@noble/ciphers
Version:
Audited & minimal JS implementation of Salsa20, ChaCha and AES
495 lines (443 loc) • 17.3 kB
text/typescript
/**
* Utilities for hex, bytes, CSPRNG.
* @module
*/
/*! noble-ciphers - MIT License (c) 2023 Paul Miller (paulmillr.com) */
/** Checks if something is Uint8Array. Be careful: nodejs Buffer will return true. */
export function isBytes(a: unknown): a is Uint8Array {
return a instanceof Uint8Array || (ArrayBuffer.isView(a) && a.constructor.name === 'Uint8Array');
}
/** Asserts something is boolean. */
export function abool(b: boolean): void {
if (typeof b !== 'boolean') throw new Error(`boolean expected, not ${b}`);
}
/** Asserts something is positive integer. */
export function anumber(n: number): void {
if (!Number.isSafeInteger(n) || n < 0) throw new Error('positive integer expected, got ' + n);
}
/** Asserts something is Uint8Array. */
export function abytes(value: Uint8Array, length?: number, title: string = ''): Uint8Array {
const bytes = isBytes(value);
const len = value?.length;
const needsLen = length !== undefined;
if (!bytes || (needsLen && len !== length)) {
const prefix = title && `"${title}" `;
const ofLen = needsLen ? ` of length ${length}` : '';
const got = bytes ? `length=${len}` : `type=${typeof value}`;
throw new Error(prefix + 'expected Uint8Array' + ofLen + ', got ' + got);
}
return value;
}
/** Asserts a hash instance has not been destroyed / finished */
export function aexists(instance: any, checkFinished = true): void {
if (instance.destroyed) throw new Error('Hash instance has been destroyed');
if (checkFinished && instance.finished) throw new Error('Hash#digest() has already been called');
}
/** Asserts output is properly-sized byte array */
export function aoutput(out: any, instance: any): void {
abytes(out, undefined, 'output');
const min = instance.outputLen;
if (out.length < min) {
throw new Error('digestInto() expects output buffer of length at least ' + min);
}
}
export type IHash = {
(data: string | Uint8Array): Uint8Array;
blockLen: number;
outputLen: number;
create: any;
};
/** Generic type encompassing 8/16/32-byte arrays - but not 64-byte. */
// prettier-ignore
export type TypedArray = Int8Array | Uint8ClampedArray | Uint8Array |
Uint16Array | Int16Array | Uint32Array | Int32Array;
/** Cast u8 / u16 / u32 to u8. */
export function u8(arr: TypedArray): Uint8Array {
return new Uint8Array(arr.buffer, arr.byteOffset, arr.byteLength);
}
/** Cast u8 / u16 / u32 to u32. */
export function u32(arr: TypedArray): Uint32Array {
return new Uint32Array(arr.buffer, arr.byteOffset, Math.floor(arr.byteLength / 4));
}
/** Zeroize a byte array. Warning: JS provides no guarantees. */
export function clean(...arrays: TypedArray[]): void {
for (let i = 0; i < arrays.length; i++) {
arrays[i].fill(0);
}
}
/** Create DataView of an array for easy byte-level manipulation. */
export function createView(arr: TypedArray): DataView {
return new DataView(arr.buffer, arr.byteOffset, arr.byteLength);
}
/** Is current platform little-endian? Most are. Big-Endian platform: IBM */
export const isLE: boolean = /* @__PURE__ */ (() =>
new Uint8Array(new Uint32Array([0x11223344]).buffer)[0] === 0x44)();
// Built-in hex conversion https://caniuse.com/mdn-javascript_builtins_uint8array_fromhex
const hasHexBuiltin: boolean = /* @__PURE__ */ (() =>
// @ts-ignore
typeof Uint8Array.from([]).toHex === 'function' && typeof Uint8Array.fromHex === 'function')();
// Array where index 0xf0 (240) is mapped to string 'f0'
const hexes = /* @__PURE__ */ Array.from({ length: 256 }, (_, i) =>
i.toString(16).padStart(2, '0')
);
/**
* Convert byte array to hex string. Uses built-in function, when available.
* @example bytesToHex(Uint8Array.from([0xca, 0xfe, 0x01, 0x23])) // 'cafe0123'
*/
export function bytesToHex(bytes: Uint8Array): string {
abytes(bytes);
// @ts-ignore
if (hasHexBuiltin) return bytes.toHex();
// pre-caching improves the speed 6x
let hex = '';
for (let i = 0; i < bytes.length; i++) {
hex += hexes[bytes[i]];
}
return hex;
}
// We use optimized technique to convert hex string to byte array
const asciis = { _0: 48, _9: 57, A: 65, F: 70, a: 97, f: 102 } as const;
function asciiToBase16(ch: number): number | undefined {
if (ch >= asciis._0 && ch <= asciis._9) return ch - asciis._0; // '2' => 50-48
if (ch >= asciis.A && ch <= asciis.F) return ch - (asciis.A - 10); // 'B' => 66-(65-10)
if (ch >= asciis.a && ch <= asciis.f) return ch - (asciis.a - 10); // 'b' => 98-(97-10)
return;
}
/**
* Convert hex string to byte array. Uses built-in function, when available.
