summaryrefslogtreecommitdiff
path: root/arch/x86/lib/memcpy_32.c
blob: 2eab7d0bfeddc6ddccee9c1318188418ebe0585e (plain) (blame)
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
#include <linux/string.h>
#include <linux/export.h>

#undef memcpy
#undef memset

__visible void *memcpy(void *to, const void *from, size_t n)
{
#if defined(CONFIG_X86_USE_3DNOW) && !defined(CONFIG_FORTIFY_SOURCE)
	return __memcpy3d(to, from, n);
#else
	return __memcpy(to, from, n);
#endif
}
EXPORT_SYMBOL(memcpy);

__visible void *memset(void *s, int c, size_t count)
{
	return __memset(s, c, count);
}
EXPORT_SYMBOL(memset);

__visible void *memmove(void *dest, const void *src, size_t n)
{
	int d0,d1,d2,d3,d4,d5;
	char *ret = dest;

	__asm__ __volatile__(
		/* Handle more 16 bytes in loop */
		"cmp $0x10, %0\n\t"
		"jb	1f\n\t"

		/* Decide forward/backward copy mode */
		"cmp %2, %1\n\t"
		"jb	2f\n\t"

		/*
		 * movs instruction have many startup latency
		 * so we handle small size by general register.
		 */
		"cmp  $680, %0\n\t"
		"jb 3f\n\t"
		/*
		 * movs instruction is only good for aligned case.
		 */
		"mov %1, %3\n\t"
		"xor %2, %3\n\t"
		"and $0xff, %3\n\t"
		"jz 4f\n\t"
		"3:\n\t"
		"sub $0x10, %0\n\t"

		/*
		 * We gobble 16 bytes forward in each loop.
		 */
		"3:\n\t"
		"sub $0x10, %0\n\t"
		"mov 0*4(%1), %3\n\t"
		"mov 1*4(%1), %4\n\t"
		"mov  %3, 0*4(%2)\n\t"
		"mov  %4, 1*4(%2)\n\t"
		"mov 2*4(%1), %3\n\t"
		"mov 3*4(%1), %4\n\t"
		"mov  %3, 2*4(%2)\n\t"
		"mov  %4, 3*4(%2)\n\t"
		"lea  0x10(%1), %1\n\t"
		"lea  0x10(%2), %2\n\t"
		"jae 3b\n\t"
		"add $0x10, %0\n\t"
		"jmp 1f\n\t"

		/*
		 * Handle data forward by movs.
		 */
		".p2align 4\n\t"
		"4:\n\t"
		"mov -4(%1, %0), %3\n\t"
		"lea -4(%2, %0), %4\n\t"
		"shr $2, %0\n\t"
		"rep movsl\n\t"
		"mov %3, (%4)\n\t"
		"jmp 11f\n\t"
		/*
		 * Handle data backward by movs.
		 */
		".p2align 4\n\t"
		"6:\n\t"
		"mov (%1), %3\n\t"
		"mov %2, %4\n\t"
		"lea -4(%1, %0), %1\n\t"
		"lea -4(%2, %0), %2\n\t"
		"shr $2, %0\n\t"
		"std\n\t"
		"rep movsl\n\t"
		"mov %3,(%4)\n\t"
		"cld\n\t"
		"jmp 11f\n\t"

		/*
		 * Start to prepare for backward copy.
		 */
		".p2align 4\n\t"
		"2:\n\t"
		"cmp  $680, %0\n\t"
		"jb 5f\n\t"
		"mov %1, %3\n\t"
		"xor %2, %3\n\t"
		"and $0xff, %3\n\t"
		"jz 6b\n\t"

		/*
		 * Calculate copy position to tail.
		 */
		"5:\n\t"
		"add %0, %1\n\t"
		"add %0, %2\n\t"
		"sub $0x10, %0\n\t"

		/*
		 * We gobble 16 bytes backward in each loop.
		 */
		"7:\n\t"
		"sub $0x10, %0\n\t"

		"mov -1*4(%1), %3\n\t"
		"mov -2*4(%1), %4\n\t"
		"mov  %3, -1*4(%2)\n\t"
		"mov  %4, -2*4(%2)\n\t"
		"mov -3*4(%1), %3\n\t"
		"mov -4*4(%1), %4\n\t"
		"mov  %3, -3*4(%2)\n\t"
		"mov  %4, -4*4(%2)\n\t"
		"lea  -0x10(%1), %1\n\t"
		"lea  -0x10(%2), %2\n\t"
		"jae 7b\n\t"
		/*
		 * Calculate copy position to head.
		 */
		"add $0x10, %0\n\t"
		"sub %0, %1\n\t"
		"sub %0, %2\n\t"

		/*
		 * Move data from 8 bytes to 15 bytes.
		 */
		".p2align 4\n\t"
		"1:\n\t"
		"cmp $8, %0\n\t"
		"jb 8f\n\t"
		"mov 0*4(%1), %3\n\t"
		"mov 1*4(%1), %4\n\t"
		"mov -2*4(%1, %0), %5\n\t"
		"mov -1*4(%1, %0), %1\n\t"

		"mov  %3, 0*4(%2)\n\t"
		"mov  %4, 1*4(%2)\n\t"
		"mov  %5, -2*4(%2, %0)\n\t"
		"mov  %1, -1*4(%2, %0)\n\t"
		"jmp 11f\n\t"

		/*
		 * Move data from 4 bytes to 7 bytes.
		 */
		".p2align 4\n\t"
		"8:\n\t"
		"cmp $4, %0\n\t"
		"jb 9f\n\t"
		"mov 0*4(%1), %3\n\t"
		"mov -1*4(%1, %0), %4\n\t"
		"mov  %3, 0*4(%2)\n\t"
		"mov  %4, -1*4(%2, %0)\n\t"
		"jmp 11f\n\t"

		/*
		 * Move data from 2 bytes to 3 bytes.
		 */
		".p2align 4\n\t"
		"9:\n\t"
		"cmp $2, %0\n\t"
		"jb 10f\n\t"
		"movw 0*2(%1), %%dx\n\t"
		"movw -1*2(%1, %0), %%bx\n\t"
		"movw %%dx, 0*2(%2)\n\t"
		"movw %%bx, -1*2(%2, %0)\n\t"
		"jmp 11f\n\t"

		/*
		 * Move data for 1 byte.
		 */
		".p2align 4\n\t"
		"10:\n\t"
		"cmp $1, %0\n\t"
		"jb 11f\n\t"
		"movb (%1), %%cl\n\t"
		"movb %%cl, (%2)\n\t"
		".p2align 4\n\t"
		"11:"
		: "=&c" (d0), "=&S" (d1), "=&D" (d2),
		  "=r" (d3),"=r" (d4), "=r"(d5)
		:"0" (n),
		 "1" (src),
		 "2" (dest)
		:"memory");

	return ret;

}
EXPORT_SYMBOL(memmove);