Crypto++ 8.9
Free C++ class library of cryptographic schemes
aria.cpp
1// aria.cpp - written and placed in the public domain by Jeffrey Walton
2
3#include "pch.h"
4#include "config.h"
5
6#include "aria.h"
7#include "misc.h"
8#include "cpu.h"
9
10NAMESPACE_BEGIN(CryptoPP)
11NAMESPACE_BEGIN(ARIATab)
12
13extern const word32 S1[256];
14extern const word32 S2[256];
15extern const word32 X1[256];
16extern const word32 X2[256];
17extern const word32 KRK[3][4];
18
19NAMESPACE_END
20NAMESPACE_END
21
22NAMESPACE_BEGIN(CryptoPP)
23
24using CryptoPP::ARIATab::S1;
25using CryptoPP::ARIATab::S2;
26using CryptoPP::ARIATab::X1;
27using CryptoPP::ARIATab::X2;
28using CryptoPP::ARIATab::KRK;
29
30inline byte ARIA_BRF(const word32 x, const int y) {
31 return static_cast<byte>(GETBYTE(x, y));
32}
33
34// Key XOR Layer. Bumps the round key pointer.
35inline const byte* ARIA_KXL(const byte rk[16], word32 t[4]) {
37 NativeBlock::Put(rk, t)(t[0])(t[1])(t[2])(t[3]);
38 return rk+16;
39}
40
41// S-Box Layer 1 + M
42inline void SBL1_M(word32& T0, word32& T1, word32& T2, word32& T3) {
43 T0=S1[ARIA_BRF(T0,3)]^S2[ARIA_BRF(T0,2)]^X1[ARIA_BRF(T0,1)]^X2[ARIA_BRF(T0,0)];
44 T1=S1[ARIA_BRF(T1,3)]^S2[ARIA_BRF(T1,2)]^X1[ARIA_BRF(T1,1)]^X2[ARIA_BRF(T1,0)];
45 T2=S1[ARIA_BRF(T2,3)]^S2[ARIA_BRF(T2,2)]^X1[ARIA_BRF(T2,1)]^X2[ARIA_BRF(T2,0)];
46 T3=S1[ARIA_BRF(T3,3)]^S2[ARIA_BRF(T3,2)]^X1[ARIA_BRF(T3,1)]^X2[ARIA_BRF(T3,0)];
47}
48
49// S-Box Layer 2 + M
50inline void SBL2_M(word32& T0, word32& T1, word32& T2, word32& T3) {
51 T0=X1[ARIA_BRF(T0,3)]^X2[ARIA_BRF(T0,2)]^S1[ARIA_BRF(T0,1)]^S2[ARIA_BRF(T0,0)];
52 T1=X1[ARIA_BRF(T1,3)]^X2[ARIA_BRF(T1,2)]^S1[ARIA_BRF(T1,1)]^S2[ARIA_BRF(T1,0)];
53 T2=X1[ARIA_BRF(T2,3)]^X2[ARIA_BRF(T2,2)]^S1[ARIA_BRF(T2,1)]^S2[ARIA_BRF(T2,0)];
54 T3=X1[ARIA_BRF(T3,3)]^X2[ARIA_BRF(T3,2)]^S1[ARIA_BRF(T3,1)]^S2[ARIA_BRF(T3,0)];
55 }
56
57inline void ARIA_P(word32& T0, word32& T1, word32& T2, word32& T3) {
58 CRYPTOPP_UNUSED(T0);
59 T1 = ((T1<< 8)&0xff00ff00) ^ ((T1>> 8)&0x00ff00ff);
60 T2 = rotrConstant<16>(T2);
61 T3 = ByteReverse((T3));
62}
63
64inline void ARIA_M(word32& X, word32& Y) {
65 Y=X<<8 ^ X>>8 ^ X<<16 ^ X>>16 ^ X<<24 ^ X>>24;
66}
67
68
69inline void ARIA_MM(word32& T0, word32& T1, word32& T2, word32& T3) {
70 T1^=T2; T2^=T3; T0^=T1;
71 T3^=T1; T2^=T0; T1^=T2;
72}
73
74inline void ARIA_FO(word32 t[4]) {
75 SBL1_M(t[0],t[1],t[2],t[3]);
76 ARIA_MM(t[0],t[1],t[2],t[3]);
77 ARIA_P(t[0],t[1],t[2],t[3]);
78 ARIA_MM(t[0],t[1],t[2],t[3]);
79}
80
81inline void ARIA_FE(word32 t[4]) {
82 SBL2_M(t[0],t[1],t[2],t[3]);
83 ARIA_MM(t[0],t[1],t[2],t[3]);
84 ARIA_P(t[2],t[3],t[0],t[1]);
85 ARIA_MM(t[0],t[1],t[2],t[3]);
86}
87
88// n-bit right shift of Y XORed to X
89template <unsigned int N>
90inline void ARIA_GSRK(const word32 X[4], const word32 Y[4], word32 RK[4])
91{
92 // MSVC is not generating a "rotate immediate". Constify to help it along.
