1 //===-- SymbolFilePDB.cpp ---------------------------------------*- C++ -*-===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 10 #include "SymbolFilePDB.h" 11 12 #include "clang/Lex/Lexer.h" 13 14 #include "lldb/Core/Module.h" 15 #include "lldb/Core/PluginManager.h" 16 #include "lldb/Symbol/ClangASTContext.h" 17 #include "lldb/Symbol/CompileUnit.h" 18 #include "lldb/Symbol/LineTable.h" 19 #include "lldb/Symbol/ObjectFile.h" 20 #include "lldb/Symbol/SymbolContext.h" 21 #include "lldb/Symbol/SymbolVendor.h" 22 #include "lldb/Symbol/TypeMap.h" 23 #include "lldb/Symbol/TypeList.h" 24 #include "lldb/Utility/RegularExpression.h" 25 26 #include "llvm/DebugInfo/PDB/GenericError.h" 27 #include "llvm/DebugInfo/PDB/IPDBDataStream.h" 28 #include "llvm/DebugInfo/PDB/IPDBEnumChildren.h" 29 #include "llvm/DebugInfo/PDB/IPDBLineNumber.h" 30 #include "llvm/DebugInfo/PDB/IPDBSourceFile.h" 31 #include "llvm/DebugInfo/PDB/IPDBTable.h" 32 #include "llvm/DebugInfo/PDB/PDBSymbol.h" 33 #include "llvm/DebugInfo/PDB/PDBSymbolBlock.h" 34 #include "llvm/DebugInfo/PDB/PDBSymbolCompiland.h" 35 #include "llvm/DebugInfo/PDB/PDBSymbolCompilandDetails.h" 36 #include "llvm/DebugInfo/PDB/PDBSymbolData.h" 37 #include "llvm/DebugInfo/PDB/PDBSymbolExe.h" 38 #include "llvm/DebugInfo/PDB/PDBSymbolFunc.h" 39 #include "llvm/DebugInfo/PDB/PDBSymbolFuncDebugEnd.h" 40 #include "llvm/DebugInfo/PDB/PDBSymbolFuncDebugStart.h" 41 #include "llvm/DebugInfo/PDB/PDBSymbolPublicSymbol.h" 42 #include "llvm/DebugInfo/PDB/PDBSymbolTypeEnum.h" 43 #include "llvm/DebugInfo/PDB/PDBSymbolTypeTypedef.h" 44 #include "llvm/DebugInfo/PDB/PDBSymbolTypeUDT.h" 45 46 #include "Plugins/Language/CPlusPlus/CPlusPlusLanguage.h" 47 #include "Plugins/SymbolFile/PDB/PDBASTParser.h" 48 49 #include <regex> 50 51 using namespace lldb; 52 using namespace lldb_private; 53 using namespace llvm::pdb; 54 55 namespace { 56 lldb::LanguageType TranslateLanguage(PDB_Lang lang) { 57 switch (lang) { 58 case PDB_Lang::Cpp: 59 return lldb::LanguageType::eLanguageTypeC_plus_plus; 60 case PDB_Lang::C: 61 return lldb::LanguageType::eLanguageTypeC; 62 default: 63 return lldb::LanguageType::eLanguageTypeUnknown; 64 } 65 } 66 67 bool ShouldAddLine(uint32_t requested_line, uint32_t actual_line, 68 uint32_t addr_length) { 69 return ((requested_line == 0 || actual_line == requested_line) && 70 addr_length > 0); 71 } 72 } 73 74 void SymbolFilePDB::Initialize() { 75 PluginManager::RegisterPlugin(GetPluginNameStatic(), 76 GetPluginDescriptionStatic(), CreateInstance, 77 DebuggerInitialize); 78 } 79 80 void SymbolFilePDB::Terminate() { 81 PluginManager::UnregisterPlugin(CreateInstance); 82 } 83 84 void SymbolFilePDB::DebuggerInitialize(lldb_private::Debugger &debugger) {} 85 86 lldb_private::ConstString SymbolFilePDB::GetPluginNameStatic() { 87 static ConstString g_name("pdb"); 88 return g_name; 89 } 90 91 const char *SymbolFilePDB::GetPluginDescriptionStatic() { 92 return "Microsoft PDB debug symbol file reader."; 93 } 94 95 lldb_private::SymbolFile * 96 SymbolFilePDB::CreateInstance(lldb_private::ObjectFile *obj_file) { 97 return new SymbolFilePDB(obj_file); 98 } 99 100 SymbolFilePDB::SymbolFilePDB(lldb_private::ObjectFile *object_file) 101 : SymbolFile(object_file), m_session_up(), m_global_scope_up(), 102 m_cached_compile_unit_count(0), m_tu_decl_ctx_up() {} 103 104 SymbolFilePDB::~SymbolFilePDB() {} 105 106 uint32_t SymbolFilePDB::CalculateAbilities() { 107 uint32_t abilities = 0; 108 if (!m_obj_file) 109 return 0; 110 111 if (!m_session_up) { 112 // Lazily load and match the PDB file, but only do this once. 113 std::string exePath = m_obj_file->GetFileSpec().GetPath(); 114 auto error = loadDataForEXE(PDB_ReaderType::DIA, llvm::StringRef(exePath), 115 m_session_up); 116 if (error) { 117 llvm::consumeError(std::move(error)); 118 auto module_sp = m_obj_file->GetModule(); 119 if (!module_sp) 120 return 0; 121 // See if any symbol file is specified through `--symfile` option. 122 FileSpec symfile = module_sp->GetSymbolFileFileSpec(); 123 if (!symfile) 124 return 0; 125 error = loadDataForPDB(PDB_ReaderType::DIA, 126 llvm::StringRef(symfile.GetPath()), 127 m_session_up); 128 if (error) { 129 llvm::consumeError(std::move(error)); 130 return 0; 131 } 132 } 133 } 134 if (!m_session_up.get()) 135 return 0; 136 137 auto enum_tables_up = m_session_up->getEnumTables(); 138 if (!enum_tables_up) 139 return 0; 140 while (auto table_up = enum_tables_up->getNext()) { 141 if (table_up->getItemCount() == 0) 142 continue; 143 auto type = table_up->getTableType(); 144 switch (type) { 145 case PDB_TableType::Symbols: 146 // This table represents a store of symbols with types listed in 147 // PDBSym_Type 148 abilities |= (CompileUnits | Functions | Blocks | 149 GlobalVariables | LocalVariables | VariableTypes); 150 break; 151 case PDB_TableType::LineNumbers: 152 abilities |= LineTables; 153 break; 154 default: break; 155 } 156 } 157 return abilities; 158 } 159 160 void SymbolFilePDB::InitializeObject() { 161 lldb::addr_t obj_load_address = m_obj_file->GetFileOffset(); 162 lldbassert(obj_load_address && 163 obj_load_address != LLDB_INVALID_ADDRESS); 164 m_session_up->setLoadAddress(obj_load_address); 165 if (!m_global_scope_up) 166 m_global_scope_up = m_session_up->getGlobalScope(); 167 lldbassert(m_global_scope_up.get()); 168 169 TypeSystem *type_system = 170 GetTypeSystemForLanguage(lldb::eLanguageTypeC_plus_plus); 171 ClangASTContext *clang_type_system = 172 llvm::dyn_cast_or_null<ClangASTContext>(type_system); 173 lldbassert(clang_type_system); 174 m_tu_decl_ctx_up = llvm::make_unique<CompilerDeclContext>( 175 type_system, clang_type_system->GetTranslationUnitDecl()); 176 } 177 178 uint32_t SymbolFilePDB::GetNumCompileUnits() { 179 if (m_cached_compile_unit_count == 0) { 180 auto compilands = m_global_scope_up->findAllChildren<PDBSymbolCompiland>(); 181 if (!compilands) 182 return 0; 183 184 // The linker could link *.dll (compiland language = LINK), or import 185 // *.dll. For example, a compiland with name `Import:KERNEL32.dll` 186 // could be found as a child of the global scope (PDB executable). 187 // Usually, such compilands contain `thunk` symbols in which we are not 188 // interested for now. However we still count them in the compiland list. 189 // If we perform any compiland related activity, like finding symbols 190 // through llvm::pdb::IPDBSession methods, such compilands will all be 191 // searched automatically no matter whether we include them or not. 192 m_cached_compile_unit_count = compilands->getChildCount(); 193 194 // The linker can inject an additional "dummy" compilation unit into the 195 // PDB. Ignore this special compile unit for our purposes, if it is there. 