include/llvm/Object/ObjectFile.h

   1 //===- ObjectFile.h - File format independent object file -------*- 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 // This file declares a file format independent ObjectFile class.
  10 //
  11 //===----------------------------------------------------------------------===//
  12
  13 #ifndef LLVM_OBJECT_OBJECTFILE_H
  14 #define LLVM_OBJECT_OBJECTFILE_H
  15
  16 #include "llvm/ADT/StringRef.h"
  17 #include "llvm/ADT/Triple.h"
  18 #include "llvm/ADT/iterator_range.h"
  19 #include "llvm/BinaryFormat/Magic.h"
  20 #include "llvm/MC/SubtargetFeature.h"
  21 #include "llvm/Object/Binary.h"
  22 #include "llvm/Object/Error.h"
  23 #include "llvm/Object/SymbolicFile.h"
  24 #include "llvm/Support/Casting.h"
  25 #include "llvm/Support/Error.h"
  26 #include "llvm/Support/FileSystem.h"
  27 #include "llvm/Support/MemoryBuffer.h"
  28 #include <cassert>
  29 #include <cstdint>
  30 #include <memory>
  31 #include <system_error>
  32
  33 namespace llvm {
  34
  35 class ARMAttributeParser;
  36
  37 namespace object {
  38
  39 class COFFObjectFile;
  40 class MachOObjectFile;
  41 class ObjectFile;
  42 class SectionRef;
  43 class SymbolRef;
  44 class symbol_iterator;
  45 class WasmObjectFile;
  46
  47 using section_iterator = content_iterator<SectionRef>;
  48
  49 /// This is a value type class that represents a single relocation in the list
  50 /// of relocations in the object file.
  51 class RelocationRef {
  52   DataRefImpl RelocationPimpl;
  53   const ObjectFile *OwningObject = nullptr;
  54
  55 public:
  56   RelocationRef() = default;
  57   RelocationRef(DataRefImpl RelocationP, const ObjectFile *Owner);
  58
  59   bool operator==(const RelocationRef &Other) const;
  60
  61   void moveNext();
  62
  63   uint64_t getOffset() const;
  64   symbol_iterator getSymbol() const;
  65   uint64_t getType() const;
  66
  67   /// Get a string that represents the type of this relocation.
  68   ///
  69   /// This is for display purposes only.
  70   void getTypeName(SmallVectorImpl<char> &Result) const;
  71
  72   DataRefImpl getRawDataRefImpl() const;
  73   const ObjectFile *getObject() const;
  74 };
  75
  76 using relocation_iterator = content_iterator<RelocationRef>;
  77
  78 /// This is a value type class that represents a single section in the list of
  79 /// sections in the object file.
  80 class SectionRef {
  81   friend class SymbolRef;
  82
  83   DataRefImpl SectionPimpl;
  84   const ObjectFile *OwningObject = nullptr;
  85
  86 public:
  87   SectionRef() = default;
  88   SectionRef(DataRefImpl SectionP, const ObjectFile *Owner);
  89
  90   bool operator==(const SectionRef &Other) const;
  91   bool operator!=(const SectionRef &Other) const;
  92   bool operator<(const SectionRef &Other) const;
  93
  94   void moveNext();
  95
  96   std::error_code getName(StringRef &Result) const;
  97   uint64_t getAddress() const;
  98   uint64_t getIndex() const;
  99   uint64_t getSize() const;
 100   std::error_code getContents(StringRef &Result) const;
 101
 102   /// Get the alignment of this section as the actual value (not log 2).
 103   uint64_t getAlignment() const;
 104
 105   bool isCompressed() const;
 106   /// Whether this section contains instructions.
 107   bool isText() const;
 108   /// Whether this section contains data, not instructions.
 109   bool isData() const;
 110   /// Whether this section contains BSS uninitialized data.
 111   bool isBSS() const;
 112   bool isVirtual() const;
 113   bool isBitcode() const;
 114   bool isStripped() const;
 115
 116   /// Whether this section will be placed in the text segment, according to the
 117   /// Berkeley size format. This is true if the section is allocatable, and
 118   /// contains either code or readonly data.
 119   bool isBerkeleyText() const;
 120   /// Whether this section will be placed in the data segment, according to the
 121   /// Berkeley size format. This is true if the section is allocatable and
 122   /// contains data (e.g. PROGBITS), but is not text.
