The build changes in the new patch look ok to me. Thanks for the explanation.
-David Coakley / AMD Open Source Compiler Engineering 2011/11/28 Huan Luo <luo_huan...@126.com>: > Hi David, > Thanks very much for your comments. > In Makefile.in, ld-new wasn't left out on purpose and is added in now. > As for osprey-gcc-4.2.0/configure.in. > +if [ "`uname -m | sed -e s/i.86/i386/`" != "i386" ] > +then > FLAGS_FOR_TARGET= > +fi > This change has nothing to do with ppc target. > In earlier revisions, somebody added these line to configure rather > than configure.in. When I regenerated configure, these lines were > lost, and without which the build process failed. So I added them to > configure.in instead. > The target-libstdc++-v3 library > At the moment, our x86-ppc cross compiler doesn't support C++ very > well. We are still working on it. > Here's the new PatchFile. Thanks again. > -- > Huan Luo > > At 2011-11-27 08:20:10,"David Coakley" <dcoak...@gmail.com> wrote: >>Hi Huan, >> >>Here are my comments on the build changes in your patch. In configure.ac, >> >>+ GCC_CONFIGURE_CFLAGS+=" -g" >>+ GCC_CONFIGURE_COMPILER+=" -m32 -DTARG_PPC32" >> >>1) You do not need to set "-g" here. There is already code in >>configure.ac that sets "-g" according to the value of BUILD_OPTIMIZE. >>2) The "-DTARG_PPC32" should be part of GCC_CONFIGURE_CFLAGS to match >>the other targets. >> >>In Makefile.in, >> >>+ifneq ($(BUILD_TARGET), PPC32) >>+BASIC_COMPONENTS += $(NATIVE_BUILD_DIR_LD)/ld/ld-new >> >>3) Why is ld-new left out for PPC32? Please add a comment. >> >>For osprey-gcc-4.2.0/configure, >> >>4) It looks like you lost some copyright information when you >>regenerated this file. >> >>For osprey-gcc-4.2.0/configure.in, >> >>+if [ "`uname -m | sed -e s/i.86/i386/`" != "i386" ] >>+then >> FLAGS_FOR_TARGET= >>+fi >> >>5) Please add a comment explaining this change -- it's not clear how >>it is related to the ppc target. >> >>I am also curious why you needed to remove the build of the >>target-libstdc++-v3 library. Don't you need it to support C++? >> >>-David Coakley / AMD Open Source Compiler Engineering >> >> >>2011/11/25 Huan Luo <luo_huan...@126.com>: >>> Hi Rayson, >>> Thanks very much for your review. >>> About the question:"Can we still host the compiler on PPC Linux?" >>> No, not at the moment. But we are working on it and I believe >>> soon we are going to submit another patch to support it. >>> After applying the two changes you've requested, the >>> PatchFile now looks like this. >>> >>> Huan Luo >>> At 2011-11-24 23:55:37,"Rayson Ho" <raysonlo...@yahoo.com> wrote: >>> >>> Hi Huan, >>> I have a question - in your patch (configure & configure.ac): >>> >>> + MACHINE_TYPE=PPC32 >>> + BUILD_ABI=I32BIT >>> + BUILD_ARCH=IA32 >>> + BUILD_HOST=IA32 >>> + BUILD_TARGET=PPC32 >>> Can we still host the compiler on PPC Linux? >>> >>> Also, 2 minor cosmetic change requests from me. >>> >>> 1) osprey/be/cg/cgemit.cxx, change: >>> >>> +#if defined(TARG_PPC32) >>> +#else >>> >>> to: >>> +#if !defined(TARG_PPC32) >>> >>> 2) osprey/be/com/wn_lower.cxx >>> - WN_INSERT_BlockLast(block, lower_if(block, ifstmt, actions)); >>> + WN_INSERT_BlockLast(block, lower_if(block, ifstmt, actions ,NULL)); >>> >>> The space should follow the ",", so it should be: >>> - WN_INSERT_BlockLast(block, lower_if(block, ifstmt, actions)); >>> + WN_INSERT_BlockLast(block, lower_if(block, ifstmt, actions, NULL)); >>> >>> Rayson >>> >>> ================================= >>> Grid Engine / Open Grid Scheduler >>> http://gridscheduler.sourceforge.net/ >>> >>> Scalable Grid Engine Support Program >>> http://www.scalablelogic.com/ >>> >>> >>> ________________________________ >>> From: Huan Luo <luo_huan...@126.com> >>> To: open64 mailing list <open64-devel@lists.sourceforge.net> >>> Sent: Thursday, November 24, 2011 6:03 AM >>> Subject: [Open64-devel] Review request x86-ppc cross compile >>> >>> Hi, >>> Could a gatekeeper help us review this patchfile? Thanks. >>> We hope to submit this diff so that the x86-ppc cross compiler can work. >>> >>> Luo >>> >>> >>> ------------------------------------------------------------------------------ >>> All the data continuously generated in your IT infrastructure >>> contains a definitive record of customers, application performance, >>> security threats, fraudulent activity, and more. Splunk takes this >>> data and makes sense of it. IT sense. And common sense. >>> http://p.sf.net/sfu/splunk-novd2d >>> _______________________________________________ >>> Open64-devel mailing list >>> Open64-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/open64-devel >>> >>> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> All the data continuously generated in your IT infrastructure >>> contains a definitive record of customers, application performance, >>> security threats, fraudulent activity, and more. Splunk takes this >>> data and makes sense of it. IT sense. And common sense. >>> http://p.sf.net/sfu/splunk-novd2d >>> _______________________________________________ >>> Open64-devel mailing list >>> Open64-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/open64-devel >>> >>> > > > ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d _______________________________________________ Open64-devel mailing list Open64-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open64-devel
