Re: (usagi-users 00320) Re: USAGI IPv6 patches
Hello!
> If you are interested in the results and need more description,
> please tell us.
Yes, I need. I reviewed the results, analysis is enclosed.
Alexey
IPV6 specs
----------
38 Fragment Reassembly (needless fragment header)
Fixed in 2.3.
42 Fragment Reassembly (Payload size = 65536)
Fixed in 2.3.
PMTU disc
---------
3 Test that accept RA with MTU Option
Indeed. Fixing now.
5 Test that host detect Path MTU increase
Sorry? This should be PASS. Apparently, tester did not wait for timeout.
Or I do not understand the problem.
ndisc
-----
2,3,4,5 and 18,20
Explanation is required. I do not understand what is meant here.
Comment is apparently wrong; states DELAY, PROBE and REACHABLE can
be moved to STALE only when the message _updates_ link info.
Otherwise, you will introduce dead lock into state machine and
entry may to stick in unreachable state forever. 18, 20 tells that
our behaviour is opposite, this is contradiction.
12,14,16
Absolutely unintelligible.
17
What the state resulting entry has?
21
What are we expected to do in this case?
23...26
1. See comment about 2...5
2. 25-2 is just nonsense. How is ndisc going to work, when even
na with solicit set does not move entry to reacheble? 8)
3. Note about invalid messages requires explanations.
27,28
8)8)
35
Something unintelligible.
37
Lie?
39
Lie?
All the missing items after #28 eplicate another ones.
addrconf
--------
All of items are two problems.
Problem #1 is fixed now.
Problem #2 requires explanations. Whats is DAD NA? And what the hell
should we answer to them? 8)
Reply to: