From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:58050)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1ZVYm6-0004X8-E9
	for qemu-devel@nongnu.org; Sat, 29 Aug 2015 01:33:59 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1ZVYm5-0002Nn-Fx
	for qemu-devel@nongnu.org; Sat, 29 Aug 2015 01:33:58 -0400
Sender: Richard Henderson <rth7680@gmail.com>
References: <1440719254-12349-1-git-send-email-afaerber@suse.de>
	<55DFE3CC.4030806@twiddle.net> <55E08AEE.3060308@suse.de>
From: Richard Henderson <rth@twiddle.net>
Message-ID: <55E1443D.1060305@twiddle.net>
Date: Fri, 28 Aug 2015 22:33:49 -0700
MIME-Version: 1.0
In-Reply-To: <55E08AEE.3060308@suse.de>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Subject: Re: [Qemu-devel] [PATCH] tcg/aarch64: Fix tcg_out_qemu_{ld,
	st} for linux-user
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?Q?Andreas_F=c3=a4rber?= <afaerber@suse.de>, qemu-devel@nongnu.org
Cc: Paolo Bonzini <pbonzini@redhat.com>, Claudio Fontana <claudio.fontana@huawei.com>, qemu-stable <qemu-stable@nongnu.org>, Aurelien Jarno <aurelien@aurel32.net>

On 08/28/2015 09:23 AM, Andreas Färber wrote:
> An otype argument is being inserted as next-to-last argument for the
> function definitions. Same for the softmmu callsites. Only in the *-user
> callsites the argument order is being changed with addr_reg and off_r
> switching order? I don't see why that should be done in this patch. If
> it was wrong before, it should've been done in a separate patch.

It wasn't "wrong" before because it didn't matter before.

The whole point of the patch is that we're changing from

	(xD = data_reg, xA = addr_reg, xG = GUEST_BASE)

	ldr	xD, [xA, xG]
to
	ldr	xD, [xG, wA, uxtw]

i.e. zero-extending xA during the address formation.

Before this patch, the order of xA and xG doesn't matter; they're both straight 
addition operands.  After this patch, if we don't switch the argument order 
it'll be xG that gets zero-extending GUEST_BASE, which is obviously incorrect.

You'll note that we also switch the argument order for the softmmu case in the 
next patch, and for the same reason.

Tomorrow I'll be home and can do any further investigation in depth.


r~