Use ghidra-native and remove the huge ghidra submodule

* 20MB repo instead of 600MB
* Apply the minimum required patches to work with r2ghidra
* Add STM8 support
* Fix crash happening in bc/6994 binary
* Bump version to 5.3.1

.github/workflows/ci.yml

@@ -73,7 +73,7 @@ jobs:
     - name: Installing Radare2
       run: sudo dpkg -i radare2/dist/debian/*/*.deb
     - name: Building r2ghidra with cmake
-      run: git submodule update --init ; mkdir build && cd build && cmake .. && make -j4 && sudo make install
+      run: git submodule update --init ; make ghidra/ghidra/Ghidra ; mkdir build && cd build && cmake .. && make -j4 && sudo make install
     - name: Packaging r2ghidra for Debian
       run: make -C scripts/debian
     - name: Testing r2ghidra oneliner
@@ -144,9 +144,9 @@ jobs:
         set PATH=%CD%\radare2\bin;%PATH%
         set ARCH=x64
         git submodule update --init
+        make ghidra/ghidra/Ghidra
         call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64
-        mkdir destdir
-        mkdir build
+        mkdir destdir build
         cd build
         cmake -GNinja ..
         cmake --build .

.github/workflows/docker.yml

@@ -1,22 +0,0 @@
-name: docker
-on:
-  push:
-    branches:
-      - master
-  pull_request:
-
-jobs:
-  build:
-    name: docker
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v2
-      - name: Checkout submodules
-        run: |
-            git submodule init
-            git submodule update
-      - name: Docker Build
-        run: cd scripts && docker build -t r2ghidra . -f Dockerfile.buster && cd ..
-      - name: Build and Test
-        run: docker run --rm -v "`pwd`:/build" -t r2ghidra /bin/bash -c "cd /build && scripts/ci-script"
-

.gitmodules

@@ -1,7 +1,3 @@
-[submodule "ghidra/ghidra"]
-	path = ghidra/ghidra
-	branch = r2ghidra
-	url = https://github.com/radareorg/ghidra.git
 [submodule "third-party/pugixml"]
 	path = third-party/pugixml
 	url = https://github.com/zeux/pugixml.git

Makefile

@@ -13,16 +13,23 @@ endif
 ghidra-processors.txt:
 	cp -f ghidra-processors.txt.default ghidra-processors.txt
 
-asan:
-	CFLAGS="-fsanitize=address" $(MAKE) -C src
+asan: ghidra/ghidra/Ghidra ghidra-processors.txt
+	#touch ghidra/ghidra/Ghidra/Features/Decompiler/src/decompile/cpp/*.cc src/*.cpp
+	touch src/*.cpp
+	CFLAGS="-fsanitize=address -g" $(MAKE) -C src -j
+	make user-install
+
+asan-run:
+	DYLD_INSERT_LIBRARIES=/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/lib/clang/12.0.0/lib/darwin/libclang_rt.asan_osx_dynamic.dylib r2 -s 9664 /Users/pancake/Downloads/usr-4/bin/bc
 
 help:
 	@echo
-	@echo "./configure       # first you need to run configure"
-	@echo "make              # build r2ghidra plugins"
-	@echo "make install      # install plugin and sleighs into prefix"
-	@echo "make user-install # install in your home"
-	@echo "make uninstall    # uninstall r2ghidra from prefix (see user-uninstall)"
+	@echo "./configure          # first you need to run configure"
+	@echo "make                 # build r2ghidra plugins"
+	@echo "make install         # install plugin and sleighs into prefix"
+	@echo "make uninstall       # uninstall r2ghidra from prefix (see user-uninstall)"
+	@echo "make user-install    # install in your home"
+	@echo "make user-uninstall  # uninstall r2ghidra from prefix (see user-uninstall)"
 	@echo
 
 clean:
@@ -53,8 +60,16 @@ user-uninstall:
 	$(MAKE) -C src sleigh-uninstall
 	rm -f $(DESTDIR)/$(BINDIR)/sleighc
 
+gclean:
+	rm -rf ghidra-native ghidra/ghidra/Ghidra
+
 ghidra/ghidra/Ghidra:
-	$(MAKE) -C ghidra
+	git clone https://github.com/radareorg/ghidra-native
+	$(MAKE) -C ghidra-native patch
+	mkdir -p ghidra/ghidra/Ghidra/Features/Decompiler/src/decompile/cpp
+	cp -rf ghidra-native/src/decompiler/* ghidra/ghidra/Ghidra/Features/Decompiler/src/decompile/cpp
+	mkdir -p ghidra/ghidra/Ghidra/Processors
+	cp -rf ghidra-native/src/Processors/* ghidra/ghidra/Ghidra/Processors/
 
 mrproper: clean
 	git submodule deinit --all -f

README.md

@@ -89,8 +89,8 @@ At the moment there is no way to select which processors to support, so it build
 First, make sure the submodule contained within this repository is fetched and up to date:
 
 ```
-git submodule init
-git submodule update
+git submodule update --init
+make ghidra/ghidra/Ghidra
 ```
 
