lilliput-ae-reference-implementation

Implementations of Lilliput-AE submitted to the NIST LWC standardization process
git clone https://git.kevinlegouguec.net/lilliput-ae-reference-implementation
Log | Files | Refs | README

commit 4582b5447826886a4ce7f5cd095daed2513d352d
parent 988fa098a026d57da624bf97b05167ae329b9f62
Author: Kévin Le Gouguec <kevin.legouguec@airbus.com>
Date:   Tue, 12 Mar 2019 10:42:10 +0100

Ajout de l'implémentation à seuil d'ordre 1

Modifications nécessaires dans l'infra :

- retrait conditionnel de test-tweakey, vu que l'API n'est pas la même
  pour l'implémentation à seuil,

- retrait  conditionnel   de  l'avertissement   "-Wparentheses",  plus
  agaçant qu'autre chose sur les calculs booléens de cipher.c, e.g.

      y_hi&3 ^ (y_hi&8)>>1

  où la  priorité est intuitive  (shifts avant AND avant  XOR).  C'est
  dommage de  perdre les avertissements sur  if (a&b == c),  mais tant
  pis… On va compter sur La Suite De Test®©™ pour nous couvrir.

Co-authored-by: Alexandre Adomnicai <a.adomnicai@trusted-objects.com>
Co-authored-by: leo <leo.reynaud17@gmail.com>

Diffstat:
Mnist/make-package.sh | 1+
Asrc/add_threshold/cipher.c | 307+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Asrc/add_threshold/cipher.h | 2++
Asrc/add_threshold/constants.h | 2++
Asrc/add_threshold/lilliput-ae-utils.h | 2++
Asrc/add_threshold/lilliput-ae.h | 2++
Asrc/add_threshold/lilliput-i.c | 2++
Asrc/add_threshold/lilliput-ii.c | 2++
Asrc/add_threshold/tweakey.c | 216+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Asrc/add_threshold/tweakey.h | 47+++++++++++++++++++++++++++++++++++++++++++++++
Mtest/check-implementation.sh | 2+-
Mtest/common.mk | 9++++++++-
12 files changed, 592 insertions(+), 2 deletions(-)

diff --git a/nist/make-package.sh b/nist/make-package.sh @@ -53,6 +53,7 @@ add-variant () implementations=( ref + add_threshold add_tweakeyloop ) diff --git a/src/add_threshold/cipher.c b/src/add_threshold/cipher.c @@ -0,0 +1,307 @@ +/* +Implementation of the Lilliput-AE tweakable block cipher. + +Authors: + Kévin Le Gouguec, + Léo Reynaud, + Alexandre Adomnicai, 2019. + +For more information, feedback or questions, refer to our website: +https://paclido.fr/lilliput-ae + +To the extent possible under law, the implementer has waived all copyright +and related or neighboring rights to the source code in this file. +http://creativecommons.org/publicdomain/zero/1.0/ + +--- + +This