1 /*
2  * Cryptographic API.
3  *
4  * Serpent Cipher Algorithm.
5  *
6  * Copyright (C) 2002 Dag Arne Osvik <osvik@ii.uib.no>
7  *               2003 Herbert Valerio Riedel <hvr@gnu.org>
8  *
9  * Added tnepres support:
10  *		Ruben Jesus Garcia Hernandez <ruben@ugr.es>, 18.10.2004
11  *              Based on code by hvr
12  *
13  * This program is free software; you can redistribute it and/or modify
14  * it under the terms of the GNU General Public License as published by
15  * the Free Software Foundation; either version 2 of the License, or
16  * (at your option) any later version.
17  */
18 
19 #include <linux/init.h>
20 #include <linux/module.h>
21 #include <linux/errno.h>
22 #include <asm/byteorder.h>
23 #include <linux/crypto.h>
24 #include <linux/types.h>
25 #include <crypto/serpent.h>
26 
27 /* Key is padded to the maximum of 256 bits before round key generation.
28  * Any key length <= 256 bits (32 bytes) is allowed by the algorithm.
29  */
30 
31 #define PHI 0x9e3779b9UL
32 
33 #define keyiter(a, b, c, d, i, j) \
34 	({ b ^= d; b ^= c; b ^= a; b ^= PHI ^ i; b = rol32(b, 11); k[j] = b; })
35 
36 #define loadkeys(x0, x1, x2, x3, i) \
37 	({ x0 = k[i]; x1 = k[i+1]; x2 = k[i+2]; x3 = k[i+3]; })
38 
39 #define storekeys(x0, x1, x2, x3, i) \
40 	({ k[i] = x0; k[i+1] = x1; k[i+2] = x2; k[i+3] = x3; })
41 
42 #define store_and_load_keys(x0, x1, x2, x3, s, l) \
43 	({ storekeys(x0, x1, x2, x3, s); loadkeys(x0, x1, x2, x3, l); })
44 
45 #define K(x0, x1, x2, x3, i) ({				\
46 	x3 ^= k[4*(i)+3];        x2 ^= k[4*(i)+2];	\
47 	x1 ^= k[4*(i)+1];        x0 ^= k[4*(i)+0];	\
48 	})
49 
50 #define LK(x0, x1, x2, x3, x4, i) ({					   \
51 							x0 = rol32(x0, 13);\
52 	x2 = rol32(x2, 3);	x1 ^= x0;		x4  = x0 << 3;	   \
53 	x3 ^= x2;		x1 ^= x2;				   \
54 	x1 = rol32(x1, 1);	x3 ^= x4;				   \
55 	x3 = rol32(x3, 7);	x4  = x1;				   \
56 	x0 ^= x1;		x4 <<= 7;		x2 ^= x3;	   \
57 	x0 ^= x3;		x2 ^= x4;		x3 ^= k[4*i+3];	   \
58 	x1 ^= k[4*i+1];		x0 = rol32(x0, 5);	x2 = rol32(x2, 22);\
59 	x0 ^= k[4*i+0];		x2 ^= k[4*i+2];				   \
60 	})
61 
62 #define KL(x0, x1, x2, x3, x4, i) ({					   \
63 	x0 ^= k[4*i+0];		x1 ^= k[4*i+1];		x2 ^= k[4*i+2];	   \
64 	x3 ^= k[4*i+3];		x0 = ror32(x0, 5);	x2 = ror32(x2, 22);\
65 	x4 =  x1;		x2 ^= x3;		x0 ^= x3;	   \
66 	x4 <<= 7;		x0 ^= x1;		x1 = ror32(x1, 1); \
67 	x2 ^= x4;		x3 = ror32(x3, 7);	x4 = x0 << 3;	   \
68 	x1 ^= x0;		x3 ^= x4;		x0 = ror32(x0, 13);\
69 	x1 ^= x2;		x3 ^= x2;		x2 = ror32(x2, 3); \
70 	})
71 
72 #define S0(x0, x1, x2, x3, x4) ({			\
73 					x4  = x3;	\
74 	x3 |= x0;	x0 ^= x4;	x4 ^= x2;	\
75 	x4 = ~x4;	x3 ^= x1;	x1 &= x0;	\
76 	x1 ^= x4;	x2 ^= x0;	x0 ^= x3;	\
77 	x4 |= x0;	x0 ^= x2;	x2 &= x1;	\
78 	x3 ^= x2;	x1 = ~x1;	x2 ^= x4;	\
79 	x1 ^= x2;					\
80 	})
81 
82 #define S1(x0, x1, x2, x3, x4) ({			\
83 					x4  = x1;	\
84 	x1 ^= x0;	x0 ^= x3;	x3 = ~x3;	\
85 	x4 &= x1;	x0 |= x1;	x3 ^= x2;	\
86 	x0 ^= x3;	x1 ^= x3;	x3 ^= x4;	\
87 	x1 |= x4;	x4 ^= x2;	x2 &= x0;	\
88 	x2 ^= x1;	x1 |= x0;	x0 = ~x0;	\
89 	x0 ^= x2;	x4 ^= x1;			\
90 	})
91 
92 #define S2(x0, x1, x2, x3, x4) ({			\
93 					x3 = ~x3;	\
94 	x1 ^= x0;	x4  = x0;	x0 &= x2;	\
95 	x0 ^= x3;	x3 |= x4;	x2 ^= x1;	\
