1 //===--- ObjectFilePCHContainerOperations.cpp -----------------------------===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 9 #include "clang/CodeGen/ObjectFilePCHContainerOperations.h" 10 #include "CGDebugInfo.h" 11 #include "CodeGenModule.h" 12 #include "clang/AST/ASTContext.h" 13 #include "clang/AST/DeclObjC.h" 14 #include "clang/AST/Expr.h" 15 #include "clang/AST/RecursiveASTVisitor.h" 16 #include "clang/Basic/CodeGenOptions.h" 17 #include "clang/Basic/Diagnostic.h" 18 #include "clang/Basic/TargetInfo.h" 19 #include "clang/CodeGen/BackendUtil.h" 20 #include "clang/Frontend/CompilerInstance.h" 21 #include "clang/Lex/HeaderSearch.h" 22 #include "clang/Lex/Preprocessor.h" 23 #include "llvm/ADT/StringRef.h" 24 #include "llvm/Bitstream/BitstreamReader.h" 25 #include "llvm/DebugInfo/DWARF/DWARFContext.h" 26 #include "llvm/IR/Constants.h" 27 #include "llvm/IR/DataLayout.h" 28 #include "llvm/IR/LLVMContext.h" 29 #include "llvm/IR/Module.h" 30 #include "llvm/MC/TargetRegistry.h" 31 #include "llvm/Object/COFF.h" 32 #include "llvm/Object/ObjectFile.h" 33 #include "llvm/Support/Path.h" 34 #include <memory> 35 #include <utility> 36 37 using namespace clang; 38 39 #define DEBUG_TYPE "pchcontainer" 40 41 namespace { 42 class PCHContainerGenerator : public ASTConsumer { 43 DiagnosticsEngine &Diags; 44 const std::string MainFileName; 45 const std::string OutputFileName; 46 ASTContext *Ctx; 47 ModuleMap &MMap; 48 const HeaderSearchOptions &HeaderSearchOpts; 49 const PreprocessorOptions &PreprocessorOpts; 50 CodeGenOptions CodeGenOpts; 51 const TargetOptions TargetOpts; 52 LangOptions LangOpts; 53 std::unique_ptr<llvm::LLVMContext> VMContext; 54 std::unique_ptr<llvm::Module> M; 55 std::unique_ptr<CodeGen::CodeGenModule> Builder; 56 std::unique_ptr<raw_pwrite_stream> OS; 57 std::shared_ptr<PCHBuffer> Buffer; 58 59 /// Visit every type and emit debug info for it. 60 struct DebugTypeVisitor : public RecursiveASTVisitor<DebugTypeVisitor> { 61 clang::CodeGen::CGDebugInfo &DI; 62 ASTContext &Ctx; 63 DebugTypeVisitor(clang::CodeGen::CGDebugInfo &DI, ASTContext &Ctx) 64 : DI(DI), Ctx(Ctx) {} 65 66 /// Determine whether this type can be represented in DWARF. 67 static bool CanRepresent(const Type *Ty) { 68 return !Ty->isDependentType() && !Ty->isUndeducedType(); 69 } 70 71 bool VisitImportDecl(ImportDecl *D) { 72 if (!D->getImportedOwningModule()) 73 DI.EmitImportDecl(*D); 74 return true; 75 } 76 77 bool VisitTypeDecl(TypeDecl *D) { 78 // TagDecls may be deferred until after all decls have been merged and we 79 // know the complete type. Pure forward declarations will be skipped, but 80 // they don't need to be emitted into the module anyway. 81 if (auto *TD = dyn_cast<TagDecl>(D)) 82 if (!TD->isCompleteDefinition()) 83 return true; 84 85 QualType QualTy = Ctx.getTypeDeclType(D); 86 if (!QualTy.isNull() && CanRepresent(QualTy.getTypePtr())) 87 DI.getOrCreateStandaloneType(QualTy, D->getLocation()); 88 return true; 89 } 90 91 bool VisitObjCInterfaceDecl(ObjCInterfaceDecl *D) { 92 QualType QualTy(D->getTypeForDecl(), 0); 93 if (!QualTy.isNull() && CanRepresent(QualTy.getTypePtr())) 94 DI.getOrCreateStandaloneType(QualTy, D->getLocation()); 95 return true; 96 } 97 98 bool VisitFunctionDecl(FunctionDecl *D) { 99 if (isa<CXXMethodDecl>(D)) 100 // This is not yet supported. Constructing the `this' argument 101 // mandates a CodeGenFunction. 