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/TypeList.h"
23 #include "lldb/Symbol/TypeMap.h"
24 #include "lldb/Symbol/Variable.h"
25 #include "lldb/Utility/RegularExpression.h"
26 
27 #include "llvm/DebugInfo/PDB/GenericError.h"
28 #include "llvm/DebugInfo/PDB/IPDBDataStream.h"
29 #include "llvm/DebugInfo/PDB/IPDBEnumChildren.h"
30 #include "llvm/DebugInfo/PDB/IPDBLineNumber.h"
31 #include "llvm/DebugInfo/PDB/IPDBSectionContrib.h"
32 #include "llvm/DebugInfo/PDB/IPDBSourceFile.h"
33 #include "llvm/DebugInfo/PDB/IPDBTable.h"
34 #include "llvm/DebugInfo/PDB/PDBSymbol.h"
35 #include "llvm/DebugInfo/PDB/PDBSymbolBlock.h"
36 #include "llvm/DebugInfo/PDB/PDBSymbolCompiland.h"
37 #include "llvm/DebugInfo/PDB/PDBSymbolCompilandDetails.h"
38 #include "llvm/DebugInfo/PDB/PDBSymbolData.h"
39 #include "llvm/DebugInfo/PDB/PDBSymbolExe.h"
40 #include "llvm/DebugInfo/PDB/PDBSymbolFunc.h"
41 #include "llvm/DebugInfo/PDB/PDBSymbolFuncDebugEnd.h"
42 #include "llvm/DebugInfo/PDB/PDBSymbolFuncDebugStart.h"
43 #include "llvm/DebugInfo/PDB/PDBSymbolPublicSymbol.h"
44 #include "llvm/DebugInfo/PDB/PDBSymbolTypeEnum.h"
45 #include "llvm/DebugInfo/PDB/PDBSymbolTypeTypedef.h"
46 #include "llvm/DebugInfo/PDB/PDBSymbolTypeUDT.h"
47 
48 #include "Plugins/Language/CPlusPlus/CPlusPlusLanguage.h" // For IsCPPMangledName
49 #include "Plugins/SymbolFile/PDB/PDBASTParser.h"
50 #include "Plugins/SymbolFile/PDB/PDBLocationToDWARFExpression.h"
51 
52 #include <regex>
53 
54 using namespace lldb;
55 using namespace lldb_private;
56 using namespace llvm::pdb;
57 
58 namespace {
59 lldb::LanguageType TranslateLanguage(PDB_Lang lang) {
60   switch (lang) {
61   case PDB_Lang::Cpp:
62     return lldb::LanguageType::eLanguageTypeC_plus_plus;
63   case PDB_Lang::C:
64     return lldb::LanguageType::eLanguageTypeC;
65   default:
66     return lldb::LanguageType::eLanguageTypeUnknown;
67   }
68 }
69 
70 bool ShouldAddLine(uint32_t requested_line, uint32_t actual_line,
71                    uint32_t addr_length) {
72   return ((requested_line == 0 || actual_line == requested_line) &&
73           addr_length > 0);
74 }
75 } // namespace
76 
77 void SymbolFilePDB::Initialize() {
78   PluginManager::RegisterPlugin(GetPluginNameStatic(),
79                                 GetPluginDescriptionStatic(), CreateInstance,
80                                 DebuggerInitialize);
81 }
82 
83 void SymbolFilePDB::Terminate() {
84   PluginManager::UnregisterPlugin(CreateInstance);
85 }
86 
87 void SymbolFilePDB::DebuggerInitialize(lldb_private::Debugger &debugger) {}
88 
89 lldb_private::ConstString SymbolFilePDB::GetPluginNameStatic() {
90   static ConstString g_name("pdb");
91   return g_name;
92 }
93 
94 const char *SymbolFilePDB::GetPluginDescriptionStatic() {
95   return "Microsoft PDB debug symbol file reader.";
96 }
97 
98 lldb_private::SymbolFile *
99 SymbolFilePDB::CreateInstance(lldb_private::ObjectFile *obj_file) {
100   return new SymbolFilePDB(obj_file);
101 }
102 
103 SymbolFilePDB::SymbolFilePDB(lldb_private::ObjectFile *object_file)
104     : SymbolFile(object_file), m_session_up(), m_global_scope_up(),
105       m_cached_compile_unit_count(0), m_tu_decl_ctx_up() {}
106 
107 SymbolFilePDB::~SymbolFilePDB() {}
108 
109 uint32_t SymbolFilePDB::CalculateAbilities() {
110   uint32_t abilities = 0;
111   if (!m_obj_file)
112     return 0;
113 
114   if (!m_session_up) {
115     // Lazily load and match the PDB file, but only do this once.
116     std::string exePath = m_obj_file->GetFileSpec().GetPath();
117     auto error = loadDataForEXE(PDB_ReaderType::DIA, llvm::StringRef(exePath),
118                                 m_session_up);
119     if (error) {
120       llvm::consumeError(std::move(error));
121       auto module_sp = m_obj_file->GetModule();
122       if (!module_sp)
123         return 0;
124       // See if any symbol file is specified through `--symfile` option.
125       FileSpec symfile = module_sp->GetSymbolFileFileSpec();
126       if (!symfile)
127         return 0;
128       error = loadDataForPDB(PDB_ReaderType::DIA,
129                              llvm::StringRef(symfile.GetPath()), m_session_up);
130       if (error) {
131         llvm::consumeError(std::move(error));
132         return 0;
133       }
134     }
135   }
136   if (!m_session_up)
137     return 0;
138 
139   auto enum_tables_up = m_session_up->getEnumTables();
140   if (!enum_tables_up)
141     return 0;
142   while (auto table_up = enum_tables_up->getNext()) {
143     if (table_up->getItemCount() == 0)
144       continue;
145     auto type = table_up->getTableType();
146     switch (type) {
147     case PDB_TableType::Symbols:
148       // This table represents a store of symbols with types listed in
149       // PDBSym_Type
150       abilities |= (CompileUnits | Functions | Blocks | GlobalVariables |
151                     LocalVariables | VariableTypes);
152       break;
153     case PDB_TableType::LineNumbers:
154       abilities |= LineTables;
155       break;
156     default:
157       break;
158     }
159   }
160   return abilities;
161 }
162 
163 void SymbolFilePDB::InitializeObject() {
164   lldb::addr_t obj_load_address = m_obj_file->GetFileOffset();
165   lldbassert(obj_load_address && obj_load_address != LLDB_INVALID_ADDRESS);
166   m_session_up->setLoadAddress(obj_load_address);
167   if (!m_global_scope_up)
168     m_global_scope_up = m_session_up->getGlobalScope();
169   lldbassert(m_global_scope_up.get());
170 
171   TypeSystem *type_system =
172       GetTypeSystemForLanguage(lldb::eLanguageTypeC_plus_plus);
173   ClangASTContext *clang_type_system =
174       llvm::dyn_cast_or_null<ClangASTContext>(type_system);
175   lldbassert(clang_type_system);
176   m_tu_decl_ctx_up = llvm::make_unique<CompilerDeclContext>(
177       type_system, clang_type_system->GetTranslationUnitDecl());
178 }
179 
180 uint32_t SymbolFilePDB::GetNumCompileUnits() {
181   if (m_cached_compile_unit_count == 0) {
182     auto compilands = m_global_scope_up->findAllChildren<PDBSymbolCompiland>();
183     if (!compilands)
184       return 0;
185 
186     // The linker could link *.dll (compiland language = LINK), or import
187     // *.dll. For example, a compiland with name `Import:KERNEL32.dll` could be
188     // found as a child of the global scope (PDB executable). Usually, such
189     // compilands contain `thunk` symbols in which we are not interested for
190     // now. However we still count them in the compiland list. If we perform
191     // any compiland related activity, like finding symbols through
192     // llvm::pdb::IPDBSession methods, such compilands will all be searched
193     // automatically no matter whether we include them or not.
194     m_cached_compile_unit_count = compilands->getChildCount();
195 
196     // The linker can inject an additional "dummy" compilation unit into the
197     // PDB. Ignore this special compile unit for our purposes, if it is there.
198     // It is always the last one.
199     auto last_compiland_up =
200         compilands->getChildAtIndex(m_cached_compile_unit_count - 1);
201     lldbassert(last_compiland_up.get());
202     std::string name = last_compiland_up->getName();
203     if (name == "* Linker *")
204       --m_cached_compile_unit_count;
205   }
206   return m_cached_compile_unit_count;
207 }
208 
209 void SymbolFilePDB::GetCompileUnitIndex(
210     const llvm::pdb::PDBSymbolCompiland &pdb_compiland, uint32_t &index) {
211   auto results_up = m_global_scope_up->findAllChildren<PDBSymbolCompiland>();
212   if (!results_up)
213     return;
214   auto uid = pdb_compiland.getSymIndexId();
215   for (uint32_t cu_idx = 0; cu_idx < GetNumCompileUnits(); ++cu_idx) {
216     auto compiland_up = results_up->getChildAtIndex(cu_idx);
217     if (!compiland_up)
218       continue;
219     if (compiland_up->getSymIndexId() == uid) {
220       index = cu_idx;
221       return;
222     }
223   }
224   index = UINT32_MAX;
225   return;
226 }
227 
228 std::unique_ptr<llvm::pdb::PDBSymbolCompiland>
229 SymbolFilePDB::GetPDBCompilandByUID(uint32_t uid) {
230   return m_session_up->getConcreteSymbolById<PDBSymbolCompiland>(uid);
231 }
232 
233 lldb::CompUnitSP SymbolFilePDB::ParseCompileUnitAtIndex(uint32_t index) {
234   if (index >= GetNumCompileUnits())
235     return CompUnitSP();
236 
237   // Assuming we always retrieve same compilands listed in same order through
238   // `PDBSymbolExe::findAllChildren` method, otherwise using `index` to get a
239   // compile unit makes no sense.