96 	x3 ^= x1;	x1 &= x0;	x0 ^= x2;	\
97 	x2 &= x3;	x3 |= x1;	x0 = ~x0;	\
98 	x3 ^= x0;	x4 ^= x0;	x0 ^= x2;	\
99 	x1 |= x2;					\
100 	})
101 
102 #define S3(x0, x1, x2, x3, x4) ({			\
103 					x4  = x1;	\
104 	x1 ^= x3;	x3 |= x0;	x4 &= x0;	\
105 	x0 ^= x2;	x2 ^= x1;	x1 &= x3;	\
106 	x2 ^= x3;	x0 |= x4;	x4 ^= x3;	\
107 	x1 ^= x0;	x0 &= x3;	x3 &= x4;	\
108 	x3 ^= x2;	x4 |= x1;	x2 &= x1;	\
109 	x4 ^= x3;	x0 ^= x3;	x3 ^= x2;	\
110 	})
111 
112 #define S4(x0, x1, x2, x3, x4) ({			\
113 					x4  = x3;	\
114 	x3 &= x0;	x0 ^= x4;			\
115 	x3 ^= x2;	x2 |= x4;	x0 ^= x1;	\
116 	x4 ^= x3;	x2 |= x0;			\
117 	x2 ^= x1;	x1 &= x0;			\
118 	x1 ^= x4;	x4 &= x2;	x2 ^= x3;	\
119 	x4 ^= x0;	x3 |= x1;	x1 = ~x1;	\
120 	x3 ^= x0;					\
121 	})
122 
123 #define S5(x0, x1, x2, x3, x4) ({			\
124 	x4  = x1;	x1 |= x0;			\
125 	x2 ^= x1;	x3 = ~x3;	x4 ^= x0;	\
126 	x0 ^= x2;	x1 &= x4;	x4 |= x3;	\
127 	x4 ^= x0;	x0 &= x3;	x1 ^= x3;	\
128 	x3 ^= x2;	x0 ^= x1;	x2 &= x4;	\
129 	x1 ^= x2;	x2 &= x0;			\
130 	x3 ^= x2;					\
131 	})
132 
133 #define S6(x0, x1, x2, x3, x4) ({			\
134 					x4  = x1;	\
135 	x3 ^= x0;	x1 ^= x2;	x2 ^= x0;	\
136 	x0 &= x3;	x1 |= x3;	x4 = ~x4;	\
137 	x0 ^= x1;	x1 ^= x2;			\
138 	x3 ^= x4;	x4 ^= x0;	x2 &= x0;	\
139 	x4 ^= x1;	x2 ^= x3;	x3 &= x1;	\
140 	x3 ^= x0;	x1 ^= x2;			\
141 	})
142 
143 #define S7(x0, x1, x2, x3, x4) ({			\
144 					x1 = ~x1;	\
145 	x4  = x1;	x0 = ~x0;	x1 &= x2;	\
146 	x1 ^= x3;	x3 |= x4;	x4 ^= x2;	\
147 	x2 ^= x3;	x3 ^= x0;	x0 |= x1;	\
148 	x2 &= x0;	x0 ^= x4;	x4 ^= x3;	\
149 	x3 &= x0;	x4 ^= x1;			\
150 	x2 ^= x4;	x3 ^= x1;	x4 |= x0;	\
151 	x4 ^= x1;					\
152 	})
153 
154 #define SI0(x0, x1, x2, x3, x4) ({			\
155 			x4  = x3;	x1 ^= x0;	\
156 	x3 |= x1;	x4 ^= x1;	x0 = ~x0;	\
157 	x2 ^= x3;	x3 ^= x0;	x0 &= x1;	\
158 	x0 ^= x2;	x2 &= x3;	x3 ^= x4;	\
159 	x2 ^= x3;	x1 ^= x3;	x3 &= x0;	\
160 	x1 ^= x0;	x0 ^= x2;	x4 ^= x3;	\
161 	})
162 
163 #define SI1(x0, x1, x2, x3, x4) ({			\
164 	x1 ^= x3;	x4  = x0;			\
165 	x0 ^= x2;	x2 = ~x2;	x4 |= x1;	\
166 	x4 ^= x3;	x3 &= x1;	x1 ^= x2;	\
167 	x2 &= x4;	x4 ^= x1;	x1 |= x3;	\
168 	x3 ^= x0;	x2 ^= x0;	x0 |= x4;	\
169 	x2 ^= x4;	x1 ^= x0;			\
170 	x4 ^= x1;					\
171 	})
172 
173 #define SI2(x0, x1, x2, x3, x4) ({			\
174 	x2 ^= x1;	x4  = x3;	x3 = ~x3;	\
175 	x3 |= x2;	x2 ^= x4;	x4 ^= x0;	\
176 	x3 ^= x1;	x1 |= x2;	x2 ^= x0;	\
177 	x1 ^= x4;	x4 |= x3;	x2 ^= x3;	\
178 	x4 ^= x2;	x2 &= x1;			\
179 	x2 ^= x3;	x3 ^= x4;	x4 ^= x0;	\
180 	})
181 
182 #define SI3(x0, x1, x2, x3, x4) ({			\
183 					x2 ^= x1;	\
184 	x4  = x1;	x1 &= x2;			\
185 	x1 ^= x0;	x0 |= x4;	x4 ^= x3;	\
186 	x0 ^= x3;	x3 |= x1;	x1 ^= x2;	\
187 	x1 ^= x3;	x0 ^= x2;	x2 ^= x3;	\
188 	x3 &= x1;	x1 ^= x0;	x0 &= x2;	\
189 	x4 ^= x3;	x3 ^= x0;	x0 ^= x1;	\
190 	})
191 
192 #define SI4(x0, x1, x2, x3, x4) ({			\
193 	x2 ^= x3;	x4  = x0;	x0 &= x1;	\
194 	x0 ^= x2;	x2 |= x3;	x4 = ~x4;	\
195 	x1 ^= x0;	x0 ^= x2;	x2 &= x4;	\
196 	x2 ^= x0;	x0 |= x4;			\
197 	x0 ^= x3;	x3 &= x2;			\
198 	x4 ^= x3;	x3 ^= x1;	x1 &= x0;	\
199 	x4 ^= x1;	x0 ^= x3;			\
200 	})
201 
202 #define SI5(x0, x1, x2, x3, x4) ({			\
203 			x4  = x1;	x1 |= x2;	\
204 	x2 ^= x4;	x1 ^= x3;	x3 &= x4;	\
205 	x2 ^= x3;	x3 |= x0;	x0 = ~x0;	\
206 	x3 ^= x2;	x2 |= x0;	x4 ^= x1;	\