196 // It is always the last one. 197 auto last_compiland_up = 198 compilands->getChildAtIndex(m_cached_compile_unit_count - 1); 199 lldbassert(last_compiland_up.get()); 200 std::string name = last_compiland_up->getName(); 201 if (name == "* Linker *") 202 --m_cached_compile_unit_count; 203 } 204 return m_cached_compile_unit_count; 205 } 206 207 void SymbolFilePDB::GetCompileUnitIndex( 208 const llvm::pdb::PDBSymbolCompiland *pdb_compiland, 209 uint32_t &index) { 210 if (!pdb_compiland) 211 return; 212 213 auto results_up = m_global_scope_up->findAllChildren<PDBSymbolCompiland>(); 214 if (!results_up) 215 return; 216 auto uid = pdb_compiland->getSymIndexId(); 217 for (uint32_t cu_idx = 0; cu_idx < GetNumCompileUnits(); ++cu_idx) { 218 auto compiland_up = results_up->getChildAtIndex(cu_idx); 219 if (!compiland_up) 220 continue; 221 if (compiland_up->getSymIndexId() == uid) { 222 index = cu_idx; 223 return; 224 } 225 } 226 index = UINT32_MAX; 227 return; 228 } 229 230 std::unique_ptr<llvm::pdb::PDBSymbolCompiland> 231 SymbolFilePDB::GetPDBCompilandByUID(uint32_t uid) { 232 return m_session_up->getConcreteSymbolById<PDBSymbolCompiland>(uid); 233 } 234 235 lldb::CompUnitSP SymbolFilePDB::ParseCompileUnitAtIndex(uint32_t index) { 236 if (index >= GetNumCompileUnits()) 237 return CompUnitSP(); 238 239 // Assuming we always retrieve same compilands listed in same order through 240 // `PDBSymbolExe::findAllChildren` method, otherwise using `index` to get a 241 // compile unit makes no sense. 242 auto results = m_global_scope_up->findAllChildren<PDBSymbolCompiland>(); 243 if (!results) 244 return CompUnitSP(); 245 auto compiland_up = results->getChildAtIndex(index); 246 if (!compiland_up) 247 return CompUnitSP(); 248 return ParseCompileUnitForUID(compiland_up->getSymIndexId(), index); 249 } 250 251 lldb::LanguageType 252 SymbolFilePDB::ParseCompileUnitLanguage(const lldb_private::SymbolContext &sc) { 253 // What fields should I expect to be filled out on the SymbolContext? Is it 254 // safe to assume that `sc.comp_unit` is valid? 255 if (!sc.comp_unit) 256 return lldb::eLanguageTypeUnknown; 257 258 auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID()); 259 if (!compiland_up) 260 return lldb::eLanguageTypeUnknown; 261 auto details = compiland_up->findOneChild<PDBSymbolCompilandDetails>(); 262 if (!details) 263 return lldb::eLanguageTypeUnknown; 264 return TranslateLanguage(details->getLanguage()); 265 } 266 267 lldb_private::Function * 268 SymbolFilePDB::ParseCompileUnitFunctionForPDBFunc( 269 const PDBSymbolFunc *pdb_func, 270 const lldb_private::SymbolContext &sc) { 271 assert(pdb_func != nullptr); 272 lldbassert(sc.comp_unit && sc.module_sp.get()); 273 274 auto file_vm_addr = pdb_func->getVirtualAddress(); 275 if (file_vm_addr == LLDB_INVALID_ADDRESS) 276 return nullptr; 277 278 auto func_length = pdb_func->getLength(); 279 AddressRange func_range = AddressRange(file_vm_addr, 280 func_length, 281 sc.module_sp->GetSectionList()); 282 if (!func_range.GetBaseAddress().IsValid()) 283 return nullptr; 284 285 user_id_t func_type_uid = pdb_func->getSignatureId(); 286 // TODO: Function symbol with invalid signature won't be handled. We'll set up 287 // a white list to trace them. 288 if (!pdb_func->getSignature()) 289 return nullptr; 290 291 lldb_private::Type* func_type = ResolveTypeUID(pdb_func->getSymIndexId()); 292 if (!func_type) 293 return nullptr; 294 295 Mangled mangled = GetMangledForPDBFunc(pdb_func); 296 297 FunctionSP func_sp = std::make_shared<Function>(sc.comp_unit, 298 pdb_func->getSymIndexId(), 299 func_type_uid, 300 mangled, 301 func_type, 302 func_range); 303 304 sc.comp_unit->AddFunction(func_sp); 305 return func_sp.get(); 306 } 307 308 size_t SymbolFilePDB::ParseCompileUnitFunctions( 309 const lldb_private::SymbolContext &sc) { 310 lldbassert(sc.comp_unit); 311 size_t func_added = 0; 312 auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID()); 313 if (!compiland_up) 314 return 0; 315 auto results_up = compiland_up->findAllChildren<PDBSymbolFunc>(); 316 if (!results_up) 317 return 0; 318 while (auto pdb_func_up = results_up->getNext()) { 319 auto func_sp = 320 sc.comp_unit->FindFunctionByUID(pdb_func_up->getSymIndexId()); 321 if (!func_sp) { 322 if (ParseCompileUnitFunctionForPDBFunc(pdb_func_up.get(), sc)) 323 ++func_added; 324 } 325 } 326 return func_added; 327 } 328 329 bool SymbolFilePDB::ParseCompileUnitLineTable( 330 const lldb_private::SymbolContext &sc) { 331 lldbassert(sc.comp_unit); 332 if (sc.comp_unit->GetLineTable()) 333 return true; 334 return ParseCompileUnitLineTable(sc, 0); 335 } 336 337 bool SymbolFilePDB::ParseCompileUnitDebugMacros( 338 const lldb_private::SymbolContext &sc) { 339 // PDB doesn't contain information about macros 340 return false; 341 } 342 343 bool SymbolFilePDB::ParseCompileUnitSupportFiles( 344 const lldb_private::SymbolContext &sc, 345 lldb_private::FileSpecList &support_files) { 346 lldbassert(sc.comp_unit); 347 348 // In theory this is unnecessary work for us, because all of this information 349 // is easily (and quickly) accessible from DebugInfoPDB, so caching it a 350 // second time seems like a waste. Unfortunately, there's no good way around 351 // this short of a moderate refactor since SymbolVendor depends on being able 352 // to cache this list. 353 auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID()); 354 if (!compiland_up) 355 return false; 356 auto files = m_session_up->getSourceFilesForCompiland(*compiland_up); 357 if (!files || files->getChildCount() == 0) 358 return false; 359 360 while (auto file = files->getNext()) { 361 FileSpec spec(file->getFileName(), false, FileSpec::ePathSyntaxWindows); 362 support_files.AppendIfUnique(spec); 363 } 364 return true; 365 } 366 367 bool SymbolFilePDB::ParseImportedModules( 368 const lldb_private::SymbolContext &sc, 369 std::vector<lldb_private::ConstString> &imported_modules) { 370 // PDB does not yet support module debug info 371 return false; 372 } 373 374 static size_t 375 ParseFunctionBlocksForPDBSymbol(const lldb_private::SymbolContext &sc, 376 uint64_t func_file_vm_addr, 377 const llvm::pdb::PDBSymbol *pdb_symbol, 378 lldb_private::Block *parent_block, 379 bool is_top_parent) { 380 assert(pdb_symbol && parent_block); 381 382 size_t num_added = 0; 383 switch (pdb_symbol->getSymTag()) { 384 case PDB_SymType::Block: 385 case PDB_SymType::Function: { 386 Block *block = nullptr; 387 auto &raw_sym = pdb_symbol->getRawSymbol(); 388 if (auto *pdb_func = llvm::dyn_cast<PDBSymbolFunc>(pdb_symbol)) { 389 if (pdb_func->hasNoInlineAttribute()) 390 break; 391 if (is_top_parent) 392 block = parent_block; 393 else 394 break; 395 } else if (llvm::dyn_cast<PDBSymbolBlock>(pdb_symbol)) { 396 auto uid = pdb_symbol->getSymIndexId(); 397 if (parent_block->FindBlockByID(uid)) 398 break; 399 if (raw_sym.getVirtualAddress() < func_file_vm_addr) 400 break; 401 402 auto block_sp = std::make_shared<Block>(pdb_symbol->getSymIndexId()); 403 parent_block->AddChild(block_sp); 404 block = block_sp.get(); 405 } else 406 llvm_unreachable("Unexpected PDB symbol!"); 407 408 block->AddRange( 409 Block::Range(raw_sym.getVirtualAddress() - func_file_vm_addr, 410 raw_sym.getLength())); 411 block->FinalizeRanges(); 412 ++num_added; 413 414 auto results_up = pdb_symbol->findAllChildren(); 415 if (!results_up) 416 break; 417 while (auto symbol_up = results_up->getNext()) { 418 num_added += ParseFunctionBlocksForPDBSymbol(sc, func_file_vm_addr, 419 symbol_up.get(), 420 block, false); 421 } 422 } break; 423 default: break; 424 } 425 return num_added; 426 } 427 428 size_t 429 SymbolFilePDB::ParseFunctionBlocks(const lldb_private::SymbolContext &sc) { 430 lldbassert(sc.comp_unit && sc.function); 431 size_t num_added = 0; 432 auto uid = sc.function->GetID(); 433 auto pdb_func_up = m_session_up->getConcreteSymbolById<PDBSymbolFunc>(uid); 434 if (!pdb_func_up) 435 return 0; 436 Block &parent_block = sc.function->GetBlock(false); 437 num_added = 438 ParseFunctionBlocksForPDBSymbol(sc, pdb_func_up->getVirtualAddress(), 439 pdb_func_up.get(), &parent_block, true); 440 return num_added; 441 } 442 443 size_t SymbolFilePDB::ParseTypes(const lldb_private::SymbolContext &sc) { 444 lldbassert(sc.module_sp.get()); 445 size_t num_added = 0; 446 auto results_up = m_session_up->getGlobalScope()->findAllChildren(); 447 if (!results_up) 448 return 0; 449 while (auto symbol_up = results_up->getNext()) { 450 switch (symbol_up->getSymTag()) { 451 case PDB_SymType::Enum: 452 case PDB_SymType::UDT: 453 case PDB_SymType::Typedef: 454 break; 455 default: 456 continue; 457 } 458 459 auto type_uid = symbol_up->getSymIndexId(); 460 if (m_types.find(type_uid) != m_types.end()) 461 continue; 462 463 // This should cause the type to get cached and stored in the `m_types` 464 // lookup. 465 if (!ResolveTypeUID(symbol_up->getSymIndexId())) 466 continue; 467 468 ++num_added; 469 } 470 return num_added; 471 } 472 473 size_t 474 SymbolFilePDB::ParseVariablesForContext(const lldb_private::SymbolContext &sc) { 475 // TODO: Implement this 476 return size_t(); 477 } 478 479 lldb_private::Type *SymbolFilePDB::ResolveTypeUID(lldb::user_id_t type_uid) { 480 auto find_result = m_types.find(type_uid); 481 if (find_result != m_types.end()) 482 return find_result->second.get(); 483 484 TypeSystem *type_system = 485 GetTypeSystemForLanguage(lldb::eLanguageTypeC_plus_plus); 486 ClangASTContext *clang_type_system = 487 llvm::dyn_cast_or_null<ClangASTContext>(type_system); 488 if (!clang_type_system) 489 return nullptr; 490 PDBASTParser *pdb = 491 llvm::dyn_cast<PDBASTParser>(clang_type_system->GetPDBParser()); 492 if (!pdb) 493 return nullptr; 494 495 auto pdb_type = m_session_up->getSymbolById(type_uid); 496 if (pdb_type == nullptr) 497 return nullptr; 498 499 lldb::TypeSP result = pdb->CreateLLDBTypeFromPDBType(*pdb_type); 500 if (result.get()) { 501 m_types.insert(std::make_pair(type_uid, result)); 502 auto type_list = GetTypeList(); 503 type_list->Insert(result); 504 } 505 return result.get(); 506 } 507 508 bool SymbolFilePDB::CompleteType(lldb_private::CompilerType &compiler_type) { 509 // TODO: Implement this 510 return false; 511 } 512 513 lldb_private::CompilerDecl SymbolFilePDB::GetDeclForUID(lldb::user_id_t uid) { 514 return lldb_private::CompilerDecl(); 515 } 516 517 lldb_private::CompilerDeclContext 518 SymbolFilePDB::GetDeclContextForUID(lldb::user_id_t uid) { 519 // PDB always uses the translation unit decl context for everything. We can 520 // improve this later but it's not easy because PDB doesn't provide a high 521 // enough level of type fidelity in this area. 522 return *m_tu_decl_ctx_up; 523 } 524 525 lldb_private::CompilerDeclContext 526 SymbolFilePDB::GetDeclContextContainingUID(lldb::user_id_t uid) { 527 return *m_tu_decl_ctx_up; 528 } 529 530 void SymbolFilePDB::ParseDeclsForContext( 531 lldb_private::CompilerDeclContext decl_ctx) {} 532 533 uint32_t 534 SymbolFilePDB::ResolveSymbolContext(const lldb_private::Address &so_addr, 535 uint32_t resolve_scope, 536 lldb_private::SymbolContext &sc) { 537 uint32_t resolved_flags = 0; 538 if (resolve_scope & eSymbolContextCompUnit || 539 resolve_scope & eSymbolContextVariable || 540 resolve_scope & eSymbolContextFunction || 541 resolve_scope & eSymbolContextBlock || 542 resolve_scope & eSymbolContextLineEntry) { 543 addr_t file_vm_addr = so_addr.GetFileAddress(); 544 auto symbol_up = 545 m_session_up->findSymbolByAddress(file_vm_addr, PDB_SymType::None); 546 if (!symbol_up) 547 return 0; 548 549 auto cu_sp = GetCompileUnitContainsAddress(so_addr); 550 if (!cu_sp) { 551 if (resolved_flags | eSymbolContextVariable) { 552 // TODO: Resolve variables 553 } 554 return 0; 555 } 556 sc.comp_unit = cu_sp.get(); 557 resolved_flags |= eSymbolContextCompUnit; 558 lldbassert(sc.module_sp == cu_sp->GetModule()); 559 560 switch (symbol_up->getSymTag()) { 561 case PDB_SymType::Function: 562 if (resolve_scope & eSymbolContextFunction) { 563 auto *pdb_func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get()); 564 assert(pdb_func); 565 auto func_uid = pdb_func->getSymIndexId(); 566 sc.function = sc.comp_unit->FindFunctionByUID(func_uid).get(); 567 if (sc.function == nullptr) 568 sc.function = ParseCompileUnitFunctionForPDBFunc(pdb_func, sc); 569 if (sc.function) { 570 resolved_flags |= eSymbolContextFunction; 571 if (resolve_scope & eSymbolContextBlock) { 572 Block &block = sc.function->GetBlock(true); 573 sc.block = block.FindBlockByID(sc.function->GetID()); 574 if (sc.block) 575 resolved_flags |= eSymbolContextBlock; 576 } 577 } 578 } 579 break; 580 default: 581 break; 582 } 583 584 if (resolve_scope & eSymbolContextLineEntry) { 585 if (auto *line_table = sc.comp_unit->GetLineTable()) { 586 Address addr(so_addr); 587 if (line_table->FindLineEntryByAddress(addr, sc.line_entry)) 588 resolved_flags |= eSymbolContextLineEntry; 589 } 590 } 591 } 592 return resolved_flags; 593 } 594 595 std::string SymbolFilePDB::GetSourceFileNameForPDBCompiland( 596 const PDBSymbolCompiland *pdb_compiland) { 597 if (!pdb_compiland) 598 return std::string(); 599 600 std::string source_file_name; 601 // `getSourceFileName` returns the basename of the original source file 602 // used to generate this compiland. It does not return the full path. 603 // Currently the only way to get that is to do a basename lookup to get the 604 // IPDBSourceFile, but this is ambiguous in the case of two source files 605 // with the same name contributing to the same compiland. This is an edge 606 // case that we ignore for now, although we need to a long-term solution. 