Merge remote-tracking branch 'origin/GT-2857_caheckman_impliedcasts'

Ghidra/Features/Decompiler/src/decompile/cpp/options.cc

@@ -394,6 +394,21 @@ string OptionNoCastPrinting::apply(Architecture *glb,const string &p1,const stri
   return "No cast printing turned "+prop;
 }
 
+/// \class OptionHideExtensions
+/// \brief Toggle whether implied extensions (ZEXT or SEXT) are printed
+string OptionHideExtensions::apply(Architecture *glb,const string &p1,const string &p2,const string &p3) const
+
+{
+  bool val = onOrOff(p1);
+  PrintC *lng = dynamic_cast<PrintC *>(glb->print);
+  if (lng == (PrintC *)0)
+    return "Can only toggle extension hiding for C language";
+  lng->setHideImpliedExts(val);
+  string prop;
+  prop = val ? "on" : "off";
+  return "Implied extension hiding turned "+prop;
+}
+
 /// \class OptionMaxLineWidth
 /// \brief Set the maximum number of characters per decompiled line
 ///

Ghidra/Features/Decompiler/src/decompile/cpp/options.hh

@@ -144,6 +144,12 @@ public:
   virtual string apply(Architecture *glb,const string &p1,const string &p2,const string &p3) const;
 };
 
+class OptionHideExtensions : public ArchOption {
+public:
+  OptionHideExtensions(void) { name="hideextensions"; }		///< Constructor
+  virtual string apply(Architecture *glb,const string &p1,const string &p2,const string &p3) const;
+};
+
 class OptionMaxLineWidth : public ArchOption {
 public:
   OptionMaxLineWidth(void) { name = "maxlinewidth"; }	///< Constructor

Ghidra/Features/Decompiler/src/decompile/cpp/printc.cc

@@ -18,6 +18,7 @@
 
 // Operator tokens for expressions
 //                        token #in prec assoc   optype       space bump
+OpToken PrintC::hidden = { "", 1, 70, false, OpToken::hiddenfunction, 0, 0, (OpToken *)0 };
 OpToken PrintC::scope = { "::", 2, 70, true, OpToken::binary, 0, 0, (OpToken *)0 };
 OpToken PrintC::object_member = { ".", 2, 66, true, OpToken::binary, 0, 0, (OpToken *)0  };
 OpToken PrintC::pointer_member = { "->", 2, 66, true, OpToken::binary, 0, 0, (OpToken *)0 };
@@ -98,6 +99,7 @@ PrintC::PrintC(Architecture *g,const string &nm) : PrintLanguage(g,nm)
   option_convention = true;
   option_nocasts = false;
   option_unplaced = false;
+  option_hide_exts = true;
   nullToken = "NULL";
   
   // Set the flip tokens
@@ -319,6 +321,21 @@ void PrintC::opTypeCast(const PcodeOp *op)
   pushVnImplied(op->getIn(0),op,mods);
 }
 
+/// The syntax represents the given op using a function with one input,
+/// where the function name is not printed. The input expression is simply printed
+/// without adornment inside the larger expression, with one minor difference.
+/// The hidden operator protects against confusing evaluation order between
+/// the operators inside and outside the hidden function.  If both the inside
+/// and outside operators are the same associative token, the hidden token
+/// makes sure the inner expression is surrounded with parentheses.
+/// \param op is the given PcodeOp
+void PrintC::opHiddenFunc(const PcodeOp *op)
+
+{
+  pushOp(&hidden,op);
+  pushVnImplied(op->getIn(0),op,mods);
+}
+
 void PrintC::opCopy(const PcodeOp *op)
 
 {
@@ -577,8 +594,12 @@ void PrintC::opReturn(const PcodeOp *op)
 void PrintC::opIntZext(const PcodeOp *op)
 
 {
-  if (castStrategy->isZextCast(op->getOut()->getHigh()->getType(),op->getIn(0)->getHigh()->getType()))
-    opTypeCast(op);
+  if (castStrategy->isZextCast(op->getOut()->getHigh()->getType(),op->getIn(0)->getHigh()->getType())) {
+    if (isExtensionCastImplied(op))
+      opHiddenFunc(op);
+    else
+      opTypeCast(op);
+  }
   else
     opFunc(op);
 }
@@ -586,8 +607,12 @@ void PrintC::opIntZext(const PcodeOp *op)
 void PrintC::opIntSext(const PcodeOp *op)
 
