# Additional SPE Instructions for Devices that Support VLE
# Freescale engineering bulletin EB689
# https://www.nxp.com/docs/en/engineering-bulletin/EB689.pdf

# additional SPE instructions that are implemented on devices with an e200z3 or e200z6 core that supports VLE
# - Vector Floating-Point Single-Precision Multiply-Add
# - Vector Floating-Point Single-Precision Multiply-Substract
# - Vector Floating-Point Single-Precision Negative Multiply-Add
# - Vector Floating-Point Single-Precision Negative Multiply-Substract
# - Floating-Point Single-Precision Multiply-Add
# - Floating-Point Single-Precision Multiply-Substract
# - Floating-Point Single-Precision Negative Multiply-Add
# - Floating-Point Single-Precision Negative Multiply-Substract



# evfsmadd rD,rA,rB
# Vector Floating-Point Single-Precision Multiply-Add
:evfsmadd D,A,B is OP=4 & D & A & B & XOP_0_10=0x282
{
	local tmpAL:4 = A:4; local tmpAH:4 = A(4);
	local tmpBL:4 = B:4; local tmpBH:4 = B(4);
	local tmpDL:4 = D:4; local tmpDH:4 = D(4);

	tmpDL = ((tmpAL f* tmpBL) f+ tmpDL);
	tmpDH = ((tmpAH f* tmpBH) f+ tmpDH);
	D = (zext(tmpDH) << 32) | zext(tmpDL);
}

# evfsmsub rD,rA,rB
# Vector Floating-Point Single-Precision Multiply-Substract
:evfsmsub D,A,B is OP=4 & D & A & B & XOP_0_10=0x283
{
	local tmpAL:4 = A:4; local tmpAH:4 = A(4);
	local tmpBL:4 = B:4; local tmpBH:4 = B(4);
	local tmpDL:4 = D:4; local tmpDH:4 = D(4);

	tmpDL = ((tmpAL f* tmpBL) f- tmpDL);
	tmpDH = ((tmpAH f* tmpBH) f- tmpDH);
	D = (zext(tmpDH) << 32) | zext(tmpDL);
}

# evfsnmadd
# Vector Floating-Point Single-Precision Negative Multiply-Add
:evfsnmadd D,A,B is OP=4 & D & A & B & XOP_0_10=0x28A
{
	local tmpAL:4 = A:4; local tmpAH:4 = A(4);
	local tmpBL:4 = B:4; local tmpBH:4 = B(4);
	local tmpDL:4 = D:4; local tmpDH:4 = D(4);

	tmpDL = f- ((tmpAL f* tmpBL) f+ tmpDL);
	tmpDH = f- ((tmpAH f* tmpBH) f+ tmpDH);
	D = (zext(tmpDH) << 32) | zext(tmpDL);
}

# evfsnmsub
# Vector Floating-Point Single-Precision Negative Multiply-Substract
:evfsnmsub D,A,B is OP=4 & D & A & B & XOP_0_10=0x28B
{
	local tmpAL:4 = A:4; local tmpAH:4 = A(4);
	local tmpBL:4 = B:4; local tmpBH:4 = B(4);
	local tmpDL:4 = D:4; local tmpDH:4 = D(4);

	tmpDL = f- ((tmpAL f* tmpBL) f- tmpDL);
	tmpDH = f- ((tmpAH f* tmpBH) f- tmpDH);
	D = (zext(tmpDH) << 32) | zext(tmpDL);
}

# efsmadd rD,rA,rB
# Floating-Point Single-Precision Multiply-Add
:efsmadd D,A,B is OP=4 & D & A & B & XOP_0_10=0x2C2
{
	local lo:4 = (A:4 f* B:4) f+ D:4;
	D = (D & 0xFFFFFFFF00000000) | zext(lo);
}

# efsmsub rD,rA,rB
# Floating-Point Single-Precision Multiply-Substract
:efsmsub D,A,B is OP=4 & D & A & B & XOP_0_10=0x2C3
{
	local lo:4 = (A:4 f* B:4) f- D:4;
	D = (D & 0xFFFFFFFF00000000) | zext(lo);
}

# efsnmadd rD,rA,rB
# Floating-Point Single-Precision Negative Multiply-Add
:efsnmadd D,A,B is OP=4 & D & A & B & XOP_0_10=0x2CA
{
	local lo:4 = f- ((A:4 f* B:4) f+ D:4);
	D = (D & 0xFFFFFFFF00000000) | zext(lo);
}

# efsnmsub rD,rA,rB
# Floating-Point Single-Precision Negative Multiply-Substract
:efsnmsub D,A,B is OP=4 & D & A & B & XOP_0_10=0x2CB
{
	local lo:4 = f- ((A:4 f* B:4) f- D:4);
	D = (D & 0xFFFFFFFF00000000) | zext(lo);
}