1 /* bmw_large_speed.c */
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_large.h"
32 #define SHL64(a,n) ((a)<<(n))
33 #define SHR64(a,n) ((a)>>(n))
34 #define ROTL64(a,n) (((a)<<(n))|((a)>>(64-(n))))
35 #define ROTR64(a,n) (((a)>>(n))|((a)<<(64-(n))))
43 void ctx_dump(const bmw_large_ctx_t* ctx){
45 cli_putstr("\r\n==== ctx dump ====");
47 cli_putstr("\r\n h[");
50 cli_hexdump_rev(&(ctx->h[i]), 8);
52 cli_putstr("\r\n counter = ");
53 cli_hexdump(&(ctx->counter), 4);
56 void dump_x(const uint64_t* q, uint8_t elements, char x){
58 cli_putstr("\r\n==== ");
60 cli_putstr(" dump ====");
61 for(i=0; i<elements;++i){
67 cli_hexdump_rev(&(q[i]), 8);
75 #define S64_0(x) ( (SHR64((x), 1)) ^ \
80 #define S64_1(x) ( (SHR64((x), 1)) ^ \
85 #define S64_2(x) ( (SHR64((x), 2)) ^ \
90 #define S64_3(x) ( (SHR64((x), 2)) ^ \
95 #define S64_4(x) ( (SHR64((x), 1)) ^ x)
97 #define S64_5(x) ( (SHR64((x), 2)) ^ x)
99 #define R64_1(x) (ROTL64((x), 5))
100 #define R64_2(x) (ROTL64((x), 11))
101 #define R64_3(x) (ROTL64((x), 27))
102 #define R64_4(x) (ROTL64((x), 32))
103 #define R64_5(x) (ROTR64((x), 27))
104 #define R64_6(x) (ROTR64((x), 21))
105 #define R64_7(x) (ROTR64((x), 11))
107 #include "f1_autogen_large.c"
110 void bmw_large_f0(uint64_t* q, uint64_t* h, const uint64_t* m){
127 // dump_x(t, 16, 'T');
128 q[ 0] = (h[ 5] - h[ 7] + h[10] + h[13] + h[14]);
129 q[ 1] = (h[ 6] - h[ 8] + h[11] + h[14] - h[15]);
130 q[ 2] = (h[ 0] + h[ 7] + h[ 9] - h[12] + h[15]);
131 q[ 3] = (h[ 0] - h[ 1] + h[ 8] - h[10] + h[13]);
132 q[ 4] = (h[ 1] + h[ 2] + h[ 9] - h[11] - h[14]);
133 q[ 5] = (h[ 3] - h[ 2] + h[10] - h[12] + h[15]);
134 q[ 6] = (h[ 4] - h[ 0] - h[ 3] - h[11] + h[13]);
135 q[ 7] = (h[ 1] - h[ 4] - h[ 5] - h[12] - h[14]);
136 q[ 8] = (h[ 2] - h[ 5] - h[ 6] + h[13] - h[15]);
137 q[ 9] = (h[ 0] - h[ 3] + h[ 6] - h[ 7] + h[14]);
138 q[10] = (h[ 8] - h[ 1] - h[ 4] - h[ 7] + h[15]);
139 q[11] = (h[ 8] - h[ 0] - h[ 2] - h[ 5] + h[ 9]);
140 q[12] = (h[ 1] + h[ 3] - h[ 6] - h[ 9] + h[10]);
141 q[13] = (h[ 2] + h[ 4] + h[ 7] + h[10] + h[11]);
142 q[14] = (h[ 3] - h[ 5] + h[ 8] - h[11] - h[12]);
143 q[15] = (h[12] - h[ 4] - h[ 6] - h[ 9] + h[13]);
146 q[ 0] = S64_0(q[ 0]) + (h[ 1] ^= m[ 1]);
147 q[ 1] = S64_1(q[ 1]) + (h[ 2] ^= m[ 2]);
148 q[ 2] = S64_2(q[ 2]) + (h[ 3] ^= m[ 3]);
149 q[ 3] = S64_3(q[ 3]) + (h[ 4] ^= m[ 4]);
150 q[ 4] = S64_4(q[ 4]) + (h[ 5] ^= m[ 5]);
