llvm/lib/ObjCopy/ELF/ELFObject.h

   1 //===- ELFObject.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_LIB_OBJCOPY_ELF_ELFOBJECT_H
  10 #define LLVM_LIB_OBJCOPY_ELF_ELFOBJECT_H
  11
  12 #include "llvm/ADT/ArrayRef.h"
  13 #include "llvm/ADT/StringRef.h"
  14 #include "llvm/ADT/Twine.h"
  15 #include "llvm/BinaryFormat/ELF.h"
  16 #include "llvm/MC/StringTableBuilder.h"
  17 #include "llvm/ObjCopy/CommonConfig.h"
  18 #include "llvm/Object/ELFObjectFile.h"
  19 #include "llvm/Support/Errc.h"
  20 #include "llvm/Support/FileOutputBuffer.h"
  21 #include "llvm/Support/MemoryBuffer.h"
  22 #include <cstddef>
  23 #include <cstdint>
  24 #include <functional>
  25 #include <memory>
  26 #include <set>
  27 #include <vector>
  28
  29 namespace llvm {
  30 enum class DebugCompressionType;
  31 namespace objcopy {
  32 namespace elf {
  33
  34 class SectionBase;
  35 class Section;
  36 class OwnedDataSection;
  37 class StringTableSection;
  38 class SymbolTableSection;
  39 class RelocationSection;
  40 class DynamicRelocationSection;
  41 class GnuDebugLinkSection;
  42 class GroupSection;
  43 class SectionIndexSection;
  44 class CompressedSection;
  45 class DecompressedSection;
  46 class Segment;
  47 class Object;
  48 struct Symbol;
  49
  50 class SectionTableRef {
  51   ArrayRef<std::unique_ptr<SectionBase>> Sections;
  52
  53 public:
  54   using iterator = pointee_iterator<const std::unique_ptr<SectionBase> *>;
  55
  56   explicit SectionTableRef(ArrayRef<std::unique_ptr<SectionBase>> Secs)
  57       : Sections(Secs) {}
  58   SectionTableRef(const SectionTableRef &) = default;
  59
  60   iterator begin() const { return iterator(Sections.data()); }
  61   iterator end() const { return iterator(Sections.data() + Sections.size()); }
  62   size_t size() const { return Sections.size(); }
  63
  64   Expected<SectionBase *> getSection(uint32_t Index, Twine ErrMsg);
  65
  66   template <class T>
  67   Expected<T *> getSectionOfType(uint32_t Index, Twine IndexErrMsg,
  68                                  Twine TypeErrMsg);
  69 };
  70
  71 enum ElfType { ELFT_ELF32LE, ELFT_ELF64LE, ELFT_ELF32BE, ELFT_ELF64BE };
  72
  73 class SectionVisitor {
  74 public:
  75   virtual ~SectionVisitor() = default;
  76
  77   virtual Error visit(const Section &Sec) = 0;
  78   virtual Error visit(const OwnedDataSection &Sec) = 0;
  79   virtual Error visit(const StringTableSection &Sec) = 0;
  80   virtual Error visit(const SymbolTableSection &Sec) = 0;
  81   virtual Error visit(const RelocationSection &Sec) = 0;
  82   virtual Error visit(const DynamicRelocationSection &Sec) = 0;
  83   virtual Error visit(const GnuDebugLinkSection &Sec) = 0;
  84   virtual Error visit(const GroupSection &Sec) = 0;
  85   virtual Error visit(const SectionIndexSection &Sec) = 0;
  86   virtual Error visit(const CompressedSection &Sec) = 0;
  87   virtual Error visit(const DecompressedSection &Sec) = 0;
  88 };
  89
  90 class MutableSectionVisitor {
  91 public:
  92   virtual ~MutableSectionVisitor() = default;
  93
  94   virtual Error visit(Section &Sec) = 0;
  95   virtual Error visit(OwnedDataSection &Sec) = 0;
  96   virtual Error visit(StringTableSection &Sec) = 0;
  97   virtual Error visit(SymbolTableSection &Sec) = 0;
  98   virtual Error visit(RelocationSection &Sec) = 0;
  99   virtual Error visit(DynamicRelocationSection &Sec) = 0;
 100   virtual Error visit(GnuDebugLinkSection &Sec) = 0;
 101   virtual Error visit(GroupSection &Sec) = 0;
 102   virtual Error visit(SectionIndexSection &Sec) = 0;
 103   virtual Error visit(CompressedSection &Sec) = 0;
 104   virtual Error visit(DecompressedSection &Sec) = 0;
 105 };
 106
 107 class SectionWriter : public SectionVisitor {
 108 protected:
 109   WritableMemoryBuffer &Out;
 110
 111 public:
 112   virtual ~SectionWriter() = default;
 113
 114   Error visit(const Section &Sec) override;
 115   Error visit(const OwnedDataSection &Sec) override;
 116   Error visit(const StringTableSection &Sec) override;
 117   Error visit(const DynamicRelocationSection &Sec) override;
 118   Error visit(const SymbolTableSection &Sec) override = 0;
 119   Error visit(const RelocationSection &Sec) override = 0;
 120   Error visit(const GnuDebugLinkSection &Sec) override = 0;
 121   Error visit(const GroupSection &Sec) override = 0;
 122   Error visit(const SectionIndexSection &Sec) override = 0;
 123   Error visit(const CompressedSection &Sec) override = 0;
 124   Error visit(const DecompressedSection &Sec) override = 0;
 125
 126   explicit SectionWriter(WritableMemoryBuffer &Buf) : Out(Buf) {}
 127 };
 128
 129 template <class ELFT> class ELFSectionWriter : public SectionWriter {
 130 private:
 131   using Elf_Word = typename ELFT::Word;
 132   using Elf_Rel = typename ELFT::Rel;
 133   using Elf_Rela = typename ELFT::Rela;
 134   using Elf_Sym = typename ELFT::Sym;
 135
 136 public:
 137   virtual ~ELFSectionWriter() {}
 138   Error visit(const SymbolTableSection &Sec) override;
 139   Error visit(const RelocationSection &Sec) override;
 140   Error visit(const GnuDebugLinkSection &Sec) override;
 141   Error visit(const GroupSection &Sec) override;
 142   Error visit(const SectionIndexSection &Sec) override;
 143   Error visit(const CompressedSection &Sec) override;
 144   Error visit(const DecompressedSection &Sec) override;
 145
 146   explicit ELFSectionWriter(WritableMemoryBuffer &Buf) : SectionWriter(Buf) {}
 147 };
