1 //===- DWARFDebugLine.cpp -------------------------------------------------===//
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 "llvm/DebugInfo/DWARF/DWARFDebugLine.h"
11 #include "llvm/ADT/SmallString.h"
12 #include "llvm/ADT/SmallVector.h"
13 #include "llvm/ADT/StringRef.h"
14 #include "llvm/BinaryFormat/Dwarf.h"
15 #include "llvm/DebugInfo/DWARF/DWARFContext.h"
16 #include "llvm/DebugInfo/DWARF/DWARFFormValue.h"
17 #include "llvm/DebugInfo/DWARF/DWARFRelocMap.h"
18 #include "llvm/Support/Format.h"
19 #include "llvm/Support/Path.h"
20 #include "llvm/Support/raw_ostream.h"
21 #include <algorithm>
22 #include <cassert>
23 #include <cinttypes>
24 #include <cstdint>
25 #include <cstdio>
26 #include <utility>
27 
28 using namespace llvm;
29 using namespace dwarf;
30 
31 using FileLineInfoKind = DILineInfoSpecifier::FileLineInfoKind;
32 
33 namespace {
34 
35 struct ContentDescriptor {
36   dwarf::LineNumberEntryFormat Type;
37   dwarf::Form Form;
38 };
39 
40 using ContentDescriptors = SmallVector<ContentDescriptor, 4>;
41 
42 } // end anonmyous namespace
43 
44 DWARFDebugLine::Prologue::Prologue() { clear(); }
45 
46 void DWARFDebugLine::Prologue::clear() {
47   TotalLength = PrologueLength = 0;
48   SegSelectorSize = 0;
49   MinInstLength = MaxOpsPerInst = DefaultIsStmt = LineBase = LineRange = 0;
50   OpcodeBase = 0;
51   FormParams = DWARFFormParams({0, 0, DWARF32});
52   StandardOpcodeLengths.clear();
53   IncludeDirectories.clear();
54   FileNames.clear();
55 }
56 
57 void DWARFDebugLine::Prologue::dump(raw_ostream &OS) const {
58   OS << "Line table prologue:\n"
59      << format("    total_length: 0x%8.8" PRIx64 "\n", TotalLength)
60      << format("         version: %u\n", getVersion());
61   if (getVersion() >= 5)
62     OS << format("    address_size: %u\n", getAddressSize())
63        << format(" seg_select_size: %u\n", SegSelectorSize);
64   OS << format(" prologue_length: 0x%8.8" PRIx64 "\n", PrologueLength)
65      << format(" min_inst_length: %u\n", MinInstLength)
66      << format(getVersion() >= 4 ? "max_ops_per_inst: %u\n" : "", MaxOpsPerInst)
67      << format(" default_is_stmt: %u\n", DefaultIsStmt)
68      << format("       line_base: %i\n", LineBase)
69      << format("      line_range: %u\n", LineRange)
70      << format("     opcode_base: %u\n", OpcodeBase);
71 
72   for (uint32_t I = 0; I != StandardOpcodeLengths.size(); ++I)
73     OS << format("standard_opcode_lengths[%s] = %u\n",
74                  LNStandardString(I + 1).data(), StandardOpcodeLengths[I]);
75 
76   if (!IncludeDirectories.empty())
77     for (uint32_t I = 0; I != IncludeDirectories.size(); ++I)
78       OS << format("include_directories[%3u] = '", I + 1)
79          << IncludeDirectories[I] << "'\n";
80 
81   if (!FileNames.empty()) {
82     OS << "                Dir  Mod Time   File Len   File Name\n"
83        << "                ---- ---------- ---------- -----------"
84           "----------------\n";
85     for (uint32_t I = 0; I != FileNames.size(); ++I) {
86       const FileNameEntry &FileEntry = FileNames[I];
87       OS << format("file_names[%3u] %4" PRIu64 " ", I + 1, FileEntry.DirIdx)
88          << format("0x%8.8" PRIx64 " 0x%8.8" PRIx64 " ", FileEntry.ModTime,
89                    FileEntry.Length)
90          << FileEntry.Name << '\n';
91     }
92   }
93 }
94 
95 // Parse v2-v4 directory and file tables.