file provides an implementation of Lilliput-TBC's tweakey schedule, +where multiplications by matrices M and M_R to the power n are performed +by functions expressing the exponentiated matrices with shifts and XORs. +*/ + +#include <stdint.h> +#include <string.h> +#include <stdio.h> + +#include "cipher.h" +#include "constants.h" +#include "tweakey.h" + + +enum permutation +{ + PERMUTATION_ENCRYPTION = 0, /* PI(i) */ + PERMUTATION_DECRYPTION = 1, /* PI^-1(i) */ + PERMUTATION_NONE +}; + +typedef enum permutation permutation; + +static const uint8_t PERMUTATIONS[2][BLOCK_BYTES] = { + [PERMUTATION_ENCRYPTION] = { 13, 9, 14, 8, 10, 11, 12, 15, 4, 5, 3, 1, 2, 6, 0, 7 }, + [PERMUTATION_DECRYPTION] = { 14, 11, 12, 10, 8, 9, 13, 15, 3, 1, 4, 5, 6, 0, 2, 7 } +}; + +static const uint8_t F[16][16] = { + {0x0, 0x2, 0x0, 0x2, 0x2, 0x0, 0x2, 0x0, 0x0, 0x2, 0x0, 0x2, 0x2, 0x0, 0x2, 0x0}, + {0x0, 0x2, 0x9, 0xb, 0x3, 0x1, 0xa, 0x8, 0xd, 0xf, 0x4, 0x6, 0xe, 0xc, 0x7, 0x5}, + {0x0, 0xb, 0x0, 0xb, 0xb, 0x0, 0xb, 0x0, 0x1, 0xa, 0x1, 0xa, 0xa, 0x1, 0xa, 0x1}, + {0x9, 0x2, 0x0, 0xb, 0x3, 0x8, 0xa, 0x1, 0x5, 0xe, 0xc, 0x7, 0xf, 0x4, 0x6, 0xd}, + {0x1, 0x2, 0x8, 0xb, 0x3, 0x0, 0xa, 0x9, 0x9, 0xa, 0x0, 0x3, 0xb, 0x8, 0x2, 0x1}, + {0x0, 0x3, 0x0, 0x3, 0x3, 0x0, 0x3, 0x0, 0x5, 0x6, 0x5, 0x6, 0x6, 0x5, 0x6, 0x5}, + {0x8, 0x2, 0x1, 0xb, 0x3, 0x9, 0xa, 0x0, 0x1, 0xb, 0x8, 0x2, 0xa, 0x0, 0x3, 0x9}, + {0x0, 0xa, 0x0, 0xa, 0xa, 0x0, 0xa, 0x0, 0x4, 0xe, 0x4, 0xe, 0xe, 0x4, 0xe, 0x4}, + {0x1, 0xe, 0x0, 0xf, 0xb, 0x4, 0xa, 0x5, 0x1, 0xe, 0x0, 0xf, 0xb, 0x4, 0xa, 0x5}, + {0xc, 0x3, 0x4, 0xb, 0x7, 0x8, 0xf, 0x0, 0x1, 0xe, 0x9, 0x6, 0xa, 0x5, 0x2, 0xd}, + {0x0, 0x6, 0x1, 0x7, 0x3, 0x5, 0x2, 0x4, 0x1, 0x7, 0x0, 0x6, 0x2, 0x4, 0x3, 0x5}, + {0x4, 0x2, 0xc, 0xa, 0x6, 0x0, 0xe, 0x8, 0x8, 0xe, 0x0, 0x6, 0xa, 0xc, 0x2, 0x4}, + {0x8, 0x6, 0x0, 0xe, 0x2, 0xc, 0xa, 0x4, 0x0, 0xe, 0x8, 0x6, 0xa, 0x4, 0x2, 0xc}, + {0x4, 0xa, 0x5, 0xb, 0xf, 0x1, 0xe, 0x0, 0x1, 0xf, 0x0, 0xe, 0xa, 0x4, 0xb, 0x5}, + {0x0, 0x7, 0x8, 0xf, 0x3, 0x4, 0xb, 0xc, 0x9, 0xe, 0x1, 0x6, 0xa, 0xd, 0x2, 0x5}, + {0x5, 0x2, 0x4, 0x3, 0x7, 0x0, 0x6, 0x1, 0x1, 0x6, 0x0, 0x7, 0x3, 0x4, 0x2, 0x5} +}; + +static const uint8_t G[4][16] = { + {0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8, 0x9, 0xa, 0xb, 0xc, 0xd, 0xe, 0xf}, + {0x0, 0x1, 0x2, 0x3, 0x5, 0x4, 0x7, 0x6, 0x8, 0x9, 0xa, 0xb, 0xd, 0xc, 0xf, 0xe}, + {0x0, 0x1, 0x3, 0x2, 0x4, 0x5, 0x7, 0x6, 0x8, 0x9, 0xb, 0xa, 