UniWav: Towards Unified Pre-training for Speech Representation Learning and Generation