serai/crypto/ed448/src/point.rs

397 lines
9.1 KiB
Rust
Raw Normal View History

FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
use core::{
ops::{Add, AddAssign, Neg, Sub, SubAssign, Mul, MulAssign},
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
iter::Sum,
};
use rand_core::RngCore;
use zeroize::Zeroize;
use subtle::{Choice, CtOption, ConstantTimeEq, ConditionallySelectable, ConditionallyNegatable};
2023-05-09 08:12:13 +00:00
use crypto_bigint::{U448, modular::constant_mod::Residue};
2023-03-07 08:06:46 +00:00
use group::{
ff::{Field, PrimeField, PrimeFieldBits},
Group, GroupEncoding,
prime::PrimeGroup,
};
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
use crate::{
backend::u8_from_bool,
scalar::Scalar,
field::{ResidueType, FieldElement, Q_4},
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
};
const D: FieldElement =
2023-05-09 08:12:13 +00:00
FieldElement(ResidueType::sub(&ResidueType::ZERO, &Residue::new(&U448::from_u16(39081))));
2023-05-09 08:12:13 +00:00
const G_Y: FieldElement = FieldElement(Residue::new(&U448::from_be_hex(concat!(
"693f46716eb6bc248876203756c9c7624bea73736ca3984087789c1e",
"05a0c2d73ad3ff1ce67c39c4fdbd132c4ed7c8ad9808795bf230fa14",
))));
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
2023-05-09 08:12:13 +00:00
const G_X: FieldElement = FieldElement(Residue::new(&U448::from_be_hex(concat!(
"4f1970c66bed0ded221d15a622bf36da9e146570470f1767ea6de324",
"a3d3a46412ae1af72ab66511433b80e18b00938e2626a82bc70cc05e",
))));
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
fn recover_x(y: FieldElement) -> CtOption<FieldElement> {
let ysq = y.square();
#[allow(non_snake_case)]
let D_ysq = D * ysq;
(D_ysq - FieldElement::ONE).invert().and_then(|inverted| {
let temp = (ysq - FieldElement::ONE) * inverted;
let mut x = temp.pow(Q_4);
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
x.conditional_negate(x.is_odd());
let xsq = x.square();
CtOption::new(x, (xsq + ysq).ct_eq(&(FieldElement::ONE + (xsq * D_ysq))))
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
})
}
2023-03-21 00:10:00 +00:00
/// Ed448 point.
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
#[derive(Clone, Copy, Debug, Zeroize)]
pub struct Point {
x: FieldElement,
y: FieldElement,
z: FieldElement,
}
const G: Point = Point { x: G_X, y: G_Y, z: FieldElement::ONE };
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
impl ConstantTimeEq for Point {
fn ct_eq(&self, other: &Self) -> Choice {
let x1 = self.x * other.z;
let x2 = other.x * self.z;
let y1 = self.y * other.z;
let y2 = other.y * self.z;
x1.ct_eq(&x2) & y1.ct_eq(&y2)
}
}
impl PartialEq for Point {
fn eq(&self, other: &Point) -> bool {
self.ct_eq(other).into()
}
}
impl Eq for Point {}
impl ConditionallySelectable for Point {
fn conditional_select(a: &Self, b: &Self, choice: Choice) -> Self {
Point {
x: FieldElement::conditional_select(&a.x, &b.x, choice),
y: FieldElement::conditional_select(&a.y, &b.y, choice),
z: FieldElement::conditional_select(&a.z, &b.z, choice),
}
}
}
impl Add for Point {
type Output = Point;
fn add(self, other: Self) -> Self {
// 12 muls, 7 additions, 4 negations
let xcp = self.x * other.x;
let ycp = self.y * other.y;
let zcp = self.z * other.z;
#[allow(non_snake_case)]
let B = zcp.square();
#[allow(non_snake_case)]
let E = D * xcp * ycp;
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
#[allow(non_snake_case)]
let F = B - E;
#[allow(non_snake_case)]
let G_ = B + E;
Point {
x: zcp * F * ((self.x + self.y) * (other.x + other.y) - xcp - ycp),
y: zcp * G_ * (ycp - xcp),
z: F * G_,
}
}
}
impl AddAssign for Point {
fn add_assign(&mut self, other: Point) {
*self = *self + other;
}
}
impl Add<&Point> for Point {
type Output = Point;
fn add(self, other: &Point) -> Point {
self + *other
}
}
impl AddAssign<&Point> for Point {
fn add_assign(&mut self, other: &Point) {
*self += *other;
}
}
impl Neg for Point {
type Output = Point;
fn neg(self) -> Self {
Point { x: -self.x, y: self.y, z: self.z }
}
}
impl Sub for Point {
