still problem building cross-compiler, x86 -> solaris2.8

still problem building cross-compiler, x86 -> solaris2.8

[Per Fransson]

Hi,

I've been trying to build a cross-compiler on an x86 with 
sparc-sun-solaris2.8 as target, to no avail. I followed the advice I got 
on this list (thank you) to put the include and lib directories from 
solaris in my prefix/sparc-sun-solaris2.8 directory. It now no longer 
gives same error message as it did before, it gives me the one below 
instead. I have already successfully installed binutils in the same 
prefix directory and configured gcc with the following command line (in 
my build directory):

$ ../gcc-3.0.4/configure --prefix=foo/myGCC/ 
--target=sparc-sun-solaris2.8 --without-newlib -v
$ make all install
...
...
...


foo/build-gcc/gcc/xgcc -Bfoo/build-gcc/gcc/ -Bfoo/myGCC//sparc-sun-solaris2.8/bin/ -Bfoo/myGCC//sparc-sun-solaris2.8/lib/ -isystem foo/myGCC//sparc-sun-solaris2.8/include -O2  -DCROSS_COMPILE -DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -isystem ./include  -fPIC -g1 -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -Dinhibit_libc -I. -I. -I../../gcc-3.0.4/gcc -I../../gcc-3.0.4/gcc/. -I../../gcc-3.0.4/gcc/config -I../../gcc-3.0.4/gcc/../include  -DL__gcc_bcmp -c ../../gcc-3.0.4/gcc/libgcc2.c -o libgcc/./__gcc_bcmp.o
rm -rf ./libgcc.a
sparc-sun-solaris2.8-ar  rc ./libgcc.a libgcc/./_muldi3.o libgcc/./_divdi3.o libgcc/./_moddi3.o libgcc/./_udivdi3.o libgcc/./_umoddi3.o libgcc/./_negdi2.o libgcc/./_lshrdi3.o libgcc/./_ashldi3.o libgcc/./_ashrdi3.o libgcc/./_ffsdi2.o libgcc/./_clz.o libgcc/./_udiv_w_sdiv.o libgcc/./_udivmoddi4.o libgcc/./_cmpdi2.o libgcc/./_ucmpdi2.o libgcc/./_floatdidf.o libgcc/./_floatdisf.o libgcc/./_fixunsdfsi.o libgcc/./_fixunssfsi.o libgcc/./_fixunsdfdi.o libgcc/./_fixdfdi.o libgcc/./_fixunssfdi.o libgcc/./_fixsfdi.o libgcc/./_fixxfdi.o libgcc/./_fixunsxfdi.o libgcc/./_floatdixf.o libgcc/./_fixunsxfsi.o libgcc/./_fixtfdi.o libgcc/./_fixunstfdi.o libgcc/./_floatditf.o libgcc/./_clear_cache.o libgcc/./_trampoline.o libgcc/./__main.o libgcc/./_exit.o libgcc/./_absvsi2.o libgcc/./_absvdi2.o libgcc/./_addvsi3.o libgcc/./_addvdi3.o libgcc/./_subvsi3.o libgcc/./_subvdi3.o libgcc/./_mulvsi3.o libgcc/./_mulvdi3.o libgcc/./_negvsi2.o libgcc/./_negvdi2.o libgcc/./_ctors.o libgcc/./unwind-dw2.o libgcc/./unwind-dw2-fde.o libgcc/./unwind-sjlj.o libgcc/./_eprintf.o libgcc/./_bb.o libgcc/./__gcc_bcmp.o
if [ -f sparc-sun-solaris2.8-ranlib ] || ( [ i686-pc-linux-gnu = sparc-sun-solaris2.8 ] && [ -f /usr/bin/ranlib -o -f /bin/ranlib ] ) ; then \
  sparc-sun-solaris2.8-ranlib ./libgcc.a ; \
else true; fi;
