1 //===-- llvm-dwp.cpp - Split DWARF merging tool for llvm ------------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // A utility for merging DWARF 5 Split DWARF .dwo files into .dwp (DWARF
10 // package files).
11 //
12 //===----------------------------------------------------------------------===//
13 #include "llvm/DWP/DWP.h"
14 #include "llvm/ADT/Twine.h"
15 #include "llvm/DWP/DWPError.h"
16 #include "llvm/MC/MCContext.h"
17 #include "llvm/MC/MCObjectFileInfo.h"
18 #include "llvm/MC/MCTargetOptionsCommandFlags.h"
19 #include "llvm/Object/Decompressor.h"
20 #include "llvm/Object/ELFObjectFile.h"
21 #include "llvm/Support/CommandLine.h"
22 #include "llvm/Support/MemoryBuffer.h"
23 #include <limits>
24 
25 using namespace llvm;
26 using namespace llvm::object;
27 
28 static mc::RegisterMCTargetOptionsFlags MCTargetOptionsFlags;
29 
30 // Returns the size of debug_str_offsets section headers in bytes.
debugStrOffsetsHeaderSize(DataExtractor StrOffsetsData,uint16_t DwarfVersion)31 static uint64_t debugStrOffsetsHeaderSize(DataExtractor StrOffsetsData,
32                                           uint16_t DwarfVersion) {
33   if (DwarfVersion <= 4)
34     return 0; // There is no header before dwarf 5.
35   uint64_t Offset = 0;
36   uint64_t Length = StrOffsetsData.getU32(&Offset);
37   if (Length == llvm::dwarf::DW_LENGTH_DWARF64)
38     return 16; // unit length: 12 bytes, version: 2 bytes, padding: 2 bytes.
39   return 8;    // unit length: 4 bytes, version: 2 bytes, padding: 2 bytes.
40 }
41 
getCUAbbrev(StringRef Abbrev,uint64_t AbbrCode)42 static uint64_t getCUAbbrev(StringRef Abbrev, uint64_t AbbrCode) {
43   uint64_t Offset = 0;
44   DataExtractor AbbrevData(Abbrev, true, 0);
45   while (AbbrevData.getULEB128(&Offset) != AbbrCode) {
46     // Tag
47     AbbrevData.getULEB128(&Offset);
48     // DW_CHILDREN
49     AbbrevData.getU8(&Offset);
50     // Attributes
51     while (AbbrevData.getULEB128(&Offset) | AbbrevData.getULEB128(&Offset))
52       ;
53   }
54   return Offset;
55 }
56 
57 static Expected<const char *>
getIndexedString(dwarf::Form Form,DataExtractor InfoData,uint64_t & InfoOffset,StringRef StrOffsets,StringRef Str,uint16_t Version)58 getIndexedString(dwarf::Form Form, DataExtractor InfoData, uint64_t &InfoOffset,
59                  StringRef StrOffsets, StringRef Str, uint16_t Version) {
60   if (Form == dwarf::DW_FORM_string)
61     return InfoData.getCStr(&InfoOffset);
62   uint64_t StrIndex;
63   switch (Form) {
64   case dwarf::DW_FORM_strx1:
65     StrIndex = InfoData.getU8(&InfoOffset);
66     break;
67   case dwarf::DW_FORM_strx2:
68     StrIndex = InfoData.getU16(&InfoOffset);
69     break;
70   case dwarf::DW_FORM_strx3:
71     StrIndex = InfoData.getU24(&InfoOffset);
72     break;
73   case dwarf::DW_FORM_strx4:
74     StrIndex = InfoData.getU32(&InfoOffset);
75     break;
76   case dwarf::DW_FORM_strx:
77   case dwarf::DW_FORM_GNU_str_index:
78     StrIndex = InfoData.getULEB128(&InfoOffset);
79     break;
80   default:
81     return make_error<DWPError>(
82         "string field must be encoded with one of the following: "
83         "DW_FORM_string, DW_FORM_strx, DW_FORM_strx1, DW_FORM_strx2, "
84         "DW_FORM_strx3, DW_FORM_strx4, or DW_FORM_GNU_str_index.");
85   }
86   DataExtractor StrOffsetsData(StrOffsets, true, 0);
87   uint64_t StrOffsetsOffset = 4 * StrIndex;
88   StrOffsetsOffset += debugStrOffsetsHeaderSize(StrOffsetsData, Version);
89 
90   uint64_t StrOffset = StrOffsetsData.getU32(&StrOffsetsOffset);
91   DataExtractor StrData(Str, true, 0);
92   return StrData.getCStr(&StrOffset);
93 }
94 
95 static Expected<CompileUnitIdentifiers>
getCUIdentifiers(InfoSectionUnitHeader & Header,StringRef Abbrev,StringRef Info,StringRef StrOffsets,StringRef Str)96 getCUIdentifiers(InfoSectionUnitHeader &Header, StringRef Abbrev,