93 static const unsigned int Q = 4-(N/32);
94 static const unsigned int R = N % 32;
95
96 RK[0] = (X[0]) ^ ((Y[(Q )%4])>>R) ^ ((Y[(Q+3)%4])<<(32-R));
97 RK[1] = (X[1]) ^ ((Y[(Q+1)%4])>>R) ^ ((Y[(Q )%4])<<(32-R));
98 RK[2] = (X[2]) ^ ((Y[(Q+2)%4])>>R) ^ ((Y[(Q+1)%4])<<(32-R));
99 RK[3] = (X[3]) ^ ((Y[(Q+3)%4])>>R) ^ ((Y[(Q+2)%4])<<(32-R));
100}
101
102void ARIA::Base::UncheckedSetKey(const byte *key, unsigned int keylen, const NameValuePairs &params)
103{
104 CRYPTOPP_UNUSED(params);
105
106 m_rk.New(4*17); // round keys
107 m_w.New(4*24); // w0, w1, w2, w3, t and u
108
109 int Q, q, R, r;
110
111 switch (keylen)
112 {
113 case 16:
114 R = r = m_rounds = 12;
115 Q = q = 0;
116 break;
117 case 32:
118 R = r = m_rounds = 16;
119 Q = q = 2;
120 break;
121 case 24:
122 R = r = m_rounds = 14;
123 Q = q = 1;
124 break;
125 default:
126 Q = q = R = r = m_rounds = 0;
128 }
129
130 // w0-w3 each has room for 4 words (16 bytes). t and u are each 4 words (16 bytes) temp areas.
131 // The storage requrements for w0-w3, t and u are 96 bytes or 24 words.
132 word32 *w0 = m_w.data(), *w1 = m_w.data()+4, *w2 = m_w.data()+8, *w3 = m_w.data()+12, *t = m_w.data()+16;
133
135 block(w0[0])(w0[1])(w0[2])(w0[3]);
136
137 t[0]=w0[0]^KRK[q][0]; t[1]=w0[1]^KRK[q][1];
138 t[2]=w0[2]^KRK[q][2]; t[3]=w0[3]^KRK[q][3];
139
140 ARIA_FO(t);
141
142 if (keylen == 32)
143 {
144 block(w1[0])(w1[1])(w1[2])(w1[3]);
145 }
146 else if (keylen == 24)
147 {
148 block(w1[0])(w1[1]); w1[2] = w1[3] = 0;
149 }
150 else
151 {
152 w1[0]=w1[1]=w1[2]=w1[3]=0;
153 }
154
155 w1[0]^=t[0]; w1[1]^=t[1]; w1[2]^=t[2]; w1[3]^=t[3];
156 std::memcpy(t, w1, 16);
157
158 q = (q==2) ? 0 : (q+1);
159 t[0]^=KRK[q][0]; t[1]^=KRK[q][1]; t[2]^=KRK[q][2]; t[3]^=KRK[q][3];
160
161 ARIA_FE(t);
162
163 t[0]^=w0[0]; t[1]^=w0[1]; t[2]^=w0[2]; t[3]^=w0[3];
164 std::memcpy(w2, t, 16);
165
166 q = (q==2) ? 0 : (q+1);
167 t[0]^=KRK[q][0]; t[1]^=KRK[q][1]; t[2]^=KRK[q][2]; t[3]^=KRK[q][3];
168
169 ARIA_FO(t);
170
171 w3[0]=t[0]^w1[0]; w3[1]=t[1]^w1[1]; w3[2]=t[2]^w1[2]; w3[3]=t[3]^w1[3];
172
173 ARIA_GSRK<19>(w0, w1, m_rk + 0);
