[SPARC64]: Fix %tstate ASI handling in start_thread{,32}()
[linux-2.6.git] / crypto / serpent.c
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: Ruben Jesus Garcia Hernandez <ruben@ugr.es>, 18.10.2004
10  *               Based on code by hvr
11  *
12  * This program is free software; you can redistribute it and/or modify
13  * it under the terms of the GNU General Public License as published by
14  * the Free Software Foundation; either version 2 of the License, or
15  * (at your option) any later version.
16  */
17
18 #include <linux/init.h>
19 #include <linux/module.h>
20 #include <linux/errno.h>
21 #include <asm/byteorder.h>
22 #include <linux/crypto.h>
23 #include <linux/types.h>
24
25 /* Key is padded to the maximum of 256 bits before round key generation.
26  * Any key length <= 256 bits (32 bytes) is allowed by the algorithm.
27  */
28
29 #define SERPENT_MIN_KEY_SIZE              0
30 #define SERPENT_MAX_KEY_SIZE             32
31 #define SERPENT_EXPKEY_WORDS            132
32 #define SERPENT_BLOCK_SIZE               16
33
34 #define PHI 0x9e3779b9UL
35
36 #define keyiter(a,b,c,d,i,j) \
37         b ^= d; b ^= c; b ^= a; b ^= PHI ^ i; b = rol32(b,11); k[j] = b;
38
39 #define loadkeys(x0,x1,x2,x3,i) \
40         x0=k[i]; x1=k[i+1]; x2=k[i+2]; x3=k[i+3];
41
42 #define storekeys(x0,x1,x2,x3,i) \
43         k[i]=x0; k[i+1]=x1; k[i+2]=x2; k[i+3]=x3;
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 #define LK(x0,x1,x2,x3,x4,i)                            \
50                                         x0=rol32(x0,13);\
51         x2=rol32(x2,3); x1 ^= x0;       x4  = x0 << 3;  \
52         x3 ^= x2;       x1 ^= x2;                       \
53         x1=rol32(x1,1); x3 ^= x4;                       \
54         x3=rol32(x3,7); x4  = x1;                       \
55         x0 ^= x1;       x4 <<= 7;       x2 ^= x3;       \
56         x0 ^= x3;       x2 ^= x4;       x3 ^= k[4*i+3]; \
57         x1 ^= k[4*i+1]; x0=rol32(x0,5); x2=rol32(x2,22);\
58         x0 ^= k[4*i+0]; x2 ^= k[4*i+2];
59
60 #define KL(x0,x1,x2,x3,x4,i)                            \
61         x0 ^= k[4*i+0]; x1 ^= k[4*i+1]; x2 ^= k[4*i+2]; \
62         x3 ^= k[4*i+3]; x0=ror32(x0,5); x2=ror32(x2,22);\
63         x4 =  x1;       x2 ^= x3;       x0 ^= x3;       \
64         x4 <<= 7;       x0 ^= x1;       x1=ror32(x1,1); \
65         x2 ^= x4;       x3=ror32(x3,7); x4 = x0 << 3;   \
66         x1 ^= x0;       x3 ^= x4;       x0=ror32(x0,13);\
67         x1 ^= x2;       x3 ^= x2;       x2=ror32(x2,3);
68
69 #define S0(x0,x1,x2,x3,x4)                              \
70                                         x4  = x3;       \
71         x3 |= x0;       x0 ^= x4;       x4 ^= x2;       \
72         x4 =~ x4;       x3 ^= x1;       x1 &= x0;       \
73         x1 ^= x4;       x2 ^= x0;       x0 ^= x3;       \
74         x4 |= x0;       x0 ^= x2;       x2 &= x1;       \
75         x3 ^= x2;       x1 =~ x1;       x2 ^= x4;       \
76         x1 ^= x2;
77
78 #define S1(x0,x1,x2,x3,x4)                              \
79                                         x4  = x1;       \
80         x1 ^= x0;       x0 ^= x3;       x3 =~ x3;       \
81         x4 &= x1;       x0 |= x1;       x3 ^= x2;       \
82         x0 ^= x3;       x1 ^= x3;       x3 ^= x4;       \
83         x1 |= x4;       x4 ^= x2;       x2 &= x0;       \
84         x2 ^= x1;       x1 |= x0;       x0 =~ x0;       \
85         x0 ^= x2;       x4 ^= x1;
86
87 #define S2(x0,x1,x2,x3,x4)                              \
88                                         x3 =~ x3;       \
