Guy Harris <[email protected]> wrote: >> We allocated a few chunks for private use years ago, and they could be >> in use internally somewhere, so we don't want to change that. But, >> maybe we should mark them as deprecated?
> "Deprecated" presumably means "if you want to use one of these points
> for internal experimental use, OK, but we strongly recommend you get an
> officially-assigned value".
Yes. And the things we marked as RESERVED_xx are really strongly for
experimentation, to be changed next time you recompile.
> I'd add "and don't ever let your internal use leak out to the outside
> world", perhaps with a parenthetical note "(I'm looking at *you*,
> Apple, and your use of DLT_USER2 for PKTAP)".
Had they used >65500, then life would have been fine if it leaked, we'd have
just laughed publically, knowing that the value can never be allocated.
That's why I think that I'd really like the 32K+ to really FCFS, and not
Expert Review. Someone needs a number, just please ask... one email business
day round trip time with IANA, and you are done.
No other checking than DoS and duplicate entry.
(I'll also note that the field in *PCAP* is actually 32-bits wide, and PCAPNG
has a 16-bit reserved field after type type. This is because some OSs did
weird stuff with the upper-16bits, and we partitioned it off years ago. We
could probably undo that and allow for 32-bits types, if someone felt that
32768->65000 was really not enough FCFS space)
--
Michael Richardson <[email protected]> . o O ( IPv6 IøT consulting )
Sandelman Software Works Inc, Ottawa and Worldwide
signature.asc
Description: PGP signature
_______________________________________________ OPSAWG mailing list [email protected] https://www.ietf.org/mailman/listinfo/opsawg
