pqc_kyber 0.2.1

A rust implementation of the post-quantum Kyber KEM algorithm
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

#![allow(dead_code)]

use core::arch::x86_64::*;
use crate::fips202::*;
use crate::keccak4x::f1600_x4;
use crate::align::{GenMatrixBuf, Eta4xBuf};

#[repr(C)]
pub struct Keccakx4State {
  s: [__m256i; 25]
}

impl Keccakx4State {
  pub fn new() -> Self {
    unsafe {Keccakx4State { s: [_mm256_setzero_si256(); 25]}}
  }
}

pub unsafe fn keccakx4_absorb_once(
  s: &mut[__m256i; 25],
  r: usize,
  in0: &[u8],
  in1: &[u8],
  in2: &[u8],
  in3: &[u8],
  mut inlen: usize,
  p: u8
)
{
  let mut pos = 0i64;
  let mut t;
  for i in 0..25 {
    s[i] = _mm256_setzero_si256();
  }
  let mut idx = _mm256_set_epi64x(
    in3.as_ptr() as i64,
    in2.as_ptr() as i64,
    in1.as_ptr() as i64,
    in0.as_ptr() as i64,
  );
  while inlen >= r {
    for i in 0..(r/8) {
      t = _mm256_i64gather_epi64(pos as *const i64, idx, 1);
      s[i] = _mm256_xor_si256(s[i], t);
      pos += 8;
    }
    inlen -= r;
    f1600_x4(s);
  }
  let end = inlen/8;
  for i in 0..end {
    t = _mm256_i64gather_epi64(pos as *const i64, idx, 1);
    s[i] = _mm256_xor_si256(s[i], t);
    pos += 8;
  }
  inlen -= 8*end;

  if inlen > 0 {
    t = _mm256_i64gather_epi64(pos  as *const i64, idx, 1);
    idx = _mm256_set1_epi64x(((1u64 << (8*inlen)) - 1) as i64);
    t = _mm256_and_si256(t, idx);
    s[end] = _mm256_xor_si256(s[end], t);
  }

  t = _mm256_set1_epi64x(((p as u64) << 8*inlen) as i64);
  s[end] = _mm256_xor_si256(s[end], t);
  t = _mm256_set1_epi64x((1u64 << 63) as i64);
  s[r/8 - 1] = _mm256_xor_si256(s[r/8 - 1], t);
}

pub unsafe fn keccakx4_squeezeblocks128(
  out: &mut [GenMatrixBuf; 4],
  mut nblocks: usize,
  r: usize,
  s: &mut [__m256i; 25]
)
{
  let mut t;
  let mut idx = 0usize;
  while nblocks > 0 {
    f1600_x4(s);
    for i in 0..(r/8) {
      t = _mm_castsi128_pd(_mm256_castsi256_si128(s[i]));
      let out0_ptr = out[0].coeffs[idx+8*i..].as_mut_ptr();
      let out1_ptr = out[1].coeffs[idx+8*i..].as_mut_ptr();
      _mm_storel_pd(out0_ptr as *mut f64, t);
      _mm_storeh_pd(out1_ptr as *mut f64, t);
      
      t = _mm_castsi128_pd(_mm256_extracti128_si256(s[i],1));
      let out2_ptr = out[2].coeffs[idx+8*i..].as_mut_ptr();
      let out3_ptr = out[3].coeffs[idx+8*i..].as_mut_ptr();
      _mm_storel_pd(out2_ptr as *mut f64, t);
      _mm_storeh_pd(out3_ptr as *mut f64, t);
    }
    idx += r;
    nblocks -= 1;
  }
}

pub unsafe fn keccakx4_squeezeblocks256(
  out: &mut [Eta4xBuf; 4],
  mut nblocks: usize,
  r: usize,
  s: &mut [__m256i; 25]
)
{
  let mut t;
  let mut idx = 0usize;
  while nblocks > 0 {
    f1600_x4(s);
    for i in 0..(r/8) {
      t = _mm_castsi128_pd(_mm256_castsi256_si128(s[i]));
      _mm_storel_pd(out[0].coeffs[idx+8*i..].as_mut_ptr() as *mut f64, t);
      _mm_storeh_pd(out[1].coeffs[idx+8*i..].as_mut_ptr() as *mut f64, t);
      t = _mm_castsi128_pd(_mm256_extracti128_si256(s[i],1));
      _mm_storel_pd(out[2].coeffs[idx+8*i..].as_mut_ptr() as *mut f64, t);
      _mm_storeh_pd(out[3].coeffs[idx+8*i..].as_mut_ptr() as *mut f64, t);
    }
    idx += r;
    nblocks -= 1;
  }
}

