[Bug tree-optimization/82732] malloc+zeroing other than memset not optimized to calloc, so asm output is malloc+memset

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82732

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |aldyh at gcc dot gnu.org,
                   |                            |jakub at gcc dot gnu.org,
                   |                            |law at gcc dot gnu.org

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So, either we need on-demand VRP for this, or
      tree niters = number_of_latch_executions (loop);
      niters = fold_convert_loc (loc, sizetype, niters);
      if (dominated_by_p (CDI_DOMINATORS, single_exit (loop)->src, bb))
        niters = size_binop_loc (loc, PLUS_EXPR, niters, size_one_node);
needs to be smarter, dunno if it should look if the loop has a guard that makes
sure the var in niters is non-zero, or something similar.  Such guards are
common though, so perhaps the niters infrastructure should be able to do that.
Have number_of_latch_executions alternative that would add one to that if
needed and take into account the possible preheaders.