Extending Whisper with prompt tuning to target-speaker ASR