gf256 0.2.0

A Rust library containing Galois-field types and utilities
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
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

//! Hardware xmul implementations if available
//!
//! These are declared here in order to be able to leverage unstable
//! features on nightly (if the feature nightly-features is provided).
//! Most of gf256 is provided as proc_macros, and those can't use unstable
//! features unless the feature is enabled with #[feature!] at the crate
//! level.
//!
//! These functions are intended to only be used by gf256's proc_macros,
//! these funcitons may or may not be available depending on target_features,
//! and may change behavior, so they shouldn't be used directly.
//!

use cfg_if::cfg_if;


/// True if carry-less multiplication instructions are available
///
/// If this is false, any carry-less multiplication operations
/// will use a more expensive bitwise implementation.
///
/// Some algorithms trade expensive division/remainder operations for
/// multiple multiplication operations, but this can backfire if
/// multiplication is also expensive. This flag allows algorithms
/// to choose the best strategy based on what's available.
///
pub const HAS_XMUL: bool = {
    cfg_if! {
        if #[cfg(any(
            all(
                not(feature="no-xmul"),
                target_arch="x86_64",
                target_feature="pclmulqdq"
            ),
            all(
                not(feature="no-xmul"),
                feature="nightly",
                target_arch="aarch64",
                target_feature="neon"
            )
        ))] {
            true
        } else {
            false
        }
    }
};


/// Widening carry-less multiplication, if hardware instructions are available
///
/// Result is a tuple (lo, hi)
///
#[cfg(any(
    all(
        not(feature="no-xmul"),
        target_arch="x86_64",
        target_feature="pclmulqdq"
    ),
    all(
        not(feature="no-xmul"),
        feature="nightly",
        target_arch="aarch64",
        target_feature="neon"
    )
))]
#[inline]
pub fn xmul8(a: u8, b: u8) -> (u8, u8) {
    cfg_if! {
        if #[cfg(all(
            not(feature="no-xmul"),
            target_arch="x86_64",
            target_feature="pclmulqdq"
        ))] {
            // x86_64 provides 64-bit xmul via the pclmulqdq instruction
            use core::arch::x86_64::*;
            unsafe {
                let a = _mm_set_epi64x(0, a as i64);
                let b = _mm_set_epi64x(0, b as i64);
                let x = _mm_clmulepi64_si128::<0>(a, b);
                let lo = _mm_extract_epi64::<0>(x) as u64;
                (lo as u8, (lo >> 8) as u8)
            }
        } else if #[cfg(all(
            not(feature="no-xmul"),
            feature="nightly",
            target_arch="aarch64",
            target_feature="neon"
        ))] {
            // aarch64 provides 64-bit xmul via the pmull instruction
            use core::arch::aarch64::*;
            unsafe {
                let x = vmull_p64(a as u64, b as u64);
                (x as u8, (x >> 8) as u8)
            }
        }
    }
}

/// Widening carry-less multiplication, if hardware instructions are available
///
/// Result is a tuple (lo, hi)
///
#[cfg(any(
    all(
        not(feature="no-xmul"),
        target_arch="x86_64",
        target_feature="pclmulqdq"
    ),
    all(
        not(feature="no-xmul"),
        feature="nightly",
        target_arch="aarch64",
        target_feature="neon"
    )
))]
#[inline]
pub fn xmul16(a: u16, b: u16) -> (u16, u16) {
    cfg_if! {
        if #[cfg(all(
            not(feature="no-xmul"),
            target_arch="x86_64",
            target_feature="pclmulqdq"
        ))] {
            // x86_64 provides 64-bit xmul via the pclmulqdq instruction
            use core::arch::x86_64::*;
            unsafe {
                let a = _mm_set_epi64x(0, a as i64);
                let b = _mm_set_epi64x(0, b as i64);
                let x = _mm_clmulepi64_si128::<0>(a, b);
                let lo = _mm_extract_epi64::<0>(x) as u64;
                (lo as u16, (lo >> 16) as u16)
            }
        } else if #[cfg(all(
            not(feature="no-xmul"),
            feature="nightly",
            target_arch="aarch64",
            target_feature="neon"
        ))] {
            // aarch64 provides 64-bit xmul via the pmull instruction
            use core::arch::aarch64::*;
            unsafe {
                let x = vmull_p64(a as u64, b as u64);
                (x as u16, (x >> 16) as u16)
            }
        }
    }
}

/// Widening carry-less multiplication, if hardware instructions are available
///
/// Result is a tuple (lo, hi)
///
#[cfg(any(
    all(
        not(feature="no-xmul"),
        target_arch="x86_64",
        target_feature="pclmulqdq"
    ),
    all(
        not(feature="no-xmul"),
        feature="nightly",
        target_arch="aarch64",
        target_feature="neon"
    )
))]
#[inline]
pub fn xmul32(a: u32, b: u32) -> (u32, u32) {
    cfg_if! {
        if #[cfg(all(
            not(feature="no-xmul"),
            target_arch="x86_64",
            target_feature="pclmulqdq"
        ))] {
            // x86_64 provides 64-bit xmul via the pclmulqdq instruction
            use core::arch::x86_64::*;
            unsafe {
                let a = _mm_set_epi64x(0, a as i64);
                let b = _mm_set_epi64x(0, b as i64);
                let x = _mm_clmulepi64_si128::<0>(a, b);
                let lo = _mm_extract_epi64::<0>(x) as u64;
                (lo as u32, (lo >> 32) as u32)
            }
        } else if #[cfg(all(
            not(feature="no-xmul"),
            feature="nightly",
            target_arch="aarch64",
            target_feature="neon"
        ))] {
            // aarch64 provides 64-bit xmul via the pmull instruction
            use core::arch::aarch64::*;
            unsafe {
                let x = vmull_p64(a as u64, b as u64);
                (x as u32, (x >> 32) as u32)
            }
        }
    }
}