240   auto results = m_global_scope_up->findAllChildren<PDBSymbolCompiland>();
241   if (!results)
242     return CompUnitSP();
243   auto compiland_up = results->getChildAtIndex(index);
244   if (!compiland_up)
245     return CompUnitSP();
246   return ParseCompileUnitForUID(compiland_up->getSymIndexId(), index);
247 }
248 
249 lldb::LanguageType
250 SymbolFilePDB::ParseCompileUnitLanguage(const lldb_private::SymbolContext &sc) {
251   // What fields should I expect to be filled out on the SymbolContext?  Is it
252   // safe to assume that `sc.comp_unit` is valid?
253   if (!sc.comp_unit)
254     return lldb::eLanguageTypeUnknown;
255 
256   auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID());
257   if (!compiland_up)
258     return lldb::eLanguageTypeUnknown;
259   auto details = compiland_up->findOneChild<PDBSymbolCompilandDetails>();
260   if (!details)
261     return lldb::eLanguageTypeUnknown;
262   return TranslateLanguage(details->getLanguage());
263 }
264 
265 lldb_private::Function *SymbolFilePDB::ParseCompileUnitFunctionForPDBFunc(
266     const PDBSymbolFunc &pdb_func, const lldb_private::SymbolContext &sc) {
267   lldbassert(sc.comp_unit && sc.module_sp.get());
268 
269   auto file_vm_addr = pdb_func.getVirtualAddress();
270   if (file_vm_addr == LLDB_INVALID_ADDRESS || file_vm_addr == 0)
271     return nullptr;
272 
273   auto func_length = pdb_func.getLength();
274   AddressRange func_range =
275       AddressRange(file_vm_addr, func_length, sc.module_sp->GetSectionList());
276   if (!func_range.GetBaseAddress().IsValid())
277     return nullptr;
278 
279   lldb_private::Type *func_type = ResolveTypeUID(pdb_func.getSymIndexId());
280   if (!func_type)
281     return nullptr;
282 
283   user_id_t func_type_uid = pdb_func.getSignatureId();
284 
285   Mangled mangled = GetMangledForPDBFunc(pdb_func);
286 
287   FunctionSP func_sp =
288       std::make_shared<Function>(sc.comp_unit, pdb_func.getSymIndexId(),
289                                  func_type_uid, mangled, func_type, func_range);
290 
291   sc.comp_unit->AddFunction(func_sp);
292   return func_sp.get();
293 }
294 
295 size_t SymbolFilePDB::ParseCompileUnitFunctions(
296     const lldb_private::SymbolContext &sc) {
297   lldbassert(sc.comp_unit);
298   size_t func_added = 0;
299   auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID());
300   if (!compiland_up)
301     return 0;
302   auto results_up = compiland_up->findAllChildren<PDBSymbolFunc>();
303   if (!results_up)
304     return 0;
305   while (auto pdb_func_up = results_up->getNext()) {
306     auto func_sp =
307         sc.comp_unit->FindFunctionByUID(pdb_func_up->getSymIndexId());
308     if (!func_sp) {
309       if (ParseCompileUnitFunctionForPDBFunc(*pdb_func_up, sc))
310         ++func_added;
311     }
312   }
313   return func_added;
314 }
315 
316 bool SymbolFilePDB::ParseCompileUnitLineTable(
317     const lldb_private::SymbolContext &sc) {
318   lldbassert(sc.comp_unit);
319   if (sc.comp_unit->GetLineTable())
320     return true;
321   return ParseCompileUnitLineTable(sc, 0);
322 }
323 
324 bool SymbolFilePDB::ParseCompileUnitDebugMacros(
325     const lldb_private::SymbolContext &sc) {
326   // PDB doesn't contain information about macros
327   return false;
328 }
329 
330 bool SymbolFilePDB::ParseCompileUnitSupportFiles(
331     const lldb_private::SymbolContext &sc,
332     lldb_private::FileSpecList &support_files) {
333   lldbassert(sc.comp_unit);
334 
335   // In theory this is unnecessary work for us, because all of this information
336   // is easily (and quickly) accessible from DebugInfoPDB, so caching it a
337   // second time seems like a waste.  Unfortunately, there's no good way around
338   // this short of a moderate refactor since SymbolVendor depends on being able
339   // to cache this list.
340   auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID());
341   if (!compiland_up)
342     return false;
343   auto files = m_session_up->getSourceFilesForCompiland(*compiland_up);
344   if (!files || files->getChildCount() == 0)
345     return false;
346 
347   while (auto file = files->getNext()) {
348     FileSpec spec(file->getFileName(), false, FileSpec::Style::windows);
349     support_files.AppendIfUnique(spec);
350   }
351 
352   // LLDB uses the DWARF-like file numeration (one based),
353   // the zeroth file is the compile unit itself
354   support_files.Insert(0, *sc.comp_unit);
355 
356   return true;
357 }
358 
359 bool SymbolFilePDB::ParseImportedModules(
360     const lldb_private::SymbolContext &sc,
361     std::vector<lldb_private::ConstString> &imported_modules) {
362   // PDB does not yet support module debug info
363   return false;
364 }
365 
366 static size_t ParseFunctionBlocksForPDBSymbol(
367     const lldb_private::SymbolContext &sc, uint64_t func_file_vm_addr,
368     const llvm::pdb::PDBSymbol *pdb_symbol, lldb_private::Block *parent_block,
369     bool is_top_parent) {
370   assert(pdb_symbol && parent_block);
371 
372   size_t num_added = 0;
373   switch (pdb_symbol->getSymTag()) {
374   case PDB_SymType::Block:
375   case PDB_SymType::Function: {
376     Block *block = nullptr;
377     auto &raw_sym = pdb_symbol->getRawSymbol();
378     if (auto *pdb_func = llvm::dyn_cast<PDBSymbolFunc>(pdb_symbol)) {
379       if (pdb_func->hasNoInlineAttribute())
380         break;
381       if (is_top_parent)
382         block = parent_block;
383       else
384         break;
385     } else if (llvm::dyn_cast<PDBSymbolBlock>(pdb_symbol)) {
386       auto uid = pdb_symbol->getSymIndexId();
387       if (parent_block->FindBlockByID(uid))
388         break;
389       if (raw_sym.getVirtualAddress() < func_file_vm_addr)
390         break;
391 
392       auto block_sp = std::make_shared<Block>(pdb_symbol->getSymIndexId());
393       parent_block->AddChild(block_sp);
394       block = block_sp.get();
395     } else
396       llvm_unreachable("Unexpected PDB symbol!");
397 
398     block->AddRange(Block::Range(
399         raw_sym.getVirtualAddress() - func_file_vm_addr, raw_sym.getLength()));
400     block->FinalizeRanges();
401     ++num_added;
402 
403     auto results_up = pdb_symbol->findAllChildren();
404     if (!results_up)
405       break;
406     while (auto symbol_up = results_up->getNext()) {
407       num_added += ParseFunctionBlocksForPDBSymbol(
408           sc, func_file_vm_addr, symbol_up.get(), block, false);
409     }
410   } break;
411   default:
412     break;
413   }
414   return num_added;
415 }
416 
417 size_t
418 SymbolFilePDB::ParseFunctionBlocks(const lldb_private::SymbolContext &sc) {
419   lldbassert(sc.comp_unit && sc.function);
420   size_t num_added = 0;
421   auto uid = sc.function->GetID();
422   auto pdb_func_up = m_session_up->getConcreteSymbolById<PDBSymbolFunc>(uid);
423   if (!pdb_func_up)
424     return 0;
425   Block &parent_block = sc.function->GetBlock(false);
426   num_added =
427       ParseFunctionBlocksForPDBSymbol(sc, pdb_func_up->getVirtualAddress(),
428                                       pdb_func_up.get(), &parent_block, true);
429   return num_added;
430 }
431 
432 size_t SymbolFilePDB::ParseTypes(const lldb_private::SymbolContext &sc) {
433   lldbassert(sc.module_sp.get());
434   if (!sc.comp_unit)
435     return 0;
436 
437   size_t num_added = 0;
438   auto compiland = GetPDBCompilandByUID(sc.comp_unit->GetID());
439   if (!compiland)
440     return 0;
441 
442   auto ParseTypesByTagFn = [&num_added, this](const PDBSymbol &raw_sym) {
443     std::unique_ptr<IPDBEnumSymbols> results;
444     PDB_SymType tags_to_search[] = {PDB_SymType::Enum, PDB_SymType::Typedef,
445                                     PDB_SymType::UDT};
446     for (auto tag : tags_to_search) {
447       results = raw_sym.findAllChildren(tag);
448       if (!results || results->getChildCount() == 0)
449         continue;
450       while (auto symbol = results->getNext()) {
451         switch (symbol->getSymTag()) {
452         case PDB_SymType::Enum:
453         case PDB_SymType::UDT:
454         case PDB_SymType::Typedef:
455           break;
456         default:
457           continue;
458         }
459 
460         // This should cause the type to get cached and stored in the `m_types`
461         // lookup.
