Pull in the IDESolver++ (aka. IterativeIDESolver) (#733)

* Pull in the IDESolver++ (aka. IterativeIDESolver) The solver is set-up in JF_N configuration from the paper "Scaling Interprocedural Static Data-Flow Analysis to Large C/C++ Applications: An Experience Report" * Some cleanup based on review * edge fact printing * more cleanup * tracking issue for c++20 * minor cleanup * Add alignment comment * Apply review comments * Remove confusing copyright notes + fix header guards for newly added files
secure-software-engineering · Nov 10, 2024 · 64a7233 · 64a7233
1 parent 96f0758
commit 64a7233
Show file tree

Hide file tree

Showing 66 changed files with 5,182 additions and 409 deletions.
diff --git a/include/phasar/ControlFlow/ICFGBase.h b/include/phasar/ControlFlow/ICFGBase.h
@@ -119,6 +119,10 @@ template <typename Derived> class ICFGBase {
     return self().getAsJsonImpl();
   }
 
+  [[nodiscard]] size_t getNumCallSites() const noexcept {
+    return self().getNumCallSitesImpl();
+  }
+
 private:
   const Derived &self() const noexcept {
     return static_cast<const Derived &>(*this);

diff --git a/include/phasar/DataFlow/IfdsIde/EdgeFunction.h b/include/phasar/DataFlow/IfdsIde/EdgeFunction.h
@@ -12,6 +12,7 @@
 
 #include "phasar/DataFlow/IfdsIde/EdgeFunctionSingletonCache.h"
 #include "phasar/Utils/ByRef.h"
+#include "phasar/Utils/EmptyBaseOptimizationUtils.h"
 #include "phasar/Utils/TypeTraits.h"
 
 #include "llvm/ADT/DenseMapInfo.h"

