serai/crypto/evrf/generalized-bulletproofs/src/arithmetic_circuit_proof.rs
Luke Parker e4e4245ee3
One Round DKG (#589)
* Upstream GBP, divisor, circuit abstraction, and EC gadgets from FCMP++

* Initial eVRF implementation

Not quite done yet. It needs to communicate the resulting points and proofs to
extract them from the Pedersen Commitments in order to return those, and then
be tested.

* Add the openings of the PCs to the eVRF as necessary

* Add implementation of secq256k1

* Make DKG Encryption a bit more flexible

No longer requires the use of an EncryptionKeyMessage, and allows pre-defined
keys for encryption.

* Make NUM_BITS an argument for the field macro

* Have the eVRF take a Zeroizing private key

* Initial eVRF-based DKG

* Add embedwards25519 curve

* Inline the eVRF into the DKG library

Due to how we're handling share encryption, we'd either need two circuits or to
dedicate this circuit to the DKG. The latter makes sense at this time.

* Add documentation to the eVRF-based DKG

* Add paragraph claiming robustness

* Update to the new eVRF proof

* Finish routing the eVRF functionality

Still needs errors and serialization, along with a few other TODOs.

* Add initial eVRF DKG test

* Improve eVRF DKG

Updates how we calculcate verification shares, improves performance when
extracting multiple sets of keys, and adds more to the test for it.

* Start using a proper error for the eVRF DKG

* Resolve various TODOs

Supports recovering multiple key shares from the eVRF DKG.

Inlines two loops to save 2**16 iterations.

Adds support for creating a constant time representation of scalars < NUM_BITS.

* Ban zero ECDH keys, document non-zero requirements

* Implement eVRF traits, all the way up to the DKG, for secp256k1/ed25519

* Add Ristretto eVRF trait impls

* Support participating multiple times in the eVRF DKG

* Only participate once per key, not once per key share

* Rewrite processor key-gen around the eVRF DKG

Still a WIP.

* Finish routing the new key gen in the processor

Doesn't touch the tests, coordinator, nor Substrate yet.
`cargo +nightly fmt && cargo +nightly-2024-07-01 clippy --all-features -p serai-processor`
does pass.

* Deduplicate and better document in processor key_gen

* Update serai-processor tests to the new key gen

* Correct amount of yx coefficients, get processor key gen test to pass

* Add embedded elliptic curve keys to Substrate

* Update processor key gen tests to the eVRF DKG

* Have set_keys take signature_participants, not removed_participants

Now no one is removed from the DKG. Only `t` people publish the key however.

Uses a BitVec for an efficient encoding of the participants.

* Update the coordinator binary for the new DKG

This does not yet update any tests.

* Add sensible Debug to key_gen::[Processor, Coordinator]Message

* Have the DKG explicitly declare how to interpolate its shares

Removes the hack for MuSig where we multiply keys by the inverse of their
lagrange interpolation factor.

* Replace Interpolation::None with Interpolation::Constant

Allows the MuSig DKG to keep the secret share as the original private key,
enabling deriving FROST nonces consistently regardless of the MuSig context.

* Get coordinator tests to pass

* Update spec to the new DKG

* Get clippy to pass across the repo

* cargo machete

* Add an extra sleep to ensure expected ordering of `Participation`s

* Update orchestration

* Remove bad panic in coordinator

It expected ConfirmationShare to be n-of-n, not t-of-n.

* Improve documentation on  functions

* Update TX size limit

We now no longer have to support the ridiculous case of having 49 DKG
participations within a 101-of-150 DKG. It does remain quite high due to
needing to _sign_ so many times. It'd may be optimal for parties with multiple
key shares to independently send their preprocesses/shares (despite the
overhead that'll cause with signatures and the transaction structure).

* Correct error in the Processor spec document

* Update a few comments in the validator-sets pallet

* Send/Recv Participation one at a time

Sending all, then attempting to receive all in an expected order, wasn't working
even with notable delays between sending messages. This points to the mempool
not working as expected...

