ObjectFile.h revision 6f9489a86f33624f9ff5388411d12359ce9cef20 
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_OBJECT_FILE_H
15#define LLVM_OBJECT_OBJECT_FILE_H
16
17#include "llvm/Object/Binary.h"
18#include "llvm/ADT/StringRef.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};
42
43template<class content_type>
44class content_iterator {
45  content_type Current;
46public:
47  content_iterator(content_type symb)
48    : Current(symb) {}
49
50  const content_type* operator->() const {
51    return &Current;
52  }
53
54  const content_type &operator*() const {
55    return Current;
56  }
57
58  bool operator==(const content_iterator &other) const {
59    return Current == other.Current;
60  }
61
62  bool operator!=(const content_iterator &other) const {
63    return !(*this == other);
64  }
65
66  content_iterator& increment(error_code &err) {
67    content_type next;
68    if (error_code ec = Current.getNext(next))
69      err = ec;
70    else
71      Current = next;
72    return *this;
73  }
74};
75
76static bool operator ==(const DataRefImpl &a, const DataRefImpl &b) {
77  // Check bitwise identical. This is the only legal way to compare a union w/o
78  // knowing which member is in use.
79  return std::memcmp(&a, &b, sizeof(DataRefImpl)) == 0;
80}
81
82static bool operator <(const DataRefImpl &a, const DataRefImpl &b) {
83  // Check bitwise identical. This is the only legal way to compare a union w/o
84  // knowing which member is in use.
85  return std::memcmp(&a, &b, sizeof(DataRefImpl)) < 0;
86}
87
88class SymbolRef;
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    std::memset(&RelocationPimpl, 0, sizeof(RelocationPimpl));
99  }
100
101  RelocationRef(DataRefImpl RelocationP, const ObjectFile *Owner);
102
103  bool operator==(const RelocationRef &Other) const;
104
105  error_code getNext(RelocationRef &Result) const;
106
107  error_code getAddress(uint64_t &Result) const;
108  error_code getOffset(uint64_t &Result) const;
109  error_code getSymbol(SymbolRef &Result) const;
110  error_code getType(uint64_t &Result) const;
111
112  /// @brief Indicates whether this relocation should hidden when listing
113  /// relocations, usually because it is the trailing part of a multipart
114  /// relocation that will be printed as part of the leading relocation.
115  error_code getHidden(bool &Result) const;
116
117  /// @brief Get a string that represents the type of this relocation.
118  ///
119  /// This is for display purposes only.
120  error_code getTypeName(SmallVectorImpl<char> &Result) const;
121  error_code getAdditionalInfo(int64_t &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};
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 {
134  friend class SymbolRef;
135  DataRefImpl SectionPimpl;
136  const ObjectFile *OwningObject;
137
138public:
139  SectionRef() : OwningObject(NULL) {
140    std::memset(&SectionPimpl, 0, sizeof(SectionPimpl));
141  }
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
163  error_code containsSymbol(SymbolRef S, bool &Result) const;
164
165  relocation_iterator begin_relocations() const;
166  relocation_iterator end_relocations() const;
167};
168typedef content_iterator<SectionRef> section_iterator;
169
170/// SymbolRef - This is a value type class that represents a single symbol in
171/// the list of symbols in the object file.