0xc, 0xd, 0xf, 0xe}, + {0x1, 0x0, 0x2, 0x3, 0x4, 0x5, 0x7, 0x6, 0x9, 0x8, 0xa, 0xb, 0xc, 0xd, 0xf, 0xe} +}; + +static const uint8_t Q[8][16] = { + {0x0, 0x4, 0x2, 0x6, 0x8, 0xc, 0xa, 0xe, 0x1, 0x5, 0x3, 0x7, 0x9, 0xd, 0xb, 0xf}, + {0x0, 0x4, 0xa, 0xe, 0x8, 0xc, 0x2, 0x6, 0x3, 0x7, 0x9, 0xd, 0xb, 0xf, 0x1, 0x5}, + {0x0, 0xc, 0x2, 0xe, 0x8, 0x4, 0xa, 0x6, 0x1, 0xd, 0x3, 0xf, 0x9, 0x5, 0xb, 0x7}, + {0x8, 0x4, 0x2, 0xe, 0x0, 0xc, 0xa, 0x6, 0xb, 0x7, 0x1, 0xd, 0x3, 0xf, 0x9, 0x5}, + {0x0, 0x6, 0x2, 0x4, 0x8, 0xe, 0xa, 0xc, 0x1, 0x7, 0x3, 0x5, 0x9, 0xf, 0xb, 0xd}, + {0x2, 0x4, 0x8, 0xe, 0xa, 0xc, 0x0, 0x6, 0x1, 0x7, 0xb, 0xd, 0x9, 0xf, 0x3, 0x5}, + {0x0, 0xe, 0x2, 0xc, 0x8, 0x6, 0xa, 0x4, 0x1, 0xf, 0x3, 0xd, 0x9, 0x7, 0xb, 0x5}, + {0xa, 0x4, 0x0, 0xe, 0x2, 0xc, 0x8, 0x6, 0x9, 0x7, 0x3, 0xd, 0x1, 0xf, 0xb, 0x5} +}; + +static const uint8_t P[16] = { + 0x0, 0x2, 0x8, 0xa, 0x4, 0X6, 0xc, 0xe, 0x1, 0x3, 0x9, 0xb, 0x5, 0x7, 0xd, 0xf +}; + +static void _state_init_TI(uint8_t X[BLOCK_BYTES], uint8_t Y[BLOCK_BYTES], uint8_t Z[BLOCK_BYTES], const uint8_t message[BLOCK_BYTES]) +{ + // To be replaced by real random numbers!!! + uint8_t SHARES_0[BLOCK_BYTES] = { + 0x0f, 0x1e, 0x2d, 0x3c, 0x4b, 0x5a, 0x69, 0x78, 0x87, 0x96, 0xa5, 0xb4, 0xc3, 0xd2, 0xe1, 0xf0 + }; + uint8_t SHARES_1[BLOCK_BYTES] = { + 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f + }; + + memcpy(X, SHARES_0, BLOCK_BYTES); + memcpy(Y, SHARES_1, BLOCK_BYTES); + for (uint8_t i=0; i<BLOCK_BYTES; i++) + { + Z[i] = message[i] ^ SHARES_0[i] ^ SHARES_1[i]; + } +} + + +static void _compute_round_tweakeys_TI( + const uint8_t key[KEY_BYTES], + const uint8_t tweak[TWEAK_BYTES], + uint8_t RTK_X[ROUNDS][ROUND_TWEAKEY_BYTES], + uint8_t RTK_Y[ROUNDS][ROUND_TWEAKEY_BYTES] +) +{ + uint8_t TK_X[TWEAKEY_BYTES]; + uint8_t TK_Y[TWEAKEY_BYTES]; + tweakey_state_init_TI(TK_X, TK_Y, key, tweak); + tweakey_state_extract_TI(TK_X, TK_Y, 0, RTK_X[0], RTK_Y[0]); + + for (uint8_t i=1; i<ROUNDS; i++) + { + tweakey_state_update_TI(TK_X, TK_Y); + tweakey_state_extract_TI(TK_X, TK_Y, i, RTK_X[i], RTK_Y[i]); + } +} + + +static void _nonlinear_layer_TI( + uint8_t X[BLOCK_BYTES], + uint8_t Y[BLOCK_BYTES], + uint8_t Z[BLOCK_BYTES], + const uint8_t RTK_X[ROUND_TWEAKEY_BYTES], + const uint8_t RTK_Y[ROUND_TWEAKEY_BYTES] +) +{ + uint8_t x_hi, y_hi, z_hi; // High nibbles for the Feistel network + uint8_t x_lo, y_lo, z_lo; // Low nibbles for the Feistel network + uint8_t