28 #if (CRYPTOPP_AESNI_AVAILABLE) 29 # include <smmintrin.h> 30 # include <wmmintrin.h> 33 #if (CRYPTOPP_ARM_NEON_AVAILABLE) 34 # include <arm_neon.h> 37 #if (CRYPTOPP_ARM_ACLE_AVAILABLE) 39 # include <arm_acle.h> 42 #if defined(CRYPTOPP_POWER8_AES_AVAILABLE) 46 #ifdef CRYPTOPP_GNU_STYLE_INLINE_ASSEMBLY 51 #ifndef EXCEPTION_EXECUTE_HANDLER 52 # define EXCEPTION_EXECUTE_HANDLER 1 56 #define M128_CAST(x) ((__m128i *)(void *)(x)) 57 #define CONST_M128_CAST(x) ((const __m128i *)(const void *)(x)) 60 extern const char RIJNDAEL_SIMD_FNAME[] = __FILE__;
66 #ifdef CRYPTOPP_GNU_STYLE_INLINE_ASSEMBLY 68 typedef void (*SigHandler)(int);
70 static jmp_buf s_jmpSIGILL;
71 static void SigIllHandler(
int)
73 longjmp(s_jmpSIGILL, 1);
76 #endif // Not CRYPTOPP_MS_STYLE_INLINE_ASSEMBLY 78 #if (CRYPTOPP_BOOL_ARM32 || CRYPTOPP_BOOL_ARMV8) 81 #if defined(CRYPTOPP_NO_CPU_FEATURE_PROBES) 83 #elif (CRYPTOPP_ARM_AES_AVAILABLE) 84 # if defined(CRYPTOPP_MS_STYLE_INLINE_ASSEMBLY) 85 volatile bool result =
true;
89 uint8x16_t data = vdupq_n_u8(0), key = vdupq_n_u8(0);
90 uint8x16_t r1 = vaeseq_u8(data, key);
91 uint8x16_t r2 = vaesdq_u8(data, key);
95 result = !!(vgetq_lane_u8(r1,0) | vgetq_lane_u8(r2,7));
97 __except (EXCEPTION_EXECUTE_HANDLER)
106 volatile bool result =
true;
108 volatile SigHandler oldHandler = signal(SIGILL, SigIllHandler);
109 if (oldHandler == SIG_ERR)
112 volatile sigset_t oldMask;
113 if (sigprocmask(0, NULLPTR, (sigset_t*)&oldMask))
116 if (setjmp(s_jmpSIGILL))
120 uint8x16_t data = vdupq_n_u8(0), key = vdupq_n_u8(0);
121 uint8x16_t r1 = vaeseq_u8(data, key);
122 uint8x16_t r2 = vaesdq_u8(data, key);
124 r2 = vaesimcq_u8(r2);
127 result = !!(vgetq_lane_u8(r1,0) | vgetq_lane_u8(r2,7));
130 sigprocmask(SIG_SETMASK, (sigset_t*)&oldMask, NULLPTR);
131 signal(SIGILL, oldHandler);
136 #endif // CRYPTOPP_ARM_AES_AVAILABLE 138 #endif // ARM32 or ARM64 142 #if (CRYPTOPP_ARM_AES_AVAILABLE) 144 ANONYMOUS_NAMESPACE_BEGIN
146 static inline void ARMV8_Enc_Block(uint64x2_t &data,
const word32 *subkeys,
unsigned int rounds)
149 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
150 uint8x16_t block = vreinterpretq_u8_u64(data);
153 block = vaeseq_u8(block, vld1q_u8(keys+0*16));
155 block = vaesmcq_u8(block);
157 for (
unsigned int i=1; i<rounds-1; i+=2)
160 block = vaeseq_u8(block, vld1q_u8(keys+i*16));
162 block = vaesmcq_u8(block);
164 block = vaeseq_u8(block, vld1q_u8(keys+(i+1)*16));
166 block = vaesmcq_u8(block);
170 block = vaeseq_u8(block, vld1q_u8(keys+(rounds-1)*16));
172 block = veorq_u8(block, vld1q_u8(keys+rounds*16));
174 data = vreinterpretq_u64_u8(block);
177 static inline void ARMV8_Enc_6_Blocks(uint64x2_t &data0, uint64x2_t &data1,
178 uint64x2_t &data2, uint64x2_t &data3, uint64x2_t &data4, uint64x2_t &data5,
179 const word32 *subkeys,
unsigned int rounds)
182 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
184 uint8x16_t block0 = vreinterpretq_u8_u64(data0);