96 static void
97 parseV2DirFileTables(const DWARFDataExtractor &DebugLineData,
98                      uint32_t *OffsetPtr, uint64_t EndPrologueOffset,
99                      std::vector<StringRef> &IncludeDirectories,
100                      std::vector<DWARFDebugLine::FileNameEntry> &FileNames) {
101   while (*OffsetPtr < EndPrologueOffset) {
102     StringRef S = DebugLineData.getCStrRef(OffsetPtr);
103     if (S.empty())
104       break;
105     IncludeDirectories.push_back(S);
106   }
107 
108   while (*OffsetPtr < EndPrologueOffset) {
109     StringRef Name = DebugLineData.getCStrRef(OffsetPtr);
110     if (Name.empty())
111       break;
112     DWARFDebugLine::FileNameEntry FileEntry;
113     FileEntry.Name = Name;
114     FileEntry.DirIdx = DebugLineData.getULEB128(OffsetPtr);
115     FileEntry.ModTime = DebugLineData.getULEB128(OffsetPtr);
116     FileEntry.Length = DebugLineData.getULEB128(OffsetPtr);
117     FileNames.push_back(FileEntry);
118   }
119 }
120 
121 // Parse v5 directory/file entry content descriptions.
122 // Returns the descriptors, or an empty vector if we did not find a path or
123 // ran off the end of the prologue.
124 static ContentDescriptors
125 parseV5EntryFormat(const DWARFDataExtractor &DebugLineData, uint32_t *OffsetPtr,
126                    uint64_t EndPrologueOffset) {
127   ContentDescriptors Descriptors;
128   int FormatCount = DebugLineData.getU8(OffsetPtr);
129   bool HasPath = false;
130   for (int I = 0; I != FormatCount; ++I) {
131     if (*OffsetPtr >= EndPrologueOffset)
132       return ContentDescriptors();
133     ContentDescriptor Descriptor;
134     Descriptor.Type =
135       dwarf::LineNumberEntryFormat(DebugLineData.getULEB128(OffsetPtr));
136     Descriptor.Form = dwarf::Form(DebugLineData.getULEB128(OffsetPtr));
137     if (Descriptor.Type == dwarf::DW_LNCT_path)
138       HasPath = true;
139     Descriptors.push_back(Descriptor);
140   }
141   return HasPath ? Descriptors : ContentDescriptors();
142 }
143 
144 static bool
145 parseV5DirFileTables(const DWARFDataExtractor &DebugLineData,
146                      uint32_t *OffsetPtr, uint64_t EndPrologueOffset,
147                      const DWARFFormParams &FormParams,
148                      std::vector<StringRef> &IncludeDirectories,
149                      std::vector<DWARFDebugLine::FileNameEntry> &FileNames) {
150   // Get the directory entry description.
151   ContentDescriptors DirDescriptors =
152     parseV5EntryFormat(DebugLineData, OffsetPtr, EndPrologueOffset);
153   if (DirDescriptors.empty())
154     return false;
155 
156   // Get the directory entries, according to the format described above.
157   int DirEntryCount = DebugLineData.getU8(OffsetPtr);
158   for (int I = 0; I != DirEntryCount; ++I) {
159     if (*OffsetPtr >= EndPrologueOffset)
160       return false;
161     for (auto Descriptor : DirDescriptors) {
162       DWARFFormValue Value(Descriptor.Form);
163       switch (Descriptor.Type) {
164       case DW_LNCT_path:
165         if (!Value.extractValue(DebugLineData, OffsetPtr, nullptr))
166           return false;
167         IncludeDirectories.push_back(Value.getAsCString().getValue());
168         break;
169       default:
170         if (!Value.skipValue(DebugLineData, OffsetPtr, FormParams))
171           return false;
172       }
173     }
174   }
175 
176   // Get the file entry description.
177   ContentDescriptors FileDescriptors =
178     parseV5EntryFormat(DebugLineData, OffsetPtr, EndPrologueOffset);
179   if (FileDescriptors.empty())
180     return false;
181 
182   // Get the file entries, according to the format described above.