diff --git a/include/phasar/DataFlow/IfdsIde/GenericFlowFunction.h b/include/phasar/DataFlow/IfdsIde/GenericFlowFunction.h
@@ -0,0 +1,108 @@
+#ifndef PHASAR_PHASARLLVM_DATAFLOWSOLVER_IFDSIDE_GENERICFLOWFUNCTION_H
+#define PHASAR_PHASARLLVM_DATAFLOWSOLVER_IFDSIDE_GENERICFLOWFUNCTION_H
+
+#include "phasar/DataFlow/IfdsIde/FlowFunctions.h"
+
+namespace psr {
+/// Encapsulates an unmanaged pointer to a FlowFunction
+template <typename D, typename Container = std::set<D>>
+class GenericFlowFunctionView {
+public:
+  using FlowFunctionType = FlowFunction<D, Container>;
+  using FlowFunctionPtrType = std::unique_ptr<FlowFunctionType>;
+
+  using container_type = Container;
+  using value_type = typename container_type::value_type;
+
+  GenericFlowFunctionView() noexcept = default;
+  GenericFlowFunctionView(FlowFunctionType *FF) noexcept : FF(FF) {}
+
+  GenericFlowFunctionView(const GenericFlowFunctionView &) noexcept = default;
+  GenericFlowFunctionView &
+  operator=(const GenericFlowFunctionView &) noexcept = default;
+
+  ~GenericFlowFunctionView() = default;
+
+  [[nodiscard]] container_type computeTargets(D Source) const {
+    assert(FF != nullptr);
+    return FF->computeTargets(std::move(Source));
+  }
+
+  explicit operator bool() const noexcept { return FF; }
+
+  [[nodiscard]] bool operator==(GenericFlowFunctionView Other) const noexcept {
+    return FF == Other.FF;
+  }
+  [[nodiscard]] bool operator==(std::nullptr_t) const noexcept {
+    return FF == nullptr;
+  }
+  [[nodiscard]] bool operator!=(GenericFlowFunctionView Other) const noexcept {
+    return !(*this == Other);
+  }
+  [[nodiscard]] bool operator!=(std::nullptr_t) const noexcept { return FF; }
+
+private:
+  FlowFunctionType *FF = nullptr;
+};
+
+/// Encapsulates a managed pointer to a FlowFunction
+template <typename D, typename Container = std::set<D>>
+class GenericFlowFunction {
+public:
+  using FlowFunctionType = FlowFunction<D, Container>;
+  using FlowFunctionPtrType = typename FlowFunctionType::FlowFunctionPtrType;
+
+  using container_type = Container;
+  using value_type = typename container_type::value_type;
+
+  GenericFlowFunction() noexcept = default;
+  GenericFlowFunction(FlowFunctionPtrType FF) noexcept : FF(std::move(FF)) {}
+  template <typename T, typename = std::enable_if_t<std::is_base_of_v<
+                            FlowFunctionType, std::decay_t<T>>>>
+  GenericFlowFunction(T &&FF)
+      : FF(std::make_unique<std::decay_t<T>>(std::forward<T>(FF))) {}
+
+  template <typename T, typename... ArgTys>
+  explicit GenericFlowFunction(std::in_place_type_t<T> /*unused*/,
+                               ArgTys &&...Args)
+      : FF(std::make_unique<T>(std::forward<ArgTys>(Args)...)) {}
+
+  GenericFlowFunction(GenericFlowFunction &&) noexcept = default;
+  GenericFlowFunction &operator=(GenericFlowFunction &&) noexcept = default;
+
+  GenericFlowFunction(const GenericFlowFunction &) = delete;
+  GenericFlowFunction &operator=(const GenericFlowFunction &) = delete;
+
+  ~GenericFlowFunction() = default;
+
+  [[nodiscard]] container_type computeTargets(D Source) const {
+    assert(FF != nullptr);
+    return FF->computeTargets(std::move(Source));
+  }
+
+  explicit operator bool() const noexcept { return FF; }
+
+  operator GenericFlowFunctionView<D, Container>() const noexcept {
+    return FF.get();
+  }
+
+  [[nodiscard]] bool
+  operator==(GenericFlowFunctionView<D, Container> Other) const noexcept {
+    return FF == Other.FF;
+  }
+  [[nodiscard]] bool operator==(std::nullptr_t) const noexcept {
+    return FF == nullptr;
+  }
+  [[nodiscard]] bool
+  operator!=(GenericFlowFunctionView<D, Container> Other) const noexcept {
+    return !(*this == Other);
+  }
+  [[nodiscard]] bool operator!=(std::nullptr_t) const noexcept { return FF; }
+
+private:
+  FlowFunctionPtrType FF;
+};
+
+} // namespace psr
+
+#endif // PHASAR_PHASARLLVM_DATAFLOWSOLVER_IFDSIDE_GENERICFLOWFUNCTION_H
diff --git a/include/phasar/DataFlow/IfdsIde/IDETabulationProblem.h b/include/phasar/DataFlow/IfdsIde/IDETabulationProblem.h
@@ -18,6 +18,7 @@
 #include "phasar/DataFlow/IfdsIde/FlowFunctions.h"
 #include "phasar/DataFlow/IfdsIde/IFDSIDESolverConfig.h"
 #include "phasar/DataFlow/IfdsIde/InitialSeeds.h"
+#include "phasar/DataFlow/IfdsIde/Solver/GenericSolverResults.h"
 #include "phasar/DataFlow/IfdsIde/SolverResults.h"
 #include "phasar/Utils/JoinLattice.h"
 #include "phasar/Utils/NullAnalysisPrinter.h"
@@ -134,15 +135,15 @@ class IDETabulationProblem : public FlowFunctions<AnalysisDomainTy, Container>,
   /// Generates a text report of the results that is written to the specified
   /// output stream.
   virtual void
-  emitTextReport([[maybe_unused]] const SolverResults<n_t, d_t, l_t> &Results,
+  emitTextReport([[maybe_unused]] GenericSolverResults<n_t, d_t, l_t> Results,
                  llvm::raw_ostream &OS = llvm::outs()) {
     OS << "No text report available!\n";
   }
 
   /// Generates a graphical report, e.g. in html or other markup languages, of
   /// the results that is written to the specified output stream.
   virtual void emitGraphicalReport(
-      [[maybe_unused]] const SolverResults<n_t, d_t, l_t> &Results,
+      [[maybe_unused]] GenericSolverResults<n_t, d_t, l_t> Results,
       llvm::raw_ostream &OS = llvm::outs()) {
     OS << "No graphical report available!\n";
   }
@@ -151,6 +152,8 @@ class IDETabulationProblem : public FlowFunctions<AnalysisDomainTy, Container>,
   /// the level of soundness is ignored. Otherwise, true.
   virtual bool setSoundness(Soundness /*S*/) { return false; }
 
