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_large.h"
33 #define SHL64(a,n) ((a)<<(n))
34 #define SHR64(a,n) ((a)>>(n))
35 #define ROTL64(a,n) (((a)<<(n))|((a)>>(64-(n))))
36 #define ROTR64(a,n) (((a)>>(n))|((a)<<(64-(n))))
44 void ctx_dump(const bmw_large_ctx_t* ctx){
46 cli_putstr_P(PSTR("\r\n==== ctx dump ===="));
48 cli_putstr_P(PSTR("\r\n h["));
50 cli_putstr_P(PSTR("] = "));
51 cli_hexdump_rev(&(ctx->h[i]), 8);
53 cli_putstr_P(PSTR("\r\n counter = "));
54 cli_hexdump(&(ctx->counter), 4);
57 void dump_x(const uint64_t* q, uint8_t elements, char x){
59 cli_putstr_P(PSTR("\r\n==== "));
61 cli_putstr_P(PSTR(" dump ===="));
62 for(i=0; i<elements;++i){
63 cli_putstr_P(PSTR("\r\n "));
65 cli_putstr_P(PSTR("["));
67 cli_putstr_P(PSTR("] = "));
68 cli_hexdump_rev(&(q[i]), 8);
76 uint64_t bmw_large_s0(uint64_t x){
85 uint64_t bmw_large_s1(uint64_t x){
94 uint64_t bmw_large_s2(uint64_t x){
103 uint64_t bmw_large_s3(uint64_t x){
112 uint64_t bmw_large_s4(uint64_t x){
119 uint64_t bmw_large_s5(uint64_t x){
126 uint64_t bmw_large_r1(uint64_t x){
132 uint64_t bmw_large_r2(uint64_t x){
138 uint64_t bmw_large_r3(uint64_t x){
144 uint64_t bmw_large_r4(uint64_t x){
150 uint64_t bmw_large_r5(uint64_t x){
152 r = ROTR64(x, 64-37);
156 uint64_t bmw_large_r6(uint64_t x){
158 r = ROTR64(x, 64-43);
162 uint64_t bmw_large_r7(uint64_t x){
164 r = ROTR64(x, 64-53);
168 #define K 0x0555555555555555LL
169 #define MASK 0xFFFFFFFFFFFFFFFFLL
171 uint64_t k_lut[] PROGMEM = {
172 16LL*K, 17LL*K, 18LL*K, 19LL*K,
173 20LL*K, 21LL*K, 22LL*K, 23LL*K,
174 24LL*K, 25LL*K, 26LL*K, 27LL*K,
175 28LL*K, 29LL*K, 30LL*K, 31LL*K };
177 /* the same as above but precomputed to avoid compiler warnings */
179 uint64_t k_lut[] PROGMEM = {
180 0x5555555555555550LL, 0x5aaaaaaaaaaaaaa5LL, 0x5ffffffffffffffaLL,
181 0x655555555555554fLL, 0x6aaaaaaaaaaaaaa4LL, 0x6ffffffffffffff9LL,
182 0x755555555555554eLL, 0x7aaaaaaaaaaaaaa3LL, 0x7ffffffffffffff8LL,
183 0x855555555555554dLL, 0x8aaaaaaaaaaaaaa2LL, 0x8ffffffffffffff7LL,
184 0x955555555555554cLL, 0x9aaaaaaaaaaaaaa1LL, 0x9ffffffffffffff6LL,
185 0xa55555555555554bLL };
187 uint64_t bmw_large_expand1(uint8_t j, const uint64_t* q, const void* m){
188 uint64_t(*s[])(uint64_t) = {bmw_large_s1, bmw_large_s2, bmw_large_s3, bmw_large_s0};
