From 070db10cc14e5f5f83ae7df708bd49e36d7433ee Mon Sep 17 00:00:00 2001 From: Peter Schwabe Date: Tue, 9 Apr 2019 14:35:44 +0200 Subject: [PATCH 01/14] Added (constant-time) AES implementation, based on BearSSL --- common/aes.c | 656 +++++++++++++++++++++++++++++++++++++++++++++++++++ common/aes.h | 46 ++++ 2 files changed, 702 insertions(+) create mode 100644 common/aes.c create mode 100644 common/aes.h diff --git a/common/aes.c b/common/aes.c new file mode 100644 index 00000000..746514fe --- /dev/null +++ b/common/aes.c @@ -0,0 +1,656 @@ +/* + * AES implementation based on code from BearSSL (https://bearssl.org/) + * by Thomas Pornin. + * + * + * Copyright (c) 2016 Thomas Pornin + * + * Permission is hereby granted, free of charge, to any person obtaining + * a copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sublicense, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice shall be + * included in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include +#include + +#include "ctaes.h" + +static inline uint32_t br_dec32le(const unsigned char *src) +{ + return (uint32_t)src[0] + | ((uint32_t)src[1] << 8) + | ((uint32_t)src[2] << 16) + | ((uint32_t)src[3] << 24); +} + + +static void br_range_dec32le(uint32_t *v, size_t num, const unsigned char *src) +{ + while (num-- > 0) { + *v ++ = br_dec32le(src); + src += 4; + } +} + + +static inline uint32_t br_swap32(uint32_t x) +{ + x = ((x & (uint32_t)0x00FF00FF) << 8) + | ((x >> 8) & (uint32_t)0x00FF00FF); + return (x << 16) | (x >> 16); +} + + +static inline void br_enc32le(unsigned char *dst, uint32_t x) +{ + dst[0] = (unsigned char)x; + dst[1] = (unsigned char)(x >> 8); + dst[2] = (unsigned char)(x >> 16); + dst[3] = (unsigned char)(x >> 24); +} + + +void br_range_enc32le(unsigned char *dst, const uint32_t *v, size_t num) +{ + while (num-- > 0) { + br_enc32le(dst, *v ++); + dst += 4; + } +} + + +static void br_aes_ct64_bitslice_Sbox(uint64_t *q) +{ + /* + * This S-box implementation is a straightforward translation of + * the circuit described by Boyar and Peralta in "A new + * combinational logic minimization technique with applications + * to cryptology" (https://eprint.iacr.org/2009/191.pdf). + * + * Note that variables x* (input) and s* (output) are numbered + * in "reverse" order (x0 is the high bit, x7 is the low bit). + */ + + uint64_t x0, x1, x2, x3, x4, x5, x6, x7; + uint64_t y1, y2, y3, y4, y5, y6, y7, y8, y9; + uint64_t y10, y11, y12, y13, y14, y15, y16, y17, y18, y19; + uint64_t y20, y21; + uint64_t z0, z1, z2, z3, z4, z5, z6, z7, z8, z9; + uint64_t z10, z11, z12, z13, z14, z15, z16, z17; + uint64_t t0, t1, t2, t3, t4, t5, t6, t7, t8, t9; + uint64_t t10, t11, t12, t13, t14, t15, t16, t17, t18, t19; + uint64_t t20, t21, t22, t23, t24, t25, t26, t27, t28, t29; + uint64_t t30, t31, t32, t33, t34, t35, t36, t37, t38, t39; + uint64_t t40, t41, t42, t43, t44, t45, t46, t47, t48, t49; + uint64_t t50, t51, t52, t53, t54, t55, t56, t57, t58, t59; + uint64_t t60, t61, t62, t63, t64, t65, t66, t67; + uint64_t s0, s1, s2, s3, s4, s5, s6, s7; + + x0 = q[7]; + x1 = q[6]; + x2 = q[5]; + x3 = q[4]; + x4 = q[3]; + x5 = q[2]; + x6 = q[1]; + x7 = q[0]; + + /* + * Top linear transformation. + */ + y14 = x3 ^ x5; + y13 = x0 ^ x6; + y9 = x0 ^ x3; + y8 = x0 ^ x5; + t0 = x1 ^ x2; + y1 = t0 ^ x7; + y4 = y1 ^ x3; + y12 = y13 ^ y14; + y2 = y1 ^ x0; + y5 = y1 ^ x6; + y3 = y5 ^ y8; + t1 = x4 ^ y12; + y15 = t1 ^ x5; + y20 = t1 ^ x1; + y6 = y15 ^ x7; + y10 = y15 ^ t0; + y11 = y20 ^ y9; + y7 = x7 ^ y11; + y17 = y10 ^ y11; + y19 = y10 ^ y8; + y16 = t0 ^ y11; + y21 = y13 ^ y16; + y18 = x0 ^ y16; + + /* + * Non-linear section. + */ + t2 = y12 & y15; + t3 = y3 & y6; + t4 = t3 ^ t2; + t5 = y4 & x7; + t6 = t5 ^ t2; + t7 = y13 & y16; + t8 = y5 & y1; + t9 = t8 ^ t7; + t10 = y2 & y7; + t11 = t10 ^ t7; + t12 = y9 & y11; + t13 = y14 & y17; + t14 = t13 ^ t12; + t15 = y8 & y10; + t16 = t15 ^ t12; + t17 = t4 ^ t14; + t18 = t6 ^ t16; + t19 = t9 ^ t14; + t20 = t11 ^ t16; + t21 = t17 ^ y20; + t22 = t18 ^ y19; + t23 = t19 ^ y21; + t24 = t20 ^ y18; + + t25 = t21 ^ t22; + t26 = t21 & t23; + t27 = t24 ^ t26; + t28 = t25 & t27; + t29 = t28 ^ t22; + t30 = t23 ^ t24; + t31 = t22 ^ t26; + t32 = t31 & t30; + t33 = t32 ^ t24; + t34 = t23 ^ t33; + t35 = t27 ^ t33; + t36 = t24 & t35; + t37 = t36 ^ t34; + t38 = t27 ^ t36; + t39 = t29 & t38; + t40 = t25 ^ t39; + + t41 = t40 ^ t37; + t42 = t29 ^ t33; + t43 = t29 ^ t40; + t44 = t33 ^ t37; + t45 = t42 ^ t41; + z0 = t44 & y15; + z1 = t37 & y6; + z2 = t33 & x7; + z3 = t43 & y16; + z4 = t40 & y1; + z5 = t29 & y7; + z6 = t42 & y11; + z7 = t45 & y17; + z8 = t41 & y10; + z9 = t44 & y12; + z10 = t37 & y3; + z11 = t33 & y4; + z12 = t43 & y13; + z13 = t40 & y5; + z14 = t29 & y2; + z15 = t42 & y9; + z16 = t45 & y14; + z17 = t41 & y8; + + /* + * Bottom linear transformation. + */ + t46 = z15 ^ z16; + t47 = z10 ^ z11; + t48 = z5 ^ z13; + t49 = z9 ^ z10; + t50 = z2 ^ z12; + t51 = z2 ^ z5; + t52 = z7 ^ z8; + t53 = z0 ^ z3; + t54 = z6 ^ z7; + t55 = z16 ^ z17; + t56 = z12 ^ t48; + t57 = t50 ^ t53; + t58 = z4 ^ t46; + t59 = z3 ^ t54; + t60 = t46 ^ t57; + t61 = z14 ^ t57; + t62 = t52 ^ t58; + t63 = t49 ^ t58; + t64 = z4 ^ t59; + t65 = t61 ^ t62; + t66 = z1 ^ t63; + s0 = t59 ^ t63; + s6 = t56 ^ ~t62; + s7 = t48 ^ ~t60; + t67 = t64 ^ t65; + s3 = t53 ^ t66; + s4 = t51 ^ t66; + s5 = t47 ^ t65; + s1 = t64 ^ ~s3; + s2 = t55 ^ ~t67; + + q[7] = s0; + q[6] = s1; + q[5] = s2; + q[4] = s3; + q[3] = s4; + q[2] = s5; + q[1] = s6; + q[0] = s7; +} + +static void br_aes_ct64_ortho(uint64_t *q) +{ +#define SWAPN(cl, ch, s, x, y) do { \ + uint64_t a, b; \ + a = (x); \ + b = (y); \ + (x) = (a & (uint64_t)cl) | ((b & (uint64_t)cl) << (s)); \ + (y) = ((a & (uint64_t)ch) >> (s)) | (b & (uint64_t)ch); \ + } while (0) + +#define SWAP2(x, y) SWAPN(0x5555555555555555, 0xAAAAAAAAAAAAAAAA, 1, x, y) +#define SWAP4(x, y) SWAPN(0x3333333333333333, 0xCCCCCCCCCCCCCCCC, 2, x, y) +#define SWAP8(x, y) SWAPN(0x0F0F0F0F0F0F0F0F, 0xF0F0F0F0F0F0F0F0, 4, x, y) + + SWAP2(q[0], q[1]); + SWAP2(q[2], q[3]); + SWAP2(q[4], q[5]); + SWAP2(q[6], q[7]); + + SWAP4(q[0], q[2]); + SWAP4(q[1], q[3]); + SWAP4(q[4], q[6]); + SWAP4(q[5], q[7]); + + SWAP8(q[0], q[4]); + SWAP8(q[1], q[5]); + SWAP8(q[2], q[6]); + SWAP8(q[3], q[7]); +} + + +static void br_aes_ct64_interleave_in(uint64_t *q0, uint64_t *q1, const uint32_t *w) +{ + uint64_t x0, x1, x2, x3; + + x0 = w[0]; + x1 = w[1]; + x2 = w[2]; + x3 = w[3]; + x0 |= (x0 << 16); + x1 |= (x1 << 16); + x2 |= (x2 << 16); + x3 |= (x3 << 16); + x0 &= (uint64_t)0x0000FFFF0000FFFF; + x1 &= (uint64_t)0x0000FFFF0000FFFF; + x2 &= (uint64_t)0x0000FFFF0000FFFF; + x3 &= (uint64_t)0x0000FFFF0000FFFF; + x0 |= (x0 << 8); + x1 |= (x1 << 8); + x2 |= (x2 << 8); + x3 |= (x3 << 8); + x0 &= (uint64_t)0x00FF00FF00FF00FF; + x1 &= (uint64_t)0x00FF00FF00FF00FF; + x2 &= (uint64_t)0x00FF00FF00FF00FF; + x3 &= (uint64_t)0x00FF00FF00FF00FF; + *q0 = x0 | (x2 << 8); + *q1 = x1 | (x3 << 8); +} + + +static void br_aes_ct64_interleave_out(uint32_t *w, uint64_t q0, uint64_t q1) +{ + uint64_t x0, x1, x2, x3; + + x0 = q0 & (uint64_t)0x00FF00FF00FF00FF; + x1 = q1 & (uint64_t)0x00FF00FF00FF00FF; + x2 = (q0 >> 8) & (uint64_t)0x00FF00FF00FF00FF; + x3 = (q1 >> 8) & (uint64_t)0x00FF00FF00FF00FF; + x0 |= (x0 >> 8); + x1 |= (x1 >> 8); + x2 |= (x2 >> 8); + x3 |= (x3 >> 8); + x0 &= (uint64_t)0x0000FFFF0000FFFF; + x1 &= (uint64_t)0x0000FFFF0000FFFF; + x2 &= (uint64_t)0x0000FFFF0000FFFF; + x3 &= (uint64_t)0x0000FFFF0000FFFF; + w[0] = (uint32_t)x0 | (uint32_t)(x0 >> 16); + w[1] = (uint32_t)x1 | (uint32_t)(x1 >> 16); + w[2] = (uint32_t)x2 | (uint32_t)(x2 >> 16); + w[3] = (uint32_t)x3 | (uint32_t)(x3 >> 16); +} + +static const unsigned char Rcon[] = { + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1B, 0x36 +}; + +static uint32_t sub_word(uint32_t x) +{ + uint64_t q[8]; + + memset(q, 0, sizeof q); + q[0] = x; + br_aes_ct64_ortho(q); + br_aes_ct64_bitslice_Sbox(q); + br_aes_ct64_ortho(q); + return (uint32_t)q[0]; +} + +static void