On powerpc 11.0-CURRENT CROSS_TOOLCHAIN=powerpc64-gcc fails: clang-tblgen use attempted before it exists

[Mark Millard]

I agree that my experiment is not what should be the 11.0-CURRENT official file contents by any means: it would break other contexts/uses for sure.

But I expect it was appropriate to identify one thing that contributes to the observed behavior of cc1plus reporting -std=c++11 as unrecognized. This does end up being separate from why gcc 4.2.1's cc1plus is in use through buildworld stage 1.2 in the first place (legacy and bootstrap-tools).


The only places that I see CC=${XCC} CXX=${XCXX} sorts of assignments for picking up and using cross compilation tools are:

> # $FreeBSD: head/Makefile.inc1 279328 2015-02-26 20:02:29Z emaste $
> ...
> WMAKEENV+=      CC="${XCC} ${XCFLAGS}" CXX="${XCXX} ${XCFLAGS} ${XCXXFLAGS}" \
> ...
> WMAKE=          ${WMAKEENV} ${MAKE} ${WORLD_FLAGS} -f Makefile.inc1 DESTDIR=${WORLDTMP}
> ...
> LIB32WMAKEFLAGS+= CC="${XCC} ${LIB32FLAGS}" \
>                 CXX="${XCXX} ${LIB32FLAGS}" \
> ...
> LIB32WMAKE=     ${LIB32WMAKEENV} ${MAKE} ${LIB32WMAKEFLAGS} \
> ...
> LIB32IMAKE=     ${LIB32WMAKE:NINSTALL=*:NDESTDIR=*:N_LDSCRIPTROOT=*} \
> ...
> KMAKEENV=       ${WMAKEENV}
> KMAKE=          ${KMAKEENV} ${MAKE} ${.MAKEFLAGS} ${KERNEL_FLAGS} KERNEL=${INSTKERNNAME}

None of these appear to be involved in any of the following...

>         ${_+_}cd ${.CURDIR}; ${BMAKE} legacy
> ...
>         ${_+_}cd ${.CURDIR}; ${BMAKE} bootstrap-tools
> ...
>         ${_+_}cd ${.CURDIR}; ${TMAKE} build-tools
> ...
>         ${_+_}cd ${.CURDIR}; ${XMAKE} cross-tools
>         ${_+_}cd ${.CURDIR}; ${XMAKE} kernel-tools
> ...
>         ${_+_}cd ${.CURDIR}; ${KTMAKE} kernel-tools
> ...
> .for _dir in lib/ncurses/ncurses lib/ncurses/ncursesw lib/libmagic
>         cd ${.CURDIR}/${_dir}; \
>             WORLDTMP=${WORLDTMP} \
>             MAKEFLAGS="-m ${.CURDIR}/tools/build/mk ${.MAKEFLAGS}" \
>             MAKEOBJDIRPREFIX=${LIB32_OBJTREE} ${MAKE} SSP_CFLAGS= DESTDIR= \
>             DIRPRFX=${_dir}/ -DNO_LINT -DNO_CPU_CFLAGS MK_WARNS=no MK_CTF=no \
>             build-tools
> .endfor



Also, unfortunately, WITHOUT_CLANG_BOOTSTRAP= mixed with WITH_CLANG= gets bootstrap-tools time frame clang-compilation activity just by the structure of Makefile.inc1 ...