183   int FileEntryCount = DebugLineData.getU8(OffsetPtr);
184   for (int I = 0; I != FileEntryCount; ++I) {
185     if (*OffsetPtr >= EndPrologueOffset)
186       return false;
187     DWARFDebugLine::FileNameEntry FileEntry;
188     for (auto Descriptor : FileDescriptors) {
189       DWARFFormValue Value(Descriptor.Form);
190       if (!Value.extractValue(DebugLineData, OffsetPtr, nullptr))
191         return false;
192       switch (Descriptor.Type) {
193       case DW_LNCT_path:
194         FileEntry.Name = Value.getAsCString().getValue();
195         break;
196       case DW_LNCT_directory_index:
197         FileEntry.DirIdx = Value.getAsUnsignedConstant().getValue();
198         break;
199       case DW_LNCT_timestamp:
200         FileEntry.ModTime = Value.getAsUnsignedConstant().getValue();
201         break;
202       case DW_LNCT_size:
203         FileEntry.Length = Value.getAsUnsignedConstant().getValue();
204         break;
205       // FIXME: Add MD5
206       default:
207         break;
208       }
209     }
210     FileNames.push_back(FileEntry);
211   }
212   return true;
213 }
214 
215 bool DWARFDebugLine::Prologue::parse(const DWARFDataExtractor &DebugLineData,
216                                      uint32_t *OffsetPtr) {
217   const uint64_t PrologueOffset = *OffsetPtr;
218 
219   clear();
220   TotalLength = DebugLineData.getU32(OffsetPtr);
221   if (TotalLength == UINT32_MAX) {
222     FormParams.Format = dwarf::DWARF64;
223     TotalLength = DebugLineData.getU64(OffsetPtr);
224   } else if (TotalLength >= 0xffffff00) {
225     return false;
226   }
227   FormParams.Version = DebugLineData.getU16(OffsetPtr);
228   if (getVersion() < 2)
229     return false;
230 
231   if (getVersion() >= 5) {
232     FormParams.AddrSize = DebugLineData.getU8(OffsetPtr);
233     assert(getAddressSize() == DebugLineData.getAddressSize() &&
234            "Line table header and data extractor disagree");
235     SegSelectorSize = DebugLineData.getU8(OffsetPtr);
236   }
237 
238   PrologueLength = DebugLineData.getUnsigned(OffsetPtr, sizeofPrologueLength());
239   const uint64_t EndPrologueOffset = PrologueLength + *OffsetPtr;
240   MinInstLength = DebugLineData.getU8(OffsetPtr);
241   if (getVersion() >= 4)
242     MaxOpsPerInst = DebugLineData.getU8(OffsetPtr);
243   DefaultIsStmt = DebugLineData.getU8(OffsetPtr);
244   LineBase = DebugLineData.getU8(OffsetPtr);
245   LineRange = DebugLineData.getU8(OffsetPtr);
246   OpcodeBase = DebugLineData.getU8(OffsetPtr);
247 
248   StandardOpcodeLengths.reserve(OpcodeBase - 1);
249   for (uint32_t I = 1; I < OpcodeBase; ++I) {
250     uint8_t OpLen = DebugLineData.getU8(OffsetPtr);
251     StandardOpcodeLengths.push_back(OpLen);
252   }
253 
254   if (getVersion() >= 5) {
255     if (!parseV5DirFileTables(DebugLineData, OffsetPtr, EndPrologueOffset,
256                               getFormParams(), IncludeDirectories, FileNames)) {
257       fprintf(stderr,
258               "warning: parsing line table prologue at 0x%8.8" PRIx64
259               " found an invalid directory or file table description at"
260               " 0x%8.8" PRIx64 "\n", PrologueOffset, (uint64_t)*OffsetPtr);
261       return false;
262     }
263   } else
264     parseV2DirFileTables(DebugLineData, OffsetPtr, EndPrologueOffset,
265                          IncludeDirectories, FileNames);
266 
267   if (*OffsetPtr != EndPrologueOffset) {
268     fprintf(stderr,
269             "warning: parsing line table prologue at 0x%8.8" PRIx64
270             " should have ended at 0x%8.8" PRIx64
271             " but it ended at 0x%8.8" PRIx64 "\n",
272             PrologueOffset, EndPrologueOffset, (uint64_t)*OffsetPtr);
273     return false;
274   }
275   return true;
276 }
277 
278 DWARFDebugLine::Row::Row(bool DefaultIsStmt) { reset(DefaultIsStmt); }
279 
280 void DWARFDebugLine::Row::postAppend() {
281   BasicBlock = false;
282   PrologueEnd = false;
283   EpilogueBegin = false;
284 }
285 
286 void DWARFDebugLine::Row::reset(bool DefaultIsStmt) {
287   Address = 0;
288   Line = 1;
289   Column = 0;
290   File = 1;
291   Isa = 0;
292   Discriminator = 0;
293   IsStmt = DefaultIsStmt;
294   BasicBlock = false;
295   EndSequence = false;
296   PrologueEnd = false;
297   EpilogueBegin = false;
298 }
299 
300 void DWARFDebugLine::Row::dumpTableHeader(raw_ostream &OS) {
301   OS << "Address            Line   Column File   ISA Discriminator Flags\n"
302      << "------------------ ------ ------ ------ --- ------------- "
303         "-------------\n";
304 }
305 
306 void DWARFDebugLine::Row::dump(raw_ostream &OS) const {
307   OS << format("0x%16.16" PRIx64 " %6u %6u", Address, Line, Column)
308      << format(" %6u %3u %13u ", File, Isa, Discriminator)
309      << (IsStmt ? " is_stmt" : "") << (BasicBlock ? " basic_block" : "")
310      << (PrologueEnd ? " prologue_end" : "")
311      << (EpilogueBegin ? " epilogue_begin" : "")
312      << (EndSequence ? " end_sequence" : "") << '\n';
313 }
314 
315 DWARFDebugLine::Sequence::Sequence() { reset(); }
316 
317 void DWARFDebugLine::Sequence::reset() {
318   LowPC = 0;
319   HighPC = 0;
320   FirstRowIndex = 0;
321   LastRowIndex = 0;
322   Empty = true;
323 }
324 
325 DWARFDebugLine::LineTable::LineTable() { clear(); }
326 
327 void DWARFDebugLine::LineTable::dump(raw_ostream &OS) const {
328   Prologue.dump(OS);
329   OS << '\n';
330 
331   if (!Rows.empty()) {
332     Row::dumpTableHeader(OS);
333     for (const Row &R : Rows) {
334       R.dump(OS);
335     }
336   }
337 }
338 
339 void DWARFDebugLine::LineTable::clear() {
340   Prologue.clear();
341   Rows.clear();
342   Sequences.clear();
343 }
344 
345 DWARFDebugLine::ParsingState::ParsingState(struct LineTable *LT)
346     : LineTable(LT) {
347   resetRowAndSequence();
348 }
349 
350 void DWARFDebugLine::ParsingState::resetRowAndSequence() {
351   Row.reset(LineTable->Prologue.DefaultIsStmt);
352   Sequence.reset();
353 }
354 
355 void DWARFDebugLine::ParsingState::appendRowToMatrix(uint32_t Offset) {
356   if (Sequence.Empty) {
357     // Record the beginning of instruction sequence.