* @example hexToBytes('cafe0123') // Uint8Array.from([0xca, 0xfe, 0x01, 0x23])
*/
export function hexToBytes(hex: string): Uint8Array {
if (typeof hex !== 'string') throw new Error('hex string expected, got ' + typeof hex);
// @ts-ignore
if (hasHexBuiltin) return Uint8Array.fromHex(hex);
const hl = hex.length;
const al = hl / 2;
if (hl % 2) throw new Error('hex string expected, got unpadded hex of length ' + hl);
const array = new Uint8Array(al);
for (let ai = 0, hi = 0; ai < al; ai++, hi += 2) {
const n1 = asciiToBase16(hex.charCodeAt(hi));
const n2 = asciiToBase16(hex.charCodeAt(hi + 1));
if (n1 === undefined || n2 === undefined) {
const char = hex[hi] + hex[hi + 1];
throw new Error('hex string expected, got non-hex character "' + char + '" at index ' + hi);
}
array[ai] = n1 * 16 + n2; // multiply first octet, e.g. 'a3' => 10*16+3 => 160 + 3 => 163
}
return array;
}
// Used in micro
export function hexToNumber(hex: string): bigint {
if (typeof hex !== 'string') throw new Error('hex string expected, got ' + typeof hex);
return BigInt(hex === '' ? '0' : '0x' + hex); // Big Endian
}
// Used in ff1
// BE: Big Endian, LE: Little Endian
export function bytesToNumberBE(bytes: Uint8Array): bigint {
return hexToNumber(bytesToHex(bytes));
}
// Used in micro, ff1
export function numberToBytesBE(n: number | bigint, len: number): Uint8Array {
return hexToBytes(n.toString(16).padStart(len * 2, '0'));
}
// Global symbols, but ts doesn't see them: https://github.com/microsoft/TypeScript/issues/31535
declare const TextEncoder: any;
declare const TextDecoder: any;
/**
* Converts string to bytes using UTF8 encoding.
* @example utf8ToBytes('abc') // new Uint8Array([97, 98, 99])
*/
export function utf8ToBytes(str: string): Uint8Array {
if (typeof str !== 'string') throw new Error('string expected');
return new Uint8Array(new TextEncoder().encode(str)); // https://bugzil.la/1681809
}
/**
* Converts bytes to string using UTF8 encoding.
* @example bytesToUtf8(new Uint8Array([97, 98, 99])) // 'abc'
*/
export function bytesToUtf8(bytes: Uint8Array): string {
return new TextDecoder().decode(bytes);
}
/**
* Checks if two U8A use same underlying buffer and overlaps.
* This is invalid and can corrupt data.
*/
export function overlapBytes(a: Uint8Array, b: Uint8Array): boolean {
return (
a.buffer === b.buffer && // best we can do, may fail with an obscure Proxy
a.byteOffset < b.byteOffset + b.byteLength && // a starts before b end
b.byteOffset < a.byteOffset + a.byteLength // b starts before a end
);
}
/**
* If input and output overlap and input starts before output, we will overwrite end of input before
* we start processing it, so this is not supported for most ciphers (except chacha/salse, which designed with this)
*/
export function complexOverlapBytes(input: Uint8Array, output: Uint8Array): void {
// This is very cursed. It works somehow, but I'm completely unsure,
// reasoning about overlapping aligned windows is very hard.
if (overlapBytes(input, output) && input.byteOffset < output.byteOffset)
throw new Error('complex overlap of input and output is not supported');
}
/**
* Copies several Uint8Arrays into one.
