Update Ghidra

CMakeLists.txt

@@ -55,7 +55,10 @@ set(CORE_SOURCE
 		src/RizinPrintC.cpp
 		src/RzCoreMutex.h
 		src/RzCoreMutex.cpp
-		src/rz_ghidra.h)
+		src/PrettyXmlEncode.h
+		src/PrettyXmlEncode.cpp
+		src/rz_ghidra.h
+		src/rz_ghidra_internal.h)
 
 if(BUILD_SLEIGH_PLUGIN)
 set(ASM_SOURCE

ghidra/CMakeLists.txt

@@ -5,6 +5,7 @@
 # Base of everything
 set(SOURCE_BASE_CXX
 		# xml.cc				// generated by yacc task
+		marshal.cc
 		space.cc
 		float.cc
 		address.cc
@@ -46,6 +47,7 @@ set(SOURCE_DECOMPILER_CXX
 	funcdata.cc
 	funcdata_block.cc
 	funcdata_varnode.cc
+	unionresolve.cc
 	funcdata_op.cc
 	pcodeinject.cc
 	heritage.cc

src/CodeXMLParse.cpp

@@ -139,32 +139,47 @@ void AnnotateCommentOffset(ANNOTATOR_PARAMS)
 void AnnotateColor(ANNOTATOR_PARAMS)
 {
 	pugi::xml_attribute attr = node.attribute("color");
-	if (attr.empty())
+	if(attr.empty())
 		return;
 
-	std::string color = attr.as_string();
-	if (color == "")
+	int color = attr.as_int(-1);
+	if(color < 0)
 		return;
 
 	RSyntaxHighlightType type;
-	if (color == "keyword")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_KEYWORD;
-	else if (color == "comment")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_COMMENT;
-	else if (color == "type")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_DATATYPE;
-	else if (color == "funcname")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_FUNCTION_NAME;
-	else if (color == "param")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_FUNCTION_PARAMETER;
-	else if (color == "var")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_LOCAL_VARIABLE;
-	else if (color == "const")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_CONSTANT_VARIABLE;
-	else if (color == "global")
-		type = RZ_SYNTAX_HIGHLIGHT_TYPE_GLOBAL_VARIABLE;
-	else
-		return;
+	switch(color)
+	{
+		case Emit::syntax_highlight::keyword_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_KEYWORD;
+			break;
+		case Emit::syntax_highlight::comment_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_COMMENT;
+			break;
+		case Emit::syntax_highlight::type_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_DATATYPE;
+			break;
+		case Emit::syntax_highlight::funcname_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_FUNCTION_NAME;
+			break;
+		case Emit::syntax_highlight::var_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_LOCAL_VARIABLE;
+			break;
+		case Emit::syntax_highlight::const_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_CONSTANT_VARIABLE;
+			break;
+		case Emit::syntax_highlight::param_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_FUNCTION_PARAMETER;
+			break;
+		case Emit::syntax_highlight::global_color:
+			type = RZ_SYNTAX_HIGHLIGHT_TYPE_GLOBAL_VARIABLE;
+			break;
+		case Emit::syntax_highlight::no_color:
+		case Emit::syntax_highlight::error_color:
+		case Emit::syntax_highlight::special_color:
+		default:
+			return;
+	}
+
 	RzCodeAnnotation annotation = {};
 	annotation.type = RZ_CODE_ANNOTATION_TYPE_SYNTAX_HIGHLIGHT;
 	annotation.syntax_highlight.type = type;

src/PrettyXmlEncode.cpp

@@ -0,0 +1,37 @@
+// SPDX-FileCopyrightText: 2023 Florian Märkl <info@florianmaerkl.de>
+// SPDX-License-Identifier: LGPL-3.0-or-later
+
+#include "PrettyXmlEncode.h"
+
+void PrettyXmlEncode::indent()
+{
+	for(int i = 0; i < depth; i++)
+		outStream << "  ";
+}
+
+void PrettyXmlEncode::openElement(const ElementId &elemId)
+{
+	if(elementTagIsOpen)
+		outStream << ">\n";
+	else
+		elementTagIsOpen = true;
+	indent();
+	depth++;
+	outStream << '<' << elemId.getName();
+}
+
+void PrettyXmlEncode::closeElement(const ElementId &elemId)
+{
+	depth--;
+	if(elementTagIsOpen)
+	{
+		outStream << "/>\n";
+		elementTagIsOpen = false;
+	}
+	else
+	{
+		indent();
+		outStream << "</" << elemId.getName() << ">\n";
+	}
+}
+

src/PrettyXmlEncode.h

@@ -0,0 +1,21 @@
+// SPDX-FileCopyrightText: 2023 Florian Märkl <info@florianmaerkl.de>
+// SPDX-License-Identifier: LGPL-3.0-or-later
+
+#ifndef PRETTY_XML_ENCODE_H
+#define PRETTY_XML_ENCODE_H
+
+#include <marshal.hh>
+
+class PrettyXmlEncode: public XmlEncode
+{
+	private:
+		int depth = 0;
+		void indent();
+
+	public:
+		PrettyXmlEncode(std::ostream &s) : XmlEncode(s) {}
+		void openElement(const ElementId &elemId) override;
+		void closeElement(const ElementId &elemId) override;
+};
+
+#endif

src/RizinArchitecture.cpp

@@ -37,7 +37,7 @@ std::string FilenameFromCore(RzCore *core)
 }
 
 RizinArchitecture::RizinArchitecture(RzCore *core, const std::string &sleigh_id)
-	: SleighArchitecture(FilenameFromCore(core), sleigh_id.empty() ? SleighIdFromCore(core) : sleigh_id, &cout),
+	: SleighArchitecture(FilenameFromCore(core), sleigh_id.empty() ? SleighIdFromCore(core) : sleigh_id, &std::cout),
 	coreMutex(core)
 {
 }

src/RizinCommentDatabase.h

@@ -29,8 +29,8 @@ class RizinCommentDatabase : public CommentDatabase
 		CommentSet::const_iterator beginComment(const Address &fad) const override;
 		CommentSet::const_iterator endComment(const Address &fad) const override;
 
