rscrypto 0.1.1

Pure Rust cryptography, hardware-accelerated: BLAKE3, SHA-2/3, AES-GCM, ChaCha20-Poly1305, Ed25519, X25519, HMAC, HKDF, Argon2, CRC. no_std, WASM, ten CPU architectures.
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

//! SHA-256 x86_64 SHA-NI kernel.
//!
//! Uses Intel SHA Extensions (`_mm_sha256rnds2_epi32`, `_mm_sha256msg1_epi32`,
//! `_mm_sha256msg2_epi32`) for hardware-accelerated compression.
//!
//! # Safety
//!
//! All functions require `sha` and `sse4.1` target features.
//! Callers must verify CPU capabilities before calling.

#![allow(unsafe_code)]
#![allow(clippy::inline_always)]

#[cfg(target_arch = "x86_64")]
use core::arch::x86_64::*;

/// Round constants K[0..64] packed as 16 groups of 4 × u32, pre-arranged
/// for direct `_mm_add_epi32` with message words.
#[cfg(target_arch = "x86_64")]
// SAFETY: [u32; 4] and __m128i have identical size/alignment; transmute is sound.
static K4: [__m128i; 16] = unsafe {
  core::mem::transmute([
    [0x428a2f98u32, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5],
    [0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5],
    [0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3],
    [0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174],
    [0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc],
    [0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da],
    [0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7],
    [0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967],
    [0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13],
    [0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85],
    [0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3],
    [0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070],
    [0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5],
    [0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3],
    [0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208],
    [0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2],
  ])
};

/// Byte-swap mask for converting little-endian loads to big-endian message words.
#[cfg(target_arch = "x86_64")]
// SAFETY: [u8; 16] and __m128i have identical size; transmute is sound.
static BSWAP_MASK: __m128i = unsafe { core::mem::transmute([3u8, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12]) };

/// SHA-256 multi-block compression using SHA-NI instructions.
///
/// Processes one or more consecutive 64-byte blocks, updating the 8-word state
/// in-place. State lives in two XMM registers (ABEF, CDGH) across blocks — no
/// reload between blocks.
///
/// # Safety
///
/// Caller must ensure `sha` and `sse4.1` CPU features are available.
#[cfg(target_arch = "x86_64")]
#[target_feature(enable = "sha,sse4.1")]
pub(crate) unsafe fn compress_blocks_sha_ni(state: &mut [u32; 8], blocks: &[u8]) {
  debug_assert_eq!(blocks.len() % 64, 0);
  if blocks.is_empty() {
    return;
  }

  // SAFETY: SHA-NI/SSE4.1 intrinsics are available via this function's #[target_feature] attribute.
  // Pointer arithmetic on `ptr` is bounded by `blocks.len()`.
  unsafe {
    // SHA-NI state layout:
    //   ABEF = [A, B, E, F]  (lanes 3,2,1,0 after shuffle)
    //   CDGH = [C, D, G, H]
    //
    // Load state and rearrange from [A,B,C,D,E,F,G,H] to ABEF/CDGH form.
    let mut abef;
    let mut cdgh;
    {
      // tmp0 = [D, C, B, A], tmp1 = [H, G, F, E]
      let tmp0 = _mm_shuffle_epi32(_mm_loadu_si128(state.as_ptr().cast()), 0xB1);
      let tmp1 = _mm_shuffle_epi32(_mm_loadu_si128(state.as_ptr().add(4).cast()), 0x1B);
      // ABEF = [A, B, E, F], CDGH = [C, D, G, H]
      abef = _mm_alignr_epi8(tmp0, tmp1, 8);
      cdgh = _mm_blend_epi16(tmp1, tmp0, 0xF0);
    }

    let num_blocks = blocks.len() / 64;
    let mut ptr = blocks.as_ptr();

    for _ in 0..num_blocks {
      let abef_save = abef;
      let cdgh_save = cdgh;

      // Load and byte-swap 4 message vectors (16 words = 1 block).
      let mut msg0 = _mm_shuffle_epi8(_mm_loadu_si128(ptr.cast()), BSWAP_MASK);
      let mut msg1 = _mm_shuffle_epi8(_mm_loadu_si128(ptr.add(16).cast()), BSWAP_MASK);
      let mut msg2 = _mm_shuffle_epi8(_mm_loadu_si128(ptr.add(32).cast()), BSWAP_MASK);
      let mut msg3 = _mm_shuffle_epi8(_mm_loadu_si128(ptr.add(48).cast()), BSWAP_MASK);

      // Rounds 0-3
      let mut tmp = _mm_add_epi32(msg0, K4[0]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);

      // Rounds 4-7
      tmp = _mm_add_epi32(msg1, K4[1]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg0 = _mm_sha256msg1_epu32(msg0, msg1);

      // Rounds 8-11
      tmp = _mm_add_epi32(msg2, K4[2]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg1 = _mm_sha256msg1_epu32(msg1, msg2);

