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
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
|
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
/*
* Implementions of functions found in $(D_PSYMBOL tanya.memory.op) for X86-64.
*
* Copyright: Eugene Wissner 2017.
* License: $(LINK2 https://www.mozilla.org/en-US/MPL/2.0/,
* Mozilla Public License, v. 2.0).
* Authors: $(LINK2 mailto:info@caraus.de, Eugene Wissner)
*/
module tanya.memory.arch.x86_64;
version (D_InlineAsm_X86_64):
pragma(inline, true)
package (tanya.memory) void copy(const void[] source, void[] target)
pure nothrow @system @nogc
{
asm pure nothrow @nogc
{
naked;
// RDI and RSI should be preserved.
mov RAX, RDI;
mov R8, RSI;
}
// Set the registers for movsb/movsq.
version (Windows) asm pure nothrow @nogc
{
// RDX - source.
// RCX - target.
mov RDI, [ RCX + 8 ];
mov RSI, [ RDX + 8 ];
mov RDX, [ RDX ];
}
else asm pure nothrow @nogc
{
// RDX - source length.
// RCX - source data.
// RDI - target length
// RSI - target data.
mov RDI, RSI;
mov RSI, RCX;
}
asm pure nothrow @nogc
{
cmp RDX, 0x08;
jc aligned_8;
test EDI, 0x07;
jz aligned_8;
naligned:
movsb;
dec RDX;
test EDI, 0x07;
jnz naligned;
aligned_8:
mov RCX, RDX;
shr RCX, 0x03;
rep;
movsq;
and EDX, 0x07;
jz end;
// Write the remaining bytes.
mov RCX, RDX;
rep;
movsb;
end: // Restore registers.
mov RSI, R8;
mov RDI, RAX;
ret;
}
}
pragma(inline, true)
package (tanya.memory) void zero(void[] memory)
pure nothrow @system @nogc
{
asm pure nothrow @nogc
{
naked;
}
version (Windows) asm pure nothrow @nogc
{
/*
* RCX - array.
*/
mov R8, [ RCX ];
mov R9, [ RCX + 8 ];
}
else asm pure nothrow @nogc
{
/*
* RSI - pointer.
* RDI - length.
*/
mov R8, RDI;
mov R9, RSI;
}
asm pure nothrow @nogc
{
// Check for zero length.
test R8, R8;
jz end;
// Set to 0.
pxor XMM0, XMM0;
// Check if the pointer is aligned to a 16-byte boundary.
and R9, -0x10;
}
// Compute the number of misaligned bytes.
version (Windows) asm pure nothrow @nogc
{
mov RAX, [ RCX + 8 ];
}
else asm pure nothrow @nogc
{
mov RAX, RSI;
}
asm pure nothrow @nogc
{
sub RAX, R9;
test RAX, RAX;
jz aligned;
// Get the number of bytes to be written until we are aligned.
mov RDX, 0x10;
sub RDX, RAX;
}
version (Windows) asm pure nothrow @nogc
{
mov R9, [ RCX + 8 ];
}
else asm pure nothrow @nogc
{
mov R9, RSI;
}
asm pure nothrow @nogc
{
// Set RAX to zero, so we can set bytes and dwords.
xor RAX, RAX;
naligned:
mov [ R9 ], AL; // Write a byte.
// Advance the pointer. Decrease the total number of bytes
// and the misaligned ones.
inc R9;
dec RDX;
dec R8;
// Checks if we are aligned.
test RDX, RDX;
jnz naligned;
aligned:
// Checks if we're done writing bytes.
test R8, R8;
jz end;
// Write 1 byte at a time.
cmp R8, 8;
jl aligned_1;
// Write 8 bytes at a time.
cmp R8, 16;
jl aligned_8;
// Write 16 bytes at a time.
cmp R8, 32;
jl aligned_16;
// Write 32 bytes at a time.
cmp R8, 64;
jl aligned_32;
aligned_64:
movdqa [ R9 ], XMM0;
movdqa [ R9 + 16 ], XMM0;
movdqa [ R9 + 32 ], XMM0;
movdqa [ R9 + 48 ], XMM0;
add R9, 64;
sub R8, 64;
cmp R8, 64;
jge aligned_64;
// Checks if we're done writing bytes.
test R8, R8;
jz end;
// Write 1 byte at a time.
cmp R8, 8;
jl aligned_1;
// Write 8 bytes at a time.
cmp R8, 16;
jl aligned_8;
// Write 16 bytes at a time.
cmp R8, 32;
jl aligned_16;
aligned_32:
movdqa [ R9 ], XMM0;
movdqa [ R9 + 16 ], XMM0;
add R9, 32;
sub R8, 32;
// Checks if we're done writing bytes.
test R8, R8;
jz end;
// Write 1 byte at a time.
cmp R8, 8;
jl aligned_1;
// Write 8 bytes at a time.
cmp R8, 16;
jl aligned_8;
aligned_16:
movdqa [ R9 ], XMM0;
add R9, 16;
sub R8, 16;
// Checks if we're done writing bytes.
test R8, R8;
jz end;
// Write 1 byte at a time.
cmp R8, 8;
jl aligned_1;
aligned_8:
mov [ R9 ], RAX;
add R9, 8;
sub R8, 8;
// Checks if we're done writing bytes.
test R8, R8;
jz end;
aligned_1:
mov [ R9 ], AL;
inc R9;
dec R8;
test R8, R8;
jnz aligned_1;
end:
ret;
}
}
|
