1 /* camellia-absorb.c
2
3 Final key setup processing for the camellia block cipher.
4
5 Copyright (C) 2006,2007 NTT
6 (Nippon Telegraph and Telephone Corporation).
7
8 Copyright (C) 2010 Niels Möller
9
10 This file is part of GNU Nettle.
11
12 GNU Nettle is free software: you can redistribute it and/or
13 modify it under the terms of either:
14
15 * the GNU Lesser General Public License as published by the Free
16 Software Foundation; either version 3 of the License, or (at your
17 option) any later version.
18
19 or
20
21 * the GNU General Public License as published by the Free
22 Software Foundation; either version 2 of the License, or (at your
23 option) any later version.
24
25 or both in parallel, as here.
26
27 GNU Nettle is distributed in the hope that it will be useful,
28 but WITHOUT ANY WARRANTY; without even the implied warranty of
29 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
30 General Public License for more details.
31
32 You should have received copies of the GNU General Public License and
33 the GNU Lesser General Public License along with this program. If
34 not, see http://www.gnu.org/licenses/.
35 */
36
37 /*
38 * Algorithm Specification
39 * http://info.isl.ntt.co.jp/crypt/eng/camellia/specifications.html
40 */
41
42 /* Based on camellia.c ver 1.2.0, see
43 http://info.isl.ntt.co.jp/crypt/eng/camellia/dl/camellia-LGPL-1.2.0.tar.gz.
44 */
45
46 #if HAVE_CONFIG_H
47 # include "config.h"
48 #endif
49
50 /* For CHAR_BIT, needed by HAVE_NATIVE_64_BIT */
51 #include <limits.h>
52
53 #include "camellia-internal.h"
54
55 #include "macros.h"
56
57 void
_nettle_camellia_absorb(unsigned nkeys,uint64_t * dst,uint64_t * subkey)58 _nettle_camellia_absorb(unsigned nkeys, uint64_t *dst, uint64_t *subkey)
59 {
60 uint64_t kw2, kw4;
61 uint32_t dw, tl, tr;
62 unsigned i;
63
64 /* At this point, the subkey array contains the subkeys as described
65 in the spec, 26 for short keys and 34 for large keys. */
66
67 /* absorb kw2 to other subkeys */
68 kw2 = subkey[1];
69
70 subkey[3] ^= kw2;
71 subkey[5] ^= kw2;
72 subkey[7] ^= kw2;
73 for (i = 8; i < nkeys; i += 8)
74 {
75 /* FIXME: gcc for x86_32 is smart enough to fetch the 32 low bits
76 and xor the result into the 32 high bits, but it still generates
77 worse code than for explicit 32-bit operations. */
78 kw2 ^= (kw2 & ~subkey[i+1]) << 32;
79 dw = (kw2 & subkey[i+1]) >> 32; kw2 ^= ROTL32(1, dw);
80
81 subkey[i+3] ^= kw2;
82 subkey[i+5] ^= kw2;
83 subkey[i+7] ^= kw2;
84 }
85 subkey[i] ^= kw2;
86
87 /* absorb kw4 to other subkeys */
88 kw4 = subkey[nkeys + 1];
89
90 for (i = nkeys - 8; i > 0; i -= 8)
91 {
92 subkey[i+6] ^= kw4;
93 subkey[i+4] ^= kw4;
94 subkey[i+2] ^= kw4;
95 kw4 ^= (kw4 & ~subkey[i]) << 32;
96 dw = (kw4 & subkey[i]) >> 32; kw4 ^= ROTL32(1, dw);
97 }
98
99 subkey[6] ^= kw4;
100 subkey[4] ^= kw4;
101 subkey[2] ^= kw4;
102 subkey[0] ^= kw4;
103
104 /* key XOR is end of F-function */
105 dst[0] = subkey[0] ^ subkey[2];
106 dst[1] = subkey[3];
107
108 dst[2] = subkey[2] ^ subkey[4];
109 dst[3] = subkey[3] ^ subkey[5];
110 dst[4] = subkey[4] ^ subkey[6];
111 dst[5] = subkey[5] ^ subkey[7];
112
113 for (i = 8; i < nkeys; i += 8)
114 {
115 tl = (subkey[i+2] >> 32) ^ (subkey[i+2] & ~subkey[i]);
116 dw = tl & (subkey[i] >> 32);
117 tr = subkey[i+2] ^ ROTL32(1, dw);
118 dst[i-2] = subkey[i-2] ^ ( ((uint64_t) tl << 32) | tr);
119
120 dst[i-1] = subkey[i];
121 dst[i] = subkey[i+1];
122
123 tl = (subkey[i-1] >> 32) ^ (subkey[i-1] & ~subkey[i+1]);
124 dw = tl & (subkey[i+1] >> 32);
125 tr = subkey[i-1] ^ ROTL32(1, dw);
126 dst[i+1] = subkey[i+3] ^ ( ((uint64_t) tl << 32) | tr);
127
128 dst[i+2] = subkey[i+2] ^ subkey[i+4];
129 dst[i+3] = subkey[i+3] ^ subkey[i+5];
130 dst[i+4] = subkey[i+4] ^ subkey[i+6];
131 dst[i+5] = subkey[i+5] ^ subkey[i+7];
132 }
133 dst[i-2] = subkey[i-2];
134 dst[i-1] = subkey[i] ^ subkey[i-1];
135
136 #if !HAVE_NATIVE_64_BIT
137 for (i = 0; i < nkeys; i += 8)
138 {
139 /* apply the inverse of the last half of F-function */
140 CAMELLIA_F_HALF_INV(dst[i+1]);
141 CAMELLIA_F_HALF_INV(dst[i+2]);
142 CAMELLIA_F_HALF_INV(dst[i+3]);
143 CAMELLIA_F_HALF_INV(dst[i+4]);
144 CAMELLIA_F_HALF_INV(dst[i+5]);
145 CAMELLIA_F_HALF_INV(dst[i+6]);
146 }
147 #endif
148
149 }
150