forked from minio/sha256-simd
-
Notifications
You must be signed in to change notification settings - Fork 0
/
sha256blockAvx_amd64.s
409 lines (374 loc) · 28.7 KB
/
sha256blockAvx_amd64.s
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
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
//+build !noasm !appengine
// SHA256 implementation for AVX
//
// Minio Cloud Storage, (C) 2016 Minio, Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
//
// This code is based on an Intel White-Paper:
// "Fast SHA-256 Implementations on Intel Architecture Processors"
//
// together with the reference implementation from the following authors:
// James Guilford <james.guilford@intel.com>
// Kirk Yap <kirk.s.yap@intel.com>
// Tim Chen <tim.c.chen@linux.intel.com>
//
// For Golang it has been converted to Plan 9 assembly with the help of
// github.com/minio/asm2plan9s to assemble Intel instructions to their Plan9
// equivalents
//
#include "textflag.h"
#define ROTATE_XS \
MOVOU X4, X15 \
MOVOU X5, X4 \
MOVOU X6, X5 \
MOVOU X7, X6 \
MOVOU X15, X7
// compute s0 four at a time and s1 two at a time
// compute W[-16] + W[-7] 4 at a time
#define FOUR_ROUNDS_AND_SCHED(a, b, c, d, e, f, g, h) \
MOVL e, R13 \ /* y0 = e */
ROLL $18, R13 \ /* y0 = e >> (25-11) */
MOVL a, R14 \ /* y1 = a */
LONG $0x0f41e3c4; WORD $0x04c6 \ // VPALIGNR XMM0,XMM7,XMM6,0x4 /* XTMP0 = W[-7] */
ROLL $23, R14 \ /* y1 = a >> (22-13) */
XORL e, R13 \ /* y0 = e ^ (e >> (25-11)) */
MOVL f, R15 \ /* y2 = f */
ROLL $27, R13 \ /* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
XORL a, R14 \ /* y1 = a ^ (a >> (22-13) */
XORL g, R15 \ /* y2 = f^g */
LONG $0xc4fef9c5 \ // VPADDD XMM0,XMM0,XMM4 /* XTMP0 = W[-7] + W[-16] */
XORL e, R13 \ /* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6) ) */
ANDL e, R15 \ /* y2 = (f^g)&e */
ROLL $21, R14 \ /* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
\ /* */
\ /* compute s0 */
\ /* */
LONG $0x0f51e3c4; WORD $0x04cc \ // VPALIGNR XMM1,XMM5,XMM4,0x4 /* XTMP1 = W[-15] */
XORL a, R14 \ /* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
ROLL $26, R13 \ /* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
XORL g, R15 \ /* y2 = CH = ((f^g)&e)^g */
ROLL $30, R14 \ /* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
ADDL R13, R15 \ /* y2 = S1 + CH */
ADDL _xfer+48(FP), R15 \ /* y2 = k + w + S1 + CH */
MOVL a, R13 \ /* y0 = a */
ADDL R15, h \ /* h = h + S1 + CH + k + w */
\ /* ROTATE_ARGS */
MOVL a, R15 \ /* y2 = a */
LONG $0xd172e9c5; BYTE $0x07 \ // VPSRLD XMM2,XMM1,0x7 /* */
ORL c, R13 \ /* y0 = a|c */