 123   bool isBerkeleyData() const;
 124
 125   bool containsSymbol(SymbolRef S) const;
 126
 127   relocation_iterator relocation_begin() const;
 128   relocation_iterator relocation_end() const;
 129   iterator_range<relocation_iterator> relocations() const {
 130     return make_range(relocation_begin(), relocation_end());
 131   }
 132   section_iterator getRelocatedSection() const;
 133
 134   DataRefImpl getRawDataRefImpl() const;
 135   const ObjectFile *getObject() const;
 136 };
 137
 138 /// This is a value type class that represents a single symbol in the list of
 139 /// symbols in the object file.
 140 class SymbolRef : public BasicSymbolRef {
 141   friend class SectionRef;
 142
 143 public:
 144   enum Type {
 145     ST_Unknown, // Type not specified
 146     ST_Data,
 147     ST_Debug,
 148     ST_File,
 149     ST_Function,
 150     ST_Other
 151   };
 152
 153   SymbolRef() = default;
 154   SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner);
 155   SymbolRef(const BasicSymbolRef &B) : BasicSymbolRef(B) {
 156     assert(isa<ObjectFile>(BasicSymbolRef::getObject()));
 157   }
 158
 159   Expected<StringRef> getName() const;
 160   /// Returns the symbol virtual address (i.e. address at which it will be
 161   /// mapped).
 162   Expected<uint64_t> getAddress() const;
 163
 164   /// Return the value of the symbol depending on the object this can be an
 165   /// offset or a virtual address.
 166   uint64_t getValue() const;
 167
 168   /// Get the alignment of this symbol as the actual value (not log 2).
 169   uint32_t getAlignment() const;
 170   uint64_t getCommonSize() const;
 171   Expected<SymbolRef::Type> getType() const;
 172
 173   /// Get section this symbol is defined in reference to. Result is
 174   /// end_sections() if it is undefined or is an absolute symbol.
 175   Expected<section_iterator> getSection() const;
 176
 177   const ObjectFile *getObject() const;
 178 };
 179
 180 class symbol_iterator : public basic_symbol_iterator {
 181 public:
 182   symbol_iterator(SymbolRef Sym) : basic_symbol_iterator(Sym) {}
 183   symbol_iterator(const basic_symbol_iterator &B)
 184       : basic_symbol_iterator(SymbolRef(B->getRawDataRefImpl(),
 185                                         cast<ObjectFile>(B->getObject()))) {}
 186
 187   const SymbolRef *operator->() const {
 188     const BasicSymbolRef &P = basic_symbol_iterator::operator *();
 189     return static_cast<const SymbolRef*>(&P);
 190   }
 191
 192   const SymbolRef &operator*() const {
 193     const BasicSymbolRef &P = basic_symbol_iterator::operator *();
 194     return static_cast<const SymbolRef&>(P);
 195   }
 196 };
 197
 198 /// This class is the base class for all object file types. Concrete instances
 199 /// of this object are created by createObjectFile, which figures out which type
 200 /// to create.
 201 class ObjectFile : public SymbolicFile {
 202   virtual void anchor();
 203
 204 protected:
 205   ObjectFile(unsigned int Type, MemoryBufferRef Source);
 206
 207   const uint8_t *base() const {
 208     return reinterpret_cast<const uint8_t *>(Data.getBufferStart());
 209   }
 210
 211   // These functions are for SymbolRef to call internally. The main goal of
 212   // this is to allow SymbolRef::SymbolPimpl to point directly to the symbol
 213   // entry in the memory mapped object file. SymbolPimpl cannot contain any
 214   // virtual functions because then it could not point into the memory mapped
 215   // file.
 216   //
 217   // Implementations assume that the DataRefImpl is valid and has not been
 218   // modified externally. It's UB otherwise.
 219   friend class SymbolRef;
 220
 221   virtual Expected<StringRef> getSymbolName(DataRefImpl Symb) const = 0;
 222   std::error_code printSymbolName(raw_ostream &OS,
 223                                   DataRefImpl Symb) const override;
 224   virtual Expected<uint64_t> getSymbolAddress(DataRefImpl Symb) const = 0;
 225   virtual uint64_t getSymbolValueImpl(DataRefImpl Symb) const = 0;
 226   virtual uint32_t getSymbolAlignment(DataRefImpl Symb) const;
 227   virtual uint64_t getCommonSymbolSizeImpl(DataRefImpl Symb) const = 0;
 228   virtual Expected<SymbolRef::Type> getSymbolType(DataRefImpl Symb) const = 0;
 229   virtual Expected<section_iterator>
 230   getSymbolSection(DataRefImpl Symb) const = 0;
 231
 232   // Same as above for SectionRef.
