1/* $OpenBSD: mul.S,v 1.6 2005/08/07 16:40:15 espie Exp $ */
2/*
3 * Copyright (c) 1992, 1993
4 *	The Regents of the University of California.  All rights reserved.
5 *
6 * This software was developed by the Computer Systems Engineering group
7 * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
8 * contributed to Berkeley.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 *    notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 *    notice, this list of conditions and the following disclaimer in the
17 *    documentation and/or other materials provided with the distribution.
18 * 3. Neither the name of the University nor the names of its contributors
19 *    may be used to endorse or promote products derived from this software
20 *    without specific prior written permission.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * SUCH DAMAGE.
33 */
34
35#include <machine/asm.h>
36
37RCSID("$MirOS: src/kern/c/sparc/mul.S,v 1.1 2008/08/01 18:31:03 tg Exp $")
38
39	.text
40
41/*
42 * Signed multiply, from Appendix E of the Sparc Version 8
43 * Architecture Manual.
44 *
45 * Returns %o0 * %o1 in %o1%o0 (i.e., %o1 holds the upper 32 bits of
46 * the 64-bit product).
47 *
48 * This code optimizes short (less than 13-bit) multiplies.
49 */
50
51#ifndef STRONG_SPARC
52.weak .mul
53#else
54FUNC(__mul)
55#endif
56FUNC(.mul)
57	mov	%o0, %y		! multiplier -> Y
58	andncc	%o0, 0xfff, %g0	! test bits 12..31
59	be	Lmul_shortway	! if zero, can do it the short way
60	andcc	%g0, %g0, %o4	! zero the partial product and clear N and V
61
62	/*
63	 * Long multiply.  32 steps, followed by a final shift step.
64	 */
65	mulscc	%o4, %o1, %o4	! 1
66	mulscc	%o4, %o1, %o4	! 2
67	mulscc	%o4, %o1, %o4	! 3
68	mulscc	%o4, %o1, %o4	! 4
69	mulscc	%o4, %o1, %o4	! 5
70	mulscc	%o4, %o1, %o4	! 6
71	mulscc	%o4, %o1, %o4	! 7
72	mulscc	%o4, %o1, %o4	! 8
73	mulscc	%o4, %o1, %o4	! 9
74	mulscc	%o4, %o1, %o4	! 10
75	mulscc	%o4, %o1, %o4	! 11
76	mulscc	%o4, %o1, %o4	! 12
77	mulscc	%o4, %o1, %o4	! 13
78	mulscc	%o4, %o1, %o4	! 14
79	mulscc	%o4, %o1, %o4	! 15
80	mulscc	%o4, %o1, %o4	! 16
81	mulscc	%o4, %o1, %o4	! 17
82	mulscc	%o4, %o1, %o4	! 18
83	mulscc	%o4, %o1, %o4	! 19
84	mulscc	%o4, %o1, %o4	! 20
85	mulscc	%o4, %o1, %o4	! 21
86	mulscc	%o4, %o1, %o4	! 22
87	mulscc	%o4, %o1, %o4	! 23
88	mulscc	%o4, %o1, %o4	! 24
89	mulscc	%o4, %o1, %o4	! 25
90	mulscc	%o4, %o1, %o4	! 26
91	mulscc	%o4, %o1, %o4	! 27
92	mulscc	%o4, %o1, %o4	! 28
93	mulscc	%o4, %o1, %o4	! 29
94	mulscc	%o4, %o1, %o4	! 30
95	mulscc	%o4, %o1, %o4	! 31
96	mulscc	%o4, %o1, %o4	! 32
97	mulscc	%o4, %g0, %o4	! final shift
98
99	! If %o0 was negative, the result is
100	!	(%o0 * %o1) + (%o1 << 32))
101	! We fix that here.
102
103	tst	%o0
104	bge	1f
105	rd	%y, %o0
106
107	! %o0 was indeed negative; fix upper 32 bits of result by subtracting
108	! %o1 (i.e., return %o4 - %o1 in %o1).
109	retl
110	 sub	%o4, %o1, %o1
111
1121:
113	retl
114	mov	%o4, %o1
115
116Lmul_shortway:
117	/*
118	 * Short multiply.  12 steps, followed by a final shift step.
119	 * The resulting bits are off by 12 and (32-12) = 20 bit positions,
120	 * but there is no problem with %o0 being negative (unlike above).
121	 */
122	mulscc	%o4, %o1, %o4	! 1
123	mulscc	%o4, %o1, %o4	! 2
124	mulscc	%o4, %o1, %o4	! 3
125	mulscc	%o4, %o1, %o4	! 4
126	mulscc	%o4, %o1, %o4	! 5
127	mulscc	%o4, %o1, %o4	! 6
128	mulscc	%o4, %o1, %o4	! 7
129	mulscc	%o4, %o1, %o4	! 8
130	mulscc	%o4, %o1, %o4	! 9
131	mulscc	%o4, %o1, %o4	! 10
132	mulscc	%o4, %o1, %o4	! 11
133	mulscc	%o4, %o1, %o4	! 12
134	mulscc	%o4, %g0, %o4	! final shift
135
136	/*
137	 *  %o4 has 20 of the bits that should be in the low part of the
138	 * result; %y has the bottom 12 (as %y's top 12).  That is:
139	 *
140	 *	  %o4		    %y
141	 * +----------------+----------------+
142	 * | -12- |   -20-  | -12- |   -20-  |
143	 * +------(---------+------)---------+
144	 *  --hi-- ----low-part----
145	 *
146	 * The upper 12 bits of %o4 should be sign-extended to form the
147	 * high part of the product (i.e., highpart = %o4 >> 20).
148	 */
149
150	rd	%y, %o5
151	sll	%o4, 12, %o0	! shift middle bits left 12
152	srl	%o5, 20, %o5	! shift low bits right 20, zero fill at left
153	or	%o5, %o0, %o0	! construct low part of result
154	retl
155	 sra	%o4, 20, %o1	! ... and extract high part of result
156