param_return: improve filtration (#155)

Fixes #155 Sets of possible argments found in definition is filtered by possible argments found in its calls only when there is only one call in such definition.
avast · Aug 1, 2020 · 9a13ab3 · 9a13ab3
1 parent 5c377e1
commit 9a13ab3
Show file tree

Hide file tree

Showing 4 changed files with 46 additions and 2 deletions.
diff --git a/include/retdec/bin2llvmir/optimizations/param_return/data_entries.h b/include/retdec/bin2llvmir/optimizations/param_return/data_entries.h
@@ -165,6 +165,9 @@ class DataFlowEntry : public FunctionEntry
 
 		void setCalledValue(llvm::Value* called);
 
+		std::size_t numberOfCalls() const;
+		bool hasBranches() const;
+
 	// Usage data.
 	//
 	public:

diff --git a/src/bin2llvmir/optimizations/param_return/data_entries.cpp b/src/bin2llvmir/optimizations/param_return/data_entries.cpp
@@ -421,6 +421,38 @@ void DataFlowEntry::setCalledValue(llvm::Value* called)
 	_calledValue = called;
 }
 
+std::size_t DataFlowEntry::numberOfCalls() const
+{
+	auto fnc = getFunction();
+	if (fnc == nullptr)
+		return 0;
+
+	std::size_t calls = 0;
+	for (auto& bb: *fnc)
+		for (auto& i: bb)
+			if (auto call = dyn_cast<CallInst>(&i)) {
+				auto* calledFnc = call->getCalledFunction();
+				if (calledFnc && !calledFnc->isIntrinsic())
+					calls++;
+			}
+
+	return calls;
+}
+
+bool DataFlowEntry::hasBranches() const
+{
+	auto fnc = getFunction();
+	if (fnc == nullptr)
+		return false;
+
+	for (auto& bb: *fnc)
+		for (auto& i: bb)
+			if (isa<BranchInst>(i))
+				return true;
+
+	return false;
+}
+
 CallEntry* DataFlowEntry::createCallEntry(CallInst* call)
 {
 	_calls.push_back(CallEntry(call, this));

diff --git a/src/bin2llvmir/optimizations/param_return/filter/filter.cpp b/src/bin2llvmir/optimizations/param_return/filter/filter.cpp
@@ -174,6 +174,10 @@ void Filter::filterCalls(DataFlowEntry* de) const
 		retTempl = callRets.front();
 	}
 
+	/* When there is definition available we should
+	 * use arguments from there. This is more reliable than
+	 * constructing intersations of used arguments in file.
+	 */
 	if (de->hasDefinition())
 	{
 		FilterableLayout defArgs;
@@ -191,12 +195,16 @@ void Filter::filterCalls(DataFlowEntry* de) const
 			// below.
 			filterArgsByKnownTypes(defArgs);
 		}
-		else if (de->args().empty())
+		else if (de->args().empty() && de->numberOfCalls() == 1 && !de->hasBranches())
 		{
+			// In this case it might be wrapper that
+			// takes arguments from call and do not modify them
+			// in definition.
 			filterCallArgsByDefLayout(defArgs, argTempl);
 			de->setArgs(createGroupedArgValues(defArgs));
 		}
-		else if (argTempl.stacks.size() > defArgs.stacks.size())
+		else if (argTempl.stacks.size() > defArgs.stacks.size()
+				&& de->numberOfCalls() == 1 && !de->hasBranches())
 		{
 			if (argTempl.gpRegisters.size() == defArgs.gpRegisters.size()
 				&& argTempl.fpRegisters.size() == defArgs.fpRegisters.size()

diff --git a/src/bin2llvmir/optimizations/param_return/param_return.cpp b/src/bin2llvmir/optimizations/param_return/param_return.cpp
@@ -579,6 +579,7 @@ void ParamReturn::dumpInfo(const DataFlowEntry& de) const
 	LOG << "\t>|config f : " << (configFnc != nullptr) << std::endl;
 	LOG << "\t>|debug f  : " << (dbgFnc != nullptr) << std::endl;
 	LOG << "\t>|wrapp c  : " << llvmObjToString(wrappedCall) << std::endl;
+	LOG << "\t>|calls cnt: " << de.numberOfCalls() << std::endl;
 	LOG << "\t>|type set : " << !de.argTypes().empty() << std::endl;
 	LOG << "\t>|ret type : " << llvmObjToString(de.getRetType()) << std::endl;
 	LOG << "\t>|ret value: " << llvmObjToString(de.getRetValue()) << std::endl;