185 uint8x16_t block1 = vreinterpretq_u8_u64(data1);
186 uint8x16_t block2 = vreinterpretq_u8_u64(data2);
187 uint8x16_t block3 = vreinterpretq_u8_u64(data3);
188 uint8x16_t block4 = vreinterpretq_u8_u64(data4);
189 uint8x16_t block5 = vreinterpretq_u8_u64(data5);
192 for (
unsigned int i=0; i<rounds-1; ++i)
194 uint8x16_t key = vld1q_u8(keys+i*16);
196 block0 = vaeseq_u8(block0, key);
198 block0 = vaesmcq_u8(block0);
200 block1 = vaeseq_u8(block1, key);
202 block1 = vaesmcq_u8(block1);
204 block2 = vaeseq_u8(block2, key);
206 block2 = vaesmcq_u8(block2);
208 block3 = vaeseq_u8(block3, key);
210 block3 = vaesmcq_u8(block3);
212 block4 = vaeseq_u8(block4, key);
214 block4 = vaesmcq_u8(block4);
216 block5 = vaeseq_u8(block5, key);
218 block5 = vaesmcq_u8(block5);
222 key = vld1q_u8(keys+(rounds-1)*16);
223 block0 = vaeseq_u8(block0, key);
224 block1 = vaeseq_u8(block1, key);
225 block2 = vaeseq_u8(block2, key);
226 block3 = vaeseq_u8(block3, key);
227 block4 = vaeseq_u8(block4, key);
228 block5 = vaeseq_u8(block5, key);
231 key = vld1q_u8(keys+rounds*16);
232 data0 = vreinterpretq_u64_u8(veorq_u8(block0, key));
233 data1 = vreinterpretq_u64_u8(veorq_u8(block1, key));
234 data2 = vreinterpretq_u64_u8(veorq_u8(block2, key));
235 data3 = vreinterpretq_u64_u8(veorq_u8(block3, key));
236 data4 = vreinterpretq_u64_u8(veorq_u8(block4, key));
237 data5 = vreinterpretq_u64_u8(veorq_u8(block5, key));
240 static inline void ARMV8_Dec_Block(uint64x2_t &data,
const word32 *subkeys,
unsigned int rounds)
243 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
244 uint8x16_t block = vreinterpretq_u8_u64(data);
247 block = vaesdq_u8(block, vld1q_u8(keys+0*16));
249 block = vaesimcq_u8(block);
251 for (
unsigned int i=1; i<rounds-1; i+=2)
254 block = vaesdq_u8(block, vld1q_u8(keys+i*16));
256 block = vaesimcq_u8(block);
258 block = vaesdq_u8(block, vld1q_u8(keys+(i+1)*16));
260 block = vaesimcq_u8(block);
264 block = vaesdq_u8(block, vld1q_u8(keys+(rounds-1)*16));
266 block = veorq_u8(block, vld1q_u8(keys+rounds*16));
268 data = vreinterpretq_u64_u8(block);
271 static inline void ARMV8_Dec_6_Blocks(uint64x2_t &data0, uint64x2_t &data1,
272 uint64x2_t &data2, uint64x2_t &data3, uint64x2_t &data4, uint64x2_t &data5,
273 const word32 *subkeys,
unsigned int rounds)
276 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
278 uint8x16_t block0 = vreinterpretq_u8_u64(data0);
279 uint8x16_t block1 = vreinterpretq_u8_u64(data1);
280 uint8x16_t block2 = vreinterpretq_u8_u64(data2);
281 uint8x16_t block3 = vreinterpretq_u8_u64(data3);
282 uint8x16_t block4 = vreinterpretq_u8_u64(data4);
283 uint8x16_t block5 = vreinterpretq_u8_u64(data5);
286 for (
unsigned int i=0; i<rounds-1; ++i)
288 key = vld1q_u8(keys+i*16);
290 block0 = vaesdq_u8(block0, key);
292 block0 = vaesimcq_u8(block0);
294 block1 = vaesdq_u8(block1, key);
296 block1 = vaesimcq_u8(block1);
298 block2 = vaesdq_u8(block2, key);
300 block2 = vaesimcq_u8(block2);
