1 //===-- ManualDWARFIndex.cpp ----------------------------------------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 
9 #include "Plugins/SymbolFile/DWARF/ManualDWARFIndex.h"
10 #include "Plugins/Language/ObjC/ObjCLanguage.h"
11 #include "Plugins/SymbolFile/DWARF/DWARFDebugInfo.h"
12 #include "Plugins/SymbolFile/DWARF/DWARFDeclContext.h"
13 #include "Plugins/SymbolFile/DWARF/LogChannelDWARF.h"
14 #include "Plugins/SymbolFile/DWARF/SymbolFileDWARFDwo.h"
15 #include "lldb/Core/Module.h"
16 #include "lldb/Host/TaskPool.h"
17 #include "lldb/Symbol/ObjectFile.h"
18 #include "lldb/Utility/Stream.h"
19 #include "lldb/Utility/Timer.h"
20 
21 using namespace lldb_private;
22 using namespace lldb;
23 
24 void ManualDWARFIndex::Index() {
25   if (!m_dwarf)
26     return;
27 
28   SymbolFileDWARF &main_dwarf = *m_dwarf;
29   m_dwarf = nullptr;
30 
31   static Timer::Category func_cat(LLVM_PRETTY_FUNCTION);
32   Timer scoped_timer(func_cat, "%p", static_cast<void *>(&main_dwarf));
33 
34   DWARFDebugInfo &main_info = main_dwarf.DebugInfo();
35   SymbolFileDWARFDwo *dwp_dwarf = main_dwarf.GetDwpSymbolFile().get();
36   DWARFDebugInfo *dwp_info = dwp_dwarf ? &dwp_dwarf->DebugInfo() : nullptr;
37 
38   std::vector<DWARFUnit *> units_to_index;
39   units_to_index.reserve(main_info.GetNumUnits() +
40                          (dwp_info ? dwp_info->GetNumUnits() : 0));
41 
42   // Process all units in the main file, as well as any type units in the dwp
43   // file. Type units in dwo files are handled when we reach the dwo file in
44   // IndexUnit.
45   for (size_t U = 0; U < main_info.GetNumUnits(); ++U) {
46     DWARFUnit *unit = main_info.GetUnitAtIndex(U);
47     if (unit && m_units_to_avoid.count(unit->GetOffset()) == 0)
48       units_to_index.push_back(unit);
49   }
50   if (dwp_info && dwp_info->ContainsTypeUnits()) {
51     for (size_t U = 0; U < dwp_info->GetNumUnits(); ++U) {
52       if (auto *tu = llvm::dyn_cast<DWARFTypeUnit>(dwp_info->GetUnitAtIndex(U)))
53         units_to_index.push_back(tu);
54     }
55   }
56 
57   if (units_to_index.empty())
58     return;
59 
60   std::vector<IndexSet> sets(units_to_index.size());
61 
62   // Keep memory down by clearing DIEs for any units if indexing
63   // caused us to load the unit's DIEs.
64   std::vector<llvm::Optional<DWARFUnit::ScopedExtractDIEs>> clear_cu_dies(
65       units_to_index.size());
66   auto parser_fn = [&](size_t cu_idx) {
67     IndexUnit(*units_to_index[cu_idx], dwp_dwarf, sets[cu_idx]);
68   };
69 
70   auto extract_fn = [&units_to_index, &clear_cu_dies](size_t cu_idx) {
71     clear_cu_dies[cu_idx] = units_to_index[cu_idx]->ExtractDIEsScoped();
72   };
73 
74   // Create a task runner that extracts dies for each DWARF unit in a
75   // separate thread
76   // First figure out which units didn't have their DIEs already
77   // parsed and remember this.  If no DIEs were parsed prior to this index
78   // function call, we are going to want to clear the CU dies after we are
79   // done indexing to make sure we don't pull in all DWARF dies, but we need
80   // to wait until all units have been indexed in case a DIE in one
81   // unit refers to another and the indexes accesses those DIEs.
82   TaskMapOverInt(0, units_to_index.size(), extract_fn);
83 
84   // Now create a task runner that can index each DWARF unit in a
85   // separate thread so we can index quickly.
