Root/
Source at commit cdde9cf73945d547acd3e96f9508c79e84ad0bf1 created 12 years 9 months ago. By Maarten ter Huurne, MMC: JZ4740: Added support for CPU frequency changing | |
---|---|
1 | /* |
2 | * SHA1 routine optimized to do word accesses rather than byte accesses, |
3 | * and to avoid unnecessary copies into the context array. |
4 | * |
5 | * This was based on the git SHA1 implementation. |
6 | */ |
7 | |
8 | #include <linux/kernel.h> |
9 | #include <linux/export.h> |
10 | #include <linux/bitops.h> |
11 | #include <linux/cryptohash.h> |
12 | #include <asm/unaligned.h> |
13 | |
14 | /* |
15 | * If you have 32 registers or more, the compiler can (and should) |
16 | * try to change the array[] accesses into registers. However, on |
17 | * machines with less than ~25 registers, that won't really work, |
18 | * and at least gcc will make an unholy mess of it. |
19 | * |
20 | * So to avoid that mess which just slows things down, we force |
21 | * the stores to memory to actually happen (we might be better off |
22 | * with a 'W(t)=(val);asm("":"+m" (W(t))' there instead, as |
23 | * suggested by Artur Skawina - that will also make gcc unable to |
24 | * try to do the silly "optimize away loads" part because it won't |
25 | * see what the value will be). |
26 | * |
27 | * Ben Herrenschmidt reports that on PPC, the C version comes close |
28 | * to the optimized asm with this (ie on PPC you don't want that |
29 | * 'volatile', since there are lots of registers). |
30 | * |
31 | * On ARM we get the best code generation by forcing a full memory barrier |
32 | * between each SHA_ROUND, otherwise gcc happily get wild with spilling and |
33 | * the stack frame size simply explode and performance goes down the drain. |
34 | */ |
35 | |
36 | #ifdef CONFIG_X86 |
37 | #define setW(x, val) (*(volatile __u32 *)&W(x) = (val)) |
38 | #elif defined(CONFIG_ARM) |
39 | #define setW(x, val) do { W(x) = (val); __asm__("":::"memory"); } while (0) |
40 | #else |
41 | #define setW(x, val) (W(x) = (val)) |
42 | #endif |
43 | |
44 | /* This "rolls" over the 512-bit array */ |
45 | #define W(x) (array[(x)&15]) |
46 | |
47 | /* |
48 | * Where do we get the source from? The first 16 iterations get it from |
49 | * the input data, the next mix it from the 512-bit array. |
50 | */ |
51 | #define SHA_SRC(t) get_unaligned_be32((__u32 *)data + t) |
52 | #define SHA_MIX(t) rol32(W(t+13) ^ W(t+8) ^ W(t+2) ^ W(t), 1) |
53 | |
54 | #define SHA_ROUND(t, input, fn, constant, A, B, C, D, E) do { \ |
55 | __u32 TEMP = input(t); setW(t, TEMP); \ |
56 | E += TEMP + rol32(A,5) + (fn) + (constant); \ |
57 | B = ror32(B, 2); } while (0) |
58 | |
59 | #define T_0_15(t, A, B, C, D, E) SHA_ROUND(t, SHA_SRC, (((C^D)&B)^D) , 0x5a827999, A, B, C, D, E ) |
60 | #define T_16_19(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, (((C^D)&B)^D) , 0x5a827999, A, B, C, D, E ) |
61 | #define T_20_39(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, (B^C^D) , 0x6ed9eba1, A, B, C, D, E ) |
62 | #define T_40_59(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, ((B&C)+(D&(B^C))) , 0x8f1bbcdc, A, B, C, D, E ) |
63 | #define T_60_79(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, (B^C^D) , 0xca62c1d6, A, B, C, D, E ) |
64 | |
65 | /** |
66 | * sha_transform - single block SHA1 transform |
67 | * |
68 | * @digest: 160 bit digest to update |
69 | * @data: 512 bits of data to hash |
70 | * @array: 16 words of workspace (see note) |
71 | * |
72 | * This function generates a SHA1 digest for a single 512-bit block. |
73 | * Be warned, it does not handle padding and message digest, do not |
74 | * confuse it with the full FIPS 180-1 digest algorithm for variable |
75 | * length messages. |
76 | * |
77 | * Note: If the hash is security sensitive, the caller should be sure |
78 | * to clear the workspace. This is left to the caller to avoid |
79 | * unnecessary clears between chained hashing operations. |
80 | */ |
81 | void sha_transform(__u32 *digest, const char *data, __u32 *array) |
82 | { |
83 | __u32 A, B, C, D, E; |
84 | |
85 | A = digest[0]; |
