[Bug tree-optimization/71872] ICE in inchash::add_expr, at tree.c:7782 - OEP_ADDRESS_OF asserted for ADDR_EXPR applied to constant

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71872

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2016-07-14
     Ever confirmed|0                           |1

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #2)
> This has nothing to do with sanitizers, it seems tree-data-ref.c is buggy
> and creates invalid trees (ADDR_EXPR of INTEGER_CST) when processing the rhs
> of
> MEM[(union gasnete_anytype32_t * {ref-all})_dst_11] =
> VIEW_CONVERT_EXPR<union gasnete_anytype32_t>(1);
> Such VCE obviously is not a DR.

I thought I have fixed that (not considering sth like this as DR...).

Ah, I excluded SSA_NAME but not constants... bah.  Hmm, somehow I remember
changing this recently but appearantly nothing in the ChangeLog.

See get_references_in_stmt.  r133311 - some time ago ;)

Note the whole function should probably be re-written in terms of
walk_stmt_load_store_ops though that doesn't know the IFN handling.