Line data Source code
1 : // Copyright (c) 2019 The Bitcoin Core developers
2 : // Distributed under the MIT software license, see the accompanying
3 : // file COPYING or http://www.opensource.org/licenses/mit-license.php.
4 :
5 : // Based on the public domain implementation by Andrew Moon
6 : // poly1305-donna-unrolled.c from https://github.com/floodyberry/poly1305-donna
7 :
8 : #include <crypto/common.h>
9 : #include <crypto/poly1305.h>
10 :
11 : #include <string.h>
12 :
13 : #define mul32x32_64(a,b) ((uint64_t)(a) * (b))
14 :
15 9521 : void poly1305_auth(unsigned char out[POLY1305_TAGLEN], const unsigned char *m, size_t inlen, const unsigned char key[POLY1305_KEYLEN]) {
16 : uint32_t t0,t1,t2,t3;
17 : uint32_t h0,h1,h2,h3,h4;
18 : uint32_t r0,r1,r2,r3,r4;
19 : uint32_t s1,s2,s3,s4;
20 : uint32_t b, nb;
21 : size_t j;
22 : uint64_t t[5];
23 : uint64_t f0,f1,f2,f3;
24 : uint64_t g0,g1,g2,g3,g4;
25 : uint64_t c;
26 9521 : unsigned char mp[16];
27 :
28 : /* clamp key */
29 9521 : t0 = ReadLE32(key+0);
30 9521 : t1 = ReadLE32(key+4);
31 9521 : t2 = ReadLE32(key+8);
32 9521 : t3 = ReadLE32(key+12);
33 :
34 : /* precompute multipliers */
35 9521 : r0 = t0 & 0x3ffffff; t0 >>= 26; t0 |= t1 << 6;
36 9521 : r1 = t0 & 0x3ffff03; t1 >>= 20; t1 |= t2 << 12;
37 9521 : r2 = t1 & 0x3ffc0ff; t2 >>= 14; t2 |= t3 << 18;
38 9521 : r3 = t2 & 0x3f03fff; t3 >>= 8;
39 9521 : r4 = t3 & 0x00fffff;
40 :
41 9521 : s1 = r1 * 5;
42 9521 : s2 = r2 * 5;
43 9521 : s3 = r3 * 5;
44 9521 : s4 = r4 * 5;
45 :
46 : /* init state */
47 : h0 = 0;
48 : h1 = 0;
49 : h2 = 0;
50 : h3 = 0;
51 : h4 = 0;
52 :
53 : /* full blocks */
54 9521 : if (inlen < 16) goto poly1305_donna_atmost15bytes;
55 : poly1305_donna_16bytes:
56 2946206 : m += 16;
57 2946206 : inlen -= 16;
58 :
59 2946206 : t0 = ReadLE32(m-16);
60 2946206 : t1 = ReadLE32(m-12);
61 2946206 : t2 = ReadLE32(m-8);
62 2946206 : t3 = ReadLE32(m-4);
63 :
64 2946206 : h0 += t0 & 0x3ffffff;
65 2946206 : h1 += ((((uint64_t)t1 << 32) | t0) >> 26) & 0x3ffffff;
66 2946206 : h2 += ((((uint64_t)t2 << 32) | t1) >> 20) & 0x3ffffff;
67 2946206 : h3 += ((((uint64_t)t3 << 32) | t2) >> 14) & 0x3ffffff;
68 2946206 : h4 += (t3 >> 8) | (1 << 24);
69 :
70 :
71 : poly1305_donna_mul:
72 2948211 : t[0] = mul32x32_64(h0,r0) + mul32x32_64(h1,s4) + mul32x32_64(h2,s3) + mul32x32_64(h3,s2) + mul32x32_64(h4,s1);
73 2948211 : t[1] = mul32x32_64(h0,r1) + mul32x32_64(h1,r0) + mul32x32_64(h2,s4) + mul32x32_64(h3,s3) + mul32x32_64(h4,s2);
74 2948211 : t[2] = mul32x32_64(h0,r2) + mul32x32_64(h1,r1) + mul32x32_64(h2,r0) + mul32x32_64(h3,s4) + mul32x32_64(h4,s3);
75 2948211 : t[3] = mul32x32_64(h0,r3) + mul32x32_64(h1,r2) + mul32x32_64(h2,r1) + mul32x32_64(h3,r0) + mul32x32_64(h4,s4);