174 ARIA_GSRK<19>(w1, w2, m_rk + 4);
175 ARIA_GSRK<19>(w2, w3, m_rk + 8);
176 ARIA_GSRK<19>(w3, w0, m_rk + 12);
177 ARIA_GSRK<31>(w0, w1, m_rk + 16);
178 ARIA_GSRK<31>(w1, w2, m_rk + 20);
179 ARIA_GSRK<31>(w2, w3, m_rk + 24);
180 ARIA_GSRK<31>(w3, w0, m_rk + 28);
181 ARIA_GSRK<67>(w0, w1, m_rk + 32);
182 ARIA_GSRK<67>(w1, w2, m_rk + 36);
183 ARIA_GSRK<67>(w2, w3, m_rk + 40);
184 ARIA_GSRK<67>(w3, w0, m_rk + 44);
185 ARIA_GSRK<97>(w0, w1, m_rk + 48);
186
187 if (keylen > 16)
188 {
189 ARIA_GSRK<97>(w1, w2, m_rk + 52);
190 ARIA_GSRK<97>(w2, w3, m_rk + 56);
191
192 if (keylen > 24)
193 {
194 ARIA_GSRK< 97>(w3, w0, m_rk + 60);
195 ARIA_GSRK<109>(w0, w1, m_rk + 64);
196 }
197 }
198
199 // Decryption operation
200 if (!IsForwardTransformation())
201 {
202 word32 *a, *z, *s;
203 r = R; q = Q;
204
205 // s reuses w0 temp area
206 a=m_rk.data(); s=m_w.data()+0; z=a+r*4;
207 std::memcpy(t, a, 16); std::memcpy(a, z, 16); std::memcpy(z, t, 16);
208
209 a+=4; z-=4;
210 for (; a<z; a+=4, z-=4)
211 {
212 ARIA_M(a[0],t[0]); ARIA_M(a[1],t[1]); ARIA_M(a[2],t[2]); ARIA_M(a[3],t[3]);
213 ARIA_MM(t[0],t[1],t[2],t[3]); ARIA_P(t[0],t[1],t[2],t[3]); ARIA_MM(t[0],t[1],t[2],t[3]);
214 std::memcpy(s, t, 16);
215
216 ARIA_M(z[0],t[0]); ARIA_M(z[1],t[1]); ARIA_M(z[2],t[2]); ARIA_M(z[3],t[3]);
217 ARIA_MM(t[0],t[1],t[2],t[3]); ARIA_P(t[0],t[1],t[2],t[3]); ARIA_MM(t[0],t[1],t[2],t[3]);
218 std::memcpy(a, t, 16); std::memcpy(z, s, 16);
219 }
220
221 ARIA_M(a[0],t[0]); ARIA_M(a[1],t[1]); ARIA_M(a[2],t[2]); ARIA_M(a[3],t[3]);
222 ARIA_MM(t[0],t[1],t[2],t[3]); ARIA_P(t[0],t[1],t[2],t[3]); ARIA_MM(t[0],t[1],t[2],t[3]);
223 std::memcpy(z, t, 16);
224 }
225
226 // Silence warnings
227 CRYPTOPP_UNUSED(Q); CRYPTOPP_UNUSED(R);
228 CRYPTOPP_UNUSED(q); CRYPTOPP_UNUSED(r);
229}
230
231void ARIA::Base::ProcessAndXorBlock(const byte *inBlock, const byte *xorBlock, byte *outBlock) const
232{
233 const byte *rk = reinterpret_cast<const byte*>(m_rk.data());
234 word32 *t = const_cast<word32*>(m_w.data()+16);
235
236 // Timing attack countermeasure. See comments in Rijndael for more details.
