Hi Remi,
Thanks you for catching this.
On 02/03/09 10:16 AM, Rémi Denis-Courmont wrote:
Overall, I think this document should go forward.
Great.
However, the recommendation part should perhaps be a bit more explicit. I
assume the recipient node should not trash packets with a matching source,
destination and fragment ID *forever*, but rather within a reasonable time
frame.
Yes. You are right. The text in the draft is misleading. After further
thought I felt that this text was extraneous anyway and can be safely
removed. So, I propose the following change to this section.
OLD:
IPv6 nodes that receive a fragment
that overlaps with a previously received fragment MUST cease the
reassembly process and MUST ignore further fragments with the same
IPv6 Source Address, IPv6 Destination Address and Fragment
Identification. It MUST also discard the previously received
fragments with the same IPv6 Source Address, IPv6 Destination Address
and Fragment Identification.
NEW:
IPv6 nodes that receive a fragment
that overlaps with a previously received fragment MUST cease the
reassembly process and MUST discard the previously received
fragments with the same IPv6 Source Address, IPv6 Destination Address
and Fragment Identification.
My reasoning was that the 60 second reassembly rule of RFC2460 would
automatically take care of "ignoring" the further fragments coming in
with the same Fragment Identification. Is this acceptable?
Thanks
Suresh
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------