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 "lldb/Core/Mangled.h"
13 #include "lldb/Core/Module.h"
14 #include "lldb/Core/Stream.h"
15 #include "lldb/Core/StreamString.h"
16 #include "lldb/Core/Timer.h"
17 #include "lldb/Host/StringConvert.h"
18 #include "lldb/Symbol/CompileUnit.h"
19 #include "lldb/Symbol/LineTable.h"
20 #include "lldb/Symbol/ObjectFile.h"
21 #include "Plugins/Language/ObjC/ObjCLanguage.h"
22 
23 #include "DWARFDebugAbbrev.h"
24 #include "DWARFDebugAranges.h"
25 #include "DWARFDebugInfo.h"
26 #include "DWARFDIECollection.h"
27 #include "DWARFFormValue.h"
28 #include "LogChannelDWARF.h"
29 #include "NameToDIE.h"
30 #include "SymbolFileDWARF.h"
31 #include "SymbolFileDWARFDwo.h"
32 #include "SymbolFileDWARFDebugMap.h"
33 
34 using namespace lldb;
35 using namespace lldb_private;
36 using namespace std;
37 
38 
39 extern int g_verbose;
40 
41 DWARFCompileUnit::DWARFCompileUnit(SymbolFileDWARF* dwarf2Data) :
42     m_dwarf2Data    (dwarf2Data),
43     m_abbrevs       (NULL),
44     m_user_data     (NULL),
45     m_die_array     (),
46     m_func_aranges_ap (),
47     m_base_addr     (0),
48     m_offset        (DW_INVALID_OFFSET),
49     m_length        (0),
50     m_version       (0),
51     m_addr_size     (DWARFCompileUnit::GetDefaultAddressSize()),
52     m_producer      (eProducerInvalid),
53     m_producer_version_major (0),
54     m_producer_version_minor (0),
55     m_producer_version_update (0),
56     m_language_type (eLanguageTypeUnknown),
57     m_is_dwarf64    (false),
58     m_is_optimized  (eLazyBoolCalculate),
59     m_addr_base (0),
60     m_base_obj_offset (DW_INVALID_OFFSET)
61 {
62 }
63 
64 DWARFCompileUnit::~DWARFCompileUnit()
65 {}
66 
67 void
68 DWARFCompileUnit::Clear()
69 {
70     m_offset        = DW_INVALID_OFFSET;
71     m_length        = 0;
72     m_version       = 0;
73     m_abbrevs       = NULL;
74     m_addr_size     = DWARFCompileUnit::GetDefaultAddressSize();
75     m_base_addr     = 0;
76     m_die_array.clear();
77     m_func_aranges_ap.reset();
78     m_user_data     = NULL;
79     m_producer      = eProducerInvalid;
80     m_language_type = eLanguageTypeUnknown;
81     m_is_dwarf64    = false;
82     m_is_optimized  = eLazyBoolCalculate;
83     m_addr_base     = 0;
84     m_base_obj_offset = DW_INVALID_OFFSET;
85 }
86 
87 bool
88 DWARFCompileUnit::Extract(const DWARFDataExtractor &debug_info, lldb::offset_t *offset_ptr)
89 {
90     Clear();
91 
92     m_offset = *offset_ptr;
93 
94     if (debug_info.ValidOffset(*offset_ptr))
95     {
96         dw_offset_t abbr_offset;
97         const DWARFDebugAbbrev *abbr = m_dwarf2Data->DebugAbbrev();
98         m_length        = debug_info.GetDWARFInitialLength(offset_ptr);
99         m_is_dwarf64    = debug_info.IsDWARF64();
100         m_version       = debug_info.GetU16(offset_ptr);
101         abbr_offset     = debug_info.GetDWARFOffset(offset_ptr);
102         m_addr_size     = debug_info.GetU8 (offset_ptr);
103 
104         bool length_OK = debug_info.ValidOffset(GetNextCompileUnitOffset()-1);
105         bool version_OK = SymbolFileDWARF::SupportedVersion(m_version);
106         bool abbr_offset_OK = m_dwarf2Data->get_debug_abbrev_data().ValidOffset(abbr_offset);
107         bool addr_size_OK = ((m_addr_size == 4) || (m_addr_size == 8));
108 
109         if (length_OK && version_OK && addr_size_OK && abbr_offset_OK && abbr != NULL)
110         {
111             m_abbrevs = abbr->GetAbbreviationDeclarationSet(abbr_offset);
112             return true;
113         }
114 
115         // reset the offset to where we tried to parse from if anything went wrong
116         *offset_ptr = m_offset;
117     }
118 
119     return false;
120 }
121 
122 
123 void
124 DWARFCompileUnit::ClearDIEs(bool keep_compile_unit_die)
125 {
126     if (m_die_array.size() > 1)
127     {
128         // std::vectors never get any smaller when resized to a smaller size,
129         // or when clear() or erase() are called, the size will report that it
130         // is smaller, but the memory allocated remains intact (call capacity()
131         // to see this). So we need to create a temporary vector and swap the
132         // contents which will cause just the internal pointers to be swapped
133         // so that when "tmp_array" goes out of scope, it will destroy the
134         // contents.
135 
136         // Save at least the compile unit DIE
137         DWARFDebugInfoEntry::collection tmp_array;
138         m_die_array.swap(tmp_array);
139         if (keep_compile_unit_die)
140             m_die_array.push_back(tmp_array.front());
141     }
142 
143     if (m_dwo_symbol_file)
144         m_dwo_symbol_file->GetCompileUnit()->ClearDIEs(keep_compile_unit_die);
145 }
146 
147 //----------------------------------------------------------------------
148 // ParseCompileUnitDIEsIfNeeded
149 //
150 // Parses a compile unit and indexes its DIEs if it hasn't already been
151 // done.
152 //----------------------------------------------------------------------
153 size_t
154 DWARFCompileUnit::ExtractDIEsIfNeeded (bool cu_die_only)
155 {
156     const size_t initial_die_array_size = m_die_array.size();
157     if ((cu_die_only && initial_die_array_size > 0) || initial_die_array_size > 1)
158         return 0; // Already parsed
159 
160     Timer scoped_timer (__PRETTY_FUNCTION__,
161                         "%8.8x: DWARFCompileUnit::ExtractDIEsIfNeeded( cu_die_only = %i )",
162                         m_offset,
163                         cu_die_only);
164 
165     // Set the offset to that of the first DIE and calculate the start of the
166     // next compilation unit header.