237 // We used Yun's 32-bit implementation, so we use words rather than bytes.
238 const int cacheLineSize = GetCacheLineSize();
239 unsigned int i;
240 volatile word32 _u = 0;
241 word32 u = _u;
242
243 for (i=0; i<COUNTOF(S1); i+=cacheLineSize/(sizeof(S1[0])))
244 u |= *(S1+i);
245 t[0] |= u;
246
247 GetBlock<word32, BigEndian>block(inBlock);
248 block(t[0])(t[1])(t[2])(t[3]);
249
250 if (m_rounds > 12) {
251 rk = ARIA_KXL(rk, t); ARIA_FO(t);
252 rk = ARIA_KXL(rk, t); ARIA_FE(t);
253 }
254
255 if (m_rounds > 14) {
256 rk = ARIA_KXL(rk, t); ARIA_FO(t);
257 rk = ARIA_KXL(rk, t); ARIA_FE(t);
258 }
259
260 rk = ARIA_KXL(rk, t); ARIA_FO(t); rk = ARIA_KXL(rk, t); ARIA_FE(t);
261 rk = ARIA_KXL(rk, t); ARIA_FO(t); rk = ARIA_KXL(rk, t); ARIA_FE(t);
262 rk = ARIA_KXL(rk, t); ARIA_FO(t); rk = ARIA_KXL(rk, t); ARIA_FE(t);
263 rk = ARIA_KXL(rk, t); ARIA_FO(t); rk = ARIA_KXL(rk, t); ARIA_FE(t);
264 rk = ARIA_KXL(rk, t); ARIA_FO(t); rk = ARIA_KXL(rk, t); ARIA_FE(t);
265 rk = ARIA_KXL(rk, t); ARIA_FO(t); rk = ARIA_KXL(rk, t);
266
267#if (CRYPTOPP_LITTLE_ENDIAN)
268 if (xorBlock)
269 {
270 outBlock[ 0] = static_cast<byte>(X1[ARIA_BRF(t[0],3)] ) ^ rk[ 3] ^ xorBlock[ 0];
271 outBlock[ 1] = static_cast<byte>(X2[ARIA_BRF(t[0],2)]>>8) ^ rk[ 2] ^ xorBlock[ 1];
272 outBlock[ 2] = static_cast<byte>(S1[ARIA_BRF(t[0],1)] ) ^ rk[ 1] ^ xorBlock[ 2];
273 outBlock[ 3] = static_cast<byte>(S2[ARIA_BRF(t[0],0)] ) ^ rk[ 0] ^ xorBlock[ 3];
274 outBlock[ 4] = static_cast<byte>(X1[ARIA_BRF(t[1],3)] ) ^ rk[ 7] ^ xorBlock[ 4];
275 outBlock[ 5] = static_cast<byte>(X2[ARIA_BRF(t[1],2)]>>8) ^ rk[ 6] ^ xorBlock[ 5];
276 outBlock[ 6] = static_cast<byte>(S1[ARIA_BRF(t[1],1)] ) ^ rk[ 5] ^ xorBlock[ 6];
277 outBlock[ 7] = static_cast<byte>(S2[ARIA_BRF(t[1],0)] ) ^ rk[ 4] ^ xorBlock[ 7];
278 outBlock[ 8] = static_cast<byte>(X1[ARIA_BRF(t[2],3)] ) ^ rk[11] ^ xorBlock[ 8];
279 outBlock[ 9] = static_cast<byte>(X2[ARIA_BRF(t[2],2)]>>8) ^ rk[10] ^ xorBlock[ 9];
280 outBlock[10] = static_cast<byte>(S1[ARIA_BRF(t[2],1)] ) ^ rk[ 9] ^ xorBlock[10];
281 outBlock[11] = static_cast<byte>(S2[ARIA_BRF(t[2],0)] ) ^ rk[ 8] ^ xorBlock[11];
282 outBlock[12] = static_cast<byte>(X1[ARIA_BRF(t[3],3)] ) ^ rk[15] ^ xorBlock[12];
283 outBlock[13] = static_cast<byte>(X2[ARIA_BRF(t[3],2)]>>8) ^ rk[14] ^ xorBlock[13];
284 outBlock[14] = static_cast<byte>(S1[ARIA_BRF(t[3],1)] ) ^ rk[13] ^ xorBlock[14];
