ObjectFile.h revision bc884fd9f7bdb64d250be639edc8dc85a20a1975 
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  // This entire union should probably be a
32  // char[max(8, sizeof(uintptr_t))] and require the impl to cast.
33  struct {
34    uint32_t a, b;
35  } d;
36  uintptr_t p;
37  DataRefImpl() {
38    std::memset(this, 0, sizeof(DataRefImpl));
39  }
40};
41
42template<class content_type>
43class content_iterator {
44  content_type Current;
45public:
46  content_iterator(content_type symb)
47    : Current(symb) {}
48
49  const content_type* operator->() const {
50    return &Current;
51  }
52
53  const content_type &operator*() const {
54    return Current;
55  }
56
57  bool operator==(const content_iterator &other) const {
58    return Current == other.Current;
59  }
60
61  bool operator!=(const content_iterator &other) const {
62    return !(*this == other);
63  }
64
65  content_iterator& increment(error_code &err) {
66    content_type next;
67    if (error_code ec = Current.getNext(next))
68      err = ec;
69    else
70      Current = next;
71    return *this;
72  }
73};
74
75inline bool operator==(const DataRefImpl &a, const DataRefImpl &b) {
76  // Check bitwise identical. This is the only legal way to compare a union w/o
77  // knowing which member is in use.
78  return std::memcmp(&a, &b, sizeof(DataRefImpl)) == 0;
79}
80
81inline bool operator<(const DataRefImpl &a, const DataRefImpl &b) {
82  // Check bitwise identical. This is the only legal way to compare a union w/o
83  // knowing which member is in use.
84  return std::memcmp(&a, &b, sizeof(DataRefImpl)) < 0;
85}
86
87class SymbolRef;
88typedef content_iterator<SymbolRef> symbol_iterator;
89
90/// RelocationRef - This is a value type class that represents a single
91/// relocation in the list of relocations in the object file.
92class RelocationRef {
93  DataRefImpl RelocationPimpl;
94  const ObjectFile *OwningObject;
95
96public:
97  RelocationRef() : OwningObject(NULL) { }
98
99  RelocationRef(DataRefImpl RelocationP, const ObjectFile *Owner);
100
101  bool operator==(const RelocationRef &Other) const;
102
103  error_code getNext(RelocationRef &Result) const;
104
105  error_code getAddress(uint64_t &Result) const;
106  error_code getOffset(uint64_t &Result) const;
107  symbol_iterator getSymbol() const;
108  error_code getType(uint64_t &Result) const;
109
110  /// @brief Indicates whether this relocation should hidden when listing
111  /// relocations, usually because it is the trailing part of a multipart
112  /// relocation that will be printed as part of the leading relocation.
113  error_code getHidden(bool &Result) const;
114
115  /// @brief Get a string that represents the type of this relocation.
116  ///
117  /// This is for display purposes only.
118  error_code getTypeName(SmallVectorImpl<char> &Result) const;
119
120  /// @brief Get a string that represents the calculation of the value of this
121  ///        relocation.
122  ///
123  /// This is for display purposes only.
124  error_code getValueString(SmallVectorImpl<char> &Result) const;
125
126  DataRefImpl getRawDataRefImpl() const;
127  const ObjectFile *getObjectFile() const;
128};
129typedef content_iterator<RelocationRef> relocation_iterator;
130
131/// SectionRef - This is a value type class that represents a single section in
132/// the list of sections in the object file.
133class SectionRef;
134typedef content_iterator<SectionRef> section_iterator;
135class SectionRef {
136  friend class SymbolRef;
137  DataRefImpl SectionPimpl;
138  const ObjectFile *OwningObject;
139
140public:
141  SectionRef() : OwningObject(NULL) { }
142
143  SectionRef(DataRefImpl SectionP, const ObjectFile *Owner);
144
145  bool operator==(const SectionRef &Other) const;
146  bool operator<(const SectionRef &Other) const;
147
148  error_code getNext(SectionRef &Result) const;
149
150  error_code getName(StringRef &Result) const;
151  error_code getAddress(uint64_t &Result) const;
152  error_code getSize(uint64_t &Result) const;
153  error_code getContents(StringRef &Result) const;
154
155  /// @brief Get the alignment of this section as the actual value (not log 2).