 Then, the radare2 plugin can be built and installed as follows:

configure

@@ -109,12 +109,12 @@ done
 : ${INSTALL_PROGRAM_STRIP:=${INSTALL} -m 755 -s}
 : ${INSTALL_MAN:=${INSTALL} -m 444}
 : ${INSTALL_LIB:=${INSTALL} -m 755 -c}
- PKGNAME='r2ghidra' ; VERSION='5.2.1' ; VERSION_MAJOR=5; VERSION_MINOR=2; VERSION_PATCH=1; VERSION_NUMBER=50201; CONTACT_MAIL="[email protected]" ; CONTACT_NAME="pancake" ; CONTACT="pancake <[email protected]>" ;
+ PKGNAME='r2ghidra' ; VERSION='5.3.1' ; VERSION_MAJOR=5; VERSION_MINOR=3; VERSION_PATCH=1; VERSION_NUMBER=50301; CONTACT_MAIL="[email protected]" ; CONTACT_NAME="pancake" ; CONTACT="pancake <[email protected]>" ;
 }
 
 show_usage() {
 cat <<EOF2
-'configure' configures r2ghidra-5.2.1 to adapt to many kinds of systems.
+'configure' configures r2ghidra-5.3.1 to adapt to many kinds of systems.
 
 Usage: ./configure [OPTION]... [VAR=VALUE]...
 
@@ -186,10 +186,10 @@ ocho() {
 
 show_version() {
 if [ "$QUIET" = 1 ]; then
-	echo "5.2.1"
+	echo "5.3.1"
 	exit 0
 fi
-echo "r2ghidra-5.2.1 configuration script done with acr v1.9.8.
+echo "r2ghidra-5.3.1 configuration script done with acr v1.9.8.
 The 'Free Software Foundation' message is only for autodetection.
 Originally written by pancake <nopcode.org>."
 exit 0
@@ -218,7 +218,7 @@ case $flag in
 	show_version ; ;;
 -r|--r|--report)
 echo "PKGNAME:   r2ghidra"
-echo "VERSION:   5.2.1"
+echo "VERSION:   5.3.1"
 echo "LANGS:     c++"
 echo "PKG-CONFIG:  r_core r_asm r_anal"
 	exit 0

configure.acr

@@ -1,5 +1,5 @@
 PKGNAME r2ghidra
-VERSION 5.2.1
+VERSION 5.3.1
 CONTACT pancake ; [email protected]
 
 LANG_CXX!

ghidra-processors.txt.all

@@ -21,6 +21,7 @@ PIC
 PowerPC
 RISCV
 Sparc
+STM8
 SuperH
 SuperH4
 TI_MSP430

meson.build

@@ -3,7 +3,7 @@ project(
   'cpp',
 license : 'LGPL3',
 meson_version : '>=0.50.1',
-version : '5.2.1',
+version : '5.3.1',
 default_options : ['c_std=c11', 'cpp_std=c++11']
 )
 

src/ArchMap.cpp

@@ -87,6 +87,7 @@ static const std::map<std::string, ArchMapper> arch_map = {
 	{ "pic24", { "PIC-24F", "default", 24 } },
 	{ "z80", { "z80", "default", 8 } },
 	{ "sparc", { "sparc" } },
+	{ "stm8", { "STM8" } },
 	{ "sh", { "SuperH4" } },
 	{ "msp430", { "TI_MSP430" } },
 	{ "m68k", {

src/CodeXMLParse.cpp

@@ -60,6 +60,7 @@ void AnnotateOpref(ANNOTATOR_PARAMS)
 	annotation.type = R_CODEMETA_TYPE_OFFSET;
 	annotation.offset.offset = op->getAddr().getOffset();
 }
+
 void AnnotateFunctionName(ANNOTATOR_PARAMS)
 {
 	const char *func_name = node.child_value();
@@ -180,10 +181,9 @@ void AnnotateLocalVariable(Symbol *symbol, std::vector<RCodeMetaItem> *out)
 		return;
 	RCodeMetaItem annotation = {};
 	annotation.variable.name = strdup(symbol->getName().c_str());
-	if(symbol->getCategory() == 0)
-		annotation.type = R_CODEMETA_TYPE_FUNCTION_PARAMETER;
-	else
-		annotation.type = R_CODEMETA_TYPE_LOCAL_VARIABLE;
+	annotation.type = (symbol->getCategory() == 0)
+		? R_CODEMETA_TYPE_FUNCTION_PARAMETER
+		: R_CODEMETA_TYPE_LOCAL_VARIABLE;
 	out->push_back(annotation);
 }
 
