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>
31 #include "bmw_small.h"
34 #define SHL32(a,n) ((a)<<(n))
35 #define SHR32(a,n) ((a)>>(n))
36 #define ROTL32(a,n) (((a)<<(n))|((a)>>(32-(n))))
37 #define ROTR32(a,n) (((a)>>(n))|((a)<<(32-(n))))
49 void ctx_dump(const bmw_small_ctx_t* ctx){
51 cli_putstr_P(PSTR("\r\n==== ctx dump ===="));
53 cli_putstr_P(PSTR("\r\n h["));
55 cli_putstr_P(PSTR("] = "));
56 cli_hexdump_rev(&(ctx->h[i]), 4);
58 cli_putstr_P(PSTR("\r\n counter = "));
59 cli_hexdump(&(ctx->counter), 4);
62 void dump_x(const uint32_t* q, uint8_t elements, char x){
64 cli_putstr_P(PSTR("\r\n==== "));
66 cli_putstr_P(PSTR(" dump ===="));
67 for(i=0; i<elements;++i){
68 cli_putstr_P(PSTR("\r\n "));
70 cli_putstr_P(PSTR("["));
72 cli_putstr_P(PSTR("] = "));
73 cli_hexdump_rev(&(q[i]), 4);
81 uint32_t bmw_small_s0(uint32_t x){
90 uint32_t bmw_small_s1(uint32_t x){
99 uint32_t bmw_small_s2(uint32_t x){
108 uint32_t bmw_small_s3(uint32_t x){
117 uint32_t bmw_small_s4(uint32_t x){
124 uint32_t bmw_small_s5(uint32_t x){
131 uint32_t bmw_small_r1(uint32_t x){
137 uint32_t bmw_small_r2(uint32_t x){
143 uint32_t bmw_small_r3(uint32_t x){
149 uint32_t bmw_small_r4(uint32_t x){
155 uint32_t bmw_small_r5(uint32_t x){
161 uint32_t bmw_small_r6(uint32_t x){
167 uint32_t bmw_small_r7(uint32_t x){
173 #define K 0x05555555L
175 uint32_t k_lut[] PROGMEM = {
176 16L*K, 17L*K, 18L*K, 19L*K, 20L*K, 21L*K, 22L*K, 23L*K,
177 24L*K, 25L*K, 26L*K, 27L*K, 28L*K, 29L*K, 30L*K, 31L*K
180 /* same as above but precomputed to avoid compiler warnings */
183 uint32_t k_lut[] PROGMEM = {
184 0x55555550L, 0x5aaaaaa5L, 0x5ffffffaL,
185 0x6555554fL, 0x6aaaaaa4L, 0x6ffffff9L,
186 0x7555554eL, 0x7aaaaaa3L, 0x7ffffff8L,
187 0x8555554dL, 0x8aaaaaa2L, 0x8ffffff7L,
188 0x9555554cL, 0x9aaaaaa1L, 0x9ffffff6L,
192 uint32_t bmw_small_expand1(uint8_t j, const uint32_t* q, const void* m, const void* h){
193 uint32_t(*s[])(uint32_t) = {bmw_small_s1, bmw_small_s2, bmw_small_s3, bmw_small_s0};
196 /* r = 0x05555555*(j+16); */
202 r += ( ROTL32(((uint32_t*)m)[j&0xf], ((j+0)&0xf)+1 )
203 + ROTL32(((uint32_t*)m)[(j+3)&0xf], ((j+3)&0xf)+1 )
204 - ROTL32(((uint32_t*)m)[(j+10)&0xf], ((j+10)&0xf)+1 )
205 + pgm_read_dword(k_lut+j)
206 ) ^ ((uint32_t*)h)[(j+7)&0xf];
208 r += ( ROTL32(((uint32_t*)m)[j&0xf], (j+1)&0xf )
209 + ROTL32(((uint32_t*)m)[(j+3)&0xf], (j+4)&0xf )
210 - ROTL32(((uint32_t*)m)[(j+10)&0xf], (j+11)&0xf )
211 + pgm_read_dword(k_lut+j)
212 ) ^ ((uint32_t*)h)[(j+7)&0xf];
215 r += pgm_read_dword(k_lut+j);
