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   lldb_private::Type* func_type = ResolveTypeUID(pdb_func->getSymIndexId());
286   if (!func_type)
287     return nullptr;
288 
289   user_id_t func_type_uid = pdb_func->getSignatureId();
290 
291   Mangled mangled = GetMangledForPDBFunc(pdb_func);
292 
293   FunctionSP func_sp = std::make_shared<Function>(sc.comp_unit,
294                                                   pdb_func->getSymIndexId(),
295                                                   func_type_uid,
296                                                   mangled,
297                                                   func_type,
298                                                   func_range);
299 
300   sc.comp_unit->AddFunction(func_sp);
301   return func_sp.get();
302 }
303 
304 size_t SymbolFilePDB::ParseCompileUnitFunctions(
305     const lldb_private::SymbolContext &sc) {
306   lldbassert(sc.comp_unit);
307   size_t func_added = 0;
308   auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID());
309   if (!compiland_up)
310     return 0;
311   auto results_up = compiland_up->findAllChildren<PDBSymbolFunc>();
312   if (!results_up)
313     return 0;
314   while (auto pdb_func_up = results_up->getNext()) {
315     auto func_sp =
316         sc.comp_unit->FindFunctionByUID(pdb_func_up->getSymIndexId());
317     if (!func_sp) {
318       if (ParseCompileUnitFunctionForPDBFunc(pdb_func_up.get(), sc))
319         ++func_added;
320     }
321   }
322   return func_added;
323 }
324 
325 bool SymbolFilePDB::ParseCompileUnitLineTable(
326     const lldb_private::SymbolContext &sc) {
327   lldbassert(sc.comp_unit);
328   if (sc.comp_unit->GetLineTable())
329     return true;
330   return ParseCompileUnitLineTable(sc, 0);
331 }
332 
333 bool SymbolFilePDB::ParseCompileUnitDebugMacros(
334     const lldb_private::SymbolContext &sc) {
335   // PDB doesn't contain information about macros
336   return false;
337 }
338 
339 bool SymbolFilePDB::ParseCompileUnitSupportFiles(
340     const lldb_private::SymbolContext &sc,
341     lldb_private::FileSpecList &support_files) {
342   lldbassert(sc.comp_unit);
343 
344   // In theory this is unnecessary work for us, because all of this information
345   // is easily (and quickly) accessible from DebugInfoPDB, so caching it a
346   // second time seems like a waste.  Unfortunately, there's no good way around
347   // this short of a moderate refactor since SymbolVendor depends on being able
348   // to cache this list.
349   auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID());
350   if (!compiland_up)
351     return false;
352   auto files = m_session_up->getSourceFilesForCompiland(*compiland_up);
353   if (!files || files->getChildCount() == 0)
354     return false;
355 
356   while (auto file = files->getNext()) {
357     FileSpec spec(file->getFileName(), false, FileSpec::ePathSyntaxWindows);
358     support_files.AppendIfUnique(spec);
359   }
360   return true;
361 }
362 
363 bool SymbolFilePDB::ParseImportedModules(
364     const lldb_private::SymbolContext &sc,
365     std::vector<lldb_private::ConstString> &imported_modules) {
366   // PDB does not yet support module debug info
367   return false;
368 }
369 
370 static size_t
371 ParseFunctionBlocksForPDBSymbol(const lldb_private::SymbolContext &sc,
372                                 uint64_t func_file_vm_addr,
373                                 const llvm::pdb::PDBSymbol *pdb_symbol,
374                                 lldb_private::Block *parent_block,
375                                 bool is_top_parent) {
376   assert(pdb_symbol && parent_block);
377 
378   size_t num_added = 0;
379   switch (pdb_symbol->getSymTag()) {
380   case PDB_SymType::Block:
381   case PDB_SymType::Function: {
382     Block *block = nullptr;
383     auto &raw_sym = pdb_symbol->getRawSymbol();
384     if (auto *pdb_func = llvm::dyn_cast<PDBSymbolFunc>(pdb_symbol)) {
385       if (pdb_func->hasNoInlineAttribute())
386         break;
387       if (is_top_parent)
388         block = parent_block;
389       else
390         break;
391     } else if (llvm::dyn_cast<PDBSymbolBlock>(pdb_symbol)) {
392       auto uid = pdb_symbol->getSymIndexId();
393       if (parent_block->FindBlockByID(uid))
394         break;
395       if (raw_sym.getVirtualAddress() < func_file_vm_addr)
396         break;
397 
398       auto block_sp = std::make_shared<Block>(pdb_symbol->getSymIndexId());
399       parent_block->AddChild(block_sp);
400       block = block_sp.get();
401     } else
402       llvm_unreachable("Unexpected PDB symbol!");
403 
404     block->AddRange(
405         Block::Range(raw_sym.getVirtualAddress() - func_file_vm_addr,
406                      raw_sym.getLength()));
407     block->FinalizeRanges();
408     ++num_added;
409 
410     auto results_up = pdb_symbol->findAllChildren();
411     if (!results_up)
412       break;
413     while (auto symbol_up = results_up->getNext()) {
414       num_added += ParseFunctionBlocksForPDBSymbol(sc, func_file_vm_addr,
415                                                    symbol_up.get(),
416                                                    block, false);
417     }
418   } break;
419   default: break;
420   }
421   return num_added;
422 }
423 
424 size_t
425 SymbolFilePDB::ParseFunctionBlocks(const lldb_private::SymbolContext &sc) {
426   lldbassert(sc.comp_unit && sc.function);
427   size_t num_added = 0;
428   auto uid = sc.function->GetID();
429   auto pdb_func_up = m_session_up->getConcreteSymbolById<PDBSymbolFunc>(uid);
430   if (!pdb_func_up)
431     return 0;
432   Block &parent_block = sc.function->GetBlock(false);
433   num_added =
434       ParseFunctionBlocksForPDBSymbol(sc, pdb_func_up->getVirtualAddress(),
435                                       pdb_func_up.get(), &parent_block, true);
436   return num_added;
437 }
438 
439 size_t SymbolFilePDB::ParseTypes(const lldb_private::SymbolContext &sc) {
440   lldbassert(sc.module_sp.get());
441   if (!sc.comp_unit)
442     return 0;
443 
444   size_t num_added = 0;
445   auto compiland = GetPDBCompilandByUID(sc.comp_unit->GetID());
446   if (!compiland)
447     return 0;
448 
449   auto ParseTypesByTagFn = [&num_added, this](const PDBSymbol &raw_sym) {
450     std::unique_ptr<IPDBEnumSymbols> results;
451     PDB_SymType tags_to_search[] = { PDB_SymType::Enum, PDB_SymType::Typedef,
452         PDB_SymType::UDT };
453     for (auto tag : tags_to_search) {
454       results = raw_sym.findAllChildren(tag);
455       if (!results || results->getChildCount() == 0)
456         continue;
457       while (auto symbol = results->getNext()) {
458         switch (symbol->getSymTag()) {
459         case PDB_SymType::Enum:
460         case PDB_SymType::UDT:
461         case PDB_SymType::Typedef:
462           break;
463         default:
464           continue;
465         }
466 
467         // This should cause the type to get cached and stored in the `m_types`
468         // lookup.