194 /* r = 0x0555555555555555LL*(j+16); */
195 r.v32[0] = pgm_read_dword(((uint8_t*)k_lut+8*j));
196 r.v32[1] = pgm_read_dword(((uint8_t*)k_lut+8*j+4));
198 r.v64 += s[i%4](q[j+i]);
200 r.v64 += ((uint64_t*)m)[j];
201 r.v64 += ((uint64_t*)m)[j+3];
202 r.v64 -= ((uint64_t*)m)[j+10];
206 uint64_t bmw_large_expand2(uint8_t j, const uint64_t* q, const void* m){
207 uint64_t(*rf[])(uint64_t) = {bmw_large_r1, bmw_large_r2, bmw_large_r3,
208 bmw_large_r4, bmw_large_r5, bmw_large_r6,
215 /* r = 0x0555555555555555LL*(j+16); */
216 r.v32[0] = pgm_read_dword(((uint8_t*)k_lut+8*j));
217 r.v32[1] = pgm_read_dword(((uint8_t*)k_lut+8*j+4));
218 for(i=0; i<14; i+=2){
221 for(i=0; i<14; i+=2){
222 r.v64 += rf[i/2](q[j+i+1]);
224 r.v64 += bmw_large_s5(q[j+14]);
225 r.v64 += bmw_large_s4(q[j+15]);
226 r.v64 += ((uint64_t*)m)[j];
227 r.v64 += ((uint64_t*)m)[(j+3)%16];
228 r.v64 -= ((uint64_t*)m)[(j+10)%16];
234 uint8_t f0_lut[] PROGMEM ={
235 5<<1, ( 7<<1)+1, (10<<1)+0, (13<<1)+0, (14<<1)+0,
236 6<<1, ( 8<<1)+1, (11<<1)+0, (14<<1)+0, (15<<1)+1,
237 0<<1, ( 7<<1)+0, ( 9<<1)+0, (12<<1)+1, (15<<1)+0,
238 0<<1, ( 1<<1)+1, ( 8<<1)+0, (10<<1)+1, (13<<1)+0,
239 1<<1, ( 2<<1)+0, ( 9<<1)+0, (11<<1)+1, (14<<1)+1,
240 3<<1, ( 2<<1)+1, (10<<1)+0, (12<<1)+1, (15<<1)+0,
241 4<<1, ( 0<<1)+1, ( 3<<1)+1, (11<<1)+1, (13<<1)+0,
242 1<<1, ( 4<<1)+1, ( 5<<1)+1, (12<<1)+1, (14<<1)+1,
243 2<<1, ( 5<<1)+1, ( 6<<1)+1, (13<<1)+0, (15<<1)+1,
244 0<<1, ( 3<<1)+1, ( 6<<1)+0, ( 7<<1)+1, (14<<1)+0,
245 8<<1, ( 1<<1)+1, ( 4<<1)+1, ( 7<<1)+1, (15<<1)+0,
246 8<<1, ( 0<<1)+1, ( 2<<1)+1, ( 5<<1)+1, ( 9<<1)+0,
247 1<<1, ( 3<<1)+0, ( 6<<1)+1, ( 9<<1)+1, (10<<1)+0,
248 2<<1, ( 4<<1)+0, ( 7<<1)+0, (10<<1)+0, (11<<1)+0,
249 3<<1, ( 5<<1)+1, ( 8<<1)+0, (11<<1)+1, (12<<1)+1,
250 12<<1, ( 4<<1)+1, ( 6<<1)+1, ( 9<<1)+1, (13<<1)+0
253 void bmw_large_f0(uint64_t* q, uint64_t* h, const void* m){
254 uint8_t i,j=-1,v,sign,l=0;
255 uint64_t(*s[])(uint64_t)={ bmw_large_s0, bmw_large_s1, bmw_large_s2,
256 bmw_large_s3, bmw_large_s4 };
258 h[i] ^= ((uint64_t*)m)[i];
261 // memset(q, 0, 4*16);
262 for(i=0; i<5*16; ++i){
263 v = pgm_read_byte(f0_lut+i);
285 void bmw_large_f0(uint64_t* q, uint64_t* h, const void* m){
287 uint64_t(*s[])(uint64_t)={ bmw_large_s0, bmw_large_s1, bmw_large_s2,