167     lldb::offset_t offset = GetFirstDIEOffset();
168     lldb::offset_t next_cu_offset = GetNextCompileUnitOffset();
169 
170     DWARFDebugInfoEntry die;
171         // Keep a flat array of the DIE for binary lookup by DIE offset
172     if (!cu_die_only)
173     {
174         Log *log (LogChannelDWARF::GetLogIfAny(DWARF_LOG_DEBUG_INFO | DWARF_LOG_LOOKUPS));
175         if (log)
176         {
177             m_dwarf2Data->GetObjectFile()->GetModule()->LogMessageVerboseBacktrace (log,
178                                                                                     "DWARFCompileUnit::ExtractDIEsIfNeeded () for compile unit at .debug_info[0x%8.8x]",
179                                                                                     GetOffset());
180         }
181     }
182 
183     uint32_t depth = 0;
184     // We are in our compile unit, parse starting at the offset
185     // we were told to parse
186     const DWARFDataExtractor& debug_info_data = m_dwarf2Data->get_debug_info_data();
187     std::vector<uint32_t> die_index_stack;
188     die_index_stack.reserve(32);
189     die_index_stack.push_back(0);
190     bool prev_die_had_children = false;
191     DWARFFormValue::FixedFormSizes fixed_form_sizes =
192         DWARFFormValue::GetFixedFormSizesForAddressSize (GetAddressByteSize(), m_is_dwarf64);
193     while (offset < next_cu_offset &&
194            die.FastExtract (debug_info_data, this, fixed_form_sizes, &offset))
195     {
196 //        if (log)
197 //            log->Printf("0x%8.8x: %*.*s%s%s",
198 //                        die.GetOffset(),
199 //                        depth * 2, depth * 2, "",
200 //                        DW_TAG_value_to_name (die.Tag()),
201 //                        die.HasChildren() ? " *" : "");
202 
203         const bool null_die = die.IsNULL();
204         if (depth == 0)
205         {
206             if (initial_die_array_size == 0)
207                 AddCompileUnitDIE(die);
208             uint64_t base_addr = die.GetAttributeValueAsAddress(m_dwarf2Data, this, DW_AT_low_pc, LLDB_INVALID_ADDRESS);
209             if (base_addr == LLDB_INVALID_ADDRESS)
210                 base_addr = die.GetAttributeValueAsAddress(m_dwarf2Data, this, DW_AT_entry_pc, 0);
211             SetBaseAddress (base_addr);
212             if (cu_die_only)
213                 return 1;
214         }
215         else
216         {
217             if (null_die)
218             {
219                 if (prev_die_had_children)
220                 {
221                     // This will only happen if a DIE says is has children
222                     // but all it contains is a NULL tag. Since we are removing
223                     // the NULL DIEs from the list (saves up to 25% in C++ code),
224                     // we need a way to let the DIE know that it actually doesn't
225                     // have children.
226                     if (!m_die_array.empty())
227                         m_die_array.back().SetEmptyChildren(true);
228                 }
229             }
230             else
231             {
232                 die.SetParentIndex(m_die_array.size() - die_index_stack[depth-1]);
233 
234                 if (die_index_stack.back())
235                     m_die_array[die_index_stack.back()].SetSiblingIndex(m_die_array.size()-die_index_stack.back());
236 
237                 // Only push the DIE if it isn't a NULL DIE
238                     m_die_array.push_back(die);
239             }
240         }
241 
242         if (null_die)
243         {
244             // NULL DIE.
245             if (!die_index_stack.empty())
246                 die_index_stack.pop_back();
247 
248             if (depth > 0)
249                 --depth;
250             if (depth == 0)
251                 break;  // We are done with this compile unit!
252 
253             prev_die_had_children = false;
254         }
255         else
256         {
257             die_index_stack.back() = m_die_array.size() - 1;
258             // Normal DIE
259             const bool die_has_children = die.HasChildren();
260             if (die_has_children)
261             {
262                 die_index_stack.push_back(0);
263                 ++depth;
264             }
265             prev_die_had_children = die_has_children;
266         }
267     }
268 
269     // Give a little bit of info if we encounter corrupt DWARF (our offset
270     // should always terminate at or before the start of the next compilation
271     // unit header).
272     if (offset > next_cu_offset)
273     {
274         m_dwarf2Data->GetObjectFile()->GetModule()->ReportWarning ("DWARF compile unit extends beyond its bounds cu 0x%8.8x at 0x%8.8" PRIx64 "\n",
275                                                                    GetOffset(),
276                                                                    offset);
277     }
278 
279     // Since std::vector objects will double their size, we really need to
280     // make a new array with the perfect size so we don't end up wasting
281     // space. So here we copy and swap to make sure we don't have any extra
282     // memory taken up.
283 
284     if (m_die_array.size () < m_die_array.capacity())
285     {
286         DWARFDebugInfoEntry::collection exact_size_die_array (m_die_array.begin(), m_die_array.end());
287         exact_size_die_array.swap (m_die_array);
288     }
289     Log *verbose_log (LogChannelDWARF::GetLogIfAll (DWARF_LOG_DEBUG_INFO | DWARF_LOG_VERBOSE));
290     if (verbose_log)
291     {
292         StreamString strm;
293         Dump(&strm);
294         if (m_die_array.empty())
295             strm.Printf("error: no DIE for compile unit");
296         else
297             m_die_array[0].Dump(m_dwarf2Data, this, strm, UINT32_MAX);
298         verbose_log->PutCString (strm.GetString().c_str());
299     }
300 
301     if (!m_dwo_symbol_file)
302         return m_die_array.size();
303 
304     DWARFCompileUnit* dwo_cu = m_dwo_symbol_file->GetCompileUnit();
305     size_t dwo_die_count = dwo_cu->ExtractDIEsIfNeeded(cu_die_only);
306     return m_die_array.size() + dwo_die_count - 1; // We have 2 CU die, but we waht to count it only as one
307 }
308 
309 void
310 DWARFCompileUnit::AddCompileUnitDIE(DWARFDebugInfoEntry& die)
311 {
312     assert (m_die_array.empty() && "Compile unit DIE already added");
313     AddDIE(die);
314 
315     const DWARFDebugInfoEntry &cu_die = m_die_array.front();
316     std::unique_ptr<SymbolFileDWARFDwo> dwo_symbol_file = m_dwarf2Data->GetDwoSymbolFileForCompileUnit(*this, cu_die);
317     if (!dwo_symbol_file)
318         return;
319 
320     DWARFCompileUnit* dwo_cu = dwo_symbol_file->GetCompileUnit();
321     if (!dwo_cu)
322         return; // Can't fetch the compile unit from the dwo file.
