xref: /openbmc/linux/crypto/serpent_generic.c (revision 2fa5ebe3)
1 // SPDX-License-Identifier: GPL-2.0-or-later
2 /*
3  * Cryptographic API.
4  *
5  * Serpent Cipher Algorithm.
6  *
7  * Copyright (C) 2002 Dag Arne Osvik <osvik@ii.uib.no>
8  */
9 
10 #include <crypto/algapi.h>
11 #include <linux/init.h>
12 #include <linux/module.h>
13 #include <linux/errno.h>
14 #include <asm/unaligned.h>
15 #include <linux/types.h>
16 #include <crypto/serpent.h>
17 
18 /* Key is padded to the maximum of 256 bits before round key generation.
19  * Any key length <= 256 bits (32 bytes) is allowed by the algorithm.
20  */
21 
22 #define PHI 0x9e3779b9UL
23 
24 #define keyiter(a, b, c, d, i, j) \
25 	({ b ^= d; b ^= c; b ^= a; b ^= PHI ^ i; b = rol32(b, 11); k[j] = b; })
26 
27 #define loadkeys(x0, x1, x2, x3, i) \
28 	({ x0 = k[i]; x1 = k[i+1]; x2 = k[i+2]; x3 = k[i+3]; })
29 
30 #define storekeys(x0, x1, x2, x3, i) \
31 	({ k[i] = x0; k[i+1] = x1; k[i+2] = x2; k[i+3] = x3; })
32 
33 #define store_and_load_keys(x0, x1, x2, x3, s, l) \
34 	({ storekeys(x0, x1, x2, x3, s); loadkeys(x0, x1, x2, x3, l); })
35 
36 #define K(x0, x1, x2, x3, i) ({				\
37 	x3 ^= k[4*(i)+3];        x2 ^= k[4*(i)+2];	\
38 	x1 ^= k[4*(i)+1];        x0 ^= k[4*(i)+0];	\
39 	})
40 
41 #define LK(x0, x1, x2, x3, x4, i) ({					   \
42 							x0 = rol32(x0, 13);\
43 	x2 = rol32(x2, 3);	x1 ^= x0;		x4  = x0 << 3;	   \
44 	x3 ^= x2;		x1 ^= x2;				   \
45 	x1 = rol32(x1, 1);	x3 ^= x4;				   \
46 	x3 = rol32(x3, 7);	x4  = x1;				   \
47 	x0 ^= x1;		x4 <<= 7;		x2 ^= x3;	   \
48 	x0 ^= x3;		x2 ^= x4;		x3 ^= k[4*i+3];	   \
49 	x1 ^= k[4*i+1];		x0 = rol32(x0, 5);	x2 = rol32(x2, 22);\
50 	x0 ^= k[4*i+0];		x2 ^= k[4*i+2];				   \
51 	})
52 
53 #define KL(x0, x1, x2, x3, x4, i) ({					   \
54 	x0 ^= k[4*i+0];		x1 ^= k[4*i+1];		x2 ^= k[4*i+2];	   \
55 	x3 ^= k[4*i+3];		x0 = ror32(x0, 5);	x2 = ror32(x2, 22);\
56 	x4 =  x1;		x2 ^= x3;		x0 ^= x3;	   \
57 	x4 <<= 7;		x0 ^= x1;		x1 = ror32(x1, 1); \
58 	x2 ^= x4;		x3 = ror32(x3, 7);	x4 = x0 << 3;	   \
59 	x1 ^= x0;		x3 ^= x4;		x0 = ror32(x0, 13);\
60 	x1 ^= x2;		x3 ^= x2;		x2 = ror32(x2, 3); \
61 	})
62 
63 #define S0(x0, x1, x2, x3, x4) ({			\
64 					x4  = x3;	\
65 	x3 |= x0;	x0 ^= x4;	x4 ^= x2;	\
66 	x4 = ~x4;	x3 ^= x1;	x1 &= x0;	\
67 	x1 ^= x4;	x2 ^= x0;	x0 ^= x3;	\
68 	x4 |= x0;	x0 ^= x2;	x2 &= x1;	\
69 	x3 ^= x2;	x1 = ~x1;	x2 ^= x4;	\
70 	x1 ^= x2;					\
71 	})
72 
73 #define S1(x0, x1, x2, x3, x4) ({			\
74 					x4  = x1;	\
75 	x1 ^= x0;	x0 ^= x3;	x3 = ~x3;	\
76 	x4 &= x1;	x0 |= x1;	x3 ^= x2;	\
77 	x0 ^= x3;	x1 ^= x3;	x3 ^= x4;	\
78 	x1 |= x4;	x4 ^= x2;	x2 &= x0;	\
79 	x2 ^= x1;	x1 |= x0;	x0 = ~x0;	\
80 	x0 ^= x2;	x4 ^= x1;			\
81 	})
82 
83 #define S2(x0, x1, x2, x3, x4) ({			\
84 					x3 = ~x3;	\
85 	x1 ^= x0;	x4  = x0;	x0 &= x2;	\
86 	x0 ^= x3;	x3 |= x4;	x2 ^= x1;	\
