xref: /titanic_44/usr/src/lib/libc/sparcv9/gen/memcpy.s (revision 90bcde942a3919300ffc73f98ea903b58386c395)
1/*
2 * CDDL HEADER START
3 *
4 * The contents of this file are subject to the terms of the
5 * Common Development and Distribution License (the "License").
6 * You may not use this file except in compliance with the License.
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 */
21
22/*
23 * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
24 * Use is subject to license terms.
25 */
26
27#pragma ident	"%Z%%M%	%I%	%E% SMI"
28
29	.file	"%M%"
30
31/*
32 * memcpy(s1, s2, len)
33 *
34 * Copy s2 to s1, always copy n bytes.
35 * Note: this does not work for overlapped copies, bcopy() does
36 *
37 * Added entry __align_cpy_1 is generally for use of the compilers.
38 *
39 *
40 * Fast assembler language version of the following C-program for memcpy
41 * which represents the `standard' for the C-library.
42 *
43 *	void *
44 *	memcpy(void *s, const void *s0, size_t n)
45 *	{
46 *		if (n != 0) {
47 *			char *s1 = s;
48 *			const char *s2 = s0;
49 *			do {
50 *				*s1++ = *s2++;
51 *			} while (--n != 0);
52 *		}
53 *		return (s);
54 *	}
55 */
56
57#include <sys/asm_linkage.h>
58
59	ANSI_PRAGMA_WEAK(memcpy,function)
60
61	ENTRY(memcpy)
62	ENTRY(__align_cpy_1)
63	mov	%o0, %g5		! save des address for return val
64	cmp	%o2, 17			! for small counts copy bytes
65	bleu,pn	%xcc, .dbytecp
66	andcc	%o1, 3, %o5		! is src word aligned
67	bz,pn	%icc, .aldst
68	cmp	%o5, 2			! is src half-word aligned
69	be,pt	%xcc, .s2algn
70	cmp	%o5, 3			! src is byte aligned
71.s1algn:ldub	[%o1], %o3		! move 1 or 3 bytes to align it
72	inc	1, %o1
73	stb	%o3, [%g5]		! move a byte to align src
74	inc	1, %g5
75	bne,pt	%icc, .s2algn
76	dec	%o2
77	b	.ald			! now go align dest
78	andcc	%g5, 3, %o5
79
80.s2algn:lduh	[%o1], %o3		! know src is 2 byte alinged
81	inc	2, %o1
82	srl	%o3, 8, %o4
83	stb	%o4, [%g5]		! have to do bytes,
84	stb	%o3, [%g5 + 1]		! don't know dst alingment
85	inc	2, %g5
86	dec	2, %o2
87
88.aldst:	andcc	%g5, 3, %o5		! align the destination address
89.ald:	bz,pn	%icc, .w4cp
90	cmp	%o5, 2
91	bz,pn	%icc, .w2cp
92	cmp	%o5, 3
93.w3cp:	lduw	[%o1], %o4
94	inc	4, %o1
95	srl	%o4, 24, %o5
96	stb	%o5, [%g5]
97	bne,pt	%icc, .w1cp
98	inc	%g5
99	dec	1, %o2
100	andn	%o2, 3, %o3		! o3 is aligned word count
101	dec	4, %o3			! avoid reading beyond tail of src
102	sub	%o1, %g5, %o1		! o1 gets the difference
103
1041:	sll	%o4, 8, %g1		! save residual bytes
105	lduw	[%o1+%g5], %o4
106	deccc	4, %o3
107	srl	%o4, 24, %o5		! merge with residual
108	or	%o5, %g1, %g1
109	st	%g1, [%g5]
110	bnz,pt	%xcc, 1b
111	inc	4, %g5
112	sub	%o1, 3, %o1		! used one byte of last word read
113	and	%o2, 3, %o2
114	b	7f
115	inc	4, %o2
116
117.w1cp:	srl	%o4, 8, %o5
118	sth	%o5, [%g5]
119	inc	2, %g5
120	dec	3, %o2
121	andn	%o2, 3, %o3		! o3 is aligned word count
122	dec	4, %o3			! avoid reading beyond tail of src
123	sub	%o1, %g5, %o1		! o1 gets the difference
124
1252:	sll	%o4, 24, %g1		! save residual bytes
126	lduw	[%o1+%g5], %o4
127	deccc	4, %o3
128	srl	%o4, 8, %o5		! merge with residual
129	or	%o5, %g1, %g1
130	st	%g1, [%g5]
131	bnz,pt	%xcc, 2b
132	inc	4, %g5
133	sub	%o1, 1, %o1		! used three bytes of last word read
134	and	%o2, 3, %o2
135	b	7f
136	inc	4, %o2
137
138.w2cp:	lduw	[%o1], %o4
139	inc	4, %o1
140	srl	%o4, 16, %o5
141	sth	%o5, [%g5]
142	inc	2, %g5
143	dec	2, %o2
144	andn	%o2, 3, %o3		! o3 is aligned word count
145	dec	4, %o3			! avoid reading beyond tail of src
146	sub	%o1, %g5, %o1		! o1 gets the difference
147
1483:	sll	%o4, 16, %g1		! save residual bytes
149	lduw	[%o1+%g5], %o4
150	deccc	4, %o3
151	srl	%o4, 16, %o5		! merge with residual
152	or	%o5, %g1, %g1
153	st	%g1, [%g5]
154	bnz,pt	%xcc, 3b
155	inc	4, %g5
156	sub	%o1, 2, %o1		! used two bytes of last word read
157	and	%o2, 3, %o2
158	b	7f
159	inc	4, %o2
160
161.w4cp:	andn	%o2, 3, %o3		! o3 is aligned word count
162	sub	%o1, %g5, %o1		! o1 gets the difference
163
1641:	lduw	[%o1+%g5], %o4		! read from address
165	deccc	4, %o3			! decrement count
166	st	%o4, [%g5]		! write at destination address
167	bgu,pt	%xcc, 1b
168	inc	4, %g5			! increment to address
169	b	7f
170	and	%o2, 3, %o2		! number of leftover bytes, if any
171
172	!
173	! differenced byte copy, works with any alignment
174	!
175.dbytecp:
176	b	7f
177	sub	%o1, %g5, %o1		! o1 gets the difference
178
1794:	stb	%o4, [%g5]		! write to address
180	inc	%g5			! inc to address
1817:	deccc	%o2			! decrement count
182	bgeu,a,pt %xcc,4b		! loop till done
183	ldub	[%o1+%g5], %o4		! read from address
184	retl
185	nop
186
187	SET_SIZE(memcpy)
188	SET_SIZE(__align_cpy_1)
189