156  error_code getAlignment(uint64_t &Result) const;
157
158  // FIXME: Move to the normalization layer when it's created.
159  error_code isText(bool &Result) const;
160  error_code isData(bool &Result) const;
161  error_code isBSS(bool &Result) const;
162  error_code isRequiredForExecution(bool &Result) const;
163  error_code isVirtual(bool &Result) const;
164  error_code isZeroInit(bool &Result) const;
165  error_code isReadOnlyData(bool &Result) const;
166
167  error_code containsSymbol(SymbolRef S, bool &Result) const;
168
169  relocation_iterator begin_relocations() const;
170  relocation_iterator end_relocations() const;
171  section_iterator getRelocatedSection() const;
172
173  DataRefImpl getRawDataRefImpl() const;
174};
175
176/// SymbolRef - This is a value type class that represents a single symbol in
177/// the list of symbols in the object file.
178class SymbolRef {
179  friend class SectionRef;
180  DataRefImpl SymbolPimpl;
181  const ObjectFile *OwningObject;
182
183public:
184  SymbolRef() : OwningObject(NULL) { }
185
186  enum Type {
187    ST_Unknown, // Type not specified
188    ST_Data,
189    ST_Debug,
190    ST_File,
191    ST_Function,
192    ST_Other
193  };
194
195  enum Flags LLVM_ENUM_INT_TYPE(unsigned) {
196    SF_None            = 0,
197    SF_Undefined       = 1U << 0,  // Symbol is defined in another object file
198    SF_Global          = 1U << 1,  // Global symbol
199    SF_Weak            = 1U << 2,  // Weak symbol
200    SF_Absolute        = 1U << 3,  // Absolute symbol
201    SF_ThreadLocal     = 1U << 4,  // Thread local symbol
202    SF_Common          = 1U << 5,  // Symbol has common linkage
203    SF_FormatSpecific  = 1U << 31  // Specific to the object file format
204                                   // (e.g. section symbols)
205  };
206
207  SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner);
208
209  bool operator==(const SymbolRef &Other) const;
210  bool operator<(const SymbolRef &Other) const;
211
212  error_code getNext(SymbolRef &Result) const;
213
214  error_code getName(StringRef &Result) const;
215  /// Returns the symbol virtual address (i.e. address at which it will be
216  /// mapped).
217  error_code getAddress(uint64_t &Result) const;
218  error_code getFileOffset(uint64_t &Result) const;
219  /// @brief Get the alignment of this symbol as the actual value (not log 2).
220  error_code getAlignment(uint32_t &Result) const;
221  error_code getSize(uint64_t &Result) const;
222  error_code getType(SymbolRef::Type &Result) const;
223
224  /// Get symbol flags (bitwise OR of SymbolRef::Flags)
225  error_code getFlags(uint32_t &Result) const;
226
227  /// @brief Get section this symbol is defined in reference to. Result is
228  /// end_sections() if it is undefined or is an absolute symbol.
229  error_code getSection(section_iterator &Result) const;
230
231  /// @brief Get value of the symbol in the symbol table.
232  error_code getValue(uint64_t &Val) const;
233
234  DataRefImpl getRawDataRefImpl() const;
235};
236
237/// LibraryRef - This is a value type class that represents a single library in
238/// the list of libraries needed by a shared or dynamic object.
239class LibraryRef {
240  friend class SectionRef;
241  DataRefImpl LibraryPimpl;
242  const ObjectFile *OwningObject;
243
244public:
245  LibraryRef() : OwningObject(NULL) { }
246
247  LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner);
248
249  bool operator==(const LibraryRef &Other) const;
250  bool operator<(const LibraryRef &Other) const;
251
252  error_code getNext(LibraryRef &Result) const;
253
254  // Get the path to this library, as stored in the object file.
