libdonut-rs 0.1.0

#![allow(non_upper_case_globals)]

use alloc::vec;
use crate::errors::{DonutError, DonutResult};
use alloc::vec::Vec;

const BLOCK_SIZE: usize = 16;
const KEY_SIZE: usize = 32;
const Nb: usize = 4;
const Nk: usize = 8;
const Nr: usize = 14;

type Block = [u8; BLOCK_SIZE];

const SBOX: [u8; 256] = [
    0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
    0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
    0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
    0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
    0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
    0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
    0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
    0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
    0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
    0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
    0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
    0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
    0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
    0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
    0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
    0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16,
];
const INV_SBOX: [u8; 256] = [
    0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38, 0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb,
    0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87, 0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb,
    0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d, 0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e,
    0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2, 0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25,
    0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16, 0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92,
    0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda, 0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84,
    0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a, 0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06,
    0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02, 0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b,
    0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea, 0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73,
    0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85, 0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e,
    0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89, 0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b,
    0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20, 0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4,
    0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31, 0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f,
    0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d, 0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef,
    0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0, 0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61,
    0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26, 0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d,
];

const RCON: [u32; 10] = [
    0x01000000, 0x02000000, 0x04000000, 0x08000000, 0x10000000, 0x20000000, 0x40000000, 0x80000000,
    0x1b000000, 0x36000000,
];

#[inline(always)]
fn load_state_from(block: &Block) -> [u8; 16] {
    *block
}

#[inline(always)]
fn store_state_to(state: [u8; 16], block: &mut Block) {
    *block = state;
}

#[inline(always)]
fn sub_word(word: u32) -> u32 {
    let b = word.to_be_bytes();
    u32::from_be_bytes([SBOX[b[0] as usize], SBOX[b[1] as usize], SBOX[b[2] as usize], SBOX[b[3] as usize]])
}

#[inline(always)]
fn rot_word(word: u32) -> u32 { word.rotate_left(8) }

#[inline(always)]
fn expand_key(key: &[u8; KEY_SIZE], w: &mut [u32; Nb * (Nr + 1)]) {
    for i in 0..Nk {
        w[i] = u32::from_be_bytes([key[4*i], key[4*i+1], key[4*i+2], key[4*i+3]]);
    }
    let mut i = Nk;
    while i < Nb * (Nr + 1) {
        let mut temp = w[i - 1];
        if i.is_multiple_of(Nk){
            temp = sub_word(rot_word(temp)) ^ RCON[(i / Nk) - 1];
        } else if i % Nk == 4 {
            temp = sub_word(temp);
        }
        w[i] = w[i - Nk] ^ temp;
        i += 1;
    }
}
#[inline(always)]
fn inv_sub_bytes(s: &mut [u8; 16]) {
    for b in s.iter_mut() {
        *b = INV_SBOX[*b as usize];
    }
}

#[unsafe(link_section = ".text")]
fn inv_shift_rows(s: &mut [u8; 16]) {
    let t = s[13];
    s[13] = s[9];
    s[9] = s[5];
    s[5] = s[1];
    s[1] = t;
    let (a, b) = (s[2], s[6]);
    s[2] = s[10];
    s[6] = s[14];
    s[10] = a;
    s[14] = b;
    let t = s[3];
    s[3] = s[7];
    s[7] = s[11];
    s[11] = s[15];
    s[15] = t;
}

#[inline(always)]
fn gf2_mul2(x: u8) -> u8 {
    if x & 0x80 != 0 {
        (x << 1) ^ 0x1b
    } else {
        x << 1
    }
}

#[inline(always)]
fn gf2_mul9(x: u8) -> u8 {
    gf2_mul2(gf2_mul2(gf2_mul2(x))) ^ x
}
#[inline(always)]
fn gf2_mul11(x: u8) -> u8 {
    gf2_mul9(x) ^ gf2_mul2(x)
}
#[inline(always)]
fn gf2_mul13(x: u8) -> u8 {
    gf2_mul9(x) ^ gf2_mul2(gf2_mul2(x))
}
#[inline(always)]
fn gf2_mul14(x: u8) -> u8 {
    gf2_mul8_4_2(x)
}

#[inline(always)]
fn gf2_mul8_4_2(x: u8) -> u8 {
    let x2 = gf2_mul2(x);
    let x4 = gf2_mul2(x2);
    let x8 = gf2_mul2(x4);
    x8 ^ x4 ^ x2
}

