It's generally assumed that a valid signature cannot be modified without the private key and remain valid. However, it is possible to modify and signature and maintain validity. One example of a system which is vulnerable to signature malleability is one in which validation as to whether an action can be executed is determined based on whether the signature has been previously used.
// UNSECURE
require(!signatureUsed[signature]);
// Validate signer and perform state modifying logic
...
signatureUsed[signature] = true;
In the above example, we can see that the signature is saved in a signatureUsed mapping after execution and validated to not exist in that mapping before execution. The problem with this is that if the signature can be modified while maintaining valididty, the transaction can be repeated by an attacker.
To understand how signature malleability works, we first need to understand a bit about elliptic curve cryptography.
An elliptic curve consists of all the points that satisfy an equation of the form:
where
Some examples:
Note that the curves are always symmetrical about the x-axis
The curve used by Ethereum is secp256k1, which looks like:
Now that we understand the basics of elliptic curve cryptography, we can dig into how signature malleability actually works on Ethereum.
Ethereum uses ECDSA as it's signature scheme. ECDSA signatures consist of a pair of numbers,
It's possible to calculate this complementary signature without knowing the private key used to produce it in the first place, which gives an attacker the ability to produce a second valid signature.
To avoid this issue, it's imperative to recognize that validating that a signature is not reused is insufficient in enforcing that the transaction is not replayed.