• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* Aes.c -- AES encryption / decryption
2 2021-05-13 : Igor Pavlov : Public domain */
3 
4 #include "Precomp.h"
5 
6 #include "CpuArch.h"
7 #include "Aes.h"
8 
9 AES_CODE_FUNC g_AesCbc_Decode;
10 #ifndef _SFX
11 AES_CODE_FUNC g_AesCbc_Encode;
12 AES_CODE_FUNC g_AesCtr_Code;
13 UInt32 g_Aes_SupportedFunctions_Flags;
14 #endif
15 
16 static UInt32 T[256 * 4];
17 static const Byte Sbox[256] = {
18   0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
19   0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
20   0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
21   0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
22   0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
23   0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
24   0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
25   0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
26   0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
27   0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
28   0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
29   0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
30   0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
31   0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
32   0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
33   0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16};
34 
35 
36 static UInt32 D[256 * 4];
37 static Byte InvS[256];
38 
39 #define xtime(x) ((((x) << 1) ^ (((x) & 0x80) != 0 ? 0x1B : 0)) & 0xFF)
40 
41 #define Ui32(a0, a1, a2, a3) ((UInt32)(a0) | ((UInt32)(a1) << 8) | ((UInt32)(a2) << 16) | ((UInt32)(a3) << 24))
42 
43 #define gb0(x) ( (x)          & 0xFF)
44 #define gb1(x) (((x) >> ( 8)) & 0xFF)
45 #define gb2(x) (((x) >> (16)) & 0xFF)
46 #define gb3(x) (((x) >> (24)))
47 
48 #define gb(n, x) gb ## n(x)
49 
50 #define TT(x) (T + (x << 8))
51 #define DD(x) (D + (x << 8))
52 
53 
54 // #define _SHOW_AES_STATUS
55 
56 #ifdef MY_CPU_X86_OR_AMD64
57   #define USE_HW_AES
58 #elif defined(MY_CPU_ARM_OR_ARM64) && defined(MY_CPU_LE)
59   #if defined(__clang__)
60     #if (__clang_major__ >= 16) // fix that check
61       #define USE_HW_AES
62     #endif
63   #elif defined(__GNUC__)
64     #if (__GNUC__ >= 6) // fix that check
65       #define USE_HW_AES
66     #endif
67   #elif defined(_MSC_VER)
68     #if _MSC_VER >= 1910
69       #define USE_HW_AES
70     #endif
71   #endif
72 #endif
73 
74 #ifdef USE_HW_AES
75 #ifdef _SHOW_AES_STATUS
76 #include <stdio.h>
77 #define _PRF(x) x
78 #else
79 #define _PRF(x)
80 #endif
81 #endif
82 
83 
AesGenTables(void)84 void AesGenTables(void)
85 {
86   unsigned i;
87   for (i = 0; i < 256; i++)
88     InvS[Sbox[i]] = (Byte)i;
89 
90   for (i = 0; i < 256; i++)
91   {
92     {
93       UInt32 a1 = Sbox[i];
94       UInt32 a2 = xtime(a1);
95       UInt32 a3 = a2 ^ a1;
96       TT(0)[i] = Ui32(a2, a1, a1, a3);
97       TT(1)[i] = Ui32(a3, a2, a1, a1);
98       TT(2)[i] = Ui32(a1, a3, a2, a1);
99       TT(3)[i] = Ui32(a1, a1, a3, a2);
100     }
101     {
102       UInt32 a1 = InvS[i];
103       UInt32 a2 = xtime(a1);
104       UInt32 a4 = xtime(a2);
105       UInt32 a8 = xtime(a4);
106       UInt32 a9 = a8 ^ a1;
107       UInt32 aB = a8 ^ a2 ^ a1;
108       UInt32 aD = a8 ^ a4 ^ a1;
109       UInt32 aE = a8 ^ a4 ^ a2;
110       DD(0)[i] = Ui32(aE, a9, aD, aB);
111       DD(1)[i] = Ui32(aB, aE, a9, aD);
112       DD(2)[i] = Ui32(aD, aB, aE, a9);
113       DD(3)[i] = Ui32(a9, aD, aB, aE);
114     }
115   }
116 
117   {
