X86 RREIL Translator

X86 Specification

- Added semantics of: MOVS/MOVSB/MOVSW/MOVSD/MOVSQ
- Bug fixes

specifications/x86/x86-pretty.ml

@@ -539,11 +539,10 @@ val show/instruction insn =
     | MOVNTQ x: "MOVNTQ" -++ show/arity2 x
     | MOVQ x: "MOVQ" -++ show/arity2 x
     | MOVQ2DQ x: "MOVQ2DQ" -++ show/arity2 x
-    | MOVSB: "MOVSB"
-    | MOVSD x: "MOVSD" -++ show/varity x
+    | MOVS x: "MOVS" -++ show/arity2 x
+    | MOVSD x: "MOVSD" -++ show/arity2 x
     | MOVSHDUP x: "MOVSHDUP" -++ show/arity2 x
     | MOVSLDUP x: "MOVSLDUP" -++ show/arity2 x
-    | MOVSQ: "MOVSQ"
     | MOVSS x: "MOVSS" -++ show/arity2 x
     | MOVSW x: "MOVSW" -++ show/arity2 x
     | MOVSX x: "MOVSX" -++ show/arity2 x

specifications/x86/x86-rreil-translator.ml

@@ -1209,6 +1209,13 @@ val sem-vmovap x = do
   commit sz dst (var temp)
 end
 
+val sem-movs x = do
+  sz <- sizeof1 x.opnd1;
+  src <- read sz x.opnd2;
+  dst <- write sz x.opnd1;
+  commit sz dst src
+end
+
 val sem-movsx x = do
   sz-dst <- sizeof1 x.opnd1;
   sz-src <- sizeof1 x.opnd2;
@@ -2170,11 +2177,10 @@ val semantics insn =
    | MOVNTQ x: sem-undef-arity2 x
    | MOVQ x: sem-undef-arity2 x
    | MOVQ2DQ x: sem-undef-arity2 x
-   | MOVSB x: sem-undef-arity0 x
-   | MOVSD x: sem-undef-varity x
+   | MOVS x: sem-movs x
+   | MOVSD x: sem-undef-arity2 x
    | MOVSHDUP x: sem-undef-arity2 x
    | MOVSLDUP x: sem-undef-arity2 x
-   | MOVSQ x: sem-undef-arity0 x
    | MOVSS x: sem-undef-arity2 x
    | MOVSW x: sem-undef-arity2 x
    | MOVSX x: sem-movsx x

specifications/x86/x86.ml

@@ -908,11 +908,10 @@ type insn =
  | MOVNTQ of arity2
  | MOVQ of arity2
  | MOVQ2DQ of arity2
- | MOVSB of arity0
- | MOVSD of varity
+ | MOVS of arity2
+ | MOVSD of arity2
  | MOVSHDUP of arity2
  | MOVSLDUP of arity2
- | MOVSQ of arity0
  | MOVSS of arity2
  | MOVSW of arity2
  | MOVSX of arity2
@@ -4072,18 +4071,18 @@ val /f3 [0x0f 0xd6 /r-reg] = binop MOVQ2DQ xmm128 mm/reg64
 
 ### MOVS/MOVSB/MOVSW/MOVSD/MOVSQ
 ###  - Move Data from String to String
-val / [0xa4] = arity0 MOVSB
+val / [0xa4] = binop MOVS (m/default/si/esi/rsi (return 8)) (m/es/di/edi/rdi (return 8))
 val / [0xa5]
- | opndsz? = arity0 MOVSB
- | rexw? = arity0 MOVSQ
- | otherwise = varity0 MOVSD
+ | opndsz? = binop MOVS (m/default/si/esi/rsi operand-size) (m/es/di/edi/rdi operand-size)
+ | rexw? = binop MOVS (m/default/si/esi/rsi operand-size) (m/es/di/edi/rdi operand-size)
+ | otherwise = binop MOVS (m/default/si/esi/rsi operand-size) (m/es/di/edi/rdi operand-size)
 
 ### MOVSD
 ###  - Move Scalar Double-Precision Floating-Point Value
-val /f2 [0x0f 0x10 /r] = varity2 MOVSD xmm128 xmm/m64
+val /f2 [0x0f 0x10 /r] = binop MOVSD xmm128 xmm/m64
 val /vex/f2/0f/vexv [0x10 /r-reg] = varity3 VMOVSD xmm128 v/xmm xmm/reg128
 val /vex/f2/0f [0x10 /r-mem] = varity2 VMOVSD xmm128 m64
-val /f2 [0x0f 0x11 /r] = varity2 MOVSD xmm/m64 xmm128
+val /f2 [0x0f 0x11 /r] = binop MOVSD xmm/m64 xmm128
 val /vex/f2/0f [0x11 /r-reg] = varity3 VMOVSD xmm/reg128 v/xmm xmm128
 val /vex/f2/0f [0x11 /r-mem] = varity2 VMOVSD m64 xmm128