{ sparc-sun-solaris2.8-nm -pg  libgcc/./_muldi3.o libgcc/./_divdi3.o libgcc/./_moddi3.o libgcc/./_udivdi3.o libgcc/./_umoddi3.o libgcc/./_negdi2.o libgcc/./_lshrdi3.o libgcc/./_ashldi3.o libgcc/./_ashrdi3.o libgcc/./_ffsdi2.o libgcc/./_clz.o libgcc/./_udiv_w_sdiv.o libgcc/./_udivmoddi4.o libgcc/./_cmpdi2.o libgcc/./_ucmpdi2.o libgcc/./_floatdidf.o libgcc/./_floatdisf.o libgcc/./_fixunsdfsi.o libgcc/./_fixunssfsi.o libgcc/./_fixunsdfdi.o libgcc/./_fixdfdi.o libgcc/./_fixunssfdi.o libgcc/./_fixsfdi.o libgcc/./_fixxfdi.o libgcc/./_fixunsxfdi.o libgcc/./_floatdixf.o libgcc/./_fixunsxfsi.o libgcc/./_fixtfdi.o libgcc/./_fixunstfdi.o libgcc/./_floatditf.o libgcc/./_clear_cache.o libgcc/./_trampoline.o libgcc/./__main.o libgcc/./_exit.o libgcc/./_absvsi2.o libgcc/./_absvdi2.o libgcc/./_addvsi3.o libgcc/./_addvdi3.o libgcc/./_subvsi3.o libgcc/./_subvdi3.o libgcc/./_mulvsi3.o libgcc/./_mulvdi3.o libgcc/./_negvsi2.o libgcc/./_negvdi2.o libgcc/./_ctors.o libgcc/./unwind-dw2.o libgcc/./unwind-dw2-fde.o libgcc/./unwind-sjlj.o; echo %%; cat ../../gcc-3.0.4/gcc/libgcc-std.ver; } | gawk -f ../../gcc-3.0.4/gcc/mkmap-symver.awk > tmp-libgcc.map
mv tmp-libgcc.map libgcc.map
foo/build-gcc/gcc/xgcc -Bfoo/build-gcc/gcc/ -Bfoo/myGCC//sparc-sun-solaris2.8/bin/ -Bfoo/myGCC//sparc-sun-solaris2.8/lib/ -isystem foo/myGCC//sparc-sun-solaris2.8/include -O2  -DCROSS_COMPILE -DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -isystem ./include  -fPIC -g1 -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -Dinhibit_libc -shared -nodefaultlibs -Wl,-h,libgcc_s.so.1 -Wl,-z,text -Wl,-z,defs -Wl,-M,libgcc.map -o libgcc_s.so   libgcc/./_muldi3.o libgcc/./_divdi3.o libgcc/./_moddi3.o libgcc/./_udivdi3.o libgcc/./_umoddi3.o libgcc/./_negdi2.o libgcc/./_lshrdi3.o libgcc/./_ashldi3.o libgcc/./_ashrdi3.o libgcc/./_ffsdi2.o libgcc/./_clz.o libgcc/./_udiv_w_sdiv.o libgcc/./_udivmoddi4.o libgcc/./_cmpdi2.o libgcc/./_ucmpdi2.o libgcc/./_floatdidf.o libgcc/./_floatdisf.o libgcc/./_fixunsdfsi.o libgcc/./_fixunssfsi.o libgcc/./_fixunsdfdi.o libgcc/./_fixdfdi.o libgcc/./_fixunssfdi.o libgcc/./_fixsfdi.o libgcc/./_fixxfdi.o libgcc/./_fixunsxfdi.o libgcc/./_floatdixf.o libgcc/./_fixunsxfsi.o libgcc/./_fixtfdi.o libgcc/./_fixunstfdi.o libgcc/./_floatditf.o libgcc/./_clear_cache.o libgcc/./_trampoline.o libgcc/./__main.o libgcc/./_exit.o libgcc/./_absvsi2.o libgcc/./_absvdi2.o libgcc/./_addvsi3.o libgcc/./_addvdi3.o libgcc/./_subvsi3.o libgcc/./_subvdi3.o libgcc/./_mulvsi3.o libgcc/./_mulvdi3.o libgcc/./_negvsi2.o libgcc/./_negvdi2.o libgcc/./_ctors.o libgcc/./unwind-dw2.o libgcc/./unwind-dw2-fde.o libgcc/./unwind-sjlj.o -lc && rm -f libgcc_s.so.1 && ln -s libgcc_s.so libgcc_s.so.1
foo/myGCC//sparc-sun-solaris2.8/bin/ld:libgcc.map: file format not recognized; treating as linker script
foo/myGCC//sparc-sun-solaris2.8/bin/ld:libgcc.map:1: parse error
collect2: ld returned 1 exit status
make[2]: *** [libgcc_s.so] Error 1
make[2]: Leaving directory `foo/build-gcc/gcc'
make[1]: *** [libgcc.a] Error 2
make[1]: Leaving directory `foo/build-gcc/gcc'
make: *** [all-gcc] Error 2




