X86: Implement the floating point media multiply instructions.

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

@@ -505,7 +505,7 @@
                 // no prefix
                 0x0: decode OPCODE_OP_BOTTOM3 {
                     0x0: Inst::ADDPS(Vo,Wo);
-                    0x1: mulps_Vo_Wo();
+                    0x1: Inst::MULPS(Vo,Wo);
                     0x2: cvtps2pd_Vo_Wq();
                     0x3: cvtdq2ps_Vo_Wo();
                     0x4: Inst::SUBPS(Vo,Wo);
@@ -516,7 +516,7 @@
                 // repe (0xF3)
                 0x4: decode OPCODE_OP_BOTTOM3 {
                     0x0: Inst::ADDSS(Vd,Wd);
-                    0x1: mulss_Vd_Wd();
+                    0x1: Inst::MULSS(Vd,Wd);
                     0x2: cvtss2sd_Vq_Wd();
                     0x3: cvttps2dq_Vo_Wo();
                     0x4: Inst::SUBSS(Vd,Wd);
@@ -527,7 +527,7 @@
                 // operand size (0x66)
                 0x1: decode OPCODE_OP_BOTTOM3 {
                     0x0: Inst::ADDPD(Vo,Wo);
-                    0x1: mulpd_Vo_Wo();
+                    0x1: Inst::MULPD(Vo,Wo);
                     0x2: cvtpd2ps_Vo_Wo();
                     0x3: cvtps2dq_Vo_Wo();
                     0x4: Inst::SUBPD(Vo,Wo);

src/arch/x86/isa/insts/simd128/floating_point/arithmetic/multiplication.py

@@ -54,22 +54,73 @@
 # Authors: Gabe Black
 
 microcode = '''
-# MULPS
-# MULPD
-# MULSS
+def macroop MULSS_XMM_XMM {
+    mmulf xmml, xmml, xmmlm, size=4, ext=1
+};
+
+def macroop MULSS_XMM_M {
+    ldfp ufp1, seg, sib, disp, dataSize=8
+    mmulf xmml, xmml, ufp1, size=4, ext=1
+};
+
+def macroop MULSS_XMM_P {
+    rdip t7
+    ldfp ufp1, seg, riprel, disp, dataSize=8
+    mmulf xmml, xmml, ufp1, size=4, ext=1
+};
 
 def macroop MULSD_XMM_XMM {
-    mulfp xmml, xmml, xmmlm
+    mmulf xmml, xmml, xmmlm, size=8, ext=1
 };
 
 def macroop MULSD_XMM_M {
     ldfp ufp1, seg, sib, disp, dataSize=8
-    mulfp xmml, xmml, ufp1
+    mmulf xmml, xmml, ufp1, size=8, ext=1
 };
 
 def macroop MULSD_XMM_P {
     rdip t7
     ldfp ufp1, seg, riprel, disp, dataSize=8
-    mulfp xmml, xmml, ufp1
+    mmulf xmml, xmml, ufp1, size=8, ext=1
+};
+
+def macroop MULPS_XMM_XMM {
+    mmulf xmml, xmml, xmmlm, size=4, ext=0
+    mmulf xmmh, xmmh, xmmhm, size=4, ext=0
+};
+
+def macroop MULPS_XMM_M {
+    ldfp ufp1, seg, sib, "DISPLACEMENT", dataSize=8
+    ldfp ufp2, seg, sib, "DISPLACEMENT + 8", dataSize=8
+    mmulf xmml, xmml, ufp1, size=4, ext=0
+    mmulf xmmh, xmmh, ufp2, size=4, ext=0
+};
+
+def macroop MULPS_XMM_P {
+    rdip t7
+    ldfp ufp1, seg, riprel, "DISPLACEMENT", dataSize=8
+    ldfp ufp2, seg, riprel, "DISPLACEMENT + 8", dataSize=8
+    mmulf xmml, xmml, ufp1, size=4, ext=0
+    mmulf xmmh, xmmh, ufp2, size=4, ext=0
+};
+
+def macroop MULPD_XMM_XMM {
+    mmulf xmml, xmml, xmmlm, size=8, ext=0
+    mmulf xmmh, xmmh, xmmhm, size=8, ext=0
+};
+
+def macroop MULPD_XMM_M {
+    ldfp ufp1, seg, sib, "DISPLACEMENT", dataSize=8
+    ldfp ufp2, seg, sib, "DISPLACEMENT + 8", dataSize=8
+    mmulf xmml, xmml, ufp1, size=8, ext=0
+    mmulf xmmh, xmmh, ufp2, size=8, ext=0
+};
+
+def macroop MULPD_XMM_P {
+    rdip t7
+    ldfp ufp1, seg, riprel, "DISPLACEMENT", dataSize=8
+    ldfp ufp2, seg, riprel, "DISPLACEMENT + 8", dataSize=8
+    mmulf xmml, xmml, ufp1, size=8, ext=0
+    mmulf xmmh, xmmh, ufp2, size=8, ext=0
 };
 '''

src/arch/x86/isa/insts/simd64/floating_point/arithmetic/multiplication.py

@@ -54,5 +54,18 @@
 # Authors: Gabe Black
 
 microcode = '''
-# PFMUL
+def macroop PFMUL_MMX_MMX {
+    mmulf mmx, mmx, mmxm, size=4, ext=0
+};
+
+def macroop PFMUL_MMX_M {
+    ldfp ufp1, seg, sib, disp, dataSize=8
+    mmulf mmx, mmx, ufp1, size=4, ext=0
+};
+
+def macroop PFMUL_MMX_P {
+    rdip t7
+    ldfp ufp1, seg, riprel, disp, dataSize=8
+    mmulf mmx, mmx, ufp1, size=4, ext=0
+};
 '''