> Can we similarly promise or say something for accesses of the > containing struct as a whole? I hadn't considered those cases (when would you want to copy a *peripheral* ?) Should be the same as before, I would think.