469         if (!ResolveTypeUID(symbol->getSymIndexId()))
470           continue;
471 
472         ++num_added;
473       }
474     }
475   };
476 
477   if (sc.function) {
478     auto pdb_func =
479         m_session_up->getConcreteSymbolById<PDBSymbolFunc>(sc.function->GetID());
480     if (!pdb_func)
481       return 0;
482     ParseTypesByTagFn(*pdb_func);
483   } else {
484     ParseTypesByTagFn(*compiland);
485 
486     // Also parse global types particularly coming from this compiland.
487     // Unfortunately, PDB has no compiland information for each global type.
488     // We have to parse them all. But ensure we only do this once.
489     static bool parse_all_global_types = false;
490     if (!parse_all_global_types) {
491       ParseTypesByTagFn(*m_global_scope_up);
492       parse_all_global_types = true;
493     }
494   }
495   return num_added;
496 }
497 
498 size_t
499 SymbolFilePDB::ParseVariablesForContext(const lldb_private::SymbolContext &sc) {
500   // TODO: Implement this
501   return size_t();
502 }
503 
504 lldb_private::Type *SymbolFilePDB::ResolveTypeUID(lldb::user_id_t type_uid) {
505   auto find_result = m_types.find(type_uid);
506   if (find_result != m_types.end())
507     return find_result->second.get();
508 
509   TypeSystem *type_system =
510       GetTypeSystemForLanguage(lldb::eLanguageTypeC_plus_plus);
511   ClangASTContext *clang_type_system =
512       llvm::dyn_cast_or_null<ClangASTContext>(type_system);
513   if (!clang_type_system)
514     return nullptr;
515   PDBASTParser *pdb =
516       llvm::dyn_cast<PDBASTParser>(clang_type_system->GetPDBParser());
517   if (!pdb)
518     return nullptr;
519 
520   auto pdb_type = m_session_up->getSymbolById(type_uid);
521   if (pdb_type == nullptr)
522     return nullptr;
523 
524   lldb::TypeSP result = pdb->CreateLLDBTypeFromPDBType(*pdb_type);
525   if (result.get()) {
526     m_types.insert(std::make_pair(type_uid, result));
527     auto type_list = GetTypeList();
528     if (type_list)
529       type_list->Insert(result);
530   }
531   return result.get();
532 }
533 
534 bool SymbolFilePDB::CompleteType(lldb_private::CompilerType &compiler_type) {
535   // TODO: Implement this
536   return false;
537 }
538 
539 lldb_private::CompilerDecl SymbolFilePDB::GetDeclForUID(lldb::user_id_t uid) {
540   return lldb_private::CompilerDecl();
541 }
542 
543 lldb_private::CompilerDeclContext
544 SymbolFilePDB::GetDeclContextForUID(lldb::user_id_t uid) {
545   // PDB always uses the translation unit decl context for everything.  We can
546   // improve this later but it's not easy because PDB doesn't provide a high
547   // enough level of type fidelity in this area.
548   return *m_tu_decl_ctx_up;
549 }
550 
551 lldb_private::CompilerDeclContext
552 SymbolFilePDB::GetDeclContextContainingUID(lldb::user_id_t uid) {
553   return *m_tu_decl_ctx_up;
554 }
555 
556 void SymbolFilePDB::ParseDeclsForContext(
557     lldb_private::CompilerDeclContext decl_ctx) {}
558 
559 uint32_t
560 SymbolFilePDB::ResolveSymbolContext(const lldb_private::Address &so_addr,
561                                     uint32_t resolve_scope,
562                                     lldb_private::SymbolContext &sc) {
563   uint32_t resolved_flags = 0;
564   if (resolve_scope & eSymbolContextCompUnit ||
565       resolve_scope & eSymbolContextVariable ||
566       resolve_scope & eSymbolContextFunction ||
567       resolve_scope & eSymbolContextBlock ||
568       resolve_scope & eSymbolContextLineEntry) {
569     addr_t file_vm_addr = so_addr.GetFileAddress();
570     auto symbol_up =
571         m_session_up->findSymbolByAddress(file_vm_addr, PDB_SymType::None);
572     if (!symbol_up)
573       return 0;
574 
575     auto cu_sp = GetCompileUnitContainsAddress(so_addr);
576     if (!cu_sp) {
577       if (resolved_flags | eSymbolContextVariable) {
578         // TODO: Resolve variables
579       }
580       return 0;
581     }
582     sc.comp_unit = cu_sp.get();
583     resolved_flags |= eSymbolContextCompUnit;
584     lldbassert(sc.module_sp == cu_sp->GetModule());
585 
586     switch (symbol_up->getSymTag()) {
587     case PDB_SymType::Function:
588       if (resolve_scope & eSymbolContextFunction) {
589         auto *pdb_func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get());
590         assert(pdb_func);
591         auto func_uid = pdb_func->getSymIndexId();
592         sc.function = sc.comp_unit->FindFunctionByUID(func_uid).get();
593         if (sc.function == nullptr)
594           sc.function = ParseCompileUnitFunctionForPDBFunc(pdb_func, sc);
595         if (sc.function) {
596           resolved_flags |= eSymbolContextFunction;
597           if (resolve_scope & eSymbolContextBlock) {
598             Block &block = sc.function->GetBlock(true);
599             sc.block = block.FindBlockByID(sc.function->GetID());
600             if (sc.block)
601               resolved_flags |= eSymbolContextBlock;
602           }
603         }
604       }
605       break;
606     default:
607       break;
608     }
609 
610     if (resolve_scope & eSymbolContextLineEntry) {
611       if (auto *line_table = sc.comp_unit->GetLineTable()) {
612         Address addr(so_addr);
613         if (line_table->FindLineEntryByAddress(addr, sc.line_entry))
614           resolved_flags |= eSymbolContextLineEntry;
615       }
616     }
617   }
618   return resolved_flags;
619 }
620 
621 std::string SymbolFilePDB::GetSourceFileNameForPDBCompiland(
622     const PDBSymbolCompiland *pdb_compiland) {
623   if (!pdb_compiland)
624     return std::string();
625 
626   std::string source_file_name;
627   // `getSourceFileName` returns the basename of the original source file
628   // used to generate this compiland.  It does not return the full path.
