Use ND_BYTES_AVAILABLE_AFTER() to calculate the number of bytes
remaining in the packet after a given pointer, rather than doing the
subtraction directly; that casts the result to a u_int (we don't handle
packets bigger than the maximum u_int value, so the difference between
the pointers will never be bigger than that value), so we don't have to
deal with it being a 64-bit value on LP64 or LLP64 systems. (It also
makes it a bit clearer what we're doing).