RFC Errata


Errata Search

 
Source of RFC  
Summary Table Full Records

RFC 4303, "IP Encapsulating Security Payload (ESP)", December 2005

Source of RFC: ipsec (sec)
See Also: RFC 4303 w/ inline errata

Errata ID: 6559
Status: Verified
Type: Technical
Publication Format(s) : TEXT

Reported By: Yaakov Stein
Date Reported: 2021-04-25
Verifier Name: Paul Wouters
Date Verified: 2022-04-10

Section 3.1.1 says:

                  AFTER APPLYING ESP
             -------------------------------------------------
       IPv4  |orig IP hdr  | ESP |     |      |   ESP   | ESP|
             |(any options)| Hdr | TCP | Data | Trailer | ICV|
             -------------------------------------------------

It should say:

                  AFTER APPLYING ESP
             ----------------------------------------------------
       IPv4  |updated IP hdr  | ESP |     |      |   ESP   | ESP|
             |(any options)   | Hdr | TCP | Data | Trailer | ICV|
             ----------------------------------------------------

Notes:

"orig" implies that the IP header is left as-is, while in fact the "protocol" field and the "total length" and the checksum must be updated. There IS appropriate text explaining this in RFC 3948 "The Total Length, Protocol, and Header Checksum (for IPv4) fields in the IP header are edited to match the resulting IP packet." but this text is missing here.

We have encountered an implementation that does not update the "total length" and the implementer claims that this is mandated by RFC 4303.

Paul / Tero:

This is updating the figure in RFC4303 (ESP) and should use "updated IP hdr" instead of "orig IP header", as the specification does in fact modify the protocol, total length and checksum fields.

In any potential future document update, text should be added that explains which fields are updated similar to what is done in the RFC3948:

The Total Length, Protocol, and Header Checksum (for IPv4) fields
in the IP header are edited to match the resulting IP packet.

As ESP is still used the IPsecME WG might want to make a RFC4303bis at some point and this fix should then be included. Perhaps the WG should think about moving it from proposed standard to internet standard at one point.

Report New Errata



Advanced Search