graviola 0.3.4

graviola is a modern, fast cryptography library
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295

// Written for Graviola by Joe Birr-Pixton, 2024.
// SPDX-License-Identifier: Apache-2.0 OR ISC OR MIT-0
// Originally from cifra, but later adopting the 32x32
// multiplication layout from poly1305-donna.

use super::blockwise::Blockwise;

pub(crate) struct Poly1305 {
    /// Current accumulator
    h: [u32; 5],

    /// Block multiplier
    r: [u32; 5],

    /// r[1..5] times 5
    r5: [u32; 4],

    /// Final XOR offset
    s: [u32; 4],

    /// Unprocessed input
    bw: Blockwise<16>,
}

impl Poly1305 {
    pub(crate) fn new(key: &[u8; 32]) -> Self {
        let h = [0; 5];
        let r = [
            read32(&key[0..4]) & 0x3ffffff,
            (read32(&key[3..7]) >> 2) & 0x3ffff03,
            (read32(&key[6..10]) >> 4) & 0x3ffc0ff,
            (read32(&key[9..13]) >> 6) & 0x3f03fff,
            (read32(&key[12..16]) >> 8) & 0x00fffff,
        ];
        let r5 = [r[1] * 5, r[2] * 5, r[3] * 5, r[4] * 5];
        let s = [
            read32(&key[16..20]),
            read32(&key[20..24]),
            read32(&key[24..28]),
            read32(&key[28..32]),
        ];
        Self {
            h,
            r,
            r5,
            s,
            bw: Blockwise::new(),
        }
    }

    pub(crate) fn add_bytes(&mut self, bytes: &[u8]) {
        let bytes = self.bw.add_leading(bytes);

        if let Some(block) = self.bw.take() {
            self.process_whole_block(&block);
        }

        let mut full_blocks = bytes.chunks_exact(16);
        for block in full_blocks.by_ref() {
            self.process_whole_block(block.try_into().unwrap());
        }

        self.bw.add_trailing(full_blocks.remainder());
    }

    pub(crate) fn finish(mut self) -> [u8; 16] {
        if let Some(block) = self.bw.clone().peek_remaining() {
            self.process_last_block(block);
        }

        full_reduce(&mut self.h);

        // redistribute into 4 words
        self.h[0] |= self.h[1] << 26;
        self.h[1] = (self.h[1] >> 6) | (self.h[2] << 20);
        self.h[2] = (self.h[2] >> 12) | (self.h[3] << 14);
        self.h[3] = (self.h[3] >> 18) | (self.h[4] << 8);

        // add s with carry
        fn add32(a: u32, b: u32) -> u64 {
            (a as u64) + (b as u64)
        }
        let f = add32(self.h[0], self.s[0]);
        self.h[0] = f as u32;
        let f = add32(self.h[1], self.s[1]) + (f >> 32);
        self.h[1] = f as u32;
        let f = add32(self.h[2], self.s[2]) + (f >> 32);
        self.h[2] = f as u32;
        let f = add32(self.h[3], self.s[3]) + (f >> 32);
        self.h[3] = f as u32;

        let mut r = [0u8; 16];
        r[0..4].copy_from_slice(&self.h[0].to_le_bytes());
        r[4..8].copy_from_slice(&self.h[1].to_le_bytes());
        r[8..12].copy_from_slice(&self.h[2].to_le_bytes());
        r[12..16].copy_from_slice(&self.h[3].to_le_bytes());

        r
    }

    fn process_whole_block(&mut self, inp: &[u8; 16]) {
        let block = [
            read32(&inp[0..4]) & 0x3ff_ffff,
            (read32(&inp[3..7]) >> 2) & 0x3ff_ffff,
            (read32(&inp[6..10]) >> 4) & 0x3ff_ffff,
            (read32(&inp[9..13]) >> 6) & 0x3ff_ffff,
            (read32(&inp[12..16]) >> 8) | (1 << 24),
        ];
        self.process_block(&block);
    }

    fn process_last_block(&mut self, inp: &[u8]) {
        let mut bytes = [0u8; 16];
        bytes[..inp.len()].copy_from_slice(inp);
        bytes[inp.len()] = 0x01;

        let block = [
            read32(&bytes[0..4]) & 0x3ff_ffff,
            (read32(&bytes[3..7]) >> 2) & 0x3ff_ffff,
            (read32(&bytes[6..10]) >> 4) & 0x3ff_ffff,
            (read32(&bytes[9..13]) >> 6) & 0x3ff_ffff,
            (read32(&bytes[12..16]) >> 8),
        ];
        self.process_block(&block);
    }

    fn process_block(&mut self, block: &[u32; 5]) {
        add(&mut self.h, block);
        mul(&mut self.h, &self.r, &self.r5);
    }
}

fn read32(bytes: &[u8]) -> u32 {
    u32::from_le_bytes(bytes.try_into().unwrap())
}

fn add(h: &mut [u32; 5], x: &[u32; 5]) {
    h[0] = h[0].wrapping_add(x[0]);
    h[1] = h[1].wrapping_add(x[1]);
    h[2] = h[2].wrapping_add(x[2]);
    h[3] = h[3].wrapping_add(x[3]);
    h[4] = h[4].wrapping_add(x[4]);
}

fn mul(h: &mut [u32; 5], r: &[u32; 5], s: &[u32; 4]) {
    fn mul32(a: u32, b: u32) -> u64 {
        u64::from(a) * u64::from(b)
    }