* Correct ThresholdKeys serialization in modular-frost test

* Updating existing TX size limit test for the new DKG parameters

* Increase time allowed for the DKG on the GH CI

* Correct construction of signature_participants in serai-client tests

Fault identified by akil.

* Further contextualize DkgConfirmer by ValidatorSet

Caught by a safety check we wouldn't reuse preprocesses across messages. That
raises the question of we were prior reusing preprocesses (reusing keys)?
Except that'd have caused a variety of signing failures (suggesting we had some
staggered timing avoiding it in practice but yes, this was possible in theory).

* Add necessary calls to set_embedded_elliptic_curve_key in coordinator set rotation tests

* Correct shimmed setting of a secq256k1 key

* cargo fmt

* Don't use `[0; 32]` for the embedded keys in the coordinator rotation test

The key_gen function expects the random values already decided.

* Big-endian secq256k1 scalars

Also restores the prior, safer, Encryption::register function.
2024-09-19 21:43:26 -04:00

679 lines
23 KiB
Rust

use rand_core::{RngCore, CryptoRng};
use zeroize::{Zeroize, ZeroizeOnDrop};
use multiexp::{multiexp, multiexp_vartime};
use ciphersuite::{group::ff::Field, Ciphersuite};
use crate::{
ScalarVector, PointVector, ProofGenerators, PedersenCommitment, PedersenVectorCommitment,
BatchVerifier,
transcript::*,
lincomb::accumulate_vector,
inner_product::{IpError, IpStatement, IpWitness, P},
};
pub use crate::lincomb::{Variable, LinComb};
/// An Arithmetic Circuit Statement.
///
/// Bulletproofs' constraints are of the form
/// `aL * aR = aO, WL * aL + WR * aR + WO * aO = WV * V + c`.
///
/// Generalized Bulletproofs modifies this to
/// `aL * aR = aO, WL * aL + WR * aR + WO * aO + WCG * C_G + WCH * C_H = WV * V + c`.
///
/// We implement the latter, yet represented (for simplicity) as
/// `aL * aR = aO, WL * aL + WR * aR + WO * aO + WCG * C_G + WCH * C_H + WV * V + c = 0`.
#[derive(Clone, Debug)]
pub struct ArithmeticCircuitStatement<'a, C: Ciphersuite> {
generators: ProofGenerators<'a, C>,
constraints: Vec<LinComb<C::F>>,
C: PointVector<C>,
V: PointVector<C>,
}
impl<'a, C: Ciphersuite> Zeroize for ArithmeticCircuitStatement<'a, C> {
fn zeroize(&mut self) {
self.constraints.zeroize();
self.C.zeroize();
self.V.zeroize();
}
}
/// The witness for an arithmetic circuit statement.
#[derive(Clone, Debug, Zeroize, ZeroizeOnDrop)]
pub struct ArithmeticCircuitWitness<C: Ciphersuite> {
aL: ScalarVector<C::F>,
aR: ScalarVector<C::F>,
aO: ScalarVector<C::F>,
c: Vec<PedersenVectorCommitment<C>>,
v: Vec<PedersenCommitment<C>>,
}
/// An error incurred during arithmetic circuit proof operations.
#[derive(Clone, Copy, PartialEq, Eq, Debug)]
pub enum AcError {
/// The vectors of scalars which are multiplied against each other were of different lengths.
DifferingLrLengths,
/// The matrices of constraints are of different lengths.
InconsistentAmountOfConstraints,
/// A constraint referred to a non-existent term.
ConstrainedNonExistentTerm,
/// A constraint referred to a non-existent commitment.
ConstrainedNonExistentCommitment,
