From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from [140.186.70.92] (port=54600 helo=eggs.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1OLiGo-0008VC-Rt
	for qemu-devel@nongnu.org; Mon, 07 Jun 2010 15:46:03 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.69)
	(envelope-from <rth7680@gmail.com>) id 1OLiGl-0003ZF-JL
	for qemu-devel@nongnu.org; Mon, 07 Jun 2010 15:46:02 -0400
Received: from mail-gy0-f173.google.com ([209.85.160.173]:51235)
	by eggs.gnu.org with esmtp (Exim 4.69)
	(envelope-from <rth7680@gmail.com>) id 1OLiGl-0003Z6-G3
	for qemu-devel@nongnu.org; Mon, 07 Jun 2010 15:45:59 -0400
Received: by gyd5 with SMTP id 5so2696662gyd.4
	for <qemu-devel@nongnu.org>; Mon, 07 Jun 2010 12:45:58 -0700 (PDT)
Sender: Richard Henderson <rth7680@gmail.com>
Message-ID: <4C0D4C5E.8030708@twiddle.net>
Date: Mon, 07 Jun 2010 12:45:34 -0700
From: Richard Henderson <rth@twiddle.net>
MIME-Version: 1.0
Subject: Re: [Qemu-devel] [PATCH 2/2] tcg-i386: Use segment registers to
	implement GUEST_BASE.
References: <1275611718-32630-1-git-send-email-rth@twiddle.net>
	<1275611718-32630-3-git-send-email-rth@twiddle.net>
	<7222640D-96FF-41DA-A41D-6D2CB6D669DB@suse.de>
In-Reply-To: <7222640D-96FF-41DA-A41D-6D2CB6D669DB@suse.de>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
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.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: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Alexander Graf <agraf@suse.de>
Cc: qemu-devel@nongnu.org, aurelien@aurel32.net

On 06/03/2010 11:35 PM, Alexander Graf wrote:
> 
> On 04.06.2010, at 02:35, Richard Henderson wrote:
> 
>> For 32-bit, using a segment override is smaller than the 4-byte 
>> immediate offset.  For 64-bit, segments can hold the entire 64-bit 
>> offset whereas the 4-byte immediate cannot.
> 
> Very nice idea indeed :). Have you found it to be faster? IIRC
> segment accesses are slower when seg_offs != 0. But then again the
> code is smaller, so it might weigh it up.

I've not yet tried to do any proper benchmarking.

I had thought that the segment access was slower only when limit != 4G,
i.e. when the segment boundaries need to be checked.  And that 64-bit
segments don't have boundaries, so the slowdown does not apply there either.

I'll put it on my queue for the weekend.


r~