1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
|
/*
* Copyright (c) 1992, 1993
* The Regents of the University of California. All rights reserved.
*
* This software was developed by the Computer Systems Engineering group
* at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
* contributed to Berkeley.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#if defined(LIBC_SCCS)
.text
.asciz "$OpenBSD: umul.S,v 1.5 2003/06/02 20:18:32 millert Exp $"
#endif /* LIBC_SCCS */
/*
* Unsigned multiply. Returns %o0 * %o1 in %o1%o0 (i.e., %o1 holds the
* upper 32 bits of the 64-bit product).
*
* This code optimizes short (less than 13-bit) multiplies. Short
* multiplies require 25 instruction cycles, and long ones require
* 45 instruction cycles.
*
* On return, overflow has occurred (%o1 is not zero) if and only if
* the Z condition code is clear, allowing, e.g., the following:
*
* call .umul
* nop
* bnz overflow (or tnz)
*/
#include "DEFS.h"
#ifndef STRONG_SPARC
.weak .umul
#else
FUNC(__umul)
#endif
FUNC(.umul)
or %o0, %o1, %o4
mov %o0, %y ! multiplier -> Y
andncc %o4, 0xfff, %g0 ! test bits 12..31 of *both* args
be Lmul_shortway ! if zero, can do it the short way
andcc %g0, %g0, %o4 ! zero the partial product and clear N and V
/*
* Long multiply. 32 steps, followed by a final shift step.
*/
mulscc %o4, %o1, %o4 ! 1
mulscc %o4, %o1, %o4 ! 2
mulscc %o4, %o1, %o4 ! 3
mulscc %o4, %o1, %o4 ! 4
mulscc %o4, %o1, %o4 ! 5
mulscc %o4, %o1, %o4 ! 6
mulscc %o4, %o1, %o4 ! 7
mulscc %o4, %o1, %o4 ! 8
mulscc %o4, %o1, %o4 ! 9
mulscc %o4, %o1, %o4 ! 10
mulscc %o4, %o1, %o4 ! 11
mulscc %o4, %o1, %o4 ! 12
mulscc %o4, %o1, %o4 ! 13
mulscc %o4, %o1, %o4 ! 14
mulscc %o4, %o1, %o4 ! 15
mulscc %o4, %o1, %o4 ! 16
mulscc %o4, %o1, %o4 ! 17
mulscc %o4, %o1, %o4 ! 18
mulscc %o4, %o1, %o4 ! 19
mulscc %o4, %o1, %o4 ! 20
mulscc %o4, %o1, %o4 ! 21
mulscc %o4, %o1, %o4 ! 22
mulscc %o4, %o1, %o4 ! 23
mulscc %o4, %o1, %o4 ! 24
mulscc %o4, %o1, %o4 ! 25
mulscc %o4, %o1, %o4 ! 26
mulscc %o4, %o1, %o4 ! 27
mulscc %o4, %o1, %o4 ! 28
mulscc %o4, %o1, %o4 ! 29
mulscc %o4, %o1, %o4 ! 30
mulscc %o4, %o1, %o4 ! 31
mulscc %o4, %o1, %o4 ! 32
mulscc %o4, %g0, %o4 ! final shift
/*
* Normally, with the shift-and-add approach, if both numbers are
* positive you get the correct result. WIth 32-bit two's-complement
* numbers, -x is represented as
*
* x 32
* ( 2 - ------ ) mod 2 * 2
* 32
* 2
*
* (the `mod 2' subtracts 1 from 1.bbbb). To avoid lots of 2^32s,
* we can treat this as if the radix point were just to the left
* of the sign bit (multiply by 2^32), and get
*
* -x = (2 - x) mod 2
*
* Then, ignoring the `mod 2's for convenience:
*
* x * y = xy
* -x * y = 2y - xy
* x * -y = 2x - xy
* -x * -y = 4 - 2x - 2y + xy
*
* For signed multiplies, we subtract (x << 32) from the partial
* product to fix this problem for negative multipliers (see mul.s).
* Because of the way the shift into the partial product is calculated
* (N xor V), this term is automatically removed for the multiplicand,
* so we don't have to adjust.
*
* But for unsigned multiplies, the high order bit wasn't a sign bit,
* and the correction is wrong. So for unsigned multiplies where the
* high order bit is one, we end up with xy - (y << 32). To fix it
* we add y << 32.
*/
tst %o1
bl,a 1f ! if %o1 < 0 (high order bit = 1),
add %o4, %o0, %o4 ! %o4 += %o0 (add y to upper half)
1: rd %y, %o0 ! get lower half of product
retl
addcc %o4, %g0, %o1 ! put upper half in place and set Z for %o1==0
Lmul_shortway:
/*
* Short multiply. 12 steps, followed by a final shift step.
* The resulting bits are off by 12 and (32-12) = 20 bit positions,
* but there is no problem with %o0 being negative (unlike above),
* and overflow is impossible (the answer is at most 24 bits long).
*/
mulscc %o4, %o1, %o4 ! 1
mulscc %o4, %o1, %o4 ! 2
mulscc %o4, %o1, %o4 ! 3
mulscc %o4, %o1, %o4 ! 4
mulscc %o4, %o1, %o4 ! 5
mulscc %o4, %o1, %o4 ! 6
mulscc %o4, %o1, %o4 ! 7
mulscc %o4, %o1, %o4 ! 8
mulscc %o4, %o1, %o4 ! 9
mulscc %o4, %o1, %o4 ! 10
mulscc %o4, %o1, %o4 ! 11
mulscc %o4, %o1, %o4 ! 12
mulscc %o4, %g0, %o4 ! final shift
/*
* %o4 has 20 of the bits that should be in the result; %y has
* the bottom 12 (as %y's top 12). That is:
*
* %o4 %y
* +----------------+----------------+
* | -12- | -20- | -12- | -20- |
* +------(---------+------)---------+
* -----result-----
*
* The 12 bits of %o4 left of the `result' area are all zero;
* in fact, all top 20 bits of %o4 are zero.
*/
rd %y, %o5
sll %o4, 12, %o0 ! shift middle bits left 12
srl %o5, 20, %o5 ! shift low bits right 20
or %o5, %o0, %o0
retl
addcc %g0, %g0, %o1 ! %o1 = zero, and set Z
|