89         x1 ^= x0;       x4  = x0;       x0 &= x2;       \
90         x0 ^= x3;       x3 |= x4;       x2 ^= x1;       \
91         x3 ^= x1;       x1 &= x0;       x0 ^= x2;       \
92         x2 &= x3;       x3 |= x1;       x0 =~ x0;       \
93         x3 ^= x0;       x4 ^= x0;       x0 ^= x2;       \
94         x1 |= x2;
95
96 #define S3(x0,x1,x2,x3,x4)                              \
97                                         x4  = x1;       \
98         x1 ^= x3;       x3 |= x0;       x4 &= x0;       \
99         x0 ^= x2;       x2 ^= x1;       x1 &= x3;       \
100         x2 ^= x3;       x0 |= x4;       x4 ^= x3;       \
101         x1 ^= x0;       x0 &= x3;       x3 &= x4;       \
102         x3 ^= x2;       x4 |= x1;       x2 &= x1;       \
103         x4 ^= x3;       x0 ^= x3;       x3 ^= x2;
104
105 #define S4(x0,x1,x2,x3,x4)                              \
106                                         x4  = x3;       \
107         x3 &= x0;       x0 ^= x4;                       \
108         x3 ^= x2;       x2 |= x4;       x0 ^= x1;       \
109         x4 ^= x3;       x2 |= x0;                       \
110         x2 ^= x1;       x1 &= x0;                       \
111         x1 ^= x4;       x4 &= x2;       x2 ^= x3;       \
112         x4 ^= x0;       x3 |= x1;       x1 =~ x1;       \
113         x3 ^= x0;
114
115 #define S5(x0,x1,x2,x3,x4)                              \
116         x4  = x1;       x1 |= x0;                       \
117         x2 ^= x1;       x3 =~ x3;       x4 ^= x0;       \
118         x0 ^= x2;       x1 &= x4;       x4 |= x3;       \
119         x4 ^= x0;       x0 &= x3;       x1 ^= x3;       \
120         x3 ^= x2;       x0 ^= x1;       x2 &= x4;       \
121         x1 ^= x2;       x2 &= x0;                       \
122         x3 ^= x2;
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 #define S7(x0,x1,x2,x3,x4)                              \
134                                         x1 =~ x1;       \
135         x4  = x1;       x0 =~ x0;       x1 &= x2;       \
136         x1 ^= x3;       x3 |= x4;       x4 ^= x2;       \
137         x2 ^= x3;       x3 ^= x0;       x0 |= x1;       \
138         x2 &= x0;       x0 ^= x4;       x4 ^= x3;       \
139         x3 &= x0;       x4 ^= x1;                       \
140         x2 ^= x4;       x3 ^= x1;       x4 |= x0;       \
141         x4 ^= x1;
142
143 #define SI0(x0,x1,x2,x3,x4)                             \
144                         x4  = x3;       x1 ^= x0;       \
145         x3 |= x1;       x4 ^= x1;       x0 =~ x0;       \
146         x2 ^= x3;       x3 ^= x0;       x0 &= x1;       \
147         x0 ^= x2;       x2 &= x3;       x3 ^= x4;       \
148         x2 ^= x3;       x1 ^= x3;       x3 &= x0;       \
149         x1 ^= x0;       x0 ^= x2;       x4 ^= x3;
150
151 #define SI1(x0,x1,x2,x3,x4)                             \
152         x1 ^= x3;       x4  = x0;                       \
153         x0 ^= x2;       x2 =~ x2;       x4 |= x1;       \
154         x4 ^= x3;       x3 &= x1;       x1 ^= x2;       \
155         x2 &= x4;       x4 ^= x1;       x1 |= x3;       \
156         x3 ^= x0;       x2 ^= x0;       x0 |= x4;       \
157         x2 ^= x4;       x1 ^= x0;                       \
158         x4 ^= x1;
159
160 #define SI2(x0,x1,x2,x3,x4)                             \
161         x2 ^= x1;       x4  = x3;       x3 =~ x3;       \
162         x3 |= x2;       x2 ^= x4;       x4 ^= x0;       \
163         x3 ^= x1;       x1 |= x2;       x2 ^= x0;       \
164         x1 ^= x4;       x4 |= x3;       x2 ^= x3;       \
