ObjectFile.h revision 167957fa095bc7200b908e6e142be3e604bcfeea 
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/Binary.h"
19#include "llvm/Support/DataTypes.h"
20#include "llvm/Support/ErrorHandling.h"
21#include "llvm/Support/MemoryBuffer.h"
22#include <cstring>
23#include <vector>
24
25namespace llvm {
26namespace object {
27
28class ObjectFile;
29
30union DataRefImpl {
31  struct {
32    // ELF needs this for relocations. This entire union should probably be a
33    // char[max(8, sizeof(uintptr_t))] and require the impl to cast.
34    uint16_t a, b;
35    uint32_t c;
36  } w;
37  struct {
38    uint32_t a, b;
39  } d;
40  uintptr_t p;
41  DataRefImpl() {
42    std::memset(this, 0, sizeof(DataRefImpl));
43  }
44};
45
46template<class content_type>
47class content_iterator {
48  content_type Current;
49public:
50  content_iterator(content_type symb)
51    : Current(symb) {}
52
53  const content_type* operator->() const {
54    return &Current;
55  }
56
57  const content_type &operator*() const {
58    return Current;
59  }
60
61  bool operator==(const content_iterator &other) const {
62    return Current == other.Current;
63  }
64
65  bool operator!=(const content_iterator &other) const {
66    return !(*this == other);
67  }
68
69  content_iterator& increment(error_code &err) {
70    content_type next;
71    if (error_code ec = Current.getNext(next))
72      err = ec;
73    else
74      Current = next;
75    return *this;
76  }
77};
78
79inline bool operator==(const DataRefImpl &a, const DataRefImpl &b) {
80  // Check bitwise identical. This is the only legal way to compare a union w/o
81  // knowing which member is in use.
82  return std::memcmp(&a, &b, sizeof(DataRefImpl)) == 0;
83}
84
85inline bool operator<(const DataRefImpl &a, const DataRefImpl &b) {
86  // Check bitwise identical. This is the only legal way to compare a union w/o
87  // knowing which member is in use.
88  return std::memcmp(&a, &b, sizeof(DataRefImpl)) < 0;
89}
90
91class SymbolRef;
92
93/// RelocationRef - This is a value type class that represents a single
94/// relocation in the list of relocations in the object file.
95class RelocationRef {
96  DataRefImpl RelocationPimpl;
97  const ObjectFile *OwningObject;
98
99public:
100  RelocationRef() : OwningObject(NULL) { }
101
102  RelocationRef(DataRefImpl RelocationP, const ObjectFile *Owner);
103
104  bool operator==(const RelocationRef &Other) const;
105
106  error_code getNext(RelocationRef &Result) const;
107
108  error_code getAddress(uint64_t &Result) const;
109  error_code getOffset(uint64_t &Result) const;
110  error_code getSymbol(SymbolRef &Result) const;
111  error_code getType(uint64_t &Result) const;
112
113  /// @brief Indicates whether this relocation should hidden when listing
114  /// relocations, usually because it is the trailing part of a multipart
115  /// relocation that will be printed as part of the leading relocation.
116  error_code getHidden(bool &Result) const;
117
118  /// @brief Get a string that represents the type of this relocation.
119  ///
120  /// This is for display purposes only.
121  error_code getTypeName(SmallVectorImpl<char> &Result) const;
122
123  /// @brief Get a string that represents the calculation of the value of this
124  ///        relocation.
125  ///
126  /// This is for display purposes only.
127  error_code getValueString(SmallVectorImpl<char> &Result) const;
128
129  DataRefImpl getRawDataRefImpl() const;
130  const ObjectFile *getObjectFile() const;
131};
132typedef content_iterator<RelocationRef> relocation_iterator;
133
134/// SectionRef - This is a value type class that represents a single section in
135/// the list of sections in the object file.