87 	x3 ^= x1;	x1 &= x0;	x0 ^= x2;	\
88 	x2 &= x3;	x3 |= x1;	x0 = ~x0;	\
89 	x3 ^= x0;	x4 ^= x0;	x0 ^= x2;	\
90 	x1 |= x2;					\
91 	})
92 
93 #define S3(x0, x1, x2, x3, x4) ({			\
94 					x4  = x1;	\
95 	x1 ^= x3;	x3 |= x0;	x4 &= x0;	\
96 	x0 ^= x2;	x2 ^= x1;	x1 &= x3;	\
97 	x2 ^= x3;	x0 |= x4;	x4 ^= x3;	\
98 	x1 ^= x0;	x0 &= x3;	x3 &= x4;	\
99 	x3 ^= x2;	x4 |= x1;	x2 &= x1;	\
100 	x4 ^= x3;	x0 ^= x3;	x3 ^= x2;	\
101 	})
102 
103 #define S4(x0, x1, x2, x3, x4) ({			\
104 					x4  = x3;	\
105 	x3 &= x0;	x0 ^= x4;			\
106 	x3 ^= x2;	x2 |= x4;	x0 ^= x1;	\
107 	x4 ^= x3;	x2 |= x0;			\
108 	x2 ^= x1;	x1 &= x0;			\
109 	x1 ^= x4;	x4 &= x2;	x2 ^= x3;	\
110 	x4 ^= x0;	x3 |= x1;	x1 = ~x1;	\
111 	x3 ^= x0;					\
112 	})
113 
114 #define S5(x0, x1, x2, x3, x4) ({			\
115 	x4  = x1;	x1 |= x0;			\
116 	x2 ^= x1;	x3 = ~x3;	x4 ^= x0;	\
117 	x0 ^= x2;	x1 &= x4;	x4 |= x3;	\
118 	x4 ^= x0;	x0 &= x3;	x1 ^= x3;	\
119 	x3 ^= x2;	x0 ^= x1;	x2 &= x4;	\
120 	x1 ^= x2;	x2 &= x0;			\
121 	x3 ^= x2;					\
122 	})
123 
124 #define S6(x0, x1, x2, x3, x4) ({			\
125 					x4  = x1;	\
126 	x3 ^= x0;	x1 ^= x2;	x2 ^= x0;	\
127 	x0 &= x3;	x1 |= x3;	x4 = ~x4;	\
128 	x0 ^= x1;	x1 ^= x2;			\
129 	x3 ^= x4;	x4 ^= x0;	x2 &= x0;	\
130 	x4 ^= x1;	x2 ^= x3;	x3 &= x1;	\
131 	x3 ^= x0;	x1 ^= x2;			\
132 	})
133 
134 #define S7(x0, x1, x2, x3, x4) ({			\
135 					x1 = ~x1;	\
136 	x4  = x1;	x0 = ~x0;	x1 &= x2;	\
137 	x1 ^= x3;	x3 |= x4;	x4 ^= x2;	\
138 	x2 ^= x3;	x3 ^= x0;	x0 |= x1;	\
139 	x2 &= x0;	x0 ^= x4;	x4 ^= x3;	\
140 	x3 &= x0;	x4 ^= x1;			\
141 	x2 ^= x4;	x3 ^= x1;	x4 |= x0;	\
142 	x4 ^= x1;					\
143 	})
144 
145 #define SI0(x0, x1, x2, x3, x4) ({			\
146 			x4  = x3;	x1 ^= x0;	\
147 	x3 |= x1;	x4 ^= x1;	x0 = ~x0;	\
148 	x2 ^= x3;	x3 ^= x0;	x0 &= x1;	\
149 	x0 ^= x2;	x2 &= x3;	x3 ^= x4;	\
150 	x2 ^= x3;	x1 ^= x3;	x3 &= x0;	\
151 	x1 ^= x0;	x0 ^= x2;	x4 ^= x3;	\
152 	})
153 
154 #define SI1(x0, x1, x2, x3, x4) ({			\
155 	x1 ^= x3;	x4  = x0;			\
156 	x0 ^= x2;	x2 = ~x2;	x4 |= x1;	\
157 	x4 ^= x3;	x3 &= x1;	x1 ^= x2;	\
158 	x2 &= x4;	x4 ^= x1;	x1 |= x3;	\
159 	x3 ^= x0;	x2 ^= x0;	x0 |= x4;	\
160 	x2 ^= x4;	x1 ^= x0;			\
161 	x4 ^= x1;					\
162 	})
163 
164 #define SI2(x0, x1, x2, x3, x4) ({			\
165 	x2 ^= x1;	x4  = x3;	x3 = ~x3;	\
166 	x3 |= x2;	x2 ^= x4;	x4 ^= x0;	\
167 	x3 ^= x1;	x1 |= x2;	x2 ^= x0;	\
168 	x1 ^= x4;	x4 |= x3;	x2 ^= x3;	\
169 	x4 ^= x2;	x2 &= x1;			\
170 	x2 ^= x3;	x3 ^= x4;	x4 ^= x0;	\
171 	})
172 
173 #define SI3(x0, x1, x2, x3, x4) ({			\
174 					x2 ^= x1;	\
175 	x4  = x1;	x1 &= x2;			\
176 	x1 ^= x0;	x0 |= x4;	x4 ^= x3;	\
177 	x0 ^= x3;	x3 |= x1;	x1 ^= x2;	\
178 	x1 ^= x3;	x0 ^= x2;	x2 ^= x3;	\
179 	x3 &= x1;	x1 ^= x0;	x0 &= x2;	\
180 	x4 ^= x3;	x3 ^= x0;	x0 ^= x1;	\
181 	})
182 
183 #define SI4(x0, x1, x2, x3, x4) ({			\
184 	x2 ^= x3;	x4  = x0;	x0 &= x1;	\