/// There weren't enough generators to prove for this statement.
NotEnoughGenerators,
/// The witness was inconsistent to the statement.
///
/// Sanity checks on the witness are always performed. If the library is compiled with debug
/// assertions on, the satisfaction of all constraints and validity of the commitmentsd is
/// additionally checked.
InconsistentWitness,
/// There was an error from the inner-product proof.
Ip(IpError),
/// The proof wasn't complete and the necessary values could not be read from the transcript.
IncompleteProof,
}
impl<C: Ciphersuite> ArithmeticCircuitWitness<C> {
/// Constructs a new witness instance.
pub fn new(
aL: ScalarVector<C::F>,
aR: ScalarVector<C::F>,
c: Vec<PedersenVectorCommitment<C>>,
v: Vec<PedersenCommitment<C>>,
) -> Result<Self, AcError> {
if aL.len() != aR.len() {
Err(AcError::DifferingLrLengths)?;
}
// The Pedersen Vector Commitments don't have their variables' lengths checked as they aren't
// paired off with each other as aL, aR are
// The PVC commit function ensures there's enough generators for their amount of terms
// If there aren't enough/the same generators when this is proven for, it'll trigger
// InconsistentWitness
let aO = aL.clone() * &aR;
Ok(ArithmeticCircuitWitness { aL, aR, aO, c, v })
}
}
struct YzChallenges<C: Ciphersuite> {
y_inv: ScalarVector<C::F>,
z: ScalarVector<C::F>,
}
impl<'a, C: Ciphersuite> ArithmeticCircuitStatement<'a, C> {
// The amount of multiplications performed.
fn n(&self) -> usize {
self.generators.len()
}
// The amount of constraints.
fn q(&self) -> usize {
self.constraints.len()
}
// The amount of Pedersen vector commitments.
fn c(&self) -> usize {
self.C.len()
}
// The amount of Pedersen commitments.
fn m(&self) -> usize {
self.V.len()
}
/// Create a new ArithmeticCircuitStatement for the specified relationship.
///
/// The `LinComb`s passed as `constraints` will be bound to evaluate to 0.
///
/// The constraints are not transcripted. They're expected to be deterministic from the context
/// and higher-level statement. If your constraints are variable, you MUST transcript them before
/// calling prove/verify.
///
/// The commitments are expected to have been transcripted extenally to this statement's
/// invocation. That's practically ensured by taking a `Commitments` struct here, which is only
/// obtainable via a transcript.
pub fn new(
generators: ProofGenerators<'a, C>,
constraints: Vec<LinComb<C::F>>,
commitments: Commitments<C>,
) -> Result<Self, AcError> {
let Commitments { C, V } = commitments;
for constraint in &constraints {
if Some(generators.len()) <= constraint.highest_a_index {
Err(AcError::ConstrainedNonExistentTerm)?;
}
if Some(C.len()) <= constraint.highest_c_index {
Err(AcError::ConstrainedNonExistentCommitment)?;
}
if Some(V.len()) <= constraint.highest_v_index {
Err(AcError::ConstrainedNonExistentCommitment)?;
}
}
Ok(Self { generators, constraints, C, V })
}
fn yz_challenges(&self, y: C::F, z_1: C::F) -> YzChallenges<C> {
let y_inv = y.invert().unwrap();
let y_inv = ScalarVector::powers(y_inv, self.n());
// Powers of z *starting with z**1*
// We could reuse powers and remove the first element, yet this is cheaper than the shift that