br_aes_ct64_keysched(uint64_t *comp_skey, const unsigned char *key, int key_len) +{ + int i, j, k, nk, nkf; + uint32_t tmp; + uint32_t skey[60]; + unsigned nrounds = 10+((key_len-16)>>2); + + nk = (int)(key_len >> 2); + nkf = (int)((nrounds + 1) << 2); + br_range_dec32le(skey, (key_len >> 2), key); + tmp = skey[(key_len >> 2) - 1]; + for (i = nk, j = 0, k = 0; i < nkf; i ++) { + if (j == 0) { + tmp = (tmp << 24) | (tmp >> 8); + tmp = sub_word(tmp) ^ Rcon[k]; + } else if (nk > 6 && j == 4) { + tmp = sub_word(tmp); + } + tmp ^= skey[i - nk]; + skey[i] = tmp; + if (++ j == nk) { + j = 0; + k ++; + } + } + + for (i = 0, j = 0; i < nkf; i += 4, j += 2) { + uint64_t q[8]; + + br_aes_ct64_interleave_in(&q[0], &q[4], skey + i); + q[1] = q[0]; + q[2] = q[0]; + q[3] = q[0]; + q[5] = q[4]; + q[6] = q[4]; + q[7] = q[4]; + br_aes_ct64_ortho(q); + comp_skey[j + 0] = + (q[0] & (uint64_t)0x1111111111111111) + | (q[1] & (uint64_t)0x2222222222222222) + | (q[2] & (uint64_t)0x4444444444444444) + | (q[3] & (uint64_t)0x8888888888888888); + comp_skey[j + 1] = + (q[4] & (uint64_t)0x1111111111111111) + | (q[5] & (uint64_t)0x2222222222222222) + | (q[6] & (uint64_t)0x4444444444444444) + | (q[7] & (uint64_t)0x8888888888888888); + } +} + +void br_aes_ct64_skey_expand(uint64_t *skey, const uint64_t *comp_skey, unsigned int nrounds) +{ + unsigned u, v, n; + + n = (nrounds + 1) << 1; + for (u = 0, v = 0; u < n; u ++, v += 4) { + uint64_t x0, x1, x2, x3; + + x0 = x1 = x2 = x3 = comp_skey[u]; + x0 &= (uint64_t)0x1111111111111111; + x1 &= (uint64_t)0x2222222222222222; + x2 &= (uint64_t)0x4444444444444444; + x3 &= (uint64_t)0x8888888888888888; + x1 >>= 1; + x2 >>= 2; + x3 >>= 3; + skey[v + 0] = (x0 << 4) - x0; + skey[v + 1] = (x1 << 4) - x1; + skey[v + 2] = (x2 << 4) - x2; + skey[v + 3] = (x3 << 4) - x3; + } +} + + +static inline void add_round_key(uint64_t *q, const uint64_t *sk) +{ + q[0] ^= sk[0]; + q[1] ^= sk[1]; + q[2] ^= sk[2]; + q[3] ^= sk[3]; + q[4] ^= sk[4]; + q[5] ^= sk[5]; + q[6] ^= sk[6]; + q[7] ^= sk[7]; +} + +static inline void shift_rows(uint64_t *q) +{ + int i; + + for (i = 0; i < 8; i ++) { + uint64_t x; + + x = q[i]; + q[i] = (x & (uint64_t)0x000000000000FFFF) + | ((x & (uint64_t)0x00000000FFF00000) >> 4) + | ((x & (uint64_t)0x00000000000F0000) << 12) + | ((x & (uint64_t)0x0000FF0000000000) >> 8) + | ((x & (uint64_t)0x000000FF00000000) << 8) + | ((x & (uint64_t)0xF000000000000000) >> 12) + | ((x & (uint64_t)0x0FFF000000000000) << 4); + } +} + +static inline uint64_t rotr32(uint64_t x) +{ + return (x << 32) | (x >> 32); +} + +static inline void mix_columns(uint64_t *q) +{ + uint64_t q0, q1, q2, q3, q4, q5, q6, q7; + uint64_t r0, r1, r2, r3, r4, r5, r6, r7; + + q0 = q[0]; + q1 = q[1]; + q2 = q[2]; + q3 = q[3]; + q4 = q[4]; + q5 = q[5]; + q6 = q[6]; + q7 = q[7]; + r0 = (q0 >> 16) | (q0 << 48); + r1 = (q1 >> 16) | (q1 << 48); + r2 = (q2 >> 16) | (q2 << 48); + r3 = (q3 >> 16) | (q3 << 48); + r4 = (q4 >> 16) | (q4 << 48); + r5 = (q5 >> 16) | (q5 << 48); + r6 = (q6 >> 16) | (q6 << 48); + r7 = (q7 >> 16) | (q7 << 48); + + q[0] = q7 ^ r7 ^ r0 ^ rotr32(q0 ^ r0); + q[1] = q0 ^ r0 ^ q7 ^ r7 ^ r1 ^ rotr32(q1 ^ r1); + q[2] = q1 ^ r1 ^ r2 ^ rotr32(q2 ^ r2); + q[3] = q2 ^ r2 ^ q7 ^ r7 ^ r3 ^ rotr32(q3 ^ r3); + q[4] = q3 ^ r3 ^ q7 ^ r7 ^ r4 ^ rotr32(q4 ^ r4); + q[5] = q4 ^ r4 ^ r5 ^ rotr32(q5 ^ r5); + q[6] = q5 ^ r5 ^ r6 ^ rotr32(q6 ^ r6); + q[7] = q6 ^ r6 ^ r7 ^ rotr32(q7 ^ r7); +} + + +static void inc4_be(uint32_t *x) +{ + uint32_t t = br_swap32(*x)+4; + *x = br_swap32(t); +} + + +static void aes_ecb4x(unsigned char out[64], const uint32_t ivw[16], const uint64_t *sk_exp, unsigned int nrounds) +{ + uint32_t w[16]; + uint64_t q[8]; + unsigned int i; + + memcpy(w, ivw, sizeof(w)); + for (i = 0; i < 4; i++) { + br_aes_ct64_interleave_in(&q[i], &q[i + 4], w + (i << 2)); + } + br_aes_ct64_ortho(q); + + + add_round_key(q, sk_exp); + for (i = 1; i < nrounds; i++) { + br_aes_ct64_bitslice_Sbox(q); + shift_rows(q); + mix_columns(q); + add_round_key(q, sk_exp + (i << 3)); + } + br_aes_ct64_bitslice_Sbox(q); + shift_rows(q); + add_round_key(q, sk_exp + 8*nrounds); + + br_aes_ct64_ortho(q); + for (i = 0; i < 4; i ++) { + br_aes_ct64_interleave_out(w + (i << 2), q[i], q[i + 4]); + } + br_range_enc32le(out, w, 16); +} + + +static void aes_ctr4x(unsigned char out[64], uint32_t ivw[16], const uint64_t *sk_exp, unsigned int nrounds) +{ + aes_ecb4x(out, ivw, sk_exp, nrounds); + + /* Increase counter for next 4 blocks */ + inc4_be(ivw+3); + inc4_be(ivw+7); + inc4_be(ivw+11); + inc4_be(ivw+15); +} + + +static void aes_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const uint64_t *rkeys, unsigned int nrounds) +{ + uint32_t blocks[16]; + unsigned char t[64]; + + while(nblocks >= 4) + { + br_range_dec32le(blocks, 16, in); + aes_ecb4x(out, blocks, rkeys, nrounds); + nblocks -= 4; + in += 64; + out += 64; + } + + if(nblocks) + { + br_range_dec32le(blocks, nblocks*4, in); + aes_ecb4x(t, blocks, rkeys, nrounds); + memcpy(out, t, nblocks*16); + } +} + + +static void aes_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const uint64_t *rkeys, unsigned int nrounds) +{ + uint32_t ivw[16]; + size_t i; + uint32_t cc = 0; + + br_range_dec32le(ivw, 3, iv); + memcpy(ivw + 4, ivw, 3 * sizeof(uint32_t)); + memcpy(ivw + 8, ivw, 3 * sizeof(uint32_t)); + memcpy(ivw + 12, ivw, 3 * sizeof(uint32_t)); + ivw[ 3] = br_swap32(cc); + ivw[ 7] = br_swap32(cc + 1); + ivw[11] = br_swap32(cc + 2); + ivw[15] = br_swap32(cc + 3); + + while (outlen > 64) { + aes_ctr4x(out, ivw, rkeys, nrounds); + out += 64; + outlen -= 64; + } + if(outlen > 0) + { + unsigned char tmp[64]; + aes_ctr4x(tmp, ivw, rkeys, nrounds); + for(i=0;isk_exp, skey, 10); +} + + +void aes192_keyexp(aes192ctx *r, const unsigned char *key) +{ + uint64_t skey[26]; + + br_aes_ct64_keysched(skey, key, 24); + br_aes_ct64_skey_expand(r->sk_exp, skey, 12); +} + + +void aes256_keyexp(aes256ctx *r, const unsigned char *key) +{ + uint64_t skey[30]; + + br_aes_ct64_keysched(skey, key, 32); + br_aes_ct64_skey_expand(r->sk_exp, skey, 14); +} + + +void aes128_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes128ctx *ctx) +{ + aes_ecb(out, in, nblocks, ctx->sk_exp, 10); +} + +void aes128_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes128ctx *ctx) +{ + aes_ctr(out, outlen, iv, ctx->sk_exp, 10); +} + +void aes192_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes192ctx *ctx) +{ + aes_ecb(out, in, nblocks, ctx->sk_exp, 12); +} + +void aes192_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes192ctx *ctx) +{ + aes_ctr(out, outlen, iv, ctx->sk_exp, 12); +} + +void aes256_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes256ctx *ctx) +{ + aes_ecb(out, in, nblocks, ctx->sk_exp, 14); +} + +void aes256_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes256ctx *ctx) +{ + aes_ctr(out, outlen, iv, ctx->sk_exp, 14); +} + diff --git a/common/aes.h b/common/aes.h new file mode 100644 index 00000000..247fe0e3 --- /dev/null +++ b/common/aes.h @@ -0,0 +1,46 @@ +#ifndef AES_H +#define AES_H + +#include + +#define AES128_KEYBYTES 16 +#define AES192_KEYBYTES 24 +#define AES256_KEYBYTES 32 +#define AESCTR_NONCEBYTES 12 + +typedef struct { + uint64_t sk_exp[88]; +} aes128ctx; + +typedef struct { + uint64_t sk_exp[104]; +} aes192ctx; + +typedef struct { + uint64_t sk_exp[120]; +} aes256ctx; + + + +void aes128_keyexp(aes128ctx *r, const unsigned char *key); + +void aes128_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes128ctx *ctx); + +void aes128_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes128ctx *ctx); + + +void aes192_keyexp(aes192ctx *r, const unsigned char *key); + +void aes192_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes192ctx *ctx); + +void aes192_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes192ctx *ctx); + + +void aes256_keyexp(aes256ctx *r, const unsigned char *key); + +void aes256_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes256ctx *ctx); + +void aes256_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes256ctx *ctx); + + +#endif From 93c24e34c60a30bf46af31c6a0953785d0834bcc Mon Sep 17 00:00:00 2001 From: Peter Schwabe Date: Tue, 9 Apr 2019 14:48:46 +0200 Subject: [PATCH 02/14] Formatted AES code properly --- common/aes.c | 974 +++++++++++++++++++++++++-------------------------- common/aes.h | 6 +- 2 files changed, 474 insertions(+), 506 deletions(-) diff --git a/common/aes.c b/common/aes.c index 746514fe..bb58ada7 100644 --- a/common/aes.c +++ b/common/aes.c @@ -5,7 +5,7 @@ * * Copyright (c) 2016 Thomas Pornin * - * Permission is hereby granted, free of charge, to any person obtaining + * Permission is hereby granted, free of charge, to any person obtaining * a copy of this software and associated documentation files (the * "Software"), to deal in the Software without restriction, including * without limitation the rights to use, copy, modify, merge, publish, @@ -13,12 +13,12 @@ * permit persons to whom the Software is furnished to do so, subject to * the following conditions: * - * The above copyright notice and this permission notice shall be + * The above copyright notice and this permission notice shall be * included in all copies or substantial portions of the Software. * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN @@ -31,626 +31,594 @@ #include "ctaes.h" -static inline uint32_t br_dec32le(const unsigned char *src) -{ - return (uint32_t)src[0] - | ((uint32_t)src[1] << 8) - | ((uint32_t)src[2] << 16) - | ((uint32_t)src[3] << 24); +static inline uint32_t br_dec32le(const unsigned char *src) { + return (uint32_t)src[0] + | ((uint32_t)src[1] << 8) + | ((uint32_t)src[2] << 16) + | ((uint32_t)src[3] << 24); } -static void br_range_dec32le(uint32_t *v, size_t num, const unsigned char *src) -{ - while (num-- > 0) { - *v ++ = br_dec32le(src); - src += 4; - } +static void br_range_dec32le(uint32_t *v, size_t num, const unsigned char *src) { + while (num-- > 0) { + *v ++ = br_dec32le(src); + src += 4; + } } -static inline uint32_t br_swap32(uint32_t x) -{ - x = ((x & (uint32_t)0x00FF00FF) << 8) - | ((x >> 8) & (uint32_t)0x00FF00FF); - return (x << 16) | (x >> 16); +static inline uint32_t br_swap32(uint32_t x) { + x = ((x & (uint32_t)0x00FF00FF) << 8) + | ((x >> 8) & (uint32_t)0x00FF00FF); + return (x << 16) | (x >> 16); } -static inline void br_enc32le(unsigned char *dst, uint32_t x) -{ - dst[0] = (unsigned char)x; - dst[1] = (unsigned char)(x >> 8); - dst[2] = (unsigned char)(x >> 16); - dst[3] = (unsigned char)(x >> 24); +static inline void br_enc32le(unsigned char *dst, uint32_t x) { + dst[0] = (unsigned char)x; + dst[1] = (unsigned char)(x >> 8); + dst[2] = (unsigned char)(x >> 16); + dst[3] = (unsigned char)(x >> 24); } -void br_range_enc32le(unsigned char *dst, const uint32_t *v, size_t num) -{ - while (num-- > 0) { - br_enc32le(dst, *v ++); - dst += 4; - } +void br_range_enc32le(unsigned char *dst, const uint32_t *v, size_t num) { + while (num-- > 0) { + br_enc32le(dst, *v ++); + dst += 4; + } } -static void br_aes_ct64_bitslice_Sbox(uint64_t *q) -{ - /* - * This S-box implementation is a straightforward translation of - * the circuit described by Boyar and Peralta in "A new - * combinational logic minimization technique with applications - * to cryptology" (https://eprint.iacr.org/2009/191.pdf). - * - * Note that variables x* (input) and s* (output) are numbered - * in "reverse" order (x0 is the high bit, x7 is the low bit). - */ +static void br_aes_ct64_bitslice_Sbox(uint64_t *q) { + /* + * This S-box implementation is a straightforward translation of + * the circuit described by Boyar and Peralta in "A new + * combinational logic minimization technique with applications + * to cryptology" (https://eprint.iacr.org/2009/191.pdf). + * + * Note that variables x* (input) and s* (output) are numbered + * in "reverse" order (x0 is the high bit, x7 is the low bit). + */ - uint64_t x0, x1, x2, x3, x4, x5, x6, x7; - uint64_t y1, y2, y3, y4, y5, y6, y7, y8, y9; - uint64_t y10, y11, y12, y13, y14, y15, y16, y17, y18, y19; - uint64_t y20, y21; - uint64_t z0, z1, z2, z3, z4, z5, z6, z7, z8, z9; - uint64_t z10, z11, z12, z13, z14, z15, z16, z17; - uint64_t t0, t1, t2, t3, t4, t5, t6, t7, t8, t9; - uint64_t t10, t11, t12, t13, t14, t15, t16, t17, t18, t19; - uint64_t t20, t21, t22, t23, t24, t25, t26, t27, t28, t29; - uint64_t t30, t31, t32, t33, t34, t35, t36, t37, t38, t39; - uint64_t t40, t41, t42, t43, t44, t45, t46, t47, t48, t49; - uint64_t t50, t51, t52, t53, t54, t55, t56, t57, t58, t59; - uint64_t t60, t61, t62, t63, t64, t65, t66, t67; - uint64_t s0, s1, s2, s3, s4, s5, s6, s7; + uint64_t x0, x1, x2, x3, x4, x5, x6, x7; + uint64_t y1, y2, y3, y4, y5, y6, y7, y8, y9; + uint64_t y10, y11, y12, y13, y14, y15, y16, y17, y18, y19; + uint64_t y20, y21; + uint64_t z0, z1, z2, z3, z4, z5, z6, z7, z8, z9; + uint64_t z10, z11, z12, z13, z14, z15, z16, z17; + uint64_t t0, t1, t2, t3, t4, t5, t6, t7, t8, t9; + uint64_t t10, t11, t12, t13, t14, t15, t16, t17, t18, t19; + uint64_t t20, t21, t22, t23, t24, t25, t26, t27, t28, t29; + uint64_t t30, t31, t32, t33, t34, t35, t36, t37, t38, t39; + uint64_t t40, t41, t42, t43, t44, t45, t46, t47, t48, t49; + uint64_t t50, t51, t52, t53, t54, t55, t56, t57, t58, t59; + uint64_t t60, t61, t62, t63, t64, t65, t66, t67; + uint64_t s0, s1, s2, s3, s4, s5, s6, s7; - x0 = q[7]; - x1 = q[6]; - x2 = q[5]; - x3 = q[4]; - x4 = q[3]; - x5 = q[2]; - x6 = q[1]; - x7 = q[0]; + x0 = q[7]; + x1 = q[6]; + x2 = q[5]; + x3 = q[4]; + x4 = q[3]; + x5 = q[2]; + x6 = q[1]; + x7 = q[0]; - /* - * Top linear transformation. - */ - y14 = x3 ^ x5; - y13 = x0 ^ x6; - y9 = x0 ^ x3; - y8 = x0 ^ x5; - t0 = x1 ^ x2; - y1 = t0 ^ x7; - y4 = y1 ^ x3; - y12 = y13 ^ y14; - y2 = y1 ^ x0; - y5 = y1 ^ x6; - y3 = y5 ^ y8; - t1 = x4 ^ y12; - y15 = t1 ^ x5; - y20 = t1 ^ x1; - y6 = y15 ^ x7; - y10 = y15 ^ t0; - y11 = y20 ^ y9; - y7 = x7 ^ y11; - y17 = y10 ^ y11; - y19 = y10 ^ y8; - y16 = t0 ^ y11; - y21 = y13 ^ y16; - y18 = x0 ^ y16; + /* + * Top linear transformation. + */ + y14 = x3 ^ x5; + y13 = x0 ^ x6; + y9 = x0 ^ x3; + y8 = x0 ^ x5; + t0 = x1 ^ x2; + y1 = t0 ^ x7; + y4 = y1 ^ x3; + y12 = y13 ^ y14; + y2 = y1 ^ x0; + y5 = y1 ^ x6; + y3 = y5 ^ y8; + t1 = x4 ^ y12; + y15 = t1 ^ x5; + y20 = t1 ^ x1; + y6 = y15 ^ x7; + y10 = y15 ^ t0; + y11 = y20 ^ y9; + y7 = x7 ^ y11; + y17 = y10 ^ y11; + y19 = y10 ^ y8; + y16 = t0 ^ y11; + y21 = y13 ^ y16; + y18 = x0 ^ y16; - /* - * Non-linear section. - */ - t2 = y12 & y15; - t3 = y3 & y6; - t4 = t3 ^ t2; - t5 = y4 & x7; - t6 = t5 ^ t2; - t7 = y13 & y16; - t8 = y5 & y1; - t9 = t8 ^ t7; - t10 = y2 & y7; - t11 = t10 ^ t7; - t12 = y9 & y11; - t13 = y14 & y17; - t14 = t13 ^ t12; - t15 = y8 & y10; - t16 = t15 ^ t12; - t17 = t4 ^ t14; - t18 = t6 ^ t16; - t19 = t9 ^ t14; - t20 = t11 ^ t16; - t21 = t17 ^ y20; - t22 = t18 ^ y19; - t23 = t19 ^ y21; - t24 = t20 ^ y18; + /* + * Non-linear section. + */ + t2 = y12 & y15; + t3 = y3 & y6; + t4 = t3 ^ t2; + t5 = y4 & x7; + t6 = t5 ^ t2; + t7 = y13 & y16; + t8 = y5 & y1; + t9 = t8 ^ t7; + t10 = y2 & y7; + t11 = t10 ^ t7; + t12 = y9 & y11; + t13 = y14 & y17; + t14 = t13 ^ t12; + t15 = y8 & y10; + t16 = t15 ^ t12; + t17 = t4 ^ t14; + t18 = t6 ^ t16; + t19 = t9 ^ t14; + t20 = t11 ^ t16; + t21 = t17 ^ y20; + t22 = t18 ^ y19; + t23 = t19 ^ y21; + t24 = t20 ^ y18; - t25 = t21 ^ t22; - t26 = t21 & t23; - t27 = t24 ^ t26; - t28 = t25 & t27; - t29 = t28 ^ t22; - t30 = t23 ^ t24; - t31 = t22 ^ t26; - t32 = t31 & t30; - t33 = t32 ^ t24; - t34 = t23 ^ t33; - t35 = t27 ^ t33; - t36 = t24 & t35; - t37 = t36 ^ t34; - t38 = t27 ^ t36; - t39 = t29 & t38; - t40 = t25 ^ t39; + t25 = t21 ^ t22; + t26 = t21 & t23; + t27 = t24 ^ t26; + t28 = t25 & t27; + t29 = t28 ^ t22; + t30 = t23 ^ t24; + t31 = t22 ^ t26; + t32 = t31 & t30; + t33 = t32 ^ t24; + t34 = t23 ^ t33; + t35 = t27 ^ t33; + t36 = t24 & t35; + t37 = t36 ^ t34; + t38 = t27 ^ t36; + t39 = t29 & t38; + t40 = t25 ^ t39; - t41 = t40 ^ t37; - t42 = t29 ^ t33; - t43 = t29 ^ t40; - t44 = t33 ^ t37; - t45 = t42 ^ t41; - z0 = t44 & y15; - z1 = t37 & y6; - z2 = t33 & x7; - z3 = t43 & y16; - z4 = t40 & y1; - z5 = t29 & y7; - z6 = t42 & y11; - z7 = t45 & y17; - z8 = t41 & y10; - z9 = t44 & y12; - z10 = t37 & y3; - z11 = t33 & y4; - z12 = t43 & y13; - z13 = t40 & y5; - z14 = t29 & y2; - z15 = t42 & y9; - z16 = t45 & y14; - z17 = t41 & y8; + t41 = t40 ^ t37; + t42 = t29 ^ t33; + t43 = t29 ^ t40; + t44 = t33 ^ t37; + t45 = t42 ^ t41; + z0 = t44 & y15; + z1 = t37 & y6; + z2 = t33 & x7; + z3 = t43 & y16; + z4 = t40 & y1; + z5 = t29 & y7; + z6 = t42 & y11; + z7 = t45 & y17; + z8 = t41 & y10; + z9 = t44 & y12; + z10 = t37 & y3; + z11 = t33 & y4; + z12 = t43 & y13; + z13 = t40 & y5; + z14 = t29 & y2; + z15 = t42 & y9; + z16 = t45 & y14; + z17 = t41 & y8; - /* - * Bottom linear transformation. - */ - t46 = z15 ^ z16; - t47 = z10 ^ z11; - t48 = z5 ^ z13; - t49 = z9 ^ z10; - t50 = z2 ^ z12; - t51 = z2 ^ z5; - t52 = z7 ^ z8; - t53 = z0 ^ z3; - t54 = z6 ^ z7; - t55 = z16 ^ z17; - t56 = z12 ^ t48; - t57 = t50 ^ t53; - t58 = z4 ^ t46; - t59 = z3 ^ t54; - t60 = t46 ^ t57; - t61 = z14 ^ t57; - t62 = t52 ^ t58; - t63 = t49 ^ t58; - t64 = z4 ^ t59; - t65 = t61 ^ t62; - t66 = z1 ^ t63; - s0 = t59 ^ t63; - s6 = t56 ^ ~t62; - s7 = t48 ^ ~t60; - t67 = t64 ^ t65; - s3 = t53 ^ t66; - s4 = t51 ^ t66; - s5 = t47 ^ t65; - s1 = t64 ^ ~s3; - s2 = t55 ^ ~t67; + /* + * Bottom linear transformation. + */ + t46 = z15 ^ z16; + t47 = z10 ^ z11; + t48 = z5 ^ z13; + t49 = z9 ^ z10; + t50 = z2 ^ z12; + t51 = z2 ^ z5; + t52 = z7 ^ z8; + t53 = z0 ^ z3; + t54 = z6 ^ z7; + t55 = z16 ^ z17; + t56 = z12 ^ t48; + t57 = t50 ^ t53; + t58 = z4 ^ t46; + t59 = z3 ^ t54; + t60 = t46 ^ t57; + t61 = z14 ^ t57; + t62 = t52 ^ t58; + t63 = t49 ^ t58; + t64 = z4 ^ t59; + t65 = t61 ^ t62; + t66 = z1 ^ t63; + s0 = t59 ^ t63; + s6 = t56 ^ ~t62; + s7 = t48 ^ ~t60; + t67 = t64 ^ t65; + s3 = t53 ^ t66; + s4 = t51 ^ t66; + s5 = t47 ^ t65; + s1 = t64 ^ ~s3; + s2 = t55 ^ ~t67; - q[7] = s0; - q[6] = s1; - q[5] = s2; - q[4] = s3; - q[3] = s4; - q[2] = s5; - q[1] = s6; - q[0] = s7; + q[7] = s0; + q[6] = s1; + q[5] = s2; + q[4] = s3; + q[3] = s4; + q[2] = s5; + q[1] = s6; + q[0] = s7; } -static void br_aes_ct64_ortho(uint64_t *q) -{ +static void br_aes_ct64_ortho(uint64_t *q) { #define SWAPN(cl, ch, s, x, y) do { \ - uint64_t a, b; \ - a = (x); \ - b = (y); \ - (x) = (a & (uint64_t)cl) | ((b & (uint64_t)cl) << (s)); \ - (y) = ((a & (uint64_t)ch) >> (s)) | (b & (uint64_t)ch); \ - } while (0) + uint64_t a, b; \ + a = (x); \ + b = (y); \ + (x) = (a & (uint64_t)cl) | ((b & (uint64_t)cl) << (s)); \ + (y) = ((a & (uint64_t)ch) >> (s)) | (b & (uint64_t)ch); \ + } while (0) #define SWAP2(x, y) SWAPN(0x5555555555555555, 0xAAAAAAAAAAAAAAAA, 1, x, y) #define SWAP4(x, y) SWAPN(0x3333333333333333, 0xCCCCCCCCCCCCCCCC, 2, x, y) #define SWAP8(x, y) SWAPN(0x0F0F0F0F0F0F0F0F, 0xF0F0F0F0F0F0F0F0, 4, x, y) - SWAP2(q[0], q[1]); - SWAP2(q[2], q[3]); - SWAP2(q[4], q[5]); - SWAP2(q[6], q[7]); + SWAP2(q[0], q[1]); + SWAP2(q[2], q[3]); + SWAP2(q[4], q[5]); + SWAP2(q[6], q[7]); - SWAP4(q[0], q[2]); - SWAP4(q[1], q[3]); - SWAP4(q[4], q[6]); - SWAP4(q[5], q[7]); + SWAP4(q[0], q[2]); + SWAP4(q[1], q[3]); + SWAP4(q[4], q[6]); + SWAP4(q[5], q[7]); - SWAP8(q[0], q[4]); - SWAP8(q[1], q[5]); - SWAP8(q[2], q[6]); - SWAP8(q[3], q[7]); + SWAP8(q[0], q[4]); + SWAP8(q[1], q[5]); + SWAP8(q[2], q[6]); + SWAP8(q[3], q[7]); } -static void br_aes_ct64_interleave_in(uint64_t *q0, uint64_t *q1, const uint32_t *w) -{ - uint64_t x0, x1, x2, x3; +static void br_aes_ct64_interleave_in(uint64_t *q0, uint64_t *q1, const uint32_t *w) { + uint64_t x0, x1, x2, x3; - x0 = w[0]; - x1 = w[1]; - x2 = w[2]; - x3 = w[3]; - x0 |= (x0 << 16); - x1 |= (x1 << 16); - x2 |= (x2 << 16); - x3 |= (x3 << 16); - x0 &= (uint64_t)0x0000FFFF0000FFFF; - x1 &= (uint64_t)0x0000FFFF0000FFFF; - x2 &= (uint64_t)0x0000FFFF0000FFFF; - x3 &= (uint64_t)0x0000FFFF0000FFFF; - x0 |= (x0 << 8); - x1 |= (x1 << 8); - x2 |= (x2 << 8); - x3 |= (x3 << 8); - x0 &= (uint64_t)0x00FF00FF00FF00FF; - x1 &= (uint64_t)0x00FF00FF00FF00FF; - x2 &= (uint64_t)0x00FF00FF00FF00FF; - x3 &= (uint64_t)0x00FF00FF00FF00FF; - *q0 = x0 | (x2 << 8); - *q1 = x1 | (x3 << 8); + x0 = w[0]; + x1 = w[1]; + x2 = w[2]; + x3 = w[3]; + x0 |= (x0 << 16); + x1 |= (x1 << 16); + x2 |= (x2 << 16); + x3 |= (x3 << 16); + x0 &= (uint64_t)0x0000FFFF0000FFFF; + x1 &= (uint64_t)0x0000FFFF0000FFFF; + x2 &= (uint64_t)0x0000FFFF0000FFFF; + x3 &= (uint64_t)0x0000FFFF0000FFFF; + x0 |= (x0 << 8); + x1 |= (x1 << 8); + x2 |= (x2 << 8); + x3 |= (x3 << 8); + x0 &= (uint64_t)0x00FF00FF00FF00FF; + x1 &= (uint64_t)0x00FF00FF00FF00FF; + x2 &= (uint64_t)0x00FF00FF00FF00FF; + x3 &= (uint64_t)0x00FF00FF00FF00FF; + *q0 = x0 | (x2 << 8); + *q1 = x1 | (x3 << 8); } -static void br_aes_ct64_interleave_out(uint32_t *w, uint64_t q0, uint64_t q1) -{ - uint64_t x0, x1, x2, x3; +static void br_aes_ct64_interleave_out(uint32_t *w, uint64_t q0, uint64_t q1) { + uint64_t x0, x1, x2, x3; - x0 = q0 & (uint64_t)0x00FF00FF00FF00FF; - x1 = q1 & (uint64_t)0x00FF00FF00FF00FF; - x2 = (q0 >> 8) & (uint64_t)0x00FF00FF00FF00FF; - x3 = (q1 >> 8) & (uint64_t)0x00FF00FF00FF00FF; - x0 |= (x0 >> 8); - x1 |= (x1 >> 8); - x2 |= (x2 >> 8); - x3 |= (x3 >> 8); - x0 &= (uint64_t)0x0000FFFF0000FFFF; - x1 &= (uint64_t)0x0000FFFF0000FFFF; - x2 &= (uint64_t)0x0000FFFF0000FFFF; - x3 &= (uint64_t)0x0000FFFF0000FFFF; - w[0] = (uint32_t)x0 | (uint32_t)(x0 >> 16); - w[1] = (uint32_t)x1 | (uint32_t)(x1 >> 16); - w[2] = (uint32_t)x2 | (uint32_t)(x2 >> 16); - w[3] = (uint32_t)x3 | (uint32_t)(x3 >> 16); + x0 = q0 & (uint64_t)0x00FF00FF00FF00FF; + x1 = q1 & (uint64_t)0x00FF00FF00FF00FF; + x2 = (q0 >> 8) & (uint64_t)0x00FF00FF00FF00FF; + x3 = (q1 >> 8) & (uint64_t)0x00FF00FF00FF00FF; + x0 |= (x0 >> 8); + x1 |= (x1 >> 8); + x2 |= (x2 >> 8); + x3 |= (x3 >> 8); + x0 &= (uint64_t)0x0000FFFF0000FFFF; + x1 &= (uint64_t)0x0000FFFF0000FFFF; + x2 &= (uint64_t)0x0000FFFF0000FFFF; + x3 &= (uint64_t)0x0000FFFF0000FFFF; + w[0] = (uint32_t)x0 | (uint32_t)(x0 >> 16); + w[1] = (uint32_t)x1 | (uint32_t)(x1 >> 16); + w[2] = (uint32_t)x2 | (uint32_t)(x2 >> 16); + w[3] = (uint32_t)x3 | (uint32_t)(x3 >> 16); } static const unsigned char Rcon[] = { - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1B, 0x36 + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1B, 0x36 }; -static uint32_t sub_word(uint32_t x) -{ - uint64_t q[8]; +static uint32_t sub_word(uint32_t x) { + uint64_t q[8]; - memset(q, 0, sizeof q); - q[0] = x; - br_aes_ct64_ortho(q); - br_aes_ct64_bitslice_Sbox(q); - br_aes_ct64_ortho(q); - return (uint32_t)q[0]; + memset(q, 0, sizeof q); + q[0] = x; + br_aes_ct64_ortho(q); + br_aes_ct64_bitslice_Sbox(q); + br_aes_ct64_ortho(q); + return (uint32_t)q[0]; } -static void br_aes_ct64_keysched(uint64_t *comp_skey, const unsigned char *key, int key_len) -{ - int i, j, k, nk, nkf; - uint32_t tmp; - uint32_t skey[60]; - unsigned nrounds = 10+((key_len-16)>>2); +static void br_aes_ct64_keysched(uint64_t *comp_skey, const unsigned char *key, int key_len) { + int i, j, k, nk, nkf; + uint32_t tmp; + uint32_t skey[60]; + unsigned nrounds = 10 + ((key_len - 16) >> 2); - nk = (int)(key_len >> 2); - nkf = (int)((nrounds + 1) << 2); - br_range_dec32le(skey, (key_len >> 2), key); - tmp = skey[(key_len >> 2) - 1]; - for (i = nk, j = 0, k = 0; i < nkf; i ++) { - if (j == 0) { - tmp = (tmp << 24) | (tmp >> 8); - tmp = sub_word(tmp) ^ Rcon[k]; - } else if (nk > 6 && j == 4) { - tmp = sub_word(tmp); - } - tmp ^= skey[i - nk]; - skey[i] = tmp; - if (++ j == nk) { - j = 0; - k ++; - } - } + nk = (int)(key_len >> 2); + nkf = (int)((nrounds + 1) << 2); + br_range_dec32le(skey, (key_len >> 2), key); + tmp = skey[(key_len >> 2) - 1]; + for (i = nk, j = 0, k = 0; i < nkf; i ++) { + if (j == 0) { + tmp = (tmp << 24) | (tmp >> 8); + tmp = sub_word(tmp) ^ Rcon[k]; + } else if (nk > 6 && j == 4) { + tmp = sub_word(tmp); + } + tmp ^= skey[i - nk]; + skey[i] = tmp; + if (++ j == nk) { + j = 0; + k ++; + } + } - for (i = 0, j = 0; i < nkf; i += 4, j += 2) { - uint64_t q[8]; + for (i = 0, j = 0; i < nkf; i += 4, j += 2) { + uint64_t q[8]; - br_aes_ct64_interleave_in(&q[0], &q[4], skey + i); - q[1] = q[0]; - q[2] = q[0]; - q[3] = q[0]; - q[5] = q[4]; - q[6] = q[4]; - q[7] = q[4]; - br_aes_ct64_ortho(q); - comp_skey[j + 0] = - (q[0] & (uint64_t)0x1111111111111111) - | (q[1] & (uint64_t)0x2222222222222222) - | (q[2] & (uint64_t)0x4444444444444444) - | (q[3] & (uint64_t)0x8888888888888888); - comp_skey[j + 1] = - (q[4] & (uint64_t)0x1111111111111111) - | (q[5] & (uint64_t)0x2222222222222222) - | (q[6] & (uint64_t)0x4444444444444444) - | (q[7] & (uint64_t)0x8888888888888888); - } + br_aes_ct64_interleave_in(&q[0], &q[4], skey + i); + q[1] = q[0]; + q[2] = q[0]; + q[3] = q[0]; + q[5] = q[4]; + q[6] = q[4]; + q[7] = q[4]; + br_aes_ct64_ortho(q); + comp_skey[j + 0] = + (q[0] & (uint64_t)0x1111111111111111) + | (q[1] & (uint64_t)0x2222222222222222) + | (q[2] & (uint64_t)0x4444444444444444) + | (q[3] & (uint64_t)0x8888888888888888); + comp_skey[j + 1] = + (q[4] & (uint64_t)0x1111111111111111) + | (q[5] & (uint64_t)0x2222222222222222) + | (q[6] & (uint64_t)0x4444444444444444) + | (q[7] & (uint64_t)0x8888888888888888); + } } -void br_aes_ct64_skey_expand(uint64_t *skey, const uint64_t *comp_skey, unsigned int nrounds) -{ - unsigned u, v, n; +void br_aes_ct64_skey_expand(uint64_t *skey, const uint64_t *comp_skey, unsigned int nrounds) { + unsigned u, v, n; - n = (nrounds + 1) << 1; - for (u = 0, v = 0; u < n; u ++, v += 4) { - uint64_t x0, x1, x2, x3; + n = (nrounds + 1) << 1; + for (u = 0, v = 0; u < n; u ++, v += 4) { + uint64_t x0, x1, x2, x3; - x0 = x1 = x2 = x3 = comp_skey[u]; - x0 &= (uint64_t)0x1111111111111111; - x1 &= (uint64_t)0x2222222222222222; - x2 &= (uint64_t)0x4444444444444444; - x3 &= (uint64_t)0x8888888888888888; - x1 >>= 1; - x2 >>= 2; - x3 >>= 3; - skey[v + 0] = (x0 << 4) - x0; - skey[v + 1] = (x1 << 4) - x1; - skey[v + 2] = (x2 << 4) - x2; - skey[v + 3] = (x3 << 4) - x3; - } + x0 = x1 = x2 = x3 = comp_skey[u]; + x0 &= (uint64_t)0x1111111111111111; + x1 &= (uint64_t)0x2222222222222222; + x2 &= (uint64_t)0x4444444444444444; + x3 &= (uint64_t)0x8888888888888888; + x1 >>= 1; + x2 >>= 2; + x3 >>= 3; + skey[v + 0] = (x0 << 4) - x0; + skey[v + 1] = (x1 << 4) - x1; + skey[v + 2] = (x2 << 4) - x2; + skey[v + 3] = (x3 << 4) - x3; + } } -static inline void add_round_key(uint64_t *q, const uint64_t *sk) -{ - q[0] ^= sk[0]; - q[1] ^= sk[1]; - q[2] ^= sk[2]; - q[3] ^= sk[3]; - q[4] ^= sk[4]; - q[5] ^= sk[5]; - q[6] ^= sk[6]; - q[7] ^= sk[7]; +static inline void add_round_key(uint64_t *q, const uint64_t *sk) { + q[0] ^= sk[0]; + q[1] ^= sk[1]; + q[2] ^= sk[2]; + q[3] ^= sk[3]; + q[4] ^= sk[4]; + q[5] ^= sk[5]; + q[6] ^= sk[6]; + q[7] ^= sk[7]; } -static inline void shift_rows(uint64_t *q) -{ - int i; +static inline void shift_rows(uint64_t *q) { + int i; - for (i = 0; i < 8; i ++) { - uint64_t x; + for (i = 0; i < 8; i ++) { + uint64_t x; - x = q[i]; - q[i] = (x & (uint64_t)0x000000000000FFFF) - | ((x & (uint64_t)0x00000000FFF00000) >> 4) - | ((x & (uint64_t)0x00000000000F0000) << 12) - | ((x & (uint64_t)0x0000FF0000000000) >> 8) - | ((x & (uint64_t)0x000000FF00000000) << 8) - | ((x & (uint64_t)0xF000000000000000) >> 12) - | ((x & (uint64_t)0x0FFF000000000000) << 4); - } + x = q[i]; + q[i] = (x & (uint64_t)0x000000000000FFFF) + | ((x & (uint64_t)0x00000000FFF00000) >> 4) + | ((x & (uint64_t)0x00000000000F0000) << 12) + | ((x & (uint64_t)0x0000FF0000000000) >> 8) + | ((x & (uint64_t)0x000000FF00000000) << 8) + | ((x & (uint64_t)0xF000000000000000) >> 12) + | ((x & (uint64_t)0x0FFF000000000000) << 4); + } } -static inline uint64_t rotr32(uint64_t x) -{ - return (x << 32) | (x >> 32); +static inline uint64_t rotr32(uint64_t x) { + return (x << 32) | (x >> 32); } -static inline void mix_columns(uint64_t *q) -{ - uint64_t q0, q1, q2, q3, q4, q5, q6, q7; - uint64_t r0, r1, r2, r3, r4, r5, r6, r7; +static inline void mix_columns(uint64_t *q) { + uint64_t q0, q1, q2, q3, q4, q5, q6, q7; + uint64_t r0, r1, r2, r3, r4, r5, r6, r7; - q0 = q[0]; - q1 = q[1]; - q2 = q[2]; - q3 = q[3]; - q4 = q[4]; - q5 = q[5]; - q6 = q[6]; - q7 = q[7]; - r0 = (q0 >> 16) | (q0 << 48); - r1 = (q1 >> 16) | (q1 << 48); - r2 = (q2 >> 16) | (q2 << 48); - r3 = (q3 >> 16) | (q3 << 48); - r4 = (q4 >> 16) | (q4 << 48); - r5 = (q5 >> 16) | (q5 << 48); - r6 = (q6 >> 16) | (q6 << 48); - r7 = (q7 >> 16) | (q7 << 48); + q0 = q[0]; + q1 = q[1]; + q2 = q[2]; + q3 = q[3]; + q4 = q[4]; + q5 = q[5]; + q6 = q[6]; + q7 = q[7]; + r0 = (q0 >> 16) | (q0 << 48); + r1 = (q1 >> 16) | (q1 << 48); + r2 = (q2 >> 16) | (q2 << 48); + r3 = (q3 >> 16) | (q3 << 48); + r4 = (q4 >> 16) | (q4 << 48); + r5 = (q5 >> 16) | (q5 << 48); + r6 = (q6 >> 16) | (q6 << 48); + r7 = (q7 >> 16) | (q7 << 48); - q[0] = q7 ^ r7 ^ r0 ^ rotr32(q0 ^ r0); - q[1] = q0 ^ r0 ^ q7 ^ r7 ^ r1 ^ rotr32(q1 ^ r1); - q[2] = q1 ^ r1 ^ r2 ^ rotr32(q2 ^ r2); - q[3] = q2 ^ r2 ^ q7 ^ r7 ^ r3 ^ rotr32(q3 ^ r3); - q[4] = q3 ^ r3 ^ q7 ^ r7 ^ r4 ^ rotr32(q4 ^ r4); - q[5] = q4 ^ r4 ^ r5 ^ rotr32(q5 ^ r5); - q[6] = q5 ^ r5 ^ r6 ^ rotr32(q6 ^ r6); - q[7] = q6 ^ r6 ^ r7 ^ rotr32(q7 ^ r7); + q[0] = q7 ^ r7 ^ r0 ^ rotr32(q0 ^ r0); + q[1] = q0 ^ r0 ^ q7 ^ r7 ^ r1 ^ rotr32(q1 ^ r1); + q[2] = q1 ^ r1 ^ r2 ^ rotr32(q2 ^ r2); + q[3] = q2 ^ r2 ^ q7 ^ r7 ^ r3 ^ rotr32(q3 ^ r3); + q[4] = q3 ^ r3 ^ q7 ^ r7 ^ r4 ^ rotr32(q4 ^ r4); + q[5] = q4 ^ r4 ^ r5 ^ rotr32(q5 ^ r5); + q[6] = q5 ^ r5 ^ r6 ^ rotr32(q6 ^ r6); + q[7] = q6 ^ r6 ^ r7 ^ rotr32(q7 ^ r7); } -static void inc4_be(uint32_t *x) -{ - uint32_t t = br_swap32(*x)+4; - *x = br_swap32(t); +static void inc4_be(uint32_t *x) { + uint32_t t = br_swap32(*x) + 4; + *x = br_swap32(t); } -static void aes_ecb4x(unsigned char out[64], const uint32_t ivw[16], const uint64_t *sk_exp, unsigned int nrounds) -{ - uint32_t w[16]; - uint64_t q[8]; - unsigned int i; +static void aes_ecb4x(unsigned char out[64], const uint32_t ivw[16], const uint64_t *sk_exp, unsigned int nrounds) { + uint32_t w[16]; + uint64_t q[8]; + unsigned int i; - memcpy(w, ivw, sizeof(w)); - for (i = 0; i < 4; i++) { - br_aes_ct64_interleave_in(&q[i], &q[i + 4], w + (i << 2)); - } - br_aes_ct64_ortho(q); + memcpy(w, ivw, sizeof(w)); + for (i = 0; i < 4; i++) { + br_aes_ct64_interleave_in(&q[i], &q[i + 4], w + (i << 2)); + } + br_aes_ct64_ortho(q); - add_round_key(q, sk_exp); - for (i = 1; i < nrounds; i++) { + add_round_key(q, sk_exp); + for (i = 1; i < nrounds; i++) { + br_aes_ct64_bitslice_Sbox(q); + shift_rows(q); + mix_columns(q); + add_round_key(q, sk_exp + (i << 3)); + } br_aes_ct64_bitslice_Sbox(q); shift_rows(q); - mix_columns(q); - add_round_key(q, sk_exp + (i << 3)); - } - br_aes_ct64_bitslice_Sbox(q); - shift_rows(q); - add_round_key(q, sk_exp + 8*nrounds); + add_round_key(q, sk_exp + 8 * nrounds); - br_aes_ct64_ortho(q); - for (i = 0; i < 4; i ++) { - br_aes_ct64_interleave_out(w + (i << 2), q[i], q[i + 4]); - } - br_range_enc32le(out, w, 16); + br_aes_ct64_ortho(q); + for (i = 0; i < 4; i ++) { + br_aes_ct64_interleave_out(w + (i << 2), q[i], q[i + 4]); + } + br_range_enc32le(out, w, 16); } -static void aes_ctr4x(unsigned char out[64], uint32_t ivw[16], const uint64_t *sk_exp, unsigned int nrounds) -{ - aes_ecb4x(out, ivw, sk_exp, nrounds); +static void aes_ctr4x(unsigned char out[64], uint32_t ivw[16], const uint64_t *sk_exp, unsigned int nrounds) { + aes_ecb4x(out, ivw, sk_exp, nrounds); - /* Increase counter for next 4 blocks */ - inc4_be(ivw+3); - inc4_be(ivw+7); - inc4_be(ivw+11); - inc4_be(ivw+15); + /* Increase counter for next 4 blocks */ + inc4_be(ivw + 3); + inc4_be(ivw + 7); + inc4_be(ivw + 11); + inc4_be(ivw + 15); } -static void aes_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const uint64_t *rkeys, unsigned int nrounds) -{ - uint32_t blocks[16]; - unsigned char t[64]; +static void aes_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const uint64_t *rkeys, unsigned int nrounds) { + uint32_t blocks[16]; + unsigned char t[64]; - while(nblocks >= 