X86: Implement PANDN, ANDNPS, and ANDNPD.

src/arch/x86/isa/decoder/two_byte_opcodes.isa

@@ -469,7 +469,7 @@
                     0x2: rqsrtps_Vo_Wo();
                     0x3: rcpps_Vo_Wo();
                     0x4: Inst::ANDPS(Vo,Wo);
-                    0x5: andnps_Vo_Wo();
+                    0x5: Inst::ANDNPS(Vo,Wo);
                     0x6: Inst::ORPS(Vo,Wo);
                     0x7: Inst::XORPS(Vo,Wo);
                 }
@@ -485,7 +485,7 @@
                     0x0: movmskpd_Gd_VRo();
                     0x1: sqrtpd_Vo_Wo();
                     0x4: Inst::ANDPD(Vo,Wo);
-                    0x5: andnpd_Vo_Wo();
+                    0x5: Inst::ANDNPD(Vo,Wo);
                     0x6: Inst::ORPD(Vo,Wo);
                     //This really should be type o, but it works on q sized
                     //chunks at a time.
@@ -990,7 +990,7 @@
                     0x4: paddusb_Pq_Qq();
                     0x5: paddusw_Pq_Qq();
                     0x6: pmaxub_Pq_Qq();
-                    0x7: pandn_Pq_Qq();
+                    0x7: Inst::PANDN(Pq,Qq);
                 }
                 // operand size (0x66)
                 0x1: decode OPCODE_OP_BOTTOM3 {
@@ -1001,7 +1001,7 @@
                     0x4: paddusb_Vo_Wo();
                     0x5: paddusw_Vo_Wo();
                     0x6: pmaxub_Vo_Wo();
-                    0x7: pandn_Vo_Wo();
+                    0x7: Inst::PANDN(Vo,Wo);
                 }
                 default: Inst::UD2();
             }

src/arch/x86/isa/insts/simd128/floating_point/logical/andp.py

@@ -97,6 +97,48 @@ def macroop ANDPD_XMM_P {
     mand xmml, xmml, ufp1
     mand xmmh, xmmh, ufp2
 };
+
+def macroop ANDNPS_XMM_XMM {
+    mandn xmml, xmml, xmmlm
+    mandn xmmh, xmmh, xmmhm
+};
+
+def macroop ANDNPS_XMM_M {
+    lea t1, seg, sib, disp, dataSize=asz
+    ldfp ufp1, seg, [1, t0, t1], dataSize=8
+    ldfp ufp2, seg, [1, t0, t1], 8, dataSize=8
+    mandn xmml, xmml, ufp1
+    mandn xmmh, xmmh, ufp2
+};
+
+def macroop ANDNPS_XMM_P {
+    rdip t7
+    lea t1, seg, riprel, disp, dataSize=asz
+    ldfp ufp1, seg, [1, t0, t1], dataSize=8
+    ldfp ufp2, seg, [1, t0, t1], 8, dataSize=8
+    mandn xmml, xmml, ufp1
+    mandn xmmh, xmmh, ufp2
+};
+
+def macroop ANDNPD_XMM_XMM {
+    mandn xmml, xmml, xmmlm
+    mandn xmmh, xmmh, xmmhm
+};
+
+def macroop ANDNPD_XMM_M {
+    lea t1, seg, sib, disp, dataSize=asz
+    ldfp ufp1, seg, [1, t0, t1], dataSize=8
+    ldfp ufp2, seg, [1, t0, t1], 8, dataSize=8
+    mandn xmml, xmml, ufp1
+    mandn xmmh, xmmh, ufp2
+};
+
+def macroop ANDNPD_XMM_P {
+    rdip t7
+    lea t1, seg, riprel, disp, dataSize=asz
+    ldfp ufp1, seg, [1, t0, t1], dataSize=8
+    ldfp ufp2, seg, [1, t0, t1], 8, dataSize=8
+    mandn xmml, xmml, ufp1
+    mandn xmmh, xmmh, ufp2
+};
 '''
-# ANDNPS
-# ANDNPD

src/arch/x86/isa/insts/simd128/integer/logical/pand.py

@@ -75,5 +75,26 @@ def macroop PAND_XMM_P {
     mand xmml, xmml, ufp1
     mand xmmh, xmmh, ufp2
 };
+
+def macroop PANDN_XMM_XMM {
+    mandn xmml, xmml, xmmlm
+    mandn xmmh, xmmh, xmmhm
+};
+
+def macroop PANDN_XMM_M {
+    lea t1, seg, sib, disp, dataSize=asz
+    ldfp ufp1, seg, [1, t0, t1], dataSize=8
+    ldfp ufp2, seg, [1, t0, t1], 8, dataSize=8
+    mandn xmml, xmml, ufp1
+    mandn xmmh, xmmh, ufp2
+};
+
+def macroop PANDN_XMM_P {
+    rdip t7
+    lea t1, seg, riprel, disp, dataSize=asz
+    ldfp ufp1, seg, [1, t0, t1], dataSize=8
+    ldfp ufp2, seg, [1, t0, t1], 8, dataSize=8
+    mandn xmml, xmml, ufp1
+    mandn xmmh, xmmh, ufp2
+};
 '''
-# PANDN

src/arch/x86/isa/insts/simd64/integer/logical/pand.py

@@ -68,5 +68,19 @@ def macroop PAND_MMX_P {
     ldfp ufp1, seg, riprel, disp, dataSize=8
     mand mmx, mmx, ufp1
 };
+
+def macroop PANDN_MMX_MMX {
+    mandn mmx, mmx, mmxm
+};
+
+def macroop PANDN_MMX_M {
+    ldfp ufp1, seg, sib, disp, dataSize=8
+    mandn mmx, mmx, ufp1
+};
+
+def macroop PANDN_MMX_P {
+    rdip t7
+    ldfp ufp1, seg, riprel, disp, dataSize=8
+    mandn mmx, mmx, ufp1
+};
 '''
-# PANDN