type Output = Point;
#[allow(clippy::suspicious_arithmetic_impl)]
fn sub(self, other: Self) -> Self {
self + other.neg()
}
}
impl SubAssign for Point {
fn sub_assign(&mut self, other: Point) {
*self = *self - other;
}
}
impl Sub<&Point> for Point {
type Output = Point;
fn sub(self, other: &Point) -> Point {
self - *other
}
}
impl SubAssign<&Point> for Point {
fn sub_assign(&mut self, other: &Point) {
*self -= *other;
}
}
impl Group for Point {
type Scalar = Scalar;
fn random(mut rng: impl RngCore) -> Self {
loop {
let mut bytes = FieldElement::random(&mut rng).to_repr();
let mut_ref: &mut [u8] = bytes.as_mut();
mut_ref[56] |= u8::try_from(rng.next_u32() % 2).unwrap() << 7;
let opt = Self::from_bytes(&bytes);
if opt.is_some().into() {
return opt.unwrap();
}
}
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
}
fn identity() -> Self {
Point { x: FieldElement::ZERO, y: FieldElement::ONE, z: FieldElement::ONE }
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
}
fn generator() -> Self {
G
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
}
fn is_identity(&self) -> Choice {
self.ct_eq(&Self::identity())
}
fn double(&self) -> Self {
// 7 muls, 7 additions, 4 negations
let xsq = self.x.square();
let ysq = self.y.square();
let zsq = self.z.square();
let xy = self.x + self.y;
#[allow(non_snake_case)]
let F = xsq + ysq;
#[allow(non_snake_case)]
let J = F - zsq.double();
Point { x: J * (xy.square() - xsq - ysq), y: F * (xsq - ysq), z: F * J }
}
}
impl Sum<Point> for Point {
fn sum<I: Iterator<Item = Point>>(iter: I) -> Point {
let mut res = Self::identity();
for i in iter {
res += i;
}
res
}
}
impl<'a> Sum<&'a Point> for Point {
fn sum<I: Iterator<Item = &'a Point>>(iter: I) -> Point {
Point::sum(iter.cloned())
}
}
impl Mul<Scalar> for Point {
type Output = Point;
fn mul(self, mut other: Scalar) -> Point {
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
// Precompute the optimal amount that's a multiple of 2
let mut table = [Point::identity(); 16];
table[1] = self;
for i in 2 .. 16 {
table[i] = table[i - 1] + self;
}
let mut res = Self::identity();
let mut bits = 0;
for (i, mut bit) in other.to_le_bits().iter_mut().rev().enumerate() {
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
bits <<= 1;
let mut bit = u8_from_bool(&mut bit);
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
bits |= bit;
bit.zeroize();
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
if ((i + 1) % 4) == 0 {
if i != 3 {
for _ in 0 .. 4 {
res = res.double();
}
}
res += table[usize::from(bits)];
bits = 0;
}
}
other.zeroize();
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
res
}
}
impl MulAssign<Scalar> for Point {
fn mul_assign(&mut self, other: Scalar) {
*self = *self * other;
}
}
impl Mul<&Scalar> for Point {
type Output = Point;
fn mul(self, other: &Scalar) -> Point {
self * *other
}
}
impl MulAssign<&Scalar> for Point {
fn mul_assign(&mut self, other: &Scalar) {
*self *= *other;
}
}
impl Point {
2023-03-21 00:10:00 +00:00
fn is_torsion_free(&self) -> Choice {
((*self * (Scalar::ZERO - Scalar::ONE)) + self).is_identity()
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
}
}
impl GroupEncoding for Point {
type Repr = <FieldElement as PrimeField>::Repr;
fn from_bytes(bytes: &Self::Repr) -> CtOption<Self> {
// Extract and clear the sign bit
let sign = Choice::from(bytes[56] >> 7);
let mut bytes = *bytes;
let mut_ref: &mut [u8] = bytes.as_mut();
mut_ref[56] &= !(1 << 7);
// Parse y, recover x
FieldElement::from_repr(bytes).and_then(|y| {
recover_x(y).and_then(|mut x| {
x.conditional_negate(x.is_odd().ct_eq(&!sign));
let not_negative_zero = !(x.is_zero() & sign);
let point = Point { x, y, z: FieldElement::ONE };
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
CtOption::new(point, not_negative_zero & point.is_torsion_free())
})
})
}
fn from_bytes_unchecked(bytes: &Self::Repr) -> CtOption<Self> {
Point::from_bytes(bytes)
}
fn to_bytes(&self) -> Self::Repr {
let z = self.z.invert().unwrap();
let x = self.x * z;