102 return true; 103 104 SmallVector<QualType, 16> ArgTypes; 105 for (auto i : D->parameters()) 106 ArgTypes.push_back(i->getType()); 107 QualType RetTy = D->getReturnType(); 108 QualType FnTy = Ctx.getFunctionType(RetTy, ArgTypes, 109 FunctionProtoType::ExtProtoInfo()); 110 if (CanRepresent(FnTy.getTypePtr())) 111 DI.EmitFunctionDecl(D, D->getLocation(), FnTy); 112 return true; 113 } 114 115 bool VisitObjCMethodDecl(ObjCMethodDecl *D) { 116 if (!D->getClassInterface()) 117 return true; 118 119 bool selfIsPseudoStrong, selfIsConsumed; 120 SmallVector<QualType, 16> ArgTypes; 121 ArgTypes.push_back(D->getSelfType(Ctx, D->getClassInterface(), 122 selfIsPseudoStrong, selfIsConsumed)); 123 ArgTypes.push_back(Ctx.getObjCSelType()); 124 for (auto i : D->parameters()) 125 ArgTypes.push_back(i->getType()); 126 QualType RetTy = D->getReturnType(); 127 QualType FnTy = Ctx.getFunctionType(RetTy, ArgTypes, 128 FunctionProtoType::ExtProtoInfo()); 129 if (CanRepresent(FnTy.getTypePtr())) 130 DI.EmitFunctionDecl(D, D->getLocation(), FnTy); 131 return true; 132 } 133 }; 134 135 public: 136 PCHContainerGenerator(CompilerInstance &CI, const std::string &MainFileName, 137 const std::string &OutputFileName, 138 std::unique_ptr<raw_pwrite_stream> OS, 139 std::shared_ptr<PCHBuffer> Buffer) 140 : Diags(CI.getDiagnostics()), MainFileName(MainFileName), 141 OutputFileName(OutputFileName), Ctx(nullptr), 142 MMap(CI.getPreprocessor().getHeaderSearchInfo().getModuleMap()), 143 HeaderSearchOpts(CI.getHeaderSearchOpts()), 144 PreprocessorOpts(CI.getPreprocessorOpts()), 145 TargetOpts(CI.getTargetOpts()), LangOpts(CI.getLangOpts()), 146 OS(std::move(OS)), Buffer(std::move(Buffer)) { 147 // The debug info output isn't affected by CodeModel and 148 // ThreadModel, but the backend expects them to be nonempty. 149 CodeGenOpts.CodeModel = "default"; 150 LangOpts.setThreadModel(LangOptions::ThreadModelKind::Single); 151 CodeGenOpts.DebugTypeExtRefs = true; 152 // When building a module MainFileName is the name of the modulemap file. 153 CodeGenOpts.MainFileName = 154 LangOpts.CurrentModule.empty() ? MainFileName : LangOpts.CurrentModule; 155 CodeGenOpts.setDebugInfo(codegenoptions::FullDebugInfo); 156 CodeGenOpts.setDebuggerTuning(CI.getCodeGenOpts().getDebuggerTuning()); 157 CodeGenOpts.DebugPrefixMap = 158 CI.getInvocation().getCodeGenOpts().DebugPrefixMap; 159 } 160 161 ~PCHContainerGenerator() override = default; 162 163 void Initialize(ASTContext &Context) override { 164 assert(!Ctx && "initialized multiple times"); 165 166 Ctx = &Context; 167 VMContext.reset(new llvm::LLVMContext()); 168 M.reset(new llvm::Module(MainFileName, *VMContext)); 169 M->setDataLayout(Ctx->getTargetInfo().getDataLayoutString()); 170 Builder.reset(new CodeGen::CodeGenModule( 171 *Ctx, HeaderSearchOpts, PreprocessorOpts, CodeGenOpts, *M, Diags)); 172 173 // Prepare CGDebugInfo to emit debug info for a clang module. 174 auto *DI = Builder->getModuleDebugInfo(); 175 StringRef ModuleName = llvm::sys::path::filename(MainFileName); 176 DI->setPCHDescriptor( 177 {ModuleName, "", OutputFileName, ASTFileSignature::createDISentinel()}); 178 DI->setModuleMap(MMap); 179 } 180 181 bool HandleTopLevelDecl(DeclGroupRef D) override { 182 if (Diags.hasErrorOccurred()) 183 return true; 184 185 // Collect debug info for all decls in this group. 