#[inline(always)]
fn inv_mix_columns(s: &mut [u8; 16]) {
    for c in 0..4 {
        let i = 4 * c;
        let a0 = s[i];
        let a1 = s[i + 1];
        let a2 = s[i + 2];
        let a3 = s[i + 3];
        s[i] = gf2_mul14(a0) ^ gf2_mul11(a1) ^ gf2_mul13(a2) ^ gf2_mul9(a3);
        s[i + 1] = gf2_mul9(a0) ^ gf2_mul14(a1) ^ gf2_mul11(a2) ^ gf2_mul13(a3);
        s[i + 2] = gf2_mul13(a0) ^ gf2_mul9(a1) ^ gf2_mul14(a2) ^ gf2_mul11(a3);
        s[i + 3] = gf2_mul11(a0) ^ gf2_mul13(a1) ^ gf2_mul9(a2) ^ gf2_mul14(a3);
    }
}

#[inline(always)]
fn add_round_key(s: &mut [u8; 16], w: &[u32], round: usize) {
    let base = round * 4;
    let k0 = w[base].to_be_bytes();
    let k1 = w[base + 1].to_be_bytes();
    let k2 = w[base + 2].to_be_bytes();
    let k3 = w[base + 3].to_be_bytes();
    s[0] ^= k0[0];
    s[1] ^= k0[1];
    s[2] ^= k0[2];
    s[3] ^= k0[3];
    s[4] ^= k1[0];
    s[5] ^= k1[1];
    s[6] ^= k1[2];
    s[7] ^= k1[3];
    s[8] ^= k2[0];
    s[9] ^= k2[1];
    s[10] ^= k2[2];
    s[11] ^= k2[3];
    s[12] ^= k3[0];
    s[13] ^= k3[1];
    s[14] ^= k3[2];
    s[15] ^= k3[3];
}
#[inline(always)]
fn aes256_decrypt_block(block: &mut Block, w: &[u32; Nb * (Nr + 1)]) {
    let mut s = load_state_from(block);

    add_round_key(&mut s, w, Nr);
    for round in (1..Nr).rev() {
        inv_shift_rows(&mut s);
        inv_sub_bytes(&mut s);
        add_round_key(&mut s, w, round);
        inv_mix_columns(&mut s);
    }
    inv_shift_rows(&mut s);
    inv_sub_bytes(&mut s);
    add_round_key(&mut s, w, 0);

    store_state_to(s, block);
}


fn aes256_cbc_decrypt(
    ciphertext: &[u8],
    plaintext: &mut [u8],
    key: &[u8; KEY_SIZE],
    iv: &Block,
) -> Option<usize> {
    if !ciphertext.len().is_multiple_of(BLOCK_SIZE) { return None; }

    let mut w = [0u32; Nb * (Nr + 1)];
    expand_key(key, &mut w);

    let mut prev = *iv;

    for (i, chunk) in ciphertext.chunks_exact(BLOCK_SIZE).enumerate() {
        let mut block: Block = chunk.try_into().unwrap();
        aes256_decrypt_block(&mut block, &w);

        xor_block_inplace(&mut block, &prev);
        let offset = i * BLOCK_SIZE;
        plaintext[offset..offset + BLOCK_SIZE].copy_from_slice(&block);

        prev.copy_from_slice(chunk);
    }

    w.fill(0);

    remove_pkcs7_padding(&plaintext[..ciphertext.len()])
}

/// Decrypts an AES-CBC-256 encrypted blob of bytes
pub fn decrypt_cbc(ciphertext: &[u8], key: &[u8], iv: &[u8]) -> DonutResult<Vec<u8>> {
    assert_eq!(key.len(), KEY_SIZE);
    assert_eq!(iv.len(), BLOCK_SIZE);

    let mut pt = vec![0u8; ciphertext.len()];
    let len = aes256_cbc_decrypt(
        ciphertext,
        &mut pt,
        key.try_into().unwrap(),
        iv.try_into().unwrap(),
    )
        .ok_or(DonutError::CryptoError)?;
    pt.truncate(len);
    Ok(pt)
}


#[inline(always)]
fn remove_pkcs7_padding(buf: &[u8]) -> Option<usize> {
    if buf.is_empty() {
        return None;
    }
    let pad_len = *buf.last()? as usize;
    if pad_len == 0 || pad_len > BLOCK_SIZE || pad_len > buf.len() {
        return None;
    }
    let mut diff = 0u8;
    for &b in &buf[buf.len() - pad_len..] {
        diff |= b ^ (pad_len as u8);
    }
    if diff == 0 {
        Some(buf.len() - pad_len)
    } else {
        None
    }
}

#[inline(always)]
fn xor_block_inplace(block: &mut Block, prev: &Block) {
    let (pfx, block_words, sfx) = unsafe { block.align_to_mut::<usize>() };
    let (pfx2, prev_words, sfx2) = unsafe { prev.align_to::<usize>() };
    if pfx.is_empty() && sfx.is_empty() && pfx2.is_empty() && sfx2.is_empty() && block_words.len()==prev_words.len() {
        for (a,b) in block_words.iter_mut().zip(prev_words.iter()) { *a ^= *b; }
    } else {
        for i in 0..BLOCK_SIZE { block[i] ^= prev[i]; }
    }
}