216 r += ((uint32_t*)m)[j&0xf];
217 r += ((uint32_t*)m)[(j+3)&0xf];
218 r -= ((uint32_t*)m)[(j+10)&0xf];
223 uint32_t bmw_small_expand2(uint8_t j, const uint32_t* q, const void* m, const void* h){
224 uint32_t(*rf[])(uint32_t) = {bmw_small_r1, bmw_small_r2, bmw_small_r3,
225 bmw_small_r4, bmw_small_r5, bmw_small_r6,
229 for(i=0; i<14; i+=2){
232 for(i=0; i<14; i+=2){
233 r += rf[i/2](q[j+i+1]);
236 r += bmw_small_s4(q[j+14]);
237 r += bmw_small_s5(q[j+15]);
239 r += bmw_small_s5(q[j+14]);
240 r += bmw_small_s4(q[j+15]);
244 r += ( ROTL32(((uint32_t*)m)[j&0xf], ((j+0)&0xf)+1 )
245 + ROTL32(((uint32_t*)m)[(j+3)&0xf], ((j+3)&0xf)+1 )
246 - ROTL32(((uint32_t*)m)[(j+10)&0xf], ((j+10)&0xf)+1 )
247 + pgm_read_dword(k_lut+j)
248 ) ^ ((uint32_t*)h)[(j+7)&0xf];
250 r += ( ROTL32(((uint32_t*)m)[j&0xf], (j+1)&0xf )
251 + ROTL32(((uint32_t*)m)[(j+3)&0xf], (j+4)&0xf )
252 - ROTL32(((uint32_t*)m)[(j+10)&0xf], (j+11)&0xf )
253 + pgm_read_dword(k_lut+j)
254 ) ^ ((uint32_t*)h)[(j+7)&0xf];
258 r += pgm_read_dword(k_lut+j);
259 r += ((uint32_t*)m)[j&0xf];
260 r += ((uint32_t*)m)[(j+3)&0xf];
261 r -= ((uint32_t*)m)[(j+10)&0xf];
268 uint8_t f0_lut[] PROGMEM = {
269 5<<1, ( 7<<1)+1, (10<<1)+0, (13<<1)+0, (14<<1)+0,
270 6<<1, ( 8<<1)+1, (11<<1)+0, (14<<1)+0, (15<<1)+1,
271 0<<1, ( 7<<1)+0, ( 9<<1)+0, (12<<1)+1, (15<<1)+0,
272 0<<1, ( 1<<1)+1, ( 8<<1)+0, (10<<1)+1, (13<<1)+0,
273 1<<1, ( 2<<1)+0, ( 9<<1)+0, (11<<1)+1, (14<<1)+1,
274 3<<1, ( 2<<1)+1, (10<<1)+0, (12<<1)+1, (15<<1)+0,
275 4<<1, ( 0<<1)+1, ( 3<<1)+1, (11<<1)+1, (13<<1)+0,
276 1<<1, ( 4<<1)+1, ( 5<<1)+1, (12<<1)+1, (14<<1)+1,
277 2<<1, ( 5<<1)+1, ( 6<<1)+1, (13<<1)+0, (15<<1)+1,
278 0<<1, ( 3<<1)+1, ( 6<<1)+0, ( 7<<1)+1, (14<<1)+0,
279 8<<1, ( 1<<1)+1, ( 4<<1)+1, ( 7<<1)+1, (15<<1)+0,
280 8<<1, ( 0<<1)+1, ( 2<<1)+1, ( 5<<1)+1, ( 9<<1)+0,
281 1<<1, ( 3<<1)+0, ( 6<<1)+1, ( 9<<1)+1, (10<<1)+0,
282 2<<1, ( 4<<1)+0, ( 7<<1)+0, (10<<1)+0, (11<<1)+0,
283 3<<1, ( 5<<1)+1, ( 8<<1)+0, (11<<1)+1, (12<<1)+1,
284 12<<1, ( 4<<1)+1, ( 6<<1)+1, ( 9<<1)+1, (13<<1)+0
287 void bmw_small_f0(uint32_t* q, const uint32_t* h, const void* m){
288 uint8_t i,j=-1,v,sign,l=0;
289 uint32_t(*s[])(uint32_t)={ bmw_small_s0, bmw_small_s1, bmw_small_s2,
290 bmw_small_s3, bmw_small_s4 };
292 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
295 // memset(q, 0, 4*16);
296 for(i=0; i<5*16; ++i){
297 v = pgm_read_byte(f0_lut+i);
318 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
321 q[i] += h[(i+1)&0xf];