tmp0, tmp1, tmp2; + uint8_t TMP_X[ROUND_TWEAKEY_BYTES]; + uint8_t TMP_Y[ROUND_TWEAKEY_BYTES]; + uint8_t TMP_Z[ROUND_TWEAKEY_BYTES]; + + // Apply the RTK to two shares + for (size_t j=0; j<ROUND_TWEAKEY_BYTES; j++) + { + TMP_X[j] = X[j] ^ RTK_X[j]; + TMP_Y[j] = Y[j] ^ RTK_Y[j]; + } + + // Threshold Implementation of the 8-bit S-box + for (size_t j=0; j<ROUND_TWEAKEY_BYTES; j++) + { + // Decomposition into nibbles + x_hi = TMP_X[j] >> 4; + x_lo = TMP_X[j] & 0xf; + y_hi = TMP_Y[j] >> 4; + y_lo = TMP_Y[j] & 0xf; + z_hi = Z[j] >> 4; + z_lo = Z[j] & 0xf; + // First 4-bit S-box + tmp0 = G[(y_lo&7)>>1][z_lo]; + tmp1 = G[(z_lo&7)>>1][x_lo]; + tmp2 = G[(x_lo&7)>>1][y_lo]; + x_hi ^= F[tmp1][tmp2]; + y_hi ^= F[tmp2][tmp0]; + z_hi ^= F[tmp0][tmp1]; + // Second 4-bit S-box + tmp0 = P[Q[y_hi&3 ^ (y_hi&8)>>1][z_hi]]; + tmp1 = P[Q[z_hi&3 ^ (z_hi&8)>>1][x_hi]]; + tmp2 = P[Q[x_hi&3 ^ (x_hi&8)>>1][y_hi]]; + x_lo ^= Q[tmp1&3 ^ (tmp1&8)>>1][tmp2]; + y_lo ^= Q[tmp2&3 ^ (tmp2&8)>>1][tmp0]; + z_lo ^= Q[tmp0&3 ^ (tmp0&8)>>1][tmp1]; + // Third 4-bit S-box + tmp0 = G[(y_lo&7)>>1][z_lo] ^ 1; + tmp1 = G[(z_lo&7)>>1][x_lo]; + tmp2 = G[(x_lo&7)>>1][y_lo]; + x_hi ^= F[tmp1][tmp2]; + y_hi ^= F[tmp2][tmp0]; + z_hi ^= F[tmp0][tmp1]; + // Build bytes from nibbles + TMP_X[j] = (x_hi << 4 | x_lo); + TMP_Y[j] = (y_hi << 4 | y_lo); + TMP_Z[j] = (z_hi << 4 | z_lo); + } + + for (size_t j=0; j<8; j++) + { + size_t dest_j = 15-j; + X[dest_j] ^= TMP_X[j]; + Y[dest_j] ^= TMP_Y[j]; + Z[dest_j] ^= TMP_Z[j]; + } +} + +static void _linear_layer(uint8_t X[BLOCK_BYTES]) +{ + X[15] ^= X[1]; + X[15] ^= X[2]; + X[15] ^= X[3]; + X[15] ^= X[4]; + X[15] ^= X[5]; + X[15] ^= X[6]; + X[15] ^= X[7]; + + X[14] ^= X[7]; + X[13] ^= X[7]; + X[12] ^= X[7]; + X[11] ^= X[7]; + X[10] ^= X[7]; + X[9] ^= X[7]; +} + +static void _permutation_layer(uint8_t X[BLOCK_BYTES], permutation p) +{ + if (p == PERMUTATION_NONE) + { + return; + } + + uint8_t X_old[BLOCK_BYTES]; + memcpy(X_old, X, BLOCK_BYTES); + + const uint8_t *pi = PERMUTATIONS[p]; + + for (size_t j=0; j<BLOCK_BYTES; j++) + { + X[pi[j]] = X_old[j]; + } +} + +static void _one_round_egfn_TI( + uint8_t X[BLOCK_BYTES], + uint8_t Y[BLOCK_BYTES], + uint8_t Z[BLOCK_BYTES], + const uint8_t RTK_X[ROUND_TWEAKEY_BYTES], + const uint8_t RTK_Y[ROUND_TWEAKEY_BYTES], + permutation p +) +{ + _nonlinear_layer_TI(X, Y, Z, RTK_X, RTK_Y); + _linear_layer(X); + _linear_layer(Y); + _linear_layer(Z); + _permutation_layer(X, p); + _permutation_layer(Y, p); + _permutation_layer(Z, p); +} + + +void