Add dprint() to io_u requeue
[fio.git] / crc / sha512.c
1 /* SHA-512 code by Jean-Luc Cooke <jlcooke@certainkey.com>
2  *
3  * Copyright (c) Jean-Luc Cooke <jlcooke@certainkey.com>
4  * Copyright (c) Andrew McDonald <andrew@mcdonald.org.uk>
5  * Copyright (c) 2003 Kyle McMartin <kyle@debian.org>
6  *
7  * This program is free software; you can redistribute it and/or modify it
8  * under the terms of the GNU General Public License as published by the
9  * Free Software Foundation; either version 2, or (at your option) any
10  * later version.
11  *
12  */
13
14 #include <string.h>
15 #include <inttypes.h>
16
17 #include "sha512.h"
18
19 #if __BYTE_ORDER == __LITTLE_ENDIAN
20 static int __be64_to_cpu(uint64_t val)
21 {
22         uint64_t c1, c2, c3, c4, c5, c6, c7, c8;
23
24         c1 = (val >> 56) & 0xff;
25         c2 = (val >> 48) & 0xff;
26         c3 = (val >> 40) & 0xff;
27         c4 = (val >> 32) & 0xff;
28         c5 = (val >> 24) & 0xff;
29         c6 = (val >> 16) & 0xff;
30         c7 = (val >> 8) & 0xff;
31         c8 = val & 0xff;
32
33         return c1 | c2 << 8 | c3 << 16 | c4 << 24 | c5 << 32 | c6 << 40 | c7 << 48 | c8 << 56;
34 }
35 #else
36 #define __be64_to_cpu(x)        (x)
37 #endif
38
39 #define SHA384_DIGEST_SIZE 48
40 #define SHA512_DIGEST_SIZE 64
41 #define SHA384_HMAC_BLOCK_SIZE 128
42 #define SHA512_HMAC_BLOCK_SIZE 128
43
44 static inline uint64_t Ch(uint64_t x, uint64_t y, uint64_t z)
45 {
46         return z ^ (x & (y ^ z));
47 }
48
49 static inline uint64_t Maj(uint64_t x, uint64_t y, uint64_t z)
50 {
51         return (x & y) | (z & (x | y));
52 }
53
54 static inline uint64_t RORuint64_t(uint64_t x, uint64_t y)
55 {
56         return (x >> y) | (x << (64 - y));
57 }
58
59 static const uint64_t sha512_K[80] = {
60         0x428a2f98d728ae22ULL, 0x7137449123ef65cdULL, 0xb5c0fbcfec4d3b2fULL,
61         0xe9b5dba58189dbbcULL, 0x3956c25bf348b538ULL, 0x59f111f1b605d019ULL,
62         0x923f82a4af194f9bULL, 0xab1c5ed5da6d8118ULL, 0xd807aa98a3030242ULL,
63         0x12835b0145706fbeULL, 0x243185be4ee4b28cULL, 0x550c7dc3d5ffb4e2ULL,
64         0x72be5d74f27b896fULL, 0x80deb1fe3b1696b1ULL, 0x9bdc06a725c71235ULL,
65         0xc19bf174cf692694ULL, 0xe49b69c19ef14ad2ULL, 0xefbe4786384f25e3ULL,
66         0x0fc19dc68b8cd5b5ULL, 0x240ca1cc77ac9c65ULL, 0x2de92c6f592b0275ULL,
67         0x4a7484aa6ea6e483ULL, 0x5cb0a9dcbd41fbd4ULL, 0x76f988da831153b5ULL,
68         0x983e5152ee66dfabULL, 0xa831c66d2db43210ULL, 0xb00327c898fb213fULL,
69         0xbf597fc7beef0ee4ULL, 0xc6e00bf33da88fc2ULL, 0xd5a79147930aa725ULL,
70         0x06ca6351e003826fULL, 0x142929670a0e6e70ULL, 0x27b70a8546d22ffcULL,
71         0x2e1b21385c26c926ULL, 0x4d2c6dfc5ac42aedULL, 0x53380d139d95b3dfULL,
72         0x650a73548baf63deULL, 0x766a0abb3c77b2a8ULL, 0x81c2c92e47edaee6ULL,
73         0x92722c851482353bULL, 0xa2bfe8a14cf10364ULL, 0xa81a664bbc423001ULL,
74         0xc24b8b70d0f89791ULL, 0xc76c51a30654be30ULL, 0xd192e819d6ef5218ULL,
