1 //===-- DWARFCompileUnit.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 "DWARFCompileUnit.h"
11 
12 #include "Plugins/Language/ObjC/ObjCLanguage.h"
13 #include "lldb/Core/Mangled.h"
14 #include "lldb/Core/Module.h"
15 #include "lldb/Core/Stream.h"
16 #include "lldb/Core/StreamString.h"
17 #include "lldb/Core/Timer.h"
18 #include "lldb/Host/StringConvert.h"
19 #include "lldb/Symbol/CompileUnit.h"
20 #include "lldb/Symbol/LineTable.h"
21 #include "lldb/Symbol/ObjectFile.h"
22 
23 #include "DWARFDIECollection.h"
24 #include "DWARFDebugAbbrev.h"
25 #include "DWARFDebugAranges.h"
26 #include "DWARFDebugInfo.h"
27 #include "DWARFFormValue.h"
28 #include "LogChannelDWARF.h"
29 #include "NameToDIE.h"
30 #include "SymbolFileDWARF.h"
31 #include "SymbolFileDWARFDebugMap.h"
32 #include "SymbolFileDWARFDwo.h"
33 
34 using namespace lldb;
35 using namespace lldb_private;
36 using namespace std;
37 
38 extern int g_verbose;
39 
40 DWARFCompileUnit::DWARFCompileUnit(SymbolFileDWARF *dwarf2Data)
41     : m_dwarf2Data(dwarf2Data), m_abbrevs(NULL), m_user_data(NULL),
42       m_die_array(), m_func_aranges_ap(), m_base_addr(0),
43       m_offset(DW_INVALID_OFFSET), m_length(0), m_version(0),
44       m_addr_size(DWARFCompileUnit::GetDefaultAddressSize()),
45       m_producer(eProducerInvalid), m_producer_version_major(0),
46       m_producer_version_minor(0), m_producer_version_update(0),
47       m_language_type(eLanguageTypeUnknown), m_is_dwarf64(false),
48       m_is_optimized(eLazyBoolCalculate), m_addr_base(0),
49       m_ranges_base(0), m_base_obj_offset(DW_INVALID_OFFSET) {}
50 
51 DWARFCompileUnit::~DWARFCompileUnit() {}
52 
53 void DWARFCompileUnit::Clear() {
54   m_offset = DW_INVALID_OFFSET;
55   m_length = 0;
56   m_version = 0;
57   m_abbrevs = NULL;
58   m_addr_size = DWARFCompileUnit::GetDefaultAddressSize();
59   m_base_addr = 0;
60   m_die_array.clear();
61   m_func_aranges_ap.reset();
62   m_user_data = NULL;
63   m_producer = eProducerInvalid;
64   m_language_type = eLanguageTypeUnknown;
65   m_is_dwarf64 = false;
66   m_is_optimized = eLazyBoolCalculate;
67   m_addr_base = 0;
68   m_base_obj_offset = DW_INVALID_OFFSET;
69 }
70 
71 bool DWARFCompileUnit::Extract(const DWARFDataExtractor &debug_info,
72                                lldb::offset_t *offset_ptr) {
73   Clear();
74 
75   m_offset = *offset_ptr;
76 
77   if (debug_info.ValidOffset(*offset_ptr)) {
78     dw_offset_t abbr_offset;
79     const DWARFDebugAbbrev *abbr = m_dwarf2Data->DebugAbbrev();
80     m_length = debug_info.GetDWARFInitialLength(offset_ptr);
81     m_is_dwarf64 = debug_info.IsDWARF64();
82     m_version = debug_info.GetU16(offset_ptr);
83     abbr_offset = debug_info.GetDWARFOffset(offset_ptr);
84     m_addr_size = debug_info.GetU8(offset_ptr);
85 
86     bool length_OK = debug_info.ValidOffset(GetNextCompileUnitOffset() - 1);
87     bool version_OK = SymbolFileDWARF::SupportedVersion(m_version);
88     bool abbr_offset_OK =
89         m_dwarf2Data->get_debug_abbrev_data().ValidOffset(abbr_offset);
90     bool addr_size_OK = ((m_addr_size == 4) || (m_addr_size == 8));
91 
92     if (length_OK && version_OK && addr_size_OK && abbr_offset_OK &&
93         abbr != NULL) {
94       m_abbrevs = abbr->GetAbbreviationDeclarationSet(abbr_offset);
95       return true;
96     }
97 
98     // reset the offset to where we tried to parse from if anything went wrong
99     *offset_ptr = m_offset;
100   }
101 
102   return false;
103 }
104 
105 void DWARFCompileUnit::ClearDIEs(bool keep_compile_unit_die) {
106   if (m_die_array.size() > 1) {
107     // std::vectors never get any smaller when resized to a smaller size,
108     // or when clear() or erase() are called, the size will report that it
109     // is smaller, but the memory allocated remains intact (call capacity()
110     // to see this). So we need to create a temporary vector and swap the
111     // contents which will cause just the internal pointers to be swapped
112     // so that when "tmp_array" goes out of scope, it will destroy the
113     // contents.
114 
115     // Save at least the compile unit DIE
116     DWARFDebugInfoEntry::collection tmp_array;
117     m_die_array.swap(tmp_array);
118     if (keep_compile_unit_die)
119       m_die_array.push_back(tmp_array.front());
120   }
121 
122   if (m_dwo_symbol_file)
123     m_dwo_symbol_file->GetCompileUnit()->ClearDIEs(keep_compile_unit_die);
124 }
125 
126 //----------------------------------------------------------------------
127 // ParseCompileUnitDIEsIfNeeded
128 //
129 // Parses a compile unit and indexes its DIEs if it hasn't already been
130 // done.
131 //----------------------------------------------------------------------
132 size_t DWARFCompileUnit::ExtractDIEsIfNeeded(bool cu_die_only) {
133   const size_t initial_die_array_size = m_die_array.size();
134   if ((cu_die_only && initial_die_array_size > 0) || initial_die_array_size > 1)
135     return 0; // Already parsed
136 
137   Timer scoped_timer(
138       LLVM_PRETTY_FUNCTION,
139       "%8.8x: DWARFCompileUnit::ExtractDIEsIfNeeded( cu_die_only = %i )",
140       m_offset, cu_die_only);
141 
142   // Set the offset to that of the first DIE and calculate the start of the
143   // next compilation unit header.
