[Bug middle-end/51782] -ftree-sra: Missing address-space information leads to wrong

["jamborm at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51782

Martin Jambor <jamborm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|tree-optimization           |middle-end

--- Comment #23 from Martin Jambor <jamborm at gcc dot gnu.org> 2012-02-21 10:35:58 UTC ---
(In reply to comment #21)
> On Mon, 20 Feb 2012, jamborm at gcc dot gnu.org wrote:
> > Perhaps get_base_address misses a DECL_P in the condition looking into
> > MEM_REFs?
> 
> No, sure not - in the above we have a component-ref inside the
> ADDR_EXPR.

I know it's invalid, I just thought it would be better to return the
MEM_REF rather than NULL_TREE anyway.  Most of checks whether the zero
argument is ADDR_EXPR are accompanied with a DECL_P of its argument
check too (e.g. the new mem_ref_refers_to_non_mem_p).  However, now I
see we don't do it e.g. in get_ref_base_and_extent so I agree we
should not do it in get_base_address either.

(In reply to comment #22)
> btw, what's the right component for the PR? tree-optimization? middle-end?

Yeah, I guess (and what it's worth, I'm changing the component to
middle-end).