186 for (auto *I : D) 187 if (!I->isFromASTFile()) { 188 DebugTypeVisitor DTV(*Builder->getModuleDebugInfo(), *Ctx); 189 DTV.TraverseDecl(I); 190 } 191 return true; 192 } 193 194 void HandleTopLevelDeclInObjCContainer(DeclGroupRef D) override { 195 HandleTopLevelDecl(D); 196 } 197 198 void HandleTagDeclDefinition(TagDecl *D) override { 199 if (Diags.hasErrorOccurred()) 200 return; 201 202 if (D->isFromASTFile()) 203 return; 204 205 // Anonymous tag decls are deferred until we are building their declcontext. 206 if (D->getName().empty()) 207 return; 208 209 // Defer tag decls until their declcontext is complete. 210 auto *DeclCtx = D->getDeclContext(); 211 while (DeclCtx) { 212 if (auto *D = dyn_cast<TagDecl>(DeclCtx)) 213 if (!D->isCompleteDefinition()) 214 return; 215 DeclCtx = DeclCtx->getParent(); 216 } 217 218 DebugTypeVisitor DTV(*Builder->getModuleDebugInfo(), *Ctx); 219 DTV.TraverseDecl(D); 220 Builder->UpdateCompletedType(D); 221 } 222 223 void HandleTagDeclRequiredDefinition(const TagDecl *D) override { 224 if (Diags.hasErrorOccurred()) 225 return; 226 227 if (const RecordDecl *RD = dyn_cast<RecordDecl>(D)) 228 Builder->getModuleDebugInfo()->completeRequiredType(RD); 229 } 230 231 void HandleImplicitImportDecl(ImportDecl *D) override { 232 if (!D->getImportedOwningModule()) 233 Builder->getModuleDebugInfo()->EmitImportDecl(*D); 234 } 235 236 /// Emit a container holding the serialized AST. 237 void HandleTranslationUnit(ASTContext &Ctx) override { 238 assert(M && VMContext && Builder); 239 // Delete these on function exit. 240 std::unique_ptr<llvm::LLVMContext> VMContext = std::move(this->VMContext); 241 std::unique_ptr<llvm::Module> M = std::move(this->M); 242 std::unique_ptr<CodeGen::CodeGenModule> Builder = std::move(this->Builder); 243 244 if (Diags.hasErrorOccurred()) 245 return; 246 247 M->setTargetTriple(Ctx.getTargetInfo().getTriple().getTriple()); 248 M->setDataLayout(Ctx.getTargetInfo().getDataLayoutString()); 249 250 // PCH files don't have a signature field in the control block, 251 // but LLVM detects DWO CUs by looking for a non-zero DWO id. 252 // We use the lower 64 bits for debug info. 253 254 uint64_t Signature = 255 Buffer->Signature ? Buffer->Signature.truncatedValue() : ~1ULL; 256 257 Builder->getModuleDebugInfo()->setDwoId(Signature); 258 259 // Finalize the Builder. 260 if (Builder) 261 Builder->Release(); 262 263 // Ensure the target exists. 264 std::string Error; 265 auto Triple = Ctx.getTargetInfo().getTriple(); 266 if (!llvm::TargetRegistry::lookupTarget(Triple.getTriple(), Error)) 267 llvm::report_fatal_error(llvm::Twine(Error)); 268 269 // Emit the serialized Clang AST into its own section. 270 assert(Buffer->IsComplete && "serialization did not complete"); 271 auto &SerializedAST = Buffer->Data; 272 auto Size = SerializedAST.size(); 273 274 if (Triple.isOSBinFormatWasm()) { 275 // Emit __clangast in custom section instead of named data segment 276 // to find it while iterating sections. 277 // This could be avoided if all data segements (the wasm sense) were 278 // represented as their own sections (in the llvm sense). 