RFC 1122 still is a part of the standard, see RFC2000.
The definition for IP was written 8 years earlier (september 1981 !!) and
is still part of the standard.
>Unfortunately, almost all this chapter in rfc is wrong,
>the mechanism appeared to be too fragile: f.e. it does not detect
>neither case when router ceases to forward nor one-way reachibility case.
>It is absolutely useless to implement these obsolete requirements.
I know the IPV4 protocol definition is neither perfect nor complete.
It is however the definition used by many vendors and enables
communication in a multivendor environment. Problems with the
protocoldefinition are well known, see RFC1025.
The requirement MUST in 3.3.1.4 leaves no room for interpretation.
My customers need implementation of 3.3.1.4 and asked for it.
I think the proposed algorithm will work and is easy to implement,
so I will give it a try.
PS, I am not on linux-net@vger.rutgers.edu. The list
linux-kernel@vger.rutgers.edu has been extremely quiet last week,
so maybe I have missed part of the discussion.
Chel