ADDL h, d \ /* d = d + h + S1 + CH + k + w */
ANDL c, R15 \ /* y2 = a&c */
LONG $0xf172e1c5; BYTE $0x19 \ // VPSLLD XMM3,XMM1,0x19 /* */
ANDL b, R13 \ /* y0 = (a|c)&b */
ADDL R14, h \ /* h = h + S1 + CH + k + w + S0 */
LONG $0xdaebe1c5 \ // VPOR XMM3,XMM3,XMM2 /* XTMP1 = W[-15] MY_ROR 7 */
ORL R15, R13 \ /* y0 = MAJ = (a|c)&b)|(a&c) */
ADDL R13, h \ /* h = h + S1 + CH + k + w + S0 + MAJ */
\ /* ROTATE_ARGS */
MOVL d, R13 \ /* y0 = e */
MOVL h, R14 \ /* y1 = a */
ROLL $18, R13 \ /* y0 = e >> (25-11) */
XORL d, R13 \ /* y0 = e ^ (e >> (25-11)) */
MOVL e, R15 \ /* y2 = f */
ROLL $23, R14 \ /* y1 = a >> (22-13) */
LONG $0xd172e9c5; BYTE $0x12 \ // VPSRLD XMM2,XMM1,0x12 /* */
XORL h, R14 \ /* y1 = a ^ (a >> (22-13) */
ROLL $27, R13 \ /* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
XORL f, R15 \ /* y2 = f^g */
LONG $0xd172b9c5; BYTE $0x03 \ // VPSRLD XMM8,XMM1,0x3 /* XTMP4 = W[-15] >> 3 */
ROLL $21, R14 \ /* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
XORL d, R13 \ /* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
ANDL d, R15 \ /* y2 = (f^g)&e */
ROLL $26, R13 \ /* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
LONG $0xf172f1c5; BYTE $0x0e \ // VPSLLD XMM1,XMM1,0xe /* */
XORL h, R14 \ /* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
XORL f, R15 \ /* y2 = CH = ((f^g)&e)^g */
LONG $0xd9efe1c5 \ // VPXOR XMM3,XMM3,XMM1 /* */
ADDL R13, R15 \ /* y2 = S1 + CH */
ADDL _xfer+52(FP), R15 \ /* y2 = k + w + S1 + CH */
ROLL $30, R14 \ /* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
LONG $0xdaefe1c5 \ // VPXOR XMM3,XMM3,XMM2 /* XTMP1 = W[-15] MY_ROR 7 ^ W[-15] MY_ROR */
MOVL h, R13 \ /* y0 = a */
ADDL R15, g \ /* h = h + S1 + CH + k + w */
MOVL h, R15 \ /* y2 = a */
LONG $0xef61c1c4; BYTE $0xc8 \ // VPXOR XMM1,XMM3,XMM8 /* XTMP1 = s0 */
ORL b, R13 \ /* y0 = a|c */
ADDL g, c \ /* d = d + h + S1 + CH + k + w */
ANDL b, R15 \ /* y2 = a&c */
\ /* */
\ /* compute low s1 */
\ /* */
LONG $0xd770f9c5; BYTE $0xfa \ // VPSHUFD XMM2,XMM7,0xfa /* XTMP2 = W[-2] {BBAA} */
ANDL a, R13 \ /* y0 = (a|c)&b */
ADDL R14, g \ /* h = h + S1 + CH + k + w + S0 */
LONG $0xc1fef9c5 \ // VPADDD XMM0,XMM0,XMM1 /* XTMP0 = W[-16] + W[-7] + s0 */
ORL R15, R13 \ /* y0 = MAJ = (a|c)&b)|(a&c) */
ADDL R13, g \ /* h = h + S1 + CH + k + w + S0 + MAJ */
\ /* ROTATE_ARGS */
MOVL c, R13 \ /* y0 = e */
MOVL g, R14 \ /* y1 = a */
ROLL $18, R13 \ /* y0 = e >> (25-11) */
XORL c, R13 \ /* y0 = e ^ (e >> (25-11)) */
ROLL $23, R14 \ /* y1 = a >> (22-13) */
MOVL d, R15 \ /* y2 = f */
XORL g, R14 \ /* y1 = a ^ (a >> (22-13) */