165         x4 ^= x2;       x2 &= x1;                       \
166         x2 ^= x3;       x3 ^= x4;       x4 ^= x0;
167
168 #define SI3(x0,x1,x2,x3,x4)                             \
169                                         x2 ^= x1;       \
170         x4  = x1;       x1 &= x2;                       \
171         x1 ^= x0;       x0 |= x4;       x4 ^= x3;       \
172         x0 ^= x3;       x3 |= x1;       x1 ^= x2;       \
173         x1 ^= x3;       x0 ^= x2;       x2 ^= x3;       \
174         x3 &= x1;       x1 ^= x0;       x0 &= x2;       \
175         x4 ^= x3;       x3 ^= x0;       x0 ^= x1;
176
177 #define SI4(x0,x1,x2,x3,x4)                             \
178         x2 ^= x3;       x4  = x0;       x0 &= x1;       \
179         x0 ^= x2;       x2 |= x3;       x4 =~ x4;       \
180         x1 ^= x0;       x0 ^= x2;       x2 &= x4;       \
181         x2 ^= x0;       x0 |= x4;                       \
182         x0 ^= x3;       x3 &= x2;                       \
183         x4 ^= x3;       x3 ^= x1;       x1 &= x0;       \
184         x4 ^= x1;       x0 ^= x3;
185
186 #define SI5(x0,x1,x2,x3,x4)                             \
187                         x4  = x1;       x1 |= x2;       \
188         x2 ^= x4;       x1 ^= x3;       x3 &= x4;       \
189         x2 ^= x3;       x3 |= x0;       x0 =~ x0;       \
190         x3 ^= x2;       x2 |= x0;       x4 ^= x1;       \
191         x2 ^= x4;       x4 &= x0;       x0 ^= x1;       \
192         x1 ^= x3;       x0 &= x2;       x2 ^= x3;       \
193         x0 ^= x2;       x2 ^= x4;       x4 ^= x3;
194
195 #define SI6(x0,x1,x2,x3,x4)                             \
196                         x0 ^= x2;                       \
197         x4  = x0;       x0 &= x3;       x2 ^= x3;       \
198         x0 ^= x2;       x3 ^= x1;       x2 |= x4;       \
199         x2 ^= x3;       x3 &= x0;       x0 =~ x0;       \
200         x3 ^= x1;       x1 &= x2;       x4 ^= x0;       \
201         x3 ^= x4;       x4 ^= x2;       x0 ^= x1;       \
202         x2 ^= x0;
203
204 #define SI7(x0,x1,x2,x3,x4)                             \
205         x4  = x3;       x3 &= x0;       x0 ^= x2;       \
206         x2 |= x4;       x4 ^= x1;       x0 =~ x0;       \
207         x1 |= x3;       x4 ^= x0;       x0 &= x2;       \
208         x0 ^= x1;       x1 &= x2;       x3 ^= x2;       \
209         x4 ^= x3;       x2 &= x3;       x3 |= x0;       \
210         x1 ^= x4;       x3 ^= x4;       x4 &= x0;       \
211         x4 ^= x2;
212
213 struct serpent_ctx {
214         u32 expkey[SERPENT_EXPKEY_WORDS];
215 };
216
217
218 static int serpent_setkey(void *ctx, const u8 *key, unsigned int keylen, u32 *flags)
219 {
220         u32 *k = ((struct serpent_ctx *)ctx)->expkey;
221         u8  *k8 = (u8 *)k;
222         u32 r0,r1,r2,r3,r4;
223         int i;
224
225         if ((keylen < SERPENT_MIN_KEY_SIZE)
226                         || (keylen > SERPENT_MAX_KEY_SIZE))
227         {
228                 *flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
229                 return -EINVAL;
230         }
231
232         /* Copy key, add padding */
233
234         for (i = 0; i < keylen; ++i)
235                 k8[i] = key[i];
236         if (i < SERPENT_MAX_KEY_SIZE)
237                 k8[i++] = 1;
238         while (i < SERPENT_MAX_KEY_SIZE)
239                 k8[i++] = 0;
240
241         /* Expand key using polynomial */
242
243         r0 = le32_to_cpu(k[3]);
244         r1 = le32_to_cpu(k[4]);
245         r2 = le32_to_cpu(k[5]);
246         r3 = le32_to_cpu(k[6]);
