From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 23162 invoked by alias); 5 May 2010 12:11:49 -0000 Received: (qmail 23152 invoked by uid 22791); 5 May 2010 12:11:48 -0000 X-SWARE-Spam-Status: No, hits=-2.2 required=5.0 tests=AWL,BAYES_00,T_RP_MATCHES_RCVD X-Spam-Check-By: sourceware.org Received: from cam-admin0.cambridge.arm.com (HELO cam-admin0.cambridge.arm.com) (217.140.96.50) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Wed, 05 May 2010 12:11:40 +0000 Received: from cam-owa1.Emea.Arm.com (cam-owa1.emea.arm.com [10.1.255.62]) by cam-admin0.cambridge.arm.com (8.12.6/8.12.6) with ESMTP id o45CBYeI023702; Wed, 5 May 2010 13:11:34 +0100 (BST) Received: from [10.1.69.50] ([10.1.255.212]) by cam-owa1.Emea.Arm.com with Microsoft SMTPSVC(6.0.3790.0); Wed, 5 May 2010 13:11:34 +0100 Subject: Re: ARM ldm/stm peepholes From: Richard Earnshaw To: Bernd Schmidt Cc: ramana.radhakrishnan@arm.com, GCC Patches In-Reply-To: <4BE15A28.1040207@codesourcery.com> References: <4BCD9301.2060605@codesourcery.com> <1272010406.6783.72.camel@e200593-lin.cambridge.arm.com> <1272027779.1977.22.camel@e200601-lin.cambridge.arm.com> <4BE15A28.1040207@codesourcery.com> Content-Type: text/plain Date: Wed, 05 May 2010 12:11:00 -0000 Message-Id: <1273061488.19300.6.camel@e102346-lin.cambridge.arm.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit X-IsSubscribed: yes Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2010-05/txt/msg00300.txt.bz2 On Wed, 2010-05-05 at 13:44 +0200, Bernd Schmidt wrote: > In the hope of moving this along, here's a smaller patch with pieces > that do not introduce functional changes, but reorganize the code a > little in preparation for the full patch. I hope this will make review > easier. > > Tested on > arm-linux-gnueabi(qemu-system-armv7{arch=armv7-a/thumb,thumb,}). Ok? > > > Bernd > Rather than need_add, I think it would be better to pass the offset. There are times when the offset can't be done with a single add insn, and we can't tell that if this is just a boolean (but it might still be profitable in those cases). Ok apart from that. R.