607 std::string file_name = pdb_compiland->getSourceFileName(); 608 if (!file_name.empty()) { 609 auto one_src_file_up = 610 m_session_up->findOneSourceFile(pdb_compiland, file_name, 611 PDB_NameSearchFlags::NS_CaseInsensitive); 612 if (one_src_file_up) 613 source_file_name = one_src_file_up->getFileName(); 614 } 615 // For some reason, source file name could be empty, so we will walk through 616 // all source files of this compiland, and determine the right source file 617 // if any that is used to generate this compiland based on language 618 // indicated in compilanddetails language field. 619 if (!source_file_name.empty()) 620 return source_file_name; 621 622 auto details_up = pdb_compiland->findOneChild<PDBSymbolCompilandDetails>(); 623 PDB_Lang pdb_lang = details_up ? details_up->getLanguage() : PDB_Lang::Cpp; 624 auto src_files_up = 625 m_session_up->getSourceFilesForCompiland(*pdb_compiland); 626 if (src_files_up) { 627 while (auto file_up = src_files_up->getNext()) { 628 FileSpec file_spec(file_up->getFileName(), false, 629 FileSpec::ePathSyntaxWindows); 630 auto file_extension = file_spec.GetFileNameExtension(); 631 if (pdb_lang == PDB_Lang::Cpp || pdb_lang == PDB_Lang::C) { 632 static const char* exts[] = { "cpp", "c", "cc", "cxx" }; 633 if (llvm::is_contained(exts, file_extension.GetStringRef().lower())) 634 source_file_name = file_up->getFileName(); 635 break; 636 } else if (pdb_lang == PDB_Lang::Masm && 637 ConstString::Compare(file_extension, ConstString("ASM"), 638 false) == 0) { 639 source_file_name = file_up->getFileName(); 640 break; 641 } 642 } 643 } 644 return source_file_name; 645 } 646 647 uint32_t SymbolFilePDB::ResolveSymbolContext( 648 const lldb_private::FileSpec &file_spec, uint32_t line, bool check_inlines, 649 uint32_t resolve_scope, lldb_private::SymbolContextList &sc_list) { 650 const size_t old_size = sc_list.GetSize(); 651 if (resolve_scope & lldb::eSymbolContextCompUnit) { 652 // Locate all compilation units with line numbers referencing the specified 653 // file. For example, if `file_spec` is <vector>, then this should return 654 // all source files and header files that reference <vector>, either 655 // directly or indirectly. 656 auto compilands = m_session_up->findCompilandsForSourceFile( 657 file_spec.GetPath(), PDB_NameSearchFlags::NS_CaseInsensitive); 658 659 if (!compilands) 660 return 0; 661 662 // For each one, either find its previously parsed data or parse it afresh 663 // and add it to the symbol context list. 664 while (auto compiland = compilands->getNext()) { 665 // If we're not checking inlines, then don't add line information for this 666 // file unless the FileSpec matches. 667 // For inline functions, we don't have to match the FileSpec since they 668 // could be defined in headers other than file specified in FileSpec. 669 if (!check_inlines) { 670 // `getSourceFileName` returns the basename of the original source file 671 // used to generate this compiland. It does not return the full path. 672 // Currently the only way to get that is to do a basename lookup to get 673 // the IPDBSourceFile, but this is ambiguous in the case of two source 674 // files with the same name contributing to the same compiland. This is 675 // a moderately extreme edge case, so we consider this OK for now, 676 // although we need to find a long-term solution. 677 std::string source_file = 678 GetSourceFileNameForPDBCompiland(compiland.get()); 679 if (source_file.empty()) 680 continue; 681 FileSpec this_spec(source_file, false, FileSpec::ePathSyntaxWindows); 682 bool need_full_match = !file_spec.GetDirectory().IsEmpty(); 683 if (FileSpec::Compare(file_spec, this_spec, need_full_match) != 0) 684 continue; 685 } 686 687 SymbolContext sc; 688 auto cu = ParseCompileUnitForUID(compiland->getSymIndexId()); 689 if (!cu.get()) 690 continue; 691 sc.comp_unit = cu.get(); 692 sc.module_sp = cu->GetModule(); 693 694 // If we were asked to resolve line entries, add all entries to the line 695 // table that match the requested line (or all lines if `line` == 0). 696 if (resolve_scope & (eSymbolContextFunction | eSymbolContextBlock | 697 eSymbolContextLineEntry)) { 698 bool has_line_table = ParseCompileUnitLineTable(sc, line); 699 700 if ((resolve_scope & eSymbolContextLineEntry) && !has_line_table) { 701 // The query asks for line entries, but we can't get them for the 702 // compile unit. This is not normal for `line` = 0. So just assert it. 703 if (line == 0) { 704 assert(0 && "Couldn't get all line entries!\n"); 705 } 706 707 // Current compiland does not have the requested line. Search next. 708 continue; 709 } 710 711 if (resolve_scope & (eSymbolContextFunction | eSymbolContextBlock)) { 712 if (!has_line_table) 713 continue; 714 715 auto *line_table = sc.comp_unit->GetLineTable(); 716 lldbassert(line_table); 717 718 uint32_t num_line_entries = line_table->GetSize(); 719 // Skip the terminal line entry. 720 --num_line_entries; 721 722 // If `line `!= 0, see if we can resolve function for each line 723 // entry in the line table. 724 for (uint32_t line_idx = 0; line && line_idx < num_line_entries; 725 ++line_idx) { 726 if (!line_table->GetLineEntryAtIndex(line_idx, sc.line_entry)) 727 continue; 728 729 auto file_vm_addr = 730 sc.line_entry.range.GetBaseAddress().GetFileAddress(); 731 if (file_vm_addr == LLDB_INVALID_ADDRESS) 732 continue; 733 734 auto symbol_up = 735 m_session_up->findSymbolByAddress(file_vm_addr, 736 PDB_SymType::Function); 737 if (symbol_up) { 738 auto func_uid = symbol_up->getSymIndexId(); 739 sc.function = sc.comp_unit->FindFunctionByUID(func_uid).get(); 740 if (sc.function == nullptr) { 741 auto pdb_func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get()); 742 assert(pdb_func); 743 sc.function = ParseCompileUnitFunctionForPDBFunc(pdb_func, sc); 744 } 745 if (sc.function && (resolve_scope & eSymbolContextBlock)) { 746 Block &block = sc.function->GetBlock(true); 747 sc.block = block.FindBlockByID(sc.function->GetID()); 748 } 749 } 750 sc_list.Append(sc); 751 } 752 } else if (has_line_table) { 753 // We can parse line table for the compile unit. But no query to 754 // resolve function or block. We append `sc` to the list anyway. 