97                  StringRef Info, StringRef StrOffsets, StringRef Str) {
98   DataExtractor InfoData(Info, true, 0);
99   uint64_t Offset = Header.HeaderSize;
100   if (Header.Version >= 5 && Header.UnitType != dwarf::DW_UT_split_compile)
101     return make_error<DWPError>(
102         std::string("unit type DW_UT_split_compile type not found in "
103                     "debug_info header. Unexpected unit type 0x" +
104                     utostr(Header.UnitType) + " found"));
105 
106   CompileUnitIdentifiers ID;
107 
108   uint32_t AbbrCode = InfoData.getULEB128(&Offset);
109   DataExtractor AbbrevData(Abbrev, true, 0);
110   uint64_t AbbrevOffset = getCUAbbrev(Abbrev, AbbrCode);
111   auto Tag = static_cast<dwarf::Tag>(AbbrevData.getULEB128(&AbbrevOffset));
112   if (Tag != dwarf::DW_TAG_compile_unit)
113     return make_error<DWPError>("top level DIE is not a compile unit");
114   // DW_CHILDREN
115   AbbrevData.getU8(&AbbrevOffset);
116   uint32_t Name;
117   dwarf::Form Form;
118   while ((Name = AbbrevData.getULEB128(&AbbrevOffset)) |
119              (Form = static_cast<dwarf::Form>(
120                   AbbrevData.getULEB128(&AbbrevOffset))) &&
121          (Name != 0 || Form != 0)) {
122     switch (Name) {
123     case dwarf::DW_AT_name: {
124       Expected<const char *> EName = getIndexedString(
125           Form, InfoData, Offset, StrOffsets, Str, Header.Version);
126       if (!EName)
127         return EName.takeError();
128       ID.Name = *EName;
129       break;
130     }
131     case dwarf::DW_AT_GNU_dwo_name:
132     case dwarf::DW_AT_dwo_name: {
133       Expected<const char *> EName = getIndexedString(
134           Form, InfoData, Offset, StrOffsets, Str, Header.Version);
135       if (!EName)
136         return EName.takeError();
137       ID.DWOName = *EName;
138       break;
139     }
140     case dwarf::DW_AT_GNU_dwo_id:
141       Header.Signature = InfoData.getU64(&Offset);
142       break;
143     default:
144       DWARFFormValue::skipValue(
145           Form, InfoData, &Offset,
146           dwarf::FormParams({Header.Version, Header.AddrSize, Header.Format}));
147     }
148   }
149   if (!Header.Signature)
150     return make_error<DWPError>("compile unit missing dwo_id");
151   ID.Signature = *Header.Signature;
152   return ID;
153 }
154 
isSupportedSectionKind(DWARFSectionKind Kind)155 static bool isSupportedSectionKind(DWARFSectionKind Kind) {
156   return Kind != DW_SECT_EXT_unknown;
157 }
158 
159 namespace llvm {
160 // Convert an internal section identifier into the index to use with
161 // UnitIndexEntry::Contributions.
getContributionIndex(DWARFSectionKind Kind,uint32_t IndexVersion)162 unsigned getContributionIndex(DWARFSectionKind Kind, uint32_t IndexVersion) {
163   assert(serializeSectionKind(Kind, IndexVersion) >= DW_SECT_INFO);
164   return serializeSectionKind(Kind, IndexVersion) - DW_SECT_INFO;
165 }
166 } // namespace llvm
167 
168 // Convert a UnitIndexEntry::Contributions index to the corresponding on-disk
169 // value of the section identifier.
getOnDiskSectionId(unsigned Index)170 static unsigned getOnDiskSectionId(unsigned Index) {
171   return Index + DW_SECT_INFO;
172 }
173 
getSubsection(StringRef Section,const DWARFUnitIndex::Entry & Entry,DWARFSectionKind Kind)174 static StringRef getSubsection(StringRef Section,
175                                const DWARFUnitIndex::Entry &Entry,
176                                DWARFSectionKind Kind) {
177   const auto *Off = Entry.getContribution(Kind);
178   if (!Off)
179     return StringRef();
180   return Section.substr(Off->getOffset(), Off->getLength());
181 }
182 
sectionOverflowErrorOrWarning(uint32_t PrevOffset,uint32_t OverflowedOffset,StringRef SectionName,OnCuIndexOverflow OverflowOptValue,bool & AnySectionOverflow)183 static Error sectionOverflowErrorOrWarning(uint32_t PrevOffset,