172class SymbolRef {
173  friend class SectionRef;
174  DataRefImpl SymbolPimpl;
175  const ObjectFile *OwningObject;
176
177public:
178  SymbolRef() : OwningObject(NULL) {
179    std::memset(&SymbolPimpl, 0, sizeof(SymbolPimpl));
180  }
181
182  enum Type {
183    ST_Unknown, // Type not specified
184    ST_Data,
185    ST_Debug,
186    ST_File,
187    ST_Function,
188    ST_Other
189  };
190
191  enum Flags {
192    SF_None            = 0,
193    SF_Undefined       = 1U << 0,  // Symbol is defined in another object file
194    SF_Global          = 1U << 1,  // Global symbol
195    SF_Weak            = 1U << 2,  // Weak symbol
196    SF_Absolute        = 1U << 3,  // Absolute symbol
197    SF_ThreadLocal     = 1U << 4,  // Thread local symbol
198    SF_Common          = 1U << 5,  // Symbol has common linkage
199    SF_FormatSpecific  = 1U << 31  // Specific to the object file format
200                                   // (e.g. section symbols)
201  };
202
203  SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner);
204
205  bool operator==(const SymbolRef &Other) const;
206  bool operator <(const SymbolRef &Other) const;
207
208  error_code getNext(SymbolRef &Result) const;
209
210  error_code getName(StringRef &Result) const;
211  error_code getAddress(uint64_t &Result) const;
212  error_code getFileOffset(uint64_t &Result) const;
213  error_code getSize(uint64_t &Result) const;
214  error_code getType(SymbolRef::Type &Result) const;
215
216  /// Returns the ascii char that should be displayed in a symbol table dump via
217  /// nm for this symbol.
218  error_code getNMTypeChar(char &Result) const;
219
220  /// Get symbol flags (bitwise OR of SymbolRef::Flags)
221  error_code getFlags(uint32_t &Result) const;
222
223  /// @brief Get section this symbol is defined in reference to. Result is
224  /// end_sections() if it is undefined or is an absolute symbol.
225  error_code getSection(section_iterator &Result) const;
226
227  DataRefImpl getRawDataRefImpl() const;
228};
229typedef content_iterator<SymbolRef> symbol_iterator;
230
231/// LibraryRef - This is a value type class that represents a single library in
232/// the list of libraries needed by a shared or dynamic object.
233class LibraryRef {
234  friend class SectionRef;
235  DataRefImpl LibraryPimpl;
236  const ObjectFile *OwningObject;
237
238public:
239  LibraryRef() : OwningObject(NULL) {
240    std::memset(&LibraryPimpl, 0, sizeof(LibraryPimpl));
241  }
242
243  LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner);
244
245  bool operator==(const LibraryRef &Other) const;
246  bool operator <(const LibraryRef &Other) const;
247
248  error_code getNext(LibraryRef &Result) const;
249
250  // Get the path to this library, as stored in the object file.
251  error_code getPath(StringRef &Result) const;
252
253  DataRefImpl getRawDataRefImpl() const;
254};
255typedef content_iterator<LibraryRef> library_iterator;
256
257const uint64_t UnknownAddressOrSize = ~0ULL;
258
259/// ObjectFile - This class is the base class for all object file types.
260/// Concrete instances of this object are created by createObjectFile, which
261/// figure out which type to create.
262class ObjectFile : public Binary {
263  virtual void anchor();
264  ObjectFile(); // = delete
265  ObjectFile(const ObjectFile &other); // = delete
266
267protected:
268  ObjectFile(unsigned int Type, MemoryBuffer *source, error_code &ec);
269
270  const uint8_t *base() const {
271    return reinterpret_cast<const uint8_t *>(Data->getBufferStart());
272  }
273
274  // These functions are for SymbolRef to call internally. The main goal of
275  // this is to allow SymbolRef::SymbolPimpl to point directly to the symbol
276  // entry in the memory mapped object file. SymbolPimpl cannot contain any
277  // virtual functions because then it could not point into the memory mapped
278  // file.
279  //
280  // Implementations assume that the DataRefImpl is valid and has not been
281  // modified externally. It's UB otherwise.
282  friend class SymbolRef;
283  virtual error_code getSymbolNext(DataRefImpl Symb, SymbolRef &Res) const = 0;
284  virtual error_code getSymbolName(DataRefImpl Symb, StringRef &Res) const = 0;
285  virtual error_code getSymbolAddress(DataRefImpl Symb, uint64_t &Res) const =0;
286  virtual error_code getSymbolFileOffset(DataRefImpl Symb, uint64_t &Res) const =0;
287  virtual error_code getSymbolSize(DataRefImpl Symb, uint64_t &Res) const = 0;
288  virtual error_code getSymbolType(DataRefImpl Symb,
289                                   SymbolRef::Type &Res) const = 0;
290  virtual error_code getSymbolNMTypeChar(DataRefImpl Symb, char &Res) const = 0;
291  virtual error_code getSymbolFlags(DataRefImpl Symb,
292                                    uint32_t &Res) const = 0;
293  virtual error_code getSymbolSection(DataRefImpl Symb,
294                                      section_iterator &Res) const = 0;
295
296  // Same as above for SectionRef.
