bluejekyll / trust-dns / 517

88%

master: 82%

push

travis-ci

Fix SIG(0) calculation (#122)

* make signatory value 0 the default value

According to RFC3007, this field SHOULD be set to 0 in KEY records, and MUST be ignored.

Therefore, a default value of 0 seems to be appropriate.

* Don't double-hash message when signing

Signer::hash_message() is only called from Signer::sign_message() and
Signer::verify_message().

Those functions then feed the result to KeyPair::sign() or
KeyPair::verify(), respectively, which already do both hashing and signing.

So, if Signer::hash_message() also hashes the message,
the result is rsa(hash(hash(message))).

TODO: This is only tested with RSA keys. Code for KeyPair::ED25519 looks
like it may not do hashing in KeyPair::sign/verify.

* Fix SIG(0) hash calculation

SIG(0) signatures are calculated over a hash which includes not only the
signed message, but also the RDATA of the SIG(0) record.

See https://tools.ietf.org/html/rfc2931#section-3.1 for details.

* Avoid code duplication in Message::sign()

Use new function `RData::SIG::set_sig(self, signature: Vec<u8>) -> Self {...}`,
instead.

64 of 64 new or added lines in 4 files covered. (100.0%)

10949 of 11922 relevant lines covered (91.84%)

0.92 hits per line