ff/lib.rs
1//! This crate provides traits for working with finite fields.
2
3#![no_std]
4#![cfg_attr(docsrs, feature(doc_cfg))]
5// Catch documentation errors caused by code changes.
6#![deny(rustdoc::broken_intra_doc_links)]
7#![forbid(unsafe_code)]
8
9#[cfg(feature = "alloc")]
10extern crate alloc;
11
12mod batch;
13pub use batch::*;
14
15pub mod helpers;
16
17#[cfg(feature = "derive")]
18#[cfg_attr(docsrs, doc(cfg(feature = "derive")))]
19pub use ff_derive::PrimeField;
20
21#[cfg(feature = "bits")]
22#[cfg_attr(docsrs, doc(cfg(feature = "bits")))]
23pub use bitvec::view::BitViewSized;
24
25#[cfg(feature = "bits")]
26use bitvec::{array::BitArray, order::Lsb0};
27
28use core::fmt;
29use core::iter::{Product, Sum};
30use core::ops::{Add, AddAssign, Mul, MulAssign, Neg, Sub, SubAssign};
31
32use rand_core::RngCore;
33use subtle::{Choice, ConditionallySelectable, ConstantTimeEq, CtOption};
34
35/// Bit representation of a field element.
36#[cfg(feature = "bits")]
37#[cfg_attr(docsrs, doc(cfg(feature = "bits")))]
38pub type FieldBits<V> = BitArray<V, Lsb0>;
39
40/// This trait represents an element of a field.
41pub trait Field:
42 Sized
43 + Eq
44 + Copy
45 + Clone
46 + Default
47 + Send
48 + Sync
49 + fmt::Debug
50 + 'static
51 + ConditionallySelectable
52 + ConstantTimeEq
53 + Neg<Output = Self>
54 + Add<Output = Self>
55 + Sub<Output = Self>
56 + Mul<Output = Self>
57 + Sum
58 + Product
59 + for<'a> Add<&'a Self, Output = Self>
60 + for<'a> Sub<&'a Self, Output = Self>
61 + for<'a> Mul<&'a Self, Output = Self>
62 + for<'a> Sum<&'a Self>
63 + for<'a> Product<&'a Self>
64 + AddAssign
65 + SubAssign
66 + MulAssign
67 + for<'a> AddAssign<&'a Self>
68 + for<'a> SubAssign<&'a Self>
69 + for<'a> MulAssign<&'a Self>
70{
71 /// The zero element of the field, the additive identity.
72 const ZERO: Self;
73
74 /// The one element of the field, the multiplicative identity.
75 const ONE: Self;
76
77 /// Returns an element chosen uniformly at random using a user-provided RNG.
78 fn random(rng: impl RngCore) -> Self;
79
80 /// Returns true iff this element is zero.
81 fn is_zero(&self) -> Choice {
82 self.ct_eq(&Self::ZERO)
83 }
84
85 /// Returns true iff this element is zero.
86 ///
87 /// # Security
88 ///
89 /// This method provides **no** constant-time guarantees. Implementors of the
90 /// `Field` trait **may** optimise this method using non-constant-time logic.
91 fn is_zero_vartime(&self) -> bool {
92 self.is_zero().into()
93 }
94
95 /// Squares this element.
96 #[must_use]
97 fn square(&self) -> Self;
98
99 /// Cubes this element.
100 #[must_use]
101 fn cube(&self) -> Self {
102 self.square() * self
103 }
104
105 /// Doubles this element.
106 #[must_use]
107 fn double(&self) -> Self;
108
109 /// Computes the multiplicative inverse of this element,
110 /// failing if the element is zero.
