From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([140.186.70.92]:44593)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <anthony@codemonkey.ws>) id 1RuW11-0001qN-G4
	for qemu-devel@nongnu.org; Mon, 06 Feb 2012 16:22:24 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <anthony@codemonkey.ws>) id 1RuW0z-0008P0-Tx
	for qemu-devel@nongnu.org; Mon, 06 Feb 2012 16:22:23 -0500
Received: from mail-pw0-f45.google.com ([209.85.160.45]:62358)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <anthony@codemonkey.ws>) id 1RuW0z-0008OI-Kz
	for qemu-devel@nongnu.org; Mon, 06 Feb 2012 16:22:21 -0500
Received: by pbaa11 with SMTP id a11so7170892pba.4
	for <qemu-devel@nongnu.org>; Mon, 06 Feb 2012 13:22:20 -0800 (PST)
Message-ID: <4F304488.1010106@codemonkey.ws>
Date: Mon, 06 Feb 2012 15:22:16 -0600
From: Anthony Liguori <anthony@codemonkey.ws>
MIME-Version: 1.0
References: <1328237992-14953-1-git-send-email-afaerber@suse.de>	<1328237992-14953-4-git-send-email-afaerber@suse.de>	<4F3028D2.2060409@codemonkey.ws>
	<4F3034AF.3000205@suse.de>
In-Reply-To: <4F3034AF.3000205@suse.de>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Subject: Re: [Qemu-devel] [PATCH RFC v3 03/21] qom: Introduce CPU class
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>
Cc: Peter Maydell <peter.maydell@linaro.org>, qemu-devel@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>

On 02/06/2012 02:14 PM, Andreas Färber wrote:
> Am 06.02.2012 20:24, schrieb Anthony Liguori:
>> On 02/02/2012 08:59 PM, Andreas Färber wrote:
>>> +/**
>>> + * cpu_common_reset:
>>> + * @cpu: The CPU whose common state is to be reset.
>>> + *
>>> + * To be used by derived classes.
>>> + */
>>> +void cpu_common_reset(CPU *cpu);
>>
>> Make this static, initialize reset = cpu_common_reset in
>> cpu_class_initfn, then in the derived class initfn, save the pointer to
>> the parent reset function so it can be called later.
>
> I don't see how that would work. To initialize, e.g., the ARMCPUClass
> with additional class fields I'm overriding the .class_init.

You're not overriding the class_init.  The class init for a type is called when 
that class is initialized for the first time.  See the documentation in object.h.

When class_init is called, the parent class type's class_init has already been 
called and the default values are set.  So:

static void arm_cpu_reset(CPUCommon *cpu)
{
    ARMCPU *s = ARM_CPU(cpu);
    ARMCPUClass *ac = ARM_CPU_GET_CLASS(s);

    // do arm specific reset

    // call super class reset
    ac->super_reset(cpu);
}

static void arm_cpu_class_initfn(ObjectClass *klass, void *data)
{
    CPUCommonClass *cc = CPU_COMMON_CLASS(klass);
    ARMCPUClass *ac = ARM_CPU_CLASS(klass);

    // cc->reset was set in CPUCommonClass's class_init
    ac->super_reset = cc->reset;
    cc->reset = arm_cpu_reset;
}

It's admittedly a little funky but this is the common idiom in gobject.  I'm not 
sure there's a better way.

Regards,

Anthony Liguori

  So in order
> to let CPUClass initialize the reset callback to its static one I'd need
> to make CPU's class_init function non-static so that I can call that
> from my derived class' class_init function, no?
>
> Andreas
>