358     Sequence.Empty = false;
359     Sequence.LowPC = Row.Address;
360     Sequence.FirstRowIndex = RowNumber;
361   }
362   ++RowNumber;
363   LineTable->appendRow(Row);
364   if (Row.EndSequence) {
365     // Record the end of instruction sequence.
366     Sequence.HighPC = Row.Address;
367     Sequence.LastRowIndex = RowNumber;
368     if (Sequence.isValid())
369       LineTable->appendSequence(Sequence);
370     Sequence.reset();
371   }
372   Row.postAppend();
373 }
374 
375 const DWARFDebugLine::LineTable *
376 DWARFDebugLine::getLineTable(uint32_t Offset) const {
377   LineTableConstIter Pos = LineTableMap.find(Offset);
378   if (Pos != LineTableMap.end())
379     return &Pos->second;
380   return nullptr;
381 }
382 
383 const DWARFDebugLine::LineTable *
384 DWARFDebugLine::getOrParseLineTable(const DWARFDataExtractor &DebugLineData,
385                                     uint32_t Offset) {
386   std::pair<LineTableIter, bool> Pos =
387       LineTableMap.insert(LineTableMapTy::value_type(Offset, LineTable()));
388   LineTable *LT = &Pos.first->second;
389   if (Pos.second) {
390     if (!LT->parse(DebugLineData, &Offset))
391       return nullptr;
392   }
393   return LT;
394 }
395 
396 bool DWARFDebugLine::LineTable::parse(const DWARFDataExtractor &DebugLineData,
397                                       uint32_t *OffsetPtr, raw_ostream *OS) {
398   const uint32_t DebugLineOffset = *OffsetPtr;
399 
400   clear();
401 
402   if (!Prologue.parse(DebugLineData, OffsetPtr)) {
403     // Restore our offset and return false to indicate failure!
404     *OffsetPtr = DebugLineOffset;
405     return false;
406   }
407 
408   if (OS)
409     Prologue.dump(*OS);
410 
411   const uint32_t EndOffset =
412       DebugLineOffset + Prologue.TotalLength + Prologue.sizeofTotalLength();
413 
414   ParsingState State(this);
415 
416   while (*OffsetPtr < EndOffset) {
417     if (OS)
418       *OS << format("0x%08.08" PRIx32 ": ", *OffsetPtr);
419 
420     uint8_t Opcode = DebugLineData.getU8(OffsetPtr);
421 
422     if (OS)
423       *OS << format("%02.02" PRIx8 " ", Opcode);
424 
425     if (Opcode == 0) {
426       // Extended Opcodes always start with a zero opcode followed by
427       // a uleb128 length so you can skip ones you don't know about
428       uint32_t ExtOffset = *OffsetPtr;
429       uint64_t Len = DebugLineData.getULEB128(OffsetPtr);
430       uint32_t ArgSize = Len - (*OffsetPtr - ExtOffset);
431 
432       uint8_t SubOpcode = DebugLineData.getU8(OffsetPtr);
433       if (OS)
434         *OS << LNExtendedString(SubOpcode);
435       switch (SubOpcode) {
436       case DW_LNE_end_sequence:
437         // Set the end_sequence register of the state machine to true and
438         // append a row to the matrix using the current values of the
439         // state-machine registers. Then reset the registers to the initial
440         // values specified above. Every statement program sequence must end
441         // with a DW_LNE_end_sequence instruction which creates a row whose
442         // address is that of the byte after the last target machine instruction
443         // of the sequence.
