1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
//! Fast, SIMD-accelerated CRC32 (IEEE) checksum computation.
//!
//! ## Usage
//!
//! ### Simple usage
//!
//! For simple use-cases, you can call the [`hash()`] convenience function to
//! directly compute the CRC32 checksum for a given byte slice:
//!
//! ```rust
//! let checksum = crc32fast::hash(b"foo bar baz");
//! ```
//!
//! ### Advanced usage
//!
//! For use-cases that require more flexibility or performance, for example when
//! processing large amounts of data, you can create and manipulate a [`Hasher`]:
//!
//! ```rust
//! use crc32fast::Hasher;
//!
//! let mut hasher = Hasher::new();
//! hasher.update(b"foo bar baz");
//! let checksum = hasher.finalize();
//! ```
//!
//! ## Performance
//!
//! This crate contains multiple CRC32 implementations:
//!
//! - A fast baseline implementation which processes up to 16 bytes per iteration
//! - An optimized implementation for modern `x86` using `sse` and `pclmulqdq` instructions
//!
//! Calling the [`Hasher::new`] constructor at runtime will perform a feature detection to select the most
//! optimal implementation for the current CPU feature set.
#![cfg_attr(not(feature = "std"), no_std)]
#![cfg_attr(
all(feature = "nightly", target_arch = "aarch64"),
feature(stdsimd, aarch64_target_feature)
)]
#[deny(missing_docs)]
#[cfg(test)]
#[macro_use]
extern crate quickcheck;
#[macro_use]
extern crate cfg_if;
#[cfg(feature = "std")]
use std as core;
use core::fmt;
use core::hash;
mod baseline;
mod combine;
mod specialized;
mod table;
/// Computes the CRC32 hash of a byte slice.
///
/// Check out [`Hasher`] for more advanced use-cases.
pub fn hash(buf: &[u8]) -> u32 {
let mut h = Hasher::new();
h.update(buf);
h.finalize()
}
#[derive(Clone)]
enum State {
Baseline(baseline::State),
Specialized(specialized::State),
}
#[derive(Clone)]
/// Represents an in-progress CRC32 computation.
pub struct Hasher {
amount: u64,
state: State,
}
const DEFAULT_INIT_STATE: u32 = 0;
impl Hasher {
/// Create a new `Hasher`.
///
/// This will perform a CPU feature detection at runtime to select the most
/// optimal implementation for the current processor architecture.
pub fn new() -> Self {
Self::new_with_initial(DEFAULT_INIT_STATE)
}
/// Create a new `Hasher` with an initial CRC32 state.
///
/// This works just like `Hasher::new`, except that it allows for an initial
/// CRC32 state to be passed in.
pub fn new_with_initial(init: u32) -> Self {
Self::new_with_initial_len(init, 0)
}
/// Create a new `Hasher` with an initial CRC32 state.
///
/// As `new_with_initial`, but also accepts a length (in bytes). The
/// resulting object can then be used with `combine` to compute `crc(a ||
/// b)` from `crc(a)`, `crc(b)`, and `len(b)`.
pub fn new_with_initial_len(init: u32, amount: u64) -> Self {
Self::internal_new_specialized(init, amount)
.unwrap_or_else(|| Self::internal_new_baseline(init, amount))
}
#[doc(hidden)]
// Internal-only API. Don't use.
pub fn internal_new_baseline(init: u32, amount: u64) -> Self {
Hasher {
amount,
state: State::Baseline(baseline::State::new(init)),
}
}
#[doc(hidden)]
// Internal-only API. Don't use.
pub fn internal_new_specialized(init: u32, amount: u64) -> Option<Self> {
{
if let Some(state) = specialized::State::new(init) {
return Some(Hasher {
amount,
state: State::Specialized(state),
});
}
}
None
}
/// Process the given byte slice and update the hash state.
pub fn update(&mut self, buf: &[u8]) {
self.amount += buf.len() as u64;
match self.state {
State::Baseline(ref mut state) => state.update(buf),
State::Specialized(ref mut state) => state.update(buf),
}
}
/// Finalize the hash state and return the computed CRC32 value.
pub fn finalize(self) -> u32 {
match self.state {
State::Baseline(state) => state.finalize(),
State::Specialized(state) => state.finalize(),
}
}
/// Reset the hash state.
pub fn reset(&mut self) {
self.amount = 0;
match self.state {
State::Baseline(ref mut state) => state.reset(),
State::Specialized(ref mut state) => state.reset(),
}
}
/// Combine the hash state with the hash state for the subsequent block of bytes.
pub fn combine(&mut self, other: &Self) {
self.amount += other.amount;
let other_crc = other.clone().finalize();
match self.state {
State::Baseline(ref mut state) => state.combine(other_crc, other.amount),
State::Specialized(ref mut state) => state.combine(other_crc, other.amount),
}
}
}
impl fmt::Debug for Hasher {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.debug_struct("crc32fast::Hasher").finish()
}
}
impl Default for Hasher {
fn default() -> Self {
Self::new()
}
}
impl hash::Hasher for Hasher {
fn write(&mut self, bytes: &[u8]) {
self.update(bytes)
}
fn finish(&self) -> u64 {
u64::from(self.clone().finalize())
}
}
#[cfg(test)]
mod test {
use super::Hasher;
quickcheck! {
fn combine(bytes_1: Vec<u8>, bytes_2: Vec<u8>) -> bool {
let mut hash_a = Hasher::new();
hash_a.update(&bytes_1);
hash_a.update(&bytes_2);
let mut hash_b = Hasher::new();
hash_b.update(&bytes_2);
let mut hash_c = Hasher::new();
hash_c.update(&bytes_1);
hash_c.combine(&hash_b);
hash_a.finalize() == hash_c.finalize()
}
fn combine_from_len(bytes_1: Vec<u8>, bytes_2: Vec<u8>) -> bool {
let mut hash_a = Hasher::new();
hash_a.update(&bytes_1);
let a = hash_a.finalize();
let mut hash_b = Hasher::new();
hash_b.update(&bytes_2);
let b = hash_b.finalize();
let mut hash_ab = Hasher::new();
hash_ab.update(&bytes_1);
hash_ab.update(&bytes_2);
let ab = hash_ab.finalize();
let mut reconstructed = Hasher::new_with_initial_len(a, bytes_1.len() as u64);
let hash_b_reconstructed = Hasher::new_with_initial_len(b, bytes_2.len() as u64);
reconstructed.combine(&hash_b_reconstructed);
reconstructed.finalize() == ab
}
}
}