author | Mark J. Nelson <Mark.J.Nelson@Sun.COM> |
Wed, 06 Aug 2008 16:29:39 -0600 | |
changeset 7298 | b69e27387f74 |
parent 6812 | febeba71273d |
child 12719 | bd9fb35d09c2 |
permissions | -rw-r--r-- |
0 | 1 |
/* |
2 |
* CDDL HEADER START |
|
3 |
* |
|
4 |
* The contents of this file are subject to the terms of the |
|
6515
10dab2b883e0
6678310 using LD_AUDIT, ld.so.1 calls shared library's .init before library is fully relocated
raf
parents:
0
diff
changeset
|
5 |
* Common Development and Distribution License (the "License"). |
10dab2b883e0
6678310 using LD_AUDIT, ld.so.1 calls shared library's .init before library is fully relocated
raf
parents:
0
diff
changeset
|
6 |
* You may not use this file except in compliance with the License. |
0 | 7 |
* |
8 |
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE |
|
9 |
* or http://www.opensolaris.org/os/licensing. |
|
10 |
* See the License for the specific language governing permissions |
|
11 |
* and limitations under the License. |
|
12 |
* |
|
13 |
* When distributing Covered Code, include this CDDL HEADER in each |
|
14 |
* file and include the License file at usr/src/OPENSOLARIS.LICENSE. |
|
15 |
* If applicable, add the following below this CDDL HEADER, with the |
|
16 |
* fields enclosed by brackets "[]" replaced with your own identifying |
|
17 |
* information: Portions Copyright [yyyy] [name of copyright owner] |
|
18 |
* |
|
19 |
* CDDL HEADER END |
|
20 |
*/ |
|
6515
10dab2b883e0
6678310 using LD_AUDIT, ld.so.1 calls shared library's .init before library is fully relocated
raf
parents:
0
diff
changeset
|
21 |
|
0 | 22 |
/* |
6515
10dab2b883e0
6678310 using LD_AUDIT, ld.so.1 calls shared library's .init before library is fully relocated
raf
parents:
0
diff
changeset
|
23 |
* Copyright 2008 Sun Microsystems, Inc. All rights reserved. |
0 | 24 |
* Use is subject to license terms. |
25 |
*/ |
|
26 |
||
7298
b69e27387f74
6733918 Teamware has retired, please welcome your new manager, Mercurial
Mark J. Nelson <Mark.J.Nelson@Sun.COM>
parents:
6812
diff
changeset
|
27 |
.file "memcpy.s" |
0 | 28 |
|
29 |
/* |
|
30 |
* memcpy(s1, s2, len) |
|
31 |
* |
|
32 |
* Copy s2 to s1, always copy n bytes. |
|
33 |
* Note: this does not work for overlapped copies, bcopy() does |
|
34 |
* |
|
35 |
* Fast assembler language version of the following C-program for memcpy |
|
36 |
* which represents the `standard' for the C-library. |
|
37 |
* |
|
38 |
* void * |
|
39 |
* memcpy(void *s, const void *s0, size_t n) |
|
40 |
* { |
|
41 |
* if (n != 0) { |
|
42 |
* char *s1 = s; |
|
43 |
* const char *s2 = s0; |
|
44 |
* do { |
|
45 |
* *s1++ = *s2++; |
|
46 |
* } while (--n != 0); |
|
47 |
* } |
|
48 |
* return (s); |
|
49 |
* } |
|
50 |
*/ |
|
51 |
||
6812 | 52 |
#include "SYS.h" |
0 | 53 |
|
54 |
ANSI_PRAGMA_WEAK(memcpy,function) |
|
55 |
||
56 |
ENTRY(memcpy) |
|
57 |
st %o0, [%sp + 68] ! save des address for return val |
|
58 |
cmp %o2, 17 ! for small counts copy bytes |
|
59 |
bleu .dbytecp |
|
60 |
andcc %o1, 3, %o5 ! is src word aligned |
|
61 |
bz .aldst |
|
62 |
cmp %o5, 2 ! is src half-word aligned |
|
63 |
be .s2algn |
|
64 |
cmp %o5, 3 ! src is byte aligned |
|
65 |
.s1algn:ldub [%o1], %o3 ! move 1 or 3 bytes to align it |
|
66 |
inc 1, %o1 |
|
67 |
stb %o3, [%o0] ! move a byte to align src |
|
68 |
inc 1, %o0 |
|
69 |
bne .s2algn |
|
70 |
dec %o2 |
|
71 |
b .ald ! now go align dest |