 {
-  if (castStrategy->isSextCast(op->getOut()->getHigh()->getType(),op->getIn(0)->getHigh()->getType()))
-    opTypeCast(op);
+  if (castStrategy->isSextCast(op->getOut()->getHigh()->getType(),op->getIn(0)->getHigh()->getType())) {
+    if (isExtensionCastImplied(op))
+      opHiddenFunc(op);
+    else
+      opTypeCast(op);
+  }
   else
     opFunc(op);
 }
@@ -1247,6 +1272,60 @@ bool PrintC::printCharacterConstant(ostream &s,const Address &addr,int4 charsize
   return res;
 }
 
+/// \brief Is the given ZEXT/SEXT cast implied by the expression its in
+///
+/// We know that the given ZEXT or SEXT op can be viewed as a natural \e cast operation.
+/// Sometimes such a cast is implied by the expression its in, and the cast itself
+/// doesn't need to be printed.
+/// \param op is the given ZEXT or SEXT PcodeOp
+/// \return \b true if the op as a cast does not need to be printed
+bool PrintC::isExtensionCastImplied(const PcodeOp *op) const
+
+{
+  if (!option_hide_exts)
+    return false;		// If hiding extensions is not on, we must always print extension
+  const Varnode *outVn = op->getOut();
+  if (outVn->isExplicit()) {
+
+  }
+  else {
+    type_metatype metatype = outVn->getHigh()->getType()->getMetatype();
+    list<PcodeOp *>::const_iterator iter;
+    for(iter=outVn->beginDescend();iter!=outVn->endDescend();++iter) {
+      PcodeOp *expOp = *iter;
+      Varnode *otherVn;
+      int4 slot;
+      switch(expOp->code()) {
+	case CPUI_PTRADD:
+	  break;
+	case CPUI_INT_ADD:
+	case CPUI_INT_SUB:
+	case CPUI_INT_MULT:
+	case CPUI_INT_DIV:
+	case CPUI_INT_AND:
+	case CPUI_INT_OR:
+	case CPUI_INT_XOR:
+	case CPUI_INT_LESS:
+	case CPUI_INT_LESSEQUAL:
+	case CPUI_INT_SLESS:
+	case CPUI_INT_SLESSEQUAL:
+	  slot = expOp->getSlot(outVn);
+	  otherVn = expOp->getIn(1-slot);
+	  // Check if the expression involves an explicit variable of the right integer type
+	  if (!otherVn->isExplicit())
+	    return false;
+	  if (otherVn->getHigh()->getType()->getMetatype() != metatype)
+	    return false;
+	  break;
+	default:
+	  return false;
+      }
+    }
+    return true;	// Everything is integer promotion
+  }
+  return false;
+}
+
 /// \brief Push a single character constant to the RPN stack
 ///
 /// For C, a character constant is usually emitted as the character in single quotes.

Ghidra/Features/Decompiler/src/decompile/cpp/printc.hh

@@ -62,6 +62,7 @@ struct PartialSymbolEntry {
 ///  - etc.
 class PrintC : public PrintLanguage {
 protected:
+  static OpToken hidden;		///< Hidden functional (that may force parentheses)
   static OpToken scope;			///< The sub-scope/namespace operator
   static OpToken object_member;		///< The \e member operator
   static OpToken pointer_member;	///< The \e points \e to \e member operator
@@ -116,6 +117,7 @@ protected:
   bool option_convention;	///< Set to \b true if we should print calling convention
   bool option_nocasts;		///< Don't print a cast if \b true
   bool option_unplaced;		///< Set to \b true if we should display unplaced comments
+  bool option_hide_exts;	///< Set to \b true if we should hide implied extension operations
   string nullToken;		///< Token to use for 'null'
   CommentSorter commsorter;	///< Container/organizer for comments in the current function
 
