1 /* Modified for SILC. -Pekka */
3 /* This is an independent implementation of the encryption algorithm: */
5 /* RIJNDAEL by Joan Daemen and Vincent Rijmen */
7 /* which is a candidate algorithm in the Advanced Encryption Standard */
8 /* programme of the US National Institute of Standards and Technology. */
10 /* Copyright in this implementation is held by Dr B R Gladman but I */
11 /* hereby give permission for its free direct or derivative use subject */
12 /* to acknowledgment of its origin and compliance with any conditions */
13 /* that the originators of the algorithm place on its exploitation. */
15 /* Dr Brian Gladman (gladman@seven77.demon.co.uk) 14th January 1999 */
17 /* Timing data for Rijndael (rijndael.c)
19 Algorithm: rijndael (rijndael.c)
22 Key Setup: 305/1389 cycles (encrypt/decrypt)
23 Encrypt: 374 cycles = 68.4 mbits/sec
24 Decrypt: 352 cycles = 72.7 mbits/sec
25 Mean: 363 cycles = 70.5 mbits/sec
28 Key Setup: 277/1595 cycles (encrypt/decrypt)
29 Encrypt: 439 cycles = 58.3 mbits/sec
30 Decrypt: 425 cycles = 60.2 mbits/sec
31 Mean: 432 cycles = 59.3 mbits/sec
34 Key Setup: 374/1960 cycles (encrypt/decrypt)
35 Encrypt: 502 cycles = 51.0 mbits/sec
36 Decrypt: 498 cycles = 51.4 mbits/sec
37 Mean: 500 cycles = 51.2 mbits/sec
41 #include "silcincludes.h"
51 u4byte ft_tab[4][256];
52 u4byte it_tab[4][256];
54 u4byte fl_tab[4][256];
55 u4byte il_tab[4][256];
59 #define ff_mult(a,b) (a && b ? pow_tab[(log_tab[a] + log_tab[b]) % 255] : 0)
61 #define f_rn(bo, bi, n, k) \
62 bo[n] = ft_tab[0][byte(bi[n],0)] ^ \
63 ft_tab[1][byte(bi[(n + 1) & 3],1)] ^ \
64 ft_tab[2][byte(bi[(n + 2) & 3],2)] ^ \
65 ft_tab[3][byte(bi[(n + 3) & 3],3)] ^ *(k + n)
67 #define i_rn(bo, bi, n, k) \
68 bo[n] = it_tab[0][byte(bi[n],0)] ^ \
69 it_tab[1][byte(bi[(n + 3) & 3],1)] ^ \
70 it_tab[2][byte(bi[(n + 2) & 3],2)] ^ \
71 it_tab[3][byte(bi[(n + 1) & 3],3)] ^ *(k + n)
76 ( fl_tab[0][byte(x, 0)] ^ \
77 fl_tab[1][byte(x, 1)] ^ \
78 fl_tab[2][byte(x, 2)] ^ \
79 fl_tab[3][byte(x, 3)] )
81 #define f_rl(bo, bi, n, k) \
82 bo[n] = fl_tab[0][byte(bi[n],0)] ^ \
83 fl_tab[1][byte(bi[(n + 1) & 3],1)] ^ \
84 fl_tab[2][byte(bi[(n + 2) & 3],2)] ^ \
85 fl_tab[3][byte(bi[(n + 3) & 3],3)] ^ *(k + n)
87 #define i_rl(bo, bi, n, k) \
88 bo[n] = il_tab[0][byte(bi[n],0)] ^ \
89 il_tab[1][byte(bi[(n + 3) & 3],1)] ^ \
90 il_tab[2][byte(bi[(n + 2) & 3],2)] ^ \
91 il_tab[3][byte(bi[(n + 1) & 3],3)] ^ *(k + n)
96 ((u4byte)sbx_tab[byte(x, 0)] << 0) ^ \
97 ((u4byte)sbx_tab[byte(x, 1)] << 8) ^ \
98 ((u4byte)sbx_tab[byte(x, 2)] << 16) ^ \
99 ((u4byte)sbx_tab[byte(x, 3)] << 24)
101 #define f_rl(bo, bi, n, k) \
102 bo[n] = (u4byte)sbx_tab[byte(bi[n],0)] ^ \
103 rotl(((u4byte)sbx_tab[byte(bi[(n + 1) & 3],1)]), 8) ^ \
104 rotl(((u4byte)sbx_tab[byte(bi[(n + 2) & 3],2)]), 16) ^ \
105 rotl(((u4byte)sbx_tab[byte(bi[(n + 3) & 3],3)]), 24) ^ *(k + n)
107 #define i_rl(bo, bi, n, k) \
108 bo[n] = (u4byte)isb_tab[byte(bi[n],0)] ^ \
109 rotl(((u4byte)isb_tab[byte(bi[(n + 3) & 3],1)]), 8) ^ \
