3 This file is part of the ARM-Crypto-Lib.
4 Copyright (C) 2006-2010 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 "bmw_small.h"
33 #define SHL32(a,n) ((a)<<(n))
34 #define SHR32(a,n) ((a)>>(n))
35 #define ROTL32(a,n) (((a)<<(n))|((a)>>(32-(n))))
36 #define ROTR32(a,n) (((a)>>(n))|((a)<<(32-(n))))
45 void ctx_dump(const bmw_small_ctx_t* ctx){
47 cli_putstr("\r\n==== ctx dump ====");
49 cli_putstr("\r\n h[");
52 cli_hexdump_rev(&(ctx->h[i]), 4);
54 cli_putstr("\r\n counter = ");
55 cli_hexdump(&(ctx->counter), 4);
58 void dump_x(const uint32_t* q, uint8_t elements, char x){
60 cli_putstr("\r\n==== ");
62 cli_putstr(" dump ====");
63 for(i=0; i<elements;++i){
69 cli_hexdump_rev(&(q[i]), 4);
77 #define S32_0(x) ( (SHR32((x), 1)) ^ \
82 #define S32_1(x) ( (SHR32((x), 1)) ^ \
87 #define S32_2(x) ( (SHR32((x), 2)) ^ \
92 #define S32_3(x) ( (SHR32((x), 2)) ^ \
97 #define S32_4(x) ( (SHR32((x), 1)) ^ (x))
99 #define S32_5(x) ( (SHR32((x), 2)) ^ (x))
101 #define R32_1(x) (ROTL32((x), 3))
102 #define R32_2(x) (ROTL32((x), 7))
103 #define R32_3(x) (ROTL32((x), 13))
104 #define R32_4(x) (ROTL32((x), 16))
105 #define R32_5(x) (ROTR32((x), 13))
106 #define R32_6(x) (ROTR32((x), 9))
107 #define R32_7(x) (ROTR32((x), 5))
110 #define K 0x05555555L
112 uint32_t k_lut[] PROGMEM = {
113 16L*K, 17L*K, 18L*K, 19L*K, 20L*K, 21L*K, 22L*K, 23L*K,
114 24L*K, 25L*K, 26L*K, 27L*K, 28L*K, 29L*K, 30L*K, 31L*K
117 /* same as above but precomputed to avoid compiler warnings */
121 0x55555550L, 0x5aaaaaa5L, 0x5ffffffaL,
122 0x6555554fL, 0x6aaaaaa4L, 0x6ffffff9L,
123 0x7555554eL, 0x7aaaaaa3L, 0x7ffffff8L,
124 0x8555554dL, 0x8aaaaaa2L, 0x8ffffff7L,
125 0x9555554cL, 0x9aaaaaa1L, 0x9ffffff6L,
130 uint32_t bmw_small_expand1(uint8_t j, const uint32_t* q, const void* m, const void* h){
132 r = ( ROTL32(((uint32_t*)m)[j], ((j+0))+1 )
133 + ROTL32(((uint32_t*)m)[(j+3)], ((j+3))+1 )
134 - ROTL32(((uint32_t*)m)[(j+10)], ((j+10))+1 )
136 ) ^ ((uint32_t*)h)[(j+7)];
137 r += S32_1(q[j+ 0]) + S32_2(q[j+ 1]) + S32_3(q[j+ 2]) + S32_0(q[j+ 3]) +
138 S32_1(q[j+ 4]) + S32_2(q[j+ 5]) + S32_3(q[j+ 6]) + S32_0(q[j+ 7]) +
139 S32_1(q[j+ 8]) + S32_2(q[j+ 9]) + S32_3(q[j+10]) + S32_0(q[j+11]) +
140 S32_1(q[j+12]) + S32_2(q[j+13]) + S32_3(q[j+14]) + S32_0(q[j+15]);
146 uint32_t bmw_small_expand2(uint8_t j, const uint32_t* q, const void* m, const void* h){
148 r = ( ROTL32(((uint32_t*)m)[j&0xf], ((j+0)&0xf)+1 )