I'm at my wit's end. Does anyone know what I can try next?

By the way, I get the same error when I try building a native gcc on the 
solaris system which is the target.

I'm thankful for any advice.

/Per

Re: still problem building cross-compiler, x86 -> solaris2.8

[Richard Henderson]

On Tue, Apr 09, 2002 at 08:35:07PM +0200, Per Fransson wrote:
> foo/myGCC//sparc-sun-solaris2.8/bin/ld:libgcc.map:1: parse error

You've found the wrong nm program, and produced an incorrect map file.


r~

Re: still problem building cross-compiler, x86 -> solaris2.8

[Per Fransson]

So which one is the right one? Are we talking about the nm program for 
solaris or the native one?

/Per

Richard Henderson wrote:

> You've found the wrong nm program, and produced an incorrect map file.

Re: still problem building cross-compiler, x86 -> solaris2.8

[Richard Henderson]

On Tue, Apr 09, 2002 at 08:55:04PM +0200, Per Fransson wrote:
> So which one is the right one? Are we talking about the nm program for 
> solaris or the native one?

The nm for solaris.


r~

Re: still problem building cross-compiler, x86 -> solaris2.8

[Hans-Peter Nilsson]

On Tue, 9 Apr 2002, Daniel Jacobowitz wrote:
> On Tue, Apr 09, 2002 at 09:31:51PM +0200, Per Fransson wrote:
> > So, while building a cross-compiler a call to "a binutil" is supposed to
> > yield the target-util? So I need to put prefix/sparc-sun-solaris28/bin
> > in PATH before the path that leads to the native binutils, is this correct?
> >
> > If not, how am I supposed to make it find the right binutil for a
> > certain point in the build?
>
> Try setting NM in your environment?  And likewise for the others.

NM_FOR_TARGET (et al.), that is.

brgds, H-P

Re: still problem building cross-compiler, x86 -> solaris2.8

[Per Fransson]

Yes they are available. I can run them from exactly the same prompt 
where I write the configure and make commands. To avoid confusion: I'm 
now talking about sparc-sun-solaris2.8-{as,nm,ar,objcopy,....}.

Again: is it out of the question that the error can have some other origin?


/Per

Re: still problem building cross-compiler, x86 -> solaris2.8

[David Edelsohn]

>>>>> Per Fransson writes:

Per> But prefix/bin/ , which contains 
Per> sparc-sun-solaris2.8-{as,nm,ar,objcopy,....} *is* in the PATH, it should 
Per> be able to find them if that is the crux. (Sorry to be bothering you 
Per> like this. I assume you'll just ignore me if it's taking up too much of 
Per> your time.)

	Are you sure that those applications are in the PATH of the
process performing the build?  Are you sure that those applications are
executable?  If the applications should be available, then you need to do
some detective work why those applications are not being recognized at the
appropriate point in the build.

David

Re: still problem building cross-compiler, x86 -> solaris2.8

[Per Fransson]

But prefix/bin/ , which contains 
sparc-sun-solaris2.8-{as,nm,ar,objcopy,....} *is* in the PATH, it should 
be able to find them if that is the crux. (Sorry to be bothering you 
like this. I assume you'll just ignore me if it's taking up too much of 
your time.)

/Per

Re: still problem building cross-compiler, x86 -> solaris2.8

[Richard Henderson]

On Tue, Apr 09, 2002 at 11:04:32PM +0200, Per Fransson wrote:
> How is it supposed to tell them apart if it needs both sets of binutils? 

NM_FOR_TARGET = ` \
  if [ -f ./nm ] ; then \
    echo ./nm ; \
  elif [ -f $(objdir)/../binutils/nm-new ] ; then \
    echo $(objdir)/../binutils/nm-new ; \
  else \
    if [ "$(host_canonical)" = "$(target)" ] ; then \
      echo nm; \
    else \
       t='$(program_transform_cross_name)'; echo nm | sed -e $$t ; \
    fi; \
  fi`

So: gcc expects the thing to be called either sparc-sun-solaris2.7-nm
or you should build inside a single tree so that it's present in 
../binutils/.


r~

Re: still problem building cross-compiler, x86 -> solaris2.8

[Per Fransson]

It results in the same error even when I make sure 'nm' yields the 
sparc-sun-solaris2.8-nm and after setting the ADDR2LINE, AR, AS, GASP, 
LD, NM, OBJCOPY, OBJDUMP, RANLIB, READELF, SIZE, STRINGS and STRIP 
environment variables to 
prefix/bin/sparc-sun-solaris2.8-addr2line,ar,as,gasp,ld,nm,objcopy,objdump,ranlib,readelf,size,stings 
and strip respectively.

I'm not really clear about this. Does it need both the native and the 
target binutils during the build but calls them both 'nm' (for example)? 
How is it supposed to tell them apart if it needs both sets of binutils? 
Can't it be something else screwing up the map file?

/Per

Re: still problem building cross-compiler, x86 -> solaris2.8

[Daniel Jacobowitz]

On Tue, Apr 09, 2002 at 09:31:51PM +0200, Per Fransson wrote:
> So, while building a cross-compiler a call to "a binutil" is supposed to 
> yield the target-util? So I need to put prefix/sparc-sun-solaris28/bin 
> in PATH before the path that leads to the native binutils, is this correct?
> 
> If not, how am I supposed to make it find the right binutil for a 
> certain point in the build?

Try setting NM in your environment?  And likewise for the others.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: still problem building cross-compiler, x86 -> solaris2.8

[Per Fransson]

So, while building a cross-compiler a call to "a binutil" is supposed to 
yield the target-util? So I need to put prefix/sparc-sun-solaris28/bin 
in PATH before the path that leads to the native binutils, is this correct?

If not, how am I supposed to make it find the right binutil for a 
certain point in the build?

/Per