1 /* gf128mul.h - GF(2^128) multiplication funct 1 /* gf128mul.h - GF(2^128) multiplication functions
2 * 2 *
3 * Copyright (c) 2003, Dr Brian Gladman, Worce 3 * Copyright (c) 2003, Dr Brian Gladman, Worcester, UK.
4 * Copyright (c) 2006 Rik Snel <rsnel@cube.dyn 4 * Copyright (c) 2006 Rik Snel <rsnel@cube.dyndns.org>
5 * 5 *
6 * Based on Dr Brian Gladman's (GPL'd) work pu 6 * Based on Dr Brian Gladman's (GPL'd) work published at
7 * http://fp.gladman.plus.com/cryptography_tec 7 * http://fp.gladman.plus.com/cryptography_technology/index.htm
8 * See the original copyright notice below. 8 * See the original copyright notice below.
9 * 9 *
10 * This program is free software; you can redi 10 * This program is free software; you can redistribute it and/or modify it
11 * under the terms of the GNU General Public L 11 * under the terms of the GNU General Public License as published by the Free
12 * Software Foundation; either version 2 of th 12 * Software Foundation; either version 2 of the License, or (at your option)
13 * any later version. 13 * any later version.
14 */ 14 */
15 /* 15 /*
16 --------------------------------------------- 16 ---------------------------------------------------------------------------
17 Copyright (c) 2003, Dr Brian Gladman, Worcest 17 Copyright (c) 2003, Dr Brian Gladman, Worcester, UK. All rights reserved.
18 18
19 LICENSE TERMS 19 LICENSE TERMS
20 20
21 The free distribution and use of this softwar 21 The free distribution and use of this software in both source and binary
22 form is allowed (with or without changes) pro 22 form is allowed (with or without changes) provided that:
23 23
24 1. distributions of this source code includ 24 1. distributions of this source code include the above copyright
25 notice, this list of conditions and the 25 notice, this list of conditions and the following disclaimer;
26 26
27 2. distributions in binary form include the 27 2. distributions in binary form include the above copyright
28 notice, this list of conditions and the 28 notice, this list of conditions and the following disclaimer
29 in the documentation and/or other associ 29 in the documentation and/or other associated materials;
30 30
31 3. the copyright holder's name is not used 31 3. the copyright holder's name is not used to endorse products
32 built using this software without specif 32 built using this software without specific written permission.
33 33
34 ALTERNATIVELY, provided that this notice is r 34 ALTERNATIVELY, provided that this notice is retained in full, this product
35 may be distributed under the terms of the GNU 35 may be distributed under the terms of the GNU General Public License (GPL),
36 in which case the provisions of the GPL apply 36 in which case the provisions of the GPL apply INSTEAD OF those given above.
37 37
38 DISCLAIMER 38 DISCLAIMER
39 39
40 This software is provided 'as is' with no exp 40 This software is provided 'as is' with no explicit or implied warranties
41 in respect of its properties, including, but 41 in respect of its properties, including, but not limited to, correctness
42 and/or fitness for purpose. 42 and/or fitness for purpose.
43 --------------------------------------------- 43 ---------------------------------------------------------------------------
44 Issue Date: 31/01/2006 44 Issue Date: 31/01/2006
45 45
46 An implementation of field multiplication in 46 An implementation of field multiplication in Galois Field GF(2^128)
47 */ 47 */
48 48
49 #ifndef _CRYPTO_GF128MUL_H 49 #ifndef _CRYPTO_GF128MUL_H
50 #define _CRYPTO_GF128MUL_H 50 #define _CRYPTO_GF128MUL_H
51 51
52 #include <asm/byteorder.h> 52 #include <asm/byteorder.h>
53 #include <crypto/b128ops.h> 53 #include <crypto/b128ops.h>
54 #include <linux/slab.h> 54 #include <linux/slab.h>
55 55
56 /* Comment by Rik: 56 /* Comment by Rik:
57 * 57 *
58 * For some background on GF(2^128) see for ex 58 * For some background on GF(2^128) see for example:
59 * http://csrc.nist.gov/groups/ST/toolkit/BCM/ 59 * http://csrc.nist.gov/groups/ST/toolkit/BCM/documents/proposedmodes/gcm/gcm-revised-spec.pdf
60 * 60 *
61 * The elements of GF(2^128) := GF(2)[X]/(X^12 61 * The elements of GF(2^128) := GF(2)[X]/(X^128-X^7-X^2-X^1-1) can
62 * be mapped to computer memory in a variety o 62 * be mapped to computer memory in a variety of ways. Let's examine
63 * three common cases. 63 * three common cases.