444         State.Row.EndSequence = true;
445         State.appendRowToMatrix(*OffsetPtr);
446         if (OS) {
447           *OS << "\n";
448           OS->indent(12);
449           State.Row.dump(*OS);
450         }
451         State.resetRowAndSequence();
452         break;
453 
454       case DW_LNE_set_address:
455         // Takes a single relocatable address as an operand. The size of the
456         // operand is the size appropriate to hold an address on the target
457         // machine. Set the address register to the value given by the
458         // relocatable address. All of the other statement program opcodes
459         // that affect the address register add a delta to it. This instruction
460         // stores a relocatable value into it instead.
461         State.Row.Address = DebugLineData.getRelocatedAddress(OffsetPtr);
462         if (OS)
463           *OS << format(" (0x%16.16" PRIx64 ")", State.Row.Address);
464         break;
465 
466       case DW_LNE_define_file:
467         // Takes 4 arguments. The first is a null terminated string containing
468         // a source file name. The second is an unsigned LEB128 number
469         // representing the directory index of the directory in which the file
470         // was found. The third is an unsigned LEB128 number representing the
471         // time of last modification of the file. The fourth is an unsigned
472         // LEB128 number representing the length in bytes of the file. The time
473         // and length fields may contain LEB128(0) if the information is not
474         // available.
475         //
476         // The directory index represents an entry in the include_directories
477         // section of the statement program prologue. The index is LEB128(0)
478         // if the file was found in the current directory of the compilation,
479         // LEB128(1) if it was found in the first directory in the
480         // include_directories section, and so on. The directory index is
481         // ignored for file names that represent full path names.
482         //
483         // The files are numbered, starting at 1, in the order in which they
484         // appear; the names in the prologue come before names defined by
485         // the DW_LNE_define_file instruction. These numbers are used in the
486         // the file register of the state machine.
487         {
488           FileNameEntry FileEntry;
489           FileEntry.Name = DebugLineData.getCStr(OffsetPtr);
490           FileEntry.DirIdx = DebugLineData.getULEB128(OffsetPtr);
491           FileEntry.ModTime = DebugLineData.getULEB128(OffsetPtr);
492           FileEntry.Length = DebugLineData.getULEB128(OffsetPtr);
493           Prologue.FileNames.push_back(FileEntry);
494           if (OS)
495             *OS << " (" << FileEntry.Name.str()
496                 << ", dir=" << FileEntry.DirIdx << ", mod_time="
497                 << format("(0x%16.16" PRIx64 ")", FileEntry.ModTime)
498                 << ", length=" << FileEntry.Length << ")";
499         }
500         break;
501 
502       case DW_LNE_set_discriminator:
503         State.Row.Discriminator = DebugLineData.getULEB128(OffsetPtr);
504         if (OS)
505           *OS << " (" << State.Row.Discriminator << ")";
506         break;
507 
508       default:
509         // Length doesn't include the zero opcode byte or the length itself, but
510         // it does include the sub_opcode, so we have to adjust for that below
511         (*OffsetPtr) += ArgSize;
512         break;
513       }
514     } else if (Opcode < Prologue.OpcodeBase) {
515       if (OS)
516         *OS << LNStandardString(Opcode);
517       switch (Opcode) {
518       // Standard Opcodes
519       case DW_LNS_copy:
520         // Takes no arguments. Append a row to the matrix using the
521         // current values of the state-machine registers. Then set
522         // the basic_block register to false.
523         State.appendRowToMatrix(*OffsetPtr);
524         if (OS) {
525           *OS << "\n";
526           OS->indent(12);
527           State.Row.dump(*OS);
528           *OS << "\n";
529         }
530         break;
531 
532       case DW_LNS_advance_pc:
533         // Takes a single unsigned LEB128 operand, multiplies it by the
534         // min_inst_length field of the prologue, and adds the
535         // result to the address register of the state machine.
536         {
537           uint64_t AddrOffset =
538               DebugLineData.getULEB128(OffsetPtr) * Prologue.MinInstLength;
539           State.Row.Address += AddrOffset;
540           if (OS)
541             *OS << " (" << AddrOffset << ")";
542         }
543         break;
544 
545       case DW_LNS_advance_line:
546         // Takes a single signed LEB128 operand and adds that value to
547         // the line register of the state machine.
548         State.Row.Line += DebugLineData.getSLEB128(OffsetPtr);
549         if (OS)
550           *OS << " (" << State.Row.Line << ")";
551         break;
552 
553       case DW_LNS_set_file:
554         // Takes a single unsigned LEB128 operand and stores it in the file
555         // register of the state machine.
