/*************************************************************************** * 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 . * * 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 method for modular square root * mrsroot.c * * Siguna Mueller's O(lg(p)^3) algorithm, Designs Codes and Cryptography, 2004 * * This is a little slower for p=1 mod 4 primes, but its not time critical, and * more importantly it doesn't pull in the large powmod code into elliptic curve programs * It does require code from mrjack.c and mrlucas.c * * If p=3 mod 4, then sqrt(a)=a^[(p+1)/4] mod p. Note that for many elliptic curves * (p+1)/4 has very low hamming weight. * * (was sqrt(a) = V_{(p+1)/4}(a+1/a,1)/(1+1/a)) * * Mueller's method is also very simple, uses very little memory, and it works just fine for p=1 mod 8 primes * (for example the "annoying" NIST modulus 2^224-2^96+1) * Also doesn't waste time on non-squares, as a jacobi test is done first * * If you know that the prime is 3 mod 4, and you know that x is almost certainly a QR * then the jacobi-dependent code can be deleted with some space savings. * * NOTE - IF p IS NOT PRIME, THIS CODE WILL FAIL SILENTLY! * */ #include #include "miracl.h" BOOL nres_sqroot(_MIPD_ big x,big w) { /* w=sqrt(x) mod p. This depends on p being prime! */ int t,js; #ifdef MR_OS_THREADS miracl *mr_mip=get_mip(); #endif if (mr_mip->ERNUM) return FALSE; copy(x,w); if (size(w)==0) return TRUE; MR_IN(100) redc(_MIPP_ w,w); /* get it back into normal form */ if (size(w)==1) /* square root of 1 is 1 */ { nres(_MIPP_ w,w); MR_OUT return TRUE; } if (size(w)==4) /* square root of 4 is 2 */ { convert(_MIPP_ 2,w); nres(_MIPP_ w,w); MR_OUT return TRUE; } if (jack(_MIPP_ w,mr_mip->modulus)!=1) { /* Jacobi test */ zero(w); MR_OUT return FALSE; } js=mr_mip->pmod8%4-2; /* 1 mod 4 or 3 mod 4 prime? */ incr(_MIPP_ mr_mip->modulus,js,mr_mip->w10); subdiv(_MIPP_ mr_mip->w10,4,mr_mip->w10); /* (p+/-1)/4 */ if (js==1) { /* 3 mod 4 primes - do a quick and dirty sqrt(x)=x^(p+1)/4 mod p */ nres(_MIPP_ w,mr_mip->w2); copy(mr_mip->one,w); forever { /* Simple Right-to-Left exponentiation */ if (mr_mip->user!=NULL) (*mr_mip->user)(); if (subdiv(_MIPP_ mr_mip->w10,2,mr_mip->w10)!=0) nres_modmult(_MIPP_ w,mr_mip->w2,w); if (mr_mip->ERNUM || size(mr_mip->w10)==0) break; nres_modmult(_MIPP_ mr_mip->w2,mr_mip->w2,mr_mip->w2); } /* nres_moddiv(_MIPP_ mr_mip->one,w,mr_mip->w11); nres_modadd(_MIPP_ mr_mip->w11,w,mr_mip->w3); nres_lucas(_MIPP_ mr_mip->w3,mr_mip->w10,w,w); nres_modadd(_MIPP_ mr_mip->w11,mr_mip->one,mr_mip->w11); nres_moddiv(_MIPP_ w,mr_mip->w11,w); */ } else { /* 1 mod 4 primes */ for (t=1; ;t++) { /* t=1.5 on average */ if (t==1) copy(w,mr_mip->w4); else { premult(_MIPP_ w,t,mr_mip->w4); divide(_MIPP_ mr_mip->w4,mr_mip->modulus,mr_mip->modulus); premult(_MIPP_ mr_mip->w4,t,mr_mip->w4); divide(_MIPP_ mr_mip->w4,mr_mip->modulus,mr_mip->modulus); } decr(_MIPP_ mr_mip->w4,4,mr_mip->w1); if (jack(_MIPP_ mr_mip->w1,mr_mip->modulus)==js) break; if (mr_mip->ERNUM) break; } decr(_MIPP_ mr_mip->w4,2,mr_mip->w3); nres(_MIPP_ mr_mip->w3,mr_mip->w3); nres_lucas(_MIPP_ mr_mip->w3,mr_mip->w10,w,w); /* heavy lifting done here */ if (t!=1) { convert(_MIPP_ t,mr_mip->w11); nres(_MIPP_ mr_mip->w11,mr_mip->w11); nres_moddiv(_MIPP_ w,mr_mip->w11,w); } } MR_OUT return TRUE; } BOOL sqroot(_MIPD_ big x,big p,big w) { /* w = sqrt(x) mod p */ #ifdef MR_OS_THREADS miracl *mr_mip=get_mip(); #endif if (mr_mip->ERNUM) return FALSE; MR_IN(101) if (subdivisible(_MIPP_ p,2)) { /* p must be odd */ zero(w); MR_OUT return FALSE; } prepare_monty(_MIPP_ p); nres(_MIPP_ x,w); if (nres_sqroot(_MIPP_ w,w)) { redc(_MIPP_ w,w); MR_OUT return TRUE; } zero(w); MR_OUT return FALSE; }