2 * Public Domain poly1305 from Andrew Moon
3 * poly1305-donna-unrolled.c from https://github.com/floodyberry/poly1305-donna
10 #define mul32x32_64(a,b) ((uint64_t)(a) * (b))
12 #define U8TO32_LE(p) \
13 (((uint32_t)((p)[0])) | \
14 ((uint32_t)((p)[1]) << 8) | \
15 ((uint32_t)((p)[2]) << 16) | \
16 ((uint32_t)((p)[3]) << 24))
18 #define U32TO8_LE(p, v) \
20 (p)[0] = (uint8_t)((v)); \
21 (p)[1] = (uint8_t)((v) >> 8); \
22 (p)[2] = (uint8_t)((v) >> 16); \
23 (p)[3] = (uint8_t)((v) >> 24); \
27 poly1305_auth(unsigned char out[POLY1305_TAGLEN], const unsigned char *m, size_t inlen, const unsigned char key[POLY1305_KEYLEN])
29 uint32_t t0, t1, t2, t3;
30 uint32_t h0, h1, h2, h3, h4;
31 uint32_t r0, r1, r2, r3, r4;
32 uint32_t s1, s2, s3, s4;
36 uint64_t f0, f1, f2, f3;
37 uint32_t g0, g1, g2, g3, g4;
42 t0 = U8TO32_LE(key + 0);
43 t1 = U8TO32_LE(key + 4);
44 t2 = U8TO32_LE(key + 8);
45 t3 = U8TO32_LE(key + 12);
47 /* precompute multipliers */
75 goto poly1305_donna_atmost15bytes;
77 poly1305_donna_16bytes:
81 t0 = U8TO32_LE(m - 16);
82 t1 = U8TO32_LE(m - 12);
83 t2 = U8TO32_LE(m - 8);
84 t3 = U8TO32_LE(m - 4);
87 h1 += ((((uint64_t) t1 << 32) | t0) >> 26) & 0x3ffffff;
88 h2 += ((((uint64_t) t2 << 32) | t1) >> 20) & 0x3ffffff;
89 h3 += ((((uint64_t) t3 << 32) | t2) >> 14) & 0x3ffffff;
90 h4 += (t3 >> 8) | (1 << 24);
93 t[0] = mul32x32_64(h0, r0) + mul32x32_64(h1, s4) + mul32x32_64(h2, s3) + mul32x32_64(h3, s2) + mul32x32_64(h4, s1);
94 t[1] = mul32x32_64(h0, r1) + mul32x32_64(h1, r0) + mul32x32_64(h2, s4) + mul32x32_64(h3, s3) + mul32x32_64(h4, s2);
95 t[2] = mul32x32_64(h0, r2) + mul32x32_64(h1, r1) + mul32x32_64(h2, r0) + mul32x32_64(h3, s4) + mul32x32_64(h4, s3);
96 t[3] = mul32x32_64(h0, r3) + mul32x32_64(h1, r2) + mul32x32_64(h2, r1) + mul32x32_64(h3, r0) + mul32x32_64(h4, s4);
97 t[4] = mul32x32_64(h0, r4) + mul32x32_64(h1, r3) + mul32x32_64(h2, r2) + mul32x32_64(h3, r1) + mul32x32_64(h4, r0);
99 h0 = (uint32_t) t[0] & 0x3ffffff;
102 h1 = (uint32_t) t[1] & 0x3ffffff;
103 b = (uint32_t) (t[1] >> 26);
105 h2 = (uint32_t) t[2] & 0x3ffffff;
106 b = (uint32_t) (t[2] >> 26);
108 h3 = (uint32_t) t[3] & 0x3ffffff;
109 b = (uint32_t) (t[3] >> 26);
111 h4 = (uint32_t) t[4] & 0x3ffffff;
112 b = (uint32_t) (t[4] >> 26);
116 goto poly1305_donna_16bytes;
119 poly1305_donna_atmost15bytes:
121 goto poly1305_donna_finish;
123 for (j = 0; j < inlen; j++)
130 t0 = U8TO32_LE(mp + 0);
131 t1 = U8TO32_LE(mp + 4);
132 t2 = U8TO32_LE(mp + 8);
133 t3 = U8TO32_LE(mp + 12);
135 h0 += t0 & 0x3ffffff;
136 h1 += ((((uint64_t) t1 << 32) | t0) >> 26) & 0x3ffffff;
137 h2 += ((((uint64_t) t2 << 32) | t1) >> 20) & 0x3ffffff;
138 h3 += ((((uint64_t) t3 << 32) | t2) >> 14) & 0x3ffffff;
141 goto poly1305_donna_mul;
143 poly1305_donna_finish:
175 g4 = h4 + b - (1 << 26);
179 h0 = (h0 & nb) | (g0 & b);
180 h1 = (h1 & nb) | (g1 & b);
181 h2 = (h2 & nb) | (g2 & b);
182 h3 = (h3 & nb) | (g3 & b);
183 h4 = (h4 & nb) | (g4 & b);
185 f0 = ((h0) | (h1 << 26)) + (uint64_t) U8TO32_LE(&key[16]);
186 f1 = ((h1 >> 6) | (h2 << 20)) + (uint64_t) U8TO32_LE(&key[20]);
187 f2 = ((h2 >> 12) | (h3 << 14)) + (uint64_t) U8TO32_LE(&key[24]);
188 f3 = ((h3 >> 18) | (h4 << 8)) + (uint64_t) U8TO32_LE(&key[28]);
190 U32TO8_LE(&out[0], f0);
192 U32TO8_LE(&out[4], f1);
194 U32TO8_LE(&out[8], f2);
196 U32TO8_LE(&out[12], f3);