[SharedCache] Add the ability to manually trigger Objective-C processing

A problem with processing Objective-C sections at the time a library is loaded is that some of the references within those sections may refer to unload sections. This results in things like selectors not being correctly typed and named.

This commit provides a way for users to manually trigger Objective-C parsing against sections for a specific library or all libraries, via the API. Combined with the previous commit a user can use the API to batch load a number of libraries and skip Objective-C processing for each one and then run it across the entire of the DSC once they are all loaded.

Further improvement would be to provide a way to trigger Objective-C processing through the UI.

view/sharedcache/api/python/_sharedcachecore.py

@@ -581,6 +581,44 @@ def BNDSCViewLoadSectionAtAddress(
 	return _BNDSCViewLoadSectionAtAddress(cache, name)
 
 
+# -------------------------------------------------------
+# _BNDSCViewProcessAllObjCSections
+
+_BNDSCViewProcessAllObjCSections = core.BNDSCViewProcessAllObjCSections
+_BNDSCViewProcessAllObjCSections.restype = None
+_BNDSCViewProcessAllObjCSections.argtypes = [
+		ctypes.POINTER(BNSharedCache),
+	]
+
+
+# noinspection PyPep8Naming
+def BNDSCViewProcessAllObjCSections(
+		cache: ctypes.POINTER(BNSharedCache)
+		) -> None:
+	return _BNDSCViewProcessAllObjCSections(cache)
+
+
+# -------------------------------------------------------
+# _BNDSCViewProcessObjCSectionsForImageWithInstallName
+
+_BNDSCViewProcessObjCSectionsForImageWithInstallName = core.BNDSCViewProcessObjCSectionsForImageWithInstallName
+_BNDSCViewProcessObjCSectionsForImageWithInstallName.restype = None
+_BNDSCViewProcessObjCSectionsForImageWithInstallName.argtypes = [
+		ctypes.POINTER(BNSharedCache),
+		ctypes.c_char_p,
+		ctypes.c_bool,
+	]
+
+
+# noinspection PyPep8Naming
+def BNDSCViewProcessObjCSectionsForImageWithInstallName(
+		cache: ctypes.POINTER(BNSharedCache), 
+		name: Optional[str], 
+		deallocName: bool
+		) -> None:
+	return _BNDSCViewProcessObjCSectionsForImageWithInstallName(cache, cstr(name), deallocName)
+
+
 # -------------------------------------------------------
 # _BNFreeSharedCacheReference
 

view/sharedcache/api/python/sharedcache.py

@@ -117,6 +117,12 @@ def load_section_at_address(self, addr):
 	def load_image_containing_address(self, addr, skipObjC = False):
 		return sccore.BNDSCViewLoadImageContainingAddress(self.handle, addr, skipObjC)
 
+	def process_objc_sections_for_image_with_install_name(self, installName):
+		return sccore.BNDSCViewProcessObjCSectionsForImageWithInstallName(self.handle, installName, False)
+
+	def process_all_objc_sections(self):
+		return sccore.BNDSCViewProcessAllObjCSections(self.handle)
+
 	@property
 	def caches(self):
 		count = ctypes.c_ulonglong()

view/sharedcache/api/sharedcache.cpp

@@ -55,6 +55,17 @@ namespace SharedCacheAPI {
 		return result;
 	}
 
+	void SharedCache::ProcessObjCSectionsForImageWithInstallName(std::string installName)
+	{
+		char* str = BNAllocString(installName.c_str());
+		BNDSCViewProcessObjCSectionsForImageWithInstallName(m_object, str, true);
+	}
+
+	void SharedCache::ProcessAllObjCSections()
+	{
+		BNDSCViewProcessAllObjCSections(m_object);
+	}
+
 	std::vector<DSCMemoryRegion> SharedCache::GetLoadedMemoryRegions()
 	{
 		size_t count;