207 	x2 ^= x4;	x4 &= x0;	x0 ^= x1;	\
208 	x1 ^= x3;	x0 &= x2;	x2 ^= x3;	\
209 	x0 ^= x2;	x2 ^= x4;	x4 ^= x3;	\
210 	})
211 
212 #define SI6(x0, x1, x2, x3, x4) ({			\
213 			x0 ^= x2;			\
214 	x4  = x0;	x0 &= x3;	x2 ^= x3;	\
215 	x0 ^= x2;	x3 ^= x1;	x2 |= x4;	\
216 	x2 ^= x3;	x3 &= x0;	x0 = ~x0;	\
217 	x3 ^= x1;	x1 &= x2;	x4 ^= x0;	\
218 	x3 ^= x4;	x4 ^= x2;	x0 ^= x1;	\
219 	x2 ^= x0;					\
220 	})
221 
222 #define SI7(x0, x1, x2, x3, x4) ({			\
223 	x4  = x3;	x3 &= x0;	x0 ^= x2;	\
224 	x2 |= x4;	x4 ^= x1;	x0 = ~x0;	\
225 	x1 |= x3;	x4 ^= x0;	x0 &= x2;	\
226 	x0 ^= x1;	x1 &= x2;	x3 ^= x2;	\
227 	x4 ^= x3;	x2 &= x3;	x3 |= x0;	\
228 	x1 ^= x4;	x3 ^= x4;	x4 &= x0;	\
229 	x4 ^= x2;					\
230 	})
231 
232 /*
233  * both gcc and clang have misoptimized this function in the past,
234  * producing horrible object code from spilling temporary variables
235  * on the stack. Forcing this part out of line avoids that.
236  */
__serpent_setkey_sbox(u32 r0,u32 r1,u32 r2,u32 r3,u32 r4,u32 * k)237 static noinline void __serpent_setkey_sbox(u32 r0, u32 r1, u32 r2,
238 					   u32 r3, u32 r4, u32 *k)
239 {
240 	k += 100;
241 	S3(r3, r4, r0, r1, r2); store_and_load_keys(r1, r2, r4, r3, 28, 24);
242 	S4(r1, r2, r4, r3, r0); store_and_load_keys(r2, r4, r3, r0, 24, 20);
243 	S5(r2, r4, r3, r0, r1); store_and_load_keys(r1, r2, r4, r0, 20, 16);
244 	S6(r1, r2, r4, r0, r3); store_and_load_keys(r4, r3, r2, r0, 16, 12);
245 	S7(r4, r3, r2, r0, r1); store_and_load_keys(r1, r2, r0, r4, 12, 8);
246 	S0(r1, r2, r0, r4, r3); store_and_load_keys(r0, r2, r4, r1, 8, 4);
247 	S1(r0, r2, r4, r1, r3); store_and_load_keys(r3, r4, r1, r0, 4, 0);
248 	S2(r3, r4, r1, r0, r2); store_and_load_keys(r2, r4, r3, r0, 0, -4);
249 	S3(r2, r4, r3, r0, r1); store_and_load_keys(r0, r1, r4, r2, -4, -8);
250 	S4(r0, r1, r4, r2, r3); store_and_load_keys(r1, r4, r2, r3, -8, -12);
251 	S5(r1, r4, r2, r3, r0); store_and_load_keys(r0, r1, r4, r3, -12, -16);
252 	S6(r0, r1, r4, r3, r2); store_and_load_keys(r4, r2, r1, r3, -16, -20);
253 	S7(r4, r2, r1, r3, r0); store_and_load_keys(r0, r1, r3, r4, -20, -24);
254 	S0(r0, r1, r3, r4, r2); store_and_load_keys(r3, r1, r4, r0, -24, -28);
255 	k -= 50;
256 	S1(r3, r1, r4, r0, r2); store_and_load_keys(r2, r4, r0, r3, 22, 18);
257 	S2(r2, r4, r0, r3, r1); store_and_load_keys(r1, r4, r2, r3, 18, 14);
258 	S3(r1, r4, r2, r3, r0); store_and_load_keys(r3, r0, r4, r1, 14, 10);
259 	S4(r3, r0, r4, r1, r2); store_and_load_keys(r0, r4, r1, r2, 10, 6);
260 	S5(r0, r4, r1, r2, r3); store_and_load_keys(r3, r0, r4, r2, 6, 2);
261 	S6(r3, r0, r4, r2, r1); store_and_load_keys(r4, r1, r0, r2, 2, -2);
262 	S7(r4, r1, r0, r2, r3); store_and_load_keys(r3, r0, r2, r4, -2, -6);
263 	S0(r3, r0, r2, r4, r1); store_and_load_keys(r2, r0, r4, r3, -6, -10);
264 	S1(r2, r0, r4, r3, r1); store_and_load_keys(r1, r4, r3, r2, -10, -14);
265 	S2(r1, r4, r3, r2, r0); store_and_load_keys(r0, r4, r1, r2, -14, -18);
266 	S3(r0, r4, r1, r2, r3); store_and_load_keys(r2, r3, r4, r0, -18, -22);
267 	k -= 50;
268 	S4(r2, r3, r4, r0, r1); store_and_load_keys(r3, r4, r0, r1, 28, 24);
269 	S5(r3, r4, r0, r1, r2); store_and_load_keys(r2, r3, r4, r1, 24, 20);
