Re: rs6000:6000 versus powerpc:commom

[Alan Modra <amodra at bigpond dot net dot au>]

On Fri, Feb 14, 2003 at 09:39:40AM -0600, Joel Sherrill wrote:
> I've now tried that and it still produces rs6000:6000.  The only place I
> see rs6000:6000 in bfd is in cpu-rs6000.c which makes me think there
> must be some tie between coff-rs6000.c and the default architecture
> that I don't have a clue about.  

Sorry, my suggestion was a bit silly given that you'll get one of
the archs in cpu-rs6000.c.

> Here is what I know about what is expected based upon the examining
> headers of old toolset objects and executables:
> 
>  + for .o's,   architecture=rs6000:6000,    flags=0x031
>  + for .exe's, architecture=powerpc:common, flags=0x133
> 
> The new toolset has it mostly correct:
> 
>  + for .o's,   architecture=rs6000:6000,    flags=0x031
>  + for .exe's, architecture=rs6000:6000,    flags=0x133

This difference might not be particularly relevant.  I changed the
bfd_mach_* values at rev 1.57 of bfd/archures.c so that zero wasn't
used.  That change also gave different values to bfd_mach_ppc and
bfd_mach_rs6k, which were previously both zero.

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre