Merge branch 'development' into f-FilteredAliasSet

secure-software-engineering · May 10, 2024 · 51ed40e · 51ed40e
2 parents 51a17e0 + 488da6a
commit 51ed40e
Show file tree

Hide file tree

Showing 36 changed files with 168 additions and 122 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -293,8 +293,12 @@ else()
 endif()
 
 # SQL
-find_path(SQLITE3_INCLUDE_DIR NAMES sqlite3.h)
-find_library(SQLITE3_LIBRARY NAMES sqlite3)
+find_package(SQLite3)
+if(SQLite3_FOUND)
+  set(PHASAR_HAS_SQLITE ON)
+else()
+  set(PHASAR_HAS_SQLITE OFF)
+endif()
 
 option(USE_LLVM_FAT_LIB  "Link against libLLVM.so instead of the individual LLVM libraries if possible (default is OFF; always on if BUILD_SHARED_LIBS is ON)" OFF)
 

diff --git a/bootstrap.sh b/bootstrap.sh
@@ -108,7 +108,7 @@ set -- "${POSITIONAL[@]}" # restore positional parameters
 
 echo "installing phasar dependencies..."
 if [ -x "$(command -v pacman)" ]; then
-    yes | sudo pacman -Syu --needed which zlib sqlite3 python3 doxygen gcc python-pip ninja cmake
+    yes | sudo pacman -Syu --needed which zlib python3 doxygen gcc ninja cmake
 else
     ./utils/InstallAptDependencies.sh
 fi

diff --git a/cmake/phasar_macros.cmake b/cmake/phasar_macros.cmake
@@ -76,6 +76,13 @@ function(generate_ll_file)
   set(GEN_C_FLAGS -fno-discard-value-names -emit-llvm -S -w)
   set(GEN_CMD_COMMENT "[LL]")
 
+  if (CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL 15)
+    list(APPEND GEN_CXX_FLAGS -Xclang -no-opaque-pointers)
+  endif()
+  if (CMAKE_C_COMPILER_VERSION VERSION_GREATER_EQUAL 15)
+    list(APPEND GEN_C_FLAGS -Xclang -no-opaque-pointers)
+  endif()
+
   if(GEN_LL_MEM2REG)
     list(APPEND GEN_CXX_FLAGS -Xclang -disable-O0-optnone)
     list(APPEND GEN_C_FLAGS -Xclang -disable-O0-optnone)
@@ -121,7 +128,7 @@ function(generate_ll_file)
     add_custom_command(
       OUTPUT ${test_code_ll_file}
       COMMAND ${GEN_CMD} ${test_code_file_path} -o ${test_code_ll_file}
-      COMMAND ${CMAKE_CXX_COMPILER_LAUNCHER} opt -mem2reg -S ${test_code_ll_file} -o ${test_code_ll_file}
+      COMMAND ${CMAKE_CXX_COMPILER_LAUNCHER} opt -mem2reg -S -opaque-pointers=0 ${test_code_ll_file} -o ${test_code_ll_file}
       COMMENT ${GEN_CMD_COMMENT}
       DEPENDS ${GEN_LL_FILE}
       VERBATIM

diff --git a/config.h.in b/config.h.in
@@ -9,4 +9,6 @@
 #cmakedefine DYNAMIC_LOG
 #cmakedefine BUILD_PHASAR_CLANG
 
+#cmakedefine PHASAR_HAS_SQLITE
+
 #endif /* PHASAR_CONFIG_CONFIG_H  */
diff --git a/include/phasar/DB.h b/include/phasar/DB.h
@@ -10,8 +10,12 @@
 #ifndef PHASAR_DB_H
 #define PHASAR_DB_H
 
-#include "phasar/DB/Hexastore.h"
+#include "phasar/Config/phasar-config.h"
 #include "phasar/DB/ProjectIRDBBase.h"
+
+#ifdef PHASAR_HAS_SQLITE
+#include "phasar/DB/Hexastore.h"
 #include "phasar/DB/Queries.h"
+#endif
 
