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
|
C nettle, low-level cryptographics library
C
C Copyright (C) 2013 Niels Möller
C
C The nettle library is free software; you can redistribute it and/or modify
C it under the terms of the GNU Lesser General Public License as published by
C the Free Software Foundation; either version 2.1 of the License, or (at your
C option) any later version.
C
C The nettle library is distributed in the hope that it will be useful, but
C WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
C or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
C License for more details.
C
C You should have received a copy of the GNU Lesser General Public License
C along with the nettle library; see the file COPYING.LIB. If not, write to
C the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
C MA 02111-1301, USA.
include_src(<arm/aes.m4>)
C Benchmarked at at 785, 914, 1051 cycles/block on cortex A9,
C for 128, 192 and 256 bit key sizes. Unclear why it is slower
C than _aes_encrypt.
define(<CTX>, <r0>)
define(<TABLE>, <r1>)
define(<LENGTH>, <r2>)
define(<DST>, <r3>)
define(<SRC>, <r12>)
define(<W0>, <r4>)
define(<W1>, <r5>)
define(<W2>, <r6>)
define(<W3>, <r7>)
define(<T0>, <r8>)
define(<KEY>, <r10>)
define(<ROUND>, <r11>)
define(<X0>, <r2>) C Overlaps LENGTH, SRC, DST
define(<X1>, <r3>)
define(<X2>, <r12>)
define(<X3>, <r14>) C lr
C AES_DECRYPT_ROUND(x0,x1,x2,x3,w0,w1,w2,w3,key)
define(<AES_DECRYPT_ROUND>, <
uxtb T0, $1
ldr $5, [TABLE, T0, lsl #2]
uxtb T0, $2
ldr $6, [TABLE, T0, lsl #2]
uxtb T0, $3
ldr $7, [TABLE, T0, lsl #2]
uxtb T0, $4
ldr $8, [TABLE, T0, lsl #2]
uxtb T0, $4, ror #8
add TABLE, TABLE, #1024
ldr T0, [TABLE, T0, lsl #2]
eor $5, $5, T0
uxtb T0, $1, ror #8
ldr T0, [TABLE, T0, lsl #2]
eor $6, $6, T0
uxtb T0, $2, ror #8
ldr T0, [TABLE, T0, lsl #2]
eor $7, $7, T0
uxtb T0, $3, ror #8
ldr T0, [TABLE, T0, lsl #2]
eor $8, $8, T0
uxtb T0, $3, ror #16
add TABLE, TABLE, #1024
ldr T0, [TABLE, T0, lsl #2]
eor $5, $5, T0
uxtb T0, $4, ror #16
ldr T0, [TABLE, T0, lsl #2]
eor $6, $6, T0
uxtb T0, $1, ror #16
ldr T0, [TABLE, T0, lsl #2]
eor $7, $7, T0
uxtb T0, $2, ror #16
ldr T0, [TABLE, T0, lsl #2]
eor $8, $8, T0
uxtb T0, $2, ror #24
add TABLE, TABLE, #1024
ldr T0, [TABLE, T0, lsl #2]
eor $5, $5, T0
uxtb T0, $3, ror #24
ldr T0, [TABLE, T0, lsl #2]
eor $6, $6, T0
uxtb T0, $4, ror #24
ldr T0, [TABLE, T0, lsl #2]
eor $7, $7, T0
uxtb T0, $1, ror #24
ldr T0, [TABLE, T0, lsl #2]
ldm $9!, {$1,$2,$3,$4}
eor $8, $8, T0
sub TABLE, TABLE, #3072
eor $5, $5, $1
eor $6, $6, $2
eor $7, $7, $3
eor $8, $8, $4
>)
.file "aes-decrypt-internal.asm"
C _aes_decrypt(struct aes_context *ctx,
C const struct aes_table *T,
C size_t length, uint8_t *dst,
C uint8_t *src)
.text
ALIGN(4)
PROLOGUE(_nettle_aes_decrypt)
teq LENGTH, #0
beq .Lend
ldr SRC, [sp]
push {r4,r5,r6,r7,r8,r10,r11,lr}
nop C For some mysterious reason, taking out this nop
C slows this function down on Cortex-A9.
ALIGN(16)
.Lblock_loop:
mov KEY, CTX
AES_LOAD(SRC,KEY,W0)
AES_LOAD(SRC,KEY,W1)
AES_LOAD(SRC,KEY,W2)
AES_LOAD(SRC,KEY,W3)
push {LENGTH, DST, SRC}
ldr ROUND, [CTX, #+AES_NROUNDS]
add TABLE, TABLE, #AES_TABLE0
b .Lentry
ALIGN(16)
.Lround_loop:
C Transform X -> W
AES_DECRYPT_ROUND(X0, X1, X2, X3, W0, W1, W2, W3, KEY)
.Lentry:
subs ROUND, ROUND,#2
C Transform W -> X
AES_DECRYPT_ROUND(W0, W1, W2, W3, X0, X1, X2, X3, KEY)
bne .Lround_loop
sub TABLE, TABLE, #AES_TABLE0
C Final round
AES_FINAL_ROUND(X0, X3, X2, X1, KEY, W0)
AES_FINAL_ROUND(X1, X0, X3, X2, KEY, W1)
AES_FINAL_ROUND(X2, X1, X0, X3, KEY, W2)
AES_FINAL_ROUND(X3, X2, X1, X0, KEY, W3)
pop {LENGTH, DST, SRC}
AES_STORE(DST,W0)
AES_STORE(DST,W1)
AES_STORE(DST,W2)
AES_STORE(DST,W3)
subs LENGTH, LENGTH, #16
bhi .Lblock_loop
pop {r4,r5,r6,r7,r8,r10,r11,pc}
.Lend:
bx lr
EPILOGUE(_nettle_aes_decrypt)
|