// pub unsafe fn keccakx4_squeezeonce128(
//   out: &mut [[u8; 168]; 4],
//   s: &mut [__m256i; 25]
// )
// {
//   let mut t;
//   f1600_x4(s);
//   for i in 0..(SHAKE128_RATE/8) {
//     t = _mm_castsi128_pd(_mm256_castsi256_si128(s[i]));
//     let out0_ptr = out[0][8*i] as *mut f64;
//     _mm_storel_pd(out0_ptr, t);
//     _mm_storeh_pd(out[1][8*i] as *mut f64, t);
//     t = _mm_castsi128_pd(_mm256_extracti128_si256(s[i],1));
//     _mm_storel_pd(out[2][8*i] as *mut f64, t);
//     _mm_storeh_pd(out[3][8*i] as *mut f64, t);
//   }
// }

// pub unsafe fn keccakx4_squeezeonce256(
//   out: &mut [[u8; 136]; 4],
//   s: &mut [__m256i; 25]
// )
// {
//   let mut t;
//   f1600_x4(s);
//   for i in 0..(SHAKE256_RATE/8) {
//     t = _mm_castsi128_pd(_mm256_castsi256_si128(s[i]));
//     _mm_storel_pd(out[0][8*i] as *mut f64, t);
//     _mm_storeh_pd(out[1][8*i] as *mut f64, t);
//     t = _mm_castsi128_pd(_mm256_extracti128_si256(s[i],1));
//     _mm_storel_pd(out[2][8*i] as *mut f64, t);
//     _mm_storeh_pd(out[3][8*i] as *mut f64, t);
//   }
// }

pub unsafe fn shake128x4_absorb_once(
  state: &mut Keccakx4State,
  in0: &[u8],
  in1: &[u8],
  in2: &[u8],
  in3: &[u8],
  inlen: usize,
)
{
  keccakx4_absorb_once(
    &mut state.s, 
    SHAKE128_RATE, 
    in0, in1, in2, in3, inlen, 
    0x1F
  )
}

pub unsafe fn shake128x4_squeezeblocks(
  out: &mut[GenMatrixBuf; 4], 
  nblocks: usize,
  state: &mut Keccakx4State
)
{
  keccakx4_squeezeblocks128(
    out, 
    nblocks, 
    SHAKE128_RATE, 
    &mut state.s
  );
}

pub unsafe fn shake256x4_absorb_once(
  state: &mut Keccakx4State,
  in0: &[u8],
  in1: &[u8],
  in2: &[u8],
  in3: &[u8],
  inlen: usize,
)
{
  keccakx4_absorb_once(
    &mut state.s, 
    SHAKE256_RATE, 
    in0, in1, in2, in3, inlen, 
    0x1F
  )
}

pub unsafe fn shake256x4_squeezeblocks(
  out: &mut[Eta4xBuf; 4],
  nblocks: usize,
  state: &mut Keccakx4State
)
{
  keccakx4_squeezeblocks256(
    out, 
    nblocks, 
    SHAKE256_RATE, 
    &mut state.s
  );
}

// pub unsafe fn shake128x4(
//   out: &mut [GenMatrixBuf; 4],
//   mut outlen: usize,
//   in0: &[u8],
//   in1: &[u8],
//   in2: &[u8],
//   in3: &[u8],
//   inlen: usize
// )
// {
//   let nblocks = outlen/SHAKE128_RATE;
//   let mut t = [[0u8; SHAKE128_RATE]; 4];
//   let mut state = Keccakx4State::new();

//   shake128x4_absorb_once(&mut state, in0, in1, in2, in3, inlen);
//   shake128x4_squeezeblocks(out, nblocks, &mut state);
//   let idx = nblocks*SHAKE128_RATE; 
//   outlen -= idx;

//   if outlen > 0 {
//     keccakx4_squeezeonce128(&mut t, &mut state.s);
//     for i in 0..outlen {
//       out[0].coeffs[idx+i] = t[0][i];
//       out[1].coeffs[idx+i] = t[1][i];
//       out[2].coeffs[idx+i] = t[2][i];
//       out[3].coeffs[idx+i] = t[3][i];
//     }
//   }
// }

// pub unsafe fn shake256x4(
//   out: &mut [Eta4xBuf; 4],
//   mut outlen: usize,
//   in0: &[u8],
//   in1: &[u8],
//   in2: &[u8],
//   in3: &[u8],
//   inlen: usize
// )
// {
//   let nblocks = outlen/SHAKE256_RATE;
//   let mut t = [[0u8; SHAKE256_RATE] ; 4];
//   let mut state = Keccakx4State::new();

//   shake256x4_absorb_once(&mut state, in0, in1, in2, in3, inlen);
//   shake256x4_squeezeblocks(out, nblocks, &mut state);

//   let idx = nblocks*SHAKE256_RATE;
//   outlen -= nblocks*SHAKE256_RATE;

//   if outlen > 0 {
//     keccakx4_squeezeonce256(&mut t, &mut state.s);
//     for i in 0..outlen {
//       out[0].coeffs[idx+i] = t[0][i];
//       out[1].coeffs[idx+i] = t[1][i];
//       out[2].coeffs[idx+i] = t[2][i];
//       out[3].coeffs[idx+i] = t[3][i];
//     }
//   }
// }