*/
export function concatBytes(...arrays: Uint8Array[]): Uint8Array {
let sum = 0;
for (let i = 0; i < arrays.length; i++) {
const a = arrays[i];
abytes(a);
sum += a.length;
}
const res = new Uint8Array(sum);
for (let i = 0, pad = 0; i < arrays.length; i++) {
const a = arrays[i];
res.set(a, pad);
pad += a.length;
}
return res;
}
// Used in ARX only
type EmptyObj = {};
export function checkOpts<T1 extends EmptyObj, T2 extends EmptyObj>(
defaults: T1,
opts: T2
): T1 & T2 {
if (opts == null || typeof opts !== 'object') throw new Error('options must be defined');
const merged = Object.assign(defaults, opts);
return merged as T1 & T2;
}
/** Compares 2 uint8array-s in kinda constant time. */
export function equalBytes(a: Uint8Array, b: Uint8Array): boolean {
if (a.length !== b.length) return false;
let diff = 0;
for (let i = 0; i < a.length; i++) diff |= a[i] ^ b[i];
return diff === 0;
}
// TODO: remove
export interface IHash2 {
blockLen: number; // Bytes per block
outputLen: number; // Bytes in output
update(buf: string | Uint8Array): this;
// Writes digest into buf
digestInto(buf: Uint8Array): void;
digest(): Uint8Array;
/**
* Resets internal state. Makes Hash instance unusable.
* Reset is impossible for keyed hashes if key is consumed into state. If digest is not consumed
* by user, they will need to manually call `destroy()` when zeroing is necessary.
*/
destroy(): void;
}
// This will allow to re-use with composable things like packed & base encoders
// Also, we probably can make tags composable
/** Sync cipher: takes byte array and returns byte array. */
export type Cipher = {
encrypt(plaintext: Uint8Array): Uint8Array;
decrypt(ciphertext: Uint8Array): Uint8Array;
};
/** Async cipher e.g. from built-in WebCrypto. */
export type AsyncCipher = {
encrypt(plaintext: Uint8Array): Promise<Uint8Array>;
decrypt(ciphertext: Uint8Array): Promise<Uint8Array>;
};
/** Cipher with `output` argument which can optimize by doing 1 less allocation. */
export type CipherWithOutput = Cipher & {
encrypt(plaintext: Uint8Array, output?: Uint8Array): Uint8Array;
decrypt(ciphertext: Uint8Array, output?: Uint8Array): Uint8Array;
};
/**
* Params are outside of return type, so it is accessible before calling constructor.
* If function support multiple nonceLength's, we return the best one.
*/
export type CipherParams = {
blockSize: number;
nonceLength?: number;
tagLength?: number;
varSizeNonce?: boolean;
};
/** ARX cipher, like salsa or chacha. */
export type ARXCipher = ((
key: Uint8Array,
nonce: Uint8Array,
AAD?: Uint8Array
) => CipherWithOutput) & {
blockSize: number;
nonceLength: number;
tagLength: number;
};
export type CipherCons<T extends any[]> = (key: Uint8Array, ...args: T) => Cipher;
/**
* Wraps a cipher: validates args, ensures encrypt() can only be called once.
* @__NO_SIDE_EFFECTS__
*/
export const wrapCipher = <C extends CipherCons<any>, P extends CipherParams>(
params: P,
constructor: C
): C & P => {
function wrappedCipher(key: Uint8Array, ...args: any[]): CipherWithOutput {
// Validate key
abytes(key, undefined, 'key');
// Big-Endian hardware is rare. Just in case someone still decides to run ciphers:
if (!isLE) throw new Error('Non little-endian hardware is not yet supported');
// Validate nonce if nonceLength is present
if (params.nonceLength !== undefined) {
const nonce = args[0];
abytes(nonce, params.varSizeNonce ? undefined : params.nonceLength, 'nonce');
}
// Validate AAD if tagLength present
const tagl = params.tagLength;
if (tagl && args[1] !== undefined) abytes(args[1], undefined, 'AAD');
const cipher = constructor(key, ...args);
const checkOutput = (fnLength: number, output?: Uint8Array) => {
if (output !== undefined) {
if (fnLength !== 2) throw new Error('cipher output not supported');
abytes(output, undefined, 'output');
}
};
// Create wrapped cipher with validation and single-use encryption
let called = false;
const wrCipher = {
encrypt(data: Uint8Array, output?: Uint8Array) {
if (called) throw new Error('cannot encrypt() twice with same key + nonce');
called = true;
abytes(data);
checkOutput(cipher.encrypt.length, output);
return (cipher as CipherWithOutput).encrypt(data, output);
},
decrypt(data: Uint8Array, output?: Uint8Array) {
abytes(data);
if (tagl && data.length < tagl)
throw new Error('"ciphertext" expected length bigger than tagLength=' + tagl);
checkOutput(cipher.decrypt.length, output);
return (cipher as CipherWithOutput).decrypt(data, output);
},
};
return wrCipher;
}
Object.assign(wrappedCipher, params);
return wrappedCipher as C & P;
};
/** Represents salsa / chacha stream. */
export type XorStream = (
key: Uint8Array,
nonce: Uint8Array,
data: Uint8Array,
output?: Uint8Array,
counter?: number
) => Uint8Array;
/**
* By default, returns u8a of length.