755 sc_list.Append(sc); 756 } 757 } else { 758 // No query for line entry, function or block. But we have a valid 759 // compile unit, append `sc` to the list. 760 sc_list.Append(sc); 761 } 762 } 763 } 764 return sc_list.GetSize() - old_size; 765 } 766 767 uint32_t SymbolFilePDB::FindGlobalVariables( 768 const lldb_private::ConstString &name, 769 const lldb_private::CompilerDeclContext *parent_decl_ctx, bool append, 770 uint32_t max_matches, lldb_private::VariableList &variables) { 771 return uint32_t(); 772 } 773 774 uint32_t 775 SymbolFilePDB::FindGlobalVariables(const lldb_private::RegularExpression ®ex, 776 bool append, uint32_t max_matches, 777 lldb_private::VariableList &variables) { 778 return uint32_t(); 779 } 780 781 bool SymbolFilePDB::ResolveFunction(llvm::pdb::PDBSymbolFunc *pdb_func, 782 bool include_inlines, 783 lldb_private::SymbolContextList &sc_list) { 784 if (!pdb_func) 785 return false; 786 lldb_private::SymbolContext sc; 787 auto file_vm_addr = pdb_func->getVirtualAddress(); 788 if (file_vm_addr == LLDB_INVALID_ADDRESS) 789 return false; 790 791 Address so_addr(file_vm_addr); 792 sc.comp_unit = GetCompileUnitContainsAddress(so_addr).get(); 793 if (!sc.comp_unit) 794 return false; 795 sc.module_sp = sc.comp_unit->GetModule(); 796 auto symbol_up = 797 m_session_up->findSymbolByAddress(file_vm_addr, PDB_SymType::Function); 798 if (!symbol_up) 799 return false; 800 801 auto *func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get()); 802 assert(func); 803 sc.function = ParseCompileUnitFunctionForPDBFunc(func, sc); 804 if (!sc.function) 805 return false; 806 807 sc_list.Append(sc); 808 return true; 809 } 810 811 bool SymbolFilePDB::ResolveFunction(uint32_t uid, bool include_inlines, 812 lldb_private::SymbolContextList &sc_list) { 813 auto pdb_func_up = 814 m_session_up->getConcreteSymbolById<PDBSymbolFunc>(uid); 815 if (!pdb_func_up && !(include_inlines && pdb_func_up->hasInlineAttribute())) 816 return false; 817 return ResolveFunction(pdb_func_up.get(), include_inlines, sc_list); 818 } 819 820 void SymbolFilePDB::CacheFunctionNames() { 821 if (!m_func_full_names.IsEmpty()) 822 return; 823 824 std::map<uint64_t, uint32_t> addr_ids; 825 826 if (auto results_up = m_global_scope_up->findAllChildren<PDBSymbolFunc>()) { 827 while (auto pdb_func_up = results_up->getNext()) { 828 auto uid = pdb_func_up->getSymIndexId(); 829 auto name = pdb_func_up->getName(); 830 auto demangled_name = pdb_func_up->getUndecoratedName(); 831 if (name.empty() && demangled_name.empty()) 832 continue; 833 if (pdb_func_up->isCompilerGenerated()) 834 continue; 835 836 if (!demangled_name.empty() && pdb_func_up->getVirtualAddress()) 837 addr_ids.insert(std::make_pair(pdb_func_up->getVirtualAddress(), uid)); 838 839 if (auto parent = pdb_func_up->getClassParent()) { 840 841 // PDB have symbols for class/struct methods or static methods in Enum 842 // Class. We won't bother to check if the parent is UDT or Enum here. 843 m_func_method_names.Append(ConstString(name), uid); 844 845 ConstString cstr_name(name); 846 847 // To search a method name, like NS::Class:MemberFunc, LLDB searches its 848 // base name, i.e. MemberFunc by default. Since PDBSymbolFunc does not 849 // have inforamtion of this, we extract base names and cache them by our 850 // own effort. 851 llvm::StringRef basename; 852 CPlusPlusLanguage::MethodName cpp_method(cstr_name); 853 if (cpp_method.IsValid()) { 854 llvm::StringRef context; 855 basename = cpp_method.GetBasename(); 856 if (basename.empty()) 857 CPlusPlusLanguage::ExtractContextAndIdentifier(name.c_str(), 858 context, basename); 859 } 860 861 if (!basename.empty()) 862 m_func_base_names.Append(ConstString(basename), uid); 863 else { 864 m_func_base_names.Append(ConstString(name), uid); 865 } 866 867 if (!demangled_name.empty()) 868 m_func_full_names.Append(ConstString(demangled_name), uid); 869 870 } else { 871 // Handle not-method symbols. 872 873 // The function name might contain namespace, or its lexical scope. It 874 // is not safe to get its base name by applying same scheme as we deal 875 // with the method names. 876 // FIXME: Remove namespace if function is static in a scope. 877 m_func_base_names.Append(ConstString(name), uid); 878 879 if (name == "main") { 880 m_func_full_names.Append(ConstString(name), uid); 881 882 if (!demangled_name.empty() && name != demangled_name) { 883 m_func_full_names.Append(ConstString(demangled_name), uid); 884 m_func_base_names.Append(ConstString(demangled_name), uid); 885 } 886 } else if (!demangled_name.empty()) { 887 m_func_full_names.Append(ConstString(demangled_name), uid); 888 } else { 889 m_func_full_names.Append(ConstString(name), uid); 890 } 891 } 892 } 893 } 894 895 if (auto results_up = 896 m_global_scope_up->findAllChildren<PDBSymbolPublicSymbol>()) { 897 while (auto pub_sym_up = results_up->getNext()) { 898 if (!pub_sym_up->isFunction()) 899 continue; 900 auto name = pub_sym_up->getName(); 901 if (name.empty()) 902 continue; 903 904 if (CPlusPlusLanguage::IsCPPMangledName(name.c_str())) { 905 auto demangled_name = pub_sym_up->getUndecoratedName(); 906 std::vector<uint32_t> ids; 907 auto vm_addr = pub_sym_up->getVirtualAddress(); 908 909 // PDB public symbol has mangled name for its associated function. 910 if (vm_addr && addr_ids.find(vm_addr) != addr_ids.end()) { 911 // Cache mangled name. 912 m_func_full_names.Append(ConstString(name), addr_ids[vm_addr]); 913 } 914 } 915 } 916 } 917 // Sort them before value searching is working properly 918 m_func_full_names.Sort(); 919 m_func_full_names.SizeToFit(); 920 m_func_method_names.Sort(); 921 m_func_method_names.SizeToFit(); 922 m_func_base_names.Sort(); 923 m_func_base_names.SizeToFit(); 924 } 925 926 uint32_t SymbolFilePDB::FindFunctions( 927 const lldb_private::ConstString &name, 928 const lldb_private::CompilerDeclContext *parent_decl_ctx, 929 uint32_t name_type_mask, bool include_inlines, bool append, 930 lldb_private::SymbolContextList &sc_list) { 931 if (!append) 932 sc_list.Clear(); 933 lldbassert((name_type_mask & eFunctionNameTypeAuto) == 0); 934 935 if (name_type_mask == eFunctionNameTypeNone) 936 return 0; 937 if (!DeclContextMatchesThisSymbolFile(parent_decl_ctx)) 938 return 0; 939 if (name.IsEmpty()) 940 return 0; 941 942 auto old_size = sc_list.GetSize(); 943 if (name_type_mask & eFunctionNameTypeFull || 944 name_type_mask & eFunctionNameTypeBase || 945 name_type_mask & eFunctionNameTypeMethod) { 946 CacheFunctionNames(); 947 948 std::set<uint32_t> resolved_ids; 949 auto ResolveFn = [include_inlines, &name, &sc_list, &resolved_ids, this] ( 950 UniqueCStringMap<uint32_t> &Names) 951 { 952 