From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-154703-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 9325 invoked by alias); 16 Jun 2009 12:56:32 -0000
Received: (qmail 9314 invoked by uid 22791); 16 Jun 2009 12:56:32 -0000
X-SWARE-Spam-Status: No, hits=-0.5 required=5.0 	tests=AWL,BAYES_50,SARE_MSGID_LONG40,SPF_PASS
X-Spam-Check-By: sourceware.org
Received: from mail-qy0-f194.google.com (HELO mail-qy0-f194.google.com) (209.85.221.194)     by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 16 Jun 2009 12:56:23 +0000
Received: by qyk32 with SMTP id 32so5730963qyk.0         for <gcc@gcc.gnu.org>; Tue, 16 Jun 2009 05:56:21 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.220.46.10 with SMTP id h10mr5564826vcf.8.1245156981405; Tue,  	16 Jun 2009 05:56:21 -0700 (PDT)
In-Reply-To: <7FB04A5C213E9943A72EE127DB74F0AD93C5706309@SJEXCHCCR02.corp.ad.broadcom.com>
References: <7FB04A5C213E9943A72EE127DB74F0AD93C5706309@SJEXCHCCR02.corp.ad.broadcom.com>
Date: Tue, 16 Jun 2009 12:56:00 -0000
Message-ID: <84fc9c000906160556h68b3ccf3k9b89ac8d0fd68211@mail.gmail.com>
Subject: Re: Questionable function renaming
From: Richard Guenther <richard.guenther@gmail.com>
To: Bingfeng Mei <bmei@broadcom.com>
Cc: "gcc@gcc.gnu.org" <gcc@gcc.gnu.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
X-IsSubscribed: yes
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
X-SW-Source: 2009-06/txt/msg00357.txt.bz2

On Tue, Jun 16, 2009 at 2:23 PM, Bingfeng Mei<bmei@broadcom.com> wrote:
> Hello,
>
> I came across a function renaming issue in GCC 4.4.
>
> ~/work/install-x86/bin/gcc -Wall -Winline -O3 -g -D_FILE_OFFSET_BITS=3D64=
 =A0-save-temps -c bzip2.c -fdump-tree-all
>
>
> Function saveInputFileMetaInfo is renamed to T.251 after ipa passes, whic=
h are not dumped into intermediate files. So compare bzip2.c.038t.release_s=
sa and bzip2.c.049t.copyrename2 file,
>
> ;; Function saveInputFileMetaInfo (saveInputFileMetaInfo)
>
> Released 3 names, 50.00%
> saveInputFileMetaInfo (Char * srcName)
> {
> =A0IntNative retVal;
>
> <bb 2>:
> =A0retVal_4 =3D __xstat (1, srcName_1(D), &fileMetaInfo);
> =A0if (retVal_4 !=3D 0)
> =A0 =A0goto <bb 3>;
> =A0else
> =A0 =A0goto <bb 4>;
>
> <bb 3>:
> =A0ioError ();
>
> <bb 4>:
> =A0return;
>
> }
>
> becomes
>
> ;; Function T.251 (T.251)
>
> T.251 ()
> {
> =A0IntNative retVal;
>
> <bb 2>:
> =A0retVal_1 =3D __xstat (1, &inName, &fileMetaInfo);
> =A0if (retVal_1 !=3D 0)
> =A0 =A0goto <bb 3>;
> =A0else
> =A0 =A0goto <bb 4>;
>
> <bb 3>:
> =A0ioError ();
>
> <bb 4>:
> =A0return;
>
> }
>
> The renamed function stays afterwards. My understanding is that ipa-cp pa=
ss did copy-propagation of the original function but still somehow create a=
 clone as suggested bzip2.c.203t.statistics.
>
> bzip2.c.203t.statistics:35 copyprop "Copies propagated" "saveInputFileMet=
aInfo" 1
> bzip2.c.203t.statistics:35 copyprop "Statements deleted" "saveInputFileMe=
taInfo" 3
>
> This seems unnecessary and causes trouble in our profiling/debugging tool=
s. I could diable this by using -fno-ipa-cp. But is there any better way? T=
hanks in advance

It constant propagated &inName here (probably not very profitable
as no further optimization is performed on the body) and removed
the then unused parameter.

Any better way to do what?

Richard.

> Cheers,
> Bingfeng Mei
>
> Boradcom UK