184                                            uint32_t OverflowedOffset,
185                                            StringRef SectionName,
186                                            OnCuIndexOverflow OverflowOptValue,
187                                            bool &AnySectionOverflow) {
188   std::string Msg =
189       (SectionName +
190        Twine(" Section Contribution Offset overflow 4G. Previous Offset ") +
191        Twine(PrevOffset) + Twine(", After overflow offset ") +
192        Twine(OverflowedOffset) + Twine("."))
193           .str();
194   if (OverflowOptValue == OnCuIndexOverflow::Continue) {
195     WithColor::defaultWarningHandler(make_error<DWPError>(Msg));
196     return Error::success();
197   } else if (OverflowOptValue == OnCuIndexOverflow::SoftStop) {
198     AnySectionOverflow = true;
199     WithColor::defaultWarningHandler(make_error<DWPError>(Msg));
200     return Error::success();
201   }
202   return make_error<DWPError>(Msg);
203 }
204 
addAllTypesFromDWP(MCStreamer & Out,MapVector<uint64_t,UnitIndexEntry> & TypeIndexEntries,const DWARFUnitIndex & TUIndex,MCSection * OutputTypes,StringRef Types,const UnitIndexEntry & TUEntry,uint32_t & TypesOffset,unsigned TypesContributionIndex,OnCuIndexOverflow OverflowOptValue,bool & AnySectionOverflow)205 static Error addAllTypesFromDWP(
206     MCStreamer &Out, MapVector<uint64_t, UnitIndexEntry> &TypeIndexEntries,
207     const DWARFUnitIndex &TUIndex, MCSection *OutputTypes, StringRef Types,
208     const UnitIndexEntry &TUEntry, uint32_t &TypesOffset,
209     unsigned TypesContributionIndex, OnCuIndexOverflow OverflowOptValue,
210     bool &AnySectionOverflow) {
211   Out.switchSection(OutputTypes);
212   for (const DWARFUnitIndex::Entry &E : TUIndex.getRows()) {
213     auto *I = E.getContributions();
214     if (!I)
215       continue;
216     auto P = TypeIndexEntries.insert(std::make_pair(E.getSignature(), TUEntry));
217     if (!P.second)
218       continue;
219     auto &Entry = P.first->second;
220     // Zero out the debug_info contribution
221     Entry.Contributions[0] = {};
222     for (auto Kind : TUIndex.getColumnKinds()) {
223       if (!isSupportedSectionKind(Kind))
224         continue;
225       auto &C =
226           Entry.Contributions[getContributionIndex(Kind, TUIndex.getVersion())];
227       C.setOffset(C.getOffset() + I->getOffset());
228       C.setLength(I->getLength());
229       ++I;
230     }
231     auto &C = Entry.Contributions[TypesContributionIndex];
232     Out.emitBytes(Types.substr(
233         C.getOffset() -
234             TUEntry.Contributions[TypesContributionIndex].getOffset(),
235         C.getLength()));
236     C.setOffset(TypesOffset);
237     uint32_t OldOffset = TypesOffset;
238     static_assert(sizeof(OldOffset) == sizeof(TypesOffset));
239     TypesOffset += C.getLength();
240     if (OldOffset > TypesOffset) {
241       if (Error Err = sectionOverflowErrorOrWarning(OldOffset, TypesOffset,
242                                                     "Types", OverflowOptValue,
243                                                     AnySectionOverflow))
244         return Err;
245       if (AnySectionOverflow) {
246         TypesOffset = OldOffset;
247         return Error::success();
248       }
249     }
250   }
251   return Error::success();
252 }
253 
addAllTypesFromTypesSection(MCStreamer & Out,MapVector<uint64_t,UnitIndexEntry> & TypeIndexEntries,MCSection * OutputTypes,const std::vector<StringRef> & TypesSections,const UnitIndexEntry & CUEntry,uint32_t & TypesOffset,OnCuIndexOverflow OverflowOptValue,bool & AnySectionOverflow)254 static Error addAllTypesFromTypesSection(
255     MCStreamer &Out, MapVector<uint64_t, UnitIndexEntry> &TypeIndexEntries,
256     MCSection *OutputTypes, const std::vector<StringRef> &TypesSections,
257     const UnitIndexEntry &CUEntry, uint32_t &TypesOffset,
258     OnCuIndexOverflow OverflowOptValue, bool &AnySectionOverflow) {
259   for (StringRef Types : TypesSections) {
260     Out.switchSection(OutputTypes);
261     uint64_t Offset = 0;
262     DataExtractor Data(Types, true, 0);
263     while (Data.isValidOffset(Offset)) {
264       UnitIndexEntry Entry = CUEntry;
265       // Zero out the debug_info contribution
266       Entry.Contributions[0] = {};
267       auto &C = Entry.Contributions[getContributionIndex(DW_SECT_EXT_TYPES, 2)];
268       C.setOffset(TypesOffset);
269       auto PrevOffset = Offset;
270       // Length of the unit, including the 4 byte length field.
271       C.setLength(Data.getU32(&Offset) + 4);
272 
273       Data.getU16(&Offset); // Version
274       Data.getU32(&Offset); // Abbrev offset
275       Data.getU8(&Offset);  // Address size
276       auto Signature = Data.getU64(&Offset);
277       Offset = PrevOffset + C.getLength32();
278 
279       auto P = TypeIndexEntries.insert(std::make_pair(Signature, Entry));
280       if (!P.second)
281         continue;
282 
283       Out.emitBytes(Types.substr(PrevOffset, C.getLength32()));
284       uint32_t OldOffset = TypesOffset;
285       TypesOffset += C.getLength32();
286       if (OldOffset > TypesOffset) {
287         if (Error Err = sectionOverflowErrorOrWarning(OldOffset, TypesOffset,
288                                                       "Types", OverflowOptValue,
289                                                       AnySectionOverflow))
290           return Err;
291         if (AnySectionOverflow) {
292           TypesOffset = OldOffset;
293           return Error::success();
294         }
295       }
296     }
297   }
298   return Error::success();
299 }
300 
buildDWODescription(StringRef Name,StringRef DWPName,StringRef DWOName)301 static std::string buildDWODescription(StringRef Name, StringRef DWPName,
302                                        StringRef DWOName) {
303   std::string Text = "\'";
304   Text += Name;
305   Text += '\'';
306   bool HasDWO = !DWOName.empty();
307   bool HasDWP = !DWPName.empty();
308   if (HasDWO || HasDWP) {
309     Text += " (from ";
310     if (HasDWO) {
311       Text += '\'';
312       Text += DWOName;
313       Text += '\'';
314     }
315     if (HasDWO && HasDWP)
316       Text += " in ";
317     if (!DWPName.empty()) {
318       Text += '\'';
319       Text += DWPName;
320       Text += '\'';
321     }
322     Text += ")";
323   }
324   return Text;
325 }
326 
createError(StringRef Name,Error E)327 static Error createError(StringRef Name, Error E) {
328   return make_error<DWPError>(
329       ("failure while decompressing compressed section: '" + Name + "', " +
330        llvm::toString(std::move(E)))
331           .str());
332 }
333 
334 static Error
handleCompressedSection(std::deque<SmallString<32>> & UncompressedSections,SectionRef Sec,StringRef Name,StringRef & Contents)335 handleCompressedSection(std::deque<SmallString<32>> &UncompressedSections,
336                         SectionRef Sec, StringRef Name, StringRef &Contents) {
337   auto *Obj = dyn_cast<ELFObjectFileBase>(Sec.getObject());
338   if (!Obj ||
339       !(static_cast<ELFSectionRef>(Sec).getFlags() & ELF::SHF_COMPRESSED))
340     return Error::success();
341   bool IsLE = isa<object::ELF32LEObjectFile>(Obj) ||
342               isa<object::ELF64LEObjectFile>(Obj);
343   bool Is64 = isa<object::ELF64LEObjectFile>(Obj) ||
344               isa<object::ELF64BEObjectFile>(Obj);
345   Expected<Decompressor> Dec = Decompressor::create(Name, Contents, IsLE, Is64);
346   if (!Dec)
347     return createError(Name, Dec.takeError());
348 
349   UncompressedSections.emplace_back();
350   if (Error E = Dec->resizeAndDecompress(UncompressedSections.back()))
351     return createError(Name, std::move(E));
352 
353   Contents = UncompressedSections.back();
354   return Error::success();
355 }
356 
357 namespace llvm {
358 // Parse and return the header of an info section compile/type unit.
