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
|
/*
* Lowest Level MPI Algorithms
* (C) 1999-2010 Jack Lloyd
* 2006 Luca Piccarreta
*
* Botan is released under the Simplified BSD License (see license.txt)
*/
#ifndef BOTAN_MP_ASM_INTERNAL_H__
#define BOTAN_MP_ASM_INTERNAL_H__
#include <botan/internal/mp_madd.h>
namespace Botan {
extern "C" {
/*
* Helper Macros for x86-64 Assembly
*/
#ifndef ASM
#define ASM(x) x "\n\t"
#endif
#define ADDSUB2_OP(OPERATION, INDEX) \
ASM("movq 8*" #INDEX "(%[y]), %[carry]") \
ASM(OPERATION " %[carry], 8*" #INDEX "(%[x])") \
#define ADDSUB3_OP(OPERATION, INDEX) \
ASM("movq 8*" #INDEX "(%[x]), %[carry]") \
ASM(OPERATION " 8*" #INDEX "(%[y]), %[carry]") \
ASM("movq %[carry], 8*" #INDEX "(%[z])") \
#define LINMUL_OP(WRITE_TO, INDEX) \
ASM("movq 8*" #INDEX "(%[x]),%%rax") \
ASM("mulq %[y]") \
ASM("addq %[carry],%%rax") \
ASM("adcq $0,%%rdx") \
ASM("movq %%rdx,%[carry]") \
ASM("movq %%rax, 8*" #INDEX "(%[" WRITE_TO "])")
#define MULADD_OP(IGNORED, INDEX) \
ASM("movq 8*" #INDEX "(%[x]),%%rax") \
ASM("mulq %[y]") \
ASM("addq %[carry],%%rax") \
ASM("adcq $0,%%rdx") \
ASM("addq 8*" #INDEX "(%[z]),%%rax") \
ASM("adcq $0,%%rdx") \
ASM("movq %%rdx,%[carry]") \
ASM("movq %%rax, 8*" #INDEX " (%[z])")
#define DO_8_TIMES(MACRO, ARG) \
MACRO(ARG, 0) \
MACRO(ARG, 1) \
MACRO(ARG, 2) \
MACRO(ARG, 3) \
MACRO(ARG, 4) \
MACRO(ARG, 5) \
MACRO(ARG, 6) \
MACRO(ARG, 7)
#define ADD_OR_SUBTRACT(CORE_CODE) \
ASM("rorq %[carry]") \
CORE_CODE \
ASM("sbbq %[carry],%[carry]") \
ASM("negq %[carry]")
/*
* Word Addition
*/
inline word word_add(word x, word y, word* carry)
{
asm(
ADD_OR_SUBTRACT(ASM("adcq %[y],%[x]"))
: [x]"=r"(x), [carry]"=r"(*carry)
: "0"(x), [y]"rm"(y), "1"(*carry)
: "cc");
return x;
}
/*
* Eight Word Block Addition, Two Argument
*/
inline word word8_add2(word x[8], const word y[8], word carry)
{
asm(
ADD_OR_SUBTRACT(DO_8_TIMES(ADDSUB2_OP, "adcq"))
: [carry]"=r"(carry)
: [x]"r"(x), [y]"r"(y), "0"(carry)
: "cc", "memory");
return carry;
}
/*
* Eight Word Block Addition, Three Argument
*/
inline word word8_add3(word z[8], const word x[8], const word y[8], word carry)
{
asm(
ADD_OR_SUBTRACT(DO_8_TIMES(ADDSUB3_OP, "adcq"))
: [carry]"=r"(carry)
: [x]"r"(x), [y]"r"(y), [z]"r"(z), "0"(carry)
: "cc", "memory");
return carry;
}
/*
* Word Subtraction
*/
inline word word_sub(word x, word y, word* carry)
{
asm(
ADD_OR_SUBTRACT(ASM("sbbq %[y],%[x]"))
: [x]"=r"(x), [carry]"=r"(*carry)
