summaryrefslogtreecommitdiffstats
path: root/vendor/ahash/src/aes_hash.rs
blob: 1c98582ceec691d29ec1abac313a72e03b4cb642 (plain)
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
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
use crate::convert::*;
#[cfg(feature = "specialize")]
use crate::fallback_hash::MULTIPLE;
use crate::operations::*;
use crate::RandomState;
use core::hash::Hasher;
use crate::random_state::PI;

/// A `Hasher` for hashing an arbitrary stream of bytes.
///
/// Instances of [`AHasher`] represent state that is updated while hashing data.
///
/// Each method updates the internal state based on the new data provided. Once
/// all of the data has been provided, the resulting hash can be obtained by calling
/// `finish()`
///
/// [Clone] is also provided in case you wish to calculate hashes for two different items that
/// start with the same data.
///
#[derive(Debug, Clone)]
pub struct AHasher {
    enc: u128,
    sum: u128,
    key: u128,
}

impl AHasher {
    /// Creates a new hasher keyed to the provided keys.
    ///
    /// Normally hashers are created via `AHasher::default()` for fixed keys or `RandomState::new()` for randomly
    /// generated keys and `RandomState::with_seeds(a,b)` for seeds that are set and can be reused. All of these work at
    /// map creation time (and hence don't have any overhead on a per-item bais).
    ///
    /// This method directly creates the hasher instance and performs no transformation on the provided seeds. This may
    /// be useful where a HashBuilder is not desired, such as for testing purposes.
    ///
    /// # Example
    ///
    /// ```
    /// use std::hash::Hasher;
    /// use ahash::AHasher;
    ///
    /// let mut hasher = AHasher::new_with_keys(1234, 5678);
    ///
    /// hasher.write_u32(1989);
    /// hasher.write_u8(11);
    /// hasher.write_u8(9);
    /// hasher.write(b"Huh?");
    ///
    /// println!("Hash is {:x}!", hasher.finish());
    /// ```
    #[inline]
    pub fn new_with_keys(key1: u128, key2: u128) -> Self {
        let pi: [u128; 2] = PI.convert();
        let key1 = key1 ^ pi[0];
        let key2 = key2 ^ pi[1];
        Self {
            enc: key1,
            sum: key2,
            key: key1 ^ key2,
        }
    }

    #[allow(unused)] // False positive
    pub(crate) fn test_with_keys(key1: u128, key2: u128) -> Self {
        Self {
            enc: key1,
            sum: key2,
            key: key1 ^ key2,
        }
    }


    #[inline]
    pub(crate) fn from_random_state(rand_state: &RandomState) -> Self {
        let key1 = [rand_state.k0, rand_state.k1].convert();
        let key2 = [rand_state.k2, rand_state.k3].convert();
        Self {
            enc: key1,
            sum: key2,
            key: key1 ^ key2,
        }
    }

    #[inline(always)]
    fn add_in_length(&mut self, length: u64) {
        //This will be scrambled by the next AES round.
        let mut enc: [u64; 2] = self.enc.convert();
        enc[0] = enc[0].wrapping_add(length);
        self.enc = enc.convert();
    }

    #[inline(always)]
    fn hash_in(&mut self, new_value: u128) {
        self.enc = aesenc(self.enc, new_value);
        self.sum = shuffle_and_add(self.sum, new_value);
    }

    #[inline(always)]
    fn hash_in_2(&mut self, v1: u128, v2: u128) {
        self.enc = aesenc(self.enc, v1);
        self.sum = shuffle_and_add(self.sum, v1);
        self.enc = aesenc(self.enc, v2);
        self.sum = shuffle_and_add(self.sum, v2);
    }

    #[inline]
    #[cfg(feature = "specialize")]
    fn short_finish(&self) -> u64 {
        let combined = aesdec(self.sum, self.enc);
        let result: [u64; 2] = aesenc(combined, combined).convert();
        result[0]
    }
}

/// Provides [Hasher] methods to hash all of the primitive types.
///
/// [Hasher]: core::hash::Hasher
impl Hasher for AHasher {
    #[inline]
    fn write_u8(&mut self, i: u8) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u16(&mut self, i: u16) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u32(&mut self, i: u32) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u128(&mut self, i: u128) {
        self.hash_in(i);
    }

    #[inline]
    #[cfg(any(target_pointer_width = "64", target_pointer_width = "32", target_pointer_width = "16"))]
    fn write_usize(&mut self, i: usize) {
        self.write_u64(i as u64);
    }

    #[inline]
    #[cfg(target_pointer_width = "128")]
    fn write_usize(&mut self, i: usize) {
        self.write_u128(i as u128);
    }

    #[inline]
    fn write_u64(&mut self, i: u64) {
        self.write_u128(i as u128);
    }