185 	x0 ^= x2;	x2 |= x3;	x4 = ~x4;	\
186 	x1 ^= x0;	x0 ^= x2;	x2 &= x4;	\
187 	x2 ^= x0;	x0 |= x4;			\
188 	x0 ^= x3;	x3 &= x2;			\
189 	x4 ^= x3;	x3 ^= x1;	x1 &= x0;	\
190 	x4 ^= x1;	x0 ^= x3;			\
191 	})
192 
193 #define SI5(x0, x1, x2, x3, x4) ({			\
194 			x4  = x1;	x1 |= x2;	\
195 	x2 ^= x4;	x1 ^= x3;	x3 &= x4;	\
196 	x2 ^= x3;	x3 |= x0;	x0 = ~x0;	\
197 	x3 ^= x2;	x2 |= x0;	x4 ^= x1;	\
198 	x2 ^= x4;	x4 &= x0;	x0 ^= x1;	\
199 	x1 ^= x3;	x0 &= x2;	x2 ^= x3;	\
200 	x0 ^= x2;	x2 ^= x4;	x4 ^= x3;	\
201 	})
202 
203 #define SI6(x0, x1, x2, x3, x4) ({			\
204 			x0 ^= x2;			\
205 	x4  = x0;	x0 &= x3;	x2 ^= x3;	\
206 	x0 ^= x2;	x3 ^= x1;	x2 |= x4;	\
207 	x2 ^= x3;	x3 &= x0;	x0 = ~x0;	\
208 	x3 ^= x1;	x1 &= x2;	x4 ^= x0;	\
209 	x3 ^= x4;	x4 ^= x2;	x0 ^= x1;	\
210 	x2 ^= x0;					\
211 	})
212 
213 #define SI7(x0, x1, x2, x3, x4) ({			\
214 	x4  = x3;	x3 &= x0;	x0 ^= x2;	\
215 	x2 |= x4;	x4 ^= x1;	x0 = ~x0;	\
216 	x1 |= x3;	x4 ^= x0;	x0 &= x2;	\
217 	x0 ^= x1;	x1 &= x2;	x3 ^= x2;	\
218 	x4 ^= x3;	x2 &= x3;	x3 |= x0;	\
219 	x1 ^= x4;	x3 ^= x4;	x4 &= x0;	\
220 	x4 ^= x2;					\
221 	})
222 
223 /*
224  * both gcc and clang have misoptimized this function in the past,
225  * producing horrible object code from spilling temporary variables
226  * on the stack. Forcing this part out of line avoids that.
227  */
228 static noinline void __serpent_setkey_sbox(u32 r0, u32 r1, u32 r2,
229 					   u32 r3, u32 r4, u32 *k)
230 {
231 	k += 100;
232 	S3(r3, r4, r0, r1, r2); store_and_load_keys(r1, r2, r4, r3, 28, 24);
233 	S4(r1, r2, r4, r3, r0); store_and_load_keys(r2, r4, r3, r0, 24, 20);
234 	S5(r2, r4, r3, r0, r1); store_and_load_keys(r1, r2, r4, r0, 20, 16);
235 	S6(r1, r2, r4, r0, r3); store_and_load_keys(r4, r3, r2, r0, 16, 12);
236 	S7(r4, r3, r2, r0, r1); store_and_load_keys(r1, r2, r0, r4, 12, 8);
237 	S0(r1, r2, r0, r4, r3); store_and_load_keys(r0, r2, r4, r1, 8, 4);
238 	S1(r0, r2, r4, r1, r3); store_and_load_keys(r3, r4, r1, r0, 4, 0);
239 	S2(r3, r4, r1, r0, r2); store_and_load_keys(r2, r4, r3, r0, 0, -4);
240 	S3(r2, r4, r3, r0, r1); store_and_load_keys(r0, r1, r4, r2, -4, -8);
241 	S4(r0, r1, r4, r2, r3); store_and_load_keys(r1, r4, r2, r3, -8, -12);
242 	S5(r1, r4, r2, r3, r0); store_and_load_keys(r0, r1, r4, r3, -12, -16);
243 	S6(r0, r1, r4, r3, r2); store_and_load_keys(r4, r2, r1, r3, -16, -20);
244 	S7(r4, r2, r1, r3, r0); store_and_load_keys(r0, r1, r3, r4, -20, -24);
245 	S0(r0, r1, r3, r4, r2); store_and_load_keys(r3, r1, r4, r0, -24, -28);
246 	k -= 50;
247 	S1(r3, r1, r4, r0, r2); store_and_load_keys(r2, r4, r0, r3, 22, 18);
248 	S2(r2, r4, r0, r3, r1); store_and_load_keys(r1, r4, r2, r3, 18, 14);
249 	S3(r1, r4, r2, r3, r0); store_and_load_keys(r3, r0, r4, r1, 14, 10);
250 	S4(r3, r0, r4, r1, r2); store_and_load_keys(r0, r4, r1, r2, 10, 6);
