ObjectFile.h revision 6c1202c459ffa6d693ad92fa84e43902bc780bca 
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 {
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  /// Returns the ascii char that should be displayed in a symbol table dump via
225  /// nm for this symbol.
226  error_code getNMTypeChar(char &Result) const;
227
228  /// Get symbol flags (bitwise OR of SymbolRef::Flags)
229  error_code getFlags(uint32_t &Result) const;
230
231  /// @brief Get section this symbol is defined in reference to. Result is
232  /// end_sections() if it is undefined or is an absolute symbol.
233  error_code getSection(section_iterator &Result) const;
234
235  /// @brief Get value of the symbol in the symbol table.
236  error_code getValue(uint64_t &Val) const;
237
238  DataRefImpl getRawDataRefImpl() const;
239};
240
241/// LibraryRef - This is a value type class that represents a single library in
242/// the list of libraries needed by a shared or dynamic object.
243class LibraryRef {
244  friend class SectionRef;
245  DataRefImpl LibraryPimpl;
246  const ObjectFile *OwningObject;
247
248public:
249  LibraryRef() : OwningObject(NULL) { }
250
251  LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner);
252
253  bool operator==(const LibraryRef &Other) const;
254  bool operator<(const LibraryRef &Other) const;
255
256  error_code getNext(LibraryRef &Result) const;
257
258  // Get the path to this library, as stored in the object file.
259  error_code getPath(StringRef &Result) const;
260
261  DataRefImpl getRawDataRefImpl() const;
262};
263typedef content_iterator<LibraryRef> library_iterator;
264
265const uint64_t UnknownAddressOrSize = ~0ULL;
266
267/// ObjectFile - This class is the base class for all object file types.
268/// Concrete instances of this object are created by createObjectFile, which
269/// figures out which type to create.
270class ObjectFile : public Binary {
271  virtual void anchor();
272  ObjectFile() LLVM_DELETED_FUNCTION;
273  ObjectFile(const ObjectFile &other) LLVM_DELETED_FUNCTION;
274
275protected:
276  ObjectFile(unsigned int Type, MemoryBuffer *source);
277
278  const uint8_t *base() const {
279    return reinterpret_cast<const uint8_t *>(Data->getBufferStart());
280  }
281
282  // These functions are for SymbolRef to call internally. The main goal of
283  // this is to allow SymbolRef::SymbolPimpl to point directly to the symbol
284  // entry in the memory mapped object file. SymbolPimpl cannot contain any
285  // virtual functions because then it could not point into the memory mapped
286  // file.
287  //
288  // Implementations assume that the DataRefImpl is valid and has not been
289  // modified externally. It's UB otherwise.
290  friend class SymbolRef;
291  virtual error_code getSymbolNext(DataRefImpl Symb, SymbolRef &Res) const = 0;
292  virtual error_code getSymbolName(DataRefImpl Symb, StringRef &Res) const = 0;
293  virtual error_code getSymbolAddress(DataRefImpl Symb, uint64_t &Res) const = 0;
294  virtual error_code getSymbolFileOffset(DataRefImpl Symb, uint64_t &Res)const=0;
295  virtual error_code getSymbolAlignment(DataRefImpl Symb, uint32_t &Res) const;
296  virtual error_code getSymbolSize(DataRefImpl Symb, uint64_t &Res) const = 0;
297  virtual error_code getSymbolType(DataRefImpl Symb,
298                                   SymbolRef::Type &Res) const = 0;
299  virtual error_code getSymbolNMTypeChar(DataRefImpl Symb, char &Res) const = 0;
300  virtual error_code getSymbolFlags(DataRefImpl Symb,
301                                    uint32_t &Res) const = 0;
302  virtual error_code getSymbolSection(DataRefImpl Symb,
303                                      section_iterator &Res) const = 0;
304  virtual error_code getSymbolValue(DataRefImpl Symb, uint64_t &Val) const = 0;
305
306  // Same as above for SectionRef.