      // Rounds 12-15
      tmp = _mm_add_epi32(msg3, K4[3]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg0 = _mm_add_epi32(msg0, _mm_alignr_epi8(msg3, msg2, 4));
      msg0 = _mm_sha256msg2_epu32(msg0, msg3);
      msg2 = _mm_sha256msg1_epu32(msg2, msg3);

      // Rounds 16-19
      tmp = _mm_add_epi32(msg0, K4[4]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg1 = _mm_add_epi32(msg1, _mm_alignr_epi8(msg0, msg3, 4));
      msg1 = _mm_sha256msg2_epu32(msg1, msg0);
      msg3 = _mm_sha256msg1_epu32(msg3, msg0);

      // Rounds 20-23
      tmp = _mm_add_epi32(msg1, K4[5]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg2 = _mm_add_epi32(msg2, _mm_alignr_epi8(msg1, msg0, 4));
      msg2 = _mm_sha256msg2_epu32(msg2, msg1);
      msg0 = _mm_sha256msg1_epu32(msg0, msg1);

      // Rounds 24-27
      tmp = _mm_add_epi32(msg2, K4[6]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg3 = _mm_add_epi32(msg3, _mm_alignr_epi8(msg2, msg1, 4));
      msg3 = _mm_sha256msg2_epu32(msg3, msg2);
      msg1 = _mm_sha256msg1_epu32(msg1, msg2);

      // Rounds 28-31
      tmp = _mm_add_epi32(msg3, K4[7]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg0 = _mm_add_epi32(msg0, _mm_alignr_epi8(msg3, msg2, 4));
      msg0 = _mm_sha256msg2_epu32(msg0, msg3);
      msg2 = _mm_sha256msg1_epu32(msg2, msg3);

      // Rounds 32-35
      tmp = _mm_add_epi32(msg0, K4[8]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg1 = _mm_add_epi32(msg1, _mm_alignr_epi8(msg0, msg3, 4));
      msg1 = _mm_sha256msg2_epu32(msg1, msg0);
      msg3 = _mm_sha256msg1_epu32(msg3, msg0);

      // Rounds 36-39
      tmp = _mm_add_epi32(msg1, K4[9]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg2 = _mm_add_epi32(msg2, _mm_alignr_epi8(msg1, msg0, 4));
      msg2 = _mm_sha256msg2_epu32(msg2, msg1);
      msg0 = _mm_sha256msg1_epu32(msg0, msg1);

      // Rounds 40-43
      tmp = _mm_add_epi32(msg2, K4[10]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg3 = _mm_add_epi32(msg3, _mm_alignr_epi8(msg2, msg1, 4));
      msg3 = _mm_sha256msg2_epu32(msg3, msg2);
      msg1 = _mm_sha256msg1_epu32(msg1, msg2);

      // Rounds 44-47
      tmp = _mm_add_epi32(msg3, K4[11]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg0 = _mm_add_epi32(msg0, _mm_alignr_epi8(msg3, msg2, 4));
      msg0 = _mm_sha256msg2_epu32(msg0, msg3);
      msg2 = _mm_sha256msg1_epu32(msg2, msg3);

      // Rounds 48-51
      tmp = _mm_add_epi32(msg0, K4[12]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg1 = _mm_add_epi32(msg1, _mm_alignr_epi8(msg0, msg3, 4));
      msg1 = _mm_sha256msg2_epu32(msg1, msg0);
      msg3 = _mm_sha256msg1_epu32(msg3, msg0);

      // Rounds 52-55
      tmp = _mm_add_epi32(msg1, K4[13]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg2 = _mm_add_epi32(msg2, _mm_alignr_epi8(msg1, msg0, 4));
      msg2 = _mm_sha256msg2_epu32(msg2, msg1);

      // Rounds 56-59
      tmp = _mm_add_epi32(msg2, K4[14]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);
      msg3 = _mm_add_epi32(msg3, _mm_alignr_epi8(msg2, msg1, 4));
      msg3 = _mm_sha256msg2_epu32(msg3, msg2);

      // Rounds 60-63
      tmp = _mm_add_epi32(msg3, K4[15]);
      cdgh = _mm_sha256rnds2_epu32(cdgh, abef, tmp);
      tmp = _mm_shuffle_epi32(tmp, 0x0E);
      abef = _mm_sha256rnds2_epu32(abef, cdgh, tmp);

      // Add saved state back.
      abef = _mm_add_epi32(abef, abef_save);
      cdgh = _mm_add_epi32(cdgh, cdgh_save);

      ptr = ptr.add(64);
    }

    // Store state back: convert from ABEF/CDGH to [A,B,C,D,E,F,G,H].
    let tmp0 = _mm_shuffle_epi32(abef, 0x1B);
    let tmp1 = _mm_shuffle_epi32(cdgh, 0xB1);
    _mm_storeu_si128(state.as_mut_ptr().cast(), _mm_blend_epi16(tmp0, tmp1, 0xF0));
    _mm_storeu_si128(state.as_mut_ptr().add(4).cast(), _mm_alignr_epi8(tmp1, tmp0, 8));
  } // unsafe
}