251 	S5(r0, r4, r1, r2, r3); store_and_load_keys(r3, r0, r4, r2, 6, 2);
252 	S6(r3, r0, r4, r2, r1); store_and_load_keys(r4, r1, r0, r2, 2, -2);
253 	S7(r4, r1, r0, r2, r3); store_and_load_keys(r3, r0, r2, r4, -2, -6);
254 	S0(r3, r0, r2, r4, r1); store_and_load_keys(r2, r0, r4, r3, -6, -10);
255 	S1(r2, r0, r4, r3, r1); store_and_load_keys(r1, r4, r3, r2, -10, -14);
256 	S2(r1, r4, r3, r2, r0); store_and_load_keys(r0, r4, r1, r2, -14, -18);
257 	S3(r0, r4, r1, r2, r3); store_and_load_keys(r2, r3, r4, r0, -18, -22);
258 	k -= 50;
259 	S4(r2, r3, r4, r0, r1); store_and_load_keys(r3, r4, r0, r1, 28, 24);
260 	S5(r3, r4, r0, r1, r2); store_and_load_keys(r2, r3, r4, r1, 24, 20);
261 	S6(r2, r3, r4, r1, r0); store_and_load_keys(r4, r0, r3, r1, 20, 16);
262 	S7(r4, r0, r3, r1, r2); store_and_load_keys(r2, r3, r1, r4, 16, 12);
263 	S0(r2, r3, r1, r4, r0); store_and_load_keys(r1, r3, r4, r2, 12, 8);
264 	S1(r1, r3, r4, r2, r0); store_and_load_keys(r0, r4, r2, r1, 8, 4);
265 	S2(r0, r4, r2, r1, r3); store_and_load_keys(r3, r4, r0, r1, 4, 0);
266 	S3(r3, r4, r0, r1, r2); storekeys(r1, r2, r4, r3, 0);
267 }
268 
269 int __serpent_setkey(struct serpent_ctx *ctx, const u8 *key,
270 		     unsigned int keylen)
271 {
272 	u32 *k = ctx->expkey;
273 	u8  *k8 = (u8 *)k;
274 	u32 r0, r1, r2, r3, r4;
275 	__le32 *lk;
276 	int i;
277 
278 	/* Copy key, add padding */
279 
280 	for (i = 0; i < keylen; ++i)
281 		k8[i] = key[i];
282 	if (i < SERPENT_MAX_KEY_SIZE)
283 		k8[i++] = 1;
284 	while (i < SERPENT_MAX_KEY_SIZE)
285 		k8[i++] = 0;
286 
287 	lk = (__le32 *)k;
288 	k[0] = le32_to_cpu(lk[0]);
289 	k[1] = le32_to_cpu(lk[1]);
290 	k[2] = le32_to_cpu(lk[2]);
291 	k[3] = le32_to_cpu(lk[3]);
292 	k[4] = le32_to_cpu(lk[4]);
293 	k[5] = le32_to_cpu(lk[5]);
294 	k[6] = le32_to_cpu(lk[6]);
295 	k[7] = le32_to_cpu(lk[7]);
296 
297 	/* Expand key using polynomial */
298 
299 	r0 = k[3];
300 	r1 = k[4];
301 	r2 = k[5];
302 	r3 = k[6];
303 	r4 = k[7];
304 
305 	keyiter(k[0], r0, r4, r2, 0, 0);
306 	keyiter(k[1], r1, r0, r3, 1, 1);
307 	keyiter(k[2], r2, r1, r4, 2, 2);
308 	keyiter(k[3], r3, r2, r0, 3, 3);
309 	keyiter(k[4], r4, r3, r1, 4, 4);
310 	keyiter(k[5], r0, r4, r2, 5, 5);
311 	keyiter(k[6], r1, r0, r3, 6, 6);
312 	keyiter(k[7], r2, r1, r4, 7, 7);
313 
314 	keyiter(k[0], r3, r2, r0, 8, 8);
315 	keyiter(k[1], r4, r3, r1, 9, 9);
316 	keyiter(k[2], r0, r4, r2, 10, 10);
317 	keyiter(k[3], r1, r0, r3, 11, 11);
318 	keyiter(k[4], r2, r1, r4, 12, 12);
319 	keyiter(k[5], r3, r2, r0, 13, 13);
320 	keyiter(k[6], r4, r3, r1, 14, 14);
321 	keyiter(k[7], r0, r4, r2, 15, 15);
322 	keyiter(k[8], r1, r0, r3, 16, 16);
323 	keyiter(k[9], r2, r1, r4, 17, 17);
324 	keyiter(k[10], r3, r2, r0, 18, 18);
325 	keyiter(k[11], r4, r3, r1, 19, 19);
326 	keyiter(k[12], r0, r4, r2, 20, 20);
327 	keyiter(k[13], r1, r0, r3, 21, 21);