4) - { - br_range_dec32le(blocks, 16, in); - aes_ecb4x(out, blocks, rkeys, nrounds); - nblocks -= 4; - in += 64; - out += 64; - } + while (nblocks >= 4) { + br_range_dec32le(blocks, 16, in); + aes_ecb4x(out, blocks, rkeys, nrounds); + nblocks -= 4; + in += 64; + out += 64; + } - if(nblocks) - { - br_range_dec32le(blocks, nblocks*4, in); - aes_ecb4x(t, blocks, rkeys, nrounds); - memcpy(out, t, nblocks*16); - } + if (nblocks) { + br_range_dec32le(blocks, nblocks * 4, in); + aes_ecb4x(t, blocks, rkeys, nrounds); + memcpy(out, t, nblocks * 16); + } } - -static void aes_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const uint64_t *rkeys, unsigned int nrounds) -{ - uint32_t ivw[16]; - size_t i; - uint32_t cc = 0; - br_range_dec32le(ivw, 3, iv); - memcpy(ivw + 4, ivw, 3 * sizeof(uint32_t)); - memcpy(ivw + 8, ivw, 3 * sizeof(uint32_t)); - memcpy(ivw + 12, ivw, 3 * sizeof(uint32_t)); - ivw[ 3] = br_swap32(cc); - ivw[ 7] = br_swap32(cc + 1); - ivw[11] = br_swap32(cc + 2); - ivw[15] = br_swap32(cc + 3); +static void aes_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const uint64_t *rkeys, unsigned int nrounds) { + uint32_t ivw[16]; + size_t i; + uint32_t cc = 0; - while (outlen > 64) { - aes_ctr4x(out, ivw, rkeys, nrounds); - out += 64; - outlen -= 64; - } - if(outlen > 0) - { - unsigned char tmp[64]; - aes_ctr4x(tmp, ivw, rkeys, nrounds); - for(i=0;i 64) { + aes_ctr4x(out, ivw, rkeys, nrounds); + out += 64; + outlen -= 64; + } + if (outlen > 0) { + unsigned char tmp[64]; + aes_ctr4x(tmp, ivw, rkeys, nrounds); + for (i = 0; i < outlen; i++) { + out[i] = tmp[i]; + } + } } -void aes128_keyexp(aes128ctx *r, const unsigned char *key) -{ - uint64_t skey[22]; +void aes128_keyexp(aes128ctx *r, const unsigned char *key) { + uint64_t skey[22]; - br_aes_ct64_keysched(skey, key, 16); - br_aes_ct64_skey_expand(r->sk_exp, skey, 10); + br_aes_ct64_keysched(skey, key, 16); + br_aes_ct64_skey_expand(r->sk_exp, skey, 10); } -void aes192_keyexp(aes192ctx *r, const unsigned char *key) -{ - uint64_t skey[26]; +void aes192_keyexp(aes192ctx *r, const unsigned char *key) { + uint64_t skey[26]; - br_aes_ct64_keysched(skey, key, 24); - br_aes_ct64_skey_expand(r->sk_exp, skey, 12); + br_aes_ct64_keysched(skey, key, 24); + br_aes_ct64_skey_expand(r->sk_exp, skey, 12); } -void aes256_keyexp(aes256ctx *r, const unsigned char *key) -{ - uint64_t skey[30]; +void aes256_keyexp(aes256ctx *r, const unsigned char *key) { + uint64_t skey[30]; - br_aes_ct64_keysched(skey, key, 32); - br_aes_ct64_skey_expand(r->sk_exp, skey, 14); + br_aes_ct64_keysched(skey, key, 32); + br_aes_ct64_skey_expand(r->sk_exp, skey, 14); } -void aes128_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes128ctx *ctx) -{ - aes_ecb(out, in, nblocks, ctx->sk_exp, 10); +void aes128_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes128ctx *ctx) { + aes_ecb(out, in, nblocks, ctx->sk_exp, 10); } -void aes128_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes128ctx *ctx) -{ - aes_ctr(out, outlen, iv, ctx->sk_exp, 10); +void aes128_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes128ctx *ctx) { + aes_ctr(out, outlen, iv, ctx->sk_exp, 10); } -void aes192_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes192ctx *ctx) -{ - aes_ecb(out, in, nblocks, ctx->sk_exp, 12); +void aes192_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes192ctx *ctx) { + aes_ecb(out, in, nblocks, ctx->sk_exp, 12); } -void aes192_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes192ctx *ctx) -{ - aes_ctr(out, outlen, iv, ctx->sk_exp, 12); +void aes192_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes192ctx *ctx) { + aes_ctr(out, outlen, iv, ctx->sk_exp, 12); } -void aes256_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes256ctx *ctx) -{ - aes_ecb(out, in, nblocks, ctx->sk_exp, 14); +void aes256_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes256ctx *ctx) { + aes_ecb(out, in, nblocks, ctx->sk_exp, 14); } -void aes256_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes256ctx *ctx) -{ - aes_ctr(out, outlen, iv, ctx->sk_exp, 14); +void aes256_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes256ctx *ctx) { + aes_ctr(out, outlen, iv, ctx->sk_exp, 14); } diff --git a/common/aes.h b/common/aes.h index 247fe0e3..b34ee06c 100644 --- a/common/aes.h +++ b/common/aes.h @@ -9,15 +9,15 @@ #define AESCTR_NONCEBYTES 12 typedef struct { - uint64_t sk_exp[88]; + uint64_t sk_exp[88]; } aes128ctx; typedef struct { - uint64_t sk_exp[104]; + uint64_t sk_exp[104]; } aes192ctx; typedef struct { - uint64_t sk_exp[120]; + uint64_t sk_exp[120]; } aes256ctx; From ed8a4ea2f4dbba57d9a8b0a86627bd388de0fef7 Mon Sep 17 00:00:00 2001 From: Peter Schwabe Date: Tue, 9 Apr 2019 14:57:52 +0200 Subject: [PATCH 03/14] Fixed wrong header name in aes.c --- common/aes.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/common/aes.c b/common/aes.c index bb58ada7..0c4fefb8 100644 --- a/common/aes.c +++ b/common/aes.c @@ -29,7 +29,7 @@ #include #include -#include "ctaes.h" +#include "aes.h" static inline uint32_t br_dec32le(const unsigned char *src) { return (uint32_t)src[0] @@ -249,8 +249,8 @@ static void br_aes_ct64_ortho(uint64_t *q) { uint64_t a, b; \ a = (x); \ b = (y); \ - (x) = (a & (uint64_t)cl) | ((b & (uint64_t)cl) << (s)); \ - (y) = ((a & (uint64_t)ch) >> (s)) | (b & (uint64_t)ch); \ + (x) = (a & (uint64_t)(cl)) | ((b & (uint64_t)(cl)) << (s)); \ + (y) = ((a & (uint64_t)(ch)) >> (s)) | (b & (uint64_t)(ch)); \ } while (0) #define SWAP2(x, y) SWAPN(0x5555555555555555, 0xAAAAAAAAAAAAAAAA, 1, x, y) From 40661601bd138ced6f92bd98e581be5a9a809a17 Mon Sep 17 00:00:00 2001 From: Peter Schwabe Date: Tue, 9 Apr 2019 15:06:23 +0200 Subject: [PATCH 04/14] Eliminated unnecessary casts --- common/aes.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/common/aes.c b/common/aes.c index 0c4fefb8..c6ff6cd6 100644 --- a/common/aes.c +++ b/common/aes.c @@ -338,14 +338,14 @@ static uint32_t sub_word(uint32_t x) { return (uint32_t)q[0]; } -static void br_aes_ct64_keysched(uint64_t *comp_skey, const unsigned char *key, int key_len) { - int i, j, k, nk, nkf; +static void br_aes_ct64_keysched(uint64_t *comp_skey, const unsigned char *key, unsigned int key_len) { + unsigned int i, j, k, nk, nkf; uint32_t tmp; uint32_t skey[60]; unsigned nrounds = 10 + ((key_len - 16) >> 2); - nk = (int)(key_len >> 2); - nkf = (int)((nrounds + 1) << 2); + nk = (key_len >> 2); + nkf = ((nrounds + 1) << 2); br_range_dec32le(skey, (key_len >> 2), key); tmp = skey[(key_len >> 2) - 1]; for (i = nk, j = 0, k = 0; i < nkf; i ++) { From 45a12595df751dcd2da4582115e628e9d91ed13a Mon Sep 17 00:00:00 2001 From: Peter Schwabe Date: Tue, 9 Apr 2019 15:41:44 +0200 Subject: [PATCH 05/14] Added const keyword to a few pointer arguments in AES --- common/aes.c | 8 ++++---- common/aes.h | 12 ++++++------ 2 files changed, 10 insertions(+), 10 deletions(-) diff --git a/common/aes.c b/common/aes.c index c6ff6cd6..718f834f 100644 --- a/common/aes.c +++ b/common/aes.c @@ -522,7 +522,7 @@ static void aes_ctr4x(unsigned char out[64], uint32_t ivw[16], const uint64_t *s } -static void aes_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const uint64_t *rkeys, unsigned int nrounds) { +static void aes_ecb(unsigned char *out, const unsigned char *in, size_t nblocks, const uint64_t *rkeys, unsigned int nrounds) { uint32_t blocks[16]; unsigned char t[64]; @@ -598,7 +598,7 @@ void aes256_keyexp(aes256ctx *r, const unsigned char *key) { } -void aes128_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes128ctx *ctx) { +void aes128_ecb(unsigned char *out, const unsigned char *in, size_t nblocks, const aes128ctx *ctx) { aes_ecb(out, in, nblocks, ctx->sk_exp, 10); } @@ -606,7 +606,7 @@ void aes128_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, cons aes_ctr(out, outlen, iv, ctx->sk_exp, 10); } -void aes192_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes192ctx *ctx) { +void aes192_ecb(unsigned char *out, const unsigned char *in, size_t nblocks, const