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/DWARFFormValue.h"
16 #include "llvm/DebugInfo/DWARF/DWARFRelocMap.h"
17 #include "llvm/Support/Format.h"
18 #include "llvm/Support/Path.h"
19 #include "llvm/Support/raw_ostream.h"
20 #include <algorithm>
21 #include <cassert>
22 #include <cinttypes>
23 #include <cstdint>
24 #include <cstdio>
25 #include <utility>
26 
27 using namespace llvm;
28 using namespace dwarf;
29 
30 using FileLineInfoKind = DILineInfoSpecifier::FileLineInfoKind;
31 
32 namespace {
33 
34 struct ContentDescriptor {
35   dwarf::LineNumberEntryFormat Type;
36   dwarf::Form Form;
37 };
38 
39 using ContentDescriptors = SmallVector<ContentDescriptor, 4>;
40 
41 } // end anonmyous namespace
42 
43 DWARFDebugLine::Prologue::Prologue() { clear(); }
44 
45 void DWARFDebugLine::Prologue::clear() {
46   TotalLength = PrologueLength = 0;
47   SegSelectorSize = 0;
48   MinInstLength = MaxOpsPerInst = DefaultIsStmt = LineBase = LineRange = 0;
49   OpcodeBase = 0;
50   FormParams = DWARFFormParams({0, 0, DWARF32});
51   HasMD5 = false;
52   StandardOpcodeLengths.clear();
53   IncludeDirectories.clear();
54   FileNames.clear();
55 }
56 
57 void DWARFDebugLine::Prologue::dump(raw_ostream &OS,
58                                     DIDumpOptions DumpOptions) const {
59   OS << "Line table prologue:\n"
60      << format("    total_length: 0x%8.8" PRIx64 "\n", TotalLength)
61      << format("         version: %u\n", getVersion());
62   if (getVersion() >= 5)
63     OS << format("    address_size: %u\n", getAddressSize())
64        << format(" seg_select_size: %u\n", SegSelectorSize);
65   OS << format(" prologue_length: 0x%8.8" PRIx64 "\n", PrologueLength)
66      << format(" min_inst_length: %u\n", MinInstLength)
67      << format(getVersion() >= 4 ? "max_ops_per_inst: %u\n" : "", MaxOpsPerInst)
68      << format(" default_is_stmt: %u\n", DefaultIsStmt)
69      << format("       line_base: %i\n", LineBase)
70      << format("      line_range: %u\n", LineRange)
71      << format("     opcode_base: %u\n", OpcodeBase);
72 
73   for (uint32_t I = 0; I != StandardOpcodeLengths.size(); ++I)
74     OS << format("standard_opcode_lengths[%s] = %u\n",
75                  LNStandardString(I + 1).data(), StandardOpcodeLengths[I]);
76 
77   if (!IncludeDirectories.empty()) {
78     // DWARF v5 starts directory indexes at 0.
79     uint32_t DirBase = getVersion() >= 5 ? 0 : 1;
80     for (uint32_t I = 0; I != IncludeDirectories.size(); ++I) {
81       OS << format("include_directories[%3u] = ", I + DirBase);
82       IncludeDirectories[I].dump(OS, DumpOptions);
83       OS << '\n';
84     }
85   }
86 
87   if (!FileNames.empty()) {
88     if (HasMD5)
89       OS << "                Dir  MD5 Checksum                     File Name\n"
90          << "                ---- -------------------------------- -----------"
91             "---------------\n";
92     else
93       OS << "                Dir  Mod Time   File Len   File Name\n"
94          << "                ---- ---------- ---------- -----------"
95             "----------------\n";
96     // DWARF v5 starts file indexes at 0.
97     uint32_t FileBase = getVersion() >= 5 ? 0 : 1;
98     for (uint32_t I = 0; I != FileNames.size(); ++I) {
99       const FileNameEntry &FileEntry = FileNames[I];
100       OS << format("file_names[%3u] %4" PRIu64 " ", I + FileBase,
101                    FileEntry.DirIdx);
102       if (HasMD5)
103         OS << FileEntry.Checksum.digest();
104       else
105         OS << format("0x%8.8" PRIx64 " 0x%8.8" PRIx64, FileEntry.ModTime,
106                      FileEntry.Length);
107       OS << ' ';
108       FileEntry.Name.dump(OS, DumpOptions);
109       OS << '\n';
110     }
111   }
112 }
113 
114 // Parse v2-v4 directory and file tables.