 148
 149 template <class ELFT> class ELFSectionSizer : public MutableSectionVisitor {
 150 private:
 151   using Elf_Rel = typename ELFT::Rel;
 152   using Elf_Rela = typename ELFT::Rela;
 153   using Elf_Sym = typename ELFT::Sym;
 154   using Elf_Word = typename ELFT::Word;
 155   using Elf_Xword = typename ELFT::Xword;
 156
 157 public:
 158   Error visit(Section &Sec) override;
 159   Error visit(OwnedDataSection &Sec) override;
 160   Error visit(StringTableSection &Sec) override;
 161   Error visit(DynamicRelocationSection &Sec) override;
 162   Error visit(SymbolTableSection &Sec) override;
 163   Error visit(RelocationSection &Sec) override;
 164   Error visit(GnuDebugLinkSection &Sec) override;
 165   Error visit(GroupSection &Sec) override;
 166   Error visit(SectionIndexSection &Sec) override;
 167   Error visit(CompressedSection &Sec) override;
 168   Error visit(DecompressedSection &Sec) override;
 169 };
 170
 171 #define MAKE_SEC_WRITER_FRIEND                                                 \
 172   friend class SectionWriter;                                                  \
 173   friend class IHexSectionWriterBase;                                          \
 174   friend class IHexSectionWriter;                                              \
 175   template <class ELFT> friend class ELFSectionWriter;                         \
 176   template <class ELFT> friend class ELFSectionSizer;
 177
 178 class BinarySectionWriter : public SectionWriter {
 179 public:
 180   virtual ~BinarySectionWriter() {}
 181
 182   Error visit(const SymbolTableSection &Sec) override;
 183   Error visit(const RelocationSection &Sec) override;
 184   Error visit(const GnuDebugLinkSection &Sec) override;
 185   Error visit(const GroupSection &Sec) override;
 186   Error visit(const SectionIndexSection &Sec) override;
 187   Error visit(const CompressedSection &Sec) override;
 188   Error visit(const DecompressedSection &Sec) override;
 189
 190   explicit BinarySectionWriter(WritableMemoryBuffer &Buf)
 191       : SectionWriter(Buf) {}
 192 };
 193
 194 using IHexLineData = SmallVector<char, 64>;
 195
 196 struct IHexRecord {
 197   // Memory address of the record.
 198   uint16_t Addr;
 199   // Record type (see below).
 200   uint16_t Type;
 201   // Record data in hexadecimal form.
 202   StringRef HexData;
 203
 204   // Helper method to get file length of the record
 205   // including newline character
 206   static size_t getLength(size_t DataSize) {
 207     // :LLAAAATT[DD...DD]CC'
 208     return DataSize * 2 + 11;
 209   }
 210
 211   // Gets length of line in a file (getLength + CRLF).
 212   static size_t getLineLength(size_t DataSize) {
 213     return getLength(DataSize) + 2;
 214   }
 215
 216   // Given type, address and data returns line which can
 217   // be written to output file.
 218   static IHexLineData getLine(uint8_t Type, uint16_t Addr,
 219                               ArrayRef<uint8_t> Data);
 220
 221   // Parses the line and returns record if possible.
 222   // Line should be trimmed from whitespace characters.
 223   static Expected<IHexRecord> parse(StringRef Line);
 224
 225   // Calculates checksum of stringified record representation
 226   // S must NOT contain leading ':' and trailing whitespace
 227   // characters
 228   static uint8_t getChecksum(StringRef S);
 229
 230   enum Type {
 231     // Contains data and a 16-bit starting address for the data.
 232     // The byte count specifies number of data bytes in the record.
 233     Data = 0,
 234     // Must occur exactly once per file in the last line of the file.
 235     // The data field is empty (thus byte count is 00) and the address
 236     // field is typically 0000.
 237     EndOfFile = 1,
 238     // The data field contains a 16-bit segment base address (thus byte
 239     // count is always 02) compatible with 80x86 real mode addressing.
 240     // The address field (typically 0000) is ignored. The segment address
 241     // from the most recent 02 record is multiplied by 16 and added to each
 242     // subsequent data record address to form the physical starting address
 243     // for the data. This allows addressing up to one megabyte of address
 244     // space.
 245     SegmentAddr = 2,
 246     // or 80x86 processors, specifies the initial content of the CS:IP
 247     // registers. The address field is 0000, the byte count is always 04,
 248     // the first two data bytes are the CS value, the latter two are the
 249     // IP value.
 250     StartAddr80x86 = 3,
 251     // Allows for 32 bit addressing (up to 4GiB). The record's address field
 252     // is ignored (typically 0000) and its byte count is always 02. The two
 253     // data bytes (big endian) specify the upper 16 bits of the 32 bit
 254     // absolute address for all subsequent type 00 records
 255     ExtendedAddr = 4,
 256     // The address field is 0000 (not used) and the byte count is always 04.
 257     // The four data bytes represent a 32-bit address value. In the case of
 258     // 80386 and higher CPUs, this address is loaded into the EIP register.
 259     StartAddr = 5,
 260     // We have no other valid types
 261     InvalidType = 6
 262   };
 263 };
 264
 265 // Base class for IHexSectionWriter. This class implements writing algorithm,
 266 // but doesn't actually write records. It is used for output buffer size
 267 // calculation in IHexWriter::finalize.