151 q[ 5] = S64_0(q[ 5]) + (h[ 6] ^= m[ 6]);
152 q[ 6] = S64_1(q[ 6]) + (h[ 7] ^= m[ 7]);
153 q[ 7] = S64_2(q[ 7]) + (h[ 8] ^= m[ 8]);
154 q[ 8] = S64_3(q[ 8]) + (h[ 9] ^= m[ 9]);
155 q[ 9] = S64_4(q[ 9]) + (h[10] ^= m[10]);
156 q[10] = S64_0(q[10]) + (h[11] ^= m[11]);
157 q[11] = S64_1(q[11]) + (h[12] ^= m[12]);
158 q[12] = S64_2(q[12]) + (h[13] ^= m[13]);
159 q[13] = S64_3(q[13]) + (h[14] ^= m[14]);
160 q[14] = S64_4(q[14]) + (h[15] ^= m[15]);
161 q[15] = S64_0(q[15]) + (h[ 0] ^= m[ 0]);
165 void bmw_large_f2(uint64_t* h, const uint64_t* q, const uint64_t* m){
168 xl = q[16] ^ q[17] ^ q[18] ^ q[19] ^ q[20] ^ q[21] ^ q[22] ^ q[23];
169 xh = xl ^ q[24] ^ q[25] ^ q[26] ^ q[27] ^ q[28] ^ q[29] ^ q[30] ^ q[31];
171 cli_putstr("\r\n XL = ");
172 cli_hexdump_rev(&xl, 8);
173 cli_putstr("\r\n XH = ");
174 cli_hexdump_rev(&xh, 8);
177 h[0] = (SHL64(xh, 5) ^ SHR64(q[16], 5) ^ m[ 0]) + (xl ^ q[24] ^ q[ 0]);
178 h[1] = (SHR64(xh, 7) ^ SHL64(q[17], 8) ^ m[ 1]) + (xl ^ q[25] ^ q[ 1]);
179 h[2] = (SHR64(xh, 5) ^ SHL64(q[18], 5) ^ m[ 2]) + (xl ^ q[26] ^ q[ 2]);
180 h[3] = (SHR64(xh, 1) ^ SHL64(q[19], 5) ^ m[ 3]) + (xl ^ q[27] ^ q[ 3]);
181 h[4] = (SHR64(xh, 3) ^ q[20] ^ m[ 4]) + (xl ^ q[28] ^ q[ 4]);
182 h[5] = (SHL64(xh, 6) ^ SHR64(q[21], 6) ^ m[ 5]) + (xl ^ q[29] ^ q[ 5]);
183 h[6] = (SHR64(xh, 4) ^ SHL64(q[22], 6) ^ m[ 6]) + (xl ^ q[30] ^ q[ 6]);
184 h[7] = (SHR64(xh,11) ^ SHL64(q[23], 2) ^ m[ 7]) + (xl ^ q[31] ^ q[ 7]);
186 h[ 8] = ROTL64(h[4], 9) + (xh ^ q[24] ^ m[ 8]) + (SHL64(xl, 8) ^ q[23] ^ q[ 8]);
187 h[ 9] = ROTL64(h[5], 10) + (xh ^ q[25] ^ m[ 9]) + (SHR64(xl, 6) ^ q[16] ^ q[ 9]);
188 h[10] = ROTL64(h[6], 11) + (xh ^ q[26] ^ m[10]) + (SHL64(xl, 6) ^ q[17] ^ q[10]);
189 h[11] = ROTL64(h[7], 12) + (xh ^ q[27] ^ m[11]) + (SHL64(xl, 4) ^ q[18] ^ q[11]);
190 h[12] = ROTL64(h[0], 13) + (xh ^ q[28] ^ m[12]) + (SHR64(xl, 3) ^ q[19] ^ q[12]);
191 h[13] = ROTL64(h[1], 14) + (xh ^ q[29] ^ m[13]) + (SHR64(xl, 4) ^ q[20] ^ q[13]);
192 h[14] = ROTL64(h[2], 15) + (xh ^ q[30] ^ m[14]) + (SHR64(xl, 7) ^ q[21] ^ q[14]);
193 h[15] = ROTL64(h[3], 16) + (xh ^ q[31] ^ m[15]) + (SHR64(xl, 2) ^ q[22] ^ q[15]);
196 void bmw_large_nextBlock(bmw_large_ctx_t* ctx, const void* block){
198 dump_x(block, 16, 'M');
199 bmw_large_f0(q, ctx->h, block);
201 bmw_large_f1(q, block, ctx->h);
203 bmw_large_f2(ctx->h, q, block);
208 void bmw_large_lastBlock(bmw_large_ctx_t* ctx, const void* block, uint16_t length_b){