 #endif // PHASAR_DB_H
diff --git a/include/phasar/DB/Hexastore.h b/include/phasar/DB/Hexastore.h
@@ -10,17 +10,20 @@
 #ifndef PHASAR_DB_HEXASTORE_H_
 #define PHASAR_DB_HEXASTORE_H_
 
-#include "phasar/DB/Queries.h"
+#include "phasar/Config/phasar-config.h"
+#ifndef PHASAR_HAS_SQLITE
+#error                                                                         \
+    "Hexastore requires SQLite3. Please install libsqlite3-dev and reconfigure PhASAR."
+#endif
 
 #include "llvm/Support/raw_ostream.h"
 
-#include "boost/format.hpp"
-#include "sqlite3.h"
-
 #include <array>
 #include <string>
 #include <vector>
 
+struct sqlite3;
+
 namespace psr {
 /**
  * @brief Holds the results of a query to the Hexastore.
@@ -51,6 +54,7 @@ struct HSResult {
            LHS.Object == RHS.Object;
   }
 };
+
 /**
  * A Hexastore is an efficient approach to store large graphs.
  * This approach is based on the paper "Database-Backed Program Analysis

diff --git a/include/phasar/DataFlow/Mono/Solver/InterMonoSolver.h b/include/phasar/DataFlow/Mono/Solver/InterMonoSolver.h
@@ -24,7 +24,6 @@
 #include <unordered_map>
 #include <unordered_set>
 #include <utility>
-#include <vector>
 
 namespace psr {
 
@@ -368,18 +367,8 @@ template <typename AnalysisDomainTy, unsigned K> class InterMonoSolver {
       }
       // Compute the data-flow facts using the respective kind of flows
       if (ICF->isCallSite(Src)) {
-        // Handle call flow(s)
-        if (!isIntraEdge(Edge)) {
-          // real call
-          for (auto &[Ctx, Facts] : Analysis[Src]) {
-            processCall(Edge); // TODO: decompose into processCall and
-                               // processCallToRet
-          }
-        } else {
-          // call-to-return
-          processCall(
-              Edge); // TODO: decompose into processCall and processCallToRet
-        }
+        // Handle call flow(s) and call-to-return flow
+        processCall(Edge);
       } else if (ICF->isExitInst(Src)) {
         // Handle return flow
         processExit(Edge);

diff --git a/include/phasar/PhasarLLVM/ControlFlow/LLVMBasedICFG.h b/include/phasar/PhasarLLVM/ControlFlow/LLVMBasedICFG.h
@@ -42,6 +42,7 @@
 namespace psr {
 class LLVMTypeHierarchy;
 class LLVMProjectIRDB;
+class Resolver;
 
 class LLVMBasedICFG;
 template <> struct CFGTraits<LLVMBasedICFG> : CFGTraits<LLVMBasedCFG> {};
@@ -87,6 +88,11 @@ class LLVMBasedICFG : public LLVMBasedCFG, public ICFGBase<LLVMBasedICFG> {
                          LLVMAliasInfoRef PT = nullptr,
                          Soundness S = Soundness::Soundy,
                          bool IncludeGlobals = true);
+  explicit LLVMBasedICFG(LLVMProjectIRDB *IRDB, Resolver &CGResolver,
+                         llvm::ArrayRef<std::string> EntryPoints = {},
+                         LLVMTypeHierarchy *TH = nullptr,
+                         Soundness S = Soundness::Soundy,
+                         bool IncludeGlobals = true);
 
   /// Creates an ICFG with an already given call-graph
   explicit LLVMBasedICFG(CallGraph<n_t, f_t> CG, LLVMProjectIRDB *IRDB,
@@ -157,6 +163,10 @@ class LLVMBasedICFG : public LLVMBasedCFG, public ICFGBase<LLVMBasedICFG> {
   [[nodiscard]] llvm::Function *buildCRuntimeGlobalCtorsDtorsModel(
       llvm::Module &M, llvm::ArrayRef<llvm::Function *> UserEntryPoints);
 