 268 class IHexSectionWriterBase : public BinarySectionWriter {
 269   // 20-bit segment address
 270   uint32_t SegmentAddr = 0;
 271   // Extended linear address
 272   uint32_t BaseAddr = 0;
 273
 274   // Write segment address corresponding to 'Addr'
 275   uint64_t writeSegmentAddr(uint64_t Addr);
 276   // Write extended linear (base) address corresponding to 'Addr'
 277   uint64_t writeBaseAddr(uint64_t Addr);
 278
 279 protected:
 280   // Offset in the output buffer
 281   uint64_t Offset = 0;
 282
 283   void writeSection(const SectionBase *Sec, ArrayRef<uint8_t> Data);
 284   virtual void writeData(uint8_t Type, uint16_t Addr, ArrayRef<uint8_t> Data);
 285
 286 public:
 287   explicit IHexSectionWriterBase(WritableMemoryBuffer &Buf)
 288       : BinarySectionWriter(Buf) {}
 289
 290   uint64_t getBufferOffset() const { return Offset; }
 291   Error visit(const Section &Sec) final;
 292   Error visit(const OwnedDataSection &Sec) final;
 293   Error visit(const StringTableSection &Sec) override;
 294   Error visit(const DynamicRelocationSection &Sec) final;
 295   using BinarySectionWriter::visit;
 296 };
 297
 298 // Real IHEX section writer
 299 class IHexSectionWriter : public IHexSectionWriterBase {
 300 public:
 301   IHexSectionWriter(WritableMemoryBuffer &Buf) : IHexSectionWriterBase(Buf) {}
 302
 303   void writeData(uint8_t Type, uint16_t Addr, ArrayRef<uint8_t> Data) override;
 304   Error visit(const StringTableSection &Sec) override;
 305 };
 306
 307 class Writer {
 308 protected:
 309   Object &Obj;
 310   std::unique_ptr<WritableMemoryBuffer> Buf;
 311   raw_ostream &Out;
 312
 313 public:
 314   virtual ~Writer();
 315   virtual Error finalize() = 0;
 316   virtual Error write() = 0;
 317
 318   Writer(Object &O, raw_ostream &Out) : Obj(O), Out(Out) {}
 319 };
 320
 321 template <class ELFT> class ELFWriter : public Writer {
 322 private:
 323   using Elf_Addr = typename ELFT::Addr;
 324   using Elf_Shdr = typename ELFT::Shdr;
 325   using Elf_Phdr = typename ELFT::Phdr;
 326   using Elf_Ehdr = typename ELFT::Ehdr;
 327
 328   void initEhdrSegment();
 329
 330   void writeEhdr();
 331   void writePhdr(const Segment &Seg);
 332   void writeShdr(const SectionBase &Sec);
 333
 334   void writePhdrs();
 335   void writeShdrs();
 336   Error writeSectionData();
 337   void writeSegmentData();
 338
 339   void assignOffsets();
 340
 341   std::unique_ptr<ELFSectionWriter<ELFT>> SecWriter;
 342
 343   size_t totalSize() const;
 344
 345 public:
 346   virtual ~ELFWriter() {}
 347   bool WriteSectionHeaders;
 348
 349   // For --only-keep-debug, select an alternative section/segment layout
 350   // algorithm.
 351   bool OnlyKeepDebug;
 352
 353   Error finalize() override;
 354   Error write() override;
 355   ELFWriter(Object &Obj, raw_ostream &Out, bool WSH, bool OnlyKeepDebug);
 356 };
 357
 358 class BinaryWriter : public Writer {
 359 private:
 360   std::unique_ptr<BinarySectionWriter> SecWriter;
 361
 362   uint64_t TotalSize = 0;
 363
 364 public:
 365   ~BinaryWriter() {}
 366   Error finalize() override;
 367   Error write() override;
 368   BinaryWriter(Object &Obj, raw_ostream &Out) : Writer(Obj, Out) {}
 369 };
 370
 371 class IHexWriter : public Writer {
 372   struct SectionCompare {
 373     bool operator()(const SectionBase *Lhs, const SectionBase *Rhs) const;
 374   };
 375
 376   std::set<const SectionBase *, SectionCompare> Sections;
 377   size_t TotalSize = 0;
 378
 379   Error checkSection(const SectionBase &Sec);
 380   uint64_t writeEntryPointRecord(uint8_t *Buf);
 381   uint64_t writeEndOfFileRecord(uint8_t *Buf);
 382
 383 public:
 384   ~IHexWriter() {}
 385   Error finalize() override;
 386   Error write() override;
 387   IHexWriter(Object &Obj, raw_ostream &Out) : Writer(Obj, Out) {}
 388 };
 389
 390 class SectionBase {
 391 public:
 392   std::string Name;
 393   Segment *ParentSegment = nullptr;
 394   uint64_t HeaderOffset = 0;
 395   uint32_t Index = 0;
 396
 397   uint32_t OriginalIndex = 0;
 398   uint64_t OriginalFlags = 0;
 399   uint64_t OriginalType = ELF::SHT_NULL;
 400   uint64_t OriginalOffset = std::numeric_limits<uint64_t>::max();
 401
 402   uint64_t Addr = 0;
 403   uint64_t Align = 1;
 404   uint32_t EntrySize = 0;
 405   uint64_t Flags = 0;
 406   uint64_t Info = 0;
 407   uint64_t Link = ELF::SHN_UNDEF;
 408   uint64_t NameIndex = 0;
 409   uint64_t Offset = 0;
 410   uint64_t Size = 0;
 411   uint64_t Type = ELF::SHT_NULL;
 412   ArrayRef<uint8_t> OriginalData;
 413   bool HasSymbol = false;
 414
 415   SectionBase() = default;
 416   SectionBase(const SectionBase &) = default;
 417
 418   virtual ~SectionBase() = default;
 419
 420   virtual Error initialize(SectionTableRef SecTable);
 421   virtual void finalize();
 422   // Remove references to these sections. The list of sections must be sorted.
