Hello Wei,

Two observations, taking as given the point already made in this thread that an AND flag does not deliver what it promises when the receiver has not implemented the stronger algorithm.

The first is that making the handling of an unimplemented algorithm a SHOULD has an undefined effect: the same flag produces AND at a verifier that implements every algorithm present, the OR outcome at one that does not and whatever local policy prefers elsewhere. So the signer cannot know what its own flag will produce.

The second point is the result values. In RFC 8601 Section 2.7.1 “policy” means the message was signed but some aspect of the signature was not acceptable to the ADMD, which is not the case of a verifier that cannot evaluate it. “softfail” is not a result of the dkim method at all and belongs to spf. The nearest existing value is “neutral”, which covers a signature that contained syntax errors or was “not otherwise able to be processed”. In DKIM1 a signature in an algorithm the verifier does not implement is not reported at all, since RFC 6376 Section 3.3.4 says that verifiers MUST ignore such signatures. The dkim2 method as defined in draft-gondwana-dkim2-authres-00 has pass, fail, permerror, temperror and none, so it offers none of these and says nothing about this case.

Leaving things as they are does not settle the question either, because the present text does not say one thing. Section 14.1 states that “A Verifier MAY cease verifying once a single failure is detected”, while Section 14.6 states that “If there is more than one signature provided then they MUST all be checked if the Verifier is able to do so. If any signature fails then an error SHOULD be reported.” Section 14.1 also leaves the form of that report open, since verifiers wishing to communicate results to other parts of the mail system “may do so in whatever manner they see fit”. There is no defined state for the error to be reported into.

What is missing is one rule in the protocol for how a hop carrying several signatures is evaluated, covering the case of a signature the verifier is unable to evaluate. Stated once, it makes the verdict on a hop the same at every verifier. One such rule is that a signature the verifier cannot evaluate is, for that verifier, in the same position as a signature that is not present, so the hop is judged on the signatures it can check. The treatment of such a signature as absent is already DKIM1’s, as noted above. What is new in DKIM2 is that the verdict is reached per hop, which is why the rule has to be stated. It also has a consequence for reporting: results that follow from a rule everyone applies can be checked by a later verifier against what it computes itself and counted together across receivers, while results produced under a rule each verifier chooses can be neither checked nor added up.

Regards,
Vittorio Moccia

Il 2026-10-25 17:57 Wei Chuang ha scritto:

For interoperability, a receiver can treat an unknown algorithm
signature as local policy passing i.e., with a SHOULD and
Authentication-Result of "policy".  If a vulnerability has been
announced or for some operator defined reason, that same local policy
can set the Authentication-Result to "softfail".
-Wei
_______________________________________________
Ietf-dkim mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to