327 void bmw_small_f0(uint32_t* q, const uint32_t* h, const void* m){
329 uint32_t(*s[])(uint32_t)={ bmw_small_s0, bmw_small_s1, bmw_small_s2,
330 bmw_small_s3, bmw_small_s4 };
332 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
335 q[ 0] = (h[ 5] - h[ 7] + h[10] + h[13] + h[14]);
336 q[ 1] = (h[ 6] - h[ 8] + h[11] + h[14] - h[15]);
337 q[ 2] = (h[ 0] + h[ 7] + h[ 9] - h[12] + h[15]);
338 q[ 3] = (h[ 0] - h[ 1] + h[ 8] - h[10] + h[13]);
339 q[ 4] = (h[ 1] + h[ 2] + h[ 9] - h[11] - h[14]);
340 q[ 5] = (h[ 3] - h[ 2] + h[10] - h[12] + h[15]);
341 q[ 6] = (h[ 4] - h[ 0] - h[ 3] - h[11] + h[13]);
342 q[ 7] = (h[ 1] - h[ 4] - h[ 5] - h[12] - h[14]);
343 q[ 8] = (h[ 2] - h[ 5] - h[ 6] + h[13] - h[15]);
344 q[ 9] = (h[ 0] - h[ 3] + h[ 6] - h[ 7] + h[14]);
345 q[10] = (h[ 8] - h[ 1] - h[ 4] - h[ 7] + h[15]);
346 q[11] = (h[ 8] - h[ 0] - h[ 2] - h[ 5] + h[ 9]);
347 q[12] = (h[ 1] + h[ 3] - h[ 6] - h[ 9] + h[10]);
348 q[13] = (h[ 2] + h[ 4] + h[ 7] + h[10] + h[11]);
349 q[14] = (h[ 3] - h[ 5] + h[ 8] - h[11] - h[12]);
350 q[15] = (h[12] - h[ 4] - h[ 6] - h[ 9] + h[13]);
357 ((uint32_t*)h)[i] ^= ((uint32_t*)m)[i];
360 q[i] += h[(i+1)&0xf];
366 void bmw_small_f1(uint32_t* q, const void* m, const void* h){
368 q[16] = bmw_small_expand1(0, q, m, h);
369 q[17] = bmw_small_expand1(1, q, m, h);
371 q[16+i] = bmw_small_expand2(i, q, m, h);
375 void bmw_small_f2(uint32_t* h, const uint32_t* q, const void* m){
386 cli_putstr_P(PSTR("\r\n XL = "));
387 cli_hexdump_rev(&xl, 4);
388 cli_putstr_P(PSTR("\r\n XH = "));
389 cli_hexdump_rev(&xh, 4);
392 h[0] ^= SHL32(xh, 5) ^ SHR32(q[16], 5);
393 h[1] ^= SHR32(xh, 7) ^ SHL32(q[17], 8);
394 h[2] ^= SHR32(xh, 5) ^ SHL32(q[18], 5);
395 h[3] ^= SHR32(xh, 1) ^ SHL32(q[19], 5);
396 h[4] ^= SHR32(xh, 3) ^ q[20];
397 h[5] ^= SHL32(xh, 6) ^ SHR32(q[21], 6);
398 h[6] ^= SHR32(xh, 4) ^ SHL32(q[22], 6);
399 h[7] ^= SHR32(xh,11) ^ SHL32(q[23], 2);
401 h[i] += xl ^ q[24+i] ^ q[i];
404 h[8+i] ^= xh ^ q[24+i];
405 h[8+i] += ROTL32(h[(4+i)%8],i+9);
407 h[ 8] += SHL32(xl, 8) ^ q[23] ^ q[ 8];
408 h[ 9] += SHR32(xl, 6) ^ q[16] ^ q[ 9];
409 h[10] += SHL32(xl, 6) ^ q[17] ^ q[10];
410 h[11] += SHL32(xl, 4) ^ q[18] ^ q[11];
411 h[12] += SHR32(xl, 3) ^ q[19] ^ q[12];
412 h[13] += SHR32(xl, 4) ^ q[20] ^ q[13];
413 h[14] += SHR32(xl, 7) ^ q[21] ^ q[14];
414 h[15] += SHR32(xl, 2) ^ q[22] ^ q[15];
417 void bmw_small_nextBlock(bmw_small_ctx_t* ctx, const void* block){
419 dump_x(block, 16, 'M');
420 bmw_small_f0(q, ctx->h, block);
422 bmw_small_f1(q, block, ctx->h);
424 bmw_small_f2(ctx->h, q, block);
