That only works if you are sure to have a 64-bit result. If you can have divisor < multiplier and need to detect overflow, it's more complicated.