#include <openssl/base.h>
#if !defined(OPENSSL_NO_ASM) && defined(OPENSSL_X86_64)
#include "rsaz_exp.h"
#include <openssl/mem.h>
#include "../internal.h"
void rsaz_1024_norm2red_avx2(void *red,const void *norm);
void rsaz_1024_mul_avx2(void *ret,const void *a,const void *b,const void *n,BN_ULONG k);
void rsaz_1024_sqr_avx2(void *ret,const void *a,const void *n,BN_ULONG k,int cnt);
void rsaz_1024_scatter5_avx2(void *tbl,const void *val,int i);
void rsaz_1024_gather5_avx2(void *val,const void *tbl,int i);
void rsaz_1024_red2norm_avx2(void *norm,const void *red);
alignas(64) static const BN_ULONG one[40] =
{1,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
alignas(64) static const BN_ULONG two80[40] =
{0,0,1<<22,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
void RSAZ_1024_mod_exp_avx2(BN_ULONG result_norm[16],
const BN_ULONG base_norm[16], const BN_ULONG exponent[16],
const BN_ULONG m_norm[16], const BN_ULONG RR[16], BN_ULONG k0)
{
alignas(64) uint8_t storage[(320 * 3) + (32 * 9 * 16)];
unsigned char *a_inv, *m, *result,
*table_s = storage + (320 * 3),
*R2 = table_s;
int index;
int wvalue;
if (((((uintptr_t)storage & 4095) + 320) >> 12) != 0) {
result = storage;
a_inv = storage + 320;
m = storage + (320 * 2);
} else {
m = storage;
result = storage + 320;
a_inv = storage + (320 * 2);
}
rsaz_1024_norm2red_avx2(m, m_norm);
rsaz_1024_norm2red_avx2(a_inv, base_norm);
rsaz_1024_norm2red_avx2(R2, RR);
rsaz_1024_mul_avx2(R2, R2, R2, m, k0);
rsaz_1024_mul_avx2(R2, R2, two80, m, k0);
rsaz_1024_mul_avx2(result, R2, one, m, k0);
rsaz_1024_mul_avx2(a_inv, a_inv, R2, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,0);
rsaz_1024_scatter5_avx2(table_s,a_inv,1);
rsaz_1024_sqr_avx2(result, a_inv, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,2);
#if 0#else
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,4);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,8);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,16);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,17);
rsaz_1024_gather5_avx2(result,table_s,2);
rsaz_1024_mul_avx2(result,result,a_inv,m,k0);
rsaz_1024_scatter5_avx2(table_s,result,3);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,6);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,12);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,24);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,25);
rsaz_1024_gather5_avx2(result,table_s,4);
rsaz_1024_mul_avx2(result,result,a_inv,m,k0);
rsaz_1024_scatter5_avx2(table_s,result,5);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,10);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,20);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,21);
rsaz_1024_gather5_avx2(result,table_s,6);
rsaz_1024_mul_avx2(result,result,a_inv,m,k0);
rsaz_1024_scatter5_avx2(table_s,result,7);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,14);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,28);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,29);
rsaz_1024_gather5_avx2(result,table_s,8);
rsaz_1024_mul_avx2(result,result,a_inv,m,k0);
rsaz_1024_scatter5_avx2(table_s,result,9);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,18);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,19);
rsaz_1024_gather5_avx2(result,table_s,10);
rsaz_1024_mul_avx2(result,result,a_inv,m,k0);
rsaz_1024_scatter5_avx2(table_s,result,11);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,22);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,23);
rsaz_1024_gather5_avx2(result,table_s,12);
rsaz_1024_mul_avx2(result,result,a_inv,m,k0);
rsaz_1024_scatter5_avx2(table_s,result,13);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,26);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,27);
rsaz_1024_gather5_avx2(result,table_s,14);
rsaz_1024_mul_avx2(result,result,a_inv,m,k0);
rsaz_1024_scatter5_avx2(table_s,result,15);
rsaz_1024_sqr_avx2(result, result, m, k0, 1);
rsaz_1024_scatter5_avx2(table_s,result,30);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_scatter5_avx2(table_s,result,31);
#endif
const uint8_t *p_str = (const uint8_t *)exponent;
wvalue = p_str[127] >> 3;
rsaz_1024_gather5_avx2(result,table_s,wvalue);
index = 1014;
while(index > -1) {
rsaz_1024_sqr_avx2(result, result, m, k0, 5);
wvalue = *((const unsigned short*)&p_str[index / 8]);
wvalue = (wvalue>> (index%8)) & 31;
index-=5;
rsaz_1024_gather5_avx2(a_inv,table_s,wvalue);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
}
rsaz_1024_sqr_avx2(result, result, m, k0, 4);
wvalue = p_str[0] & 15;
rsaz_1024_gather5_avx2(a_inv,table_s,wvalue);
rsaz_1024_mul_avx2(result, result, a_inv, m, k0);
rsaz_1024_mul_avx2(result, result, one, m, k0);
rsaz_1024_red2norm_avx2(result_norm, result);
}
#endif