From patchwork Tue Apr 12 08:05:42 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ramana Radhakrishnan X-Patchwork-Id: 90737 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from sourceware.org (server1.sourceware.org [209.132.180.131]) by ozlabs.org (Postfix) with SMTP id AF402B6F16 for ; Tue, 12 Apr 2011 18:06:43 +1000 (EST) Received: (qmail 4461 invoked by alias); 12 Apr 2011 08:06:21 -0000 Received: (qmail 4422 invoked by uid 22791); 12 Apr 2011 08:06:10 -0000 X-SWARE-Spam-Status: No, hits=-2.6 required=5.0 tests=AWL, BAYES_00, RCVD_IN_DNSWL_LOW X-Spam-Check-By: sourceware.org Received: from mail-wy0-f175.google.com (HELO mail-wy0-f175.google.com) (74.125.82.175) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 12 Apr 2011 08:05:52 +0000 Received: by wye20 with SMTP id 20so6468583wye.20 for ; Tue, 12 Apr 2011 01:05:47 -0700 (PDT) Received: by 10.227.37.197 with SMTP id y5mr6317266wbd.4.1302595546839; Tue, 12 Apr 2011 01:05:46 -0700 (PDT) Received: from [192.168.32.37] (fw-lnat.cambridge.arm.com [217.140.96.63]) by mx.google.com with ESMTPS id z13sm3858858wbd.29.2011.04.12.01.05.46 (version=SSLv3 cipher=OTHER); Tue, 12 Apr 2011 01:05:46 -0700 (PDT) Message-ID: <4DA407D6.6@linaro.org> Date: Tue, 12 Apr 2011 09:05:42 +0100 From: Ramana Radhakrishnan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.14) Gecko/20110223 Lightning/1.0b2 Thunderbird/3.1.8 MIME-Version: 1.0 To: gcc-patches@gcc.gnu.org CC: Patch Tracking Subject: [Patch ARM] Fix PR48090 X-IsSubscribed: yes Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Delivered-To: mailing list gcc-patches@gcc.gnu.org Hi, This should fix PR48090 and should be applied to all release branches. The first alternative doesn't need an early clobber since it is tied to operand 0 - the second alternative however does need one. This is a bug that manifests itself with a particular set of command line options and it is interesting to note that this hasn't been caught in such a long time. Testing in progress . Ok to commit to all release branches ? cheers Ramana 2011-04-11 Ramana Radhakrishnan PR target/48090 * config/arm/arm.md (*arm_negdi2): Fix early clobber constraints. "TARGET_ARM" Index: gcc/config/arm/arm.md =================================================================== --- gcc/config/arm/arm.md (revision 172252) +++ gcc/config/arm/arm.md (working copy) @@ -3554,7 +3554,7 @@ ;; The constraints here are to prevent a *partial* overlap (where %Q0 == %R1). ;; The first alternative allows the common case of a *full* overlap. (define_insn "*arm_negdi2" - [(set (match_operand:DI 0 "s_register_operand" "=&r,r") + [(set (match_operand:DI 0 "s_register_operand" "=r,&r") (neg:DI (match_operand:DI 1 "s_register_operand" "0,r"))) (clobber (reg:CC CC_REGNUM))]