More build oddities

Ralf Corsepius ralf.corsepius at rtems.org
Thu Apr 10 04:17:14 UTC 2008


On Wed, 2008-04-09 at 21:51 -0600, Daron Chabot wrote:
> On 9-Apr-08, at 9:23 PM, Nigel Spon wrote:
> >
> > On 10/04/2008, at 3:06 PM, Daron Chabot wrote:
> >> hmmm... try un-defining CC, then re-run your problematic script. I  
> >> vaguely recall some issue(s) when CC is defined...

Never set CC, CFLAGS etc. unless you know what you are doing ;)

Some people (esp. Windows user) are used to set these globally in their
environment, some broken IDE-wanna-be's also tend to do so, but ...
... don't do so, unless you really know what you are doing.


> > Well, thats interesting. I tried:
> 
> "interesting" is good... :-)
> 
> >
> >
> > 	unset CC
> > 	PATH="$BUILD_ROOT/bin:$PATH"
> > 	export PATH
> > 	cd b-rtems
> > 	make RTEMS_BSP="$BSP" all install
> >
> >
> > and got:
> >
> >
> > configure: creating make/virtex.cache
> > checking for RTEMS_CPU_MODEL... ppc405
> > checking for RTEMS_BSP_FAMILY... (cached) virtex
> > checking for CPU_CFLAGS... (cached) -mcpu=403 -Dppc405 -Dvirtex
> > checking for CFLAGS_OPTIMIZE_V... (cached) -O2 -g -fno-keep-inline- 
> > functions
> > checking for style of include used by make... GNU
> > checking for powerpc-rtems4.8-gcc... powerpc-rtems4.8-gcc
> > checking for powerpc-rtems4.8-gcc... (cached) powerpc-rtems4.8-gcc
> > checking for C compiler default output file name...
> > configure: error: C compiler cannot create executables
So far, this looks good.

> > But in config.log we now see:
> >
> >
> > configure:2950: checking for CFLAGS_OPTIMIZE_V
> > configure:2957: result: -O2 -g -fno-keep-inline-functions
> > configure:2978: checking for style of include used by make
> > configure:3006: result: GNU
> > configure:3038: checking for powerpc-rtems4.8-gcc
> > configure:3054: found /Users/frederickspon/Documents/RTEMSX_4.8/ 
> > RTEMS_Xcode/Build~~/bin/powerpc-rtems4.8-gcc
> > configure:3065: result: powerpc-rtems4.8-gcc
> > configure:3086: checking for powerpc-rtems4.8-gcc
> > configure:3113: result: powerpc-rtems4.8-gcc
> > configure:3391: checking for C compiler version
> > configure:3398: powerpc-rtems4.8-gcc --version >&5
> > powerpc-rtems4.8-gcc (GCC) 4.1.1
> > Copyright (C) 2006 Free Software Foundation, Inc.
> > This is free software; see the source for copying conditions.   
> > There is NO
> > warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR  
> > PURPOSE.
> >
> > configure:3401: $? = 0
> > configure:3408: powerpc-rtems4.8-gcc -v >&5
> > Using built-in specs.
> > Target: powerpc-rtems4.8
> > Configured with: /Users/frederickspon/Documents/RTEMSX_4.8/ 
> > RTEMS_Xcode/../tools/gcc-4.1.1/configure --target=powerpc-rtems4.8  
> > --prefix=/Users/frederickspon/Documents/RTEMSX_4.8/RTEMS_Xcode/ 
> > Build~~ --with-gnu-as --with-gnu-ld --with-newlib --verbose -- 
> > enable-threads --enable-languages=c,c++
> > Thread model: rtems
> > gcc version 4.1.1
> > configure:3411: $? = 0
> > configure:3418: powerpc-rtems4.8-gcc -V >&5
> > powerpc-rtems4.8-gcc: '-V' option must have argument
> > configure:3421: $? = 1
> > configure:3444: checking for C compiler default output file name
> > configure:3471: powerpc-rtems4.8-gcc -mcpu=403 -Dppc405 -Dvirtex - 
> > O2 -g -fno-keep-inline-functions   conftest.c  >&5
> > /Users/frederickspon/Documents/RTEMSX_4.8/RTEMS_Xcode/Build~~/lib/ 
> > gcc/powerpc-rtems4.8/4.1.1/../../../../powerpc-rtems4.8/bin/ld:  
> > crt0.o: No such file: No such file or directory
RTEMS this crt0.o should  come from newlib.

=> Your cross-toolchain is likely broken. Very likely, something related
to newlib and/or multilibs is broken. 


- Did you build newlib simultaneously to gcc?
- Did you apply rtems's gcc and newlib patches?
  For powerpc targets, they matter.
- Did you build gcc multilib'ed?


> > collect2: ld returned 1 exit status
> 
> The "~~" in that path worries  
> me...	                                                              ^^^
I fail to understand why anybody would use something like this ;)
It definitely means asking for trouble ...
... so far... I haven't spotted  anythin in Nigel's logs which is
clearly related to it.

Ralf





More information about the users mailing list