parseInfoSectionUnitHeader(StringRef Info)359 Expected<InfoSectionUnitHeader> parseInfoSectionUnitHeader(StringRef Info) {
360   InfoSectionUnitHeader Header;
361   Error Err = Error::success();
362   uint64_t Offset = 0;
363   DWARFDataExtractor InfoData(Info, true, 0);
364   std::tie(Header.Length, Header.Format) =
365       InfoData.getInitialLength(&Offset, &Err);
366   if (Err)
367     return make_error<DWPError>("cannot parse compile unit length: " +
368                                 llvm::toString(std::move(Err)));
369 
370   if (!InfoData.isValidOffset(Offset + (Header.Length - 1))) {
371     return make_error<DWPError>(
372         "compile unit exceeds .debug_info section range: " +
373         utostr(Offset + Header.Length) + " >= " + utostr(InfoData.size()));
374   }
375 
376   Header.Version = InfoData.getU16(&Offset, &Err);
377   if (Err)
378     return make_error<DWPError>("cannot parse compile unit version: " +
379                                 llvm::toString(std::move(Err)));
380 
381   uint64_t MinHeaderLength;
382   if (Header.Version >= 5) {
383     // Size: Version (2), UnitType (1), AddrSize (1), DebugAbbrevOffset (4),
384     // Signature (8)
385     MinHeaderLength = 16;
386   } else {
387     // Size: Version (2), DebugAbbrevOffset (4), AddrSize (1)
388     MinHeaderLength = 7;
389   }
390   if (Header.Length < MinHeaderLength) {
391     return make_error<DWPError>("unit length is too small: expected at least " +
392                                 utostr(MinHeaderLength) + " got " +
393                                 utostr(Header.Length) + ".");
394   }
395   if (Header.Version >= 5) {
396     Header.UnitType = InfoData.getU8(&Offset);
397     Header.AddrSize = InfoData.getU8(&Offset);
398     Header.DebugAbbrevOffset = InfoData.getU32(&Offset);
399     Header.Signature = InfoData.getU64(&Offset);
400     if (Header.UnitType == dwarf::DW_UT_split_type) {
401       // Type offset.
402       MinHeaderLength += 4;
403       if (Header.Length < MinHeaderLength)
404         return make_error<DWPError>("type unit is missing type offset");
405       InfoData.getU32(&Offset);
406     }
407   } else {
408     // Note that, address_size and debug_abbrev_offset fields have switched
409     // places between dwarf version 4 and 5.
410     Header.DebugAbbrevOffset = InfoData.getU32(&Offset);
411     Header.AddrSize = InfoData.getU8(&Offset);
412   }
413 
414   Header.HeaderSize = Offset;
415   return Header;
416 }
417 
writeStringsAndOffsets(MCStreamer & Out,DWPStringPool & Strings,MCSection * StrOffsetSection,StringRef CurStrSection,StringRef CurStrOffsetSection,uint16_t Version)418 void writeStringsAndOffsets(MCStreamer &Out, DWPStringPool &Strings,
419                             MCSection *StrOffsetSection,
420                             StringRef CurStrSection,
421                             StringRef CurStrOffsetSection, uint16_t Version) {
422   // Could possibly produce an error or warning if one of these was non-null but
423   // the other was null.
424   if (CurStrSection.empty() || CurStrOffsetSection.empty())
425     return;
426 
427   DenseMap<uint64_t, uint32_t> OffsetRemapping;
428 
429   DataExtractor Data(CurStrSection, true, 0);
430   uint64_t LocalOffset = 0;
431   uint64_t PrevOffset = 0;
432   while (const char *S = Data.getCStr(&LocalOffset)) {
433     OffsetRemapping[PrevOffset] =
434         Strings.getOffset(S, LocalOffset - PrevOffset);
435     PrevOffset = LocalOffset;
436   }
437 
438   Data = DataExtractor(CurStrOffsetSection, true, 0);
439 
440   Out.switchSection(StrOffsetSection);
441 
442   uint64_t HeaderSize = debugStrOffsetsHeaderSize(Data, Version);
443   uint64_t Offset = 0;
444   uint64_t Size = CurStrOffsetSection.size();
445   // FIXME: This can be caused by bad input and should be handled as such.
446   assert(HeaderSize <= Size && "StrOffsetSection size is less than its header");
447   // Copy the header to the output.