std::vector<uint32_t> ids; 953 if (Names.GetValues(name, ids)) { 954 for (auto id : ids) { 955 if (resolved_ids.find(id) == resolved_ids.end()) { 956 if (ResolveFunction(id, include_inlines, sc_list)) 957 resolved_ids.insert(id); 958 } 959 } 960 } 961 }; 962 if (name_type_mask & eFunctionNameTypeFull) { 963 ResolveFn(m_func_full_names); 964 } 965 if (name_type_mask & eFunctionNameTypeBase) { 966 ResolveFn(m_func_base_names); 967 } 968 if (name_type_mask & eFunctionNameTypeMethod) { 969 ResolveFn(m_func_method_names); 970 } 971 } 972 return sc_list.GetSize() - old_size; 973 } 974 975 uint32_t 976 SymbolFilePDB::FindFunctions(const lldb_private::RegularExpression ®ex, 977 bool include_inlines, bool append, 978 lldb_private::SymbolContextList &sc_list) { 979 if (!append) 980 sc_list.Clear(); 981 if (!regex.IsValid()) 982 return 0; 983 984 auto old_size = sc_list.GetSize(); 985 CacheFunctionNames(); 986 987 std::set<uint32_t> resolved_ids; 988 auto ResolveFn = [®ex, include_inlines, &sc_list, &resolved_ids, this] ( 989 UniqueCStringMap<uint32_t> &Names) 990 { 991 std::vector<uint32_t> ids; 992 if (Names.GetValues(regex, ids)) { 993 for (auto id : ids) { 994 if (resolved_ids.find(id) == resolved_ids.end()) 995 if (ResolveFunction(id, include_inlines, sc_list)) 996 resolved_ids.insert(id); 997 } 998 } 999 }; 1000 ResolveFn(m_func_full_names); 1001 ResolveFn(m_func_base_names); 1002 1003 return sc_list.GetSize() - old_size; 1004 } 1005 1006 void SymbolFilePDB::GetMangledNamesForFunction( 1007 const std::string &scope_qualified_name, 1008 std::vector<lldb_private::ConstString> &mangled_names) {} 1009 1010 uint32_t SymbolFilePDB::FindTypes( 1011 const lldb_private::SymbolContext &sc, 1012 const lldb_private::ConstString &name, 1013 const lldb_private::CompilerDeclContext *parent_decl_ctx, bool append, 1014 uint32_t max_matches, 1015 llvm::DenseSet<lldb_private::SymbolFile *> &searched_symbol_files, 1016 lldb_private::TypeMap &types) { 1017 if (!append) 1018 types.Clear(); 1019 if (!name) 1020 return 0; 1021 if (!DeclContextMatchesThisSymbolFile(parent_decl_ctx)) 1022 return 0; 1023 1024 searched_symbol_files.clear(); 1025 searched_symbol_files.insert(this); 1026 1027 std::string name_str = name.AsCString(); 1028 1029 // There is an assumption 'name' is not a regex 1030 FindTypesByName(name_str, max_matches, types); 1031 1032 return types.GetSize(); 1033 } 1034 1035 void 1036 SymbolFilePDB::FindTypesByRegex(const lldb_private::RegularExpression ®ex, 1037 uint32_t max_matches, 1038 lldb_private::TypeMap &types) { 1039 // When searching by regex, we need to go out of our way to limit the search 1040 // space as much as possible since this searches EVERYTHING in the PDB, 1041 // manually doing regex comparisons. PDB library isn't optimized for regex 1042 // searches or searches across multiple symbol types at the same time, so the 1043 // best we can do is to search enums, then typedefs, then classes one by one, 1044 // and do a regex comparison against each of them. 1045 PDB_SymType tags_to_search[] = {PDB_SymType::Enum, PDB_SymType::Typedef, 1046 PDB_SymType::UDT}; 1047 std::unique_ptr<IPDBEnumSymbols> results; 1048 1049 uint32_t matches = 0; 1050 1051 for (auto tag : tags_to_search) { 1052 results = m_global_scope_up->findAllChildren(tag); 1053 if (!results) 1054 continue; 1055 1056 while (auto result = results->getNext()) { 1057 if (max_matches > 0 && matches >= max_matches) 1058 break; 1059 1060 std::string type_name; 1061 if (auto enum_type = llvm::dyn_cast<PDBSymbolTypeEnum>(result.get())) 1062 type_name = enum_type->getName(); 1063 else if (auto typedef_type = 1064 llvm::dyn_cast<PDBSymbolTypeTypedef>(result.get())) 1065 type_name = typedef_type->getName(); 1066 else if (auto class_type = llvm::dyn_cast<PDBSymbolTypeUDT>(result.get())) 1067 type_name = class_type->getName(); 1068 else { 1069 // We're looking only for types that have names. Skip symbols, as well 1070 // as unnamed types such as arrays, pointers, etc. 1071 continue; 1072 } 1073 1074 if (!regex.Execute(type_name)) 1075 continue; 1076 1077 // This should cause the type to get cached and stored in the `m_types` 1078 // lookup. 1079 if (!ResolveTypeUID(result->getSymIndexId())) 1080 continue; 1081 1082 auto iter = m_types.find(result->getSymIndexId()); 1083 if (iter == m_types.end()) 1084 continue; 1085 types.Insert(iter->second); 1086 ++matches; 1087 } 1088 } 1089 } 1090 1091 void SymbolFilePDB::FindTypesByName(const std::string &name, 1092 uint32_t max_matches, 1093 lldb_private::TypeMap &types) { 1094 std::unique_ptr<IPDBEnumSymbols> results; 1095 results = m_global_scope_up->findChildren(PDB_SymType::None, name, 1096 PDB_NameSearchFlags::NS_Default); 1097 if (!results) 1098 return; 1099 1100 uint32_t matches = 0; 1101 1102 while (auto result = results->getNext()) { 1103 if (max_matches > 0 && matches >= max_matches) 1104 break; 1105 switch (result->getSymTag()) { 1106 case PDB_SymType::Enum: 1107 case PDB_SymType::UDT: 1108 case PDB_SymType::Typedef: 1109 break; 1110 default: 1111 // We're looking only for types that have names. Skip symbols, as well as 1112 // unnamed types such as arrays, pointers, etc. 1113 continue; 1114 } 1115 1116 // This should cause the type to get cached and stored in the `m_types` 1117 // lookup. 1118 if (!ResolveTypeUID(result->getSymIndexId())) 1119 continue; 1120 1121 auto iter = m_types.find(result->getSymIndexId()); 1122 if (iter == m_types.end()) 1123 continue; 1124 types.Insert(iter->second); 1125 ++matches; 1126 } 1127 } 1128 1129 size_t SymbolFilePDB::FindTypes( 1130 const std::vector<lldb_private::CompilerContext> &contexts, bool append, 1131 lldb_private::TypeMap &types) { 1132 return 0; 1133 } 1134 1135 lldb_private::TypeList *SymbolFilePDB::GetTypeList() { 1136 return m_obj_file->GetModule()->GetTypeList(); 1137 } 1138 1139 void 1140 SymbolFilePDB::GetTypesForPDBSymbol(const llvm::pdb::PDBSymbol *pdb_symbol, 1141 uint32_t type_mask, 1142 TypeCollection &type_collection) { 1143 if (!pdb_symbol) 1144 return; 1145 1146 bool can_parse = false; 1147 switch (pdb_symbol->getSymTag()) { 1148 case PDB_SymType::ArrayType: 1149 can_parse = ((type_mask & eTypeClassArray) != 0); 1150 break; 1151 case PDB_SymType::BuiltinType: 1152 can_parse = ((type_mask & eTypeClassBuiltin) != 0); 1153 break; 1154 case PDB_SymType::Enum: 1155 can_parse = ((type_mask & eTypeClassEnumeration) != 0); 1156 break; 1157 case PDB_SymType::Function: 1158 case PDB_SymType::FunctionSig: 1159 can_parse = ((type_mask & eTypeClassFunction) != 0); 1160 break; 1161 case PDB_SymType::PointerType: 1162 can_parse = ((type_mask & (eTypeClassPointer | eTypeClassBlockPointer | 1163 eTypeClassMemberPointer)) != 0); 1164 break; 1165 case PDB_SymType::Typedef: 1166 can_parse = ((type_mask & eTypeClassTypedef) != 0); 1167 break; 1168 case PDB_SymType::UDT: { 1169 auto *udt = llvm::dyn_cast<PDBSymbolTypeUDT>(pdb_symbol); 1170 assert(udt); 1171 can_parse = (udt->getUdtKind() != PDB_UdtType::Interface && 1172 ((type_mask & (eTypeClassClass | eTypeClassStruct | 1173 eTypeClassUnion)) != 0)); 1174 } break; 1175 default:break; 1176 } 1177 1178 if (can_parse) { 1179 if (auto *type = ResolveTypeUID(pdb_symbol->getSymIndexId())) { 1180 auto result = 1181 std::find(type_collection.begin(), type_collection.end(), type); 1182 if (result == type_collection.end()) 1183 type_collection.push_back(type); 1184 } 1185 } 1186 1187 auto results_up = pdb_symbol->findAllChildren(); 1188 while (auto symbol_up = results_up->getNext()) 1189 GetTypesForPDBSymbol(symbol_up.get(), type_mask, type_collection); 1190 } 1191 1192 size_t SymbolFilePDB::GetTypes(lldb_private::SymbolContextScope *sc_scope, 1193 uint32_t type_mask, 1194 lldb_private::TypeList &type_list) { 1195 TypeCollection type_collection; 1196 uint32_t old_size = type_list.GetSize(); 1197 CompileUnit *cu = sc_scope ? 1198 sc_scope->CalculateSymbolContextCompileUnit() : nullptr; 1199 if (cu) { 1200 auto compiland_up = GetPDBCompilandByUID(cu->GetID()); 1201 GetTypesForPDBSymbol(compiland_up.get(), type_mask, type_collection); 1202 } else { 1203 for (uint32_t cu_idx = 0; cu_idx < GetNumCompileUnits(); ++cu_idx) { 1204 auto cu_sp = ParseCompileUnitAtIndex(cu_idx); 1205 if (cu_sp.get()) { 1206 auto compiland_up = GetPDBCompilandByUID(cu_sp->GetID()); 1207 GetTypesForPDBSymbol(compiland_up.get(), type_mask, type_collection); 1208 } 1209 } 1210 } 1211 1212 for (auto type : type_collection) { 1213 type->GetForwardCompilerType(); 1214 type_list.Insert(type->shared_from_this()); 1215 } 1216 return type_list.GetSize() - old_size; 1217 } 1218 1219 lldb_private::TypeSystem * 1220 SymbolFilePDB::GetTypeSystemForLanguage(lldb::LanguageType language) { 1221 auto type_system = 1222 m_obj_file->GetModule()->GetTypeSystemForLanguage(language); 1223 if (type_system) 1224 type_system->SetSymbolFile(this); 1225 return type_system; 1226 } 1227 1228 lldb_private::CompilerDeclContext SymbolFilePDB::FindNamespace( 1229 const lldb_private::SymbolContext &sc, 1230 const lldb_private::ConstString &name, 1231 const lldb_private::CompilerDeclContext *parent_decl_ctx) { 1232 return lldb_private::CompilerDeclContext(); 1233 } 1234 1235 lldb_private::ConstString SymbolFilePDB::GetPluginName() { 1236 static ConstString g_name("pdb"); 1237 return g_name; 1238 } 1239 1240 uint32_t SymbolFilePDB::GetPluginVersion() { return 1; } 1241 1242 IPDBSession &SymbolFilePDB::GetPDBSession() { return *m_session_up; } 1243 1244 const IPDBSession &SymbolFilePDB::GetPDBSession() const { 1245 return *m_session_up; 1246 } 1247 1248 lldb::CompUnitSP 1249 SymbolFilePDB::ParseCompileUnitForUID(uint32_t id, uint32_t index) { 1250 auto found_cu = m_comp_units.find(id); 1251 if (found_cu != m_comp_units.end()) 1252 return found_cu->second; 1253 1254 auto compiland_up = GetPDBCompilandByUID(id); 1255 if (!compiland_up) 1256 return CompUnitSP(); 1257 std::string path = GetSourceFileNameForPDBCompiland(compiland_up.get()); 1258 if (path.empty()) 1259 return CompUnitSP(); 1260 1261 lldb::LanguageType lang; 1262 auto details = compiland_up->findOneChild<PDBSymbolCompilandDetails>(); 1263 if (!details) 1264 lang = lldb::eLanguageTypeC_plus_plus; 1265 else 1266 lang = TranslateLanguage(details->getLanguage()); 1267 1268 // Don't support optimized code for now, DebugInfoPDB does not return this 1269 // information. 1270 LazyBool optimized = eLazyBoolNo; 1271 auto cu_sp = std::make_shared<CompileUnit>( 1272 m_obj_file->GetModule(), nullptr, path.c_str(), id, lang, optimized); 1273 1274 if (!cu_sp) 1275 return CompUnitSP(); 1276 1277 m_comp_units.insert(std::make_pair(id, cu_sp)); 1278 if (index == UINT32_MAX) 1279 GetCompileUnitIndex(compiland_up.get(), index); 1280 lldbassert(index != UINT32_MAX); 1281 m_obj_file->GetModule()->GetSymbolVendor()->SetCompileUnitAtIndex( 1282 index, cu_sp); 1283 return cu_sp; 1284 } 1285 1286 bool SymbolFilePDB::ParseCompileUnitLineTable( 1287 const lldb_private::SymbolContext &sc, uint32_t match_line) { 1288 lldbassert(sc.comp_unit); 1289 1290 auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID()); 1291 if (!compiland_up) 1292 return false; 1293 1294 // LineEntry needs the *index* of the file into the list of support files 1295 // returned by ParseCompileUnitSupportFiles. But the underlying SDK gives us 1296 // a globally unique idenfitifier in the namespace of the PDB. So, we have to 1297 // do a mapping so that we can hand out indices. 1298 llvm::DenseMap<uint32_t, uint32_t> index_map; 1299 BuildSupportFileIdToSupportFileIndexMap(*compiland_up, index_map); 1300 auto line_table = llvm::make_unique<LineTable>(sc.comp_unit); 1301 1302 // Find contributions to `compiland` from all source and header files. 1303 std::string path = sc.comp_unit->GetPath(); 1304 auto files = m_session_up->getSourceFilesForCompiland(*compiland_up); 1305 if (!files) 1306 return false; 1307 1308 // For each source and header file, create a LineSequence for contributions to 1309 // the compiland from that file, and add the sequence. 1310 while (auto file = files->getNext()) { 1311 std::unique_ptr<LineSequence> sequence( 1312 line_table->CreateLineSequenceContainer()); 1313 auto lines = m_session_up->findLineNumbers(*compiland_up, *file); 1314 if (!lines) 1315 continue; 1316 int entry_count = lines->getChildCount(); 1317 1318 uint64_t prev_addr; 1319 uint32_t prev_length; 1320 uint32_t prev_line; 1321 uint32_t prev_source_idx; 1322 1323 for (int i = 0; i < entry_count; ++i) { 1324 auto line = lines->getChildAtIndex(i); 1325 1326 uint64_t lno = line->getLineNumber(); 1327 uint64_t addr = line->getVirtualAddress(); 1328 uint32_t length = line->getLength(); 1329 uint32_t source_id = line->getSourceFileId(); 1330 uint32_t col = line->getColumnNumber(); 1331 uint32_t source_idx = index_map[source_id]; 1332 1333 // There was a gap between the current entry and the previous entry if the 1334 // addresses don't perfectly line up. 1335 bool is_gap = (i > 0) && (prev_addr + prev_length < addr); 1336 1337 // Before inserting the current entry, insert a terminal entry at the end 1338 // of the previous entry's address range if the current entry resulted in 1339 // a gap from the previous entry. 