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.i"
110 void bmw_large_f0(uint64_t* q, uint64_t* h, const uint64_t* m){
112 uint64_t tr0, tr1, tr2;
113 t[ 0] = h[ 0] ^ m[ 0];
114 t[ 1] = h[ 1] ^ m[ 1];
115 t[ 2] = h[ 2] ^ m[ 2];
116 t[ 3] = h[ 3] ^ m[ 3];
117 t[ 4] = h[ 4] ^ m[ 4];
118 t[ 5] = h[ 5] ^ m[ 5];
119 t[ 6] = h[ 6] ^ m[ 6];
120 t[ 7] = h[ 7] ^ m[ 7];
121 t[ 8] = h[ 8] ^ m[ 8];
122 t[ 9] = h[ 9] ^ m[ 9];
123 t[10] = h[10] ^ m[10];
124 t[11] = h[11] ^ m[11];
125 t[12] = h[12] ^ m[12];
126 t[13] = h[13] ^ m[13];
127 t[14] = h[14] ^ m[14];
128 t[15] = h[15] ^ m[15];
129 // dump_x(t, 16, 'T');
131 q[ 0] = (t[ 5] - t[ 7] + t[10] + t[13] + t[14]);
132 q[ 1] = (t[ 6] - t[ 8] + t[11] + t[14] - t[15]);
133 q[ 2] = (t[ 0] + t[ 7] + t[ 9] - t[12] + t[15]);
134 q[ 3] = (t[ 0] - t[ 1] + t[ 8] - t[10] + t[13]);
135 q[ 4] = (t[ 1] + t[ 2] + t[ 9] - t[11] - t[14]);
136 q[ 5] = (t[ 3] - t[ 2] + t[10] - t[12] + t[15]);
137 q[ 6] = (t[ 4] - t[ 0] - t[ 3] - t[11] + t[13]);
138 q[ 7] = (t[ 1] - t[ 4] - t[ 5] - t[12] - t[14]);
139 q[ 8] = (t[ 2] - t[ 5] - t[ 6] + t[13] - t[15]);
140 q[ 9] = (t[ 0] - t[ 3] + t[ 6] - t[ 7] + t[14]);
141 q[10] = (t[ 8] - t[ 1] - t[ 4] - t[ 7] + t[15]);
142 q[11] = (t[ 8] - t[ 0] - t[ 2] - t[ 5] + t[ 9]);
143 q[12] = (t[ 1] + t[ 3] - t[ 6] - t[ 9] + t[10]);
144 q[13] = (t[ 2] + t[ 4] + t[ 7] + t[10] + t[11]);
145 q[14] = (t[ 3] - t[ 5] + t[ 8] - t[11] - t[12]);
146 q[15] = (t[12] - t[ 4] - t[ 6] - t[ 9] + t[13]);
148 q[ 0] = +t[ 5] +t[10] +t[13] +(tr1=-t[ 7]+t[14]) ;
149 q[ 3] = +t[ 8] +t[13] +t[ 0] +(tr2=-t[ 1]-t[10]) ;
150 q[ 6] = -t[11] +t[13] -t[ 0] -t[ 3] +t[ 4] ;
151 q[ 9] = +t[ 0] +(tr0=-t[ 3]+t[ 6]) +(tr1) ;
152 q[12] = -t[ 9] -(tr0) -(tr2) ;
153 q[15] = -t[ 4] +(tr0=-t[ 9]+t[12]) +(tr1=-t[ 6]+t[13]) ;
154 q[ 2] = +t[ 7] +t[15] +t[ 0] -(tr0) ;
155 q[ 5] = +t[10] +(tr0=-t[ 2]+t[15]) +(tr2=+t[ 3]-t[12]) ;
156 q[ 8] = -t[ 5] -(tr0) +(tr1) ;
157 q[11] = -t[ 0] -t[ 2] +t[ 9] +(tr0=-t[ 5]+t[ 8]) ;
158 q[14] = -t[11] +(tr0) +(tr2) ;
159 q[ 1] = +t[ 6] +(tr0=+t[11]+t[14]) +(tr1=-t[ 8]-t[15]) ;
160 q[ 4] = +t[ 9] +t[ 1] +t[ 2] -(tr0) ;
161 q[ 7] = -t[12] -t[14] +t[ 1] -t[ 4] -t[ 5] ;
162 q[10] = -t[ 1] +(tr0=-t[ 4]-t[ 7]) -(tr1) ;
163 q[13] = +t[ 2] +t[10] +t[11] -(tr0) ;
166 q[ 0] = S64_0(q[ 0]) + h[ 1];
167 q[ 1] = S64_1(q[ 1]) + h[ 2];
168 q[ 2] = S64_2(q[ 2]) + h[ 3];