448   Out.emitBytes(Data.getBytes(&Offset, HeaderSize));
449   while (Offset < Size) {
450     auto OldOffset = Data.getU32(&Offset);
451     auto NewOffset = OffsetRemapping[OldOffset];
452     Out.emitIntValue(NewOffset, 4);
453   }
454 }
455 
456 enum AccessField { Offset, Length };
writeIndexTable(MCStreamer & Out,ArrayRef<unsigned> ContributionOffsets,const MapVector<uint64_t,UnitIndexEntry> & IndexEntries,const AccessField & Field)457 void writeIndexTable(MCStreamer &Out, ArrayRef<unsigned> ContributionOffsets,
458                      const MapVector<uint64_t, UnitIndexEntry> &IndexEntries,
459                      const AccessField &Field) {
460   for (const auto &E : IndexEntries)
461     for (size_t I = 0; I != std::size(E.second.Contributions); ++I)
462       if (ContributionOffsets[I])
463         Out.emitIntValue((Field == AccessField::Offset
464                               ? E.second.Contributions[I].getOffset32()
465                               : E.second.Contributions[I].getLength32()),
466                          4);
467 }
468 
writeIndex(MCStreamer & Out,MCSection * Section,ArrayRef<unsigned> ContributionOffsets,const MapVector<uint64_t,UnitIndexEntry> & IndexEntries,uint32_t IndexVersion)469 void writeIndex(MCStreamer &Out, MCSection *Section,
470                 ArrayRef<unsigned> ContributionOffsets,
471                 const MapVector<uint64_t, UnitIndexEntry> &IndexEntries,
472                 uint32_t IndexVersion) {
473   if (IndexEntries.empty())
474     return;
475 
476   unsigned Columns = 0;
477   for (auto &C : ContributionOffsets)
478     if (C)
479       ++Columns;
480 
481   std::vector<unsigned> Buckets(NextPowerOf2(3 * IndexEntries.size() / 2));
482   uint64_t Mask = Buckets.size() - 1;
483   size_t I = 0;
484   for (const auto &P : IndexEntries) {
485     auto S = P.first;
486     auto H = S & Mask;
487     auto HP = ((S >> 32) & Mask) | 1;
488     while (Buckets[H]) {
489       assert(S != IndexEntries.begin()[Buckets[H] - 1].first &&
490              "Duplicate unit");
491       H = (H + HP) & Mask;
492     }
493     Buckets[H] = I + 1;
494     ++I;
495   }
496 
497   Out.switchSection(Section);
498   Out.emitIntValue(IndexVersion, 4);        // Version
499   Out.emitIntValue(Columns, 4);             // Columns
500   Out.emitIntValue(IndexEntries.size(), 4); // Num Units
501   Out.emitIntValue(Buckets.size(), 4);      // Num Buckets
502 
503   // Write the signatures.
504   for (const auto &I : Buckets)
505     Out.emitIntValue(I ? IndexEntries.begin()[I - 1].first : 0, 8);
506 
507   // Write the indexes.
508   for (const auto &I : Buckets)
509     Out.emitIntValue(I, 4);
510 
511   // Write the column headers (which sections will appear in the table)
512   for (size_t I = 0; I != ContributionOffsets.size(); ++I)
513     if (ContributionOffsets[I])
514       Out.emitIntValue(getOnDiskSectionId(I), 4);
515 
516   // Write the offsets.
517   writeIndexTable(Out, ContributionOffsets, IndexEntries, AccessField::Offset);
518 
519   // Write the lengths.
520   writeIndexTable(Out, ContributionOffsets, IndexEntries, AccessField::Length);
521 }
522 
buildDuplicateError(const std::pair<uint64_t,UnitIndexEntry> & PrevE,const CompileUnitIdentifiers & ID,StringRef DWPName)523 Error buildDuplicateError(const std::pair<uint64_t, UnitIndexEntry> &PrevE,
524                           const CompileUnitIdentifiers &ID, StringRef DWPName) {
525   return make_error<DWPError>(
526       std::string("duplicate DWO ID (") + utohexstr(PrevE.first) + ") in " +
527       buildDWODescription(PrevE.second.Name, PrevE.second.DWPName,
528                           PrevE.second.DWOName) +
529       " and " + buildDWODescription(ID.Name, DWPName, ID.DWOName));
530 }
531 
handleSection(const StringMap<std::pair<MCSection *,DWARFSectionKind>> & KnownSections,const MCSection * StrSection,const MCSection * StrOffsetSection,const MCSection * TypesSection,const MCSection * CUIndexSection,const MCSection * TUIndexSection,const MCSection * InfoSection,const SectionRef & Section,MCStreamer & Out,std::deque<SmallString<32>> & UncompressedSections,uint32_t (& ContributionOffsets)[8],UnitIndexEntry & CurEntry,StringRef & CurStrSection,StringRef & CurStrOffsetSection,std::vector<StringRef> & CurTypesSection,std::vector<StringRef> & CurInfoSection,StringRef & AbbrevSection,StringRef & CurCUIndexSection,StringRef & CurTUIndexSection,std::vector<std::pair<DWARFSectionKind,uint32_t>> & SectionLength)532 Error handleSection(
533     const StringMap<std::pair<MCSection *, DWARFSectionKind>> &KnownSections,
534     const MCSection *StrSection, const MCSection *StrOffsetSection,
535     const MCSection *TypesSection, const MCSection *CUIndexSection,
536     const MCSection *TUIndexSection, const MCSection *InfoSection,
537     const SectionRef &Section, MCStreamer &Out,
538     std::deque<SmallString<32>> &UncompressedSections,
539     uint32_t (&ContributionOffsets)[8], UnitIndexEntry &CurEntry,
540     StringRef &CurStrSection, StringRef &CurStrOffsetSection,
541     std::vector<StringRef> &CurTypesSection,
542     std::vector<StringRef> &CurInfoSection, StringRef &AbbrevSection,
543     StringRef &CurCUIndexSection, StringRef &CurTUIndexSection,