aes192ctx *ctx) { aes_ecb(out, in, nblocks, ctx->sk_exp, 12); } @@ -614,7 +614,7 @@ void aes192_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, cons aes_ctr(out, outlen, iv, ctx->sk_exp, 12); } -void aes256_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes256ctx *ctx) { +void aes256_ecb(unsigned char *out, const unsigned char *in, size_t nblocks, const aes256ctx *ctx) { aes_ecb(out, in, nblocks, ctx->sk_exp, 14); } diff --git a/common/aes.h b/common/aes.h index b34ee06c..6b3fca3d 100644 --- a/common/aes.h +++ b/common/aes.h @@ -24,23 +24,23 @@ typedef struct { void aes128_keyexp(aes128ctx *r, const unsigned char *key); -void aes128_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes128ctx *ctx); +void aes128_ecb(unsigned char *out, const unsigned char *in, size_t nblocks, const aes128ctx *ctx); -void aes128_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes128ctx *ctx); +void aes128_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes128ctx *ctx); void aes192_keyexp(aes192ctx *r, const unsigned char *key); -void aes192_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes192ctx *ctx); +void aes192_ecb(unsigned char *out, const unsigned char *in, size_t nblocks, const aes192ctx *ctx); -void aes192_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes192ctx *ctx); +void aes192_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes192ctx *ctx); void aes256_keyexp(aes256ctx *r, const unsigned char *key); -void aes256_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes256ctx *ctx); +void aes256_ecb(unsigned char *out, const unsigned char *in, size_t nblocks, const aes256ctx *ctx); -void aes256_ecb(unsigned char *out, unsigned char *in, size_t nblocks, const aes256ctx *ctx); +void aes256_ctr(unsigned char *out, size_t outlen, const unsigned char *iv, const aes256ctx *ctx); #endif From 4970379bafa6628e979232604dc32c144333daaf Mon Sep 17 00:00:00 2001 From: Peter Schwabe Date: Wed, 10 Apr 2019 13:24:42 +0200 Subject: [PATCH 06/14] Added simple test for AES --- common/aes.h | 1 + test/common/aes.c | 93 +++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 94 insertions(+) create mode 100644 test/common/aes.c diff --git a/common/aes.h b/common/aes.h index 6b3fca3d..dd5a1019 100644 --- a/common/aes.h +++ b/common/aes.h @@ -2,6 +2,7 @@ #define AES_H #include +#include #define AES128_KEYBYTES 16 #define AES192_KEYBYTES 24 diff --git a/test/common/aes.c b/test/common/aes.c new file mode 100644 index 00000000..b7fc4cb1 --- /dev/null +++ b/test/common/aes.c @@ -0,0 +1,93 @@ +#include "aes.h" + +#include +#include + +const unsigned char msg[48] = "The quick brown fox jumps over the lazy dog!!!!"; + +const unsigned char stream128[67] = { + 0x36, 0x88, 0x7b, 0x28, 0x99, 0x8f, 0x4d, 0x2b, 0x37, 0xff, 0x06, 0x63, 0xfc, 0x5c, 0xef, 0x2f, + 0x43, 0xeb, 0xeb, 0x7e, 0xc1, 0x58, 0xe4, 0xb4, 0x27, 0x78, 0x2e, 0xa7, 0x90, 0x00, 0x09, 0x91, + 0xcd, 0x0e, 0x40, 0x18, 0x8b, 0x2b, 0x52, 0x1e, 0x8d, 0xfb, 0x0e, 0x7b, 0x80, 0xe7, 0xb6, 0xd4, + 0xba, 0x48, 0xf5, 0x19, 0xb4, 0xbf, 0xbf, 0xb1, 0x52, 0x6f, 0x12, 0xe2, 0x7b, 0x61, 0x90, 0x0d, + 0x95, 0x83, 0x84}; + +const unsigned char stream192[67] = { + 0xb0, 0x13, 0x08, 0x01, 0xfa, 0x88, 0x28, 0x98, 0xac, 0x77, 0x72, 0xaf, 0x2a, 0x8d, 0x17, 0xf3, + 0xba, 0x37, 0xd6, 0x8e, 0x01, 0x5b, 0x25, 0xb5, 0x51, 0x40, 0xf9, 0x55, 0x73, 0x89, 0xbe, 0xd3, + 0x5f, 0x3c, 0xde, 0xc2, 0x6b, 0xf6, 0xfc, 0x26, 0x97, 0x22, 0xfa, 0x8e, 0xcb, 0x62, 0xe8, 0xa5, + 0x03, 0xd2, 0x8a, 0x7e, 0xaf, 0xe3, 0xe5, 0x63, 0x2e, 0xb8, 0x48, 0x20, 0x1c, 0x48, 0xd5, 0xc2, + 0x27, 0x1e, 0x1e}; + +const unsigned char stream256[67] = { + 0x12, 0x39, 0x28, 0xd0, 0xda, 0xd1, 0xfd, 0xe7, 0x64, 0x74, 0x10, 0x5a, 0x29, 0x17, 0x3c, 0x62, + 0x05, 0xde, 0x4c, 0x98, 0x22, 0x9b, 0xad, 0x16, 0x32, 0x75, 0xbf, 0x26, 0xe5, 0x84, 0x7e, 0xc8, + 0x4e, 0x73, 0x68, 0xce, 0x9a, 0x11, 0xb6, 0x55, 0x53, 0x05, 0x39, 0xa1, 0xa7, 0x1f, 0x16, 0x55, + 0x4a, 0xd3, 0x6c, 0xc6, 0x2c, 0xb4, 0x55, 0x9f, 0x5f, 0xa3, 0xe8, 0x39, 0xfa, 0x9d, 0x96, 0xb6, + 0xb7, 0xc9, 0xc5}; + +const unsigned char ct128[48] = { +0x10, 0xdc, 0x43, 0x2b, 0x15, 0x11, 0x81, 0x36, 0x3f, 0x00, 0x51, 0x74, 0x81, 0x7c, 0x22, 0x87, +0x3a, 0x3b, 0xfe, 0xd7, 0xb9, 0xa6, 0xf2, 0x3c, 0x81, 0x00, 0x63, 0xef, 0xe5, 0xb8, 0xbd, 0x36, +0x11, 0xcc, 0xc9, 0xdf, 0x2b, 0xea, 0xbc, 0xe6, 0x11, 0x1c, 0x34, 0x79, 0xf9, 0x6b, 0x47, 0x7b}; + +const unsigned char ct192[48] = { +0x63, 0xc6, 0xde, 0x28, 0x36, 0xb4, 0x29, 0xbf, 0xbe, 0x9d, 0x15, 0x8e, 0x83, 0x04, 0xa3, 0x18, +0x34, 0x79, 0xe8, 0x02, 0x8a, 0x34, 0x50, 0x7c, 0xa9, 0x08, 0x48, 0x47, 0xee, 0x90, 0x79, 0x13, +0x66, 0x2d, 0xa4, 0xf1, 0x3e, 0x8b, 0x76, 0xa9, 0x50, 0xf7, 0x6e, 0xa8, 0xbf, 0x29, 0xaf, 0x84}; + +const unsigned char ct256[48] = { +0xdb, 0xb5, 0x44, 0x70, 0x68, 0xe6, 0xad, 0x6a, 0x09, 0xdf, 0xa6, 0xef, 0x85, 0x73, 0xff, 0xc0, +0xc2, 0x91, 0x38, 0xbd, 0xd7, 0xd0, 0x22, 0x7e, 0x79, 0x71, 0xa1, 0x98, 0x6f, 0xd5, 0x80, 0xa8, +0x1e, 0x97, 0xd7, 0x6d, 0xd2, 0x6b, 0x0e, 0x7b, 0x79, 0x76, 0x75, 0x86, 0xa5, 0x2f, 0x76, 0x0b}; + +const unsigned char key[AES256_KEYBYTES] = { + 0x66, 0xd9, 0xb7, 0x60, 0x0e, 0xda, 0xaa, 0x81, 0x42, 0xa2, 0xd6, 0x3d, 0x8f, 0x51, 0x6c, 0x6f, + 0xb6, 0xdf, 0x5b, 0x97, 0xf3, 0xf1, 0xf7, 0x0e, 0xeb, 0xe0, 0x40, 0x4d, 0xc5, 0x24, 0xa1, 0xfa}; +const unsigned char nonce[AESCTR_NONCEBYTES] = { + 0x9d, 0x2d, 0x3e, 0x6e, 0x48, 0x5c, 0xf6, 0x6b, 0xb2, 0xb9, 0x25, 0xf4}; + +int main(void) +{ + unsigned char ct[67]; + aes128ctx ctx128; + aes192ctx ctx192; + aes256ctx ctx256; + + aes128_keyexp(&ctx128, key); + aes192_keyexp(&ctx192, key); + aes256_keyexp(&ctx256, key); + + aes128_ctr(ct, 67, nonce, &ctx128); + if(memcmp(ct, stream128, 67)) { + printf("ERROR AES128CTR output does not match test vector.\n"); + } + + aes192_ctr(ct, 67, nonce, &ctx192); + if(memcmp(ct, stream192, 67)) { + printf("ERROR AES192CTR output does not match test vector.\n"); + } + + aes256_ctr(ct, 67, nonce, &ctx256); + if(memcmp(ct, stream256, 67)) { + printf("ERROR AES256CTR output does not match test vector.\n"); + } + + + aes128_ecb(ct, msg, 3, &ctx128); + if(memcmp(ct, ct128, 48)) { + printf("ERROR AES128ECB output does not match test vector.\n"); + } + + aes192_ecb(ct, msg, 3, &ctx192); + if(memcmp(ct, ct192, 48)) { + printf("ERROR AES192ECB output does not match test vector.\n"); + } + + aes256_ecb(ct, msg, 3, &ctx256); + if(memcmp(ct, ct256, 48)) { + printf("ERROR AES256ECB output does not match test vector.\n"); + } + + return 0; +} From 28337843b925af0c8867d4d72e29485333f6281c Mon Sep 17 00:00:00 2001 From: Peter Schwabe Date: Wed, 10 Apr 2019 16:02:45 +0200 Subject: [PATCH 07/14] Added python test running all unit tests of common crypto functions --- test/Makefile | 10 +++------- test/common/aes.c | 9 ++++++++- test/test_common.py | 29 +++++++++++++++++++++++++++++ 3 files changed, 40 insertions(+), 8 deletions(-) create mode 100644 test/test_common.py diff --git a/test/Makefile b/test/Makefile index 8b938246..51da5b8b 100644 --- a/test/Makefile +++ b/test/Makefile @@ -10,8 +10,8 @@ SCHEME_UPPERCASE=$(shell echo $(SCHEME) | tr a-z A-Z | sed 's/-//g') IMPLEMENTATION_UPPERCASE=$(shell echo $(IMPLEMENTATION) | tr a-z A-Z | sed 's/-//g') COMMON_DIR=../common -COMMON_FILES=$(COMMON_DIR)/fips202.c $(COMMON_DIR)/sha2.c -COMMON_HEADERS=$(COMMON_DIR)/fips202.h $(COMMON_DIR)/randombytes.h $(COMMON_DIR)/sha2.h +COMMON_FILES=$(COMMON_DIR)/aes.c $(COMMON_DIR)/sha2.c $(COMMON_DIR)/fips202.c +COMMON_HEADERS=$(COMMON_DIR)/*.h DEST_DIR=../bin # This -Wall was supported by the European Commission through the ERC Starting Grant 805031 (EPOQUE) @@ -38,11 +38,7 @@ testvectors: $(DEST_DIR)/testvectors_$(SCHEME)_$(IMPLEMENTATION) .PHONY: printparams printparams: $(DEST_DIR)/printparams_$(SCHEME)_$(IMPLEMENTATION) -$(DEST_DIR)/test_fips202: common/fips202.c $(COMMON_FILES) - mkdir -p $(DEST_DIR) - $(CC) $(CFLAGS) $< $(COMMON_FILES) -o $@ - -$(DEST_DIR)/test_sha2: common/sha2.c $(COMMON_FILES) +$(DEST_DIR)/test_%: common/%.c $(COMMON_FILES) mkdir -p $(DEST_DIR) $(CC) $(CFLAGS) $< $(COMMON_FILES) -o $@ diff --git a/test/common/aes.c b/test/common/aes.c index b7fc4cb1..dcfe8029 100644 --- a/test/common/aes.c +++ b/test/common/aes.c @@ -50,6 +50,7 @@ const unsigned char nonce[AESCTR_NONCEBYTES] = { int main(void) { unsigned char ct[67]; + int r = 0; aes128ctx ctx128; aes192ctx ctx192; aes256ctx ctx256; @@ -61,33 +62,39 @@ int main(void) aes128_ctr(ct, 67, nonce, &ctx128); if(memcmp(ct, stream128, 67)) { printf("ERROR AES128CTR output does not match test vector.