86 
87   TaskMapOverInt(0, units_to_index.size(), parser_fn);
88 
89   auto finalize_fn = [this, &sets](NameToDIE(IndexSet::*index)) {
90     NameToDIE &result = m_set.*index;
91     for (auto &set : sets)
92       result.Append(set.*index);
93     result.Finalize();
94   };
95 
96   TaskPool::RunTasks([&]() { finalize_fn(&IndexSet::function_basenames); },
97                      [&]() { finalize_fn(&IndexSet::function_fullnames); },
98                      [&]() { finalize_fn(&IndexSet::function_methods); },
99                      [&]() { finalize_fn(&IndexSet::function_selectors); },
100                      [&]() { finalize_fn(&IndexSet::objc_class_selectors); },
101                      [&]() { finalize_fn(&IndexSet::globals); },
102                      [&]() { finalize_fn(&IndexSet::types); },
103                      [&]() { finalize_fn(&IndexSet::namespaces); });
104 }
105 
106 void ManualDWARFIndex::IndexUnit(DWARFUnit &unit, SymbolFileDWARFDwo *dwp,
107                                  IndexSet &set) {
108   Log *log = LogChannelDWARF::GetLogIfAll(DWARF_LOG_LOOKUPS);
109 
110   if (log) {
111     m_module.LogMessage(
112         log, "ManualDWARFIndex::IndexUnit for unit at .debug_info[0x%8.8x]",
113         unit.GetOffset());
114   }
115 
116   const LanguageType cu_language = SymbolFileDWARF::GetLanguage(unit);
117 
118   IndexUnitImpl(unit, cu_language, set);
119 
120   if (SymbolFileDWARFDwo *dwo_symbol_file = unit.GetDwoSymbolFile()) {
121     // Type units in a dwp file are indexed separately, so we just need to
122     // process the split unit here. However, if the split unit is in a dwo file,
123     // then we need to process type units here.
124     if (dwo_symbol_file == dwp) {
125       IndexUnitImpl(unit.GetNonSkeletonUnit(), cu_language, set);
126     } else {
127       DWARFDebugInfo &dwo_info = dwo_symbol_file->DebugInfo();
128       for (size_t i = 0; i < dwo_info.GetNumUnits(); ++i)
129         IndexUnitImpl(*dwo_info.GetUnitAtIndex(i), cu_language, set);
130     }
131   }
132 }
133 
134 void ManualDWARFIndex::IndexUnitImpl(DWARFUnit &unit,
135                                      const LanguageType cu_language,
136                                      IndexSet &set) {
137   for (const DWARFDebugInfoEntry &die : unit.dies()) {
138     const dw_tag_t tag = die.Tag();
139 
140     switch (tag) {
141     case DW_TAG_array_type:
142     case DW_TAG_base_type:
143     case DW_TAG_class_type:
144     case DW_TAG_constant:
145     case DW_TAG_enumeration_type:
146     case DW_TAG_inlined_subroutine:
147     case DW_TAG_namespace:
148     case DW_TAG_string_type:
149     case DW_TAG_structure_type:
150     case DW_TAG_subprogram:
151     case DW_TAG_subroutine_type:
152     case DW_TAG_typedef:
153     case DW_TAG_union_type:
154     case DW_TAG_unspecified_type:
155     case DW_TAG_variable:
156       break;
157 
158     default:
159       continue;
160     }
161 
162     DWARFAttributes attributes;
163     const char *name = nullptr;
164     const char *mangled_cstr = nullptr;
165     bool is_declaration = false;
166     // bool is_artificial = false;
167     bool has_address = false;
168     bool has_location_or_const_value = false;
169     bool is_global_or_static_variable = false;
170 
171     DWARFFormValue specification_die_form;
172     const size_t num_attributes = die.GetAttributes(&unit, attributes);
173     if (num_attributes > 0) {
174       for (uint32_t i = 0; i < num_attributes; ++i) {
175         dw_attr_t attr = attributes.AttributeAtIndex(i);
176         DWARFFormValue form_value;
177         switch (attr) {
178         case DW_AT_name:
179           if (attributes.ExtractFormValueAtIndex(i, form_value))
180             name = form_value.AsCString();
181           break;
182 
183         case DW_AT_declaration:
184           if (attributes.ExtractFormValueAtIndex(i, form_value))
185             is_declaration = form_value.Unsigned() != 0;
186           break;
187 
188         case DW_AT_MIPS_linkage_name:
189         case DW_AT_linkage_name:
190           if (attributes.ExtractFormValueAtIndex(i, form_value))
191             mangled_cstr = form_value.AsCString();
192           break;
193 
194         case DW_AT_low_pc:
195         case DW_AT_high_pc:
196         case DW_AT_ranges:
197           has_address = true;
198           break;
199 
200         case DW_AT_entry_pc:
201           has_address = true;
202           break;
203 
204         case DW_AT_location:
205         case DW_AT_const_value:
206           has_location_or_const_value = true;
207           is_global_or_static_variable = die.IsGlobalOrStaticVariable();
208 
209           break;
210 
211         case DW_AT_specification:
212           if (attributes.ExtractFormValueAtIndex(i, form_value))
213             specification_die_form = form_value;
214           break;
215         }
216       }
217     }
218 
219     DIERef ref = *DWARFDIE(&unit, &die).GetDIERef();
220     switch (tag) {
221     case DW_TAG_inlined_subroutine:
222     case DW_TAG_subprogram:
223       if (has_address) {
224         if (name) {
225           bool is_objc_method = false;
226           if (cu_language == eLanguageTypeObjC ||
227               cu_language == eLanguageTypeObjC_plus_plus) {
228             ObjCLanguage::MethodName objc_method(name, true);
229             if (objc_method.IsValid(true)) {
230               is_objc_method = true;
231               ConstString class_name_with_category(
232                   objc_method.GetClassNameWithCategory());
233               ConstString objc_selector_name(objc_method.GetSelector());
234               ConstString objc_fullname_no_category_name(
235                   objc_method.GetFullNameWithoutCategory(true));
236               ConstString class_name_no_category(objc_method.GetClassName());
237               set.function_fullnames.Insert(ConstString(name), ref);
238               if (class_name_with_category)
239                 set.objc_class_selectors.Insert(class_name_with_category, ref);
240               if (class_name_no_category &&
241                   class_name_no_category != class_name_with_category)
242                 set.objc_class_selectors.Insert(class_name_no_category, ref);
243               if (objc_selector_name)
244                 set.function_selectors.Insert(objc_selector_name, ref);
245               if (objc_fullname_no_category_name)
246                 set.function_fullnames.Insert(objc_fullname_no_category_name,
247                                               ref);
248             }
249           }
250           // If we have a mangled name, then the DW_AT_name attribute is
251           // usually the method name without the class or any parameters
252           bool is_method = DWARFDIE(&unit, &die).IsMethod();
253 
254           if (is_method)
255             set.function_methods.Insert(ConstString(name), ref);
256           else
257             set.function_basenames.Insert(ConstString(name), ref);
258 
259           if (!is_method && !mangled_cstr && !is_objc_method)
260             set.function_fullnames.Insert(ConstString(name), ref);
261         }
262         if (mangled_cstr) {
263           // Make sure our mangled name isn't the same string table entry as
264           // our name. If it starts with '_', then it is ok, else compare the
265           // string to make sure it isn't the same and we don't end up with
266           // duplicate entries
267           if (name && name != mangled_cstr &&
268               ((mangled_cstr[0] == '_') ||
269                (::strcmp(name, mangled_cstr) != 0))) {
270             set.function_fullnames.Insert(ConstString(mangled_cstr), ref);
271           }
272         }
273       }
274       break;
275 
276     case DW_TAG_array_type:
277     case DW_TAG_base_type:
278     case DW_TAG_class_type:
279     case DW_TAG_constant:
280     case DW_TAG_enumeration_type:
281     case DW_TAG_string_type:
282     case DW_TAG_structure_type:
283     case DW_TAG_subroutine_type:
284     case DW_TAG_typedef:
285     case DW_TAG_union_type:
286     case DW_TAG_unspecified_type:
287       if (name && !is_declaration)
288         set.types.Insert(ConstString(name), ref);
289       if (mangled_cstr && !is_declaration)
290         set.types.Insert(ConstString(mangled_cstr), ref);
291       break;
292 
293     case DW_TAG_namespace:
294       if (name)
295         set.namespaces.Insert(ConstString(name), ref);
296       break;
297 
298     case DW_TAG_variable:
299       if (name && has_location_or_const_value && is_global_or_static_variable) {
300         set.globals.Insert(ConstString(name), ref);
301         // Be sure to include variables by their mangled and demangled names if
302         // they have any since a variable can have a basename "i", a mangled
303         // named "_ZN12_GLOBAL__N_11iE" and a demangled mangled name
304         // "(anonymous namespace)::i"...