1340 if (is_gap && ShouldAddLine(match_line, prev_line, prev_length)) { 1341 line_table->AppendLineEntryToSequence( 1342 sequence.get(), prev_addr + prev_length, prev_line, 0, 1343 prev_source_idx, false, false, false, false, true); 1344 } 1345 1346 if (ShouldAddLine(match_line, lno, length)) { 1347 bool is_statement = line->isStatement(); 1348 bool is_prologue = false; 1349 bool is_epilogue = false; 1350 auto func = 1351 m_session_up->findSymbolByAddress(addr, PDB_SymType::Function); 1352 if (func) { 1353 auto prologue = func->findOneChild<PDBSymbolFuncDebugStart>(); 1354 if (prologue) 1355 is_prologue = (addr == prologue->getVirtualAddress()); 1356 1357 auto epilogue = func->findOneChild<PDBSymbolFuncDebugEnd>(); 1358 if (epilogue) 1359 is_epilogue = (addr == epilogue->getVirtualAddress()); 1360 } 1361 1362 line_table->AppendLineEntryToSequence(sequence.get(), addr, lno, col, 1363 source_idx, is_statement, false, 1364 is_prologue, is_epilogue, false); 1365 } 1366 1367 prev_addr = addr; 1368 prev_length = length; 1369 prev_line = lno; 1370 prev_source_idx = source_idx; 1371 } 1372 1373 if (entry_count > 0 && ShouldAddLine(match_line, prev_line, prev_length)) { 1374 // The end is always a terminal entry, so insert it regardless. 1375 line_table->AppendLineEntryToSequence( 1376 sequence.get(), prev_addr + prev_length, prev_line, 0, 1377 prev_source_idx, false, false, false, false, true); 1378 } 1379 1380 line_table->InsertSequence(sequence.release()); 1381 } 1382 1383 if (line_table->GetSize()) { 1384 sc.comp_unit->SetLineTable(line_table.release()); 1385 return true; 1386 } 1387 return false; 1388 } 1389 1390 void SymbolFilePDB::BuildSupportFileIdToSupportFileIndexMap( 1391 const PDBSymbolCompiland &compiland, 1392 llvm::DenseMap<uint32_t, uint32_t> &index_map) const { 1393 // This is a hack, but we need to convert the source id into an index into the 1394 // support files array. We don't want to do path comparisons to avoid 1395 // basename / full path issues that may or may not even be a problem, so we 1396 // use the globally unique source file identifiers. Ideally we could use the 1397 // global identifiers everywhere, but LineEntry currently assumes indices. 1398 auto source_files = m_session_up->getSourceFilesForCompiland(compiland); 1399 if (!source_files) 1400 return; 1401 int index = 0; 1402 1403 while (auto file = source_files->getNext()) { 1404 uint32_t source_id = file->getUniqueId(); 1405 index_map[source_id] = index++; 1406 } 1407 } 1408 1409 lldb::CompUnitSP SymbolFilePDB::GetCompileUnitContainsAddress( 1410 const lldb_private::Address &so_addr) { 1411 lldb::addr_t file_vm_addr = so_addr.GetFileAddress(); 1412 if (file_vm_addr == LLDB_INVALID_ADDRESS) 1413 return nullptr; 1414 1415 auto lines_up = 1416 m_session_up->findLineNumbersByAddress(file_vm_addr, /*Length=*/200); 1417 if (!lines_up) 1418 return nullptr; 1419 1420 auto first_line_up = lines_up->getNext(); 1421 if (!first_line_up) 1422 return nullptr; 1423 auto compiland_up = GetPDBCompilandByUID(first_line_up->getCompilandId()); 1424 if (compiland_up) { 1425 return ParseCompileUnitForUID(compiland_up->getSymIndexId()); 1426 } 1427 1428 return nullptr; 1429 } 1430 1431 Mangled 1432 SymbolFilePDB::GetMangledForPDBFunc(const llvm::pdb::PDBSymbolFunc *pdb_func) { 1433 Mangled mangled; 1434 if (!pdb_func) 1435 return mangled; 1436 1437 auto func_name = pdb_func->getName(); 1438 auto func_undecorated_name = pdb_func->getUndecoratedName(); 1439 std::string func_decorated_name; 1440 1441 // Seek from public symbols for non-static function's decorated name if any. 1442 // For static functions, they don't have undecorated names and aren't exposed 1443 // in Public Symbols either. 1444 if (!func_undecorated_name.empty()) { 1445 auto result_up = 1446 m_global_scope_up->findChildren(PDB_SymType::PublicSymbol, 1447 func_undecorated_name, 1448 PDB_NameSearchFlags::NS_UndecoratedName); 1449 if (result_up) { 1450 while (auto symbol_up = result_up->getNext()) { 1451 // For a public symbol, it is unique. 1452 lldbassert(result_up->getChildCount() == 1); 1453 if (auto *pdb_public_sym = 1454 llvm::dyn_cast<PDBSymbolPublicSymbol>(symbol_up.get())) { 1455 if (pdb_public_sym->isFunction()) { 1456 func_decorated_name = pdb_public_sym->getName(); 1457 } 1458 } 1459 } 1460 } 1461 } 1462 if (!func_decorated_name.empty()) { 1463 mangled.SetMangledName(ConstString(func_decorated_name)); 1464 1465 // For MSVC, format of C funciton's decorated name depends on calling 1466 // conventon. Unfortunately none of the format is recognized by current 1467 // LLDB. For example, `_purecall` is a __cdecl C function. From PDB, 1468 // `__purecall` is retrieved as both its decorated and 1469 // undecorated name (using PDBSymbolFunc::getUndecoratedName method). 1470 // However `__purecall` string is not treated as mangled in LLDB 1471 // (neither `?` nor `_Z` prefix). Mangled::GetDemangledName method 1472 // will fail internally and caches an empty string as its undecorated 1473 // name. So we will face a contradition here for the same symbol: 1474 // non-empty undecorated name from PDB 1475 // empty undecorated name from LLDB 1476 if (!func_undecorated_name.empty() && 1477 mangled.GetDemangledName(mangled.GuessLanguage()).IsEmpty()) 1478 mangled.SetDemangledName(ConstString(func_undecorated_name)); 1479 1480 // LLDB uses several flags to control how a C++ decorated name is 1481 // undecorated for MSVC. See `safeUndecorateName` in Class Mangled. 1482 // So the yielded name could be different from what we retrieve from 1483 // PDB source unless we also apply same flags in getting undecorated 1484 // name through PDBSymbolFunc::getUndecoratedNameEx method. 1485 if (!func_undecorated_name.empty() && 1486 mangled.GetDemangledName(mangled.GuessLanguage()) != 1487 ConstString(func_undecorated_name)) 1488 mangled.SetDemangledName(ConstString(func_undecorated_name)); 1489 } else if (!func_undecorated_name.empty()) { 1490 mangled.SetDemangledName(ConstString(func_undecorated_name)); 1491 } else if (!func_name.empty()) 1492 mangled.SetValue(ConstString(func_name), false); 1493 1494 return mangled; 1495 } 1496 1497 bool SymbolFilePDB::DeclContextMatchesThisSymbolFile( 1498 const lldb_private::CompilerDeclContext *decl_ctx) { 1499 if (decl_ctx == nullptr || !decl_ctx->IsValid()) 1500 return true; 1501 1502 TypeSystem *decl_ctx_type_system = decl_ctx->GetTypeSystem(); 1503 if (!decl_ctx_type_system) 1504 return false; 1505 TypeSystem *type_system = GetTypeSystemForLanguage( 1506 decl_ctx_type_system->GetMinimumLanguage(nullptr)); 1507 if (decl_ctx_type_system == type_system) 1508 return true; // The type systems match, return true 1509 1510 return false; 1511 } 1512