ecc-384-modp.asm 3.93 KB
Newer Older
1
2
3
C x86_64/ecc-384-modp.asm

ifelse(<
4
   Copyright (C) 2013, 2015 Niels Möller
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

   This file is part of GNU Nettle.

   GNU Nettle is free software: you can redistribute it and/or
   modify it under the terms of either:

     * the GNU Lesser General Public License as published by the Free
       Software Foundation; either version 3 of the License, or (at your
       option) any later version.

   or

     * the GNU General Public License as published by the Free
       Software Foundation; either version 2 of the License, or (at your
       option) any later version.

   or both in parallel, as here.

   GNU Nettle 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
   General Public License for more details.

   You should have received copies of the GNU General Public License and
   the GNU Lesser General Public License along with this program.  If
   not, see http://www.gnu.org/licenses/.
>)
32
33
34
35

	.file "ecc-384-modp.asm"

define(<RP>, <%rsi>)
36
define(<D5>, <%rax>)
37
38
39
40
41
42
43
44
45
46
47
48
49
50
define(<T0>, <%rbx>)
define(<T1>, <%rcx>)
define(<T2>, <%rdx>)
define(<T3>, <%rbp>)
define(<T4>, <%rdi>)
define(<T5>, <%r8>)
define(<H0>, <%r9>)
define(<H1>, <%r10>)
define(<H2>, <%r11>)
define(<H3>, <%r12>)
define(<H4>, <%r13>)
define(<H5>, <%r14>)
define(<C2>, <%r15>)
define(<C0>, H5)	C Overlap
51
52
define(<TMP>, RP)	C Overlap

53
54
55
56
57
58
59
60
61
62
63

PROLOGUE(nettle_ecc_384_modp)
	W64_ENTRY(2, 0)

	push	%rbx
	push	%rbp
	push	%r12
	push	%r13
	push	%r14
	push	%r15

64
65
66
	C First get top 2 limbs, which need folding twice.
	C B^10 = B^6 + B^4 + 2^32 (B-1)B^4.
	C We handle the terms as follow:
67
	C
68
	C B^6: Folded immediatly.
69
	C
70
	C B^4: Delayed, added in in the next folding.
71
	C
72
73
74
75
76
77
78
79
80
81
82
	C 2^32(B-1) B^4: Low half limb delayed until the next
	C folding. Top 1.5 limbs subtracted and shifter now, resulting
	C in 2.5 limbs. The low limb saved in D5, high 1.5 limbs added
	C in.

	mov	80(RP), H4
	mov	88(RP), H5
	C Shift right 32 bits, into H1, H0
	mov	H4, H0
	mov	H5, H1
	mov	H5, D5
83
	shr	$32, H1
84
85
86
87
88
89
90
91
92
93
94
95
	shl	$32, D5
	shr	$32, H0
	or	D5, H0

	C	H1 H0
	C       -  H1 H0
	C       --------
	C       H1 H0 D5
	mov	H0, D5
	neg	D5
	sbb	H1, H0
	sbb	$0, H1
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

	xor	C2, C2
	add	H4, H0
	adc	H5, H1
	adc	$0, C2

	C Add in to high part
	add	48(RP), H0
	adc	56(RP), H1
	adc	$0, C2		C Do C2 later

	C +1 term
	mov	(RP), T0
	add	H0, T0
	mov	8(RP), T1
	adc	H1, T1
	mov	16(RP), T2
	mov	64(RP), H2
	adc	H2, T2
	mov	24(RP), T3
	mov	72(RP), H3
	adc	H3, T3
	mov	32(RP), T4
	adc	H4, T4
	mov	40(RP), T5
	adc	H5, T5
	sbb	C0, C0
	neg	C0		C FIXME: Switch sign of C0?

	push	RP

	C +B^2 term
	add	H0, T2
	adc	H1, T3
	adc	H2, T4
	adc	H3, T5
	adc	$0, C0

134
	C Shift left, including low half of H4
135
	mov	H3, TMP
136
	shl	$32, H4
137
	shr	$32, TMP
138
	or	TMP, H4
139
140

	mov	H2, TMP
141
	shl	$32, H3
142
143
144
145
	shr	$32, TMP
	or	TMP, H3

	mov	H1, TMP
146
	shl	$32, H2
147
148
149
150
	shr	$32, TMP
	or	TMP, H2

	mov	H0, TMP
151
	shl	$32, H1
152
153
154
	shr	$32, TMP
	or	TMP, H1

155
156
157
158
159
160
	shl	$32, H0

	C   H4 H3 H2 H1 H0  0
	C  -   H4 H3 H2 H1 H0
	C  ---------------
	C   H4 H3 H2 H1 H0 TMP
161

162
163
164
165
166
167
168
169
170
	mov	H0, TMP
	neg	TMP
	sbb	H1, H0
	sbb	H2, H1
	sbb	H3, H2
	sbb	H4, H3
	sbb	$0, H4

	add	TMP, T0
171
172
173
174
	adc	H0, T1
	adc	H1, T2
	adc	H2, T3
	adc	H3, T4
175
	adc	H4, T5
176
177
178
	adc	$0, C0

	C Remains to add in C2 and C0
179
	C Set H1, H0 = (2^96 - 2^32 + 1) C0
180
181
182
	mov	C0, H0
	mov	C0, H1
	shl	$32, H1
183
	sub	H1, H0
184
185
	sbb	$0, H1

186
187
188
189
190
191
192
193
194
	C Set H3, H2 = (2^96 - 2^32 + 1) C2
	mov	C2, H2
	mov	C2, H3
	shl	$32, H3
	sub	H3, H2
	sbb	$0, H3
	add	C0, H2		C No carry. Could use lea trick

	xor	C0, C0
195
196
	add	H0, T0
	adc	H1, T1
197
198
199
200
201
	adc	H2, T2
	adc	H3, T3
	adc	C2, T4
	adc	D5, T5		C Value delayed from initial folding
	adc	$0, C0		C Use sbb and switch sign?
202
203
204
205
206
207
208
209
210
211

	C Final unlikely carry
	mov	C0, H0
	mov	C0, H1
	shl	$32, H1
	sub	H1, H0
	sbb	$0, H1

	pop	RP

212
	add	H0, T0
213
	mov	T0, (RP)
214
	adc	H1, T1
215
	mov	T1, 8(RP)
216
	adc	C0, T2
217
	mov	T2, 16(RP)
218
	adc	$0, T3
219
	mov	T3, 24(RP)
220
	adc	$0, T4
221
	mov	T4, 32(RP)
222
	adc	$0, T5
223
224
225
226
227
228
229
230
231
232
233
234
	mov	T5, 40(RP)

	pop	%r15
	pop	%r14
	pop	%r13
	pop	%r12
	pop	%rbp
	pop	%rbx

	W64_EXIT(2, 0)
	ret
EPILOGUE(nettle_ecc_384_modp)