150 lines
5.1 KiB
C
150 lines
5.1 KiB
C
|
|
/***************************************************************************
|
|
*
|
|
Copyright 2013 CertiVox UK Ltd. *
|
|
*
|
|
This file is part of CertiVox MIRACL Crypto SDK. *
|
|
*
|
|
The CertiVox MIRACL Crypto SDK provides developers with an *
|
|
extensive and efficient set of cryptographic functions. *
|
|
For further information about its features and functionalities please *
|
|
refer to http://www.certivox.com *
|
|
*
|
|
* The CertiVox MIRACL Crypto SDK is free software: you can *
|
|
redistribute it and/or modify it under the terms of the *
|
|
GNU Affero General Public License as published by the *
|
|
Free Software Foundation, either version 3 of the License, *
|
|
or (at your option) any later version. *
|
|
*
|
|
* The CertiVox MIRACL Crypto SDK is distributed in the hope *
|
|
that it will be useful, but WITHOUT ANY WARRANTY; without even the *
|
|
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. *
|
|
See the GNU Affero General Public License for more details. *
|
|
*
|
|
* You should have received a copy of the GNU Affero General Public *
|
|
License along with CertiVox MIRACL Crypto SDK. *
|
|
If not, see <http://www.gnu.org/licenses/>. *
|
|
*
|
|
You can be released from the requirements of the license by purchasing *
|
|
a commercial license. Buying such a license is mandatory as soon as you *
|
|
develop commercial activities involving the CertiVox MIRACL Crypto SDK *
|
|
without disclosing the source code of your own applications, or shipping *
|
|
the CertiVox MIRACL Crypto SDK with a closed source product. *
|
|
*
|
|
***************************************************************************/
|
|
/*
|
|
* MIRACL Chinese Remainder Thereom routines (for use with big moduli)
|
|
* mrcrt.c
|
|
*/
|
|
|
|
#include <stdlib.h>
|
|
#include "miracl.h"
|
|
|
|
#ifndef MR_STATIC
|
|
|
|
BOOL crt_init(_MIPD_ big_chinese *c,int r,big *moduli)
|
|
{ /* calculate CRT constants */
|
|
int i,j,k;
|
|
#ifdef MR_OS_THREADS
|
|
miracl *mr_mip=get_mip();
|
|
#endif
|
|
if (r<2 || mr_mip->ERNUM) return FALSE;
|
|
for (i=0;i<r;i++) if (size(moduli[i])<2) return FALSE;
|
|
|
|
MR_IN(73)
|
|
|
|
c->M=(big *)mr_alloc(_MIPP_ r,sizeof(big));
|
|
if (c->M==NULL)
|
|
{
|
|
mr_berror(_MIPP_ MR_ERR_OUT_OF_MEMORY);
|
|
MR_OUT
|
|
return FALSE;
|
|
}
|
|
c->C=(big *)mr_alloc(_MIPP_ r*(r-1)/2,sizeof(big));
|
|
if (c->C==NULL)
|
|
{
|
|
mr_free(c->M);
|
|
mr_berror(_MIPP_ MR_ERR_OUT_OF_MEMORY);
|
|
MR_OUT
|
|
return FALSE;
|
|
}
|
|
c->V=(big *)mr_alloc(_MIPP_ r,sizeof(big));
|
|
if (c->V==NULL)
|
|
{
|
|
mr_free(c->M);
|
|
mr_free(c->C);
|
|
mr_berror(_MIPP_ MR_ERR_OUT_OF_MEMORY);
|
|
MR_OUT
|
|
return FALSE;
|
|
}
|
|
for (k=0,i=0;i<r;i++)
|
|
{
|
|
c->V[i]=mirvar(_MIPP_ 0);
|
|
c->M[i]=mirvar(_MIPP_ 0);
|
|
copy(moduli[i],c->M[i]);
|
|
for (j=0;j<i;j++,k++)
|
|
{
|
|
c->C[k]=mirvar(_MIPP_ 0);
|
|
invmodp(_MIPP_ c->M[j],c->M[i],c->C[k]);
|
|
}
|
|
}
|
|
c->NP=r;
|
|
MR_OUT
|
|
return TRUE;
|
|
}
|
|
|
|
void crt_end(big_chinese *c)
|
|
{ /* clean up after CRT */
|
|
int i,j,k;
|
|
if (c->NP<2) return;
|
|
for (k=0,i=0;i<c->NP;i++)
|
|
{
|
|
mirkill(c->M[i]);
|
|
for (j=0;j<i;j++)
|
|
mirkill(c->C[k++]);
|
|
mirkill(c->V[i]);
|
|
}
|
|
mr_free(c->M);
|
|
mr_free(c->V);
|
|
mr_free(c->C);
|
|
c->NP=0;
|
|
}
|
|
|
|
#endif
|
|
|
|
void crt(_MIPD_ big_chinese *c,big *u,big x)
|
|
{ /* Chinese Remainder Thereom *
|
|
* Calculate x given remainders u[i] mod M[i] */
|
|
int i,j,k;
|
|
#ifdef MR_OS_THREADS
|
|
miracl *mr_mip=get_mip();
|
|
#endif
|
|
if (c->NP<2 || mr_mip->ERNUM) return;
|
|
|
|
MR_IN(74)
|
|
|
|
copy(u[0],c->V[0]);
|
|
for (k=0,i=1;i<c->NP;i++)
|
|
{ /* Knuth page 274 */
|
|
subtract(_MIPP_ u[i],c->V[0],c->V[i]);
|
|
mad(_MIPP_ c->V[i],c->C[k],c->C[k],c->M[i],c->M[i],c->V[i]);
|
|
k++;
|
|
for (j=1;j<i;j++,k++)
|
|
{
|
|
subtract(_MIPP_ c->V[i],c->V[j],c->V[i]);
|
|
mad(_MIPP_ c->V[i],c->C[k],c->C[k],c->M[i],c->M[i],c->V[i]);
|
|
}
|
|
if (size(c->V[i])<0) add(_MIPP_ c->V[i],c->M[i],c->V[i]);
|
|
}
|
|
zero(x);
|
|
convert(_MIPP_ 1,mr_mip->w1);
|
|
for (i=0;i<c->NP;i++)
|
|
{
|
|
multiply(_MIPP_ mr_mip->w1,c->V[i],mr_mip->w2);
|
|
add(_MIPP_ x,mr_mip->w2,x);
|
|
multiply(_MIPP_ mr_mip->w1,c->M[i],mr_mip->w1);
|
|
}
|
|
MR_OUT
|
|
}
|
|
|