[interp] Register runtime symbols for clang-repl

lib/CppInterOp/CppInterOp.cpp

@@ -90,6 +90,34 @@
 #include <unistd.h>
 #endif // WIN32
 
+//  Runtime symbols required if the library using JIT (Cpp::Evaluate) does not
+//  link to llvm
+#if !defined(CPPINTEROP_USE_CLING) && !defined(EMSCRIPTEN)
+struct __clang_Interpreter_NewTag {
+} __ci_newtag;
+#if CLANG_VERSION_MAJOR > 21
+extern "C" void* __clang_Interpreter_SetValueWithAlloc(void* This, void* OutVal,
+                                                       void* OpaqueType)
+#else
+void* __clang_Interpreter_SetValueWithAlloc(void* This, void* OutVal,
+                                            void* OpaqueType);
+#endif
+
+#if CLANG_VERSION_MAJOR > 18
+    extern "C" void __clang_Interpreter_SetValueNoAlloc(void* This,
+                                                        void* OutVal,
+                                                        void* OpaqueType, ...);
+#else
+void __clang_Interpreter_SetValueNoAlloc(void*, void*, void*);
+void __clang_Interpreter_SetValueNoAlloc(void*, void*, void*, void*);
+void __clang_Interpreter_SetValueNoAlloc(void*, void*, void*, float);
+void __clang_Interpreter_SetValueNoAlloc(void*, void*, void*, double);
+void __clang_Interpreter_SetValueNoAlloc(void*, void*, void*, long double);
+void __clang_Interpreter_SetValueNoAlloc(void*, void*, void*,
+                                         unsigned long long);
+#endif
+#endif // CPPINTEROP_USE_CLING
+
 namespace Cpp {
 
 using namespace clang;
@@ -3330,6 +3358,30 @@
 }
 
 namespace {
+#if !defined(CPPINTEROP_USE_CLING) && !defined(EMSCRIPTEN)
+bool DefineAbsoluteSymbol(compat::Interpreter& I,
+                          const char* linker_mangled_name, uint64_t address) {
+  using namespace llvm;
+  using namespace llvm::orc;
+
+  llvm::orc::LLJIT& Jit = *compat::getExecutionEngine(I);
+  llvm::orc::ExecutionSession& ES = Jit.getExecutionSession();
+  JITDylib& DyLib = *Jit.getProcessSymbolsJITDylib().get();
+
+  llvm::orc::SymbolMap InjectedSymbols{
+      {ES.intern(linker_mangled_name), ExecutorSymbolDef(ExecutorAddr(address),
+                               JITSymbolFlags::Exported)}
+  };
+
+  if (Error Err = DyLib.define(absoluteSymbols(InjectedSymbols))) {
+    logAllUnhandledErrors(std::move(Err), errs(),
+                          "DefineAbsoluteSymbol error: ");
+    return true;
+  }
+  return false;
+}
+#endif
+
 static std::string MakeResourcesPath() {
   StringRef Dir;
 #ifdef LLVM_BINARY_DIR
@@ -3442,6 +3494,85 @@
        sInterpreters->back().get()});
 