255  error_code getPath(StringRef &Result) const;
256
257  DataRefImpl getRawDataRefImpl() const;
258};
259typedef content_iterator<LibraryRef> library_iterator;
260
261const uint64_t UnknownAddressOrSize = ~0ULL;
262
263/// ObjectFile - This class is the base class for all object file types.
264/// Concrete instances of this object are created by createObjectFile, which
265/// figures out which type to create.
266class ObjectFile : public Binary {
267  virtual void anchor();
268  ObjectFile() LLVM_DELETED_FUNCTION;
269  ObjectFile(const ObjectFile &other) LLVM_DELETED_FUNCTION;
270
271protected:
272  ObjectFile(unsigned int Type, MemoryBuffer *source);
273
274  const uint8_t *base() const {
275    return reinterpret_cast<const uint8_t *>(Data->getBufferStart());
276  }
277
278  // These functions are for SymbolRef to call internally. The main goal of
279  // this is to allow SymbolRef::SymbolPimpl to point directly to the symbol
280  // entry in the memory mapped object file. SymbolPimpl cannot contain any
281  // virtual functions because then it could not point into the memory mapped
282  // file.
283  //
284  // Implementations assume that the DataRefImpl is valid and has not been
285  // modified externally. It's UB otherwise.
286  friend class SymbolRef;
287  virtual error_code getSymbolNext(DataRefImpl Symb, SymbolRef &Res) const = 0;
288  virtual error_code getSymbolName(DataRefImpl Symb, StringRef &Res) const = 0;
289  virtual error_code getSymbolAddress(DataRefImpl Symb, uint64_t &Res) const = 0;
290  virtual error_code getSymbolFileOffset(DataRefImpl Symb, uint64_t &Res)const=0;
291  virtual error_code getSymbolAlignment(DataRefImpl Symb, uint32_t &Res) const;
292  virtual error_code getSymbolSize(DataRefImpl Symb, uint64_t &Res) const = 0;
293  virtual error_code getSymbolType(DataRefImpl Symb,
294                                   SymbolRef::Type &Res) const = 0;
295  virtual error_code getSymbolFlags(DataRefImpl Symb,
296                                    uint32_t &Res) const = 0;
297  virtual error_code getSymbolSection(DataRefImpl Symb,
298                                      section_iterator &Res) const = 0;
299  virtual error_code getSymbolValue(DataRefImpl Symb, uint64_t &Val) const = 0;
300
301  // Same as above for SectionRef.
302  friend class SectionRef;
303  virtual error_code getSectionNext(DataRefImpl Sec, SectionRef &Res) const = 0;
304  virtual error_code getSectionName(DataRefImpl Sec, StringRef &Res) const = 0;
305  virtual error_code getSectionAddress(DataRefImpl Sec, uint64_t &Res) const =0;
306  virtual error_code getSectionSize(DataRefImpl Sec, uint64_t &Res) const = 0;
307  virtual error_code getSectionContents(DataRefImpl Sec, StringRef &Res)const=0;
308  virtual error_code getSectionAlignment(DataRefImpl Sec, uint64_t &Res)const=0;
309  virtual error_code isSectionText(DataRefImpl Sec, bool &Res) const = 0;
310  virtual error_code isSectionData(DataRefImpl Sec, bool &Res) const = 0;
311  virtual error_code isSectionBSS(DataRefImpl Sec, bool &Res) const = 0;
312  virtual error_code isSectionRequiredForExecution(DataRefImpl Sec,
313                                                   bool &Res) const = 0;
314  // A section is 'virtual' if its contents aren't present in the object image.