|
72 |
andcc %o0, 3, %o5 |
|
73 |
||
74 |
.s2algn:lduh [%o1], %o3 ! know src is 2 byte alinged |
|
75 |
inc 2, %o1 |
|
76 |
srl %o3, 8, %o4 |
|
77 |
stb %o4, [%o0] ! have to do bytes, |
|
78 |
stb %o3, [%o0 + 1] ! don't know dst alingment |
|
79 |
inc 2, %o0 |
|
80 |
dec 2, %o2 |
|
81 |
||
82 |
.aldst: andcc %o0, 3, %o5 ! align the destination address |
|
83 |
.ald: bz .w4cp |
|
84 |
cmp %o5, 2 |
|
85 |
bz .w2cp |
|
86 |
cmp %o5, 3 |
|
87 |
.w3cp: ld [%o1], %o4 |
|
88 |
inc 4, %o1 |
|
89 |
srl %o4, 24, %o5 |
|
90 |
stb %o5, [%o0] |
|
91 |
bne .w1cp |
|
92 |
inc %o0 |
|
93 |
dec 1, %o2 |
|
94 |
andn %o2, 3, %o3 ! o3 is aligned word count |
|
95 |
dec 4, %o3 ! avoid reading beyond tail of src |
|
96 |
sub %o1, %o0, %o1 ! o1 gets the difference |
|
97 |
||
98 |
1: sll %o4, 8, %g1 ! save residual bytes |
|
99 |
ld [%o1+%o0], %o4 |
|
100 |
deccc 4, %o3 |
|
101 |
srl %o4, 24, %o5 ! merge with residual |
|
102 |
or %o5, %g1, %g1 |
|
103 |
st %g1, [%o0] |
|
104 |
bnz 1b |
|
105 |
inc 4, %o0 |
|
106 |
sub %o1, 3, %o1 ! used one byte of last word read |
|
107 |
and %o2, 3, %o2 |
|
108 |
b 7f |
|
109 |
inc 4, %o2 |
|
110 |
||
111 |
.w1cp: srl %o4, 8, %o5 |
|
112 |
sth %o5, [%o0] |
|
113 |
inc 2, %o0 |
|
114 |
dec 3, %o2 |
|
115 |
andn %o2, 3, %o3 ! o3 is aligned word count |
|
116 |
dec 4, %o3 ! avoid reading beyond tail of src |
|
117 |
sub %o1, %o0, %o1 ! o1 gets the difference |
|
118 |
||
119 |
2: sll %o4, 24, %g1 ! save residual bytes |
|
120 |
ld [%o1+%o0], %o4 |
|
121 |
deccc 4, %o3 |
|
122 |
srl %o4, 8, %o5 ! merge with residual |
|
123 |
or %o5, %g1, %g1 |
|
124 |
st %g1, [%o0] |
|
125 |
bnz 2b |
|
126 |
inc 4, %o0 |
|
127 |
sub %o1, 1, %o1 ! used three bytes of last word read |
|
128 |
and %o2, 3, %o2 |
|
129 |
b 7f |
|
130 |
inc 4, %o2 |
|
131 |
||
132 |
.w2cp: ld [%o1], %o4 |
|
133 |
inc 4, %o1 |
|
134 |
srl %o4, 16, %o5 |
|
135 |
sth %o5, [%o0] |
|
136 |
inc 2, %o0 |
|
137 |
dec 2, %o2 |
|
138 |
andn %o2, 3, %o3 ! o3 is aligned word count |
|
139 |
dec 4, %o3 ! avoid reading beyond tail of src |
|
140 |
sub %o1, %o0, %o1 ! o1 gets the difference |
|
141 |
||
142 |
3: sll %o4, 16, %g1 ! save residual bytes |
|
143 |
ld [%o1+%o0], %o4 |
|
144 |
deccc 4, %o3 |
|
145 |
srl %o4, 16, %o5 ! merge with residual |
|
146 |
or %o5, %g1, %g1 |
|
147 |
st %g1, [%o0] |
|
148 |
bnz 3b |
|
149 |
inc 4, %o0 |
|
150 |
sub %o1, 2, %o1 ! used two bytes of last word read |
|
151 |
and %o2, 3, %o2 |
|
152 |
b 7f |
|
153 |
inc 4, %o2 |
|
154 |
||
155 |
.w4cp: andn %o2, 3, %o3 ! o3 is aligned word count |
|
156 |
sub %o1, %o0, %o1 ! o1 gets the difference |
|
157 |
||
158 |
1: ld [%o1+%o0], %o4 ! read from address |
|
159 |
deccc 4, %o3 ! decrement count |
|
160 |
st %o4, [%o0] ! write at destination address |
|
161 |
bgu 1b |
|
162 |
inc 4, %o0 ! increment to address |
|
163 |
b 7f |
|
164 |
and %o2, 3, %o2 ! number of leftover bytes, if any |
|
165 |
||
166 |
! |
|
167 |
! differenced byte copy, works with any alignment |
|
168 |
! |
|
169 |
.dbytecp: |
|
170 |
b 7f |
|
171 |
sub %o1, %o0, %o1 ! o1 gets the difference |
|
172 |
||
173 |
4: stb %o4, [%o0] ! write to address |
|
174 |
inc %o0 ! inc to address |
|
175 |
7: deccc %o2 ! decrement count |
|
176 |
bgeu,a 4b ! loop till done |
|
177 |
ldub [%o1+%o0], %o4 ! read from address |
|
178 |
retl |
|
179 |
ld [%sp + 68], %o0 ! return s1, destination address |
|
180 |
||
181 |
SET_SIZE(memcpy) |