323 
324     DWARFDIE dwo_cu_die = dwo_cu->GetCompileUnitDIEOnly();
325     if (!dwo_cu_die.IsValid())
326         return; // Can't fetch the compile unit DIE from the dwo file.
327 
328     uint64_t main_dwo_id = cu_die.GetAttributeValueAsUnsigned(m_dwarf2Data,
329                                                               this,
330                                                               DW_AT_GNU_dwo_id,
331                                                               0);
332     uint64_t sub_dwo_id = dwo_cu_die.GetAttributeValueAsUnsigned(DW_AT_GNU_dwo_id, 0);
333     if (main_dwo_id != sub_dwo_id)
334         return; // The 2 dwo ID isn't match. Don't use the dwo file as it belongs to a differectn compilation.
335 
336     m_dwo_symbol_file = std::move(dwo_symbol_file);
337 
338     dw_addr_t addr_base = cu_die.GetAttributeValueAsUnsigned(m_dwarf2Data,
339                                                              this,
340                                                              DW_AT_GNU_addr_base,
341                                                              0);
342     dwo_cu->SetAddrBase(addr_base, m_offset);
343 }
344 
345 dw_offset_t
346 DWARFCompileUnit::GetAbbrevOffset() const
347 {
348     return m_abbrevs ? m_abbrevs->GetOffset() : DW_INVALID_OFFSET;
349 }
350 
351 
352 
353 bool
354 DWARFCompileUnit::Verify(Stream *s) const
355 {
356     const DWARFDataExtractor& debug_info = m_dwarf2Data->get_debug_info_data();
357     bool valid_offset = debug_info.ValidOffset(m_offset);
358     bool length_OK = debug_info.ValidOffset(GetNextCompileUnitOffset()-1);
359     bool version_OK = SymbolFileDWARF::SupportedVersion(m_version);
360     bool abbr_offset_OK = m_dwarf2Data->get_debug_abbrev_data().ValidOffset(GetAbbrevOffset());
361     bool addr_size_OK = ((m_addr_size == 4) || (m_addr_size == 8));
362     bool verbose = s->GetVerbose();
363     if (valid_offset && length_OK && version_OK && addr_size_OK && abbr_offset_OK)
364     {
365         if (verbose)
366             s->Printf("    0x%8.8x: OK\n", m_offset);
367         return true;
368     }
369     else
370     {
371         s->Printf("    0x%8.8x: ", m_offset);
372 
373         m_dwarf2Data->get_debug_info_data().Dump (s, m_offset, lldb::eFormatHex, 1, Size(), 32, LLDB_INVALID_ADDRESS, 0, 0);
374         s->EOL();
375         if (valid_offset)
376         {
377             if (!length_OK)
378                 s->Printf("        The length (0x%8.8x) for this compile unit is too large for the .debug_info provided.\n", m_length);
379             if (!version_OK)
380                 s->Printf("        The 16 bit compile unit header version is not supported.\n");
381             if (!abbr_offset_OK)
382                 s->Printf("        The offset into the .debug_abbrev section (0x%8.8x) is not valid.\n", GetAbbrevOffset());
383             if (!addr_size_OK)
384                 s->Printf("        The address size is unsupported: 0x%2.2x\n", m_addr_size);
385         }
386         else
387             s->Printf("        The start offset of the compile unit header in the .debug_info is invalid.\n");
388     }
389     return false;
390 }
391 
392 
393 void
394 DWARFCompileUnit::Dump(Stream *s) const
395 {
396     s->Printf("0x%8.8x: Compile Unit: length = 0x%8.8x, version = 0x%4.4x, abbr_offset = 0x%8.8x, addr_size = 0x%2.2x (next CU at {0x%8.8x})\n",
397                 m_offset, m_length, m_version, GetAbbrevOffset(), m_addr_size, GetNextCompileUnitOffset());
398 }
399 
400 
401 static uint8_t g_default_addr_size = 4;
402 
403 uint8_t
404 DWARFCompileUnit::GetAddressByteSize(const DWARFCompileUnit* cu)
405 {
406     if (cu)
407         return cu->GetAddressByteSize();
408     return DWARFCompileUnit::GetDefaultAddressSize();
409 }
410 
411 bool
412 DWARFCompileUnit::IsDWARF64(const DWARFCompileUnit* cu)
413 {
414     if (cu)
415         return cu->IsDWARF64();
416     return false;
417 }
418 
419 uint8_t
420 DWARFCompileUnit::GetDefaultAddressSize()
421 {
422     return g_default_addr_size;
423 }
424 
425 void
426 DWARFCompileUnit::SetDefaultAddressSize(uint8_t addr_size)
427 {
428     g_default_addr_size = addr_size;
429 }
430 
431 lldb::user_id_t
432 DWARFCompileUnit::GetID () const
433 {
434     dw_offset_t local_id = m_base_obj_offset != DW_INVALID_OFFSET ? m_base_obj_offset : m_offset;
435     if (m_dwarf2Data)
436         return DIERef(local_id, local_id).GetUID(m_dwarf2Data);
437     else
438         return local_id;
439 }
440 
441 void
442 DWARFCompileUnit::BuildAddressRangeTable (SymbolFileDWARF* dwarf2Data,
443                                           DWARFDebugAranges* debug_aranges)
444 {
445     // This function is usually called if there in no .debug_aranges section
446     // in order to produce a compile unit level set of address ranges that
447     // is accurate.