297  friend class SectionRef;
298  virtual error_code getSectionNext(DataRefImpl Sec, SectionRef &Res) const = 0;
299  virtual error_code getSectionName(DataRefImpl Sec, StringRef &Res) const = 0;
300  virtual error_code getSectionAddress(DataRefImpl Sec, uint64_t &Res) const =0;
301  virtual error_code getSectionSize(DataRefImpl Sec, uint64_t &Res) const = 0;
302  virtual error_code getSectionContents(DataRefImpl Sec, StringRef &Res)const=0;
303  virtual error_code getSectionAlignment(DataRefImpl Sec, uint64_t &Res)const=0;
304  virtual error_code isSectionText(DataRefImpl Sec, bool &Res) const = 0;
305  virtual error_code isSectionData(DataRefImpl Sec, bool &Res) const = 0;
306  virtual error_code isSectionBSS(DataRefImpl Sec, bool &Res) const = 0;
307  virtual error_code sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb,
308                                           bool &Result) const = 0;
309  virtual relocation_iterator getSectionRelBegin(DataRefImpl Sec) const = 0;
310  virtual relocation_iterator getSectionRelEnd(DataRefImpl Sec) const = 0;
311
312
313  // Same as above for RelocationRef.
314  friend class RelocationRef;
315  virtual error_code getRelocationNext(DataRefImpl Rel,
316                                       RelocationRef &Res) const = 0;
317  virtual error_code getRelocationAddress(DataRefImpl Rel,
318                                          uint64_t &Res) const =0;
319  virtual error_code getRelocationOffset(DataRefImpl Rel,
320                                         uint64_t &Res) const =0;
321  virtual error_code getRelocationSymbol(DataRefImpl Rel,
322                                         SymbolRef &Res) const = 0;
323  virtual error_code getRelocationType(DataRefImpl Rel,
324                                       uint64_t &Res) const = 0;
325  virtual error_code getRelocationTypeName(DataRefImpl Rel,
326                                       SmallVectorImpl<char> &Result) const = 0;
327  virtual error_code getRelocationAdditionalInfo(DataRefImpl Rel,
328                                                 int64_t &Res) const = 0;
329  virtual error_code getRelocationValueString(DataRefImpl Rel,
330                                       SmallVectorImpl<char> &Result) const = 0;
331  virtual error_code getRelocationHidden(DataRefImpl Rel, bool &Result) const {
332    Result = false;
333    return object_error::success;
334  }
335
336  // Same for LibraryRef
337  friend class LibraryRef;
338  virtual error_code getLibraryNext(DataRefImpl Lib, LibraryRef &Res) const = 0;
339  virtual error_code getLibraryPath(DataRefImpl Lib, StringRef &Res) const = 0;
340
341public:
342
343  virtual symbol_iterator begin_symbols() const = 0;
344  virtual symbol_iterator end_symbols() const = 0;
345
346  virtual symbol_iterator begin_dynamic_symbols() const = 0;
347  virtual symbol_iterator end_dynamic_symbols() const = 0;
348
349  virtual section_iterator begin_sections() const = 0;
350  virtual section_iterator end_sections() const = 0;
351
352  virtual library_iterator begin_libraries_needed() const = 0;
353  virtual library_iterator end_libraries_needed() const = 0;
354
355  /// @brief The number of bytes used to represent an address in this object
356  ///        file format.
357  virtual uint8_t getBytesInAddress() const = 0;
358
359  virtual StringRef getFileFormatName() const = 0;
360  virtual /* Triple::ArchType */ unsigned getArch() const = 0;
361
362  /// For shared objects, returns the name which this object should be
363  /// loaded from at runtime. This corresponds to DT_SONAME on ELF and
364  /// LC_ID_DYLIB (install name) on MachO.
365  virtual StringRef getLoadName() const = 0;
366
367  /// @returns Pointer to ObjectFile subclass to handle this type of object.
