Struct schnorrkel::musig::MuSig
source · [−]pub struct MuSig<T: SigningTranscript + Clone, S> { /* private fields */ }
Expand description
Schnorr multi-signature (MuSig) container generic over its session types
Implementations
Iterates over public keys.
If require_reveal=true
then we count only public key that revealed their R
values.
Aggregate public key given currently revealed R
values
Aggregate public key expected if all currently committed nodes fully participate
We permit extending the transcript whenever you like, so
that say the message may be agreed upon in parallel to the
commitments. We advise against doing so however, as this
requires absolute faith in your random number generator,
usually rand::thread_rng()
.
Initialize a multi-signature aka cosignature protocol run.
We encurage borrowing the Keypair
to minimize copies of
the private key, so we provide the Keypair::musig
method
for the K = &'k Keypair
case. You could use Rc
or Arc
with this MuSig::new
method, or even pass in an owned copy.
Our commitment to our R
to send to all other cosigners
pub fn add_their_commitment(
&mut self,
them: PublicKey,
theirs: Commitment
) -> SignatureResult<()>
pub fn add_their_commitment(
&mut self,
them: PublicKey,
theirs: Commitment
) -> SignatureResult<()>
Add a new cosigner’s public key and associated R
bypassing our commitmewnt phase.
Commit to reveal phase transition.
Reveal our R
contribution to send to all other cosigners
Include a revealed R
value from a previously committed cosigner
Add a new cosigner’s public key and associated R
bypassing our
commitmewnt phase.
We implemented defenses that reduce the risks posed by this method, but anyone who wishes provable security should heed the advice below:
Avoid using this due to the attack described in “On the Provable Security of Two-Round Multi-Signatures” by Manu Drijvers, Kasra Edalatnejad, Bryan Ford, and Gregory Neven https://eprint.iacr.org/2018/417 Avoid using this for public keys held by networked devices in particular.
There are however limited scenarios in which using this appears secure, primarily if the trusted device is (a) air gapped, (b) stateful, and (c) infrequently used, via some constrained channel like manually scanning QR code. Almost all hardware wallets designs fail (b), but non-hardware wallets fail (a), with the middle ground being only something like Pairty Signer. Also, any public keys controlled by an organization likely fail (c) too, making this only useful for individuals.
Reveal to cosign phase transition.
Reveals our signature contribution
pub fn add_their_cosignature(
&mut self,
them: PublicKey,
theirs: Cosignature
) -> SignatureResult<()>
pub fn add_their_cosignature(
&mut self,
them: PublicKey,
theirs: Cosignature
) -> SignatureResult<()>
Include a cosignature from another cosigner
Interate over the cosigners who successfully revaled and later cosigned.
Interate over the possible cosigners who successfully committed and revaled, but actually cosigned.
pub fn add(
&mut self,
them: PublicKey,
their_reveal: Reveal,
their_cosignature: Cosignature
) -> SignatureResult<()>
pub fn add(
&mut self,
them: PublicKey,
their_reveal: Reveal,
their_cosignature: Cosignature
) -> SignatureResult<()>
Adds revealed R
and cosignature into a cosignature collector