115 static void
116 parseV2DirFileTables(const DWARFDataExtractor &DebugLineData,
117                      uint32_t *OffsetPtr, uint64_t EndPrologueOffset,
118                      std::vector<DWARFFormValue> &IncludeDirectories,
119                      std::vector<DWARFDebugLine::FileNameEntry> &FileNames) {
120   while (*OffsetPtr < EndPrologueOffset) {
121     StringRef S = DebugLineData.getCStrRef(OffsetPtr);
122     if (S.empty())
123       break;
124     DWARFFormValue Dir(dwarf::DW_FORM_string);
125     Dir.setPValue(S.data());
126     IncludeDirectories.push_back(Dir);
127   }
128 
129   while (*OffsetPtr < EndPrologueOffset) {
130     StringRef Name = DebugLineData.getCStrRef(OffsetPtr);
131     if (Name.empty())
132       break;
133     DWARFDebugLine::FileNameEntry FileEntry;
134     FileEntry.Name.setForm(dwarf::DW_FORM_string);
135     FileEntry.Name.setPValue(Name.data());
136     FileEntry.DirIdx = DebugLineData.getULEB128(OffsetPtr);
137     FileEntry.ModTime = DebugLineData.getULEB128(OffsetPtr);
138     FileEntry.Length = DebugLineData.getULEB128(OffsetPtr);
139     FileNames.push_back(FileEntry);
140   }
141 }
142 
143 // Parse v5 directory/file entry content descriptions.
144 // Returns the descriptors, or an empty vector if we did not find a path or
145 // ran off the end of the prologue.
146 static ContentDescriptors
147 parseV5EntryFormat(const DWARFDataExtractor &DebugLineData, uint32_t *OffsetPtr,
148                    uint64_t EndPrologueOffset, bool *HasMD5) {
149   ContentDescriptors Descriptors;
150   int FormatCount = DebugLineData.getU8(OffsetPtr);
151   bool HasPath = false;
152   for (int I = 0; I != FormatCount; ++I) {
153     if (*OffsetPtr >= EndPrologueOffset)
154       return ContentDescriptors();
155     ContentDescriptor Descriptor;
156     Descriptor.Type =
157       dwarf::LineNumberEntryFormat(DebugLineData.getULEB128(OffsetPtr));
158     Descriptor.Form = dwarf::Form(DebugLineData.getULEB128(OffsetPtr));
159     if (Descriptor.Type == dwarf::DW_LNCT_path)
160       HasPath = true;
161     else if (Descriptor.Type == dwarf::DW_LNCT_MD5 && HasMD5)
162       *HasMD5 = true;
163     Descriptors.push_back(Descriptor);
164   }
165   return HasPath ? Descriptors : ContentDescriptors();
166 }
167 
168 static bool
169 parseV5DirFileTables(const DWARFDataExtractor &DebugLineData,
170                      uint32_t *OffsetPtr, uint64_t EndPrologueOffset,
171                      const DWARFFormParams &FormParams, const DWARFContext &Ctx,
172                      const DWARFUnit *U, bool &HasMD5,
173                      std::vector<DWARFFormValue> &IncludeDirectories,
174                      std::vector<DWARFDebugLine::FileNameEntry> &FileNames) {
175   // Get the directory entry description.
176   ContentDescriptors DirDescriptors =
177       parseV5EntryFormat(DebugLineData, OffsetPtr, EndPrologueOffset, nullptr);
178   if (DirDescriptors.empty())
179     return false;
180 
181   // Get the directory entries, according to the format described above.
182   int DirEntryCount = DebugLineData.getU8(OffsetPtr);
183   for (int I = 0; I != DirEntryCount; ++I) {
184     if (*OffsetPtr >= EndPrologueOffset)
185       return false;
186     for (auto Descriptor : DirDescriptors) {
187       DWARFFormValue Value(Descriptor.Form);
188       switch (Descriptor.Type) {
189       case DW_LNCT_path:
190         if (!Value.extractValue(DebugLineData, OffsetPtr, FormParams, &Ctx, U))
191           return false;
192         IncludeDirectories.push_back(Value);
193         break;
194       default:
195         if (!Value.skipValue(DebugLineData, OffsetPtr, FormParams))
196           return false;
197       }
198     }
199   }
200 
201   // Get the file entry description.
202   ContentDescriptors FileDescriptors =
203       parseV5EntryFormat(DebugLineData, OffsetPtr, EndPrologueOffset, &HasMD5);
204   if (FileDescriptors.empty())
205     return false;
206 
207   // Get the file entries, according to the format described above.