302 block3 = vaesdq_u8(block3, key);
304 block3 = vaesimcq_u8(block3);
306 block4 = vaesdq_u8(block4, key);
308 block4 = vaesimcq_u8(block4);
310 block5 = vaesdq_u8(block5, key);
312 block5 = vaesimcq_u8(block5);
316 key = vld1q_u8(keys+(rounds-1)*16);
317 block0 = vaesdq_u8(block0, key);
318 block1 = vaesdq_u8(block1, key);
319 block2 = vaesdq_u8(block2, key);
320 block3 = vaesdq_u8(block3, key);
321 block4 = vaesdq_u8(block4, key);
322 block5 = vaesdq_u8(block5, key);
325 key = vld1q_u8(keys+rounds*16);
326 data0 = vreinterpretq_u64_u8(veorq_u8(block0, key));
327 data1 = vreinterpretq_u64_u8(veorq_u8(block1, key));
328 data2 = vreinterpretq_u64_u8(veorq_u8(block2, key));
329 data3 = vreinterpretq_u64_u8(veorq_u8(block3, key));
330 data4 = vreinterpretq_u64_u8(veorq_u8(block4, key));
331 data5 = vreinterpretq_u64_u8(veorq_u8(block5, key));
334 ANONYMOUS_NAMESPACE_END
336 size_t Rijndael_Enc_AdvancedProcessBlocks_ARMV8(
const word32 *subKeys,
size_t rounds,
337 const byte *inBlocks,
const byte *xorBlocks, byte *outBlocks,
size_t length, word32 flags)
339 return AdvancedProcessBlocks128_6x1_NEON(ARMV8_Enc_Block, ARMV8_Enc_6_Blocks,
340 subKeys, rounds, inBlocks, xorBlocks, outBlocks, length, flags);
343 size_t Rijndael_Dec_AdvancedProcessBlocks_ARMV8(
const word32 *subKeys,
size_t rounds,
344 const byte *inBlocks,
const byte *xorBlocks, byte *outBlocks,
size_t length, word32 flags)
346 return AdvancedProcessBlocks128_6x1_NEON(ARMV8_Dec_Block, ARMV8_Dec_6_Blocks,
347 subKeys, rounds, inBlocks, xorBlocks, outBlocks, length, flags);
350 #endif // CRYPTOPP_ARM_AES_AVAILABLE 354 #if (CRYPTOPP_AESNI_AVAILABLE) 356 ANONYMOUS_NAMESPACE_BEGIN
359 CRYPTOPP_ALIGN_DATA(16)
360 const word32 s_rconLE[] = {
361 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1B, 0x36
364 static inline void AESNI_Enc_Block(__m128i &block, MAYBE_CONST word32 *subkeys,
unsigned int rounds)
366 const __m128i* skeys =
reinterpret_cast<const __m128i*
>(subkeys);
368 block = _mm_xor_si128(block, skeys[0]);
369 for (
unsigned int i=1; i<rounds-1; i+=2)
371 block = _mm_aesenc_si128(block, skeys[i]);
372 block = _mm_aesenc_si128(block, skeys[i+1]);
374 block = _mm_aesenc_si128(block, skeys[rounds-1]);
375 block = _mm_aesenclast_si128(block, skeys[rounds]);
378 static inline void AESNI_Enc_4_Blocks(__m128i &block0, __m128i &block1, __m128i &block2, __m128i &block3,
379 MAYBE_CONST word32 *subkeys,
unsigned int rounds)
381 const __m128i* skeys =
reinterpret_cast<const __m128i*
>(subkeys);
383 __m128i rk = skeys[0];
384 block0 = _mm_xor_si128(block0, rk);
385 block1 = _mm_xor_si128(block1, rk);
386 block2 = _mm_xor_si128(block2, rk);
387 block3 = _mm_xor_si128(block3, rk);
388 for (
unsigned int i=1; i<rounds; i++)
391 block0 = _mm_aesenc_si128(block0, rk);
392 block1 = _mm_aesenc_si128(block1, rk);
393 block2 = _mm_aesenc_si128(block2, rk);
394 block3 = _mm_aesenc_si128(block3, rk);
397 block0 = _mm_aesenclast_si128(block0, rk);
