native.intel.cbc.cbc128.c Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of bcprov-lts8on Show documentation
Show all versions of bcprov-lts8on Show documentation
The Long Term Stable (LTS) Bouncy Castle Crypto package is a Java implementation of cryptographic algorithms. This jar contains the JCA/JCE provider and low-level API for the BC LTS version 2.73.6 for Java 8 and later.
The newest version!
#include
#include
#include "cbc.h"
//
// AVX or 128b single block implementation.
//
static inline void aesdec_8_blocks_128b(__m128i *b1, __m128i *b2, __m128i *b3, __m128i *b4,
__m128i *b5, __m128i *b6, __m128i *b7, __m128i *b8,
const __m128i *round_keys, const int num_rounds,
const int num_blocks) {
const __m128i rk_ark = round_keys[0];
*b1 = _mm_xor_si128(*b1, rk_ark);
if (num_blocks > 1)
*b2 = _mm_xor_si128(*b2, rk_ark);
if (num_blocks > 2)
*b3 = _mm_xor_si128(*b3, rk_ark);
if (num_blocks > 3)
*b4 = _mm_xor_si128(*b4, rk_ark);
if (num_blocks > 4)
*b5 = _mm_xor_si128(*b5, rk_ark);
if (num_blocks > 5)
*b6 = _mm_xor_si128(*b6, rk_ark);
if (num_blocks > 6)
*b7 = _mm_xor_si128(*b7, rk_ark);
if (num_blocks > 7)
*b8 = _mm_xor_si128(*b8, rk_ark);
int round;
for (round = 1; round < num_rounds; round++) {
const __m128i rk = round_keys[round];
*b1 = _mm_aesdec_si128(*b1, rk);
if (num_blocks > 1)
*b2 = _mm_aesdec_si128(*b2, rk);
if (num_blocks > 2)
*b3 = _mm_aesdec_si128(*b3, rk);
if (num_blocks > 3)
*b4 = _mm_aesdec_si128(*b4, rk);
if (num_blocks > 4)
*b5 = _mm_aesdec_si128(*b5, rk);
if (num_blocks > 5)
*b6 = _mm_aesdec_si128(*b6, rk);
if (num_blocks > 6)
*b7 = _mm_aesdec_si128(*b7, rk);
if (num_blocks > 7)
*b8 = _mm_aesdec_si128(*b8, rk);
}
const __m128i rk_last = round_keys[round];
*b1 = _mm_aesdeclast_si128(*b1, rk_last);
if (num_blocks > 1)
*b2 = _mm_aesdeclast_si128(*b2, rk_last);
if (num_blocks > 2)
*b3 = _mm_aesdeclast_si128(*b3, rk_last);
if (num_blocks > 3)
*b4 = _mm_aesdeclast_si128(*b4, rk_last);
if (num_blocks > 4)
*b5 = _mm_aesdeclast_si128(*b5, rk_last);
if (num_blocks > 5)
*b6 = _mm_aesdeclast_si128(*b6, rk_last);
if (num_blocks > 6)
*b7 = _mm_aesdeclast_si128(*b7, rk_last);
if (num_blocks > 7)
*b8 = _mm_aesdeclast_si128(*b8, rk_last);
}
static inline void aes_cbc_dec_blocks_128b(unsigned char *in, unsigned char *out,
__m128i *chainblock,
const __m128i *roundKeys,
const int num_rounds, const uint32_t num_blocks) {
if (num_blocks == 8) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
__m128i d1 = _mm_loadu_si128((__m128i *) &in[1 * 16]);
__m128i d2 = _mm_loadu_si128((__m128i *) &in[2 * 16]);
__m128i d3 = _mm_loadu_si128((__m128i *) &in[3 * 16]);
__m128i d4 = _mm_loadu_si128((__m128i *) &in[4 * 16]);
__m128i d5 = _mm_loadu_si128((__m128i *) &in[5 * 16]);
__m128i d6 = _mm_loadu_si128((__m128i *) &in[6 * 16]);
__m128i d7 = _mm_loadu_si128((__m128i *) &in[7 * 16]);
const __m128i iv0 = *chainblock;
const __m128i iv1 = d0;
const __m128i iv2 = d1;
const __m128i iv3 = d2;
const __m128i iv4 = d3;
const __m128i iv5 = d4;
const __m128i iv6 = d5;
const __m128i iv7 = d6;
*chainblock = d7;
aesdec_8_blocks_128b(&d0, &d1, &d2, &d3, &d4, &d5, &d6, &d7, roundKeys, num_rounds, 8);
d0 = _mm_xor_si128(d0, iv0);
d1 = _mm_xor_si128(d1, iv1);
d2 = _mm_xor_si128(d2, iv2);