64 * 64 *
65 * Take a look at the 16 binary octets below i 65 * Take a look at the 16 binary octets below in memory order. The msb's
66 * are left and the lsb's are right. char b[16 66 * are left and the lsb's are right. char b[16] is an array and b[0] is
67 * the first octet. 67 * the first octet.
68 * 68 *
69 * 10000000 00000000 00000000 00000000 .... 00 69 * 10000000 00000000 00000000 00000000 .... 00000000 00000000 00000000
70 * b[0] b[1] b[2] b[3] 70 * b[0] b[1] b[2] b[3] b[13] b[14] b[15]
71 * 71 *
72 * Every bit is a coefficient of some power of 72 * Every bit is a coefficient of some power of X. We can store the bits
73 * in every byte in little-endian order and th 73 * in every byte in little-endian order and the bytes themselves also in
74 * little endian order. I will call this lle ( 74 * little endian order. I will call this lle (little-little-endian).
75 * The above buffer represents the polynomial 75 * The above buffer represents the polynomial 1, and X^7+X^2+X^1+1 looks
76 * like 11100001 00000000 .... 00000000 = { 0x 76 * like 11100001 00000000 .... 00000000 = { 0xE1, 0x00, }.
77 * This format was originally implemented in g 77 * This format was originally implemented in gf128mul and is used
78 * in GCM (Galois/Counter mode) and in ABL (Ar 78 * in GCM (Galois/Counter mode) and in ABL (Arbitrary Block Length).
79 * 79 *
80 * Another convention says: store the bits in 80 * Another convention says: store the bits in bigendian order and the
81 * bytes also. This is bbe (big-big-endian). N 81 * bytes also. This is bbe (big-big-endian). Now the buffer above
82 * represents X^127. X^7+X^2+X^1+1 looks like 82 * represents X^127. X^7+X^2+X^1+1 looks like 00000000 .... 10000111,
83 * b[15] = 0x87 and the rest is 0. LRW uses th 83 * b[15] = 0x87 and the rest is 0. LRW uses this convention and bbe
84 * is partly implemented. 84 * is partly implemented.
85 * 85 *
86 * Both of the above formats are easy to imple 86 * Both of the above formats are easy to implement on big-endian
87 * machines. 87 * machines.
88 * 88 *
89 * XTS and EME (the latter of which is patent 89 * XTS and EME (the latter of which is patent encumbered) use the ble
90 * format (bits are stored in big endian order 90 * format (bits are stored in big endian order and the bytes in little
91 * endian). The above buffer represents X^7 in 91 * endian). The above buffer represents X^7 in this case and the
92 * primitive polynomial is b[0] = 0x87. 92 * primitive polynomial is b[0] = 0x87.
93 * 93 *
94 * The common machine word-size is smaller tha 94 * The common machine word-size is smaller than 128 bits, so to make
95 * an efficient implementation we must split i 95 * an efficient implementation we must split into machine word sizes.
96 * This implementation uses 64-bit words for t 96 * This implementation uses 64-bit words for the moment. Machine
97 * endianness comes into play. The lle format 97 * endianness comes into play. The lle format in relation to machine
98 * endianness is discussed below by the origin 98 * endianness is discussed below by the original author of gf128mul Dr
99 * Brian Gladman. 99 * Brian Gladman.
