include/llvm/DebugInfo/DWARF/DWARFUnit.h

   1 //===- DWARFUnit.h ----------------------------------------------*- C++ -*-===//
   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 #ifndef LLVM_DEBUGINFO_DWARF_DWARFUNIT_H
  10 #define LLVM_DEBUGINFO_DWARF_DWARFUNIT_H
  11
  12 #include "llvm/ADT/Optional.h"
  13 #include "llvm/ADT/STLExtras.h"
  14 #include "llvm/ADT/SmallVector.h"
  15 #include "llvm/ADT/StringRef.h"
  16 #include "llvm/ADT/iterator_range.h"
  17 #include "llvm/BinaryFormat/Dwarf.h"
  18 #include "llvm/DebugInfo/DWARF/DWARFDebugInfoEntry.h"
  19 #include "llvm/DebugInfo/DWARF/DWARFDebugRangeList.h"
  20 #include "llvm/DebugInfo/DWARF/DWARFDebugRnglists.h"
  21 #include "llvm/DebugInfo/DWARF/DWARFDie.h"
  22 #include "llvm/DebugInfo/DWARF/DWARFFormValue.h"
  23 #include "llvm/DebugInfo/DWARF/DWARFRelocMap.h"
  24 #include "llvm/DebugInfo/DWARF/DWARFSection.h"
  25 #include "llvm/DebugInfo/DWARF/DWARFUnitIndex.h"
  26 #include "llvm/Support/DataExtractor.h"
  27 #include <algorithm>
  28 #include <cassert>
  29 #include <cstddef>
  30 #include <cstdint>
  31 #include <map>
  32 #include <memory>
  33 #include <utility>
  34 #include <vector>
  35
  36 namespace llvm {
  37
  38 class DWARFAbbreviationDeclarationSet;
  39 class DWARFContext;
  40 class DWARFDebugAbbrev;
  41 class DWARFUnit;
  42
  43 /// Base class describing the header of any kind of "unit."  Some information
  44 /// is specific to certain unit types.  We separate this class out so we can
  45 /// parse the header before deciding what specific kind of unit to construct.
  46 class DWARFUnitHeader {
  47   // Offset within section.
  48   uint32_t Offset = 0;
  49   // Version, address size, and DWARF format.
  50   dwarf::FormParams FormParams;
  51   uint32_t Length = 0;
  52   uint64_t AbbrOffset = 0;
  53
  54   // For DWO units only.
  55   const DWARFUnitIndex::Entry *IndexEntry = nullptr;
  56
  57   // For type units only.
  58   uint64_t TypeHash = 0;
  59   uint32_t TypeOffset = 0;
  60
  61   // For v5 split or skeleton compile units only.
  62   Optional<uint64_t> DWOId;
  63
  64   // Unit type as parsed, or derived from the section kind.
  65   uint8_t UnitType = 0;
  66
  67   // Size as parsed. uint8_t for compactness.
  68   uint8_t Size = 0;
  69
  70 public:
  71   /// Parse a unit header from \p debug_info starting at \p offset_ptr.
  72   bool extract(DWARFContext &Context, const DWARFDataExtractor &debug_info,
  73                uint32_t *offset_ptr, DWARFSectionKind Kind = DW_SECT_INFO,
  74                const DWARFUnitIndex *Index = nullptr,
  75                const DWARFUnitIndex::Entry *Entry = nullptr);
  76   uint32_t getOffset() const { return Offset; }
  77   const dwarf::FormParams &getFormParams() const { return FormParams; }
  78   uint16_t getVersion() const { return FormParams.Version; }
  79   dwarf::DwarfFormat getFormat() const { return FormParams.Format; }
  80   uint8_t getAddressByteSize() const { return FormParams.AddrSize; }
  81   uint8_t getRefAddrByteSize() const { return FormParams.getRefAddrByteSize(); }
  82   uint8_t getDwarfOffsetByteSize() const {
  83     return FormParams.getDwarfOffsetByteSize();
  84   }
  85   uint32_t getLength() const { return Length; }
  86   uint64_t getAbbrOffset() const { return AbbrOffset; }
  87   Optional<uint64_t> getDWOId() const { return DWOId; }
  88   void setDWOId(uint64_t Id) {
  89     assert((!DWOId || *DWOId == Id) && "setting DWOId to a different value");
  90     DWOId = Id;
  91   }
  92   const DWARFUnitIndex::Entry *getIndexEntry() const { return IndexEntry; }
  93   uint64_t getTypeHash() const { return TypeHash; }
  94   uint32_t getTypeOffset() const { return TypeOffset; }
  95   uint8_t getUnitType() const { return UnitType; }
  96   bool isTypeUnit() const {
  97     return UnitType == dwarf::DW_UT_type || UnitType == dwarf::DW_UT_split_type;
  98   }
  99   uint8_t getSize() const { return Size; }
 100   // FIXME: Support DWARF64.
