[Bug fortran/82995] Segmentation fault passing optional argument to intrinsic sum function

[tkoenig at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82995

Thomas Koenig <tkoenig at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #44790|0                           |1
        is obsolete|                            |

--- Comment #11 from Thomas Koenig <tkoenig at gcc dot gnu.org> ---
Created attachment 45291
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=45291&action=edit
Next attempt

This patch handles the library side correctly, as far as cursory testing
reveals.

However, there still is something in there that I do not understand.

For the test case

module y 
    implicit none 
contains 

  function my_sum(input,mask) 
    logical, intent(in), optional :: mask(:) 
    integer, intent(in) :: input(:) 
    integer :: my_sum 
    my_sum = sum(input,mask) 
  end function my_sum
end module y

program test_my_sum 
  use y
  implicit none 
  integer :: input(3) = [1,2,3] 
  integer :: res
  res = my_sum(input) 
  print *,res

end program test_my_sum

I get, on the caller's side

  {
    struct array01_integer(kind=4) parm.10;

    parm.10.span = 4;
    parm.10.dtype = {.elem_len=4, .rank=1, .type=1};
    parm.10.dim[0].lbound = 1;
    parm.10.dim[0].ubound = 3;
    parm.10.dim[0].stride = 1;
    parm.10.data = (void *) &input[0];
    parm.10.offset = -1;
    res = my_sum (&parm.10, 0B);
  }

which looks good - there is a NULL pointer being passed to my_sum.

In my_sum, I get

  {
    integer(kind=4) val.8;
    integer(kind=8) D.3900;
    integer(kind=8) D.3901;
    integer(kind=8) D.3902;
    integer(kind=8) D.3903;

    val.8 = 0;
    D.3900 = offset.6;
    D.3901 = ubound.4;
    D.3902 = offset.2;
    D.3903 = ubound.0;
    {
      integer(kind=8) S.9;
      integer(kind=8) D.3905;
      integer(kind=8) D.3906;

      D.3905 = stride.5;
      D.3906 = stride.1;
      S.9 = 1;
      while (1)
        {
          if (S.9 > D.3901) goto L.1;
          if ((mask == 0B || (logical(kind=4)[0:] * restrict) mask->data == 0B)
|| (*mask.0)[S.9 * D.3905 + D.3900])
            {
              val.8 = (*input.0)[S.9 * D.3906 + D.3902] + val.8;
            }
          S.9 = S.9 + 1;
        }
      L.1:;
    }
    __result_my_sum = val.8;
  }
  return __result_my_sum;
}

which looks quite OK.

Still, I get zero as a result.

Something to do with arguments being passed wrong, maybe?