629   // Currently the only way to get that is to do a basename lookup to get the
630   // IPDBSourceFile, but this is ambiguous in the case of two source files
631   // with the same name contributing to the same compiland. This is an edge
632   // case that we ignore for now, although we need to a long-term solution.
633   std::string file_name = pdb_compiland->getSourceFileName();
634   if (!file_name.empty()) {
635     auto one_src_file_up =
636         m_session_up->findOneSourceFile(pdb_compiland, file_name,
637                                         PDB_NameSearchFlags::NS_CaseInsensitive);
638     if (one_src_file_up)
639       source_file_name = one_src_file_up->getFileName();
640   }
641   // For some reason, source file name could be empty, so we will walk through
642   // all source files of this compiland, and determine the right source file
643   // if any that is used to generate this compiland based on language
644   // indicated in compilanddetails language field.
645   if (!source_file_name.empty())
646     return source_file_name;
647 
648   auto details_up = pdb_compiland->findOneChild<PDBSymbolCompilandDetails>();
649   PDB_Lang pdb_lang = details_up ? details_up->getLanguage() : PDB_Lang::Cpp;
650   auto src_files_up =
651       m_session_up->getSourceFilesForCompiland(*pdb_compiland);
652   if (src_files_up) {
653     while (auto file_up = src_files_up->getNext()) {
654       FileSpec file_spec(file_up->getFileName(), false,
655                          FileSpec::ePathSyntaxWindows);
656       auto file_extension = file_spec.GetFileNameExtension();
657       if (pdb_lang == PDB_Lang::Cpp || pdb_lang == PDB_Lang::C) {
658         static const char* exts[] = { "cpp", "c", "cc", "cxx" };
659         if (llvm::is_contained(exts, file_extension.GetStringRef().lower())) {
660           source_file_name = file_up->getFileName();
661           break;
662         }
663       } else if (pdb_lang == PDB_Lang::Masm &&
664                  ConstString::Compare(file_extension, ConstString("ASM"),
665                                       false) == 0) {
666         source_file_name = file_up->getFileName();
667         break;
668       }
669     }
670   }
671   return source_file_name;
672 }
673 
674 uint32_t SymbolFilePDB::ResolveSymbolContext(
675     const lldb_private::FileSpec &file_spec, uint32_t line, bool check_inlines,
676     uint32_t resolve_scope, lldb_private::SymbolContextList &sc_list) {
677   const size_t old_size = sc_list.GetSize();
678   if (resolve_scope & lldb::eSymbolContextCompUnit) {
679     // Locate all compilation units with line numbers referencing the specified
680     // file.  For example, if `file_spec` is <vector>, then this should return
681     // all source files and header files that reference <vector>, either
682     // directly or indirectly.
683     auto compilands = m_session_up->findCompilandsForSourceFile(
684         file_spec.GetPath(), PDB_NameSearchFlags::NS_CaseInsensitive);
685 
686     if (!compilands)
687       return 0;
688 
689     // For each one, either find its previously parsed data or parse it afresh
690     // and add it to the symbol context list.
691     while (auto compiland = compilands->getNext()) {
692       // If we're not checking inlines, then don't add line information for this
693       // file unless the FileSpec matches.
694       // For inline functions, we don't have to match the FileSpec since they
695       // could be defined in headers other than file specified in FileSpec.
696       if (!check_inlines) {
697         // `getSourceFileName` returns the basename of the original source file
698         // used to generate this compiland.  It does not return the full path.
699         // Currently the only way to get that is to do a basename lookup to get
700         // the IPDBSourceFile, but this is ambiguous in the case of two source
701         // files with the same name contributing to the same compiland.  This is
702         // a moderately extreme edge case, so we consider this OK for now,
703         // although we need to find a long-term solution.
704         std::string source_file =
705             GetSourceFileNameForPDBCompiland(compiland.get());
706         if (source_file.empty())
707           continue;
708         FileSpec this_spec(source_file, false, FileSpec::ePathSyntaxWindows);
709         bool need_full_match = !file_spec.GetDirectory().IsEmpty();
710         if (FileSpec::Compare(file_spec, this_spec, need_full_match) != 0)
711           continue;
712       }
713 
714       SymbolContext sc;
715       auto cu = ParseCompileUnitForUID(compiland->getSymIndexId());
716       if (!cu.get())
717         continue;
718       sc.comp_unit = cu.get();
719       sc.module_sp = cu->GetModule();
720 
721       // If we were asked to resolve line entries, add all entries to the line
722       // table that match the requested line (or all lines if `line` == 0).
723       if (resolve_scope & (eSymbolContextFunction | eSymbolContextBlock |
724                            eSymbolContextLineEntry)) {
725         bool has_line_table = ParseCompileUnitLineTable(sc, line);
726 
727         if ((resolve_scope & eSymbolContextLineEntry) && !has_line_table) {
728           // The query asks for line entries, but we can't get them for the
729           // compile unit. This is not normal for `line` = 0. So just assert it.
730           assert(line && "Couldn't get all line entries!\n");
731 
732           // Current compiland does not have the requested line. Search next.
733           continue;
734         }
735 
736         if (resolve_scope & (eSymbolContextFunction | eSymbolContextBlock)) {
737           if (!has_line_table)
738             continue;
739 
740           auto *line_table = sc.comp_unit->GetLineTable();
741           lldbassert(line_table);
742 
743           uint32_t num_line_entries = line_table->GetSize();
744           // Skip the terminal line entry.
745           --num_line_entries;
746 
747           // If `line `!= 0, see if we can resolve function for each line
748           // entry in the line table.