462         if (auto type = ResolveTypeUID(symbol->getSymIndexId())) {
463           // Resolve the type completely to avoid a completion
464           // (and so a list change, which causes an iterators invalidation)
465           // during a TypeList dumping
466           type->GetFullCompilerType();
467           ++num_added;
468         }
469       }
470     }
471   };
472 
473   if (sc.function) {
474     auto pdb_func = m_session_up->getConcreteSymbolById<PDBSymbolFunc>(
475         sc.function->GetID());
476     if (!pdb_func)
477       return 0;
478     ParseTypesByTagFn(*pdb_func);
479   } else {
480     ParseTypesByTagFn(*compiland);
481 
482     // Also parse global types particularly coming from this compiland.
483     // Unfortunately, PDB has no compiland information for each global type. We
484     // have to parse them all. But ensure we only do this once.
485     static bool parse_all_global_types = false;
486     if (!parse_all_global_types) {
487       ParseTypesByTagFn(*m_global_scope_up);
488       parse_all_global_types = true;
489     }
490   }
491   return num_added;
492 }
493 
494 size_t
495 SymbolFilePDB::ParseVariablesForContext(const lldb_private::SymbolContext &sc) {
496   if (!sc.comp_unit)
497     return 0;
498 
499   size_t num_added = 0;
500   if (sc.function) {
501     auto pdb_func = m_session_up->getConcreteSymbolById<PDBSymbolFunc>(
502         sc.function->GetID());
503     if (!pdb_func)
504       return 0;
505 
506     num_added += ParseVariables(sc, *pdb_func);
507     sc.function->GetBlock(false).SetDidParseVariables(true, true);
508   } else if (sc.comp_unit) {
509     auto compiland = GetPDBCompilandByUID(sc.comp_unit->GetID());
510     if (!compiland)
511       return 0;
512 
513     if (sc.comp_unit->GetVariableList(false))
514       return 0;
515 
516     auto results = m_global_scope_up->findAllChildren<PDBSymbolData>();
517     if (results && results->getChildCount()) {
518       while (auto result = results->getNext()) {
519         auto cu_id = result->getCompilandId();
520         // FIXME: We are not able to determine variable's compile unit.
521         if (cu_id == 0)
522           continue;
523 
524         if (cu_id == sc.comp_unit->GetID())
525           num_added += ParseVariables(sc, *result);
526       }
527     }
528 
529     // FIXME: A `file static` or `global constant` variable appears both in
530     // compiland's children and global scope's children with unexpectedly
531     // different symbol's Id making it ambiguous.
532 
533     // FIXME: 'local constant', for example, const char var[] = "abc", declared
534     // in a function scope, can't be found in PDB.
535 
536     // Parse variables in this compiland.
537     num_added += ParseVariables(sc, *compiland);
538   }
539 
540   return num_added;
541 }
542 
543 lldb_private::Type *SymbolFilePDB::ResolveTypeUID(lldb::user_id_t type_uid) {
544   auto find_result = m_types.find(type_uid);
545   if (find_result != m_types.end())
546     return find_result->second.get();
547 
548   TypeSystem *type_system =
549       GetTypeSystemForLanguage(lldb::eLanguageTypeC_plus_plus);
550   ClangASTContext *clang_type_system =
551       llvm::dyn_cast_or_null<ClangASTContext>(type_system);
552   if (!clang_type_system)
553     return nullptr;
554   PDBASTParser *pdb =
555       llvm::dyn_cast<PDBASTParser>(clang_type_system->GetPDBParser());
556   if (!pdb)
557     return nullptr;
558 
559   auto pdb_type = m_session_up->getSymbolById(type_uid);
560   if (pdb_type == nullptr)
561     return nullptr;
562 
563   lldb::TypeSP result = pdb->CreateLLDBTypeFromPDBType(*pdb_type);
564   if (result) {
565     m_types.insert(std::make_pair(type_uid, result));
566     auto type_list = GetTypeList();
567     if (type_list)
568       type_list->Insert(result);
569   }
570   return result.get();
571 }
572 
573 bool SymbolFilePDB::CompleteType(lldb_private::CompilerType &compiler_type) {
574   std::lock_guard<std::recursive_mutex> guard(
575       GetObjectFile()->GetModule()->GetMutex());
576 
577   ClangASTContext *clang_ast_ctx = llvm::dyn_cast_or_null<ClangASTContext>(
578       GetTypeSystemForLanguage(lldb::eLanguageTypeC_plus_plus));
579   if (!clang_ast_ctx)
580     return false;
581 
582   PDBASTParser *pdb =
583       llvm::dyn_cast<PDBASTParser>(clang_ast_ctx->GetPDBParser());
584   if (!pdb)
585     return false;
586 
587   return pdb->CompleteTypeFromPDB(compiler_type);
588 }
589 
590 lldb_private::CompilerDecl SymbolFilePDB::GetDeclForUID(lldb::user_id_t uid) {
591   return lldb_private::CompilerDecl();
592 }
593 
594 lldb_private::CompilerDeclContext
595 SymbolFilePDB::GetDeclContextForUID(lldb::user_id_t uid) {
596   // PDB always uses the translation unit decl context for everything.  We can
597   // improve this later but it's not easy because PDB doesn't provide a high
598   // enough level of type fidelity in this area.
599   return *m_tu_decl_ctx_up;
600 }
601 
602 lldb_private::CompilerDeclContext
603 SymbolFilePDB::GetDeclContextContainingUID(lldb::user_id_t uid) {
604   return *m_tu_decl_ctx_up;
605 }
606 
607 void SymbolFilePDB::ParseDeclsForContext(
608     lldb_private::CompilerDeclContext decl_ctx) {}
609 
610 uint32_t
611 SymbolFilePDB::ResolveSymbolContext(const lldb_private::Address &so_addr,
612                                     uint32_t resolve_scope,
613                                     lldb_private::SymbolContext &sc) {
614   uint32_t resolved_flags = 0;
615   if (resolve_scope & eSymbolContextCompUnit ||
616       resolve_scope & eSymbolContextVariable ||
617       resolve_scope & eSymbolContextFunction ||
618       resolve_scope & eSymbolContextBlock ||
619       resolve_scope & eSymbolContextLineEntry) {
620     auto cu_sp = GetCompileUnitContainsAddress(so_addr);
621     if (!cu_sp) {
622       if (resolved_flags | eSymbolContextVariable) {
623         // TODO: Resolve variables
624       }
625       return 0;
626     }
627     sc.comp_unit = cu_sp.get();
628     resolved_flags |= eSymbolContextCompUnit;
629     lldbassert(sc.module_sp == cu_sp->GetModule());
630   }
631 
632   if (resolve_scope & eSymbolContextFunction) {
633     addr_t file_vm_addr = so_addr.GetFileAddress();
634     auto symbol_up =
635         m_session_up->findSymbolByAddress(file_vm_addr, PDB_SymType::Function);
636     if (symbol_up) {
637       auto *pdb_func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get());
638       assert(pdb_func);
639       auto func_uid = pdb_func->getSymIndexId();
640       sc.function = sc.comp_unit->FindFunctionByUID(func_uid).get();
641       if (sc.function == nullptr)
642         sc.function = ParseCompileUnitFunctionForPDBFunc(*pdb_func, sc);
643       if (sc.function) {
644         resolved_flags |= eSymbolContextFunction;
645         if (resolve_scope & eSymbolContextBlock) {
646           Block &block = sc.function->GetBlock(true);
647           sc.block = block.FindBlockByID(sc.function->GetID());
648           if (sc.block)
649             resolved_flags |= eSymbolContextBlock;
650         }
651       }
652     }
653   }
654 
655   if (resolve_scope & eSymbolContextLineEntry) {
656     if (auto *line_table = sc.comp_unit->GetLineTable()) {
657       Address addr(so_addr);
658       if (line_table->FindLineEntryByAddress(addr, sc.line_entry))
659         resolved_flags |= eSymbolContextLineEntry;
660     }
661   }
662 
663   return resolved_flags;
664 }
665 
666 uint32_t SymbolFilePDB::ResolveSymbolContext(
667     const lldb_private::FileSpec &file_spec, uint32_t line, bool check_inlines,
668     uint32_t resolve_scope, lldb_private::SymbolContextList &sc_list) {
669   const size_t old_size = sc_list.GetSize();
670   if (resolve_scope & lldb::eSymbolContextCompUnit) {
671     // Locate all compilation units with line numbers referencing the specified
672     // file.  For example, if `file_spec` is <vector>, then this should return
673     // all source files and header files that reference <vector>, either
674     // directly or indirectly.