429 void bmw_small_lastBlock(bmw_small_ctx_t* ctx, const void* block, uint16_t length_b){
431 while(length_b >= BMW_SMALL_BLOCKSIZE){
432 bmw_small_nextBlock(ctx, block);
433 length_b -= BMW_SMALL_BLOCKSIZE;
434 block = (uint8_t*)block + BMW_SMALL_BLOCKSIZE_B;
436 memset(buffer, 0, 64);
437 memcpy(buffer, block, (length_b+7)/8);
438 buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
439 if(length_b+1>64*8-64){
440 bmw_small_nextBlock(ctx, buffer);
441 memset(buffer, 0, 64-8);
444 *((uint64_t*)&(buffer[64-8])) = (uint64_t)(ctx->counter*512LL)+(uint64_t)length_b;
445 bmw_small_nextBlock(ctx, buffer);
449 memset(buffer, 0xaa, 64);
451 buffer[i*4] = i+0xa0;
453 // dump_x(buffer, 16, 'A');
454 dump_x(ctx->h, 16, 'M');
455 bmw_small_f0(q, (uint32_t*)buffer, ctx->h);
456 dump_x(buffer, 16, 'a');
458 bmw_small_f1(q, ctx->h, (uint32_t*)buffer);
460 bmw_small_f2((uint32_t*)buffer, q, ctx->h);
461 memcpy(ctx->h, buffer, 64);
465 void bmw224_init(bmw224_ctx_t* ctx){
467 ctx->h[0] = 0x00010203;
469 ctx->h[i] = ctx->h[i-1]+ 0x04040404;
472 ctx->h[13] = 0x24353637;
478 void bmw256_init(bmw256_ctx_t* ctx){
480 ctx->h[0] = 0x40414243;
482 ctx->h[i] = ctx->h[i-1]+ 0x04040404;
488 void bmw224_nextBlock(bmw224_ctx_t* ctx, const void* block){
489 bmw_small_nextBlock(ctx, block);
492 void bmw256_nextBlock(bmw256_ctx_t* ctx, const void* block){
493 bmw_small_nextBlock(ctx, block);
496 void bmw224_lastBlock(bmw224_ctx_t* ctx, const void* block, uint16_t length_b){
497 bmw_small_lastBlock(ctx, block, length_b);
500 void bmw256_lastBlock(bmw256_ctx_t* ctx, const void* block, uint16_t length_b){
501 bmw_small_lastBlock(ctx, block, length_b);
504 void bmw224_ctx2hash(void* dest, const bmw224_ctx_t* ctx){
505 memcpy(dest, &(ctx->h[9]), 224/8);
508 void bmw256_ctx2hash(void* dest, const bmw256_ctx_t* ctx){
509 memcpy(dest, &(ctx->h[8]), 256/8);
512 void bmw224(void* dest, const void* msg, uint32_t length_b){
515 while(length_b>=BMW_SMALL_BLOCKSIZE){
516 bmw_small_nextBlock(&ctx, msg);
517 length_b -= BMW_SMALL_BLOCKSIZE;
518 msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
520 bmw_small_lastBlock(&ctx, msg, length_b);
521 bmw224_ctx2hash(dest, &ctx);
524 void bmw256(void* dest, const void* msg, uint32_t length_b){
527 while(length_b>=BMW_SMALL_BLOCKSIZE){
528 bmw_small_nextBlock(&ctx, msg);
529 length_b -= BMW_SMALL_BLOCKSIZE;
530 msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
532 bmw_small_lastBlock(&ctx, msg, length_b);
533 bmw256_ctx2hash(dest, &ctx);