110 rotl(((u4byte)isb_tab[byte(bi[(n + 2) & 3],2)]), 16) ^ \
111 rotl(((u4byte)isb_tab[byte(bi[(n + 1) & 3],3)]), 24) ^ *(k + n)
119 /* log and power tables for GF(2**8) finite field with */
120 /* 0x11b as modular polynomial - the simplest prmitive */
121 /* root is 0x11, used here to generate the tables */
123 for(i = 0,p = 1; i < 256; ++i)
125 pow_tab[i] = (u1byte)p; log_tab[p] = (u1byte)i;
127 p = p ^ (p << 1) ^ (p & 0x80 ? 0x01b : 0);
130 log_tab[1] = 0; p = 1;
132 for(i = 0; i < 10; ++i)
136 p = (p << 1) ^ (p & 0x80 ? 0x1b : 0);
139 /* note that the affine byte transformation matrix in */
140 /* rijndael specification is in big endian format with */
141 /* bit 0 as the most significant bit. In the remainder */
142 /* of the specification the bits are numbered from the */
143 /* least significant end of a byte. */
145 for(i = 0; i < 256; ++i)
147 p = (i ? pow_tab[255 - log_tab[i]] : 0); q = p;
148 q = (q >> 7) | (q << 1); p ^= q;
149 q = (q >> 7) | (q << 1); p ^= q;
150 q = (q >> 7) | (q << 1); p ^= q;
151 q = (q >> 7) | (q << 1); p ^= q ^ 0x63;
152 sbx_tab[i] = (u1byte)p; isb_tab[p] = (u1byte)i;
155 for(i = 0; i < 256; ++i)
161 t = p; fl_tab[0][i] = t;
162 fl_tab[1][i] = rotl(t, 8);
163 fl_tab[2][i] = rotl(t, 16);
164 fl_tab[3][i] = rotl(t, 24);
166 t = ((u4byte)ff_mult(2, p)) |
169 ((u4byte)ff_mult(3, p) << 24);
172 ft_tab[1][i] = rotl(t, 8);
173 ft_tab[2][i] = rotl(t, 16);
174 ft_tab[3][i] = rotl(t, 24);
180 t = p; il_tab[0][i] = t;
181 il_tab[1][i] = rotl(t, 8);
182 il_tab[2][i] = rotl(t, 16);
183 il_tab[3][i] = rotl(t, 24);
185 t = ((u4byte)ff_mult(14, p)) |
186 ((u4byte)ff_mult( 9, p) << 8) |
187 ((u4byte)ff_mult(13, p) << 16) |
188 ((u4byte)ff_mult(11, p) << 24);
191 it_tab[1][i] = rotl(t, 8);
192 it_tab[2][i] = rotl(t, 16);
193 it_tab[3][i] = rotl(t, 24);
199 #define star_x(x) (((x) & 0x7f7f7f7f) << 1) ^ ((((x) & 0x80808080) >> 7) * 0x1b)
201 #define imix_col(y,x) \
207 (y) ^= rotr(u ^ t, 8) ^ \
211 /* initialise the key schedule from the user supplied key */
215 t = ls_box(rotr(t, 8)) ^ rco_tab[i]; \
216 t ^= e_key[4 * i]; e_key[4 * i + 4] = t; \
217 t ^= e_key[4 * i + 1]; e_key[4 * i + 5] = t; \
218 t ^= e_key[4 * i + 2]; e_key[4 * i + 6] = t; \
219 t ^= e_key[4 * i + 3]; e_key[4 * i + 7] = t; \
223 { t = ls_box(rotr(t, 8)) ^ rco_tab[i]; \
224 t ^= e_key[6 * i]; e_key[6 * i + 6] = t; \
225 t ^= e_key[6 * i + 1]; e_key[6 * i + 7] = t; \
226 t ^= e_key[6 * i + 2]; e_key[6 * i + 8] = t; \
227 t ^= e_key[6 * i + 3]; e_key[6 * i + 9] = t; \
228 t ^= e_key[6 * i + 4]; e_key[6 * i + 10] = t; \
229 t ^= e_key[6 * i + 5]; e_key[6 * i + 11] = t; \
233 { t = ls_box(rotr(t, 8)) ^ rco_tab[i]; \
234 t ^= e_key[8 * i]; e_key[8 * i + 8] = t; \
235 t ^= e_key[8 * i + 1]; e_key[8 * i + 9] = t; \
236 t ^= e_key[8 * i + 2]; e_key[8 * i + 10] = t; \
237 t ^= e_key[8 * i + 3]; e_key[8 * i + 11] = t; \
238 t = e_key[8 * i + 4] ^ ls_box(t); \
239 e_key[8 * i + 12] = t; \
240 t ^= e_key[8 * i + 5]; e_key[8 * i + 13] = t; \
241 t ^= e_key[8 * i + 6]; e_key[8 * i + 14] = t; \
