On successive pole assignment by linear-quadratic optimal feedbacks