270 	S6(r2, r3, r4, r1, r0); store_and_load_keys(r4, r0, r3, r1, 20, 16);
271 	S7(r4, r0, r3, r1, r2); store_and_load_keys(r2, r3, r1, r4, 16, 12);
272 	S0(r2, r3, r1, r4, r0); store_and_load_keys(r1, r3, r4, r2, 12, 8);
273 	S1(r1, r3, r4, r2, r0); store_and_load_keys(r0, r4, r2, r1, 8, 4);
274 	S2(r0, r4, r2, r1, r3); store_and_load_keys(r3, r4, r0, r1, 4, 0);
275 	S3(r3, r4, r0, r1, r2); storekeys(r1, r2, r4, r3, 0);
276 }
277 
__serpent_setkey(struct serpent_ctx * ctx,const u8 * key,unsigned int keylen)278 int __serpent_setkey(struct serpent_ctx *ctx, const u8 *key,
279 		     unsigned int keylen)
280 {
281 	u32 *k = ctx->expkey;
282 	u8  *k8 = (u8 *)k;
283 	u32 r0, r1, r2, r3, r4;
284 	int i;
285 
286 	/* Copy key, add padding */
287 
288 	for (i = 0; i < keylen; ++i)
289 		k8[i] = key[i];
290 	if (i < SERPENT_MAX_KEY_SIZE)
291 		k8[i++] = 1;
292 	while (i < SERPENT_MAX_KEY_SIZE)
293 		k8[i++] = 0;
294 
295 	/* Expand key using polynomial */
296 
297 	r0 = le32_to_cpu(k[3]);
298 	r1 = le32_to_cpu(k[4]);
299 	r2 = le32_to_cpu(k[5]);
300 	r3 = le32_to_cpu(k[6]);
301 	r4 = le32_to_cpu(k[7]);
302 
303 	keyiter(le32_to_cpu(k[0]), r0, r4, r2, 0, 0);
304 	keyiter(le32_to_cpu(k[1]), r1, r0, r3, 1, 1);
305 	keyiter(le32_to_cpu(k[2]), r2, r1, r4, 2, 2);
306 	keyiter(le32_to_cpu(k[3]), r3, r2, r0, 3, 3);
307 	keyiter(le32_to_cpu(k[4]), r4, r3, r1, 4, 4);
308 	keyiter(le32_to_cpu(k[5]), r0, r4, r2, 5, 5);
309 	keyiter(le32_to_cpu(k[6]), r1, r0, r3, 6, 6);
310 	keyiter(le32_to_cpu(k[7]), r2, r1, r4, 7, 7);
311 
312 	keyiter(k[0], r3, r2, r0, 8, 8);
313 	keyiter(k[1], r4, r3, r1, 9, 9);
314 	keyiter(k[2], r0, r4, r2, 10, 10);
315 	keyiter(k[3], r1, r0, r3, 11, 11);
316 	keyiter(k[4], r2, r1, r4, 12, 12);
317 	keyiter(k[5], r3, r2, r0, 13, 13);
318 	keyiter(k[6], r4, r3, r1, 14, 14);
319 	keyiter(k[7], r0, r4, r2, 15, 15);
320 	keyiter(k[8], r1, r0, r3, 16, 16);
321 	keyiter(k[9], r2, r1, r4, 17, 17);
322 	keyiter(k[10], r3, r2, r0, 18, 18);
323 	keyiter(k[11], r4, r3, r1, 19, 19);
324 	keyiter(k[12], r0, r4, r2, 20, 20);
325 	keyiter(k[13], r1, r0, r3, 21, 21);
326 	keyiter(k[14], r2, r1, r4, 22, 22);
327 	keyiter(k[15], r3, r2, r0, 23, 23);
328 	keyiter(k[16], r4, r3, r1, 24, 24);
329 	keyiter(k[17], r0, r4, r2, 25, 25);
330 	keyiter(k[18], r1, r0, r3, 26, 26);
331 	keyiter(k[19], r2, r1, r4, 27, 27);
332 	keyiter(k[20], r3, r2, r0, 28, 28);
333 	keyiter(k[21], r4, r3, r1, 29, 29);
334 	keyiter(k[22], r0, r4, r2, 30, 30);
335 	keyiter(k[23], r1, r0, r3, 31, 31);
336 
337 	k += 50;
338 
339 	keyiter(k[-26], r2, r1, r4, 32, -18);
340 	keyiter(k[-25], r3, r2, r0, 33, -17);
341 	keyiter(k[-24], r4, r3, r1, 34, -16);
342 	keyiter(k[-23], r0, r4, r2, 35, -15);
343 	keyiter(k[-22], r1, r0, r3, 36, -14);
344 	keyiter(k[-21], r2, r1, r4, 37, -13);
345 	keyiter(k[-20], r3, r2, r0, 38, -12);
346 	keyiter(k[-19], r4, r3, r1, 39, -11);
347 	keyiter(k[-18], r0, r4, r2, 40, -10);
348 	keyiter(k[-17], r1, r0, r3, 41, -9);
349 	keyiter(k[-16], r2, r1, r4, 42, -8);
350 	keyiter(k[-15], r3, r2, r0, 43, -7);
351 	keyiter(k[-14], r4, r3, r1, 44, -6);
352 	keyiter(k[-13], r0, r4, r2, 45, -5);
353 	keyiter(k[-12], r1, r0, r3, 46, -4);
354 	keyiter(k[-11], r2, r1, r4, 47, -3);