448 
449     size_t num_debug_aranges = debug_aranges->GetNumRanges();
450 
451     // First get the compile unit DIE only and check if it has a DW_AT_ranges
452     const DWARFDebugInfoEntry* die = GetCompileUnitDIEPtrOnly();
453 
454     const dw_offset_t cu_offset = GetOffset();
455     if (die)
456     {
457         DWARFRangeList ranges;
458         const size_t num_ranges = die->GetAttributeAddressRanges(dwarf2Data, this, ranges, false);
459         if (num_ranges > 0)
460         {
461             // This compile unit has DW_AT_ranges, assume this is correct if it
462             // is present since clang no longer makes .debug_aranges by default
463             // and it emits DW_AT_ranges for DW_TAG_compile_units. GCC also does
464             // this with recent GCC builds.
465             for (size_t i=0; i<num_ranges; ++i)
466             {
467                 const DWARFRangeList::Entry &range = ranges.GetEntryRef(i);
468                 debug_aranges->AppendRange(cu_offset, range.GetRangeBase(), range.GetRangeEnd());
469             }
470 
471             return; // We got all of our ranges from the DW_AT_ranges attribute
472         }
473     }
474     // We don't have a DW_AT_ranges attribute, so we need to parse the DWARF
475 
476     // If the DIEs weren't parsed, then we don't want all dies for all compile units
477     // to stay loaded when they weren't needed. So we can end up parsing the DWARF
478     // and then throwing them all away to keep memory usage down.
479     const bool clear_dies = ExtractDIEsIfNeeded (false) > 1;
480 
481     die = DIEPtr();
482     if (die)
483         die->BuildAddressRangeTable(dwarf2Data, this, debug_aranges);
484 
485     if (debug_aranges->GetNumRanges() == num_debug_aranges)
486     {
487         // We got nothing from the functions, maybe we have a line tables only
488         // situation. Check the line tables and build the arange table from this.
489         SymbolContext sc;
490         sc.comp_unit = dwarf2Data->GetCompUnitForDWARFCompUnit(this);
491         if (sc.comp_unit)
492         {
493             SymbolFileDWARFDebugMap *debug_map_sym_file = m_dwarf2Data->GetDebugMapSymfile();
494             if (debug_map_sym_file == NULL)
495             {
496                 LineTable *line_table = sc.comp_unit->GetLineTable();
497 
498                 if (line_table)
499                 {
500                     LineTable::FileAddressRanges file_ranges;
501                     const bool append = true;
502                     const size_t num_ranges = line_table->GetContiguousFileAddressRanges (file_ranges, append);
503                     for (uint32_t idx=0; idx<num_ranges; ++idx)
504                     {
505                         const LineTable::FileAddressRanges::Entry &range = file_ranges.GetEntryRef(idx);
506                         debug_aranges->AppendRange(cu_offset, range.GetRangeBase(), range.GetRangeEnd());
507                     }
508                 }
509             }
510             else
511                 debug_map_sym_file->AddOSOARanges(dwarf2Data,debug_aranges);
512         }
513     }
514 
515     if (debug_aranges->GetNumRanges() == num_debug_aranges)
516     {
517         // We got nothing from the functions, maybe we have a line tables only
518         // situation. Check the line tables and build the arange table from this.
519         SymbolContext sc;
520         sc.comp_unit = dwarf2Data->GetCompUnitForDWARFCompUnit(this);
521         if (sc.comp_unit)
522         {
523             LineTable *line_table = sc.comp_unit->GetLineTable();
524 
525             if (line_table)
526             {
527                 LineTable::FileAddressRanges file_ranges;
528                 const bool append = true;
529                 const size_t num_ranges = line_table->GetContiguousFileAddressRanges (file_ranges, append);
530                 for (uint32_t idx=0; idx<num_ranges; ++idx)
531                 {
532                     const LineTable::FileAddressRanges::Entry &range = file_ranges.GetEntryRef(idx);
533                     debug_aranges->AppendRange(GetOffset(), range.GetRangeBase(), range.GetRangeEnd());
534                 }
535             }
536         }
537     }
538 
539     // Keep memory down by clearing DIEs if this generate function
540     // caused them to be parsed
541     if (clear_dies)
542         ClearDIEs (true);
543 
544 }
545 
546 
547 const DWARFDebugAranges &
548 DWARFCompileUnit::GetFunctionAranges ()
549 {
550     if (m_func_aranges_ap.get() == NULL)
551     {
552         m_func_aranges_ap.reset (new DWARFDebugAranges());
553         Log *log (LogChannelDWARF::GetLogIfAll(DWARF_LOG_DEBUG_ARANGES));
554 
555         if (log)
556         {
557             m_dwarf2Data->GetObjectFile()->GetModule()->LogMessage (log,
558                                                                     "DWARFCompileUnit::GetFunctionAranges() for compile unit at .debug_info[0x%8.8x]",
559                                                                     GetOffset());
560         }
561         const DWARFDebugInfoEntry* die = DIEPtr();
562         if (die)
563             die->BuildFunctionAddressRangeTable (m_dwarf2Data, this, m_func_aranges_ap.get());
564 
565         if (m_dwo_symbol_file)
566         {
567             DWARFCompileUnit* dwo_cu = m_dwo_symbol_file->GetCompileUnit();
568             const DWARFDebugInfoEntry* dwo_die = dwo_cu->DIEPtr();
569             if (dwo_die)
570                 dwo_die->BuildFunctionAddressRangeTable (m_dwo_symbol_file.get(),
571                                                          dwo_cu,
572                                                          m_func_aranges_ap.get());
573         }
574 
575         const bool minimize = false;
576         m_func_aranges_ap->Sort(minimize);
577     }
578     return *m_func_aranges_ap.get();
579 }
580 
581 DWARFDIE
582 DWARFCompileUnit::LookupAddress (const dw_addr_t address)
583 {
584     if (DIE())
585     {
586         const DWARFDebugAranges &func_aranges = GetFunctionAranges ();
587 
588         // Re-check the aranges auto pointer contents in case it was created above
589         if (!func_aranges.IsEmpty())
590             return GetDIE(func_aranges.FindAddress(address));
591     }
592     return DWARFDIE();
593 }
594 
595 //----------------------------------------------------------------------
596 // Compare function DWARFDebugAranges::Range structures
597 //----------------------------------------------------------------------
598 static bool CompareDIEOffset (const DWARFDebugInfoEntry& die, const dw_offset_t die_offset)
599 {
600     return die.GetOffset() < die_offset;
601 }
602 
603 //----------------------------------------------------------------------
604 // GetDIE()
605 //
606 // Get the DIE (Debug Information Entry) with the specified offset by
607 // first checking if the DIE is contained within this compile unit and
608 // grabbing the DIE from this compile unit. Otherwise we grab the DIE
609 // from the DWARF file.
