SourceCode-based IFDS/IDE UnitTests (#789)

* Add utility function to aid src-code based unittests

* half of xtaint test, stuck on 12

* Fixed all but 09 and 12

* fixed all enabled ext taint tests

* reworked test structure

* refractor + more tests. Weird errors

* debug commit

* Fix psr::getDILocation()

* std::variant for SrcCodeLocationEntry

* bug fixed + ext taint test works

* compile tests as dbg + ground truths

* Fix Gt for UninitTest_02_SHOULD_LEAK

* Add TestingSrcLocation

* GLCA nearly done, questions for meeting

* taint double free tests fail

* Fixed IFDSConstAnalysisTest

* fixed tainttest and segfaults + refactoring

* Fix AnalysisPrinterTest

* Convert AnalysisPrinterTest to src-code-based ground-truth

* Start converting InstInterationAnalysis tests

* Convert some more IIA test cases

* Add more IIA tests

* Convert FIIA-test + fix use-after-free in IIA test

* HandleTypeState_03 not working

* All but three fileio tests work

* UninitVarTest fixed all but one

* fixed all but 03 of TSAnalysisFile

* Fixed all but one test

* Fixed UninitTest_21

* Enable modeling of instruction-operands in TestingSrcLocation + fix 20th uninit test

* reworked test 06

* cleanup

* readded LIBCPP_GTEST_SKIP

* fixed use after scope in taint analysis test

* Make IDEExtendedTaintAnalysisTest more compact

* Convert IIAFlowFact tests +make GLCA test more compact

* Refactor rest of IfdsIde tests

* Move everything in SrcCodeLocationEntry to psr::unittest + some cleanup

* Some cleanup

---------

Co-authored-by: Fabian Schiebel <fabian.schiebel@iem.fraunhofer.de>
Co-authored-by: Fabian Schiebel <fabianbs@mail.upb.de>

Author: mxHuber

.clang-tidy

@@ -61,7 +61,7 @@ CheckOptions:
   - key:             readability-identifier-naming.ParameterIgnoredRegexp
     value:           (d|d1|d2|d3|d4|d5|eP|f|n)
   - key:             readability-identifier-naming.FunctionIgnoredRegexp
-    value:           (try_emplace|from_json|to_json|equal_to|to_string|DToString|NToString|FToString|LToString|hash_value)
+    value:           (try_emplace|from_json|to_json|equal_to|to_string|DToString|NToString|FToString|LToString|hash_value|dyn_cast)
   - key:             cppcoreguidelines-special-member-functions.AllowSoleDefaultDtor
     value:           1
   - key:             cppcoreguidelines-special-member-functions.AllowMissingMoveFunctions

include/phasar/DataFlow/IfdsIde/Solver/IDESolver.h

@@ -1616,7 +1616,7 @@ class IDESolver
         } else {
           // Get the fact-ID
           D1FactId = G.getFactID(D1Fact);
-          std::string D1Label = DToString(D1Fact);
+          std::string D1Label{DToString(D1Fact)};
 
           // Get or create the fact subgraph
           D1FSG = FG->getOrCreateFactSG(D1FactId, D1Label);
@@ -1634,7 +1634,7 @@ class IDESolver
           if (!IDEProblem.isZeroValue(D2Fact)) {
             // Get the fact-ID
             D2FactId = G.getFactID(D2Fact);
-            std::string D2Label = DToString(D2Fact);
+            std::string D2Label{DToString(D2Fact)};
             DOTNode D2 = {FuncName, D2Label, N2StmtId, D2FactId, false, true};
             std::string EFLabel;
             auto EFVec = IntermediateEdgeFunctions[std::make_tuple(
@@ -1719,7 +1719,7 @@ class IDESolver
         } else {
           // Get the fact-ID
           D1FactId = G.getFactID(D1Fact);
-          std::string D1Label = DToString(D1Fact);
+          std::string D1Label{DToString(D1Fact)};
           D1 = {FNameOfN1, D1Label, N1StmtId, D1FactId, false, true};
           // FG should already exist even for single statement functions
           if (!G.containsFactSG(FNameOfN1, D1FactId)) {
@@ -1738,7 +1738,7 @@ class IDESolver
           } else {
             // Get the fact-ID
             D2FactId = G.getFactID(D2Fact);
-            std::string D2Label = DToString(D2Fact);
+            std::string D2Label{DToString(D2Fact)};
             D2 = {FNameOfN2, D2Label, N2StmtId, D2FactId, false, true};
             // FG should already exist even for single statement functions
             if (!G.containsFactSG(FNameOfN2, D2FactId)) {

include/phasar/Domain/LatticeDomain.h

@@ -11,6 +11,7 @@
 #define PHASAR_DOMAIN_LATTICEDOMAIN_H
 
 #include "phasar/Utils/ByRef.h"
+#include "phasar/Utils/DebugOutput.h"
 #include "phasar/Utils/JoinLattice.h"
 #include "phasar/Utils/TypeTraits.h"
 
@@ -100,9 +101,7 @@ struct LatticeDomain : public std::variant<Top, L, Bottom> {
   }
 };
 
