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
410
411
412
413
414
415
416
417
|
/* Optimized memrchr with sse2
Copyright (C) 2011 Free Software Foundation, Inc.
Contributed by Intel Corporation.
This file is part of the GNU C Library.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#ifndef NOT_IN_libc
# include <sysdep.h>
# define CFI_PUSH(REG) \
cfi_adjust_cfa_offset (4); \
cfi_rel_offset (REG, 0)
# define CFI_POP(REG) \
cfi_adjust_cfa_offset (-4); \
cfi_restore (REG)
# define PUSH(REG) pushl REG; CFI_PUSH (REG)
# define POP(REG) popl REG; CFI_POP (REG)
# define PARMS 4
# define STR1 PARMS
# define STR2 STR1+4
# define LEN STR2+4
# define MEMCHR __memrchr_sse2_bsf
.text
ENTRY (MEMCHR)
mov STR1(%esp), %ecx
movd STR2(%esp), %xmm1
mov LEN(%esp), %edx
sub $16, %edx
jbe L(length_less16)
punpcklbw %xmm1, %xmm1
add %edx, %ecx
punpcklbw %xmm1, %xmm1
movdqu (%ecx), %xmm0
pshufd $0, %xmm1, %xmm1
pcmpeqb %xmm1, %xmm0
/* Check if there is a match. */
pmovmskb %xmm0, %eax
test %eax, %eax
jnz L(matches0)
sub $64, %ecx
mov %ecx, %eax
and $15, %eax
jz L(loop_prolog)
add $16, %ecx
add $16, %edx
sub %eax, %ecx
sub %eax, %edx
.p2align 4
/* Loop start on aligned string. */
L(loop_prolog):
sub $64, %edx
jbe L(exit_loop)
movdqa 48(%ecx), %xmm0
pcmpeqb %xmm1, %xmm0
pmovmskb %xmm0, %eax
test %eax, %eax
jnz L(matches48)
movdqa 32(%ecx), %xmm2
pcmpeqb %xmm1, %xmm2
pmovmskb %xmm2, %eax
test %eax, %eax
jnz L(matches32)
movdqa 16(%ecx), %xmm3
pcmpeqb %xmm1, %xmm3
pmovmskb %xmm3, %eax
test %eax, %eax
jnz L(matches16)
movdqa (%ecx), %xmm4
pcmpeqb %xmm1, %xmm4
pmovmskb %xmm4, %eax
test %eax, %eax
jnz L(matches0)
sub $64, %ecx
sub $64, %edx
jbe L(exit_loop)
movdqa 48(%ecx), %xmm0
pcmpeqb %xmm1, %xmm0
pmovmskb %xmm0, %eax
test %eax, %eax
jnz L(matches48)
movdqa 32(%ecx), %xmm2
pcmpeqb %xmm1, %xmm2
pmovmskb %xmm2, %eax
test %eax, %eax
jnz L(matches32)
movdqa 16(%ecx), %xmm3
pcmpeqb %xmm1, %xmm3
pmovmskb %xmm3, %eax
test %eax, %eax
jnz L(matches16)
movdqa (%ecx), %xmm3
pcmpeqb %xmm1, %xmm3
pmovmskb %xmm3, %eax
test %eax, %eax
jnz L(matches0)
mov %ecx, %eax
and $63, %eax
test %eax, %eax
jz L(align64_loop)
add $64, %ecx
add $64, %edx
sub %eax, %ecx
sub %eax, %edx
.p2align 4
L(align64_loop):
sub $64, %ecx
sub $64, %edx
jbe L(exit_loop)
movdqa (%ecx), %xmm0
movdqa 16(%ecx), %xmm2
movdqa 32(%ecx), %xmm3
movdqa 48(%ecx), %xmm4
pcmpeqb %xmm1, %xmm0
pcmpeqb %xmm1, %xmm2
pcmpeqb %xmm1, %xmm3
pcmpeqb %xmm1, %xmm4
pmaxub %xmm3, %xmm0
pmaxub %xmm4, %xmm2
pmaxub %xmm0, %xmm2
pmovmskb %xmm2, %eax
test %eax, %eax
jz L(align64_loop)
pmovmskb %xmm4, %eax
test %eax, %eax
jnz L(matches48)
pmovmskb %xmm3, %eax
test %eax, %eax
jnz L(matches32)
movdqa 16(%ecx), %xmm2
pcmpeqb %xmm1, %xmm2
pcmpeqb (%ecx), %xmm1
pmovmskb %xmm2, %eax
test %eax, %eax
jnz L(matches16)
pmovmskb %xmm1, %eax
bsr %eax, %eax
add %ecx, %eax
ret
.p2align 4
L(exit_loop):