@@ -153,9 +155,11 @@ protected:
   void emitCommentGroup(const PcodeOp *inst);		///< Emit comments associated with a given statement
   void emitCommentFuncHeader(const Funcdata *fd);	///< Emit comments in the given function's header
   void opFunc(const PcodeOp *op);			///< Push a \e functional expression based on the given p-code op to the RPN stack
-  void opTypeCast(const PcodeOp *op);			///< Pus the given p-code op using type-cast syntax to the RPN stack
+  void opTypeCast(const PcodeOp *op);			///< Push the given p-code op using type-cast syntax to the RPN stack
+  void opHiddenFunc(const PcodeOp *op);			///< Push the given p-code op as a hidden token
   static bool hasCharTerminator(uint1 *buffer,int4 size,int4 charsize);
   bool printCharacterConstant(ostream &s,const Address &addr,int4 charsize) const;
+  bool isExtensionCastImplied(const PcodeOp *op) const;
   virtual void pushConstant(uintb val,const Datatype *ct,
 			    const Varnode *vn,const PcodeOp *op);
   virtual bool pushEquate(uintb val,int4 sz,const EquateSymbol *sym,
@@ -195,6 +199,7 @@ public:
   void setCStyleComments(void) { setCommentDelimeter("/* "," */",false); }	///< Set c-style "/* */" comment delimiters
   void setCPlusPlusStyleComments(void) { setCommentDelimeter("// ","",true); }	///< Set c++-style "//" comment delimiters
   void setDisplayUnplaced(bool val) { option_unplaced = val; }	///< Toggle whether \e unplaced comments are displayed in the header
+  void setHideImpliedExts(bool val) { option_hide_exts = val; }	///< Toggle whether implied extensions are hidden
   virtual ~PrintC(void) {}
   virtual void adjustTypeOperators(void);
   virtual void setCommentStyle(const string &nm);

Ghidra/Features/Decompiler/src/decompile/cpp/printlanguage.cc

@@ -52,55 +52,6 @@ PrintLanguageCapability *PrintLanguageCapability::findCapability(const string &n
   return (PrintLanguageCapability *)0;
 }
 
-/// \brief Determine if the given token should be emitted in its own parenthetic expression
-///
-/// This token is being emitted. Check if its input expression, ending with the given
-/// operator token, needs to be surrounded by parentheses to convey the proper meaning.
-/// \param op2 is the input token to \b this operator
-/// \param stage is the stage of \b this operator currently being printed
-/// \return \b true if \b op2 (as input to \b this) should be parenthesized
-bool OpToken::parentheses(const OpToken &op2,int4 stage) const
-
-{
-  switch(type) {
-  case space:
-  case binary:
-    if (precedence > op2.precedence) return true;
-    if (precedence < op2.precedence) return false;
-    if (associative && (this == &op2)) return false;
-    // If operators are adjacent to each other, the
-    // operator printed first must be evaluated first
-    // In this case op2 must be evaluated first, so we
-    // check if it is printed first (in first stage of binary)
-    if ((op2.type==postsurround)&&(stage==0)) return false;
-    return true;
-  case unary_prefix:
-    if (precedence > op2.precedence) return true;
-    if (precedence < op2.precedence) return false;
-    //    if (associative && (this == &op2)) return false;
-    if ((op2.type==unary_prefix)||(op2.type==presurround)) return false;
-    return true;
-  case postsurround:
-    if (stage==1) return false;	// Inside the surround
-    if (precedence > op2.precedence) return true;
-    if (precedence < op2.precedence) return false;
-    // If the precedences are equal, we know this postsurround
-    // comes after, so op2 being first doesn't need parens
-    if ((op2.type==postsurround)||(op2.type==binary)) return false;
-    //    if (associative && (this == &op2)) return false;
-    return true;
-  case presurround:
-    if (stage==0) return false;	// Inside the surround
-    if (precedence > op2.precedence) return true;
-    if (precedence < op2.precedence) return false;
-    //    if (associative && (this == &op2)) return false;
-    if ((op2.type==unary_prefix)||(op2.type==presurround)) return false;
-    return true;
-  }
-
-  return true;
-}
-
 /// \param g is the Architecture that owns and will use this PrintLanguage
 /// \param nm is the formal name of the language
 PrintLanguage::PrintLanguage(Architecture *g,const string &nm)
@@ -178,7 +129,7 @@ void PrintLanguage::pushOp(const OpToken *tok,const PcodeOp *op)
   }
   else {
     emitOp(revpol.back());
-    paren = revpol.back().tok->parentheses(*tok,revpol.back().visited);
+    paren = parentheses(tok);
     if (paren)
       id = emit->openParen('(');
     else
@@ -308,6 +259,68 @@ void PrintLanguage::pushVnLHS(const Varnode *vn,const PcodeOp *op)
   }
 }
 