355 	keyiter(k[-10], r3, r2, r0, 48, -2);
356 	keyiter(k[-9], r4, r3, r1, 49, -1);
357 	keyiter(k[-8], r0, r4, r2, 50, 0);
358 	keyiter(k[-7], r1, r0, r3, 51, 1);
359 	keyiter(k[-6], r2, r1, r4, 52, 2);
360 	keyiter(k[-5], r3, r2, r0, 53, 3);
361 	keyiter(k[-4], r4, r3, r1, 54, 4);
362 	keyiter(k[-3], r0, r4, r2, 55, 5);
363 	keyiter(k[-2], r1, r0, r3, 56, 6);
364 	keyiter(k[-1], r2, r1, r4, 57, 7);
365 	keyiter(k[0], r3, r2, r0, 58, 8);
366 	keyiter(k[1], r4, r3, r1, 59, 9);
367 	keyiter(k[2], r0, r4, r2, 60, 10);
368 	keyiter(k[3], r1, r0, r3, 61, 11);
369 	keyiter(k[4], r2, r1, r4, 62, 12);
370 	keyiter(k[5], r3, r2, r0, 63, 13);
371 	keyiter(k[6], r4, r3, r1, 64, 14);
372 	keyiter(k[7], r0, r4, r2, 65, 15);
373 	keyiter(k[8], r1, r0, r3, 66, 16);
374 	keyiter(k[9], r2, r1, r4, 67, 17);
375 	keyiter(k[10], r3, r2, r0, 68, 18);
376 	keyiter(k[11], r4, r3, r1, 69, 19);
377 	keyiter(k[12], r0, r4, r2, 70, 20);
378 	keyiter(k[13], r1, r0, r3, 71, 21);
379 	keyiter(k[14], r2, r1, r4, 72, 22);
380 	keyiter(k[15], r3, r2, r0, 73, 23);
381 	keyiter(k[16], r4, r3, r1, 74, 24);
382 	keyiter(k[17], r0, r4, r2, 75, 25);
383 	keyiter(k[18], r1, r0, r3, 76, 26);
384 	keyiter(k[19], r2, r1, r4, 77, 27);
385 	keyiter(k[20], r3, r2, r0, 78, 28);
386 	keyiter(k[21], r4, r3, r1, 79, 29);
387 	keyiter(k[22], r0, r4, r2, 80, 30);
388 	keyiter(k[23], r1, r0, r3, 81, 31);
389 
390 	k += 50;
391 
392 	keyiter(k[-26], r2, r1, r4, 82, -18);
393 	keyiter(k[-25], r3, r2, r0, 83, -17);
394 	keyiter(k[-24], r4, r3, r1, 84, -16);
395 	keyiter(k[-23], r0, r4, r2, 85, -15);
396 	keyiter(k[-22], r1, r0, r3, 86, -14);
397 	keyiter(k[-21], r2, r1, r4, 87, -13);
398 	keyiter(k[-20], r3, r2, r0, 88, -12);
399 	keyiter(k[-19], r4, r3, r1, 89, -11);
400 	keyiter(k[-18], r0, r4, r2, 90, -10);
401 	keyiter(k[-17], r1, r0, r3, 91, -9);
402 	keyiter(k[-16], r2, r1, r4, 92, -8);
403 	keyiter(k[-15], r3, r2, r0, 93, -7);
404 	keyiter(k[-14], r4, r3, r1, 94, -6);
405 	keyiter(k[-13], r0, r4, r2, 95, -5);
406 	keyiter(k[-12], r1, r0, r3, 96, -4);
407 	keyiter(k[-11], r2, r1, r4, 97, -3);
408 	keyiter(k[-10], r3, r2, r0, 98, -2);
409 	keyiter(k[-9], r4, r3, r1, 99, -1);
410 	keyiter(k[-8], r0, r4, r2, 100, 0);
411 	keyiter(k[-7], r1, r0, r3, 101, 1);
412 	keyiter(k[-6], r2, r1, r4, 102, 2);
413 	keyiter(k[-5], r3, r2, r0, 103, 3);
414 	keyiter(k[-4], r4, r3, r1, 104, 4);
415 	keyiter(k[-3], r0, r4, r2, 105, 5);
416 	keyiter(k[-2], r1, r0, r3, 106, 6);
417 	keyiter(k[-1], r2, r1, r4, 107, 7);
418 	keyiter(k[0], r3, r2, r0, 108, 8);
419 	keyiter(k[1], r4, r3, r1, 109, 9);
420 	keyiter(k[2], r0, r4, r2, 110, 10);
421 	keyiter(k[3], r1, r0, r3, 111, 11);
422 	keyiter(k[4], r2, r1, r4, 112, 12);
423 	keyiter(k[5], r3, r2, r0, 113, 13);
424 	keyiter(k[6], r4, r3, r1, 114, 14);
425 	keyiter(k[7], r0, r4, r2, 115, 15);
426 	keyiter(k[8], r1, r0, r3, 116, 16);
427 	keyiter(k[9], r2, r1, r4, 117, 17);
428 	keyiter(k[10], r3, r2, r0, 118, 18);
429 	keyiter(k[11], r4, r3, r1, 119, 19);
430 	keyiter(k[12], r0, r4, r2, 120, 20);
431 	keyiter(k[13], r1, r0, r3, 121, 21);
432 	keyiter(k[14], r2, r1, r4, 122, 22);
433 	keyiter(k[15], r3, r2, r0, 123, 23);
434 	keyiter(k[16], r4, r3, r1, 124, 24);