136class SectionRef {
137  friend class SymbolRef;
138  DataRefImpl SectionPimpl;
139  const ObjectFile *OwningObject;
140
141public:
142  SectionRef() : OwningObject(NULL) { }
143
144  SectionRef(DataRefImpl SectionP, const ObjectFile *Owner);
145
146  bool operator==(const SectionRef &Other) const;
147  bool operator<(const SectionRef &Other) const;
148
149  error_code getNext(SectionRef &Result) const;
150
151  error_code getName(StringRef &Result) const;
152  error_code getAddress(uint64_t &Result) const;
153  error_code getSize(uint64_t &Result) const;
154  error_code getContents(StringRef &Result) const;
155
156  /// @brief Get the alignment of this section as the actual value (not log 2).
157  error_code getAlignment(uint64_t &Result) const;
158
159  // FIXME: Move to the normalization layer when it's created.
160  error_code isText(bool &Result) const;
161  error_code isData(bool &Result) const;
162  error_code isBSS(bool &Result) const;
163  error_code isRequiredForExecution(bool &Result) const;
164  error_code isVirtual(bool &Result) const;
165  error_code isZeroInit(bool &Result) const;
166  error_code isReadOnlyData(bool &Result) const;
167
168  error_code containsSymbol(SymbolRef S, bool &Result) const;
169
170  relocation_iterator begin_relocations() const;
171  relocation_iterator end_relocations() const;
172
173  DataRefImpl getRawDataRefImpl() const;
174};
175typedef content_iterator<SectionRef> section_iterator;
176
177/// SymbolRef - This is a value type class that represents a single symbol in
178/// the list of symbols in the object file.
179class SymbolRef {
180  friend class SectionRef;
181  DataRefImpl SymbolPimpl;
182  const ObjectFile *OwningObject;
183
184public:
185  SymbolRef() : OwningObject(NULL) { }
186
187  enum Type {
188    ST_Unknown, // Type not specified
189    ST_Data,
190    ST_Debug,
191    ST_File,
192    ST_Function,
193    ST_Other
194  };
195
196  enum Flags {
197    SF_None            = 0,
198    SF_Undefined       = 1U << 0,  // Symbol is defined in another object file
199    SF_Global          = 1U << 1,  // Global symbol
200    SF_Weak            = 1U << 2,  // Weak symbol
201    SF_Absolute        = 1U << 3,  // Absolute symbol
202    SF_ThreadLocal     = 1U << 4,  // Thread local symbol
203    SF_Common          = 1U << 5,  // Symbol has common linkage
204    SF_FormatSpecific  = 1U << 31  // Specific to the object file format
205                                   // (e.g. section symbols)
206  };
207
208  SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner);
209
210  bool operator==(const SymbolRef &Other) const;
211  bool operator<(const SymbolRef &Other) const;
212
213  error_code getNext(SymbolRef &Result) const;
214
215  error_code getName(StringRef &Result) const;
216  /// Returns the symbol virtual address (i.e. address at which it will be
217  /// mapped).
218  error_code getAddress(uint64_t &Result) const;
219  error_code getFileOffset(uint64_t &Result) const;
220  /// @brief Get the alignment of this symbol as the actual value (not log 2).
221  error_code getAlignment(uint32_t &Result) const;
222  error_code getSize(uint64_t &Result) const;
223  error_code getType(SymbolRef::Type &Result) const;
224
225  /// Returns the ascii char that should be displayed in a symbol table dump via
226  /// nm for this symbol.
227  error_code getNMTypeChar(char &Result) const;
228
229  /// Get symbol flags (bitwise OR of SymbolRef::Flags)
230  error_code getFlags(uint32_t &Result) const;
231
232  /// @brief Get section this symbol is defined in reference to. Result is
233  /// end_sections() if it is undefined or is an absolute symbol.
234  error_code getSection(section_iterator &Result) const;
235
236  /// @brief Get value of the symbol in the symbol table.
237  error_code getValue(uint64_t &Val) const;
238
239  DataRefImpl getRawDataRefImpl() const;
240};
241typedef content_iterator<SymbolRef> symbol_iterator;
242
243/// LibraryRef - This is a value type class that represents a single library in
244/// the list of libraries needed by a shared or dynamic object.