lilliput_tbc_encrypt( + const uint8_t key[KEY_BYTES], + const uint8_t tweak[TWEAK_BYTES], + const uint8_t message[BLOCK_BYTES], + uint8_t ciphertext[BLOCK_BYTES] +) +{ + uint8_t X[BLOCK_BYTES]; + uint8_t Y[BLOCK_BYTES]; + uint8_t Z[BLOCK_BYTES]; + _state_init_TI(X, Y, Z, message); + + uint8_t RTK_X[ROUNDS][ROUND_TWEAKEY_BYTES]; + uint8_t RTK_Y[ROUNDS][ROUND_TWEAKEY_BYTES]; + _compute_round_tweakeys_TI(key, tweak, RTK_X, RTK_Y); + + + for (uint8_t i=0; i<ROUNDS-1; i++) + { + _one_round_egfn_TI(X, Y, Z, RTK_X[i], RTK_Y[i], PERMUTATION_ENCRYPTION); + } + + _one_round_egfn_TI(X, Y, Z, RTK_X[ROUNDS-1], RTK_Y[ROUNDS-1], PERMUTATION_NONE); + + + for (size_t i=0; i<BLOCK_BYTES; i++) + { + ciphertext[i] = X[i] ^ Y[i] ^ Z[i]; + } +} + +void lilliput_tbc_decrypt( + const uint8_t key[KEY_BYTES], + const uint8_t tweak[TWEAK_BYTES], + const uint8_t ciphertext[BLOCK_BYTES], + uint8_t message[BLOCK_BYTES] +) +{ + uint8_t X[BLOCK_BYTES]; + uint8_t Y[BLOCK_BYTES]; + uint8_t Z[BLOCK_BYTES]; + _state_init_TI(X, Y, Z, ciphertext); + + uint8_t RTK_X[ROUNDS][ROUND_TWEAKEY_BYTES]; + uint8_t RTK_Y[ROUNDS][ROUND_TWEAKEY_BYTES]; + _compute_round_tweakeys_TI(key, tweak, RTK_X, RTK_Y); + + for (uint8_t i=0; i<ROUNDS-1; i++) + { + _one_round_egfn_TI(X, Y, Z, RTK_X[ROUNDS-1-i], RTK_Y[ROUNDS-1-i], PERMUTATION_DECRYPTION); + } + + _one_round_egfn_TI(X, Y, Z, RTK_X[0], RTK_Y[0], PERMUTATION_NONE); + + for (size_t i=0; i<BLOCK_BYTES; i++) + { + message[i] = X[i] ^ Y[i] ^ Z[i]; + } +} diff --git a/src/add_threshold/cipher.h b/src/add_threshold/cipher.h @@ -0,0 +1 @@ +../ref/cipher.h +\ No newline at end of file diff --git a/src/add_threshold/constants.h b/src/add_threshold/constants.h @@ -0,0 +1 @@ +../ref/constants.h +\ No newline at end of file diff --git a/src/add_threshold/lilliput-ae-utils.h b/src/add_threshold/lilliput-ae-utils.h @@ -0,0 +1 @@ +../ref/lilliput-ae-utils.h +\ No newline at end of file diff --git a/src/add_threshold/lilliput-ae.h b/src/add_threshold/lilliput-ae.h @@ -0,0 +1 @@ +../ref/lilliput-ae.h +\ No newline at end of file diff --git a/src/add_threshold/lilliput-i.c b/src/add_threshold/lilliput-i.c @@ -0,0 +1 @@ +../ref/lilliput-i.c +\ No newline at end of file diff --git a/src/add_threshold/lilliput-ii.c b/src/add_threshold/lilliput-ii.c @@ -0,0 +1 @@ +../ref/lilliput-ii.c +\ No newline at end of file diff --git