Date: Fri, 4 Jun 2010 01:50:03 GMT From: Garrett Cooper <yanefbsd@gmail.com> To: freebsd-bugs@FreeBSD.org Subject: Re: bin/147451: "gcc -g -O2 -march=nocona" loops compiling devel/qt4-script Message-ID: <201006040150.o541o3Vk093863@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR gnu/147451; it has been noted by GNATS. From: Garrett Cooper <yanefbsd@gmail.com> To: Peter Jeremy <peterjeremy@acm.org> Cc: FreeBSD-gnats-submit@freebsd.org Subject: Re: bin/147451: "gcc -g -O2 -march=nocona" loops compiling devel/qt4-script Date: Thu, 3 Jun 2010 18:49:35 -0700 On Thu, Jun 3, 2010 at 2:06 PM, Peter Jeremy <peterjeremy@acm.org> wrote: > >>Number: =A0 =A0 =A0 =A0 147451 >>Category: =A0 =A0 =A0 bin >>Synopsis: =A0 =A0 =A0 "gcc -g -O2 -march=3Dnocona" loops compiling devel/= qt4-script >>Confidential: =A0 no >>Severity: =A0 =A0 =A0 non-critical >>Priority: =A0 =A0 =A0 low >>Responsible: =A0 =A0freebsd-bugs >>State: =A0 =A0 =A0 =A0 =A0open >>Quarter: >>Keywords: >>Date-Required: >>Class: =A0 =A0 =A0 =A0 =A0sw-bug >>Submitter-Id: =A0 current-users >>Arrival-Date: =A0 Thu Jun 03 21:10:02 UTC 2010 >>Closed-Date: >>Last-Modified: >>Originator: =A0 =A0 Peter Jeremy >>Release: =A0 =A0 =A0 =A0FreeBSD 8.0-STABLE amd64 >>Organization: > n/a >>Environment: > System: FreeBSD server.vk2pj.dyndns.org 8.0-STABLE FreeBSD 8.0-STABLE #3:= Mon Apr 5 10:18:04 EST 2010 root@server.vk2pj.dyndns.org:/var/obj/usr/src/= sys/server amd64 > > GNU C++ version 4.2.1 20070719 =A0[FreeBSD] (amd64-undermydesk-freebsd) > =A0 =A0 =A0 =A0compiled by GNU C version 4.2.1 20070719 =A0[FreeBSD]. > > Also occurs on 8.1-PRERELEASE from early June. > >>Description: > =A0 =A0 =A0 =A0The compilation of .../JavaScriptCore/interpreter/Interpre= ter.cpp > =A0 =A0 =A0 =A0in ports/devel/qt4-script loops, with cc1plus slowly eatin= g memory > =A0 =A0 =A0 =A0until it hits ulimit or runs out of swap, when compiled wi= th > =A0 =A0 =A0 =A0'-g -O2 -march=3Dnocona' > > =A0 =A0 =A0 =A0Removing either '-g' or '-march-nocona' causes the compila= tion to > =A0 =A0 =A0 =A0succeed normally in 30 seconds or less. =A0With both speci= fied, > =A0 =A0 =A0 =A0I saw the cc1plus VSZ grow to over 1.5GB and use over 70 m= inutes > =A0 =A0 =A0 =A0of CPU before I gave up watching. > >>How-To-Repeat: > =A0 =A0 =A0 =A0cd /usr/ports/devel/qt4-script > =A0 =A0 =A0 =A0make clean > =A0 =A0 =A0 =A0make configure > =A0 =A0 =A0 =A0cd WRKDIRPREFIX/usr/ports/devel/qt4-script/work/qt-everywh= ere-opensource-src-4.6.2/src/script > =A0 =A0 =A0 =A0c++ -c -O2 -g -march=3Dnocona -pipe -fno-strict-aliasing -= Wreturn-type -fno-strict-aliasing -O2 -fvisibility=3Dhidden -fvisibility-in= lines-hidden -fPIC -DQT_SHARED -DJSC=3DQTJSC -Djscyyparse=3Dqtjscyyparse -D= jscyylex=3Dqtjscyylex -Djscyyerror=3Dqtjscyyerror -DWTF=3DQTWTF -DQT_BUILD_= SCRIPT_LIB -DQT_NO_USING_NAMESPACE -DQLALR_NO_QSCRIPTGRAMMAR_DEBUG_INFO -DQ= T_NO_CAST_TO_ASCII -DQT_ASCII_CAST_WARNINGS -DQT3_SUPPORT -DQT_MOC_COMPAT -= DBUILDING_QT__=3D1 -DNDEBUG -DBUILDING_QT__ -DBUILDING_JavaScriptCore -DBUI= LDING_WTF -DWTF_USE_JAVASCRIPTCORE_BINDINGS=3D1 -DWTF_CHANGES=3D1 -DNDEBUG = -DJS_EXPORT=3D -DJS_EXPORTDATA=3D -DQT_NO_DEBUG -DQT_CORE_LIB -D_LARGEFILE6= 4_SOURCE -D_LARGEFILE_SOURCE -I/usr/local/share/qt4/mkspecs/freebsd-g++ -I.