285 outBlock[15] = static_cast<byte>(S2[ARIA_BRF(t[3],0)] ) ^ rk[12] ^ xorBlock[15];
286 }
287 else
288 {
289 outBlock[ 0] = static_cast<byte>(X1[ARIA_BRF(t[0],3)] ) ^ rk[ 3];
290 outBlock[ 1] = static_cast<byte>(X2[ARIA_BRF(t[0],2)]>>8) ^ rk[ 2];
291 outBlock[ 2] = static_cast<byte>(S1[ARIA_BRF(t[0],1)] ) ^ rk[ 1];
292 outBlock[ 3] = static_cast<byte>(S2[ARIA_BRF(t[0],0)] ) ^ rk[ 0];
293 outBlock[ 4] = static_cast<byte>(X1[ARIA_BRF(t[1],3)] ) ^ rk[ 7];
294 outBlock[ 5] = static_cast<byte>(X2[ARIA_BRF(t[1],2)]>>8) ^ rk[ 6];
295 outBlock[ 6] = static_cast<byte>(S1[ARIA_BRF(t[1],1)] ) ^ rk[ 5];
296 outBlock[ 7] = static_cast<byte>(S2[ARIA_BRF(t[1],0)] ) ^ rk[ 4];
297 outBlock[ 8] = static_cast<byte>(X1[ARIA_BRF(t[2],3)] ) ^ rk[11];
298 outBlock[ 9] = static_cast<byte>(X2[ARIA_BRF(t[2],2)]>>8) ^ rk[10];
299 outBlock[10] = static_cast<byte>(S1[ARIA_BRF(t[2],1)] ) ^ rk[ 9];
300 outBlock[11] = static_cast<byte>(S2[ARIA_BRF(t[2],0)] ) ^ rk[ 8];
301 outBlock[12] = static_cast<byte>(X1[ARIA_BRF(t[3],3)] ) ^ rk[15];
302 outBlock[13] = static_cast<byte>(X2[ARIA_BRF(t[3],2)]>>8) ^ rk[14];
303 outBlock[14] = static_cast<byte>(S1[ARIA_BRF(t[3],1)] ) ^ rk[13];
304 outBlock[15] = static_cast<byte>(S2[ARIA_BRF(t[3],0)] ) ^ rk[12];
305 }
306#else
307 if (xorBlock)
308 {
309 outBlock[ 0] = static_cast<byte>(X1[ARIA_BRF(t[0],3)] ) ^ rk[ 0] ^ xorBlock[ 0];
310 outBlock[ 1] = static_cast<byte>(X2[ARIA_BRF(t[0],2)]>>8) ^ rk[ 1] ^ xorBlock[ 1];
311 outBlock[ 2] = static_cast<byte>(S1[ARIA_BRF(t[0],1)] ) ^ rk[ 2] ^ xorBlock[ 2];
312 outBlock[ 3] = static_cast<byte>(S2[ARIA_BRF(t[0],0)] ) ^ rk[ 3] ^ xorBlock[ 3];
313 outBlock[ 4] = static_cast<byte>(X1[ARIA_BRF(t[1],3)] ) ^ rk[ 4] ^ xorBlock[ 4];
314 outBlock[ 5] = static_cast<byte>(X2[ARIA_BRF(t[1],2)]>>8) ^ rk[ 5] ^ xorBlock[ 5];
315 outBlock[ 6] = static_cast<byte>(S1[ARIA_BRF(t[1],1)] ) ^ rk[ 6] ^ xorBlock[ 6];
316 outBlock[ 7] = static_cast<byte>(S2[ARIA_BRF(t[1],0)] ) ^ rk[ 7] ^ xorBlock[ 7];
317 outBlock[ 8] = static_cast<byte>(X1[ARIA_BRF(t[2],3)] ) ^ rk[ 8] ^ xorBlock[ 8];
318 outBlock[ 9] = static_cast<byte>(X2[ARIA_BRF(t[2],2)]>>8) ^ rk[ 9] ^ xorBlock[ 9];
319 outBlock[10] = static_cast<byte>(S1[ARIA_BRF(t[2],1)] ) ^ rk[10] ^ xorBlock[10];
320 outBlock[11] = static_cast<byte>(S2[ARIA_BRF(t[2],0)] ) ^ rk[11] ^ xorBlock[11];
321 outBlock[12] = static_cast<byte>(X1[ARIA_BRF(t[3],3)] ) ^ rk[12] ^ xorBlock[12];