d3 = _mm_xor_si128(d3, iv3);
d4 = _mm_xor_si128(d4, iv4);
d5 = _mm_xor_si128(d5, iv5);
d6 = _mm_xor_si128(d6, iv6);
d7 = _mm_xor_si128(d7, iv7);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
_mm_storeu_si128((__m128i *) &out[1 * 16], d1);
_mm_storeu_si128((__m128i *) &out[2 * 16], d2);
_mm_storeu_si128((__m128i *) &out[3 * 16], d3);
_mm_storeu_si128((__m128i *) &out[4 * 16], d4);
_mm_storeu_si128((__m128i *) &out[5 * 16], d5);
_mm_storeu_si128((__m128i *) &out[6 * 16], d6);
_mm_storeu_si128((__m128i *) &out[7 * 16], d7);
} else if (num_blocks == 7) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
__m128i d1 = _mm_loadu_si128((__m128i *) &in[1 * 16]);
__m128i d2 = _mm_loadu_si128((__m128i *) &in[2 * 16]);
__m128i d3 = _mm_loadu_si128((__m128i *) &in[3 * 16]);
__m128i d4 = _mm_loadu_si128((__m128i *) &in[4 * 16]);
__m128i d5 = _mm_loadu_si128((__m128i *) &in[5 * 16]);
__m128i d6 = _mm_loadu_si128((__m128i *) &in[6 * 16]);
const __m128i iv0 = *chainblock;
const __m128i iv1 = d0;
const __m128i iv2 = d1;
const __m128i iv3 = d2;
const __m128i iv4 = d3;
const __m128i iv5 = d4;
const __m128i iv6 = d5;
*chainblock = d6;
aesdec_8_blocks_128b(&d0, &d1, &d2, &d3, &d4, &d5, &d6, &d6, roundKeys, num_rounds, 7);
d0 = _mm_xor_si128(d0, iv0);
d1 = _mm_xor_si128(d1, iv1);
d2 = _mm_xor_si128(d2, iv2);
d3 = _mm_xor_si128(d3, iv3);
d4 = _mm_xor_si128(d4, iv4);
d5 = _mm_xor_si128(d5, iv5);
d6 = _mm_xor_si128(d6, iv6);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
_mm_storeu_si128((__m128i *) &out[1 * 16], d1);
_mm_storeu_si128((__m128i *) &out[2 * 16], d2);
_mm_storeu_si128((__m128i *) &out[3 * 16], d3);
_mm_storeu_si128((__m128i *) &out[4 * 16], d4);
_mm_storeu_si128((__m128i *) &out[5 * 16], d5);
_mm_storeu_si128((__m128i *) &out[6 * 16], d6);
} else if (num_blocks == 6) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
__m128i d1 = _mm_loadu_si128((__m128i *) &in[1 * 16]);
__m128i d2 = _mm_loadu_si128((__m128i *) &in[2 * 16]);
__m128i d3 = _mm_loadu_si128((__m128i *) &in[3 * 16]);
__m128i d4 = _mm_loadu_si128((__m128i *) &in[4 * 16]);
__m128i d5 = _mm_loadu_si128((__m128i *) &in[5 * 16]);
const __m128i iv0 = *chainblock;
const __m128i iv1 = d0;
const __m128i iv2 = d1;
const __m128i iv3 = d2;
const __m128i iv4 = d3;
const __m128i iv5 = d4;
*chainblock = d5;
aesdec_8_blocks_128b(&d0, &d1, &d2, &d3, &d4, &d5, &d5, &d5, roundKeys, num_rounds, 6);
d0 = _mm_xor_si128(d0, iv0);
d1 = _mm_xor_si128(d1, iv1);
d2 = _mm_xor_si128(d2, iv2);
d3 = _mm_xor_si128(d3, iv3);
d4 = _mm_xor_si128(d4, iv4);
d5 = _mm_xor_si128(d5, iv5);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
_mm_storeu_si128((__m128i *) &out[1 * 16], d1);
_mm_storeu_si128((__m128i *) &out[2 * 16], d2);
_mm_storeu_si128((__m128i *) &out[3 * 16], d3);
_mm_storeu_si128((__m128i *) &out[4 * 16], d4);
_mm_storeu_si128((__m128i *) &out[5 * 16], d5);
} else if (num_blocks == 5) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
__m128i d1 = _mm_loadu_si128((__m128i *) &in[1 * 16]);
__m128i d2 = _mm_loadu_si128((__m128i *) &in[2 * 16]);
__m128i d3 = _mm_loadu_si128((__m128i *) &in[3 * 16]);
__m128i d4 = _mm_loadu_si128((__m128i *) &in[4 * 16]);
const __m128i iv0 = *chainblock;
const __m128i iv1 = d0;