328 	keyiter(k[14], r2, r1, r4, 22, 22);
329 	keyiter(k[15], r3, r2, r0, 23, 23);
330 	keyiter(k[16], r4, r3, r1, 24, 24);
331 	keyiter(k[17], r0, r4, r2, 25, 25);
332 	keyiter(k[18], r1, r0, r3, 26, 26);
333 	keyiter(k[19], r2, r1, r4, 27, 27);
334 	keyiter(k[20], r3, r2, r0, 28, 28);
335 	keyiter(k[21], r4, r3, r1, 29, 29);
336 	keyiter(k[22], r0, r4, r2, 30, 30);
337 	keyiter(k[23], r1, r0, r3, 31, 31);
338 
339 	k += 50;
340 
341 	keyiter(k[-26], r2, r1, r4, 32, -18);
342 	keyiter(k[-25], r3, r2, r0, 33, -17);
343 	keyiter(k[-24], r4, r3, r1, 34, -16);
344 	keyiter(k[-23], r0, r4, r2, 35, -15);
345 	keyiter(k[-22], r1, r0, r3, 36, -14);
346 	keyiter(k[-21], r2, r1, r4, 37, -13);
347 	keyiter(k[-20], r3, r2, r0, 38, -12);
348 	keyiter(k[-19], r4, r3, r1, 39, -11);
349 	keyiter(k[-18], r0, r4, r2, 40, -10);
350 	keyiter(k[-17], r1, r0, r3, 41, -9);
351 	keyiter(k[-16], r2, r1, r4, 42, -8);
352 	keyiter(k[-15], r3, r2, r0, 43, -7);
353 	keyiter(k[-14], r4, r3, r1, 44, -6);
354 	keyiter(k[-13], r0, r4, r2, 45, -5);
355 	keyiter(k[-12], r1, r0, r3, 46, -4);
356 	keyiter(k[-11], r2, r1, r4, 47, -3);
357 	keyiter(k[-10], r3, r2, r0, 48, -2);
358 	keyiter(k[-9], r4, r3, r1, 49, -1);
359 	keyiter(k[-8], r0, r4, r2, 50, 0);
360 	keyiter(k[-7], r1, r0, r3, 51, 1);
361 	keyiter(k[-6], r2, r1, r4, 52, 2);
362 	keyiter(k[-5], r3, r2, r0, 53, 3);
363 	keyiter(k[-4], r4, r3, r1, 54, 4);
364 	keyiter(k[-3], r0, r4, r2, 55, 5);
365 	keyiter(k[-2], r1, r0, r3, 56, 6);
366 	keyiter(k[-1], r2, r1, r4, 57, 7);
367 	keyiter(k[0], r3, r2, r0, 58, 8);
368 	keyiter(k[1], r4, r3, r1, 59, 9);
369 	keyiter(k[2], r0, r4, r2, 60, 10);
370 	keyiter(k[3], r1, r0, r3, 61, 11);
371 	keyiter(k[4], r2, r1, r4, 62, 12);
372 	keyiter(k[5], r3, r2, r0, 63, 13);
373 	keyiter(k[6], r4, r3, r1, 64, 14);
374 	keyiter(k[7], r0, r4, r2, 65, 15);
375 	keyiter(k[8], r1, r0, r3, 66, 16);
376 	keyiter(k[9], r2, r1, r4, 67, 17);
377 	keyiter(k[10], r3, r2, r0, 68, 18);
378 	keyiter(k[11], r4, r3, r1, 69, 19);
379 	keyiter(k[12], r0, r4, r2, 70, 20);
380 	keyiter(k[13], r1, r0, r3, 71, 21);
381 	keyiter(k[14], r2, r1, r4, 72, 22);
382 	keyiter(k[15], r3, r2, r0, 73, 23);
383 	keyiter(k[16], r4, r3, r1, 74, 24);
384 	keyiter(k[17], r0, r4, r2, 75, 25);
385 	keyiter(k[18], r1, r0, r3, 76, 26);
386 	keyiter(k[19], r2, r1, r4, 77, 27);
387 	keyiter(k[20], r3, r2, r0, 78, 28);
388 	keyiter(k[21], r4, r3, r1, 79, 29);
389 	keyiter(k[22], r0, r4, r2, 80, 30);
390 	keyiter(k[23], r1, r0, r3, 81, 31);
391 
392 	k += 50;
393 
394 	keyiter(k[-26], r2, r1, r4, 82, -18);
395 	keyiter(k[-25], r3, r2, r0, 83, -17);
396 	keyiter(k[-24], r4, r3, r1, 84, -16);
397 	keyiter(k[-23], r0, r4, r2, 85, -15);
398 	keyiter(k[-22], r1, r0, r3, 86, -14);
399 	keyiter(k[-21], r2, r1, r4, 87, -13);
400 	keyiter(k[-20], r3, r2, r0, 88, -12);