208   int FileEntryCount = DebugLineData.getU8(OffsetPtr);
209   for (int I = 0; I != FileEntryCount; ++I) {
210     if (*OffsetPtr >= EndPrologueOffset)
211       return false;
212     DWARFDebugLine::FileNameEntry FileEntry;
213     for (auto Descriptor : FileDescriptors) {
214       DWARFFormValue Value(Descriptor.Form);
215       if (!Value.extractValue(DebugLineData, OffsetPtr, FormParams, &Ctx, U))
216         return false;
217       switch (Descriptor.Type) {
218       case DW_LNCT_path:
219         FileEntry.Name = Value;
220         break;
221       case DW_LNCT_directory_index:
222         FileEntry.DirIdx = Value.getAsUnsignedConstant().getValue();
223         break;
224       case DW_LNCT_timestamp:
225         FileEntry.ModTime = Value.getAsUnsignedConstant().getValue();
226         break;
227       case DW_LNCT_size:
228         FileEntry.Length = Value.getAsUnsignedConstant().getValue();
229         break;
230       case DW_LNCT_MD5:
231         assert(Value.getAsBlock().getValue().size() == 16);
232         std::uninitialized_copy_n(Value.getAsBlock().getValue().begin(), 16,
233                                   FileEntry.Checksum.Bytes.begin());
234         break;
235       default:
236         break;
237       }
238     }
239     FileNames.push_back(FileEntry);
240   }
241   return true;
242 }
243 
244 bool DWARFDebugLine::Prologue::parse(const DWARFDataExtractor &DebugLineData,
245                                      uint32_t *OffsetPtr,
246                                      const DWARFContext &Ctx,
247                                      const DWARFUnit *U) {
248   const uint64_t PrologueOffset = *OffsetPtr;
249 
250   clear();
251   TotalLength = DebugLineData.getU32(OffsetPtr);
252   if (TotalLength == UINT32_MAX) {
253     FormParams.Format = dwarf::DWARF64;
254     TotalLength = DebugLineData.getU64(OffsetPtr);
255   } else if (TotalLength >= 0xffffff00) {
256     return false;
257   }
258   FormParams.Version = DebugLineData.getU16(OffsetPtr);
259   if (getVersion() < 2)
260     return false;
261 
262   if (getVersion() >= 5) {
263     FormParams.AddrSize = DebugLineData.getU8(OffsetPtr);
264     assert((DebugLineData.getAddressSize() == 0 ||
265             DebugLineData.getAddressSize() == getAddressSize()) &&
266            "Line table header and data extractor disagree");
267     SegSelectorSize = DebugLineData.getU8(OffsetPtr);
268   }
269 
270   PrologueLength = DebugLineData.getUnsigned(OffsetPtr, sizeofPrologueLength());
271   const uint64_t EndPrologueOffset = PrologueLength + *OffsetPtr;
272   MinInstLength = DebugLineData.getU8(OffsetPtr);
273   if (getVersion() >= 4)
274     MaxOpsPerInst = DebugLineData.getU8(OffsetPtr);
275   DefaultIsStmt = DebugLineData.getU8(OffsetPtr);
276   LineBase = DebugLineData.getU8(OffsetPtr);
277   LineRange = DebugLineData.getU8(OffsetPtr);
278   OpcodeBase = DebugLineData.getU8(OffsetPtr);
279 
280   StandardOpcodeLengths.reserve(OpcodeBase - 1);
281   for (uint32_t I = 1; I < OpcodeBase; ++I) {
282     uint8_t OpLen = DebugLineData.getU8(OffsetPtr);
283     StandardOpcodeLengths.push_back(OpLen);
284   }
285 
286   if (getVersion() >= 5) {
287     if (!parseV5DirFileTables(DebugLineData, OffsetPtr, EndPrologueOffset,
288                               FormParams, Ctx, U, HasMD5, IncludeDirectories,
289                               FileNames)) {
290       fprintf(stderr,
291               "warning: parsing line table prologue at 0x%8.8" PRIx64
292               " found an invalid directory or file table description at"
293               " 0x%8.8" PRIx64 "\n", PrologueOffset, (uint64_t)*OffsetPtr);
294       return false;
295     }
296   } else
297     parseV2DirFileTables(DebugLineData, OffsetPtr, EndPrologueOffset,
298                          IncludeDirectories, FileNames);
299 
300   if (*OffsetPtr != EndPrologueOffset) {
301     fprintf(stderr,
302             "warning: parsing line table prologue at 0x%8.8" PRIx64
303             " should have ended at 0x%8.8" PRIx64
304             " but it ended at 0x%8.8" PRIx64 "\n",
305             PrologueOffset, EndPrologueOffset, (uint64_t)*OffsetPtr);
306     return false;
307   }
308   return true;
309 }
310 
311 DWARFDebugLine::Row::Row(bool DefaultIsStmt) { reset(DefaultIsStmt); }
312 
313 void DWARFDebugLine::Row::postAppend() {
314   BasicBlock = false;
315   PrologueEnd = false;
316   EpilogueBegin = false;
317 }
318 
319 void DWARFDebugLine::Row::reset(bool DefaultIsStmt) {
320   Address = 0;
321   Line = 1;
322   Column = 0;
323   File = 1;
324   Isa = 0;
325   Discriminator = 0;
326   IsStmt = DefaultIsStmt;
327   BasicBlock = false;
328   EndSequence = false;
329   PrologueEnd = false;
330   EpilogueBegin = false;
331 }
332 
333 void DWARFDebugLine::Row::dumpTableHeader(raw_ostream &OS) {
334   OS << "Address            Line   Column File   ISA Discriminator Flags\n"
335      << "------------------ ------ ------ ------ --- ------------- "
336         "-------------\n";
337 }
338 
339 void DWARFDebugLine::Row::dump(raw_ostream &OS) const {
340   OS << format("0x%16.16" PRIx64 " %6u %6u", Address, Line, Column)
341      << format(" %6u %3u %13u ", File, Isa, Discriminator)
342      << (IsStmt ? " is_stmt" : "") << (BasicBlock ? " basic_block" : "")
343      << (PrologueEnd ? " prologue_end" : "")
344      << (EpilogueBegin ? " epilogue_begin" : "")
345      << (EndSequence ? " end_sequence" : "") << '\n';
346 }
347 
348 DWARFDebugLine::Sequence::Sequence() { reset(); }
349 
350 void DWARFDebugLine::Sequence::reset() {
351   LowPC = 0;
352   HighPC = 0;
353   FirstRowIndex = 0;
354   LastRowIndex = 0;
355   Empty = true;
356 }
357 
358 DWARFDebugLine::LineTable::LineTable() { clear(); }
359 
360 void DWARFDebugLine::LineTable::dump(raw_ostream &OS,
361                                      DIDumpOptions DumpOptions) const {
362   Prologue.dump(OS, DumpOptions);
363   OS << '\n';
364 
365   if (!Rows.empty()) {
366     Row::dumpTableHeader(OS);
367     for (const Row &R : Rows) {
368       R.dump(OS);
369     }
370   }
371 }
372 
373 void DWARFDebugLine::LineTable::clear() {
374   Prologue.clear();
375   Rows.clear();
376   Sequences.clear();
377 }
378 
379 DWARFDebugLine::ParsingState::ParsingState(struct LineTable *LT)
380     : LineTable(LT) {
381   resetRowAndSequence();
382 }
383 
384 void DWARFDebugLine::ParsingState::resetRowAndSequence() {
385   Row.reset(LineTable->Prologue.DefaultIsStmt);
386   Sequence.reset();
387 }
388 
389 void DWARFDebugLine::ParsingState::appendRowToMatrix(uint32_t Offset) {
390   if (Sequence.Empty) {
391     // Record the beginning of instruction sequence.
