Text file src/runtime/memmove_plan9_386.s
1 // Inferno's libkern/memmove-386.s
2 // https://bitbucket.org/inferno-os/inferno-os/src/default/libkern/memmove-386.s
3 //
4 // Copyright © 1994-1999 Lucent Technologies Inc. All rights reserved.
5 // Revisions Copyright © 2000-2007 Vita Nuova Holdings Limited (www.vitanuova.com). All rights reserved.
6 // Portions Copyright 2009 The Go Authors. All rights reserved.
7 //
8 // Permission is hereby granted, free of charge, to any person obtaining a copy
9 // of this software and associated documentation files (the "Software"), to deal
10 // in the Software without restriction, including without limitation the rights
11 // to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
12 // copies of the Software, and to permit persons to whom the Software is
13 // furnished to do so, subject to the following conditions:
14 //
15 // The above copyright notice and this permission notice shall be included in
16 // all copies or substantial portions of the Software.
17 //
18 // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
19 // IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
20 // FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
21 // AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
22 // LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
23 // OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
24 // THE SOFTWARE.
25
26 #include "textflag.h"
27
28 // func memmove(to, from unsafe.Pointer, n uintptr)
29 TEXT runtime·memmove(SB), NOSPLIT, $0-12
30 MOVL to+0(FP), DI
31 MOVL from+4(FP), SI
32 MOVL n+8(FP), BX
33
34 // REP instructions have a high startup cost, so we handle small sizes
35 // with some straightline code. The REP MOVSL instruction is really fast
36 // for large sizes. The cutover is approximately 1K.
37 tail:
38 TESTL BX, BX
39 JEQ move_0
40 CMPL BX, $2
41 JBE move_1or2
42 CMPL BX, $4
43 JB move_3
44 JE move_4
45 CMPL BX, $8
46 JBE move_5through8
47 CMPL BX, $16
48 JBE move_9through16
49
50 /*
51 * check and set for backwards
52 */
53 CMPL SI, DI
54 JLS back
55
56 /*
57 * forward copy loop
58 */
59 forward:
60 MOVL BX, CX
61 SHRL $2, CX
62 ANDL $3, BX
63
64 REP; MOVSL
65 JMP tail
66 /*
67 * check overlap
68 */
69 back:
70 MOVL SI, CX
71 ADDL BX, CX
72 CMPL CX, DI
73 JLS forward
74 /*
75 * whole thing backwards has
76 * adjusted addresses
77 */
78
79 ADDL BX, DI
80 ADDL BX, SI
81 STD
82
83 /*
84 * copy
85 */
86 MOVL BX, CX
87 SHRL $2, CX
88 ANDL $3, BX
89
90 SUBL $4, DI
91 SUBL $4, SI
92 REP; MOVSL
93
94 CLD
95 ADDL $4, DI
96 ADDL $4, SI
97 SUBL BX, DI
98 SUBL BX, SI
99 JMP tail
100
101 move_1or2:
102 MOVB (SI), AX
103 MOVB -1(SI)(BX*1), CX
104 MOVB AX, (DI)
105 MOVB CX, -1(DI)(BX*1)
106 RET
107 move_0:
108 RET
109 move_3:
110 MOVW (SI), AX
111 MOVB 2(SI), CX
112 MOVW AX, (DI)
113 MOVB CX, 2(DI)
114 RET
115 move_4:
116 // We need a separate case for 4 to make sure we write pointers atomically.
117 MOVL (SI), AX
118 MOVL AX, (DI)
119 RET
120 move_5through8:
121 MOVL (SI), AX
122 MOVL -4(SI)(BX*1), CX
123 MOVL AX, (DI)
124 MOVL CX, -4(DI)(BX*1)
125 RET
126 move_9through16:
127 MOVL (SI), AX
128 MOVL 4(SI), CX
129 MOVL -8(SI)(BX*1), DX
130 MOVL -4(SI)(BX*1), BP
131 MOVL AX, (DI)
132 MOVL CX, 4(DI)
133 MOVL DX, -8(DI)(BX*1)
134 MOVL BP, -4(DI)(BX*1)
135 RET
View as plain text