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
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
|
/* SPDX-License-Identifier: GPL-2.0-or-later */
/***************************************************************************
* Copyright (C) 2006 by Joachim Fritschi, <jfritschi@freenet.de> *
* *
***************************************************************************/
.file "twofish-x86_64-asm.S"
.text
#include <linux/linkage.h>
#include <asm/asm-offsets.h>
#define a_offset 0
#define b_offset 4
#define c_offset 8
#define d_offset 12
/* Structure of the crypto context struct*/
#define s0 0 /* S0 Array 256 Words each */
#define s1 1024 /* S1 Array */
#define s2 2048 /* S2 Array */
#define s3 3072 /* S3 Array */
#define w 4096 /* 8 whitening keys (word) */
#define k 4128 /* key 1-32 ( word ) */
/* define a few register aliases to allow macro substitution */
#define R0 %rax
#define R0D %eax
#define R0B %al
#define R0H %ah
#define R1 %rbx
#define R1D %ebx
#define R1B %bl
#define R1H %bh
#define R2 %rcx
#define R2D %ecx
#define R2B %cl
#define R2H %ch
#define R3 %rdx
#define R3D %edx
#define R3B %dl
#define R3H %dh
/* performs input whitening */
#define input_whitening(src,context,offset)\
xor w+offset(context), src;
/* performs input whitening */
#define output_whitening(src,context,offset)\
xor w+16+offset(context), src;
/*
* a input register containing a (rotated 16)
* b input register containing b
* c input register containing c
* d input register containing d (already rol $1)
* operations on a and b are interleaved to increase performance
*/
#define encrypt_round(a,b,c,d,round)\
movzx b ## B, %edi;\
mov s1(%r11,%rdi,4),%r8d;\
movzx a ## B, %edi;\
mov s2(%r11,%rdi,4),%r9d;\
movzx b ## H, %edi;\
ror $16, b ## D;\
xor s2(%r11,%rdi,4),%r8d;\
movzx a ## H, %edi;\
ror $16, a ## D;\
xor s3(%r11,%rdi,4),%r9d;\
movzx b ## B, %edi;\
xor s3(%r11,%rdi,4),%r8d;\
movzx a ## B, %edi;\
xor (%r11,%rdi,4), %r9d;\
movzx b ## H, %edi;\
ror $15, b ## D;\
xor (%r11,%rdi,4), %r8d;\
movzx a ## H, %edi;\
xor s1(%r11,%rdi,4),%r9d;\
add %r8d, %r9d;\
add %r9d, %r8d;\
add k+round(%r11), %r9d;\
xor %r9d, c ## D;\
rol $15, c ## D;\
add k+4+round(%r11),%r8d;\
xor %r8d, d ## D;
/*
* a input register containing a(rotated 16)
* b input register containing b
* c input register containing c
* d input register containing d (already rol $1)
* operations on a and b are interleaved to increase performance
* during the round a and b are prepared for the output whitening
*/
#define encrypt_last_round(a,b,c,d,round)\
mov b ## D, %r10d;\
shl $32, %r10;\
movzx b ## B, %edi;\
mov s1(%r11,%rdi,4),%r8d;\
movzx a ## B, %edi;\
mov s2(%r11,%rdi,4),%r9d;\
movzx b ## H, %edi;\
ror $16, b ## D;\
xor s2(%r11,%rdi,4),%r8d;\
movzx a ## H, %edi;\
ror $16, a ## D;\
xor s3(%r11,%rdi,4),%r9d;\
movzx b ## B, %edi;\
xor s3(%r11,%rdi,4),%r8d;\
movzx a ## B, %edi;\
xor (%r11,%rdi,4), %r9d;\
xor a, %r10;\
movzx b ## H, %edi;\
xor (%r11,%rdi,4), %r8d;\
movzx a ## H, %edi;\
xor s1(%r11,%rdi,4),%r9d;\
add %r8d, %r9d;\
add %r9d, %r8d;\
add k+round(%r11), %r9d;\
xor %r9d, c ## D;\
ror $1, c ## D;\
add k+4+round(%r11),%r8d;\
xor %r8d, d ## D
/*
* a input register containing a
* b input register containing b (rotated 16)
* c input register containing c (already rol $1)
* d input register containing d