749           for (uint32_t line_idx = 0; line && line_idx < num_line_entries;
750                ++line_idx) {
751             if (!line_table->GetLineEntryAtIndex(line_idx, sc.line_entry))
752               continue;
753 
754             auto file_vm_addr =
755                 sc.line_entry.range.GetBaseAddress().GetFileAddress();
756             if (file_vm_addr == LLDB_INVALID_ADDRESS)
757               continue;
758 
759             auto symbol_up =
760                 m_session_up->findSymbolByAddress(file_vm_addr,
761                                                   PDB_SymType::Function);
762             if (symbol_up) {
763               auto func_uid = symbol_up->getSymIndexId();
764               sc.function = sc.comp_unit->FindFunctionByUID(func_uid).get();
765               if (sc.function == nullptr) {
766                 auto pdb_func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get());
767                 assert(pdb_func);
768                 sc.function = ParseCompileUnitFunctionForPDBFunc(pdb_func, sc);
769               }
770               if (sc.function && (resolve_scope & eSymbolContextBlock)) {
771                 Block &block = sc.function->GetBlock(true);
772                 sc.block = block.FindBlockByID(sc.function->GetID());
773               }
774             }
775             sc_list.Append(sc);
776           }
777         } else if (has_line_table) {
778           // We can parse line table for the compile unit. But no query to
779           // resolve function or block. We append `sc` to the list anyway.
780           sc_list.Append(sc);
781         }
782       } else {
783         // No query for line entry, function or block. But we have a valid
784         // compile unit, append `sc` to the list.
785         sc_list.Append(sc);
786       }
787     }
788   }
789   return sc_list.GetSize() - old_size;
790 }
791 
792 uint32_t SymbolFilePDB::FindGlobalVariables(
793     const lldb_private::ConstString &name,
794     const lldb_private::CompilerDeclContext *parent_decl_ctx, bool append,
795     uint32_t max_matches, lldb_private::VariableList &variables) {
796   return uint32_t();
797 }
798 
799 uint32_t
800 SymbolFilePDB::FindGlobalVariables(const lldb_private::RegularExpression &regex,
801                                    bool append, uint32_t max_matches,
802                                    lldb_private::VariableList &variables) {
803   return uint32_t();
804 }
805 
806 bool SymbolFilePDB::ResolveFunction(llvm::pdb::PDBSymbolFunc *pdb_func,
807                                     bool include_inlines,
808                                     lldb_private::SymbolContextList &sc_list) {
809   if (!pdb_func)
810     return false;
811   lldb_private::SymbolContext sc;
812   auto file_vm_addr = pdb_func->getVirtualAddress();
813   if (file_vm_addr == LLDB_INVALID_ADDRESS)
814     return false;
815 
816   Address so_addr(file_vm_addr);
817   sc.comp_unit = GetCompileUnitContainsAddress(so_addr).get();
818   if (!sc.comp_unit)
819     return false;
820   sc.module_sp = sc.comp_unit->GetModule();
821   auto symbol_up =
822       m_session_up->findSymbolByAddress(file_vm_addr, PDB_SymType::Function);
823   if (!symbol_up)
824     return false;
825 
826   auto *func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get());
827   assert(func);
828   sc.function = ParseCompileUnitFunctionForPDBFunc(func, sc);
829   if (!sc.function)
830     return false;
831 
832   sc_list.Append(sc);
833   return true;
834 }
835 
836 bool SymbolFilePDB::ResolveFunction(uint32_t uid, bool include_inlines,
837                                     lldb_private::SymbolContextList &sc_list) {
838   auto pdb_func_up =
839       m_session_up->getConcreteSymbolById<PDBSymbolFunc>(uid);
840   if (!pdb_func_up && !(include_inlines && pdb_func_up->hasInlineAttribute()))
841     return false;
842   return ResolveFunction(pdb_func_up.get(), include_inlines, sc_list);
843 }
844 
845 void SymbolFilePDB::CacheFunctionNames() {
846   if (!m_func_full_names.IsEmpty())
847     return;
848 
849   std::map<uint64_t, uint32_t> addr_ids;
850 
851   if (auto results_up = m_global_scope_up->findAllChildren<PDBSymbolFunc>()) {
852     while (auto pdb_func_up = results_up->getNext()) {
853       if (pdb_func_up->isCompilerGenerated())
854         continue;
855 
856       auto name = pdb_func_up->getName();
857       auto demangled_name = pdb_func_up->getUndecoratedName();
858       if (name.empty() && demangled_name.empty())
859         continue;
860 
861       auto uid = pdb_func_up->getSymIndexId();
862       if (!demangled_name.empty() && pdb_func_up->getVirtualAddress())
863         addr_ids.insert(std::make_pair(pdb_func_up->getVirtualAddress(), uid));
864 
865       if (auto parent = pdb_func_up->getClassParent()) {
866 
867         // PDB have symbols for class/struct methods or static methods in Enum
868         // Class. We won't bother to check if the parent is UDT or Enum here.
869         m_func_method_names.Append(ConstString(name), uid);
870 
871         ConstString cstr_name(name);
872 
873         // To search a method name, like NS::Class:MemberFunc, LLDB searches its
874         // base name, i.e. MemberFunc by default. Since PDBSymbolFunc does not
875         // have inforamtion of this, we extract base names and cache them by our
876         // own effort.
877         llvm::StringRef basename;
878         CPlusPlusLanguage::MethodName cpp_method(cstr_name);
879         if (cpp_method.IsValid()) {
880           llvm::StringRef context;
881           basename = cpp_method.GetBasename();
882           if (basename.empty())
883             CPlusPlusLanguage::ExtractContextAndIdentifier(name.c_str(),
884                                                            context, basename);
885         }
886 
887         if (!basename.empty())
888           m_func_base_names.Append(ConstString(basename), uid);
889         else {
890           m_func_base_names.Append(ConstString(name), uid);
891         }
892 
893         if (!demangled_name.empty())
894           m_func_full_names.Append(ConstString(demangled_name), uid);
895 
896       } else {
897         // Handle not-method symbols.
898 
899         // The function name might contain namespace, or its lexical scope. It
900         // is not safe to get its base name by applying same scheme as we deal
901         // with the method names.
902         // FIXME: Remove namespace if function is static in a scope.
903         m_func_base_names.Append(ConstString(name), uid);
904 
905         if (name == "main") {
906           m_func_full_names.Append(ConstString(name), uid);
907 
908           if (!demangled_name.empty() && name != demangled_name) {
909             m_func_full_names.Append(ConstString(demangled_name), uid);
910             m_func_base_names.Append(ConstString(demangled_name), uid);
911           }
912         } else if (!demangled_name.empty()) {
913           m_func_full_names.Append(ConstString(demangled_name), uid);
914         } else {
915           m_func_full_names.Append(ConstString(name), uid);
916         }
917       }
918     }
919   }
920 
921   if (auto results_up =
922       m_global_scope_up->findAllChildren<PDBSymbolPublicSymbol>()) {
923     while (auto pub_sym_up = results_up->getNext()) {
924       if (!pub_sym_up->isFunction())
925         continue;
926       auto name = pub_sym_up->getName();
927       if (name.empty())
928         continue;
929 
930       if (CPlusPlusLanguage::IsCPPMangledName(name.c_str())) {
931         auto vm_addr = pub_sym_up->getVirtualAddress();
932 
933         // PDB public symbol has mangled name for its associated function.
