add sqrt but not enable

include/mcl/fp.hpp

@@ -75,6 +75,7 @@ public:
 		assert(sizeof(mp_limb_t) == sizeof(Unit));
 		// set default wrapper function
 		op_.neg = negW;
+		op_.sqr = sqrW;
 		op_.add = addW;
 		op_.sub = subW;
 		op_.mul = mulW;
@@ -277,7 +278,7 @@ public:
 	static inline void mul(FpT& z, const FpT& x, const FpT& y) { op_.mul(z.v_, x.v_, y.v_); }
 	static inline void inv(FpT& y, const FpT& x) { op_.invOp(y.v_, x.v_, op_); }
 	static inline void neg(FpT& y, const FpT& x) { op_.neg(y.v_, x.v_); }
-	static inline void square(FpT& y, const FpT& x) { mul(y, x, x); }
+	static inline void square(FpT& y, const FpT& x) { op_.mul(y.v_, x.v_, x.v_); }
 	static inline void div(FpT& z, const FpT& x, const FpT& y)
 	{
 		FpT rev;
@@ -430,6 +431,13 @@ public:
 		op_.mulPreP(xy, x, y);
 		op_.modP(z, xy, op_.p);
 	}
+	static inline void sqrW(Unit *y, const Unit *x)
+	{
+//		Unit xx[maxSize * 2];
+//		op_.sqrPreP(xx, x);
+//		op_.modP(y, xx, op_.p);
+		mulW(y, x, x);
+	}
 	static inline void negW(Unit *y, const Unit *x)
 	{
 		op_.negP(y, x, op_.p);

include/mcl/op.hpp

@@ -75,6 +75,7 @@ struct Op {
 	void2u copy;
 	// not require p(function having p)
 	void2u neg;
+	void2u sqr;
 	void3u add;
 	void3u sub;
 	void3u mul;
@@ -86,6 +87,7 @@ struct Op {
 	void (*mont)(Unit *z, const Unit *x, const Unit *y, const Unit *p, Unit rp);
 	// require p
 	void3u negP;
+	void2u sqrPreP;
 	void2uOp invOp;
 	void4u addP;
 	void4u subP;
@@ -95,10 +97,10 @@ struct Op {
 	Op()
 		: N(0), bitSize(0)
 		, isZero(0), clear(0), copy(0)
-		, neg(0), add(0), sub(0), mul(0)
+		, neg(0), sqr(0), add(0), sub(0), mul(0)
 		, useMont(false), preInv(0)
 		, rp(0), mont(0)
-		, negP(0), invOp(0), addP(0), subP(0), mulPreP(0), modP(0)
+		, negP(0), sqrPreP(0), invOp(0), addP(0), subP(0), mulPreP(0), modP(0)
 		, fg(createFpGenerator())
 	{
 	}

src/fp.cpp

@@ -131,6 +131,15 @@ struct OpeFunc {
 		mpz_mul(mz, mx, my);
 		Gmp::getArray(z, N * 2, mz);
 	}
+	static inline void sqrPreC(Unit *y, const Unit *x)
+	{
+//		mpz_t mx, my;
+//		set_zero(my, y, N * 2);
+//		set_mpz_t(mx, x);
+//		mpz_mul(my, mx, mx);
+//		Gmp::getArray(y, N * 2, my);
+		mulPreC(y, x, x);
+	}
 	// x[N * 2] -> y[N]
 	static inline void modC(Unit *y, const Unit *x, const Unit *p)
 	{
@@ -172,6 +181,10 @@ struct OpeFunc {
 		}
 		subC(y, p, x, p);
 	}
+	static inline void sqrC(Unit *y, const Unit *x, const Unit *p)
+	{
+		sqrC(y, p, x, p);
+	}
 };
 
 #ifdef MCL_USE_LLVM
@@ -200,8 +213,9 @@ struct OpeFunc {
 		addP = OpeFunc<n>::addC; \
 		subP = OpeFunc<n>::subC; \
 		mulPreP = OpeFunc<n>::mulPreC; \
+		sqrPreP = OpeFunc<n>::sqrPreC; \
 		modP = OpeFunc<n>::modC; \
-		SET_OP_LLVM(n) 
+		SET_OP_LLVM(n)
 
 #ifdef MCL_USE_XBYAK
 inline void invOpForMont(Unit *y, const Unit *x, const Op& op)
@@ -256,6 +270,9 @@ static void initForMont(Op& op, const Unit *p, Mode mode)
 	op.add = Xbyak::CastTo<void3u>(fg->add_);
 	op.sub = Xbyak::CastTo<void3u>(fg->sub_);
 	op.mul = Xbyak::CastTo<void3u>(fg->mul_);
+//	if (fg->sqr_) {
+//		op.sqr = Xbyak::CastTo<void2u>(fg->sqr_);
+//	}
 	op.preInv = Xbyak::CastTo<int2u>(op.fg->preInv_);
 	op.invOp = &invOpForMont;