ROLL $27, R13 \ /* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
LONG $0xd272b9c5; BYTE $0x0a \ // VPSRLD XMM8,XMM2,0xa /* XTMP4 = W[-2] >> 10 {BBAA} */
XORL e, R15 \ /* y2 = f^g */
LONG $0xd273e1c5; BYTE $0x13 \ // VPSRLQ XMM3,XMM2,0x13 /* XTMP3 = W[-2] MY_ROR 19 {xBxA} */
XORL c, R13 \ /* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
ANDL c, R15 \ /* y2 = (f^g)&e */
LONG $0xd273e9c5; BYTE $0x11 \ // VPSRLQ XMM2,XMM2,0x11 /* XTMP2 = W[-2] MY_ROR 17 {xBxA} */
ROLL $21, R14 \ /* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
XORL g, R14 \ /* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
XORL e, R15 \ /* y2 = CH = ((f^g)&e)^g */
ROLL $26, R13 \ /* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
LONG $0xd3efe9c5 \ // VPXOR XMM2,XMM2,XMM3 /* */
ADDL R13, R15 \ /* y2 = S1 + CH */
ROLL $30, R14 \ /* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
ADDL _xfer+56(FP), R15 \ /* y2 = k + w + S1 + CH */
LONG $0xc2ef39c5 \ // VPXOR XMM8,XMM8,XMM2 /* XTMP4 = s1 {xBxA} */
MOVL g, R13 \ /* y0 = a */
ADDL R15, f \ /* h = h + S1 + CH + k + w */
MOVL g, R15 \ /* y2 = a */
LONG $0x003942c4; BYTE $0xc2 \ // VPSHUFB XMM8,XMM8,XMM10 /* XTMP4 = s1 {00BA} */
ORL a, R13 \ /* y0 = a|c */
ADDL f, b \ /* d = d + h + S1 + CH + k + w */
ANDL a, R15 \ /* y2 = a&c */
LONG $0xfe79c1c4; BYTE $0xc0 \ // VPADDD XMM0,XMM0,XMM8 /* XTMP0 = {..., ..., W[1], W[0]} */
ANDL h, R13 \ /* y0 = (a|c)&b */
ADDL R14, f \ /* h = h + S1 + CH + k + w + S0 */
\ /* */
\ /* compute high s1 */
\ /* */
LONG $0xd070f9c5; BYTE $0x50 \ // VPSHUFD XMM2,XMM0,0x50 /* XTMP2 = W[-2] {DDCC} */
ORL R15, R13 \ /* y0 = MAJ = (a|c)&b)|(a&c) */
ADDL R13, f \ /* h = h + S1 + CH + k + w + S0 + MAJ */
\ /* ROTATE_ARGS */
MOVL b, R13 \ /* y0 = e */
ROLL $18, R13 \ /* y0 = e >> (25-11) */
MOVL f, R14 \ /* y1 = a */
ROLL $23, R14 \ /* y1 = a >> (22-13) */
XORL b, R13 \ /* y0 = e ^ (e >> (25-11)) */
MOVL c, R15 \ /* y2 = f */
ROLL $27, R13 \ /* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
LONG $0xd272a1c5; BYTE $0x0a \ // VPSRLD XMM11,XMM2,0xa /* XTMP5 = W[-2] >> 10 {DDCC} */
XORL f, R14 \ /* y1 = a ^ (a >> (22-13) */
XORL d, R15 \ /* y2 = f^g */
LONG $0xd273e1c5; BYTE $0x13 \ // VPSRLQ XMM3,XMM2,0x13 /* XTMP3 = W[-2] MY_ROR 19 {xDxC} */
XORL b, R13 \ /* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
ANDL b, R15 \ /* y2 = (f^g)&e */
ROLL $21, R14 \ /* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
LONG $0xd273e9c5; BYTE $0x11 \ // VPSRLQ XMM2,XMM2,0x11 /* XTMP2 = W[-2] MY_ROR 17 {xDxC} */
XORL f, R14 \ /* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