245class LibraryRef {
246  friend class SectionRef;
247  DataRefImpl LibraryPimpl;
248  const ObjectFile *OwningObject;
249
250public:
251  LibraryRef() : OwningObject(NULL) { }
252
253  LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner);
254
255  bool operator==(const LibraryRef &Other) const;
256  bool operator<(const LibraryRef &Other) const;
257
258  error_code getNext(LibraryRef &Result) const;
259
260  // Get the path to this library, as stored in the object file.
261  error_code getPath(StringRef &Result) const;
262
263  DataRefImpl getRawDataRefImpl() const;
264};
265typedef content_iterator<LibraryRef> library_iterator;
266
267const uint64_t UnknownAddressOrSize = ~0ULL;
268
269/// ObjectFile - This class is the base class for all object file types.
270/// Concrete instances of this object are created by createObjectFile, which
271/// figures out which type to create.
272class ObjectFile : public Binary {
273  virtual void anchor();
274  ObjectFile() LLVM_DELETED_FUNCTION;
275  ObjectFile(const ObjectFile &other) LLVM_DELETED_FUNCTION;
276
277protected:
278  ObjectFile(unsigned int Type, MemoryBuffer *source);
279
280  const uint8_t *base() const {
281    return reinterpret_cast<const uint8_t *>(Data->getBufferStart());
282  }
283
284  // These functions are for SymbolRef to call internally. The main goal of
285  // this is to allow SymbolRef::SymbolPimpl to point directly to the symbol
286  // entry in the memory mapped object file. SymbolPimpl cannot contain any
287  // virtual functions because then it could not point into the memory mapped
288  // file.
289  //
290  // Implementations assume that the DataRefImpl is valid and has not been
291  // modified externally. It's UB otherwise.
292  friend class SymbolRef;
293  virtual error_code getSymbolNext(DataRefImpl Symb, SymbolRef &Res) const = 0;
294  virtual error_code getSymbolName(DataRefImpl Symb, StringRef &Res) const = 0;
295  virtual error_code getSymbolAddress(DataRefImpl Symb, uint64_t &Res) const = 0;
296  virtual error_code getSymbolFileOffset(DataRefImpl Symb, uint64_t &Res)const=0;
297  virtual error_code getSymbolAlignment(DataRefImpl Symb, uint32_t &Res) const;
298  virtual error_code getSymbolSize(DataRefImpl Symb, uint64_t &Res) const = 0;
299  virtual error_code getSymbolType(DataRefImpl Symb,
300                                   SymbolRef::Type &Res) const = 0;
301  virtual error_code getSymbolNMTypeChar(DataRefImpl Symb, char &Res) const = 0;
302  virtual error_code getSymbolFlags(DataRefImpl Symb,
303                                    uint32_t &Res) const = 0;
304  virtual error_code getSymbolSection(DataRefImpl Symb,
305                                      section_iterator &Res) const = 0;
306  virtual error_code getSymbolValue(DataRefImpl Symb, uint64_t &Val) const = 0;
307
308  // Same as above for SectionRef.
309  friend class SectionRef;
310  virtual error_code getSectionNext(DataRefImpl Sec, SectionRef &Res) const = 0;
311  virtual error_code getSectionName(DataRefImpl Sec, StringRef &Res) const = 0;
312  virtual error_code getSectionAddress(DataRefImpl Sec, uint64_t &Res) const =0;
313  virtual error_code getSectionSize(DataRefImpl Sec, uint64_t &Res) const = 0;
314  virtual error_code getSectionContents(DataRefImpl Sec, StringRef &Res)const=0;
315  virtual error_code getSectionAlignment(DataRefImpl Sec, uint64_t &Res)const=0;
316  virtual error_code isSectionText(DataRefImpl Sec, bool &Res) const = 0;
317  virtual error_code isSectionData(DataRefImpl Sec, bool &Res) const = 0;
318  virtual error_code isSectionBSS(DataRefImpl Sec, bool &Res) const = 0;
319  virtual error_code isSectionRequiredForExecution(DataRefImpl Sec,
320                                                   bool &Res) const = 0;
321  // A section is 'virtual' if its contents aren't present in the object image.