247         r4 = le32_to_cpu(k[7]);
248
249         keyiter(le32_to_cpu(k[0]),r0,r4,r2,0,0);
250         keyiter(le32_to_cpu(k[1]),r1,r0,r3,1,1);
251         keyiter(le32_to_cpu(k[2]),r2,r1,r4,2,2);
252         keyiter(le32_to_cpu(k[3]),r3,r2,r0,3,3);
253         keyiter(le32_to_cpu(k[4]),r4,r3,r1,4,4);
254         keyiter(le32_to_cpu(k[5]),r0,r4,r2,5,5);
255         keyiter(le32_to_cpu(k[6]),r1,r0,r3,6,6);
256         keyiter(le32_to_cpu(k[7]),r2,r1,r4,7,7);
257
258         keyiter(k[  0],r3,r2,r0,  8,  8); keyiter(k[  1],r4,r3,r1,  9,  9);
259         keyiter(k[  2],r0,r4,r2, 10, 10); keyiter(k[  3],r1,r0,r3, 11, 11);
260         keyiter(k[  4],r2,r1,r4, 12, 12); keyiter(k[  5],r3,r2,r0, 13, 13);
261         keyiter(k[  6],r4,r3,r1, 14, 14); keyiter(k[  7],r0,r4,r2, 15, 15);
262         keyiter(k[  8],r1,r0,r3, 16, 16); keyiter(k[  9],r2,r1,r4, 17, 17);
263         keyiter(k[ 10],r3,r2,r0, 18, 18); keyiter(k[ 11],r4,r3,r1, 19, 19);
264         keyiter(k[ 12],r0,r4,r2, 20, 20); keyiter(k[ 13],r1,r0,r3, 21, 21);
265         keyiter(k[ 14],r2,r1,r4, 22, 22); keyiter(k[ 15],r3,r2,r0, 23, 23);
266         keyiter(k[ 16],r4,r3,r1, 24, 24); keyiter(k[ 17],r0,r4,r2, 25, 25);
267         keyiter(k[ 18],r1,r0,r3, 26, 26); keyiter(k[ 19],r2,r1,r4, 27, 27);
268         keyiter(k[ 20],r3,r2,r0, 28, 28); keyiter(k[ 21],r4,r3,r1, 29, 29);
269         keyiter(k[ 22],r0,r4,r2, 30, 30); keyiter(k[ 23],r1,r0,r3, 31, 31);
270
271         k += 50;
272
273         keyiter(k[-26],r2,r1,r4, 32,-18); keyiter(k[-25],r3,r2,r0, 33,-17);
274         keyiter(k[-24],r4,r3,r1, 34,-16); keyiter(k[-23],r0,r4,r2, 35,-15);
275         keyiter(k[-22],r1,r0,r3, 36,-14); keyiter(k[-21],r2,r1,r4, 37,-13);
276         keyiter(k[-20],r3,r2,r0, 38,-12); keyiter(k[-19],r4,r3,r1, 39,-11);
277         keyiter(k[-18],r0,r4,r2, 40,-10); keyiter(k[-17],r1,r0,r3, 41, -9);
278         keyiter(k[-16],r2,r1,r4, 42, -8); keyiter(k[-15],r3,r2,r0, 43, -7);
279         keyiter(k[-14],r4,r3,r1, 44, -6); keyiter(k[-13],r0,r4,r2, 45, -5);
280         keyiter(k[-12],r1,r0,r3, 46, -4); keyiter(k[-11],r2,r1,r4, 47, -3);
281         keyiter(k[-10],r3,r2,r0, 48, -2); keyiter(k[ -9],r4,r3,r1, 49, -1);
282         keyiter(k[ -8],r0,r4,r2, 50,  0); keyiter(k[ -7],r1,r0,r3, 51,  1);
283         keyiter(k[ -6],r2,r1,r4, 52,  2); keyiter(k[ -5],r3,r2,r0, 53,  3);
284         keyiter(k[ -4],r4,r3,r1, 54,  4); keyiter(k[ -3],r0,r4,r2, 55,  5);
285         keyiter(k[ -2],r1,r0,r3, 56,  6); keyiter(k[ -1],r2,r1,r4, 57,  7);
286         keyiter(k[  0],r3,r2,r0, 58,  8); keyiter(k[  1],r4,r3,r1, 59,  9);
287         keyiter(k[  2],r0,r4,r2, 60, 10); keyiter(k[  3],r1,r0,r3, 61, 11);
288         keyiter(k[  4],r2,r1,r4, 62, 12); keyiter(k[  5],r3,r2,r0, 63, 13);
289         keyiter(k[  6],r4,r3,r1, 64, 14); keyiter(k[  7],r0,r4,r2, 65, 15);
290         keyiter(k[  8],r1,r0,r3, 66, 16); keyiter(k[  9],r2,r1,r4, 67, 17);
291         keyiter(k[ 10],r3,r2,r0, 68, 18); keyiter(k[ 11],r4,r3,r1, 69, 19);
292         keyiter(k[ 12],r0,r4,r2, 70, 20); keyiter(k[ 13],r1,r0,r3, 71, 21);
293         keyiter(k[ 14],r2,r1,r4, 72, 22); keyiter(k[ 15],r3,r2,r0, 73, 23);
294         keyiter(k[ 16],r4,r3,r1, 74, 24); keyiter(k[ 17],r0,r4,r2, 75, 25);
295         keyiter(k[ 18],r1,r0,r3, 76, 26); keyiter(k[ 19],r2,r1,r4, 77, 27);
296         keyiter(k[ 20],r3,r2,r0, 78, 28); keyiter(k[ 21],r4,r3,r1, 79, 29);
297         keyiter(k[ 22],r0,r4,r2, 80, 30); keyiter(k[ 23],r1,r0,r3, 81, 31);
298
299         k += 50;
300
301         keyiter(k[-26],r2,r1,r4, 82,-18); keyiter(k[-25],r3,r2,r0, 83,-17);
