An-0002VS-KC
	for qemu-devel@nongnu.org; Wed, 05 Mar 2014 04:31:33 -0500
Message-ID: <5316EEEF.30901@redhat.com>
Date: Wed, 05 Mar 2014 10:31:27 +0100
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1393969512-9468-1-git-send-email-stefanha@redhat.com>
	<5316DCD3.3000306@redhat.com> <5316ED59.8010400@suse.de>
In-Reply-To: <5316ED59.8010400@suse.de>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Qemu-devel] [PATCH 0/4] qom: link property fixes
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: =?UTF-8?B?QW5kcmVhcyBGw6RyYmVy?= <afaerber@suse.de>, Stefan Hajnoczi <stefanha@redhat.com>, qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>, Anthony Liguori <aliguori@amazon.com>

Il 05/03/2014 10:24, Andreas F=C3=A4rber ha scritto:
> Am 05.03.2014 09:14, schrieb Paolo Bonzini:
>> Il 04/03/2014 22:45, Stefan Hajnoczi ha scritto:
>>> There are two problems with QOM link properties:
>>>
>>> 1. There are refcount leaks in many object_property_add_link() caller=
s.
>>> 2. There is no way to prevent link properties from being set after a
>>> device is
>>>    realized.
>>>
>>> This patch series fixes these issues by extending the
>>> object_property_add_link() function.
>>
>> Thanks for looking into this!  Just two (mostly) cosmetic comments:
>>
>> 1) in patch 2 we could use a flags argument instead of a "bool".  This
>> is more easily extensible and self-documenting.
>
> I wonder if we can do this more cleverly: The two ways a link<> propert=
y
> can get set that I can think of are either a) assigning the field a new
> value in C code or b) using object_property_set_link() with a textual
> path. Can't the latter simply save an unref flag within the property's
> opaque when used?

Sounds too magic...

Paolo