sigmah 0.5.0

Create and scan binary signature in Rust efficiently
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

use core::simd::{cmp::SimdPartialEq, LaneCount, Mask, Simd, SupportedLaneCount};

use crate::utils::{pad_zeroes_slice_unchecked, simd::SimdBits};
use multiversion::multiversion;

/// Computes the result with the given the formula:
/// ```
/// pat:    [1, 2, 3, 4]
/// data:   [1, 1, 3, 3]
/// mask:   [1, 0, 1, 1] # 1 indicates equal, 0 indicates optional/don't care
/// result: [1, 1, 1, 0]
/// ```
///
/// This is an alternative algorithm for result:
/// ```
/// for i in range(len(pat)):
///   result[i] = mask[i] ? pat[i] == data[i] : 1
/// all(matches(x, true) for x in result)
/// ```
#[inline(always)]
#[multiversion(targets(
    "x86_64+avx512vl",
    "x86_64+avx2",
    "x86_64+avx",
    "x86_64+sse2",
    "x86_64+sse",
    "x86+avx512bw",
    "x86+avx2",
    "x86+avx",
    "x86+sse2",
    "x86+sse",
    "aarch64+sve2",
    "aarch64+sve",
    "aarch64+neon",
    "arm+neon",
    "arm+vfp4",
    "arm+vfp3",
    "arm+vfp2",
))]
pub fn match_simd_select_core<const N: usize>(
    data: [u8; N],
    pattern: [u8; N],
    mask: impl Into<u64>,
) -> bool
where
    LaneCount<N>: SupportedLaneCount,
{
    Mask::from_bitmask(mask.into())
        .select_mask(
            Simd::from_array(data).simd_eq(Simd::from_array(pattern)),
            Mask::from_bitmask(u64::MAX),
        )
        .all()
}

#[inline(always)]
#[multiversion(targets(
    "x86_64+avx512bw",
    "x86_64+avx2",
    "x86_64+avx",
    "x86_64+sse2",
    "x86_64+sse",
    "x86+avx512bw",
    "x86+avx2",
    "x86+avx",
    "x86+sse2",
    "x86+sse",
    "aarch64+sve2",
    "aarch64+sve",
    "aarch64+neon",
    "arm+neon",
    "arm+vfp4",
    "arm+vfp3",
    "arm+vfp2",
))]
pub fn match_simd_core<const N: usize>(
    data: [u8; N],
    pattern: [u8; N],
    mask: impl Into<u64>,
) -> bool
where
    LaneCount<N>: SupportedLaneCount,
{
    (!Mask::from_bitmask(mask.into()) | Simd::from_array(data).simd_eq(Simd::from_array(pattern)))
        .all()
}

#[inline(always)]
#[multiversion(targets(
    "x86_64+avx512vl",
    "x86_64+avx2",
    "x86_64+avx",
    "x86_64+sse2",
    "x86_64+sse",
    "x86+avx512bw",
    "x86+avx2",
    "x86+avx",
    "x86+sse2",
    "x86+sse",
    "aarch64+sve2",
    "aarch64+sve",
    "aarch64+neon",
    "arm+neon",
    "arm+vfp4",
    "arm+vfp3",
    "arm+vfp2",
))]
pub fn equal_then_find_first_position_simd_core<const N: usize>(
    first: Simd<u8, N>,
    window: [u8; N],
) -> Option<usize>
where
    LaneCount<N>: SupportedLaneCount,
{
    Simd::from_array(window).simd_eq(first).first_set()
}

#[inline(always)]
#[multiversion(targets(
    "x86_64+avx512vl",
    "x86_64+avx2",
    "x86_64+avx",
    "x86_64+sse2",
    "x86_64+sse",
    "x86+avx512bw",
    "x86+avx2",
    "x86+avx",
    "x86+sse2",
    "x86+sse",
    "aarch64+sve2",
    "aarch64+sve",
    "aarch64+neon",
    "arm+neon",
    "arm+vfp4",
    "arm+vfp3",
    "arm+vfp2",
))]
pub fn equal_then_find_first_position_simd<T: SimdBits>(first: u8, window: &[u8]) -> Option<usize>
where
    LaneCount<{ T::LANES }>: SupportedLaneCount,
{
    let first_splat = Simd::splat(first);
    window
        .chunks(T::LANES)
        .map(|x| unsafe { pad_zeroes_slice_unchecked(x) })
        .enumerate()
        .filter_map(|(stride, window)| {
            equal_then_find_first_position_simd_core(first_splat, window)
                .map(|position| T::LANES * stride + position)
        })
        .next()
}

#[inline(always)]
#[multiversion(targets(
    "x86_64+avx512vl",
    "x86_64+avx2",
    "x86_64+avx",
    "x86_64+sse2",
    "x86_64+sse",
    "x86+avx512bw",
    "x86+avx2",
    "x86+avx",
    "x86+sse2",
    "x86+sse",
    "aarch64+sve2",
    "aarch64+sve",
    "aarch64+neon",
    "arm+neon",
    "arm+vfp4",
    "arm+vfp3",
    "arm+vfp2",
))]
pub fn equal_then_find_second_position_simd_core<const N: usize>(
    first: Simd<u8, N>,
    window: [u8; N],
) -> Option<usize>
where
    LaneCount<N>: SupportedLaneCount,
{
    (Simd::from_array(window).simd_eq(first) & Mask::from_bitmask((i64::MAX - 1) as u64))
        .first_set()
}

#[inline(always)]
#[multiversion(targets(
    "x86_64+avx512vl",
    "x86_64+avx2",
    "x86_64+avx",
    "x86_64+sse2",
    "x86_64+sse",
    "x86+avx512bw",
    "x86+avx2",
    "x86+avx",
    "x86+sse2",
    "x86+sse",
    "aarch64+sve2",
    "aarch64+sve",
    "aarch64+neon",
    "arm+neon",
    "arm+vfp4",
    "arm+vfp3",
    "arm+vfp2",
))]
pub fn equal_then_find_second_position_simd_core_simd<T: SimdBits>(
    first: u8,
    window: &[u8],
) -> Option<usize>
where
    LaneCount<{ T::LANES }>: SupportedLaneCount,
{
    let first_splat = Simd::splat(first);
    window
        .chunks(T::LANES)
        .map(|x| unsafe { pad_zeroes_slice_unchecked(x) })
        .enumerate()
        .filter_map(|(stride, window)| {
            equal_then_find_second_position_simd_core(first_splat, window)
                .map(|position| T::LANES * stride + position)
        })
        .next()
}