ObjectFile.h revision 36b56886974eae4f9c5ebc96befd3e7bfe5de338 
1//===- ObjectFile.h - File format independent object file -------*- C++ -*-===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This file declares a file format independent ObjectFile class.
11//
12//===----------------------------------------------------------------------===//
13
14#ifndef LLVM_OBJECT_OBJECTFILE_H
15#define LLVM_OBJECT_OBJECTFILE_H
16
17#include "llvm/ADT/StringRef.h"
18#include "llvm/Object/SymbolicFile.h"
19#include "llvm/Support/DataTypes.h"
20#include "llvm/Support/ErrorHandling.h"
21#include "llvm/Support/FileSystem.h"
22#include "llvm/Support/MemoryBuffer.h"
23#include <cstring>
24#include <vector>
25
26namespace llvm {
27namespace object {
28
29class ObjectFile;
30
31class SymbolRef;
32class symbol_iterator;
33
34/// RelocationRef - This is a value type class that represents a single
35/// relocation in the list of relocations in the object file.
36class RelocationRef {
37  DataRefImpl RelocationPimpl;
38  const ObjectFile *OwningObject;
39
40public:
41  RelocationRef() : OwningObject(NULL) { }
42
43  RelocationRef(DataRefImpl RelocationP, const ObjectFile *Owner);
44
45  bool operator==(const RelocationRef &Other) const;
46
47  void moveNext();
48
49  error_code getAddress(uint64_t &Result) const;
50  error_code getOffset(uint64_t &Result) const;
51  symbol_iterator getSymbol() const;
52  error_code getType(uint64_t &Result) const;
53
54  /// @brief Indicates whether this relocation should hidden when listing
55  /// relocations, usually because it is the trailing part of a multipart
56  /// relocation that will be printed as part of the leading relocation.
57  error_code getHidden(bool &Result) const;
58
59  /// @brief Get a string that represents the type of this relocation.
60  ///
61  /// This is for display purposes only.
62  error_code getTypeName(SmallVectorImpl<char> &Result) const;
63
64  /// @brief Get a string that represents the calculation of the value of this
65  ///        relocation.
66  ///
67  /// This is for display purposes only.
68  error_code getValueString(SmallVectorImpl<char> &Result) const;
69
70  DataRefImpl getRawDataRefImpl() const;
71  const ObjectFile *getObjectFile() const;
72};
73typedef content_iterator<RelocationRef> relocation_iterator;
74
75/// SectionRef - This is a value type class that represents a single section in
76/// the list of sections in the object file.
77class SectionRef;
78typedef content_iterator<SectionRef> section_iterator;
79class SectionRef {
80  friend class SymbolRef;
81  DataRefImpl SectionPimpl;
82  const ObjectFile *OwningObject;
83
84public:
85  SectionRef() : OwningObject(NULL) { }
86
87  SectionRef(DataRefImpl SectionP, const ObjectFile *Owner);
88
89  bool operator==(const SectionRef &Other) const;
90  bool operator!=(const SectionRef &Other) const;
91  bool operator<(const SectionRef &Other) const;
92
93  void moveNext();
94
95  error_code getName(StringRef &Result) const;
96  error_code getAddress(uint64_t &Result) const;
97  error_code getSize(uint64_t &Result) const;
98  error_code getContents(StringRef &Result) const;
99
100  /// @brief Get the alignment of this section as the actual value (not log 2).
101  error_code getAlignment(uint64_t &Result) const;
102
103  // FIXME: Move to the normalization layer when it's created.