    #[inline]
    #[allow(clippy::collapsible_if)]
    fn write(&mut self, input: &[u8]) {
        let mut data = input;
        let length = data.len();
        self.add_in_length(length as u64);
        //A 'binary search' on sizes reduces the number of comparisons.
        if data.len() <= 8 {
            let value = read_small(data);
            self.hash_in(value.convert());
        } else {
            if data.len() > 32 {
                if data.len() > 64 {
                    let tail = data.read_last_u128x4();
                    let mut current: [u128; 4] = [self.key; 4];
                    current[0] = aesenc(current[0], tail[0]);
                    current[1] = aesenc(current[1], tail[1]);
                    current[2] = aesenc(current[2], tail[2]);
                    current[3] = aesenc(current[3], tail[3]);
                    let mut sum: [u128; 2] = [self.key, self.key];
                    sum[0] = add_by_64s(sum[0].convert(), tail[0].convert()).convert();
                    sum[1] = add_by_64s(sum[1].convert(), tail[1].convert()).convert();
                    sum[0] = shuffle_and_add(sum[0], tail[2]);
                    sum[1] = shuffle_and_add(sum[1], tail[3]);
                    while data.len() > 64 {
                        let (blocks, rest) = data.read_u128x4();
                        current[0] = aesenc(current[0], blocks[0]);
                        current[1] = aesenc(current[1], blocks[1]);
                        current[2] = aesenc(current[2], blocks[2]);
                        current[3] = aesenc(current[3], blocks[3]);
                        sum[0] = shuffle_and_add(sum[0], blocks[0]);
                        sum[1] = shuffle_and_add(sum[1], blocks[1]);
                        sum[0] = shuffle_and_add(sum[0], blocks[2]);
                        sum[1] = shuffle_and_add(sum[1], blocks[3]);
                        data = rest;
                    }
                    self.hash_in_2(aesenc(current[0], current[1]), aesenc(current[2], current[3]));
                    self.hash_in(add_by_64s(sum[0].convert(), sum[1].convert()).convert());
                } else {
                    //len 33-64
                    let (head, _) = data.read_u128x2();
                    let tail = data.read_last_u128x2();
                    self.hash_in_2(head[0], head[1]);
                    self.hash_in_2(tail[0], tail[1]);
                }
            } else {
                if data.len() > 16 {
                    //len 17-32
                    self.hash_in_2(data.read_u128().0, data.read_last_u128());
                } else {
                    //len 9-16
                    let value: [u64; 2] = [data.read_u64().0, data.read_last_u64()];
                    self.hash_in(value.convert());
                }
            }
        }
    }
    #[inline]
    fn finish(&self) -> u64 {
        let combined = aesdec(self.sum, self.enc);
        let result: [u64; 2] = aesenc(aesenc(combined, self.key), combined).convert();
        result[0]
    }
}

#[cfg(feature = "specialize")]
pub(crate) struct AHasherU64 {
    pub(crate) buffer: u64,
    pub(crate) pad: u64,
}

/// A specialized hasher for only primitives under 64 bits.
#[cfg(feature = "specialize")]
impl Hasher for AHasherU64 {
    #[inline]
    fn finish(&self) -> u64 {
        let rot = (self.pad & 63) as u32;
        self.buffer.rotate_left(rot)
    }

    #[inline]
    fn write(&mut self, _bytes: &[u8]) {
        unreachable!("Specialized hasher was called with a different type of object")
    }

    #[inline]
    fn write_u8(&mut self, i: u8) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u16(&mut self, i: u16) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u32(&mut self, i: u32) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u64(&mut self, i: u64) {
        self.buffer = folded_multiply(i ^ self.buffer, MULTIPLE);
    }

    #[inline]
    fn write_u128(&mut self, _i: u128) {
        unreachable!("Specialized hasher was called with a different type of object")
    }

    #[inline]
    fn write_usize(&mut self, _i: usize) {
        unreachable!("Specialized hasher was called with a different type of object")
    }
}

#[cfg(feature = "specialize")]
pub(crate) struct AHasherFixed(pub AHasher);

/// A specialized hasher for fixed size primitives larger than 64 bits.
#[cfg(feature = "specialize")]
impl Hasher for AHasherFixed {
    #[inline]
    fn finish(&self) -> u64 {
        self.0.short_finish()
    }

    #[inline]
    fn write(&mut self, bytes: &[u8]) {
        self.0.write(bytes)
    }

    #[inline]
    fn write_u8(&mut self, i: u8) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u16(&mut self, i: u16) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u32(&mut self, i: u32) {
        self.write_u64(i as u64);
    }

    #[inline]
    fn write_u64(&mut self, i: u64) {
        self.0.write_u64(i);
    }