307  friend class SectionRef;
308  virtual error_code getSectionNext(DataRefImpl Sec, SectionRef &Res) const = 0;
309  virtual error_code getSectionName(DataRefImpl Sec, StringRef &Res) const = 0;
310  virtual error_code getSectionAddress(DataRefImpl Sec, uint64_t &Res) const =0;
311  virtual error_code getSectionSize(DataRefImpl Sec, uint64_t &Res) const = 0;
312  virtual error_code getSectionContents(DataRefImpl Sec, StringRef &Res)const=0;
313  virtual error_code getSectionAlignment(DataRefImpl Sec, uint64_t &Res)const=0;
314  virtual error_code isSectionText(DataRefImpl Sec, bool &Res) const = 0;
315  virtual error_code isSectionData(DataRefImpl Sec, bool &Res) const = 0;
316  virtual error_code isSectionBSS(DataRefImpl Sec, bool &Res) const = 0;
317  virtual error_code isSectionRequiredForExecution(DataRefImpl Sec,
318                                                   bool &Res) const = 0;
319  // A section is 'virtual' if its contents aren't present in the object image.
320  virtual error_code isSectionVirtual(DataRefImpl Sec, bool &Res) const = 0;
321  virtual error_code isSectionZeroInit(DataRefImpl Sec, bool &Res) const = 0;
322  virtual error_code isSectionReadOnlyData(DataRefImpl Sec, bool &Res) const =0;
323  virtual error_code sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb,
324                                           bool &Result) const = 0;
325  virtual relocation_iterator getSectionRelBegin(DataRefImpl Sec) const = 0;
326  virtual relocation_iterator getSectionRelEnd(DataRefImpl Sec) const = 0;
327  virtual section_iterator getRelocatedSection(DataRefImpl Sec) const;
328
329  // Same as above for RelocationRef.
330  friend class RelocationRef;
331  virtual error_code getRelocationNext(DataRefImpl Rel,
332                                       RelocationRef &Res) const = 0;
333  virtual error_code getRelocationAddress(DataRefImpl Rel,
334                                          uint64_t &Res) const =0;
335  virtual error_code getRelocationOffset(DataRefImpl Rel,
336                                         uint64_t &Res) const =0;
337  virtual symbol_iterator getRelocationSymbol(DataRefImpl Rel) const = 0;
338  virtual error_code getRelocationType(DataRefImpl Rel,
339                                       uint64_t &Res) const = 0;
340  virtual error_code getRelocationTypeName(DataRefImpl Rel,
341                                       SmallVectorImpl<char> &Result) const = 0;
342  virtual error_code getRelocationValueString(DataRefImpl Rel,
343                                       SmallVectorImpl<char> &Result) const = 0;
344  virtual error_code getRelocationHidden(DataRefImpl Rel, bool &Result) const {
345    Result = false;
346    return object_error::success;
347  }
348
349  // Same for LibraryRef
350  friend class LibraryRef;
351  virtual error_code getLibraryNext(DataRefImpl Lib, LibraryRef &Res) const = 0;
352  virtual error_code getLibraryPath(DataRefImpl Lib, StringRef &Res) const = 0;
353
354public:
355
356  virtual symbol_iterator begin_symbols() const = 0;
357  virtual symbol_iterator end_symbols() const = 0;
358
359  virtual symbol_iterator begin_dynamic_symbols() const = 0;
360  virtual symbol_iterator end_dynamic_symbols() const = 0;
361
362  virtual section_iterator begin_sections() const = 0;
363  virtual section_iterator end_sections() const = 0;
364
365  virtual library_iterator begin_libraries_needed() const = 0;
366  virtual library_iterator end_libraries_needed() const = 0;
367
368  /// @brief The number of bytes used to represent an address in this object
369  ///        file format.
370  virtual uint8_t getBytesInAddress() const = 0;
371
372  virtual StringRef getFileFormatName() const = 0;
373  virtual /* Triple::ArchType */ unsigned getArch() const = 0;
374
375  /// For shared objects, returns the name which this object should be
376  /// loaded from at runtime. This corresponds to DT_SONAME on ELF and
377  /// LC_ID_DYLIB (install name) on MachO.
378  virtual StringRef getLoadName() const = 0;
379
380  /// @returns Pointer to ObjectFile subclass to handle this type of object.
381  /// @param ObjectPath The path to the object file. ObjectPath.isObject must
382  ///        return true.
383  /// @brief Create ObjectFile from path.
384  static ObjectFile *createObjectFile(StringRef ObjectPath);
385  static ObjectFile *createObjectFile(MemoryBuffer *Object);
386
387  static inline bool classof(const Binary *v) {
388    return v->isObject();
389  }
390
391public:
392  static ObjectFile *createCOFFObjectFile(MemoryBuffer *Object);
393  static ObjectFile *createELFObjectFile(MemoryBuffer *Object);
394  static ObjectFile *createMachOObjectFile(MemoryBuffer *Object);
395};
396
397// Inline function definitions.
