Maximal Controllability for Boundary Control Problems