242 t ^= e_key[8 * i + 7]; e_key[8 * i + 15] = t; \
245 u4byte *rijndael_set_key(RijndaelContext *ctx,
246 const u4byte in_key[], const u4byte key_len)
248 u4byte i, t, u, v, w;
249 u4byte *e_key = ctx->e_key;
250 u4byte *d_key = ctx->d_key;
256 k_len = ctx->k_len = (key_len + 31) / 32;
258 e_key[0] = in_key[0]; e_key[1] = in_key[1];
259 e_key[2] = in_key[2]; e_key[3] = in_key[3];
263 case 4: t = e_key[3];
264 for(i = 0; i < 10; ++i)
268 case 6: e_key[4] = in_key[4]; t = e_key[5] = in_key[5];
269 for(i = 0; i < 8; ++i)
273 case 8: e_key[4] = in_key[4]; e_key[5] = in_key[5];
274 e_key[6] = in_key[6]; t = e_key[7] = in_key[7];
275 for(i = 0; i < 7; ++i)
280 d_key[0] = e_key[0]; d_key[1] = e_key[1];
281 d_key[2] = e_key[2]; d_key[3] = e_key[3];
283 for(i = 4; i < 4 * k_len + 24; ++i)
285 imix_col(d_key[i], e_key[i]);
291 /* encrypt a block of text */
293 #define f_nround(bo, bi, k) \
294 f_rn(bo, bi, 0, k); \
295 f_rn(bo, bi, 1, k); \
296 f_rn(bo, bi, 2, k); \
297 f_rn(bo, bi, 3, k); \
300 #define f_lround(bo, bi, k) \
301 f_rl(bo, bi, 0, k); \
302 f_rl(bo, bi, 1, k); \
303 f_rl(bo, bi, 2, k); \
306 void rijndael_encrypt(RijndaelContext *ctx,
307 const u4byte in_blk[4], u4byte out_blk[4])
309 u4byte b0[4], b1[4], *kp;
310 u4byte *e_key = ctx->e_key;
311 u4byte k_len = ctx->k_len;
313 b0[0] = in_blk[0] ^ e_key[0]; b0[1] = in_blk[1] ^ e_key[1];
314 b0[2] = in_blk[2] ^ e_key[2]; b0[3] = in_blk[3] ^ e_key[3];
320 f_nround(b1, b0, kp); f_nround(b0, b1, kp);
325 f_nround(b1, b0, kp); f_nround(b0, b1, kp);
328 f_nround(b1, b0, kp); f_nround(b0, b1, kp);
329 f_nround(b1, b0, kp); f_nround(b0, b1, kp);
330 f_nround(b1, b0, kp); f_nround(b0, b1, kp);
331 f_nround(b1, b0, kp); f_nround(b0, b1, kp);
332 f_nround(b1, b0, kp); f_lround(b0, b1, kp);
334 out_blk[0] = b0[0]; out_blk[1] = b0[1];
335 out_blk[2] = b0[2]; out_blk[3] = b0[3];
338 /* decrypt a block of text */
340 #define i_nround(bo, bi, k) \
341 i_rn(bo, bi, 0, k); \
342 i_rn(bo, bi, 1, k); \
343 i_rn(bo, bi, 2, k); \
344 i_rn(bo, bi, 3, k); \
347 #define i_lround(bo, bi, k) \
348 i_rl(bo, bi, 0, k); \
349 i_rl(bo, bi, 1, k); \
350 i_rl(bo, bi, 2, k); \
353 void rijndael_decrypt(RijndaelContext *ctx,
354 const u4byte in_blk[4], u4byte out_blk[4])
356 u4byte b0[4], b1[4], *kp;
357 u4byte *e_key = ctx->e_key;
358 u4byte *d_key = ctx->d_key;
359 u4byte k_len = ctx->k_len;
361 b0[0] = in_blk[0] ^ e_key[4 * k_len + 24]; b0[1] = in_blk[1] ^ e_key[4 * k_len + 25];
362 b0[2] = in_blk[2] ^ e_key[4 * k_len + 26]; b0[3] = in_blk[3] ^ e_key[4 * k_len + 27];
364 kp = d_key + 4 * (k_len + 5);
368 i_nround(b1, b0, kp); i_nround(b0, b1, kp);
373 i_nround(b1, b0, kp); i_nround(b0, b1, kp);
376 i_nround(b1, b0, kp); i_nround(b0, b1, kp);
377 i_nround(b1, b0, kp); i_nround(b0, b1, kp);
378 i_nround(b1, b0, kp); i_nround(b0, b1, kp);
379 i_nround(b1, b0, kp); i_nround(b0, b1, kp);
380 i_nround(b1, b0, kp); i_lround(b0, b1, kp);
382 out_blk[0] = b0[0]; out_blk[1] = b0[1];
383 out_blk[2] = b0[2]; out_blk[3] = b0[3];