/// Widening carry-less multiplication, if hardware instructions are available
///
/// Result is a tuple (lo, hi)
///
#[cfg(any(
    all(
        not(feature="no-xmul"),
        target_arch="x86_64",
        target_feature="pclmulqdq"
    ),
    all(
        not(feature="no-xmul"),
        feature="nightly",
        target_arch="aarch64",
        target_feature="neon"
    )
))]
#[inline]
pub fn xmul64(a: u64, b: u64) -> (u64, u64) {
    cfg_if! {
        if #[cfg(all(
            not(feature="no-xmul"),
            target_arch="x86_64",
            target_feature="pclmulqdq"
        ))] {
            // x86_64 provides 64-bit xmul via the pclmulqdq instruction
            use core::arch::x86_64::*;
            unsafe {
                let a = _mm_set_epi64x(0, a as i64);
                let b = _mm_set_epi64x(0, b as i64);
                let x = _mm_clmulepi64_si128::<0>(a, b);
                let lo = _mm_extract_epi64::<0>(x) as u64;
                let hi = _mm_extract_epi64::<1>(x) as u64;
                (lo, hi)
            }
        } else if #[cfg(all(
            not(feature="no-xmul"),
            feature="nightly",
            target_arch="aarch64",
            target_feature="neon"
        ))] {
            // aarch64 provides 64-bit xmul via the pmull instruction
            use core::arch::aarch64::*;
            unsafe {
                let x = vmull_p64(a as u64, b as u64);
                (x as u64, (x >> 64) as u64)
            }
        }
    }
}

/// Widening carry-less multiplication, if hardware instructions are available
///
/// Result is a tuple (lo, hi)
///
#[cfg(any(
    all(
        not(feature="no-xmul"),
        target_arch="x86_64",
        target_feature="pclmulqdq"
    ),
    all(
        not(feature="no-xmul"),
        feature="nightly",
        target_arch="aarch64",
        target_feature="neon"
    )
))]
#[inline]
pub fn xmul128(a: u128, b: u128) -> (u128, u128) {
    cfg_if! {
        if #[cfg(all(
            not(feature="no-xmul"),
            target_arch="x86_64",
            target_feature="pclmulqdq"
        ))] {
            // x86_64 provides 64-bit xmul via the pclmulqdq instruction
            use core::arch::x86_64::*;
            unsafe {
                let a = _mm_set_epi64x((a >> 64) as i64, a as i64);
                let b = _mm_set_epi64x((b >> 64) as i64, b as i64);
                let x = _mm_clmulepi64_si128::<0x00>(a, b);
                let y = _mm_clmulepi64_si128::<0x01>(a, b);
                let z = _mm_clmulepi64_si128::<0x10>(a, b);
                let w = _mm_clmulepi64_si128::<0x11>(a, b);
                let lolo = _mm_extract_epi64::<0>(x) as u64;
                let lohi = (_mm_extract_epi64::<1>(x) as u64)
                    ^ (_mm_extract_epi64::<0>(y) as u64)
                    ^ (_mm_extract_epi64::<0>(z) as u64);
                let hilo = (_mm_extract_epi64::<0>(w) as u64)
                    ^ (_mm_extract_epi64::<1>(y) as u64)
                    ^ (_mm_extract_epi64::<1>(z) as u64);
                let hihi = _mm_extract_epi64::<1>(w) as u64;
                let lo = ((lohi as u128) << 64) | (lolo as u128);
                let hi = ((hihi as u128) << 64) | (hilo as u128);
                (lo, hi)
            }
        } else if #[cfg(all(
            not(feature="no-xmul"),
            feature="nightly",
            target_arch="aarch64",
            target_feature="neon"
        ))] {
            // aarch64 provides 64-bit xmul via the pmull instruction
            use core::arch::aarch64::*;
            unsafe {
                let x = vmull_p64(a as u64, b as u64);
                let y = vmull_p64((a >> 64) as u64, (b >>  0) as u64);
                let z = vmull_p64((a >>  0) as u64, (b >> 64) as u64);
                let w = vmull_p64((a >> 64) as u64, (b >> 64) as u64);
                (x ^ (y << 64) ^ (z << 64), w ^ (y >> 64) ^ (z >> 64))
            }
        }
    }
}


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

    #[cfg(any(
        all(
            not(feature="no-xmul"),
            target_arch="x86_64",
            target_feature="pclmulqdq"
        ),
        all(
            not(feature="no-xmul"),
            feature="nightly",
            target_arch="aarch64",
            target_feature="neon"
        )
    ))]
    #[test]
    fn xmul() {
        assert_eq!(xmul8(0x12, 0x12), (0x04, 0x01));
        assert_eq!(xmul16(0x1234, 0x1234), (0x0510, 0x0104));
        assert_eq!(xmul32(0x12345678, 0x12345678), (0x11141540, 0x01040510));
        assert_eq!(xmul64(0x123456789abcdef1, 0x123456789abcdef1), (0x4144455051545501, 0x0104051011141540));
        assert_eq!(xmul128(0x123456789abcdef123456789abcdef12, 0x123456789abcdef123456789abcdef12), (0x04051011141540414445505154550104, 0x01040510111415404144455051545501));
    }
}