118   AES_CODE_FUNC d = AesCbc_Decode;
119   #ifndef _SFX
120   AES_CODE_FUNC e = AesCbc_Encode;
121   AES_CODE_FUNC c = AesCtr_Code;
122   UInt32 flags = 0;
123   #endif
124 
125   #ifdef USE_HW_AES
126   if (CPU_IsSupported_AES())
127   {
128     // #pragma message ("AES HW")
129     _PRF(printf("\n===AES HW\n"));
130     d = AesCbc_Decode_HW;
131 
132     #ifndef _SFX
133     e = AesCbc_Encode_HW;
134     c = AesCtr_Code_HW;
135     flags = k_Aes_SupportedFunctions_HW;
136     #endif
137 
138     #ifdef MY_CPU_X86_OR_AMD64
139     if (CPU_IsSupported_VAES_AVX2())
140     {
141       _PRF(printf("\n===vaes avx2\n"));
142       d = AesCbc_Decode_HW_256;
143       #ifndef _SFX
144       c = AesCtr_Code_HW_256;
145       flags |= k_Aes_SupportedFunctions_HW_256;
146       #endif
147     }
148     #endif
149   }
150   #endif
151 
152   g_AesCbc_Decode = d;
153   #ifndef _SFX
154   g_AesCbc_Encode = e;
155   g_AesCtr_Code = c;
156   g_Aes_SupportedFunctions_Flags = flags;
157   #endif
158   }
159 }
160 
161 
162 #define HT(i, x, s) TT(x)[gb(x, s[(i + x) & 3])]
163 
164 #define HT4(m, i, s, p) m[i] = \
165     HT(i, 0, s) ^ \
166     HT(i, 1, s) ^ \
167     HT(i, 2, s) ^ \
168     HT(i, 3, s) ^ w[p + i]
169 
170 #define HT16(m, s, p) \
171     HT4(m, 0, s, p); \
172     HT4(m, 1, s, p); \
173     HT4(m, 2, s, p); \
174     HT4(m, 3, s, p); \
175 
176 #define FT(i, x) Sbox[gb(x, m[(i + x) & 3])]
177 #define FT4(i) dest[i] = Ui32(FT(i, 0), FT(i, 1), FT(i, 2), FT(i, 3)) ^ w[i];
178 
179 
180 #define HD(i, x, s) DD(x)[gb(x, s[(i - x) & 3])]
181 
182 #define HD4(m, i, s, p) m[i] = \
183     HD(i, 0, s) ^ \
184     HD(i, 1, s) ^ \
185     HD(i, 2, s) ^ \
186     HD(i, 3, s) ^ w[p + i];
187 
188 #define HD16(m, s, p) \
189     HD4(m, 0, s, p); \
190     HD4(m, 1, s, p); \
191     HD4(m, 2, s, p); \
192     HD4(m, 3, s, p); \
193 
194 #define FD(i, x) InvS[gb(x, m[(i - x) & 3])]
195 #define FD4(i) dest[i] = Ui32(FD(i, 0), FD(i, 1), FD(i, 2), FD(i, 3)) ^ w[i];
196 
Aes_SetKey_Enc(UInt32 * w,const Byte * key,unsigned keySize)197 void MY_FAST_CALL Aes_SetKey_Enc(UInt32 *w, const Byte *key, unsigned keySize)
198 {
199   unsigned i, m;
200   const UInt32 *wLim;
201   UInt32 t;
202   UInt32 rcon = 1;
203 
204   keySize /= 4;
205   w[0] = ((UInt32)keySize / 2) + 3;
206   w += 4;
207 
208   for (i = 0; i < keySize; i++, key += 4)
209     w[i] = GetUi32(key);
210 
211   t = w[(size_t)keySize - 1];
212   wLim = w + (size_t)keySize * 3 + 28;
213   m = 0;
214   do
215   {
216     if (m == 0)
217     {
218       t = Ui32(Sbox[gb1(t)] ^ rcon, Sbox[gb2(t)], Sbox[gb3(t)], Sbox[gb0(t)]);
219       rcon <<= 1;
220       if (rcon & 0x100)
221         rcon = 0x1b;
222       m = keySize;
223     }
224     else if (m == 4 && keySize > 6)
225       t = Ui32(Sbox[gb0(t)], Sbox[gb1(t)], Sbox[gb2(t)], Sbox[gb3(t)]);
226     m--;
227     t ^= w[0];
228     w[keySize] = t;
229   }
230   while (++w != wLim);
231 }
232 
Aes_SetKey_Dec(UInt32 * w,const Byte * key,unsigned keySize)233 void MY_FAST_CALL Aes_SetKey_Dec(UInt32 *w, const Byte *key, unsigned keySize)
234 {
235   unsigned i, num;
236   Aes_SetKey_Enc(w, key, keySize);
237   num = keySize + 20;
238   w += 8;
239   for (i = 0; i < num; i++)
240   {
241     UInt32 r = w[i];
242     w[i] =
243       DD(0)[Sbox[gb0(r)]] ^
244       DD(1)[Sbox[gb1(r)]] ^
245       DD(2)[Sbox[gb2(r)]] ^
246       DD(3)[Sbox[gb3(r)]];
247   }
248 }
249 
250 /* Aes_Encode and Aes_Decode functions work with little-endian words.
251   src and dest are pointers to 4 UInt32 words.