934         if (vm_addr && addr_ids.find(vm_addr) != addr_ids.end()) {
935           // Cache mangled name.
936           m_func_full_names.Append(ConstString(name), addr_ids[vm_addr]);
937         }
938       }
939     }
940   }
941   // Sort them before value searching is working properly
942   m_func_full_names.Sort();
943   m_func_full_names.SizeToFit();
944   m_func_method_names.Sort();
945   m_func_method_names.SizeToFit();
946   m_func_base_names.Sort();
947   m_func_base_names.SizeToFit();
948 }
949 
950 uint32_t SymbolFilePDB::FindFunctions(
951     const lldb_private::ConstString &name,
952     const lldb_private::CompilerDeclContext *parent_decl_ctx,
953     uint32_t name_type_mask, bool include_inlines, bool append,
954     lldb_private::SymbolContextList &sc_list) {
955   if (!append)
956     sc_list.Clear();
957   lldbassert((name_type_mask & eFunctionNameTypeAuto) == 0);
958 
959   if (name_type_mask == eFunctionNameTypeNone)
960     return 0;
961   if (!DeclContextMatchesThisSymbolFile(parent_decl_ctx))
962     return 0;
963   if (name.IsEmpty())
964     return 0;
965 
966   auto old_size = sc_list.GetSize();
967   if (name_type_mask & eFunctionNameTypeFull ||
968       name_type_mask & eFunctionNameTypeBase ||
969       name_type_mask & eFunctionNameTypeMethod) {
970     CacheFunctionNames();
971 
972     std::set<uint32_t> resolved_ids;
973     auto ResolveFn = [include_inlines, &name, &sc_list, &resolved_ids, this] (
974         UniqueCStringMap<uint32_t> &Names)
975     {
976       std::vector<uint32_t> ids;
977       if (Names.GetValues(name, ids)) {
978         for (auto id : ids) {
979           if (resolved_ids.find(id) == resolved_ids.end()) {
980             if (ResolveFunction(id, include_inlines, sc_list))
981               resolved_ids.insert(id);
982           }
983         }
984       }
985     };
986     if (name_type_mask & eFunctionNameTypeFull) {
987       ResolveFn(m_func_full_names);
988     }
989     if (name_type_mask & eFunctionNameTypeBase) {
990       ResolveFn(m_func_base_names);
991     }
992     if (name_type_mask & eFunctionNameTypeMethod) {
993       ResolveFn(m_func_method_names);
994     }
995   }
996   return sc_list.GetSize() - old_size;
997 }
998 
999 uint32_t
1000 SymbolFilePDB::FindFunctions(const lldb_private::RegularExpression &regex,
1001                              bool include_inlines, bool append,
1002                              lldb_private::SymbolContextList &sc_list) {
1003   if (!append)
1004     sc_list.Clear();
1005   if (!regex.IsValid())
1006     return 0;
1007 
1008   auto old_size = sc_list.GetSize();
1009   CacheFunctionNames();
1010 
1011   std::set<uint32_t> resolved_ids;
1012   auto ResolveFn = [&regex, include_inlines, &sc_list, &resolved_ids, this] (
1013       UniqueCStringMap<uint32_t> &Names)
1014   {
1015     std::vector<uint32_t> ids;
1016     if (Names.GetValues(regex, ids)) {
1017       for (auto id : ids) {
1018         if (resolved_ids.find(id) == resolved_ids.end())
1019           if (ResolveFunction(id, include_inlines, sc_list))
1020             resolved_ids.insert(id);
1021       }
1022     }
1023   };
1024   ResolveFn(m_func_full_names);
1025   ResolveFn(m_func_base_names);
1026 
1027   return sc_list.GetSize() - old_size;
1028 }
1029 
1030 void SymbolFilePDB::GetMangledNamesForFunction(
1031     const std::string &scope_qualified_name,
1032     std::vector<lldb_private::ConstString> &mangled_names) {}
1033 
1034 uint32_t SymbolFilePDB::FindTypes(
1035     const lldb_private::SymbolContext &sc,
1036     const lldb_private::ConstString &name,
1037     const lldb_private::CompilerDeclContext *parent_decl_ctx, bool append,
1038     uint32_t max_matches,
1039     llvm::DenseSet<lldb_private::SymbolFile *> &searched_symbol_files,
1040     lldb_private::TypeMap &types) {
1041   if (!append)
1042     types.Clear();
1043   if (!name)
1044     return 0;
1045   if (!DeclContextMatchesThisSymbolFile(parent_decl_ctx))
1046     return 0;
1047 
1048   searched_symbol_files.clear();
1049   searched_symbol_files.insert(this);
1050 
1051   std::string name_str = name.AsCString();
1052 
1053   // There is an assumption 'name' is not a regex
1054   FindTypesByName(name_str, max_matches, types);
1055 
1056   return types.GetSize();
1057 }
1058 
1059 void
1060 SymbolFilePDB::FindTypesByRegex(const lldb_private::RegularExpression &regex,
1061                                 uint32_t max_matches,
1062                                 lldb_private::TypeMap &types) {
1063   // When searching by regex, we need to go out of our way to limit the search
1064   // space as much as possible since this searches EVERYTHING in the PDB,
1065   // manually doing regex comparisons.  PDB library isn't optimized for regex
1066   // searches or searches across multiple symbol types at the same time, so the
1067   // best we can do is to search enums, then typedefs, then classes one by one,
1068   // and do a regex comparison against each of them.
1069   PDB_SymType tags_to_search[] = {PDB_SymType::Enum, PDB_SymType::Typedef,
1070                                   PDB_SymType::UDT};
1071   std::unique_ptr<IPDBEnumSymbols> results;
1072 
1073   uint32_t matches = 0;
1074 
1075   for (auto tag : tags_to_search) {
1076     results = m_global_scope_up->findAllChildren(tag);
1077     if (!results)
1078       continue;
1079 
1080     while (auto result = results->getNext()) {
1081       if (max_matches > 0 && matches >= max_matches)
1082         break;
1083 
1084       std::string type_name;
1085       if (auto enum_type = llvm::dyn_cast<PDBSymbolTypeEnum>(result.get()))
1086         type_name = enum_type->getName();
1087       else if (auto typedef_type =
1088                llvm::dyn_cast<PDBSymbolTypeTypedef>(result.get()))
1089         type_name = typedef_type->getName();
1090       else if (auto class_type = llvm::dyn_cast<PDBSymbolTypeUDT>(result.get()))
1091         type_name = class_type->getName();
1092       else {
1093         // We're looking only for types that have names.  Skip symbols, as well
1094         // as unnamed types such as arrays, pointers, etc.