210 while(length_b >= BMW_LARGE_BLOCKSIZE){
211 bmw_large_nextBlock(ctx, block);
212 length_b -= BMW_LARGE_BLOCKSIZE;
213 block = (uint8_t*)block + BMW_LARGE_BLOCKSIZE_B;
215 memset(buffer, 0, 128);
216 memcpy(buffer, block, (length_b+7)/8);
217 buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
218 if(length_b+1>128*8-64){
219 bmw_large_nextBlock(ctx, buffer);
220 memset(buffer, 0, 128-8);
223 *((uint64_t*)&(buffer[128-8])) = (uint64_t)(ctx->counter*1024LL)+(uint64_t)length_b;
224 bmw_large_nextBlock(ctx, buffer);
227 memset(buffer, 0xaa, 128);
229 buffer[8*i] = i + 0xa0;
231 bmw_large_f0(q, (uint64_t*)buffer, ctx->h);
232 bmw_large_f1(q, ctx->h, (uint64_t*)buffer);
233 bmw_large_f2((uint64_t*)buffer, q, ctx->h);
234 memcpy(ctx->h, buffer, 128);
237 void bmw384_init(bmw384_ctx_t* ctx){
239 ctx->h[0] = 0x0001020304050607LL;
241 ctx->h[i] = ctx->h[i-1]+ 0x0808080808080808LL;
244 ctx->h[6] = 0x3031323324353637LL;
250 void bmw512_init(bmw512_ctx_t* ctx){
252 ctx->h[0] = 0x8081828384858687LL;
254 ctx->h[i] = ctx->h[i-1]+ 0x0808080808080808LL;
260 void bmw384_nextBlock(bmw384_ctx_t* ctx, const void* block){
261 bmw_large_nextBlock(ctx, block);
264 void bmw512_nextBlock(bmw512_ctx_t* ctx, const void* block){
265 bmw_large_nextBlock(ctx, block);
268 void bmw384_lastBlock(bmw384_ctx_t* ctx, const void* block, uint16_t length_b){
269 bmw_large_lastBlock(ctx, block, length_b);
272 void bmw512_lastBlock(bmw512_ctx_t* ctx, const void* block, uint16_t length_b){
273 bmw_large_lastBlock(ctx, block, length_b);
276 void bmw384_ctx2hash(void* dest, const bmw384_ctx_t* ctx){
277 memcpy(dest, &(ctx->h[10]), 384/8);
280 void bmw512_ctx2hash(void* dest, const bmw512_ctx_t* ctx){
281 memcpy(dest, &(ctx->h[8]), 512/8);
284 void bmw384(void* dest, const void* msg, uint32_t length_b){
287 while(length_b>=BMW_LARGE_BLOCKSIZE){
288 bmw_large_nextBlock(&ctx, msg);
289 length_b -= BMW_LARGE_BLOCKSIZE;
290 msg = (uint8_t*)msg + BMW_LARGE_BLOCKSIZE_B;
292 bmw_large_lastBlock(&ctx, msg, length_b);
293 bmw384_ctx2hash(dest, &ctx);
296 void bmw512(void* dest, const void* msg, uint32_t length_b){
299 while(length_b>=BMW_LARGE_BLOCKSIZE){
300 bmw_large_nextBlock(&ctx, msg);
301 length_b -= BMW_LARGE_BLOCKSIZE;
302 msg = (uint8_t*)msg + BMW_LARGE_BLOCKSIZE_B;
304 bmw_large_lastBlock(&ctx, msg, length_b);
305 bmw512_ctx2hash(dest, &ctx);