I'm examining the subject.
From: Markku Savela <[EMAIL PROTECTED]>
Subject: Redirect ICMP and TAHI test?
Date: Mon, 19 Feb 2001 14:38:06 +0200
> In TAHI Neighbor Discovery test 62, it introduces two routers, then
> sends a redirect from "one" of them, and expects the redirect to have
> an effect...
He reported to me that his problem had come from
not the tests but elsewhere.
However, he also pointed out potential bugs of the tests:
ncStateByRedirect4Nonce.seq
ncStateByRedirect4Incomplete.seq
ncStateByRedirect4Reachable.seq
ncStateByRedirect4Stale.seq
ncStateByRedirect4Probe.seq
Those tests assumes that
- a NUT does not have round-robin fashion
in the default router selection.
- a NUT selects a same router as the default.
- a NUT selects the router whose RA comes first
becomes the default router.
In other word, some of the assumptions are broken,
the tests may be failed.
From: "Richard Draves" <[EMAIL PROTECTED]>
Subject: RE: Redirect ICMP and TAHI test?
Date: Mon, 19 Feb 2001 08:25:44 -0800
> Yes, the Redirect TAHI tests fail on the MS implementation, for the same
> reason. I think the test is broken.
I guess that one of the above assumptions is broken
in the case of MS implementation. This one always
update the default router when accepting RA.
I will be able to debug/fix the test
if I have an implementation that behaves the same.
Thank you for your feedback, again.
--- nobuo
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