 233   friend class SectionRef;
 234
 235   virtual void moveSectionNext(DataRefImpl &Sec) const = 0;
 236   virtual std::error_code getSectionName(DataRefImpl Sec,
 237                                          StringRef &Res) const = 0;
 238   virtual uint64_t getSectionAddress(DataRefImpl Sec) const = 0;
 239   virtual uint64_t getSectionIndex(DataRefImpl Sec) const = 0;
 240   virtual uint64_t getSectionSize(DataRefImpl Sec) const = 0;
 241   virtual std::error_code getSectionContents(DataRefImpl Sec,
 242                                              StringRef &Res) const = 0;
 243   virtual uint64_t getSectionAlignment(DataRefImpl Sec) const = 0;
 244   virtual bool isSectionCompressed(DataRefImpl Sec) const = 0;
 245   virtual bool isSectionText(DataRefImpl Sec) const = 0;
 246   virtual bool isSectionData(DataRefImpl Sec) const = 0;
 247   virtual bool isSectionBSS(DataRefImpl Sec) const = 0;
 248   // A section is 'virtual' if its contents aren't present in the object image.
 249   virtual bool isSectionVirtual(DataRefImpl Sec) const = 0;
 250   virtual bool isSectionBitcode(DataRefImpl Sec) const;
 251   virtual bool isSectionStripped(DataRefImpl Sec) const;
 252   virtual bool isBerkeleyText(DataRefImpl Sec) const;
 253   virtual bool isBerkeleyData(DataRefImpl Sec) const;
 254   virtual relocation_iterator section_rel_begin(DataRefImpl Sec) const = 0;
 255   virtual relocation_iterator section_rel_end(DataRefImpl Sec) const = 0;
 256   virtual section_iterator getRelocatedSection(DataRefImpl Sec) const;
 257
 258   // Same as above for RelocationRef.
 259   friend class RelocationRef;
 260   virtual void moveRelocationNext(DataRefImpl &Rel) const = 0;
 261   virtual uint64_t getRelocationOffset(DataRefImpl Rel) const = 0;
 262   virtual symbol_iterator getRelocationSymbol(DataRefImpl Rel) const = 0;
 263   virtual uint64_t getRelocationType(DataRefImpl Rel) const = 0;
 264   virtual void getRelocationTypeName(DataRefImpl Rel,
 265                                      SmallVectorImpl<char> &Result) const = 0;
 266
 267   uint64_t getSymbolValue(DataRefImpl Symb) const;
 268
 269 public:
 270   ObjectFile() = delete;
 271   ObjectFile(const ObjectFile &other) = delete;
 272
 273   uint64_t getCommonSymbolSize(DataRefImpl Symb) const {
 274     assert(getSymbolFlags(Symb) & SymbolRef::SF_Common);
 275     return getCommonSymbolSizeImpl(Symb);
 276   }
 277
 278   virtual std::vector<SectionRef> dynamic_relocation_sections() const {
 279     return std::vector<SectionRef>();
 280   }
 281
 282   using symbol_iterator_range = iterator_range<symbol_iterator>;
 283   symbol_iterator_range symbols() const {
 284     return symbol_iterator_range(symbol_begin(), symbol_end());
 285   }
 286
 287   virtual section_iterator section_begin() const = 0;
 288   virtual section_iterator section_end() const = 0;
 289
 290   using section_iterator_range = iterator_range<section_iterator>;
 291   section_iterator_range sections() const {
 292     return section_iterator_range(section_begin(), section_end());
 293   }
 294
 295   /// The number of bytes used to represent an address in this object
 296   ///        file format.
 297   virtual uint8_t getBytesInAddress() const = 0;
 298
 299   virtual StringRef getFileFormatName() const = 0;
 300   virtual Triple::ArchType getArch() const = 0;
 301   virtual SubtargetFeatures getFeatures() const = 0;
 302   virtual void setARMSubArch(Triple &TheTriple) const { }
 303   virtual Expected<uint64_t> getStartAddress() const {
 304     return errorCodeToError(object_error::parse_failed);
 305   };
 306
 307   /// Create a triple from the data in this object file.