+  void initialize(LLVMProjectIRDB *IRDB, Resolver &CGResolver,
+                  llvm::ArrayRef<std::string> EntryPoints,
+                  LLVMTypeHierarchy *TH, Soundness S, bool IncludeGlobals);
+
   // ---
 
   CallGraph<const llvm::Instruction *, const llvm::Function *> CG;

diff --git a/include/phasar/PhasarLLVM/ControlFlow/Resolver/OTFResolver.h b/include/phasar/PhasarLLVM/ControlFlow/Resolver/OTFResolver.h
@@ -35,16 +35,14 @@ class Value;
 
 namespace psr {
 
-class LLVMBasedICFG;
 class LLVMTypeHierarchy;
 
 class OTFResolver : public Resolver {
 protected:
-  LLVMBasedICFG &ICF;
   LLVMAliasInfoRef PT;
 
 public:
-  OTFResolver(LLVMProjectIRDB &IRDB, LLVMTypeHierarchy &TH, LLVMBasedICFG &ICF,
+  OTFResolver(LLVMProjectIRDB &IRDB, LLVMTypeHierarchy &TH,
               LLVMAliasInfoRef PT);
 
   ~OTFResolver() override = default;

diff --git a/include/phasar/PhasarLLVM/ControlFlow/Resolver/Resolver.h b/include/phasar/PhasarLLVM/ControlFlow/Resolver/Resolver.h
@@ -36,7 +36,6 @@ namespace psr {
 class LLVMProjectIRDB;
 class LLVMTypeHierarchy;
 enum class CallGraphAnalysisType;
-class LLVMBasedICFG;
 class LLVMPointsToInfo;
 
 [[nodiscard]] std::optional<unsigned>
@@ -83,9 +82,10 @@ class Resolver {
 
   [[nodiscard]] virtual std::string str() const = 0;
 
-  static std::unique_ptr<Resolver>
-  create(CallGraphAnalysisType Ty, LLVMProjectIRDB *IRDB, LLVMTypeHierarchy *TH,
-         LLVMBasedICFG *ICF = nullptr, LLVMAliasInfoRef PT = nullptr);
+  static std::unique_ptr<Resolver> create(CallGraphAnalysisType Ty,
+                                          LLVMProjectIRDB *IRDB,
+                                          LLVMTypeHierarchy *TH,
+                                          LLVMAliasInfoRef PT = nullptr);
 };
 } // namespace psr
 

diff --git a/include/phasar/PhasarLLVM/DataFlow/IfdsIde/Problems/IDEGeneralizedLCA/EdgeValue.h b/include/phasar/PhasarLLVM/DataFlow/IfdsIde/Problems/IDEGeneralizedLCA/EdgeValue.h
@@ -10,6 +10,7 @@
 #ifndef PHASAR_PHASARLLVM_DATAFLOW_IFDSIDE_PROBLEMS_IDEGENERALIZEDLCA_EDGEVALUE_H
 #define PHASAR_PHASARLLVM_DATAFLOW_IFDSIDE_PROBLEMS_IDEGENERALIZEDLCA_EDGEVALUE_H
 
+#include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/APSInt.h"
 #include "llvm/ADT/Twine.h"
 #include "llvm/IR/Constant.h"

diff --git a/lib/AnalysisStrategy/CMakeLists.txt b/lib/AnalysisStrategy/CMakeLists.txt
@@ -2,6 +2,5 @@ file(GLOB_RECURSE ANALYSIS_STRATEGY_SRC *.h *.cpp)
 
 add_phasar_library(phasar_analysis_strategy
   ${ANALYSIS_STRATEGY_SRC}
-  LINKS phasar_db phasar_controlflow
   LLVM_LINK_COMPONENTS Support
 )
diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt
@@ -18,29 +18,35 @@ if(PHASAR_BUILD_DYNLIB)
     set(PHASAR_DYNLIB_KIND SHARED)
 endif()
 