315  virtual error_code isSectionVirtual(DataRefImpl Sec, bool &Res) const = 0;
316  virtual error_code isSectionZeroInit(DataRefImpl Sec, bool &Res) const = 0;
317  virtual error_code isSectionReadOnlyData(DataRefImpl Sec, bool &Res) const =0;
318  virtual error_code sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb,
319                                           bool &Result) const = 0;
320  virtual relocation_iterator section_rel_begin(DataRefImpl Sec) const = 0;
321  virtual relocation_iterator section_rel_end(DataRefImpl Sec) const = 0;
322  virtual section_iterator getRelocatedSection(DataRefImpl Sec) const;
323
324  // Same as above for RelocationRef.
325  friend class RelocationRef;
326  virtual error_code getRelocationNext(DataRefImpl Rel,
327                                       RelocationRef &Res) const = 0;
328  virtual error_code getRelocationAddress(DataRefImpl Rel,
329                                          uint64_t &Res) const =0;
330  virtual error_code getRelocationOffset(DataRefImpl Rel,
331                                         uint64_t &Res) const =0;
332  virtual symbol_iterator getRelocationSymbol(DataRefImpl Rel) const = 0;
333  virtual error_code getRelocationType(DataRefImpl Rel,
334                                       uint64_t &Res) const = 0;
335  virtual error_code getRelocationTypeName(DataRefImpl Rel,
336                                       SmallVectorImpl<char> &Result) const = 0;
337  virtual error_code getRelocationValueString(DataRefImpl Rel,
338                                       SmallVectorImpl<char> &Result) const = 0;
339  virtual error_code getRelocationHidden(DataRefImpl Rel, bool &Result) const {
340    Result = false;
341    return object_error::success;
342  }
343
344  // Same for LibraryRef
345  friend class LibraryRef;
346  virtual error_code getLibraryNext(DataRefImpl Lib, LibraryRef &Res) const = 0;
347  virtual error_code getLibraryPath(DataRefImpl Lib, StringRef &Res) const = 0;
348
349public:
350
351  virtual symbol_iterator begin_symbols() const = 0;
352  virtual symbol_iterator end_symbols() const = 0;
353
354  virtual symbol_iterator begin_dynamic_symbols() const = 0;
355  virtual symbol_iterator end_dynamic_symbols() const = 0;
356
357  virtual section_iterator begin_sections() const = 0;
358  virtual section_iterator end_sections() const = 0;
359
360  virtual library_iterator begin_libraries_needed() const = 0;
361  virtual library_iterator end_libraries_needed() const = 0;
362
363  /// @brief The number of bytes used to represent an address in this object
364  ///        file format.
365  virtual uint8_t getBytesInAddress() const = 0;
366
367  virtual StringRef getFileFormatName() const = 0;
368  virtual /* Triple::ArchType */ unsigned getArch() const = 0;
369
370  /// For shared objects, returns the name which this object should be
371  /// loaded from at runtime. This corresponds to DT_SONAME on ELF and
372  /// LC_ID_DYLIB (install name) on MachO.
373  virtual StringRef getLoadName() const = 0;
374
375  /// @returns Pointer to ObjectFile subclass to handle this type of object.
376  /// @param ObjectPath The path to the object file. ObjectPath.isObject must
377  ///        return true.
378  /// @brief Create ObjectFile from path.
379  static ObjectFile *createObjectFile(StringRef ObjectPath);
380  static ObjectFile *createObjectFile(MemoryBuffer *Object);
381
382  static inline bool classof(const Binary *v) {
383    return v->isObject();
384  }
385
386public:
387  static ObjectFile *createCOFFObjectFile(MemoryBuffer *Object);
388  static ObjectFile *createELFObjectFile(MemoryBuffer *Object);
389  static ObjectFile *createMachOObjectFile(MemoryBuffer *Object);
390};
391
392// Inline function definitions.
