Re: [babel] Shepherd Review of draft-ietf-babel-source-specific-04

[Toke Høiland-Jørgensen <[e-mail]>]

Juliusz Chroboczek <[e-mail]> writes:

>> It seems to me that you need to unambiguously specify what is done if
>> there is a [...] TLV with multiple Source Prefix sub-TLVs. Such TLVs
>> could always be ignored or the first or the last Source Prefix sut-TLV
>> could be the effective sub-TLV with other Source Prefix sub-TLVs
>> ignored. But whatever is specified, it needs to be mandatory unambiguous
>> behavior.
>
> [...]
>
>> What do you do when you get a malformed Source Prefix sub-TLV with a zero
>> Source Plen?
>
> [...]
>
>> Continuing the theme of some of the comments above, what do you do with
>> a wildcard request that does contain a Source Prefix sub-TLV?
>
> Toke, I could use some advice here.  Please have a look at your code and
> let me know whether requiring (MUST) dropping the enclosing TLV in all of
> these cases complicates your implementation or not.  (I'll do the same on
> my side.)

Dropping the enclosing TLV is no problem, skipping the source TLV would
need patches... For reference, Bird currently does the following[0]:

- Malformed sub-TLV or Source prefix plen == 0: Parse error, drop the whole packet.

- Source sub-TLV on wildcard update or route request: Ignore enclosing TLV.

- Source sub-TLV on any other TLV: Parse error, drop the whole packet.

- Source-specific routing not configured: Ignore enclosing TLV.

- Duplicate source prefix sub-TLV: Ignore enclosing TLV.

-Toke

[0]: https://gitlab.labs.nic.cz/labs/bird/blob/master/proto/babel/packets.c#L998