111 fn invert(&self) -> CtOption<Self>;
112
113 /// Computes:
114 ///
115 /// - $(\textsf{true}, \sqrt{\textsf{num}/\textsf{div}})$, if $\textsf{num}$ and
116 /// $\textsf{div}$ are nonzero and $\textsf{num}/\textsf{div}$ is a square in the
117 /// field;
118 /// - $(\textsf{true}, 0)$, if $\textsf{num}$ is zero;
119 /// - $(\textsf{false}, 0)$, if $\textsf{num}$ is nonzero and $\textsf{div}$ is zero;
120 /// - $(\textsf{false}, \sqrt{G_S \cdot \textsf{num}/\textsf{div}})$, if
121 /// $\textsf{num}$ and $\textsf{div}$ are nonzero and $\textsf{num}/\textsf{div}$ is
122 /// a nonsquare in the field;
123 ///
124 /// where $G_S$ is a non-square.
125 ///
126 /// # Warnings
127 ///
128 /// - The choice of root from `sqrt` is unspecified.
129 /// - The value of $G_S$ is unspecified, and cannot be assumed to have any specific
130 /// value in a generic context.
131 fn sqrt_ratio(num: &Self, div: &Self) -> (Choice, Self);
132
133 /// Equivalent to `Self::sqrt_ratio(self, one())`.
134 ///
135 /// The provided method is implemented in terms of [`Self::sqrt_ratio`].
136 fn sqrt_alt(&self) -> (Choice, Self) {
137 Self::sqrt_ratio(self, &Self::ONE)
138 }
139
140 /// Returns the square root of the field element, if it is
141 /// quadratic residue.
142 ///
143 /// The provided method is implemented in terms of [`Self::sqrt_ratio`].
144 fn sqrt(&self) -> CtOption<Self> {
145 let (is_square, res) = Self::sqrt_ratio(self, &Self::ONE);
146 CtOption::new(res, is_square)
147 }
148
149 /// Exponentiates `self` by `exp`, where `exp` is a little-endian order integer
150 /// exponent.
151 ///
152 /// # Guarantees
153 ///
154 /// This operation is constant time with respect to `self`, for all exponents with the
155 /// same number of digits (`exp.as_ref().len()`). It is variable time with respect to
156 /// the number of digits in the exponent.
157 fn pow<S: AsRef<[u64]>>(&self, exp: S) -> Self {
158 let mut res = Self::ONE;
159 for e in exp.as_ref().iter().rev() {
160 for i in (0..64).rev() {
161 res = res.square();
162 let mut tmp = res;
163 tmp *= self;
164 res.conditional_assign(&tmp, (((*e >> i) & 1) as u8).into());
165 }
166 }
167 res
168 }
169
170 /// Exponentiates `self` by `exp`, where `exp` is a little-endian order integer
171 /// exponent.
172 ///
173 /// # Guarantees
174 ///
175 /// **This operation is variable time with respect to `self`, for all exponent.** If
176 /// the exponent is fixed, this operation is effectively constant time. However, for
177 /// stronger constant-time guarantees, [`Field::pow`] should be used.
178 fn pow_vartime<S: AsRef<[u64]>>(&self, exp: S) -> Self {
179 let mut res = Self::ONE;
180 for e in exp.as_ref().iter().rev() {
181 for i in (0..64).rev() {
182 res = res.square();
183
184 if ((*e >> i) & 1) == 1 {
185 res.mul_assign(self);
186 }
187 }
188 }
189
190 res
191 }
192}
193
194/// This represents an element of a non-binary prime field.
195pub trait PrimeField: Field + From<u64> {
196 /// The prime field can be converted back and forth into this binary
197 /// representation.
198 type Repr: Copy + Default + Send + Sync + 'static + AsRef<[u8]> + AsMut<[u8]>;
199
200 /// Interpret a string of numbers as a (congruent) prime field element.
201 /// Does not accept unnecessary leading zeroes or a blank string.
202 ///
203 /// # Security
204 ///
205 /// This method provides **no** constant-time guarantees.