544     std::vector<std::pair<DWARFSectionKind, uint32_t>> &SectionLength) {
545   if (Section.isBSS())
546     return Error::success();
547 
548   if (Section.isVirtual())
549     return Error::success();
550 
551   Expected<StringRef> NameOrErr = Section.getName();
552   if (!NameOrErr)
553     return NameOrErr.takeError();
554   StringRef Name = *NameOrErr;
555 
556   Expected<StringRef> ContentsOrErr = Section.getContents();
557   if (!ContentsOrErr)
558     return ContentsOrErr.takeError();
559   StringRef Contents = *ContentsOrErr;
560 
561   if (auto Err = handleCompressedSection(UncompressedSections, Section, Name,
562                                          Contents))
563     return Err;
564 
565   Name = Name.substr(Name.find_first_not_of("._"));
566 
567   auto SectionPair = KnownSections.find(Name);
568   if (SectionPair == KnownSections.end())
569     return Error::success();
570 
571   if (DWARFSectionKind Kind = SectionPair->second.second) {
572     if (Kind != DW_SECT_EXT_TYPES && Kind != DW_SECT_INFO) {
573       SectionLength.push_back(std::make_pair(Kind, Contents.size()));
574     }
575 
576     if (Kind == DW_SECT_ABBREV) {
577       AbbrevSection = Contents;
578     }
579   }
580 
581   MCSection *OutSection = SectionPair->second.first;
582   if (OutSection == StrOffsetSection)
583     CurStrOffsetSection = Contents;
584   else if (OutSection == StrSection)
585     CurStrSection = Contents;
586   else if (OutSection == TypesSection)
587     CurTypesSection.push_back(Contents);
588   else if (OutSection == CUIndexSection)
589     CurCUIndexSection = Contents;
590   else if (OutSection == TUIndexSection)
591     CurTUIndexSection = Contents;
592   else if (OutSection == InfoSection)
593     CurInfoSection.push_back(Contents);
594   else {
595     Out.switchSection(OutSection);
596     Out.emitBytes(Contents);
597   }
598   return Error::success();
599 }
600 
write(MCStreamer & Out,ArrayRef<std::string> Inputs,OnCuIndexOverflow OverflowOptValue)601 Error write(MCStreamer &Out, ArrayRef<std::string> Inputs,
602             OnCuIndexOverflow OverflowOptValue) {
603   const auto &MCOFI = *Out.getContext().getObjectFileInfo();
604   MCSection *const StrSection = MCOFI.getDwarfStrDWOSection();
605   MCSection *const StrOffsetSection = MCOFI.getDwarfStrOffDWOSection();
606   MCSection *const TypesSection = MCOFI.getDwarfTypesDWOSection();
607   MCSection *const CUIndexSection = MCOFI.getDwarfCUIndexSection();
608   MCSection *const TUIndexSection = MCOFI.getDwarfTUIndexSection();
609   MCSection *const InfoSection = MCOFI.getDwarfInfoDWOSection();
610   const StringMap<std::pair<MCSection *, DWARFSectionKind>> KnownSections = {
611       {"debug_info.dwo", {InfoSection, DW_SECT_INFO}},
612       {"debug_types.dwo", {MCOFI.getDwarfTypesDWOSection(), DW_SECT_EXT_TYPES}},
613       {"debug_str_offsets.dwo", {StrOffsetSection, DW_SECT_STR_OFFSETS}},
614       {"debug_str.dwo", {StrSection, static_cast<DWARFSectionKind>(0)}},
615       {"debug_loc.dwo", {MCOFI.getDwarfLocDWOSection(), DW_SECT_EXT_LOC}},
616       {"debug_line.dwo", {MCOFI.getDwarfLineDWOSection(), DW_SECT_LINE}},
617       {"debug_macro.dwo", {MCOFI.getDwarfMacroDWOSection(), DW_SECT_MACRO}},
618       {"debug_abbrev.dwo", {MCOFI.getDwarfAbbrevDWOSection(), DW_SECT_ABBREV}},
619       {"debug_loclists.dwo",
620        {MCOFI.getDwarfLoclistsDWOSection(), DW_SECT_LOCLISTS}},
621       {"debug_rnglists.dwo",
622        {MCOFI.getDwarfRnglistsDWOSection(), DW_SECT_RNGLISTS}},
623       {"debug_cu_index", {CUIndexSection, static_cast<DWARFSectionKind>(0)}},
624       {"debug_tu_index", {TUIndexSection, static_cast<DWARFSectionKind>(0)}}};
625 
626   MapVector<uint64_t, UnitIndexEntry> IndexEntries;
627   MapVector<uint64_t, UnitIndexEntry> TypeIndexEntries;
628 
629   uint32_t ContributionOffsets[8] = {};
630   uint16_t Version = 0;
631   uint32_t IndexVersion = 0;
632   bool AnySectionOverflow = false;
633 
634   DWPStringPool Strings(Out, StrSection);
635 
636   SmallVector<OwningBinary<object::ObjectFile>, 128> Objects;
637   Objects.reserve(Inputs.size());
638 
639   std::deque<SmallString<32>> UncompressedSections;
640 
641   for (const auto &Input : Inputs) {
642     auto ErrOrObj = object::ObjectFile::createObjectFile(Input);
643     if (!ErrOrObj) {
644       return handleErrors(ErrOrObj.takeError(),
645                           [&](std::unique_ptr<ECError> EC) -> Error {
646                             return createFileError(Input, Error(std::move(EC)));
647                           });
648     }
649 
650     auto &Obj = *ErrOrObj->getBinary();
651     Objects.push_back(std::move(*ErrOrObj));
652 
653     UnitIndexEntry CurEntry = {};
654 
655     StringRef CurStrSection;
656     StringRef CurStrOffsetSection;
657     std::vector<StringRef> CurTypesSection;
658     std::vector<StringRef> CurInfoSection;
659     StringRef AbbrevSection;
660     StringRef CurCUIndexSection;
661     StringRef CurTUIndexSection;
662 
663     // This maps each section contained in this file to its length.
664     // This information is later on used to calculate the contributions,
665     // i.e. offset and length, of each compile/type unit to a section.