675     auto compilands = m_session_up->findCompilandsForSourceFile(
676         file_spec.GetPath(), PDB_NameSearchFlags::NS_CaseInsensitive);
677 
678     if (!compilands)
679       return 0;
680 
681     // For each one, either find its previously parsed data or parse it afresh
682     // and add it to the symbol context list.
683     while (auto compiland = compilands->getNext()) {
684       // If we're not checking inlines, then don't add line information for
685       // this file unless the FileSpec matches. For inline functions, we don't
686       // have to match the FileSpec since they could be defined in headers
687       // other than file specified in FileSpec.
688       if (!check_inlines) {
689         std::string source_file = compiland->getSourceFileFullPath();
690         if (source_file.empty())
691           continue;
692         FileSpec this_spec(source_file, false, FileSpec::Style::windows);
693         bool need_full_match = !file_spec.GetDirectory().IsEmpty();
694         if (FileSpec::Compare(file_spec, this_spec, need_full_match) != 0)
695           continue;
696       }
697 
698       SymbolContext sc;
699       auto cu = ParseCompileUnitForUID(compiland->getSymIndexId());
700       if (!cu)
701         continue;
702       sc.comp_unit = cu.get();
703       sc.module_sp = cu->GetModule();
704 
705       // If we were asked to resolve line entries, add all entries to the line
706       // table that match the requested line (or all lines if `line` == 0).
707       if (resolve_scope & (eSymbolContextFunction | eSymbolContextBlock |
708                            eSymbolContextLineEntry)) {
709         bool has_line_table = ParseCompileUnitLineTable(sc, line);
710 
711         if ((resolve_scope & eSymbolContextLineEntry) && !has_line_table) {
712           // The query asks for line entries, but we can't get them for the
713           // compile unit. This is not normal for `line` = 0. So just assert
714           // it.
715           assert(line && "Couldn't get all line entries!\n");
716 
717           // Current compiland does not have the requested line. Search next.
718           continue;
719         }
720 
721         if (resolve_scope & (eSymbolContextFunction | eSymbolContextBlock)) {
722           if (!has_line_table)
723             continue;
724 
725           auto *line_table = sc.comp_unit->GetLineTable();
726           lldbassert(line_table);
727 
728           uint32_t num_line_entries = line_table->GetSize();
729           // Skip the terminal line entry.
730           --num_line_entries;
731 
732           // If `line `!= 0, see if we can resolve function for each line entry
733           // in the line table.
734           for (uint32_t line_idx = 0; line && line_idx < num_line_entries;
735                ++line_idx) {
736             if (!line_table->GetLineEntryAtIndex(line_idx, sc.line_entry))
737               continue;
738 
739             auto file_vm_addr =
740                 sc.line_entry.range.GetBaseAddress().GetFileAddress();
741             if (file_vm_addr == LLDB_INVALID_ADDRESS || file_vm_addr == 0)
742               continue;
743 
744             auto symbol_up = m_session_up->findSymbolByAddress(
745                 file_vm_addr, PDB_SymType::Function);
746             if (symbol_up) {
747               auto func_uid = symbol_up->getSymIndexId();
748               sc.function = sc.comp_unit->FindFunctionByUID(func_uid).get();
749               if (sc.function == nullptr) {
750                 auto pdb_func = llvm::dyn_cast<PDBSymbolFunc>(symbol_up.get());
751                 assert(pdb_func);
752                 sc.function = ParseCompileUnitFunctionForPDBFunc(*pdb_func, sc);
753               }
754               if (sc.function && (resolve_scope & eSymbolContextBlock)) {
755                 Block &block = sc.function->GetBlock(true);
756                 sc.block = block.FindBlockByID(sc.function->GetID());
757               }
758             }
759             sc_list.Append(sc);
760           }
761         } else if (has_line_table) {
762           // We can parse line table for the compile unit. But no query to
763           // resolve function or block. We append `sc` to the list anyway.
764           sc_list.Append(sc);
765         }
766       } else {
767         // No query for line entry, function or block. But we have a valid
768         // compile unit, append `sc` to the list.
769         sc_list.Append(sc);
770       }
771     }
772   }
773   return sc_list.GetSize() - old_size;
774 }
775 
776 std::string SymbolFilePDB::GetMangledForPDBData(const PDBSymbolData &pdb_data) {
777   std::string decorated_name;
778   auto vm_addr = pdb_data.getVirtualAddress();
779   if (vm_addr != LLDB_INVALID_ADDRESS && vm_addr) {
780     auto result_up =
781         m_global_scope_up->findAllChildren(PDB_SymType::PublicSymbol);
782     if (result_up) {
783       while (auto symbol_up = result_up->getNext()) {
784         if (symbol_up->getRawSymbol().getVirtualAddress() == vm_addr) {
785           decorated_name = symbol_up->getRawSymbol().getName();
786           break;
787         }
788       }
789     }
790   }
791   if (!decorated_name.empty())
792     return decorated_name;
793 
794   return std::string();
795 }
796 
797 VariableSP SymbolFilePDB::ParseVariableForPDBData(
798     const lldb_private::SymbolContext &sc,
799     const llvm::pdb::PDBSymbolData &pdb_data) {
800   VariableSP var_sp;
801   uint32_t var_uid = pdb_data.getSymIndexId();
802   auto result = m_variables.find(var_uid);
803   if (result != m_variables.end())
804     return result->second;
805 
806   ValueType scope = eValueTypeInvalid;
807   bool is_static_member = false;
808   bool is_external = false;
809   bool is_artificial = false;
810 
811   switch (pdb_data.getDataKind()) {
812   case PDB_DataKind::Global:
813     scope = eValueTypeVariableGlobal;
814     is_external = true;
815     break;
816   case PDB_DataKind::Local:
817     scope = eValueTypeVariableLocal;
818     break;
819   case PDB_DataKind::FileStatic:
820     scope = eValueTypeVariableStatic;
821     break;
822   case PDB_DataKind::StaticMember:
823     is_static_member = true;
824     scope = eValueTypeVariableStatic;
825     break;
826   case PDB_DataKind::Member:
827     scope = eValueTypeVariableStatic;
828     break;
829   case PDB_DataKind::Param:
830     scope = eValueTypeVariableArgument;
831     break;
832   case PDB_DataKind::Constant:
833     scope = eValueTypeConstResult;
834     break;
835   default:
836     break;
837   }
838 
839   switch (pdb_data.getLocationType()) {
840   case PDB_LocType::TLS:
841     scope = eValueTypeVariableThreadLocal;
842     break;
843   case PDB_LocType::RegRel: {
844     // It is a `this` pointer.
845     if (pdb_data.getDataKind() == PDB_DataKind::ObjectPtr) {
846       scope = eValueTypeVariableArgument;
847       is_artificial = true;
848     }
849   } break;
850   default:
851     break;
852   }
853 
854   Declaration decl;
855   if (!is_artificial && !pdb_data.isCompilerGenerated()) {
856     if (auto lines = pdb_data.getLineNumbers()) {
857       if (auto first_line = lines->getNext()) {
858         uint32_t src_file_id = first_line->getSourceFileId();
859         auto src_file = m_session_up->getSourceFileById(src_file_id);
860         if (src_file) {
861           FileSpec spec(src_file->getFileName(), /*resolve_path*/ false);
862           decl.SetFile(spec);
863           decl.SetColumn(first_line->getColumnNumber());
864           decl.SetLine(first_line->getLineNumber());
865         }
866       }
867     }
868   }
869 
870   Variable::RangeList ranges;
871   SymbolContextScope *context_scope = sc.comp_unit;
872   if (scope == eValueTypeVariableLocal) {
873     if (sc.function) {
874       context_scope = sc.function->GetBlock(true).FindBlockByID(
875           pdb_data.getClassParentId());
876       if (context_scope == nullptr)
877         context_scope = sc.function;
878     }
879   }
880 
881   SymbolFileTypeSP type_sp =
882       std::make_shared<SymbolFileType>(*this, pdb_data.getTypeId());
883 
884   auto var_name = pdb_data.getName();
885   auto mangled = GetMangledForPDBData(pdb_data);
886   auto mangled_cstr = mangled.empty() ? nullptr : mangled.c_str();
887 
888   bool is_constant;
889   DWARFExpression location = ConvertPDBLocationToDWARFExpression(
890       GetObjectFile()->GetModule(), pdb_data, is_constant);
891 
892   var_sp = std::make_shared<Variable>(
893       var_uid, var_name.c_str(), mangled_cstr, type_sp, scope, context_scope,
894       ranges, &decl, location, is_external, is_artificial, is_static_member);
895   var_sp->SetLocationIsConstantValueData(is_constant);
896 
897   m_variables.insert(std::make_pair(var_uid, var_sp));
898   return var_sp;
899 }
900 
901 size_t
902 SymbolFilePDB::ParseVariables(const lldb_private::SymbolContext &sc,
903                               const llvm::pdb::PDBSymbol &pdb_symbol,
904                               lldb_private::VariableList *variable_list) {
905   size_t num_added = 0;
906 
907   if (auto pdb_data = llvm::dyn_cast<PDBSymbolData>(&pdb_symbol)) {
908     VariableListSP local_variable_list_sp;
909 
910     auto result = m_variables.find(pdb_data->getSymIndexId());
911     if (result != m_variables.end()) {
912       if (variable_list)
913         variable_list->AddVariableIfUnique(result->second);
914     } else {
915       // Prepare right VariableList for this variable.