@@ -297,13 +297,8 @@ static void ParseNode(pugi::xml_node node, ParseCodeXMLContext *ctx, std::ostrea
 	for(auto &annotation : annotations)
 	{
 		annotation.end = stream.tellp();
-#if R2_VERSION_MAJOR == 5 && R2_VERSION_MINOR == 2 && R2_VERSION_PATCH == 0
-		r_codemeta_add_annotation (code, &annotation);
-#else
 		r_codemeta_add_item (code, &annotation);
-#endif
 	}
-
 #ifdef TEST_UNKNOWN_NODES
 	if(close_test)
 		stream << "</" << node.name() << ">";

src/R2PrintC.cpp

@@ -24,7 +24,7 @@ R2PrintC::R2PrintC(Architecture *g, const string &nm)
 	: PrintC(g, nm)
 {
  	option_NULL = true;
-	option_space_after_comma = true;
+//	option_space_after_comma = true;
 // 	option_nocasts = true;
 ///  option_convention = true;
 ///  option_hide_exts = true;

src/R2Scope.cpp

@@ -123,7 +123,7 @@ FunctionSymbol *R2Scope::registerFunction(RAnalFunction *fcn) const
 				auto flag = reinterpret_cast<RFlagItem *>(pos);
 				if(flag->space && flag->space->name && !strcmp(flag->space->name, R_FLAGS_FS_SECTIONS))
 					continue;
-				if (flag->realname && *flag->realname) {
+				if (!R_STR_ISEMPTY (flag->realname)) {
 					fcn_name = flag->realname;
 					break;
 				}
@@ -180,10 +180,7 @@ FunctionSymbol *R2Scope::registerFunction(RAnalFunction *fcn) const
 			{
 				uintb off;
 				int delta = var->delta + fcn->bp_off - extraPop; // not 100% sure if extraPop is correct here
-				if(delta >= 0)
-					off = delta;
-				else
-					off = stackSpace->getHighest() + delta + 1;
+				off = (delta >= 0)? delta: stackSpace->getHighest() + delta + 1;
 				return Address(stackSpace, off);
 			}
 			case R_ANAL_VAR_KIND_REG:
@@ -496,7 +493,7 @@ Symbol *R2Scope::queryR2Absolute(ut64 addr, bool contain) const
 	RCoreLock core(arch->getCore());
 
 	RAnalFunction *fcn = r_anal_get_function_at(core->anal, addr);
-#if 0
+#if 1
 	// This can cause functions to be registered twice (hello-arm test)
 	if(!fcn && contain)
 	{
@@ -642,5 +639,7 @@ LabSymbol *R2Scope::findCodeLabel(const Address &addr) const
 
 Funcdata *R2Scope::resolveExternalRefFunction(ExternRefSymbol *sym) const
 {
+	if (!sym)
+		return nullptr;
 	return queryFunction(sym->getRefAddr());
 }

src/SleighInstruction.cpp

@@ -43,7 +43,7 @@ SleighInstruction *R2Sleigh::getInstruction(Address &addr)
 void SleighParserContext::setPrototype(SleighInstructionPrototype *p)
 {
 	prototype = p;
-	*getBaseState() = &prototype->rootState;
+ 	*getBaseState() = &prototype->rootState;
 }
 
 void R2Sleigh::reconstructContext(ParserContext &protoContext)

src/asm_ghidra.cpp

@@ -86,7 +86,7 @@ static RAsmPlugin r_asm_plugin_ghidra = {
 	KV(.arch, "sleigh"),
 	KV(.author, "FXTi"),
 	KV(.version, nullptr),
-	KV(.cpus, "6502,6805,8051,arm,avr,cr16,dalvik,hppa,java,m68k,m8c,mips,mcs96,msp430,pic24,ppc,sh,sparc,tricore,toy,v850,x86,z80"),
+	KV(.cpus, "6502,6805,8051,arm,avr,cr16,dalvik,hppa,java,m68k,m8c,mips,mcs96,msp430,pic24,ppc,sh,sparc,stm8,tricore,toy,v850,x86,z80"),
 	KV(.desc, "SLEIGH Disassembler from Ghidra"),
 	KV(.license, "GPL3"),
 	KV(.user, nullptr),

src/core_ghidra.cpp

@@ -19,6 +19,8 @@
 #include <vector>
 #include <mutex>
 
+#undef DEBUG_EXCEPTIONS
+
 #define CMD_PREFIX "pdg"
 #define CFG_PREFIX "r2ghidra"
 
