Distributional reinforcement learning with linear function approximation