149 + ROTL32(((uint32_t*)m)[(j+3)&0xf], ((j+3)&0xf)+1 )
150 - ROTL32(((uint32_t*)m)[(j+10)&0xf], ((j+10)&0xf)+1 )
152 ) ^ ((uint32_t*)h)[(j+7)&0xf];
153 r += (q[j+ 0]) + R32_1(q[j+ 1]) + (q[j+ 2]) + R32_2(q[j+ 3]) +
154 (q[j+ 4]) + R32_3(q[j+ 5]) + (q[j+ 6]) + R32_4(q[j+ 7]) +
155 (q[j+ 8]) + R32_5(q[j+ 9]) + (q[j+10]) + R32_6(q[j+11]) +
156 (q[j+12]) + R32_7(q[j+13]) + S32_4(q[j+14]) + S32_5(q[j+15]);
161 void bmw_small_f1(uint32_t* q, const void* m, const void* h);
164 void bmw_small_f0(uint32_t* q, uint32_t* h, const uint32_t* m){
183 q[ 0] = (h[ 5] - h[ 7] + h[10] + h[13] + h[14]);
184 q[ 3] = (h[ 0] - h[ 1] + h[ 8] - h[10] + h[13]);
185 q[ 6] = (h[ 4] - h[ 0] - h[ 3] - h[11] + h[13]);
186 q[ 9] = (h[ 0] - h[ 3] + h[ 6] - h[ 7] + h[14]);
187 q[12] = (h[ 1] + h[ 3] - h[ 6] - h[ 9] + h[10]);
188 q[15] = (h[12] - h[ 4] - h[ 6] - h[ 9] + h[13]);
189 q[ 2] = (h[ 0] + h[ 7] + h[ 9] - h[12] + h[15]);
190 q[ 5] = (h[ 3] - h[ 2] + h[10] - h[12] + h[15]);
191 q[ 8] = (h[ 2] - h[ 5] - h[ 6] + h[13] - h[15]);
192 q[11] = (h[ 8] - h[ 0] - h[ 2] - h[ 5] + h[ 9]);
193 q[14] = (h[ 3] - h[ 5] + h[ 8] - h[11] - h[12]);
194 q[ 1] = (h[ 6] - h[ 8] + h[11] + h[14] - h[15]);
195 q[ 4] = (h[ 1] + h[ 2] + h[ 9] - h[11] - h[14]);
196 q[ 7] = (h[ 1] - h[ 4] - h[ 5] - h[12] - h[14]);
197 q[10] = (h[ 8] - h[ 1] - h[ 4] - h[ 7] + h[15]);
198 q[13] = (h[ 2] + h[ 4] + h[ 7] + h[10] + h[11]);
201 q[ 0] = S32_0(q[ 0]); q[ 1] = S32_1(q[ 1]); q[ 2] = S32_2(q[ 2]); q[ 3] = S32_3(q[ 3]); q[ 4] = S32_4(q[ 4]);
202 q[ 5] = S32_0(q[ 5]); q[ 6] = S32_1(q[ 6]); q[ 7] = S32_2(q[ 7]); q[ 8] = S32_3(q[ 8]); q[ 9] = S32_4(q[ 9]);
203 q[10] = S32_0(q[10]); q[11] = S32_1(q[11]); q[12] = S32_2(q[12]); q[13] = S32_3(q[13]); q[14] = S32_4(q[14]);
204 q[15] = S32_0(q[15]);
205 q[ 0] += h[ 1] ^= m[ 1];
206 q[ 1] += h[ 2] ^= m[ 2];
207 q[ 2] += h[ 3] ^= m[ 3];
208 q[ 3] += h[ 4] ^= m[ 4];
209 q[ 4] += h[ 5] ^= m[ 5];
210 q[ 5] += h[ 6] ^= m[ 6];
211 q[ 6] += h[ 7] ^= m[ 7];
212 q[ 7] += h[ 8] ^= m[ 8];
213 q[ 8] += h[ 9] ^= m[ 9];
214 q[ 9] += h[10] ^= m[10];
215 q[10] += h[11] ^= m[11];
216 q[11] += h[12] ^= m[12];
217 q[12] += h[13] ^= m[13];
218 q[13] += h[14] ^= m[14];
219 q[14] += h[15] ^= m[15];
220 q[15] += h[ 0] ^= m[ 0];
222 q[ 0] = S32_0(q[ 0]) + (h[ 1] ^= m[ 1]);
223 q[ 1] = S32_1(q[ 1]) + (h[ 2] ^= m[ 2]);
224 q[ 2] = S32_2(q[ 2]) + (h[ 3] ^= m[ 3]);
225 q[ 3] = S32_3(q[ 3]) + (h[ 4] ^= m[ 4]);
226 q[ 4] = S32_4(q[ 4]) + (h[ 5] ^= m[ 5]);
