tk_schedule.c 16.7 KB
Newer Older
Alexandre Adomnicai committed
1 2 3 4 5 6
/******************************************************************************
* Implementation of the SKINNY tweakey schedule to match fixslicing.
* 
* @author	Alexandre Adomnicai, Nanyang Technological University,
*			alexandre.adomnicai@ntu.edu.sg
*
Alexandre Adomnicai committed
7
* @date		June 2020
Alexandre Adomnicai committed
8
******************************************************************************/
Alexandre Adomnicai committed
9
#include <string.h>
Alexandre Adomnicai committed
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268
#include "skinny128.h"

/******************************************************************************
* The round constants according to the new representation.
******************************************************************************/
u32 rconst_32_bs[224] = {
	0x00000004, 0xffffffbf, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
	0x10000100, 0xfffffeff, 0x44000000, 0xfbffffff, 0x00000000, 0x04000000,
	0x00100000, 0x00100000, 0x00100001, 0xffefffff, 0x00440000, 0xffafffff,
	0x00400000, 0x00400000, 0x01000000, 0x01000000, 0x01401000, 0xffbfffff,
	0x01004000, 0xfefffbff, 0x00000400, 0x00000400, 0x00000010, 0x00000000,
	0x00010410, 0xfffffbef, 0x00000054, 0xffffffaf, 0x00000000, 0x00000040,
	0x00000100, 0x00000100, 0x10000140, 0xfffffeff, 0x44000000, 0xfffffeff,
	0x04000000, 0x04000000, 0x00100000, 0x00100000, 0x04000001, 0xfbffffff,
	0x00140000, 0xffafffff, 0x00400000, 0x00000000, 0x00000000, 0x00000000,
	0x01401000, 0xfebfffff, 0x01004400, 0xfffffbff, 0x00000000, 0x00000400,
	0x00000010, 0x00000010, 0x00010010, 0xffffffff, 0x00000004, 0xffffffaf,
	0x00000040, 0x00000040, 0x00000100, 0x00000000, 0x10000140, 0xffffffbf,
	0x40000100, 0xfbfffeff, 0x00000000, 0x04000000, 0x00100000, 0x00000000,
	0x04100001, 0xffefffff, 0x00440000, 0xffefffff, 0x00000000, 0x00400000,
	0x01000000, 0x01000000, 0x00401000, 0xffffffff, 0x00004000, 0xfeffffff, 
	0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00010400, 0xfffffbff,
	0x00000014, 0xffffffbf, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
	0x10000100, 0xffffffff, 0x40000000, 0xfbffffff, 0x00000000, 0x04000000,
	0x00100000, 0x00000000, 0x00100001, 0xffefffff, 0x00440000, 0xffafffff,
	0x00000000, 0x00400000, 0x01000000, 0x01000000, 0x01401000, 0xffffffff,
	0x00004000, 0xfeffffff, 0x00000400, 0x00000400, 0x00000010, 0x00000000,
	0x00010400, 0xfffffbff, 0x00000014, 0xffffffaf, 0x00000000, 0x00000000,
	0x00000000, 0x00000000, 0x10000140, 0xfffffeff, 0x44000000, 0xffffffff,
	0x00000000, 0x04000000, 0x00100000, 0x00100000, 0x00000001, 0xffefffff,
	0x00440000, 0xffafffff, 0x00400000, 0x00000000, 0x00000000, 0x01000000,
	0x01401000, 0xffbfffff, 0x01004000, 0xfffffbff, 0x00000400, 0x00000400,
	0x00000010, 0x00000000, 0x00010010, 0xfffffbff, 0x00000014, 0xffffffef,
	0x00000000, 0x00000040, 0x00000100, 0x00000000, 0x10000040, 0xfffffeff,
	0x44000000, 0xfffffeff, 0x00000000, 0x00000000, 0x00000000, 0x00100000,
	0x04000001, 0xffffffff, 0x00040000, 0xffffffff, 0x00400000, 0x00000000,
	0x00000000, 0x00000000, 0x00001000, 0xfebfffff, 0x01004400, 0xffffffff,
	0x00000000, 0x00000000, 0x00000000, 0x00000010, 0x00010000, 0xffffffff,
	0x00000004, 0xffffffbf, 0x00000040, 0x00000000, 0x00000000, 0x00000000,
	0x10000100, 0xfffffebf, 0x44000100, 0xffffffff, 0x00000000, 0x04000000,
	0x00100000, 0x00100000, 0x00000001, 0xffffffff, 0x00040000, 0xffafffff,
	0x00400000, 0x00000000, 0x00000000, 0x00000000, 0x01401000, 0xffbfffff,
	0x01004000, 0xfffffbff, 0x00000000, 0x00000400, 0x00000010, 0x00000000,
	0x00010010, 0xffffffff
};