> # $FreeBSD: head/Makefile.inc1 279328 2015-02-26 20:02:29Z emaste $
> ...
> _bt=            _bootstrap-tools
> ...
> # We need to build tblgen when we're building clang either as
> # the bootstrap compiler, or as the part of the normal build.
> .if ${MK_CLANG_BOOTSTRAP} != "no" || ${MK_CLANG} != "no"
> _clang_tblgen= \
>         lib/clang/libllvmsupport \
>         lib/clang/libllvmtablegen \
>         usr.bin/clang/tblgen \
>         usr.bin/clang/clang-tblgen
> 
> ${_bt}-usr.bin/clang/clang-tblgen: ${_bt}-lib/clang/libllvmtablegen ${_bt}-lib/clang/libllvmsupport
> ${_bt}-usr.bin/clang/tblgen: ${_bt}-lib/clang/libllvmtablegen ${_bt}-lib/clang/libllvmsupport
> .endif
> ...
> .for _tool in \
>     ${_clang_tblgen} \
> ...
> ${_bt}-${_tool}: .PHONY .MAKE
>         ${_+_}@${ECHODIR} "===> ${_tool} (obj,depend,all,install)"; \
>                 cd ${.CURDIR}/${_tool} && \
>                 ${MAKE} DIRPRFX=${_tool}/ obj && \
>                 ${MAKE} DIRPRFX=${_tool}/ depend && \
>                 ${MAKE} DIRPRFX=${_tool}/ all && \
>                 ${MAKE} DIRPRFX=${_tool}/ DESTDIR=${MAKEOBJDIRPREFIX}/legacy install
> 
> bootstrap-tools: ${_bt}-${_tool}
> .endfor


I've reverted to trying CROSS_TOOLCHAIN=powerpc64-gcc using WITHOUT_CLANG_BOOTSTRAP= and WITHOUT_CLANG= since even without the -std=c++11 issue (via my experiment) it tries gcc 4.2.1 for compiling part of clang during stage 1.2 --and that fails.




===
Mark Millard
markmi at dsl-only.net

On 2015-Mar-16, at 04:51 AM, Warner Losh <imp at bsdimp.com> wrote:


> On Mar 16, 2015, at 7:02 PM, Dimitry Andric <dim at freebsd.org> wrote:
> 
> On 16 Mar 2015, at 09:02, Mark Millard <markmi at dsl-only.net> wrote:
>> 
>> I found why gcc 4.2.1's cc1plus was getting -std=c++11 for the CROSS_TOOLCHAIN=powerpc64-gcc compiles that involve WITH_CLANG= . (WITHOUT_CLANG= does not get the "unrecognized" notices.) There is a global assignment to CXXFLAGS for all compilers whenever clang.build.mk is in use (showing my experimental change...):
>> 
>> # svnlite diff /usr/srcC/lib/clang/clang.build.mk
>> Index: /usr/srcC/lib/clang/clang.build.mk
>> ===================================================================
>> --- /usr/srcC/lib/clang/clang.build.mk	(revision 279514)
>> +++ /usr/srcC/lib/clang/clang.build.mk	(working copy)
>> @@ -34,8 +34,8 @@
>> CFLAGS+=	-DLLVM_DEFAULT_TARGET_TRIPLE=\"${TARGET_TRIPLE}\" \
>> 		-DLLVM_HOST_TRIPLE=\"${BUILD_TRIPLE}\" \
>> 		-DDEFAULT_SYSROOT=\"${TOOLS_PREFIX}\"
>> -CXXFLAGS+=	-std=c++11 -fno-exceptions -fno-rtti
>> -CXXFLAGS.clang+= -stdlib=libc++
>> +CXXFLAGS+=	-fno-exceptions -fno-rtti
>> +CXXFLAGS.clang+= -std=c++11 -stdlib=libc++
>> 
>> .PATH:	${LLVM_SRCS}/${SRCDIR}
>> 
>> It may be that the "-fno-exceptions -fno-rtti" are also suspect for being not limited to clang contexts.
> 
> This is incorrect.  Clang needs -std=c++11, otherwise it cannot compile.
> 
> I suspect you also need WITHOUT_CLANG_BOOTSTRAP (and WITHOUT_GCC_BOOTSTRAP, probably).

From earlier debugging, I’m pretty sure that the wrong g++ is being used in the CROSS_TOOLCHAIN case. I’m in Japan right now on travel, so I haven’t been able to track it down. WITHOUT_GCC_BOOTSTRAP should be implied in that case, but if not, that might explain why...

Warner