100 * 100 *
101 * Let's look at the bbe and ble format on a l 101 * Let's look at the bbe and ble format on a little endian machine.
102 * 102 *
103 * bbe on a little endian machine u32 x[4]: 103 * bbe on a little endian machine u32 x[4]:
104 * 104 *
105 * MS x[0] LS MS 105 * MS x[0] LS MS x[1] LS
106 * ms ls ms ls ms ls ms ls ms ls m 106 * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
107 * 103..96 111.104 119.112 127.120 71...64 7 107 * 103..96 111.104 119.112 127.120 71...64 79...72 87...80 95...88
108 * 108 *
109 * MS x[2] LS MS 109 * MS x[2] LS MS x[3] LS
110 * ms ls ms ls ms ls ms ls ms ls m 110 * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
111 * 39...32 47...40 55...48 63...56 07...00 1 111 * 39...32 47...40 55...48 63...56 07...00 15...08 23...16 31...24
112 * 112 *
113 * ble on a little endian machine 113 * ble on a little endian machine
114 * 114 *
115 * MS x[0] LS MS 115 * MS x[0] LS MS x[1] LS
116 * ms ls ms ls ms ls ms ls ms ls m 116 * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
117 * 31...24 23...16 15...08 07...00 63...56 5 117 * 31...24 23...16 15...08 07...00 63...56 55...48 47...40 39...32
118 * 118 *
119 * MS x[2] LS MS 119 * MS x[2] LS MS x[3] LS
120 * ms ls ms ls ms ls ms ls ms ls m 120 * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
121 * 95...88 87...80 79...72 71...64 127.120 1 121 * 95...88 87...80 79...72 71...64 127.120 199.112 111.104 103..96
122 * 122 *
123 * Multiplications in GF(2^128) are mostly bit 123 * Multiplications in GF(2^128) are mostly bit-shifts, so you see why
124 * ble (and lbe also) are easier to implement 124 * ble (and lbe also) are easier to implement on a little-endian
125 * machine than on a big-endian machine. The c 125 * machine than on a big-endian machine. The converse holds for bbe
126 * and lle. 126 * and lle.
127 * 127 *
128 * Note: to have good alignment, it seems to m 128 * Note: to have good alignment, it seems to me that it is sufficient
129 * to keep elements of GF(2^128) in type u64[2 129 * to keep elements of GF(2^128) in type u64[2]. On 32-bit wordsize
130 * machines this will automatically aligned to 130 * machines this will automatically aligned to wordsize and on a 64-bit
131 * machine also. 131 * machine also.
132 */ 132 */
133 /* Multiply a GF(2^128) field element by 133 /* Multiply a GF(2^128) field element by x. Field elements are
134 held in arrays of bytes in which field bit 134 held in arrays of bytes in which field bits 8n..8n + 7 are held in
135 byte[n], with lower indexed bits placed in 135 byte[n], with lower indexed bits placed in the more numerically
136 significant bit positions within bytes. 136 significant bit positions within bytes.