288 bmw_large_s3, bmw_large_s4 };
290 h[i] ^= ((uint64_t*)m)[i];
293 q[ 0] = (h[ 5] - h[ 7] + h[10] + h[13] + h[14]);
294 q[ 1] = (h[ 6] - h[ 8] + h[11] + h[14] - h[15]);
295 q[ 2] = (h[ 0] + h[ 7] + h[ 9] - h[12] + h[15]);
296 q[ 3] = (h[ 0] - h[ 1] + h[ 8] - h[10] + h[13]);
297 q[ 4] = (h[ 1] + h[ 2] + h[ 9] - h[11] - h[14]);
298 q[ 5] = (h[ 3] - h[ 2] + h[10] - h[12] + h[15]);
299 q[ 6] = (h[ 4] - h[ 0] - h[ 3] - h[11] + h[13]);
300 q[ 7] = (h[ 1] - h[ 4] - h[ 5] - h[12] - h[14]);
301 q[ 8] = (h[ 2] - h[ 5] - h[ 6] + h[13] - h[15]);
302 q[ 9] = (h[ 0] - h[ 3] + h[ 6] - h[ 7] + h[14]);
303 q[10] = (h[ 8] - h[ 1] - h[ 4] - h[ 7] + h[15]);
304 q[11] = (h[ 8] - h[ 0] - h[ 2] - h[ 5] + h[ 9]);
305 q[12] = (h[ 1] + h[ 3] - h[ 6] - h[ 9] + h[10]);
306 q[13] = (h[ 2] + h[ 4] + h[ 7] + h[10] + h[11]);
307 q[14] = (h[ 3] - h[ 5] + h[ 8] - h[11] - h[12]);
308 q[15] = (h[12] - h[ 4] - h[ 6] - h[ 9] + h[13]);
316 void bmw_large_f1(uint64_t* q, const void* m){
318 q[16] = bmw_large_expand1(0, q, m);
319 q[17] = bmw_large_expand1(1, q, m);
321 q[16+i] = bmw_large_expand2(i, q, m);
325 void bmw_large_f2(uint64_t* h, const uint64_t* q, const void* m){
336 cli_putstr_P(PSTR("\r\n XL = "));
337 cli_hexdump_rev(&xl, 4);
338 cli_putstr_P(PSTR("\r\n XH = "));
339 cli_hexdump_rev(&xh, 4);
342 h[0] ^= SHL64(xh, 5) ^ SHR64(q[16], 5);
343 h[1] ^= SHR64(xh, 7) ^ SHL64(q[17], 8);
344 h[2] ^= SHR64(xh, 5) ^ SHL64(q[18], 5);
345 h[3] ^= SHR64(xh, 1) ^ SHL64(q[19], 5);
346 h[4] ^= SHR64(xh, 3) ^ q[20];
347 h[5] ^= SHL64(xh, 6) ^ SHR64(q[21], 6);
348 h[6] ^= SHR64(xh, 4) ^ SHL64(q[22], 6);
349 h[7] ^= SHR64(xh,11) ^ SHL64(q[23], 2);
351 h[i] += xl ^ q[24+i] ^ q[i];
354 h[8+i] ^= xh ^ q[24+i];
355 h[8+i] += ROTL64(h[(4+i)%8],i+9);
357 h[ 8] += SHL64(xl, 8) ^ q[23] ^ q[ 8];
358 h[ 9] += SHR64(xl, 6) ^ q[16] ^ q[ 9];
359 h[10] += SHL64(xl, 6) ^ q[17] ^ q[10];
360 h[11] += SHL64(xl, 4) ^ q[18] ^ q[11];
361 h[12] += SHR64(xl, 3) ^ q[19] ^ q[12];
362 h[13] += SHR64(xl, 4) ^ q[20] ^ q[13];
363 h[14] += SHR64(xl, 7) ^ q[21] ^ q[14];
364 h[15] += SHR64(xl, 2) ^ q[22] ^ q[15];
367 void bmw_large_nextBlock(bmw_large_ctx_t* ctx, const void* block){
369 dump_x(block, 16, 'M');
370 bmw_large_f0(q, ctx->h, block);
