From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1MvoqR-00058A-9J
	for qemu-devel@nongnu.org; Thu, 08 Oct 2009 04:59:31 -0400
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1MvoqM-00056y-1d
	for qemu-devel@nongnu.org; Thu, 08 Oct 2009 04:59:30 -0400
Received: from [199.232.76.173] (port=53869 helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1MvoqL-00056r-QL
	for qemu-devel@nongnu.org; Thu, 08 Oct 2009 04:59:25 -0400
Received: from mx1.redhat.com ([209.132.183.28]:49995)
	by monty-python.gnu.org with esmtp (Exim 4.60)
	(envelope-from <mst@redhat.com>) id 1MvoqL-0004Cd-C1
	for qemu-devel@nongnu.org; Thu, 08 Oct 2009 04:59:25 -0400
Date: Thu, 8 Oct 2009 10:57:20 +0200
From: "Michael S. Tsirkin" <mst@redhat.com>
Message-ID: <20091008085720.GB5269@redhat.com>
References: <20091007123348.GA31537@redhat.com> <4ACC9133.9060903@redhat.com>
	<20091007134853.GA9769@redhat.com> <4ACCA4AD.7050606@redhat.com>
	<20091007192127.GA3796@redhat.com> <4ACDA18D.809@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4ACDA18D.809@redhat.com>
Subject: [Qemu-devel] Re: [PATCH] qemu/pci: optimize pci config handling
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.gnu.org/pipermail/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: Isaku Yamahata <yamahata@valinux.co.jp>, qemu-devel@nongnu.org

On Thu, Oct 08, 2009 at 10:23:41AM +0200, Paolo Bonzini wrote:
>
>>>> What you porpose here is certainly more code than we had.
>>>> So why is this a good idea?
>>>
>>> Because it avoids the memcpy/memcmp most of the time (when the memcmp
>>> would surely succeed).
>>
>> Yes :) But at the cost of more code.  I don't think speed
>> matters there, so less code is good.
>
> Fine.
>
>>>   I supposed that would also matter more as the
>>> config space size increases---correct me and dismiss the patch if I am
>>> mistaken.
>>
>> No, we'll always only look need to look at the header, whatever the size
>> of the config space.  That's the point of the patch I posted - future
>> proof against config space size increases, not optimization.
>
> But fewer reads on average will not modify the header, so there will be  
> even fewer memcpy with my patch when the config space will be 4k.

Oh, I see. I was worrying about us adding some side effect where you
write into PCI extended register and header changes.  Maybe that won't
ever happen.

> Paolo