Seems like an editing error resulted in this dangling sentence. The corresponding text in draft -04 read as shown below, and I suggest we reinstate the "discard the option" behaviour:
The MAP algorithm, as per Section 5.1 <http://tools.ietf.org/html/draft-ietf-softwire-map-dhcp-04#section-5.1> [I-D.ietf-softwire-map <http://tools.ietf.org/html/draft-ietf-softwire-map-dhcp-04#ref-I-D.ietf-softwire-map>], allows the port-indexing algorithm to be parametrized in terms of an excluded port range (a-bits) and an explicit Port Set ID (PSID) value, which can be conveyed via the S46 Port Parameters Option . The following behaviour is expected when processing this option by MAP clients: A client receiving a S46 Port parameters Option in an OPTION_S46_BRULE or OPTION_S46_FRULE whose derived IPv4 address is not 32 bit-long MUST discard the option. The formula for this check is "prefix4-len + ea-len == 32". This is to ensure that the explicit PSID is only applied to configurations with a completely formed IPv4 address. On 22 November 2013 03:00, Tom Taylor <[email protected]> wrote: > Just realized that the option will be provided only if no EA bits were > present in the BMR. Hence the test as it stands is correct. The sentence > fragment still needs expansion or deletion. Wonder if we need text saying > what to do if prefix4-len + ea-len > 32 and the option is also present? > > Tom > > On 21/11/2013 8:54 PM, Tom Taylor wrote: > >> The text describing client validation of OPTION_S46_PORTPARAMS in the >> second-last paragraph of Section 4.5 currently reads as follows: >> >> When receiving the Port Parameters option with an explicit PSID, the >> client MUST use this explicit PSID in configuring its MAP interface. >> If the conveyed IPv4 address is not 32 bit-long. The formula for >> this check is "prefix4-len + ea-len = 32" and serves to ensure >> that the explicit PSID is only applied to configurations with a >> completely formed IPv4 address. >> >> I don't think the sentence fragment "If the conveyed IPv4 address ..." >> should be there. In the subsequent sentence, I believe the check must be: >> >> "prefix4-len + ea-len > 32" >> >> Otherwise no port values are conveyed by the option. See Section 5.2 of >> the MAP document, particularly Figure 6 and associated text. >> >> I would also add "shared and" before "completely formed" at the end of >> the paragraph. >> >> Tom Taylor >> > _______________________________________________ > Softwires mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/softwires >
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