/******************************************************************************
* 	Pack the input into the bitsliced representation
* 	24 28 56 60 88 92 120 124 | ... | 0 4 32 36 64 68 96 100
* 	25 29 57 61 89 93 121 125 | ... | 1 5 33 37 65 69 97 101
* 	26 30 58 62 90 94 122 126 | ... | 2 6 34 38 66 70 98 102
* 	27 31 59 63 91 95 123 127 | ... | 3 7 35 39 67 71 99 103
******************************************************************************/
void packing(u32* out, const u8* in) {
	u32 tmp;
	LE_LOAD(out, in);
	LE_LOAD(out + 1, in + 8);
	LE_LOAD(out + 2, in + 4);
	LE_LOAD(out + 3, in + 12);
	SWAPMOVE(out[0], out[0], 0x0a0a0a0a, 3);
	SWAPMOVE(out[1], out[1], 0x0a0a0a0a, 3);
	SWAPMOVE(out[2], out[2], 0x0a0a0a0a, 3);
	SWAPMOVE(out[3], out[3], 0x0a0a0a0a, 3);
	SWAPMOVE(out[2], out[0], 0x30303030, 2);
	SWAPMOVE(out[1], out[0], 0x0c0c0c0c, 4);
	SWAPMOVE(out[3], out[0], 0x03030303, 6);
	SWAPMOVE(out[1], out[2], 0x0c0c0c0c, 2);
	SWAPMOVE(out[3], out[2], 0x03030303, 4);
	SWAPMOVE(out[3], out[1], 0x03030303, 2);
}

/******************************************************************************
* Unpack the input to a byte-wise representation
******************************************************************************/
void unpacking(u8* out, u32 *in) {
	u32 tmp;
	SWAPMOVE(in[3], in[1], 0x03030303, 2);
	SWAPMOVE(in[3], in[2], 0x03030303, 4);
	SWAPMOVE(in[1], in[2], 0x0c0c0c0c, 2);
	SWAPMOVE(in[3], in[0], 0x03030303, 6);
	SWAPMOVE(in[1], in[0], 0x0c0c0c0c, 4);
	SWAPMOVE(in[2], in[0], 0x30303030, 2);
	SWAPMOVE(in[0], in[0], 0x0a0a0a0a, 3);
	SWAPMOVE(in[1], in[1], 0x0a0a0a0a, 3);
	SWAPMOVE(in[2], in[2], 0x0a0a0a0a, 3);
	SWAPMOVE(in[3], in[3], 0x0a0a0a0a, 3);
	LE_STORE(out, in[0]);
	LE_STORE(out + 8, in[1]);
	LE_STORE(out + 4, in[2]);
	LE_STORE(out + 12, in[3]);
}

/******************************************************************************
* 	0 4        1 5
* 	1 5  --->  2 6
* 	2 6        3 7
* 	3 7        4 0
******************************************************************************/
void lfsr2_bs(u32* tk) {
	u32 tmp;
	tmp = tk[0] ^ (tk[2] & 0xaaaaaaaa);
	tmp = ((tmp & 0xaaaaaaaa) >> 1) | ((tmp << 1) & 0xaaaaaaaa);
	tk[0] = tk[1];
	tk[1] = tk[2];
	tk[2] = tk[3];
	tk[3] = tmp;
}

/******************************************************************************
* 	0 4        7 3
* 	1 5  --->  0 4
* 	2 6        1 5
* 	3 7        2 6
******************************************************************************/
void lfsr3_bs(u32* tk) {
	u32 tmp;
	tmp = tk[3] ^ ((tk[1] & 0xaaaaaaaa) >> 1);
	tmp = ((tmp & 0xaaaaaaaa) >> 1) | ((tmp << 1) & 0xaaaaaaaa);
	tk[3] = tk[2];
	tk[2] = tk[1];
	tk[1] = tk[0];
	tk[0] = tmp;
}

/******************************************************************************
* Apply the permutation in a bitsliced manner, twice
******************************************************************************/
void permute_tk_2(u32* tk) {
	u32 tmp;
	for(int i =0; i < 4; i++) {
		tmp = tk[i];
		tk[i] = ROR(tmp,14) & 0xcc00cc00;
		tk[i] |= (tmp & 0x000000ff) << 16;
		tk[i] |= (tmp & 0xcc000000)>> 2;
		tk[i] |= (tmp & 0x0033cc00) >> 8;
		tk[i] |= (tmp & 0x00cc0000) >>18;
	}
}

/******************************************************************************
* Apply the permutation in a bitsliced manner, 4 times
******************************************************************************/
void permute_tk_4(u32* tk) {
	u32 tmp;
	for(int i =0; i < 4; i++) {
		tmp = tk[i];
		tk[i] = ROR(tmp,22) & 0xcc0000cc;
		tk[i] |= ROR(tmp,16) & 0x3300cc00;
		tk[i] |= ROR(tmp, 24) & 0x00cc3300;
		tk[i] |= (tmp & 0x00cc00cc) >> 2;
	}
}

/******************************************************************************
* Apply the permutation in a bitsliced manner, 6 times
******************************************************************************/
void permute_tk_6(u32* tk) {
	u32 tmp;
	for(int i =0; i < 4; i++) {
		tmp = tk[i];
		tk[i] = ROR(tmp,6) & 0xcccc0000;
		tk[i] |= ROR(tmp,24) & 0x330000cc;
		tk[i] |= ROR(tmp,10) & 0x3333;
		tk[i] |= (tmp & 0xcc) << 14;
		tk[i] |= (tmp & 0x3300) << 2;
	}
}

/******************************************************************************
* Apply the permutation in a bitsliced manner, 8 times
******************************************************************************/
void permute_tk_8(u32* tk) {
	u32 tmp;
	for(int i =0; i < 4; i++) {
		tmp = tk[i];
		tk[i] = ROR(tmp,24) & 0xcc000033;
		tk[i] |= ROR(tmp,8) & 0x33cc0000;
		tk[i] |= ROR(tmp,26) & 0x00333300;
		tk[i] |= (tmp & 0x00333300) >> 6;
	}
}

/******************************************************************************
* Apply the permutation in a bitsliced manner, 10 times
******************************************************************************/
void permute_tk_10(u32* tk) {
	u32 tmp;
	for(int i =0; i < 4; i++) {
		tmp = tk[i];
		tk[i] = ROR(tmp,8) & 0xcc330000;
		tk[i] |= ROR(tmp,26) & 0x33000033;
		tk[i] |= ROR(tmp,22) & 0x00cccc00;
		tk[i] |= (tmp & 0x00330000) >> 14;
		tk[i] |= (tmp & 0xcc00) >> 2;
	}
}

/******************************************************************************
* Apply the permutation in a bitsliced manner, 12 times
******************************************************************************/
void permute_tk_12(u32* tk) {
	u32 tmp;
	for(int i =0; i < 4; i++) {
		tmp = tk[i];
		tk[i] = ROR(tmp,8) & 0xcc33;
		tk[i] |= ROR(tmp,30) & 0x00cc00cc;
		tk[i] |= ROR(tmp,10) & 0x33330000;
		tk[i] |= ROR(tmp,16) & 0xcc003300;
	}
}

/******************************************************************************
* Apply the permutation in a bitsliced manner, 14 times
******************************************************************************/
void permute_tk_14(u32* tk) {
	u32 tmp;
	for(int i =0; i < 4; i++) {
		tmp = tk[i];
		tk[i] = ROR(tmp,24) & 0x0033cc00;
		tk[i] |= ROR(tmp,14) & 0x00cc0000;
		tk[i] |= ROR(tmp,30) & 0xcc000000;
		tk[i] |= ROR(tmp,16) & 0x000000ff;
		tk[i] |= ROR(tmp,18) & 0x33003300;
	}
}

/******************************************************************************
* Precompute all LFSRs on TK2
******************************************************************************/
void precompute_lfsr_tk2(u32* tk, const u8* key, const int rounds) {
	u32 tk2[4];
	packing(tk2, key);
	memcpy(tk, tk2, 16);
	for(int i = 0 ; i < rounds; i+=2) {
		lfsr2_bs(tk2);
		memcpy(tk+i*4+4, tk2, 16);
	}
}

/******************************************************************************
* Precompute all LFSRs on TK3
******************************************************************************/
void precompute_lfsr_tk3(u32* tk, const u8* key, const int rounds) {
	u32 tk3[4];
	packing(tk3, key);
	tk[0] ^= tk3[0];
	tk[1] ^= tk3[1];
	tk[2] ^= tk3[2];
	tk[3] ^= tk3[3];
	for(int i = 0 ; i < rounds; i+=2) {
		lfsr3_bs(tk3);
		tk[i*4+4] ^= tk3[0];
		tk[i*4+5] ^= tk3[1];
		tk[i*4+6] ^= tk3[2];
		tk[i*4+7] ^= tk3[3];
	}
}

/******************************************************************************
Alexandre Adomnicai committed
269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
* Precompute LFSR2(TK2) ^ LFSR3(TK3) for all round tweakeys.
* It is equivalent to the following 2 function calls:
* - precompute_lfsr_tk2(tk, t2, SKINNY128_384_ROUNDS);
* - precompute_lfsr_tk3(tk, t3, SKINNY128_384_ROUNDS);
* However 'precompute_lfsr_tk2_3' can allow to save cycles on some platform.
* On ARMv7 one should observe a gain of ~1k cycles per function call. It can be
* explained by the fact that less memory accesses to 'tk' are computed.
* 
* To save some code size, the loop can be replaced by the following one:
*	for(int i = 0 ; i < rounds; i+=2) {
*		lfsr2_bs(tk2);
*		lfsr3_bs(tk3);
*		tk[i*4+4] = tk2[0] ^ tk3[0];
*		tk[i*4+5] = tk2[1] ^ tk3[1];
*		tk[i*4+6] = tk2[2] ^ tk3[2];
*		tk[i*4+7] = tk2[3] ^ tk3[3];
*	}
* at the cost of some cycles (~260 on ARM Cortex-M).
******************************************************************************/
void precompute_lfsr_tk2_3(u32* tk, const u8* t2, const u8* t3, const int rounds) {
	u32 tk2[4], tk3[4];
	packing(tk2, t2);
	packing(tk3, t3);
	tk[0] = tk2[0] ^ tk3[0];
	tk[1] = tk2[1] ^ tk3[1];
	tk[2] = tk2[2] ^ tk3[2];
	tk[3] = tk2[3] ^ tk3[3];
	for(int i = 0 ; i < rounds; i+=8) {
		tk2[0] ^= (tk2[2] & 0xaaaaaaaa);
		tk2[0] = ((tk2[0] & 0xaaaaaaaa) >> 1) | ((tk2[0] << 1) & 0xaaaaaaaa);
		tk3[3] ^= ((tk3[1] & 0xaaaaaaaa) >> 1);
		tk3[3] = ((tk3[3] & 0xaaaaaaaa) >> 1) | ((tk3[3] << 1) & 0xaaaaaaaa);
		tk[i*4+4] = tk2[1] ^ tk3[3];
		tk[i*4+5] = tk2[2] ^ tk3[0];
		tk[i*4+6] = tk2[3] ^ tk3[1];
		tk[i*4+7] = tk2[0] ^ tk3[2];
		tk2[1] ^= (tk2[3] & 0xaaaaaaaa);
		tk2[1] = ((tk2[1] & 0xaaaaaaaa) >> 1) | ((tk2[1] << 1) & 0xaaaaaaaa);
		tk3[2] ^= ((tk3[0] & 0xaaaaaaaa) >> 1);
		tk3[2] = ((tk3[2] & 0xaaaaaaaa) >> 1) | ((tk3[2] << 1) & 0xaaaaaaaa);
		tk[i*4+12] = tk2[2] ^ tk3[2];
		tk[i*4+13] = tk2[3] ^ tk3[3];
		tk[i*4+14] = tk2[0] ^ tk3[0];
		tk[i*4+15] = tk2[1] ^ tk3[1];
		tk2[2] ^= (tk2[0] & 0xaaaaaaaa);
		tk2[2] = ((tk2[2] & 0xaaaaaaaa) >> 1) | ((tk2[2] << 1) & 0xaaaaaaaa);
		tk3[1] ^= ((tk3[3] & 0xaaaaaaaa) >> 1);
		tk3[1] = ((tk3[1] & 0xaaaaaaaa) >> 1) | ((tk3[1] << 1) & 0xaaaaaaaa);
		tk[i*4+20] = tk2[3] ^ tk3[1];
		tk[i*4+21] = tk2[0] ^ tk3[2];
		tk[i*4+22] = tk2[1] ^ tk3[3];
		tk[i*4+23] = tk2[2] ^ tk3[0];
		tk2[3] ^= (tk2[1] & 0xaaaaaaaa);
		tk2[3] = ((tk2[3] & 0xaaaaaaaa) >> 1) | ((tk2[3] << 1) & 0xaaaaaaaa);
		tk3[0] ^= ((tk3[2] & 0xaaaaaaaa) >> 1);
		tk3[0] = ((tk3[0] & 0xaaaaaaaa) >> 1) | ((tk3[0] << 1) & 0xaaaaaaaa);
		tk[i*4+28] = tk2[0] ^ tk3[0];
		tk[i*4+29] = tk2[1] ^ tk3[1];
		tk[i*4+30] = tk2[2] ^ tk3[2];
		tk[i*4+31] = tk2[3] ^ tk3[3];
	}
}

/******************************************************************************
Alexandre Adomnicai committed
333 334 335 336 337 338 339
* XOR TK with TK1 before applying the permutations.
* The key is then rearranged to match the barrel shiftrows representation.
******************************************************************************/
void permute_tk(u32* tk, const u8* key, const int rounds) {
	u32 test;
	u32 tk1[4], tmp[4];
	packing(tk1, key);
Alexandre Adomnicai committed
340 341 342 343
	tmp[0] = tk[0] ^ tk1[0];
	tmp[1] = tk[1] ^ tk1[1];
	tmp[2] = tk[2] ^ tk1[2];
	tmp[3] = tk[3] ^ tk1[3];
Alexandre Adomnicai committed
344 345 346 347 348 349
	for(int i = 0 ; i < rounds; i += 8) {
		test = (i % 16 < 8) ? 1 : 0; 			//to apply the right power of P
		tk[i*4] = tmp[2] & 0xf0f0f0f0;
		tk[i*4+1] = tmp[3] & 0xf0f0f0f0;
		tk[i*4+2] = tmp[0] & 0xf0f0f0f0;
		tk[i*4+3] = tmp[1] & 0xf0f0f0f0;
Alexandre Adomnicai committed
350 351 352 353
		tmp[0] = tk[i*4+4] ^ tk1[0];
		tmp[1] = tk[i*4+5] ^ tk1[1];
		tmp[2] = tk[i*4+6] ^ tk1[2];
		tmp[3] = tk[i*4+7] ^ tk1[3];
Alexandre Adomnicai committed
354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369
		if (test)
			permute_tk_2(tmp); 					// applies P^2
		else
			permute_tk_10(tmp); 				// applies P^10
		tk[i*4+4] = ROR(tmp[0],26) & 0xc3c3c3c3;
		tk[i*4+5] = ROR(tmp[1],26) & 0xc3c3c3c3;
		tk[i*4+6] = ROR(tmp[2],26) & 0xc3c3c3c3;
		tk[i*4+7] = ROR(tmp[3],26) & 0xc3c3c3c3;
		tk[i*4+8] = ROR(tmp[2],28) & 0x03030303;
		tk[i*4+8] |= ROR(tmp[2],12) & 0x0c0c0c0c;
		tk[i*4+9] = ROR(tmp[3],28) & 0x03030303;
		tk[i*4+9] |= ROR(tmp[3],12) & 0x0c0c0c0c;
		tk[i*4+10] = ROR(tmp[0],28) & 0x03030303;
		tk[i*4+10] |= ROR(tmp[0],12) & 0x0c0c0c0c;
		tk[i*4+11] = ROR(tmp[1],28) & 0x03030303;
		tk[i*4+11] |= ROR(tmp[1],12) & 0x0c0c0c0c;
Alexandre Adomnicai committed
370 371 372 373
		tmp[0] = tk[i*4+12] ^ tk1[0];
		tmp[1] = tk[i*4+13] ^ tk1[1];
		tmp[2] = tk[i*4+14] ^ tk1[2];
		tmp[3] = tk[i*4+15] ^ tk1[3];
Alexandre Adomnicai committed
374 375 376 377 378 379 380 381 382 383 384 385
		if (test)
			permute_tk_4(tmp); 					// applies P^4
		else
			permute_tk_12(tmp); 				// applies P^12
		for(int j = 0; j < 4; j++) {
			tk[i*4+12+j] = ROR(tmp[j],14) & 0x30303030;
			tk[i*4+12+j] |= ROR(tmp[j],6) & 0x0c0c0c0c;
		}
		tk[i*4+16] = ROR(tmp[2], 16) & 0xf0f0f0f0;
		tk[i*4+17] = ROR(tmp[3], 16) & 0xf0f0f0f0;
		tk[i*4+18] = ROR(tmp[0], 16) & 0xf0f0f0f0;
		tk[i*4+19] = ROR(tmp[1], 16) & 0xf0f0f0f0;
Alexandre Adomnicai committed
386 387 388 389
		tmp[0] = tk[i*4+20] ^ tk1[0];
