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
|
/* memset with AVX2
Copyright (C) 2014-2015 Free Software Foundation, Inc.
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/>. */
#include <sysdep.h>
#if IS_IN (libc)
#include "asm-syntax.h"
#ifndef MEMSET
# define MEMSET __memset_avx2
# define MEMSET_CHK __memset_chk_avx2
#endif
.section .text.avx2,"ax",@progbits
#if defined PIC
ENTRY (MEMSET_CHK)
cmpq %rdx, %rcx
jb HIDDEN_JUMPTARGET (__chk_fail)
END (MEMSET_CHK)
#endif
ENTRY (MEMSET)
vpxor %xmm0, %xmm0, %xmm0
vmovd %esi, %xmm1
lea (%rdi, %rdx), %rsi
mov %rdi, %rax
vpshufb %xmm0, %xmm1, %xmm0
cmp $16, %rdx
jb L(less_16bytes)
cmp $256, %rdx
jae L(256bytesormore)
cmp $128, %dl
jb L(less_128bytes)
vmovdqu %xmm0, (%rdi)
vmovdqu %xmm0, 0x10(%rdi)
vmovdqu %xmm0, 0x20(%rdi)
vmovdqu %xmm0, 0x30(%rdi)
vmovdqu %xmm0, 0x40(%rdi)
vmovdqu %xmm0, 0x50(%rdi)
vmovdqu %xmm0, 0x60(%rdi)
vmovdqu %xmm0, 0x70(%rdi)
vmovdqu %xmm0, -0x80(%rsi)
vmovdqu %xmm0, -0x70(%rsi)
vmovdqu %xmm0, -0x60(%rsi)
vmovdqu %xmm0, -0x50(%rsi)
vmovdqu %xmm0, -0x40(%rsi)
vmovdqu %xmm0, -0x30(%rsi)
vmovdqu %xmm0, -0x20(%rsi)
vmovdqu %xmm0, -0x10(%rsi)
ret
.p2align 4
L(less_128bytes):
cmp $64, %dl
jb L(less_64bytes)
vmovdqu %xmm0, (%rdi)
vmovdqu %xmm0, 0x10(%rdi)
vmovdqu %xmm0, 0x20(%rdi)
vmovdqu %xmm0, 0x30(%rdi)
vmovdqu %xmm0, -0x40(%rsi)
vmovdqu %xmm0, -0x30(%rsi)
vmovdqu %xmm0, -0x20(%rsi)
vmovdqu %xmm0, -0x10(%rsi)
ret
.p2align 4
L(less_64bytes):
cmp $32, %dl
jb L(less_32bytes)
vmovdqu %xmm0, (%rdi)
vmovdqu %xmm0, 0x10(%rdi)
vmovdqu %xmm0, -0x20(%rsi)
vmovdqu %xmm0, -0x10(%rsi)
ret
.p2align 4
L(less_32bytes):
vmovdqu %xmm0, (%rdi)
vmovdqu %xmm0, -0x10(%rsi)
ret
.p2align 4
L(less_16bytes):
cmp $8, %dl
jb L(less_8bytes)
vmovq %xmm0, (%rdi)
vmovq %xmm0, -0x08(%rsi)
ret
.p2align 4
L(less_8bytes):
vmovd %xmm0, %ecx
cmp $4, %dl
jb L(less_4bytes)
mov %ecx, (%rdi)
mov %ecx, -0x04(%rsi)
ret
.p2align 4
L(less_4bytes):
cmp $2, %dl
jb L(less_2bytes)
mov %cx, (%rdi)
mov %cx, -0x02(%rsi)
ret
.p2align 4
L(less_2bytes):
cmp $1, %dl
jb L(less_1bytes)
mov %cl, (%rdi)
L(less_1bytes):
ret
.p2align 4
L(256bytesormore):
vinserti128 $1, %xmm0, %ymm0, %ymm0
and $-0x20, %rdi
add $0x20, %rdi
vmovdqu %ymm0, (%rax)
sub %rdi, %rax
lea -0x80(%rax, %rdx), %rcx
cmp $4096, %rcx
ja L(gobble_data)
L(gobble_128_loop):
vmovdqa %ymm0, (%rdi)
vmovdqa %ymm0, 0x20(%rdi)
vmovdqa %ymm0, 0x40(%rdi)
vmovdqa %ymm0, 0x60(%rdi)
sub $-0x80, %rdi
add $-0x80, %ecx
jb L(gobble_128_loop)
mov %rsi, %rax
vmovdqu %ymm0, -0x80(%rsi)
vmovdqu %ymm0, -0x60(%rsi)
vmovdqu %ymm0, -0x40(%rsi)
vmovdqu %ymm0, -0x20(%rsi)
sub %rdx, %rax
vzeroupper
ret
.p2align 4
L(gobble_data):
sub $-0x80, %rcx
vmovd %xmm0, %eax
rep stosb
mov %rsi, %rax
sub %rdx, %rax
vzeroupper
ret
END (MEMSET)
#endif
|