1095         continue;
1096       }
1097 
1098       if (!regex.Execute(type_name))
1099         continue;
1100 
1101       // This should cause the type to get cached and stored in the `m_types`
1102       // lookup.
1103       if (!ResolveTypeUID(result->getSymIndexId()))
1104         continue;
1105 
1106       auto iter = m_types.find(result->getSymIndexId());
1107       if (iter == m_types.end())
1108         continue;
1109       types.Insert(iter->second);
1110       ++matches;
1111     }
1112   }
1113 }
1114 
1115 void SymbolFilePDB::FindTypesByName(const std::string &name,
1116                                     uint32_t max_matches,
1117                                     lldb_private::TypeMap &types) {
1118   std::unique_ptr<IPDBEnumSymbols> results;
1119   if (name.empty())
1120     return;
1121   results = m_global_scope_up->findChildren(PDB_SymType::None, name,
1122                                             PDB_NameSearchFlags::NS_Default);
1123   if (!results)
1124     return;
1125 
1126   uint32_t matches = 0;
1127 
1128   while (auto result = results->getNext()) {
1129     if (max_matches > 0 && matches >= max_matches)
1130       break;
1131     switch (result->getSymTag()) {
1132     case PDB_SymType::Enum:
1133     case PDB_SymType::UDT:
1134     case PDB_SymType::Typedef:
1135       break;
1136     default:
1137       // We're looking only for types that have names.  Skip symbols, as well as
1138       // unnamed types such as arrays, pointers, etc.
1139       continue;
1140     }
1141 
1142     // This should cause the type to get cached and stored in the `m_types`
1143     // lookup.
1144     if (!ResolveTypeUID(result->getSymIndexId()))
1145       continue;
1146 
1147     auto iter = m_types.find(result->getSymIndexId());
1148     if (iter == m_types.end())
1149       continue;
1150     types.Insert(iter->second);
1151     ++matches;
1152   }
1153 }
1154 
1155 size_t SymbolFilePDB::FindTypes(
1156     const std::vector<lldb_private::CompilerContext> &contexts, bool append,
1157     lldb_private::TypeMap &types) {
1158   return 0;
1159 }
1160 
1161 lldb_private::TypeList *SymbolFilePDB::GetTypeList() {
1162   return m_obj_file->GetModule()->GetTypeList();
1163 }
1164 
1165 void
1166 SymbolFilePDB::GetTypesForPDBSymbol(const llvm::pdb::PDBSymbol *pdb_symbol,
1167                                     uint32_t type_mask,
1168                                     TypeCollection &type_collection) {
1169   if (!pdb_symbol)
1170     return;
1171 
1172   bool can_parse = false;
1173   switch (pdb_symbol->getSymTag()) {
1174   case PDB_SymType::ArrayType:
1175     can_parse = ((type_mask & eTypeClassArray) != 0);
1176     break;
1177   case PDB_SymType::BuiltinType:
1178     can_parse = ((type_mask & eTypeClassBuiltin) != 0);
1179     break;
1180   case PDB_SymType::Enum:
1181     can_parse = ((type_mask & eTypeClassEnumeration) != 0);
1182     break;
1183   case PDB_SymType::Function:
1184   case PDB_SymType::FunctionSig:
1185     can_parse = ((type_mask & eTypeClassFunction) != 0);
1186     break;
1187   case PDB_SymType::PointerType:
1188     can_parse = ((type_mask & (eTypeClassPointer | eTypeClassBlockPointer |
1189                                eTypeClassMemberPointer)) != 0);
1190     break;
1191   case PDB_SymType::Typedef:
1192     can_parse = ((type_mask & eTypeClassTypedef) != 0);
1193     break;
1194   case PDB_SymType::UDT: {
1195     auto *udt = llvm::dyn_cast<PDBSymbolTypeUDT>(pdb_symbol);
1196     assert(udt);
1197     can_parse = (udt->getUdtKind() != PDB_UdtType::Interface &&
1198         ((type_mask & (eTypeClassClass | eTypeClassStruct |
1199                        eTypeClassUnion)) != 0));
1200   } break;
1201   default:break;
1202   }
1203 
1204   if (can_parse) {
1205     if (auto *type = ResolveTypeUID(pdb_symbol->getSymIndexId())) {
1206       auto result =
1207           std::find(type_collection.begin(), type_collection.end(), type);
1208       if (result == type_collection.end())
1209         type_collection.push_back(type);
1210     }
1211   }
1212 
1213   auto results_up = pdb_symbol->findAllChildren();
1214   while (auto symbol_up = results_up->getNext())
1215     GetTypesForPDBSymbol(symbol_up.get(), type_mask, type_collection);
1216 }
1217 
1218 size_t SymbolFilePDB::GetTypes(lldb_private::SymbolContextScope *sc_scope,
1219                                uint32_t type_mask,
1220                                lldb_private::TypeList &type_list) {
1221   TypeCollection type_collection;
1222   uint32_t old_size = type_list.GetSize();
