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
|
// Copyright 2017 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
#include "textflag.h"
DATA rotInvSRows<>+0x00(SB)/8, $0x080f0205040b0e01
DATA rotInvSRows<>+0x08(SB)/8, $0x00070a0d0c030609
GLOBL rotInvSRows<>(SB), (NOPTR+RODATA), $16
DATA invSRows<>+0x00(SB)/8, $0x0b0e0104070a0d00
DATA invSRows<>+0x08(SB)/8, $0x0306090c0f020508
GLOBL invSRows<>(SB), (NOPTR+RODATA), $16
// func encryptBlockAsm(nr int, xk *uint32, dst, src *byte)
TEXT ·encryptBlockAsm(SB),NOSPLIT,$0
MOVD nr+0(FP), R9
MOVD xk+8(FP), R10
MOVD dst+16(FP), R11
MOVD src+24(FP), R12
VLD1 (R12), [V0.B16]
CMP $12, R9
BLT enc128
BEQ enc196
enc256:
VLD1.P 32(R10), [V1.B16, V2.B16]
AESE V1.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V2.B16, V0.B16
AESMC V0.B16, V0.B16
enc196:
VLD1.P 32(R10), [V3.B16, V4.B16]
AESE V3.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V4.B16, V0.B16
AESMC V0.B16, V0.B16
enc128:
VLD1.P 64(R10), [V5.B16, V6.B16, V7.B16, V8.B16]
VLD1.P 64(R10), [V9.B16, V10.B16, V11.B16, V12.B16]
VLD1.P 48(R10), [V13.B16, V14.B16, V15.B16]
AESE V5.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V6.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V7.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V8.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V9.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V10.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V11.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V12.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V13.B16, V0.B16
AESMC V0.B16, V0.B16
AESE V14.B16, V0.B16
VEOR V0.B16, V15.B16, V0.B16
VST1 [V0.B16], (R11)
RET
// func decryptBlockAsm(nr int, xk *uint32, dst, src *byte)
TEXT ·decryptBlockAsm(SB),NOSPLIT,$0
MOVD nr+0(FP), R9
MOVD xk+8(FP), R10
MOVD dst+16(FP), R11
MOVD src+24(FP), R12
VLD1 (R12), [V0.B16]
CMP $12, R9
BLT dec128
BEQ dec196
dec256:
VLD1.P 32(R10), [V1.B16, V2.B16]
AESD V1.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V2.B16, V0.B16
AESIMC V0.B16, V0.B16
dec196:
VLD1.P 32(R10), [V3.B16, V4.B16]
AESD V3.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V4.B16, V0.B16
AESIMC V0.B16, V0.B16
dec128:
VLD1.P 64(R10), [V5.B16, V6.B16, V7.B16, V8.B16]
VLD1.P 64(R10), [V9.B16, V10.B16, V11.B16, V12.B16]
VLD1.P 48(R10), [V13.B16, V14.B16, V15.B16]
AESD V5.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V6.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V7.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V8.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V9.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V10.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V11.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V12.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V13.B16, V0.B16
AESIMC V0.B16, V0.B16
AESD V14.B16, V0.B16
VEOR V0.B16, V15.B16, V0.B16
VST1 [V0.B16], (R11)
RET
// func expandKeyAsm(nr int, key *byte, enc, dec *uint32) {
// Note that round keys are stored in uint128 format, not uint32
TEXT ·expandKeyAsm(SB),NOSPLIT,$0
MOVD nr+0(FP), R8
MOVD key+8(FP), R9
MOVD enc+16(FP), R10
MOVD dec+24(FP), R11
LDP rotInvSRows<>(SB), (R0, R1)
VMOV R0, V3.D[0]
VMOV R1, V3.D[1]
VEOR V0.B16, V0.B16, V0.B16 // All zeroes
MOVW $1, R13
TBZ $1, R8, ks192
TBNZ $2, R8, ks256
LDPW (R9), (R4, R5)
LDPW 8(R9), (R6, R7)
STPW.P (R4, R5), 8(R10)
STPW.P (R6, R7), 8(R10)
MOVW $0x1b, R14
ks128Loop:
VMOV R7, V2.S[0]
WORD $0x4E030042 // TBL V3.B16, [V2.B16], V2.B16
