! Check that for parallel do, reduction is only processed for the loop ! RUN: bbc -fopenmp -emit-hlfir %s -o - | FileCheck %s ! RUN: flang-new -fc1 -fopenmp -emit-hlfir %s -o - | FileCheck %s ! CHECK: omp.parallel { ! CHECK: omp.wsloop reduction(@add_reduction_i_32 subroutine sb integer :: x x = 0 !$omp parallel do reduction(+:x) do i=1,100 x = x + 1 end do !$omp end parallel do end subroutine