Add new optimized C 'block-sha1' routines
[git/git.git] / block-sha1 / sha1.c
1 /*
2 * Based on the Mozilla SHA1 (see mozilla-sha1/sha1.c),
3 * optimized to do word accesses rather than byte accesses,
4 * and to avoid unnecessary copies into the context array.
5 */
6
7 #include <string.h>
8 #include <arpa/inet.h>
9
10 #include "sha1.h"
11
12 /* Hash one 64-byte block of data */
13 static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data);
14
15 void blk_SHA1_Init(blk_SHA_CTX *ctx)
16 {
17 ctx->lenW = 0;
18 ctx->size = 0;
19
20 /* Initialize H with the magic constants (see FIPS180 for constants)
21 */
22 ctx->H[0] = 0x67452301;
23 ctx->H[1] = 0xefcdab89;
24 ctx->H[2] = 0x98badcfe;
25 ctx->H[3] = 0x10325476;
26 ctx->H[4] = 0xc3d2e1f0;
27 }
28
29
30 void blk_SHA1_Update(blk_SHA_CTX *ctx, const void *data, unsigned long len)
31 {
32 int lenW = ctx->lenW;
33
34 ctx->size += len;
35
36 /* Read the data into W and process blocks as they get full
37 */
38 if (lenW) {
39 int left = 64 - lenW;
40 if (len < left)
41 left = len;
42 memcpy(lenW + (char *)ctx->W, data, left);
43 lenW = (lenW + left) & 63;
44 len -= left;
45 data += left;
46 ctx->lenW = lenW;
47 if (lenW)
48 return;
49 blk_SHA1Block(ctx, ctx->W);
50 }
51 while (len >= 64) {
52 blk_SHA1Block(ctx, data);
53 data += 64;
54 len -= 64;
55 }
56 if (len) {
57 memcpy(ctx->W, data, len);
58 ctx->lenW = len;
59 }
60 }
61
62
63 void blk_SHA1_Final(unsigned char hashout[20], blk_SHA_CTX *ctx)
64 {
65 static const unsigned char pad[64] = { 0x80 };
66 unsigned int padlen[2];
67 int i;
68
69 /* Pad with a binary 1 (ie 0x80), then zeroes, then length
70 */
71 padlen[0] = htonl(ctx->size >> (32 - 3));
72 padlen[1] = htonl(ctx->size << 3);
73
74 blk_SHA1_Update(ctx, pad, 1+ (63 & (55 - ctx->lenW)));
75 blk_SHA1_Update(ctx, padlen, 8);
76
77 /* Output hash
78 */
79 for (i = 0; i < 5; i++)
80 ((unsigned int *)hashout)[i] = htonl(ctx->H[i]);
81 }
82
83 #define SHA_ROT(X,n) (((X) << (n)) | ((X) >> (32-(n))))
84
85 static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
86 {
87 int t;
88 unsigned int A,B,C,D,E,TEMP;
89 unsigned int W[80];
90
91 for (t = 0; t < 16; t++)
92 W[t] = htonl(data[t]);
93
94 /* Unroll it? */
95 for (t = 16; t <= 79; t++)
96 W[t] = SHA_ROT(W[t-3] ^ W[t-8] ^ W[t-14] ^ W[t-16], 1);
97
98 A = ctx->H[0];
99 B = ctx->H[1];
100 C = ctx->H[2];
101 D = ctx->H[3];
102 E = ctx->H[4];
103
104 #define T_0_19(t) \
105 TEMP = SHA_ROT(A,5) + (((C^D)&B)^D) + E + W[t] + 0x5a827999; \
106 E = D; D = C; C = SHA_ROT(B, 30); B = A; A = TEMP;
107
108 T_0_19( 0); T_0_19( 1); T_0_19( 2); T_0_19( 3); T_0_19( 4);
109 T_0_19( 5); T_0_19( 6); T_0_19( 7); T_0_19( 8); T_0_19( 9);
110 T_0_19(10); T_0_19(11); T_0_19(12); T_0_19(13); T_0_19(14);
111 T_0_19(15); T_0_19(16); T_0_19(17); T_0_19(18); T_0_19(19);
112
113 #define T_20_39(t) \
114 TEMP = SHA_ROT(A,5) + (B^C^D) + E + W[t] + 0x6ed9eba1; \
115 E = D; D = C; C = SHA_ROT(B, 30); B = A; A = TEMP;
116
117 T_20_39(20); T_20_39(21); T_20_39(22); T_20_39(23); T_20_39(24);
118 T_20_39(25); T_20_39(26); T_20_39(27); T_20_39(28); T_20_39(29);
119 T_20_39(30); T_20_39(31); T_20_39(32); T_20_39(33); T_20_39(34);
120 T_20_39(35); T_20_39(36); T_20_39(37); T_20_39(38); T_20_39(39);
121
122 #define T_40_59(t) \
123 TEMP = SHA_ROT(A,5) + ((B&C)|(D&(B|C))) + E + W[t] + 0x8f1bbcdc; \
124 E = D; D = C; C = SHA_ROT(B, 30); B = A; A = TEMP;
125
126 T_40_59(40); T_40_59(41); T_40_59(42); T_40_59(43); T_40_59(44);
127 T_40_59(45); T_40_59(46); T_40_59(47); T_40_59(48); T_40_59(49);
128 T_40_59(50); T_40_59(51); T_40_59(52); T_40_59(53); T_40_59(54);
129 T_40_59(55); T_40_59(56); T_40_59(57); T_40_59(58); T_40_59(59);
130
131 #define T_60_79(t) \
132 TEMP = SHA_ROT(A,5) + (B^C^D) + E + W[t] + 0xca62c1d6; \
133 E = D; D = C; C = SHA_ROT(B, 30); B = A; A = TEMP;
134
135 T_60_79(60); T_60_79(61); T_60_79(62); T_60_79(63); T_60_79(64);
136 T_60_79(65); T_60_79(66); T_60_79(67); T_60_79(68); T_60_79(69);
137 T_60_79(70); T_60_79(71); T_60_79(72); T_60_79(73); T_60_79(74);
138 T_60_79(75); T_60_79(76); T_60_79(77); T_60_79(78); T_60_79(79);
139
140 ctx->H[0] += A;
141 ctx->H[1] += B;
142 ctx->H[2] += C;
143 ctx->H[3] += D;
144 ctx->H[4] += E;
145 }