435 	keyiter(k[17], r0, r4, r2, 125, 25);
436 	keyiter(k[18], r1, r0, r3, 126, 26);
437 	keyiter(k[19], r2, r1, r4, 127, 27);
438 	keyiter(k[20], r3, r2, r0, 128, 28);
439 	keyiter(k[21], r4, r3, r1, 129, 29);
440 	keyiter(k[22], r0, r4, r2, 130, 30);
441 	keyiter(k[23], r1, r0, r3, 131, 31);
442 
443 	/* Apply S-boxes */
444 	__serpent_setkey_sbox(r0, r1, r2, r3, r4, ctx->expkey);
445 
446 	return 0;
447 }
448 EXPORT_SYMBOL_GPL(__serpent_setkey);
449 
serpent_setkey(struct crypto_tfm * tfm,const u8 * key,unsigned int keylen)450 int serpent_setkey(struct crypto_tfm *tfm, const u8 *key, unsigned int keylen)
451 {
452 	return __serpent_setkey(crypto_tfm_ctx(tfm), key, keylen);
453 }
454 EXPORT_SYMBOL_GPL(serpent_setkey);
455 
__serpent_encrypt(struct serpent_ctx * ctx,u8 * dst,const u8 * src)456 void __serpent_encrypt(struct serpent_ctx *ctx, u8 *dst, const u8 *src)
457 {
458 	const u32 *k = ctx->expkey;
459 	const __le32 *s = (const __le32 *)src;
460 	__le32	*d = (__le32 *)dst;
461 	u32	r0, r1, r2, r3, r4;
462 
463 /*
464  * Note: The conversions between u8* and u32* might cause trouble
465  * on architectures with stricter alignment rules than x86
466  */
467 
468 	r0 = le32_to_cpu(s[0]);
469 	r1 = le32_to_cpu(s[1]);
470 	r2 = le32_to_cpu(s[2]);
471 	r3 = le32_to_cpu(s[3]);
472 
473 					K(r0, r1, r2, r3, 0);
474 	S0(r0, r1, r2, r3, r4);		LK(r2, r1, r3, r0, r4, 1);
475 	S1(r2, r1, r3, r0, r4);		LK(r4, r3, r0, r2, r1, 2);
476 	S2(r4, r3, r0, r2, r1);		LK(r1, r3, r4, r2, r0, 3);
477 	S3(r1, r3, r4, r2, r0);		LK(r2, r0, r3, r1, r4, 4);
478 	S4(r2, r0, r3, r1, r4);		LK(r0, r3, r1, r4, r2, 5);
479 	S5(r0, r3, r1, r4, r2);		LK(r2, r0, r3, r4, r1, 6);
480 	S6(r2, r0, r3, r4, r1);		LK(r3, r1, r0, r4, r2, 7);
481 	S7(r3, r1, r0, r4, r2);		LK(r2, r0, r4, r3, r1, 8);
482 	S0(r2, r0, r4, r3, r1);		LK(r4, r0, r3, r2, r1, 9);
483 	S1(r4, r0, r3, r2, r1);		LK(r1, r3, r2, r4, r0, 10);
484 	S2(r1, r3, r2, r4, r0);		LK(r0, r3, r1, r4, r2, 11);
485 	S3(r0, r3, r1, r4, r2);		LK(r4, r2, r3, r0, r1, 12);
486 	S4(r4, r2, r3, r0, r1);		LK(r2, r3, r0, r1, r4, 13);
487 	S5(r2, r3, r0, r1, r4);		LK(r4, r2, r3, r1, r0, 14);
488 	S6(r4, r2, r3, r1, r0);		LK(r3, r0, r2, r1, r4, 15);
489 	S7(r3, r0, r2, r1, r4);		LK(r4, r2, r1, r3, r0, 16);
490 	S0(r4, r2, r1, r3, r0);		LK(r1, r2, r3, r4, r0, 17);
491 	S1(r1, r2, r3, r4, r0);		LK(r0, r3, r4, r1, r2, 18);
492 	S2(r0, r3, r4, r1, r2);		LK(r2, r3, r0, r1, r4, 19);
493 	S3(r2, r3, r0, r1, r4);		LK(r1, r4, r3, r2, r0, 20);
494 	S4(r1, r4, r3, r2, r0);		LK(r4, r3, r2, r0, r1, 21);
495 	S5(r4, r3, r2, r0, r1);		LK(r1, r4, r3, r0, r2, 22);
496 	S6(r1, r4, r3, r0, r2);		LK(r3, r2, r4, r0, r1, 23);
497 	S7(r3, r2, r4, r0, r1);		LK(r1, r4, r0, r3, r2, 24);
498 	S0(r1, r4, r0, r3, r2);		LK(r0, r4, r3, r1, r2, 25);
499 	S1(r0, r4, r3, r1, r2);		LK(r2, r3, r1, r0, r4, 26);
500 	S2(r2, r3, r1, r0, r4);		LK(r4, r3, r2, r0, r1, 27);
501 	S3(r4, r3, r2, r0, r1);		LK(r0, r1, r3, r4, r2, 28);
502 	S4(r0, r1, r3, r4, r2);		LK(r1, r3, r4, r2, r0, 29);
503 	S5(r1, r3, r4, r2, r0);		LK(r0, r1, r3, r2, r4, 30);
504 	S6(r0, r1, r3, r2, r4);		LK(r3, r4, r1, r2, r0, 31);
505 	S7(r3, r4, r1, r2, r0);		K(r0, r1, r2, r3, 32);