398 block1 = _mm_aesenclast_si128(block1, rk);
399 block2 = _mm_aesenclast_si128(block2, rk);
400 block3 = _mm_aesenclast_si128(block3, rk);
403 static inline void AESNI_Dec_Block(__m128i &block, MAYBE_CONST word32 *subkeys,
unsigned int rounds)
405 const __m128i* skeys =
reinterpret_cast<const __m128i*
>(subkeys);
407 block = _mm_xor_si128(block, skeys[0]);
408 for (
unsigned int i=1; i<rounds-1; i+=2)
410 block = _mm_aesdec_si128(block, skeys[i]);
411 block = _mm_aesdec_si128(block, skeys[i+1]);
413 block = _mm_aesdec_si128(block, skeys[rounds-1]);
414 block = _mm_aesdeclast_si128(block, skeys[rounds]);
417 static inline void AESNI_Dec_4_Blocks(__m128i &block0, __m128i &block1, __m128i &block2, __m128i &block3,
418 MAYBE_CONST word32 *subkeys,
unsigned int rounds)
420 const __m128i* skeys =
reinterpret_cast<const __m128i*
>(subkeys);
422 __m128i rk = skeys[0];
423 block0 = _mm_xor_si128(block0, rk);
424 block1 = _mm_xor_si128(block1, rk);
425 block2 = _mm_xor_si128(block2, rk);
426 block3 = _mm_xor_si128(block3, rk);
427 for (
unsigned int i=1; i<rounds; i++)
430 block0 = _mm_aesdec_si128(block0, rk);
431 block1 = _mm_aesdec_si128(block1, rk);
432 block2 = _mm_aesdec_si128(block2, rk);
433 block3 = _mm_aesdec_si128(block3, rk);
436 block0 = _mm_aesdeclast_si128(block0, rk);
437 block1 = _mm_aesdeclast_si128(block1, rk);
438 block2 = _mm_aesdeclast_si128(block2, rk);
439 block3 = _mm_aesdeclast_si128(block3, rk);
442 ANONYMOUS_NAMESPACE_END
444 void Rijndael_UncheckedSetKey_SSE4_AESNI(
const byte *userKey,
size_t keyLen, word32 *rk)
446 const size_t rounds = keyLen / 4 + 6;
447 const word32 *rc = s_rconLE;
449 __m128i temp = _mm_loadu_si128(M128_CAST(userKey+keyLen-16));
450 std::memcpy(rk, userKey, keyLen);
453 const size_t keySize = 4*(rounds+1);
454 const word32* end = rk + keySize;
458 rk[keyLen/4] = rk[0] ^ _mm_extract_epi32(_mm_aeskeygenassist_si128(temp, 0), 3) ^ *(rc++);
459 rk[keyLen/4+1] = rk[1] ^ rk[keyLen/4];
460 rk[keyLen/4+2] = rk[2] ^ rk[keyLen/4+1];
461 rk[keyLen/4+3] = rk[3] ^ rk[keyLen/4+2];
463 if (rk + keyLen/4 + 4 == end)
468 rk[10] = rk[ 4] ^ rk[ 9];
469 rk[11] = rk[ 5] ^ rk[10];
470 temp = _mm_insert_epi32(temp, rk[11], 3);
472 else if (keyLen == 32)
474 temp = _mm_insert_epi32(temp, rk[11], 3);
475 rk[12] = rk[ 4] ^ _mm_extract_epi32(_mm_aeskeygenassist_si128(temp, 0), 2);
476 rk[13] = rk[ 5] ^ rk[12];
477 rk[14] = rk[ 6] ^ rk[13];
478 rk[15] = rk[ 7] ^ rk[14];
479 temp = _mm_insert_epi32(temp, rk[15], 3);
483 temp = _mm_insert_epi32(temp, rk[7], 3);
490 void Rijndael_UncheckedSetKeyRev_AESNI(word32 *key,
unsigned int rounds)
495 vec_swap(*M128_CAST(key), *M128_CAST(key+4*rounds));
497 for (i = 4, j = 4*rounds-4; i < j; i += 4, j -= 4)
499 temp = _mm_aesimc_si128(*M128_CAST(key+i));
500 *M128_CAST(key+i) = _mm_aesimc_si128(*M128_CAST(key+j));
501 *M128_CAST(key+j) = temp;
504 *M128_CAST(key+i) = _mm_aesimc_si128(*M128_CAST(key+i));
507 size_t Rijndael_Enc_AdvancedProcessBlocks_AESNI(