 423   virtual Error
 424   removeSectionReferences(bool AllowBrokenLinks,
 425                           function_ref<bool(const SectionBase *)> ToRemove);
 426   virtual Error removeSymbols(function_ref<bool(const Symbol &)> ToRemove);
 427   virtual Error accept(SectionVisitor &Visitor) const = 0;
 428   virtual Error accept(MutableSectionVisitor &Visitor) = 0;
 429   virtual void markSymbols();
 430   virtual void
 431   replaceSectionReferences(const DenseMap<SectionBase *, SectionBase *> &);
 432   virtual bool hasContents() const { return false; }
 433   // Notify the section that it is subject to removal.
 434   virtual void onRemove();
 435
 436   virtual void restoreSymTabLink(SymbolTableSection &) {}
 437 };
 438
 439 class Segment {
 440 private:
 441   struct SectionCompare {
 442     bool operator()(const SectionBase *Lhs, const SectionBase *Rhs) const {
 443       // Some sections might have the same address if one of them is empty. To
 444       // fix this we can use the lexicographic ordering on ->Addr and the
 445       // original index.
 446       if (Lhs->OriginalOffset == Rhs->OriginalOffset)
 447         return Lhs->OriginalIndex < Rhs->OriginalIndex;
 448       return Lhs->OriginalOffset < Rhs->OriginalOffset;
 449     }
 450   };
 451
 452 public:
 453   uint32_t Type = 0;
 454   uint32_t Flags = 0;
 455   uint64_t Offset = 0;
 456   uint64_t VAddr = 0;
 457   uint64_t PAddr = 0;
 458   uint64_t FileSize = 0;
 459   uint64_t MemSize = 0;
 460   uint64_t Align = 0;
 461
 462   uint32_t Index = 0;
 463   uint64_t OriginalOffset = 0;
 464   Segment *ParentSegment = nullptr;
 465   ArrayRef<uint8_t> Contents;
 466   std::set<const SectionBase *, SectionCompare> Sections;
 467
 468   explicit Segment(ArrayRef<uint8_t> Data) : Contents(Data) {}
 469   Segment() = default;
 470
 471   const SectionBase *firstSection() const {
 472     if (!Sections.empty())
 473       return *Sections.begin();
 474     return nullptr;
 475   }
 476
 477   void removeSection(const SectionBase *Sec) { Sections.erase(Sec); }
 478   void addSection(const SectionBase *Sec) { Sections.insert(Sec); }
 479
 480   ArrayRef<uint8_t> getContents() const { return Contents; }
 481 };
 482
 483 class Section : public SectionBase {
 484   MAKE_SEC_WRITER_FRIEND
 485
 486   ArrayRef<uint8_t> Contents;
 487   SectionBase *LinkSection = nullptr;
 488   bool HasSymTabLink = false;
 489
 490 public:
 491   explicit Section(ArrayRef<uint8_t> Data) : Contents(Data) {}
 492
 493   Error accept(SectionVisitor &Visitor) const override;
 494   Error accept(MutableSectionVisitor &Visitor) override;
 495   Error removeSectionReferences(
 496       bool AllowBrokenLinks,
 497       function_ref<bool(const SectionBase *)> ToRemove) override;
 498   Error initialize(SectionTableRef SecTable) override;
 499   void finalize() override;
 500   bool hasContents() const override {
 501     return Type != ELF::SHT_NOBITS && Type != ELF::SHT_NULL;
 502   }
 503   void restoreSymTabLink(SymbolTableSection &SymTab) override;
 504 };
 505
 506 class OwnedDataSection : public SectionBase {
 507   MAKE_SEC_WRITER_FRIEND
 508
 509   std::vector<uint8_t> Data;
 510
 511 public:
 512   OwnedDataSection(StringRef SecName, ArrayRef<uint8_t> Data)
 513       : Data(std::begin(Data), std::end(Data)) {
 514     Name = SecName.str();
 515     Type = OriginalType = ELF::SHT_PROGBITS;
 516     Size = Data.size();
 517     OriginalOffset = std::numeric_limits<uint64_t>::max();
 518   }
 519
 520   OwnedDataSection(const Twine &SecName, uint64_t SecAddr, uint64_t SecFlags,
 521                    uint64_t SecOff) {
 522     Name = SecName.str();
 523     Type = OriginalType = ELF::SHT_PROGBITS;
 524     Addr = SecAddr;
 525     Flags = OriginalFlags = SecFlags;
 526     OriginalOffset = SecOff;
 527   }
 528
 529   OwnedDataSection(SectionBase &S, ArrayRef<uint8_t> Data)
 530       : SectionBase(S), Data(std::begin(Data), std::end(Data)) {
 531     Size = Data.size();
 532   }
 533
 534   void appendHexData(StringRef HexData);
 535   Error accept(SectionVisitor &Sec) const override;
 536   Error accept(MutableSectionVisitor &Visitor) override;
 537   bool hasContents() const override { return true; }
 538 };
 539
 540 class CompressedSection : public SectionBase {
 541   MAKE_SEC_WRITER_FRIEND
 542
 543   uint32_t ChType = 0;
 544   DebugCompressionType CompressionType;
 545   uint64_t DecompressedSize;
 546   uint64_t DecompressedAlign;
 547   SmallVector<uint8_t, 128> CompressedData;
 548
 549 public:
 550   CompressedSection(const SectionBase &Sec,
 551     DebugCompressionType CompressionType, bool Is64Bits);
 552   CompressedSection(ArrayRef<uint8_t> CompressedData, uint32_t ChType,
 553                     uint64_t DecompressedSize, uint64_t DecompressedAlign);
 554
 555   uint64_t getDecompressedSize() const { return DecompressedSize; }
 556   uint64_t getDecompressedAlign() const { return DecompressedAlign; }
 557   uint64_t getChType() const { return ChType; }
 558
 559   Error accept(SectionVisitor &Visitor) const override;
 560   Error accept(MutableSectionVisitor &Visitor) override;
 561
 562   static bool classof(const SectionBase *S) {
 563     return S->OriginalFlags & ELF::SHF_COMPRESSED;
 564   }
 565 };
 566
 567 class DecompressedSection : public SectionBase {
 568   MAKE_SEC_WRITER_FRIEND
 569
 570 public:
 571   uint32_t ChType;
 572   explicit DecompressedSection(const CompressedSection &Sec)
 573       : SectionBase(Sec), ChType(Sec.getChType()) {
 574     Size = Sec.getDecompressedSize();
 575     Align = Sec.getDecompressedAlign();
 576     Flags = OriginalFlags = (Flags & ~ELF::SHF_COMPRESSED);
 577   }
 578
 579   Error accept(SectionVisitor &Visitor) const override;
 580   Error accept(MutableSectionVisitor &Visitor) override;
 581 };
 582
 583 // There are two types of string tables that can exist, dynamic and not dynamic.
