1 /* $NetBSD: mul.S,v 1.3 1997/07/16 14:37:42 christos Exp $ */
4 * Copyright (c) 1992, 1993
5 * The Regents of the University of California. All rights reserved.
7 * This software was developed by the Computer Systems Engineering group
8 * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
9 * contributed to Berkeley.
11 * Redistribution and use in source and binary forms, with or without
12 * modification, are permitted provided that the following conditions
14 * 1. Redistributions of source code must retain the above copyright
15 * notice, this list of conditions and the following disclaimer.
16 * 2. Redistributions in binary form must reproduce the above copyright
17 * notice, this list of conditions and the following disclaimer in the
18 * documentation and/or other materials provided with the distribution.
19 * 3. All advertising materials mentioning features or use of this software
20 * must display the following acknowledgement:
21 * This product includes software developed by the University of
22 * California, Berkeley and its contributors.
23 * 4. Neither the name of the University nor the names of its contributors
24 * may be used to endorse or promote products derived from this software
25 * without specific prior written permission.
27 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
28 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
31 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
32 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
33 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
34 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
35 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
36 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
39 * from: Header: mul.s,v 1.5 92/06/25 13:24:03 torek Exp
42 #include <machine/asm.h>
43 #if defined(LIBC_SCCS) && !defined(lint)
45 .asciz "@(#)mul.s 8.1 (Berkeley) 6/4/93"
47 RCSID("$NetBSD: mul.S,v 1.3 1997/07/16 14:37:42 christos Exp $")
49 #endif /* LIBC_SCCS and not lint */
52 * Signed multiply, from Appendix E of the Sparc Version 8
53 * Architecture Manual.
55 * Returns %o0 * %o1 in %o1%o0 (i.e., %o1 holds the upper 32 bits of
56 * the 64-bit product).
58 * This code optimizes short (less than 13-bit) multiplies.
62 mov %o0, %y ! multiplier -> Y
63 andncc %o0, 0xfff, %g0 ! test bits 12..31
64 be Lmul_shortway ! if zero, can do it the short way
65 andcc %g0, %g0, %o4 ! zero the partial product and clear N and V
68 * Long multiply. 32 steps, followed by a final shift step.
70 mulscc %o4, %o1, %o4 ! 1
71 mulscc %o4, %o1, %o4 ! 2
72 mulscc %o4, %o1, %o4 ! 3
73 mulscc %o4, %o1, %o4 ! 4
74 mulscc %o4, %o1, %o4 ! 5
75 mulscc %o4, %o1, %o4 ! 6
76 mulscc %o4, %o1, %o4 ! 7
77 mulscc %o4, %o1, %o4 ! 8
78 mulscc %o4, %o1, %o4 ! 9
79 mulscc %o4, %o1, %o4 ! 10
80 mulscc %o4, %o1, %o4 ! 11
81 mulscc %o4, %o1, %o4 ! 12
82 mulscc %o4, %o1, %o4 ! 13
83 mulscc %o4, %o1, %o4 ! 14
84 mulscc %o4, %o1, %o4 ! 15
85 mulscc %o4, %o1, %o4 ! 16
86 mulscc %o4, %o1, %o4 ! 17
87 mulscc %o4, %o1, %o4 ! 18
88 mulscc %o4, %o1, %o4 ! 19
89 mulscc %o4, %o1, %o4 ! 20
90 mulscc %o4, %o1, %o4 ! 21
91 mulscc %o4, %o1, %o4 ! 22
92 mulscc %o4, %o1, %o4 ! 23
93 mulscc %o4, %o1, %o4 ! 24
94 mulscc %o4, %o1, %o4 ! 25
95 mulscc %o4, %o1, %o4 ! 26
96 mulscc %o4, %o1, %o4 ! 27
97 mulscc %o4, %o1, %o4 ! 28
98 mulscc %o4, %o1, %o4 ! 29
99 mulscc %o4, %o1, %o4 ! 30
100 mulscc %o4, %o1, %o4 ! 31
101 mulscc %o4, %o1, %o4 ! 32
102 mulscc %o4, %g0, %o4 ! final shift
104 ! If %o0 was negative, the result is
105 ! (%o0 * %o1) + (%o1 << 32))
112 ! %o0 was indeed negative; fix upper 32 bits of result by subtracting
113 ! %o1 (i.e., return %o4 - %o1 in %o1).
123 * Short multiply. 12 steps, followed by a final shift step.
124 * The resulting bits are off by 12 and (32-12) = 20 bit positions,
125 * but there is no problem with %o0 being negative (unlike above).
127 mulscc %o4, %o1, %o4 ! 1
128 mulscc %o4, %o1, %o4 ! 2
129 mulscc %o4, %o1, %o4 ! 3
130 mulscc %o4, %o1, %o4 ! 4
131 mulscc %o4, %o1, %o4 ! 5
132 mulscc %o4, %o1, %o4 ! 6
133 mulscc %o4, %o1, %o4 ! 7
134 mulscc %o4, %o1, %o4 ! 8
135 mulscc %o4, %o1, %o4 ! 9
136 mulscc %o4, %o1, %o4 ! 10
137 mulscc %o4, %o1, %o4 ! 11
138 mulscc %o4, %o1, %o4 ! 12
139 mulscc %o4, %g0, %o4 ! final shift
142 * %o4 has 20 of the bits that should be in the low part of the
143 * result; %y has the bottom 12 (as %y's top 12). That is:
146 * +----------------+----------------+
147 * | -12- | -20- | -12- | -20- |
148 * +------(---------+------)---------+
149 * --hi-- ----low-part----
151 * The upper 12 bits of %o4 should be sign-extended to form the
152 * high part of the product (i.e., highpart = %o4 >> 20).
156 sll %o4, 12, %o0 ! shift middle bits left 12
157 srl %o5, 20, %o5 ! shift low bits right 20, zero fill at left
158 or %o5, %o0, %o0 ! construct low part of result
160 sra %o4, 20, %o1 ! ... and extract high part of result