ROLL $26, R13 \ /* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
XORL d, R15 \ /* y2 = CH = ((f^g)&e)^g */
LONG $0xd3efe9c5 \ // VPXOR XMM2,XMM2,XMM3 /* */
ROLL $30, R14 \ /* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
ADDL R13, R15 \ /* y2 = S1 + CH */
ADDL _xfer+60(FP), R15 \ /* y2 = k + w + S1 + CH */
LONG $0xdaef21c5 \ // VPXOR XMM11,XMM11,XMM2 /* XTMP5 = s1 {xDxC} */
MOVL f, R13 \ /* y0 = a */
ADDL R15, e \ /* h = h + S1 + CH + k + w */
MOVL f, R15 \ /* y2 = a */
LONG $0x002142c4; BYTE $0xdc \ // VPSHUFB XMM11,XMM11,XMM12 /* XTMP5 = s1 {DC00} */
ORL h, R13 \ /* y0 = a|c */
ADDL e, a \ /* d = d + h + S1 + CH + k + w */
ANDL h, R15 \ /* y2 = a&c */
LONG $0xe0fea1c5 \ // VPADDD XMM4,XMM11,XMM0 /* X0 = {W[3], W[2], W[1], W[0]} */
ANDL g, R13 \ /* y0 = (a|c)&b */
ADDL R14, e \ /* h = h + S1 + CH + k + w + S0 */
ORL R15, R13 \ /* y0 = MAJ = (a|c)&b)|(a&c) */
ADDL R13, e \ /* h = h + S1 + CH + k + w + S0 + MAJ */
\ /* ROTATE_ARGS */
ROTATE_XS
#define DO_ROUND(a, b, c, d, e, f, g, h, offset) \
MOVL e, R13 \ /* y0 = e */
ROLL $18, R13 \ /* y0 = e >> (25-11) */
MOVL a, R14 \ /* y1 = a */
XORL e, R13 \ /* y0 = e ^ (e >> (25-11)) */
ROLL $23, R14 \ /* y1 = a >> (22-13) */
MOVL f, R15 \ /* y2 = f */
XORL a, R14 \ /* y1 = a ^ (a >> (22-13) */
ROLL $27, R13 \ /* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
XORL g, R15 \ /* y2 = f^g */
XORL e, R13 \ /* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
ROLL $21, R14 \ /* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
ANDL e, R15 \ /* y2 = (f^g)&e */
XORL a, R14 \ /* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
ROLL $26, R13 \ /* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
XORL g, R15 \ /* y2 = CH = ((f^g)&e)^g */
ADDL R13, R15 \ /* y2 = S1 + CH */
ROLL $30, R14 \ /* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
ADDL _xfer+offset(FP), R15 \ /* y2 = k + w + S1 + CH */
MOVL a, R13 \ /* y0 = a */
ADDL R15, h \ /* h = h + S1 + CH + k + w */
MOVL a, R15 \ /* y2 = a */
ORL c, R13 \ /* y0 = a|c */
ADDL h, d \ /* d = d + h + S1 + CH + k + w */
ANDL c, R15 \ /* y2 = a&c */
ANDL b, R13 \ /* y0 = (a|c)&b */
ADDL R14, h \ /* h = h + S1 + CH + k + w + S0 */
ORL R15, R13 \ /* y0 = MAJ = (a|c)&b)|(a&c) */
ADDL R13, h /* h = h + S1 + CH + k + w + S0 + MAJ */
// func blockAvx(h []uint32, message []uint8, reserved0, reserved1, reserved2, reserved3 uint64)
TEXT ·blockAvx(SB), 7, $0
MOVQ h+0(FP), SI // SI: &h
MOVQ message+24(FP), R8 // &message
MOVQ lenmessage+32(FP), R9 // length of message
CMPQ R9, $0
JEQ done_hash
ADDQ R8, R9
