Update to the Cryptographic Message Syntax (CMS) for Algorithm Identifier Protection

This document updates the Cryptographic Message Syntax (CMS) to ensure that algorithm identifiers in signed-data and authenticated-data content types are adequately protected. The CMS signed-data Content Type , unlike X.509 certificates , can be vulnerable to algorithm substitution attacks. In an algorithm substitution attack, the attacker changes either the algorithm identifier or the parameters associated with the algorithm identifier to change the verification process used by the recipient. The X.509 certificate structure protects the algorithm identifier and the associated parameters by signing them. In an algorithm substitution attack, the attacker looks for a different algorithm that produces the same result as the algorithm used by the originator. As an example, if the signer of a message used SHA-256 as the digest algorithm to hash the message content, then the attacker looks for a weaker hash algorithm that produces a result that is of the same length. The attacker’s goal is to find a different message that results in the same hash value, which is called a cross-algorithm collision. Today, there are many hash functions that produce 256-bit results. One of them may be found to be weak in the future. Further, when a digest algorithm produces a larger result than is needed by a digital signature algorithm, the digest value is reduced to the size needed by the signature algorithm. This can be done both by truncation and modulo operations, with the simplest being straightforward truncation. In this situation, the attacker needs to find a collision with the reduced digest value. As an example, if the message signer uses SHA-512 as the digest algorithm and ECDSA with the P-256 curve as the signature algorithm, then the attacker needs to find a collision with the first half of the digest. Similar attacks can be mounted against parameterized algorithm identifiers. When looking at randomized hash functions, such as the example in , the algorithm identifier parameter includes a random value that can be manipulated by an attacker looking for collisions. Some other algorithm identifiers include complex parameter structures, and each value provides another opportunity for manipulation by an attacker. This document makes two updates to CMS to provide protection for the algorithm identifier. First, it mandates a convention followed by many implementations by requiring the originator to use the same hash algorithm to compute the digest of the message content and the digest of signed attributes. Second, it recommends that the originator include the CMSAlgorithmProtection attribute .

The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “NOT RECOMMENDED”, “MAY”, and “OPTIONAL” in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here.

This section updates to require the originator to use the same hash algorithm to compute the digest of the message content and the digest of signed attributes.

Change the paragraph describing the digestAlgorithm as follows: OLD: digestAlgorithm identifies the message digest algorithm, and any associated parameters, used by the signer. The message digest is computed on either the content being signed or the content together with the signed attributes using the process described in Section 5.4. The message digest algorithm SHOULD be among those listed in the digestAlgorithms field of the associated SignerData. Implementations MAY fail to validate signatures that use a digest algorithm that is not included in the SignedData digestAlgorithms set. NEW: digestAlgorithm identifies the message digest algorithm, and any associated parameters, used by the signer. The message digest is computed on either the content being signed or the content together with the signedAttrs using the process described in Section 5.4. The message digest algorithm SHOULD be among those listed in the digestAlgorithms field of the associated SignerData. If the signedAttrs field is present in the SignerInfo, then the same digest algorithm MUST be used to compute both the digest of the SignedData encapContentInfo eContent, which is carried in the message-digest attribute, and the digest of the DER-encoded signedAttrs, which is passed to the signature algorithm. Implementations MAY fail to validate signatures that use a digest algorithm that is not included in the SignedData digestAlgorithms set.

Add the following paragraph as the second paragraph in Section 5.4: ADD: When the signedAttrs field is present, the same digest algorithm MUST be used to compute the digest of the encapContentInfo eContent OCTET STRING, which is carried in the message-digest attribute, and the digest of the collection of attributes that are signed. nit: there may be a grammar nit here, relating to the parallelism of “compute the digest of” – I think “the collection of attributes that are signed” should also have an “of” or “digest of” prefix.