 101   uint32_t getNextUnitOffset() const { return Offset + Length + 4; }
 102 };
 103
 104 const DWARFUnitIndex &getDWARFUnitIndex(DWARFContext &Context,
 105                                         DWARFSectionKind Kind);
 106
 107 /// Describe a collection of units. Intended to hold all units either from
 108 /// .debug_info and .debug_types, or from .debug_info.dwo and .debug_types.dwo.
 109 class DWARFUnitVector final : public SmallVector<std::unique_ptr<DWARFUnit>, 1> {
 110   std::function<std::unique_ptr<DWARFUnit>(uint32_t, DWARFSectionKind,
 111                                            const DWARFSection *,
 112                                            const DWARFUnitIndex::Entry *)>
 113       Parser;
 114   int NumInfoUnits = -1;
 115
 116 public:
 117   using UnitVector = SmallVectorImpl<std::unique_ptr<DWARFUnit>>;
 118   using iterator = typename UnitVector::iterator;
 119   using iterator_range = llvm::iterator_range<typename UnitVector::iterator>;
 120
 121   DWARFUnit *getUnitForOffset(uint32_t Offset) const;
 122   DWARFUnit *getUnitForIndexEntry(const DWARFUnitIndex::Entry &E);
 123
 124   /// Read units from a .debug_info or .debug_types section.  Calls made
 125   /// before finishedInfoUnits() are assumed to be for .debug_info sections,
 126   /// calls after finishedInfoUnits() are for .debug_types sections.  Caller
 127   /// must not mix calls to addUnitsForSection and addUnitsForDWOSection.
 128   void addUnitsForSection(DWARFContext &C, const DWARFSection &Section,
 129                           DWARFSectionKind SectionKind);
 130   /// Read units from a .debug_info.dwo or .debug_types.dwo section.  Calls
 131   /// made before finishedInfoUnits() are assumed to be for .debug_info.dwo
 132   /// sections, calls after finishedInfoUnits() are for .debug_types.dwo
 133   /// sections.  Caller must not mix calls to addUnitsForSection and
 134   /// addUnitsForDWOSection.
 135   void addUnitsForDWOSection(DWARFContext &C, const DWARFSection &DWOSection,
 136                              DWARFSectionKind SectionKind, bool Lazy = false);
 137
 138   /// Add an existing DWARFUnit to this UnitVector. This is used by the DWARF
 139   /// verifier to process unit separately.
 140   DWARFUnit *addUnit(std::unique_ptr<DWARFUnit> Unit);
 141
 142   /// Returns number of all units held by this instance.
 143   unsigned getNumUnits() const { return size(); }
 144   /// Returns number of units from all .debug_info[.dwo] sections.
 145   unsigned getNumInfoUnits() const {
 146     return NumInfoUnits == -1 ? size() : NumInfoUnits;
 147   }
 148   /// Returns number of units from all .debug_types[.dwo] sections.
 149   unsigned getNumTypesUnits() const { return size() - NumInfoUnits; }
 150   /// Indicate that parsing .debug_info[.dwo] is done, and remaining units
 151   /// will be from .debug_types[.dwo].
 152   void finishedInfoUnits() { NumInfoUnits = size(); }
 153
 154 private:
 155   void addUnitsImpl(DWARFContext &Context, const DWARFObject &Obj,
 156                     const DWARFSection &Section, const DWARFDebugAbbrev *DA,
 157                     const DWARFSection *RS, const DWARFSection *LocSection,
 158                     StringRef SS, const DWARFSection &SOS,
 159                     const DWARFSection *AOS, const DWARFSection &LS, bool LE,
 160                     bool IsDWO, bool Lazy, DWARFSectionKind SectionKind);
 161 };
 162
 163 /// Represents base address of the CU.