302         keyiter(k[-24],r4,r3,r1, 84,-16); keyiter(k[-23],r0,r4,r2, 85,-15);
303         keyiter(k[-22],r1,r0,r3, 86,-14); keyiter(k[-21],r2,r1,r4, 87,-13);
304         keyiter(k[-20],r3,r2,r0, 88,-12); keyiter(k[-19],r4,r3,r1, 89,-11);
305         keyiter(k[-18],r0,r4,r2, 90,-10); keyiter(k[-17],r1,r0,r3, 91, -9);
306         keyiter(k[-16],r2,r1,r4, 92, -8); keyiter(k[-15],r3,r2,r0, 93, -7);
307         keyiter(k[-14],r4,r3,r1, 94, -6); keyiter(k[-13],r0,r4,r2, 95, -5);
308         keyiter(k[-12],r1,r0,r3, 96, -4); keyiter(k[-11],r2,r1,r4, 97, -3);
309         keyiter(k[-10],r3,r2,r0, 98, -2); keyiter(k[ -9],r4,r3,r1, 99, -1);
310         keyiter(k[ -8],r0,r4,r2,100,  0); keyiter(k[ -7],r1,r0,r3,101,  1);
311         keyiter(k[ -6],r2,r1,r4,102,  2); keyiter(k[ -5],r3,r2,r0,103,  3);
312         keyiter(k[ -4],r4,r3,r1,104,  4); keyiter(k[ -3],r0,r4,r2,105,  5);
313         keyiter(k[ -2],r1,r0,r3,106,  6); keyiter(k[ -1],r2,r1,r4,107,  7);
314         keyiter(k[  0],r3,r2,r0,108,  8); keyiter(k[  1],r4,r3,r1,109,  9);
315         keyiter(k[  2],r0,r4,r2,110, 10); keyiter(k[  3],r1,r0,r3,111, 11);
316         keyiter(k[  4],r2,r1,r4,112, 12); keyiter(k[  5],r3,r2,r0,113, 13);
317         keyiter(k[  6],r4,r3,r1,114, 14); keyiter(k[  7],r0,r4,r2,115, 15);
318         keyiter(k[  8],r1,r0,r3,116, 16); keyiter(k[  9],r2,r1,r4,117, 17);
319         keyiter(k[ 10],r3,r2,r0,118, 18); keyiter(k[ 11],r4,r3,r1,119, 19);
320         keyiter(k[ 12],r0,r4,r2,120, 20); keyiter(k[ 13],r1,r0,r3,121, 21);
321         keyiter(k[ 14],r2,r1,r4,122, 22); keyiter(k[ 15],r3,r2,r0,123, 23);
322         keyiter(k[ 16],r4,r3,r1,124, 24); keyiter(k[ 17],r0,r4,r2,125, 25);
323         keyiter(k[ 18],r1,r0,r3,126, 26); keyiter(k[ 19],r2,r1,r4,127, 27);
324         keyiter(k[ 20],r3,r2,r0,128, 28); keyiter(k[ 21],r4,r3,r1,129, 29);
325         keyiter(k[ 22],r0,r4,r2,130, 30); keyiter(k[ 23],r1,r0,r3,131, 31);
326
327         /* Apply S-boxes */
328
329         S3(r3,r4,r0,r1,r2); storekeys(r1,r2,r4,r3, 28); loadkeys(r1,r2,r4,r3, 24);
330         S4(r1,r2,r4,r3,r0); storekeys(r2,r4,r3,r0, 24); loadkeys(r2,r4,r3,r0, 20);
331         S5(r2,r4,r3,r0,r1); storekeys(r1,r2,r4,r0, 20); loadkeys(r1,r2,r4,r0, 16);
332         S6(r1,r2,r4,r0,r3); storekeys(r4,r3,r2,r0, 16); loadkeys(r4,r3,r2,r0, 12);
333         S7(r4,r3,r2,r0,r1); storekeys(r1,r2,r0,r4, 12); loadkeys(r1,r2,r0,r4,  8);
334         S0(r1,r2,r0,r4,r3); storekeys(r0,r2,r4,r1,  8); loadkeys(r0,r2,r4,r1,  4);
335         S1(r0,r2,r4,r1,r3); storekeys(r3,r4,r1,r0,  4); loadkeys(r3,r4,r1,r0,  0);
336         S2(r3,r4,r1,r0,r2); storekeys(r2,r4,r3,r0,  0); loadkeys(r2,r4,r3,r0, -4);
337         S3(r2,r4,r3,r0,r1); storekeys(r0,r1,r4,r2, -4); loadkeys(r0,r1,r4,r2, -8);
338         S4(r0,r1,r4,r2,r3); storekeys(r1,r4,r2,r3, -8); loadkeys(r1,r4,r2,r3,-12);
339         S5(r1,r4,r2,r3,r0); storekeys(r0,r1,r4,r3,-12); loadkeys(r0,r1,r4,r3,-16);
340         S6(r0,r1,r4,r3,r2); storekeys(r4,r2,r1,r3,-16); loadkeys(r4,r2,r1,r3,-20);
341         S7(r4,r2,r1,r3,r0); storekeys(r0,r1,r3,r4,-20); loadkeys(r0,r1,r3,r4,-24);
342         S0(r0,r1,r3,r4,r2); storekeys(r3,r1,r4,r0,-24); loadkeys(r3,r1,r4,r0,-28);
343         k -= 50;
344         S1(r3,r1,r4,r0,r2); storekeys(r2,r4,r0,r3, 22); loadkeys(r2,r4,r0,r3, 18);
345         S2(r2,r4,r0,r3,r1); storekeys(r1,r4,r2,r3, 18); loadkeys(r1,r4,r2,r3, 14);
346         S3(r1,r4,r2,r3,r0); storekeys(r3,r0,r4,r1, 14); loadkeys(r3,r0,r4,r1, 10);
347         S4(r3,r0,r4,r1,r2); storekeys(r0,r4,r1,r2, 10); loadkeys(r0,r4,r1,r2,  6);