401 	keyiter(k[-19], r4, r3, r1, 89, -11);
402 	keyiter(k[-18], r0, r4, r2, 90, -10);
403 	keyiter(k[-17], r1, r0, r3, 91, -9);
404 	keyiter(k[-16], r2, r1, r4, 92, -8);
405 	keyiter(k[-15], r3, r2, r0, 93, -7);
406 	keyiter(k[-14], r4, r3, r1, 94, -6);
407 	keyiter(k[-13], r0, r4, r2, 95, -5);
408 	keyiter(k[-12], r1, r0, r3, 96, -4);
409 	keyiter(k[-11], r2, r1, r4, 97, -3);
410 	keyiter(k[-10], r3, r2, r0, 98, -2);
411 	keyiter(k[-9], r4, r3, r1, 99, -1);
412 	keyiter(k[-8], r0, r4, r2, 100, 0);
413 	keyiter(k[-7], r1, r0, r3, 101, 1);
414 	keyiter(k[-6], r2, r1, r4, 102, 2);
415 	keyiter(k[-5], r3, r2, r0, 103, 3);
416 	keyiter(k[-4], r4, r3, r1, 104, 4);
417 	keyiter(k[-3], r0, r4, r2, 105, 5);
418 	keyiter(k[-2], r1, r0, r3, 106, 6);
419 	keyiter(k[-1], r2, r1, r4, 107, 7);
420 	keyiter(k[0], r3, r2, r0, 108, 8);
421 	keyiter(k[1], r4, r3, r1, 109, 9);
422 	keyiter(k[2], r0, r4, r2, 110, 10);
423 	keyiter(k[3], r1, r0, r3, 111, 11);
424 	keyiter(k[4], r2, r1, r4, 112, 12);
425 	keyiter(k[5], r3, r2, r0, 113, 13);
426 	keyiter(k[6], r4, r3, r1, 114, 14);
427 	keyiter(k[7], r0, r4, r2, 115, 15);
428 	keyiter(k[8], r1, r0, r3, 116, 16);
429 	keyiter(k[9], r2, r1, r4, 117, 17);
430 	keyiter(k[10], r3, r2, r0, 118, 18);
431 	keyiter(k[11], r4, r3, r1, 119, 19);
432 	keyiter(k[12], r0, r4, r2, 120, 20);
433 	keyiter(k[13], r1, r0, r3, 121, 21);
434 	keyiter(k[14], r2, r1, r4, 122, 22);
435 	keyiter(k[15], r3, r2, r0, 123, 23);
436 	keyiter(k[16], r4, r3, r1, 124, 24);
437 	keyiter(k[17], r0, r4, r2, 125, 25);
438 	keyiter(k[18], r1, r0, r3, 126, 26);
439 	keyiter(k[19], r2, r1, r4, 127, 27);
440 	keyiter(k[20], r3, r2, r0, 128, 28);
441 	keyiter(k[21], r4, r3, r1, 129, 29);
442 	keyiter(k[22], r0, r4, r2, 130, 30);
443 	keyiter(k[23], r1, r0, r3, 131, 31);
444 
445 	/* Apply S-boxes */
446 	__serpent_setkey_sbox(r0, r1, r2, r3, r4, ctx->expkey);
447 
448 	return 0;
449 }
450 EXPORT_SYMBOL_GPL(__serpent_setkey);
451 
452 int serpent_setkey(struct crypto_tfm *tfm, const u8 *key, unsigned int keylen)
453 {
454 	return __serpent_setkey(crypto_tfm_ctx(tfm), key, keylen);
455 }
456 EXPORT_SYMBOL_GPL(serpent_setkey);
457 
458 void __serpent_encrypt(const void *c, u8 *dst, const u8 *src)
459 {
460 	const struct serpent_ctx *ctx = c;
461 	const u32 *k = ctx->expkey;
462 	u32	r0, r1, r2, r3, r4;
463 
464 	r0 = get_unaligned_le32(src);
465 	r1 = get_unaligned_le32(src + 4);
466 	r2 = get_unaligned_le32(src + 8);
467 	r3 = get_unaligned_le32(src + 12);
468 
469 					K(r0, r1, r2, r3, 0);
470 	S0(r0, r1, r2, r3, r4);		LK(r2, r1, r3, r0, r4, 1);
471 	S1(r2, r1, r3, r0, r4);		LK(r4, r3, r0, r2, r1, 2);
472 	S2(r4, r3, r0, r2, r1);		LK(r1, r3, r4, r2, r0, 3);
473 	S3(r1, r3, r4, r2, r0);		LK(r2, r0, r3, r1, r4, 4);
474 	S4(r2, r0, r3, r1, r4);		LK(r0, r3, r1, r4, r2, 5);