1223   CompileUnit *cu = sc_scope ?
1224       sc_scope->CalculateSymbolContextCompileUnit() : nullptr;
1225   if (cu) {
1226     auto compiland_up = GetPDBCompilandByUID(cu->GetID());
1227     GetTypesForPDBSymbol(compiland_up.get(), type_mask, type_collection);
1228   } else {
1229     for (uint32_t cu_idx = 0; cu_idx < GetNumCompileUnits(); ++cu_idx) {
1230       auto cu_sp = ParseCompileUnitAtIndex(cu_idx);
1231       if (cu_sp.get()) {
1232         auto compiland_up = GetPDBCompilandByUID(cu_sp->GetID());
1233         GetTypesForPDBSymbol(compiland_up.get(), type_mask, type_collection);
1234       }
1235     }
1236   }
1237 
1238   for (auto type : type_collection) {
1239     type->GetForwardCompilerType();
1240     type_list.Insert(type->shared_from_this());
1241   }
1242   return type_list.GetSize() - old_size;
1243 }
1244 
1245 lldb_private::TypeSystem *
1246 SymbolFilePDB::GetTypeSystemForLanguage(lldb::LanguageType language) {
1247   auto type_system =
1248       m_obj_file->GetModule()->GetTypeSystemForLanguage(language);
1249   if (type_system)
1250     type_system->SetSymbolFile(this);
1251   return type_system;
1252 }
1253 
1254 lldb_private::CompilerDeclContext SymbolFilePDB::FindNamespace(
1255     const lldb_private::SymbolContext &sc,
1256     const lldb_private::ConstString &name,
1257     const lldb_private::CompilerDeclContext *parent_decl_ctx) {
1258   return lldb_private::CompilerDeclContext();
1259 }
1260 
1261 lldb_private::ConstString SymbolFilePDB::GetPluginName() {
1262   static ConstString g_name("pdb");
1263   return g_name;
1264 }
1265 
1266 uint32_t SymbolFilePDB::GetPluginVersion() { return 1; }
1267 
1268 IPDBSession &SymbolFilePDB::GetPDBSession() { return *m_session_up; }
1269 
1270 const IPDBSession &SymbolFilePDB::GetPDBSession() const {
1271   return *m_session_up;
1272 }
1273 
1274 lldb::CompUnitSP
1275 SymbolFilePDB::ParseCompileUnitForUID(uint32_t id, uint32_t index) {
1276   auto found_cu = m_comp_units.find(id);
1277   if (found_cu != m_comp_units.end())
1278     return found_cu->second;
1279 
1280   auto compiland_up = GetPDBCompilandByUID(id);
1281   if (!compiland_up)
1282     return CompUnitSP();
1283 
1284   lldb::LanguageType lang;
1285   auto details = compiland_up->findOneChild<PDBSymbolCompilandDetails>();
1286   if (!details)
1287     lang = lldb::eLanguageTypeC_plus_plus;
1288   else
1289     lang = TranslateLanguage(details->getLanguage());
1290 
1291   if (lang == lldb::LanguageType::eLanguageTypeUnknown)
1292     return CompUnitSP();
1293 
1294   std::string path = GetSourceFileNameForPDBCompiland(compiland_up.get());
1295   if (path.empty())
1296     return CompUnitSP();
1297 
1298   // Don't support optimized code for now, DebugInfoPDB does not return this
1299   // information.
1300   LazyBool optimized = eLazyBoolNo;
1301   auto cu_sp = std::make_shared<CompileUnit>(
1302       m_obj_file->GetModule(), nullptr, path.c_str(), id, lang, optimized);
1303 
1304   if (!cu_sp)
1305     return CompUnitSP();
1306 
1307   m_comp_units.insert(std::make_pair(id, cu_sp));
1308   if (index == UINT32_MAX)
1309     GetCompileUnitIndex(compiland_up.get(), index);
1310   lldbassert(index != UINT32_MAX);
1311   m_obj_file->GetModule()->GetSymbolVendor()->SetCompileUnitAtIndex(
1312       index, cu_sp);
1313   return cu_sp;
1314 }
1315 
1316 bool SymbolFilePDB::ParseCompileUnitLineTable(
1317     const lldb_private::SymbolContext &sc, uint32_t match_line) {
1318   lldbassert(sc.comp_unit);
1319 
1320   auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID());
1321   if (!compiland_up)
1322     return false;
1323 
1324   // LineEntry needs the *index* of the file into the list of support files
1325   // returned by ParseCompileUnitSupportFiles.  But the underlying SDK gives us
1326   // a globally unique idenfitifier in the namespace of the PDB.  So, we have to
1327   // do a mapping so that we can hand out indices.
1328   llvm::DenseMap<uint32_t, uint32_t> index_map;
1329   BuildSupportFileIdToSupportFileIndexMap(*compiland_up, index_map);
1330   auto line_table = llvm::make_unique<LineTable>(sc.comp_unit);
1331 
1332   // Find contributions to `compiland` from all source and header files.
1333   std::string path = sc.comp_unit->GetPath();
1334   auto files = m_session_up->getSourceFilesForCompiland(*compiland_up);
1335   if (!files)
1336     return false;
1337 
1338   // For each source and header file, create a LineSequence for contributions to
1339   // the compiland from that file, and add the sequence.
1340   while (auto file = files->getNext()) {
1341     std::unique_ptr<LineSequence> sequence(
1342         line_table->CreateLineSequenceContainer());
1343     auto lines = m_session_up->findLineNumbers(*compiland_up, *file);
1344     if (!lines)
1345       continue;
1346     int entry_count = lines->getChildCount();
1347 
1348     uint64_t prev_addr;
1349     uint32_t prev_length;
1350     uint32_t prev_line;
1351     uint32_t prev_source_idx;
1352 
1353     for (int i = 0; i < entry_count; ++i) {
1354       auto line = lines->getChildAtIndex(i);
1355 
1356       uint64_t lno = line->getLineNumber();
1357       uint64_t addr = line->getVirtualAddress();
1358       uint32_t length = line->getLength();
1359       uint32_t source_id = line->getSourceFileId();
1360       uint32_t col = line->getColumnNumber();
1361       uint32_t source_idx = index_map[source_id];
1362 
1363       // There was a gap between the current entry and the previous entry if the
1364       // addresses don't perfectly line up.
1365       bool is_gap = (i > 0) && (prev_addr + prev_length < addr);
1366 
1367       // Before inserting the current entry, insert a terminal entry at the end
1368       // of the previous entry's address range if the current entry resulted in
1369       // a gap from the previous entry.
1370       if (is_gap && ShouldAddLine(match_line, prev_line, prev_length)) {
1371         line_table->AppendLineEntryToSequence(
1372             sequence.get(), prev_addr + prev_length, prev_line, 0,
1373             prev_source_idx, false, false, false, false, true);
1374       }
1375 
1376       if (ShouldAddLine(match_line, lno, length)) {
1377         bool is_statement = line->isStatement();
1378         bool is_prologue = false;
1379         bool is_epilogue = false;
1380         auto func =
1381             m_session_up->findSymbolByAddress(addr, PDB_SymType::Function);
1382         if (func) {
1383           auto prologue = func->findOneChild<PDBSymbolFuncDebugStart>();
1384           if (prologue)
1385             is_prologue = (addr == prologue->getVirtualAddress());
1386 
1387           auto epilogue = func->findOneChild<PDBSymbolFuncDebugEnd>();
1388           if (epilogue)
1389             is_epilogue = (addr == epilogue->getVirtualAddress());
1390         }
1391 
1392         line_table->AppendLineEntryToSequence(sequence.get(), addr, lno, col,
1393                                               source_idx, is_statement, false,
1394                                               is_prologue, is_epilogue, false);
1395       }
1396 
1397       prev_addr = addr;
1398       prev_length = length;
1399       prev_line = lno;
1400       prev_source_idx = source_idx;
1401     }
1402 
1403     if (entry_count > 0 && ShouldAddLine(match_line, prev_line, prev_length)) {
1404       // The end is always a terminal entry, so insert it regardless.