104  error_code isText(bool &Result) const;
105  error_code isData(bool &Result) const;
106  error_code isBSS(bool &Result) const;
107  error_code isRequiredForExecution(bool &Result) const;
108  error_code isVirtual(bool &Result) const;
109  error_code isZeroInit(bool &Result) const;
110  error_code isReadOnlyData(bool &Result) const;
111
112  error_code containsSymbol(SymbolRef S, bool &Result) const;
113
114  relocation_iterator relocation_begin() const;
115  relocation_iterator relocation_end() const;
116  typedef iterator_range<relocation_iterator> relocation_iterator_range;
117  relocation_iterator_range relocations() const {
118    return relocation_iterator_range(relocation_begin(), relocation_end());
119  }
120  bool relocation_empty() const;
121  section_iterator getRelocatedSection() const;
122
123  DataRefImpl getRawDataRefImpl() const;
124};
125
126/// SymbolRef - This is a value type class that represents a single symbol in
127/// the list of symbols in the object file.
128class SymbolRef : public BasicSymbolRef {
129  friend class SectionRef;
130
131public:
132  SymbolRef() : BasicSymbolRef() {}
133
134  enum Type {
135    ST_Unknown, // Type not specified
136    ST_Data,
137    ST_Debug,
138    ST_File,
139    ST_Function,
140    ST_Other
141  };
142
143  SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner);
144
145  error_code getName(StringRef &Result) const;
146  /// Returns the symbol virtual address (i.e. address at which it will be
147  /// mapped).
148  error_code getAddress(uint64_t &Result) const;
149  error_code getFileOffset(uint64_t &Result) const;
150  /// @brief Get the alignment of this symbol as the actual value (not log 2).
151  error_code getAlignment(uint32_t &Result) const;
152  error_code getSize(uint64_t &Result) const;
153  error_code getType(SymbolRef::Type &Result) const;
154
155  /// @brief Get section this symbol is defined in reference to. Result is
156  /// end_sections() if it is undefined or is an absolute symbol.
157  error_code getSection(section_iterator &Result) const;
158
159  const ObjectFile *getObject() const;
160};
161
162class symbol_iterator : public basic_symbol_iterator {
163public:
164  symbol_iterator(SymbolRef Sym) : basic_symbol_iterator(Sym) {}
165  symbol_iterator(const basic_symbol_iterator &B)
166      : basic_symbol_iterator(SymbolRef(B->getRawDataRefImpl(),
167                                        cast<ObjectFile>(B->getObject()))) {}
168
169  const SymbolRef *operator->() const {
170    const BasicSymbolRef &P = basic_symbol_iterator::operator *();
171    return static_cast<const SymbolRef*>(&P);
172  }
173
174  const SymbolRef &operator*() const {
175    const BasicSymbolRef &P = basic_symbol_iterator::operator *();
176    return static_cast<const SymbolRef&>(P);
177  }
178};
179
180/// LibraryRef - This is a value type class that represents a single library in
181/// the list of libraries needed by a shared or dynamic object.
182class LibraryRef {
183  friend class SectionRef;
184  DataRefImpl LibraryPimpl;
185  const ObjectFile *OwningObject;
186
187public:
188  LibraryRef() : OwningObject(NULL) { }
189
190  LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner);
191
192  bool operator==(const LibraryRef &Other) const;
193  bool operator<(const LibraryRef &Other) const;
194
195  error_code getNext(LibraryRef &Result) const;
196
197  // Get the path to this library, as stored in the object file.
198  error_code getPath(StringRef &Result) const;
199
200  DataRefImpl getRawDataRefImpl() const;
201};
202typedef content_iterator<LibraryRef> library_iterator;
203
204/// ObjectFile - This class is the base class for all object file types.
205/// Concrete instances of this object are created by createObjectFile, which
206/// figures out which type to create.
207class ObjectFile : public SymbolicFile {
208  virtual void anchor();
209  ObjectFile() LLVM_DELETED_FUNCTION;
210  ObjectFile(const ObjectFile &other) LLVM_DELETED_FUNCTION;
211
212protected:
213  ObjectFile(unsigned int Type, MemoryBuffer *Source, bool BufferOwned = true);
214
215  const uint8_t *base() const {
216    return reinterpret_cast<const uint8_t *>(Data->getBufferStart());
217  }
218
219  // These functions are for SymbolRef to call internally. The main goal of
220  // this is to allow SymbolRef::SymbolPimpl to point directly to the symbol
221  // entry in the memory mapped object file. SymbolPimpl cannot contain any
222  // virtual functions because then it could not point into the memory mapped
223  // file.