let y = self.y * z;
let mut bytes = y.to_repr();
let mut_ref: &mut [u8] = bytes.as_mut();
mut_ref[56] |= x.is_odd().unwrap_u8() << 7;
bytes
}
}
impl PrimeGroup for Point {}
#[test]
fn test_group() {
ff_group_tests::group::test_prime_group_bits::<_, Point>(&mut rand_core::OsRng);
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
}
#[test]
fn generator() {
assert!(G.x == G_X);
assert!(G.y == G_Y);
assert!(recover_x(G.y).unwrap() == G.x);
}
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
#[test]
fn torsion() {
use generic_array::GenericArray;
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
// Uses the originally suggested generator which had torsion
let old_y = FieldElement::from_repr(*GenericArray::from_slice(
&hex::decode(
"\
12796c1532041525945f322e414d434467cfd5c57c9a9af2473b2775\
8c921c4828b277ca5f2891fc4f3d79afdf29a64c72fb28b59c16fa51\
00",
)
.unwrap(),
))
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
.unwrap();
let old = Point { x: -recover_x(old_y).unwrap(), y: old_y, z: FieldElement::ONE };
FROST Ed448 (#107) * Theoretical ed448 impl * Fixes * Basic tests * More efficient scalarmul Precomputes a table to minimize additions required. * Add a torsion test * Split into a constant and variable time backend The variable time one is still far too slow, at 53s for the tests (~5s a scalarmul). It should be usable as a PoC though. * Rename unsafe Ed448 It's not only unworthy of the Serai branding and deserves more clarity in the name. * Add wide reduction to ed448 * Add Zeroize to Ed448 * Rename Ed448 group.rs to point.rs * Minor lint to FROST * Ed448 ciphersuite with 8032 test vector * Macro out the backend fields * Slight efficiency improvement to point decompression * Disable the multiexp test in FROST for Ed448 * fmt + clippy ed448 * Fix an infinite loop in the constant time ed448 backend * Add b"chal" to the 8032 context string for Ed448 Successfully tests against proposed vectors for the FROST IETF draft. * Fix fmt and clippy * Use a tabled pow algorithm in ed448's const backend * Slight tweaks to variable time backend Stop from_repr(MODULUS) from passing. * Use extended points Almost two orders of magnitude faster. * Efficient ed448 doubling * Remove the variable time backend With the recent performance improvements, the constant time backend is now 4x faster than the variable time backend was. While the variable time backend remains much faster, and the constant time backend is still slow compared to other libraries, it's sufficiently performant now. The FROST test, which runs a series of multiexps over the curve, does take 218.26s while Ristretto takes 1 and secp256k1 takes 4.57s. While 50x slower than secp256k1 is horrible, it's ~1.5 orders of magntiude, which is close enough to the desire stated in https://github.com/serai-dex/serai/issues/108 to meet it. Largely makes this library safe to use. * Correct constants in ed448 * Rename unsafe-ed448 to minimal-ed448 Enables all FROST tests against it. * No longer require the hazmat feature to use ed448 * Remove extraneous as_refs
2022-08-29 07:32:59 +00:00
assert!(bool::from(!old.is_torsion_free()));
}
#[test]
fn vector() {
use generic_array::GenericArray;
assert_eq!(
Point::generator().double(),
Point::from_bytes(GenericArray::from_slice(
&hex::decode(
"\
ed8693eacdfbeada6ba0cdd1beb2bcbb98302a3a8365650db8c4d88a\
726de3b7d74d8835a0d76e03b0c2865020d659b38d04d74a63e905ae\
80"
)
.unwrap()
))
.unwrap()
);
assert_eq!(
Point::generator() *
Scalar::from_repr(*GenericArray::from_slice(
&hex::decode(
"\
6298e1eef3c379392caaed061ed8a31033c9e9e3420726f23b404158\
a401cd9df24632adfe6b418dc942d8a091817dd8bd70e1c72ba52f3c\
00"
)
.unwrap()
))
.unwrap(),
Point::from_bytes(GenericArray::from_slice(
&hex::decode(
"\
3832f82fda00ff5365b0376df705675b63d2a93c24c6e81d40801ba2\
65632be10f443f95968fadb70d10786827f30dc001c8d0f9b7c1d1b0\
00"
)
.unwrap()
))
.unwrap()
);
}
// Checks random won't infinitely loop
#[test]
fn random() {
Point::random(&mut rand_core::OsRng);
}