322  virtual error_code isSectionVirtual(DataRefImpl Sec, bool &Res) const = 0;
323  virtual error_code isSectionZeroInit(DataRefImpl Sec, bool &Res) const = 0;
324  virtual error_code isSectionReadOnlyData(DataRefImpl Sec, bool &Res) const =0;
325  virtual error_code sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb,
326                                           bool &Result) const = 0;
327  virtual relocation_iterator getSectionRelBegin(DataRefImpl Sec) const = 0;
328  virtual relocation_iterator getSectionRelEnd(DataRefImpl Sec) const = 0;
329
330
331  // Same as above for RelocationRef.
332  friend class RelocationRef;
333  virtual error_code getRelocationNext(DataRefImpl Rel,
334                                       RelocationRef &Res) const = 0;
335  virtual error_code getRelocationAddress(DataRefImpl Rel,
336                                          uint64_t &Res) const =0;
337  virtual error_code getRelocationOffset(DataRefImpl Rel,
338                                         uint64_t &Res) const =0;
339  virtual error_code getRelocationSymbol(DataRefImpl Rel,
340                                         SymbolRef &Res) const = 0;
341  virtual error_code getRelocationType(DataRefImpl Rel,
342                                       uint64_t &Res) const = 0;
343  virtual error_code getRelocationTypeName(DataRefImpl Rel,
344                                       SmallVectorImpl<char> &Result) const = 0;
345  virtual error_code getRelocationValueString(DataRefImpl Rel,
346                                       SmallVectorImpl<char> &Result) const = 0;
347  virtual error_code getRelocationHidden(DataRefImpl Rel, bool &Result) const {
348    Result = false;
349    return object_error::success;
350  }
351
352  // Same for LibraryRef
353  friend class LibraryRef;
354  virtual error_code getLibraryNext(DataRefImpl Lib, LibraryRef &Res) const = 0;
355  virtual error_code getLibraryPath(DataRefImpl Lib, StringRef &Res) const = 0;
356
357public:
358
359  virtual symbol_iterator begin_symbols() const = 0;
360  virtual symbol_iterator end_symbols() const = 0;
361
362  virtual symbol_iterator begin_dynamic_symbols() const = 0;
363  virtual symbol_iterator end_dynamic_symbols() const = 0;
364
365  virtual section_iterator begin_sections() const = 0;
366  virtual section_iterator end_sections() const = 0;
367
368  virtual library_iterator begin_libraries_needed() const = 0;
369  virtual library_iterator end_libraries_needed() const = 0;
370
371  /// @brief The number of bytes used to represent an address in this object
372  ///        file format.
373  virtual uint8_t getBytesInAddress() const = 0;
374
375  virtual StringRef getFileFormatName() const = 0;
376  virtual /* Triple::ArchType */ unsigned getArch() const = 0;
377
378  /// For shared objects, returns the name which this object should be
379  /// loaded from at runtime. This corresponds to DT_SONAME on ELF and
380  /// LC_ID_DYLIB (install name) on MachO.
381  virtual StringRef getLoadName() const = 0;
382
383  /// @returns Pointer to ObjectFile subclass to handle this type of object.
384  /// @param ObjectPath The path to the object file. ObjectPath.isObject must
385  ///        return true.
386  /// @brief Create ObjectFile from path.
387  static ObjectFile *createObjectFile(StringRef ObjectPath);
388  static ObjectFile *createObjectFile(MemoryBuffer *Object);
389
390  static inline bool classof(const Binary *v) {
391    return v->isObject();
392  }
393
394public:
395  static ObjectFile *createCOFFObjectFile(MemoryBuffer *Object);
396  static ObjectFile *createELFObjectFile(MemoryBuffer *Object);
397  static ObjectFile *createMachOObjectFile(MemoryBuffer *Object);
398};
399
400// Inline function definitions.
