On the ineffectiveness of 1/m-based interference bounds in the analysis of global EDF and FIFO scheduling