392     Sequence.Empty = false;
393     Sequence.LowPC = Row.Address;
394     Sequence.FirstRowIndex = RowNumber;
395   }
396   ++RowNumber;
397   LineTable->appendRow(Row);
398   if (Row.EndSequence) {
399     // Record the end of instruction sequence.
400     Sequence.HighPC = Row.Address;
401     Sequence.LastRowIndex = RowNumber;
402     if (Sequence.isValid())
403       LineTable->appendSequence(Sequence);
404     Sequence.reset();
405   }
406   Row.postAppend();
407 }
408 
409 const DWARFDebugLine::LineTable *
410 DWARFDebugLine::getLineTable(uint32_t Offset) const {
411   LineTableConstIter Pos = LineTableMap.find(Offset);
412   if (Pos != LineTableMap.end())
413     return &Pos->second;
414   return nullptr;
415 }
416 
417 const DWARFDebugLine::LineTable *
418 DWARFDebugLine::getOrParseLineTable(DWARFDataExtractor &DebugLineData,
419                                     uint32_t Offset, const DWARFContext &Ctx,
420                                     const DWARFUnit *U) {
421   std::pair<LineTableIter, bool> Pos =
422       LineTableMap.insert(LineTableMapTy::value_type(Offset, LineTable()));
423   LineTable *LT = &Pos.first->second;
424   if (Pos.second) {
425     if (!LT->parse(DebugLineData, &Offset, Ctx, U))
426       return nullptr;
427   }
428   return LT;
429 }
430 
431 bool DWARFDebugLine::LineTable::parse(DWARFDataExtractor &DebugLineData,
432                                       uint32_t *OffsetPtr,
433                                       const DWARFContext &Ctx,
434                                       const DWARFUnit *U, raw_ostream *OS) {
435   const uint32_t DebugLineOffset = *OffsetPtr;
436 
437   clear();
438 
439   if (!Prologue.parse(DebugLineData, OffsetPtr, Ctx, U)) {
440     // Restore our offset and return false to indicate failure!
441     *OffsetPtr = DebugLineOffset;
442     return false;
443   }
444 
445   if (OS) {
446     // The presence of OS signals verbose dumping.
447     DIDumpOptions DumpOptions;
448     DumpOptions.Verbose = true;
449     Prologue.dump(*OS, DumpOptions);
450   }
451 
452   const uint32_t EndOffset =
453       DebugLineOffset + Prologue.TotalLength + Prologue.sizeofTotalLength();
454 
455   // See if we should tell the data extractor the address size.
456   if (DebugLineData.getAddressSize() == 0)
457     DebugLineData.setAddressSize(Prologue.getAddressSize());
458   else
459     assert(Prologue.getAddressSize() == 0 ||
460            Prologue.getAddressSize() == DebugLineData.getAddressSize());
461 
462   ParsingState State(this);
463 
464   while (*OffsetPtr < EndOffset) {
465     if (OS)
466       *OS << format("0x%08.08" PRIx32 ": ", *OffsetPtr);
467 
468     uint8_t Opcode = DebugLineData.getU8(OffsetPtr);
469 
470     if (OS)
471       *OS << format("%02.02" PRIx8 " ", Opcode);
472 
473     if (Opcode == 0) {
474       // Extended Opcodes always start with a zero opcode followed by
475       // a uleb128 length so you can skip ones you don't know about
476       uint64_t Len = DebugLineData.getULEB128(OffsetPtr);
477       uint32_t ExtOffset = *OffsetPtr;
478 
479       // Tolerate zero-length; assume length is correct and soldier on.