const __m128i iv2 = d1;
const __m128i iv3 = d2;
const __m128i iv4 = d3;
*chainblock = d4;
aesdec_8_blocks_128b(&d0, &d1, &d2, &d3, &d4, &d4, &d4, &d4, roundKeys, num_rounds, 5);
d0 = _mm_xor_si128(d0, iv0);
d1 = _mm_xor_si128(d1, iv1);
d2 = _mm_xor_si128(d2, iv2);
d3 = _mm_xor_si128(d3, iv3);
d4 = _mm_xor_si128(d4, iv4);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
_mm_storeu_si128((__m128i *) &out[1 * 16], d1);
_mm_storeu_si128((__m128i *) &out[2 * 16], d2);
_mm_storeu_si128((__m128i *) &out[3 * 16], d3);
_mm_storeu_si128((__m128i *) &out[4 * 16], d4);
} else if (num_blocks == 4) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
__m128i d1 = _mm_loadu_si128((__m128i *) &in[1 * 16]);
__m128i d2 = _mm_loadu_si128((__m128i *) &in[2 * 16]);
__m128i d3 = _mm_loadu_si128((__m128i *) &in[3 * 16]);
const __m128i iv0 = *chainblock;
const __m128i iv1 = d0;
const __m128i iv2 = d1;
const __m128i iv3 = d2;
*chainblock = d3;
aesdec_8_blocks_128b(&d0, &d1, &d2, &d3, &d3, &d3, &d3, &d3, roundKeys, num_rounds, 4);
d0 = _mm_xor_si128(d0, iv0);
d1 = _mm_xor_si128(d1, iv1);
d2 = _mm_xor_si128(d2, iv2);
d3 = _mm_xor_si128(d3, iv3);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
_mm_storeu_si128((__m128i *) &out[1 * 16], d1);
_mm_storeu_si128((__m128i *) &out[2 * 16], d2);
_mm_storeu_si128((__m128i *) &out[3 * 16], d3);
} else if (num_blocks == 3) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
__m128i d1 = _mm_loadu_si128((__m128i *) &in[1 * 16]);
__m128i d2 = _mm_loadu_si128((__m128i *) &in[2 * 16]);
const __m128i iv0 = *chainblock;
const __m128i iv1 = d0;
const __m128i iv2 = d1;
*chainblock = d2;
aesdec_8_blocks_128b(&d0, &d1, &d2, &d2, &d2, &d2, &d2, &d2, roundKeys, num_rounds, 3);
d0 = _mm_xor_si128(d0, iv0);
d1 = _mm_xor_si128(d1, iv1);
d2 = _mm_xor_si128(d2, iv2);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
_mm_storeu_si128((__m128i *) &out[1 * 16], d1);
_mm_storeu_si128((__m128i *) &out[2 * 16], d2);
} else if (num_blocks == 2) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
__m128i d1 = _mm_loadu_si128((__m128i *) &in[1 * 16]);
const __m128i iv0 = *chainblock;
const __m128i iv1 = d0;
*chainblock = d1;
aesdec_8_blocks_128b(&d0, &d1, &d1, &d1, &d1, &d1, &d1, &d1, roundKeys, num_rounds, 2);
d0 = _mm_xor_si128(d0, iv0);
d1 = _mm_xor_si128(d1, iv1);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
_mm_storeu_si128((__m128i *) &out[1 * 16], d1);
} else if (num_blocks == 1) {
__m128i d0 = _mm_loadu_si128((__m128i *) &in[0 * 16]);
const __m128i iv0 = *chainblock;
*chainblock = d0;
aesdec_8_blocks_128b(&d0, &d0, &d0, &d0, &d0, &d0, &d0, &d0, roundKeys, num_rounds, 1);
d0 = _mm_xor_si128(d0, iv0);
_mm_storeu_si128((__m128i *) &out[0 * 16], d0);
}
}
size_t cbc_decrypt(cbc_ctx *cbc, unsigned char *src, uint32_t blocks, unsigned char *dest) {
assert(cbc != NULL);
unsigned char *destStart = dest;
while (blocks >= 8) {
aes_cbc_dec_blocks_128b(src, dest, &cbc->chainblock, cbc->roundKeys, cbc->num_rounds, 8);
blocks -= 8;
src += CBC_BLOCK_SIZE * 8;
dest += CBC_BLOCK_SIZE * 8;
}
aes_cbc_dec_blocks_128b(src, dest, &cbc->chainblock, cbc->roundKeys, cbc->num_rounds, blocks);
dest += blocks * CBC_BLOCK_SIZE;
return (size_t) (dest - destStart);
}