169 q[ 3] = S64_3(q[ 3]) + h[ 4];
170 q[ 4] = S64_4(q[ 4]) + h[ 5];
171 q[ 5] = S64_0(q[ 5]) + h[ 6];
172 q[ 6] = S64_1(q[ 6]) + h[ 7];
173 q[ 7] = S64_2(q[ 7]) + h[ 8];
174 q[ 8] = S64_3(q[ 8]) + h[ 9];
175 q[ 9] = S64_4(q[ 9]) + h[10];
176 q[10] = S64_0(q[10]) + h[11];
177 q[11] = S64_1(q[11]) + h[12];
178 q[12] = S64_2(q[12]) + h[13];
179 q[13] = S64_3(q[13]) + h[14];
180 q[14] = S64_4(q[14]) + h[15];
181 q[15] = S64_0(q[15]) + h[ 0];
185 void bmw_large_f2(uint64_t* h, const uint64_t* q, const uint64_t* m){
188 xl = q[16] ^ q[17] ^ q[18] ^ q[19] ^ q[20] ^ q[21] ^ q[22] ^ q[23];
189 xh = xl ^ q[24] ^ q[25] ^ q[26] ^ q[27] ^ q[28] ^ q[29] ^ q[30] ^ q[31];
191 cli_putstr("\r\n XL = ");
192 cli_hexdump_rev(&xl, 8);
193 cli_putstr("\r\n XH = ");
194 cli_hexdump_rev(&xh, 8);
197 h[0] = (SHL64(xh, 5) ^ SHR64(q[16], 5) ^ m[ 0]) + (xl ^ q[24] ^ q[ 0]);
198 h[1] = (SHR64(xh, 7) ^ SHL64(q[17], 8) ^ m[ 1]) + (xl ^ q[25] ^ q[ 1]);
199 h[2] = (SHR64(xh, 5) ^ SHL64(q[18], 5) ^ m[ 2]) + (xl ^ q[26] ^ q[ 2]);
200 h[3] = (SHR64(xh, 1) ^ SHL64(q[19], 5) ^ m[ 3]) + (xl ^ q[27] ^ q[ 3]);
201 h[4] = (SHR64(xh, 3) ^ q[20] ^ m[ 4]) + (xl ^ q[28] ^ q[ 4]);
202 h[5] = (SHL64(xh, 6) ^ SHR64(q[21], 6) ^ m[ 5]) + (xl ^ q[29] ^ q[ 5]);
203 h[6] = (SHR64(xh, 4) ^ SHL64(q[22], 6) ^ m[ 6]) + (xl ^ q[30] ^ q[ 6]);
204 h[7] = (SHR64(xh,11) ^ SHL64(q[23], 2) ^ m[ 7]) + (xl ^ q[31] ^ q[ 7]);
206 h[ 8] = ROTL64(h[4], 9) + (xh ^ q[24] ^ m[ 8]) + (SHL64(xl, 8) ^ q[23] ^ q[ 8]);
207 h[ 9] = ROTL64(h[5], 10) + (xh ^ q[25] ^ m[ 9]) + (SHR64(xl, 6) ^ q[16] ^ q[ 9]);
208 h[10] = ROTL64(h[6], 11) + (xh ^ q[26] ^ m[10]) + (SHL64(xl, 6) ^ q[17] ^ q[10]);
209 h[11] = ROTL64(h[7], 12) + (xh ^ q[27] ^ m[11]) + (SHL64(xl, 4) ^ q[18] ^ q[11]);
210 h[12] = ROTL64(h[0], 13) + (xh ^ q[28] ^ m[12]) + (SHR64(xl, 3) ^ q[19] ^ q[12]);
211 h[13] = ROTL64(h[1], 14) + (xh ^ q[29] ^ m[13]) + (SHR64(xl, 4) ^ q[20] ^ q[13]);
212 h[14] = ROTL64(h[2], 15) + (xh ^ q[30] ^ m[14]) + (SHR64(xl, 7) ^ q[21] ^ q[14]);
213 h[15] = ROTL64(h[3], 16) + (xh ^ q[31] ^ m[15]) + (SHR64(xl, 2) ^ q[22] ^ q[15]);
216 void bmw_large_nextBlock(bmw_large_ctx_t* ctx, const void* block){
218 dump_x(block, 16, 'M');
219 bmw_large_f0(q, ctx->h, block);
221 bmw_large_f1(q, block, ctx->h);
223 bmw_large_f2(ctx->h, q, block);
228 void bmw_large_lastBlock(bmw_large_ctx_t* ctx, const void* block, uint16_t length_b){