a/src/add_threshold/tweakey.c b/src/add_threshold/tweakey.c @@ -0,0 +1,216 @@ +/* +Implementation of the Lilliput-AE tweakable block cipher. + +Authors: + Kévin Le Gouguec, + Léo Reynaud, + Alexandre Adomnicai, 2019. + +For more information, feedback or questions, refer to our website: +https://paclido.fr/lilliput-ae + +To the extent possible under law, the implementer has waived all copyright +and related or neighboring rights to the source code in this file. +http://creativecommons.org/publicdomain/zero/1.0/ + +--- + +This file provides an implementation of Lilliput-TBC's tweakey schedule, +where multiplications by matrices M and M_R to the power n are performed +by functions expressing the exponentiated matrices with shifts and XORs. +*/ + +#include <stdint.h> +#include <string.h> + +#include "constants.h" +#include "tweakey.h" + + +#define LANE_BITS 64 +#define LANE_BYTES (LANE_BITS/8) +#define LANES_NB (TWEAKEY_BYTES/LANE_BYTES) + + +void tweakey_state_init_TI( + uint8_t TK_X[TWEAKEY_BYTES], + uint8_t TK_Y[KEY_BYTES], + const uint8_t key[KEY_BYTES], + const uint8_t tweak[TWEAK_BYTES] +) +{ + // To be replaced by real random numbers!!! + uint8_t SHARES_0[KEY_BYTES] = { + 0x0f, 0x1e, 0x2d, 0x3c, 0x4b, 0x5a, 0x69, 0x78, 0x87, 0x96, 0xa5, 0xb4, 0xc3, 0xd2, 0xe1, 0xf0 + }; + + memcpy(TK_Y, SHARES_0, KEY_BYTES); + memcpy(TK_X, tweak, TWEAK_BYTES); + + for (size_t i=0; i<KEY_BYTES; i++){ + TK_X[i+TWEAK_BYTES] = key[i] ^ SHARES_0[i] ; + } +} + + +void tweakey_state_extract_TI( + const uint8_t TK_X[TWEAKEY_BYTES], + const uint8_t TK_Y[KEY_BYTES], + uint8_t round_constant, + uint8_t round_tweakey_X[ROUND_TWEAKEY_BYTES], + uint8_t round_tweakey_Y[ROUND_TWEAKEY_BYTES] +) +{ + memset(round_tweakey_X, 0, ROUND_TWEAKEY_BYTES); + memset(round_tweakey_Y, 0, ROUND_TWEAKEY_BYTES); + + for (size_t j=0; j<LANES_NB; j++) + { + const uint8_t *TKj_X = TK_X + j*LANE_BYTES; + + for (size_t k=0; k<LANE_BYTES; k++) + { + round_tweakey_X[k] ^= TKj_X[k]; + } + } + + + for (size_t j=0; j<(KEY_BYTES / LANE_BYTES); j++) + { + const uint8_t *TKj_Y = TK_Y + j*LANE_BYTES; + + for (size_t k=0; k<LANE_BYTES; k++) + { + round_tweakey_Y[k] ^= TKj_Y[k]; + } + } + + round_tweakey_X[0] ^= round_constant; +} + + +static void _multiply_M(const uint8_t x[LANE_BYTES], uint8_t y[LANE_BYTES]) +{ + y[7] = x[6]; + y[6] = x[5]; + y[5] = x[5]<<3 ^ x[4]; + y[4] = x[4]>>3 ^ x[3]; + y[3] = x[2]; + y[2] = x[6]<<2 ^ x[1]; + y[1] = x[0]; + y[0] = x[7]; +} + +static