Change the paragraph discussing the signed attributes as follows: OLD: The recipient MUST NOT rely on any message digest values computed by the originator. If the SignedData signerInfo includes signedAttributes, then the content message digest MUST be calculated as described in Section 5.4. For the signature to be valid, the message digest value calculated by the recipient MUST be the same as the value of the messageDigest attribute included in the signedAttributes of the SignedData signerInfo. NEW: The recipient MUST NOT rely on any message digest values computed by the originator. If the SignedData signerInfo includes the signedAttrs field, then the content message digest MUST be calculated as described in Section 5.4, using the same digest algorithm to compute the digest of the encapContentInfo eContent OCTET STRING and the message-digest attribute. For the signature to be valid, the message digest value calculated by the recipient MUST be the same as the value of the messageDigest attribute included in the signedAttrs field of the SignedData signerInfo.

The new requirement introduced above might lead to incompatibility with an implementation that allowed different digest algorithms to be used to compute the digest of the message content and the digest of signed attributes. The signatures produced by such an implementation when two different digest algorithms are used will be considered invalid by an implementation that follows this specification. However, most, if not all, implementations already require the originator to use the same digest algorithm for both operations.

The new requirement introduced above might lead to compatibility issues for timestamping systems when the originator does not wish to share the message content with the Time Stamp Authority (TSA) . In this situation, the originator sends a TimeStampReq to the TSA that includes a MessageImprint, which consists of a digest algorithm identifier and a digest value, then the TSA uses the originator-provided digest in the MessageImprint. When producing the TimeStampToken, the TSA MUST use the same digest algorithm to compute the digest of the encapContentInfo eContent, which is an OCTET STRING that contains the TSTInfo, and the message-digest attribute within the SignerInfo. To ensure that TimeStampToken values that were generated before this update remain valid, no requirement is placed on a TSA to ensure that the digest algorithm for the TimeStampToken matches the digest algorithm for the MessageImprint embedded within the TSTInfo.

This section updates to recommend that the originator include the CMSAlgorithmProtection attribute whenever signed attributes or authenticated attributes are present.

Add the following paragraph as the eighth paragraph in Section 14: ADD: While there are no known algorithm substitution attacks today, the inclusion of the algorithm identifiers used by the originator as a signed attribute or an authenticated attribute makes such an attack significantly more difficult. Therefore, the originator of a signed-data content type that includes signed attributes SHOULD include the CMSAlgorithmProtection attribute as one of the signed attributes. Likewise, the originator of an authenticated-data content type that includes authenticated attributes SHOULD include the CMSAlgorithmProtection attribute as one of the authenticated attributes.

This document makes no requests of the IANA.

The security properties of the CMS signed-data and authenticated-data content types are updated to offer protection for algorithm identifiers, which makes algorithm substitution attacks significantly more difficult. For the signed-data content type, the improvements specified in this document force an attacker to mount a hash algorithm substitution attack on the overall signature, not just on the message digest of the encapContentInfo eContent. Some digital signature algorithms have prevented hash function substitutions by including a digest algorithm identifier as an input to the signature algorithm. As discussed in , such a “firewall” may not be effective or even possible with newer signature algorithms. For example, RSASSA-PKCS1-v1_5 protects the digest algorithm identifier, but RSASSA-PSS does not. Therefore, it remains important that a signer have a way to signal to a recipient which digest algorithms are allowed to be used in conjunction with the verification of an overall signature. This signaling can be done as part of the specification of the signature algorithm, in an X.509v3 certificate extension , or some other means. The Digital Signature Standard (DSS) takes the first approach by requiring the use of an “approved” one-way hash algorithm. For the authenticated-data content type, the improvements specified in this document force an attacker to mount a MAC algorithm substitution attack, which is difficult because the attacker does not know the authentication key. The CMSAlgorithmProtection attribute offers protection for the algorithm identifiers used in the signed-data and authenticated-data content types. However, no protection is provided for the algorithm identifiers in the enveloped-data, digested-data, or encrypted-data content types. Likewise, The CMSAlgorithmProtection attribute provides no protection for the algorithm identifiers used in the authenticated-enveloped-data content type defined in . A mechanism for algorithm identifier protection for these content types is work for the future.

Many thanks to Jim Schaad and Peter Gutmann; without knowing it, they motivated me to write this document. Thanks to Roman Danyliw, Ben Kaduk, and Peter Yee for their careful review and editorial suggestions.