506 
507 	d[0] = cpu_to_le32(r0);
508 	d[1] = cpu_to_le32(r1);
509 	d[2] = cpu_to_le32(r2);
510 	d[3] = cpu_to_le32(r3);
511 }
512 EXPORT_SYMBOL_GPL(__serpent_encrypt);
513 
serpent_encrypt(struct crypto_tfm * tfm,u8 * dst,const u8 * src)514 static void serpent_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
515 {
516 	struct serpent_ctx *ctx = crypto_tfm_ctx(tfm);
517 
518 	__serpent_encrypt(ctx, dst, src);
519 }
520 
__serpent_decrypt(struct serpent_ctx * ctx,u8 * dst,const u8 * src)521 void __serpent_decrypt(struct serpent_ctx *ctx, u8 *dst, const u8 *src)
522 {
523 	const u32 *k = ctx->expkey;
524 	const __le32 *s = (const __le32 *)src;
525 	__le32	*d = (__le32 *)dst;
526 	u32	r0, r1, r2, r3, r4;
527 
528 	r0 = le32_to_cpu(s[0]);
529 	r1 = le32_to_cpu(s[1]);
530 	r2 = le32_to_cpu(s[2]);
531 	r3 = le32_to_cpu(s[3]);
532 
533 					K(r0, r1, r2, r3, 32);
534 	SI7(r0, r1, r2, r3, r4);	KL(r1, r3, r0, r4, r2, 31);
535 	SI6(r1, r3, r0, r4, r2);	KL(r0, r2, r4, r1, r3, 30);
536 	SI5(r0, r2, r4, r1, r3);	KL(r2, r3, r0, r4, r1, 29);
537 	SI4(r2, r3, r0, r4, r1);	KL(r2, r0, r1, r4, r3, 28);
538 	SI3(r2, r0, r1, r4, r3);	KL(r1, r2, r3, r4, r0, 27);
539 	SI2(r1, r2, r3, r4, r0);	KL(r2, r0, r4, r3, r1, 26);
540 	SI1(r2, r0, r4, r3, r1);	KL(r1, r0, r4, r3, r2, 25);
541 	SI0(r1, r0, r4, r3, r2);	KL(r4, r2, r0, r1, r3, 24);
542 	SI7(r4, r2, r0, r1, r3);	KL(r2, r1, r4, r3, r0, 23);
543 	SI6(r2, r1, r4, r3, r0);	KL(r4, r0, r3, r2, r1, 22);
544 	SI5(r4, r0, r3, r2, r1);	KL(r0, r1, r4, r3, r2, 21);
545 	SI4(r0, r1, r4, r3, r2);	KL(r0, r4, r2, r3, r1, 20);
546 	SI3(r0, r4, r2, r3, r1);	KL(r2, r0, r1, r3, r4, 19);
547 	SI2(r2, r0, r1, r3, r4);	KL(r0, r4, r3, r1, r2, 18);
548 	SI1(r0, r4, r3, r1, r2);	KL(r2, r4, r3, r1, r0, 17);
549 	SI0(r2, r4, r3, r1, r0);	KL(r3, r0, r4, r2, r1, 16);
550 	SI7(r3, r0, r4, r2, r1);	KL(r0, r2, r3, r1, r4, 15);
551 	SI6(r0, r2, r3, r1, r4);	KL(r3, r4, r1, r0, r2, 14);
552 	SI5(r3, r4, r1, r0, r2);	KL(r4, r2, r3, r1, r0, 13);
553 	SI4(r4, r2, r3, r1, r0);	KL(r4, r3, r0, r1, r2, 12);
554 	SI3(r4, r3, r0, r1, r2);	KL(r0, r4, r2, r1, r3, 11);
555 	SI2(r0, r4, r2, r1, r3);	KL(r4, r3, r1, r2, r0, 10);
556 	SI1(r4, r3, r1, r2, r0);	KL(r0, r3, r1, r2, r4, 9);
557 	SI0(r0, r3, r1, r2, r4);	KL(r1, r4, r3, r0, r2, 8);
558 	SI7(r1, r4, r3, r0, r2);	KL(r4, r0, r1, r2, r3, 7);
559 	SI6(r4, r0, r1, r2, r3);	KL(r1, r3, r2, r4, r0, 6);
560 	SI5(r1, r3, r2, r4, r0);	KL(r3, r0, r1, r2, r4, 5);
561 	SI4(r3, r0, r1, r2, r4);	KL(r3, r1, r4, r2, r0, 4);
562 	SI3(r3, r1, r4, r2, r0);	KL(r4, r3, r0, r2, r1, 3);
563 	SI2(r4, r3, r0, r2, r1);	KL(r3, r1, r2, r0, r4, 2);
564 	SI1(r3, r1, r2, r0, r4);	KL(r4, r1, r2, r0, r3, 1);
565 	SI0(r4, r1, r2, r0, r3);	K(r2, r3, r1, r4, 0);
566 
567 	d[0] = cpu_to_le32(r2);
568 	d[1] = cpu_to_le32(r3);
569 	d[2] = cpu_to_le32(r1);
570 	d[3] = cpu_to_le32(r4);
571 }
572 EXPORT_SYMBOL_GPL(__serpent_decrypt);
573 
serpent_decrypt(struct crypto_tfm * tfm,u8 * dst,const u8 * src)574 static void serpent_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
575 {
576 	struct serpent_ctx *ctx = crypto_tfm_ctx(tfm);
577 
578 	__serpent_decrypt(ctx, dst, src);
579 }
580 
