From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1M6kiV-0001eA-5X
	for qemu-devel@nongnu.org; Wed, 20 May 2009 08:16:15 -0400
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1M6kiU-0001dj-KM
	for qemu-devel@nongnu.org; Wed, 20 May 2009 08:16:14 -0400
Received: from [199.232.76.173] (port=43593 helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1M6kiU-0001dR-Gq
	for qemu-devel@nongnu.org; Wed, 20 May 2009 08:16:14 -0400
Received: from fg-out-1718.google.com ([72.14.220.157]:38046)
	by monty-python.gnu.org with esmtp (Exim 4.60)
	(envelope-from <laurent.desnogues@gmail.com>) id 1M6kiU-0003YO-3i
	for qemu-devel@nongnu.org; Wed, 20 May 2009 08:16:14 -0400
Received: by fg-out-1718.google.com with SMTP id e21so145354fga.8
	for <qemu-devel@nongnu.org>; Wed, 20 May 2009 05:16:12 -0700 (PDT)
MIME-Version: 1.0
In-Reply-To: <200905201148.43631.paul@codesourcery.com>
References: <1242745197.24234.7.camel@peak10.cs.hut.fi>
	<200905201148.43631.paul@codesourcery.com>
Date: Wed, 20 May 2009 14:16:12 +0200
Message-ID: <761ea48b0905200516g47713089g5d0b06f6f94bcd1a@mail.gmail.com>
Subject: Re: [Qemu-devel] Instruction counting instrumentation for ARM +
	initial patch
From: Laurent Desnogues <laurent.desnogues@gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
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: =?ISO-8859-1?Q?Timo_T=F6yry?= <ttoyry@cs.hut.fi>
Cc: qemu-devel@nongnu.org

On Wed, May 20, 2009 at 12:48 PM, Paul Brook <paul@codesourcery.com> wrote:
> On Tuesday 19 May 2009, Timo T=F6yry wrote:
>> Hi,
>>
>> The attached patch implements instruction counting instrumentation for
>> ARM usermode linux. Currently the patch is a working draft but we
>> intend to clean it up for possible inclusion. We'd appreciate any
>> comments. Below is a more detailed description.
>
> Ewww. This looks pretty much impossible to maintain.
> If you really want to do this I recommend dumping an execution trace, and
> having a third party utility do the counting after the fact.

My advice would be similar to Paul's:  for my instruction counter,
instead of counting instructions, I count translation blocks.  I can
either output a TB trace (which can be huge, but useful to do
trace analysis) or an execution count for each TB.  As a postpass
you can do whatever you want.  The benefits, on top of being easier
to maintain, are that it's more flexible, it has a very small impact on
QEMU execution time, and it can be trivially ported to any front end.
The problem is to play nicely with TB flushes :-)

HTH,

Laurent