 584 // In the dynamic case the string table is allocated. Changing a dynamic string
 585 // table would mean altering virtual addresses and thus the memory image. So
 586 // dynamic string tables should not have an interface to modify them or
 587 // reconstruct them. This type lets us reconstruct a string table. To avoid
 588 // this class being used for dynamic string tables (which has happened) the
 589 // classof method checks that the particular instance is not allocated. This
 590 // then agrees with the makeSection method used to construct most sections.
 591 class StringTableSection : public SectionBase {
 592   MAKE_SEC_WRITER_FRIEND
 593
 594   StringTableBuilder StrTabBuilder;
 595
 596 public:
 597   StringTableSection() : StrTabBuilder(StringTableBuilder::ELF) {
 598     Type = OriginalType = ELF::SHT_STRTAB;
 599   }
 600
 601   void addString(StringRef Name);
 602   uint32_t findIndex(StringRef Name) const;
 603   void prepareForLayout();
 604   Error accept(SectionVisitor &Visitor) const override;
 605   Error accept(MutableSectionVisitor &Visitor) override;
 606
 607   static bool classof(const SectionBase *S) {
 608     if (S->OriginalFlags & ELF::SHF_ALLOC)
 609       return false;
 610     return S->OriginalType == ELF::SHT_STRTAB;
 611   }
 612 };
 613
 614 // Symbols have a st_shndx field that normally stores an index but occasionally
 615 // stores a different special value. This enum keeps track of what the st_shndx
 616 // field means. Most of the values are just copies of the special SHN_* values.
 617 // SYMBOL_SIMPLE_INDEX means that the st_shndx is just an index of a section.
 618 enum SymbolShndxType {
 619   SYMBOL_SIMPLE_INDEX = 0,
 620   SYMBOL_ABS = ELF::SHN_ABS,
 621   SYMBOL_COMMON = ELF::SHN_COMMON,
 622   SYMBOL_LOPROC = ELF::SHN_LOPROC,
 623   SYMBOL_AMDGPU_LDS = ELF::SHN_AMDGPU_LDS,
 624   SYMBOL_HEXAGON_SCOMMON = ELF::SHN_HEXAGON_SCOMMON,
 625   SYMBOL_HEXAGON_SCOMMON_2 = ELF::SHN_HEXAGON_SCOMMON_2,
 626   SYMBOL_HEXAGON_SCOMMON_4 = ELF::SHN_HEXAGON_SCOMMON_4,
 627   SYMBOL_HEXAGON_SCOMMON_8 = ELF::SHN_HEXAGON_SCOMMON_8,
 628   SYMBOL_MIPS_ACOMMON = ELF::SHN_MIPS_ACOMMON,
 629   SYMBOL_MIPS_TEXT = ELF::SHN_MIPS_TEXT,
 630   SYMBOL_MIPS_DATA = ELF::SHN_MIPS_DATA,
 631   SYMBOL_MIPS_SCOMMON = ELF::SHN_MIPS_SCOMMON,
 632   SYMBOL_MIPS_SUNDEFINED = ELF::SHN_MIPS_SUNDEFINED,
 633   SYMBOL_HIPROC = ELF::SHN_HIPROC,
 634   SYMBOL_LOOS = ELF::SHN_LOOS,
 635   SYMBOL_HIOS = ELF::SHN_HIOS,
 636   SYMBOL_XINDEX = ELF::SHN_XINDEX,
 637 };
 638
 639 struct Symbol {
 640   uint8_t Binding;
 641   SectionBase *DefinedIn = nullptr;
 642   SymbolShndxType ShndxType;
 643   uint32_t Index;
 644   std::string Name;
 645   uint32_t NameIndex;
 646   uint64_t Size;
 647   uint8_t Type;
 648   uint64_t Value;
 649   uint8_t Visibility;
 650   bool Referenced = false;
 651
 652   uint16_t getShndx() const;
 653   bool isCommon() const;
 654 };
 655
 656 class SectionIndexSection : public SectionBase {
 657   MAKE_SEC_WRITER_FRIEND
 658
 659 private:
 660   std::vector<uint32_t> Indexes;
 661   SymbolTableSection *Symbols = nullptr;
 662
 663 public:
 664   virtual ~SectionIndexSection() {}
 665   void addIndex(uint32_t Index) {
 666     assert(Size > 0);
 667     Indexes.push_back(Index);
 668   }
 669
 670   void reserve(size_t NumSymbols) {
 671     Indexes.reserve(NumSymbols);
 672     Size = NumSymbols * 4;
 673   }
 674   void setSymTab(SymbolTableSection *SymTab) { Symbols = SymTab; }
 675   Error initialize(SectionTableRef SecTable) override;
 676   void finalize() override;
 677   Error accept(SectionVisitor &Visitor) const override;
 678   Error accept(MutableSectionVisitor &Visitor) override;
 679
 680   SectionIndexSection() {
 681     Name = ".symtab_shndx";
 682     Align = 4;
 683     EntrySize = 4;
 684     Type = OriginalType = ELF::SHT_SYMTAB_SHNDX;
 685   }
 686 };
 687
 688 class SymbolTableSection : public SectionBase {
 689   MAKE_SEC_WRITER_FRIEND
 690
 691   void setStrTab(StringTableSection *StrTab) { SymbolNames = StrTab; }
 692   void assignIndices();
 693
 694 protected:
 695   std::vector<std::unique_ptr<Symbol>> Symbols;
 696   StringTableSection *SymbolNames = nullptr;
 697   SectionIndexSection *SectionIndexTable = nullptr;
 698   bool IndicesChanged = false;
 699
 700   using SymPtr = std::unique_ptr<Symbol>;
 701
 702 public:
 703   SymbolTableSection() { Type = OriginalType = ELF::SHT_SYMTAB; }
 704
 705   void addSymbol(Twine Name, uint8_t Bind, uint8_t Type, SectionBase *DefinedIn,
 706                  uint64_t Value, uint8_t Visibility, uint16_t Shndx,
 707                  uint64_t SymbolSize);
 708   void prepareForLayout();
 709   // An 'empty' symbol table still contains a null symbol.