305 
306         // Make sure our mangled name isn't the same string table entry as our
307         // name. If it starts with '_', then it is ok, else compare the string
308         // to make sure it isn't the same and we don't end up with duplicate
309         // entries
310         if (mangled_cstr && name != mangled_cstr &&
311             ((mangled_cstr[0] == '_') || (::strcmp(name, mangled_cstr) != 0))) {
312           set.globals.Insert(ConstString(mangled_cstr), ref);
313         }
314       }
315       break;
316 
317     default:
318       continue;
319     }
320   }
321 }
322 
323 void ManualDWARFIndex::GetGlobalVariables(
324     ConstString basename, llvm::function_ref<bool(DIERef ref)> callback) {
325   Index();
326   m_set.globals.Find(basename, callback);
327 }
328 
329 void ManualDWARFIndex::GetGlobalVariables(
330     const RegularExpression &regex,
331     llvm::function_ref<bool(DIERef ref)> callback) {
332   Index();
333   m_set.globals.Find(regex, callback);
334 }
335 
336 void ManualDWARFIndex::GetGlobalVariables(
337     const DWARFUnit &unit, llvm::function_ref<bool(DIERef ref)> callback) {
338   Index();
339   m_set.globals.FindAllEntriesForUnit(unit, callback);
340 }
341 
342 void ManualDWARFIndex::GetObjCMethods(
343     ConstString class_name, llvm::function_ref<bool(DIERef ref)> callback) {
344   Index();
345   m_set.objc_class_selectors.Find(class_name, callback);
346 }
347 
348 void ManualDWARFIndex::GetCompleteObjCClass(
349     ConstString class_name, bool must_be_implementation,
350     llvm::function_ref<bool(DIERef ref)> callback) {
351   Index();
352   m_set.types.Find(class_name, callback);
353 }
354 
355 void ManualDWARFIndex::GetTypes(ConstString name,
356                                 llvm::function_ref<bool(DIERef ref)> callback) {
357   Index();
358   m_set.types.Find(name, callback);
359 }
360 
361 void ManualDWARFIndex::GetTypes(const DWARFDeclContext &context,
362                                 llvm::function_ref<bool(DIERef ref)> callback) {
363   Index();
364   m_set.types.Find(ConstString(context[0].name), callback);
365 }
366 
367 void ManualDWARFIndex::GetNamespaces(
368     ConstString name, llvm::function_ref<bool(DIERef ref)> callback) {
369   Index();
370   m_set.namespaces.Find(name, callback);
371 }
372 
373 void ManualDWARFIndex::GetFunctions(
374     ConstString name, SymbolFileDWARF &dwarf,
375     const CompilerDeclContext &parent_decl_ctx, uint32_t name_type_mask,
376     llvm::function_ref<bool(DWARFDIE die)> callback) {
377   Index();
378 
379   if (name_type_mask & eFunctionNameTypeFull) {
380     if (!m_set.function_fullnames.Find(name, [&](DIERef die_ref) {
381           DWARFDIE die = dwarf.GetDIE(die_ref);
382           if (!die)
383             return true;
384           if (!SymbolFileDWARF::DIEInDeclContext(parent_decl_ctx, die))
385             return true;
386           return callback(die);
387         }))
388       return;
389   }
390   if (name_type_mask & eFunctionNameTypeBase) {
391     if (!m_set.function_basenames.Find(name, [&](DIERef die_ref) {
392           DWARFDIE die = dwarf.GetDIE(die_ref);
393           if (!die)
394             return true;
395           if (!SymbolFileDWARF::DIEInDeclContext(parent_decl_ctx, die))
396             return true;
397           return callback(die);
398         }))
399       return;
400   }
401 
402   if (name_type_mask & eFunctionNameTypeMethod && !parent_decl_ctx.IsValid()) {
403     if (!m_set.function_methods.Find(name, [&](DIERef die_ref) {
404           DWARFDIE die = dwarf.GetDIE(die_ref);
405           if (!die)
406             return true;
407           return callback(die);
408         }))
409       return;
410   }
411 
412   if (name_type_mask & eFunctionNameTypeSelector &&
413       !parent_decl_ctx.IsValid()) {
414     if (!m_set.function_selectors.Find(name, [&](DIERef die_ref) {
415           DWARFDIE die = dwarf.GetDIE(die_ref);
416           if (!die)
417             return true;
418           return callback(die);
419         }))
420       return;
421   }
422 }
423 
424 void ManualDWARFIndex::GetFunctions(
425     const RegularExpression &regex,
426     llvm::function_ref<bool(DIERef ref)> callback) {
427   Index();
428 
429   if (!m_set.function_basenames.Find(regex, callback))
430     return;
431   if (!m_set.function_fullnames.Find(regex, callback))
432     return;
433 }
434 
435 void ManualDWARFIndex::Dump(Stream &s) {
436   s.Format("Manual DWARF index for ({0}) '{1:F}':",
437            m_module.GetArchitecture().GetArchitectureName(),
438            m_module.GetObjectFile()->GetFileSpec());
439   s.Printf("\nFunction basenames:\n");
440   m_set.function_basenames.Dump(&s);
441   s.Printf("\nFunction fullnames:\n");
442   m_set.function_fullnames.Dump(&s);
443   s.Printf("\nFunction methods:\n");
444   m_set.function_methods.Dump(&s);
445   s.Printf("\nFunction selectors:\n");
446   m_set.function_selectors.Dump(&s);
447   s.Printf("\nObjective-C class selectors:\n");
448   m_set.objc_class_selectors.Dump(&s);
449   s.Printf("\nGlobals and statics:\n");
450   m_set.globals.Dump(&s);
451   s.Printf("\nTypes:\n");
452   m_set.types.Dump(&s);
453   s.Printf("\nNamespaces:\n");
454   m_set.namespaces.Dump(&s);
455 }
456