393inline SymbolRef::SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner)
394  : SymbolPimpl(SymbolP)
395  , OwningObject(Owner) {}
396
397inline bool SymbolRef::operator==(const SymbolRef &Other) const {
398  return SymbolPimpl == Other.SymbolPimpl;
399}
400
401inline bool SymbolRef::operator<(const SymbolRef &Other) const {
402  return SymbolPimpl < Other.SymbolPimpl;
403}
404
405inline error_code SymbolRef::getNext(SymbolRef &Result) const {
406  return OwningObject->getSymbolNext(SymbolPimpl, Result);
407}
408
409inline error_code SymbolRef::getName(StringRef &Result) const {
410  return OwningObject->getSymbolName(SymbolPimpl, Result);
411}
412
413inline error_code SymbolRef::getAddress(uint64_t &Result) const {
414  return OwningObject->getSymbolAddress(SymbolPimpl, Result);
415}
416
417inline error_code SymbolRef::getFileOffset(uint64_t &Result) const {
418  return OwningObject->getSymbolFileOffset(SymbolPimpl, Result);
419}
420
421inline error_code SymbolRef::getAlignment(uint32_t &Result) const {
422  return OwningObject->getSymbolAlignment(SymbolPimpl, Result);
423}
424
425inline error_code SymbolRef::getSize(uint64_t &Result) const {
426  return OwningObject->getSymbolSize(SymbolPimpl, Result);
427}
428
429inline error_code SymbolRef::getFlags(uint32_t &Result) const {
430  return OwningObject->getSymbolFlags(SymbolPimpl, Result);
431}
432
433inline error_code SymbolRef::getSection(section_iterator &Result) const {
434  return OwningObject->getSymbolSection(SymbolPimpl, Result);
435}
436
437inline error_code SymbolRef::getType(SymbolRef::Type &Result) const {
438  return OwningObject->getSymbolType(SymbolPimpl, Result);
439}
440
441inline error_code SymbolRef::getValue(uint64_t &Val) const {
442  return OwningObject->getSymbolValue(SymbolPimpl, Val);
443}
444
445inline DataRefImpl SymbolRef::getRawDataRefImpl() const {
446  return SymbolPimpl;
447}
448
449
450/// SectionRef
451inline SectionRef::SectionRef(DataRefImpl SectionP,
452                              const ObjectFile *Owner)
453  : SectionPimpl(SectionP)
454  , OwningObject(Owner) {}
455
456inline bool SectionRef::operator==(const SectionRef &Other) const {
457  return SectionPimpl == Other.SectionPimpl;
458}
459
460inline bool SectionRef::operator<(const SectionRef &Other) const {
461  return SectionPimpl < Other.SectionPimpl;
462}
463
464inline error_code SectionRef::getNext(SectionRef &Result) const {
465  return OwningObject->getSectionNext(SectionPimpl, Result);
466}
467
468inline error_code SectionRef::getName(StringRef &Result) const {
469  return OwningObject->getSectionName(SectionPimpl, Result);
470}
471
472inline error_code SectionRef::getAddress(uint64_t &Result) const {
473  return OwningObject->getSectionAddress(SectionPimpl, Result);
474}
475
476inline error_code SectionRef::getSize(uint64_t &Result) const {
477  return OwningObject->getSectionSize(SectionPimpl, Result);
478}
479
480inline error_code SectionRef::getContents(StringRef &Result) const {
481  return OwningObject->getSectionContents(SectionPimpl, Result);
482}
483
484inline error_code SectionRef::getAlignment(uint64_t &Result) const {
485  return OwningObject->getSectionAlignment(SectionPimpl, Result);
486}
487
488inline error_code SectionRef::isText(bool &Result) const {
489  return OwningObject->isSectionText(SectionPimpl, Result);
490}
491
492inline error_code SectionRef::isData(bool &Result) const {
493  return OwningObject->isSectionData(SectionPimpl, Result);
494}
495
496inline error_code SectionRef::isBSS(bool &Result) const {
497  return OwningObject->isSectionBSS(SectionPimpl, Result);
498}
499
500inline error_code SectionRef::isRequiredForExecution(bool &Result) const {
501  return OwningObject->isSectionRequiredForExecution(SectionPimpl, Result);