398inline SymbolRef::SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner)
399  : SymbolPimpl(SymbolP)
400  , OwningObject(Owner) {}
401
402inline bool SymbolRef::operator==(const SymbolRef &Other) const {
403  return SymbolPimpl == Other.SymbolPimpl;
404}
405
406inline bool SymbolRef::operator<(const SymbolRef &Other) const {
407  return SymbolPimpl < Other.SymbolPimpl;
408}
409
410inline error_code SymbolRef::getNext(SymbolRef &Result) const {
411  return OwningObject->getSymbolNext(SymbolPimpl, Result);
412}
413
414inline error_code SymbolRef::getName(StringRef &Result) const {
415  return OwningObject->getSymbolName(SymbolPimpl, Result);
416}
417
418inline error_code SymbolRef::getAddress(uint64_t &Result) const {
419  return OwningObject->getSymbolAddress(SymbolPimpl, Result);
420}
421
422inline error_code SymbolRef::getFileOffset(uint64_t &Result) const {
423  return OwningObject->getSymbolFileOffset(SymbolPimpl, Result);
424}
425
426inline error_code SymbolRef::getAlignment(uint32_t &Result) const {
427  return OwningObject->getSymbolAlignment(SymbolPimpl, Result);
428}
429
430inline error_code SymbolRef::getSize(uint64_t &Result) const {
431  return OwningObject->getSymbolSize(SymbolPimpl, Result);
432}
433
434inline error_code SymbolRef::getNMTypeChar(char &Result) const {
435  return OwningObject->getSymbolNMTypeChar(SymbolPimpl, Result);
436}
437
438inline error_code SymbolRef::getFlags(uint32_t &Result) const {
439  return OwningObject->getSymbolFlags(SymbolPimpl, Result);
440}
441
442inline error_code SymbolRef::getSection(section_iterator &Result) const {
443  return OwningObject->getSymbolSection(SymbolPimpl, Result);
444}
445
446inline error_code SymbolRef::getType(SymbolRef::Type &Result) const {
447  return OwningObject->getSymbolType(SymbolPimpl, Result);
448}
449
450inline error_code SymbolRef::getValue(uint64_t &Val) const {
451  return OwningObject->getSymbolValue(SymbolPimpl, Val);
452}
453
454inline DataRefImpl SymbolRef::getRawDataRefImpl() const {
455  return SymbolPimpl;
456}
457
458
459/// SectionRef
460inline SectionRef::SectionRef(DataRefImpl SectionP,
461                              const ObjectFile *Owner)
462  : SectionPimpl(SectionP)
463  , OwningObject(Owner) {}
464
465inline bool SectionRef::operator==(const SectionRef &Other) const {
466  return SectionPimpl == Other.SectionPimpl;
467}
468
469inline bool SectionRef::operator<(const SectionRef &Other) const {
470  return SectionPimpl < Other.SectionPimpl;
471}
472
473inline error_code SectionRef::getNext(SectionRef &Result) const {
474  return OwningObject->getSectionNext(SectionPimpl, Result);
475}
476
477inline error_code SectionRef::getName(StringRef &Result) const {
478  return OwningObject->getSectionName(SectionPimpl, Result);
479}
480
481inline error_code SectionRef::getAddress(uint64_t &Result) const {
482  return OwningObject->getSectionAddress(SectionPimpl, Result);
483}
484
485inline error_code SectionRef::getSize(uint64_t &Result) const {
486  return OwningObject->getSectionSize(SectionPimpl, Result);
487}
488
489inline error_code SectionRef::getContents(StringRef &Result) const {
490  return OwningObject->getSectionContents(SectionPimpl, Result);
491}
492
493inline error_code SectionRef::getAlignment(uint64_t &Result) const {
494  return OwningObject->getSectionAlignment(SectionPimpl, Result);
495}
496
497inline error_code SectionRef::isText(bool &Result) const {
498  return OwningObject->isSectionText(SectionPimpl, Result);
499}
500
501inline error_code SectionRef::isData(bool &Result) const {
502  return OwningObject->isSectionData(SectionPimpl, Result);
503}
504
505inline error_code SectionRef::isBSS(bool &Result) const {
506  return OwningObject->isSectionBSS(SectionPimpl, Result);
507}
508
509inline error_code SectionRef::isRequiredForExecution(bool &Result) const {
