3 This file is part of the AVR-Crypto-Lib.
4 Copyright (C) 2009 Daniel Otte (daniel.otte@rub.de)
6 This program is free software: you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation, either version 3 of the License, or
9 (at your option) any later version.
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program. If not, see <http://www.gnu.org/licenses/>.
22 * \email daniel.otte@rub.de
24 * \license GPLv3 or later
30 #include <avr/pgmspace.h>
32 #include "bmw_small.h"
35 #define SHL32(a,n) ((a)<<(n))
36 #define SHR32(a,n) ((a)>>(n))
37 #define ROTL32(a,n) (((a)<<(n))|((a)>>(32-(n))))
38 #define ROTR32(a,n) (((a)>>(n))|((a)<<(32-(n))))
59 void ctx_dump(const bmw_small_ctx_t* ctx){
61 cli_putstr_P(PSTR("\r\n==== ctx dump ===="));
63 cli_putstr_P(PSTR("\r\n h["));
65 cli_putstr_P(PSTR("] = "));
66 cli_hexdump_rev(&(ctx->h[i]), 4);
68 cli_putstr_P(PSTR("\r\n counter = "));
69 cli_hexdump(&(ctx->counter), 4);
72 void dump_x(const uint32_t* q, uint8_t elements, char x){
74 cli_putstr_P(PSTR("\r\n==== "));
76 cli_putstr_P(PSTR(" dump ===="));
77 for(i=0; i<elements;++i){
78 cli_putstr_P(PSTR("\r\n "));
80 cli_putstr_P(PSTR("["));
82 cli_putstr_P(PSTR("] = "));
83 cli_hexdump_rev(&(q[i]), 4);
92 uint32_t bmw_small_s0(uint32_t x){
102 uint32_t bmw_small_s1(uint32_t x){
112 uint32_t bmw_small_s2(uint32_t x){
122 uint32_t bmw_small_s3(uint32_t x){
132 uint32_t bmw_small_s4(uint32_t x){
140 uint32_t bmw_small_s5(uint32_t x){
148 uint32_t bmw_small_r1(uint32_t x){
155 uint32_t bmw_small_r2(uint32_t x){
162 uint32_t bmw_small_r3(uint32_t x){
169 uint32_t bmw_small_r4(uint32_t x){
176 uint32_t bmw_small_r5(uint32_t x){
183 uint32_t bmw_small_r6(uint32_t x){
190 uint32_t bmw_small_r7(uint32_t x){
196 #define K 0x05555555L
198 uint32_t k_lut[] PROGMEM = {
199 16L*K, 17L*K, 18L*K, 19L*K, 20L*K, 21L*K, 22L*K, 23L*K,
200 24L*K, 25L*K, 26L*K, 27L*K, 28L*K, 29L*K, 30L*K, 31L*K
203 /* same as above but precomputed to avoid compiler warnings */
206 uint32_t k_lut[] PROGMEM = {
207 0x55555550L, 0x5aaaaaa5L, 0x5ffffffaL,
208 0x6555554fL, 0x6aaaaaa4L, 0x6ffffff9L,
209 0x7555554eL, 0x7aaaaaa3L, 0x7ffffff8L,
210 0x8555554dL, 0x8aaaaaa2L, 0x8ffffff7L,
211 0x9555554cL, 0x9aaaaaa1L, 0x9ffffff6L,
215 uint32_t bmw_small_expand1(uint8_t j, const uint32_t* q, const void* m, const void* h){
216 uint32_t(*s[])(uint32_t) = {bmw_small_s1, bmw_small_s2, bmw_small_s3, bmw_small_s0};
219 /* r = 0x05555555*(j+16); */
222 r = ( ROTL32(((uint32_t*)m)[j&0xf], ((j+0)&0xf)+1 )
223 + ROTL32(((uint32_t*)m)[(j+3)&0xf], ((j+3)&0xf)+1 )
224 - ROTL32(((uint32_t*)m)[(j+10)&0xf], ((j+10)&0xf)+1 )
225 + pgm_read_dword(k_lut+j)
226 ) ^ ((uint32_t*)h)[(j+7)&0xf];
228 r = pgm_read_dword(k_lut+j);
229 r += ((uint32_t*)m)[j&0xf];
230 r += ((uint32_t*)m)[(j+3)&0xf];
231 r -= ((uint32_t*)m)[(j+10)&0xf];
240 uint32_t bmw_small_expand2(uint8_t j, const uint32_t* q, const void* m, const void* h){
241 uint32_t(*rf[])(uint32_t) = {bmw_small_r1, bmw_small_r2, bmw_small_r3,
242 bmw_small_r4, bmw_small_r5, bmw_small_r6,
246 for(i=0; i<14; i+=2){
249 for(i=0; i<14; i+=2){
250 r += rf[i/2](q[j+i+1]);
253 r += bmw_small_s4(q[j+14]);
254 r += bmw_small_s5(q[j+15]);
256 r += bmw_small_s5(q[j+14]);
257 r += bmw_small_s4(q[j+15]);
260 r += ( ROTL32(((uint32_t*)m)[j&0xf], ((j+0)&0xf)+1 )
261 + ROTL32(((uint32_t*)m)[(j+3)&0xf], ((j+3)&0xf)+1 )
262 - ROTL32(((uint32_t*)m)[(j+10)&0xf], ((j+10)&0xf)+1 )
263 + pgm_read_dword(k_lut+j)
264 ) ^ ((uint32_t*)h)[(j+7)&0xf];
266 r += pgm_read_dword(k_lut+j);
267 r += ((uint32_t*)m)[j&0xf];
268 r += ((uint32_t*)m)[(j+3)&0xf];
269 r -= ((uint32_t*)m)[(j+10)&0xf];
275 /* to understand this implementation take a look at f0-opt-table.txt */
276 static uint16_t hack_table[5] PROGMEM = { 0x0311, 0xDDB3, 0x2A79, 0x07AA, 0x51C2 };
277 static uint8_t offset_table[5] PROGMEM = { 4+16, 6+16, 9+16, 12+16, 13+16 };
280 void bmw_small_f0(uint32_t* q, uint32_t* h, const void* m){
283 uint32_t(*s[])(uint32_t)={ bmw_small_s0, bmw_small_s1, bmw_small_s2,
284 bmw_small_s3, bmw_small_s4 };
286 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
293 j=pgm_read_byte(offset_table+c);
294 hack_reg=pgm_read_word(&(hack_table[c]));
311 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
314 q[i] += h[(i+1)&0xf];
318 #endif /* F0_HACK==2*/
322 uint8_t f0_lut[] PROGMEM = {
323 5<<1, ( 7<<1)+1, (10<<1)+0, (13<<1)+0, (14<<1)+0,
324 6<<1, ( 8<<1)+1, (11<<1)+0, (14<<1)+0, (15<<1)+1,
325 0<<1, ( 7<<1)+0, ( 9<<1)+0, (12<<1)+1, (15<<1)+0,
326 0<<1, ( 1<<1)+1, ( 8<<1)+0, (10<<1)+1, (13<<1)+0,
327 1<<1, ( 2<<1)+0, ( 9<<1)+0, (11<<1)+1, (14<<1)+1,
328 3<<1, ( 2<<1)+1, (10<<1)+0, (12<<1)+1, (15<<1)+0,
329 4<<1, ( 0<<1)+1, ( 3<<1)+1, (11<<1)+1, (13<<1)+0,
330 1<<1, ( 4<<1)+1, ( 5<<1)+1, (12<<1)+1, (14<<1)+1,
331 2<<1, ( 5<<1)+1, ( 6<<1)+1, (13<<1)+0, (15<<1)+1,
332 0<<1, ( 3<<1)+1, ( 6<<1)+0, ( 7<<1)+1, (14<<1)+0,
333 8<<1, ( 1<<1)+1, ( 4<<1)+1, ( 7<<1)+1, (15<<1)+0,
334 8<<1, ( 0<<1)+1, ( 2<<1)+1, ( 5<<1)+1, ( 9<<1)+0,
335 1<<1, ( 3<<1)+0, ( 6<<1)+1, ( 9<<1)+1, (10<<1)+0,
336 2<<1, ( 4<<1)+0, ( 7<<1)+0, (10<<1)+0, (11<<1)+0,
337 3<<1, ( 5<<1)+1, ( 8<<1)+0, (11<<1)+1, (12<<1)+1,
338 12<<1, ( 4<<1)+1, ( 6<<1)+1, ( 9<<1)+1, (13<<1)+0
342 void bmw_small_f0(uint32_t* q, uint32_t* h, const void* m){
343 uint8_t i,j=-1,v,sign,l=0;
344 uint32_t(*s[])(uint32_t)={ bmw_small_s0, bmw_small_s1, bmw_small_s2,
345 bmw_small_s3, bmw_small_s4 };
347 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
350 // memset(q, 0, 4*16);
351 for(i=0; i<5*16; ++i){
352 v = pgm_read_byte(f0_lut+i);