144   lldb::offset_t offset = GetFirstDIEOffset();
145   lldb::offset_t next_cu_offset = GetNextCompileUnitOffset();
146 
147   DWARFDebugInfoEntry die;
148   // Keep a flat array of the DIE for binary lookup by DIE offset
149   if (!cu_die_only) {
150     Log *log(
151         LogChannelDWARF::GetLogIfAny(DWARF_LOG_DEBUG_INFO | DWARF_LOG_LOOKUPS));
152     if (log) {
153       m_dwarf2Data->GetObjectFile()->GetModule()->LogMessageVerboseBacktrace(
154           log, "DWARFCompileUnit::ExtractDIEsIfNeeded () for compile unit at "
155                ".debug_info[0x%8.8x]",
156           GetOffset());
157     }
158   }
159 
160   uint32_t depth = 0;
161   // We are in our compile unit, parse starting at the offset
162   // we were told to parse
163   const DWARFDataExtractor &debug_info_data =
164       m_dwarf2Data->get_debug_info_data();
165   std::vector<uint32_t> die_index_stack;
166   die_index_stack.reserve(32);
167   die_index_stack.push_back(0);
168   bool prev_die_had_children = false;
169   DWARFFormValue::FixedFormSizes fixed_form_sizes =
170       DWARFFormValue::GetFixedFormSizesForAddressSize(GetAddressByteSize(),
171                                                       m_is_dwarf64);
172   while (offset < next_cu_offset &&
173          die.FastExtract(debug_info_data, this, fixed_form_sizes, &offset)) {
174     //        if (log)
175     //            log->Printf("0x%8.8x: %*.*s%s%s",
176     //                        die.GetOffset(),
177     //                        depth * 2, depth * 2, "",
178     //                        DW_TAG_value_to_name (die.Tag()),
179     //                        die.HasChildren() ? " *" : "");
180 
181     const bool null_die = die.IsNULL();
182     if (depth == 0) {
183       if (initial_die_array_size == 0)
184         AddCompileUnitDIE(die);
185       uint64_t base_addr = die.GetAttributeValueAsAddress(
186           m_dwarf2Data, this, DW_AT_low_pc, LLDB_INVALID_ADDRESS);
187       if (base_addr == LLDB_INVALID_ADDRESS)
188         base_addr = die.GetAttributeValueAsAddress(m_dwarf2Data, this,
189                                                    DW_AT_entry_pc, 0);
190       SetBaseAddress(base_addr);
191       if (cu_die_only)
192         return 1;
193     } else {
194       if (null_die) {
195         if (prev_die_had_children) {
196           // This will only happen if a DIE says is has children
197           // but all it contains is a NULL tag. Since we are removing
198           // the NULL DIEs from the list (saves up to 25% in C++ code),
199           // we need a way to let the DIE know that it actually doesn't
200           // have children.
201           if (!m_die_array.empty())
202             m_die_array.back().SetEmptyChildren(true);
203         }
204       } else {
205         die.SetParentIndex(m_die_array.size() - die_index_stack[depth - 1]);
206 
207         if (die_index_stack.back())
208           m_die_array[die_index_stack.back()].SetSiblingIndex(
209               m_die_array.size() - die_index_stack.back());
210 
211         // Only push the DIE if it isn't a NULL DIE
212         m_die_array.push_back(die);
213       }
214     }
215 
216     if (null_die) {
217       // NULL DIE.
218       if (!die_index_stack.empty())
219         die_index_stack.pop_back();
220 
221       if (depth > 0)
222         --depth;
223       if (depth == 0)
224         break; // We are done with this compile unit!
225 
226       prev_die_had_children = false;
227     } else {
228       die_index_stack.back() = m_die_array.size() - 1;
229       // Normal DIE
230       const bool die_has_children = die.HasChildren();
231       if (die_has_children) {
232         die_index_stack.push_back(0);
233         ++depth;
234       }
235       prev_die_had_children = die_has_children;
236     }
237   }
238 
239   // Give a little bit of info if we encounter corrupt DWARF (our offset
240   // should always terminate at or before the start of the next compilation
241   // unit header).
242   if (offset > next_cu_offset) {
243     m_dwarf2Data->GetObjectFile()->GetModule()->ReportWarning(
244         "DWARF compile unit extends beyond its bounds cu 0x%8.8x at "
245         "0x%8.8" PRIx64 "\n",
246         GetOffset(), offset);
247   }
248 
249   // Since std::vector objects will double their size, we really need to
250   // make a new array with the perfect size so we don't end up wasting
251   // space. So here we copy and swap to make sure we don't have any extra
252   // memory taken up.
253 
254   if (m_die_array.size() < m_die_array.capacity()) {
255     DWARFDebugInfoEntry::collection exact_size_die_array(m_die_array.begin(),
256                                                          m_die_array.end());
257     exact_size_die_array.swap(m_die_array);
258   }
259   Log *verbose_log(
260       LogChannelDWARF::GetLogIfAll(DWARF_LOG_DEBUG_INFO | DWARF_LOG_VERBOSE));
261   if (verbose_log) {
262     StreamString strm;
263     Dump(&strm);
264     if (m_die_array.empty())
265       strm.Printf("error: no DIE for compile unit");
266     else
267       m_die_array[0].Dump(m_dwarf2Data, this, strm, UINT32_MAX);
268     verbose_log->PutCString(strm.GetString().c_str());
269   }
270 
271   if (!m_dwo_symbol_file)
272     return m_die_array.size();
273 
274   DWARFCompileUnit *dwo_cu = m_dwo_symbol_file->GetCompileUnit();
275   size_t dwo_die_count = dwo_cu->ExtractDIEsIfNeeded(cu_die_only);
276   return m_die_array.size() + dwo_die_count -
277          1; // We have 2 CU die, but we want to count it only as one
278 }
279 
280 void DWARFCompileUnit::AddCompileUnitDIE(DWARFDebugInfoEntry &die) {
281   assert(m_die_array.empty() && "Compile unit DIE already added");
282   AddDIE(die);
283 
284   const DWARFDebugInfoEntry &cu_die = m_die_array.front();
285   std::unique_ptr<SymbolFileDWARFDwo> dwo_symbol_file =
286       m_dwarf2Data->GetDwoSymbolFileForCompileUnit(*this, cu_die);
287   if (!dwo_symbol_file)
288     return;
289 
290   DWARFCompileUnit *dwo_cu = dwo_symbol_file->GetCompileUnit();
291   if (!dwo_cu)
292     return; // Can't fetch the compile unit from the dwo file.