tnepres_setkey(struct crypto_tfm * tfm,const u8 * key,unsigned int keylen)581 static int tnepres_setkey(struct crypto_tfm *tfm, const u8 *key,
582 			  unsigned int keylen)
583 {
584 	u8 rev_key[SERPENT_MAX_KEY_SIZE];
585 	int i;
586 
587 	for (i = 0; i < keylen; ++i)
588 		rev_key[keylen - i - 1] = key[i];
589 
590 	return serpent_setkey(tfm, rev_key, keylen);
591 }
592 
tnepres_encrypt(struct crypto_tfm * tfm,u8 * dst,const u8 * src)593 static void tnepres_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
594 {
595 	const u32 * const s = (const u32 * const)src;
596 	u32 * const d = (u32 * const)dst;
597 
598 	u32 rs[4], rd[4];
599 
600 	rs[0] = swab32(s[3]);
601 	rs[1] = swab32(s[2]);
602 	rs[2] = swab32(s[1]);
603 	rs[3] = swab32(s[0]);
604 
605 	serpent_encrypt(tfm, (u8 *)rd, (u8 *)rs);
606 
607 	d[0] = swab32(rd[3]);
608 	d[1] = swab32(rd[2]);
609 	d[2] = swab32(rd[1]);
610 	d[3] = swab32(rd[0]);
611 }
612 
tnepres_decrypt(struct crypto_tfm * tfm,u8 * dst,const u8 * src)613 static void tnepres_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
614 {
615 	const u32 * const s = (const u32 * const)src;
616 	u32 * const d = (u32 * const)dst;
617 
618 	u32 rs[4], rd[4];
619 
620 	rs[0] = swab32(s[3]);
621 	rs[1] = swab32(s[2]);
622 	rs[2] = swab32(s[1]);
623 	rs[3] = swab32(s[0]);
624 
625 	serpent_decrypt(tfm, (u8 *)rd, (u8 *)rs);
626 
627 	d[0] = swab32(rd[3]);
628 	d[1] = swab32(rd[2]);
629 	d[2] = swab32(rd[1]);
630 	d[3] = swab32(rd[0]);
631 }
632 
633 static struct crypto_alg srp_algs[2] = { {
634 	.cra_name		=	"serpent",
635 	.cra_driver_name	=	"serpent-generic",
636 	.cra_priority		=	100,
637 	.cra_flags		=	CRYPTO_ALG_TYPE_CIPHER,
638 	.cra_blocksize		=	SERPENT_BLOCK_SIZE,
639 	.cra_ctxsize		=	sizeof(struct serpent_ctx),
640 	.cra_alignmask		=	3,
641 	.cra_module		=	THIS_MODULE,
642 	.cra_u			=	{ .cipher = {
643 	.cia_min_keysize	=	SERPENT_MIN_KEY_SIZE,
644 	.cia_max_keysize	=	SERPENT_MAX_KEY_SIZE,
645 	.cia_setkey		=	serpent_setkey,
646 	.cia_encrypt		=	serpent_encrypt,
647 	.cia_decrypt		=	serpent_decrypt } }
648 }, {
649 	.cra_name		=	"tnepres",
650 	.cra_flags		=	CRYPTO_ALG_TYPE_CIPHER,
651 	.cra_blocksize		=	SERPENT_BLOCK_SIZE,
652 	.cra_ctxsize		=	sizeof(struct serpent_ctx),
653 	.cra_alignmask		=	3,
654 	.cra_module		=	THIS_MODULE,
655 	.cra_u			=	{ .cipher = {
656 	.cia_min_keysize	=	SERPENT_MIN_KEY_SIZE,
657 	.cia_max_keysize	=	SERPENT_MAX_KEY_SIZE,
658 	.cia_setkey		=	tnepres_setkey,
659 	.cia_encrypt		=	tnepres_encrypt,
660 	.cia_decrypt		=	tnepres_decrypt } }
661 } };
662 
serpent_mod_init(void)663 static int __init serpent_mod_init(void)
664 {
665 	return crypto_register_algs(srp_algs, ARRAY_SIZE(srp_algs));
666 }
667 
serpent_mod_fini(void)668 static void __exit serpent_mod_fini(void)
669 {
670 	crypto_unregister_algs(srp_algs, ARRAY_SIZE(srp_algs));
671 }
672 
673 module_init(serpent_mod_init);
674 module_exit(serpent_mod_fini);
675 
676 MODULE_LICENSE("GPL");
677 MODULE_DESCRIPTION("Serpent and tnepres (kerneli compatible serpent reversed) Cipher Algorithm");
678 MODULE_AUTHOR("Dag Arne Osvik <osvik@ii.uib.no>");
679 MODULE_ALIAS_CRYPTO("tnepres");
680 MODULE_ALIAS_CRYPTO("serpent");
681 MODULE_ALIAS_CRYPTO("serpent-generic");
682