The only pros of step gradient I can think of is concentration of your sample/protein and speed. The pros of linear gradient are better resolution (with step gradient you probably get several proteins in each peak), sharper peaks (smaller tailing).
I think first it is better to optimise eluting conditions in linear gradient condition. then you can use step gradient in further routine purification, becuase you will know what is the exact ratio of buffers when your protein elutes.