const word32 *subKeys,
size_t rounds,
508 const byte *inBlocks,
const byte *xorBlocks, byte *outBlocks,
size_t length, word32 flags)
511 MAYBE_CONST word32* sk = MAYBE_UNCONST_CAST(word32*, subKeys);
512 MAYBE_CONST byte* ib = MAYBE_UNCONST_CAST(byte*, inBlocks);
513 MAYBE_CONST byte* xb = MAYBE_UNCONST_CAST(byte*, xorBlocks);
515 return AdvancedProcessBlocks128_4x1_SSE(AESNI_Enc_Block, AESNI_Enc_4_Blocks,
516 sk, rounds, ib, xb, outBlocks, length, flags);
519 size_t Rijndael_Dec_AdvancedProcessBlocks_AESNI(
const word32 *subKeys,
size_t rounds,
520 const byte *inBlocks,
const byte *xorBlocks, byte *outBlocks,
size_t length, word32 flags)
522 MAYBE_CONST word32* sk = MAYBE_UNCONST_CAST(word32*, subKeys);
523 MAYBE_CONST byte* ib = MAYBE_UNCONST_CAST(byte*, inBlocks);
524 MAYBE_CONST byte* xb = MAYBE_UNCONST_CAST(byte*, xorBlocks);
526 return AdvancedProcessBlocks128_4x1_SSE(AESNI_Dec_Block, AESNI_Dec_4_Blocks,
527 sk, rounds, ib, xb, outBlocks, length, flags);
530 #endif // CRYPTOPP_AESNI_AVAILABLE 534 #if (CRYPTOPP_POWER8_AES_AVAILABLE) 536 ANONYMOUS_NAMESPACE_BEGIN
539 CRYPTOPP_ALIGN_DATA(16)
540 static const uint32_t s_rconBE[] = {
541 0x01000000, 0x02000000, 0x04000000, 0x08000000,
542 0x10000000, 0x20000000, 0x40000000, 0x80000000,
543 0x1B000000, 0x36000000
546 static inline void POWER8_Enc_Block(
uint32x4_p &block,
const word32 *subkeys,
unsigned int rounds)
549 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
554 for (
size_t i=1; i<rounds-1; i+=2)
566 uint32x4_p &block5,
const word32 *subkeys,
unsigned int rounds)
569 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
572 block0 =
VecXor(block0, k);
573 block1 =
VecXor(block1, k);
574 block2 =
VecXor(block2, k);
575 block3 =
VecXor(block3, k);
576 block4 =
VecXor(block4, k);
577 block5 =
VecXor(block5, k);
579 for (
size_t i=1; i<rounds; ++i)
599 static inline void POWER8_Dec_Block(
uint32x4_p &block,
const word32 *subkeys,
unsigned int rounds)
602 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
607 for (
size_t i=rounds-1; i>1; i-=2)
619 uint32x4_p &block5,
const word32 *subkeys,
unsigned int rounds)
622 const byte *keys =
reinterpret_cast<const byte*
>(subkeys);
625 block0 =
VecXor(block0, k);
626 block1 =
VecXor(block1, k);
627 block2 =
VecXor(block2, k);
628 block3 =
VecXor(block3, k);
629 block4 =
VecXor(block4, k);
630 block5 =
VecXor(block5, k);
632 for (
size_t i=rounds-1; i>0; --i)
652 ANONYMOUS_NAMESPACE_END
654 void Rijndael_UncheckedSetKey_POWER8(
const byte* userKey,
size_t keyLen, word32* rk,
const byte* Se)
656 const size_t rounds = keyLen / 4 + 6;
657 const word32 *rc = s_rconBE;
658 word32 *rkey = rk, temp;
663 const size_t keySize = 4*(rounds+1);
664 const word32* end = rkey + keySize;
668 temp = rkey[keyLen/4-1];
669 word32 x = (word32(Se[GETBYTE(temp, 2)]) << 24) ^ (word32(Se[GETBYTE(temp, 1)]) << 16) ^
670 (word32(Se[GETBYTE(temp, 0)]) << 8) ^ Se[GETBYTE(temp, 3)];
671 rkey[keyLen/4] = rkey[0] ^ x ^ *(rc++);
