1/* memrchr optimized with AVX2.
2 Copyright (C) 2017-2024 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4
5 The GNU C Library is free software; you can redistribute it and/or
6 modify it under the terms of the GNU Lesser General Public
7 License as published by the Free Software Foundation; either
8 version 2.1 of the License, or (at your option) any later version.
9
10 The GNU C Library is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 Lesser General Public License for more details.
14
15 You should have received a copy of the GNU Lesser General Public
16 License along with the GNU C Library; if not, see
17 <https://www.gnu.org/licenses/>. */
18
19#include <isa-level.h>
20
21#if ISA_SHOULD_BUILD (3)
22
23# include <sysdep.h>
24
25# ifndef MEMRCHR
26# define MEMRCHR __memrchr_avx2
27# endif
28
29# ifndef VZEROUPPER
30# define VZEROUPPER vzeroupper
31# endif
32
33# ifndef SECTION
34# define SECTION(p) p##.avx
35# endif
36
37# define VEC_SIZE 32
38# define PAGE_SIZE 4096
39 .section SECTION(.text), "ax", @progbits
40ENTRY_P2ALIGN(MEMRCHR, 6)
41# ifdef __ILP32__
42 /* Clear upper bits. */
43 and %RDX_LP, %RDX_LP
44# else
45 test %RDX_LP, %RDX_LP
46# endif
47 jz L(zero_0)
48
49 vmovd %esi, %xmm0
50 /* Get end pointer. Minus one for two reasons. 1) It is necessary for a
51 correct page cross check and 2) it correctly sets up end ptr to be
52 subtract by lzcnt aligned. */
53 leaq -1(%rdx, %rdi), %rax
54
55 vpbroadcastb %xmm0, %ymm0
56
57 /* Check if we can load 1x VEC without cross a page. */
58 testl $(PAGE_SIZE - VEC_SIZE), %eax
59 jz L(page_cross)
60
61 vpcmpeqb -(VEC_SIZE - 1)(%rax), %ymm0, %ymm1
62 vpmovmskb %ymm1, %ecx
63 cmpq $VEC_SIZE, %rdx
64 ja L(more_1x_vec)
65
66L(ret_vec_x0_test):
67 /* If ecx is zero (no matches) lzcnt will set it 32 (VEC_SIZE) which
68 will guarantee edx (len) is less than it. */
69 lzcntl %ecx, %ecx
70
71 /* Hoist vzeroupper (not great for RTM) to save code size. This allows
72 all logic for edx (len) <= VEC_SIZE to fit in first cache line. */
73 COND_VZEROUPPER
74 cmpl %ecx, %edx
75 jle L(zero_0)
76 subq %rcx, %rax
77 ret
78
79 /* Fits in aligning bytes of first cache line. */
80L(zero_0):
81 xorl %eax, %eax
82 ret
83
84 .p2align 4,, 9
85L(ret_vec_x0):
86 lzcntl %ecx, %ecx
87 subq %rcx, %rax
88L(return_vzeroupper):
89 ZERO_UPPER_VEC_REGISTERS_RETURN
90
91 .p2align 4,, 10
92L(more_1x_vec):
93 testl %ecx, %ecx
94 jnz L(ret_vec_x0)
95
96 /* Align rax (string pointer). */
97 andq $-VEC_SIZE, %rax
98
99 /* Recompute remaining length after aligning. */
100 movq %rax, %rdx
101 /* Need this comparison next no matter what. */
102 vpcmpeqb -(VEC_SIZE)(%rax), %ymm0, %ymm1
103 subq %rdi, %rdx
104 decq %rax
105 vpmovmskb %ymm1, %ecx
106 /* Fall through for short (hotter than length). */
107 cmpq $(VEC_SIZE * 2), %rdx