206 fn from_str_vartime(s: &str) -> Option<Self> {
207 if s.is_empty() {
208 return None;
209 }
210
211 if s == "0" {
212 return Some(Self::ZERO);
213 }
214
215 let mut res = Self::ZERO;
216
217 let ten = Self::from(10);
218
219 let mut first_digit = true;
220
221 for c in s.chars() {
222 match c.to_digit(10) {
223 Some(c) => {
224 if first_digit {
225 if c == 0 {
226 return None;
227 }
228
229 first_digit = false;
230 }
231
232 res.mul_assign(&ten);
233 res.add_assign(&Self::from(u64::from(c)));
234 }
235 None => {
236 return None;
237 }
238 }
239 }
240
241 Some(res)
242 }
243
244 /// Obtains a field element congruent to the integer `v`.
245 ///
246 /// For fields where `Self::CAPACITY >= 128`, this is injective and will produce a
247 /// unique field element.
248 ///
249 /// For fields where `Self::CAPACITY < 128`, this is surjective; some field elements
250 /// will be produced by multiple values of `v`.
251 ///
252 /// If you want to deterministically sample a field element representing a value, use
253 /// [`FromUniformBytes`] instead.
254 fn from_u128(v: u128) -> Self {
255 let lower = v as u64;
256 let upper = (v >> 64) as u64;
257 let mut tmp = Self::from(upper);
258 for _ in 0..64 {
259 tmp = tmp.double();
260 }
261 tmp + Self::from(lower)
262 }
263
264 /// Attempts to convert a byte representation of a field element into an element of
265 /// this prime field, failing if the input is not canonical (is not smaller than the
266 /// field's modulus).
267 ///
268 /// The byte representation is interpreted with the same endianness as elements
269 /// returned by [`PrimeField::to_repr`].
270 fn from_repr(repr: Self::Repr) -> CtOption<Self>;
271
272 /// Attempts to convert a byte representation of a field element into an element of
273 /// this prime field, failing if the input is not canonical (is not smaller than the
274 /// field's modulus).
275 ///
276 /// The byte representation is interpreted with the same endianness as elements
277 /// returned by [`PrimeField::to_repr`].
278 ///
279 /// # Security
280 ///
281 /// This method provides **no** constant-time guarantees. Implementors of the
282 /// `PrimeField` trait **may** optimise this method using non-constant-time logic.
283 fn from_repr_vartime(repr: Self::Repr) -> Option<Self> {
284 Self::from_repr(repr).into()
285 }
286
287 /// Converts an element of the prime field into the standard byte representation for
288 /// this field.
289 ///
290 /// The endianness of the byte representation is implementation-specific. Generic
291 /// encodings of field elements should be treated as opaque.
292 fn to_repr(&self) -> Self::Repr;
293
294 /// Returns true iff this element is odd.
295 fn is_odd(&self) -> Choice;
296
297 /// Returns true iff this element is even.
298 #[inline(always)]
299 fn is_even(&self) -> Choice {
300 !self.is_odd()
301 }
302
303 /// Modulus of the field written as a string for debugging purposes.
304 ///
305 /// The encoding of the modulus is implementation-specific. Generic users of the
306 /// `PrimeField` trait should treat this string as opaque.
307 const MODULUS: &'static str;
308
309 /// How many bits are needed to represent an element of this field.
310 const NUM_BITS: u32;
311
312 /// How many bits of information can be reliably stored in the field element.
313 ///
314 /// This is usually `Self::NUM_BITS - 1`.
315 const CAPACITY: u32;
316
317 /// Inverse of $2$ in the field.
318 const TWO_INV: Self;
319
320 /// A fixed multiplicative generator of `modulus - 1` order. This element must also be
321 /// a quadratic nonresidue.
322 ///
323 /// It can be calculated using [SageMath] as `GF(modulus).primitive_element()`.
324 ///
325 /// Implementations of this trait MUST ensure that this is the generator used to
326 /// derive `Self::ROOT_OF_UNITY`.
327 ///
328 /// [SageMath]: https://www.sagemath.org/
329 const MULTIPLICATIVE_GENERATOR: Self;
330
331 /// An integer `s` satisfying the equation `2^s * t = modulus - 1` with `t` odd.