75         0xd69906245565a910ULL, 0xf40e35855771202aULL, 0x106aa07032bbd1b8ULL,
76         0x19a4c116b8d2d0c8ULL, 0x1e376c085141ab53ULL, 0x2748774cdf8eeb99ULL,
77         0x34b0bcb5e19b48a8ULL, 0x391c0cb3c5c95a63ULL, 0x4ed8aa4ae3418acbULL,
78         0x5b9cca4f7763e373ULL, 0x682e6ff3d6b2b8a3ULL, 0x748f82ee5defb2fcULL,
79         0x78a5636f43172f60ULL, 0x84c87814a1f0ab72ULL, 0x8cc702081a6439ecULL,
80         0x90befffa23631e28ULL, 0xa4506cebde82bde9ULL, 0xbef9a3f7b2c67915ULL,
81         0xc67178f2e372532bULL, 0xca273eceea26619cULL, 0xd186b8c721c0c207ULL,
82         0xeada7dd6cde0eb1eULL, 0xf57d4f7fee6ed178ULL, 0x06f067aa72176fbaULL,
83         0x0a637dc5a2c898a6ULL, 0x113f9804bef90daeULL, 0x1b710b35131c471bULL,
84         0x28db77f523047d84ULL, 0x32caab7b40c72493ULL, 0x3c9ebe0a15c9bebcULL,
85         0x431d67c49c100d4cULL, 0x4cc5d4becb3e42b6ULL, 0x597f299cfc657e2aULL,
86         0x5fcb6fab3ad6faecULL, 0x6c44198c4a475817ULL,
87 };
88
89 #define e0(x)       (RORuint64_t(x,28) ^ RORuint64_t(x,34) ^ RORuint64_t(x,39))
90 #define e1(x)       (RORuint64_t(x,14) ^ RORuint64_t(x,18) ^ RORuint64_t(x,41))
91 #define s0(x)       (RORuint64_t(x, 1) ^ RORuint64_t(x, 8) ^ (x >> 7))
92 #define s1(x)       (RORuint64_t(x,19) ^ RORuint64_t(x,61) ^ (x >> 6))
93
94 /* H* initial state for SHA-512 */
95 #define H0         0x6a09e667f3bcc908ULL
96 #define H1         0xbb67ae8584caa73bULL
97 #define H2         0x3c6ef372fe94f82bULL
98 #define H3         0xa54ff53a5f1d36f1ULL
99 #define H4         0x510e527fade682d1ULL
100 #define H5         0x9b05688c2b3e6c1fULL
101 #define H6         0x1f83d9abfb41bd6bULL
102 #define H7         0x5be0cd19137e2179ULL
103
104 /* H'* initial state for SHA-384 */
105 #define HP0 0xcbbb9d5dc1059ed8ULL
106 #define HP1 0x629a292a367cd507ULL
107 #define HP2 0x9159015a3070dd17ULL
108 #define HP3 0x152fecd8f70e5939ULL
109 #define HP4 0x67332667ffc00b31ULL
110 #define HP5 0x8eb44a8768581511ULL
111 #define HP6 0xdb0c2e0d64f98fa7ULL
112 #define HP7 0x47b5481dbefa4fa4ULL
113
114 static inline void LOAD_OP(int I, uint64_t *W, const uint8_t *input)
115 {
116         W[I] = __be64_to_cpu( ((uint64_t *)(input))[I] );
117 }
118
119 static inline void BLEND_OP(int I, uint64_t *W)
120 {
121         W[I] = s1(W[I-2]) + W[I-7] + s0(W[I-15]) + W[I-16];
122 }
123
124 static void sha512_transform(uint64_t *state, uint64_t *W, const uint8_t *input)
125 {
126         uint64_t a, b, c, d, e, f, g, h, t1, t2;
127
128         int i;
129
130         /* load the input */
131         for (i = 0; i < 16; i++)
132                 LOAD_OP(i, W, input);
133
134         for (i = 16; i < 80; i++)
135                 BLEND_OP(i, W);
136
137         /* load the state into our registers */
138         a=state[0];   b=state[1];   c=state[2];   d=state[3];
139         e=state[4];   f=state[5];   g=state[6];   h=state[7];
140
141         /* now iterate */
142         for (i=0; i<80; i+=8) {
143                 t1 = h + e1(e) + Ch(e,f,g) + sha512_K[i  ] + W[i  ];
144                 t2 = e0(a) + Maj(a,b,c);    d+=t1;    h=t1+t2;