401inline SymbolRef::SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner)
402  : SymbolPimpl(SymbolP)
403  , OwningObject(Owner) {}
404
405inline bool SymbolRef::operator==(const SymbolRef &Other) const {
406  return SymbolPimpl == Other.SymbolPimpl;
407}
408
409inline bool SymbolRef::operator<(const SymbolRef &Other) const {
410  return SymbolPimpl < Other.SymbolPimpl;
411}
412
413inline error_code SymbolRef::getNext(SymbolRef &Result) const {
414  return OwningObject->getSymbolNext(SymbolPimpl, Result);
415}
416
417inline error_code SymbolRef::getName(StringRef &Result) const {
418  return OwningObject->getSymbolName(SymbolPimpl, Result);
419}
420
421inline error_code SymbolRef::getAddress(uint64_t &Result) const {
422  return OwningObject->getSymbolAddress(SymbolPimpl, Result);
423}
424
425inline error_code SymbolRef::getFileOffset(uint64_t &Result) const {
426  return OwningObject->getSymbolFileOffset(SymbolPimpl, Result);
427}
428
429inline error_code SymbolRef::getAlignment(uint32_t &Result) const {
430  return OwningObject->getSymbolAlignment(SymbolPimpl, Result);
431}
432
433inline error_code SymbolRef::getSize(uint64_t &Result) const {
434  return OwningObject->getSymbolSize(SymbolPimpl, Result);
435}
436
437inline error_code SymbolRef::getNMTypeChar(char &Result) const {
438  return OwningObject->getSymbolNMTypeChar(SymbolPimpl, Result);
439}
440
441inline error_code SymbolRef::getFlags(uint32_t &Result) const {
442  return OwningObject->getSymbolFlags(SymbolPimpl, Result);
443}
444
445inline error_code SymbolRef::getSection(section_iterator &Result) const {
446  return OwningObject->getSymbolSection(SymbolPimpl, Result);
447}
448
449inline error_code SymbolRef::getType(SymbolRef::Type &Result) const {
450  return OwningObject->getSymbolType(SymbolPimpl, Result);
451}
452
453inline error_code SymbolRef::getValue(uint64_t &Val) const {
454  return OwningObject->getSymbolValue(SymbolPimpl, Val);
455}
456
457inline DataRefImpl SymbolRef::getRawDataRefImpl() const {
458  return SymbolPimpl;
459}
460
461
462/// SectionRef
463inline SectionRef::SectionRef(DataRefImpl SectionP,
464                              const ObjectFile *Owner)
465  : SectionPimpl(SectionP)
466  , OwningObject(Owner) {}
467
468inline bool SectionRef::operator==(const SectionRef &Other) const {
469  return SectionPimpl == Other.SectionPimpl;
470}
471
472inline bool SectionRef::operator<(const SectionRef &Other) const {
473  return SectionPimpl < Other.SectionPimpl;
474}
475
476inline error_code SectionRef::getNext(SectionRef &Result) const {
477  return OwningObject->getSectionNext(SectionPimpl, Result);
478}
479
480inline error_code SectionRef::getName(StringRef &Result) const {
481  return OwningObject->getSectionName(SectionPimpl, Result);
482}
483
484inline error_code SectionRef::getAddress(uint64_t &Result) const {
485  return OwningObject->getSectionAddress(SectionPimpl, Result);
486}
487
488inline error_code SectionRef::getSize(uint64_t &Result) const {
489  return OwningObject->getSectionSize(SectionPimpl, Result);
490}
491
492inline error_code SectionRef::getContents(StringRef &Result) const {
493  return OwningObject->getSectionContents(SectionPimpl, Result);
494}
495
496inline error_code SectionRef::getAlignment(uint64_t &Result) const {
497  return OwningObject->getSectionAlignment(SectionPimpl, Result);
498}
499
500inline error_code SectionRef::isText(bool &Result) const {
501  return OwningObject->isSectionText(SectionPimpl, Result);
502}
503
504inline error_code SectionRef::isData(bool &Result) const {
505  return OwningObject->isSectionData(SectionPimpl, Result);
506}
507
508inline error_code SectionRef::isBSS(bool &Result) const {
509  return OwningObject->isSectionBSS(SectionPimpl, Result);
510}
511