556         State.Row.File = DebugLineData.getULEB128(OffsetPtr);
557         if (OS)
558           *OS << " (" << State.Row.File << ")";
559         break;
560 
561       case DW_LNS_set_column:
562         // Takes a single unsigned LEB128 operand and stores it in the
563         // column register of the state machine.
564         State.Row.Column = DebugLineData.getULEB128(OffsetPtr);
565         if (OS)
566           *OS << " (" << State.Row.Column << ")";
567         break;
568 
569       case DW_LNS_negate_stmt:
570         // Takes no arguments. Set the is_stmt register of the state
571         // machine to the logical negation of its current value.
572         State.Row.IsStmt = !State.Row.IsStmt;
573         break;
574 
575       case DW_LNS_set_basic_block:
576         // Takes no arguments. Set the basic_block register of the
577         // state machine to true
578         State.Row.BasicBlock = true;
579         break;
580 
581       case DW_LNS_const_add_pc:
582         // Takes no arguments. Add to the address register of the state
583         // machine the address increment value corresponding to special
584         // opcode 255. The motivation for DW_LNS_const_add_pc is this:
585         // when the statement program needs to advance the address by a
586         // small amount, it can use a single special opcode, which occupies
587         // a single byte. When it needs to advance the address by up to
588         // twice the range of the last special opcode, it can use
589         // DW_LNS_const_add_pc followed by a special opcode, for a total
590         // of two bytes. Only if it needs to advance the address by more
591         // than twice that range will it need to use both DW_LNS_advance_pc
592         // and a special opcode, requiring three or more bytes.
593         {
594           uint8_t AdjustOpcode = 255 - Prologue.OpcodeBase;
595           uint64_t AddrOffset =
596               (AdjustOpcode / Prologue.LineRange) * Prologue.MinInstLength;
597           State.Row.Address += AddrOffset;
598           if (OS)
599             *OS
600                 << format(" (0x%16.16" PRIx64 ")", AddrOffset);
601         }
602         break;
603 
604       case DW_LNS_fixed_advance_pc:
605         // Takes a single uhalf operand. Add to the address register of
606         // the state machine the value of the (unencoded) operand. This
607         // is the only extended opcode that takes an argument that is not
608         // a variable length number. The motivation for DW_LNS_fixed_advance_pc
609         // is this: existing assemblers cannot emit DW_LNS_advance_pc or
610         // special opcodes because they cannot encode LEB128 numbers or
611         // judge when the computation of a special opcode overflows and
612         // requires the use of DW_LNS_advance_pc. Such assemblers, however,
613         // can use DW_LNS_fixed_advance_pc instead, sacrificing compression.
614         {
615           uint16_t PCOffset = DebugLineData.getU16(OffsetPtr);
616           State.Row.Address += PCOffset;
617           if (OS)
618             *OS
619                 << format(" (0x%16.16" PRIx64 ")", PCOffset);
620         }
621         break;
622 
623       case DW_LNS_set_prologue_end:
624         // Takes no arguments. Set the prologue_end register of the
625         // state machine to true
626         State.Row.PrologueEnd = true;
627         break;
628 
629       case DW_LNS_set_epilogue_begin:
630         // Takes no arguments. Set the basic_block register of the
631         // state machine to true
632         State.Row.EpilogueBegin = true;
633         break;
634 
635       case DW_LNS_set_isa:
636         // Takes a single unsigned LEB128 operand and stores it in the
637         // column register of the state machine.
638         State.Row.Isa = DebugLineData.getULEB128(OffsetPtr);
639         if (OS)
640           *OS << " (" << State.Row.Isa << ")";
641         break;
642 
643       default:
644         // Handle any unknown standard opcodes here. We know the lengths
645         // of such opcodes because they are specified in the prologue
646         // as a multiple of LEB128 operands for each opcode.
647         {
648           assert(Opcode - 1U < Prologue.StandardOpcodeLengths.size());
649           uint8_t OpcodeLength = Prologue.StandardOpcodeLengths[Opcode - 1];
650           for (uint8_t I = 0; I < OpcodeLength; ++I) {
651             uint64_t Value = DebugLineData.getULEB128(OffsetPtr);
652             if (OS)
653               *OS << format("Skipping ULEB128 value: 0x%16.16" PRIx64 ")\n",
654                             Value);
655           }
656         }
657         break;
658       }
659     } else {
660       // Special Opcodes
661 
662       // A special opcode value is chosen based on the amount that needs
663       // to be added to the line and address registers. The maximum line
664       // increment for a special opcode is the value of the line_base
665       // field in the header, plus the value of the line_range field,
666       // minus 1 (line base + line range - 1). If the desired line
667       // increment is greater than the maximum line increment, a standard
668       // opcode must be used instead of a special opcode. The "address
669       // advance" is calculated by dividing the desired address increment
670       // by the minimum_instruction_length field from the header. The
671       // special opcode is then calculated using the following formula:
672       //
673       //  opcode = (desired line increment - line_base) +
674       //           (line_range * address advance) + opcode_base
675       //
676       // If the resulting opcode is greater than 255, a standard opcode
677       // must be used instead.
