Exploring Efficient Waveform Diffusion Models for Foley Sound Generation