Patch to make gcc-3.2.2 compile for i386-rtems target

[Joel Sherrill]

Valette Eric wrote:
> 
> As joel said, in theory there was not patches needed for gcc 3.2.2, I
> decided to give it a try as well a the latest newlib.

Sorry.  This one is needed.  Someone broke i386-elf and i386-rtems after
my testing finished.  I knew about this one and it actually had a PR #
which I fixed and closed.

> I do not know if ralf has anything pending but when trying to compile
> gcc-3.2.2 to try to hunt the bug, I found the following patch was
> necessary (allthough I do not rellay understand why it is not in
> i386elf.h...

Just a slip up with a last minute patch.  I was guilty of breaking
Ada builds on the ia64-linux target.  Apparently, there wasn't enough
of a "hands off testing" period before 3.2.2 went out the door.  This
is being resolved with a 2 week no change period for 3.2.3.

> 
> --
>     __
>    /  `                         Eric Valette
>   /--   __  o _.                6 rue Paul Le Flem
> (___, / (_(_(__                 35740 Pace
> 
> Tel: +33 (0)2 99 85 26 76       Fax: +33 (0)2 99 85 26 76
> E-mail: eric.valette at free.fr
> 
>   ------------------------------------------------------------------------
> *** gcc-3.2.2/gcc/config/i386/rtemself.h~       Fri Apr 12 15:09:28 2002
> --- gcc-3.2.2/gcc/config/i386/rtemself.h        Mon Mar 17 12:01:54 2003
> ***************
> *** 29,31 ****
> --- 29,35 ----
> 
>   #undef CPP_SPEC
>   #define CPP_SPEC "%(cpp_cpu) %{msoft-float:-D_SOFT_FLOAT}"
> + /* The svr4 ABI for the i386 says that records and unions are returned
> +    in memory.  */
> + #undef DEFAULT_PCC_STRUCT_RETURN
> + #define DEFAULT_PCC_STRUCT_RETURN 1