From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1HRxEj-0008Me-Ac
	for qemu-devel@nongnu.org; Thu, 15 Mar 2007 17:11:49 -0400
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1HRxEh-0008MK-Vc
	for qemu-devel@nongnu.org; Thu, 15 Mar 2007 17:11:49 -0400
Received: from [199.232.76.173] (helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1HRxEh-0008MD-Nx
	for qemu-devel@nongnu.org; Thu, 15 Mar 2007 16:11:47 -0500
Received: from mail.codesourcery.com ([65.74.133.4])
	by monty-python.gnu.org with esmtps
	(TLS-1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.60)
	(envelope-from <paul@codesourcery.com>) id 1HRxDe-00026C-Oe
	for qemu-devel@nongnu.org; Thu, 15 Mar 2007 17:10:43 -0400
From: Paul Brook <paul@codesourcery.com>
Subject: Re: [Qemu-devel] qemu-arm: wrong execution of post-indexed loads when
	Rm and Rd are the same register
Date: Thu, 15 Mar 2007 21:10:39 +0000
References: <1173987324.9939.0.camel@edgy-laptop>
	<200703152003.21276.paul@codesourcery.com>
	<bdb0025f0703151332j36d31aa7i6ee2cea9bfb29a94@mail.gmail.com>
In-Reply-To: <bdb0025f0703151332j36d31aa7i6ee2cea9bfb29a94@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200703152110.39814.paul@codesourcery.com>
Reply-To: qemu-devel@nongnu.org
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: qemu-devel@nongnu.org
Cc: Rodrigo Vivi <rodrigo.vivi@gmail.com>, Lauro Ramos Venancio <lauro.venancio@indt.org.br>

> > This is still wrong.
>
> So, is this a known bug?

Still wrong implies it's a bug, and your patch does not fix it properly.

> > The writeback must happen after the load.
>
> We code like this because
> - we didn't find this restriction in arm reference manual

It's the Abort model section you mention below.

> - the LLVM uses this instruction expecting a result like this

The compiler knows nothing about the abort behavior. The difference is only 
visible if the load faults.

> - That was the result that we got running these instructions in an OMAP1710

I suggest you check again. I'm fairly sure the arm926 implements the Base 
Restored abort model.

Paul