MOVQ R9, _inp_end+64(FP) // store end of message
// Register definition
// a --> eax
// b --> ebx
// c --> ecx
// d --> r8d
// e --> edx
// f --> r9d
// g --> r10d
// h --> r11d
//
// y0 --> r13d
// y1 --> r14d
// y2 --> r15d
MOVL (0*4)(SI), AX // a = H0
MOVL (1*4)(SI), BX // b = H1
MOVL (2*4)(SI), CX // c = H2
MOVL (3*4)(SI), R8 // d = H3
MOVL (4*4)(SI), DX // e = H4
MOVL (5*4)(SI), R9 // f = H5
MOVL (6*4)(SI), R10 // g = H6
MOVL (7*4)(SI), R11 // h = H7
MOVOU bflipMask<>(SB), X13
MOVOU shuf00BA<>(SB), X10 // shuffle xBxA -> 00BA
MOVOU shufDC00<>(SB), X12 // shuffle xDxC -> DC00
MOVQ message+24(FP), SI // SI: &message
loop0:
LEAQ constants<>(SB), BP
// byte swap first 16 dwords
MOVOU 0*16(SI), X4
LONG $0x0059c2c4; BYTE $0xe5 // VPSHUFB XMM4, XMM4, XMM13
MOVOU 1*16(SI), X5
LONG $0x0051c2c4; BYTE $0xed // VPSHUFB XMM5, XMM5, XMM13
MOVOU 2*16(SI), X6
LONG $0x0049c2c4; BYTE $0xf5 // VPSHUFB XMM6, XMM6, XMM13
MOVOU 3*16(SI), X7
LONG $0x0041c2c4; BYTE $0xfd // VPSHUFB XMM7, XMM7, XMM13
MOVQ SI, _inp+72(FP)
MOVD $0x3, DI
// schedule 48 input dwords, by doing 3 rounds of 16 each
loop1:
LONG $0x4dfe59c5; BYTE $0x00 // VPADDD XMM9, XMM4, 0[RBP] /* Add 1st constant to first part of message */
MOVOU X9, _xfer+48(FP)
FOUR_ROUNDS_AND_SCHED(AX, BX, CX, R8, DX, R9, R10, R11)
LONG $0x4dfe59c5; BYTE $0x10 // VPADDD XMM9, XMM4, 16[RBP] /* Add 2nd constant to message */
MOVOU X9, _xfer+48(FP)
FOUR_ROUNDS_AND_SCHED(DX, R9, R10, R11, AX, BX, CX, R8)
LONG $0x4dfe59c5; BYTE $0x20 // VPADDD XMM9, XMM4, 32[RBP] /* Add 3rd constant to message */
MOVOU X9, _xfer+48(FP)
FOUR_ROUNDS_AND_SCHED(AX, BX, CX, R8, DX, R9, R10, R11)
LONG $0x4dfe59c5; BYTE $0x30 // VPADDD XMM9, XMM4, 48[RBP] /* Add 4th constant to message */
MOVOU X9, _xfer+48(FP)
ADDQ $64, BP
FOUR_ROUNDS_AND_SCHED(DX, R9, R10, R11, AX, BX, CX, R8)
SUBQ $1, DI
JNE loop1
MOVD $0x2, DI
loop2:
LONG $0x4dfe59c5; BYTE $0x00 // VPADDD XMM9, XMM4, 0[RBP] /* Add 1st constant to first part of message */
MOVOU X9, _xfer+48(FP)
DO_ROUND( AX, BX, CX, R8, DX, R9, R10, R11, 48)
DO_ROUND(R11, AX, BX, CX, R8, DX, R9, R10, 52)
DO_ROUND(R10, R11, AX, BX, CX, R8, DX, R9, 56)
DO_ROUND( R9, R10, R11, AX, BX, CX, R8, DX, 60)
LONG $0x4dfe51c5; BYTE $0x10 // VPADDD XMM9, XMM5, 16[RBP] /* Add 2nd constant to message */
MOVOU X9, _xfer+48(FP)
ADDQ $32, BP
DO_ROUND( DX, R9, R10, R11, AX, BX, CX, R8, 48)
DO_ROUND( R8, DX, R9, R10, R11, AX, BX, CX, 52)
DO_ROUND( CX, R8, DX, R9, R10, R11, AX, BX, 56)
DO_ROUND( BX, CX, R8, DX, R9, R10, R11, AX, 60)
MOVOU X6, X4
MOVOU X7, X5
SUBQ $1, DI
JNE loop2