502}
503
504inline error_code SectionRef::isVirtual(bool &Result) const {
505  return OwningObject->isSectionVirtual(SectionPimpl, Result);
506}
507
508inline error_code SectionRef::isZeroInit(bool &Result) const {
509  return OwningObject->isSectionZeroInit(SectionPimpl, Result);
510}
511
512inline error_code SectionRef::isReadOnlyData(bool &Result) const {
513  return OwningObject->isSectionReadOnlyData(SectionPimpl, Result);
514}
515
516inline error_code SectionRef::containsSymbol(SymbolRef S, bool &Result) const {
517  return OwningObject->sectionContainsSymbol(SectionPimpl, S.SymbolPimpl,
518                                             Result);
519}
520
521inline relocation_iterator SectionRef::begin_relocations() const {
522  return OwningObject->section_rel_begin(SectionPimpl);
523}
524
525inline relocation_iterator SectionRef::end_relocations() const {
526  return OwningObject->section_rel_end(SectionPimpl);
527}
528
529inline section_iterator SectionRef::getRelocatedSection() const {
530  return OwningObject->getRelocatedSection(SectionPimpl);
531}
532
533inline DataRefImpl SectionRef::getRawDataRefImpl() const {
534  return SectionPimpl;
535}
536
537/// RelocationRef
538inline RelocationRef::RelocationRef(DataRefImpl RelocationP,
539                              const ObjectFile *Owner)
540  : RelocationPimpl(RelocationP)
541  , OwningObject(Owner) {}
542
543inline bool RelocationRef::operator==(const RelocationRef &Other) const {
544  return RelocationPimpl == Other.RelocationPimpl;
545}
546
547inline error_code RelocationRef::getNext(RelocationRef &Result) const {
548  return OwningObject->getRelocationNext(RelocationPimpl, Result);
549}
550
551inline error_code RelocationRef::getAddress(uint64_t &Result) const {
552  return OwningObject->getRelocationAddress(RelocationPimpl, Result);
553}
554
555inline error_code RelocationRef::getOffset(uint64_t &Result) const {
556  return OwningObject->getRelocationOffset(RelocationPimpl, Result);
557}
558
559inline symbol_iterator RelocationRef::getSymbol() const {
560  return OwningObject->getRelocationSymbol(RelocationPimpl);
561}
562
563inline error_code RelocationRef::getType(uint64_t &Result) const {
564  return OwningObject->getRelocationType(RelocationPimpl, Result);
565}
566
567inline error_code RelocationRef::getTypeName(SmallVectorImpl<char> &Result)
568  const {
569  return OwningObject->getRelocationTypeName(RelocationPimpl, Result);
570}
571
572inline error_code RelocationRef::getValueString(SmallVectorImpl<char> &Result)
573  const {
574  return OwningObject->getRelocationValueString(RelocationPimpl, Result);
575}
576
577inline error_code RelocationRef::getHidden(bool &Result) const {
578  return OwningObject->getRelocationHidden(RelocationPimpl, Result);
579}
580
581inline DataRefImpl RelocationRef::getRawDataRefImpl() const {
582  return RelocationPimpl;
583}
584
585inline const ObjectFile *RelocationRef::getObjectFile() const {
586  return OwningObject;
587}
588
589// Inline function definitions.
590inline LibraryRef::LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner)
591  : LibraryPimpl(LibraryP)
592  , OwningObject(Owner) {}
593
594inline bool LibraryRef::operator==(const LibraryRef &Other) const {
595  return LibraryPimpl == Other.LibraryPimpl;
596}
597
598inline bool LibraryRef::operator<(const LibraryRef &Other) const {
599  return LibraryPimpl < Other.LibraryPimpl;
600}
601
602inline error_code LibraryRef::getNext(LibraryRef &Result) const {
603  return OwningObject->getLibraryNext(LibraryPimpl, Result);
604}
605
606inline error_code LibraryRef::getPath(StringRef &Result) const {
607  return OwningObject->getLibraryPath(LibraryPimpl, Result);
608}
609
610} // end namespace object
611} // end namespace llvm
612
613#endif
614