view/sharedcache/api/sharedcacheapi.h

@@ -261,6 +261,9 @@ namespace SharedCacheAPI {
 		bool LoadSectionAtAddress(uint64_t addr);
 		bool LoadImageContainingAddress(uint64_t addr, bool skipObjC = false);
 		std::vector<std::string> GetAvailableImages();
+
+		void ProcessObjCSectionsForImageWithInstallName(std::string installName);
+		void ProcessAllObjCSections();
 
 		std::vector<DSCSymbol> LoadAllSymbolsAndWait();
 

view/sharedcache/api/sharedcachecore.h

@@ -123,6 +123,9 @@ extern "C"
 	SHAREDCACHE_FFI_API bool BNDSCViewLoadImageWithInstallName(BNSharedCache* cache, char* name, bool skipObjC);
 	SHAREDCACHE_FFI_API bool BNDSCViewLoadSectionAtAddress(BNSharedCache* cache, uint64_t name);
 	SHAREDCACHE_FFI_API bool BNDSCViewLoadImageContainingAddress(BNSharedCache* cache, uint64_t address, bool skipObjC);
+
+	SHAREDCACHE_FFI_API void BNDSCViewProcessObjCSectionsForImageWithInstallName(BNSharedCache* cache, char* name, bool deallocName);
+	SHAREDCACHE_FFI_API void BNDSCViewProcessAllObjCSections(BNSharedCache* cache);
 
 	SHAREDCACHE_FFI_API char* BNDSCViewGetNameForAddress(BNSharedCache* cache, uint64_t address);
 	SHAREDCACHE_FFI_API char* BNDSCViewGetImageNameForAddress(BNSharedCache* cache, uint64_t address);

view/sharedcache/core/SharedCache.cpp

@@ -1727,6 +1727,81 @@ bool SharedCache::LoadSectionAtAddress(uint64_t address)
 	return true;
 }
 
+static void GetObjCSettings(Ref<BinaryView> view, bool* processObjCMetadata, bool* processCFStrings)
+{
+	auto settings = view->GetLoadSettings(VIEW_NAME);
+	*processCFStrings = true;
+	*processObjCMetadata = true;
+	if (settings && settings->Contains("loader.dsc.processCFStrings"))
+		*processCFStrings = settings->Get<bool>("loader.dsc.processCFStrings", view);
+	if (settings && settings->Contains("loader.dsc.processObjC"))
+		*processObjCMetadata = settings->Get<bool>("loader.dsc.processObjC", view);
+}
+
+static void ProcessObjCSectionsForImageWithName(std::string baseName, std::shared_ptr<VM> vm, std::shared_ptr<DSCObjC::DSCObjCProcessor> objc, bool processCFStrings, bool processObjCMetadata, Ref<Logger> logger)
+{
+	try
+	{
+		if (processObjCMetadata)
+			objc->ProcessObjCData(vm, baseName);
+		if (processCFStrings)
+			objc->ProcessCFStrings(vm, baseName);
+	}
+	catch (const std::exception& ex)
+	{
+		logger->LogWarn("Error processing ObjC data for image %s: %s", baseName.c_str(), ex.what());
+	}
+	catch (...)
+	{
+		logger->LogWarn("Error processing ObjC data for image %s", baseName.c_str());
+	}
+}
+
+void SharedCache::ProcessObjCSectionsForImageWithInstallName(std::string installName)
+{
+	bool processCFStrings;
+	bool processObjCMetadata;
+	GetObjCSettings(m_dscView, &processCFStrings, &processObjCMetadata);
+
+	if (!processObjCMetadata && !processCFStrings)
+		return;
+
+	auto objc = std::make_shared<DSCObjC::DSCObjCProcessor>(m_dscView, this, false);
+	auto vm = GetVMMap();
+
+	ProcessObjCSectionsForImageWithName(base_name(installName), vm, objc, processCFStrings, processObjCMetadata, m_logger);
+}
+
+void SharedCache::ProcessAllObjCSections()
+{
+	bool processCFStrings;
+	bool processObjCMetadata;
+	GetObjCSettings(m_dscView, &processCFStrings, &processObjCMetadata);
+
+	if (!processObjCMetadata && !processCFStrings)
+		return;
+
+	auto objc = std::make_shared<DSCObjC::DSCObjCProcessor>(m_dscView, this, false);
+	auto vm = GetVMMap();
+
+	std::set<uint64_t> processedImageHeaders;
+	for (auto region : GetMappedRegions())
+	{
+		if (!region.loaded)
+			continue;
+
+		// Don't repeat the same images multiple times
+		auto header = HeaderForAddress(region.start);
+		if (!header)
+			continue;
+		if (processedImageHeaders.find(header->textBase) != processedImageHeaders.end())
+			continue;
+		processedImageHeaders.insert(header->textBase);
+
+		ProcessObjCSectionsForImageWithName(header->identifierPrefix, vm, objc, processCFStrings, processObjCMetadata, m_logger);
+	}
+}
+
 bool SharedCache::LoadImageWithInstallName(std::string installName, bool skipObjC)
 {
 	auto settings = m_dscView->GetLoadSettings(VIEW_NAME);
@@ -1839,29 +1914,11 @@ bool SharedCache::LoadImageWithInstallName(std::string installName, bool skipObj
 