373 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
376 q[i] += h[(i+1)&0xf];
380 #endif /* F0_HACK==1 */
384 void bmw_small_f0(uint32_t* q, uint32_t* h, const void* m){
386 uint32_t(*s[])(uint32_t)={ bmw_small_s0, bmw_small_s1, bmw_small_s2,
387 bmw_small_s3, bmw_small_s4 };
389 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
392 q[ 0] = (h[ 5] - h[ 7] + h[10] + h[13] + h[14]);
393 q[ 1] = (h[ 6] - h[ 8] + h[11] + h[14] - h[15]);
394 q[ 2] = (h[ 0] + h[ 7] + h[ 9] - h[12] + h[15]);
395 q[ 3] = (h[ 0] - h[ 1] + h[ 8] - h[10] + h[13]);
396 q[ 4] = (h[ 1] + h[ 2] + h[ 9] - h[11] - h[14]);
397 q[ 5] = (h[ 3] - h[ 2] + h[10] - h[12] + h[15]);
398 q[ 6] = (h[ 4] - h[ 0] - h[ 3] - h[11] + h[13]);
399 q[ 7] = (h[ 1] - h[ 4] - h[ 5] - h[12] - h[14]);
400 q[ 8] = (h[ 2] - h[ 5] - h[ 6] + h[13] - h[15]);
401 q[ 9] = (h[ 0] - h[ 3] + h[ 6] - h[ 7] + h[14]);
402 q[10] = (h[ 8] - h[ 1] - h[ 4] - h[ 7] + h[15]);
403 q[11] = (h[ 8] - h[ 0] - h[ 2] - h[ 5] + h[ 9]);
404 q[12] = (h[ 1] + h[ 3] - h[ 6] - h[ 9] + h[10]);
405 q[13] = (h[ 2] + h[ 4] + h[ 7] + h[10] + h[11]);
406 q[14] = (h[ 3] - h[ 5] + h[ 8] - h[11] - h[12]);
407 q[15] = (h[12] - h[ 4] - h[ 6] - h[ 9] + h[13]);
414 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
417 q[i] += h[(i+1)&0xf];
421 #endif /* F0_HACK==0 */
424 void bmw_small_f1(uint32_t* q, const void* m, const void* h){
426 q[16] = bmw_small_expand1(0, q, m, h);
427 q[17] = bmw_small_expand1(1, q, m, h);
429 q[16+i] = bmw_small_expand2(i, q, m, h);
434 void bmw_small_f2(uint32_t* h, uint32_t* q, const void* m){
445 cli_putstr_P(PSTR("\r\n XL = "));
446 cli_hexdump_rev(&xl, 4);
447 cli_putstr_P(PSTR("\r\n XH = "));
448 cli_hexdump_rev(&xh, 4);
451 h[0] ^= SHL32(xh, 5) ^ SHR32(q[16], 5);
452 h[1] ^= SHR32(xh, 7) ^ SHL32(q[17], 8);
453 h[2] ^= SHR32(xh, 5) ^ SHL32(q[18], 5);
454 h[3] ^= SHR32(xh, 1) ^ SHL32(q[19], 5);
455 h[4] ^= SHR32(xh, 3) ^ q[20];
456 h[5] ^= SHL32(xh, 6) ^ SHR32(q[21], 6);
457 h[6] ^= SHR32(xh, 4) ^ SHL32(q[22], 6);
458 h[7] ^= SHR32(xh,11) ^ SHL32(q[23], 2);
460 h[i] += xl ^ q[24+i] ^ q[i];
463 h[8+i] ^= xh ^ q[24+i];
464 h[8+i] += ROTL32(h[(4+i)%8],i+9);
467 h[ 8] += SHL32(xl, 8) ^ q[23] ^ q[ 8];
468 h[ 9] += SHR32(xl, 6) ^ q[16] ^ q[ 9];
469 h[10] += SHL32(xl, 6) ^ q[17] ^ q[10];
470 h[11] += SHL32(xl, 4) ^ q[18] ^ q[11];
471 h[12] += SHR32(xl, 3) ^ q[19] ^ q[12];
472 h[13] += SHR32(xl, 4) ^ q[20] ^ q[13];
473 h[14] += SHR32(xl, 7) ^ q[21] ^ q[14];
474 h[15] += SHR32(xl, 2) ^ q[22] ^ q[15];
476 memxor(q+9, q+16, 7*4);
478 h[ 8] += SHL32(xl, 8) ^ q[ 8];
479 h[ 9] += SHR32(xl, 6) ^ q[ 9];
480 h[10] += SHL32(xl, 6) ^ q[10];
481 h[11] += SHL32(xl, 4) ^ q[11];
482 h[12] += SHR32(xl, 3) ^ q[12];
483 h[13] += SHR32(xl, 4) ^ q[13];