+  const ProjectIRDBBase<db_t> *getProjectIRDB() const noexcept { return IRDB; }
+
 protected:
   typename FlowFunctions<AnalysisDomainTy, Container>::FlowFunctionPtrType
   generateFromZero(d_t FactToGenerate) {

diff --git a/include/phasar/DataFlow/IfdsIde/Solver/Compressor.h b/include/phasar/DataFlow/IfdsIde/Solver/Compressor.h
@@ -0,0 +1,191 @@
+#ifndef PHASAR_DATAFLOW_IFDSIDE_SOLVER_COMPRESSOR_H
+#define PHASAR_DATAFLOW_IFDSIDE_SOLVER_COMPRESSOR_H
+
+#include "phasar/DB/ProjectIRDBBase.h"
+#include "phasar/Utils/ByRef.h"
+#include "phasar/Utils/TypeTraits.h"
+
+#include "llvm/ADT/DenseMap.h"
+#include "llvm/ADT/DenseMapInfo.h"
+#include "llvm/ADT/SmallVector.h"
+
+#include <cstdint>
+#include <deque>
+#include <functional>
+#include <optional>
+#include <type_traits>
+
+namespace psr {
+template <typename T, typename Enable = void> class Compressor;
+
+template <typename T>
+class Compressor<T, std::enable_if_t<CanEfficientlyPassByValue<T>>> {
+public:
+  void reserve(size_t Capacity) {
+    assert(Capacity <= UINT32_MAX);
+    ToInt.reserve(Capacity);
+    FromInt.reserve(Capacity);
+  }
+
+  uint32_t getOrInsert(T Elem) {
+    auto [It, Inserted] = ToInt.try_emplace(Elem, ToInt.size());
+    if (Inserted) {
+      FromInt.push_back(Elem);
+    }
+    return It->second;
+  }
+
+  std::optional<uint32_t> getOrNull(T Elem) const {
+    if (auto It = ToInt.find(Elem); It != ToInt.end()) {
+      return It->second;
+    }
+    return std::nullopt;
+  }
+
+  T operator[](size_t Idx) const noexcept {
+    assert(Idx < FromInt.size());
+    return FromInt[Idx];
+  }
+
+  [[nodiscard]] size_t size() const noexcept { return FromInt.size(); }
+  [[nodiscard]] size_t capacity() const noexcept {
+    return FromInt.capacity() +
+           ToInt.getMemorySize() / sizeof(typename decltype(ToInt)::value_type);
+  }
+
+  auto begin() const noexcept { return FromInt.begin(); }
+  auto end() const noexcept { return FromInt.end(); }
+
+private:
+  llvm::DenseMap<T, uint32_t> ToInt;
+  llvm::SmallVector<T, 0> FromInt;
+};
+
+template <typename T>
+class Compressor<T, std::enable_if_t<!CanEfficientlyPassByValue<T>>> {
+public:
+  void reserve(size_t Capacity) {
+    assert(Capacity <= UINT32_MAX);
+    ToInt.reserve(Capacity);
+  }
+
+  uint32_t getOrInsert(const T &Elem) {
+    if (auto It = ToInt.find(&Elem); It != ToInt.end()) {
+      return It->second;
+    }
+    auto Ret = FromInt.size();
+    auto *Ins = &FromInt.emplace_back(Elem);
+    ToInt[Ins] = Ret;
+    return Ret;
+  }
+
+  uint32_t getOrInsert(T &&Elem) {
+    if (auto It = ToInt.find(&Elem); It != ToInt.end()) {
+      return It->second;
+    }
+    auto Ret = FromInt.size();
+    auto *Ins = &FromInt.emplace_back(std::move(Elem));
+    ToInt[Ins] = Ret;
+    return Ret;
+  }
+
+  std::optional<uint32_t> getOrNull(const T &Elem) const {
+    if (auto It = ToInt.find(&Elem); It != ToInt.end()) {
+      return It->second;
+    }
+    return std::nullopt;
+  }
+
+  const T &operator[](size_t Idx) const noexcept {
+    assert(Idx < FromInt.size());
+    return FromInt[Idx];
+  }
+
+  [[nodiscard]] size_t size() const noexcept { return FromInt.size(); }
+  [[nodiscard]] size_t capacity() const noexcept {
+    return FromInt.size() +
+           ToInt.getMemorySize() / sizeof(typename decltype(ToInt)::value_type);
+  }
+
