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
|
/* Currently not working and not used. */
/*
Copyright (C) 1999 Free Software Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP Library is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or (at your
option) any later version.
The GNU MP Library is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
License for more details.
You should have received a copy of the GNU Lesser General Public License
along with the GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA.
*/
#define SAVE_BORROW_RESTORE_CARRY(r) adcl r,r; shll $31,r
#define SAVE_CARRY_RESTORE_BORROW(r) adcl r,r
.globl mpn_addsub_n_0
.globl mpn_addsub_n_1
/* Cute i386/i486/p6 addsub loop for the "full overlap" case r1==s2,r2==s1.
We let subtraction and addition alternate in being two limbs
ahead of the other, thereby avoiding some SAVE_RESTORE. */
// r1 = r2 + r1 edi = esi + edi
// r2 = r2 - r1 esi = esi - edi
// s1 s2
// r2 r1
// eax,ebx,ecx,edx,esi,edi,ebp
mpn_addsub_n_0:
pushl %edi
pushl %esi
pushl %ebx
pushl %ebp
movl 20(%esp),%edi /* res_ptr */
movl 24(%esp),%esi /* s1_ptr */
movl 36(%esp),%ebp /* size */
shrl $2,%ebp
xorl %edx,%edx
.align 4
Loop0: // L=load E=execute S=store
movl (%esi),%ebx // sub 0 L
movl 4(%esi),%ecx // sub 1 L
sbbl (%edi),%ebx // sub 0 LE
sbbl 4(%edi),%ecx // sub 1 LE
// SAVE_BORROW_RESTORE_CARRY(%edx)
movl (%esi),%eax // add 0 L
adcl %eax,(%edi) // add 0 LES
movl 4(%esi),%eax // add 1 L
adcl %eax,4(%edi) // add 1 LES
movl %ebx,(%esi) // sub 0 S
movl %ecx,4(%esi) // sub 1 S
movl 8(%esi),%ebx // add 2 L
adcl 8(%edi),%ebx // add 2 LE
movl 12(%esi),%ecx // add 3 L
adcl 12(%edi),%ecx // add 3 LE
// SAVE_CARRY_RESTORE_BORROW(%edx)
movl 8(%edi),%eax // sub 2 L
sbbl %eax,8(%esi) // sub 2 LES
movl 12(%edi),%eax // sub 3 L
sbbl %eax,12(%esi) // sub 3 LES
movl %ebx,8(%edi) // add 2 S
movl %ecx,12(%edi) // add 3 S
leal 16(%esi),%esi
leal 16(%edi),%edi
decl %ebp
jnz Loop0
popl %ebp
popl %ebx
popl %esi
popl %edi
ret
/* Cute i386/i486/p6 addsub loop for the "full overlap" case r1==s1,r2==s2.
We let subtraction and addition alternate in being two limbs
ahead of the other, thereby avoiding some SAVE_RESTORE. */
// r1 = r1 + r2 edi = edi + esi
// r2 = r1 - r2 esi = edi - esi
// s2 s1
// r2 r1
// eax,ebx,ecx,edx,esi,edi,ebp
mpn_addsub_n_1:
pushl %edi
pushl %esi
pushl %ebx
pushl %ebp
movl 20(%esp),%edi /* res_ptr */
movl 24(%esp),%esi /* s1_ptr */
movl 36(%esp),%ebp /* size */
shrl $2,%ebp
xorl %edx,%edx
.align 4
Loop1: // L=load E=execute S=store
movl (%edi),%ebx // sub 0 L
sbbl (%esi),%ebx // sub 0 LE
movl 4(%edi),%ecx // sub 1 L
sbbl 4(%esi),%ecx // sub 1 LE
// SAVE_BORROW_RESTORE_CARRY(%edx)
movl (%esi),%eax // add 0 L
adcl %eax,(%edi) // add 0 LES
movl 4(%esi),%eax // add 1 L
adcl %eax,4(%edi) // add 1 LES
movl %ebx,(%esi) // sub 0 S
movl %ecx,4(%esi) // sub 1 S
movl 8(%esi),%ebx // add 2 L
adcl 8(%edi),%ebx // add 2 LE
movl 12(%esi),%ecx // add 3 L
adcl 12(%edi),%ecx // add 3 LE
// SAVE_CARRY_RESTORE_BORROW(%edx)
movl 8(%edi),%eax // sub 2 L
sbbl 8(%esi),%eax // sub 2 LES
movl %eax,8(%esi) // sub 2 S
movl 12(%edi),%eax // sub 3 L
sbbl 12(%esi),%eax // sub 3 LE
movl %eax,12(%esi) // sub 3 S
movl %ebx,8(%edi) // add 2 S
movl %ecx,12(%edi) // add 3 S
leal 16(%esi),%esi
leal 16(%edi),%edi
decl %ebp
jnz Loop1
popl %ebp
popl %ebx
popl %esi
popl %edi
ret
.globl mpn_copy
mpn_copy:
pushl %edi
pushl %esi
pushl %ebx
pushl %ebp
movl 20(%esp),%edi /* res_ptr */
movl 24(%esp),%esi /* s1_ptr */
movl 28(%esp),%ebp /* size */
shrl $2,%ebp
.align 4
Loop2:
movl (%esi),%eax
movl 4(%esi),%ebx
movl %eax,(%edi)
movl %ebx,4(%edi)
movl 8(%esi),%eax
movl 12(%esi),%ebx
movl %eax,8(%edi)
movl %ebx,12(%edi)
leal 16(%esi),%esi
leal 16(%edi),%edi
decl %ebp
jnz Loop2
popl %ebp
popl %ebx
popl %esi
popl %edi
ret
|