332 ///
333 /// This is the number of leading zero bits in the little-endian bit representation of
334 /// `modulus - 1`.
335 const S: u32;
336
337 /// The `2^s` root of unity.
338 ///
339 /// It can be calculated by exponentiating `Self::MULTIPLICATIVE_GENERATOR` by `t`,
340 /// where `t = (modulus - 1) >> Self::S`.
341 const ROOT_OF_UNITY: Self;
342
343 /// Inverse of [`Self::ROOT_OF_UNITY`].
344 const ROOT_OF_UNITY_INV: Self;
345
346 /// Generator of the `t-order` multiplicative subgroup.
347 ///
348 /// It can be calculated by exponentiating [`Self::MULTIPLICATIVE_GENERATOR`] by `2^s`,
349 /// where `s` is [`Self::S`].
350 const DELTA: Self;
351}
352
353/// The subset of prime-order fields such that `(modulus - 1)` is divisible by `N`.
354///
355/// If `N` is prime, there will be `N - 1` valid choices of [`Self::ZETA`]. Similarly to
356/// [`PrimeField::MULTIPLICATIVE_GENERATOR`], the specific choice does not matter, as long
357/// as the choice is consistent across all uses of the field.
358pub trait WithSmallOrderMulGroup<const N: u8>: PrimeField {
359 /// A field element of small multiplicative order $N$.
360 ///
361 /// The presence of this element allows you to perform (certain types of)
362 /// endomorphisms on some elliptic curves.
363 ///
364 /// It can be calculated using [SageMath] as
365 /// `GF(modulus).primitive_element() ^ ((modulus - 1) // N)`.
366 /// Choosing the element of order $N$ that is smallest, when considered
367 /// as an integer, may help to ensure consistency.
368 ///
369 /// [SageMath]: https://www.sagemath.org/
370 const ZETA: Self;
371}
372
373/// Trait for constructing a [`PrimeField`] element from a fixed-length uniform byte
374/// array.
375///
376/// "Uniform" means that the byte array's contents must be indistinguishable from the
377/// [discrete uniform distribution]. Suitable byte arrays can be obtained:
378/// - from a cryptographically-secure randomness source (which makes this constructor
379/// equivalent to [`Field::random`]).
380/// - from a cryptographic hash function output, which enables a "random" field element to
381/// be selected deterministically. This is the primary use case for `FromUniformBytes`.
382///
383/// The length `N` of the byte array is chosen by the trait implementer such that the loss
384/// of uniformity in the mapping from byte arrays to field elements is cryptographically
385/// negligible.
386///
387/// [discrete uniform distribution]: https://en.wikipedia.org/wiki/Discrete_uniform_distribution
388///
389/// # Examples
390///
391/// ```
392/// # #[cfg(feature = "derive")] {
393/// # // Fake this so we don't actually need a dev-dependency on bls12_381.
394/// # mod bls12_381 {
395/// # use ff::{Field, PrimeField};
396/// #
397/// # #[derive(PrimeField)]
398/// # #[PrimeFieldModulus = "52435875175126190479447740508185965837690552500527637822603658699938581184513"]
399/// # #[PrimeFieldGenerator = "7"]
400/// # #[PrimeFieldReprEndianness = "little"]
401/// # pub struct Scalar([u64; 4]);
402/// #
403/// # impl ff::FromUniformBytes<64> for Scalar {
404/// # fn from_uniform_bytes(_bytes: &[u8; 64]) -> Self {
405/// # // Fake impl for doctest
406/// # Scalar::ONE
407/// # }
408/// # }
409/// # }
410/// #
411/// use blake2b_simd::blake2b;
412/// use bls12_381::Scalar;
413/// use ff::FromUniformBytes;
414///
415/// // `bls12_381::Scalar` implements `FromUniformBytes<64>`, and BLAKE2b (by default)
416/// // produces a 64-byte hash.
