openipc-core 0.1.33

Shared OpenIPC FPV packet, RTP, and Realtek RX parsing logic
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

//! Architecture-specific Galois-field multiply/XOR loops.
//!
//! Callers provide two 16-byte lookup tables containing the products for the
//! low and high nibble of each source byte. Bounds checks and CPU feature
//! selection stay in safe Rust; only the vector loads and stores are unsafe.

const LANES: usize = 16;

#[inline(always)]
pub(crate) fn addmul(dst: &mut [u8], src: &[u8], low: &[u8; LANES], high: &[u8; LANES]) -> usize {
    let len = dst.len().min(src.len());

    #[cfg(target_arch = "aarch64")]
    let processed = {
        // NEON is part of the AArch64 baseline.
        unsafe { aarch64::addmul(&mut dst[..len], &src[..len], low, high) }
    };

    #[cfg(any(target_arch = "x86", target_arch = "x86_64"))]
    let processed = {
        if std::arch::is_x86_feature_detected!("ssse3") {
            unsafe { x86::addmul(&mut dst[..len], &src[..len], low, high) }
        } else {
            0
        }
    };

    #[cfg(all(target_arch = "wasm32", target_feature = "simd128"))]
    let processed = unsafe { wasm::addmul(&mut dst[..len], &src[..len], low, high) };

    #[cfg(not(any(
        target_arch = "aarch64",
        target_arch = "x86",
        target_arch = "x86_64",
        all(target_arch = "wasm32", target_feature = "simd128")
    )))]
    let processed = 0;

    processed
}

#[cfg(target_arch = "aarch64")]
mod aarch64 {
    use core::arch::aarch64::*;

    pub(super) unsafe fn addmul(
        dst: &mut [u8],
        src: &[u8],
        low: &[u8; 16],
        high: &[u8; 16],
    ) -> usize {
        let vector_len = dst.len() & !15;
        let unrolled_len = dst.len() & !63;
        let low_table = vld1q_u8(low.as_ptr());
        let high_table = vld1q_u8(high.as_ptr());
        let nibble_mask = vdupq_n_u8(0x0f);

        for base in (0..unrolled_len).step_by(64) {
            for offset in [base, base + 16, base + 32, base + 48] {
                let input = vld1q_u8(src.as_ptr().add(offset));
                let low_product = vqtbl1q_u8(low_table, vandq_u8(input, nibble_mask));
                let high_product = vqtbl1q_u8(high_table, vshrq_n_u8(input, 4));
                let output = vld1q_u8(dst.as_ptr().add(offset));
                vst1q_u8(
                    dst.as_mut_ptr().add(offset),
                    veorq_u8(output, veorq_u8(low_product, high_product)),
                );
            }
        }
        for offset in (unrolled_len..vector_len).step_by(16) {
            let input = vld1q_u8(src.as_ptr().add(offset));
            let product = veorq_u8(
                vqtbl1q_u8(low_table, vandq_u8(input, nibble_mask)),
                vqtbl1q_u8(high_table, vshrq_n_u8(input, 4)),
            );
            let output = vld1q_u8(dst.as_ptr().add(offset));
            vst1q_u8(dst.as_mut_ptr().add(offset), veorq_u8(output, product));
        }

        vector_len
    }
}

#[cfg(any(target_arch = "x86", target_arch = "x86_64"))]
mod x86 {
    #[cfg(target_arch = "x86")]
    use core::arch::x86::*;
    #[cfg(target_arch = "x86_64")]
    use core::arch::x86_64::*;

    #[target_feature(enable = "ssse3")]
    pub(super) unsafe fn addmul(
        dst: &mut [u8],
        src: &[u8],
        low: &[u8; 16],
        high: &[u8; 16],
    ) -> usize {
        let vector_len = dst.len() & !15;
        let unrolled_len = dst.len() & !63;
        let low_table = _mm_loadu_si128(low.as_ptr().cast());
        let high_table = _mm_loadu_si128(high.as_ptr().cast());
        let nibble_mask = _mm_set1_epi8(0x0f);

        for base in (0..unrolled_len).step_by(64) {
            for offset in [base, base + 16, base + 32, base + 48] {
                let input = _mm_loadu_si128(src.as_ptr().add(offset).cast());
                let low_product = _mm_shuffle_epi8(low_table, _mm_and_si128(input, nibble_mask));
                let high_indexes = _mm_and_si128(_mm_srli_epi16(input, 4), nibble_mask);
                let high_product = _mm_shuffle_epi8(high_table, high_indexes);
                let output = _mm_loadu_si128(dst.as_ptr().add(offset).cast());
                _mm_storeu_si128(
                    dst.as_mut_ptr().add(offset).cast(),
                    _mm_xor_si128(output, _mm_xor_si128(low_product, high_product)),
                );
            }
        }
        for offset in (unrolled_len..vector_len).step_by(16) {
            let input = _mm_loadu_si128(src.as_ptr().add(offset).cast());
            let low_product = _mm_shuffle_epi8(low_table, _mm_and_si128(input, nibble_mask));
            let high_indexes = _mm_and_si128(_mm_srli_epi16(input, 4), nibble_mask);
            let high_product = _mm_shuffle_epi8(high_table, high_indexes);
            let output = _mm_loadu_si128(dst.as_ptr().add(offset).cast());
            _mm_storeu_si128(
                dst.as_mut_ptr().add(offset).cast(),
                _mm_xor_si128(output, _mm_xor_si128(low_product, high_product)),
            );
        }

        vector_len
    }
}

#[cfg(all(target_arch = "wasm32", target_feature = "simd128"))]
mod wasm {
    use core::arch::wasm32::*;

    pub(super) unsafe fn addmul(
        dst: &mut [u8],
        src: &[u8],
        low: &[u8; 16],
        high: &[u8; 16],
    ) -> usize {
        let vector_len = dst.len() & !15;
        let unrolled_len = dst.len() & !63;
        let low_table = v128_load(low.as_ptr().cast());
        let high_table = v128_load(high.as_ptr().cast());
        let nibble_mask = u8x16_splat(0x0f);

        for base in (0..unrolled_len).step_by(64) {
            for offset in [base, base + 16, base + 32, base + 48] {
                let input = v128_load(src.as_ptr().add(offset).cast());
                let low_product = i8x16_swizzle(low_table, v128_and(input, nibble_mask));
                let high_product = i8x16_swizzle(high_table, u8x16_shr(input, 4));
                let output = v128_load(dst.as_ptr().add(offset).cast());
                v128_store(
                    dst.as_mut_ptr().add(offset).cast(),
                    v128_xor(output, v128_xor(low_product, high_product)),
                );
            }
        }
        for offset in (unrolled_len..vector_len).step_by(16) {
            let input = v128_load(src.as_ptr().add(offset).cast());
            let low_product = i8x16_swizzle(low_table, v128_and(input, nibble_mask));
            let high_product = i8x16_swizzle(high_table, u8x16_shr(input, 4));
            let output = v128_load(dst.as_ptr().add(offset).cast());
            v128_store(
                dst.as_mut_ptr().add(offset).cast(),
                v128_xor(output, v128_xor(low_product, high_product)),
            );
        }

        vector_len
    }
}