Re: biarch gcc

[Peter Bergner <bergner at vnet dot ibm dot com>]

Andreas Jaeger wrote:
> Alan Modra <amodra@bigpond.net.au> writes:
> > Secondly, getting header files right is not so easy.  Typically, some 32
> > bit headers are different from the corresponding 64 bit header.  It's
> > possible to hack around this problem by creating lots of stubs that test
> > __powerpc64__ and then include the real header, or by using esoteric
> > -isystem gcc options.  You also need to deal with patched headers
> > installed by fixincludes, which might get in the way of such hacks.
>
>
> The Glibc headers for sparc, s390 and x86_64 were written with that in
> mind and do not have any problems that I'm aware of.  What problems
> exactly do you have?

I believe Alan was talking about kernel headers here and not GLIBC headers.
The stub workaround Alan speaks of is the same workaround you (ie, SuSE) used
on SLES8.  For example:

    cannon% cat /usr/include/asm/unistd.h
    #ifndef __PPC_ASM_STUB__UNISTD_H__
    #define __PPC_ASM_STUB__UNISTD_H__
    #ifdef __powerpc64__
    #include <asm-ppc64/unistd.h>
    #else
    #include <asm-ppc/unistd.h>
    #endif
    #endif


Peter