= -I../../include/QtCore -I../../include -I../3rdparty/javascriptcore/JavaSc= riptCore -I../3rdparty/javascriptcore -I../3rdparty/javascriptcore/JavaScri= ptCore/assembler -I../3rdparty/javascriptcore/JavaScriptCore/bytecode -I../= 3rdparty/javascriptcore/JavaScriptCore/bytecompiler -I! > =A0../3rdparty/javascriptcore/JavaScriptCore/debugger -I../3rdparty/javas= criptcore/JavaScriptCore/interpreter -I../3rdparty/javascriptcore/JavaScrip= tCore/jit -I../3rdparty/javascriptcore/JavaScriptCore/parser -I../3rdparty/= javascriptcore/JavaScriptCore/profiler -I../3rdparty/javascriptcore/JavaScr= iptCore/runtime -I../3rdparty/javascriptcore/JavaScriptCore/wrec -I../3rdpa= rty/javascriptcore/JavaScriptCore/wtf -I../3rdparty/javascriptcore/JavaScri= ptCore/wtf/unicode -I../3rdparty/javascriptcore/JavaScriptCore/yarr -I../3r= dparty/javascriptcore/JavaScriptCore/API -I../3rdparty/javascriptcore/JavaS= criptCore/ForwardingHeaders -Igenerated -I../../include/QtScript -I.rcc/rel= ease-shared -I/usr/ports/work/usr/ports/devel/qt4-script/work/qt-everywhere= -opensource-src-4.6.2/src/3rdparty/javascriptcore/WebKit/qt/Api -I../3rdpar= ty/javascriptcore/JavaScriptCore/pcre -I/usr/ports/work/usr/ports/devel/qt4= -script/work/qt-everywhere-opensource-src-4.6.2/src/3rdparty/javascriptcore= /WebKitBuild/Relea! > =A0se/JavaScriptCore/tmp -I../3rdparty/javascriptcore/JavaScriptC! > =A0ore -I../3rdparty/javascriptcore/JavaScriptCore/parser -I../3rdparty/j= avascriptcore/JavaScriptCore/bytecompiler -I../3rdparty/javascriptcore/Java= ScriptCore/debugger -I../3rdparty/javascriptcore/JavaScriptCore/runtime -I.= ./3rdparty/javascriptcore/JavaScriptCore/wtf -I/usr/ports/work/usr/ports/de= vel/qt4-script/work/qt-everywhere-opensource-src-4.6.2/src/script/../3rdpar= ty/javascriptcore/JavaScriptCore/unicode -I../3rdparty/javascriptcore/JavaS= criptCore/interpreter -I../3rdparty/javascriptcore/JavaScriptCore/jit -I../= 3rdparty/javascriptcore/JavaScriptCore/profiler -I../3rdparty/javascriptcor= e/JavaScriptCore/wrec -I../3rdparty/javascriptcore/JavaScriptCore/API -I../= 3rdparty/javascriptcore/JavaScriptCore/bytecode -I../3rdparty/javascriptcor= e/JavaScriptCore/assembler -I../3rdparty/javascriptcore/JavaScriptCore/gene= rated -I. -Iparser -I.moc/release-shared -I/usr/local/include -o obj/releas= e/Interpreter.o ../3rdparty/javascriptcore/JavaScriptCore/interpreter/Inter= preter.cpp > > =A0 =A0 =A0 =A0[The above command line was extracted from the actual buil= d log > =A0 =A0 =A0 =A0and run manually with different options to check the probl= em] Worked for me on this machine, but I have over 12GB of RAM... so not really a fair comparison. $ ls -d /var/db/pkg/qt4-script-4.6.2/; uname -a; sysctl hw.physmem /var/db/pkg/qt4-script-4.6.2/ FreeBSD bioshock.cisco.com 9.0-CURRENT FreeBSD 9.0-CURRENT #2 r206173: Tue May 25 13:21:21 PDT 2010 root@bioshock.cisco.com:/usr/obj/usr/src/sys/BIOSHOCK amd64 hw.physmem: 12867121152 This is probably one of those ridiculous optimizer bugs like what happened with x11-servers/xorg-server back in the day when 4.2.1 was initially imported... Thanks, -Garrett
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201006040150.o541o3Vk093863>