86 | B = digest[1]; |
87 | C = digest[2]; |
88 | D = digest[3]; |
89 | E = digest[4]; |
90 | |
91 | /* Round 1 - iterations 0-16 take their input from 'data' */ |
92 | T_0_15( 0, A, B, C, D, E); |
93 | T_0_15( 1, E, A, B, C, D); |
94 | T_0_15( 2, D, E, A, B, C); |
95 | T_0_15( 3, C, D, E, A, B); |
96 | T_0_15( 4, B, C, D, E, A); |
97 | T_0_15( 5, A, B, C, D, E); |
98 | T_0_15( 6, E, A, B, C, D); |
99 | T_0_15( 7, D, E, A, B, C); |
100 | T_0_15( 8, C, D, E, A, B); |
101 | T_0_15( 9, B, C, D, E, A); |
102 | T_0_15(10, A, B, C, D, E); |
103 | T_0_15(11, E, A, B, C, D); |
104 | T_0_15(12, D, E, A, B, C); |
105 | T_0_15(13, C, D, E, A, B); |
106 | T_0_15(14, B, C, D, E, A); |
107 | T_0_15(15, A, B, C, D, E); |
108 | |
109 | /* Round 1 - tail. Input from 512-bit mixing array */ |
110 | T_16_19(16, E, A, B, C, D); |
111 | T_16_19(17, D, E, A, B, C); |
112 | T_16_19(18, C, D, E, A, B); |
113 | T_16_19(19, B, C, D, E, A); |
114 | |
115 | /* Round 2 */ |
116 | T_20_39(20, A, B, C, D, E); |
117 | T_20_39(21, E, A, B, C, D); |
118 | T_20_39(22, D, E, A, B, C); |
119 | T_20_39(23, C, D, E, A, B); |
120 | T_20_39(24, B, C, D, E, A); |
121 | T_20_39(25, A, B, C, D, E); |
122 | T_20_39(26, E, A, B, C, D); |
123 | T_20_39(27, D, E, A, B, C); |
124 | T_20_39(28, C, D, E, A, B); |
125 | T_20_39(29, B, C, D, E, A); |
126 | T_20_39(30, A, B, C, D, E); |
127 | T_20_39(31, E, A, B, C, D); |
128 | T_20_39(32, D, E, A, B, C); |
129 | T_20_39(33, C, D, E, A, B); |
130 | T_20_39(34, B, C, D, E, A); |
131 | T_20_39(35, A, B, C, D, E); |
132 | T_20_39(36, E, A, B, C, D); |
133 | T_20_39(37, D, E, A, B, C); |
134 | T_20_39(38, C, D, E, A, B); |
135 | T_20_39(39, B, C, D, E, A); |
136 | |
137 | /* Round 3 */ |
138 | T_40_59(40, A, B, C, D, E); |
139 | T_40_59(41, E, A, B, C, D); |
140 | T_40_59(42, D, E, A, B, C); |
141 | T_40_59(43, C, D, E, A, B); |
142 | T_40_59(44, B, C, D, E, A); |
143 | T_40_59(45, A, B, C, D, E); |
144 | T_40_59(46, E, A, B, C, D); |
145 | T_40_59(47, D, E, A, B, C); |
146 | T_40_59(48, C, D, E, A, B); |
147 | T_40_59(49, B, C, D, E, A); |
148 | T_40_59(50, A, B, C, D, E); |
149 | T_40_59(51, E, A, B, C, D); |
150 | T_40_59(52, D, E, A, B, C); |
151 | T_40_59(53, C, D, E, A, B); |
152 | T_40_59(54, B, C, D, E, A); |
153 | T_40_59(55, A, B, C, D, E); |
154 | T_40_59(56, E, A, B, C, D); |
155 | T_40_59(57, D, E, A, B, C); |
156 | T_40_59(58, C, D, E, A, B); |
157 | T_40_59(59, B, C, D, E, A); |
158 | |
159 | /* Round 4 */ |
160 | T_60_79(60, A, B, C, D, E); |
161 | T_60_79(61, E, A, B, C, D); |
162 | T_60_79(62, D, E, A, B, C); |
163 | T_60_79(63, C, D, E, A, B); |
164 | T_60_79(64, B, C, D, E, A); |
165 | T_60_79(65, A, B, C, D, E); |
166 | T_60_79(66, E, A, B, C, D); |
167 | T_60_79(67, D, E, A, B, C); |
168 | T_60_79(68, C, D, E, A, B); |
169 | T_60_79(69, B, C, D, E, A); |
170 | T_60_79(70, A, B, C, D, E); |
171 | T_60_79(71, E, A, B, C, D); |
172 | T_60_79(72, D, E, A, B, C); |
173 | T_60_79(73, C, D, E, A, B); |
174 | T_60_79(74, B, C, D, E, A); |
175 | T_60_79(75, A, B, C, D, E); |
176 | T_60_79(76, E, A, B, C, D); |
177 | T_60_79(77, D, E, A, B, C); |
178 | T_60_79(78, C, D, E, A, B); |
179 | T_60_79(79, B, C, D, E, A); |
180 | |
181 | digest[0] += A; |
182 | digest[1] += B; |
183 | digest[2] += C; |
184 | digest[3] += D; |
185 | digest[4] += E; |
186 | } |
187 | EXPORT_SYMBOL(sha_transform); |
188 | |
189 | /** |
190 | * sha_init - initialize the vectors for a SHA1 digest |
191 | * @buf: vector to initialize |
192 | */ |
193 | void sha_init(__u32 *buf) |
194 | { |
195 | buf[0] = 0x67452301; |
196 | buf[1] = 0xefcdab89; |
197 | buf[2] = 0x98badcfe; |
198 | buf[3] = 0x10325476; |
199 | buf[4] = 0xc3d2e1f0; |
200 | } |
201 |
Branches:
ben-wpan
ben-wpan-stefan
javiroman/ks7010
jz-2.6.34
jz-2.6.34-rc5
jz-2.6.34-rc6
jz-2.6.34-rc7
jz-2.6.35
jz-2.6.36
jz-2.6.37
jz-2.6.38
jz-2.6.39
jz-3.0
jz-3.1
jz-3.11
jz-3.12
jz-3.13
jz-3.15
jz-3.16
jz-3.18-dt
jz-3.2
jz-3.3
jz-3.4
jz-3.5
jz-3.6
jz-3.6-rc2-pwm
jz-3.9
jz-3.9-clk
jz-3.9-rc8
jz47xx
jz47xx-2.6.38
master
Tags:
od-2011-09-04
od-2011-09-18
v2.6.34-rc5
v2.6.34-rc6
v2.6.34-rc7
v3.9