forked from torvalds/linux
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathglue_helper-asm-avx2.S
180 lines (169 loc) · 5.22 KB
/
glue_helper-asm-avx2.S
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
/*
* Shared glue code for 128bit block ciphers, AVX2 assembler macros
*
* Copyright © 2012-2013 Jussi Kivilinna <[email protected]>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of 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.
*
*/
#define load_16way(src, x0, x1, x2, x3, x4, x5, x6, x7) \
vmovdqu (0*32)(src), x0; \
vmovdqu (1*32)(src), x1; \
vmovdqu (2*32)(src), x2; \
vmovdqu (3*32)(src), x3; \
vmovdqu (4*32)(src), x4; \
vmovdqu (5*32)(src), x5; \
vmovdqu (6*32)(src), x6; \
vmovdqu (7*32)(src), x7;
#define store_16way(dst, x0, x1, x2, x3, x4, x5, x6, x7) \
vmovdqu x0, (0*32)(dst); \
vmovdqu x1, (1*32)(dst); \
vmovdqu x2, (2*32)(dst); \
vmovdqu x3, (3*32)(dst); \
vmovdqu x4, (4*32)(dst); \
vmovdqu x5, (5*32)(dst); \
vmovdqu x6, (6*32)(dst); \
vmovdqu x7, (7*32)(dst);
#define store_cbc_16way(src, dst, x0, x1, x2, x3, x4, x5, x6, x7, t0) \
vpxor t0, t0, t0; \
vinserti128 $1, (src), t0, t0; \
vpxor t0, x0, x0; \
vpxor (0*32+16)(src), x1, x1; \
vpxor (1*32+16)(src), x2, x2; \
vpxor (2*32+16)(src), x3, x3; \
vpxor (3*32+16)(src), x4, x4; \
vpxor (4*32+16)(src), x5, x5; \
vpxor (5*32+16)(src), x6, x6; \
vpxor (6*32+16)(src), x7, x7; \
store_16way(dst, x0, x1, x2, x3, x4, x5, x6, x7);
#define inc_le128(x, minus_one, tmp) \
vpcmpeqq minus_one, x, tmp; \
vpsubq minus_one, x, x; \
vpslldq $8, tmp, tmp; \
vpsubq tmp, x, x;
#define add2_le128(x, minus_one, minus_two, tmp1, tmp2) \
vpcmpeqq minus_one, x, tmp1; \
vpcmpeqq minus_two, x, tmp2; \
vpsubq minus_two, x, x; \
vpor tmp2, tmp1, tmp1; \
vpslldq $8, tmp1, tmp1; \
vpsubq tmp1, x, x;
#define load_ctr_16way(iv, bswap, x0, x1, x2, x3, x4, x5, x6, x7, t0, t0x, t1, \
t1x, t2, t2x, t3, t3x, t4, t5) \
vpcmpeqd t0, t0, t0; \
vpsrldq $8, t0, t0; /* ab: -1:0 ; cd: -1:0 */ \
vpaddq t0, t0, t4; /* ab: -2:0 ; cd: -2:0 */\
\
/* load IV and byteswap */ \
vmovdqu (iv), t2x; \
vmovdqa t2x, t3x; \
inc_le128(t2x, t0x, t1x); \
vbroadcasti128 bswap, t1; \
vinserti128 $1, t2x, t3, t2; /* ab: le0 ; cd: le1 */ \
vpshufb t1, t2, x0; \
\
/* construct IVs */ \
add2_le128(t2, t0, t4, t3, t5); /* ab: le2 ; cd: le3 */ \
vpshufb t1, t2, x1; \
add2_le128(t2, t0, t4, t3, t5); \
vpshufb t1, t2, x2; \
add2_le128(t2, t0, t4, t3, t5); \
vpshufb t1, t2, x3; \
add2_le128(t2, t0, t4, t3, t5); \
vpshufb t1, t2, x4; \
add2_le128(t2, t0, t4, t3, t5); \
vpshufb t1, t2, x5; \
add2_le128(t2, t0, t4, t3, t5); \
vpshufb t1, t2, x6; \