475 	S5(r0, r3, r1, r4, r2);		LK(r2, r0, r3, r4, r1, 6);
476 	S6(r2, r0, r3, r4, r1);		LK(r3, r1, r0, r4, r2, 7);
477 	S7(r3, r1, r0, r4, r2);		LK(r2, r0, r4, r3, r1, 8);
478 	S0(r2, r0, r4, r3, r1);		LK(r4, r0, r3, r2, r1, 9);
479 	S1(r4, r0, r3, r2, r1);		LK(r1, r3, r2, r4, r0, 10);
480 	S2(r1, r3, r2, r4, r0);		LK(r0, r3, r1, r4, r2, 11);
481 	S3(r0, r3, r1, r4, r2);		LK(r4, r2, r3, r0, r1, 12);
482 	S4(r4, r2, r3, r0, r1);		LK(r2, r3, r0, r1, r4, 13);
483 	S5(r2, r3, r0, r1, r4);		LK(r4, r2, r3, r1, r0, 14);
484 	S6(r4, r2, r3, r1, r0);		LK(r3, r0, r2, r1, r4, 15);
485 	S7(r3, r0, r2, r1, r4);		LK(r4, r2, r1, r3, r0, 16);
486 	S0(r4, r2, r1, r3, r0);		LK(r1, r2, r3, r4, r0, 17);
487 	S1(r1, r2, r3, r4, r0);		LK(r0, r3, r4, r1, r2, 18);
488 	S2(r0, r3, r4, r1, r2);		LK(r2, r3, r0, r1, r4, 19);
489 	S3(r2, r3, r0, r1, r4);		LK(r1, r4, r3, r2, r0, 20);
490 	S4(r1, r4, r3, r2, r0);		LK(r4, r3, r2, r0, r1, 21);
491 	S5(r4, r3, r2, r0, r1);		LK(r1, r4, r3, r0, r2, 22);
492 	S6(r1, r4, r3, r0, r2);		LK(r3, r2, r4, r0, r1, 23);
493 	S7(r3, r2, r4, r0, r1);		LK(r1, r4, r0, r3, r2, 24);
494 	S0(r1, r4, r0, r3, r2);		LK(r0, r4, r3, r1, r2, 25);
495 	S1(r0, r4, r3, r1, r2);		LK(r2, r3, r1, r0, r4, 26);
496 	S2(r2, r3, r1, r0, r4);		LK(r4, r3, r2, r0, r1, 27);
497 	S3(r4, r3, r2, r0, r1);		LK(r0, r1, r3, r4, r2, 28);
498 	S4(r0, r1, r3, r4, r2);		LK(r1, r3, r4, r2, r0, 29);
499 	S5(r1, r3, r4, r2, r0);		LK(r0, r1, r3, r2, r4, 30);
500 	S6(r0, r1, r3, r2, r4);		LK(r3, r4, r1, r2, r0, 31);
501 	S7(r3, r4, r1, r2, r0);		K(r0, r1, r2, r3, 32);
502 
503 	put_unaligned_le32(r0, dst);
504 	put_unaligned_le32(r1, dst + 4);
505 	put_unaligned_le32(r2, dst + 8);
506 	put_unaligned_le32(r3, dst + 12);
507 }
508 EXPORT_SYMBOL_GPL(__serpent_encrypt);
509 
510 static void serpent_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
511 {
512 	struct serpent_ctx *ctx = crypto_tfm_ctx(tfm);
513 
514 	__serpent_encrypt(ctx, dst, src);
515 }
516 
517 void __serpent_decrypt(const void *c, u8 *dst, const u8 *src)
518 {
519 	const struct serpent_ctx *ctx = c;
520 	const u32 *k = ctx->expkey;
521 	u32	r0, r1, r2, r3, r4;
522 
523 	r0 = get_unaligned_le32(src);
524 	r1 = get_unaligned_le32(src + 4);
525 	r2 = get_unaligned_le32(src + 8);
526 	r3 = get_unaligned_le32(src + 12);
527 
528 					K(r0, r1, r2, r3, 32);
529 	SI7(r0, r1, r2, r3, r4);	KL(r1, r3, r0, r4, r2, 31);
530 	SI6(r1, r3, r0, r4, r2);	KL(r0, r2, r4, r1, r3, 30);
531 	SI5(r0, r2, r4, r1, r3);	KL(r2, r3, r0, r4, r1, 29);
532 	SI4(r2, r3, r0, r4, r1);	KL(r2, r0, r1, r4, r3, 28);
533 	SI3(r2, r0, r1, r4, r3);	KL(r1, r2, r3, r4, r0, 27);
534 	SI2(r1, r2, r3, r4, r0);	KL(r2, r0, r4, r3, r1, 26);
535 	SI1(r2, r0, r4, r3, r1);	KL(r1, r0, r4, r3, r2, 25);
536 	SI0(r1, r0, r4, r3, r2);	KL(r4, r2, r0, r1, r3, 24);