348         S5(r0,r4,r1,r2,r3); storekeys(r3,r0,r4,r2,  6); loadkeys(r3,r0,r4,r2,  2);
349         S6(r3,r0,r4,r2,r1); storekeys(r4,r1,r0,r2,  2); loadkeys(r4,r1,r0,r2, -2);
350         S7(r4,r1,r0,r2,r3); storekeys(r3,r0,r2,r4, -2); loadkeys(r3,r0,r2,r4, -6);
351         S0(r3,r0,r2,r4,r1); storekeys(r2,r0,r4,r3, -6); loadkeys(r2,r0,r4,r3,-10);
352         S1(r2,r0,r4,r3,r1); storekeys(r1,r4,r3,r2,-10); loadkeys(r1,r4,r3,r2,-14);
353         S2(r1,r4,r3,r2,r0); storekeys(r0,r4,r1,r2,-14); loadkeys(r0,r4,r1,r2,-18);
354         S3(r0,r4,r1,r2,r3); storekeys(r2,r3,r4,r0,-18); loadkeys(r2,r3,r4,r0,-22);
355         k -= 50;
356         S4(r2,r3,r4,r0,r1); storekeys(r3,r4,r0,r1, 28); loadkeys(r3,r4,r0,r1, 24);
357         S5(r3,r4,r0,r1,r2); storekeys(r2,r3,r4,r1, 24); loadkeys(r2,r3,r4,r1, 20);
358         S6(r2,r3,r4,r1,r0); storekeys(r4,r0,r3,r1, 20); loadkeys(r4,r0,r3,r1, 16);
359         S7(r4,r0,r3,r1,r2); storekeys(r2,r3,r1,r4, 16); loadkeys(r2,r3,r1,r4, 12);
360         S0(r2,r3,r1,r4,r0); storekeys(r1,r3,r4,r2, 12); loadkeys(r1,r3,r4,r2,  8);
361         S1(r1,r3,r4,r2,r0); storekeys(r0,r4,r2,r1,  8); loadkeys(r0,r4,r2,r1,  4);
362         S2(r0,r4,r2,r1,r3); storekeys(r3,r4,r0,r1,  4); loadkeys(r3,r4,r0,r1,  0);
363         S3(r3,r4,r0,r1,r2); storekeys(r1,r2,r4,r3,  0);
364
365         return 0;
366 }
367
368 static void serpent_encrypt(void *ctx, u8 *dst, const u8 *src)
369 {
370         const u32
371                 *k = ((struct serpent_ctx *)ctx)->expkey,
372                 *s = (const u32 *)src;
373         u32     *d = (u32 *)dst,
374                 r0, r1, r2, r3, r4;
375
376 /*
377  * Note: The conversions between u8* and u32* might cause trouble
378  * on architectures with stricter alignment rules than x86
379  */
380
381         r0 = le32_to_cpu(s[0]);
382         r1 = le32_to_cpu(s[1]);
383         r2 = le32_to_cpu(s[2]);
384         r3 = le32_to_cpu(s[3]);
385
386                                  K(r0,r1,r2,r3,0);
387         S0(r0,r1,r2,r3,r4);     LK(r2,r1,r3,r0,r4,1);
388         S1(r2,r1,r3,r0,r4);     LK(r4,r3,r0,r2,r1,2);
389         S2(r4,r3,r0,r2,r1);     LK(r1,r3,r4,r2,r0,3);
390         S3(r1,r3,r4,r2,r0);     LK(r2,r0,r3,r1,r4,4);
391         S4(r2,r0,r3,r1,r4);     LK(r0,r3,r1,r4,r2,5);
392         S5(r0,r3,r1,r4,r2);     LK(r2,r0,r3,r4,r1,6);
393         S6(r2,r0,r3,r4,r1);     LK(r3,r1,r0,r4,r2,7);
394         S7(r3,r1,r0,r4,r2);     LK(r2,r0,r4,r3,r1,8);
395         S0(r2,r0,r4,r3,r1);     LK(r4,r0,r3,r2,r1,9);
396         S1(r4,r0,r3,r2,r1);     LK(r1,r3,r2,r4,r0,10);
397         S2(r1,r3,r2,r4,r0);     LK(r0,r3,r1,r4,r2,11);
398         S3(r0,r3,r1,r4,r2);     LK(r4,r2,r3,r0,r1,12);
399         S4(r4,r2,r3,r0,r1);     LK(r2,r3,r0,r1,r4,13);
400         S5(r2,r3,r0,r1,r4);     LK(r4,r2,r3,r1,r0,14);
401         S6(r4,r2,r3,r1,r0);     LK(r3,r0,r2,r1,r4,15);
402         S7(r3,r0,r2,r1,r4);     LK(r4,r2,r1,r3,r0,16);
403         S0(r4,r2,r1,r3,r0);     LK(r1,r2,r3,r4,r0,17);
404         S1(r1,r2,r3,r4,r0);     LK(r0,r3,r4,r1,r2,18);
405         S2(r0,r3,r4,r1,r2);     LK(r2,r3,r0,r1,r4,19);
406         S3(r2,r3,r0,r1,r4);     LK(r1,r4,r3,r2,r0,20);
407         S4(r1,r4,r3,r2,r0);     LK(r4,r3,r2,r0,r1,21);
408         S5(r4,r3,r2,r0,r1);     LK(r1,r4,r3,r0,r2,22);
409         S6(r1,r4,r3,r0,r2);     LK(r3,r2,r4,r0,r1,23);
410         S7(r3,r2,r4,r0,r1);     LK(r1,r4,r0,r3,r2,24);
411         S0(r1,r4,r0,r3,r2);     LK(r0,r4,r3,r1,r2,25);
412         S1(r0,r4,r3,r1,r2);     LK(r2,r3,r1,r0,r4,26);
413         S2(r2,r3,r1,r0,r4);     LK(r4,r3,r2,r0,r1,27);