610 //----------------------------------------------------------------------
611 DWARFDIE
612 DWARFCompileUnit::GetDIE (dw_offset_t die_offset)
613 {
614     if (die_offset != DW_INVALID_OFFSET)
615     {
616         if (m_dwo_symbol_file)
617             return m_dwo_symbol_file->GetCompileUnit()->GetDIE(die_offset);
618 
619         if (ContainsDIEOffset(die_offset))
620         {
621             ExtractDIEsIfNeeded (false);
622             DWARFDebugInfoEntry::iterator end = m_die_array.end();
623             DWARFDebugInfoEntry::iterator pos = lower_bound(m_die_array.begin(), end, die_offset, CompareDIEOffset);
624             if (pos != end)
625             {
626                 if (die_offset == (*pos).GetOffset())
627                     return DWARFDIE(this, &(*pos));
628             }
629         }
630         else
631         {
632             // Don't specify the compile unit offset as we don't know it because the DIE belongs to
633             // a different compile unit in the same symbol file.
634             return m_dwarf2Data->DebugInfo()->GetDIEForDIEOffset(die_offset);
635         }
636     }
637     return DWARFDIE(); // Not found
638 }
639 
640 size_t
641 DWARFCompileUnit::AppendDIEsWithTag (const dw_tag_t tag, DWARFDIECollection& dies, uint32_t depth) const
642 {
643     size_t old_size = dies.Size();
644     DWARFDebugInfoEntry::const_iterator pos;
645     DWARFDebugInfoEntry::const_iterator end = m_die_array.end();
646     for (pos = m_die_array.begin(); pos != end; ++pos)
647     {
648         if (pos->Tag() == tag)
649             dies.Append (DWARFDIE(this, &(*pos)));
650     }
651 
652     // Return the number of DIEs added to the collection
653     return dies.Size() - old_size;
654 }
655 
656 //void
657 //DWARFCompileUnit::AddGlobalDIEByIndex (uint32_t die_idx)
658 //{
659 //    m_global_die_indexes.push_back (die_idx);
660 //}
661 //
662 //
663 //void
664 //DWARFCompileUnit::AddGlobal (const DWARFDebugInfoEntry* die)
665 //{
666 //    // Indexes to all file level global and static variables
667 //    m_global_die_indexes;
668 //
669 //    if (m_die_array.empty())
670 //        return;
671 //
672 //    const DWARFDebugInfoEntry* first_die = &m_die_array[0];
673 //    const DWARFDebugInfoEntry* end = first_die + m_die_array.size();
674 //    if (first_die <= die && die < end)
675 //        m_global_die_indexes.push_back (die - first_die);
676 //}
677 
678 
679 void
680 DWARFCompileUnit::Index (NameToDIE& func_basenames,
681                          NameToDIE& func_fullnames,
682                          NameToDIE& func_methods,
683                          NameToDIE& func_selectors,
684                          NameToDIE& objc_class_selectors,
685                          NameToDIE& globals,
686                          NameToDIE& types,
687                          NameToDIE& namespaces)
688 {
689     Log *log (LogChannelDWARF::GetLogIfAll (DWARF_LOG_LOOKUPS));
690 
691     if (log)
692     {
693         m_dwarf2Data->GetObjectFile()->GetModule()->LogMessage (log,
694                                                                 "DWARFCompileUnit::Index() for compile unit at .debug_info[0x%8.8x]",
695                                                                 GetOffset());
696     }
697 
698     const LanguageType cu_language = GetLanguageType();
699     DWARFFormValue::FixedFormSizes fixed_form_sizes =
700         DWARFFormValue::GetFixedFormSizesForAddressSize (GetAddressByteSize(), m_is_dwarf64);
701 
702     IndexPrivate(this,
703                  cu_language,
704                  fixed_form_sizes,
705                  GetOffset(),
706                  func_basenames,
707                  func_fullnames,
708                  func_methods,
709                  func_selectors,
710                  objc_class_selectors,
711                  globals,
712                  types,
713                  namespaces);
714 
715     SymbolFileDWARFDwo* dwo_symbol_file = GetDwoSymbolFile();
716     if (dwo_symbol_file)
717     {
718         IndexPrivate(dwo_symbol_file->GetCompileUnit(),
719                      cu_language,
720                      fixed_form_sizes,
721                      GetOffset(),
722                      func_basenames,
723                      func_fullnames,
724                      func_methods,
725                      func_selectors,
726                      objc_class_selectors,
727                      globals,
728                      types,
729                      namespaces);
730     }
731 }
732 
733 void
734 DWARFCompileUnit::IndexPrivate (DWARFCompileUnit* dwarf_cu,
735                                 const LanguageType cu_language,
736                                 const DWARFFormValue::FixedFormSizes& fixed_form_sizes,
737                                 const dw_offset_t cu_offset,
738                                 NameToDIE& func_basenames,
739                                 NameToDIE& func_fullnames,
740                                 NameToDIE& func_methods,
741                                 NameToDIE& func_selectors,
742                                 NameToDIE& objc_class_selectors,
743                                 NameToDIE& globals,
744                                 NameToDIE& types,
745                                 NameToDIE& namespaces)
746 {
747     DWARFDebugInfoEntry::const_iterator pos;
748     DWARFDebugInfoEntry::const_iterator begin = dwarf_cu->m_die_array.begin();
749     DWARFDebugInfoEntry::const_iterator end = dwarf_cu->m_die_array.end();
750     for (pos = begin; pos != end; ++pos)
751     {
752         const DWARFDebugInfoEntry &die = *pos;
753 
754         const dw_tag_t tag = die.Tag();
755 
756         switch (tag)
757         {
758         case DW_TAG_subprogram:
759         case DW_TAG_inlined_subroutine:
760         case DW_TAG_base_type:
761         case DW_TAG_class_type:
762         case DW_TAG_constant:
763         case DW_TAG_enumeration_type:
764         case DW_TAG_string_type:
765         case DW_TAG_subroutine_type:
766         case DW_TAG_structure_type:
767         case DW_TAG_union_type:
768         case DW_TAG_typedef:
769         case DW_TAG_namespace:
770         case DW_TAG_variable:
771         case DW_TAG_unspecified_type:
772             break;
773 
774         default:
775             continue;
776         }
777 
778         DWARFAttributes attributes;
779         const char *name = NULL;
780         const char *mangled_cstr = NULL;
781         bool is_declaration = false;
782         //bool is_artificial = false;
783         bool has_address = false;
784         bool has_location_or_const_value = false;
785         bool is_global_or_static_variable = false;
786 
787         DWARFFormValue specification_die_form;
788         const size_t num_attributes = die.GetAttributes(dwarf_cu, fixed_form_sizes, attributes);
789         if (num_attributes > 0)
790         {
