From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 42074 invoked by alias); 22 Sep 2015 19:20:33 -0000 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 Received: (qmail 42062 invoked by uid 89); 22 Sep 2015 19:20:32 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=0.1 required=5.0 tests=AWL,BAYES_50,KAM_LAZY_DOMAIN_SECURITY,RP_MATCHES_RCVD,SPF_HELO_PASS autolearn=no version=3.3.2 X-HELO: mx1.redhat.com Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with (AES256-GCM-SHA384 encrypted) ESMTPS; Tue, 22 Sep 2015 19:20:32 +0000 Received: from int-mx14.intmail.prod.int.phx2.redhat.com (int-mx14.intmail.prod.int.phx2.redhat.com [10.5.11.27]) by mx1.redhat.com (Postfix) with ESMTPS id ED9648E6E3; Tue, 22 Sep 2015 19:20:30 +0000 (UTC) Received: from anchor.twiddle.net (vpn-225-169.phx2.redhat.com [10.3.225.169]) by int-mx14.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id t8MJKUsl004212; Tue, 22 Sep 2015 15:20:30 -0400 Subject: Re: [PATCH] Explicit register earlyclobber fix To: Jeff Law , Andrew Jenner , GCC Patches References: <56017AA6.20709@codesourcery.com> <5601A2DD.5090408@redhat.com> <5601A82C.7060307@redhat.com> <5601A8EF.4050803@redhat.com> <5601A978.9040107@redhat.com> From: Richard Henderson Message-ID: <5601A9FD.3020302@redhat.com> Date: Tue, 22 Sep 2015 19:31:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.2.0 MIME-Version: 1.0 In-Reply-To: <5601A978.9040107@redhat.com> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-IsSubscribed: yes X-SW-Source: 2015-09/txt/msg01688.txt.bz2 On 09/22/2015 12:18 PM, Jeff Law wrote: > On 09/22/2015 01:15 PM, Richard Henderson wrote: >>> And doesn't it have to be =& and always at the start for an asm constraint? >>> For operands in patterns, we have to consider operand alternatives, but that's >>> not on issue here. >> >> Technically we do support alternatives on asms. Not usefully except for CISC, >> since all alternatives have to use the same output pattern, but one can write >> things like >> >> asm("add %0,%1" : "=r,m"(x) : "rim,ri"(y)) > Wow, learn something every day. In that case, I think we do have to search a > little harder than just peeking at the first two letters of the constraint :( We wouldn't be. He's adding this check at the beginning of each alternative. So why would we want to search the entire rest of the string now? r~