417/// let hash = blake2b(b"Some message");
418/// let val = Scalar::from_uniform_bytes(hash.as_array());
419/// # }
420/// ```
421///
422/// # Implementing `FromUniformBytes`
423///
424/// [`Self::from_uniform_bytes`] should always be implemented by interpreting the provided
425/// byte array as the little endian unsigned encoding of an integer, and then reducing that
426/// integer modulo the field modulus.
427///
428/// For security, `N` must be chosen so that `N * 8 >= Self::NUM_BITS + 128`. A larger
429/// value of `N` may be chosen for convenience; for example, for a field with a 255-bit
430/// modulus, `N = 64` is convenient as it matches the output length of several common
431/// cryptographic hash functions (such as SHA-512 and BLAKE2b).
432///
433/// ## Trait design
434///
435/// This trait exists because `PrimeField::from_uniform_bytes([u8; N])` cannot currently
436/// exist (trait methods cannot use associated constants in the const positions of their
437/// type signature, and we do not want `PrimeField` to require a generic const parameter).
438/// However, this has the side-effect that `FromUniformBytes` can be implemented multiple
439/// times for different values of `N`. Most implementations of [`PrimeField`] should only
440/// need to implement `FromUniformBytes` trait for one value of `N` (chosen following the
441/// above considerations); if you find yourself needing to implement it multiple times,
442/// please [let us know about your use case](https://github.com/zkcrypto/ff/issues/new) so
443/// we can take it into consideration for future evolutions of the `ff` traits.
444pub trait FromUniformBytes<const N: usize>: PrimeField {
445 /// Returns a field element that is congruent to the provided little endian unsigned
446 /// byte representation of an integer.
447 fn from_uniform_bytes(bytes: &[u8; N]) -> Self;
448}
449
450/// This represents the bits of an element of a prime field.
451#[cfg(feature = "bits")]
452#[cfg_attr(docsrs, doc(cfg(feature = "bits")))]
453pub trait PrimeFieldBits: PrimeField {
454 /// The backing store for a bit representation of a prime field element.
455 type ReprBits: BitViewSized + Send + Sync;
456
457 /// Converts an element of the prime field into a little-endian sequence of bits.
458 fn to_le_bits(&self) -> FieldBits<Self::ReprBits>;
459
460 /// Returns the bits of the field characteristic (the modulus) in little-endian order.
461 fn char_le_bits() -> FieldBits<Self::ReprBits>;
462}
463
464/// Functions and re-exported crates used by the [`PrimeField`] derive macro.
465#[cfg(feature = "derive")]
466#[cfg_attr(docsrs, doc(cfg(feature = "derive")))]
467pub mod derive {
468 pub use crate::arith_impl::*;
469
470 pub use {byteorder, rand_core, subtle};
471
472 #[cfg(feature = "bits")]
473 pub use bitvec;
474}
475
476#[cfg(feature = "derive")]
477mod arith_impl {
478 /// Computes `a - (b + borrow)`, returning the result and the new borrow.
479 #[inline(always)]
480 pub const fn sbb(a: u64, b: u64, borrow: u64) -> (u64, u64) {
481 let ret = (a as u128).wrapping_sub((b as u128) + ((borrow >> 63) as u128));
482 (ret as u64, (ret >> 64) as u64)
483 }
484
485 /// Computes `a + b + carry`, returning the result and the new carry over.
486 #[inline(always)]
487 pub const fn adc(a: u64, b: u64, carry: u64) -> (u64, u64) {
488 let ret = (a as u128) + (b as u128) + (carry as u128);
489 (ret as u64, (ret >> 64) as u64)
490 }
491
492 /// Computes `a + (b * c) + carry`, returning the result and the new carry over.
493 #[inline(always)]
494 pub const fn mac(a: u64, b: u64, c: u64, carry: u64) -> (u64, u64) {
495 let ret = (a as u128) + ((b as u128) * (c as u128)) + (carry as u128);
496 (ret as u64, (ret >> 64) as u64)
497 }
498}