916       if (auto lexical_parent = pdb_data->getLexicalParent()) {
917         switch (lexical_parent->getSymTag()) {
918         case PDB_SymType::Exe:
919           assert(sc.comp_unit);
920           LLVM_FALLTHROUGH;
921         case PDB_SymType::Compiland: {
922           if (sc.comp_unit) {
923             local_variable_list_sp = sc.comp_unit->GetVariableList(false);
924             if (!local_variable_list_sp) {
925               local_variable_list_sp = std::make_shared<VariableList>();
926               sc.comp_unit->SetVariableList(local_variable_list_sp);
927             }
928           }
929         } break;
930         case PDB_SymType::Block:
931         case PDB_SymType::Function: {
932           if (sc.function) {
933             Block *block = sc.function->GetBlock(true).FindBlockByID(
934                 lexical_parent->getSymIndexId());
935             if (block) {
936               local_variable_list_sp = block->GetBlockVariableList(false);
937               if (!local_variable_list_sp) {
938                 local_variable_list_sp = std::make_shared<VariableList>();
939                 block->SetVariableList(local_variable_list_sp);
940               }
941             }
942           }
943         } break;
944         default:
945           break;
946         }
947       }
948 
949       if (local_variable_list_sp) {
950         if (auto var_sp = ParseVariableForPDBData(sc, *pdb_data)) {
951           local_variable_list_sp->AddVariableIfUnique(var_sp);
952           if (variable_list)
953             variable_list->AddVariableIfUnique(var_sp);
954           ++num_added;
955         }
956       }
957     }
958   }
959 
960   if (auto results = pdb_symbol.findAllChildren()) {
961     while (auto result = results->getNext())
962       num_added += ParseVariables(sc, *result, variable_list);
963   }
964 
965   return num_added;
966 }
967 
968 uint32_t SymbolFilePDB::FindGlobalVariables(
969     const lldb_private::ConstString &name,
970     const lldb_private::CompilerDeclContext *parent_decl_ctx,
971     uint32_t max_matches, lldb_private::VariableList &variables) {
972   if (!DeclContextMatchesThisSymbolFile(parent_decl_ctx))
973     return 0;
974   if (name.IsEmpty())
975     return 0;
976 
977   auto results =
978       m_global_scope_up->findChildren(PDB_SymType::Data, name.GetStringRef(),
979                                       PDB_NameSearchFlags::NS_CaseSensitive);
980   if (!results)
981     return 0;
982 
983   uint32_t matches = 0;
984   size_t old_size = variables.GetSize();
985   while (auto result = results->getNext()) {
986     auto pdb_data = llvm::dyn_cast<PDBSymbolData>(result.get());
987     if (max_matches > 0 && matches >= max_matches)
988       break;
989 
990     SymbolContext sc;
991     sc.module_sp = m_obj_file->GetModule();
992     lldbassert(sc.module_sp.get());
993 
994     sc.comp_unit = ParseCompileUnitForUID(pdb_data->getCompilandId()).get();
995     // FIXME: We are not able to determine the compile unit.
996     if (sc.comp_unit == nullptr)
997       continue;
998 
999     ParseVariables(sc, *pdb_data, &variables);
1000     matches = variables.GetSize() - old_size;
1001   }
1002 
1003   return matches;
1004 }
1005 
1006 uint32_t
1007 SymbolFilePDB::FindGlobalVariables(const lldb_private::RegularExpression &regex,
1008                                    uint32_t max_matches,
1009                                    lldb_private::VariableList &variables) {
1010   if (!regex.IsValid())
1011     return 0;
1012   auto results = m_global_scope_up->findAllChildren<PDBSymbolData>();
1013   if (!results)
1014     return 0;
1015 
1016   uint32_t matches = 0;
1017   size_t old_size = variables.GetSize();
1018   while (auto pdb_data = results->getNext()) {
1019     if (max_matches > 0 && matches >= max_matches)
1020       break;
1021 
1022     auto var_name = pdb_data->getName();
1023     if (var_name.empty())
1024       continue;
1025     if (!regex.Execute(var_name))
1026       continue;
1027     SymbolContext sc;
1028     sc.module_sp = m_obj_file->GetModule();
1029     lldbassert(sc.module_sp.get());
1030 
1031     sc.comp_unit = ParseCompileUnitForUID(pdb_data->getCompilandId()).get();
1032     // FIXME: We are not able to determine the compile unit.
1033     if (sc.comp_unit == nullptr)
1034       continue;
1035 
1036     ParseVariables(sc, *pdb_data, &variables);
1037     matches = variables.GetSize() - old_size;
1038   }
1039 
1040   return matches;
1041 }
1042 
1043 bool SymbolFilePDB::ResolveFunction(const llvm::pdb::PDBSymbolFunc &pdb_func,
1044                                     bool include_inlines,
1045                                     lldb_private::SymbolContextList &sc_list) {
1046   lldb_private::SymbolContext sc;
1047   sc.comp_unit = ParseCompileUnitForUID(pdb_func.getCompilandId()).get();
1048   if (!sc.comp_unit)
1049     return false;
1050   sc.module_sp = sc.comp_unit->GetModule();
1051   sc.function = ParseCompileUnitFunctionForPDBFunc(pdb_func, sc);
1052   if (!sc.function)
1053     return false;
1054 
1055   sc_list.Append(sc);
1056   return true;
1057 }
1058 
1059 bool SymbolFilePDB::ResolveFunction(uint32_t uid, bool include_inlines,
1060                                     lldb_private::SymbolContextList &sc_list) {
1061   auto pdb_func_up = m_session_up->getConcreteSymbolById<PDBSymbolFunc>(uid);
1062   if (!pdb_func_up && !(include_inlines && pdb_func_up->hasInlineAttribute()))
1063     return false;
1064   return ResolveFunction(*pdb_func_up, include_inlines, sc_list);
1065 }
1066 
1067 void SymbolFilePDB::CacheFunctionNames() {
1068   if (!m_func_full_names.IsEmpty())
1069     return;
1070 
1071   std::map<uint64_t, uint32_t> addr_ids;
1072 
1073   if (auto results_up = m_global_scope_up->findAllChildren<PDBSymbolFunc>()) {
1074     while (auto pdb_func_up = results_up->getNext()) {
1075       if (pdb_func_up->isCompilerGenerated())
1076         continue;
1077 
1078       auto name = pdb_func_up->getName();
1079       auto demangled_name = pdb_func_up->getUndecoratedName();
1080       if (name.empty() && demangled_name.empty())
1081         continue;
1082 
1083       auto uid = pdb_func_up->getSymIndexId();
1084       if (!demangled_name.empty() && pdb_func_up->getVirtualAddress())
1085         addr_ids.insert(std::make_pair(pdb_func_up->getVirtualAddress(), uid));
1086 
1087       if (auto parent = pdb_func_up->getClassParent()) {
1088 
1089         // PDB have symbols for class/struct methods or static methods in Enum
1090         // Class. We won't bother to check if the parent is UDT or Enum here.
1091         m_func_method_names.Append(ConstString(name), uid);
1092 
1093         ConstString cstr_name(name);
1094 
1095         // To search a method name, like NS::Class:MemberFunc, LLDB searches
1096         // its base name, i.e. MemberFunc by default. Since PDBSymbolFunc does
1097         // not have inforamtion of this, we extract base names and cache them
1098         // by our own effort.
1099         llvm::StringRef basename;
1100         CPlusPlusLanguage::MethodName cpp_method(cstr_name);
1101         if (cpp_method.IsValid()) {
1102           llvm::StringRef context;
1103           basename = cpp_method.GetBasename();
1104           if (basename.empty())
1105             CPlusPlusLanguage::ExtractContextAndIdentifier(name.c_str(),
1106                                                            context, basename);
1107         }
1108 
1109         if (!basename.empty())
1110           m_func_base_names.Append(ConstString(basename), uid);
1111         else {
1112           m_func_base_names.Append(ConstString(name), uid);
1113         }
1114 
1115         if (!demangled_name.empty())
1116           m_func_full_names.Append(ConstString(demangled_name), uid);
1117 
1118       } else {
1119         // Handle not-method symbols.
1120 
1121         // The function name might contain namespace, or its lexical scope. It
1122         // is not safe to get its base name by applying same scheme as we deal
1123         // with the method names.
1124         // FIXME: Remove namespace if function is static in a scope.
