Fix tlbsx instruction for PPC440

[Alan Modra]

On Mon, Sep 27, 2004 at 01:27:39PM -0500, Kumar Gala wrote:
> 
> On Sep 27, 2004, at 1:18 PM, Zack Weinberg wrote:
> 
> >Alan Modra <amodra@bigpond.net.au> writes:
> >
> >> On Sat, Sep 25, 2004 at 11:52:13AM -0700, Zack Weinberg wrote:
> > >> It seems simplest to me to give this entire instruction family an
> > >> optional RT register, and not worry about which instructions are
> >>> supported on which subarchitectures - hence PPC403|PPC440|BOOKE.
> >>
> > > Works for me.  The architecture book says that where the result is
> > > stored is implementation dependent.
> >
> >Thanks.  Here's a patch against binutils HEAD.  I made the tests a bit
> > more comprehensive and put them with the other tlbsx tests (hence the
> > much longer delta to booke.d).  I'm still wondering whether it's
> > necessary to say PPC403|PPC440|BOOKE, or whether some subset of that
> > would do.
> 
> I think you could drop the PPC440 and just make it PPC403|BOOKE.

Agreed.  Patch OK with that change.

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre