https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65386
--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> --- The test looks bogus to me. Citing the standard: final task A task that forces all of its child tasks to become final and included tasks. None of the explicit tasks (final or not) create any child tasks, so nothing requires any explicit task to be included, so I see no reason why there should be any restriction on where the tasks are scheduled. IMHO the spec just says that all the children of the final task will be included (i.e. tied to the final parent task).