484 h[14] += SHR32(xl, 7) ^ q[14];
485 h[15] += SHR32(xl, 2) ^ q[15];
489 void bmw_small_nextBlock(bmw_small_ctx_t* ctx, const void* block){
491 dump_x(block, 16, 'M');
492 bmw_small_f0(q, ctx->h, block);
494 bmw_small_f1(q, block, ctx->h);
496 bmw_small_f2(ctx->h, q, block);
501 void bmw_small_lastBlock(bmw_small_ctx_t* ctx, const void* block, uint16_t length_b){
503 while(length_b >= BMW_SMALL_BLOCKSIZE){
504 bmw_small_nextBlock(ctx, block);
505 length_b -= BMW_SMALL_BLOCKSIZE;
506 block = (uint8_t*)block + BMW_SMALL_BLOCKSIZE_B;
508 memset(buffer, 0, 64);
509 memcpy(buffer, block, (length_b+7)/8);
510 buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
511 if(length_b+1>64*8-64){
512 bmw_small_nextBlock(ctx, buffer);
513 memset(buffer, 0, 64-8);
516 *((uint64_t*)&(buffer[64-8])) = (uint64_t)(ctx->counter*512LL)+(uint64_t)length_b;
517 bmw_small_nextBlock(ctx, buffer);
521 memset(buffer, 0xaa, 64);
523 buffer[i*4] = i+0xa0;
525 // dump_x(buffer, 16, 'A');
526 dump_x(ctx->h, 16, 'M');
527 bmw_small_f0(q, (uint32_t*)buffer, ctx->h);
528 dump_x(buffer, 16, 'a');
530 bmw_small_f1(q, ctx->h, (uint32_t*)buffer);
532 bmw_small_f2((uint32_t*)buffer, q, ctx->h);
533 memcpy(ctx->h, buffer, 64);
537 void bmw224_init(bmw224_ctx_t* ctx){
539 ctx->h[0] = 0x00010203;
541 ctx->h[i] = ctx->h[i-1]+ 0x04040404;
544 ctx->h[13] = 0x24353637;
550 void bmw256_init(bmw256_ctx_t* ctx){
552 ctx->h[0] = 0x40414243;
554 ctx->h[i] = ctx->h[i-1]+ 0x04040404;
560 void bmw224_nextBlock(bmw224_ctx_t* ctx, const void* block){
561 bmw_small_nextBlock(ctx, block);
564 void bmw256_nextBlock(bmw256_ctx_t* ctx, const void* block){
565 bmw_small_nextBlock(ctx, block);
568 void bmw224_lastBlock(bmw224_ctx_t* ctx, const void* block, uint16_t length_b){
569 bmw_small_lastBlock(ctx, block, length_b);
572 void bmw256_lastBlock(bmw256_ctx_t* ctx, const void* block, uint16_t length_b){
573 bmw_small_lastBlock(ctx, block, length_b);
576 void bmw224_ctx2hash(void* dest, const bmw224_ctx_t* ctx){
577 memcpy(dest, &(ctx->h[9]), 224/8);
580 void bmw256_ctx2hash(void* dest, const bmw256_ctx_t* ctx){
581 memcpy(dest, &(ctx->h[8]), 256/8);
584 void bmw224(void* dest, const void* msg, uint32_t length_b){
587 while(length_b>=BMW_SMALL_BLOCKSIZE){
588 bmw_small_nextBlock(&ctx, msg);
589 length_b -= BMW_SMALL_BLOCKSIZE;
590 msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
592 bmw_small_lastBlock(&ctx, msg, length_b);
593 bmw224_ctx2hash(dest, &ctx);
596 void bmw256(void* dest, const void* msg, uint32_t length_b){
599 while(length_b>=BMW_SMALL_BLOCKSIZE){
600 bmw_small_nextBlock(&ctx, msg);
601 length_b -= BMW_SMALL_BLOCKSIZE;
602 msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
604 bmw_small_lastBlock(&ctx, msg, length_b);
605 bmw256_ctx2hash(dest, &ctx);