1405       line_table->AppendLineEntryToSequence(
1406           sequence.get(), prev_addr + prev_length, prev_line, 0,
1407           prev_source_idx, false, false, false, false, true);
1408     }
1409 
1410     line_table->InsertSequence(sequence.release());
1411   }
1412 
1413   if (line_table->GetSize()) {
1414     sc.comp_unit->SetLineTable(line_table.release());
1415     return true;
1416   }
1417   return false;
1418 }
1419 
1420 void SymbolFilePDB::BuildSupportFileIdToSupportFileIndexMap(
1421     const PDBSymbolCompiland &compiland,
1422     llvm::DenseMap<uint32_t, uint32_t> &index_map) const {
1423   // This is a hack, but we need to convert the source id into an index into the
1424   // support files array.  We don't want to do path comparisons to avoid
1425   // basename / full path issues that may or may not even be a problem, so we
1426   // use the globally unique source file identifiers.  Ideally we could use the
1427   // global identifiers everywhere, but LineEntry currently assumes indices.
1428   auto source_files = m_session_up->getSourceFilesForCompiland(compiland);
1429   if (!source_files)
1430     return;
1431   int index = 0;
1432 
1433   while (auto file = source_files->getNext()) {
1434     uint32_t source_id = file->getUniqueId();
1435     index_map[source_id] = index++;
1436   }
1437 }
1438 
1439 lldb::CompUnitSP SymbolFilePDB::GetCompileUnitContainsAddress(
1440      const lldb_private::Address &so_addr) {
1441   lldb::addr_t file_vm_addr = so_addr.GetFileAddress();
1442   if (file_vm_addr == LLDB_INVALID_ADDRESS)
1443     return nullptr;
1444 
1445   auto lines_up =
1446       m_session_up->findLineNumbersByAddress(file_vm_addr, /*Length=*/200);
1447   if (!lines_up)
1448     return nullptr;
1449 
1450   auto first_line_up = lines_up->getNext();
1451   if (!first_line_up)
1452     return nullptr;
1453   auto compiland_up = GetPDBCompilandByUID(first_line_up->getCompilandId());
1454   if (compiland_up) {
1455     return ParseCompileUnitForUID(compiland_up->getSymIndexId());
1456   }
1457 
1458   return nullptr;
1459 }
1460 
1461 Mangled
1462 SymbolFilePDB::GetMangledForPDBFunc(const llvm::pdb::PDBSymbolFunc *pdb_func) {
1463   Mangled mangled;
1464   if (!pdb_func)
1465     return mangled;
1466 
1467   auto func_name = pdb_func->getName();
1468   auto func_undecorated_name = pdb_func->getUndecoratedName();
1469   std::string func_decorated_name;
1470 
1471   // Seek from public symbols for non-static function's decorated name if any.
1472   // For static functions, they don't have undecorated names and aren't exposed
1473   // in Public Symbols either.
1474   if (!func_undecorated_name.empty()) {
1475     auto result_up =
1476         m_global_scope_up->findChildren(PDB_SymType::PublicSymbol,
1477                                         func_undecorated_name,
1478                                         PDB_NameSearchFlags::NS_UndecoratedName);
1479     if (result_up) {
1480       while (auto symbol_up = result_up->getNext()) {
1481         // For a public symbol, it is unique.
1482         lldbassert(result_up->getChildCount() == 1);
1483         if (auto *pdb_public_sym =
1484             llvm::dyn_cast_or_null<PDBSymbolPublicSymbol>(symbol_up.get())) {
1485           if (pdb_public_sym->isFunction()) {
1486             func_decorated_name = pdb_public_sym->getName();
1487             break;
1488           }
1489         }
1490       }
1491     }
1492   }
1493   if (!func_decorated_name.empty()) {
1494     mangled.SetMangledName(ConstString(func_decorated_name));
1495 
1496     // For MSVC, format of C funciton's decorated name depends on calling
1497     // conventon. Unfortunately none of the format is recognized by current
1498     // LLDB. For example, `_purecall` is a __cdecl C function. From PDB,
1499     // `__purecall` is retrieved as both its decorated and
1500     // undecorated name (using PDBSymbolFunc::getUndecoratedName method).
1501     // However `__purecall` string is not treated as mangled in LLDB
1502     // (neither `?` nor `_Z` prefix). Mangled::GetDemangledName method
1503     // will fail internally and caches an empty string as its undecorated
1504     // name. So we will face a contradition here for the same symbol:
1505     //   non-empty undecorated name from PDB
1506     //   empty undecorated name from LLDB
1507     if (!func_undecorated_name.empty() &&
1508         mangled.GetDemangledName(mangled.GuessLanguage()).IsEmpty())
1509       mangled.SetDemangledName(ConstString(func_undecorated_name));
1510 
1511     // LLDB uses several flags to control how a C++ decorated name is
1512     // undecorated for MSVC. See `safeUndecorateName` in Class Mangled.
1513     // So the yielded name could be different from what we retrieve from
1514     // PDB source unless we also apply same flags in getting undecorated
1515     // name through PDBSymbolFunc::getUndecoratedNameEx method.
1516     if (!func_undecorated_name.empty() &&
1517         mangled.GetDemangledName(mangled.GuessLanguage()) !=
1518             ConstString(func_undecorated_name))
1519       mangled.SetDemangledName(ConstString(func_undecorated_name));
1520   } else if (!func_undecorated_name.empty()) {
1521     mangled.SetDemangledName(ConstString(func_undecorated_name));
1522   } else if (!func_name.empty())
1523     mangled.SetValue(ConstString(func_name), false);
1524 
1525   return mangled;
1526 }
1527 
1528 bool SymbolFilePDB::DeclContextMatchesThisSymbolFile(
1529     const lldb_private::CompilerDeclContext *decl_ctx) {
1530   if (decl_ctx == nullptr || !decl_ctx->IsValid())
1531     return true;
1532 
1533   TypeSystem *decl_ctx_type_system = decl_ctx->GetTypeSystem();
1534   if (!decl_ctx_type_system)
1535     return false;
1536   TypeSystem *type_system = GetTypeSystemForLanguage(
1537       decl_ctx_type_system->GetMinimumLanguage(nullptr));
1538   if (decl_ctx_type_system == type_system)
1539     return true; // The type systems match, return true
1540 
1541   return false;
1542 }
1543