State Feedback l1 Optimal Controllers Can be Dynamic