Combine simplify_set WORD_REGISTER_OPERATIONS

[Alan Modra]

On Sun, Jan 31, 2016 at 06:02:35PM -0600, Segher Boessenkool wrote:
> On Mon, Feb 01, 2016 at 08:46:42AM +1030, Alan Modra wrote:
> > The comment says this test is supposed to prevent "a narrower
> > operation than requested", but it actually only allows a larger
> > subreg, not one the same size.  Fix that.
> > 
> > Bootstrapped and regression tested powerpc64-linux.  OK for stage1?
> > 
> > Note that this bug was found when investigating why gcc-6 does not
> > suffer from pr69548, ie. this bug was masking a powerpc backend bug.
> 
> It sounds like you have a testcase, can we see it please?

The testcase in pr69548 will show changes in rtl..

> And, just a missed optimisation, not a bug, right?

Yes, not a bug, and only presumed a missed optimisation.  I don't
actually have a testcase that shows worse code.  All I have is a
comment that makes sense to me, that doesn't agree exactly with the
code, and some understanding how the code may have been accidentally
written the way it is.

-- 
Alan Modra
Australia Development Lab, IBM