[Bug tree-optimization/93964] [8/9/10 Regression] [graphite] ICE in assign_parameter_index_in_region, at graphite-scop-detection.c:1104

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93964

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2020-03-02
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot 
gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.

1104      gcc_assert (TREE_CODE (name) == SSA_NAME
1105                  && INTEGRAL_TYPE_P (TREE_TYPE (name))
1106                  && ! defined_in_sese_p (name, region->region));

the parameter has pointer type.  Some disconnect between analyses predicates.

Ouch.  It's from niters which niter analysis expanded that far.
(unsigned int) (long int) _1 + 4294967295

So we end up using _1 instead of _2 here which would have been fine :/