-template <typename L,
-          typename = std::void_t<decltype(std::declval<llvm::raw_ostream &>()
-                                          << std::declval<L>())>>
+template <typename L>
 inline llvm::raw_ostream &operator<<(llvm::raw_ostream &OS,
                                      const LatticeDomain<L> &LD) {
   if (LD.isBottom()) {
@@ -114,7 +113,11 @@ inline llvm::raw_ostream &operator<<(llvm::raw_ostream &OS,
 
   const auto *Val = LD.getValueOrNull();
   assert(Val && "Only alternative remaining is L");
-  return OS << *Val;
+  if constexpr (is_llvm_printable_v<L>) {
+    return OS << *Val;
+  } else {
+    return OS << PrettyPrinter{*Val};
+  }
 }
 
 template <typename L>

include/phasar/Utils/BitVectorSet.h

@@ -292,9 +292,9 @@ class BitVectorSet {
 
   [[nodiscard]] size_t size() const noexcept { return Bits.count(); }
 
-  [[nodiscard]] const BitVectorTy &getBits() const &noexcept { return Bits; }
-  [[nodiscard]] BitVectorTy &getBits() &noexcept { return Bits; }
-  [[nodiscard]] BitVectorTy &&getBits() &&noexcept { return std::move(Bits); }
+  [[nodiscard]] const BitVectorTy &getBits() const & noexcept { return Bits; }
+  [[nodiscard]] BitVectorTy &getBits() & noexcept { return Bits; }
+  [[nodiscard]] BitVectorTy &&getBits() && noexcept { return std::move(Bits); }
 
   friend bool operator==(const BitVectorSet &Lhs, const BitVectorSet &Rhs) {
     bool LeftEmpty = Lhs.empty();
@@ -392,6 +392,11 @@ class BitVectorSet {
     EndIter.setBits(Bits);
     return EndIter;
   }
+
+  static void clearPosition() {
+    Position.left.clear();
+    Position.right.clear();
+  }
 };
 
 // Overloads with the other intersectWith functions from Utilities.h

lib/PhasarLLVM/Utils/LLVMIRToSrc.cpp

@@ -101,15 +101,19 @@ static llvm::DISubprogram *getDISubprogram(const llvm::Value *V) {
 
 llvm::DILocation *psr::getDILocation(const llvm::Value *V) {
   // Arguments and Instruction such as AllocaInst
-  if (auto *DbgIntr = getDbgVarIntrinsic(V)) {
-    if (auto *MN = DbgIntr->getMetadata(llvm::LLVMContext::MD_dbg)) {
+
+  if (const auto *I = llvm::dyn_cast<llvm::Instruction>(V)) {
+    if (auto *MN = I->getMetadata(llvm::LLVMContext::MD_dbg)) {
       return llvm::dyn_cast<llvm::DILocation>(MN);
     }
-  } else if (const auto *I = llvm::dyn_cast<llvm::Instruction>(V)) {
-    if (auto *MN = I->getMetadata(llvm::LLVMContext::MD_dbg)) {
+  }
+
+  if (auto *DbgIntr = getDbgVarIntrinsic(V)) {
+    if (auto *MN = DbgIntr->getMetadata(llvm::LLVMContext::MD_dbg)) {
       return llvm::dyn_cast<llvm::DILocation>(MN);
     }
   }
+
   return nullptr;
 }
 

test/llvm_test_code/general_linear_constant/CMakeLists.txt

@@ -14,5 +14,5 @@ set(Sources
 )
 
 foreach(TEST_SRC ${Sources})
-  generate_ll_file(FILE ${TEST_SRC})
+  generate_ll_file(FILE ${TEST_SRC} DEBUG)
 endforeach(TEST_SRC)

test/llvm_test_code/general_linear_constant/NullTest.c

@@ -2,4 +2,7 @@ char *foo(char *str) { return str; }
 
 extern void puts(const char *);
 
-int main() { puts(foo(0)); }
+int main() {
+  puts(foo(0));
+  return 0;
+}

test/llvm_test_code/inst_interaction/CMakeLists.txt

@@ -41,9 +41,9 @@ set(SourcesOpt
 )
 
 foreach(TEST_SRC ${Sources})
-  generate_ll_file(FILE ${TEST_SRC})
+  generate_ll_file(FILE ${TEST_SRC} DEBUG)
 endforeach(TEST_SRC)
 
 foreach(TEST_SRC ${SourcesOpt})
-  generate_ll_file(FILE ${TEST_SRC} O1)
+  generate_ll_file(FILE ${TEST_SRC} DEBUG O1)
 endforeach(TEST_SRC)

test/llvm_test_code/openssl/key_derivation/CMakeLists.txt

@@ -2,5 +2,5 @@ file(GLOB openssl_files *.c *.cpp)
 
 foreach(TEST_SRC ${openssl_files})
   get_filename_component(TEST_SRC_FILE ${TEST_SRC} NAME)
-  generate_ll_file(FILE ${TEST_SRC_FILE})
+  generate_ll_file(FILE ${TEST_SRC_FILE} DEBUG)
 endforeach(TEST_SRC)

test/llvm_test_code/openssl/secure_heap/CMakeLists.txt

@@ -2,5 +2,5 @@ file(GLOB openssl_files *.c *.cpp)
 
 foreach(TEST_SRC ${openssl_files})
   get_filename_component(TEST_SRC_FILE ${TEST_SRC} NAME)
-  generate_ll_file(FILE ${TEST_SRC_FILE})
+  generate_ll_file(FILE ${TEST_SRC_FILE} DEBUG)
 endforeach(TEST_SRC)