// would require
let q = self.q();
let mut z = ScalarVector(Vec::with_capacity(q));
z.0.push(z_1);
for _ in 1 .. q {
z.0.push(*z.0.last().unwrap() * z_1);
}
z.0.truncate(q);
YzChallenges { y_inv, z }
}
/// Prove for this statement/witness.
pub fn prove<R: RngCore + CryptoRng>(
self,
rng: &mut R,
transcript: &mut Transcript,
mut witness: ArithmeticCircuitWitness<C>,
) -> Result<(), AcError> {
let n = self.n();
let c = self.c();
let m = self.m();
// Check the witness length and pad it to the necessary power of two
if witness.aL.len() > n {
Err(AcError::NotEnoughGenerators)?;
}
while witness.aL.len() < n {
witness.aL.0.push(C::F::ZERO);
witness.aR.0.push(C::F::ZERO);
witness.aO.0.push(C::F::ZERO);
}
for c in &mut witness.c {
if c.g_values.len() > n {
Err(AcError::NotEnoughGenerators)?;
}
if c.h_values.len() > n {
Err(AcError::NotEnoughGenerators)?;
}
// The Pedersen vector commitments internally have n terms
while c.g_values.len() < n {
c.g_values.0.push(C::F::ZERO);
}
while c.h_values.len() < n {
c.h_values.0.push(C::F::ZERO);
}
}
// Check the witness's consistency with the statement
if (c != witness.c.len()) || (m != witness.v.len()) {
Err(AcError::InconsistentWitness)?;
}
#[cfg(debug_assertions)]
{
for (commitment, opening) in self.V.0.iter().zip(witness.v.iter()) {
if *commitment != opening.commit(self.generators.g(), self.generators.h()) {
Err(AcError::InconsistentWitness)?;
}
}
for (commitment, opening) in self.C.0.iter().zip(witness.c.iter()) {
if Some(*commitment) !=
opening.commit(
self.generators.g_bold_slice(),
self.generators.h_bold_slice(),
self.generators.h(),
)
{
Err(AcError::InconsistentWitness)?;
}
}
for constraint in &self.constraints {
let eval =
constraint
.WL
.iter()
.map(|(i, weight)| *weight * witness.aL[*i])
.chain(constraint.WR.iter().map(|(i, weight)| *weight * witness.aR[*i]))
.chain(constraint.WO.iter().map(|(i, weight)| *weight * witness.aO[*i]))
.chain(
constraint.WCG.iter().zip(&witness.c).flat_map(|(weights, c)| {
weights.iter().map(|(j, weight)| *weight * c.g_values[*j])
}),
)
.chain(
constraint.WCH.iter().zip(&witness.c).flat_map(|(weights, c)| {
weights.iter().map(|(j, weight)| *weight * c.h_values[*j])
}),
)
.chain(constraint.WV.iter().map(|(i, weight)| *weight * witness.v[*i].value))
.chain(core::iter::once(constraint.c))
.sum::<C::F>();
if eval != C::F::ZERO {
Err(AcError::InconsistentWitness)?;
}
}
}
let alpha = C::F::random(&mut *rng);
let beta = C::F::random(&mut *rng);
let rho = C::F::random(&mut *rng);
let AI = {
let alg = witness.aL.0.iter().enumerate().map(|(i, aL)| (*aL, self.generators.g_bold(i)));
let arh = witness.aR.0.iter().enumerate().map(|(i, aR)| (*aR, self.generators.h_bold(i)));
let ah = core::iter::once((alpha, self.generators.h()));
let mut AI_terms = alg.chain(arh).chain(ah).collect::<Vec<_>>();
let AI = multiexp(&AI_terms);
AI_terms.zeroize();
AI
};
let AO = {
let aog = witness.aO.0.iter().enumerate().map(|(i, aO)| (*aO, self.generators.g_bold(i)));
let bh = core::iter::once((beta, self.generators.h()));
let mut AO_terms = aog.chain(bh).collect::<Vec<_>>();
let AO = multiexp(&AO_terms);
AO_terms.zeroize();
AO
};