224  //
225  // Implementations assume that the DataRefImpl is valid and has not been
226  // modified externally. It's UB otherwise.
227  friend class SymbolRef;
228  virtual error_code getSymbolName(DataRefImpl Symb, StringRef &Res) const = 0;
229  error_code printSymbolName(raw_ostream &OS, DataRefImpl Symb) const override;
230  virtual error_code getSymbolAddress(DataRefImpl Symb, uint64_t &Res) const = 0;
231  virtual error_code getSymbolAlignment(DataRefImpl Symb, uint32_t &Res) const;
232  virtual error_code getSymbolSize(DataRefImpl Symb, uint64_t &Res) const = 0;
233  virtual error_code getSymbolType(DataRefImpl Symb,
234                                   SymbolRef::Type &Res) const = 0;
235  virtual error_code getSymbolSection(DataRefImpl Symb,
236                                      section_iterator &Res) const = 0;
237
238  // Same as above for SectionRef.
239  friend class SectionRef;
240  virtual void moveSectionNext(DataRefImpl &Sec) const = 0;
241  virtual error_code getSectionName(DataRefImpl Sec, StringRef &Res) const = 0;
242  virtual error_code getSectionAddress(DataRefImpl Sec, uint64_t &Res) const =0;
243  virtual error_code getSectionSize(DataRefImpl Sec, uint64_t &Res) const = 0;
244  virtual error_code getSectionContents(DataRefImpl Sec, StringRef &Res)const=0;
245  virtual error_code getSectionAlignment(DataRefImpl Sec, uint64_t &Res)const=0;
246  virtual error_code isSectionText(DataRefImpl Sec, bool &Res) const = 0;
247  virtual error_code isSectionData(DataRefImpl Sec, bool &Res) const = 0;
248  virtual error_code isSectionBSS(DataRefImpl Sec, bool &Res) const = 0;
249  virtual error_code isSectionRequiredForExecution(DataRefImpl Sec,
250                                                   bool &Res) const = 0;
251  // A section is 'virtual' if its contents aren't present in the object image.
252  virtual error_code isSectionVirtual(DataRefImpl Sec, bool &Res) const = 0;
253  virtual error_code isSectionZeroInit(DataRefImpl Sec, bool &Res) const = 0;
254  virtual error_code isSectionReadOnlyData(DataRefImpl Sec, bool &Res) const =0;
255  virtual error_code sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb,
256                                           bool &Result) const = 0;
257  virtual relocation_iterator section_rel_begin(DataRefImpl Sec) const = 0;
258  virtual relocation_iterator section_rel_end(DataRefImpl Sec) const = 0;
259  virtual bool section_rel_empty(DataRefImpl Sec) const = 0;
260  virtual section_iterator getRelocatedSection(DataRefImpl Sec) const;
261
262  // Same as above for RelocationRef.