480       if (Len == 0) {
481         if (OS)
482           *OS << "Badly formed extended line op (length 0)\n";
483         continue;
484       }
485 
486       uint8_t SubOpcode = DebugLineData.getU8(OffsetPtr);
487       if (OS)
488         *OS << LNExtendedString(SubOpcode);
489       switch (SubOpcode) {
490       case DW_LNE_end_sequence:
491         // Set the end_sequence register of the state machine to true and
492         // append a row to the matrix using the current values of the
493         // state-machine registers. Then reset the registers to the initial
494         // values specified above. Every statement program sequence must end
495         // with a DW_LNE_end_sequence instruction which creates a row whose
496         // address is that of the byte after the last target machine instruction
497         // of the sequence.
498         State.Row.EndSequence = true;
499         State.appendRowToMatrix(*OffsetPtr);
500         if (OS) {
501           *OS << "\n";
502           OS->indent(12);
503           State.Row.dump(*OS);
504         }
505         State.resetRowAndSequence();
506         break;
507 
508       case DW_LNE_set_address:
509         // Takes a single relocatable address as an operand. The size of the
510         // operand is the size appropriate to hold an address on the target
511         // machine. Set the address register to the value given by the
512         // relocatable address. All of the other statement program opcodes
513         // that affect the address register add a delta to it. This instruction
514         // stores a relocatable value into it instead.
515         //
516         // Make sure the extractor knows the address size.  If not, infer it
517         // from the size of the operand.
518         if (DebugLineData.getAddressSize() == 0)
519           DebugLineData.setAddressSize(Len - 1);
520         else
521           assert(DebugLineData.getAddressSize() == Len - 1);
522         State.Row.Address = DebugLineData.getRelocatedAddress(OffsetPtr);
523         if (OS)
524           *OS << format(" (0x%16.16" PRIx64 ")", State.Row.Address);
525         break;
526 
527       case DW_LNE_define_file:
528         // Takes 4 arguments. The first is a null terminated string containing
529         // a source file name. The second is an unsigned LEB128 number
530         // representing the directory index of the directory in which the file
531         // was found. The third is an unsigned LEB128 number representing the
532         // time of last modification of the file. The fourth is an unsigned
533         // LEB128 number representing the length in bytes of the file. The time
534         // and length fields may contain LEB128(0) if the information is not
535         // available.
536         //
537         // The directory index represents an entry in the include_directories
538         // section of the statement program prologue. The index is LEB128(0)
539         // if the file was found in the current directory of the compilation,
540         // LEB128(1) if it was found in the first directory in the
541         // include_directories section, and so on. The directory index is
542         // ignored for file names that represent full path names.
543         //
544         // The files are numbered, starting at 1, in the order in which they
545         // appear; the names in the prologue come before names defined by
546         // the DW_LNE_define_file instruction. These numbers are used in the
547         // the file register of the state machine.
548         {
549           FileNameEntry FileEntry;
550           const char *Name = DebugLineData.getCStr(OffsetPtr);
551           FileEntry.Name.setForm(dwarf::DW_FORM_string);
552           FileEntry.Name.setPValue(Name);
553           FileEntry.DirIdx = DebugLineData.getULEB128(OffsetPtr);
554           FileEntry.ModTime = DebugLineData.getULEB128(OffsetPtr);
555           FileEntry.Length = DebugLineData.getULEB128(OffsetPtr);
556           Prologue.FileNames.push_back(FileEntry);
557           if (OS)
558             *OS << " (" << Name << ", dir=" << FileEntry.DirIdx << ", mod_time="
559                 << format("(0x%16.16" PRIx64 ")", FileEntry.ModTime)
560                 << ", length=" << FileEntry.Length << ")";
561         }
562         break;
563 
564       case DW_LNE_set_discriminator:
565         State.Row.Discriminator = DebugLineData.getULEB128(OffsetPtr);
566         if (OS)
567           *OS << " (" << State.Row.Discriminator << ")";
568         break;
569 
570       default:
571         if (OS)
572           *OS << format("Unrecognized extended op 0x%02.02" PRIx8, SubOpcode)
573               << format(" length %" PRIx64, Len);
574         // Len doesn't include the zero opcode byte or the length itself, but
575         // it does include the sub_opcode, so we have to adjust for that.
576         (*OffsetPtr) += Len - 1;
577         break;
578       }
579       // Make sure the stated and parsed lengths are the same.
580       // Otherwise we have an unparseable line-number program.
581       if (*OffsetPtr - ExtOffset != Len) {
582         fprintf(stderr, "Unexpected line op length at offset 0x%8.8" PRIx32
583                 " expected 0x%2.2" PRIx64 " found 0x%2.2" PRIx32 "\n",
584                 ExtOffset, Len, *OffsetPtr - ExtOffset);
585         // Skip the rest of the line-number program.
586         *OffsetPtr = EndOffset;
587         return false;
588       }
589     } else if (Opcode < Prologue.OpcodeBase) {
590       if (OS)
591         *OS << LNStandardString(Opcode);
592       switch (Opcode) {
593       // Standard Opcodes
594       case DW_LNS_copy:
595         // Takes no arguments. Append a row to the matrix using the
596         // current values of the state-machine registers. Then set
597         // the basic_block register to false.