+/// The token at the top of the stack is being emitted. Check if its input expression,
+/// ending with the given operator token, needs to be surrounded by parentheses to convey
+/// the proper meaning.
+/// \param op2 is the input token to \b this operator
+/// \param stage is the stage of \b this operator currently being printed
+/// \return \b true if \b op2 (as input to \b this) should be parenthesized
+bool PrintLanguage::parentheses(const OpToken *op2)
+
+{
+  ReversePolish &top( revpol.back() );
+  const OpToken *topToken = top.tok;
+  int4 stage = top.visited;
+  switch(topToken->type) {
+  case OpToken::space:
+  case OpToken::binary:
+    if (topToken->precedence > op2->precedence) return true;
+    if (topToken->precedence < op2->precedence) return false;
+    if (topToken->associative && (topToken == op2)) return false;
+    // If operators are adjacent to each other, the
+    // operator printed first must be evaluated first
+    // In this case op2 must be evaluated first, so we
+    // check if it is printed first (in first stage of binary)
+    if ((op2->type==OpToken::postsurround)&&(stage==0)) return false;
+    return true;
+  case OpToken::unary_prefix:
+    if (topToken->precedence > op2->precedence) return true;
+    if (topToken->precedence < op2->precedence) return false;
+    //    if (associative && (this == &op2)) return false;
+    if ((op2->type==OpToken::unary_prefix)||(op2->type==OpToken::presurround)) return false;
+    return true;
+  case OpToken::postsurround:
+    if (stage==1) return false;	// Inside the surround
+    if (topToken->precedence > op2->precedence) return true;
+    if (topToken->precedence < op2->precedence) return false;
+    // If the precedences are equal, we know this postsurround
+    // comes after, so op2 being first doesn't need parens
+    if ((op2->type==OpToken::postsurround)||(op2->type==OpToken::binary)) return false;
+    //    if (associative && (this == &op2)) return false;
+    return true;
+  case OpToken::presurround:
+    if (stage==0) return false;	// Inside the surround
+    if (topToken->precedence > op2->precedence) return true;
+    if (topToken->precedence < op2->precedence) return false;
+    //    if (associative && (this == &op2)) return false;
+    if ((op2->type==OpToken::unary_prefix)||(op2->type==OpToken::presurround)) return false;
+    return true;
+  case OpToken::hiddenfunction:
+    if ((stage==0)&&(revpol.size() > 1)) {	// If there is an unresolved previous token
+      // New token is printed next to the previous token.
+      const OpToken *prevToken = revpol[revpol.size()-2].tok;
+      if (prevToken->type != OpToken::binary && prevToken->type != OpToken::unary_prefix)
+	return false;
+      if (prevToken->precedence < op2->precedence) return false;
+      // If precedence is equal, make sure we don't treat two tokens as associative,
+      // i.e. we should have parentheses
+    }
+    return true;
+  }
+
+  return true;
+}
+
 /// An OpToken directly from the RPN is sent to the low-level emitter,
 /// resolving any final spacing or parentheses.
 /// \param entry is the RPN entry to be emitted
@@ -351,6 +364,8 @@ void PrintLanguage::emitOp(const ReversePolish &entry)
     if (entry.visited != 1) return;
     emit->spaces(entry.tok->spacing,entry.tok->bump);
     break;
+  case OpToken::hiddenfunction:
+    return;			// Never directly prints anything
   }
 }
 

Ghidra/Features/Decompiler/src/decompile/cpp/printlanguage.hh

@@ -87,6 +87,7 @@ public:
     postsurround,		///< Function or array operator form
     presurround,		///< Modifier form (like a cast operation)
     space,			///< No explicitly printed token
+    hiddenfunction		///< Operation that isn't explicitly printed
   };
   const char *print;		///< Printing characters for the token
   int4 stage;			///< Additional elements consumed from the RPN stack when emitting this token
@@ -96,7 +97,6 @@ public:
   int4 spacing;			///< Spaces to print around operator
   int4 bump;			///< Spaces to indent if we break here
   OpToken *negate;		///< The token representing the negation of this token
-  bool parentheses(const OpToken &op2,int4 stage) const;
 };
 
 /// \brief The base class API for emitting a high-level language
@@ -263,6 +263,7 @@ protected:
   void pushVnExplicit(const Varnode *vn,const PcodeOp *op);		///< Push an explicit variable onto the RPN stack
   void pushVnLHS(const Varnode *vn,const PcodeOp *op);			///< Push a variable as the left-hand side of an expression
 
+  bool parentheses(const OpToken *op2);	///< Determine if the given token should be emitted in its own parenthetic expression
   void emitOp(const ReversePolish &entry);				///< Send an operator token from the RPN to the emitter
   void emitAtom(const Atom &atom);					///< Send an variable token from the RPN to the emitter
   static bool unicodeNeedsEscape(int4 codepoint);			///< Determine if the given codepoint needs to be escaped