791             for (uint32_t i=0; i<num_attributes; ++i)
792             {
793                 dw_attr_t attr = attributes.AttributeAtIndex(i);
794                 DWARFFormValue form_value;
795                 switch (attr)
796                 {
797                 case DW_AT_name:
798                     if (attributes.ExtractFormValueAtIndex(i, form_value))
799                         name = form_value.AsCString();
800                     break;
801 
802                 case DW_AT_declaration:
803                     if (attributes.ExtractFormValueAtIndex(i, form_value))
804                         is_declaration = form_value.Unsigned() != 0;
805                     break;
806 
807 //                case DW_AT_artificial:
808 //                    if (attributes.ExtractFormValueAtIndex(i, form_value))
809 //                        is_artificial = form_value.Unsigned() != 0;
810 //                    break;
811 
812                 case DW_AT_MIPS_linkage_name:
813                 case DW_AT_linkage_name:
814                     if (attributes.ExtractFormValueAtIndex(i, form_value))
815                         mangled_cstr = form_value.AsCString();
816                     break;
817 
818                 case DW_AT_low_pc:
819                 case DW_AT_high_pc:
820                 case DW_AT_ranges:
821                     has_address = true;
822                     break;
823 
824                 case DW_AT_entry_pc:
825                     has_address = true;
826                     break;
827 
828                 case DW_AT_location:
829                 case DW_AT_const_value:
830                     has_location_or_const_value = true;
831                     if (tag == DW_TAG_variable)
832                     {
833                         const DWARFDebugInfoEntry* parent_die = die.GetParent();
834                         while ( parent_die != NULL )
835                         {
836                             switch (parent_die->Tag())
837                             {
838                             case DW_TAG_subprogram:
839                             case DW_TAG_lexical_block:
840                             case DW_TAG_inlined_subroutine:
841                                 // Even if this is a function level static, we don't add it. We could theoretically
842                                 // add these if we wanted to by introspecting into the DW_AT_location and seeing
843                                 // if the location describes a hard coded address, but we dont want the performance
844                                 // penalty of that right now.
845                                 is_global_or_static_variable = false;
846 //                              if (attributes.ExtractFormValueAtIndex(dwarf2Data, i, form_value))
847 //                              {
848 //                                  // If we have valid block data, then we have location expression bytes
849 //                                  // that are fixed (not a location list).
850 //                                  const uint8_t *block_data = form_value.BlockData();
851 //                                  if (block_data)
852 //                                  {
853 //                                      uint32_t block_length = form_value.Unsigned();
854 //                                      if (block_length == 1 + attributes.CompileUnitAtIndex(i)->GetAddressByteSize())
855 //                                      {
856 //                                          if (block_data[0] == DW_OP_addr)
857 //                                              add_die = true;
858 //                                      }
859 //                                  }
860 //                              }
861                                 parent_die = NULL;  // Terminate the while loop.
862                                 break;
863 
864                             case DW_TAG_compile_unit:
865                                 is_global_or_static_variable = true;
866                                 parent_die = NULL;  // Terminate the while loop.
867                                 break;
868 
869                             default:
870                                 parent_die = parent_die->GetParent();   // Keep going in the while loop.
871                                 break;
872                             }
873                         }
874                     }
875                     break;
876 
877                 case DW_AT_specification:
878                     if (attributes.ExtractFormValueAtIndex(i, form_value))
879                         specification_die_form = form_value;
880                     break;
881                 }
882             }
883         }
884 
885         switch (tag)
886         {
887         case DW_TAG_subprogram:
888             if (has_address)
889             {
890                 if (name)
891                 {
892                     ObjCLanguage::MethodName objc_method(name, true);
893                     if (objc_method.IsValid(true))
894                     {
895                         ConstString objc_class_name_with_category (objc_method.GetClassNameWithCategory());
896                         ConstString objc_selector_name (objc_method.GetSelector());
897                         ConstString objc_fullname_no_category_name (objc_method.GetFullNameWithoutCategory(true));
898                         ConstString objc_class_name_no_category (objc_method.GetClassName());
899                         func_fullnames.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
900                         if (objc_class_name_with_category)
901                             objc_class_selectors.Insert(objc_class_name_with_category, DIERef(cu_offset, die.GetOffset()));
902                         if (objc_class_name_no_category && objc_class_name_no_category != objc_class_name_with_category)
903                             objc_class_selectors.Insert(objc_class_name_no_category, DIERef(cu_offset, die.GetOffset()));
904                         if (objc_selector_name)
905                             func_selectors.Insert (objc_selector_name, DIERef(cu_offset, die.GetOffset()));
906                         if (objc_fullname_no_category_name)
907                             func_fullnames.Insert (objc_fullname_no_category_name, DIERef(cu_offset, die.GetOffset()));
908                     }
909                     // If we have a mangled name, then the DW_AT_name attribute
910                     // is usually the method name without the class or any parameters
911                     const DWARFDebugInfoEntry *parent = die.GetParent();
912                     bool is_method = false;
913                     if (parent)
914                     {
915                         dw_tag_t parent_tag = parent->Tag();
916                         if (parent_tag == DW_TAG_class_type || parent_tag == DW_TAG_structure_type)