537 	SI7(r4, r2, r0, r1, r3);	KL(r2, r1, r4, r3, r0, 23);
538 	SI6(r2, r1, r4, r3, r0);	KL(r4, r0, r3, r2, r1, 22);
539 	SI5(r4, r0, r3, r2, r1);	KL(r0, r1, r4, r3, r2, 21);
540 	SI4(r0, r1, r4, r3, r2);	KL(r0, r4, r2, r3, r1, 20);
541 	SI3(r0, r4, r2, r3, r1);	KL(r2, r0, r1, r3, r4, 19);
542 	SI2(r2, r0, r1, r3, r4);	KL(r0, r4, r3, r1, r2, 18);
543 	SI1(r0, r4, r3, r1, r2);	KL(r2, r4, r3, r1, r0, 17);
544 	SI0(r2, r4, r3, r1, r0);	KL(r3, r0, r4, r2, r1, 16);
545 	SI7(r3, r0, r4, r2, r1);	KL(r0, r2, r3, r1, r4, 15);
546 	SI6(r0, r2, r3, r1, r4);	KL(r3, r4, r1, r0, r2, 14);
547 	SI5(r3, r4, r1, r0, r2);	KL(r4, r2, r3, r1, r0, 13);
548 	SI4(r4, r2, r3, r1, r0);	KL(r4, r3, r0, r1, r2, 12);
549 	SI3(r4, r3, r0, r1, r2);	KL(r0, r4, r2, r1, r3, 11);
550 	SI2(r0, r4, r2, r1, r3);	KL(r4, r3, r1, r2, r0, 10);
551 	SI1(r4, r3, r1, r2, r0);	KL(r0, r3, r1, r2, r4, 9);
552 	SI0(r0, r3, r1, r2, r4);	KL(r1, r4, r3, r0, r2, 8);
553 	SI7(r1, r4, r3, r0, r2);	KL(r4, r0, r1, r2, r3, 7);
554 	SI6(r4, r0, r1, r2, r3);	KL(r1, r3, r2, r4, r0, 6);
555 	SI5(r1, r3, r2, r4, r0);	KL(r3, r0, r1, r2, r4, 5);
556 	SI4(r3, r0, r1, r2, r4);	KL(r3, r1, r4, r2, r0, 4);
557 	SI3(r3, r1, r4, r2, r0);	KL(r4, r3, r0, r2, r1, 3);
558 	SI2(r4, r3, r0, r2, r1);	KL(r3, r1, r2, r0, r4, 2);
559 	SI1(r3, r1, r2, r0, r4);	KL(r4, r1, r2, r0, r3, 1);
560 	SI0(r4, r1, r2, r0, r3);	K(r2, r3, r1, r4, 0);
561 
562 	put_unaligned_le32(r2, dst);
563 	put_unaligned_le32(r3, dst + 4);
564 	put_unaligned_le32(r1, dst + 8);
565 	put_unaligned_le32(r4, dst + 12);
566 }
567 EXPORT_SYMBOL_GPL(__serpent_decrypt);
568 
569 static void serpent_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
570 {
571 	struct serpent_ctx *ctx = crypto_tfm_ctx(tfm);
572 
573 	__serpent_decrypt(ctx, dst, src);
574 }
575 
576 static struct crypto_alg srp_alg = {
577 	.cra_name		=	"serpent",
578 	.cra_driver_name	=	"serpent-generic",
579 	.cra_priority		=	100,
580 	.cra_flags		=	CRYPTO_ALG_TYPE_CIPHER,
581 	.cra_blocksize		=	SERPENT_BLOCK_SIZE,
582 	.cra_ctxsize		=	sizeof(struct serpent_ctx),
583 	.cra_module		=	THIS_MODULE,
584 	.cra_u			=	{ .cipher = {
585 	.cia_min_keysize	=	SERPENT_MIN_KEY_SIZE,
586 	.cia_max_keysize	=	SERPENT_MAX_KEY_SIZE,
587 	.cia_setkey		=	serpent_setkey,
588 	.cia_encrypt		=	serpent_encrypt,
589 	.cia_decrypt		=	serpent_decrypt } }
590 };
591 
592 static int __init serpent_mod_init(void)
593 {
594 	return crypto_register_alg(&srp_alg);
595 }
596 
597 static void __exit serpent_mod_fini(void)
598 {
599 	crypto_unregister_alg(&srp_alg);
600 }
601 
602 subsys_initcall(serpent_mod_init);
603 module_exit(serpent_mod_fini);
604 
605 MODULE_LICENSE("GPL");
606 MODULE_DESCRIPTION("Serpent Cipher Algorithm");
607 MODULE_AUTHOR("Dag Arne Osvik <osvik@ii.uib.no>");
608 MODULE_ALIAS_CRYPTO("serpent");
609 MODULE_ALIAS_CRYPTO("serpent-generic");
610