>> and the kernel does not know what kind of alignment userland have
>> used!
>But - since this is all a hypothetical assuming that the padding is
>counted into cmsg_len - it can find out, by looking at the cmsg_len
>values.

	if we always pad cmsg_len to padding boundary, you lose ability to pass
	shorter data to kernel.  the only way i can think of is to add an
	additioanl field, cmsg_offsettonext (or whatever) which includes
	padding (keep cmsg_len as is).

itojun