    #[inline]
    fn write_u128(&mut self, i: u128) {
        self.0.write_u128(i);
    }

    #[inline]
    fn write_usize(&mut self, i: usize) {
        self.0.write_usize(i);
    }
}

#[cfg(feature = "specialize")]
pub(crate) struct AHasherStr(pub AHasher);

/// A specialized hasher for strings
/// Note that the other types don't panic because the hash impl for String tacks on an unneeded call. (As does vec)
#[cfg(feature = "specialize")]
impl Hasher for AHasherStr {
    #[inline]
    fn finish(&self) -> u64 {
        let result : [u64; 2] = self.0.enc.convert();
        result[0]
    }

    #[inline]
    fn write(&mut self, bytes: &[u8]) {
        if bytes.len() > 8 {
            self.0.write(bytes);
            self.0.enc = aesdec(self.0.sum, self.0.enc);
            self.0.enc = aesenc(aesenc(self.0.enc, self.0.key), self.0.enc);
        } else {
            self.0.add_in_length(bytes.len() as u64);
            let value = read_small(bytes).convert();
            self.0.sum = shuffle_and_add(self.0.sum, value);
            self.0.enc = aesdec(self.0.sum, self.0.enc);
            self.0.enc = aesenc(aesenc(self.0.enc, self.0.key), self.0.enc);
        }
    }

    #[inline]
    fn write_u8(&mut self, _i: u8) {}

    #[inline]
    fn write_u16(&mut self, _i: u16) {}

    #[inline]
    fn write_u32(&mut self, _i: u32) {}

    #[inline]
    fn write_u64(&mut self, _i: u64) {}

    #[inline]
    fn write_u128(&mut self, _i: u128) {}

    #[inline]
    fn write_usize(&mut self, _i: usize) {}
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::convert::Convert;
    use crate::operations::aesenc;
    use crate::RandomState;
    use std::hash::{BuildHasher, Hasher};
    #[test]
    fn test_sanity() {
        let mut hasher = RandomState::with_seeds(1, 2, 3, 4).build_hasher();
        hasher.write_u64(0);
        let h1 = hasher.finish();
        hasher.write(&[1, 0, 0, 0, 0, 0, 0, 0]);
        let h2 = hasher.finish();
        assert_ne!(h1, h2);
    }

    #[cfg(feature = "compile-time-rng")]
    #[test]
    fn test_builder() {
        use std::collections::HashMap;
        use std::hash::BuildHasherDefault;

        let mut map = HashMap::<u32, u64, BuildHasherDefault<AHasher>>::default();
        map.insert(1, 3);
    }

    #[cfg(feature = "compile-time-rng")]
    #[test]
    fn test_default() {
        let hasher_a = AHasher::default();
        let a_enc: [u64; 2] = hasher_a.enc.convert();
        let a_sum: [u64; 2] = hasher_a.sum.convert();
        assert_ne!(0, a_enc[0]);
        assert_ne!(0, a_enc[1]);
        assert_ne!(0, a_sum[0]);
        assert_ne!(0, a_sum[1]);
        assert_ne!(a_enc[0], a_enc[1]);
        assert_ne!(a_sum[0], a_sum[1]);
        assert_ne!(a_enc[0], a_sum[0]);
        assert_ne!(a_enc[1], a_sum[1]);
        let hasher_b = AHasher::default();
        let b_enc: [u64; 2] = hasher_b.enc.convert();
        let b_sum: [u64; 2] = hasher_b.sum.convert();
        assert_eq!(a_enc[0], b_enc[0]);
        assert_eq!(a_enc[1], b_enc[1]);
        assert_eq!(a_sum[0], b_sum[0]);
        assert_eq!(a_sum[1], b_sum[1]);
    }

    #[test]
    fn test_hash() {
        let mut result: [u64; 2] = [0x6c62272e07bb0142, 0x62b821756295c58d];
        let value: [u64; 2] = [1 << 32, 0xFEDCBA9876543210];
        result = aesenc(value.convert(), result.convert()).convert();
        result = aesenc(result.convert(), result.convert()).convert();
        let mut result2: [u64; 2] = [0x6c62272e07bb0142, 0x62b821756295c58d];
        let value2: [u64; 2] = [1, 0xFEDCBA9876543210];
        result2 = aesenc(value2.convert(), result2.convert()).convert();
        result2 = aesenc(result2.convert(), result.convert()).convert();
        let result: [u8; 16] = result.convert();
        let result2: [u8; 16] = result2.convert();
        assert_ne!(hex::encode(result), hex::encode(result2));
    }

    #[test]
    fn test_conversion() {
        let input: &[u8] = "dddddddd".as_bytes();
        let bytes: u64 = as_array!(input, 8).convert();
        assert_eq!(bytes, 0x6464646464646464);
    }
}