MOVQ h+0(FP), SI // SI: &h
ADDL (0*4)(SI), AX // H0 = a + H0
MOVL AX, (0*4)(SI)
ADDL (1*4)(SI), BX // H1 = b + H1
MOVL BX, (1*4)(SI)
ADDL (2*4)(SI), CX // H2 = c + H2
MOVL CX, (2*4)(SI)
ADDL (3*4)(SI), R8 // H3 = d + H3
MOVL R8, (3*4)(SI)
ADDL (4*4)(SI), DX // H4 = e + H4
MOVL DX, (4*4)(SI)
ADDL (5*4)(SI), R9 // H5 = f + H5
MOVL R9, (5*4)(SI)
ADDL (6*4)(SI), R10 // H6 = g + H6
MOVL R10, (6*4)(SI)
ADDL (7*4)(SI), R11 // H7 = h + H7
MOVL R11, (7*4)(SI)
MOVQ _inp+72(FP), SI
ADDQ $64, SI
CMPQ _inp_end+64(FP), SI
JNE loop0
done_hash:
RET
// Constants table
DATA constants<>+0x0(SB)/8, $0x71374491428a2f98
DATA constants<>+0x8(SB)/8, $0xe9b5dba5b5c0fbcf
DATA constants<>+0x10(SB)/8, $0x59f111f13956c25b
DATA constants<>+0x18(SB)/8, $0xab1c5ed5923f82a4
DATA constants<>+0x20(SB)/8, $0x12835b01d807aa98
DATA constants<>+0x28(SB)/8, $0x550c7dc3243185be
DATA constants<>+0x30(SB)/8, $0x80deb1fe72be5d74
DATA constants<>+0x38(SB)/8, $0xc19bf1749bdc06a7
DATA constants<>+0x40(SB)/8, $0xefbe4786e49b69c1
DATA constants<>+0x48(SB)/8, $0x240ca1cc0fc19dc6
DATA constants<>+0x50(SB)/8, $0x4a7484aa2de92c6f
DATA constants<>+0x58(SB)/8, $0x76f988da5cb0a9dc
DATA constants<>+0x60(SB)/8, $0xa831c66d983e5152
DATA constants<>+0x68(SB)/8, $0xbf597fc7b00327c8
DATA constants<>+0x70(SB)/8, $0xd5a79147c6e00bf3
DATA constants<>+0x78(SB)/8, $0x1429296706ca6351
DATA constants<>+0x80(SB)/8, $0x2e1b213827b70a85
DATA constants<>+0x88(SB)/8, $0x53380d134d2c6dfc
DATA constants<>+0x90(SB)/8, $0x766a0abb650a7354
DATA constants<>+0x98(SB)/8, $0x92722c8581c2c92e
DATA constants<>+0xa0(SB)/8, $0xa81a664ba2bfe8a1
DATA constants<>+0xa8(SB)/8, $0xc76c51a3c24b8b70
DATA constants<>+0xb0(SB)/8, $0xd6990624d192e819
DATA constants<>+0xb8(SB)/8, $0x106aa070f40e3585
DATA constants<>+0xc0(SB)/8, $0x1e376c0819a4c116
DATA constants<>+0xc8(SB)/8, $0x34b0bcb52748774c
DATA constants<>+0xd0(SB)/8, $0x4ed8aa4a391c0cb3
DATA constants<>+0xd8(SB)/8, $0x682e6ff35b9cca4f
DATA constants<>+0xe0(SB)/8, $0x78a5636f748f82ee
DATA constants<>+0xe8(SB)/8, $0x8cc7020884c87814
DATA constants<>+0xf0(SB)/8, $0xa4506ceb90befffa
DATA constants<>+0xf8(SB)/8, $0xc67178f2bef9a3f7
DATA bflipMask<>+0x00(SB)/8, $0x0405060700010203
DATA bflipMask<>+0x08(SB)/8, $0x0c0d0e0f08090a0b
DATA shuf00BA<>+0x00(SB)/8, $0x0b0a090803020100
DATA shuf00BA<>+0x08(SB)/8, $0xFFFFFFFFFFFFFFFF
DATA shufDC00<>+0x00(SB)/8, $0xFFFFFFFFFFFFFFFF
DATA shufDC00<>+0x08(SB)/8, $0x0b0a090803020100
GLOBL constants<>(SB), 8, $256
GLOBL bflipMask<>(SB), (NOPTR+RODATA), $16
GLOBL shuf00BA<>(SB), (NOPTR+RODATA), $16
GLOBL shufDC00<>(SB), (NOPTR+RODATA), $16