598         State.appendRowToMatrix(*OffsetPtr);
599         if (OS) {
600           *OS << "\n";
601           OS->indent(12);
602           State.Row.dump(*OS);
603           *OS << "\n";
604         }
605         break;
606 
607       case DW_LNS_advance_pc:
608         // Takes a single unsigned LEB128 operand, multiplies it by the
609         // min_inst_length field of the prologue, and adds the
610         // result to the address register of the state machine.
611         {
612           uint64_t AddrOffset =
613               DebugLineData.getULEB128(OffsetPtr) * Prologue.MinInstLength;
614           State.Row.Address += AddrOffset;
615           if (OS)
616             *OS << " (" << AddrOffset << ")";
617         }
618         break;
619 
620       case DW_LNS_advance_line:
621         // Takes a single signed LEB128 operand and adds that value to
622         // the line register of the state machine.
623         State.Row.Line += DebugLineData.getSLEB128(OffsetPtr);
624         if (OS)
625           *OS << " (" << State.Row.Line << ")";
626         break;
627 
628       case DW_LNS_set_file:
629         // Takes a single unsigned LEB128 operand and stores it in the file
630         // register of the state machine.
631         State.Row.File = DebugLineData.getULEB128(OffsetPtr);
632         if (OS)
633           *OS << " (" << State.Row.File << ")";
634         break;
635 
636       case DW_LNS_set_column:
637         // Takes a single unsigned LEB128 operand and stores it in the
638         // column register of the state machine.
639         State.Row.Column = DebugLineData.getULEB128(OffsetPtr);
640         if (OS)
641           *OS << " (" << State.Row.Column << ")";
642         break;
643 
644       case DW_LNS_negate_stmt:
645         // Takes no arguments. Set the is_stmt register of the state
646         // machine to the logical negation of its current value.
647         State.Row.IsStmt = !State.Row.IsStmt;
648         break;
649 
650       case DW_LNS_set_basic_block:
651         // Takes no arguments. Set the basic_block register of the
652         // state machine to true
653         State.Row.BasicBlock = true;
654         break;
655 
656       case DW_LNS_const_add_pc:
657         // Takes no arguments. Add to the address register of the state
658         // machine the address increment value corresponding to special
659         // opcode 255. The motivation for DW_LNS_const_add_pc is this:
660         // when the statement program needs to advance the address by a
661         // small amount, it can use a single special opcode, which occupies
662         // a single byte. When it needs to advance the address by up to
663         // twice the range of the last special opcode, it can use
664         // DW_LNS_const_add_pc followed by a special opcode, for a total
665         // of two bytes. Only if it needs to advance the address by more
666         // than twice that range will it need to use both DW_LNS_advance_pc
667         // and a special opcode, requiring three or more bytes.
668         {
669           uint8_t AdjustOpcode = 255 - Prologue.OpcodeBase;
670           uint64_t AddrOffset =
671               (AdjustOpcode / Prologue.LineRange) * Prologue.MinInstLength;
672           State.Row.Address += AddrOffset;
673           if (OS)
674             *OS
675                 << format(" (0x%16.16" PRIx64 ")", AddrOffset);
676         }
677         break;
678 
679       case DW_LNS_fixed_advance_pc:
680         // Takes a single uhalf operand. Add to the address register of
681         // the state machine the value of the (unencoded) operand. This
682         // is the only extended opcode that takes an argument that is not
683         // a variable length number. The motivation for DW_LNS_fixed_advance_pc
684         // is this: existing assemblers cannot emit DW_LNS_advance_pc or
685         // special opcodes because they cannot encode LEB128 numbers or
686         // judge when the computation of a special opcode overflows and
687         // requires the use of DW_LNS_advance_pc. Such assemblers, however,
688         // can use DW_LNS_fixed_advance_pc instead, sacrificing compression.
689         {
690           uint16_t PCOffset = DebugLineData.getU16(OffsetPtr);
691           State.Row.Address += PCOffset;
692           if (OS)
693             *OS
694                 << format(" (0x%16.16" PRIx64 ")", PCOffset);
695         }
696         break;
697 
698       case DW_LNS_set_prologue_end:
699         // Takes no arguments. Set the prologue_end register of the
700         // state machine to true
701         State.Row.PrologueEnd = true;
702         break;
703 
704       case DW_LNS_set_epilogue_begin:
705         // Takes no arguments. Set the basic_block register of the
706         // state machine to true
707         State.Row.EpilogueBegin = true;
708         break;
709 
710       case DW_LNS_set_isa:
711         // Takes a single unsigned LEB128 operand and stores it in the
712         // column register of the state machine.
713         State.Row.Isa = DebugLineData.getULEB128(OffsetPtr);
714         if (OS)
715           *OS << " (" << State.Row.Isa << ")";
716         break;
717 
718       default:
719         // Handle any unknown standard opcodes here. We know the lengths
720         // of such opcodes because they are specified in the prologue
721         // as a multiple of LEB128 operands for each opcode.