 710   bool empty() const { return Symbols.size() == 1; }
 711   bool indicesChanged() const { return IndicesChanged; }
 712   void setShndxTable(SectionIndexSection *ShndxTable) {
 713     SectionIndexTable = ShndxTable;
 714   }
 715   const SectionIndexSection *getShndxTable() const { return SectionIndexTable; }
 716   void fillShndxTable();
 717   const SectionBase *getStrTab() const { return SymbolNames; }
 718   Expected<const Symbol *> getSymbolByIndex(uint32_t Index) const;
 719   Expected<Symbol *> getSymbolByIndex(uint32_t Index);
 720   void updateSymbols(function_ref<void(Symbol &)> Callable);
 721
 722   Error removeSectionReferences(
 723       bool AllowBrokenLinks,
 724       function_ref<bool(const SectionBase *)> ToRemove) override;
 725   Error initialize(SectionTableRef SecTable) override;
 726   void finalize() override;
 727   Error accept(SectionVisitor &Visitor) const override;
 728   Error accept(MutableSectionVisitor &Visitor) override;
 729   Error removeSymbols(function_ref<bool(const Symbol &)> ToRemove) override;
 730   void replaceSectionReferences(
 731       const DenseMap<SectionBase *, SectionBase *> &FromTo) override;
 732
 733   static bool classof(const SectionBase *S) {
 734     return S->OriginalType == ELF::SHT_SYMTAB;
 735   }
 736 };
 737
 738 struct Relocation {
 739   Symbol *RelocSymbol = nullptr;
 740   uint64_t Offset;
 741   uint64_t Addend;
 742   uint32_t Type;
 743 };
 744
 745 // All relocation sections denote relocations to apply to another section.
 746 // However, some relocation sections use a dynamic symbol table and others use
 747 // a regular symbol table. Because the types of the two symbol tables differ in
 748 // our system (because they should behave differently) we can't uniformly
 749 // represent all relocations with the same base class if we expose an interface
 750 // that mentions the symbol table type. So we split the two base types into two
 751 // different classes, one which handles the section the relocation is applied to
 752 // and another which handles the symbol table type. The symbol table type is
 753 // taken as a type parameter to the class (see RelocSectionWithSymtabBase).
 754 class RelocationSectionBase : public SectionBase {
 755 protected:
 756   SectionBase *SecToApplyRel = nullptr;
 757
 758 public:
 759   const SectionBase *getSection() const { return SecToApplyRel; }
 760   void setSection(SectionBase *Sec) { SecToApplyRel = Sec; }
 761
 762   StringRef getNamePrefix() const;
 763
 764   static bool classof(const SectionBase *S) {
 765     return S->OriginalType == ELF::SHT_REL || S->OriginalType == ELF::SHT_RELA;
 766   }
 767 };
 768
 769 // Takes the symbol table type to use as a parameter so that we can deduplicate
 770 // that code between the two symbol table types.
 771 template <class SymTabType>
 772 class RelocSectionWithSymtabBase : public RelocationSectionBase {
 773   void setSymTab(SymTabType *SymTab) { Symbols = SymTab; }
 774
 775 protected:
 776   RelocSectionWithSymtabBase() = default;
 777
 778   SymTabType *Symbols = nullptr;
 779
 780 public:
 781   Error initialize(SectionTableRef SecTable) override;
 782   void finalize() override;
 783 };
 784
 785 class RelocationSection
 786     : public RelocSectionWithSymtabBase<SymbolTableSection> {
 787   MAKE_SEC_WRITER_FRIEND
 788
 789   std::vector<Relocation> Relocations;
 790   const Object &Obj;
 791
 792 public:
 793   RelocationSection(const Object &O) : Obj(O) {}
 794   void addRelocation(Relocation Rel) { Relocations.push_back(Rel); }
 795   Error accept(SectionVisitor &Visitor) const override;
 796   Error accept(MutableSectionVisitor &Visitor) override;
 797   Error removeSectionReferences(
 798       bool AllowBrokenLinks,
 799       function_ref<bool(const SectionBase *)> ToRemove) override;
 800   Error removeSymbols(function_ref<bool(const Symbol &)> ToRemove) override;
 801   void markSymbols() override;
 802   void replaceSectionReferences(
 803       const DenseMap<SectionBase *, SectionBase *> &FromTo) override;
 804   const Object &getObject() const { return Obj; }
 805
 806   static bool classof(const SectionBase *S) {
 807     if (S->OriginalFlags & ELF::SHF_ALLOC)
 808       return false;
 809     return S->OriginalType == ELF::SHT_REL || S->OriginalType == ELF::SHT_RELA;
 810   }
 811 };
 812
 813 // TODO: The way stripping and groups interact is complicated
 814 // and still needs to be worked on.
 815
 816 class GroupSection : public SectionBase {
 817   MAKE_SEC_WRITER_FRIEND
 818   const SymbolTableSection *SymTab = nullptr;
 819   Symbol *Sym = nullptr;
 820   ELF::Elf32_Word FlagWord;
 821   SmallVector<SectionBase *, 3> GroupMembers;
 822
 823 public:
 824   // TODO: Contents is present in several classes of the hierarchy.
 825   // This needs to be refactored to avoid duplication.