@@ -60,8 +62,10 @@ static const ConfigVar cfg_var_sleighhome   ("sleighhome",  "",         "SLEIGHH
 static const ConfigVar cfg_var_sleighid     ("lang",        "",         "Custom Sleigh ID to override auto-detection (e.g. x86:LE:32:default)");
 static const ConfigVar cfg_var_cmt_cpp      ("cmt.cpp",     "true",     "C++ comment style");
 static const ConfigVar cfg_var_cmt_indent   ("cmt.indent",  "4",        "Comment indent");
+#if 0
 static const ConfigVar cfg_var_nl_brace     ("nl.brace",    "false",    "Newline before opening '{'");
 static const ConfigVar cfg_var_nl_else      ("nl.else",     "false",    "Newline before else");
+#endif
 static const ConfigVar cfg_var_indent       ("indent",      "4",        "Indent increment");
 static const ConfigVar cfg_var_linelen      ("linelen",     "120",      "Max line length");
 static const ConfigVar cfg_var_maximplref   ("maximplref",  "2",        "Maximum number of references to an expression before showing an explicit variable.");
@@ -116,8 +120,6 @@ static void PrintUsage(const RCore *const core)
 
 enum class DecompileMode { DEFAULT, XML, DEBUG_XML, OFFSET, STATEMENTS, JSON };
 
-//#define DEBUG_EXCEPTIONS
-
 static void ApplyPrintCConfig(RConfig *cfg, PrintC *print_c)
 {
 	if(!print_c)
@@ -128,11 +130,13 @@ static void ApplyPrintCConfig(RConfig *cfg, PrintC *print_c)
 	else
 		print_c->setCStyleComments();
 
+/*
 	print_c->setSpaceAfterComma(true);
 
 	print_c->setNewlineBeforeOpeningBrace(cfg_var_nl_brace.GetBool(cfg));
 	print_c->setNewlineBeforeElse(cfg_var_nl_else.GetBool(cfg));
 	print_c->setNewlineAfterPrototype(false);
+*/
 	print_c->setIndentIncrement(cfg_var_indent.GetInt(cfg));
 	print_c->setLineCommentIndent(cfg_var_cmt_indent.GetInt(cfg));
 	print_c->setMaxLineSize(cfg_var_linelen.GetInt(cfg));
@@ -144,18 +148,20 @@ static void Decompile(RCore *core, ut64 addr, DecompileMode mode, std::stringstr
 	if(!function)
 		throw LowlevelError("No function at this offset");
 	R2Architecture arch(core, cfg_var_sleighid.GetString(core->config));
-	DocumentStorage store;
+	DocumentStorage store = DocumentStorage();
 	arch.max_implied_ref = cfg_var_maximplref.GetInt(core->config);
 	arch.setRawPtr(cfg_var_rawptr.GetBool(core->config));
 	arch.init(store);
-	Funcdata *func = arch.symboltab->getGlobalScope()->findFunction(Address(arch.getDefaultCodeSpace(), function->addr));
+
+	auto faddr = Address(arch.getDefaultCodeSpace(), function->addr);
+	Funcdata *func = arch.symboltab->getGlobalScope()->findFunction(faddr);
 	arch.print->setOutputStream(&out_stream);
 	arch.setPrintLanguage("r2-c-language");
 	auto r2c = dynamic_cast<R2PrintC *>(arch.print);
 	bool showCasts = cfg_var_casts.GetBool(core->config);
 	r2c->setOptionNoCasts(!showCasts);
 	ApplyPrintCConfig(core->config, dynamic_cast<PrintC *>(arch.print));
-	if(!func)
+	if(func == nullptr)
 		throw LowlevelError("No function in Scope");
 	arch.getCore()->sleepBegin();
 	auto action = arch.allacts.getCurrent();
@@ -545,7 +551,7 @@ static int r2ghidra_cmd(void *user, const char *input)
 	RCore *core = (RCore *) user;
 	if (!strncmp (input, CMD_PREFIX, strlen(CMD_PREFIX)))
 	{
-		_cmd (core, input + 3);
+		_cmd (core, input + strlen(CMD_PREFIX));
 		return true;
 	}
 	return false;