293 
294   DWARFDIE dwo_cu_die = dwo_cu->GetCompileUnitDIEOnly();
295   if (!dwo_cu_die.IsValid())
296     return; // Can't fetch the compile unit DIE from the dwo file.
297 
298   uint64_t main_dwo_id = cu_die.GetAttributeValueAsUnsigned(
299       m_dwarf2Data, this, DW_AT_GNU_dwo_id, 0);
300   uint64_t sub_dwo_id =
301       dwo_cu_die.GetAttributeValueAsUnsigned(DW_AT_GNU_dwo_id, 0);
302   if (main_dwo_id != sub_dwo_id)
303     return; // The 2 dwo ID isn't match. Don't use the dwo file as it belongs to
304             // a differectn compilation.
305 
306   m_dwo_symbol_file = std::move(dwo_symbol_file);
307 
308   dw_addr_t addr_base = cu_die.GetAttributeValueAsUnsigned(
309       m_dwarf2Data, this, DW_AT_GNU_addr_base, 0);
310   dw_addr_t ranges_base = cu_die.GetAttributeValueAsUnsigned(
311       m_dwarf2Data, this, DW_AT_GNU_ranges_base, 0);
312   dwo_cu->SetAddrBase(addr_base, ranges_base, m_offset);
313 }
314 
315 dw_offset_t DWARFCompileUnit::GetAbbrevOffset() const {
316   return m_abbrevs ? m_abbrevs->GetOffset() : DW_INVALID_OFFSET;
317 }
318 
319 bool DWARFCompileUnit::Verify(Stream *s) const {
320   const DWARFDataExtractor &debug_info = m_dwarf2Data->get_debug_info_data();
321   bool valid_offset = debug_info.ValidOffset(m_offset);
322   bool length_OK = debug_info.ValidOffset(GetNextCompileUnitOffset() - 1);
323   bool version_OK = SymbolFileDWARF::SupportedVersion(m_version);
324   bool abbr_offset_OK =
325       m_dwarf2Data->get_debug_abbrev_data().ValidOffset(GetAbbrevOffset());
326   bool addr_size_OK = ((m_addr_size == 4) || (m_addr_size == 8));
327   bool verbose = s->GetVerbose();
328   if (valid_offset && length_OK && version_OK && addr_size_OK &&
329       abbr_offset_OK) {
330     if (verbose)
331       s->Printf("    0x%8.8x: OK\n", m_offset);
332     return true;
333   } else {
334     s->Printf("    0x%8.8x: ", m_offset);
335 
336     m_dwarf2Data->get_debug_info_data().Dump(s, m_offset, lldb::eFormatHex, 1,
337                                              Size(), 32, LLDB_INVALID_ADDRESS,
338                                              0, 0);
339     s->EOL();
340     if (valid_offset) {
341       if (!length_OK)
342         s->Printf("        The length (0x%8.8x) for this compile unit is too "
343                   "large for the .debug_info provided.\n",
344                   m_length);
345       if (!version_OK)
346         s->Printf("        The 16 bit compile unit header version is not "
347                   "supported.\n");
348       if (!abbr_offset_OK)
349         s->Printf("        The offset into the .debug_abbrev section (0x%8.8x) "
350                   "is not valid.\n",
351                   GetAbbrevOffset());
352       if (!addr_size_OK)
353         s->Printf("        The address size is unsupported: 0x%2.2x\n",
354                   m_addr_size);
355     } else
356       s->Printf("        The start offset of the compile unit header in the "
357                 ".debug_info is invalid.\n");
358   }
359   return false;
360 }
361 
362 void DWARFCompileUnit::Dump(Stream *s) const {
363   s->Printf("0x%8.8x: Compile Unit: length = 0x%8.8x, version = 0x%4.4x, "
364             "abbr_offset = 0x%8.8x, addr_size = 0x%2.2x (next CU at "
365             "{0x%8.8x})\n",
366             m_offset, m_length, m_version, GetAbbrevOffset(), m_addr_size,
367             GetNextCompileUnitOffset());
368 }
369 
370 static uint8_t g_default_addr_size = 4;
371 
372 uint8_t DWARFCompileUnit::GetAddressByteSize(const DWARFCompileUnit *cu) {
373   if (cu)
374     return cu->GetAddressByteSize();
375   return DWARFCompileUnit::GetDefaultAddressSize();
376 }
377 
378 bool DWARFCompileUnit::IsDWARF64(const DWARFCompileUnit *cu) {
379   if (cu)
380     return cu->IsDWARF64();
381   return false;
382 }
383 
384 uint8_t DWARFCompileUnit::GetDefaultAddressSize() {
385   return g_default_addr_size;
386 }
387 
388 void DWARFCompileUnit::SetDefaultAddressSize(uint8_t addr_size) {
389   g_default_addr_size = addr_size;
390 }
391 
392 lldb::user_id_t DWARFCompileUnit::GetID() const {
393   dw_offset_t local_id =
394       m_base_obj_offset != DW_INVALID_OFFSET ? m_base_obj_offset : m_offset;
395   if (m_dwarf2Data)
396     return DIERef(local_id, local_id).GetUID(m_dwarf2Data);
397   else
398     return local_id;
399 }
400 
401 void DWARFCompileUnit::BuildAddressRangeTable(
402     SymbolFileDWARF *dwarf2Data, DWARFDebugAranges *debug_aranges) {
403   // This function is usually called if there in no .debug_aranges section
404   // in order to produce a compile unit level set of address ranges that
405   // is accurate.