* operations on a and b are interleaved to increase performance
*/
#define decrypt_round(a,b,c,d,round)\
movzx a ## B, %edi;\
mov (%r11,%rdi,4), %r9d;\
movzx b ## B, %edi;\
mov s3(%r11,%rdi,4),%r8d;\
movzx a ## H, %edi;\
ror $16, a ## D;\
xor s1(%r11,%rdi,4),%r9d;\
movzx b ## H, %edi;\
ror $16, b ## D;\
xor (%r11,%rdi,4), %r8d;\
movzx a ## B, %edi;\
xor s2(%r11,%rdi,4),%r9d;\
movzx b ## B, %edi;\
xor s1(%r11,%rdi,4),%r8d;\
movzx a ## H, %edi;\
ror $15, a ## D;\
xor s3(%r11,%rdi,4),%r9d;\
movzx b ## H, %edi;\
xor s2(%r11,%rdi,4),%r8d;\
add %r8d, %r9d;\
add %r9d, %r8d;\
add k+round(%r11), %r9d;\
xor %r9d, c ## D;\
add k+4+round(%r11),%r8d;\
xor %r8d, d ## D;\
rol $15, d ## D;
/*
* a input register containing a
* b input register containing b
* c input register containing c (already rol $1)
* d input register containing d
* operations on a and b are interleaved to increase performance
* during the round a and b are prepared for the output whitening
*/
#define decrypt_last_round(a,b,c,d,round)\
movzx a ## B, %edi;\
mov (%r11,%rdi,4), %r9d;\
movzx b ## B, %edi;\
mov s3(%r11,%rdi,4),%r8d;\
movzx b ## H, %edi;\
ror $16, b ## D;\
xor (%r11,%rdi,4), %r8d;\
movzx a ## H, %edi;\
mov b ## D, %r10d;\
shl $32, %r10;\
xor a, %r10;\
ror $16, a ## D;\
xor s1(%r11,%rdi,4),%r9d;\
movzx b ## B, %edi;\
xor s1(%r11,%rdi,4),%r8d;\
movzx a ## B, %edi;\
xor s2(%r11,%rdi,4),%r9d;\
movzx b ## H, %edi;\
xor s2(%r11,%rdi,4),%r8d;\
movzx a ## H, %edi;\
xor s3(%r11,%rdi,4),%r9d;\
add %r8d, %r9d;\
add %r9d, %r8d;\
add k+round(%r11), %r9d;\
xor %r9d, c ## D;\
add k+4+round(%r11),%r8d;\
xor %r8d, d ## D;\
ror $1, d ## D;
SYM_FUNC_START(twofish_enc_blk)
pushq R1
/* %rdi contains the ctx address */
/* %rsi contains the output address */
/* %rdx contains the input address */
/* ctx address is moved to free one non-rex register
as target for the 8bit high operations */
mov %rdi, %r11
movq (R3), R1
movq 8(R3), R3
input_whitening(R1,%r11,a_offset)
input_whitening(R3,%r11,c_offset)
mov R1D, R0D
rol $16, R0D
shr $32, R1
mov R3D, R2D
shr $32, R3
rol $1, R3D
encrypt_round(R0,R1,R2,R3,0);
encrypt_round(R2,R3,R0,R1,8);
encrypt_round(R0,R1,R2,R3,2*8);
encrypt_round(R2,R3,R0,R1,3*8);
encrypt_round(R0,R1,R2,R3,4*8);
encrypt_round(R2,R3,R0,R1,5*8);
encrypt_round(R0,R1,R2,R3,6*8);
encrypt_round(R2,R3,R0,R1,7*8);
encrypt_round(R0,R1,R2,R3,8*8);
encrypt_round(R2,R3,R0,R1,9*8);
encrypt_round(R0,R1,R2,R3,10*8);
encrypt_round(R2,R3,R0,R1,11*8);
encrypt_round(R0,R1,R2,R3,12*8);
encrypt_round(R2,R3,R0,R1,13*8);
encrypt_round(R0,R1,R2,R3,14*8);
encrypt_last_round(R2,R3,R0,R1,15*8);
output_whitening(%r10,%r11,a_offset)
movq %r10, (%rsi)
shl $32, R1
xor R0, R1
output_whitening(R1,%r11,c_offset)
movq R1, 8(%rsi)
popq R1
movl $1,%eax
ret
SYM_FUNC_END(twofish_enc_blk)
SYM_FUNC_START(twofish_dec_blk)
pushq R1
/* %rdi contains the ctx address */
/* %rsi contains the output address */
/* %rdx contains the input address */
/* ctx address is moved to free one non-rex register
as target for the 8bit high operations */
mov %rdi, %r11
movq (R3), R1
movq 8(R3), R3
output_whitening(R1,%r11,a_offset)
output_whitening(R3,%r11,c_offset)
mov R1D, R0D
shr $32, R1
rol $16, R1D
mov R3D, R2D
shr $32, R3
rol $1, R2D
decrypt_round(R0,R1,R2,R3,15*8);
decrypt_round(R2,R3,R0,R1,14*8);
decrypt_round(R0
|