722         {
723           assert(Opcode - 1U < Prologue.StandardOpcodeLengths.size());
724           uint8_t OpcodeLength = Prologue.StandardOpcodeLengths[Opcode - 1];
725           for (uint8_t I = 0; I < OpcodeLength; ++I) {
726             uint64_t Value = DebugLineData.getULEB128(OffsetPtr);
727             if (OS)
728               *OS << format("Skipping ULEB128 value: 0x%16.16" PRIx64 ")\n",
729                             Value);
730           }
731         }
732         break;
733       }
734     } else {
735       // Special Opcodes
736 
737       // A special opcode value is chosen based on the amount that needs
738       // to be added to the line and address registers. The maximum line
739       // increment for a special opcode is the value of the line_base
740       // field in the header, plus the value of the line_range field,
741       // minus 1 (line base + line range - 1). If the desired line
742       // increment is greater than the maximum line increment, a standard
743       // opcode must be used instead of a special opcode. The "address
744       // advance" is calculated by dividing the desired address increment
745       // by the minimum_instruction_length field from the header. The
746       // special opcode is then calculated using the following formula:
747       //
748       //  opcode = (desired line increment - line_base) +
749       //           (line_range * address advance) + opcode_base
750       //
751       // If the resulting opcode is greater than 255, a standard opcode
752       // must be used instead.
753       //
754       // To decode a special opcode, subtract the opcode_base from the
755       // opcode itself to give the adjusted opcode. The amount to
756       // increment the address register is the result of the adjusted
757       // opcode divided by the line_range multiplied by the
758       // minimum_instruction_length field from the header. That is:
759       //
760       //  address increment = (adjusted opcode / line_range) *
761       //                      minimum_instruction_length
762       //
763       // The amount to increment the line register is the line_base plus
764       // the result of the adjusted opcode modulo the line_range. That is:
765       //
766       // line increment = line_base + (adjusted opcode % line_range)
767 
768       uint8_t AdjustOpcode = Opcode - Prologue.OpcodeBase;
769       uint64_t AddrOffset =
770           (AdjustOpcode / Prologue.LineRange) * Prologue.MinInstLength;
771       int32_t LineOffset =
772           Prologue.LineBase + (AdjustOpcode % Prologue.LineRange);
773       State.Row.Line += LineOffset;
774       State.Row.Address += AddrOffset;
775 
776       if (OS) {
777         *OS << "address += " << ((uint32_t)AdjustOpcode)
778             << ",  line += " << LineOffset << "\n";
779         OS->indent(12);
780         State.Row.dump(*OS);
781       }
782 
783       State.appendRowToMatrix(*OffsetPtr);
784       // Reset discriminator to 0.
785       State.Row.Discriminator = 0;
786     }
787     if(OS)
788       *OS << "\n";
789   }
790 
791   if (!State.Sequence.Empty) {
792     fprintf(stderr, "warning: last sequence in debug line table is not"
793                     "terminated!\n");
794   }
795 
796   // Sort all sequences so that address lookup will work faster.
797   if (!Sequences.empty()) {
798     std::sort(Sequences.begin(), Sequences.end(), Sequence::orderByLowPC);
799     // Note: actually, instruction address ranges of sequences should not
800     // overlap (in shared objects and executables). If they do, the address
801     // lookup would still work, though, but result would be ambiguous.
802     // We don't report warning in this case. For example,
803     // sometimes .so compiled from multiple object files contains a few
804     // rudimentary sequences for address ranges [0x0, 0xsomething).
805   }
806 
807   return EndOffset;
808 }
809 
810 uint32_t
811 DWARFDebugLine::LineTable::findRowInSeq(const DWARFDebugLine::Sequence &Seq,
812                                         uint64_t Address) const {
813   if (!Seq.containsPC(Address))
814     return UnknownRowIndex;
815   // Search for instruction address in the rows describing the sequence.
816   // Rows are stored in a vector, so we may use arithmetical operations with
817   // iterators.
818   DWARFDebugLine::Row Row;
819   Row.Address = Address;
820   RowIter FirstRow = Rows.begin() + Seq.FirstRowIndex;
821   RowIter LastRow = Rows.begin() + Seq.LastRowIndex;
822   LineTable::RowIter RowPos = std::lower_bound(
823       FirstRow, LastRow, Row, DWARFDebugLine::Row::orderByAddress);
824   if (RowPos == LastRow) {
825     return Seq.LastRowIndex - 1;
826   }
827   uint32_t Index = Seq.FirstRowIndex + (RowPos - FirstRow);
828   if (RowPos->Address > Address) {
829     if (RowPos == FirstRow)
830       return UnknownRowIndex;
831     else
832       Index--;
833   }
834   return Index;
835 }
836 
837 uint32_t DWARFDebugLine::LineTable::lookupAddress(uint64_t Address) const {
838   if (Sequences.empty())
839     return UnknownRowIndex;
840   // First, find an instruction sequence containing the given address.