230 while(length_b >= BMW_LARGE_BLOCKSIZE){
231 bmw_large_nextBlock(ctx, block);
232 length_b -= BMW_LARGE_BLOCKSIZE;
233 block = (uint8_t*)block + BMW_LARGE_BLOCKSIZE_B;
235 memset(buffer, 0, 128);
236 memcpy(buffer, block, (length_b+7)/8);
237 buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
238 if(length_b+1>128*8-64){
239 bmw_large_nextBlock(ctx, buffer);
240 memset(buffer, 0, 128-8);
243 *((uint64_t*)&(buffer[128-8])) = (uint64_t)(ctx->counter*1024LL)+(uint64_t)length_b;
244 bmw_large_nextBlock(ctx, buffer);
247 memset(buffer, 0xaa, 128);
249 buffer[8*i] = i + 0xa0;
251 bmw_large_f0(q, (uint64_t*)buffer, ctx->h);
252 bmw_large_f1(q, ctx->h, (uint64_t*)buffer);
253 bmw_large_f2((uint64_t*)buffer, q, ctx->h);
254 memcpy(ctx->h, buffer, 128);
257 void bmw384_init(bmw384_ctx_t* ctx){
259 ctx->h[0] = 0x0001020304050607LL;
261 ctx->h[i] = ctx->h[i-1]+ 0x0808080808080808LL;
264 ctx->h[6] = 0x3031323324353637LL;
270 void bmw512_init(bmw512_ctx_t* ctx){
272 ctx->h[0] = 0x8081828384858687LL;
274 ctx->h[i] = ctx->h[i-1]+ 0x0808080808080808LL;
280 void bmw384_nextBlock(bmw384_ctx_t* ctx, const void* block){
281 bmw_large_nextBlock(ctx, block);
284 void bmw512_nextBlock(bmw512_ctx_t* ctx, const void* block){
285 bmw_large_nextBlock(ctx, block);
288 void bmw384_lastBlock(bmw384_ctx_t* ctx, const void* block, uint16_t length_b){
289 bmw_large_lastBlock(ctx, block, length_b);
292 void bmw512_lastBlock(bmw512_ctx_t* ctx, const void* block, uint16_t length_b){
293 bmw_large_lastBlock(ctx, block, length_b);
296 void bmw384_ctx2hash(void* dest, const bmw384_ctx_t* ctx){
297 memcpy(dest, &(ctx->h[10]), 384/8);
300 void bmw512_ctx2hash(void* dest, const bmw512_ctx_t* ctx){
301 memcpy(dest, &(ctx->h[8]), 512/8);
304 void bmw384(void* dest, const void* msg, uint32_t length_b){
307 while(length_b>=BMW_LARGE_BLOCKSIZE){
308 bmw_large_nextBlock(&ctx, msg);
309 length_b -= BMW_LARGE_BLOCKSIZE;
310 msg = (uint8_t*)msg + BMW_LARGE_BLOCKSIZE_B;
312 bmw_large_lastBlock(&ctx, msg, length_b);
313 bmw384_ctx2hash(dest, &ctx);
316 void bmw512(void* dest, const void* msg, uint32_t length_b){
319 while(length_b>=BMW_LARGE_BLOCKSIZE){
320 bmw_large_nextBlock(&ctx, msg);
321 length_b -= BMW_LARGE_BLOCKSIZE;
322 msg = (uint8_t*)msg + BMW_LARGE_BLOCKSIZE_B;
324 bmw_large_lastBlock(&ctx, msg, length_b);
325 bmw512_ctx2hash(dest, &ctx);