372 bmw_large_f1(q, block);
374 bmw_large_f2(ctx->h, q, block);
379 void bmw_large_lastBlock(bmw_large_ctx_t* ctx, const void* block, uint16_t length_b){
381 while(length_b >= BMW_LARGE_BLOCKSIZE){
382 bmw_large_nextBlock(ctx, block);
383 length_b -= BMW_LARGE_BLOCKSIZE;
384 block = (uint8_t*)block + BMW_LARGE_BLOCKSIZE_B;
386 memset(buffer, 0, 128);
387 memcpy(buffer, block, (length_b+7)/8);
388 buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
389 if(length_b+1>128*8-64){
390 bmw_large_nextBlock(ctx, buffer);
391 memset(buffer, 0, 128-8);
394 *((uint64_t*)&(buffer[128-8])) = (uint64_t)(ctx->counter*1024LL)+(uint64_t)length_b;
395 bmw_large_nextBlock(ctx, buffer);
398 void bmw384_init(bmw384_ctx_t* ctx){
400 ctx->h[0] = 0x0001020304050607LL;
402 ctx->h[i] = ctx->h[i-1]+ 0x0808080808080808LL;
405 ctx->h[6] = 0x3031323324353637LL;
411 void bmw512_init(bmw512_ctx_t* ctx){
413 ctx->h[0] = 0x8081828384858687LL;
415 ctx->h[i] = ctx->h[i-1]+ 0x0808080808080808LL;
421 void bmw384_nextBlock(bmw384_ctx_t* ctx, const void* block){
422 bmw_large_nextBlock(ctx, block);
425 void bmw512_nextBlock(bmw512_ctx_t* ctx, const void* block){
426 bmw_large_nextBlock(ctx, block);
429 void bmw384_lastBlock(bmw384_ctx_t* ctx, const void* block, uint16_t length_b){
430 bmw_large_lastBlock(ctx, block, length_b);
433 void bmw512_lastBlock(bmw512_ctx_t* ctx, const void* block, uint16_t length_b){
434 bmw_large_lastBlock(ctx, block, length_b);
437 void bmw384_ctx2hash(void* dest, const bmw384_ctx_t* ctx){
438 memcpy(dest, &(ctx->h[10]), 384/8);
441 void bmw512_ctx2hash(void* dest, const bmw512_ctx_t* ctx){
442 memcpy(dest, &(ctx->h[8]), 512/8);
445 void bmw384(void* dest, const void* msg, uint32_t length_b){
448 while(length_b>=BMW_LARGE_BLOCKSIZE){
449 bmw_large_nextBlock(&ctx, msg);
450 length_b -= BMW_LARGE_BLOCKSIZE;
451 msg = (uint8_t*)msg + BMW_LARGE_BLOCKSIZE_B;
453 bmw_large_lastBlock(&ctx, msg, length_b);
454 bmw384_ctx2hash(dest, &ctx);
457 void bmw512(void* dest, const void* msg, uint32_t length_b){
460 while(length_b>=BMW_LARGE_BLOCKSIZE){
461 bmw_large_nextBlock(&ctx, msg);
462 length_b -= BMW_LARGE_BLOCKSIZE;
463 msg = (uint8_t*)msg + BMW_LARGE_BLOCKSIZE_B;
465 bmw_large_lastBlock(&ctx, msg, length_b);
466 bmw512_ctx2hash(dest, &ctx);