On Fri, 2020-01-31 at 14:41 +0000, Roberto Sassu wrote: > I thought that using a stronger algorithm for hash collision detection but > doing remote attestation with the weaker would not bring additional value. > > If there is a hash collision on SHA1, an attacker can still replace the data of > one of the two entries in the measurement list with the data of the other > without being detected (without additional countermeasures). > > If the verifier additionally checks for duplicate template digests, he could > detect the attack (IMA would not add a new measurement entry with the > same template digest of previous entries). > > Ok, I will use ima_hash_algo for hash collision detection. Thanks! Mimi