263  friend class RelocationRef;
264  virtual void moveRelocationNext(DataRefImpl &Rel) const = 0;
265  virtual error_code getRelocationAddress(DataRefImpl Rel,
266                                          uint64_t &Res) const =0;
267  virtual error_code getRelocationOffset(DataRefImpl Rel,
268                                         uint64_t &Res) const =0;
269  virtual symbol_iterator getRelocationSymbol(DataRefImpl Rel) const = 0;
270  virtual error_code getRelocationType(DataRefImpl Rel,
271                                       uint64_t &Res) const = 0;
272  virtual error_code getRelocationTypeName(DataRefImpl Rel,
273                                       SmallVectorImpl<char> &Result) const = 0;
274  virtual error_code getRelocationValueString(DataRefImpl Rel,
275                                       SmallVectorImpl<char> &Result) const = 0;
276  virtual error_code getRelocationHidden(DataRefImpl Rel, bool &Result) const {
277    Result = false;
278    return object_error::success;
279  }
280
281  // Same for LibraryRef
282  friend class LibraryRef;
283  virtual error_code getLibraryNext(DataRefImpl Lib, LibraryRef &Res) const = 0;
284  virtual error_code getLibraryPath(DataRefImpl Lib, StringRef &Res) const = 0;
285
286public:
287  typedef iterator_range<symbol_iterator> symbol_iterator_range;
288  symbol_iterator_range symbols() const {
289    return symbol_iterator_range(symbol_begin(), symbol_end());
290  }
291
292  virtual section_iterator section_begin() const = 0;
293  virtual section_iterator section_end() const = 0;
294
295  typedef iterator_range<section_iterator> section_iterator_range;
296  section_iterator_range sections() const {
297    return section_iterator_range(section_begin(), section_end());
298  }
299
300  virtual library_iterator needed_library_begin() const = 0;
301  virtual library_iterator needed_library_end() const = 0;
302
303  /// @brief The number of bytes used to represent an address in this object
304  ///        file format.
305  virtual uint8_t getBytesInAddress() const = 0;
306
307  virtual StringRef getFileFormatName() const = 0;
308  virtual /* Triple::ArchType */ unsigned getArch() const = 0;
309
310  /// For shared objects, returns the name which this object should be
311  /// loaded from at runtime. This corresponds to DT_SONAME on ELF and
312  /// LC_ID_DYLIB (install name) on MachO.
313  virtual StringRef getLoadName() const = 0;
314
315  /// @returns Pointer to ObjectFile subclass to handle this type of object.
316  /// @param ObjectPath The path to the object file. ObjectPath.isObject must
317  ///        return true.
318  /// @brief Create ObjectFile from path.
319  static ErrorOr<ObjectFile *> createObjectFile(StringRef ObjectPath);
320  static ErrorOr<ObjectFile *> createObjectFile(MemoryBuffer *Object,
321                                                bool BufferOwned,
322                                                sys::fs::file_magic Type);
323  static ErrorOr<ObjectFile *> createObjectFile(MemoryBuffer *Object) {
324    return createObjectFile(Object, true, sys::fs::file_magic::unknown);
325  }
326
327
328  static inline bool classof(const Binary *v) {
329    return v->isObject();
330  }
331
332public:
333  static ErrorOr<ObjectFile *> createCOFFObjectFile(MemoryBuffer *Object,
334                                                    bool BufferOwned = true);
335  static ErrorOr<ObjectFile *> createELFObjectFile(MemoryBuffer *Object,
336                                                   bool BufferOwned = true);
337  static ErrorOr<ObjectFile *> createMachOObjectFile(MemoryBuffer *Object,
338                                                     bool BufferOwned = true);
339};
340
341// Inline function definitions.
342inline SymbolRef::SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner)
343    : BasicSymbolRef(SymbolP, Owner) {}
344
345inline error_code SymbolRef::getName(StringRef &Result) const {
346  return getObject()->getSymbolName(getRawDataRefImpl(), Result);
347}
348
349inline error_code SymbolRef::getAddress(uint64_t &Result) const {
350  return getObject()->getSymbolAddress(getRawDataRefImpl(), Result);
351}
352
353inline error_code SymbolRef::getFileOffset(uint64_t &Result) const {
354  uint64_t Address;
355  if (error_code EC = getAddress(Address))
356    return EC;
357  if (Address == UnknownAddressOrSize) {
358    Result = UnknownAddressOrSize;
359    return object_error::success;
360  }
361
362  const ObjectFile *Obj = getObject();
363  section_iterator SecI(Obj->section_begin());
364  if (error_code EC = getSection(SecI))
365    return EC;
366
367  if (SecI == Obj->section_end()) {
368    Result = UnknownAddressOrSize;
369    return object_error::success;
370  }
371
372  uint64_t SectionAddress;
373  if (error_code EC = SecI->getAddress(SectionAddress))
374    return EC;
375
376  uint64_t OffsetInSection = Address - SectionAddress;
377
378  StringRef SecContents;
379  if (error_code EC = SecI->getContents(SecContents))
380    return EC;
381
382  // FIXME: this is a hack.