322 outBlock[13] = static_cast<byte>(X2[ARIA_BRF(t[3],2)]>>8) ^ rk[13] ^ xorBlock[13];
323 outBlock[14] = static_cast<byte>(S1[ARIA_BRF(t[3],1)] ) ^ rk[14] ^ xorBlock[14];
324 outBlock[15] = static_cast<byte>(S2[ARIA_BRF(t[3],0)] ) ^ rk[15] ^ xorBlock[15];
325 }
326 else
327 {
328 outBlock[ 0] = static_cast<byte>(X1[ARIA_BRF(t[0],3)] ) ^ rk[ 0];
329 outBlock[ 1] = static_cast<byte>(X2[ARIA_BRF(t[0],2)]>>8) ^ rk[ 1];
330 outBlock[ 2] = static_cast<byte>(S1[ARIA_BRF(t[0],1)] ) ^ rk[ 2];
331 outBlock[ 3] = static_cast<byte>(S2[ARIA_BRF(t[0],0)] ) ^ rk[ 3];
332 outBlock[ 4] = static_cast<byte>(X1[ARIA_BRF(t[1],3)] ) ^ rk[ 4];
333 outBlock[ 5] = static_cast<byte>(X2[ARIA_BRF(t[1],2)]>>8) ^ rk[ 5];
334 outBlock[ 6] = static_cast<byte>(S1[ARIA_BRF(t[1],1)] ) ^ rk[ 6];
335 outBlock[ 7] = static_cast<byte>(S2[ARIA_BRF(t[1],0)] ) ^ rk[ 7];
336 outBlock[ 8] = static_cast<byte>(X1[ARIA_BRF(t[2],3)] ) ^ rk[ 8];
337 outBlock[ 9] = static_cast<byte>(X2[ARIA_BRF(t[2],2)]>>8) ^ rk[ 9];
338 outBlock[10] = static_cast<byte>(S1[ARIA_BRF(t[2],1)] ) ^ rk[10];
339 outBlock[11] = static_cast<byte>(S2[ARIA_BRF(t[2],0)] ) ^ rk[11];
340 outBlock[12] = static_cast<byte>(X1[ARIA_BRF(t[3],3)] ) ^ rk[12];
341 outBlock[13] = static_cast<byte>(X2[ARIA_BRF(t[3],2)]>>8) ^ rk[13];
342 outBlock[14] = static_cast<byte>(S1[ARIA_BRF(t[3],1)] ) ^ rk[14];
343 outBlock[15] = static_cast<byte>(S2[ARIA_BRF(t[3],0)] ) ^ rk[15];
344 }
345#endif // CRYPTOPP_LITTLE_ENDIAN
346}
347
348NAMESPACE_END
Classes for the ARIA block cipher.
Access a block of memory.
Definition misc.h:2975
Interface for retrieving values given their names.
Definition cryptlib.h:327
A::pointer data()
Provides a pointer to the first element in the memory block.
Definition secblock.h:857
void New(size_type newSize)
Change size without preserving contents.
Definition secblock.h:1126
Library configuration file.
unsigned int word32
32-bit unsigned datatype
Definition config_int.h:72
Functions for CPU features and intrinsics.
Utility functions for the Crypto++ library.
byte ByteReverse(byte value)
Reverses bytes in a 8-bit value.
Definition misc.h:2231
#define COUNTOF(arr)
Counts elements in an array.
Definition misc.h:193
Crypto++ library namespace.
Precompiled header file.
Access a block of memory.
Definition misc.h:3053
#define CRYPTOPP_ASSERT(exp)
Debugging and diagnostic assertion.
Definition trap.h:68