406 
407   size_t num_debug_aranges = debug_aranges->GetNumRanges();
408 
409   // First get the compile unit DIE only and check if it has a DW_AT_ranges
410   const DWARFDebugInfoEntry *die = GetCompileUnitDIEPtrOnly();
411 
412   const dw_offset_t cu_offset = GetOffset();
413   if (die) {
414     DWARFRangeList ranges;
415     const size_t num_ranges =
416         die->GetAttributeAddressRanges(dwarf2Data, this, ranges, false);
417     if (num_ranges > 0) {
418       // This compile unit has DW_AT_ranges, assume this is correct if it
419       // is present since clang no longer makes .debug_aranges by default
420       // and it emits DW_AT_ranges for DW_TAG_compile_units. GCC also does
421       // this with recent GCC builds.
422       for (size_t i = 0; i < num_ranges; ++i) {
423         const DWARFRangeList::Entry &range = ranges.GetEntryRef(i);
424         debug_aranges->AppendRange(cu_offset, range.GetRangeBase(),
425                                    range.GetRangeEnd());
426       }
427 
428       return; // We got all of our ranges from the DW_AT_ranges attribute
429     }
430   }
431   // We don't have a DW_AT_ranges attribute, so we need to parse the DWARF
432 
433   // If the DIEs weren't parsed, then we don't want all dies for all compile
434   // units
435   // to stay loaded when they weren't needed. So we can end up parsing the DWARF
436   // and then throwing them all away to keep memory usage down.
437   const bool clear_dies = ExtractDIEsIfNeeded(false) > 1;
438 
439   die = DIEPtr();
440   if (die)
441     die->BuildAddressRangeTable(dwarf2Data, this, debug_aranges);
442 
443   if (debug_aranges->GetNumRanges() == num_debug_aranges) {
444     // We got nothing from the functions, maybe we have a line tables only
445     // situation. Check the line tables and build the arange table from this.
446     SymbolContext sc;
447     sc.comp_unit = dwarf2Data->GetCompUnitForDWARFCompUnit(this);
448     if (sc.comp_unit) {
449       SymbolFileDWARFDebugMap *debug_map_sym_file =
450           m_dwarf2Data->GetDebugMapSymfile();
451       if (debug_map_sym_file == NULL) {
452         LineTable *line_table = sc.comp_unit->GetLineTable();
453 
454         if (line_table) {
455           LineTable::FileAddressRanges file_ranges;
456           const bool append = true;
457           const size_t num_ranges =
458               line_table->GetContiguousFileAddressRanges(file_ranges, append);
459           for (uint32_t idx = 0; idx < num_ranges; ++idx) {
460             const LineTable::FileAddressRanges::Entry &range =
461                 file_ranges.GetEntryRef(idx);
462             debug_aranges->AppendRange(cu_offset, range.GetRangeBase(),
463                                        range.GetRangeEnd());
464           }
465         }
466       } else
467         debug_map_sym_file->AddOSOARanges(dwarf2Data, debug_aranges);
468     }
469   }
470 
471   if (debug_aranges->GetNumRanges() == num_debug_aranges) {
472     // We got nothing from the functions, maybe we have a line tables only
473     // situation. Check the line tables and build the arange table from this.
474     SymbolContext sc;
475     sc.comp_unit = dwarf2Data->GetCompUnitForDWARFCompUnit(this);
476     if (sc.comp_unit) {
477       LineTable *line_table = sc.comp_unit->GetLineTable();
478 
479       if (line_table) {
480         LineTable::FileAddressRanges file_ranges;
481         const bool append = true;
482         const size_t num_ranges =
483             line_table->GetContiguousFileAddressRanges(file_ranges, append);
484         for (uint32_t idx = 0; idx < num_ranges; ++idx) {
485           const LineTable::FileAddressRanges::Entry &range =
486               file_ranges.GetEntryRef(idx);
487           debug_aranges->AppendRange(GetOffset(), range.GetRangeBase(),
488                                      range.GetRangeEnd());
489         }
490       }
491     }
492   }
493 
494   // Keep memory down by clearing DIEs if this generate function
495   // caused them to be parsed
496   if (clear_dies)
497     ClearDIEs(true);
498 }
499 
500 const DWARFDebugAranges &DWARFCompileUnit::GetFunctionAranges() {
501   if (m_func_aranges_ap.get() == NULL) {
502     m_func_aranges_ap.reset(new DWARFDebugAranges());
503     Log *log(LogChannelDWARF::GetLogIfAll(DWARF_LOG_DEBUG_ARANGES));
504 
505     if (log) {
506       m_dwarf2Data->GetObjectFile()->GetModule()->LogMessage(
507           log, "DWARFCompileUnit::GetFunctionAranges() for compile unit at "
508                ".debug_info[0x%8.8x]",
509           GetOffset());
510     }
511     const DWARFDebugInfoEntry *die = DIEPtr();
512     if (die)
513       die->BuildFunctionAddressRangeTable(m_dwarf2Data, this,
514                                           m_func_aranges_ap.get());
515 
516     if (m_dwo_symbol_file) {
517       DWARFCompileUnit *dwo_cu = m_dwo_symbol_file->GetCompileUnit();
518       const DWARFDebugInfoEntry *dwo_die = dwo_cu->DIEPtr();
519       if (dwo_die)
520         dwo_die->BuildFunctionAddressRangeTable(m_dwo_symbol_file.get(), dwo_cu,
521                                                 m_func_aranges_ap.get());
522     }
523 
524     const bool minimize = false;
525     m_func_aranges_ap->Sort(minimize);
526   }
527   return *m_func_aranges_ap.get();
528 }
529 
530 DWARFDIE
531 DWARFCompileUnit::LookupAddress(const dw_addr_t address) {
532   if (DIE()) {
533     const DWARFDebugAranges &func_aranges = GetFunctionAranges();
534 
535     // Re-check the aranges auto pointer contents in case it was created above
536     if (!func_aranges.IsEmpty())
537       return GetDIE(func_aranges.FindAddress(address));
538   }
539   return DWARFDIE();
540 }
541 
542 //----------------------------------------------------------------------
543 // Compare function DWARFDebugAranges::Range structures
544 //----------------------------------------------------------------------
545 static bool CompareDIEOffset(const DWARFDebugInfoEntry &die,
546                              const dw_offset_t die_offset) {
547   return die.GetOffset() < die_offset;
548 }
549 
550 //----------------------------------------------------------------------
551 // GetDIE()
552 //
553 // Get the DIE (Debug Information Entry) with the specified offset by
554 // first checking if the DIE is contained within this compile unit and
555 // grabbing the DIE from this compile unit. Otherwise we grab the DIE
556 // from the DWARF file.