+set(PHASAR_LINK_LIBS
+    phasar_utils
+    phasar_passes
+    phasar_config
+    phasar_pointer
+    phasar_controlflow
+
+    phasar_llvm_utils
+    phasar_llvm_db
+    phasar_llvm_pointer
+    phasar_llvm_typehierarchy
+    phasar_llvm_controlflow
+
+    phasar_taintconfig
+    phasar_mono
+    phasar_llvm
+    phasar_llvm_ifdside
+    phasar_analysis_strategy
+    phasar_controller
+)
+if(SQLite3_FOUND)
+    list(APPEND PHASAR_LINK_LIBS phasar_db)
+endif()
+
 add_phasar_library(phasar ${PHASAR_DYNLIB_KIND}
     FILES
         LibPhasar.cpp
     LINKS
-        phasar_utils
-        phasar_passes
-        phasar_config
-        phasar_db
-        phasar_pointer
-        phasar_controlflow
-
-        phasar_llvm_utils
-        phasar_llvm_db
-        phasar_llvm_pointer
-        phasar_llvm_typehierarchy
-        phasar_llvm_controlflow
-
-        phasar_taintconfig
-        phasar_mono
-        phasar_llvm
-        phasar_llvm_ifdside
-        phasar_analysis_strategy
-        phasar_controller
+        ${PHASAR_LINK_LIBS}
     LINK_PRIVATE
         ${Boost_LIBRARIES}
     LLVM_LINK_COMPONENTS

diff --git a/lib/Controller/CMakeLists.txt b/lib/Controller/CMakeLists.txt
@@ -15,6 +15,7 @@ add_phasar_library(phasar_controller
     phasar_utils
     phasar_analysis_strategy
     phasar_taintconfig
+    phasar_passes
 
   LLVM_LINK_COMPONENTS
     Core

diff --git a/lib/DB/CMakeLists.txt b/lib/DB/CMakeLists.txt
@@ -1,12 +1,10 @@
-file(GLOB_RECURSE DB_SRC *.h *.cpp)
+if(SQLite3_FOUND)
+  file(GLOB_RECURSE DB_SRC *.h *.cpp)
 
-add_phasar_library(phasar_db
-  ${DB_SRC}
-  LINKS phasar_passes phasar_utils
-  LLVM_LINK_COMPONENTS Support
-  LINK_PRIVATE ${SQLITE3_LIBRARY}
-)
-
-target_include_directories(phasar_db
-  PRIVATE ${SQLITE3_INCLUDE_DIR}
-)
+  add_phasar_library(phasar_db
+    ${DB_SRC}
+    LINKS phasar_passes phasar_utils
+    LLVM_LINK_COMPONENTS Support
+    LINK_PRIVATE SQLite::SQLite3
+  )
+endif()
diff --git a/lib/DB/Hexastore.cpp b/lib/DB/Hexastore.cpp
@@ -9,11 +9,16 @@
 
 #include "phasar/DB/Hexastore.h"
 
+#include "phasar/DB/Queries.h"
+
+#include "boost/format.hpp"
+#include "sqlite3.h"
+
 namespace psr {
 
 Hexastore::Hexastore(const std::string &Filename) {
   sqlite3_open(Filename.c_str(), &HSInternalDB);
-  const std::string Query = INIT;
+  const std::string &Query = INIT;
   char *Err;
   sqlite3_exec(HSInternalDB, Query.c_str(), callback, nullptr, &Err);
   if (Err != nullptr) {

diff --git a/lib/PhasarLLVM/ControlFlow/CMakeLists.txt b/lib/PhasarLLVM/ControlFlow/CMakeLists.txt
@@ -7,7 +7,6 @@ add_phasar_library(phasar_llvm_controlflow
     phasar_llvm_pointer
     phasar_llvm_typehierarchy
     phasar_llvm_utils
-    phasar_db
     phasar_controlflow
 
   LLVM_LINK_COMPONENTS