666     std::vector<std::pair<DWARFSectionKind, uint32_t>> SectionLength;
667 
668     for (const auto &Section : Obj.sections())
669       if (auto Err = handleSection(
670               KnownSections, StrSection, StrOffsetSection, TypesSection,
671               CUIndexSection, TUIndexSection, InfoSection, Section, Out,
672               UncompressedSections, ContributionOffsets, CurEntry,
673               CurStrSection, CurStrOffsetSection, CurTypesSection,
674               CurInfoSection, AbbrevSection, CurCUIndexSection,
675               CurTUIndexSection, SectionLength))
676         return Err;
677 
678     if (CurInfoSection.empty())
679       continue;
680 
681     Expected<InfoSectionUnitHeader> HeaderOrErr =
682         parseInfoSectionUnitHeader(CurInfoSection.front());
683     if (!HeaderOrErr)
684       return HeaderOrErr.takeError();
685     InfoSectionUnitHeader &Header = *HeaderOrErr;
686 
687     if (Version == 0) {
688       Version = Header.Version;
689       IndexVersion = Version < 5 ? 2 : 5;
690     } else if (Version != Header.Version) {
691       return make_error<DWPError>("incompatible DWARF compile unit versions.");
692     }
693 
694     writeStringsAndOffsets(Out, Strings, StrOffsetSection, CurStrSection,
695                            CurStrOffsetSection, Header.Version);
696 
697     for (auto Pair : SectionLength) {
698       auto Index = getContributionIndex(Pair.first, IndexVersion);
699       CurEntry.Contributions[Index].setOffset(ContributionOffsets[Index]);
700       CurEntry.Contributions[Index].setLength(Pair.second);
701       uint32_t OldOffset = ContributionOffsets[Index];
702       ContributionOffsets[Index] += CurEntry.Contributions[Index].getLength32();
703       if (OldOffset > ContributionOffsets[Index]) {
704         uint32_t SectionIndex = 0;
705         for (auto &Section : Obj.sections()) {
706           if (SectionIndex == Index) {
707             if (Error Err = sectionOverflowErrorOrWarning(
708                     OldOffset, ContributionOffsets[Index], *Section.getName(),
709                     OverflowOptValue, AnySectionOverflow))
710               return Err;
711           }
712           ++SectionIndex;
713         }
714         if (AnySectionOverflow)
715           break;
716       }
717     }
718 
719     uint32_t &InfoSectionOffset =
720         ContributionOffsets[getContributionIndex(DW_SECT_INFO, IndexVersion)];
721     if (CurCUIndexSection.empty()) {
722       bool FoundCUUnit = false;
723       Out.switchSection(InfoSection);
724       for (StringRef Info : CurInfoSection) {
725         uint64_t UnitOffset = 0;
726         while (Info.size() > UnitOffset) {
727           Expected<InfoSectionUnitHeader> HeaderOrError =
728               parseInfoSectionUnitHeader(Info.substr(UnitOffset, Info.size()));
729           if (!HeaderOrError)
730             return HeaderOrError.takeError();
731           InfoSectionUnitHeader &Header = *HeaderOrError;
732 
733           UnitIndexEntry Entry = CurEntry;
734           auto &C = Entry.Contributions[getContributionIndex(DW_SECT_INFO,
735                                                              IndexVersion)];
736           C.setOffset(InfoSectionOffset);
737           C.setLength(Header.Length + 4);
738 
739           if (std::numeric_limits<uint32_t>::max() - InfoSectionOffset <
740               C.getLength32()) {
741             if (Error Err = sectionOverflowErrorOrWarning(
742                     InfoSectionOffset, InfoSectionOffset + C.getLength32(),
743                     "debug_info", OverflowOptValue, AnySectionOverflow))
744               return Err;
745             if (AnySectionOverflow) {
746               if (Header.Version < 5 ||
747                   Header.UnitType == dwarf::DW_UT_split_compile)
748                 FoundCUUnit = true;
749               break;
750             }
751           }
752 
753           UnitOffset += C.getLength32();
754           if (Header.Version < 5 ||
755               Header.UnitType == dwarf::DW_UT_split_compile) {
756             Expected<CompileUnitIdentifiers> EID = getCUIdentifiers(
757                 Header, AbbrevSection,
758                 Info.substr(UnitOffset - C.getLength32(), C.getLength32()),
759                 CurStrOffsetSection, CurStrSection);
760 
761             if (!EID)
762               return createFileError(Input, EID.takeError());
763             const auto &ID = *EID;
764             auto P = IndexEntries.insert(std::make_pair(ID.Signature, Entry));
765             if (!P.second)
766               return buildDuplicateError(*P.first, ID, "");
767             P.first->second.Name = ID.Name;
768             P.first->second.DWOName = ID.DWOName;
769 
770             FoundCUUnit = true;
771           } else if (Header.UnitType == dwarf::DW_UT_split_type) {
772             auto P = TypeIndexEntries.insert(
773                 std::make_pair(*Header.Signature, Entry));
774             if (!P.second)
775               continue;
776           }
777           Out.emitBytes(
778               Info.substr(UnitOffset - C.getLength32(), C.getLength32()));
779           InfoSectionOffset += C.getLength32();
780         }
781         if (AnySectionOverflow)
782           break;
783       }
784 
785       if (!FoundCUUnit)
786         return make_error<DWPError>("no compile unit found in file: " + Input);
787 
788       if (IndexVersion == 2) {
789         // Add types from the .debug_types section from DWARF < 5.