 308   Triple makeTriple() const;
 309
 310   virtual std::error_code
 311     getBuildAttributes(ARMAttributeParser &Attributes) const {
 312       return std::error_code();
 313     }
 314
 315   /// Maps a debug section name to a standard DWARF section name.
 316   virtual StringRef mapDebugSectionName(StringRef Name) const { return Name; }
 317
 318   /// True if this is a relocatable object (.o/.obj).
 319   virtual bool isRelocatableObject() const = 0;
 320
 321   /// @returns Pointer to ObjectFile subclass to handle this type of object.
 322   /// @param ObjectPath The path to the object file. ObjectPath.isObject must
 323   ///        return true.
 324   /// Create ObjectFile from path.
 325   static Expected<OwningBinary<ObjectFile>>
 326   createObjectFile(StringRef ObjectPath);
 327
 328   static Expected<std::unique_ptr<ObjectFile>>
 329   createObjectFile(MemoryBufferRef Object, llvm::file_magic Type);
 330   static Expected<std::unique_ptr<ObjectFile>>
 331   createObjectFile(MemoryBufferRef Object) {
 332     return createObjectFile(Object, llvm::file_magic::unknown);
 333   }
 334
 335   static bool classof(const Binary *v) {
 336     return v->isObject();
 337   }
 338
 339   static Expected<std::unique_ptr<COFFObjectFile>>
 340   createCOFFObjectFile(MemoryBufferRef Object);
 341
 342   static Expected<std::unique_ptr<ObjectFile>>
 343   createELFObjectFile(MemoryBufferRef Object);
 344
 345   static Expected<std::unique_ptr<MachOObjectFile>>
 346   createMachOObjectFile(MemoryBufferRef Object,
 347                         uint32_t UniversalCputype = 0,
 348                         uint32_t UniversalIndex = 0);
 349
 350   static Expected<std::unique_ptr<WasmObjectFile>>
 351   createWasmObjectFile(MemoryBufferRef Object);
 352 };
 353
 354 // Inline function definitions.
 355 inline SymbolRef::SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner)
 356     : BasicSymbolRef(SymbolP, Owner) {}
 357
 358 inline Expected<StringRef> SymbolRef::getName() const {
 359   return getObject()->getSymbolName(getRawDataRefImpl());
 360 }
 361
 362 inline Expected<uint64_t> SymbolRef::getAddress() const {
 363   return getObject()->getSymbolAddress(getRawDataRefImpl());
 364 }
 365
 366 inline uint64_t SymbolRef::getValue() const {
 367   return getObject()->getSymbolValue(getRawDataRefImpl());
 368 }
 369
 370 inline uint32_t SymbolRef::getAlignment() const {
 371   return getObject()->getSymbolAlignment(getRawDataRefImpl());
 372 }
 373
 374 inline uint64_t SymbolRef::getCommonSize() const {
 375   return getObject()->getCommonSymbolSize(getRawDataRefImpl());
 376 }
 377
 378 inline Expected<section_iterator> SymbolRef::getSection() const {
 379   return getObject()->getSymbolSection(getRawDataRefImpl());
 380 }
 381
 382 inline Expected<SymbolRef::Type> SymbolRef::getType() const {
 383   return getObject()->getSymbolType(getRawDataRefImpl());
 384 }
 385
 386 inline const ObjectFile *SymbolRef::getObject() const {
 387   const SymbolicFile *O = BasicSymbolRef::getObject();
 388   return cast<ObjectFile>(O);
 389 }
 390
 391 /// SectionRef
 392 inline SectionRef::SectionRef(DataRefImpl SectionP,
 393                               const ObjectFile *Owner)
 394   : SectionPimpl(SectionP)
 395   , OwningObject(Owner) {}
 396
 397 inline bool SectionRef::operator==(const SectionRef &Other) const {
 398   return SectionPimpl == Other.SectionPimpl;
 399 }
 400
 401 inline bool SectionRef::operator!=(const SectionRef &Other) const {
 402   return SectionPimpl != Other.SectionPimpl;
 403 }
 404
 405 inline bool SectionRef::operator<(const SectionRef &Other) const {
 406   return SectionPimpl < Other.SectionPimpl;
 407 }
 408
 409 inline void SectionRef::moveNext() {
 410   return OwningObject->moveSectionNext(SectionPimpl);
 411 }
 412
 413 inline std::error_code SectionRef::getName(StringRef &Result) const {