368  /// @param ObjectPath The path to the object file. ObjectPath.isObject must
369  ///        return true.
370  /// @brief Create ObjectFile from path.
371  static ObjectFile *createObjectFile(StringRef ObjectPath);
372  static ObjectFile *createObjectFile(MemoryBuffer *Object);
373
374  static inline bool classof(const Binary *v) {
375    return v->isObject();
376  }
377  static inline bool classof(const ObjectFile *v) { return true; }
378
379public:
380  static ObjectFile *createCOFFObjectFile(MemoryBuffer *Object);
381  static ObjectFile *createELFObjectFile(MemoryBuffer *Object);
382  static ObjectFile *createMachOObjectFile(MemoryBuffer *Object);
383};
384
385// Inline function definitions.
386inline SymbolRef::SymbolRef(DataRefImpl SymbolP, const ObjectFile *Owner)
387  : SymbolPimpl(SymbolP)
388  , OwningObject(Owner) {}
389
390inline bool SymbolRef::operator==(const SymbolRef &Other) const {
391  return SymbolPimpl == Other.SymbolPimpl;
392}
393
394inline bool SymbolRef::operator <(const SymbolRef &Other) const {
395  return SymbolPimpl < Other.SymbolPimpl;
396}
397
398inline error_code SymbolRef::getNext(SymbolRef &Result) const {
399  return OwningObject->getSymbolNext(SymbolPimpl, Result);
400}
401
402inline error_code SymbolRef::getName(StringRef &Result) const {
403  return OwningObject->getSymbolName(SymbolPimpl, Result);
404}
405
406inline error_code SymbolRef::getAddress(uint64_t &Result) const {
407  return OwningObject->getSymbolAddress(SymbolPimpl, Result);
408}
409
410inline error_code SymbolRef::getFileOffset(uint64_t &Result) const {
411  return OwningObject->getSymbolFileOffset(SymbolPimpl, Result);
412}
413
414inline error_code SymbolRef::getSize(uint64_t &Result) const {
415  return OwningObject->getSymbolSize(SymbolPimpl, Result);
416}
417
418inline error_code SymbolRef::getNMTypeChar(char &Result) const {
419  return OwningObject->getSymbolNMTypeChar(SymbolPimpl, Result);
420}
421
422inline error_code SymbolRef::getFlags(uint32_t &Result) const {
423  return OwningObject->getSymbolFlags(SymbolPimpl, Result);
424}
425
426inline error_code SymbolRef::getSection(section_iterator &Result) const {
427  return OwningObject->getSymbolSection(SymbolPimpl, Result);
428}
429
430inline error_code SymbolRef::getType(SymbolRef::Type &Result) const {
431  return OwningObject->getSymbolType(SymbolPimpl, Result);
432}
433
434inline DataRefImpl SymbolRef::getRawDataRefImpl() const {
435  return SymbolPimpl;
436}
437
438
439/// SectionRef
440inline SectionRef::SectionRef(DataRefImpl SectionP,
441                              const ObjectFile *Owner)
442  : SectionPimpl(SectionP)
443  , OwningObject(Owner) {}
444
445inline bool SectionRef::operator==(const SectionRef &Other) const {
446  return SectionPimpl == Other.SectionPimpl;
447}
448
449inline bool SectionRef::operator <(const SectionRef &Other) const {
450  return SectionPimpl < Other.SectionPimpl;
451}
452
453inline error_code SectionRef::getNext(SectionRef &Result) const {
454  return OwningObject->getSectionNext(SectionPimpl, Result);
455}
456
457inline error_code SectionRef::getName(StringRef &Result) const {
458  return OwningObject->getSectionName(SectionPimpl, Result);
459}
460
461inline error_code SectionRef::getAddress(uint64_t &Result) const {
462  return OwningObject->getSectionAddress(SectionPimpl, Result);
463}
464
465inline error_code SectionRef::getSize(uint64_t &Result) const {
466  return OwningObject->getSectionSize(SectionPimpl, Result);
467}
468
469inline error_code SectionRef::getContents(StringRef &Result) const {
470  return OwningObject->getSectionContents(SectionPimpl, Result);