void _multiply_M2(const uint8_t x[LANE_BYTES], uint8_t y[LANE_BYTES]) +{ + uint8_t x_M_5 = x[5]<<3 ^ x[4]; + uint8_t x_M_4 = x[4]>>3 ^ x[3]; + + y[7] = x[5]; + y[6] = x_M_5; + y[5] = x_M_5<<3 ^ x_M_4; + y[4] = x_M_4>>3 ^ x[2]; + y[3] = x[6]<<2 ^ x[1]; + y[2] = x[5]<<2 ^ x[0]; + y[1] = x[7]; + y[0] = x[6]; +} + +static void _multiply_M3(const uint8_t x[LANE_BYTES], uint8_t y[LANE_BYTES]) +{ + uint8_t x_M_5 = x[5]<<3 ^ x[4]; + uint8_t x_M_4 = x[4]>>3 ^ x[3]; + uint8_t x_M2_5 = x_M_5<<3 ^ x_M_4; + uint8_t x_M2_4 = x_M_4>>3 ^ x[2]; + + y[7] = x_M_5; + y[6] = x_M2_5; + y[5] = x_M2_5<<3 ^ x_M2_4; + y[4] = x_M2_4>>3 ^ x[6]<<2 ^ x[1]; + y[3] = x[5]<<2 ^ x[0]; + y[2] = x_M_5<<2 ^ x[7]; + y[1] = x[6]; + y[0] = x[5]; +} + +static void _multiply_MR(const uint8_t x[LANE_BYTES], uint8_t y[LANE_BYTES]) +{ + y[0] = x[1]; + y[1] = x[2]; + y[2] = x[3] ^ x[4]>>3; + y[3] = x[4]; + y[4] = x[5] ^ x[6]<<3; + y[5] = x[3]<<2 ^ x[6]; + y[6] = x[7]; + y[7] = x[0]; +} + +static void _multiply_MR2(const uint8_t x[LANE_BYTES], uint8_t y[LANE_BYTES]) +{ + uint8_t x_MR_4 = x[5] ^ x[6]<<3; + + y[0] = x[2]; + y[1] = x[3] ^ x[4]>>3; + y[2] = x[4] ^ x_MR_4>>3; + y[3] = x_MR_4; + y[4] = x[3]<<2 ^ x[6] ^ x[7]<<3; + y[5] = x[4]<<2 ^ x[7]; + y[6] = x[0]; + y[7] = x[1]; +} + +static void _multiply_MR3(const uint8_t x[LANE_BYTES], uint8_t y[LANE_BYTES]) +{ + uint8_t x_MR_4 = x[5] ^ x[6]<<3; + uint8_t x_MR2_4 = x[3]<<2 ^ x[6] ^ x[7]<<3; + + y[0] = x[3] ^ x[4]>>3; + y[1] = x[4] ^ x_MR_4>>3; + y[2] = x_MR_4 ^ x_MR2_4>>3; + y[3] = x_MR2_4; + y[4] = x[0]<<3 ^ x[4]<<2 ^ x[7]; + y[5] = x_MR_4<<2 ^ x[0]; + y[6] = x[1]; + y[7] = x[2]; +} + +typedef void (*matrix_multiplication)(const uint8_t x[LANE_BYTES], uint8_t y[LANE_BYTES]); + +static const matrix_multiplication ALPHAS[6] = { + _multiply_M, + _multiply_M2, + _multiply_M3, + _multiply_MR, + _multiply_MR2, + _multiply_MR3 +}; + + +void tweakey_state_update_TI(uint8_t TK_X[TWEAKEY_BYTES], uint8_t TK_Y[KEY_BYTES]) +{ + /* Skip lane 0, as it is multiplied by the identity matrix. */ + + for (size_t j=1; j<(TWEAK_BYTES/LANE_BYTES); j++) + { + uint8_t *TKj_X = TK_X + j*LANE_BYTES; + + uint8_t TKj_old_X[LANE_BYTES]; + memcpy(TKj_old_X, TKj_X, LANE_BYTES); + + ALPHAS[j-1](TKj_old_X, TKj_X); + } + + for (size_t j=0; j<(KEY_BYTES/LANE_BYTES); j++) + { + uint8_t *TKj_X = TK_X + (j + (TWEAK_BYTES/LANE_BYTES))*LANE_BYTES; + uint8_t *TKj_Y = TK_Y + j*LANE_BYTES; + + uint8_t