672 rkey[keyLen/4+1] = rkey[1] ^ rkey[keyLen/4];
673 rkey[keyLen/4+2] = rkey[2] ^ rkey[keyLen/4+1];
674 rkey[keyLen/4+3] = rkey[3] ^ rkey[keyLen/4+2];
676 if (rkey + keyLen/4 + 4 == end)
681 rkey[10] = rkey[ 4] ^ rkey[ 9];
682 rkey[11] = rkey[ 5] ^ rkey[10];
684 else if (keyLen == 32)
687 rkey[12] = rkey[ 4] ^ (word32(Se[GETBYTE(temp, 3)]) << 24) ^ (word32(Se[GETBYTE(temp, 2)]) << 16) ^ (word32(Se[GETBYTE(temp, 1)]) << 8) ^ Se[GETBYTE(temp, 0)];
688 rkey[13] = rkey[ 5] ^ rkey[12];
689 rkey[14] = rkey[ 6] ^ rkey[13];
690 rkey[15] = rkey[ 7] ^ rkey[14];
695 #if (CRYPTOPP_LITTLE_ENDIAN) 697 const uint8x16_p mask = ((
uint8x16_p){12,13,14,15, 8,9,10,11, 4,5,6,7, 0,1,2,3});
701 for (i=0; i<rounds; i+=2, rkey+=8)
703 const uint8x16_p d1 = vec_vsx_ld( 0, (uint8_t*)rkey);
704 const uint8x16_p d2 = vec_vsx_ld(16, (uint8_t*)rkey);
705 vec_vsx_st(
VecPermute(d1, zero, mask), 0, (uint8_t*)rkey);
706 vec_vsx_st(
VecPermute(d2, zero, mask), 16, (uint8_t*)rkey);
709 for ( ; i<rounds+1; i++, rkey+=4)
711 const uint8x16_p d = vec_vsx_ld( 0, (uint8_t*)rkey);
712 vec_vsx_st(
VecPermute(d, zero, mask), 0, (uint8_t*)rkey);
717 size_t Rijndael_Enc_AdvancedProcessBlocks128_6x1_ALTIVEC(
const word32 *subKeys,
size_t rounds,
718 const byte *inBlocks,
const byte *xorBlocks, byte *outBlocks,
size_t length, word32 flags)
720 return AdvancedProcessBlocks128_6x1_ALTIVEC(POWER8_Enc_Block, POWER8_Enc_6_Blocks,
721 subKeys, rounds, inBlocks, xorBlocks, outBlocks, length, flags);
724 size_t Rijndael_Dec_AdvancedProcessBlocks128_6x1_ALTIVEC(
const word32 *subKeys,
size_t rounds,
725 const byte *inBlocks,
const byte *xorBlocks, byte *outBlocks,
size_t length, word32 flags)
727 return AdvancedProcessBlocks128_6x1_ALTIVEC(POWER8_Dec_Block, POWER8_Dec_6_Blocks,
728 subKeys, rounds, inBlocks, xorBlocks, outBlocks, length, flags);
731 #endif // CRYPTOPP_POWER8_AES_AVAILABLE T1 VecDecryptLast(const T1 state, const T2 key)
Final round of AES decryption.
Utility functions for the Crypto++ library.
Library configuration file.
bool IsAlignedOn(const void *ptr, unsigned int alignment)
Determines whether ptr is aligned to a minimum value.
T1 VecPermute(const T1 vec, const T2 mask)
Permutes a vector.
__vector unsigned int uint32x4_p
Vector of 32-bit elements.
Support functions for PowerPC and vector operations.
Template for AdvancedProcessBlocks and SIMD processing.
#define CRYPTOPP_ASSERT(exp)
Debugging and diagnostic assertion.
T1 VecXor(const T1 vec1, const T2 vec2)
XOR two vectors.
T1 VecEncryptLast(const T1 state, const T2 key)
Final round of AES encryption.
Crypto++ library namespace.
T1 VecDecrypt(const T1 state, const T2 key)
One round of AES decryption.
uint32x4_p VecLoad(const byte src[16])
Loads a vector from a byte array.
__vector unsigned char uint8x16_p
Vector of 8-bit elements.
T1 VecEncrypt(const T1 state, const T2 key)
One round of AES encryption.
void vec_swap(T &a, T &b)
Swaps two variables which are arrays.