Re: [rfc5225] Positioning of IPX dynamic chain in profile 0x0104 (IP ONLY)
Carl Knutsson <[email protected]> Wed, 04 Nov 2009 09:46:35 +0100
| Newsgroups | gmane.ietf.rohc |
|---|---|
| Message-ID | <[email protected]> |
Klaus, Suneeta, Comments inline... Klaus Warnke wrote: > Hi Suneeta, > > I think the dynamic chain ends with the _dynamic format of the last > extension header after the inner ip header, like ah_dynamic. > This makes sense for me. > > I have two question too. > > For what profile is the ipv4_regular_innermost_dynamic format? > This format can be used for a innermost ipv4 header in a profile not > ip-only. Yes, for all profiles except ROHCv2 IP the compressed format 'ipv4_regular_innermost_dynamic' will fail. > And, additionally, there is no counterpart for ipv6 defined. > Correct. Only two *_dynamic formats is needed for ipv6. > And, is there a special reason, why the field ip_version for the > udp_baseheader is static encoded in DEFAULT, but for the other > _baseheader encoding methods not? No special reason. > > br > Klaus > > At Tue, 13 Oct 2009 04:49:47 -0700 (PDT), > Suneeta Rao wrote: >> Hi ROHCWG, >> >> I have a doubt regarding placing of IPX dynamic chain in IP ONLY >> profile. >> >> As per my understanding, under normal circumstances i.e.when IPX >> headers are absent, the dynamic chain shall terminate with IP >> dynamic chain as per RFC5225 ROHC-FNCOMPRESSED >> ipv4_endpoint_innermost_dynamic (for IPV4) Pg 60. >> >> What should be behavior in presence of IPX headers (AH/GRE for IPV4 >> and HOP-BY-HOP/ROUTING/DESTINATION for IPV6)? >> >> Should the chain terminate with IP dynamic chain(as per ROHCFN named >> above) followed by IPX dynamic chain. or Alternatively should it >> terminate in IP dynamic chain and IPX headers cannot be compressed >> in this profile. >> >> The confusion is because the ROHCFN says "ipv4_endpoint" for >> innermost IP header in IP ONLY profile. *_endpoint_* have no special meaning. Only compressed formats that ends with *_irregular, *_static, *_dynamic or *_list_item comes with special rules. "ENFORCE((is_innermost == 1) && .." The ROHC FN statement above tells us that it is the innermost or last IP header in the dynamic chain. Not necessarily the last item in the dynamic. cheers, /Calle