917                         {
918                             is_method = true;
919                         }
920                         else
921                         {
922                             if (specification_die_form.IsValid())
923                             {
924                                 DWARFDIE specification_die = dwarf_cu->GetSymbolFileDWARF()->DebugInfo()->GetDIE (DIERef(specification_die_form));
925                                 if (specification_die.GetParent().IsStructOrClass())
926                                     is_method = true;
927                             }
928                         }
929                     }
930 
931 
932                     if (is_method)
933                         func_methods.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
934                     else
935                         func_basenames.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
936 
937                     if (!is_method && !mangled_cstr && !objc_method.IsValid(true))
938                         func_fullnames.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
939                 }
940                 if (mangled_cstr)
941                 {
942                     // Make sure our mangled name isn't the same string table entry
943                     // as our name. If it starts with '_', then it is ok, else compare
944                     // the string to make sure it isn't the same and we don't end up
945                     // with duplicate entries
946                     if (name != mangled_cstr && ((mangled_cstr[0] == '_') || (name && ::strcmp(name, mangled_cstr) != 0)))
947                     {
948                         Mangled mangled (ConstString(mangled_cstr), true);
949                         func_fullnames.Insert (mangled.GetMangledName(), DIERef(cu_offset, die.GetOffset()));
950                         ConstString demangled = mangled.GetDemangledName(cu_language);
951                         if (demangled)
952                             func_fullnames.Insert (demangled, DIERef(cu_offset, die.GetOffset()));
953                     }
954                 }
955             }
956             break;
957 
958         case DW_TAG_inlined_subroutine:
959             if (has_address)
960             {
961                 if (name)
962                     func_basenames.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
963                 if (mangled_cstr)
964                 {
965                     // Make sure our mangled name isn't the same string table entry
966                     // as our name. If it starts with '_', then it is ok, else compare
967                     // the string to make sure it isn't the same and we don't end up
968                     // with duplicate entries
969                     if (name != mangled_cstr && ((mangled_cstr[0] == '_') || (::strcmp(name, mangled_cstr) != 0)))
970                     {
971                         Mangled mangled (ConstString(mangled_cstr), true);
972                         func_fullnames.Insert (mangled.GetMangledName(), DIERef(cu_offset, die.GetOffset()));
973                         ConstString demangled = mangled.GetDemangledName(cu_language);
974                         if (demangled)
975                             func_fullnames.Insert (demangled, DIERef(cu_offset, die.GetOffset()));
976                     }
977                 }
978                 else
979                     func_fullnames.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
980             }
981             break;
982 
983         case DW_TAG_base_type:
984         case DW_TAG_class_type:
985         case DW_TAG_constant:
986         case DW_TAG_enumeration_type:
987         case DW_TAG_string_type:
988         case DW_TAG_subroutine_type:
989         case DW_TAG_structure_type:
990         case DW_TAG_union_type:
991         case DW_TAG_typedef:
992         case DW_TAG_unspecified_type:
993             if (name && !is_declaration)
994                 types.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
995             if (mangled_cstr && !is_declaration)
996                 types.Insert (ConstString(mangled_cstr), DIERef(cu_offset, die.GetOffset()));
997             break;
998 
999         case DW_TAG_namespace:
1000             if (name)
1001                 namespaces.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
1002             break;
1003 
1004         case DW_TAG_variable:
1005             if (name && has_location_or_const_value && is_global_or_static_variable)
1006             {
1007                 globals.Insert (ConstString(name), DIERef(cu_offset, die.GetOffset()));
1008                 // Be sure to include variables by their mangled and demangled
1009                 // names if they have any since a variable can have a basename
1010                 // "i", a mangled named "_ZN12_GLOBAL__N_11iE" and a demangled
1011                 // mangled name "(anonymous namespace)::i"...
1012 
1013                 // Make sure our mangled name isn't the same string table entry
1014                 // as our name. If it starts with '_', then it is ok, else compare
1015                 // the string to make sure it isn't the same and we don't end up
1016                 // with duplicate entries
1017                 if (mangled_cstr && name != mangled_cstr && ((mangled_cstr[0] == '_') || (::strcmp(name, mangled_cstr) != 0)))
1018                 {
1019                     Mangled mangled (ConstString(mangled_cstr), true);
1020                     globals.Insert (mangled.GetMangledName(), DIERef(cu_offset, die.GetOffset()));
1021                     ConstString demangled = mangled.GetDemangledName(cu_language);
1022                     if (demangled)
1023                         globals.Insert (demangled, DIERef(cu_offset, die.GetOffset()));
1024                 }
1025             }
1026             break;
1027 
1028         default:
1029             continue;
1030         }
1031     }
1032 }
1033 
1034 bool
1035 DWARFCompileUnit::Supports_unnamed_objc_bitfields ()
1036 {
1037     if (GetProducer() == eProducerClang)
1038     {
1039         const uint32_t major_version = GetProducerVersionMajor();
1040         if (major_version > 425 || (major_version == 425 && GetProducerVersionUpdate() >= 13))
1041             return true;
1042         else
1043             return false;
1044     }
1045     return true; // Assume all other compilers didn't have incorrect ObjC bitfield info
1046 }
1047 
1048 bool
1049 DWARFCompileUnit::Supports_DW_AT_APPLE_objc_complete_type ()
1050 {
1051     if (GetProducer() == eProducerLLVMGCC)
1052         return false;
1053     return true;
1054 }
1055 
1056 bool
1057 DWARFCompileUnit::DW_AT_decl_file_attributes_are_invalid()
1058 {
1059     // llvm-gcc makes completely invalid decl file attributes and won't ever
1060     // be fixed, so we need to know to ignore these.