227 q[ 5] = S32_0(q[ 5]) + (h[ 6] ^= m[ 6]);
228 q[ 6] = S32_1(q[ 6]) + (h[ 7] ^= m[ 7]);
229 q[ 7] = S32_2(q[ 7]) + (h[ 8] ^= m[ 8]);
230 q[ 8] = S32_3(q[ 8]) + (h[ 9] ^= m[ 9]);
231 q[ 9] = S32_4(q[ 9]) + (h[10] ^= m[10]);
232 q[10] = S32_0(q[10]) + (h[11] ^= m[11]);
233 q[11] = S32_1(q[11]) + (h[12] ^= m[12]);
234 q[12] = S32_2(q[12]) + (h[13] ^= m[13]);
235 q[13] = S32_3(q[13]) + (h[14] ^= m[14]);
236 q[14] = S32_4(q[14]) + (h[15] ^= m[15]);
237 q[15] = S32_0(q[15]) + (h[ 0] ^= m[ 0]);
241 void bmw_small_f2(uint32_t* h, uint32_t* q, const uint32_t* m){
243 xl = q[16] ^ q[17] ^ q[18] ^ q[19] ^ q[20] ^ q[21] ^ q[22] ^ q[23];
244 xh = xl ^ q[24] ^ q[25] ^ q[26] ^ q[27] ^ q[28] ^ q[29] ^ q[30] ^ q[31];
246 cli_putstr("\r\n XL = ");
247 cli_hexdump_rev(&xl, 4);
248 cli_putstr("\r\n XH = ");
249 cli_hexdump_rev(&xh, 4);
252 h[0] = (SHL32(xh, 5) ^ SHR32(q[16], 5) ^ m[ 0]) + (xl ^ q[24] ^ q[ 0]);
253 h[1] = (SHR32(xh, 7) ^ SHL32(q[17], 8) ^ m[ 1]) + (xl ^ q[25] ^ q[ 1]);
254 h[2] = (SHR32(xh, 5) ^ SHL32(q[18], 5) ^ m[ 2]) + (xl ^ q[26] ^ q[ 2]);
255 h[3] = (SHR32(xh, 1) ^ SHL32(q[19], 5) ^ m[ 3]) + (xl ^ q[27] ^ q[ 3]);
256 h[4] = (SHR32(xh, 3) ^ q[20] ^ m[ 4]) + (xl ^ q[28] ^ q[ 4]);
257 h[5] = (SHL32(xh, 6) ^ SHR32(q[21], 6) ^ m[ 5]) + (xl ^ q[29] ^ q[ 5]);
258 h[6] = (SHR32(xh, 4) ^ SHL32(q[22], 6) ^ m[ 6]) + (xl ^ q[30] ^ q[ 6]);
259 h[7] = (SHR32(xh,11) ^ SHL32(q[23], 2) ^ m[ 7]) + (xl ^ q[31] ^ q[ 7]);
261 h[ 8] = ROTL32(h[4], 9) + (xh ^ q[24] ^ m[ 8]) + (SHL32(xl, 8) ^ q[23] ^ q[ 8]);
262 h[ 9] = ROTL32(h[5], 10) + (xh ^ q[25] ^ m[ 9]) + (SHR32(xl, 6) ^ q[16] ^ q[ 9]);
263 h[10] = ROTL32(h[6], 11) + (xh ^ q[26] ^ m[10]) + (SHL32(xl, 6) ^ q[17] ^ q[10]);
264 h[11] = ROTL32(h[7], 12) + (xh ^ q[27] ^ m[11]) + (SHL32(xl, 4) ^ q[18] ^ q[11]);
265 h[12] = ROTL32(h[0], 13) + (xh ^ q[28] ^ m[12]) + (SHR32(xl, 3) ^ q[19] ^ q[12]);
266 h[13] = ROTL32(h[1], 14) + (xh ^ q[29] ^ m[13]) + (SHR32(xl, 4) ^ q[20] ^ q[13]);
267 h[14] = ROTL32(h[2], 15) + (xh ^ q[30] ^ m[14]) + (SHR32(xl, 7) ^ q[21] ^ q[14]);
268 h[15] = ROTL32(h[3], 16) + (xh ^ q[31] ^ m[15]) + (SHR32(xl, 2) ^ q[22] ^ q[15]);
271 void bmw_small_nextBlock(bmw_small_ctx_t* ctx, const void* block){
273 dump_x(block, 16, 'M');
274 bmw_small_f0(q, ctx->h, block);
276 bmw_small_f1(q, block, ctx->h);
277 dump_x(q+16, 16, 'Q');
278 bmw_small_f2(ctx->h, q, block);
283 void bmw_small_lastBlock(bmw_small_ctx_t* ctx, const void* block, uint16_t length_b){