383  uint64_t SectionOffset = (uint64_t)SecContents.data() - (uint64_t)Obj->base();
384
385  Result = SectionOffset + OffsetInSection;
386  return object_error::success;
387}
388
389inline error_code SymbolRef::getAlignment(uint32_t &Result) const {
390  return getObject()->getSymbolAlignment(getRawDataRefImpl(), Result);
391}
392
393inline error_code SymbolRef::getSize(uint64_t &Result) const {
394  return getObject()->getSymbolSize(getRawDataRefImpl(), Result);
395}
396
397inline error_code SymbolRef::getSection(section_iterator &Result) const {
398  return getObject()->getSymbolSection(getRawDataRefImpl(), Result);
399}
400
401inline error_code SymbolRef::getType(SymbolRef::Type &Result) const {
402  return getObject()->getSymbolType(getRawDataRefImpl(), Result);
403}
404
405inline const ObjectFile *SymbolRef::getObject() const {
406  const SymbolicFile *O = BasicSymbolRef::getObject();
407  return cast<ObjectFile>(O);
408}
409
410
411/// SectionRef
412inline SectionRef::SectionRef(DataRefImpl SectionP,
413                              const ObjectFile *Owner)
414  : SectionPimpl(SectionP)
415  , OwningObject(Owner) {}
416
417inline bool SectionRef::operator==(const SectionRef &Other) const {
418  return SectionPimpl == Other.SectionPimpl;
419}
420
421inline bool SectionRef::operator!=(const SectionRef &Other) const {
422  return SectionPimpl != Other.SectionPimpl;
423}
424
425inline bool SectionRef::operator<(const SectionRef &Other) const {
426  return SectionPimpl < Other.SectionPimpl;
427}
428
429inline void SectionRef::moveNext() {
430  return OwningObject->moveSectionNext(SectionPimpl);
431}
432
433inline error_code SectionRef::getName(StringRef &Result) const {
434  return OwningObject->getSectionName(SectionPimpl, Result);
435}
436
437inline error_code SectionRef::getAddress(uint64_t &Result) const {
438  return OwningObject->getSectionAddress(SectionPimpl, Result);
439}
440
441inline error_code SectionRef::getSize(uint64_t &Result) const {
442  return OwningObject->getSectionSize(SectionPimpl, Result);
443}
444
445inline error_code SectionRef::getContents(StringRef &Result) const {
446  return OwningObject->getSectionContents(SectionPimpl, Result);
447}
448
449inline error_code SectionRef::getAlignment(uint64_t &Result) const {
450  return OwningObject->getSectionAlignment(SectionPimpl, Result);
451}
452
453inline error_code SectionRef::isText(bool &Result) const {
454  return OwningObject->isSectionText(SectionPimpl, Result);
455}
456
457inline error_code SectionRef::isData(bool &Result) const {
458  return OwningObject->isSectionData(SectionPimpl, Result);
459}
460
461inline error_code SectionRef::isBSS(bool &Result) const {
462  return OwningObject->isSectionBSS(SectionPimpl, Result);
463}
464
465inline error_code SectionRef::isRequiredForExecution(bool &Result) const {
466  return OwningObject->isSectionRequiredForExecution(SectionPimpl, Result);
467}
468
469inline error_code SectionRef::isVirtual(bool &Result) const {
470  return OwningObject->isSectionVirtual(SectionPimpl, Result);
471}
472
473inline error_code SectionRef::isZeroInit(bool &Result) const {
474  return OwningObject->isSectionZeroInit(SectionPimpl, Result);
475}
476
477inline error_code SectionRef::isReadOnlyData(bool &Result) const {
478  return OwningObject->isSectionReadOnlyData(SectionPimpl, Result);
479}
480