 	if (!skipObjC)
 	{
-		try
-		{
-			auto objc = std::make_unique<DSCObjC::DSCObjCProcessor>(m_dscView, this, false);
+		bool processCFStrings;
+		bool processObjCMetadata;
+		GetObjCSettings(m_dscView, &processCFStrings, &processObjCMetadata);
 
-			bool processCFStrings = true;
-			bool processObjCMetadata = true;
-			if (settings && settings->Contains("loader.dsc.processCFStrings"))
-				processCFStrings = settings->Get<bool>("loader.dsc.processCFStrings", m_dscView);
-			if (settings && settings->Contains("loader.dsc.processObjC"))
-				processObjCMetadata = settings->Get<bool>("loader.dsc.processObjC", m_dscView);
-			if (processObjCMetadata)
-				objc->ProcessObjCData(vm, h->identifierPrefix);
-			if (processCFStrings)
-				objc->ProcessCFStrings(vm, h->identifierPrefix);
-		}
-		catch (const std::exception& ex)
-		{
-			m_logger->LogWarn("Error processing ObjC data: %s", ex.what());
-		}
-		catch (...)
-		{
-			m_logger->LogWarn("Error processing ObjC data");
-		}
+		ProcessObjCSectionsForImageWithName(h->identifierPrefix, vm, std::make_shared<DSCObjC::DSCObjCProcessor>(m_dscView, this, false), processCFStrings, processObjCMetadata, m_logger);
 	}
 
 	m_dscView->AddAnalysisOption("linearsweep");
@@ -3067,6 +3124,22 @@ extern "C"
 		return false;
 	}
 
+	void BNDSCViewProcessObjCSectionsForImageWithInstallName(BNSharedCache* cache, char* name, bool deallocName)
+	{
+		std::string imageName = std::string(name);
+		if (deallocName)
+			BNFreeString(name);
+
+		if (cache->object)
+			cache->object->ProcessObjCSectionsForImageWithInstallName(imageName);
+	}
+
+	void BNDSCViewProcessAllObjCSections(BNSharedCache* cache)
+	{
+		if (cache->object)
+			cache->object->ProcessAllObjCSections();
+	}
+
 	char** BNDSCViewGetInstallNames(BNSharedCache* cache, size_t* count)
 	{
 		if (cache->object)

view/sharedcache/core/SharedCache.h

@@ -574,6 +574,8 @@ namespace SharedCacheCore {
 		bool LoadImageWithInstallName(std::string installName, bool skipObjC);
 		bool LoadSectionAtAddress(uint64_t address);
 		bool LoadImageContainingAddress(uint64_t address, bool skipObjC);
+		void ProcessObjCSectionsForImageWithInstallName(std::string installName);
+		void ProcessAllObjCSections();
 		std::string NameForAddress(uint64_t address);
 		std::string ImageNameForAddress(uint64_t address);
 		std::vector<std::string> GetAvailableImages();