Thanks very much for your useful comments.
And I've made a few adjustments accordingly:
1) In Makefile.in, the duplication of the "install_compiler" rule was
deleted but the ppc install_fe remained. Currently ppc doesn't
support c++, and it doesn't build the related components. ppc install
skips libgcc and libstdc++-v3 to avoid trying to install things doesn't
exist.
2) In osprey-gcc-4.2.0/gcc/Makefile.in, unused code was deleted.
3) In osprey-gcc-4.2.0/gcc/Makefile.in, the rule "install-common"
is duplicated so that ppc doesn't have to handle EXTRA_PARTS.
It's fine for x86 native but there's problem for x86-ppc cross
because EXTRA_PARTS isn't built.

The new patch is attached. Thanks again.




--

Best wishes.

Huan Luo
 



At 2012-04-09 15:58:24,"David Coakley" <dcoak...@gmail.com> wrote:
>I am glad you are adopting the configure-based build system for ppc.
>
>Here are my review comments:
>
>1) In Makefile.in, there is no need to duplicate the "install" rule.
>It looks like the "install_compiler" rule could be shared by rewriting
>it as:
>
>install_compiler:
>        export TOOLROOT=@prefix@; \
>        export TOP_SRCDIR=$(TOP_SRCDIR); \
>                $(TOP_SRCDIR)/install_compiler.sh $(MACHINE_TYPE) $(if
>$(BUILD_FORTRAN),"YES","NO") @CROSS_TARGET@
>
>
>Just curious, why does the ppc install skip libgcc and libstdc++-v3?
>
>2) In osprey-gcc-4.2.0/gcc/Makefile.in, my suggestion is to delete
>unused code rather than commenting it out.
>
>3) In osprey-gcc-4.2.0/gcc/Makefile.in, most of the large rule
>"install-common" is duplicated to add the EXTRA_PARTS handling.  Do
>you know of a problem with using EXTRA_PARTS for all targets?  I think
>it is fine for x86.
>
>-David Coakley / AMD Open Source Compiler Engineering
>
>2012/4/7 Huan Luo <luo_huan...@126.com>:
>>  Hi,
>>     This is a patch to trunk-r3898 which helps open64 x86-ppc cross compiler
>> generates its own gcc preprocessor instead using gcc of the native system.
>> Also it changes the way in which the cross compiler is installed.
>>     The test result on a x86 machine is available.
>>     Would you please help review this patch? Thanks very much.
>>
>> --
>> Best wishes.
>>
>> Huan Luo
>>
>>
>>

Attachment: patch1
Description: Binary data

------------------------------------------------------------------------------
Better than sec? Nothing is better than sec when it comes to
monitoring Big Data applications. Try Boundary one-second 
resolution app monitoring today. Free.
http://p.sf.net/sfu/Boundary-dev2dev
_______________________________________________
Open64-devel mailing list
Open64-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open64-devel

Reply via email to