512inline error_code SectionRef::isRequiredForExecution(bool &Result) const {
513  return OwningObject->isSectionRequiredForExecution(SectionPimpl, Result);
514}
515
516inline error_code SectionRef::isVirtual(bool &Result) const {
517  return OwningObject->isSectionVirtual(SectionPimpl, Result);
518}
519
520inline error_code SectionRef::isZeroInit(bool &Result) const {
521  return OwningObject->isSectionZeroInit(SectionPimpl, Result);
522}
523
524inline error_code SectionRef::isReadOnlyData(bool &Result) const {
525  return OwningObject->isSectionReadOnlyData(SectionPimpl, Result);
526}
527
528inline error_code SectionRef::containsSymbol(SymbolRef S, bool &Result) const {
529  return OwningObject->sectionContainsSymbol(SectionPimpl, S.SymbolPimpl,
530                                             Result);
531}
532
533inline relocation_iterator SectionRef::begin_relocations() const {
534  return OwningObject->getSectionRelBegin(SectionPimpl);
535}
536
537inline relocation_iterator SectionRef::end_relocations() const {
538  return OwningObject->getSectionRelEnd(SectionPimpl);
539}
540
541inline DataRefImpl SectionRef::getRawDataRefImpl() const {
542  return SectionPimpl;
543}
544
545/// RelocationRef
546inline RelocationRef::RelocationRef(DataRefImpl RelocationP,
547                              const ObjectFile *Owner)
548  : RelocationPimpl(RelocationP)
549  , OwningObject(Owner) {}
550
551inline bool RelocationRef::operator==(const RelocationRef &Other) const {
552  return RelocationPimpl == Other.RelocationPimpl;
553}
554
555inline error_code RelocationRef::getNext(RelocationRef &Result) const {
556  return OwningObject->getRelocationNext(RelocationPimpl, Result);
557}
558
559inline error_code RelocationRef::getAddress(uint64_t &Result) const {
560  return OwningObject->getRelocationAddress(RelocationPimpl, Result);
561}
562
563inline error_code RelocationRef::getOffset(uint64_t &Result) const {
564  return OwningObject->getRelocationOffset(RelocationPimpl, Result);
565}
566
567inline error_code RelocationRef::getSymbol(SymbolRef &Result) const {
568  return OwningObject->getRelocationSymbol(RelocationPimpl, Result);
569}
570
571inline error_code RelocationRef::getType(uint64_t &Result) const {
572  return OwningObject->getRelocationType(RelocationPimpl, Result);
573}
574
575inline error_code RelocationRef::getTypeName(SmallVectorImpl<char> &Result)
576  const {
577  return OwningObject->getRelocationTypeName(RelocationPimpl, Result);
578}
579
580inline error_code RelocationRef::getValueString(SmallVectorImpl<char> &Result)
581  const {
582  return OwningObject->getRelocationValueString(RelocationPimpl, Result);
583}
584
585inline error_code RelocationRef::getHidden(bool &Result) const {
586  return OwningObject->getRelocationHidden(RelocationPimpl, Result);
587}
588
589inline DataRefImpl RelocationRef::getRawDataRefImpl() const {
590  return RelocationPimpl;
591}
592
593inline const ObjectFile *RelocationRef::getObjectFile() const {
594  return OwningObject;
595}
596
597// Inline function definitions.
598inline LibraryRef::LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner)
599  : LibraryPimpl(LibraryP)
600  , OwningObject(Owner) {}
601
602inline bool LibraryRef::operator==(const LibraryRef &Other) const {
603  return LibraryPimpl == Other.LibraryPimpl;
604}
605
606inline bool LibraryRef::operator<(const LibraryRef &Other) const {
607  return LibraryPimpl < Other.LibraryPimpl;
608}
609
610inline error_code LibraryRef::getNext(LibraryRef &Result) const {
611  return OwningObject->getLibraryNext(LibraryPimpl, Result);
612}
613
614inline error_code LibraryRef::getPath(StringRef &Result) const {
615  return OwningObject->getLibraryPath(LibraryPimpl, Result);
616}
617
618} // end namespace object
619} // end namespace llvm
620
621#endif
622