 164 /// Represents a unit's contribution to the string offsets table.
 165 struct StrOffsetsContributionDescriptor {
 166   uint64_t Base = 0;
 167   /// The contribution size not including the header.
 168   uint64_t Size = 0;
 169   /// Format and version.
 170   dwarf::FormParams FormParams = {0, 0, dwarf::DwarfFormat::DWARF32};
 171
 172   StrOffsetsContributionDescriptor(uint64_t Base, uint64_t Size,
 173                                    uint8_t Version, dwarf::DwarfFormat Format)
 174       : Base(Base), Size(Size), FormParams({Version, 0, Format}) {}
 175
 176   uint8_t getVersion() const { return FormParams.Version; }
 177   dwarf::DwarfFormat getFormat() const { return FormParams.Format; }
 178   uint8_t getDwarfOffsetByteSize() const {
 179     return FormParams.getDwarfOffsetByteSize();
 180   }
 181   /// Determine whether a contribution to the string offsets table is
 182   /// consistent with the relevant section size and that its length is
 183   /// a multiple of the size of one of its entries.
 184   Optional<StrOffsetsContributionDescriptor>
 185   validateContributionSize(DWARFDataExtractor &DA);
 186 };
 187
 188 class DWARFUnit {
 189   DWARFContext &Context;
 190   /// Section containing this DWARFUnit.
 191   const DWARFSection &InfoSection;
 192
 193   DWARFUnitHeader Header;
 194   const DWARFDebugAbbrev *Abbrev;
 195   const DWARFSection *RangeSection;
 196   uint32_t RangeSectionBase;
 197   /// We either keep track of the location list section or its data, depending
 198   /// on whether we are handling a split DWARF section or not.
 199   union {
 200     const DWARFSection *LocSection;
 201     StringRef LocSectionData;
 202   };
 203   const DWARFSection &LineSection;
 204   StringRef StringSection;
 205   const DWARFSection &StringOffsetSection;
 206   const DWARFSection *AddrOffsetSection;
 207   uint32_t AddrOffsetSectionBase = 0;
 208   bool isLittleEndian;
 209   bool IsDWO;
 210   const DWARFUnitVector &UnitVector;
 211
 212   /// Start, length, and DWARF format of the unit's contribution to the string
 213   /// offsets table (DWARF v5).
 214   Optional<StrOffsetsContributionDescriptor> StringOffsetsTableContribution;
 215
 216   /// A table of range lists (DWARF v5 and later).
 217   Optional<DWARFDebugRnglistTable> RngListTable;
 218
 219   mutable const DWARFAbbreviationDeclarationSet *Abbrevs;
 220   llvm::Optional<SectionedAddress> BaseAddr;
 221   /// The compile unit debug information entry items.
 222   std::vector<DWARFDebugInfoEntry> DieArray;
 223
 224   /// Map from range's start address to end address and corresponding DIE.
 225   /// IntervalMap does not support range removal, as a result, we use the
 226   /// std::map::upper_bound for address range lookup.
 227   std::map<uint64_t, std::pair<uint64_t, DWARFDie>> AddrDieMap;
 228
 229   using die_iterator_range =
 230       iterator_range<std::vector<DWARFDebugInfoEntry>::iterator>;
 231
 232   std::shared_ptr<DWARFUnit> DWO;
 233
 234   uint32_t getDIEIndex(const DWARFDebugInfoEntry *Die) {
 235     auto First = DieArray.data();
 236     assert(Die >= First && Die < First + DieArray.size());
 237     return Die - First;
 238   }
 239
 240 protected:
 241   const DWARFUnitHeader &getHeader() const { return Header; }
 242
 243   /// Size in bytes of the parsed unit header.
 244   uint32_t getHeaderSize() const { return Header.getSize(); }
 245
 246   /// Find the unit's contribution to the string offsets table and determine its
 247   /// length and form. The given offset is expected to be derived from the unit
 248   /// DIE's DW_AT_str_offsets_base attribute.