414         S3(r4,r3,r2,r0,r1);     LK(r0,r1,r3,r4,r2,28);
415         S4(r0,r1,r3,r4,r2);     LK(r1,r3,r4,r2,r0,29);
416         S5(r1,r3,r4,r2,r0);     LK(r0,r1,r3,r2,r4,30);
417         S6(r0,r1,r3,r2,r4);     LK(r3,r4,r1,r2,r0,31);
418         S7(r3,r4,r1,r2,r0);      K(r0,r1,r2,r3,32);
419
420         d[0] = cpu_to_le32(r0);
421         d[1] = cpu_to_le32(r1);
422         d[2] = cpu_to_le32(r2);
423         d[3] = cpu_to_le32(r3);
424 }
425
426 static void serpent_decrypt(void *ctx, u8 *dst, const u8 *src)
427 {
428         const u32
429                 *k = ((struct serpent_ctx *)ctx)->expkey,
430                 *s = (const u32 *)src;
431         u32     *d = (u32 *)dst,
432                 r0, r1, r2, r3, r4;
433
434         r0 = le32_to_cpu(s[0]);
435         r1 = le32_to_cpu(s[1]);
436         r2 = le32_to_cpu(s[2]);
437         r3 = le32_to_cpu(s[3]);
438
439                                 K(r0,r1,r2,r3,32);
440         SI7(r0,r1,r2,r3,r4);    KL(r1,r3,r0,r4,r2,31);
441         SI6(r1,r3,r0,r4,r2);    KL(r0,r2,r4,r1,r3,30);
442         SI5(r0,r2,r4,r1,r3);    KL(r2,r3,r0,r4,r1,29);
443         SI4(r2,r3,r0,r4,r1);    KL(r2,r0,r1,r4,r3,28);
444         SI3(r2,r0,r1,r4,r3);    KL(r1,r2,r3,r4,r0,27);
445         SI2(r1,r2,r3,r4,r0);    KL(r2,r0,r4,r3,r1,26);
446         SI1(r2,r0,r4,r3,r1);    KL(r1,r0,r4,r3,r2,25);
447         SI0(r1,r0,r4,r3,r2);    KL(r4,r2,r0,r1,r3,24);
448         SI7(r4,r2,r0,r1,r3);    KL(r2,r1,r4,r3,r0,23);
449         SI6(r2,r1,r4,r3,r0);    KL(r4,r0,r3,r2,r1,22);
450         SI5(r4,r0,r3,r2,r1);    KL(r0,r1,r4,r3,r2,21);
451         SI4(r0,r1,r4,r3,r2);    KL(r0,r4,r2,r3,r1,20);
452         SI3(r0,r4,r2,r3,r1);    KL(r2,r0,r1,r3,r4,19);
453         SI2(r2,r0,r1,r3,r4);    KL(r0,r4,r3,r1,r2,18);
454         SI1(r0,r4,r3,r1,r2);    KL(r2,r4,r3,r1,r0,17);
455         SI0(r2,r4,r3,r1,r0);    KL(r3,r0,r4,r2,r1,16);
456         SI7(r3,r0,r4,r2,r1);    KL(r0,r2,r3,r1,r4,15);
457         SI6(r0,r2,r3,r1,r4);    KL(r3,r4,r1,r0,r2,14);
458         SI5(r3,r4,r1,r0,r2);    KL(r4,r2,r3,r1,r0,13);
459         SI4(r4,r2,r3,r1,r0);    KL(r4,r3,r0,r1,r2,12);
460         SI3(r4,r3,r0,r1,r2);    KL(r0,r4,r2,r1,r3,11);
461         SI2(r0,r4,r2,r1,r3);    KL(r4,r3,r1,r2,r0,10);
462         SI1(r4,r3,r1,r2,r0);    KL(r0,r3,r1,r2,r4,9);
463         SI0(r0,r3,r1,r2,r4);    KL(r1,r4,r3,r0,r2,8);
464         SI7(r1,r4,r3,r0,r2);    KL(r4,r0,r1,r2,r3,7);
465         SI6(r4,r0,r1,r2,r3);    KL(r1,r3,r2,r4,r0,6);
466         SI5(r1,r3,r2,r4,r0);    KL(r3,r0,r1,r2,r4,5);
467         SI4(r3,r0,r1,r2,r4);    KL(r3,r1,r4,r2,r0,4);
468         SI3(r3,r1,r4,r2,r0);    KL(r4,r3,r0,r2,r1,3);
469         SI2(r4,r3,r0,r2,r1);    KL(r3,r1,r2,r0,r4,2);
470         SI1(r3,r1,r2,r0,r4);    KL(r4,r1,r2,r0,r3,1);
471         SI0(r4,r1,r2,r0,r3);    K(r2,r3,r1,r4,0);
472
473         d[0] = cpu_to_le32(r2);
474         d[1] = cpu_to_le32(r3);
475         d[2] = cpu_to_le32(r1);
476         d[3] = cpu_to_le32(r4);
477 }
478
479 static struct crypto_alg serpent_alg = {
480         .cra_name               =       "serpent",
481         .cra_flags              =       CRYPTO_ALG_TYPE_CIPHER,
482         .cra_blocksize          =       SERPENT_BLOCK_SIZE,
483         .cra_ctxsize            =       sizeof(struct serpent_ctx),
484         .cra_module             =       THIS_MODULE,
485         .cra_list               =       LIST_HEAD_INIT(serpent_alg.cra_list),
486         .cra_u                  =       { .cipher = {
487         .cia_min_keysize        =       SERPENT_MIN_KEY_SIZE,