 826   ArrayRef<uint8_t> Contents;
 827
 828   explicit GroupSection(ArrayRef<uint8_t> Data) : Contents(Data) {}
 829
 830   void setSymTab(const SymbolTableSection *SymTabSec) { SymTab = SymTabSec; }
 831   void setSymbol(Symbol *S) { Sym = S; }
 832   void setFlagWord(ELF::Elf32_Word W) { FlagWord = W; }
 833   void addMember(SectionBase *Sec) { GroupMembers.push_back(Sec); }
 834
 835   Error accept(SectionVisitor &) const override;
 836   Error accept(MutableSectionVisitor &Visitor) override;
 837   void finalize() override;
 838   Error removeSectionReferences(
 839       bool AllowBrokenLinks,
 840       function_ref<bool(const SectionBase *)> ToRemove) override;
 841   Error removeSymbols(function_ref<bool(const Symbol &)> ToRemove) override;
 842   void markSymbols() override;
 843   void replaceSectionReferences(
 844       const DenseMap<SectionBase *, SectionBase *> &FromTo) override;
 845   void onRemove() override;
 846
 847   static bool classof(const SectionBase *S) {
 848     return S->OriginalType == ELF::SHT_GROUP;
 849   }
 850 };
 851
 852 class DynamicSymbolTableSection : public Section {
 853 public:
 854   explicit DynamicSymbolTableSection(ArrayRef<uint8_t> Data) : Section(Data) {}
 855
 856   static bool classof(const SectionBase *S) {
 857     return S->OriginalType == ELF::SHT_DYNSYM;
 858   }
 859 };
 860
 861 class DynamicSection : public Section {
 862 public:
 863   explicit DynamicSection(ArrayRef<uint8_t> Data) : Section(Data) {}
 864
 865   static bool classof(const SectionBase *S) {
 866     return S->OriginalType == ELF::SHT_DYNAMIC;
 867   }
 868 };
 869
 870 class DynamicRelocationSection
 871     : public RelocSectionWithSymtabBase<DynamicSymbolTableSection> {
 872   MAKE_SEC_WRITER_FRIEND
 873
 874 private:
 875   ArrayRef<uint8_t> Contents;
 876
 877 public:
 878   explicit DynamicRelocationSection(ArrayRef<uint8_t> Data) : Contents(Data) {}
 879
 880   Error accept(SectionVisitor &) const override;
 881   Error accept(MutableSectionVisitor &Visitor) override;
 882   Error removeSectionReferences(
 883       bool AllowBrokenLinks,
 884       function_ref<bool(const SectionBase *)> ToRemove) override;
 885
 886   static bool classof(const SectionBase *S) {
 887     if (!(S->OriginalFlags & ELF::SHF_ALLOC))
 888       return false;
 889     return S->OriginalType == ELF::SHT_REL || S->OriginalType == ELF::SHT_RELA;
 890   }
 891 };
 892
 893 class GnuDebugLinkSection : public SectionBase {
 894   MAKE_SEC_WRITER_FRIEND
 895
 896 private:
 897   StringRef FileName;
 898   uint32_t CRC32;
 899
 900   void init(StringRef File);
 901
 902 public:
 903   // If we add this section from an external source we can use this ctor.
 904   explicit GnuDebugLinkSection(StringRef File, uint32_t PrecomputedCRC);
 905   Error accept(SectionVisitor &Visitor) const override;
 906   Error accept(MutableSectionVisitor &Visitor) override;
 907 };
 908
 909 class Reader {
 910 public:
 911   virtual ~Reader();
 912   virtual Expected<std::unique_ptr<Object>> create(bool EnsureSymtab) const = 0;
 913 };
 914
 915 using object::Binary;
 916 using object::ELFFile;
 917 using object::ELFObjectFile;
 918 using object::OwningBinary;
 919
 920 class BasicELFBuilder {
 921 protected:
 922   std::unique_ptr<Object> Obj;
 923
 924   void initFileHeader();
 925   void initHeaderSegment();
 926   StringTableSection *addStrTab();
 927   SymbolTableSection *addSymTab(StringTableSection *StrTab);
 928   Error initSections();
 929
 930 public:
 931   BasicELFBuilder() : Obj(std::make_unique<Object>()) {}
 932 };
 933
 934 class BinaryELFBuilder : public BasicELFBuilder {
 935   MemoryBuffer *MemBuf;
 936   uint8_t NewSymbolVisibility;
 937   void addData(SymbolTableSection *SymTab);
 938
 939 public:
 940   BinaryELFBuilder(MemoryBuffer *MB, uint8_t NewSymbolVisibility)
 941       : MemBuf(MB), NewSymbolVisibility(NewSymbolVisibility) {}
 942
 943   Expected<std::unique_ptr<Object>> build();
 944 };
 945
 946 class IHexELFBuilder : public BasicELFBuilder {
 947   const std::vector<IHexRecord> &Records;
 948
 949   void addDataSections();
 950
 951 public:
 952   IHexELFBuilder(const std::vector<IHexRecord> &Records) : Records(Records) {}
 953
 954   Expected<std::unique_ptr<Object>> build();
 955 };
 956
 957 template <class ELFT> class ELFBuilder {
 958 private:
 959   using Elf_Addr = typename ELFT::Addr;
 960   using Elf_Shdr = typename ELFT::Shdr;
 961   using Elf_Word = typename ELFT::Word;
 962
 963   const ELFFile<ELFT> &ElfFile;
 964   Object &Obj;
 965   size_t EhdrOffset = 0;
 966   std::optional<StringRef> ExtractPartition;
 967
 968   void setParentSegment(Segment &Child);
 969   Error readProgramHeaders(const ELFFile<ELFT> &HeadersFile);
 970   Error initGroupSection(GroupSection *GroupSec);
 971   Error initSymbolTable(SymbolTableSection *SymTab);
 972   Error readSectionHeaders();
 973   Error readSections(bool EnsureSymtab);
 974   Error findEhdrOffset();
 975   Expected<SectionBase &> makeSection(const Elf_Shdr &Shdr);
 976
 977 public:
 978   ELFBuilder(const ELFObjectFile<ELFT> &ElfObj, Object &Obj,
 979              std::optional<StringRef> ExtractPartition);
 980
 981   Error build(bool EnsureSymtab);
 982 };
 983
