# Due to limitations on variable length matching that preclude opcode matching afterwards, all memory addressing forms of PCLMULQDQ are decoded to PCLMULQDQ, not the macro names.
# Display is non-standard, but semantics, and de-compilation should be correct.

macro pclmul(src1, src2, dest) {
    local i:4 = 0:4;
    local temp:16 = 0;

    <start>
    if (i > 63:4) goto <end>;
        if ((src1 & (1 << i)) == 0) goto <skip>;   
            temp = temp ^ (src2 << i);
    <skip>
        i = i+1;
        goto <start>;
    <end>
    dest = temp;
}

:PCLMULLQLQDQ XmmReg1, XmmReg2  is vexMode=0 & $(PRE_66) & byte=0x0f; byte=0x3a; byte=0x44; xmmmod=3 & XmmReg1 & XmmReg2; byte=0x00
{
    local src1:16 = zext(XmmReg1[0,64]);
    local src2:16 = zext(XmmReg2[0,64]);
    pclmul(src1,src2,XmmReg1);
}

:PCLMULHQLQDQ XmmReg1, XmmReg2  is vexMode=0 & $(PRE_66) & byte=0x0f; byte=0x3a; byte=0x44; xmmmod=3 & XmmReg1 & XmmReg2; byte=0x01
{
    local src1:16 = zext(XmmReg1[64,64]);
    local src2:16 = zext(XmmReg2[0,64]);
    pclmul(src1,src2,XmmReg1);
}

:PCLMULLQHQDQ XmmReg1, XmmReg2  is vexMode=0 & $(PRE_66) & byte=0x0f; byte=0x3a; byte=0x44; xmmmod=3 & XmmReg1 & XmmReg2; byte=0x10
{
    local src1:16 = zext(XmmReg1[0,64]);
    local src2:16 = zext(XmmReg2[64,64]);
    pclmul(src1,src2,XmmReg1);
}

:PCLMULHQHQDQ XmmReg1, XmmReg2  is vexMode=0 & $(PRE_66) & byte=0x0f; byte=0x3a; byte=0x44; xmmmod=3 & XmmReg1 & XmmReg2; byte=0x11
{
    local src1:16 = zext(XmmReg1[64,64]);
    local src2:16 = zext(XmmReg2[64,64]);
    pclmul(src1,src2,XmmReg1);
}

:PCLMULQDQ XmmReg1, XmmReg2, imm8  is vexMode=0 & $(PRE_66) & byte=0x0f; byte=0x3a; byte=0x44; xmmmod=3 & XmmReg1 & XmmReg2; imm8 & imm8_4 & imm8_0
{
	if (imm8_0:1) goto <src1_b>;
		src1:16 = zext(XmmReg1[0,64]);
		goto <done1>;
		
	<src1_b>
		src1 = zext(XmmReg1[64,64]);
	
	<done1>
		
	if (imm8_4:1) goto <src2_b>;
		src2:16 = zext(XmmReg2[0,64]);
		goto <done2>;
		
	<src2_b>
		src2 = zext(XmmReg2[64,64]);
		
	<done2>
	
    pclmul(src1,src2,XmmReg1); 
}

:PCLMULQDQ XmmReg, m128, imm8  is vexMode=0 & $(PRE_66) & byte=0x0f; byte=0x3a; byte=0x44; XmmReg ... & m128; imm8 & imm8_4 & imm8_0
{
	if (imm8_0:1) goto <src1_b>;
		src1:16 = zext(XmmReg[0,64]);
		goto <done1>;
		
	<src1_b>
		src1 = zext(XmmReg[64,64]);
	
	<done1>
        local m:16 = m128;		
	if (imm8_4:1) goto <src2_b>;
		src2:16 = zext(m[0,64]);
		goto <done2>;
		
	<src2_b>
		src2 = zext(m[64,64]);
		
	<done2>
	
     pclmul(src1,src2,XmmReg);
}

:VPCLMULLQLQDQ XmmReg1, vexVVVV_XmmReg, XmmReg2 is $(VEX_NDS) & $(VEX_L128) & $(VEX_PRE_66) & $(VEX_0F3A) & $(VEX_WIG) & vexVVVV_XmmReg; byte=0x44; xmmmod=3 & (XmmReg1 & YmmReg1) & XmmReg2; byte=0x00
{
    local src1:16 = zext(vexVVVV_XmmReg[0,64]);
    local src2:16 = zext(XmmReg2[0,64]);
    pclmul(src1,src2,XmmReg1);
	YmmReg1 = zext(XmmReg1);
}