790         if (Error Err = addAllTypesFromTypesSection(
791                 Out, TypeIndexEntries, TypesSection, CurTypesSection, CurEntry,
792                 ContributionOffsets[getContributionIndex(DW_SECT_EXT_TYPES, 2)],
793                 OverflowOptValue, AnySectionOverflow))
794           return Err;
795       }
796       if (AnySectionOverflow)
797         break;
798       continue;
799     }
800 
801     if (CurInfoSection.size() != 1)
802       return make_error<DWPError>("expected exactly one occurrence of a debug "
803                                   "info section in a .dwp file");
804     StringRef DwpSingleInfoSection = CurInfoSection.front();
805 
806     DWARFUnitIndex CUIndex(DW_SECT_INFO);
807     DataExtractor CUIndexData(CurCUIndexSection, Obj.isLittleEndian(), 0);
808     if (!CUIndex.parse(CUIndexData))
809       return make_error<DWPError>("failed to parse cu_index");
810     if (CUIndex.getVersion() != IndexVersion)
811       return make_error<DWPError>("incompatible cu_index versions, found " +
812                                   utostr(CUIndex.getVersion()) +
813                                   " and expecting " + utostr(IndexVersion));
814 
815     Out.switchSection(InfoSection);
816     for (const DWARFUnitIndex::Entry &E : CUIndex.getRows()) {
817       auto *I = E.getContributions();
818       if (!I)
819         continue;
820       auto P = IndexEntries.insert(std::make_pair(E.getSignature(), CurEntry));
821       StringRef CUInfoSection =
822           getSubsection(DwpSingleInfoSection, E, DW_SECT_INFO);
823       Expected<InfoSectionUnitHeader> HeaderOrError =
824           parseInfoSectionUnitHeader(CUInfoSection);
825       if (!HeaderOrError)
826         return HeaderOrError.takeError();
827       InfoSectionUnitHeader &Header = *HeaderOrError;
828 
829       Expected<CompileUnitIdentifiers> EID = getCUIdentifiers(
830           Header, getSubsection(AbbrevSection, E, DW_SECT_ABBREV),
831           CUInfoSection,
832           getSubsection(CurStrOffsetSection, E, DW_SECT_STR_OFFSETS),
833           CurStrSection);
834       if (!EID)
835         return createFileError(Input, EID.takeError());
836       const auto &ID = *EID;
837       if (!P.second)
838         return buildDuplicateError(*P.first, ID, Input);
839       auto &NewEntry = P.first->second;
840       NewEntry.Name = ID.Name;
841       NewEntry.DWOName = ID.DWOName;
842       NewEntry.DWPName = Input;
843       for (auto Kind : CUIndex.getColumnKinds()) {
844         if (!isSupportedSectionKind(Kind))
845           continue;
846         auto &C =
847             NewEntry.Contributions[getContributionIndex(Kind, IndexVersion)];
848         C.setOffset(C.getOffset() + I->getOffset());
849         C.setLength(I->getLength());
850         ++I;
851       }
852       unsigned Index = getContributionIndex(DW_SECT_INFO, IndexVersion);
853       auto &C = NewEntry.Contributions[Index];
854       Out.emitBytes(CUInfoSection);
855       C.setOffset(InfoSectionOffset);
856       InfoSectionOffset += C.getLength32();
857     }
858 
859     if (!CurTUIndexSection.empty()) {
860       llvm::DWARFSectionKind TUSectionKind;
861       MCSection *OutSection;
862       StringRef TypeInputSection;
863       // Write type units into debug info section for DWARFv5.
864       if (Version >= 5) {
865         TUSectionKind = DW_SECT_INFO;
866         OutSection = InfoSection;
867         TypeInputSection = DwpSingleInfoSection;
868       } else {
869         // Write type units into debug types section for DWARF < 5.
870         if (CurTypesSection.size() != 1)
871           return make_error<DWPError>(
872               "multiple type unit sections in .dwp file");
873 
874         TUSectionKind = DW_SECT_EXT_TYPES;
875         OutSection = TypesSection;
876         TypeInputSection = CurTypesSection.front();
877       }
878 
879       DWARFUnitIndex TUIndex(TUSectionKind);
880       DataExtractor TUIndexData(CurTUIndexSection, Obj.isLittleEndian(), 0);
881       if (!TUIndex.parse(TUIndexData))
882         return make_error<DWPError>("failed to parse tu_index");
883       if (TUIndex.getVersion() != IndexVersion)
884         return make_error<DWPError>("incompatible tu_index versions, found " +
885                                     utostr(TUIndex.getVersion()) +
886                                     " and expecting " + utostr(IndexVersion));
887 
888       unsigned TypesContributionIndex =
889           getContributionIndex(TUSectionKind, IndexVersion);
890       if (Error Err = addAllTypesFromDWP(
891               Out, TypeIndexEntries, TUIndex, OutSection, TypeInputSection,
892               CurEntry, ContributionOffsets[TypesContributionIndex],
893               TypesContributionIndex, OverflowOptValue, AnySectionOverflow))
894         return Err;
895     }
896     if (AnySectionOverflow)
897       break;
898   }
899 
900   if (Version < 5) {
901     // Lie about there being no info contributions so the TU index only includes
902     // the type unit contribution for DWARF < 5. In DWARFv5 the TU index has a
903     // contribution to the info section, so we do not want to lie about it.
904     ContributionOffsets[0] = 0;
905   }
906   writeIndex(Out, MCOFI.getDwarfTUIndexSection(), ContributionOffsets,
907              TypeIndexEntries, IndexVersion);
908 
909   if (Version < 5) {
910     // Lie about the type contribution for DWARF < 5. In DWARFv5 the type
911     // section does not exist, so no need to do anything about this.
912     ContributionOffsets[getContributionIndex(DW_SECT_EXT_TYPES, 2)] = 0;
913     // Unlie about the info contribution
914     ContributionOffsets[0] = 1;
915   }
916 
917   writeIndex(Out, MCOFI.getDwarfCUIndexSection(), ContributionOffsets,
918              IndexEntries, IndexVersion);
919 
920   return Error::success();
921 }
922 } // namespace llvm
923