 984 class BinaryReader : public Reader {
 985   MemoryBuffer *MemBuf;
 986   uint8_t NewSymbolVisibility;
 987
 988 public:
 989   BinaryReader(MemoryBuffer *MB, const uint8_t NewSymbolVisibility)
 990       : MemBuf(MB), NewSymbolVisibility(NewSymbolVisibility) {}
 991   Expected<std::unique_ptr<Object>> create(bool EnsureSymtab) const override;
 992 };
 993
 994 class IHexReader : public Reader {
 995   MemoryBuffer *MemBuf;
 996
 997   Expected<std::vector<IHexRecord>> parse() const;
 998   Error parseError(size_t LineNo, Error E) const {
 999     return LineNo == -1U
1000                ? createFileError(MemBuf->getBufferIdentifier(), std::move(E))
1001                : createFileError(MemBuf->getBufferIdentifier(), LineNo,
1002                                  std::move(E));
1003   }
1004   template <typename... Ts>
1005   Error parseError(size_t LineNo, char const *Fmt, const Ts &...Vals) const {
1006     Error E = createStringError(errc::invalid_argument, Fmt, Vals...);
1007     return parseError(LineNo, std::move(E));
1008   }
1009
1010 public:
1011   IHexReader(MemoryBuffer *MB) : MemBuf(MB) {}
1012
1013   Expected<std::unique_ptr<Object>> create(bool EnsureSymtab) const override;
1014 };
1015
1016 class ELFReader : public Reader {
1017   Binary *Bin;
1018   std::optional<StringRef> ExtractPartition;
1019
1020 public:
1021   Expected<std::unique_ptr<Object>> create(bool EnsureSymtab) const override;
1022   explicit ELFReader(Binary *B, std::optional<StringRef> ExtractPartition)
1023       : Bin(B), ExtractPartition(ExtractPartition) {}
1024 };
1025
1026 class Object {
1027 private:
1028   using SecPtr = std::unique_ptr<SectionBase>;
1029   using SegPtr = std::unique_ptr<Segment>;
1030
1031   std::vector<SecPtr> Sections;
1032   std::vector<SegPtr> Segments;
1033   std::vector<SecPtr> RemovedSections;
1034   DenseMap<SectionBase *, std::vector<uint8_t>> UpdatedSections;
1035
1036   static bool sectionIsAlloc(const SectionBase &Sec) {
1037     return Sec.Flags & ELF::SHF_ALLOC;
1038   };
1039
1040 public:
1041   template <class T>
1042   using ConstRange = iterator_range<pointee_iterator<
1043       typename std::vector<std::unique_ptr<T>>::const_iterator>>;
1044
1045   // It is often the case that the ELF header and the program header table are
1046   // not present in any segment. This could be a problem during file layout,
1047   // because other segments may get assigned an offset where either of the
1048   // two should reside, which will effectively corrupt the resulting binary.
1049   // Other than that we use these segments to track program header offsets
1050   // when they may not follow the ELF header.
1051   Segment ElfHdrSegment;
1052   Segment ProgramHdrSegment;
1053
1054   bool Is64Bits;
1055   uint8_t OSABI;
1056   uint8_t ABIVersion;
1057   uint64_t Entry;
1058   uint64_t SHOff;
1059   uint32_t Type;
1060   uint32_t Machine;
1061   uint32_t Version;
1062   uint32_t Flags;
1063
1064   bool HadShdrs = true;
1065   bool MustBeRelocatable = false;
1066   StringTableSection *SectionNames = nullptr;
1067   SymbolTableSection *SymbolTable = nullptr;
1068   SectionIndexSection *SectionIndexTable = nullptr;
1069
1070   bool IsMips64EL = false;
1071
1072   SectionTableRef sections() const { return SectionTableRef(Sections); }
1073   iterator_range<
1074       filter_iterator<pointee_iterator<std::vector<SecPtr>::const_iterator>,
1075                       decltype(&sectionIsAlloc)>>
1076   allocSections() const {
1077     return make_filter_range(make_pointee_range(Sections), sectionIsAlloc);
1078   }
1079
1080   const auto &getUpdatedSections() const { return UpdatedSections; }
1081   Error updateSection(StringRef Name, ArrayRef<uint8_t> Data);
1082
1083   SectionBase *findSection(StringRef Name) {
1084     auto SecIt =
1085         find_if(Sections, [&](const SecPtr &Sec) { return Sec->Name == Name; });
1086     return SecIt == Sections.end() ? nullptr : SecIt->get();
1087   }
1088   SectionTableRef removedSections() { return SectionTableRef(RemovedSections); }
1089
1090   ConstRange<Segment> segments() const { return make_pointee_range(Segments); }
1091
1092   Error removeSections(bool AllowBrokenLinks,
1093                        std::function<bool(const SectionBase &)> ToRemove);
1094   Error replaceSections(const DenseMap<SectionBase *, SectionBase *> &FromTo);
1095   Error removeSymbols(function_ref<bool(const Symbol &)> ToRemove);
1096   template <class T, class... Ts> T &addSection(Ts &&...Args) {
1097     auto Sec = std::make_unique<T>(std::forward<Ts>(Args)...);
1098     auto Ptr = Sec.get();
1099     MustBeRelocatable |= isa<RelocationSection>(*Ptr);
1100     Sections.emplace_back(std::move(Sec));
1101     Ptr->Index = Sections.size();
1102     return *Ptr;
1103   }
1104   Error addNewSymbolTable();
1105   Segment &addSegment(ArrayRef<uint8_t> Data) {
1106     Segments.emplace_back(std::make_unique<Segment>(Data));
1107     return *Segments.back();
1108   }
1109   bool isRelocatable() const {
1110     return (Type != ELF::ET_DYN && Type != ELF::ET_EXEC) || MustBeRelocatable;
1111   }
1112 };
1113
1114 } // end namespace elf
1115 } // end namespace objcopy
1116 } // end namespace llvm
1117
1118 #endif // LLVM_LIB_OBJCOPY_ELF_ELFOBJECT_H