Bizarre error with mixed C and Fortran

[Bernd Lütkenhöner <bernd dot lutkenhoner at uni-muenster dot de>]

Dear All,

After a system upgrade, I had to replace the G77 compiler by GCC
(version 4.3.2). Unfortunately, some programs show bizarre errors now.
I finally managed to isolate one of these errors and to reproduce
it in a very simple test environment, consisting of one C and
one Fortran module.

Here is a scheme of the calling sequence (routines F* written
in Fortran, routines C* written in C):

    main()  --->  F1  --->  C1  --->  F2  --->  CX

The output generated by routine C1 is:

1.000000 2.000000 3.000000   0.000000 0.000000 0.000000 -0.000000
nan 2.000000 3.000000   0.000000 0.000000 0.000000 -0.000000 -0.000000
1.000000 2.000000 3.000000   0.000000 0.000000 0.000000 -0.000000 -0.000000

Note the NaN in the second line (the last two lines should be
identical).

If routine CX is replaced by FORTRAN routine FX, i.e.,

    main()  --->  F1  --->  C1  --->  F2  --->  FX

the error disappears. There seems to be a bug either in the compiler
or the linker. Any workaround (e.g. a compiler option)?

Many thanks,
   Bernd


########################################################################
Here are the two modules (these two files, and the Makefile, could
also be made available in a TAR or ZIP file).



/***************************************/
/* The C module                        */
/***************************************/
#include <stdio.h>
#include <stdlib.h>

extern void f1_(void);
extern void f2_(double *r);


int main(void) {
   f1_();   /* FORTRAN routine */
}


void c1_(void)   /* called by FORTRAN routine f1 */
{
   double r[3],x[5];

   f2_(r);   /* FORTRAN routine */

   /* works fine (uncommenting has no influence on subsequent error) */
   printf ("%f %f %f   %f %f %f %f\n",
        r[0],r[1],r[2],x[0],x[1],x[2],x[3]);

   /*!!!!!! prints NaN for r[0] !!!!!!*/
   printf ("%f %f %f   %f %f %f %f %f\n",
        r[0],r[1],r[2],x[0],x[1],x[2],x[3],x[4]);

   /* works fine (same statement as before, just called a second time) */
   printf ("%f %f %f   %f %f %f %f %f\n",
        r[0],r[1],r[2],x[0],x[1],x[2],x[3],x[4]);
}


double cx_(double *r)      /* called by FORTRAN routine f2 */
{
   r[0] = 1;
   r[1] = 2;
   r[2] = 3;
}



c***************************************/
c*  The FORTRAN module                 */
c***************************************/

      subroutine f1
      call c1
      return
      end


      SUBROUTINE f2 (r)
      REAL*8 r(3)

c  calling C-routine cx causes a problem,
      call cx (r)

c  no problem if routine fx (see below) is called instead of C-routine cx
c     call fx (r)

      return
      END


      SUBROUTINE fx (r)
      REAL*8 r(3)
      r(1) = 1
      r(2) = 2
      r(3) = 3
      return
      end


########################################################################
Finally, some information about the version of GCC that I used:

>gcc -v
Using built-in specs.
Target: i586-suse-linux
Configured with: ../configure --prefix=/usr --infodir=/usr/share/info
--mandir=/usr/share/man --libdir=/usr/lib --libexecdir=/usr/lib
--enable-languages=c,c++,objc,fortran,obj-c++,java,ada
--enable-checking=release --with-gxx-include-dir=/usr/include/c++/4.3
--enable-ssp --disable-libssp --with-bugurl=http://bugs.opensuse.org/
--with-pkgversion='SUSE Linux' --disable-libgcj --disable-libmudflap
--with-slibdir=/lib --with-system-zlib --enable-__cxa_atexit
--enable-libstdcxx-allocator=new --disable-libstdcxx-pch
--enable-version-specific-runtime-libs --program-suffix=-4.3
--enable-linux-futex --without-system-libunwind --with-cpu=generic
--build=i586-suse-linux
Thread model: posix
gcc version 4.3.2 [gcc-4_3-branch revision 141291] (SUSE Linux)