From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 22837 invoked by alias); 3 Aug 2011 15:53:52 -0000 Received: (qmail 22819 invoked by uid 22791); 3 Aug 2011 15:53:51 -0000 X-SWARE-Spam-Status: No, hits=-6.7 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_HI,RP_MATCHES_RCVD,SPF_HELO_PASS X-Spam-Check-By: sourceware.org Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Wed, 03 Aug 2011 15:53:38 +0000 Received: from int-mx12.intmail.prod.int.phx2.redhat.com (int-mx12.intmail.prod.int.phx2.redhat.com [10.5.11.25]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id p73FrbFp003535 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 3 Aug 2011 11:53:37 -0400 Received: from anchor.twiddle.net (vpn-228-63.phx2.redhat.com [10.3.228.63]) by int-mx12.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id p73FrbBR008935; Wed, 3 Aug 2011 11:53:37 -0400 Message-ID: <4E396F00.7090109@redhat.com> Date: Wed, 03 Aug 2011 15:53:00 -0000 From: Richard Henderson User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:5.0) Gecko/20110707 Thunderbird/5.0 MIME-Version: 1.0 To: Georg-Johann Lay CC: GCC Patches , Denis Chertykov Subject: Re: [RFC] Cleanup DW_CFA_GNU_args_size handling References: <4E36D44A.6020607@redhat.com> <4E36F3AF.3070300@gjlay.de> <4E36F48F.4040106@redhat.com> <4E393BBB.8030300@gjlay.de> <4E395612.8060806@gjlay.de> <4E396589.8070601@redhat.com> <4E396D9C.8040502@gjlay.de> In-Reply-To: <4E396D9C.8040502@gjlay.de> Content-Type: text/plain; charset=UTF-8 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: 2011-08/txt/msg00361.txt.bz2 On 08/03/2011 08:47 AM, Georg-Johann Lay wrote: > With ACCUMULATE_OUTGOING_ARGS it looks much better: there is just > one such block in the prologue/epilogue. > > I think ACCUMULATE_OUTGOING_ARGS would be a win definitely. That's what I thought too, but with the test case in PR49881 I couldn't make A_O_A come out smaller than PUSHes. The reason being the function didn't otherwise need a frame pointer and we got to use REG_Y for something more useful. Perhaps that test case isn't the rule for real-world code. That's why I suggested implementing the command-line switch, to give the user an option of trying both and selecting the smaller option for their particular code. r~