let mut sL = ScalarVector(Vec::with_capacity(n));
let mut sR = ScalarVector(Vec::with_capacity(n));
for _ in 0 .. n {
sL.0.push(C::F::random(&mut *rng));
sR.0.push(C::F::random(&mut *rng));
}
let S = {
let slg = sL.0.iter().enumerate().map(|(i, sL)| (*sL, self.generators.g_bold(i)));
let srh = sR.0.iter().enumerate().map(|(i, sR)| (*sR, self.generators.h_bold(i)));
let rh = core::iter::once((rho, self.generators.h()));
let mut S_terms = slg.chain(srh).chain(rh).collect::<Vec<_>>();
let S = multiexp(&S_terms);
S_terms.zeroize();
S
};
transcript.push_point(AI);
transcript.push_point(AO);
transcript.push_point(S);
let y = transcript.challenge();
let z = transcript.challenge();
let YzChallenges { y_inv, z } = self.yz_challenges(y, z);
let y = ScalarVector::powers(y, n);
// t is a n'-term polynomial
// While Bulletproofs discuss it as a 6-term polynomial, Generalized Bulletproofs re-defines it
// as `2(n' + 1)`-term, where `n'` is `2 (c + 1)`.
// When `c = 0`, `n' = 2`, and t is `6` (which lines up with Bulletproofs having a 6-term
// polynomial).
// ni = n'
let ni = 2 * (c + 1);
// These indexes are from the Generalized Bulletproofs paper
#[rustfmt::skip]
let ilr = ni / 2; // 1 if c = 0
#[rustfmt::skip]
let io = ni; // 2 if c = 0
#[rustfmt::skip]
let is = ni + 1; // 3 if c = 0
#[rustfmt::skip]
let jlr = ni / 2; // 1 if c = 0
#[rustfmt::skip]
let jo = 0; // 0 if c = 0
#[rustfmt::skip]
let js = ni + 1; // 3 if c = 0
// If c = 0, these indexes perfectly align with the stated powers of X from the Bulletproofs
// paper for the following coefficients
// Declare the l and r polynomials, assigning the traditional coefficients to their positions
let mut l = vec![];
let mut r = vec![];
for _ in 0 .. (is + 1) {
l.push(ScalarVector::new(0));
r.push(ScalarVector::new(0));
}
let mut l_weights = ScalarVector::new(n);
let mut r_weights = ScalarVector::new(n);
let mut o_weights = ScalarVector::new(n);
for (constraint, z) in self.constraints.iter().zip(&z.0) {
accumulate_vector(&mut l_weights, &constraint.WL, *z);
accumulate_vector(&mut r_weights, &constraint.WR, *z);
accumulate_vector(&mut o_weights, &constraint.WO, *z);
}
l[ilr] = (r_weights * &y_inv) + &witness.aL;
l[io] = witness.aO.clone();
l[is] = sL;
r[jlr] = l_weights + &(witness.aR.clone() * &y);
r[jo] = o_weights - &y;
r[js] = sR * &y;
// Pad as expected
for l in &mut l {
debug_assert!((l.len() == 0) || (l.len() == n));
if l.len() == 0 {
*l = ScalarVector::new(n);
}
}
for r in &mut r {
debug_assert!((r.len() == 0) || (r.len() == n));
if r.len() == 0 {
*r = ScalarVector::new(n);
}
}
// We now fill in the vector commitments
// We use unused coefficients of l increasing from 0 (skipping ilr), and unused coefficients of
// r decreasing from n' (skipping jlr)
let mut cg_weights = Vec::with_capacity(witness.c.len());
let mut ch_weights = Vec::with_capacity(witness.c.len());
for i in 0 .. witness.c.len() {
let mut cg = ScalarVector::new(n);
let mut ch = ScalarVector::new(n);
for (constraint, z) in self.constraints.iter().zip(&z.0) {
if let Some(WCG) = constraint.WCG.get(i) {
accumulate_vector(&mut cg, WCG, *z);
}
if let Some(WCH) = constraint.WCH.get(i) {