76 2948211 : t[4] = mul32x32_64(h0,r4) + mul32x32_64(h1,r3) + mul32x32_64(h2,r2) + mul32x32_64(h3,r1) + mul32x32_64(h4,r0);
77 :
78 2948211 : h0 = (uint32_t)t[0] & 0x3ffffff; c = (t[0] >> 26);
79 2948211 : t[1] += c; h1 = (uint32_t)t[1] & 0x3ffffff; b = (uint32_t)(t[1] >> 26);
80 2948211 : t[2] += b; h2 = (uint32_t)t[2] & 0x3ffffff; b = (uint32_t)(t[2] >> 26);
81 2948211 : t[3] += b; h3 = (uint32_t)t[3] & 0x3ffffff; b = (uint32_t)(t[3] >> 26);
82 2948211 : t[4] += b; h4 = (uint32_t)t[4] & 0x3ffffff; b = (uint32_t)(t[4] >> 26);
83 2948211 : h0 += b * 5;
84 :
85 2948211 : if (inlen >= 16) goto poly1305_donna_16bytes;
86 :
87 : /* final bytes */
88 : poly1305_donna_atmost15bytes:
89 11526 : if (!inlen) goto poly1305_donna_finish;
90 :
91 32051 : for (j = 0; j < inlen; j++) mp[j] = m[j];
92 2005 : mp[j++] = 1;
93 2034 : for (; j < 16; j++) mp[j] = 0;
94 : inlen = 0;
95 :
96 2005 : t0 = ReadLE32(mp+0);
97 2005 : t1 = ReadLE32(mp+4);
98 2005 : t2 = ReadLE32(mp+8);
99 2005 : t3 = ReadLE32(mp+12);
100 :
101 2005 : h0 += t0 & 0x3ffffff;
102 2005 : h1 += ((((uint64_t)t1 << 32) | t0) >> 26) & 0x3ffffff;
103 2005 : h2 += ((((uint64_t)t2 << 32) | t1) >> 20) & 0x3ffffff;
104 2005 : h3 += ((((uint64_t)t3 << 32) | t2) >> 14) & 0x3ffffff;
105 2005 : h4 += (t3 >> 8);
106 :
107 2005 : goto poly1305_donna_mul;
108 :
109 : poly1305_donna_finish:
110 9521 : b = h0 >> 26; h0 = h0 & 0x3ffffff;
111 9521 : h1 += b; b = h1 >> 26; h1 = h1 & 0x3ffffff;
112 9521 : h2 += b; b = h2 >> 26; h2 = h2 & 0x3ffffff;
113 9521 : h3 += b; b = h3 >> 26; h3 = h3 & 0x3ffffff;
114 9521 : h4 += b; b = h4 >> 26; h4 = h4 & 0x3ffffff;
115 9521 : h0 += b * 5; b = h0 >> 26; h0 = h0 & 0x3ffffff;
116 9521 : h1 += b;
117 :
118 9521 : g0 = h0 + 5; b = g0 >> 26; g0 &= 0x3ffffff;
119 9521 : g1 = h1 + b; b = g1 >> 26; g1 &= 0x3ffffff;
120 9521 : g2 = h2 + b; b = g2 >> 26; g2 &= 0x3ffffff;
121 9521 : g3 = h3 + b; b = g3 >> 26; g3 &= 0x3ffffff;
122 9521 : g4 = h4 + b - (1 << 26);
123 :
124 9521 : b = (g4 >> 31) - 1;
125 9521 : nb = ~b;
126 9521 : h0 = (h0 & nb) | (g0 & b);
127 9521 : h1 = (h1 & nb) | (g1 & b);
128 9521 : h2 = (h2 & nb) | (g2 & b);
129 9521 : h3 = (h3 & nb) | (g3 & b);
130 9521 : h4 = (h4 & nb) | (g4 & b);
131 :
132 9521 : f0 = ((h0 ) | (h1 << 26)) + (uint64_t)ReadLE32(&key[16]);
133 9521 : f1 = ((h1 >> 6) | (h2 << 20)) + (uint64_t)ReadLE32(&key[20]);
134 9521 : f2 = ((h2 >> 12) | (h3 << 14)) + (uint64_t)ReadLE32(&key[24]);
135 9521 : f3 = ((h3 >> 18) | (h4 << 8)) + (uint64_t)ReadLE32(&key[28]);
136 :
137 9521 : WriteLE32(&out[ 0], f0); f1 += (f0 >> 32);
138 9521 : WriteLE32(&out[ 4], f1); f2 += (f1 >> 32);
139 9521 : WriteLE32(&out[ 8], f2); f3 += (f2 >> 32);
140 9521 : WriteLE32(&out[12], f3);
141 9521 : }
|