557 //----------------------------------------------------------------------
558 DWARFDIE
559 DWARFCompileUnit::GetDIE(dw_offset_t die_offset) {
560   if (die_offset != DW_INVALID_OFFSET) {
561     if (m_dwo_symbol_file)
562       return m_dwo_symbol_file->GetCompileUnit()->GetDIE(die_offset);
563 
564     if (ContainsDIEOffset(die_offset)) {
565       ExtractDIEsIfNeeded(false);
566       DWARFDebugInfoEntry::iterator end = m_die_array.end();
567       DWARFDebugInfoEntry::iterator pos =
568           lower_bound(m_die_array.begin(), end, die_offset, CompareDIEOffset);
569       if (pos != end) {
570         if (die_offset == (*pos).GetOffset())
571           return DWARFDIE(this, &(*pos));
572       }
573     } else {
574       // Don't specify the compile unit offset as we don't know it because the
575       // DIE belongs to
576       // a different compile unit in the same symbol file.
577       return m_dwarf2Data->DebugInfo()->GetDIEForDIEOffset(die_offset);
578     }
579   }
580   return DWARFDIE(); // Not found
581 }
582 
583 size_t DWARFCompileUnit::AppendDIEsWithTag(const dw_tag_t tag,
584                                            DWARFDIECollection &dies,
585                                            uint32_t depth) const {
586   size_t old_size = dies.Size();
587   DWARFDebugInfoEntry::const_iterator pos;
588   DWARFDebugInfoEntry::const_iterator end = m_die_array.end();
589   for (pos = m_die_array.begin(); pos != end; ++pos) {
590     if (pos->Tag() == tag)
591       dies.Append(DWARFDIE(this, &(*pos)));
592   }
593 
594   // Return the number of DIEs added to the collection
595   return dies.Size() - old_size;
596 }
597 
598 // void
599 // DWARFCompileUnit::AddGlobalDIEByIndex (uint32_t die_idx)
600 //{
601 //    m_global_die_indexes.push_back (die_idx);
602 //}
603 //
604 //
605 // void
606 // DWARFCompileUnit::AddGlobal (const DWARFDebugInfoEntry* die)
607 //{
608 //    // Indexes to all file level global and static variables
609 //    m_global_die_indexes;
610 //
611 //    if (m_die_array.empty())
612 //        return;
613 //
614 //    const DWARFDebugInfoEntry* first_die = &m_die_array[0];
615 //    const DWARFDebugInfoEntry* end = first_die + m_die_array.size();
616 //    if (first_die <= die && die < end)
617 //        m_global_die_indexes.push_back (die - first_die);
618 //}
619 
620 void DWARFCompileUnit::Index(NameToDIE &func_basenames,
621                              NameToDIE &func_fullnames, NameToDIE &func_methods,
622                              NameToDIE &func_selectors,
623                              NameToDIE &objc_class_selectors,
624                              NameToDIE &globals, NameToDIE &types,
625                              NameToDIE &namespaces) {
626   Log *log(LogChannelDWARF::GetLogIfAll(DWARF_LOG_LOOKUPS));
627 
628   if (log) {
629     m_dwarf2Data->GetObjectFile()->GetModule()->LogMessage(
630         log,
631         "DWARFCompileUnit::Index() for compile unit at .debug_info[0x%8.8x]",
632         GetOffset());
633   }
634 
635   const LanguageType cu_language = GetLanguageType();
636   DWARFFormValue::FixedFormSizes fixed_form_sizes =
637       DWARFFormValue::GetFixedFormSizesForAddressSize(GetAddressByteSize(),
638                                                       m_is_dwarf64);
639 
640   IndexPrivate(this, cu_language, fixed_form_sizes, GetOffset(), func_basenames,
641                func_fullnames, func_methods, func_selectors,
642                objc_class_selectors, globals, types, namespaces);
643 
644   SymbolFileDWARFDwo *dwo_symbol_file = GetDwoSymbolFile();
645   if (dwo_symbol_file) {
646     IndexPrivate(dwo_symbol_file->GetCompileUnit(), cu_language,
647                  fixed_form_sizes, GetOffset(), func_basenames, func_fullnames,
648                  func_methods, func_selectors, objc_class_selectors, globals,
649                  types, namespaces);
650   }
651 }
652 
653 void DWARFCompileUnit::IndexPrivate(
654     DWARFCompileUnit *dwarf_cu, const LanguageType cu_language,
655     const DWARFFormValue::FixedFormSizes &fixed_form_sizes,
656     const dw_offset_t cu_offset, NameToDIE &func_basenames,
657     NameToDIE &func_fullnames, NameToDIE &func_methods,
658     NameToDIE &func_selectors, NameToDIE &objc_class_selectors,
659     NameToDIE &globals, NameToDIE &types, NameToDIE &namespaces) {
660   DWARFDebugInfoEntry::const_iterator pos;
661   DWARFDebugInfoEntry::const_iterator begin = dwarf_cu->m_die_array.begin();
662   DWARFDebugInfoEntry::const_iterator end = dwarf_cu->m_die_array.end();
663   for (pos = begin; pos != end; ++pos) {
664     const DWARFDebugInfoEntry &die = *pos;
665 
666     const dw_tag_t tag = die.Tag();
667 
668     switch (tag) {
669     case DW_TAG_array_type:
670     case DW_TAG_base_type:
671     case DW_TAG_class_type:
672     case DW_TAG_constant:
673     case DW_TAG_enumeration_type:
674     case DW_TAG_inlined_subroutine:
675     case DW_TAG_namespace:
676     case DW_TAG_string_type:
677     case DW_TAG_structure_type:
678     case DW_TAG_subprogram:
679     case DW_TAG_subroutine_type:
680     case DW_TAG_typedef:
681     case DW_TAG_union_type:
682     case DW_TAG_unspecified_type:
683     case DW_TAG_variable:
684       break;
685 
686     default:
687       continue;
688     }
689 
690     DWARFAttributes attributes;
691     const char *name = NULL;
692     const char *mangled_cstr = NULL;
693     bool is_declaration = false;
694     // bool is_artificial = false;
695     bool has_address = false;
696     bool has_location_or_const_value = false;
697     bool is_global_or_static_variable = false;
698 
699     DWARFFormValue specification_die_form;
700     const size_t num_attributes =
701         die.GetAttributes(dwarf_cu, fixed_form_sizes, attributes);
702     if (num_attributes > 0) {
703       for (uint32_t i = 0; i < num_attributes; ++i) {
704         dw_attr_t attr = attributes.AttributeAtIndex(i);
705         DWARFFormValue form_value;
706         switch (attr) {
707         case DW_AT_name:
708           if (attributes.ExtractFormValueAtIndex(i, form_value))
709             name = form_value.AsCString();
710           break;
711 
712         case DW_AT_declaration:
713           if (attributes.ExtractFormValueAtIndex(i, form_value))
714             is_declaration = form_value.Unsigned() != 0;
715           break;
716 
717         //                case DW_AT_artificial:
718         //                    if (attributes.ExtractFormValueAtIndex(i,
719         //                    form_value))
720         //                        is_artificial = form_value.Unsigned() != 0;
721         //                    break;
722 
723         case DW_AT_MIPS_linkage_name:
724         case DW_AT_linkage_name:
725           if (attributes.ExtractFormValueAtIndex(i, form_value))
726             mangled_cstr = form_value.AsCString();
727           break;
728 
729         case DW_AT_low_pc:
730         case DW_AT_high_pc:
731         case DW_AT_ranges:
732           has_address = true;
733           break;
734 
735         case DW_AT_entry_pc:
736           has_address = true;
737           break;
738 
739         case DW_AT_location:
740         case DW_AT_const_value:
741           has_location_or_const_value = true;
742           if (tag == DW_TAG_variable) {
743             const DWARFDebugInfoEntry *parent_die = die.GetParent();
744             while (parent_die != NULL) {
745               switch (parent_die->Tag()) {
746               case DW_TAG_subprogram:
747               case DW_TAG_lexical_block:
748               case DW_TAG_inlined_subroutine:
749                 // Even if this is a function level static, we don't add it. We
750                 // could theoretically
751                 // add these if we wanted to by introspecting into the
752                 // DW_AT_location and seeing
753                 // if the location describes a hard coded address, but we dont
754                 // want the performance
755                 // penalty of that right now.
756                 is_global_or_static_variable = false;
757                 //                              if
758                 //                              (attributes.ExtractFormValueAtIndex(dwarf2Data,
759                 //                              i, form_value))
760                 //                              {
761                 //                                  // If we have valid block
762                 //                                  data, then we have location
763                 //                                  expression bytes
764                 //                                  // that are fixed (not a
765                 //                                  location list).
766                 //                                  const uint8_t *block_data =
767                 //                                  form_value.BlockData();
768                 //                                  if (block_data)
769                 //                                  {
770                 //                                      uint32_t block_length =
771                 //                                      form_value.Unsigned();
772                 //                                      if (block_length == 1 +
773                 //                                      attributes.CompileUnitAtIndex(i)->GetAddressByteSize())
774                 //                                      {
775                 //                                          if (block_data[0] ==
776                 //                                          DW_OP_addr)
777                 //                                              add_die = true;
778                 //                                      }
779                 //                                  }
780                 //                              }
781                 parent_die = NULL; // Terminate the while loop.
782                 break;
783 
784               case DW_TAG_compile_unit:
785                 is_global_or_static_variable = true;
786                 parent_die = NULL; // Terminate the while loop.
787                 break;
788 
789               default:
790                 parent_die =
791                     parent_die->GetParent(); // Keep going in the while loop.
792                 break;
793               }
794             }
795           }
796           break;
797 
798         case DW_AT_specification:
799           if (attributes.ExtractFormValueAtIndex(i, form_value))
800             specification_die_form = form_value;
801           break;
802         }
803       }
804     }
805 
806     switch (tag) {
807     case DW_TAG_subprogram:
808       if (has_address) {
809         if (name) {
810           ObjCLanguage::MethodName objc_method(name, true);
811           if (objc_method.IsValid(true)) {
812             ConstString objc_class_name_with_category(
813                 objc_method.GetClassNameWithCategory());
814             ConstString objc_selector_name(objc_method.GetSelector());
815             ConstString objc_fullname_no_category_name(
816                 objc_method.GetFullNameWithoutCategory(true));
817             ConstString objc_class_name_no_category(objc_method.GetClassName());
818             func_fullnames.Insert(ConstString(name),
819                                   DIERef(cu_offset, die.GetOffset()));
820             if (objc_class_name_with_category)
821               objc_class_selectors.Insert(objc_class_name_with_category,
822                                           DIERef(cu_offset, die.GetOffset()));
823             if (objc_class_name_no_category &&
824                 objc_class_name_no_category != objc_class_name_with_category)
825               objc_class_selectors.Insert(objc_class_name_no_category,
826                                           DIERef(cu_offset, die.GetOffset()));
827             if (objc_selector_name)
828               func_selectors.Insert(objc_selector_name,
829                                     DIERef(cu_offset, die.GetOffset()));
830             if (objc_fullname_no_category_name)
831               func_fullnames.Insert(objc_fullname_no_category_name,
832                                     DIERef(cu_offset, die.GetOffset()));
833           }
834           // If we have a mangled name, then the DW_AT_name attribute
835           // is usually the method name without the class or any parameters
836           const DWARFDebugInfoEntry *parent = die.GetParent();
837           bool is_method = false;
838           if (parent) {
839             dw_tag_t parent_tag = parent->Tag();
840             if (parent_tag == DW_TAG_class_type ||
841                 parent_tag == DW_TAG_structure_type) {