1125         m_func_base_names.Append(ConstString(name), uid);
1126 
1127         if (name == "main") {
1128           m_func_full_names.Append(ConstString(name), uid);
1129 
1130           if (!demangled_name.empty() && name != demangled_name) {
1131             m_func_full_names.Append(ConstString(demangled_name), uid);
1132             m_func_base_names.Append(ConstString(demangled_name), uid);
1133           }
1134         } else if (!demangled_name.empty()) {
1135           m_func_full_names.Append(ConstString(demangled_name), uid);
1136         } else {
1137           m_func_full_names.Append(ConstString(name), uid);
1138         }
1139       }
1140     }
1141   }
1142 
1143   if (auto results_up =
1144           m_global_scope_up->findAllChildren<PDBSymbolPublicSymbol>()) {
1145     while (auto pub_sym_up = results_up->getNext()) {
1146       if (!pub_sym_up->isFunction())
1147         continue;
1148       auto name = pub_sym_up->getName();
1149       if (name.empty())
1150         continue;
1151 
1152       if (CPlusPlusLanguage::IsCPPMangledName(name.c_str())) {
1153         auto vm_addr = pub_sym_up->getVirtualAddress();
1154 
1155         // PDB public symbol has mangled name for its associated function.
1156         if (vm_addr && addr_ids.find(vm_addr) != addr_ids.end()) {
1157           // Cache mangled name.
1158           m_func_full_names.Append(ConstString(name), addr_ids[vm_addr]);
1159         }
1160       }
1161     }
1162   }
1163   // Sort them before value searching is working properly
1164   m_func_full_names.Sort();
1165   m_func_full_names.SizeToFit();
1166   m_func_method_names.Sort();
1167   m_func_method_names.SizeToFit();
1168   m_func_base_names.Sort();
1169   m_func_base_names.SizeToFit();
1170 }
1171 
1172 uint32_t SymbolFilePDB::FindFunctions(
1173     const lldb_private::ConstString &name,
1174     const lldb_private::CompilerDeclContext *parent_decl_ctx,
1175     uint32_t name_type_mask, bool include_inlines, bool append,
1176     lldb_private::SymbolContextList &sc_list) {
1177   if (!append)
1178     sc_list.Clear();
1179   lldbassert((name_type_mask & eFunctionNameTypeAuto) == 0);
1180 
1181   if (name_type_mask == eFunctionNameTypeNone)
1182     return 0;
1183   if (!DeclContextMatchesThisSymbolFile(parent_decl_ctx))
1184     return 0;
1185   if (name.IsEmpty())
1186     return 0;
1187 
1188   auto old_size = sc_list.GetSize();
1189   if (name_type_mask & eFunctionNameTypeFull ||
1190       name_type_mask & eFunctionNameTypeBase ||
1191       name_type_mask & eFunctionNameTypeMethod) {
1192     CacheFunctionNames();
1193 
1194     std::set<uint32_t> resolved_ids;
1195     auto ResolveFn = [include_inlines, &name, &sc_list, &resolved_ids,
1196                       this](UniqueCStringMap<uint32_t> &Names) {
1197       std::vector<uint32_t> ids;
1198       if (Names.GetValues(name, ids)) {
1199         for (auto id : ids) {
1200           if (resolved_ids.find(id) == resolved_ids.end()) {
1201             if (ResolveFunction(id, include_inlines, sc_list))
1202               resolved_ids.insert(id);
1203           }
1204         }
1205       }
1206     };
1207     if (name_type_mask & eFunctionNameTypeFull) {
1208       ResolveFn(m_func_full_names);
1209     }
1210     if (name_type_mask & eFunctionNameTypeBase) {
1211       ResolveFn(m_func_base_names);
1212     }
1213     if (name_type_mask & eFunctionNameTypeMethod) {
1214       ResolveFn(m_func_method_names);
1215     }
1216   }
1217   return sc_list.GetSize() - old_size;
1218 }
1219 
1220 uint32_t
1221 SymbolFilePDB::FindFunctions(const lldb_private::RegularExpression &regex,
1222                              bool include_inlines, bool append,
1223                              lldb_private::SymbolContextList &sc_list) {
1224   if (!append)
1225     sc_list.Clear();
1226   if (!regex.IsValid())
1227     return 0;
1228 
1229   auto old_size = sc_list.GetSize();
1230   CacheFunctionNames();
1231 
1232   std::set<uint32_t> resolved_ids;
1233   auto ResolveFn = [&regex, include_inlines, &sc_list, &resolved_ids,
1234                     this](UniqueCStringMap<uint32_t> &Names) {
1235     std::vector<uint32_t> ids;
1236     if (Names.GetValues(regex, ids)) {
1237       for (auto id : ids) {
1238         if (resolved_ids.find(id) == resolved_ids.end())
1239           if (ResolveFunction(id, include_inlines, sc_list))
1240             resolved_ids.insert(id);
1241       }
1242     }
1243   };
1244   ResolveFn(m_func_full_names);
1245   ResolveFn(m_func_base_names);
1246 
1247   return sc_list.GetSize() - old_size;
1248 }
1249 
1250 void SymbolFilePDB::GetMangledNamesForFunction(
1251     const std::string &scope_qualified_name,
1252     std::vector<lldb_private::ConstString> &mangled_names) {}
1253 
1254 uint32_t SymbolFilePDB::FindTypes(
1255     const lldb_private::SymbolContext &sc,
1256     const lldb_private::ConstString &name,
1257     const lldb_private::CompilerDeclContext *parent_decl_ctx, bool append,
1258     uint32_t max_matches,
1259     llvm::DenseSet<lldb_private::SymbolFile *> &searched_symbol_files,
1260     lldb_private::TypeMap &types) {
1261   if (!append)
1262     types.Clear();
1263   if (!name)
1264     return 0;
1265   if (!DeclContextMatchesThisSymbolFile(parent_decl_ctx))
1266     return 0;
1267 
1268   searched_symbol_files.clear();
1269   searched_symbol_files.insert(this);
1270 
1271   std::string name_str = name.AsCString();
1272 
1273   // There is an assumption 'name' is not a regex
1274   FindTypesByName(name_str, max_matches, types);
1275 
1276   return types.GetSize();
1277 }
1278 
1279 void SymbolFilePDB::FindTypesByRegex(
1280     const lldb_private::RegularExpression &regex, uint32_t max_matches,
1281     lldb_private::TypeMap &types) {
1282   // When searching by regex, we need to go out of our way to limit the search
1283   // space as much as possible since this searches EVERYTHING in the PDB,
1284   // manually doing regex comparisons.  PDB library isn't optimized for regex
1285   // searches or searches across multiple symbol types at the same time, so the
1286   // best we can do is to search enums, then typedefs, then classes one by one,
1287   // and do a regex comparison against each of them.
1288   PDB_SymType tags_to_search[] = {PDB_SymType::Enum, PDB_SymType::Typedef,
1289                                   PDB_SymType::UDT};
1290   std::unique_ptr<IPDBEnumSymbols> results;
1291 
1292   uint32_t matches = 0;
1293 
1294   for (auto tag : tags_to_search) {
1295     results = m_global_scope_up->findAllChildren(tag);
1296     if (!results)
1297       continue;
1298 
1299     while (auto result = results->getNext()) {
1300       if (max_matches > 0 && matches >= max_matches)
1301         break;
1302 
1303       std::string type_name;
1304       if (auto enum_type = llvm::dyn_cast<PDBSymbolTypeEnum>(result.get()))
1305         type_name = enum_type->getName();
1306       else if (auto typedef_type =
1307                    llvm::dyn_cast<PDBSymbolTypeTypedef>(result.get()))
1308         type_name = typedef_type->getName();
1309       else if (auto class_type = llvm::dyn_cast<PDBSymbolTypeUDT>(result.get()))
1310         type_name = class_type->getName();
1311       else {
1312         // We're looking only for types that have names.  Skip symbols, as well
1313         // as unnamed types such as arrays, pointers, etc.
1314         continue;
1315       }
1316 
1317       if (!regex.Execute(type_name))
1318         continue;
1319 
1320       // This should cause the type to get cached and stored in the `m_types`
1321       // lookup.
1322       if (!ResolveTypeUID(result->getSymIndexId()))
1323         continue;
1324 
1325       auto iter = m_types.find(result->getSymIndexId());
1326       if (iter == m_types.end())
1327         continue;
1328       types.Insert(iter->second);
1329       ++matches;
1330     }
1331   }
1332 }
1333 
1334 void SymbolFilePDB::FindTypesByName(const std::string &name,
1335                                     uint32_t max_matches,
1336                                     lldb_private::TypeMap &types) {
1337   std::unique_ptr<IPDBEnumSymbols> results;
1338   if (name.empty())
1339     return;
1340   results = m_global_scope_up->findChildren(PDB_SymType::None, name,
1341                                             PDB_NameSearchFlags::NS_Default);
1342   if (!results)
1343     return;
1344 
1345   uint32_t matches = 0;
1346 
1347   while (auto result = results->getNext()) {
1348     if (max_matches > 0 && matches >= max_matches)
1349       break;
1350     switch (result->getSymTag()) {
1351     case PDB_SymType::Enum:
1352     case PDB_SymType::UDT:
1353     case PDB_SymType::Typedef:
1354       break;
1355     default:
1356       // We're looking only for types that have names.  Skip symbols, as well
1357       // as unnamed types such as arrays, pointers, etc.