:VPCLMULHQLQDQ XmmReg1, vexVVVV_XmmReg, XmmReg2 is $(VEX_NDS) & $(VEX_L128) & $(VEX_PRE_66) & $(VEX_0F3A) & $(VEX_WIG) & vexVVVV_XmmReg; byte=0x44; xmmmod=3 & (XmmReg1 & YmmReg1) & XmmReg2; byte=0x01
{
    local src1:16 = zext(vexVVVV_XmmReg[64,64]);
    local src2:16 = zext(XmmReg2[0,64]);
    pclmul(src1,src2,XmmReg1);
	YmmReg1 = zext(XmmReg1);
}

:VPCLMULLQHQDQ XmmReg1, vexVVVV_XmmReg, XmmReg2 is $(VEX_NDS) & $(VEX_L128) & $(VEX_PRE_66) & $(VEX_0F3A) & $(VEX_WIG) & vexVVVV_XmmReg; byte=0x44; xmmmod=3 & (XmmReg1 & YmmReg1) & XmmReg2; byte=0x10
{
    local src1:16 = zext(vexVVVV_XmmReg[0,64]);
    local src2:16 = zext(XmmReg2[64,64]);
    pclmul(src1,src2,XmmReg1);
	YmmReg1 = zext(XmmReg1);
}

:VPCLMULHQHQDQ XmmReg1, vexVVVV_XmmReg, XmmReg2 is $(VEX_NDS) & $(VEX_L128) & $(VEX_PRE_66) & $(VEX_0F3A) & $(VEX_WIG) & vexVVVV_XmmReg; byte=0x44; xmmmod=3 & (XmmReg1 & YmmReg1) & XmmReg2; byte=0x11
{
    local src1:16 = zext(vexVVVV_XmmReg[64,64]);
    local src2:16 = zext(XmmReg2[64,64]);
    pclmul(src1,src2,XmmReg1);
	YmmReg1 = zext(XmmReg1);
}

:VPCLMULQDQ XmmReg1, vexVVVV_XmmReg, XmmReg2, imm8 is $(VEX_NDS) & $(VEX_L128) & $(VEX_PRE_66) & $(VEX_0F3A) & $(VEX_WIG) & vexVVVV_XmmReg; byte=0x44; xmmmod=3 & (XmmReg1 & YmmReg1) & XmmReg2; imm8 & imm8_4 & imm8_0
{
	if (imm8_0:1) goto <src1_b>;
		src1:16 = zext(vexVVVV_XmmReg[0,64]);
		goto <done1>;
		
	<src1_b>
		src1 = zext(vexVVVV_XmmReg[64,64]);
	
	<done1>
		
	if (imm8_4:1) goto <src2_b>;
		src2:16 = zext(XmmReg2[0,64]);
		goto <done2>;
		
	<src2_b>
		src2 = zext(XmmReg2[64,64]);
		
	<done2>
	
    pclmul(src1,src2,XmmReg1);
	YmmReg1 = zext(XmmReg1);
}

:VPCLMULQDQ XmmReg1, vexVVVV_XmmReg, m128, imm8 is $(VEX_NDS) & $(VEX_L128) & $(VEX_PRE_66) & $(VEX_0F3A) & $(VEX_WIG) & vexVVVV_XmmReg; byte=0x44; (XmmReg1 & YmmReg1) ... & m128; imm8 & imm8_4 & imm8_0
{
	if (imm8_0:1) goto <src1_b>;
		src1:16 = zext(vexVVVV_XmmReg[0,64]);
		goto <done1>;
		
	<src1_b>
		src1 = zext(vexVVVV_XmmReg[64,64]);
	
	<done1>

        local m:16 = m128;
	if (imm8_4:1) goto <src2_b>;
		src2:16 = zext(m[0,64]);
		goto <done2>;
		
	<src2_b>
		src2 = zext(m[64,64]);
		
	<done2>
	
    pclmul(src1,src2,XmmReg1);
	YmmReg1 = zext(XmmReg1);
}