842               is_method = true;
843             } else {
844               if (specification_die_form.IsValid()) {
845                 DWARFDIE specification_die =
846                     dwarf_cu->GetSymbolFileDWARF()->DebugInfo()->GetDIE(
847                         DIERef(specification_die_form));
848                 if (specification_die.GetParent().IsStructOrClass())
849                   is_method = true;
850               }
851             }
852           }
853 
854           if (is_method)
855             func_methods.Insert(ConstString(name),
856                                 DIERef(cu_offset, die.GetOffset()));
857           else
858             func_basenames.Insert(ConstString(name),
859                                   DIERef(cu_offset, die.GetOffset()));
860 
861           if (!is_method && !mangled_cstr && !objc_method.IsValid(true))
862             func_fullnames.Insert(ConstString(name),
863                                   DIERef(cu_offset, die.GetOffset()));
864         }
865         if (mangled_cstr) {
866           // Make sure our mangled name isn't the same string table entry
867           // as our name. If it starts with '_', then it is ok, else compare
868           // the string to make sure it isn't the same and we don't end up
869           // with duplicate entries
870           if (name && name != mangled_cstr &&
871               ((mangled_cstr[0] == '_') ||
872                (::strcmp(name, mangled_cstr) != 0))) {
873             Mangled mangled(ConstString(mangled_cstr), true);
874             func_fullnames.Insert(mangled.GetMangledName(),
875                                   DIERef(cu_offset, die.GetOffset()));
876             ConstString demangled = mangled.GetDemangledName(cu_language);
877             if (demangled)
878               func_fullnames.Insert(demangled,
879                                     DIERef(cu_offset, die.GetOffset()));
880           }
881         }
882       }
883       break;
884 
885     case DW_TAG_inlined_subroutine:
886       if (has_address) {
887         if (name)
888           func_basenames.Insert(ConstString(name),
889                                 DIERef(cu_offset, die.GetOffset()));
890         if (mangled_cstr) {
891           // Make sure our mangled name isn't the same string table entry
892           // as our name. If it starts with '_', then it is ok, else compare
893           // the string to make sure it isn't the same and we don't end up
894           // with duplicate entries
895           if (name && name != mangled_cstr &&
896               ((mangled_cstr[0] == '_') ||
897                (::strcmp(name, mangled_cstr) != 0))) {
898             Mangled mangled(ConstString(mangled_cstr), true);
899             func_fullnames.Insert(mangled.GetMangledName(),
900                                   DIERef(cu_offset, die.GetOffset()));
901             ConstString demangled = mangled.GetDemangledName(cu_language);
902             if (demangled)
903               func_fullnames.Insert(demangled,
904                                     DIERef(cu_offset, die.GetOffset()));
905           }
906         } else
907           func_fullnames.Insert(ConstString(name),
908                                 DIERef(cu_offset, die.GetOffset()));
909       }
910       break;
911 
912     case DW_TAG_array_type:
913     case DW_TAG_base_type:
914     case DW_TAG_class_type:
915     case DW_TAG_constant:
916     case DW_TAG_enumeration_type:
917     case DW_TAG_string_type:
918     case DW_TAG_structure_type:
919     case DW_TAG_subroutine_type:
920     case DW_TAG_typedef:
921     case DW_TAG_union_type:
922     case DW_TAG_unspecified_type:
923       if (name && !is_declaration)
924         types.Insert(ConstString(name), DIERef(cu_offset, die.GetOffset()));
925       if (mangled_cstr && !is_declaration)
926         types.Insert(ConstString(mangled_cstr),
927                      DIERef(cu_offset, die.GetOffset()));
928       break;
929 
930     case DW_TAG_namespace:
931       if (name)
932         namespaces.Insert(ConstString(name),
933                           DIERef(cu_offset, die.GetOffset()));
934       break;
935 
936     case DW_TAG_variable:
937       if (name && has_location_or_const_value && is_global_or_static_variable) {
938         globals.Insert(ConstString(name), DIERef(cu_offset, die.GetOffset()));
939         // Be sure to include variables by their mangled and demangled
940         // names if they have any since a variable can have a basename
941         // "i", a mangled named "_ZN12_GLOBAL__N_11iE" and a demangled
942         // mangled name "(anonymous namespace)::i"...
943 
944         // Make sure our mangled name isn't the same string table entry
945         // as our name. If it starts with '_', then it is ok, else compare
946         // the string to make sure it isn't the same and we don't end up
947         // with duplicate entries
948         if (mangled_cstr && name != mangled_cstr &&
949             ((mangled_cstr[0] == '_') || (::strcmp(name, mangled_cstr) != 0))) {
950           Mangled mangled(ConstString(mangled_cstr), true);
951           globals.Insert(mangled.GetMangledName(),
952                          DIERef(cu_offset, die.GetOffset()));
953           ConstString demangled = mangled.GetDemangledName(cu_language);
954           if (demangled)
955             globals.Insert(demangled, DIERef(cu_offset, die.GetOffset()));
956         }
957       }
958       break;
959 
960     default:
961       continue;
962     }
963   }
964 }
965 
966 bool DWARFCompileUnit::Supports_unnamed_objc_bitfields() {
967   if (GetProducer() == eProducerClang) {
968     const uint32_t major_version = GetProducerVersionMajor();
969     if (major_version > 425 ||
970         (major_version == 425 && GetProducerVersionUpdate() >= 13))
971       return true;
972     else
973       return false;
974   }
975   return true; // Assume all other compilers didn't have incorrect ObjC bitfield
976                // info
977 }
978 
979 bool DWARFCompileUnit::Supports_DW_AT_APPLE_objc_complete_type() {
980   if (GetProducer() == eProducerLLVMGCC)
981     return false;
982   return true;
983 }
984 
985 bool DWARFCompileUnit::DW_AT_decl_file_attributes_are_invalid() {
986   // llvm-gcc makes completely invalid decl file attributes and won't ever
987   // be fixed, so we need to know to ignore these.