510  return OwningObject->isSectionRequiredForExecution(SectionPimpl, Result);
511}
512
513inline error_code SectionRef::isVirtual(bool &Result) const {
514  return OwningObject->isSectionVirtual(SectionPimpl, Result);
515}
516
517inline error_code SectionRef::isZeroInit(bool &Result) const {
518  return OwningObject->isSectionZeroInit(SectionPimpl, Result);
519}
520
521inline error_code SectionRef::isReadOnlyData(bool &Result) const {
522  return OwningObject->isSectionReadOnlyData(SectionPimpl, Result);
523}
524
525inline error_code SectionRef::containsSymbol(SymbolRef S, bool &Result) const {
526  return OwningObject->sectionContainsSymbol(SectionPimpl, S.SymbolPimpl,
527                                             Result);
528}
529
530inline relocation_iterator SectionRef::begin_relocations() const {
531  return OwningObject->getSectionRelBegin(SectionPimpl);
532}
533
534inline relocation_iterator SectionRef::end_relocations() const {
535  return OwningObject->getSectionRelEnd(SectionPimpl);
536}
537
538inline section_iterator SectionRef::getRelocatedSection() const {
539  return OwningObject->getRelocatedSection(SectionPimpl);
540}
541
542inline DataRefImpl SectionRef::getRawDataRefImpl() const {
543  return SectionPimpl;
544}
545
546/// RelocationRef
547inline RelocationRef::RelocationRef(DataRefImpl RelocationP,
548                              const ObjectFile *Owner)
549  : RelocationPimpl(RelocationP)
550  , OwningObject(Owner) {}
551
552inline bool RelocationRef::operator==(const RelocationRef &Other) const {
553  return RelocationPimpl == Other.RelocationPimpl;
554}
555
556inline error_code RelocationRef::getNext(RelocationRef &Result) const {
557  return OwningObject->getRelocationNext(RelocationPimpl, Result);
558}
559
560inline error_code RelocationRef::getAddress(uint64_t &Result) const {
561  return OwningObject->getRelocationAddress(RelocationPimpl, Result);
562}
563
564inline error_code RelocationRef::getOffset(uint64_t &Result) const {
565  return OwningObject->getRelocationOffset(RelocationPimpl, Result);
566}
567
568inline symbol_iterator RelocationRef::getSymbol() const {
569  return OwningObject->getRelocationSymbol(RelocationPimpl);
570}
571
572inline error_code RelocationRef::getType(uint64_t &Result) const {
573  return OwningObject->getRelocationType(RelocationPimpl, Result);
574}
575
576inline error_code RelocationRef::getTypeName(SmallVectorImpl<char> &Result)
577  const {
578  return OwningObject->getRelocationTypeName(RelocationPimpl, Result);
579}
580
581inline error_code RelocationRef::getValueString(SmallVectorImpl<char> &Result)
582  const {
583  return OwningObject->getRelocationValueString(RelocationPimpl, Result);
584}
585
586inline error_code RelocationRef::getHidden(bool &Result) const {
587  return OwningObject->getRelocationHidden(RelocationPimpl, Result);
588}
589
590inline DataRefImpl RelocationRef::getRawDataRefImpl() const {
591  return RelocationPimpl;
592}
593
594inline const ObjectFile *RelocationRef::getObjectFile() const {
595  return OwningObject;
596}
597
598// Inline function definitions.
599inline LibraryRef::LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner)
600  : LibraryPimpl(LibraryP)
601  , OwningObject(Owner) {}
602
603inline bool LibraryRef::operator==(const LibraryRef &Other) const {
604  return LibraryPimpl == Other.LibraryPimpl;
605}
606
607inline bool LibraryRef::operator<(const LibraryRef &Other) const {
608  return LibraryPimpl < Other.LibraryPimpl;
609}
610
611inline error_code LibraryRef::getNext(LibraryRef &Result) const {
612  return OwningObject->getLibraryNext(LibraryPimpl, Result);
613}
614
615inline error_code LibraryRef::getPath(StringRef &Result) const {
616  return OwningObject->getLibraryPath(LibraryPimpl, Result);
617}
618
619} // end namespace object
620} // end namespace llvm
621
622#endif
623