* When out is available, it checks it for validity and uses it.
*/
export function getOutput(
expectedLength: number,
out?: Uint8Array,
onlyAligned = true
): Uint8Array {
if (out === undefined) return new Uint8Array(expectedLength);
if (out.length !== expectedLength)
throw new Error(
'"output" expected Uint8Array of length ' + expectedLength + ', got: ' + out.length
);
if (onlyAligned && !isAligned32(out)) throw new Error('invalid output, must be aligned');
return out;
}
export function u64Lengths(dataLength: number, aadLength: number, isLE: boolean): Uint8Array {
abool(isLE);
const num = new Uint8Array(16);
const view = createView(num);
view.setBigUint64(0, BigInt(aadLength), isLE);
view.setBigUint64(8, BigInt(dataLength), isLE);
return num;
}
// Is byte array aligned to 4 byte offset (u32)?
export function isAligned32(bytes: Uint8Array): boolean {
return bytes.byteOffset % 4 === 0;
}
// copy bytes to new u8a (aligned). Because Buffer.slice is broken.
export function copyBytes(bytes: Uint8Array): Uint8Array {
return Uint8Array.from(bytes);
}
/** Cryptographically secure PRNG. Uses internal OS-level `crypto.getRandomValues`. */
export function randomBytes(bytesLength = 32): Uint8Array {
const cr = typeof globalThis === 'object' ? (globalThis as any).crypto : null;
if (typeof cr?.getRandomValues !== 'function')
throw new Error('crypto.getRandomValues must be defined');
return cr.getRandomValues(new Uint8Array(bytesLength));
}
/**
* The pseudorandom number generator doesn't wipe current state:
* instead, it generates new one based on previous state + entropy.
* Not reseed/rekey, since AES CTR DRBG does rekey on each randomBytes,
* which is in fact `reseed`, since it changes counter too.
*/
export interface PRG {
addEntropy(seed: Uint8Array): void;
randomBytes(length: number): Uint8Array;
clean(): void;
}
type RemoveNonceInner<T extends any[], Ret> = ((...args: T) => Ret) extends (
arg0: any,
arg1: any,
...rest: infer R
) => any
? (key: Uint8Array, ...args: R) => Ret
: never;
export type RemoveNonce<T extends (...args: any) => any> = RemoveNonceInner<
Parameters<T>,
ReturnType<T>
>;
export type CipherWithNonce = ((
key: Uint8Array,
nonce: Uint8Array,
...args: any[]
) => Cipher | AsyncCipher) & {
nonceLength: number;
};
/**
* Uses CSPRG for nonce, nonce injected in ciphertext.
* For `encrypt`, a `nonceBytes`-length buffer is fetched from CSPRNG and
* prepended to encrypted ciphertext. For `decrypt`, first `nonceBytes` of ciphertext
* are treated as nonce.
*
* NOTE: Under the same key, using random nonces (e.g. `managedNonce`) with AES-GCM and ChaCha
* should be limited to `2**23` (8M) messages to get a collision chance of `2**-50`. Stretching to * `2**32` (4B) messages, chance would become `2**-33` - still negligible, but creeping up.
* @example
* const gcm = managedNonce(aes.gcm);
* const ciphr = gcm(key).encrypt(data);
* const plain = gcm(key).decrypt(ciph);
*/
export function managedNonce<T extends CipherWithNonce>(
fn: T,
randomBytes_: typeof randomBytes = randomBytes
): RemoveNonce<T> {
const { nonceLength } = fn;
anumber(nonceLength);
const addNonce = (nonce: Uint8Array, ciphertext: Uint8Array) => {
const out = concatBytes(nonce, ciphertext);
ciphertext.fill(0);
return out;
};
// NOTE: we cannot support DST here, it would be mistake:
// - we don't know how much dst length cipher requires
// - nonce may unalign dst and break everything
// - we create new u8a anyway (concatBytes)
// - previously we passed all args to cipher, but that was mistake!
return ((key: Uint8Array, ...args: any[]): any => ({
encrypt(plaintext: Uint8Array) {
abytes(plaintext);
const nonce = randomBytes_(nonceLength);
const encrypted = fn(key, nonce, ...args).encrypt(plaintext);
// @ts-ignore
if (encrypted instanceof Promise) return encrypted.then((ct) => addNonce(nonce, ct));
return addNonce(nonce, encrypted);
},
decrypt(ciphertext: Uint8Array) {
abytes(ciphertext);
const nonce = ciphertext.subarray(0, nonceLength);
const decrypted = ciphertext.subarray(nonceLength);
return fn(key, nonce, ...args).decrypt(decrypted);
},
})) as RemoveNonce<T>;
}