471}
472
473inline error_code SectionRef::getAlignment(uint64_t &Result) const {
474  return OwningObject->getSectionAlignment(SectionPimpl, Result);
475}
476
477inline error_code SectionRef::isText(bool &Result) const {
478  return OwningObject->isSectionText(SectionPimpl, Result);
479}
480
481inline error_code SectionRef::isData(bool &Result) const {
482  return OwningObject->isSectionData(SectionPimpl, Result);
483}
484
485inline error_code SectionRef::isBSS(bool &Result) const {
486  return OwningObject->isSectionBSS(SectionPimpl, Result);
487}
488
489inline error_code SectionRef::containsSymbol(SymbolRef S, bool &Result) const {
490  return OwningObject->sectionContainsSymbol(SectionPimpl, S.SymbolPimpl,
491                                             Result);
492}
493
494inline relocation_iterator SectionRef::begin_relocations() const {
495  return OwningObject->getSectionRelBegin(SectionPimpl);
496}
497
498inline relocation_iterator SectionRef::end_relocations() const {
499  return OwningObject->getSectionRelEnd(SectionPimpl);
500}
501
502
503/// RelocationRef
504inline RelocationRef::RelocationRef(DataRefImpl RelocationP,
505                              const ObjectFile *Owner)
506  : RelocationPimpl(RelocationP)
507  , OwningObject(Owner) {}
508
509inline bool RelocationRef::operator==(const RelocationRef &Other) const {
510  return RelocationPimpl == Other.RelocationPimpl;
511}
512
513inline error_code RelocationRef::getNext(RelocationRef &Result) const {
514  return OwningObject->getRelocationNext(RelocationPimpl, Result);
515}
516
517inline error_code RelocationRef::getAddress(uint64_t &Result) const {
518  return OwningObject->getRelocationAddress(RelocationPimpl, Result);
519}
520
521inline error_code RelocationRef::getOffset(uint64_t &Result) const {
522  return OwningObject->getRelocationOffset(RelocationPimpl, Result);
523}
524
525inline error_code RelocationRef::getSymbol(SymbolRef &Result) const {
526  return OwningObject->getRelocationSymbol(RelocationPimpl, Result);
527}
528
529inline error_code RelocationRef::getType(uint64_t &Result) const {
530  return OwningObject->getRelocationType(RelocationPimpl, Result);
531}
532
533inline error_code RelocationRef::getTypeName(SmallVectorImpl<char> &Result)
534  const {
535  return OwningObject->getRelocationTypeName(RelocationPimpl, Result);
536}
537
538inline error_code RelocationRef::getAdditionalInfo(int64_t &Result) const {
539  return OwningObject->getRelocationAdditionalInfo(RelocationPimpl, Result);
540}
541
542inline error_code RelocationRef::getValueString(SmallVectorImpl<char> &Result)
543  const {
544  return OwningObject->getRelocationValueString(RelocationPimpl, Result);
545}
546
547inline error_code RelocationRef::getHidden(bool &Result) const {
548  return OwningObject->getRelocationHidden(RelocationPimpl, Result);
549}
550// Inline function definitions.
551inline LibraryRef::LibraryRef(DataRefImpl LibraryP, const ObjectFile *Owner)
552  : LibraryPimpl(LibraryP)
553  , OwningObject(Owner) {}
554
555inline bool LibraryRef::operator==(const LibraryRef &Other) const {
556  return LibraryPimpl == Other.LibraryPimpl;
557}
558
559inline bool LibraryRef::operator <(const LibraryRef &Other) const {
560  return LibraryPimpl < Other.LibraryPimpl;
561}
562
563inline error_code LibraryRef::getNext(LibraryRef &Result) const {
564  return OwningObject->getLibraryNext(LibraryPimpl, Result);
565}
566
567inline error_code LibraryRef::getPath(StringRef &Result) const {
568  return OwningObject->getLibraryPath(LibraryPimpl, Result);
569}
570
571} // end namespace object
572} // end namespace llvm
573
574#endif
575