 249   Optional<StrOffsetsContributionDescriptor>
 250   determineStringOffsetsTableContribution(DWARFDataExtractor &DA);
 251
 252   /// Find the unit's contribution to the string offsets table and determine its
 253   /// length and form. The given offset is expected to be 0 in a dwo file or,
 254   /// in a dwp file, the start of the unit's contribution to the string offsets
 255   /// table section (as determined by the index table).
 256   Optional<StrOffsetsContributionDescriptor>
 257   determineStringOffsetsTableContributionDWO(DWARFDataExtractor &DA);
 258
 259 public:
 260   DWARFUnit(DWARFContext &Context, const DWARFSection &Section,
 261             const DWARFUnitHeader &Header, const DWARFDebugAbbrev *DA,
 262             const DWARFSection *RS, const DWARFSection *LocSection,
 263             StringRef SS, const DWARFSection &SOS, const DWARFSection *AOS,
 264             const DWARFSection &LS, bool LE, bool IsDWO,
 265             const DWARFUnitVector &UnitVector);
 266
 267   virtual ~DWARFUnit();
 268
 269   bool isDWOUnit() const { return IsDWO; }
 270   DWARFContext& getContext() const { return Context; }
 271   const DWARFSection &getInfoSection() const { return InfoSection; }
 272   const DWARFSection *getLocSection() const { return LocSection; }
 273   StringRef getLocSectionData() const { return LocSectionData; }
 274   uint32_t getOffset() const { return Header.getOffset(); }
 275   const dwarf::FormParams &getFormParams() const {
 276     return Header.getFormParams();
 277   }
 278   uint16_t getVersion() const { return Header.getVersion(); }
 279   uint8_t getAddressByteSize() const { return Header.getAddressByteSize(); }
 280   uint8_t getRefAddrByteSize() const { return Header.getRefAddrByteSize(); }
 281   uint8_t getDwarfOffsetByteSize() const {
 282     return Header.getDwarfOffsetByteSize();
 283   }
 284   uint32_t getLength() const { return Header.getLength(); }
 285   uint8_t getUnitType() const { return Header.getUnitType(); }
 286   bool isTypeUnit() const { return Header.isTypeUnit(); }
 287   uint32_t getNextUnitOffset() const { return Header.getNextUnitOffset(); }
 288   const DWARFSection &getLineSection() const { return LineSection; }
 289   StringRef getStringSection() const { return StringSection; }
 290   const DWARFSection &getStringOffsetSection() const {
 291     return StringOffsetSection;
 292   }
 293
 294   void setAddrOffsetSection(const DWARFSection *AOS, uint32_t Base) {
 295     AddrOffsetSection = AOS;
 296     AddrOffsetSectionBase = Base;
 297   }
 298
 299   /// Recursively update address to Die map.
 300   void updateAddressDieMap(DWARFDie Die);
 301
 302   void setRangesSection(const DWARFSection *RS, uint32_t Base) {
 303     RangeSection = RS;
 304     RangeSectionBase = Base;
 305   }
 306
 307   Optional<SectionedAddress> getAddrOffsetSectionItem(uint32_t Index) const;
 308   Optional<uint64_t> getStringOffsetSectionItem(uint32_t Index) const;
 309
 310   DWARFDataExtractor getDebugInfoExtractor() const;
 311
 312   DataExtractor getStringExtractor() const {
 313     return DataExtractor(StringSection, false, 0);
 314   }
 315
 316   /// Extract the range list referenced by this compile unit from the
 317   /// .debug_ranges section. If the extraction is unsuccessful, an error
 318   /// is returned. Successful extraction requires that the compile unit
 319   /// has already been extracted.