841   DWARFDebugLine::Sequence Sequence;
842   Sequence.LowPC = Address;
843   SequenceIter FirstSeq = Sequences.begin();
844   SequenceIter LastSeq = Sequences.end();
845   SequenceIter SeqPos = std::lower_bound(
846       FirstSeq, LastSeq, Sequence, DWARFDebugLine::Sequence::orderByLowPC);
847   DWARFDebugLine::Sequence FoundSeq;
848   if (SeqPos == LastSeq) {
849     FoundSeq = Sequences.back();
850   } else if (SeqPos->LowPC == Address) {
851     FoundSeq = *SeqPos;
852   } else {
853     if (SeqPos == FirstSeq)
854       return UnknownRowIndex;
855     FoundSeq = *(SeqPos - 1);
856   }
857   return findRowInSeq(FoundSeq, Address);
858 }
859 
860 bool DWARFDebugLine::LineTable::lookupAddressRange(
861     uint64_t Address, uint64_t Size, std::vector<uint32_t> &Result) const {
862   if (Sequences.empty())
863     return false;
864   uint64_t EndAddr = Address + Size;
865   // First, find an instruction sequence containing the given address.
866   DWARFDebugLine::Sequence Sequence;
867   Sequence.LowPC = Address;
868   SequenceIter FirstSeq = Sequences.begin();
869   SequenceIter LastSeq = Sequences.end();
870   SequenceIter SeqPos = std::lower_bound(
871       FirstSeq, LastSeq, Sequence, DWARFDebugLine::Sequence::orderByLowPC);
872   if (SeqPos == LastSeq || SeqPos->LowPC != Address) {
873     if (SeqPos == FirstSeq)
874       return false;
875     SeqPos--;
876   }
877   if (!SeqPos->containsPC(Address))
878     return false;
879 
880   SequenceIter StartPos = SeqPos;
881 
882   // Add the rows from the first sequence to the vector, starting with the
883   // index we just calculated
884 
885   while (SeqPos != LastSeq && SeqPos->LowPC < EndAddr) {
886     const DWARFDebugLine::Sequence &CurSeq = *SeqPos;
887     // For the first sequence, we need to find which row in the sequence is the
888     // first in our range.
889     uint32_t FirstRowIndex = CurSeq.FirstRowIndex;
890     if (SeqPos == StartPos)
891       FirstRowIndex = findRowInSeq(CurSeq, Address);
892 
893     // Figure out the last row in the range.
894     uint32_t LastRowIndex = findRowInSeq(CurSeq, EndAddr - 1);
895     if (LastRowIndex == UnknownRowIndex)
896       LastRowIndex = CurSeq.LastRowIndex - 1;
897 
898     assert(FirstRowIndex != UnknownRowIndex);
899     assert(LastRowIndex != UnknownRowIndex);
900 
901     for (uint32_t I = FirstRowIndex; I <= LastRowIndex; ++I) {
902       Result.push_back(I);
903     }
904 
905     ++SeqPos;
906   }
907 
908   return true;
909 }
910 
911 bool DWARFDebugLine::LineTable::hasFileAtIndex(uint64_t FileIndex) const {
912   return FileIndex != 0 && FileIndex <= Prologue.FileNames.size();
913 }
914 
915 bool DWARFDebugLine::LineTable::getFileNameByIndex(uint64_t FileIndex,
916                                                    const char *CompDir,
917                                                    FileLineInfoKind Kind,
918                                                    std::string &Result) const {
919   if (Kind == FileLineInfoKind::None || !hasFileAtIndex(FileIndex))
920     return false;
921   const FileNameEntry &Entry = Prologue.FileNames[FileIndex - 1];
922   StringRef FileName = Entry.Name.getAsCString().getValue();
923   if (Kind != FileLineInfoKind::AbsoluteFilePath ||
924       sys::path::is_absolute(FileName)) {
925     Result = FileName;
926     return true;
927   }
928 
929   SmallString<16> FilePath;
930   uint64_t IncludeDirIndex = Entry.DirIdx;
931   StringRef IncludeDir;
932   // Be defensive about the contents of Entry.
933   if (IncludeDirIndex > 0 &&
934       IncludeDirIndex <= Prologue.IncludeDirectories.size())
935     IncludeDir = Prologue.IncludeDirectories[IncludeDirIndex - 1]
936                      .getAsCString()
937                      .getValue();
938 
939   // We may still need to append compilation directory of compile unit.
940   // We know that FileName is not absolute, the only way to have an
941   // absolute path at this point would be if IncludeDir is absolute.
942   if (CompDir && Kind == FileLineInfoKind::AbsoluteFilePath &&
943       sys::path::is_relative(IncludeDir))
944     sys::path::append(FilePath, CompDir);
945 
946   // sys::path::append skips empty strings.
947   sys::path::append(FilePath, IncludeDir, FileName);
948   Result = FilePath.str();
949   return true;
950 }
951 
952 bool DWARFDebugLine::LineTable::getFileLineInfoForAddress(
953     uint64_t Address, const char *CompDir, FileLineInfoKind Kind,
954     DILineInfo &Result) const {
955   // Get the index of row we're looking for in the line table.
956   uint32_t RowIndex = lookupAddress(Address);
957   if (RowIndex == -1U)
958     return false;
959   // Take file number and line/column from the row.
960   const auto &Row = Rows[RowIndex];
961   if (!getFileNameByIndex(Row.File, CompDir, Kind, Result.FileName))
962     return false;
963   Result.Line = Row.Line;
964   Result.Column = Row.Column;
965   Result.Discriminator = Row.Discriminator;
966   return true;
967 }
968