From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:46584)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1VWSZQ-0001oU-Cn
	for qemu-devel@nongnu.org; Wed, 16 Oct 2013 10:59:38 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1VWSZK-0008Kr-6C
	for qemu-devel@nongnu.org; Wed, 16 Oct 2013 10:59:32 -0400
Received: from mx1.redhat.com ([209.132.183.28]:4514)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1VWSZJ-0008KX-UW
	for qemu-devel@nongnu.org; Wed, 16 Oct 2013 10:59:26 -0400
Message-ID: <525EA9C2.90805@redhat.com>
Date: Wed, 16 Oct 2013 16:59:14 +0200
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <525E8DC3.70409@borntraeger.net> <525E8F53.3090603@redhat.com>
	<525EA150.3060107@de.ibm.com>
In-Reply-To: <525EA150.3060107@de.ibm.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] RFC: KVM _CREATE_DEVICE considered harmful?
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: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: "qemu-devel@nongnu.org" <qemu-devel@nongnu.org>, Jens Freimann <jfrei@linux.vnet.ibm.com>, Alexander Graf <agraf@suse.de>, Gleb Natapov <gleb@redhat.com>, KVM list <kvm@vger.kernel.org>

Il 16/10/2013 16:23, Christian Borntraeger ha scritto:
> That would work, but it would change the ioctl number of KVM_*_DEVICE_ADDR,
> due to the changed size of struct kvm_device_attr. We would then need compat
>  handlers in the kernel.

Actually I did that on purpose :) but perhaps you're right that the
complication would be too high.

> If we could encode it in the existing interface the impact would be smaller.
> e.g. 
> 
> #define ATTR_ATTR_MASK 0xffffffffULL
> #define ATTR_LEN_MASK 0xffffffff00000000ULL
> 
> 
>     switch (attr->attr) {
> ---> 
>     switch (attr->attr & ATTR_ATTR_MASK) {
> 
> Then we could keep the device model abstraction.

That makes sense too.

Paolo