    let d0 = mul32(h[0], r[0])
        + mul32(h[1], s[3])
        + mul32(h[2], s[2])
        + mul32(h[3], s[1])
        + mul32(h[4], s[0]);
    let d1 = mul32(h[0], r[1])
        + mul32(h[1], r[0])
        + mul32(h[2], s[3])
        + mul32(h[3], s[2])
        + mul32(h[4], s[1]);
    let d2 = mul32(h[0], r[2])
        + mul32(h[1], r[1])
        + mul32(h[2], r[0])
        + mul32(h[3], s[3])
        + mul32(h[4], s[2]);
    let d3 = mul32(h[0], r[3])
        + mul32(h[1], r[2])
        + mul32(h[2], r[1])
        + mul32(h[3], r[0])
        + mul32(h[4], s[3]);
    let d4 = mul32(h[0], r[4])
        + mul32(h[1], r[3])
        + mul32(h[2], r[2])
        + mul32(h[3], r[1])
        + mul32(h[4], r[0]);

    // partial reduction
    let carry = d0 >> 26;
    h[0] = (d0 & 0x3ff_ffff) as u32;
    let d1 = d1 + carry;
    let carry = d1 >> 26;
    h[1] = (d1 & 0x3ff_ffff) as u32;
    let d2 = d2 + carry;
    let carry = d2 >> 26;
    h[2] = (d2 & 0x3ff_ffff) as u32;
    let d3 = d3 + carry;
    let carry = d3 >> 26;
    h[3] = (d3 & 0x3ff_ffff) as u32;
    let d4 = d4 + carry;
    let carry = (d4 >> 26) as u32;
    h[4] = (d4 & 0x3ff_ffff) as u32;
    h[0] += carry * 5;
    let carry = h[0] >> 26;
    h[0] &= 0x3ff_ffff;
    h[1] += carry;
}

fn full_reduce(h: &mut [u32; 5]) {
    min_reduce(h);
    maybe_sub_130_5(h);
}

fn min_reduce(h: &mut [u32; 5]) {
    let carry = h[1] >> 26;
    h[1] &= 0x3ffffff;
    h[2] = h[2].wrapping_add(carry);

    let carry = h[2] >> 26;
    h[2] &= 0x3ffffff;
    h[3] = h[3].wrapping_add(carry);

    let carry = h[3] >> 26;
    h[3] &= 0x3ffffff;
    h[4] = h[4].wrapping_add(carry);

    let carry = h[4] >> 26;
    h[4] &= 0x3ffffff;
    h[0] = h[0].wrapping_add(carry.wrapping_mul(5));

    let carry = h[0] >> 26;
    h[0] &= 0x3ffffff;
    h[1] = h[1].wrapping_add(carry);
}

fn maybe_sub_130_5(h: &mut [u32; 5]) {
    let g0 = h[0].wrapping_add(5);
    let carry = g0 >> 26;
    let g0 = g0 & 0x3ffffff;

    let g1 = h[1].wrapping_add(carry);
    let carry = g1 >> 26;
    let g1 = g1 & 0x3ffffff;

    let g2 = h[2].wrapping_add(carry);
    let carry = g2 >> 26;
    let g2 = g2 & 0x3ffffff;

    let g3 = h[3].wrapping_add(carry);
    let carry = g3 >> 26;
    let g3 = g3 & 0x3ffffff;

    let g4 = h[4].wrapping_add(carry).wrapping_sub(1 << 26);

    let negative_mask = equal_mask(g4 & 0x80000000, 0x80000000);
    let positive_mask = !negative_mask;

    h[0] = (h[0] & negative_mask) | (g0 & positive_mask);
    h[1] = (h[1] & negative_mask) | (g1 & positive_mask);
    h[2] = (h[2] & negative_mask) | (g2 & positive_mask);
    h[3] = (h[3] & negative_mask) | (g3 & positive_mask);
    h[4] = (h[4] & negative_mask) | (g4 & positive_mask);
}

/// Produce 0xffffffff if x == y, zero
fn equal_mask(x: u32, y: u32) -> u32 {
    let diff = x ^ y;
    let diff_is_zero = !diff & diff.wrapping_sub(1);
    0u32.wrapping_sub(diff_is_zero >> 31)
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_vectors() {
        // From draft-agl-tls-chacha20poly1305-04 section 7
        let key = &[
            0x74, 0x68, 0x69, 0x73, 0x20, 0x69, 0x73, 0x20, 0x33, 0x32, 0x2d, 0x62, 0x79, 0x74,
            0x65, 0x20, 0x6b, 0x65, 0x79, 0x20, 0x66, 0x6f, 0x72, 0x20, 0x50, 0x6f, 0x6c, 0x79,
            0x31, 0x33, 0x30, 0x35,
        ];

        let mut p = Poly1305::new(key);
        p.add_bytes(&[0u8; 32]);
        assert_eq!(
            p.finish(),
            [
                0x49, 0xec, 0x78, 0x09, 0x0e, 0x48, 0x1e, 0xc6, 0xc2, 0x6b, 0x33, 0xb9, 0x1c, 0xcc,
                0x03, 0x07
            ]
        );

        let mut p = Poly1305::new(key);
        p.add_bytes(&[
            0x48, 0x65, 0x6c, 0x6c, 0x6f, 0x20, 0x77, 0x6f, 0x72, 0x6c, 0x64, 0x21,
        ]);
        assert_eq!(
            p.finish(),
            [
                0xa6, 0xf7, 0x45, 0x00, 0x8f, 0x81, 0xc9, 0x16, 0xa2, 0x0d, 0xcc, 0x74, 0xee, 0xf2,
                0xb2, 0xf0
            ]
        );
    }
}