 320   Error extractRangeList(uint32_t RangeListOffset,
 321                          DWARFDebugRangeList &RangeList) const;
 322   void clear();
 323
 324   const Optional<StrOffsetsContributionDescriptor> &
 325   getStringOffsetsTableContribution() const {
 326     return StringOffsetsTableContribution;
 327   }
 328
 329   uint8_t getDwarfStringOffsetsByteSize() const {
 330     assert(StringOffsetsTableContribution);
 331     return StringOffsetsTableContribution->getDwarfOffsetByteSize();
 332   }
 333
 334   uint64_t getStringOffsetsBase() const {
 335     assert(StringOffsetsTableContribution);
 336     return StringOffsetsTableContribution->Base;
 337   }
 338
 339   const DWARFAbbreviationDeclarationSet *getAbbreviations() const;
 340
 341   static bool isMatchingUnitTypeAndTag(uint8_t UnitType, dwarf::Tag Tag) {
 342     switch (UnitType) {
 343     case dwarf::DW_UT_compile:
 344       return Tag == dwarf::DW_TAG_compile_unit;
 345     case dwarf::DW_UT_type:
 346       return Tag == dwarf::DW_TAG_type_unit;
 347     case dwarf::DW_UT_partial:
 348       return Tag == dwarf::DW_TAG_partial_unit;
 349     case dwarf::DW_UT_skeleton:
 350       return Tag == dwarf::DW_TAG_skeleton_unit;
 351     case dwarf::DW_UT_split_compile:
 352     case dwarf::DW_UT_split_type:
 353       return dwarf::isUnitType(Tag);
 354     }
 355     return false;
 356   }
 357
 358   /// Return the number of bytes for the header of a unit of
 359   /// UnitType type.
 360   ///
 361   /// This function must be called with a valid unit type which in
 362   /// DWARF5 is defined as one of the following six types.
 363   static uint32_t getDWARF5HeaderSize(uint8_t UnitType) {
 364     switch (UnitType) {
 365     case dwarf::DW_UT_compile:
 366     case dwarf::DW_UT_partial:
 367       return 12;
 368     case dwarf::DW_UT_skeleton:
 369     case dwarf::DW_UT_split_compile:
 370       return 20;
 371     case dwarf::DW_UT_type:
 372     case dwarf::DW_UT_split_type:
 373       return 24;
 374     }
 375     llvm_unreachable("Invalid UnitType.");
 376   }
 377
 378   llvm::Optional<SectionedAddress> getBaseAddress();
 379
 380   DWARFDie getUnitDIE(bool ExtractUnitDIEOnly = true) {
 381     extractDIEsIfNeeded(ExtractUnitDIEOnly);
 382     if (DieArray.empty())
 383       return DWARFDie();
 384     return DWARFDie(this, &DieArray[0]);
 385   }
 386
 387   const char *getCompilationDir();
 388   Optional<uint64_t> getDWOId() {
 389     extractDIEsIfNeeded(/*CUDieOnly*/ true);
 390     return getHeader().getDWOId();
 391   }
 392   void setDWOId(uint64_t NewID) { Header.setDWOId(NewID); }
 393
 394   /// Return a vector of address ranges resulting from a (possibly encoded)
 395   /// range list starting at a given offset in the appropriate ranges section.
 396   Expected<DWARFAddressRangesVector> findRnglistFromOffset(uint32_t Offset);
 397
 398   /// Return a vector of address ranges retrieved from an encoded range
 399   /// list whose offset is found via a table lookup given an index (DWARF v5
 400   /// and later).
 401   Expected<DWARFAddressRangesVector> findRnglistFromIndex(uint32_t Index);
 402
 403   /// Return a rangelist's offset based on an index. The index designates
 404   /// an entry in the rangelist table's offset array and is supplied by
 405   /// DW_FORM_rnglistx.
 406   Optional<uint32_t> getRnglistOffset(uint32_t Index) {
 407     if (RngListTable)
 408       return RngListTable->getOffsetEntry(Index);
 409     return None;
 410   }
 411
 412   Expected<DWARFAddressRangesVector> collectAddressRanges();
 413
 414   /// Returns subprogram DIE with address range encompassing the provided
 415   /// address. The pointer is alive as long as parsed compile unit DIEs are not
 416   /// cleared.
 417   DWARFDie getSubroutineForAddress(uint64_t Address);
 418
 419   /// getInlinedChainForAddress - fetches inlined chain for a given address.
 420   /// Returns empty chain if there is no subprogram containing address. The
 421   /// chain is valid as long as parsed compile unit DIEs are not cleared.
