[PATCH] PPC Floating point symbols defined both in libgcc and glibc

[Ranjani Murthy]

Some floating point symbols are defined in both libgcc and libc and
they produce mulitple definition errors while compiling. This patch
removes them from the libc library.  Patch and test case are below:


--- libc/sysdeps/powerpc/nofpu/Makefile.orig     
+++ libc/sysdeps/powerpc/nofpu/Makefile  
@@ -1,7 +1,7 @@
  # Makefile fragment for PowerPC with no FPU.
                                                                                
  ifeq ($(subdir),soft-fp)
-sysdep_routines += $(gcc-single-routines) $(gcc-double-routines) sim-full
+sysdep_routines += sim-full
  endif
                                                                                
  ifeq ($(subdir),math)


Test case to reproduce the problem:

#include <stdio.h>
                                                                                
extern int __ledf2(double, double);
                                                                                
int main()
{
        return __ledf2(0, 0);
}