279 // TODO: https://github.com/WebAssembly/tool-conventions/issues/138 280 llvm::NamedMDNode *MD = 281 M->getOrInsertNamedMetadata("wasm.custom_sections"); 282 llvm::Metadata *Ops[2] = { 283 llvm::MDString::get(*VMContext, "__clangast"), 284 llvm::MDString::get(*VMContext, 285 StringRef(SerializedAST.data(), Size))}; 286 auto *NameAndContent = llvm::MDTuple::get(*VMContext, Ops); 287 MD->addOperand(NameAndContent); 288 } else { 289 auto Int8Ty = llvm::Type::getInt8Ty(*VMContext); 290 auto *Ty = llvm::ArrayType::get(Int8Ty, Size); 291 auto *Data = llvm::ConstantDataArray::getString( 292 *VMContext, StringRef(SerializedAST.data(), Size), 293 /*AddNull=*/false); 294 auto *ASTSym = new llvm::GlobalVariable( 295 *M, Ty, /*constant*/ true, llvm::GlobalVariable::InternalLinkage, 296 Data, "__clang_ast"); 297 // The on-disk hashtable needs to be aligned. 298 ASTSym->setAlignment(llvm::Align(8)); 299 300 // Mach-O also needs a segment name. 301 if (Triple.isOSBinFormatMachO()) 302 ASTSym->setSection("__CLANG,__clangast"); 303 // COFF has an eight character length limit. 304 else if (Triple.isOSBinFormatCOFF()) 305 ASTSym->setSection("clangast"); 306 else 307 ASTSym->setSection("__clangast"); 308 } 309 310 LLVM_DEBUG({ 311 // Print the IR for the PCH container to the debug output. 312 llvm::SmallString<0> Buffer; 313 clang::EmitBackendOutput( 314 Diags, HeaderSearchOpts, CodeGenOpts, TargetOpts, LangOpts, 315 Ctx.getTargetInfo().getDataLayoutString(), M.get(), 316 BackendAction::Backend_EmitLL, 317 std::make_unique<llvm::raw_svector_ostream>(Buffer)); 318 llvm::dbgs() << Buffer; 319 }); 320 321 // Use the LLVM backend to emit the pch container. 322 clang::EmitBackendOutput(Diags, HeaderSearchOpts, CodeGenOpts, TargetOpts, 323 LangOpts, 324 Ctx.getTargetInfo().getDataLayoutString(), M.get(), 325 BackendAction::Backend_EmitObj, std::move(OS)); 326 327 // Free the memory for the temporary buffer. 328 llvm::SmallVector<char, 0> Empty; 329 SerializedAST = std::move(Empty); 330 } 331 }; 332 333 } // anonymous namespace 334 335 std::unique_ptr<ASTConsumer> 336 ObjectFilePCHContainerWriter::CreatePCHContainerGenerator( 337 CompilerInstance &CI, const std::string &MainFileName, 338 const std::string &OutputFileName, 339 std::unique_ptr<llvm::raw_pwrite_stream> OS, 340 std::shared_ptr<PCHBuffer> Buffer) const { 341 return std::make_unique<PCHContainerGenerator>( 342 CI, MainFileName, OutputFileName, std::move(OS), Buffer); 343 } 344 345 StringRef 346 ObjectFilePCHContainerReader::ExtractPCH(llvm::MemoryBufferRef Buffer) const { 347 StringRef PCH; 348 auto OFOrErr = llvm::object::ObjectFile::createObjectFile(Buffer); 349 if (OFOrErr) { 350 auto &OF = OFOrErr.get(); 351 bool IsCOFF = isa<llvm::object::COFFObjectFile>(*OF); 352 // Find the clang AST section in the container. 353 for (auto &Section : OF->sections()) { 354 StringRef Name; 355 if (Expected<StringRef> NameOrErr = Section.getName()) 356 Name = *NameOrErr; 357 else 358 consumeError(NameOrErr.takeError()); 359 360 if ((!IsCOFF && Name == "__clangast") || (IsCOFF && Name == "clangast")) { 361 if (Expected<StringRef> E = Section.getContents()) 362 return *E; 363 else { 364 handleAllErrors(E.takeError(), [&](const llvm::ErrorInfoBase &EIB) { 365 EIB.log(llvm::errs()); 366 }); 367 return ""; 368 } 369 } 370 } 371 } 372 handleAllErrors(OFOrErr.takeError(), [&](const llvm::ErrorInfoBase &EIB) { 373 if (EIB.convertToErrorCode() == 374 llvm::object::object_error::invalid_file_type) 375 // As a fallback, treat the buffer as a raw AST. 376 PCH = Buffer.getBuffer(); 377 else 378 EIB.log(llvm::errs()); 379 }); 380 return PCH; 381 } 382