RFC Errata
RFC 2131, "Dynamic Host Configuration Protocol", March 1997
Note: This RFC has been updated by RFC 3396, RFC 4361, RFC 5494, RFC 6842
Source of RFC: dhc (int)
Errata ID: 7776
Status: Rejected
Type: Technical
Publication Format(s) : TEXT
Reported By: Imrane
Date Reported: 2024-01-23
Rejected by: Eric Vyncke
Date Rejected: 2024-04-23
Section 4.3.1 says:
Client identifier MUST NOT MUST NOT MAY
It should say:
Client identifier MUST NOT MUST NOT MUST NOT
Notes:
In the "Options" list in Table 3 ("Fields and options used by DHCP server"), the "Client identifier" option has "MUST NOT" for both DHCPOFFER and DHCPACK; however, for DHCPNAK, it has "MAY".
"Client identifier" should be a "MUST NOT" for DHCPNAK as well.
It seems that the field should only be used by a client and never by a server, and if that's true for the OFFER and ACK, then it should be even more correct for the NAK.
"Vendor class identifier" has a MAY for all three messages, so maybe it was a typo in the previous option because of the repetitive input in the next one.
--VERIFIER NOTES--
RFC 6842 has addressed this problem with more information than a simple errata.
I.e., the problem in RFC 2131 exists indeed but has been fixed.