678       //
679       // To decode a special opcode, subtract the opcode_base from the
680       // opcode itself to give the adjusted opcode. The amount to
681       // increment the address register is the result of the adjusted
682       // opcode divided by the line_range multiplied by the
683       // minimum_instruction_length field from the header. That is:
684       //
685       //  address increment = (adjusted opcode / line_range) *
686       //                      minimum_instruction_length
687       //
688       // The amount to increment the line register is the line_base plus
689       // the result of the adjusted opcode modulo the line_range. That is:
690       //
691       // line increment = line_base + (adjusted opcode % line_range)
692 
693       uint8_t AdjustOpcode = Opcode - Prologue.OpcodeBase;
694       uint64_t AddrOffset =
695           (AdjustOpcode / Prologue.LineRange) * Prologue.MinInstLength;
696       int32_t LineOffset =
697           Prologue.LineBase + (AdjustOpcode % Prologue.LineRange);
698       State.Row.Line += LineOffset;
699       State.Row.Address += AddrOffset;
700 
701       if (OS) {
702         *OS << "address += " << ((uint32_t)AdjustOpcode)
703             << ",  line += " << LineOffset << "\n";
704         OS->indent(12);
705         State.Row.dump(*OS);
706       }
707 
708       State.appendRowToMatrix(*OffsetPtr);
709       // Reset discriminator to 0.
710       State.Row.Discriminator = 0;
711     }
712     if(OS)
713       *OS << "\n";
714   }
715 
716   if (!State.Sequence.Empty) {
717     fprintf(stderr, "warning: last sequence in debug line table is not"
718                     "terminated!\n");
719   }
720 
721   // Sort all sequences so that address lookup will work faster.
722   if (!Sequences.empty()) {
723     std::sort(Sequences.begin(), Sequences.end(), Sequence::orderByLowPC);
724     // Note: actually, instruction address ranges of sequences should not
725     // overlap (in shared objects and executables). If they do, the address
726     // lookup would still work, though, but result would be ambiguous.
727     // We don't report warning in this case. For example,
728     // sometimes .so compiled from multiple object files contains a few
729     // rudimentary sequences for address ranges [0x0, 0xsomething).
730   }
731 
732   return EndOffset;
733 }
734 
735 uint32_t
736 DWARFDebugLine::LineTable::findRowInSeq(const DWARFDebugLine::Sequence &Seq,
737                                         uint64_t Address) const {
738   if (!Seq.containsPC(Address))
739     return UnknownRowIndex;
740   // Search for instruction address in the rows describing the sequence.
741   // Rows are stored in a vector, so we may use arithmetical operations with
742   // iterators.
743   DWARFDebugLine::Row Row;
744   Row.Address = Address;
745   RowIter FirstRow = Rows.begin() + Seq.FirstRowIndex;
746   RowIter LastRow = Rows.begin() + Seq.LastRowIndex;
747   LineTable::RowIter RowPos = std::lower_bound(
748       FirstRow, LastRow, Row, DWARFDebugLine::Row::orderByAddress);
749   if (RowPos == LastRow) {
750     return Seq.LastRowIndex - 1;
751   }
752   uint32_t Index = Seq.FirstRowIndex + (RowPos - FirstRow);
753   if (RowPos->Address > Address) {
754     if (RowPos == FirstRow)
755       return UnknownRowIndex;
756     else
757       Index--;
758   }
759   return Index;
760 }
761 
762 uint32_t DWARFDebugLine::LineTable::lookupAddress(uint64_t Address) const {
763   if (Sequences.empty())
764     return UnknownRowIndex;
765   // First, find an instruction sequence containing the given address.
766   DWARFDebugLine::Sequence Sequence;
767   Sequence.LowPC = Address;
768   SequenceIter FirstSeq = Sequences.begin();
769   SequenceIter LastSeq = Sequences.end();
770   SequenceIter SeqPos = std::lower_bound(
771       FirstSeq, LastSeq, Sequence, DWARFDebugLine::Sequence::orderByLowPC);
772   DWARFDebugLine::Sequence FoundSeq;
773   if (SeqPos == LastSeq) {
774     FoundSeq = Sequences.back();
775   } else if (SeqPos->LowPC == Address) {
776     FoundSeq = *SeqPos;
777   } else {
778     if (SeqPos == FirstSeq)
779       return UnknownRowIndex;
780     FoundSeq = *(SeqPos - 1);
781   }
782   return findRowInSeq(FoundSeq, Address);
783 }
784 
785 bool DWARFDebugLine::LineTable::lookupAddressRange(
786     uint64_t Address, uint64_t Size, std::vector<uint32_t> &Result) const {
787   if (Sequences.empty())
788     return false;
789   uint64_t EndAddr = Address + Size;
790   // First, find an instruction sequence containing the given address.