 422   void getInlinedChainForAddress(uint64_t Address,
 423                                  SmallVectorImpl<DWARFDie> &InlinedChain);
 424
 425   /// Return the DWARFUnitVector containing this unit.
 426   const DWARFUnitVector &getUnitVector() const { return UnitVector; }
 427
 428   /// Returns the number of DIEs in the unit. Parses the unit
 429   /// if necessary.
 430   unsigned getNumDIEs() {
 431     extractDIEsIfNeeded(false);
 432     return DieArray.size();
 433   }
 434
 435   /// Return the index of a DIE inside the unit's DIE vector.
 436   ///
 437   /// It is illegal to call this method with a DIE that hasn't be
 438   /// created by this unit. In other word, it's illegal to call this
 439   /// method on a DIE that isn't accessible by following
 440   /// children/sibling links starting from this unit's getUnitDIE().
 441   uint32_t getDIEIndex(const DWARFDie &D) {
 442     return getDIEIndex(D.getDebugInfoEntry());
 443   }
 444
 445   /// Return the DIE object at the given index.
 446   DWARFDie getDIEAtIndex(unsigned Index) {
 447     assert(Index < DieArray.size());
 448     return DWARFDie(this, &DieArray[Index]);
 449   }
 450
 451   DWARFDie getParent(const DWARFDebugInfoEntry *Die);
 452   DWARFDie getSibling(const DWARFDebugInfoEntry *Die);
 453   DWARFDie getPreviousSibling(const DWARFDebugInfoEntry *Die);
 454   DWARFDie getFirstChild(const DWARFDebugInfoEntry *Die);
 455   DWARFDie getLastChild(const DWARFDebugInfoEntry *Die);
 456
 457   /// Return the DIE object for a given offset inside the
 458   /// unit's DIE vector.
 459   ///
 460   /// The unit needs to have its DIEs extracted for this method to work.
 461   DWARFDie getDIEForOffset(uint32_t Offset) {
 462     extractDIEsIfNeeded(false);
 463     assert(!DieArray.empty());
 464     auto it = std::lower_bound(
 465         DieArray.begin(), DieArray.end(), Offset,
 466         [](const DWARFDebugInfoEntry &LHS, uint32_t Offset) {
 467           return LHS.getOffset() < Offset;
 468         });
 469     if (it != DieArray.end() && it->getOffset() == Offset)
 470       return DWARFDie(this, &*it);
 471     return DWARFDie();
 472   }
 473
 474   uint32_t getLineTableOffset() const {
 475     if (auto IndexEntry = Header.getIndexEntry())
 476       if (const auto *Contrib = IndexEntry->getOffset(DW_SECT_LINE))
 477         return Contrib->Offset;
 478     return 0;
 479   }
 480
 481   die_iterator_range dies() {
 482     extractDIEsIfNeeded(false);
 483     return die_iterator_range(DieArray.begin(), DieArray.end());
 484   }
 485
 486   virtual void dump(raw_ostream &OS, DIDumpOptions DumpOpts) = 0;
 487 private:
 488   /// Size in bytes of the .debug_info data associated with this compile unit.
 489   size_t getDebugInfoSize() const {
 490     return Header.getLength() + 4 - getHeaderSize();
 491   }
 492
 493   /// extractDIEsIfNeeded - Parses a compile unit and indexes its DIEs if it
 494   /// hasn't already been done. Returns the number of DIEs parsed at this call.
 495   size_t extractDIEsIfNeeded(bool CUDieOnly);
 496
 497   /// extractDIEsToVector - Appends all parsed DIEs to a vector.
 498   void extractDIEsToVector(bool AppendCUDie, bool AppendNonCUDIEs,
 499                            std::vector<DWARFDebugInfoEntry> &DIEs) const;
 500
 501   /// clearDIEs - Clear parsed DIEs to keep memory usage low.
 502   void clearDIEs(bool KeepCUDie);
 503
 504   /// parseDWO - Parses .dwo file for current compile unit. Returns true if
 505   /// it was actually constructed.
 506   bool parseDWO();
 507 };
 508
 509 } // end namespace llvm
 510
 511 #endif // LLVM_DEBUGINFO_DWARF_DWARFUNIT_H