988   return GetProducer() == eProducerLLVMGCC;
989 }
990 
991 void DWARFCompileUnit::ParseProducerInfo() {
992   m_producer_version_major = UINT32_MAX;
993   m_producer_version_minor = UINT32_MAX;
994   m_producer_version_update = UINT32_MAX;
995 
996   const DWARFDebugInfoEntry *die = GetCompileUnitDIEPtrOnly();
997   if (die) {
998 
999     const char *producer_cstr = die->GetAttributeValueAsString(
1000         m_dwarf2Data, this, DW_AT_producer, NULL);
1001     if (producer_cstr) {
1002       RegularExpression llvm_gcc_regex(
1003           llvm::StringRef("^4\\.[012]\\.[01] \\(Based on Apple "
1004                           "Inc\\. build [0-9]+\\) \\(LLVM build "
1005                           "[\\.0-9]+\\)$"));
1006       if (llvm_gcc_regex.Execute(llvm::StringRef(producer_cstr))) {
1007         m_producer = eProducerLLVMGCC;
1008       } else if (strstr(producer_cstr, "clang")) {
1009         static RegularExpression g_clang_version_regex(
1010             llvm::StringRef("clang-([0-9]+)\\.([0-9]+)\\.([0-9]+)"));
1011         RegularExpression::Match regex_match(3);
1012         if (g_clang_version_regex.Execute(llvm::StringRef(producer_cstr),
1013                                           &regex_match)) {
1014           std::string str;
1015           if (regex_match.GetMatchAtIndex(producer_cstr, 1, str))
1016             m_producer_version_major =
1017                 StringConvert::ToUInt32(str.c_str(), UINT32_MAX, 10);
1018           if (regex_match.GetMatchAtIndex(producer_cstr, 2, str))
1019             m_producer_version_minor =
1020                 StringConvert::ToUInt32(str.c_str(), UINT32_MAX, 10);
1021           if (regex_match.GetMatchAtIndex(producer_cstr, 3, str))
1022             m_producer_version_update =
1023                 StringConvert::ToUInt32(str.c_str(), UINT32_MAX, 10);
1024         }
1025         m_producer = eProducerClang;
1026       } else if (strstr(producer_cstr, "GNU"))
1027         m_producer = eProducerGCC;
1028     }
1029   }
1030   if (m_producer == eProducerInvalid)
1031     m_producer = eProcucerOther;
1032 }
1033 
1034 DWARFCompileUnit::Producer DWARFCompileUnit::GetProducer() {
1035   if (m_producer == eProducerInvalid)
1036     ParseProducerInfo();
1037   return m_producer;
1038 }
1039 
1040 uint32_t DWARFCompileUnit::GetProducerVersionMajor() {
1041   if (m_producer_version_major == 0)
1042     ParseProducerInfo();
1043   return m_producer_version_major;
1044 }
1045 
1046 uint32_t DWARFCompileUnit::GetProducerVersionMinor() {
1047   if (m_producer_version_minor == 0)
1048     ParseProducerInfo();
1049   return m_producer_version_minor;
1050 }
1051 
1052 uint32_t DWARFCompileUnit::GetProducerVersionUpdate() {
1053   if (m_producer_version_update == 0)
1054     ParseProducerInfo();
1055   return m_producer_version_update;
1056 }
1057 
1058 LanguageType DWARFCompileUnit::LanguageTypeFromDWARF(uint64_t val) {
1059   // Note: user languages between lo_user and hi_user
1060   // must be handled explicitly here.
1061   switch (val) {
1062   case DW_LANG_Mips_Assembler:
1063     return eLanguageTypeMipsAssembler;
1064   case DW_LANG_GOOGLE_RenderScript:
1065     return eLanguageTypeExtRenderScript;
1066   default:
1067     return static_cast<LanguageType>(val);
1068   }
1069 }
1070 
1071 LanguageType DWARFCompileUnit::GetLanguageType() {
1072   if (m_language_type != eLanguageTypeUnknown)
1073     return m_language_type;
1074 
1075   const DWARFDebugInfoEntry *die = GetCompileUnitDIEPtrOnly();
1076   if (die)
1077     m_language_type = LanguageTypeFromDWARF(die->GetAttributeValueAsUnsigned(
1078         m_dwarf2Data, this, DW_AT_language, 0));
1079   return m_language_type;
1080 }
1081 
1082 bool DWARFCompileUnit::IsDWARF64() const { return m_is_dwarf64; }
1083 
1084 bool DWARFCompileUnit::GetIsOptimized() {
1085   if (m_is_optimized == eLazyBoolCalculate) {
1086     const DWARFDebugInfoEntry *die = GetCompileUnitDIEPtrOnly();
1087     if (die) {
1088       m_is_optimized = eLazyBoolNo;
1089       if (die->GetAttributeValueAsUnsigned(m_dwarf2Data, this,
1090                                            DW_AT_APPLE_optimized, 0) == 1) {
1091         m_is_optimized = eLazyBoolYes;
1092       }
1093     }
1094   }
1095   if (m_is_optimized == eLazyBoolYes) {
1096     return true;
1097   } else {
1098     return false;
1099   }
1100 }
1101 
1102 DWARFFormValue::FixedFormSizes DWARFCompileUnit::GetFixedFormSizes() {
1103   return DWARFFormValue::GetFixedFormSizesForAddressSize(GetAddressByteSize(),
1104                                                          IsDWARF64());
1105 }
1106 
1107 TypeSystem *DWARFCompileUnit::GetTypeSystem() {
1108   if (m_dwarf2Data)
1109     return m_dwarf2Data->GetTypeSystemForLanguage(GetLanguageType());
1110   else
1111     return nullptr;
1112 }
1113 
1114 void DWARFCompileUnit::SetUserData(void *d) {
1115   m_user_data = d;
1116   if (m_dwo_symbol_file)
1117     m_dwo_symbol_file->GetCompileUnit()->SetUserData(d);
1118 }
1119 
1120 void DWARFCompileUnit::SetAddrBase(dw_addr_t addr_base,
1121                                    dw_addr_t ranges_base,
1122                                    dw_offset_t base_obj_offset) {
1123   m_addr_base = addr_base;
1124   m_ranges_base = ranges_base;
1125   m_base_obj_offset = base_obj_offset;
1126 }
1127 
1128 lldb::ByteOrder DWARFCompileUnit::GetByteOrder() const {
1129   return m_dwarf2Data->GetObjectFile()->GetByteOrder();
1130 }
1131