791   DWARFDebugLine::Sequence Sequence;
792   Sequence.LowPC = Address;
793   SequenceIter FirstSeq = Sequences.begin();
794   SequenceIter LastSeq = Sequences.end();
795   SequenceIter SeqPos = std::lower_bound(
796       FirstSeq, LastSeq, Sequence, DWARFDebugLine::Sequence::orderByLowPC);
797   if (SeqPos == LastSeq || SeqPos->LowPC != Address) {
798     if (SeqPos == FirstSeq)
799       return false;
800     SeqPos--;
801   }
802   if (!SeqPos->containsPC(Address))
803     return false;
804 
805   SequenceIter StartPos = SeqPos;
806 
807   // Add the rows from the first sequence to the vector, starting with the
808   // index we just calculated
809 
810   while (SeqPos != LastSeq && SeqPos->LowPC < EndAddr) {
811     const DWARFDebugLine::Sequence &CurSeq = *SeqPos;
812     // For the first sequence, we need to find which row in the sequence is the
813     // first in our range.
814     uint32_t FirstRowIndex = CurSeq.FirstRowIndex;
815     if (SeqPos == StartPos)
816       FirstRowIndex = findRowInSeq(CurSeq, Address);
817 
818     // Figure out the last row in the range.
819     uint32_t LastRowIndex = findRowInSeq(CurSeq, EndAddr - 1);
820     if (LastRowIndex == UnknownRowIndex)
821       LastRowIndex = CurSeq.LastRowIndex - 1;
822 
823     assert(FirstRowIndex != UnknownRowIndex);
824     assert(LastRowIndex != UnknownRowIndex);
825 
826     for (uint32_t I = FirstRowIndex; I <= LastRowIndex; ++I) {
827       Result.push_back(I);
828     }
829 
830     ++SeqPos;
831   }
832 
833   return true;
834 }
835 
836 bool DWARFDebugLine::LineTable::hasFileAtIndex(uint64_t FileIndex) const {
837   return FileIndex != 0 && FileIndex <= Prologue.FileNames.size();
838 }
839 
840 bool DWARFDebugLine::LineTable::getFileNameByIndex(uint64_t FileIndex,
841                                                    const char *CompDir,
842                                                    FileLineInfoKind Kind,
843                                                    std::string &Result) const {
844   if (Kind == FileLineInfoKind::None || !hasFileAtIndex(FileIndex))
845     return false;
846   const FileNameEntry &Entry = Prologue.FileNames[FileIndex - 1];
847   StringRef FileName = Entry.Name;
848   if (Kind != FileLineInfoKind::AbsoluteFilePath ||
849       sys::path::is_absolute(FileName)) {
850     Result = FileName;
851     return true;
852   }
853 
854   SmallString<16> FilePath;
855   uint64_t IncludeDirIndex = Entry.DirIdx;
856   StringRef IncludeDir;
857   // Be defensive about the contents of Entry.
858   if (IncludeDirIndex > 0 &&
859       IncludeDirIndex <= Prologue.IncludeDirectories.size())
860     IncludeDir = Prologue.IncludeDirectories[IncludeDirIndex - 1];
861 
862   // We may still need to append compilation directory of compile unit.
863   // We know that FileName is not absolute, the only way to have an
864   // absolute path at this point would be if IncludeDir is absolute.
865   if (CompDir && Kind == FileLineInfoKind::AbsoluteFilePath &&
866       sys::path::is_relative(IncludeDir))
867     sys::path::append(FilePath, CompDir);
868 
869   // sys::path::append skips empty strings.
870   sys::path::append(FilePath, IncludeDir, FileName);
871   Result = FilePath.str();
872   return true;
873 }
874 
875 bool DWARFDebugLine::LineTable::getFileLineInfoForAddress(
876     uint64_t Address, const char *CompDir, FileLineInfoKind Kind,
877     DILineInfo &Result) const {
878   // Get the index of row we're looking for in the line table.
879   uint32_t RowIndex = lookupAddress(Address);
880   if (RowIndex == -1U)
881     return false;
882   // Take file number and line/column from the row.
883   const auto &Row = Rows[RowIndex];
884   if (!getFileNameByIndex(Row.File, CompDir, Kind, Result.FileName))
885     return false;
886   Result.Line = Row.Line;
887   Result.Column = Row.Column;
888   Result.Discriminator = Row.Discriminator;
889   return true;
890 }
891