add $64, %edx
cmp $32, %edx
jbe L(exit_loop_32)
movdqa 48(%ecx), %xmm0
pcmpeqb %xmm1, %xmm0
pmovmskb %xmm0, %eax
test %eax, %eax
jnz L(matches48)
movdqa 32(%ecx), %xmm2
pcmpeqb %xmm1, %xmm2
pmovmskb %xmm2, %eax
test %eax, %eax
jnz L(matches32)
movdqa 16(%ecx), %xmm3
pcmpeqb %xmm1, %xmm3
pmovmskb %xmm3, %eax
test %eax, %eax
jnz L(matches16_1)
cmp $48, %edx
jbe L(return_null)
pcmpeqb (%ecx), %xmm1
pmovmskb %xmm1, %eax
test %eax, %eax
jnz L(matches0_1)
xor %eax, %eax
ret
.p2align 4
L(exit_loop_32):
movdqa 48(%ecx), %xmm0
pcmpeqb %xmm1, %xmm0
pmovmskb %xmm0, %eax
test %eax, %eax
jnz L(matches48_1)
cmp $16, %edx
jbe L(return_null)
pcmpeqb 32(%ecx), %xmm1
pmovmskb %xmm1, %eax
test %eax, %eax
jnz L(matches32_1)
xor %eax, %eax
ret
.p2align 4
L(matches0):
bsr %eax, %eax
add %ecx, %eax
ret
.p2align 4
L(matches16):
bsr %eax, %eax
lea 16(%eax, %ecx), %eax
ret
.p2align 4
L(matches32):
bsr %eax, %eax
lea 32(%eax, %ecx), %eax
ret
.p2align 4
L(matches48):
bsr %eax, %eax
lea 48(%eax, %ecx), %eax
ret
.p2align 4
L(matches0_1):
bsr %eax, %eax
sub $64, %edx
add %eax, %edx
jl L(return_null)
add %ecx, %eax
ret
.p2align 4
L(matches16_1):
bsr %eax, %eax
sub $48, %edx
add %eax, %edx
jl L(return_null)
lea 16(%ecx, %eax), %eax
ret
.p2align 4
L(matches32_1):
bsr %eax, %eax
sub $32, %edx
add %eax, %edx
jl L(return_null)
lea 32(%ecx, %eax), %eax
ret
.p2align 4
L(matches48_1):
bsr %eax, %eax
sub $16, %edx
add %eax, %edx
jl L(return_null)
lea 48(%ecx, %eax), %eax
ret
.p2align 4
L(return_null):
xor %eax, %eax
ret
.p2align 4
L(length_less16_offset0):
mov %dl, %cl
pcmpeqb (%eax), %xmm1
mov $1, %edx
sal %cl, %edx
sub $1, %edx
mov %edx, %ecx
pmovmskb %xmm1, %edx
and %ecx, %edx
test %edx, %edx
jz L(return_null)
bsr %edx, %ecx
add %ecx, %eax
ret
.p2align 4
L(length_less16):
punpcklbw %xmm1, %xmm1
mov %ecx, %eax
punpcklbw %xmm1, %xmm1
add $16, %edx
jz L(return_null)
pshufd $0, %xmm1, %xmm1
and $15, %ecx
jz L(length_less16_offset0)
PUSH (%edi)
mov %cl, %dh
add %dl, %dh
and $-16, %eax
sub $16, %dh
ja L(length_less16_part2)
pcmpeqb (%eax), %xmm1
pmovmskb %xmm1, %edi
sar %cl, %edi
add %ecx, %eax
mov %dl, %cl
mov $1, %edx
sal %cl, %edx
sub $1, %edx
and %edx, %edi
test %edi, %edi
jz L(ret_null)
bsr %edi, %edi
add %edi, %eax
POP (%edi)
ret
CFI_PUSH (%edi)
.p2align 4
L(length_less16_part2):
movdqa 16(%eax), %xmm2
pcmpeqb %xmm1, %xmm2
pmovmskb %xmm2, %edi
mov %cl, %ch
mov %dh, %cl
mov $1, %edx
sal %cl, %edx
sub $1, %edx
and %edx, %edi
test %edi, %edi
jnz L(length_less16_part2_return)
pcmpeqb (%eax), %xmm1
pmovmskb %xmm1, %edi
mov %ch, %cl
sar %cl, %edi
test %edi, %edi
jz L(ret_null)
bsr %edi, %edi
add %edi, %eax
xor %ch, %ch
add %ecx, %eax
POP (%edi)
ret
CFI_PUSH (%edi)
.p2align 4
L(length_less16_part2_return):
bsr %edi, %edi
lea 16(%eax, %edi), %eax
POP (%edi)
ret
CFI_PUSH (%edi)
.p2align 4
L(ret_null):
xor %eax, %eax
POP (%edi)
ret
END (MEMCHR)
#endif
|