TKj_X_old[LANE_BYTES]; + uint8_t TKj_Y_old[LANE_BYTES]; + memcpy(TKj_X_old, TKj_X, LANE_BYTES); + memcpy(TKj_Y_old, TKj_Y, LANE_BYTES); + + ALPHAS[j-1 + (TWEAK_BYTES/LANE_BYTES)](TKj_X_old, TKj_X); + ALPHAS[j-1 + (TWEAK_BYTES/LANE_BYTES)](TKj_Y_old, TKj_Y); + } +} diff --git a/src/add_threshold/tweakey.h b/src/add_threshold/tweakey.h @@ -0,0 +1,47 @@ +/* +Implementation of the Lilliput-AE tweakable block cipher. + +Authors: + Kévin Le Gouguec, + Léo Reynaud, + Alexandre Adomnicai, 2019. + +For more information, feedback or questions, refer to our website: +https://paclido.fr/lilliput-ae + +To the extent possible under law, the implementer has waived all copyright +and related or neighboring rights to the source code in this file. +http://creativecommons.org/publicdomain/zero/1.0/ + +--- + +This file provides the interface for Lilliput-TBC's tweakey schedule. +*/ + +#ifndef TWEAKEY_H +#define TWEAKEY_H + +#include <stdint.h> + +#include "constants.h" + + +void tweakey_state_init_TI( + uint8_t TK_X[TWEAKEY_BYTES], + uint8_t TK_Y[TWEAKEY_BYTES], + const uint8_t key[KEY_BYTES], + const uint8_t tweak[TWEAK_BYTES] +); + +void tweakey_state_extract_TI( + const uint8_t TK_X[TWEAKEY_BYTES], + const uint8_t TK_Y[KEY_BYTES], + uint8_t round_constant, + uint8_t round_tweakey_X[ROUND_TWEAKEY_BYTES], + uint8_t round_tweakey_Y[ROUND_TWEAKEY_BYTES] +); + +void tweakey_state_update_TI(uint8_t TK_X[TWEAKEY_BYTES], uint8_t TK_Y[KEY_BYTES]); + + +#endif /* TWEAKEY_H */ diff --git a/test/check-implementation.sh b/test/check-implementation.sh @@ -46,7 +46,7 @@ run-genkat () local nist_flags=(-std=c99 -Wall -Wextra -Wshadow -fsanitize=address,undefined -O2) - gcc ${nist_flags[@]} -Werror -I${genkat_dir} ${genkat_dir}/*.c -o ${genkat} + gcc ${nist_flags[@]} -I${genkat_dir} ${genkat_dir}/*.c -o ${genkat} ${genkat} mv LWC_AEAD_KAT*.txt ${vectors_dir}/${mode}-${keylen} diff --git a/test/common.mk b/test/common.mk @@ -11,9 +11,16 @@ endif IMPLEMENTATION = ref +ifeq "$(IMPLEMENTATION)" "add_threshold" +# Filter out tests on tweakey schedule, as the thresholded API differs. +tests = $(filter-out test-tweakey,$(basename $(wildcard test-*.c))) +# Don't trigger warnings for "a&b ^ c". +CFLAGS += -Wno-parentheses +else tests = $(basename $(wildcard test-*.c)) -traces = $(basename $(wildcard traces-*.c)) +endif +traces = $(basename $(wildcard traces-*.c)) test_dir = $(dir $(lastword $(MAKEFILE_LIST))) root_dir = $(test_dir)..