   assert(sInterpreters->size() == sInterpreterASTMap->size());
+
+// Define runtime symbols in the JIT dylib for clang-repl
+#if !defined(CPPINTEROP_USE_CLING) && !defined(EMSCRIPTEN)
+  DefineAbsoluteSymbol(*I, "__ci_newtag", (uint64_t)&__ci_newtag);
+// llvm > 22 has this defined as a C symbol that does not require mangling
+#if CLANG_VERSION_MAJOR >= 22
+  DefineAbsoluteSymbol(*I, "__clang_Interpreter_SetValueWithAlloc",
+                       (uint64_t)&__clang_Interpreter_SetValueWithAlloc);
+#else
+  // obtain mangled name
+  auto* D = static_cast<clang::Decl*>(
+      Cpp::GetNamed("__clang_Interpreter_SetValueWithAlloc"));
+  if (auto* FD = llvm::dyn_cast<FunctionDecl>(D)) {
+    auto GD = GlobalDecl(FD);
+    std::string mangledName;
+    compat::maybeMangleDeclName(GD, mangledName);
+    DefineAbsoluteSymbol(*I, mangledName.c_str(),
+                         (uint64_t)&__clang_Interpreter_SetValueWithAlloc);
+  }
+#endif
+// llvm < 19 has multiple overloads of __clang_Interpreter_SetValueNoAlloc
+#if CLANG_VERSION_MAJOR < 19
+  // obtain all 6 candidates, and obtain the correct Decl for each overload
+  // using BestOverloadFunctionMatch. We then map the decl to the correct
+  // function pointer (force the compiler to find the right declarion by casting
+  // to the corresponding function pointer signature) and then register it.
+  const std::vector<TCppFunction_t> Methods = Cpp::GetFunctionsUsingName(
+      Cpp::GetGlobalScope(), "__clang_Interpreter_SetValueNoAlloc");
+  std::string mangledName;
+  ASTContext& Ctxt = I->getSema().getASTContext();
+  auto* TAI = Ctxt.VoidPtrTy.getAsOpaquePtr();
+
+  // possible parameter lists for __clang_Interpreter_SetValueNoAlloc overloads
+  // in LLVM 18
+  const std::vector<std::vector<Cpp::TemplateArgInfo>> a_params = {
+      {TAI, TAI, TAI},
+      {TAI, TAI, TAI, TAI},
+      {TAI, TAI, TAI, Ctxt.FloatTy.getAsOpaquePtr()},
+      {TAI, TAI, TAI, Ctxt.DoubleTy.getAsOpaquePtr()},
+      {TAI, TAI, TAI, Ctxt.LongDoubleTy.getAsOpaquePtr()},
+      {TAI, TAI, TAI, Ctxt.UnsignedLongLongTy.getAsOpaquePtr()}};
+
+  using FP0 = void (*)(void*, void*, void*);
+  using FP1 = void (*)(void*, void*, void*, void*);
+  using FP2 = void (*)(void*, void*, void*, float);
+  using FP3 = void (*)(void*, void*, void*, double);
+  using FP4 = void (*)(void*, void*, void*, long double);
+  using FP5 = void (*)(void*, void*, void*, unsigned long long);
+
+  const std::vector<void*> func_pointers = {
+      reinterpret_cast<void*>(
+          static_cast<FP0>(&__clang_Interpreter_SetValueNoAlloc)),
+      reinterpret_cast<void*>(
+          static_cast<FP1>(&__clang_Interpreter_SetValueNoAlloc)),
+      reinterpret_cast<void*>(
+          static_cast<FP2>(&__clang_Interpreter_SetValueNoAlloc)),
+      reinterpret_cast<void*>(
+          static_cast<FP3>(&__clang_Interpreter_SetValueNoAlloc)),
+      reinterpret_cast<void*>(
+          static_cast<FP4>(&__clang_Interpreter_SetValueNoAlloc)),
+      reinterpret_cast<void*>(
+          static_cast<FP5>(&__clang_Interpreter_SetValueNoAlloc))};
+
+  // these symbols are not externed, so we need to mangle their names
+  for (size_t i = 0; i < a_params.size(); ++i) {
+    auto* decl = static_cast<clang::Decl*>(
+        Cpp::BestOverloadFunctionMatch(Methods, {}, a_params[i]));
+    if (auto* fd = llvm::dyn_cast<clang::FunctionDecl>(decl)) {
+      auto gd = clang::GlobalDecl(fd);
+      compat::maybeMangleDeclName(gd, mangledName);
+      DefineAbsoluteSymbol(*I, mangledName.c_str(),
+                           reinterpret_cast<uint64_t>(func_pointers[i]));
+    }
+  }
+#else
+  DefineAbsoluteSymbol(*I, "__clang_Interpreter_SetValueNoAlloc",
+                       (uint64_t)&__clang_Interpreter_SetValueNoAlloc);
+#endif
+#endif
   return I;
 }
 

unittests/CppInterOp/InterpreterTest.cpp

@@ -84,6 +84,38 @@ TEST(InterpreterTest, Evaluate) {
   EXPECT_FALSE(HadError) ;
 }
 
+TEST(InterpreterTest, EvaluateExtensive) {
+#ifdef EMSCRIPTEN
+  GTEST_SKIP() << "Test fails for Emscipten builds";
+#endif
+#ifdef _WIN32
+  GTEST_SKIP() << "Disabled on Windows. Needs fixing.";
+#endif
+  if (llvm::sys::RunningOnValgrind())
+    GTEST_SKIP() << "XFAIL due to Valgrind report";
+  Cpp::CreateInterpreter();
+  EXPECT_TRUE(Cpp::Evaluate("__cplusplus") == 201402);
+
+  bool HadError;
+  EXPECT_TRUE(Cpp::Evaluate("#error", &HadError) == (intptr_t)~0UL);
+  EXPECT_TRUE(HadError);
+// for llvm < 19 this tests all different overloads of __clang_Interpreter_SetValueNoAlloc
+  EXPECT_EQ(Cpp::Evaluate("int i = 11; ++i", &HadError), 12);
+  EXPECT_FALSE(HadError) ;
+  EXPECT_EQ(Cpp::Evaluate("double a = 12.; a", &HadError), 12.);
+  EXPECT_FALSE(HadError) ;
+  EXPECT_EQ(Cpp::Evaluate("float b = 13.; b", &HadError), 13.);
+  EXPECT_FALSE(HadError) ;
+  EXPECT_EQ(Cpp::Evaluate("long double c = 14.; c", &HadError), 14.);
+  EXPECT_FALSE(HadError) ;
+  EXPECT_EQ(Cpp::Evaluate("long double d = 15.; d", &HadError), 15.);
+  EXPECT_FALSE(HadError);
+  EXPECT_EQ(Cpp::Evaluate("unsigned long long e = 16; e", &HadError), 16);
+  EXPECT_FALSE(HadError) ;
+  EXPECT_NE(Cpp::Evaluate("struct S{} s; s", &HadError), (intptr_t)~0UL);
+  EXPECT_FALSE(HadError) ;
+}
+
 TEST(InterpreterTest, DeleteInterpreter) {
   auto* I1 = Cpp::CreateInterpreter();
   auto* I2 = Cpp::CreateInterpreter();