Francisco Jerez schrieb:
> Chí-Thanh Christopher Nguyễn <chith...@gentoo.org> writes:
>
>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68504 Bugzilla:
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58241 CC:
>> mesa-sta...@lists.freedesktop.org --- configure.ac
>> | 11 +++++++++++ src/gallium/state_trackers/clover/Makefile.am | 2
>> +- 2 files changed, 12 insertions(+), 1 deletion(-)
>>
>> diff --git a/configure.ac b/configure.ac index 40e052c..15831fd
>> 100644 --- a/configure.ac +++ b/configure.ac @@ -1390,6 +1390,17 @@ if
>> test "x$enable_opencl" = xyes; then AC_MSG_ERROR([pkg-config cannot
>> use libclc.pc which is required to build clover]) fi
>>
>> +dnl +dnl Workaround for altivec failure with GNU C++ and -std=c++0x,
>> PR58241 +dnl + + if test "x${ac_cv_cxx_compiler_gnu}" = xyes ;
>> then + CLOVER_CXX_STD="gnu++0x" + else +
>> CLOVER_CXX_STD="c++0x" + fi + AC_SUBST([CLOVER_CXX_STD]) +
>> GALLIUM_STATE_TRACKERS_DIRS="$GALLIUM_STATE_TRACKERS_DIRS clover"
>> GALLIUM_TARGET_DIRS="$GALLIUM_TARGET_DIRS opencl"
>> enable_gallium_loader=yes diff --git
>> a/src/gallium/state_trackers/clover/Makefile.am
>> b/src/gallium/state_trackers/clover/Makefile.am index b4c197a..e339c1f
>> 100644 --- a/src/gallium/state_trackers/clover/Makefile.am +++
>> b/src/gallium/state_trackers/clover/Makefile.am @@ -30,7 +30,7 @@
>> libclllvm_la_SOURCES = \ llvm/invocation.cpp
>>
>> libclover_la_CXXFLAGS = \ - -std=c++0x + -std=$(CLOVER_CXX_STD)
>>
>> libclover_la_LIBADD = \ libcltgsi.la libclllvm.la
>
> How about disabling AltiVec code generation by adding "-mno-altivec" to
> CXXFLAGS? AFAICT what's going on is that the "-maltivec" option (which
> is implied by your "-mcpu=7450") enables a number of language
> extensions as a side effect that collide with the C++ standard, which
> is exacerbated by GCC's seemingly broken implementation.
Then what is the purpose of the altivec code in cl_platform.h?
> I wonder if they could have any undesirable side effects on components
> other than clover: not including "altivec.h" doesn't seem to guarantee
> that the problematic preprocessor definitions ("bool" and "vector")
> aren't going to be present.
>
> Switching the dialect we're compiling with to GNU C++ isn't especially
> reassuring, clover doesn't need GNU's C++ extensions and we don't want
> to start using any of them by accident -- we don't use the AltiVec
> extensions either so disabling the latter seems like the most
> reasonable solution to me.
Unintentional use of GNU C++ extensions could be mitigated by adding
-Wpedantic (although the gcc manpage admits that this check is not complete).
I have no strong opinion either way.
Best regards,
Chí-Thanh Christopher Nguyễn
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
