summaryrefslogtreecommitdiffstats
path: root/arch/arm/lib/string.S
blob: 17daa9d2639b720097f1f12531ef3029cdcbad93 (plain)
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
/*
 * linux/arch/arm/lib/string.S
 *
 * Copyright (C) 1995-1998 Russell King
 *
 * This is commonly used to clear the frame buffer and the frame
 * backing buffer.  As such, it will be rarely called with r2 < 32.
 *
 * Optimisations by Matthew Wilcox
 */
#include <linux/linkage.h>
#include <asm/assembler.h>
		.text
# Prototype: char *strrchr(const char *s,char c);

/*
 * Prototype: void memzero(void *d, size_t n)
 */
ENTRY(memzero)
		mov	r2, r1
		mov	r1, #0
/*
 * Prototype: void memsetl(unsigned long *d, unsigned long c, size_t n)
 */
ENTRY(memsetl)
		teq	r2, #0
		RETINSTR(moveq,pc,lr)
		stmfd	sp!, {lr}
		mov	lr, r1
		mov	r3, r1
		mov	ip, r1

		@ r2 = {32 ... 4}

1:		subs	r2, r2, #32
		stmgeia	r0!, {r1, r3, ip, lr}
		stmgeia	r0!, {r1, r3, ip, lr}
		bgt	1b
		LOADREGS(eqfd, sp!, {pc})

		@ r2 can be {-4 ... -28}

		cmp	r2, #-16
		stmgeia	r0!, {r1, r3, ip, lr}
		addlts	r2, r2, #16
		LOADREGS(eqfd, sp!, {pc})

		@ r2 can be {-4 ... -12}

		cmp	r2, #-8
		stmgeia	r0!, {r1, r3}
		strne	r1, [r0]
		LOADREGS(fd, sp!, {pc})

		.global	__page_memcpy
__page_memcpy:	stmfd	sp!, {r4, r5, lr}
1:		subs	r2, r2, #4*8
		ldmgeia	r1!, {r3, r4, r5, ip}
		stmgeia	r0!, {r3, r4, r5, ip}
		ldmgeia	r1!, {r3, r4, r5, ip}
		stmgeia	r0!, {r3, r4, r5, ip}
		bgt	1b
		LOADREGS(fd, sp!, {r4, r5, pc})

		.global	memset
memset:		mov	r3, r0
		cmp	r2, #16
		blt	6f
		ands	ip, r3, #3
		beq	1f
		cmp	ip, #2
		strltb	r1, [r3], #1			@ Align destination
		strleb	r1, [r3], #1
		strb	r1, [r3], #1
		rsb	ip, ip, #4
		sub	r2, r2, ip
1:		orr	r1, r1, r1, lsl #8
		orr	r1, r1, r1, lsl #16
		cmp	r2, #256
		blt	4f
		stmfd	sp!, {r4, r5, lr}
		mov	r4, r1
		mov	r5, r1
		mov	lr, r1
		mov	ip, r2, lsr #6
		sub	r2, r2, ip, lsl #6
2:		stmia	r3!, {r1, r4, r5, lr}		@ 64 bytes at a time.
		stmia	r3!, {r1, r4, r5, lr}
		stmia	r3!, {r1, r4, r5, lr}
		stmia	r3!, {r1, r4, r5, lr}
		subs	ip, ip, #1
		bne	2b
		teq	r2, #0
		LOADREGS(eqfd, sp!, {r4, r5, pc})	@ Now <64 bytes to go.
		tst	r2, #32
		stmneia	r3!, {r1, r4, r5, lr}
		stmneia	r3!, {r1, r4, r5, lr}
		tst	r2, #16
		stmneia	r3!, {r1, r4, r5, lr}
		ldmia	sp!, {r4, r5}
3:		tst	r2, #8
		stmneia	r3!, {r1, lr}
		tst	r2, #4
		strne	r1, [r3], #4
		tst	r2, #2
		strneb	r1, [r3], #1
		strneb	r1, [r3], #1
		tst	r2, #1
		strneb	r1, [r3], #1
		LOADREGS(fd, sp!, {pc})

4:		movs	ip, r2, lsr #3
		beq	3b
		sub	r2, r2, ip, lsl #3
		stmfd	sp!, {lr}
		mov	lr, r1
		subs	ip, ip, #4
5:		stmgeia	r3!, {r1, lr}
		stmgeia	r3!, {r1, lr}
		stmgeia	r3!, {r1, lr}
		stmgeia	r3!, {r1, lr}
		subges	ip, ip, #4
		bge	5b
		tst	ip, #2
		stmneia	r3!, {r1, lr}
		stmneia	r3!, {r1, lr}
		tst	ip, #1
		stmneia	r3!, {r1, lr}
		teq	r2, #0
		LOADREGS(eqfd, sp!, {pc})
		b	3b

6:		subs	r2, r2, #1
		strgeb	r1, [r3], #1
		bgt	6b
		RETINSTR(mov, pc, lr)

ENTRY(strrchr)
		stmfd	sp!, {lr}
		mov	r3, #0
1:		ldrb	r2, [r0], #1
		teq	r2, r1
		moveq	r3, r0
		teq	r2, #0
		bne	1b
		mov	r0, r3
		LOADREGS(fd, sp!, {pc})

ENTRY(strchr)
		stmfd	sp!,{lr}
		mov	r3, #0
1:		ldrb	r2, [r0], #1
		teq	r2, r1
		teqne	r2, #0
		bne	1b
		teq	r2, #0
		moveq	r0, #0
		subne	r0, r0, #1
		LOADREGS(fd, sp!, {pc})

ENTRY(memchr)
		stmfd	sp!, {lr}
1:		ldrb	r3, [r0], #1
		teq	r3, r1
		beq	2f
		subs	r2, r2, #1
		bpl	1b
2:		movne	r0, #0
		subeq	r0, r0, #1
		LOADREGS(fd, sp!, {pc})