forked from cronokirby/saferith
-
Notifications
You must be signed in to change notification settings - Fork 0
/
arith_arm.s
284 lines (243 loc) · 4.2 KB
/
arith_arm.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
// Copyright 2009 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_go file.
//go:build !math_big_pure_go
// +build !math_big_pure_go
#include "textflag.h"
// This file provides fast assembly versions for the elementary
// arithmetic operations on vectors implemented in arith.go.
// func addVV(z, x, y []Word) (c Word)
TEXT ·addVV(SB),NOSPLIT,$0
ADD.S $0, R0 // clear carry flag
MOVW z+0(FP), R1
MOVW z_len+4(FP), R4
MOVW x+12(FP), R2
MOVW y+24(FP), R3
ADD R4<<2, R1, R4
B E1
L1:
MOVW.P 4(R2), R5
MOVW.P 4(R3), R6
ADC.S R6, R5
MOVW.P R5, 4(R1)
E1:
TEQ R1, R4
BNE L1
MOVW $0, R0
MOVW.CS $1, R0
MOVW R0, c+36(FP)
RET
// func subVV(z, x, y []Word) (c Word)
// (same as addVV except for SBC instead of ADC and label names)
TEXT ·subVV(SB),NOSPLIT,$0
SUB.S $0, R0 // clear borrow flag
MOVW z+0(FP), R1
MOVW z_len+4(FP), R4
MOVW x+12(FP), R2
MOVW y+24(FP), R3
ADD R4<<2, R1, R4
B E2
L2:
MOVW.P 4(R2), R5
MOVW.P 4(R3), R6
SBC.S R6, R5
MOVW.P R5, 4(R1)
E2:
TEQ R1, R4
BNE L2
MOVW $0, R0
MOVW.CC $1, R0
MOVW R0, c+36(FP)
RET
// func addVW(z, x []Word, y Word) (c Word)
TEXT ·addVW(SB),NOSPLIT,$0
MOVW z+0(FP), R1
MOVW z_len+4(FP), R4
MOVW x+12(FP), R2
MOVW y+24(FP), R3
ADD R4<<2, R1, R4
TEQ R1, R4
BNE L3a
MOVW R3, c+28(FP)
RET
L3a:
MOVW.P 4(R2), R5
ADD.S R3, R5
MOVW.P R5, 4(R1)
B E3
L3:
MOVW.P 4(R2), R5
ADC.S $0, R5
MOVW.P R5, 4(R1)
E3:
TEQ R1, R4
BNE L3
MOVW $0, R0
MOVW.CS $1, R0
MOVW R0, c+28(FP)
RET
// func subVW(z, x []Word, y Word) (c Word)
TEXT ·subVW(SB),NOSPLIT,$0
MOVW z+0(FP), R1
MOVW z_len+4(FP), R4
MOVW x+12(FP), R2
MOVW y+24(FP), R3
ADD R4<<2, R1, R4
TEQ R1, R4
BNE L4a
MOVW R3, c+28(FP)
RET
L4a:
MOVW.P 4(R2), R5
SUB.S R3, R5
MOVW.P R5, 4(R1)
B E4
L4:
MOVW.P 4(R2), R5
SBC.S $0, R5
MOVW.P R5, 4(R1)
E4:
TEQ R1, R4
BNE L4
MOVW $0, R0
MOVW.CC $1, R0
MOVW R0, c+28(FP)
RET
// func shlVU(z, x []Word, s uint) (c Word)
TEXT ·shlVU(SB),NOSPLIT,$0
MOVW z_len+4(FP), R5
TEQ $0, R5
BEQ X7
MOVW z+0(FP), R1
MOVW x+12(FP), R2
ADD R5<<2, R2, R2
ADD R5<<2, R1, R5
MOVW s+24(FP), R3
TEQ $0, R3 // shift 0 is special
BEQ Y7
ADD $4, R1 // stop one word early
MOVW $32, R4
SUB R3, R4
MOVW $0, R7
MOVW.W -4(R2), R6
MOVW R6<<R3, R7
MOVW R6>>R4, R6
MOVW R6, c+28(FP)
B E7
L7:
MOVW.W -4(R2), R6
ORR R6>>R4, R7
MOVW.W R7, -4(R5)
MOVW R6<<R3, R7
E7:
TEQ R1, R5
BNE L7
MOVW R7, -4(R5)
RET
Y7: // copy loop, because shift 0 == shift 32
MOVW.W -4(R2), R6
MOVW.W R6, -4(R5)
TEQ R1, R5
BNE Y7
X7:
MOVW $0, R1
MOVW R1, c+28(FP)
RET
// func shrVU(z, x []Word, s uint) (c Word)
TEXT ·shrVU(SB),NOSPLIT,$0
MOVW z_len+4(FP), R5
TEQ $0, R5
BEQ X6
MOVW z+0(FP), R1
MOVW x+12(FP), R2
ADD R5<<2, R1, R5
MOVW s+24(FP), R3
TEQ $0, R3 // shift 0 is special
BEQ Y6
SUB $4, R5 // stop one word early
MOVW $32, R4
SUB R3, R4
MOVW $0, R7
// first word
MOVW.P 4(R2), R6
MOVW R6>>R3, R7
MOVW R6<<R4, R6
MOVW R6, c+28(FP)
B E6
// word loop
L6:
MOVW.P 4(R2), R6
ORR R6<<R4, R7
MOVW.P R7, 4(R1)
MOVW R6>>R3, R7
E6:
TEQ R1, R5
BNE L6
MOVW R7, 0(R1)
RET
Y6: // copy loop, because shift 0 == shift 32
MOVW.P 4(R2), R6
MOVW.P R6, 4(R1)
TEQ R1, R5
BNE Y6
X6:
MOVW $0, R1
MOVW R1, c+28(FP)
RET
// func mulAddVWW(z, x []Word, y, r Word) (c Word)
TEXT ·mulAddVWW(SB),NOSPLIT,$0
MOVW $0, R0
MOVW z+0(FP), R1
MOVW z_len+4(FP), R5
MOVW x+12(FP), R2
MOVW y+24(FP), R3
MOVW r+28(FP), R4
ADD R5<<2, R1, R5
B E8
// word loop
L8:
MOVW.P 4(R2), R6
MULLU R6, R3, (R7, R6)
ADD.S R4, R6
ADC R0, R7
MOVW.P R6, 4(R1)
MOVW R7, R4
E8:
TEQ R1, R5
BNE L8
MOVW R4, c+32(FP)
RET
// func addMulVVW(z, x []Word, y Word) (c Word)
TEXT ·addMulVVW(SB),NOSPLIT,$0
MOVW $0, R0
MOVW z+0(FP), R1
MOVW z_len+4(FP), R5
MOVW x+12(FP), R2
MOVW y+24(FP), R3
ADD R5<<2, R1, R5
MOVW $0, R4
B E9
// word loop
L9:
MOVW.P 4(R2), R6
MULLU R6, R3, (R7, R6)
ADD.S R4, R6
ADC R0, R7
MOVW 0(R1), R4
ADD.S R4, R6
ADC R0, R7
MOVW.P R6, 4(R1)
MOVW R7, R4
E9:
TEQ R1, R5
BNE L9
MOVW R4, c+28(FP)
RET
// func mulWW(x, y Word) (z1, z0 Word)
TEXT ·mulWW(SB),NOSPLIT,$0
MOVW x+0(FP), R1
MOVW y+4(FP), R2
MULLU R1, R2, (R4, R3)
MOVW R4, z1+8(FP)
MOVW R3, z0+12(FP)
RET