AESE V0.B16, V2.B16 // Use AES to compute the SBOX
EORW R13, R4
LSLW $1, R13 // Compute next Rcon
ANDSW $0x100, R13, ZR
CSELW NE, R14, R13, R13 // Fake modulo
SUBS $1, R8
VMOV V2.S[0], R0
EORW R0, R4
EORW R4, R5
EORW R5, R6
EORW R6, R7
STPW.P (R4, R5), 8(R10)
STPW.P (R6, R7), 8(R10)
BNE ks128Loop
CBZ R11, ksDone // If dec is nil we are done
SUB $176, R10
// Decryption keys are encryption keys with InverseMixColumns applied
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
VMOV V0.B16, V7.B16
AESIMC V1.B16, V6.B16
AESIMC V2.B16, V5.B16
AESIMC V3.B16, V4.B16
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
AESIMC V0.B16, V11.B16
AESIMC V1.B16, V10.B16
AESIMC V2.B16, V9.B16
AESIMC V3.B16, V8.B16
VLD1 (R10), [V0.B16, V1.B16, V2.B16]
AESIMC V0.B16, V14.B16
AESIMC V1.B16, V13.B16
VMOV V2.B16, V12.B16
VST1.P [V12.B16, V13.B16, V14.B16], 48(R11)
VST1.P [V8.B16, V9.B16, V10.B16, V11.B16], 64(R11)
VST1 [V4.B16, V5.B16, V6.B16, V7.B16], (R11)
B ksDone
ks192:
LDPW (R9), (R2, R3)
LDPW 8(R9), (R4, R5)
LDPW 16(R9), (R6, R7)
STPW.P (R2, R3), 8(R10)
STPW.P (R4, R5), 8(R10)
SUB $4, R8
ks192Loop:
STPW.P (R6, R7), 8(R10)
VMOV R7, V2.S[0]
WORD $0x4E030042 //TBL V3.B16, [V2.B16], V2.B16
AESE V0.B16, V2.B16
EORW R13, R2
LSLW $1, R13
SUBS $1, R8
VMOV V2.S[0], R0
EORW R0, R2
EORW R2, R3
EORW R3, R4
EORW R4, R5
EORW R5, R6
EORW R6, R7
STPW.P (R2, R3), 8(R10)
STPW.P (R4, R5), 8(R10)
BNE ks192Loop
CBZ R11, ksDone
SUB $208, R10
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
VMOV V0.B16, V7.B16
AESIMC V1.B16, V6.B16
AESIMC V2.B16, V5.B16
AESIMC V3.B16, V4.B16
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
AESIMC V0.B16, V11.B16
AESIMC V1.B16, V10.B16
AESIMC V2.B16, V9.B16
AESIMC V3.B16, V8.B16
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
AESIMC V0.B16, V15.B16
AESIMC V1.B16, V14.B16
AESIMC V2.B16, V13.B16
AESIMC V3.B16, V12.B16
VLD1 (R10), [V0.B16]
VST1.P [V0.B16], 16(R11)
VST1.P [V12.B16, V13.B16, V14.B16, V15.B16], 64(R11)
VST1.P [V8.B16, V9.B16, V10.B16, V11.B16], 64(R11)
VST1 [V4.B16, V5.B16, V6.B16, V7.B16], (R11)
B ksDone
ks256:
LDP invSRows<>(SB), (R0, R1)
VMOV R0, V4.D[0]
VMOV R1, V4.D[1]
LDPW (R9), (R0, R1)
LDPW 8(R9), (R2, R3)
LDPW 16(R9), (R4, R5)
LDPW 24(R9), (R6, R7)
STPW.P (R0, R1), 8(R10)
STPW.P (R2, R3), 8(R10)
SUB $7, R8
ks256Loop:
STPW.P (R4, R5), 8(R10)
STPW.P (R6, R7), 8(R10)
VMOV R7, V2.S[0]
WORD $0x4E030042 //TBL V3.B16, [V2.B16], V2.B16
AESE V0.B16, V2.B16
EORW R13, R0
LSLW $1, R13
SUBS $1, R8
VMOV V2.S[0], R9
EORW R9, R0
EORW R0, R1
EORW R1, R2
EORW R2, R3
VMOV R3, V2.S[0]
WORD $0x4E040042 //TBL V3.B16, [V2.B16], V2.B16
AESE V0.B16, V2.B16
VMOV V2.S[0], R9
EORW R9, R4
EORW R4, R5
EORW R5, R6
EORW R6, R7
STPW.P (R0, R1), 8(R10)
STPW.P (R2, R3), 8(R10)
BNE ks256Loop
CBZ R11, ksDone
SUB $240, R10
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
VMOV V0.B16, V7.B16
AESIMC V1.B16, V6.B16
AESIMC V2.B16, V5.B16
AESIMC V3.B16, V4.B16
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
AESIMC V0.B16, V11.B16
AESIMC V1.B16, V10.B16
AESIMC V2.B16, V9.B16
AESIMC V3.B16, V8.B16
VLD1.P 64(R10), [V0.B16, V1.B16, V2.B16, V3.B16]
AESIMC V0.B16, V15.B16
AESIMC V1.B16, V14.B16
AESIMC V2.B16, V13.B16
AESIMC V3.B16, V12.B16
VLD1 (R10), [V0.B16, V1.B16, V2.B16]
AESIMC V0.B16, V18.B16
AESIMC V1.B16, V17.B16
VMOV V2.B16, V16.B16
VST1.P [V16.B16, V17.B16, V18.B16], 48(R11)
VST1.P [V12.B16, V13.B16, V14.B16, V15.B16], 64(R11)
VST1.P [V8.B16, V9.B16, V10.B16, V11.B16], 64(R11)
VST1 [V4.B16, V5.B16, V6.B16, V7.B16], (R11)
ksDone:
RET
|