488         .cia_max_keysize        =       SERPENT_MAX_KEY_SIZE,
489         .cia_setkey             =       serpent_setkey,
490         .cia_encrypt            =       serpent_encrypt,
491         .cia_decrypt            =       serpent_decrypt } }
492 };
493
494 static int tnepres_setkey(void *ctx, const u8 *key, unsigned int keylen, u32 *flags)
495 {
496         u8 rev_key[SERPENT_MAX_KEY_SIZE];
497         int i;
498
499         if ((keylen < SERPENT_MIN_KEY_SIZE)
500             || (keylen > SERPENT_MAX_KEY_SIZE)) {
501                 *flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
502                 return -EINVAL;
503         } 
504
505         for (i = 0; i < keylen; ++i)
506                 rev_key[keylen - i - 1] = key[i];
507  
508         return serpent_setkey(ctx, rev_key, keylen, flags);
509 }
510
511 static void tnepres_encrypt(void *ctx, u8 *dst, const u8 *src)
512 {
513         const u32 * const s = (const u32 * const)src;
514         u32 * const d = (u32 * const)dst;
515
516         u32 rs[4], rd[4];
517
518         rs[0] = swab32(s[3]);
519         rs[1] = swab32(s[2]);
520         rs[2] = swab32(s[1]);
521         rs[3] = swab32(s[0]);
522
523         serpent_encrypt(ctx, (u8 *)rd, (u8 *)rs);
524
525         d[0] = swab32(rd[3]);
526         d[1] = swab32(rd[2]);
527         d[2] = swab32(rd[1]);
528         d[3] = swab32(rd[0]);
529 }
530
531 static void tnepres_decrypt(void *ctx, u8 *dst, const u8 *src)
532 {
533         const u32 * const s = (const u32 * const)src;
534         u32 * const d = (u32 * const)dst;
535
536         u32 rs[4], rd[4];
537
538         rs[0] = swab32(s[3]);
539         rs[1] = swab32(s[2]);
540         rs[2] = swab32(s[1]);
541         rs[3] = swab32(s[0]);
542
543         serpent_decrypt(ctx, (u8 *)rd, (u8 *)rs);
544
545         d[0] = swab32(rd[3]);
546         d[1] = swab32(rd[2]);
547         d[2] = swab32(rd[1]);
548         d[3] = swab32(rd[0]);
549 }
550
551 static struct crypto_alg tnepres_alg = {
552         .cra_name               =       "tnepres",
553         .cra_flags              =       CRYPTO_ALG_TYPE_CIPHER,
554         .cra_blocksize          =       SERPENT_BLOCK_SIZE,
555         .cra_ctxsize            =       sizeof(struct serpent_ctx),
556         .cra_alignmask          =       3,
557         .cra_module             =       THIS_MODULE,
558         .cra_list               =       LIST_HEAD_INIT(serpent_alg.cra_list),
559         .cra_u                  =       { .cipher = {
560         .cia_min_keysize        =       SERPENT_MIN_KEY_SIZE,
561         .cia_max_keysize        =       SERPENT_MAX_KEY_SIZE,
562         .cia_setkey             =       tnepres_setkey,
563         .cia_encrypt            =       tnepres_encrypt,
564         .cia_decrypt            =       tnepres_decrypt } }
565 };
566
567 static int __init init(void)
568 {
569         int ret = crypto_register_alg(&serpent_alg);
570
571         if (ret)
572                 return ret;
573
574         ret = crypto_register_alg(&tnepres_alg);
575
576         if (ret)
577                 crypto_unregister_alg(&serpent_alg);
578
579         return ret;
580 }
581
582 static void __exit fini(void)
583 {
584         crypto_unregister_alg(&tnepres_alg);
585         crypto_unregister_alg(&serpent_alg);
586 }
587
588 module_init(init);
589 module_exit(fini);
590
591 MODULE_LICENSE("GPL");
592 MODULE_DESCRIPTION("Serpent and tnepres (kerneli compatible serpent reversed) Cipher Algorithm");
593 MODULE_AUTHOR("Dag Arne Osvik <osvik@ii.uib.no>");
594 MODULE_ALIAS("tnepres");