-		void saveXml(ostream &s) const override { cache.saveXml(s); }
-		void restoreXml(const Element *el, const AddrSpaceManager *trans) override { throw LowlevelError("commentdb::restoreXml unimplemented"); }
+		void encode(Encoder &encoder) const override { cache.encode(encoder); }
+		void decode(Decoder &decoder) override { throw LowlevelError("CommentDatabaseGhidra::decode unimplemented"); }
 };
 
 #endif //RZ_GHIDRA_RizinCOMMENTDATABASE_H

src/RizinScope.cpp

@@ -225,6 +225,7 @@ FunctionSymbol *RizinScope::registerFunction(RzAnalysisFunction *fcn) const
 		params.registerTrial(addr, type->getSize());
 		int4 i = params.whichTrial(addr, type->getSize());
 		params.getTrial(i).markActive();
+		params.getTrial(i).markUsed();
 		return true;
 	});
 
@@ -298,12 +299,20 @@ FunctionSymbol *RizinScope::registerFunction(RzAnalysisFunction *fcn) const
 				return true;
 			}
 
-			paramIndex = params.whichTrial(addr, type->getSize());
-			if(paramIndex < 0)
+			int4 paramTrialIndex = params.whichTrial(addr, type->getSize());
+			if(paramTrialIndex < 0)
 			{
 				arch->addWarning("Failed to determine arg index of " + to_string(var->name));
 				return true;
 			}
+
+			paramIndex = 0;
+			for(int4 i = 0; i < paramTrialIndex; i++)
+			{
+				if(!params.getTrial(i).isUsed())
+					continue;
+				paramIndex++;
+			}
 		}
 
 		varRanges.insertRange(addr.getSpace(), addr.getOffset(), last);
@@ -406,7 +415,8 @@ FunctionSymbol *RizinScope::registerFunction(RzAnalysisFunction *fcn) const
 
 	child(&doc, "rangelist");
 
-	auto sym = cache->addMapSym(&doc);
+	XmlDecode dec(arch, &doc);
+	auto sym = cache->addMapSym(dec);
 	return dynamic_cast<FunctionSymbol *>(sym);
 }
 

src/RizinScope.h

@@ -87,8 +87,8 @@ class RizinScope : public Scope
 		void renameSymbol(Symbol *sym,const string &newname) override	{ throw LowlevelError("renameSymbol unimplemented"); }
 		void retypeSymbol(Symbol *sym,Datatype *ct) override			{ throw LowlevelError("retypeSymbol unimplemented"); }
 		string makeNameUnique(const string &nm) const override			{ throw LowlevelError("makeNameUnique unimplemented"); }
-		void saveXml(ostream &s) const override							{ cache->saveXml(s); }
-		void restoreXml(const Element *el) override						{ throw LowlevelError("restoreXml unimplemented"); }
+		void encode(Encoder &encoder) const override					{ cache->encode(encoder); }
+		void decode(Decoder &decoder) override							{ throw LowlevelError("decode unimplemented"); }
 		void printEntries(ostream &s) const override					{ throw LowlevelError("printEntries unimplemented"); }
 		int4 getCategorySize(int4 cat) const override					{ throw LowlevelError("getCategorySize unimplemented"); }
 		Symbol *getCategorySymbol(int4 cat,int4 ind) const override		{ throw LowlevelError("getCategorySymbol unimplemented"); }

src/RizinTypeFactory.cpp

@@ -53,11 +53,13 @@ Datatype *RizinTypeFactory::addRizinStruct(RzBaseType *type, StackTypes &stack_t
 			// if(elements > 0)
 			// 	memberType = getTypeArray(elements, memberType);
 
-			fields.push_back({
+			TypeField tf = {
+				(int4)offset, // id = offset by default
 				(int4)offset, // Currently, this is 0 most of the time: member->offset,
 				std::string(member->name),
 				member_type
-			});
+			};
+			fields.push_back(tf);
 
 			// TODO: right now, we track member offset ourselves
 			// which means all structs are assumed to be packed.
@@ -110,7 +112,7 @@ Datatype *RizinTypeFactory::addRizinTypedef(RzBaseType *type, StackTypes &stack_
 	Datatype *resolved = fromRzTypeInternal(type->type, nullptr, &stack_types, true, false); // use prototype=true to avoid recursion
 	if(!resolved)
 		return nullptr;
-	Datatype *typedefd = getTypedef(resolved, type->name, 0);
+	Datatype *typedefd = getTypedef(resolved, type->name, 0, 0);
 	fromRzTypeInternal(type->type, nullptr, &stack_types, false, false); // fully create the type after querying with prototype=true before
 	return typedefd;
 }