137 137
138 On little endian machines the bit indexes 138 On little endian machines the bit indexes translate into the bit
139 positions within four 32-bit words in the 139 positions within four 32-bit words in the following way
140 140
141 MS x[0] LS MS 141 MS x[0] LS MS x[1] LS
142 ms ls ms ls ms ls ms ls ms ls m 142 ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
143 24...31 16...23 08...15 00...07 56...63 4 143 24...31 16...23 08...15 00...07 56...63 48...55 40...47 32...39
144 144
145 MS x[2] LS MS 145 MS x[2] LS MS x[3] LS
146 ms ls ms ls ms ls ms ls ms ls m 146 ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
147 88...95 80...87 72...79 64...71 120.127 1 147 88...95 80...87 72...79 64...71 120.127 112.119 104.111 96..103
148 148
149 On big endian machines the bit indexes tra 149 On big endian machines the bit indexes translate into the bit
150 positions within four 32-bit words in the 150 positions within four 32-bit words in the following way
151 151
152 MS x[0] LS MS 152 MS x[0] LS MS x[1] LS
153 ms ls ms ls ms ls ms ls ms ls m 153 ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
154 00...07 08...15 16...23 24...31 32...39 4 154 00...07 08...15 16...23 24...31 32...39 40...47 48...55 56...63
155 155
156 MS x[2] LS MS 156 MS x[2] LS MS x[3] LS
157 ms ls ms ls ms ls ms ls ms ls m 157 ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls
158 64...71 72...79 80...87 88...95 96..103 1 158 64...71 72...79 80...87 88...95 96..103 104.111 112.119 120.127
159 */ 159 */
160 160
161 /* A slow generic version of gf_mul, impl 161 /* A slow generic version of gf_mul, implemented for lle and bbe
162 * It multiplies a and b and puts the res 162 * It multiplies a and b and puts the result in a */
163 void gf128mul_lle(be128 *a, const be128 *b); 163 void gf128mul_lle(be128 *a, const be128 *b);
164 164
165 void gf128mul_bbe(be128 *a, const be128 *b); 165 void gf128mul_bbe(be128 *a, const be128 *b);
166 166
167 /* 167 /*
168 * The following functions multiply a field el 168 * The following functions multiply a field element by x in
169 * the polynomial field representation. They 169 * the polynomial field representation. They use 64-bit word operations
170 * to gain speed but compensate for machine en 170 * to gain speed but compensate for machine endianness and hence work
171 * correctly on both styles of machine. 171 * correctly on both styles of machine.
172 * 172 *
173 * They are defined here for performance. 173 * They are defined here for performance.
174 */ 174 */
175 175
176 static inline u64 gf128mul_mask_from_bit(u64 x 176 static inline u64 gf128mul_mask_from_bit(u64 x, int which)
177 { 177 {
178 /* a constant-time version of 'x & ((u 178 /* a constant-time version of 'x & ((u64)1 << which) ? (u64)-1 : 0' */
179 return ((s64)(x << (63 - which)) >> 63 179 return ((s64)(x << (63 - which)) >> 63);
180 } 180 }
181 181
182 static inline void gf128mul_x_lle(be128 *r, co 182 static inline void gf128mul_x_lle(be128 *r, const be128 *x)
183 { 183 {
184 u64 a = be64_to_cpu(x->a); 184 u64 a = be64_to_cpu(x->a);
185 u64 b = be64_to_cpu(x->b); 185 u64 b = be64_to_cpu(x->b);
186 186
187 /* equivalent to gf128mul_table_le[(b 187 /* equivalent to gf128mul_table_le[(b << 7) & 0xff] << 48
188 * (see crypto/gf128mul.c): */ 188 * (see crypto/gf128mul.c): */
189 u64 _tt = gf128mul_mask_from_bit(b, 0) 189 u64 _tt = gf128mul_mask_from_bit(b, 0) & ((u64)0xe1 << 56);
190 190
191 r->b = cpu_to_be64((b >> 1) | (a << 63 191 r->b = cpu_to_be64((b >> 1) | (a << 63));
192 r->a = cpu_to_be64((a >> 1) ^ _tt); 192 r->a = cpu_to_be64((a >> 1) ^ _tt);