108 ja L(more_2x_vec)
109L(last_2x_vec):
110 cmpl $VEC_SIZE, %edx
111 jbe L(ret_vec_x0_test)
112
113 testl %ecx, %ecx
114 jnz L(ret_vec_x0)
115
116 vpcmpeqb -(VEC_SIZE * 2 - 1)(%rax), %ymm0, %ymm1
117 vpmovmskb %ymm1, %ecx
118 /* 64-bit lzcnt. This will naturally add 32 to position. */
119 lzcntq %rcx, %rcx
120 COND_VZEROUPPER
121 cmpl %ecx, %edx
122 jle L(zero_0)
123 subq %rcx, %rax
124 ret
125
126
127 /* Inexpensive place to put this regarding code size / target alignments
128 / ICache NLP. Necessary for 2-byte encoding of jump to page cross
129 case which in turn is necessary for hot path (len <= VEC_SIZE) to fit
130 in first cache line. */
131L(page_cross):
132 movq %rax, %rsi
133 andq $-VEC_SIZE, %rsi
134 vpcmpeqb (%rsi), %ymm0, %ymm1
135 vpmovmskb %ymm1, %ecx
136 /* Shift out negative alignment (because we are starting from endptr and
137 working backwards). */
138 movl %eax, %r8d
139 /* notl because eax already has endptr - 1. (-x = ~(x - 1)). */
140 notl %r8d
141 shlxl %r8d, %ecx, %ecx
142 cmpq %rdi, %rsi
143 ja L(more_1x_vec)
144 lzcntl %ecx, %ecx
145 COND_VZEROUPPER
146 cmpl %ecx, %edx
147 jle L(zero_0)
148 subq %rcx, %rax
149 ret
150 .p2align 4,, 11
151L(ret_vec_x1):
152 /* This will naturally add 32 to position. */
153 lzcntq %rcx, %rcx
154 subq %rcx, %rax
155 VZEROUPPER_RETURN
156 .p2align 4,, 10
157L(more_2x_vec):
158 testl %ecx, %ecx
159 jnz L(ret_vec_x0)
160
161 vpcmpeqb -(VEC_SIZE * 2 - 1)(%rax), %ymm0, %ymm1
162 vpmovmskb %ymm1, %ecx
163 testl %ecx, %ecx
164 jnz L(ret_vec_x1)
165
166
167 /* Needed no matter what. */
168 vpcmpeqb -(VEC_SIZE * 3 - 1)(%rax), %ymm0, %ymm1
169 vpmovmskb %ymm1, %ecx
170
171 subq $(VEC_SIZE * 4), %rdx
172 ja L(more_4x_vec)
173
174 cmpl $(VEC_SIZE * -1), %edx
175 jle L(ret_vec_x2_test)
176
177L(last_vec):
178 testl %ecx, %ecx
179 jnz L(ret_vec_x2)
180
181 /* Needed no matter what. */
182 vpcmpeqb -(VEC_SIZE * 4 - 1)(%rax), %ymm0, %ymm1
183 vpmovmskb %ymm1, %ecx
184 lzcntl %ecx, %ecx
185 subq $(VEC_SIZE * 3), %rax
186 COND_VZEROUPPER
187 subq %rcx, %rax
188 cmpq %rax, %rdi
189 ja L(zero_2)
190 ret
191
192 /* First in aligning bytes. */
193L(zero_2):
194 xorl %eax, %eax
195 ret
196
197 .p2align 4,, 4
198L(ret_vec_x2_test):
199 lzcntl %ecx, %ecx
200 subq $(VEC_SIZE * 2), %rax
201 COND_VZEROUPPER
202 subq %rcx, %rax
203 cmpq %rax, %rdi
204 ja L(zero_2)
205 ret
206
207
208 .p2align 4,, 11
209L(ret_vec_x2):
210 /* ecx must be non-zero. */
211 bsrl %ecx, %ecx
212 leaq (VEC_SIZE * -3 + 1)(%rcx, %rax), %rax
213 VZEROUPPER_RETURN
214
215 .p2align 4,, 14
216L(ret_vec_x3):
217 /* ecx must be non-zero. */
218 bsrl %ecx, %ecx
219 leaq (VEC_SIZE * -4 + 1)(%rcx, %rax), %rax
220 VZEROUPPER_RETURN
221
222
223
224 .p2align 4
225L(more_4x_vec):
226 testl %ecx, %ecx
227 jnz L(ret_vec_x2)
228
229 vpcmpeqb -(VEC_SIZE * 4 - 1)(%rax), %ymm0, %ymm1