285 while(length_b >= BMW_SMALL_BLOCKSIZE){
286 bmw_small_nextBlock(ctx, block);
287 length_b -= BMW_SMALL_BLOCKSIZE;
288 block = (uint8_t*)block + BMW_SMALL_BLOCKSIZE_B;
290 memset(buffer, 0, 64);
291 memcpy(buffer, block, (length_b+7)/8);
292 buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
293 if(length_b+1>64*8-64){
294 bmw_small_nextBlock(ctx, buffer);
295 memset(buffer, 0, 64-8);
298 *((uint64_t*)&(buffer[64-8])) = (uint64_t)(ctx->counter*512LL)+(uint64_t)length_b;
299 bmw_small_nextBlock(ctx, buffer);
302 memset(buffer, 0xaa, 64);
304 buffer[i*4] = i+0xa0;
306 // dump_x(buffer, 16, 'A');
307 dump_x(ctx->h, 16, 'M');
308 bmw_small_f0(q, (uint32_t*)buffer, ctx->h);
309 dump_x(buffer, 16, 'a');
311 bmw_small_f1(q, ctx->h, (uint32_t*)buffer);
313 bmw_small_f2((uint32_t*)buffer, q, ctx->h);
314 memcpy(ctx->h, buffer, 64);
317 void bmw224_init(bmw224_ctx_t* ctx){
319 ctx->h[0] = 0x00010203;
321 ctx->h[i] = ctx->h[i-1]+ 0x04040404;
327 void bmw256_init(bmw256_ctx_t* ctx){
329 ctx->h[0] = 0x40414243;
331 ctx->h[i] = ctx->h[i-1]+ 0x04040404;
337 void bmw224_nextBlock(bmw224_ctx_t* ctx, const void* block){
338 bmw_small_nextBlock(ctx, block);
341 void bmw256_nextBlock(bmw256_ctx_t* ctx, const void* block){
342 bmw_small_nextBlock(ctx, block);
345 void bmw224_lastBlock(bmw224_ctx_t* ctx, const void* block, uint16_t length_b){
346 bmw_small_lastBlock(ctx, block, length_b);
349 void bmw256_lastBlock(bmw256_ctx_t* ctx, const void* block, uint16_t length_b){
350 bmw_small_lastBlock(ctx, block, length_b);
353 void bmw224_ctx2hash(void* dest, const bmw224_ctx_t* ctx){
354 memcpy(dest, &(ctx->h[9]), 224/8);
357 void bmw256_ctx2hash(void* dest, const bmw256_ctx_t* ctx){
358 memcpy(dest, &(ctx->h[8]), 256/8);
361 void bmw224(void* dest, const void* msg, uint32_t length_b){
364 while(length_b>=BMW_SMALL_BLOCKSIZE){
365 bmw_small_nextBlock(&ctx, msg);
366 length_b -= BMW_SMALL_BLOCKSIZE;
367 msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
369 bmw_small_lastBlock(&ctx, msg, length_b);
370 bmw224_ctx2hash(dest, &ctx);
373 void bmw256(void* dest, const void* msg, uint32_t length_b){
376 while(length_b>=BMW_SMALL_BLOCKSIZE){
377 bmw_small_nextBlock(&ctx, msg);
378 length_b -= BMW_SMALL_BLOCKSIZE;
379 msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
381 bmw_small_lastBlock(&ctx, msg, length_b);
382 bmw256_ctx2hash(dest, &ctx);
385 #include "f1_autogen.c"