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
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
|
{
This file is part of the Free Pascal run time library.
Copyright (c) 1999-2000 by the Free Pascal development team
This file contains some helper routines for int64 and qword
See the file COPYING.FPC, included in this distribution,
for details about the copyright.
This program 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.
**********************************************************************}
{$Q- no overflow checking }
{$R- no range checking }
{$define FPC_SYSTEM_HAS_DIV_QWORD}
function fpc_div_qword(n,z : qword) : qword;assembler;[public,alias: 'FPC_DIV_QWORD']; compilerproc;
var
shift,lzz,lzn : longint;
saveebx,saveedi : longint;
asm
movl %ebx,saveebx
movl %edi,saveedi
{ the following piece of code is taken from the }
{ AMD Athlon Processor x86 Code Optimization manual }
movl n+4,%ecx
movl n,%ebx
movl %ecx,%eax
orl %ebx,%eax
jnz .Lnodivzero
movl %ebp,%edx
movl $200,%eax
call HandleErrorFrame
jmp .Lexit
.Lnodivzero:
movl z+4,%edx
movl z,%eax
testl %ecx,%ecx
jnz .Lqworddivbigdivisor
cmpl %ebx,%edx
jae .Lqworddivtwo_divs
divl %ebx
movl %ecx,%edx
jmp .Lexit
.Lqworddivtwo_divs:
movl %eax,%ecx
movl %edx,%eax
xorl %edx,%edx
divl %ebx
xchgl %ecx,%eax
divl %ebx
movl %ecx,%edx
jmp .Lexit
.Lqworddivbigdivisor:
movl %ecx,%edi
shrl $1,%edx
rcrl $1,%eax
rorl $1,%edi
rcrl $1,%ebx
bsrl %ecx,%ecx
shrdl %cl,%edi,%ebx
shrdl %cl,%edx,%eax
shrl %cl,%edx
roll $1,%edi
divl %ebx
movl z,%ebx
movl %eax,%ecx
imull %eax,%edi
mull n
addl %edi,%edx
subl %eax,%ebx
movl %ecx,%eax
movl z+4,%ecx
sbbl %edx,%ecx
sbbl $0,%eax
xorl %edx,%edx
.Lexit:
movl saveebx,%ebx
movl saveedi,%edi
end;
{$define FPC_SYSTEM_HAS_MOD_QWORD}
function fpc_mod_qword(n,z : qword) : qword;assembler;[public,alias: 'FPC_MOD_QWORD']; compilerproc;
var
shift,lzz,lzn : longint;
saveebx,saveedi : longint;
asm
movl %ebx,saveebx
movl %edi,saveedi
{ the following piece of code is taken from the }
{ AMD Athlon Processor x86 Code Optimization manual }
movl n+4,%ecx
movl n,%ebx
movl %ecx,%eax
orl %ebx,%eax
jnz .Lnodivzero
movl %ebp,%edx
movl $200,%eax
call HandleErrorFrame
jmp .Lexit
.Lnodivzero:
movl z+4,%edx
movl z,%eax
testl %ecx,%ecx
jnz .Lqwordmodr_big_divisior
cmpl %ebx,%edx
jae .Lqwordmodr_two_divs
divl %ebx
movl %edx,%eax
movl %ecx,%edx
jmp .Lexit
.Lqwordmodr_two_divs:
movl %eax,%ecx
movl %edx,%eax
xorl %edx,%edx
divl %ebx
movl %ecx,%eax
divl %ebx
movl %edx,%eax
xorl %edx,%edx
jmp .Lexit
.Lqwordmodr_big_divisior:
movl %ecx,%edi
shrl $1,%edx
rcrl $1,%eax
rorl $1,%edi
rcrl $1,%ebx
bsrl %ecx,%ecx
shrdl %cl,%edi,%ebx
shrdl %cl,%edx,%eax
shrl %cl,%edx
roll $1,%edi
divl %ebx
movl z,%ebx
movl %eax,%ecx
imull %eax,%edi
mull n
addl %edi,%edx
subl %eax,%ebx
movl z+4,%ecx
movl n,%eax
sbbl %edx,%ecx
sbbl %edx,%edx
andl %edx,%eax
andl n+4,%edx
addl %ebx,%eax
adcl %ecx,%edx
.Lexit:
movl saveebx,%ebx
movl saveedi,%edi
end;
{$define FPC_SYSTEM_HAS_MUL_QWORD}
{ multiplies two qwords
the longbool for checkoverflow avoids a misaligned stack
}
function fpc_mul_qword(f1,f2 : qword;checkoverflow : longbool) : qword;[public,alias: 'FPC_MUL_QWORD']; compilerproc;
var
r : qword;
overflowed : boolean;
begin
overflowed:=false;
{ the following piece of code is taken from the
AMD Athlon Processor x86 Code Optimization manual }
asm
movl f1+4,%edx
movl f2+4,%ecx
cmpl $0,checkoverflow
jnz .Loverflowchecked
orl %ecx,%edx
movl f2,%edx
movl f1,%eax
jnz .Lqwordmultwomul
mull %edx
jmp .Lqwordmulready
.Lqwordmultwomul:
imul f1+4,%edx
imul %eax,%ecx
addl %edx,%ecx
mull f2
add %ecx,%edx
.Lqwordmulready:
movl %eax,r
movl %edx,r+4
jmp .Lend
.Loverflowchecked:
{ if both upper dwords are <>0 then it overflows always }
or %ecx,%ecx
jz .Loverok1
or %edx,%edx
jnz .Loverflowed
.Loverok1:
{ overflow checked code }
orl %ecx,%edx
movl f2,%edi
movl f1,%esi
jnz .Lqwordmultwomul2
movl %edi,%eax
mull %esi
movl %eax,%esi
movl %edx,%edi
jmp .Lqwordmulready2
.Lqwordmultwomul2:
movl f1+4,%eax
mull %edi
movl %eax,%edi
jc .Loverflowed
movl %esi,%eax
mull %ecx
movl %eax,%ecx
jc .Loverflowed
addl %edi,%ecx
jc .Loverflowed
movl f2,%eax
mull %esi
movl %eax,%esi
movl %edx,%edi
addl %ecx,%edi
jc .Loverflowed
.Lqwordmulready2:
movl %esi,r
movl %edi,r+4
jmp .Lend
.Loverflowed:
movb $1,overflowed
.Lend:
end [ 'eax','edx','ecx','edi','esi' ];
fpc_mul_qword:=r;
if overflowed then
HandleErrorFrame(215,get_frame);
end;
|