193 } 193 }
194 194
195 static inline void gf128mul_x_bbe(be128 *r, co 195 static inline void gf128mul_x_bbe(be128 *r, const be128 *x)
196 { 196 {
197 u64 a = be64_to_cpu(x->a); 197 u64 a = be64_to_cpu(x->a);
198 u64 b = be64_to_cpu(x->b); 198 u64 b = be64_to_cpu(x->b);
199 199
200 /* equivalent to gf128mul_table_be[a > 200 /* equivalent to gf128mul_table_be[a >> 63] (see crypto/gf128mul.c): */
201 u64 _tt = gf128mul_mask_from_bit(a, 63 201 u64 _tt = gf128mul_mask_from_bit(a, 63) & 0x87;
202 202
203 r->a = cpu_to_be64((a << 1) | (b >> 63 203 r->a = cpu_to_be64((a << 1) | (b >> 63));
204 r->b = cpu_to_be64((b << 1) ^ _tt); 204 r->b = cpu_to_be64((b << 1) ^ _tt);
205 } 205 }
206 206
207 /* needed by XTS */ 207 /* needed by XTS */
208 static inline void gf128mul_x_ble(le128 *r, co 208 static inline void gf128mul_x_ble(le128 *r, const le128 *x)
209 { 209 {
210 u64 a = le64_to_cpu(x->a); 210 u64 a = le64_to_cpu(x->a);
211 u64 b = le64_to_cpu(x->b); 211 u64 b = le64_to_cpu(x->b);
212 212
213 /* equivalent to gf128mul_table_be[b > 213 /* equivalent to gf128mul_table_be[b >> 63] (see crypto/gf128mul.c): */
214 u64 _tt = gf128mul_mask_from_bit(a, 63 214 u64 _tt = gf128mul_mask_from_bit(a, 63) & 0x87;
215 215
216 r->a = cpu_to_le64((a << 1) | (b >> 63 216 r->a = cpu_to_le64((a << 1) | (b >> 63));
217 r->b = cpu_to_le64((b << 1) ^ _tt); 217 r->b = cpu_to_le64((b << 1) ^ _tt);
218 } 218 }
219 219
220 /* 4k table optimization */ 220 /* 4k table optimization */
221 221
222 struct gf128mul_4k { 222 struct gf128mul_4k {
223 be128 t[256]; 223 be128 t[256];
224 }; 224 };
225 225
226 struct gf128mul_4k *gf128mul_init_4k_lle(const 226 struct gf128mul_4k *gf128mul_init_4k_lle(const be128 *g);
227 struct gf128mul_4k *gf128mul_init_4k_bbe(const 227 struct gf128mul_4k *gf128mul_init_4k_bbe(const be128 *g);
228 void gf128mul_4k_lle(be128 *a, const struct gf 228 void gf128mul_4k_lle(be128 *a, const struct gf128mul_4k *t);
229 void gf128mul_4k_bbe(be128 *a, const struct gf 229 void gf128mul_4k_bbe(be128 *a, const struct gf128mul_4k *t);
230 void gf128mul_x8_ble(le128 *r, const le128 *x) 230 void gf128mul_x8_ble(le128 *r, const le128 *x);
231 static inline void gf128mul_free_4k(struct gf1 231 static inline void gf128mul_free_4k(struct gf128mul_4k *t)
232 { 232 {
233 kfree_sensitive(t); !! 233 kzfree(t);
234 } 234 }
235 235
236 236
237 /* 64k table optimization, implemented for bbe 237 /* 64k table optimization, implemented for bbe */
238 238
239 struct gf128mul_64k { 239 struct gf128mul_64k {
240 struct gf128mul_4k *t[16]; 240 struct gf128mul_4k *t[16];
241 }; 241 };
242 242
243 /* First initialize with the constant factor w 243 /* First initialize with the constant factor with which you
244 * want to multiply and then call gf128mul_64k 244 * want to multiply and then call gf128mul_64k_bbe with the other
245 * factor in the first argument, and the table 245 * factor in the first argument, and the table in the second.
246 * Afterwards, the result is stored in *a. 246 * Afterwards, the result is stored in *a.
247 */ 247 */
248 struct gf128mul_64k *gf128mul_init_64k_bbe(con 248 struct gf128mul_64k *gf128mul_init_64k_bbe(const be128 *g);
249 void gf128mul_free_64k(struct gf128mul_64k *t) 249 void gf128mul_free_64k(struct gf128mul_64k *t);
250 void gf128mul_64k_bbe(be128 *a, const struct g 250 void gf128mul_64k_bbe(be128 *a, const struct gf128mul_64k *t);
251 251
252 #endif /* _CRYPTO_GF128MUL_H */ 252 #endif /* _CRYPTO_GF128MUL_H */
253 253
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.