		tmp[1] = tk[i*4+21] ^ tk1[1];
		tmp[2] = tk[i*4+22] ^ tk1[2];
		tmp[3] = tk[i*4+23] ^ tk1[3];
Alexandre Adomnicai committed
390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
		if (test)
			permute_tk_6(tmp); 					//	applies P^6
		else
			permute_tk_14(tmp); 				// applies P^14
		tk[i*4+20] = ROR(tmp[0], 10) & 0xc3c3c3c3;
		tk[i*4+21] = ROR(tmp[1], 10) & 0xc3c3c3c3;
		tk[i*4+22] = ROR(tmp[2], 10) & 0xc3c3c3c3;
		tk[i*4+23] = ROR(tmp[3], 10) & 0xc3c3c3c3;
		tk[i*4+24] = ROR(tmp[2],12) & 0x03030303;
		tk[i*4+24] |= ROR(tmp[2],28) & 0x0c0c0c0c;
		tk[i*4+25] = ROR(tmp[3],12) & 0x03030303;
		tk[i*4+25] |= ROR(tmp[3],28) & 0x0c0c0c0c;
		tk[i*4+26] = ROR(tmp[0],12) & 0x03030303;
		tk[i*4+26] |= ROR(tmp[0],28) & 0x0c0c0c0c;
		tk[i*4+27] = ROR(tmp[1],12) & 0x03030303;
		tk[i*4+27] |= ROR(tmp[1],28) & 0x0c0c0c0c;
Alexandre Adomnicai committed
406 407 408 409
		tmp[0] = tk[i*4+28] ^ tk1[0];
		tmp[1] = tk[i*4+29] ^ tk1[1];
		tmp[2] = tk[i*4+30] ^ tk1[2];
		tmp[3] = tk[i*4+31] ^ tk1[3];
Alexandre Adomnicai committed
410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429
		if (test)
			permute_tk_8(tmp); 					// applies P^8
		for(int j = 0; j < 4; j++) {
			tk[i*4+28+j] = ROR(tmp[j],30) & 0x30303030;
			tk[i*4+28+j] |= ROR(tmp[j],22) & 0x0c0c0c0c;
		}
		if (test && (i+8 < rounds)) { 			//only if next loop iteration
			tk[i*4+32] = tmp[2] & 0xf0f0f0f0;
			tk[i*4+33] = tmp[3] & 0xf0f0f0f0;
			tk[i*4+34] = tmp[0] & 0xf0f0f0f0;
			tk[i*4+35] = tmp[1] & 0xf0f0f0f0;
		}
	}
}

/******************************************************************************
* Precompute LFSR2(TK2) ^ LFSR3(TK3) ^ rconst.
******************************************************************************/
void precompute_rtk2_3(u32* rtk, const u8* tk2, const u8 * tk3) {
	memset(rtk, 0x00, 16*SKINNY128_384_ROUNDS);
Alexandre Adomnicai committed
430
	precompute_lfsr_tk2_3(rtk, tk2, tk3, SKINNY128_384_ROUNDS);
Alexandre Adomnicai committed
431 432 433 434 435 436 437 438 439 440 441 442 443
	permute_tk(rtk, (u8*)(rtk+8), SKINNY128_384_ROUNDS);	// rtk+8 is NULL
	for(int i = 0; i < SKINNY128_384_ROUNDS; i++) {			// add rconsts
		for(int j = 0; j < 4; j++)
			rtk[i*4+j] ^= rconst_32_bs[i*4+j];
	}
}

/******************************************************************************
* Precompute RTK1.
******************************************************************************/
void precompute_rtk1(u32* rtk1, const u8* tk1) {
	memset(rtk1, 0x00, 16*16);
	permute_tk(rtk1, tk1, 16);
Alexandre Adomnicai committed
444
}