481inline error_code SectionRef::containsSymbol(SymbolRef S, bool &Result) const {
482  return OwningObject->sectionContainsSymbol(SectionPimpl,
483                                             S.getRawDataRefImpl(), Result);
484}
485
486inline relocation_iterator SectionRef::relocation_begin() const {
487  return OwningObject->section_rel_begin(SectionPimpl);
488}
489
490inline relocation_iterator SectionRef::relocation_end() const {
491  return OwningObject->section_rel_end(SectionPimpl);
492}
493
494inline bool SectionRef::relocation_empty() const {
495  return OwningObject->section_rel_empty(SectionPimpl);
496}
497
498inline section_iterator SectionRef::getRelocatedSection() const {
499  return OwningObject->getRelocatedSection(SectionPimpl);
500}
501
502inline DataRefImpl SectionRef::getRawDataRefImpl() const {
503  return SectionPimpl;
504}
505
506/// RelocationRef
507inline RelocationRef::RelocationRef(DataRefImpl RelocationP,
508                              const ObjectFile *Owner)
509  : RelocationPimpl(RelocationP)
510  , OwningObject(Owner) {}
511
512inline bool RelocationRef::operator==(const RelocationRef &Other) const {
513  return RelocationPimpl == Other.RelocationPimpl;
514}
515
516inline void RelocationRef::moveNext() {
517  return OwningObject->moveRelocationNext(RelocationPimpl);
518}
519
520inline error_code RelocationRef::getAddress(uint64_t &Result) const {
521  return OwningObject->getRelocationAddress(RelocationPimpl, Result);
522}
523
524inline error_code RelocationRef::getOffset(uint64_t &Result) const {
525  return OwningObject->getRelocationOffset(RelocationPimpl, Result);
526}
527
528inline symbol_iterator RelocationRef::getSymbol() const {
529  return OwningObject->getRelocationSymbol(RelocationPimpl);
530}
531
532inline error_code RelocationRef::getType(uint64_t &Result) const {
533  return OwningObject->getRelocationType(RelocationPimpl, Result);
534}
535
536inline error_code RelocationRef::getTypeName(SmallVectorImpl<char> &Result)
537  const {
538  return OwningObject->getRelocationTypeName(RelocationPimpl, Result);
539}
540
541inline error_code RelocationRef::getValueString(SmallVectorImpl<char> &Result)
542  const {
543  return OwningObject->getRelocationValueString(RelocationPimpl, Result);
544}
545
546inline error_code RelocationRef::getHidden(bool &Result) const {
547  return OwningObject->getRelocationHidden(RelocationPimpl, Result);
548}
549
550inline DataRefImpl RelocationRef::getRawDataRefImpl() const {
551  return RelocationPimpl;
552}
553
554inline const ObjectFile *RelocationRef::getObjectFile() const {
555  return OwningObject;
556}
557
558// Inline function definitions.
559inline LibraryRef::LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner)
560  : LibraryPimpl(LibraryP)
561  , OwningObject(Owner) {}
562
563inline bool LibraryRef::operator==(const LibraryRef &Other) const {
564  return LibraryPimpl == Other.LibraryPimpl;
565}
566
567inline bool LibraryRef::operator<(const LibraryRef &Other) const {
568  return LibraryPimpl < Other.LibraryPimpl;
569}
570
571inline error_code LibraryRef::getNext(LibraryRef &Result) const {
572  return OwningObject->getLibraryNext(LibraryPimpl, Result);
573}
574
575inline error_code LibraryRef::getPath(StringRef &Result) const {
576  return OwningObject->getLibraryPath(LibraryPimpl, Result);
577}
578
579} // end namespace object
580} // end namespace llvm
581
582#endif
583