1358       continue;
1359     }
1360 
1361     // This should cause the type to get cached and stored in the `m_types`
1362     // lookup.
1363     if (!ResolveTypeUID(result->getSymIndexId()))
1364       continue;
1365 
1366     auto iter = m_types.find(result->getSymIndexId());
1367     if (iter == m_types.end())
1368       continue;
1369     types.Insert(iter->second);
1370     ++matches;
1371   }
1372 }
1373 
1374 size_t SymbolFilePDB::FindTypes(
1375     const std::vector<lldb_private::CompilerContext> &contexts, bool append,
1376     lldb_private::TypeMap &types) {
1377   return 0;
1378 }
1379 
1380 lldb_private::TypeList *SymbolFilePDB::GetTypeList() {
1381   return m_obj_file->GetModule()->GetTypeList();
1382 }
1383 
1384 void SymbolFilePDB::GetTypesForPDBSymbol(const llvm::pdb::PDBSymbol &pdb_symbol,
1385                                          uint32_t type_mask,
1386                                          TypeCollection &type_collection) {
1387   bool can_parse = false;
1388   switch (pdb_symbol.getSymTag()) {
1389   case PDB_SymType::ArrayType:
1390     can_parse = ((type_mask & eTypeClassArray) != 0);
1391     break;
1392   case PDB_SymType::BuiltinType:
1393     can_parse = ((type_mask & eTypeClassBuiltin) != 0);
1394     break;
1395   case PDB_SymType::Enum:
1396     can_parse = ((type_mask & eTypeClassEnumeration) != 0);
1397     break;
1398   case PDB_SymType::Function:
1399   case PDB_SymType::FunctionSig:
1400     can_parse = ((type_mask & eTypeClassFunction) != 0);
1401     break;
1402   case PDB_SymType::PointerType:
1403     can_parse = ((type_mask & (eTypeClassPointer | eTypeClassBlockPointer |
1404                                eTypeClassMemberPointer)) != 0);
1405     break;
1406   case PDB_SymType::Typedef:
1407     can_parse = ((type_mask & eTypeClassTypedef) != 0);
1408     break;
1409   case PDB_SymType::UDT: {
1410     auto *udt = llvm::dyn_cast<PDBSymbolTypeUDT>(&pdb_symbol);
1411     assert(udt);
1412     can_parse = (udt->getUdtKind() != PDB_UdtType::Interface &&
1413                  ((type_mask & (eTypeClassClass | eTypeClassStruct |
1414                                 eTypeClassUnion)) != 0));
1415   } break;
1416   default:
1417     break;
1418   }
1419 
1420   if (can_parse) {
1421     if (auto *type = ResolveTypeUID(pdb_symbol.getSymIndexId())) {
1422       auto result =
1423           std::find(type_collection.begin(), type_collection.end(), type);
1424       if (result == type_collection.end())
1425         type_collection.push_back(type);
1426     }
1427   }
1428 
1429   auto results_up = pdb_symbol.findAllChildren();
1430   while (auto symbol_up = results_up->getNext())
1431     GetTypesForPDBSymbol(*symbol_up, type_mask, type_collection);
1432 }
1433 
1434 size_t SymbolFilePDB::GetTypes(lldb_private::SymbolContextScope *sc_scope,
1435                                uint32_t type_mask,
1436                                lldb_private::TypeList &type_list) {
1437   TypeCollection type_collection;
1438   uint32_t old_size = type_list.GetSize();
1439   CompileUnit *cu =
1440       sc_scope ? sc_scope->CalculateSymbolContextCompileUnit() : nullptr;
1441   if (cu) {
1442     auto compiland_up = GetPDBCompilandByUID(cu->GetID());
1443     if (!compiland_up)
1444       return 0;
1445     GetTypesForPDBSymbol(*compiland_up, type_mask, type_collection);
1446   } else {
1447     for (uint32_t cu_idx = 0; cu_idx < GetNumCompileUnits(); ++cu_idx) {
1448       auto cu_sp = ParseCompileUnitAtIndex(cu_idx);
1449       if (cu_sp) {
1450         if (auto compiland_up = GetPDBCompilandByUID(cu_sp->GetID()))
1451           GetTypesForPDBSymbol(*compiland_up, type_mask, type_collection);
1452       }
1453     }
1454   }
1455 
1456   for (auto type : type_collection) {
1457     type->GetForwardCompilerType();
1458     type_list.Insert(type->shared_from_this());
1459   }
1460   return type_list.GetSize() - old_size;
1461 }
1462 
1463 lldb_private::TypeSystem *
1464 SymbolFilePDB::GetTypeSystemForLanguage(lldb::LanguageType language) {
1465   auto type_system =
1466       m_obj_file->GetModule()->GetTypeSystemForLanguage(language);
1467   if (type_system)
1468     type_system->SetSymbolFile(this);
1469   return type_system;
1470 }
1471 
1472 lldb_private::CompilerDeclContext SymbolFilePDB::FindNamespace(
1473     const lldb_private::SymbolContext &sc,
1474     const lldb_private::ConstString &name,
1475     const lldb_private::CompilerDeclContext *parent_decl_ctx) {
1476   return lldb_private::CompilerDeclContext();
1477 }
1478 
1479 lldb_private::ConstString SymbolFilePDB::GetPluginName() {
1480   static ConstString g_name("pdb");
1481   return g_name;
1482 }
1483 
1484 uint32_t SymbolFilePDB::GetPluginVersion() { return 1; }
1485 
1486 IPDBSession &SymbolFilePDB::GetPDBSession() { return *m_session_up; }
1487 
1488 const IPDBSession &SymbolFilePDB::GetPDBSession() const {
1489   return *m_session_up;
1490 }
1491 
1492 lldb::CompUnitSP SymbolFilePDB::ParseCompileUnitForUID(uint32_t id,
1493                                                        uint32_t index) {
1494   auto found_cu = m_comp_units.find(id);
1495   if (found_cu != m_comp_units.end())
1496     return found_cu->second;
1497 
1498   auto compiland_up = GetPDBCompilandByUID(id);
1499   if (!compiland_up)
1500     return CompUnitSP();
1501 
1502   lldb::LanguageType lang;
1503   auto details = compiland_up->findOneChild<PDBSymbolCompilandDetails>();
1504   if (!details)
1505     lang = lldb::eLanguageTypeC_plus_plus;
1506   else
1507     lang = TranslateLanguage(details->getLanguage());
1508 
1509   if (lang == lldb::LanguageType::eLanguageTypeUnknown)
1510     return CompUnitSP();
1511 
1512   std::string path = compiland_up->getSourceFileFullPath();
1513   if (path.empty())
1514     return CompUnitSP();
1515 
1516   // Don't support optimized code for now, DebugInfoPDB does not return this
1517   // information.
1518   LazyBool optimized = eLazyBoolNo;
1519   auto cu_sp = std::make_shared<CompileUnit>(m_obj_file->GetModule(), nullptr,
1520                                              path.c_str(), id, lang, optimized);
1521 
1522   if (!cu_sp)
1523     return CompUnitSP();
1524 
1525   m_comp_units.insert(std::make_pair(id, cu_sp));
1526   if (index == UINT32_MAX)
1527     GetCompileUnitIndex(*compiland_up, index);
1528   lldbassert(index != UINT32_MAX);
1529   m_obj_file->GetModule()->GetSymbolVendor()->SetCompileUnitAtIndex(index,
1530                                                                     cu_sp);
1531   return cu_sp;
1532 }
1533 
1534 bool SymbolFilePDB::ParseCompileUnitLineTable(
1535     const lldb_private::SymbolContext &sc, uint32_t match_line) {
1536   lldbassert(sc.comp_unit);
1537 
1538   auto compiland_up = GetPDBCompilandByUID(sc.comp_unit->GetID());
1539   if (!compiland_up)
1540     return false;
1541 
1542   // LineEntry needs the *index* of the file into the list of support files
1543   // returned by ParseCompileUnitSupportFiles.  But the underlying SDK gives us
1544   // a globally unique idenfitifier in the namespace of the PDB.  So, we have
1545   // to do a mapping so that we can hand out indices.
1546   llvm::DenseMap<uint32_t, uint32_t> index_map;
1547   BuildSupportFileIdToSupportFileIndexMap(*compiland_up, index_map);
1548   auto line_table = llvm::make_unique<LineTable>(sc.comp_unit);
1549 
1550   // Find contributions to `compiland` from all source and header files.
1551   std::string path = sc.comp_unit->GetPath();
1552   auto files = m_session_up->getSourceFilesForCompiland(*compiland_up);
1553   if (!files)
1554     return false;
1555 
1556   // For each source and header file, create a LineSequence for contributions
1557   // to the compiland from that file, and add the sequence.