\n"); + r = 1; } aes192_ctr(ct, 67, nonce, &ctx192); if(memcmp(ct, stream192, 67)) { printf("ERROR AES192CTR output does not match test vector.\n"); + r = 1; } aes256_ctr(ct, 67, nonce, &ctx256); if(memcmp(ct, stream256, 67)) { printf("ERROR AES256CTR output does not match test vector.\n"); + r = 1; } aes128_ecb(ct, msg, 3, &ctx128); if(memcmp(ct, ct128, 48)) { printf("ERROR AES128ECB output does not match test vector.\n"); + r = 1; } aes192_ecb(ct, msg, 3, &ctx192); if(memcmp(ct, ct192, 48)) { printf("ERROR AES192ECB output does not match test vector.\n"); + r = 1; } aes256_ecb(ct, msg, 3, &ctx256); if(memcmp(ct, ct256, 48)) { printf("ERROR AES256ECB output does not match test vector.\n"); + r = 1; } - return 0; + return r; } diff --git a/test/test_common.py b/test/test_common.py new file mode 100644 index 00000000..65dc50a6 --- /dev/null +++ b/test/test_common.py @@ -0,0 +1,29 @@ +""" +Runs functional tests for common crypto functions (e.g., fips202, sha2, aes). +""" + +import os +import re +import platform +import unittest + +import pqclean +import helpers + + +def test_common(): + if os.name != 'nt': # Cannot build on Windows at the moment + for d in os.listdir('common'): + primitive = re.sub("\.c$", "", d) + binname = os.path.join('..', 'bin', 'test_'+primitive) + helpers.make(binname) + helpers.run_subprocess(binname) + + +if __name__ == '__main__': + try: + import nose2 + nose2.main() + except ImportError: + import nose + nose.runmodule() From 280bc60b1b1624d9afc54d333555e9e5869a7ffc Mon Sep 17 00:00:00 2001 From: Joost Rijneveld Date: Wed, 10 Apr 2019 16:34:44 +0200 Subject: [PATCH 08/14] Use standard helper to skip Windows for /common --- test/test_common.py | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/test/test_common.py b/test/test_common.py index 65dc50a6..b303df5d 100644 --- a/test/test_common.py +++ b/test/test_common.py @@ -11,13 +11,13 @@ import pqclean import helpers +@helpers.skip_windows() def test_common(): - if os.name != 'nt': # Cannot build on Windows at the moment - for d in os.listdir('common'): - primitive = re.sub("\.c$", "", d) - binname = os.path.join('..', 'bin', 'test_'+primitive) - helpers.make(binname) - helpers.run_subprocess(binname) + for d in os.listdir('common'): + primitive = re.sub(r"\.c$", "", d) + binname = os.path.join('..', 'bin', 'test_'+primitive) + helpers.make(binname) + helpers.run_subprocess(binname) if __name__ == '__main__': From 03eb577a40c7011cea11b8e731cac8d52ea7b403 Mon Sep 17 00:00:00 2001 From: Joost Rijneveld Date: Wed, 10 Apr 2019 16:37:02 +0200 Subject: [PATCH 09/14] Pass list instead of single command to subprocess --- test/test_common.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/test/test_common.py b/test/test_common.py index b303df5d..d57ad4a5 100644 --- a/test/test_common.py +++ b/test/test_common.py @@ -17,7 +17,7 @@ def test_common(): primitive = re.sub(r"\.c$", "", d) binname = os.path.join('..', 'bin', 'test_'+primitive) helpers.make(binname) - helpers.run_subprocess(binname) + helpers.run_subprocess([binname]) if __name__ == '__main__': From f4607265c6c8299b51c7b365f7b7a0f3813e1e9d Mon Sep 17 00:00:00 2001 From: Joost Rijneveld Date: Wed, 10 Apr 2019 16:37:59 +0200 Subject: [PATCH 10/14] Remove unused imports --- test/test_common.py | 3 --- 1 file changed, 3 deletions(-) diff --git a/test/test_common.py b/test/test_common.py index d57ad4a5..0a96986b 100644 --- a/test/test_common.py +++ b/test/test_common.py @@ -4,10 +4,7 @@ Runs functional tests for common crypto functions (e.g., fips202, sha2, aes). import os import re -import platform -import unittest -import pqclean import helpers From 5d600a25436525aecd70252f247393978ae74e32 Mon Sep 17 00:00:00 2001 From: Joost Rijneveld Date: Wed, 10 Apr 2019 16:42:28 +0200 Subject: [PATCH 11/14] Split tests into primitives This ensures separate errors for separate common files --- test/test_common.py | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/test/test_common.py b/test/test_common.py index 0a96986b..ec7a593f 100644 --- a/test/test_common.py +++ b/test/test_common.py @@ -12,9 +12,13 @@ import helpers def test_common(): for d in os.listdir('common'): primitive = re.sub(r"\.c$", "", d) - binname = os.path.join('..', 'bin', 'test_'+primitive) - helpers.make(binname) - helpers.run_subprocess([binname]) + yield check_common, primitive + + +def check_common(primitive): + binname = os.path.join('..', 'bin', 'test_'+primitive) + helpers.make(binname) + helpers.run_subprocess([binname]) if __name__ == '__main__': From e766700a81f08cbe946494f48e4b89a15472f029 Mon Sep 17 00:00:00 2001 From: Joost Rijneveld Date: Wed, 10 Apr 2019 17:43:04 +0200 Subject: [PATCH 12/14] Add test_aes to `make clean` --- test/Makefile | 1 + 1 file changed, 1 insertion(+) diff --git a/test/Makefile b/test/Makefile index 51da5b8b..167d9dbc 100644 --- a/test/Makefile +++ b/test/Makefile @@ -58,5 +58,6 @@ $(DEST_DIR)/printparams_$(SCHEME)_$(IMPLEMENTATION): build-scheme crypto_$(TYPE) clean: $(RM) $(DEST_DIR)/functest_$(SCHEME)_$(IMPLEMENTATION) $(RM) $(DEST_DIR)/testvectors_$(SCHEME)_$(IMPLEMENTATION) + $(RM) $(DEST_DIR)/test_aes $(RM) $(DEST_DIR)/test_fips202 $(RM) $(DEST_DIR)/test_sha2 From 69b74706c192452491aa4be6153ba0e0fc437ee0 Mon Sep 17 00:00:00 2001 From: Joost Rijneveld Date: Wed, 10 Apr 2019 17:54:52 +0200 Subject: [PATCH 13/14] Fix SHA3 test --- test/common/fips202.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/test/common/fips202.c b/test/common/fips202.c index 7c28672f..e7d017d8 100644 --- a/test/common/fips202.c +++ b/test/common/fips202.c @@ -54,8 +54,8 @@ const unsigned char expected[512] = { static int test_sha3_256_incremental(void) { unsigned char input[512]; - unsigned char check[512]; - unsigned char output[512]; + unsigned char check[32]; + unsigned char output[32]; uint64_t s_inc[26]; int i; int absorbed; @@ -78,15 +78,15 @@ static int test_sha3_256_incremental(void) { sha3_256_inc_finalize(output, s_inc); - if (memcmp(check, output, 512)) { + if (memcmp(check, output, 32)) { printf("ERROR sha3_256 incremental did not match sha3_256.\n"); printf(" Expected: "); - for (i = 0; i < 512; i++) { + for (i = 0; i < 32; i++) { printf("%02X", check[i]); } printf("\n"); printf(" Received: "); - for (i = 0; i < 512; i++) { + for (i = 0; i < 32; i++) { printf("%02X", output[i]); } printf("\n"); From a7c147f4101c6617d5ad9544ea2c303e904c22d2 Mon Sep 17 00:00:00 2001 From: Joost Rijneveld Date: Thu, 11 Apr 2019 09:54:52 +0200 Subject: [PATCH 14/14] Include 'common' in name of common test binaries --- test/Makefile | 2 +- test/test_common.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/test/Makefile b/test/Makefile index 167d9dbc..bb1ea5ce 100644 --- a/test/Makefile +++ b/test/Makefile @@ -38,7 +38,7 @@ testvectors: $(DEST_DIR)/testvectors_$(SCHEME)_$(IMPLEMENTATION) .PHONY: printparams printparams: $(DEST_DIR)/printparams_$(SCHEME)_$(IMPLEMENTATION) -$(DEST_DIR)/test_%: common/%.c $(COMMON_FILES) +$(DEST_DIR)/test_common_%: common/%.c $(COMMON_FILES) mkdir -p $(DEST_DIR) $(CC) $(CFLAGS) $< $(COMMON_FILES) -o $@ diff --git a/test/test_common.py b/test/test_common.py index ec7a593f..3640bb3f 100644 --- a/test/test_common.py +++ b/test/test_common.py @@ -16,7 +16,7 @@ def test_common(): def check_common(primitive): - binname = os.path.join('..', 'bin', 'test_'+primitive) + binname = os.path.join('..', 'bin', 'test_common_'+primitive) helpers.make(binname) helpers.run_subprocess([binname])