accumulate_vector(&mut ch, WCH, *z);
}
}
cg_weights.push(cg);
ch_weights.push(ch);
}
for (i, (c, (cg_weights, ch_weights))) in
witness.c.iter().zip(cg_weights.into_iter().zip(ch_weights)).enumerate()
{
let i = i + 1;
let j = ni - i;
l[i] = c.g_values.clone();
l[j] = ch_weights * &y_inv;
r[j] = cg_weights;
r[i] = (c.h_values.clone() * &y) + &r[i];
}
// Multiply them to obtain t
let mut t = ScalarVector::new(1 + (2 * (l.len() - 1)));
for (i, l) in l.iter().enumerate() {
for (j, r) in r.iter().enumerate() {
let new_coeff = i + j;
t[new_coeff] += l.inner_product(r.0.iter());
}
}
// Per Bulletproofs, calculate masks tau for each t where (i > 0) && (i != 2)
// Per Generalized Bulletproofs, calculate masks tau for each t where i != n'
// With Bulletproofs, t[0] is zero, hence its omission, yet Generalized Bulletproofs uses it
let mut tau_before_ni = vec![];
for _ in 0 .. ni {
tau_before_ni.push(C::F::random(&mut *rng));
}
let mut tau_after_ni = vec![];
for _ in 0 .. t.0[(ni + 1) ..].len() {
tau_after_ni.push(C::F::random(&mut *rng));
}
// Calculate commitments to the coefficients of t, blinded by tau
debug_assert_eq!(t.0[0 .. ni].len(), tau_before_ni.len());
for (t, tau) in t.0[0 .. ni].iter().zip(tau_before_ni.iter()) {
transcript.push_point(multiexp(&[(*t, self.generators.g()), (*tau, self.generators.h())]));
}
debug_assert_eq!(t.0[(ni + 1) ..].len(), tau_after_ni.len());
for (t, tau) in t.0[(ni + 1) ..].iter().zip(tau_after_ni.iter()) {
transcript.push_point(multiexp(&[(*t, self.generators.g()), (*tau, self.generators.h())]));
}
let x: ScalarVector<C::F> = ScalarVector::powers(transcript.challenge(), t.len());
let poly_eval = |poly: &[ScalarVector<C::F>], x: &ScalarVector<_>| -> ScalarVector<_> {
let mut res = ScalarVector::<C::F>::new(poly[0].0.len());
for (i, coeff) in poly.iter().enumerate() {
res = res + &(coeff.clone() * x[i]);
}
res
};
let l = poly_eval(&l, &x);
let r = poly_eval(&r, &x);
let t_caret = l.inner_product(r.0.iter());
let mut V_weights = ScalarVector::new(self.V.len());
for (constraint, z) in self.constraints.iter().zip(&z.0) {
// We use `-z`, not `z`, as we write our constraint as `... + WV V = 0` not `= WV V + ..`
// This means we need to subtract `WV V` from both sides, which we accomplish here
accumulate_vector(&mut V_weights, &constraint.WV, -*z);
}
let tau_x = {
let mut tau_x_poly = vec![];
tau_x_poly.extend(tau_before_ni);
tau_x_poly.push(V_weights.inner_product(witness.v.iter().map(|v| &v.mask)));
tau_x_poly.extend(tau_after_ni);
let mut tau_x = C::F::ZERO;
for (i, coeff) in tau_x_poly.into_iter().enumerate() {
tau_x += coeff * x[i];
}
tau_x
};
// Calculate u for the powers of x variable to ilr/io/is
let u = {
// Calculate the first part of u
let mut u = (alpha * x[ilr]) + (beta * x[io]) + (rho * x[is]);
// Incorporate the commitment masks multiplied by the associated power of x
for (i, commitment) in witness.c.iter().enumerate() {
let i = i + 1;
u += x[i] * commitment.mask;
}
u
};
// Use the Inner-Product argument to prove for this
// P = t_caret * g + l * g_bold + r * (y_inv * h_bold)
let mut P_terms = Vec::with_capacity(1 + (2 * self.generators.len()));
debug_assert_eq!(l.len(), r.len());