1061     return GetProducer() == eProducerLLVMGCC;
1062 }
1063 
1064 void
1065 DWARFCompileUnit::ParseProducerInfo ()
1066 {
1067     m_producer_version_major = UINT32_MAX;
1068     m_producer_version_minor = UINT32_MAX;
1069     m_producer_version_update = UINT32_MAX;
1070 
1071     const DWARFDebugInfoEntry *die = GetCompileUnitDIEPtrOnly();
1072     if (die)
1073     {
1074 
1075         const char *producer_cstr = die->GetAttributeValueAsString(m_dwarf2Data, this, DW_AT_producer, NULL);
1076         if (producer_cstr)
1077         {
1078             RegularExpression llvm_gcc_regex("^4\\.[012]\\.[01] \\(Based on Apple Inc\\. build [0-9]+\\) \\(LLVM build [\\.0-9]+\\)$");
1079             if (llvm_gcc_regex.Execute (producer_cstr))
1080             {
1081                 m_producer = eProducerLLVMGCC;
1082             }
1083             else if (strstr(producer_cstr, "clang"))
1084             {
1085                 static RegularExpression g_clang_version_regex("clang-([0-9]+)\\.([0-9]+)\\.([0-9]+)");
1086                 RegularExpression::Match regex_match(3);
1087                 if (g_clang_version_regex.Execute (producer_cstr, &regex_match))
1088                 {
1089                     std::string str;
1090                     if (regex_match.GetMatchAtIndex (producer_cstr, 1, str))
1091                         m_producer_version_major = StringConvert::ToUInt32(str.c_str(), UINT32_MAX, 10);
1092                     if (regex_match.GetMatchAtIndex (producer_cstr, 2, str))
1093                         m_producer_version_minor = StringConvert::ToUInt32(str.c_str(), UINT32_MAX, 10);
1094                     if (regex_match.GetMatchAtIndex (producer_cstr, 3, str))
1095                         m_producer_version_update = StringConvert::ToUInt32(str.c_str(), UINT32_MAX, 10);
1096                 }
1097                 m_producer = eProducerClang;
1098             }
1099             else if (strstr(producer_cstr, "GNU"))
1100                 m_producer = eProducerGCC;
1101         }
1102     }
1103     if (m_producer == eProducerInvalid)
1104         m_producer = eProcucerOther;
1105 }
1106 
1107 DWARFCompileUnit::Producer
1108 DWARFCompileUnit::GetProducer ()
1109 {
1110     if (m_producer == eProducerInvalid)
1111         ParseProducerInfo ();
1112     return m_producer;
1113 }
1114 
1115 
1116 uint32_t
1117 DWARFCompileUnit::GetProducerVersionMajor()
1118 {
1119     if (m_producer_version_major == 0)
1120         ParseProducerInfo ();
1121     return m_producer_version_major;
1122 }
1123 
1124 uint32_t
1125 DWARFCompileUnit::GetProducerVersionMinor()
1126 {
1127     if (m_producer_version_minor == 0)
1128         ParseProducerInfo ();
1129     return m_producer_version_minor;
1130 }
1131 
1132 uint32_t
1133 DWARFCompileUnit::GetProducerVersionUpdate()
1134 {
1135     if (m_producer_version_update == 0)
1136         ParseProducerInfo ();
1137     return m_producer_version_update;
1138 }
1139 
1140 LanguageType
1141 DWARFCompileUnit::LanguageTypeFromDWARF(uint64_t val)
1142 {
1143     // Note: user languages between lo_user and hi_user
1144     // must be handled explicitly here.
1145     switch (val)
1146     {
1147     case DW_LANG_Mips_Assembler:
1148         return eLanguageTypeMipsAssembler;
1149     case DW_LANG_GOOGLE_RenderScript:
1150         return eLanguageTypeExtRenderScript;
1151     default:
1152         return static_cast<LanguageType>(val);
1153     }
1154 }
1155 
1156 LanguageType
1157 DWARFCompileUnit::GetLanguageType()
1158 {
1159     if (m_language_type != eLanguageTypeUnknown)
1160         return m_language_type;
1161 
1162     const DWARFDebugInfoEntry *die = GetCompileUnitDIEPtrOnly();
1163     if (die)
1164         m_language_type = LanguageTypeFromDWARF(die->GetAttributeValueAsUnsigned(m_dwarf2Data, this, DW_AT_language, 0));
1165     return m_language_type;
1166 }
1167 
1168 bool
1169 DWARFCompileUnit::IsDWARF64() const
1170 {
1171     return m_is_dwarf64;
1172 }
1173 
1174 bool
1175 DWARFCompileUnit::GetIsOptimized ()
1176 {
1177     if (m_is_optimized == eLazyBoolCalculate)
1178     {
1179         const DWARFDebugInfoEntry *die = GetCompileUnitDIEPtrOnly();
1180         if (die)
1181         {
1182             m_is_optimized = eLazyBoolNo;
1183             if (die->GetAttributeValueAsUnsigned (m_dwarf2Data, this, DW_AT_APPLE_optimized, 0) == 1)
1184             {
1185                 m_is_optimized = eLazyBoolYes;
1186             }
1187         }
1188     }
1189     if (m_is_optimized == eLazyBoolYes)
1190     {
1191         return true;
1192     }
1193     else
1194     {
1195         return false;
1196     }
1197 }
1198 
1199 DWARFFormValue::FixedFormSizes
1200 DWARFCompileUnit::GetFixedFormSizes ()
1201 {
1202     return DWARFFormValue::GetFixedFormSizesForAddressSize (GetAddressByteSize(), IsDWARF64());
1203 }
1204 
1205 TypeSystem *
1206 DWARFCompileUnit::GetTypeSystem ()
1207 {
1208     if (m_dwarf2Data)
1209         return m_dwarf2Data->GetTypeSystemForLanguage(GetLanguageType());
1210     else
1211         return nullptr;
1212 }
1213 
1214 void
1215 DWARFCompileUnit::SetUserData(void *d)
1216 {
1217     m_user_data = d;
1218     if (m_dwo_symbol_file)
1219         m_dwo_symbol_file->GetCompileUnit()->SetUserData(d);
1220 }
1221 
1222 void
1223 DWARFCompileUnit::SetAddrBase(dw_addr_t addr_base, dw_offset_t base_obj_offset)
1224 {
1225     m_addr_base = addr_base;
1226     m_base_obj_offset = base_obj_offset;
1227 }
1228 
1229 lldb::ByteOrder
1230 DWARFCompileUnit::GetByteOrder() const
1231 {
1232     return m_dwarf2Data->GetObjectFile()->GetByteOrder();
1233 }
1234