add2_le128(t2, t0, t4, t3, t5); \
vpshufb t1, t2, x7; \
vextracti128 $1, t2, t2x; \
inc_le128(t2x, t0x, t3x); \
vmovdqu t2x, (iv);
#define store_ctr_16way(src, dst, x0, x1, x2, x3, x4, x5, x6, x7) \
vpxor (0*32)(src), x0, x0; \
vpxor (1*32)(src), x1, x1; \
vpxor (2*32)(src), x2, x2; \
vpxor (3*32)(src), x3, x3; \
vpxor (4*32)(src), x4, x4; \
vpxor (5*32)(src), x5, x5; \
vpxor (6*32)(src), x6, x6; \
vpxor (7*32)(src), x7, x7; \
store_16way(dst, x0, x1, x2, x3, x4, x5, x6, x7);
#define gf128mul_x_ble(iv, mask, tmp) \
vpsrad $31, iv, tmp; \
vpaddq iv, iv, iv; \
vpshufd $0x13, tmp, tmp; \
vpand mask, tmp, tmp; \
vpxor tmp, iv, iv;
#define gf128mul_x2_ble(iv, mask1, mask2, tmp0, tmp1) \
vpsrad $31, iv, tmp0; \
vpaddq iv, iv, tmp1; \
vpsllq $2, iv, iv; \
vpshufd $0x13, tmp0, tmp0; \
vpsrad $31, tmp1, tmp1; \
vpand mask2, tmp0, tmp0; \
vpshufd $0x13, tmp1, tmp1; \
vpxor tmp0, iv, iv; \
vpand mask1, tmp1, tmp1; \
vpxor tmp1, iv, iv;
#define load_xts_16way(iv, src, dst, x0, x1, x2, x3, x4, x5, x6, x7, tiv, \
tivx, t0, t0x, t1, t1x, t2, t2x, t3, \
xts_gf128mul_and_shl1_mask_0, \
xts_gf128mul_and_shl1_mask_1) \
vbroadcasti128 xts_gf128mul_and_shl1_mask_0, t1; \
\
/* load IV and construct second IV */ \
vmovdqu (iv), tivx; \
vmovdqa tivx, t0x; \
gf128mul_x_ble(tivx, t1x, t2x); \
vbroadcasti128 xts_gf128mul_and_shl1_mask_1, t2; \
vinserti128 $1, tivx, t0, tiv; \
vpxor (0*32)(src), tiv, x0; \
vmovdqu tiv, (0*32)(dst); \
\
/* construct and store IVs, also xor with source */ \
gf128mul_x2_ble(tiv, t1, t2, t0, t3); \
vpxor (1*32)(src), tiv, x1; \
vmovdqu tiv, (1*32)(dst); \
\
gf128mul_x2_ble(tiv, t1, t2, t0, t3); \
vpxor (2*32)(src), tiv, x2; \
vmovdqu tiv, (2*32)(dst); \
\
gf128mul_x2_ble(tiv, t1, t2, t0, t3); \
vpxor (3*32)(src), tiv, x3; \
vmovdqu tiv, (3*32)(dst); \
\
gf128mul_x2_ble(tiv, t1, t2, t0, t3); \
vpxor (4*32)(src), tiv, x4; \
vmovdqu tiv, (4*32)(dst); \
\
gf128mul_x2_ble(tiv, t1, t2, t0, t3); \
vpxor (5*32)(src), tiv, x5; \
vmovdqu tiv, (5*32)(dst); \
\
gf128mul_x2_ble(tiv, t1, t2, t0, t3); \
vpxor (6*32)(src), tiv, x6; \
vmovdqu tiv, (6*32)(dst); \
\
gf128mul_x2_ble(tiv, t1, t2, t0, t3); \
vpxor (7*32)(src), tiv, x7; \
vmovdqu tiv, (7*32)(dst); \
\
vextracti128 $1, tiv, tivx; \
gf128mul_x_ble(tivx, t1x, t2x); \
vmovdqu tivx, (iv);
#define store_xts_16way(dst, x0, x1, x2, x3, x4, x5, x6, x7) \
vpxor (0*32)(dst), x0, x0; \
vpxor (1*32)(dst), x1, x1; \
vpxor (2*32)(dst), x2, x2; \
vpxor (3*32)(dst), x3, x3; \
vpxor (4*32)(dst), x4, x4; \
vpxor (5*32)(dst), x5, x5; \
vpxor (6*32)(dst), x6, x6; \
vpxor (7*32)(dst), x7, x7; \
store_16way(dst, x0, x1, x2, x3, x4, x5, x6, x7);