for (i, (l, r)) in l.0.iter().zip(r.0.iter()).enumerate() {
P_terms.push((*l, self.generators.g_bold(i)));
P_terms.push((y_inv[i] * r, self.generators.h_bold(i)));
}
// Protocol 1, inlined, since our IpStatement is for Protocol 2
transcript.push_scalar(tau_x);
transcript.push_scalar(u);
transcript.push_scalar(t_caret);
let ip_x = transcript.challenge();
P_terms.push((ip_x * t_caret, self.generators.g()));
IpStatement::new(
self.generators,
y_inv,
ip_x,
// Safe since IpStatement isn't a ZK proof
P::Prover(multiexp_vartime(&P_terms)),
)
.unwrap()
.prove(transcript, IpWitness::new(l, r).unwrap())
.map_err(AcError::Ip)
}
/// Verify a proof for this statement.
pub fn verify<R: RngCore + CryptoRng>(
self,
rng: &mut R,
verifier: &mut BatchVerifier<C>,
transcript: &mut VerifierTranscript,
) -> Result<(), AcError> {
let n = self.n();
let c = self.c();
let ni = 2 * (c + 1);
let ilr = ni / 2;
let io = ni;
let is = ni + 1;
let jlr = ni / 2;
let l_r_poly_len = 1 + ni + 1;
let t_poly_len = (2 * l_r_poly_len) - 1;
let AI = transcript.read_point::<C>().map_err(|_| AcError::IncompleteProof)?;
let AO = transcript.read_point::<C>().map_err(|_| AcError::IncompleteProof)?;
let S = transcript.read_point::<C>().map_err(|_| AcError::IncompleteProof)?;
let y = transcript.challenge();
let z = transcript.challenge();
let YzChallenges { y_inv, z } = self.yz_challenges(y, z);
let mut l_weights = ScalarVector::new(n);
let mut r_weights = ScalarVector::new(n);
let mut o_weights = ScalarVector::new(n);
for (constraint, z) in self.constraints.iter().zip(&z.0) {
accumulate_vector(&mut l_weights, &constraint.WL, *z);
accumulate_vector(&mut r_weights, &constraint.WR, *z);
accumulate_vector(&mut o_weights, &constraint.WO, *z);
}
let r_weights = r_weights * &y_inv;
let delta = r_weights.inner_product(l_weights.0.iter());
let mut T_before_ni = Vec::with_capacity(ni);
let mut T_after_ni = Vec::with_capacity(t_poly_len - ni - 1);
for _ in 0 .. ni {
T_before_ni.push(transcript.read_point::<C>().map_err(|_| AcError::IncompleteProof)?);
}
for _ in 0 .. (t_poly_len - ni - 1) {
T_after_ni.push(transcript.read_point::<C>().map_err(|_| AcError::IncompleteProof)?);
}
let x: ScalarVector<C::F> = ScalarVector::powers(transcript.challenge(), t_poly_len);
let tau_x = transcript.read_scalar::<C>().map_err(|_| AcError::IncompleteProof)?;
let u = transcript.read_scalar::<C>().map_err(|_| AcError::IncompleteProof)?;
let t_caret = transcript.read_scalar::<C>().map_err(|_| AcError::IncompleteProof)?;
// Lines 88-90, modified per Generalized Bulletproofs as needed w.r.t. t
{
let verifier_weight = C::F::random(&mut *rng);
// lhs of the equation, weighted to enable batch verification
verifier.g += t_caret * verifier_weight;
verifier.h += tau_x * verifier_weight;
let mut V_weights = ScalarVector::new(self.V.len());
for (constraint, z) in self.constraints.iter().zip(&z.0) {
// We use `-z`, not `z`, as we write our constraint as `... + WV V = 0` not `= WV V + ..`
// This means we need to subtract `WV V` from both sides, which we accomplish here
accumulate_vector(&mut V_weights, &constraint.WV, -*z);
}
V_weights = V_weights * x[ni];