 414   return OwningObject->getSectionName(SectionPimpl, Result);
 415 }
 416
 417 inline uint64_t SectionRef::getAddress() const {
 418   return OwningObject->getSectionAddress(SectionPimpl);
 419 }
 420
 421 inline uint64_t SectionRef::getIndex() const {
 422   return OwningObject->getSectionIndex(SectionPimpl);
 423 }
 424
 425 inline uint64_t SectionRef::getSize() const {
 426   return OwningObject->getSectionSize(SectionPimpl);
 427 }
 428
 429 inline std::error_code SectionRef::getContents(StringRef &Result) const {
 430   return OwningObject->getSectionContents(SectionPimpl, Result);
 431 }
 432
 433 inline uint64_t SectionRef::getAlignment() const {
 434   return OwningObject->getSectionAlignment(SectionPimpl);
 435 }
 436
 437 inline bool SectionRef::isCompressed() const {
 438   return OwningObject->isSectionCompressed(SectionPimpl);
 439 }
 440
 441 inline bool SectionRef::isText() const {
 442   return OwningObject->isSectionText(SectionPimpl);
 443 }
 444
 445 inline bool SectionRef::isData() const {
 446   return OwningObject->isSectionData(SectionPimpl);
 447 }
 448
 449 inline bool SectionRef::isBSS() const {
 450   return OwningObject->isSectionBSS(SectionPimpl);
 451 }
 452
 453 inline bool SectionRef::isVirtual() const {
 454   return OwningObject->isSectionVirtual(SectionPimpl);
 455 }
 456
 457 inline bool SectionRef::isBitcode() const {
 458   return OwningObject->isSectionBitcode(SectionPimpl);
 459 }
 460
 461 inline bool SectionRef::isStripped() const {
 462   return OwningObject->isSectionStripped(SectionPimpl);
 463 }
 464
 465 inline bool SectionRef::isBerkeleyText() const {
 466   return OwningObject->isBerkeleyText(SectionPimpl);
 467 }
 468
 469 inline bool SectionRef::isBerkeleyData() const {
 470   return OwningObject->isBerkeleyData(SectionPimpl);
 471 }
 472
 473 inline relocation_iterator SectionRef::relocation_begin() const {
 474   return OwningObject->section_rel_begin(SectionPimpl);
 475 }
 476
 477 inline relocation_iterator SectionRef::relocation_end() const {
 478   return OwningObject->section_rel_end(SectionPimpl);
 479 }
 480
 481 inline section_iterator SectionRef::getRelocatedSection() const {
 482   return OwningObject->getRelocatedSection(SectionPimpl);
 483 }
 484
 485 inline DataRefImpl SectionRef::getRawDataRefImpl() const {
 486   return SectionPimpl;
 487 }
 488
 489 inline const ObjectFile *SectionRef::getObject() const {
 490   return OwningObject;
 491 }
 492
 493 /// RelocationRef
 494 inline RelocationRef::RelocationRef(DataRefImpl RelocationP,
 495                               const ObjectFile *Owner)
 496   : RelocationPimpl(RelocationP)
 497   , OwningObject(Owner) {}
 498
 499 inline bool RelocationRef::operator==(const RelocationRef &Other) const {
 500   return RelocationPimpl == Other.RelocationPimpl;
 501 }
 502
 503 inline void RelocationRef::moveNext() {
 504   return OwningObject->moveRelocationNext(RelocationPimpl);
 505 }
 506
 507 inline uint64_t RelocationRef::getOffset() const {
 508   return OwningObject->getRelocationOffset(RelocationPimpl);
 509 }
 510
 511 inline symbol_iterator RelocationRef::getSymbol() const {
 512   return OwningObject->getRelocationSymbol(RelocationPimpl);
 513 }
 514
 515 inline uint64_t RelocationRef::getType() const {
 516   return OwningObject->getRelocationType(RelocationPimpl);
 517 }
 518
 519 inline void RelocationRef::getTypeName(SmallVectorImpl<char> &Result) const {
 520   return OwningObject->getRelocationTypeName(RelocationPimpl, Result);
 521 }
 522
 523 inline DataRefImpl RelocationRef::getRawDataRefImpl() const {
 524   return RelocationPimpl;
 525 }
 526
 527 inline const ObjectFile *RelocationRef::getObject() const {
 528   return OwningObject;
 529 }
 530
 531 } // end namespace object
 532
 533 } // end namespace llvm
 534
 535 #endif // LLVM_OBJECT_OBJECTFILE_H