[Bug fortran/62131] New: Openmp Regression 4.9.1 : Subobject of an allocatable array not allowed in OMP ATOMIC

["vladimir.fuka at gmail dot com" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62131

            Bug ID: 62131
           Summary: Openmp Regression 4.9.1 : Subobject of an allocatable
                    array not allowed in OMP ATOMIC
           Product: gcc
           Version: 4.9.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vladimir.fuka at gmail dot com

layzrc raised this problem on comp.lang.fortran
https://groups.google.com/forum/#!topic/comp.lang.fortran/lVzeHW_X1aw

module storage
  integer,allocatable :: nerrs(:,:)
end module

program atomic
  use storage
    allocate(nerrs(10,10))
!$omp parallel do
    do k=1,10
      call uperrs(k,1)
    enddo
!$omp end parallel do
  stop
contains
  subroutine uperrs(i,io)
  integer,intent(in) :: i,io
!$omp atomic
    nerrs(i,io)=nerrs(i,io)+1
  end subroutine
end

>gfortran -fopenmp atomic.f90
atomic.f90:18.29:

    nerrs(i,io)=nerrs(i,io)+1
                             1
Error: !$OMP ATOMIC with ALLOCATABLE variable at (1)



I believe this is a  wrong and causes a bad regression, because `nerrs(i,io)`
is not an allocatable variable. I believe the section 2.12.6 of OpenMP 4
concerns situations when the whole variable being updated or assigned is
allocatable (e.g. allocatable scalars) and it causes possible re-allocation.