// rhs of the equation, negated to cause a sum to zero
// `delta - z...`, instead of `delta + z...`, is done for the same reason as in the above WV
// matrix transform
verifier.g -= verifier_weight *
x[ni] *
(delta - z.inner_product(self.constraints.iter().map(|constraint| &constraint.c)));
for pair in V_weights.0.into_iter().zip(self.V.0) {
verifier.additional.push((-verifier_weight * pair.0, pair.1));
}
for (i, T) in T_before_ni.into_iter().enumerate() {
verifier.additional.push((-verifier_weight * x[i], T));
}
for (i, T) in T_after_ni.into_iter().enumerate() {
verifier.additional.push((-verifier_weight * x[ni + 1 + i], T));
}
}
let verifier_weight = C::F::random(&mut *rng);
// Multiply `x` by `verifier_weight` as this effects `verifier_weight` onto most scalars and
// saves a notable amount of operations
let x = x * verifier_weight;
// This following block effectively calculates P, within the multiexp
{
verifier.additional.push((x[ilr], AI));
verifier.additional.push((x[io], AO));
// h' ** y is equivalent to h as h' is h ** y_inv
let mut log2_n = 0;
while (1 << log2_n) != n {
log2_n += 1;
}
verifier.h_sum[log2_n] -= verifier_weight;
verifier.additional.push((x[is], S));
// Lines 85-87 calculate WL, WR, WO
// We preserve them in terms of g_bold and h_bold for a more efficient multiexp
let mut h_bold_scalars = l_weights * x[jlr];
for (i, wr) in (r_weights * x[jlr]).0.into_iter().enumerate() {
verifier.g_bold[i] += wr;
}
// WO is weighted by x**jo where jo == 0, hence why we can ignore the x term
h_bold_scalars = h_bold_scalars + &(o_weights * verifier_weight);
let mut cg_weights = Vec::with_capacity(self.C.len());
let mut ch_weights = Vec::with_capacity(self.C.len());
for i in 0 .. self.C.len() {
let mut cg = ScalarVector::new(n);
let mut ch = ScalarVector::new(n);
for (constraint, z) in self.constraints.iter().zip(&z.0) {
if let Some(WCG) = constraint.WCG.get(i) {
accumulate_vector(&mut cg, WCG, *z);
}
if let Some(WCH) = constraint.WCH.get(i) {
accumulate_vector(&mut ch, WCH, *z);
}
}
cg_weights.push(cg);
ch_weights.push(ch);
}
// Push the terms for C, which increment from 0, and the terms for WC, which decrement from
// n'
for (i, (C, (WCG, WCH))) in
self.C.0.into_iter().zip(cg_weights.into_iter().zip(ch_weights)).enumerate()
{
let i = i + 1;
let j = ni - i;
verifier.additional.push((x[i], C));
h_bold_scalars = h_bold_scalars + &(WCG * x[j]);
for (i, scalar) in (WCH * &y_inv * x[j]).0.into_iter().enumerate() {
verifier.g_bold[i] += scalar;
}
}
// All terms for h_bold here have actually been for h_bold', h_bold * y_inv
h_bold_scalars = h_bold_scalars * &y_inv;
for (i, scalar) in h_bold_scalars.0.into_iter().enumerate() {
verifier.h_bold[i] += scalar;
}
// Remove u * h from P
verifier.h -= verifier_weight * u;
}
// Prove for lines 88, 92 with an Inner-Product statement
// This inlines Protocol 1, as our IpStatement implements Protocol 2
let ip_x = transcript.challenge();
// P is amended with this additional term
verifier.g += verifier_weight * ip_x * t_caret;
IpStatement::new(self.generators, y_inv, ip_x, P::Verifier { verifier_weight })
.unwrap()
.verify(verifier, transcript)
.map_err(AcError::Ip)?;
Ok(())
}
}