252   src and dest can point to same block */
253 
254 // MY_FORCE_INLINE
Aes_Encode(const UInt32 * w,UInt32 * dest,const UInt32 * src)255 static void Aes_Encode(const UInt32 *w, UInt32 *dest, const UInt32 *src)
256 {
257   UInt32 s[4];
258   UInt32 m[4];
259   UInt32 numRounds2 = w[0];
260   w += 4;
261   s[0] = src[0] ^ w[0];
262   s[1] = src[1] ^ w[1];
263   s[2] = src[2] ^ w[2];
264   s[3] = src[3] ^ w[3];
265   w += 4;
266   for (;;)
267   {
268     HT16(m, s, 0);
269     if (--numRounds2 == 0)
270       break;
271     HT16(s, m, 4);
272     w += 8;
273   }
274   w += 4;
275   FT4(0); FT4(1); FT4(2); FT4(3);
276 }
277 
278 MY_FORCE_INLINE
Aes_Decode(const UInt32 * w,UInt32 * dest,const UInt32 * src)279 static void Aes_Decode(const UInt32 *w, UInt32 *dest, const UInt32 *src)
280 {
281   UInt32 s[4];
282   UInt32 m[4];
283   UInt32 numRounds2 = w[0];
284   w += 4 + numRounds2 * 8;
285   s[0] = src[0] ^ w[0];
286   s[1] = src[1] ^ w[1];
287   s[2] = src[2] ^ w[2];
288   s[3] = src[3] ^ w[3];
289   for (;;)
290   {
291     w -= 8;
292     HD16(m, s, 4);
293     if (--numRounds2 == 0)
294       break;
295     HD16(s, m, 0);
296   }
297   FD4(0); FD4(1); FD4(2); FD4(3);
298 }
299 
AesCbc_Init(UInt32 * p,const Byte * iv)300 void AesCbc_Init(UInt32 *p, const Byte *iv)
301 {
302   unsigned i;
303   for (i = 0; i < 4; i++)
304     p[i] = GetUi32(iv + i * 4);
305 }
306 
AesCbc_Encode(UInt32 * p,Byte * data,size_t numBlocks)307 void MY_FAST_CALL AesCbc_Encode(UInt32 *p, Byte *data, size_t numBlocks)
308 {
309   for (; numBlocks != 0; numBlocks--, data += AES_BLOCK_SIZE)
310   {
311     p[0] ^= GetUi32(data);
312     p[1] ^= GetUi32(data + 4);
313     p[2] ^= GetUi32(data + 8);
314     p[3] ^= GetUi32(data + 12);
315 
316     Aes_Encode(p + 4, p, p);
317 
318     SetUi32(data,      p[0]);
319     SetUi32(data + 4,  p[1]);
320     SetUi32(data + 8,  p[2]);
321     SetUi32(data + 12, p[3]);
322   }
323 }
324 
AesCbc_Decode(UInt32 * p,Byte * data,size_t numBlocks)325 void MY_FAST_CALL AesCbc_Decode(UInt32 *p, Byte *data, size_t numBlocks)
326 {
327   UInt32 in[4], out[4];
328   for (; numBlocks != 0; numBlocks--, data += AES_BLOCK_SIZE)
329   {
330     in[0] = GetUi32(data);
331     in[1] = GetUi32(data + 4);
332     in[2] = GetUi32(data + 8);
333     in[3] = GetUi32(data + 12);
334 
335     Aes_Decode(p + 4, out, in);
336 
337     SetUi32(data,      p[0] ^ out[0]);
338     SetUi32(data + 4,  p[1] ^ out[1]);
339     SetUi32(data + 8,  p[2] ^ out[2]);
340     SetUi32(data + 12, p[3] ^ out[3]);
341 
342     p[0] = in[0];
343     p[1] = in[1];
344     p[2] = in[2];
345     p[3] = in[3];
346   }
347 }
348 
AesCtr_Code(UInt32 * p,Byte * data,size_t numBlocks)349 void MY_FAST_CALL AesCtr_Code(UInt32 *p, Byte *data, size_t numBlocks)
350 {
351   for (; numBlocks != 0; numBlocks--)
352   {
353     UInt32 temp[4];
354     unsigned i;
355 
356     if (++p[0] == 0)
357       p[1]++;
358 
359     Aes_Encode(p + 4, temp, p);
360 
361     for (i = 0; i < 4; i++, data += 4)
362     {
363       UInt32 t = temp[i];
364 
365       #ifdef MY_CPU_LE_UNALIGN
366         *((UInt32 *)(void *)data) ^= t;
367       #else
368         data[0] = (Byte)(data[0] ^ (t & 0xFF));
369         data[1] = (Byte)(data[1] ^ ((t >> 8) & 0xFF));
370         data[2] = (Byte)(data[2] ^ ((t >> 16) & 0xFF));
371         data[3] = (Byte)(data[3] ^ ((t >> 24)));
372       #endif
373     }
374   }
375 }
376