: "0"(x), [y]"rm"(y), "1"(*carry)
: "cc");
return x;
}
/*
* Eight Word Block Subtraction, Two Argument
*/
inline word word8_sub2(word x[8], const word y[8], word carry)
{
asm(
ADD_OR_SUBTRACT(DO_8_TIMES(ADDSUB2_OP, "sbbq"))
: [carry]"=r"(carry)
: [x]"r"(x), [y]"r"(y), "0"(carry)
: "cc", "memory");
return carry;
}
/*
* Eight Word Block Subtraction, Two Argument
*/
inline word word8_sub2_rev(word x[8], const word y[8], word carry)
{
asm(
ADD_OR_SUBTRACT(DO_8_TIMES(ADDSUB3_OP, "sbbq"))
: [carry]"=r"(carry)
: [x]"r"(y), [y]"r"(x), [z]"r"(x), "0"(carry)
: "cc", "memory");
return carry;
}
/*
* Eight Word Block Subtraction, Three Argument
*/
inline word word8_sub3(word z[8], const word x[8], const word y[8], word carry)
{
asm(
ADD_OR_SUBTRACT(DO_8_TIMES(ADDSUB3_OP, "sbbq"))
: [carry]"=r"(carry)
: [x]"r"(x), [y]"r"(y), [z]"r"(z), "0"(carry)
: "cc", "memory");
return carry;
}
/*
* Eight Word Block Linear Multiplication
*/
inline word word8_linmul2(word x[8], word y, word carry)
{
asm(
DO_8_TIMES(LINMUL_OP, "x")
: [carry]"=r"(carry)
: [x]"r"(x), [y]"rm"(y), "0"(carry)
: "cc", "%rax", "%rdx");
return carry;
}
/*
* Eight Word Block Linear Multiplication
*/
inline word word8_linmul3(word z[8], const word x[8], word y, word carry)
{
asm(
DO_8_TIMES(LINMUL_OP, "z")
: [carry]"=r"(carry)
: [z]"r"(z), [x]"r"(x), [y]"rm"(y), "0"(carry)
: "cc", "%rax", "%rdx");
return carry;
}
/*
* Eight Word Block Multiply/Add
*/
inline word word8_madd3(word z[8], const word x[8], word y, word carry)
{
asm(
DO_8_TIMES(MULADD_OP, "")
: [carry]"=r"(carry)
: [z]"r"(z), [x]"r"(x), [y]"rm"(y), "0"(carry)
: "cc", "%rax", "%rdx");
return carry;
}
/*
* Multiply-Add Accumulator
*/
inline void word3_muladd(word* w2, word* w1, word* w0, word x, word y)
{
asm(
ASM("mulq %[y]")
ASM("addq %[x],%[w0]")
ASM("adcq %[y],%[w1]")
ASM("adcq $0,%[w2]")
: [w0]"=r"(*w0), [w1]"=r"(*w1), [w2]"=r"(*w2)
: [x]"a"(x), [y]"d"(y), "0"(*w0), "1"(*w1), "2"(*w2)
: "cc");
}
/*
* Multiply-Add Accumulator
*/
inline void word3_muladd_2(word* w2, word* w1, word* w0, word x, word y)
{
asm(
ASM("mulq %[y]")
ASM("addq %[x],%[w0]")
ASM("adcq %[y],%[w1]")
ASM("adcq $0,%[w2]")
ASM("addq %[x],%[w0]")
ASM("adcq %[y],%[w1]")
ASM("adcq $0,%[w2]")
: [w0]"=r"(*w0), [w1]"=r"(*w1), [w2]"=r"(*w2)
: [x]"a"(x), [y]"d"(y), "0"(*w0), "1"(*w1), "2"(*w2)
: "cc");
}
#undef ASM
#undef DO_8_TIMES
#undef ADD_OR_SUBTRACT
#undef ADDSUB2_OP
#undef ADDSUB3_OP
#undef LINMUL_OP
#undef MULADD_OP
}
}
#endif
|