1558   while (auto file = files->getNext()) {
1559     std::unique_ptr<LineSequence> sequence(
1560         line_table->CreateLineSequenceContainer());
1561     auto lines = m_session_up->findLineNumbers(*compiland_up, *file);
1562     if (!lines)
1563       continue;
1564     int entry_count = lines->getChildCount();
1565 
1566     uint64_t prev_addr;
1567     uint32_t prev_length;
1568     uint32_t prev_line;
1569     uint32_t prev_source_idx;
1570 
1571     for (int i = 0; i < entry_count; ++i) {
1572       auto line = lines->getChildAtIndex(i);
1573 
1574       uint64_t lno = line->getLineNumber();
1575       uint64_t addr = line->getVirtualAddress();
1576       uint32_t length = line->getLength();
1577       uint32_t source_id = line->getSourceFileId();
1578       uint32_t col = line->getColumnNumber();
1579       uint32_t source_idx = index_map[source_id];
1580 
1581       // There was a gap between the current entry and the previous entry if
1582       // the addresses don't perfectly line up.
1583       bool is_gap = (i > 0) && (prev_addr + prev_length < addr);
1584 
1585       // Before inserting the current entry, insert a terminal entry at the end
1586       // of the previous entry's address range if the current entry resulted in
1587       // a gap from the previous entry.
1588       if (is_gap && ShouldAddLine(match_line, prev_line, prev_length)) {
1589         line_table->AppendLineEntryToSequence(
1590             sequence.get(), prev_addr + prev_length, prev_line, 0,
1591             prev_source_idx, false, false, false, false, true);
1592 
1593         line_table->InsertSequence(sequence.release());
1594         sequence.reset(line_table->CreateLineSequenceContainer());
1595       }
1596 
1597       if (ShouldAddLine(match_line, lno, length)) {
1598         bool is_statement = line->isStatement();
1599         bool is_prologue = false;
1600         bool is_epilogue = false;
1601         auto func =
1602             m_session_up->findSymbolByAddress(addr, PDB_SymType::Function);
1603         if (func) {
1604           auto prologue = func->findOneChild<PDBSymbolFuncDebugStart>();
1605           if (prologue)
1606             is_prologue = (addr == prologue->getVirtualAddress());
1607 
1608           auto epilogue = func->findOneChild<PDBSymbolFuncDebugEnd>();
1609           if (epilogue)
1610             is_epilogue = (addr == epilogue->getVirtualAddress());
1611         }
1612 
1613         line_table->AppendLineEntryToSequence(sequence.get(), addr, lno, col,
1614                                               source_idx, is_statement, false,
1615                                               is_prologue, is_epilogue, false);
1616       }
1617 
1618       prev_addr = addr;
1619       prev_length = length;
1620       prev_line = lno;
1621       prev_source_idx = source_idx;
1622     }
1623 
1624     if (entry_count > 0 && ShouldAddLine(match_line, prev_line, prev_length)) {
1625       // The end is always a terminal entry, so insert it regardless.
1626       line_table->AppendLineEntryToSequence(
1627           sequence.get(), prev_addr + prev_length, prev_line, 0,
1628           prev_source_idx, false, false, false, false, true);
1629     }
1630 
1631     line_table->InsertSequence(sequence.release());
1632   }
1633 
1634   if (line_table->GetSize()) {
1635     sc.comp_unit->SetLineTable(line_table.release());
1636     return true;
1637   }
1638   return false;
1639 }
1640 
1641 void SymbolFilePDB::BuildSupportFileIdToSupportFileIndexMap(
1642     const PDBSymbolCompiland &compiland,
1643     llvm::DenseMap<uint32_t, uint32_t> &index_map) const {
1644   // This is a hack, but we need to convert the source id into an index into
1645   // the support files array.  We don't want to do path comparisons to avoid
1646   // basename / full path issues that may or may not even be a problem, so we
1647   // use the globally unique source file identifiers.  Ideally we could use the
1648   // global identifiers everywhere, but LineEntry currently assumes indices.
1649   auto source_files = m_session_up->getSourceFilesForCompiland(compiland);
1650   if (!source_files)
1651     return;
1652 
1653   // LLDB uses the DWARF-like file numeration (one based)
1654   int index = 1;
1655 
1656   while (auto file = source_files->getNext()) {
1657     uint32_t source_id = file->getUniqueId();
1658     index_map[source_id] = index++;
1659   }
1660 }
1661 
1662 lldb::CompUnitSP SymbolFilePDB::GetCompileUnitContainsAddress(
1663     const lldb_private::Address &so_addr) {
1664   lldb::addr_t file_vm_addr = so_addr.GetFileAddress();
1665   if (file_vm_addr == LLDB_INVALID_ADDRESS || file_vm_addr == 0)
1666     return nullptr;
1667 
1668   // If it is a PDB function's vm addr, this is the first sure bet.
1669   if (auto lines =
1670           m_session_up->findLineNumbersByAddress(file_vm_addr, /*Length=*/1)) {
1671     if (auto first_line = lines->getNext())
1672       return ParseCompileUnitForUID(first_line->getCompilandId());
1673   }
1674 
1675   // Otherwise we resort to section contributions.
1676   if (auto sec_contribs = m_session_up->getSectionContribs()) {
1677     while (auto section = sec_contribs->getNext()) {
1678       auto va = section->getVirtualAddress();
1679       if (file_vm_addr >= va && file_vm_addr < va + section->getLength())
1680         return ParseCompileUnitForUID(section->getCompilandId());
1681     }
1682   }
1683   return nullptr;
1684 }
1685 
1686 Mangled
1687 SymbolFilePDB::GetMangledForPDBFunc(const llvm::pdb::PDBSymbolFunc &pdb_func) {
1688   Mangled mangled;
1689   auto func_name = pdb_func.getName();
1690   auto func_undecorated_name = pdb_func.getUndecoratedName();
1691   std::string func_decorated_name;
1692 
1693   // Seek from public symbols for non-static function's decorated name if any.
1694   // For static functions, they don't have undecorated names and aren't exposed
1695   // in Public Symbols either.
1696   if (!func_undecorated_name.empty()) {
1697     auto result_up = m_global_scope_up->findChildren(
1698         PDB_SymType::PublicSymbol, func_undecorated_name,
1699         PDB_NameSearchFlags::NS_UndecoratedName);
1700     if (result_up) {
1701       while (auto symbol_up = result_up->getNext()) {
1702         // For a public symbol, it is unique.
1703         lldbassert(result_up->getChildCount() == 1);
1704         if (auto *pdb_public_sym =
1705                 llvm::dyn_cast_or_null<PDBSymbolPublicSymbol>(
1706                     symbol_up.get())) {
1707           if (pdb_public_sym->isFunction()) {
1708             func_decorated_name = pdb_public_sym->getName();
1709             break;
1710           }
1711         }
1712       }
1713     }
1714   }
1715   if (!func_decorated_name.empty()) {
1716     mangled.SetMangledName(ConstString(func_decorated_name));
1717 
1718     // For MSVC, format of C funciton's decorated name depends on calling
1719     // conventon. Unfortunately none of the format is recognized by current
1720     // LLDB. For example, `_purecall` is a __cdecl C function. From PDB,
1721     // `__purecall` is retrieved as both its decorated and undecorated name
1722     // (using PDBSymbolFunc::getUndecoratedName method). However `__purecall`
1723     // string is not treated as mangled in LLDB (neither `?` nor `_Z` prefix).
1724     // Mangled::GetDemangledName method will fail internally and caches an
1725     // empty string as its undecorated name. So we will face a contradition
1726     // here for the same symbol:
1727     //   non-empty undecorated name from PDB
1728     //   empty undecorated name from LLDB
1729     if (!func_undecorated_name.empty() &&
1730         mangled.GetDemangledName(mangled.GuessLanguage()).IsEmpty())
1731       mangled.SetDemangledName(ConstString(func_undecorated_name));
1732 
1733     // LLDB uses several flags to control how a C++ decorated name is
1734     // undecorated for MSVC. See `safeUndecorateName` in Class Mangled. So the
1735     // yielded name could be different from what we retrieve from
1736     // PDB source unless we also apply same flags in getting undecorated
1737     // name through PDBSymbolFunc::getUndecoratedNameEx method.
1738     if (!func_undecorated_name.empty() &&
1739         mangled.GetDemangledName(mangled.GuessLanguage()) !=
1740             ConstString(func_undecorated_name))
1741       mangled.SetDemangledName(ConstString(func_undecorated_name));
1742   } else if (!func_undecorated_name.empty()) {
1743     mangled.SetDemangledName(ConstString(func_undecorated_name));
1744   } else if (!func_name.empty())
1745     mangled.SetValue(ConstString(func_name), false);
1746 
1747   return mangled;
1748 }
1749 
1750 bool SymbolFilePDB::DeclContextMatchesThisSymbolFile(
1751     const lldb_private::CompilerDeclContext *decl_ctx) {
1752   if (decl_ctx == nullptr || !decl_ctx->IsValid())
1753     return true;
1754 
1755   TypeSystem *decl_ctx_type_system = decl_ctx->GetTypeSystem();
1756   if (!decl_ctx_type_system)
1757     return false;
1758   TypeSystem *type_system = GetTypeSystemForLanguage(
1759       decl_ctx_type_system->GetMinimumLanguage(nullptr));
1760   if (decl_ctx_type_system == type_system)
1761     return true; // The type systems match, return true
1762 
1763   return false;
1764 }
1765