145                 t1 = g + e1(d) + Ch(d,e,f) + sha512_K[i+1] + W[i+1];
146                 t2 = e0(h) + Maj(h,a,b);    c+=t1;    g=t1+t2;
147                 t1 = f + e1(c) + Ch(c,d,e) + sha512_K[i+2] + W[i+2];
148                 t2 = e0(g) + Maj(g,h,a);    b+=t1;    f=t1+t2;
149                 t1 = e + e1(b) + Ch(b,c,d) + sha512_K[i+3] + W[i+3];
150                 t2 = e0(f) + Maj(f,g,h);    a+=t1;    e=t1+t2;
151                 t1 = d + e1(a) + Ch(a,b,c) + sha512_K[i+4] + W[i+4];
152                 t2 = e0(e) + Maj(e,f,g);    h+=t1;    d=t1+t2;
153                 t1 = c + e1(h) + Ch(h,a,b) + sha512_K[i+5] + W[i+5];
154                 t2 = e0(d) + Maj(d,e,f);    g+=t1;    c=t1+t2;
155                 t1 = b + e1(g) + Ch(g,h,a) + sha512_K[i+6] + W[i+6];
156                 t2 = e0(c) + Maj(c,d,e);    f+=t1;    b=t1+t2;
157                 t1 = a + e1(f) + Ch(f,g,h) + sha512_K[i+7] + W[i+7];
158                 t2 = e0(b) + Maj(b,c,d);    e+=t1;    a=t1+t2;
159         }
160
161         state[0] += a; state[1] += b; state[2] += c; state[3] += d;
162         state[4] += e; state[5] += f; state[6] += g; state[7] += h;
163
164         /* erase our data */
165         a = b = c = d = e = f = g = h = t1 = t2 = 0;
166 }
167
168 void sha512_init(struct sha512_ctx *sctx)
169 {
170         sctx->state[0] = H0;
171         sctx->state[1] = H1;
172         sctx->state[2] = H2;
173         sctx->state[3] = H3;
174         sctx->state[4] = H4;
175         sctx->state[5] = H5;
176         sctx->state[6] = H6;
177         sctx->state[7] = H7;
178         sctx->count[0] = sctx->count[1] = sctx->count[2] = sctx->count[3] = 0;
179 }
180
181 void sha512_update(struct sha512_ctx *sctx, const uint8_t *data,
182                    unsigned int len)
183 {
184         unsigned int i, index, part_len;
185
186         /* Compute number of bytes mod 128 */
187         index = (unsigned int)((sctx->count[0] >> 3) & 0x7F);
188         
189         /* Update number of bits */
190         if ((sctx->count[0] += (len << 3)) < (len << 3)) {
191                 if ((sctx->count[1] += 1) < 1)
192                         if ((sctx->count[2] += 1) < 1)
193                                 sctx->count[3]++;
194                 sctx->count[1] += (len >> 29);
195         }
196         
197         part_len = 128 - index;
198         
199         /* Transform as many times as possible. */
200         if (len >= part_len) {
201                 memcpy(&sctx->buf[index], data, part_len);
202                 sha512_transform(sctx->state, sctx->W, sctx->buf);
203
204                 for (i = part_len; i + 127 < len; i+=128)
205                         sha512_transform(sctx->state, sctx->W, &data[i]);
206
207                 index = 0;
208         } else {
209                 i = 0;
210         }
211
212         /* Buffer remaining input */
213         memcpy(&sctx->buf[index], &data[i], len - i);
214
215         /* erase our data */
216         memset(sctx->W, 0, sizeof(sctx->W));
217 }