[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
(usagi-users 01419) Re: (KAME-snap 6306) IPv6 header's Payload Length field
- To: snap-users@xxxxxxxx
- Subject: (usagi-users 01419) Re: (KAME-snap 6306) IPv6 header's Payload Length field
- From: JINMEI Tatuya / 神明達哉 <jinmei@xxxxxxxx>
- Date: Wed, 01 May 2002 12:26:33 +0900
- Cc: usagi-users@xxxxxxxxxxxxxx
- In-reply-to: <Pine.GSO.4.44.0204301050270.6226-100000@gate.ispras.ru>
- Organization: Research & Development Center, Toshiba Corp., Kawasaki, Japan.
- References: <Pine.GSO.4.44.0204301050270.6226-100000@gate.ispras.ru>
- Reply-to: usagi-users@xxxxxxxxxxxxxx
- User-agent: Wanderlust/2.6.1 (Upside Down) Emacs/21.1 Mule/5.0 (SAKAKI)
>>>>> On Tue, 30 Apr 2002 11:01:22 +0400 (MSK),
>>>>> Grigory Kljuchnikov <grn@xxxxxxxxx> said:
> I have some questions about the behaviour of an implementation in cases
> when the IPv6 header's Payload Length + size of IPv6 header isn't equal the size
> of the real recieved data.
These are basically protocol questions, so I'd recommend you to ask
these (and further) questions in the IETF IPv6 wg list. But I'll try
to give some short answers.
> 1) What must an implementation do in the case if the Payload Length field
> of the IPv6 header indicates presence of octets past the end of the upper layer
> packet? Must it generate ICMP Parameter Problem message or
> pass superfluous data unchanged (in end-point case and in router case)?
> Is it implementation depended?
I don't understand the question. How do you decide the "end of the
upper layer packet" without the payload length field of the IPv6
header?
> 2) If IPv6 header's Payload Length is zero, but the size of the real recieved
> data is longer then size of the IPv6 header (40 bytes). Is it an error?
Yes, unless the packet has a jumbo payload option.
> 3) If IPv6 header's Payload Length > 0, but the size of the real recieved
> data is longer then IPv6 header's Payload Length + size of IPv6 header.
> May an implementation pass superfluous data unchanged?
This can happen if a lower layer add a trailing pad to the packet. So
the implementation drop the pad before passing it to the upper layer.
> 4) If IPv6 header's Payload Length + size of IPv6 header is longer then
> size of the real recieved data. Is it an error case or implementation depended?
The packet should be discarded. I guess the specification does not
say anything about this case.
> 5) Which is the behaviour of an implementation in the case if it's IP included
> packet of ICMP Error message?
I don't understand the question, but if you mean what if an error
occurs against an ICMPv6 error message, the implementation should drop
the packet.
> 6) And if IP included packet of ICMP Error message is less then size of IPv6
> header?
Sorry, I don't understand the question.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
jinmei@xxxxxxxxxxxxxxxxxxxxx