+  auto begin() const noexcept { return FromInt.begin(); }
+  auto end() const noexcept { return FromInt.end(); }
+
+private:
+  struct DSI : llvm::DenseMapInfo<const T *> {
+    static auto getHashValue(const T *Elem) noexcept {
+      assert(Elem != nullptr);
+      if constexpr (has_llvm_dense_map_info<T>) {
+        return llvm::DenseMapInfo<T>::getHashValue(*Elem);
+      } else {
+        return std::hash<T>{}(*Elem);
+      }
+    }
+    static auto isEqual(const T *LHS, const T *RHS) noexcept {
+      if (LHS == RHS) {
+        return true;
+      }
+      if (LHS == DSI::getEmptyKey() || LHS == DSI::getTombstoneKey() ||
+          RHS == DSI::getEmptyKey() || RHS == DSI::getTombstoneKey()) {
+        return false;
+      }
+      if constexpr (has_llvm_dense_map_info<T>) {
+        return llvm::DenseMapInfo<T>::isEqual(*LHS, *RHS);
+      } else {
+        return *LHS == *RHS;
+      }
+    }
+  };
+
+  std::deque<T> FromInt;
+  llvm::DenseMap<const T *, uint32_t, DSI> ToInt;
+};
+
+struct NoneCompressor final {
+  constexpr NoneCompressor() noexcept = default;
+
+  template <typename T,
+            typename = std::enable_if_t<!std::is_same_v<NoneCompressor, T>>>
+  constexpr NoneCompressor(const T & /*unused*/) noexcept {}
+
+  template <typename T>
+  [[nodiscard]] decltype(auto) getOrInsert(T &&Val) const noexcept {
+    return std::forward<T>(Val);
+  }
+  template <typename T>
+  [[nodiscard]] decltype(auto) operator[](T &&Val) const noexcept {
+    return std::forward<T>(Val);
+  }
+  void reserve(size_t /*unused*/) const noexcept {}
+
+  [[nodiscard]] size_t size() const noexcept { return 0; }
+  [[nodiscard]] size_t capacity() const noexcept { return 0; }
+};
+
+class LLVMProjectIRDB;
+
+/// Once we have fast instruction IDs (as we already have in IntelliSecPhasar),
+/// we might want to create a specialization for T/const llvm::Value * that uses
+/// the IDs from the IRDB
+template <typename T> struct NodeCompressorTraits {
+  using type = Compressor<T>;
+
+  static type create(const ProjectIRDBBase<LLVMProjectIRDB>
+                         * /*IRDB*/) noexcept(noexcept(type())) {
+    return type();
+  }
+};
+
+template <typename T, typename = void> struct ValCompressorTraits {
+  using type = Compressor<T>;
+  using id_type = uint32_t;
+};
+
+template <typename T>
+struct ValCompressorTraits<T, std::enable_if_t<CanEfficientlyPassByValue<T>>> {
+  using type = NoneCompressor;
+  using id_type = T;
+};
+
+} // namespace psr
+
+#endif // PHASAR_DATAFLOW_IFDSIDE_SOLVER_COMPRESSOR_H
diff --git a/include/phasar/DataFlow/IfdsIde/Solver/ESGEdgeKind.h b/include/phasar/DataFlow/IfdsIde/Solver/ESGEdgeKind.h
@@ -1,3 +1,6 @@
+#ifndef PHASAR_DATAFLOW_IFDSIDE_SOLVER_ESGEDGEKIND_H
+#define PHASAR_DATAFLOW_IFDSIDE_SOLVER_ESGEDGEKIND_H
+
 /******************************************************************************
  * Copyright (c) 2022 Philipp Schubert.
  * All rights reserved. This program and the accompanying materials are made
@@ -7,9 +10,6 @@
  *     Fabian Schiebel and others
  *****************************************************************************/
 
-#ifndef PHASAR_DATAFLOW_IFDSIDE_SOLVER_ESGEDGEKIND_H
-#define PHASAR_DATAFLOW_IFDSIDE_SOLVER_ESGEDGEKIND_H
-
 namespace psr {
 enum class ESGEdgeKind { Normal, Call, CallToRet, SkipUnknownFn, Ret, Summary };