Yes this was exactly my doubt ... and I actually thought that such boundary
must be sub-TLV

Maybe Tony had in mind sub-TLV which consists of TLVs (sub-sub-TLVs) such
boundary could also work.

But this sentence:

"it cannot be required to have parts of another TLV in order to correctly
parse any sub-TLV."

is a bit cryptic. Can you pls provide some real life example(s) ?

Thx

On Fri, Feb 21, 2025 at 6:13 PM Les Ginsberg (ginsberg) <[email protected]>
wrote:

> Robert/Tony -
>
> I think this relates to a point discussed during Adrian's RTGDIR review.
> Adrian made the point that:
>
> "... that the split between component TLVs MUST be done at a sub-TLV or
> other unit boundary."
>
> I had suggested text (but not actually added it to the latest version of
> the draft):
>
> “Each TLV that is part of an MP-TLV MUST be parsable independent of other
> TLVs in the MP-TLV.
> Breaking of a single sub-TLV or other data unit across TLVs MUST NOT be
> done.”
>
> Perhaps adding that text would help.
>
> Tony, of course, is also correct. He is saying just as we can split a top
> level TLV into multiple TLVs, we can also split any level of sub-TLV into
> multiple TLVs - the restriction being that the splitting MUST be done on a
> parsable boundary i.e., it cannot be required to have parts of another TLV
> in order to correctly parse any sub-TLV.
>
>    Les
>
> > -----Original Message-----
> > From: Tony Li <[email protected]> On Behalf Of Tony Li
> > Sent: Friday, February 21, 2025 8:58 AM
> > To: Robert Raszuk <[email protected]>
> > Cc: Les Ginsberg (ginsberg) <[email protected]>; lsr <[email protected]>
> > Subject: Re: [Lsr] [Last-Call] 答复: Re: 【Can you concatenate several
> pieces
> > together without one "explicit key" to identify them belong to the same
> > segment】Re: Rtgdir last call review of draft-ietf-lsr-multi-tlv-08
> >
> >
> > Robert,
> >
> > > So it was very clear that we MUST skip what we do not recognize :)  I
> was not
> > sure if we should at that point bail out from further parsing of a given
> TLV or
> > trying next sub-TLV. I guess there is no mandate of that in the spec and
> > implementations should/may try to continue to parse.
> >
> >
> > You always continue to parse.  Production implementations are not test
> suites
> > looking for bugs.  If there is unrecognized data, then the presumption
> is that it
> > is new functionality that the implementation does not yet understand.
> > Stopping parsing would break legacy behavior.
> >
> >
> > > Also is my understanding correct that the subject draft does not allow
> to
> > split sub-TLVs itself ? Meaning that any sub-TLV must fit one part of
> the TLV. I
> > found some text allowing duplication of sub-TLVs in multiple parts of
> TLV if
> > sender choose to do such thing - but I assume this says that sub-TLV is
> still a
> > complete one in each part ?
> >
> >
> > I’m not sure that I can parse that. :-)
> >
> > Sub-TLVs can be split into multiple sub-TLVs. If the parent TLV is full,
> then parts
> > of the sub-TLV may be carried in separate instances of the parent TLV.
> >
> > In other words, the solution is fully recursive.
> >
> > T
> >
>
>
_______________________________________________
Lsr mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to