230 vpmovmskb %ymm1, %ecx
231
232 testl %ecx, %ecx
233 jnz L(ret_vec_x3)
234
235 /* Check if near end before re-aligning (otherwise might do an
236 unnecessary loop iteration). */
237 addq $-(VEC_SIZE * 4), %rax
238 cmpq $(VEC_SIZE * 4), %rdx
239 jbe L(last_4x_vec)
240
241 /* Align rax to (VEC_SIZE - 1). */
242 orq $(VEC_SIZE * 4 - 1), %rax
243 movq %rdi, %rdx
244 /* Get endptr for loop in rdx. NB: Can't just do while rax > rdi because
245 lengths that overflow can be valid and break the comparison. */
246 orq $(VEC_SIZE * 4 - 1), %rdx
247
248 .p2align 4
249L(loop_4x_vec):
250 /* Need this comparison next no matter what. */
251 vpcmpeqb -(VEC_SIZE * 1 - 1)(%rax), %ymm0, %ymm1
252 vpcmpeqb -(VEC_SIZE * 2 - 1)(%rax), %ymm0, %ymm2
253 vpcmpeqb -(VEC_SIZE * 3 - 1)(%rax), %ymm0, %ymm3
254 vpcmpeqb -(VEC_SIZE * 4 - 1)(%rax), %ymm0, %ymm4
255
256 vpor %ymm1, %ymm2, %ymm2
257 vpor %ymm3, %ymm4, %ymm4
258 vpor %ymm2, %ymm4, %ymm4
259 vpmovmskb %ymm4, %esi
260
261 testl %esi, %esi
262 jnz L(loop_end)
263
264 addq $(VEC_SIZE * -4), %rax
265 cmpq %rdx, %rax
266 jne L(loop_4x_vec)
267
268 subl %edi, %edx
269 incl %edx
270
271L(last_4x_vec):
272 /* Used no matter what. */
273 vpcmpeqb -(VEC_SIZE * 1 - 1)(%rax), %ymm0, %ymm1
274 vpmovmskb %ymm1, %ecx
275
276 cmpl $(VEC_SIZE * 2), %edx
277 jbe L(last_2x_vec)
278
279 testl %ecx, %ecx
280 jnz L(ret_vec_x0_end)
281
282 vpcmpeqb -(VEC_SIZE * 2 - 1)(%rax), %ymm0, %ymm1
283 vpmovmskb %ymm1, %ecx
284 testl %ecx, %ecx
285 jnz L(ret_vec_x1_end)
286
287 /* Used no matter what. */
288 vpcmpeqb -(VEC_SIZE * 3 - 1)(%rax), %ymm0, %ymm1
289 vpmovmskb %ymm1, %ecx
290
291 cmpl $(VEC_SIZE * 3), %edx
292 ja L(last_vec)
293
294 lzcntl %ecx, %ecx
295 subq $(VEC_SIZE * 2), %rax
296 COND_VZEROUPPER
297 subq %rcx, %rax
298 cmpq %rax, %rdi
299 jbe L(ret0)
300 xorl %eax, %eax
301L(ret0):
302 ret
303
304
305 .p2align 4
306L(loop_end):
307 vpmovmskb %ymm1, %ecx
308 testl %ecx, %ecx
309 jnz L(ret_vec_x0_end)
310
311 vpmovmskb %ymm2, %ecx
312 testl %ecx, %ecx
313 jnz L(ret_vec_x1_end)
314
315 vpmovmskb %ymm3, %ecx
316 /* Combine last 2 VEC matches. If ecx (VEC3) is zero (no CHAR in VEC3)
317 then it won't affect the result in esi (VEC4). If ecx is non-zero
318 then CHAR in VEC3 and bsrq will use that position. */
319 salq $32, %rcx
320 orq %rsi, %rcx
321 bsrq %rcx, %rcx
322 leaq (VEC_SIZE * -4 + 1)(%rcx, %rax), %rax
323 VZEROUPPER_RETURN
324
325 .p2align 4,, 4
326L(ret_vec_x1_end):
327 /* 64-bit version will automatically add 32 (VEC_SIZE). */
328 lzcntq %rcx, %rcx
329 subq %rcx, %rax
330 VZEROUPPER_RETURN
331
332 .p2align 4,, 4
333L(ret_vec_x0_end):
334 lzcntl %ecx, %ecx
335 subq %rcx, %rax
336 VZEROUPPER_RETURN
337
338 /* 2 bytes until next cache line. */
339END(MEMRCHR)
340#endif
341

source code of glibc/sysdeps/x86_64/multiarch/memrchr-avx2.S