• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 //===- ModuleFile.h - Module file description -------------------*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 //  This file defines the Module class, which describes a module that has
10 //  been loaded from an AST file.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #ifndef LLVM_CLANG_SERIALIZATION_MODULEFILE_H
15 #define LLVM_CLANG_SERIALIZATION_MODULEFILE_H
16 
17 #include "clang/Basic/FileManager.h"
18 #include "clang/Basic/Module.h"
19 #include "clang/Basic/SourceLocation.h"
20 #include "clang/Serialization/ASTBitCodes.h"
21 #include "clang/Serialization/ContinuousRangeMap.h"
22 #include "clang/Serialization/ModuleFileExtension.h"
23 #include "llvm/ADT/DenseMap.h"
24 #include "llvm/ADT/PointerIntPair.h"
25 #include "llvm/ADT/SetVector.h"
26 #include "llvm/ADT/SmallVector.h"
27 #include "llvm/ADT/StringRef.h"
28 #include "llvm/Bitstream/BitstreamReader.h"
29 #include "llvm/Support/Endian.h"
30 #include <cassert>
31 #include <cstdint>
32 #include <memory>
33 #include <string>
34 #include <vector>
35 
36 namespace clang {
37 
38 namespace serialization {
39 
40 /// Specifies the kind of module that has been loaded.
41 enum ModuleKind {
42   /// File is an implicitly-loaded module.
43   MK_ImplicitModule,
44 
45   /// File is an explicitly-loaded module.
46   MK_ExplicitModule,
47 
48   /// File is a PCH file treated as such.
49   MK_PCH,
50 
51   /// File is a PCH file treated as the preamble.
52   MK_Preamble,
53 
54   /// File is a PCH file treated as the actual main file.
55   MK_MainFile,
56 
57   /// File is from a prebuilt module path.
58   MK_PrebuiltModule
59 };
60 
61 /// The input file that has been loaded from this AST file, along with
62 /// bools indicating whether this was an overridden buffer or if it was
63 /// out-of-date or not-found.
64 class InputFile {
65   enum {
66     Overridden = 1,
67     OutOfDate = 2,
68     NotFound = 3
69   };
70   llvm::PointerIntPair<const FileEntryRef::MapEntry *, 2, unsigned> Val;
71 
72 public:
73   InputFile() = default;
74 
75   InputFile(FileEntryRef File, bool isOverridden = false,
76             bool isOutOfDate = false) {
77     assert(!(isOverridden && isOutOfDate) &&
78            "an overridden cannot be out-of-date");
79     unsigned intVal = 0;
80     if (isOverridden)
81       intVal = Overridden;
82     else if (isOutOfDate)
83       intVal = OutOfDate;
84     Val.setPointerAndInt(&File.getMapEntry(), intVal);
85   }
86 
getNotFound()87   static InputFile getNotFound() {
88     InputFile File;
89     File.Val.setInt(NotFound);
90     return File;
91   }
92 
getFile()93   OptionalFileEntryRefDegradesToFileEntryPtr getFile() const {
94     if (auto *P = Val.getPointer())
95       return FileEntryRef(*P);
96     return None;
97   }
isOverridden()98   bool isOverridden() const { return Val.getInt() == Overridden; }
isOutOfDate()99   bool isOutOfDate() const { return Val.getInt() == OutOfDate; }
isNotFound()100   bool isNotFound() const { return Val.getInt() == NotFound; }
101 };
102 
103 /// Information about a module that has been loaded by the ASTReader.
104 ///
105 /// Each instance of the Module class corresponds to a single AST file, which
106 /// may be a precompiled header, precompiled preamble, a module, or an AST file
107 /// of some sort loaded as the main file, all of which are specific formulations
108 /// of the general notion of a "module". A module may depend on any number of
109 /// other modules.
110 class ModuleFile {
111 public:
ModuleFile(ModuleKind Kind,unsigned Generation)112   ModuleFile(ModuleKind Kind, unsigned Generation)
113       : Kind(Kind), Generation(Generation) {}
114   ~ModuleFile();
115 
116   // === General information ===
117 
118   /// The index of this module in the list of modules.
119   unsigned Index = 0;
120 
121   /// The type of this module.
122   ModuleKind Kind;
123 
124   /// The file name of the module file.
125   std::string FileName;
126 
127   /// The name of the module.
128   std::string ModuleName;
129 
130   /// The base directory of the module.
131   std::string BaseDirectory;
132 
getTimestampFilename()133   std::string getTimestampFilename() const {
134     return FileName + ".timestamp";
135   }
136 
137   /// The original source file name that was used to build the
138   /// primary AST file, which may have been modified for
139   /// relocatable-pch support.
140   std::string OriginalSourceFileName;
141 
142   /// The actual original source file name that was used to
143   /// build this AST file.
144   std::string ActualOriginalSourceFileName;
145 
146   /// The file ID for the original source file that was used to
147   /// build this AST file.
148   FileID OriginalSourceFileID;
149 
150   /// The directory that the PCH was originally created in. Used to
151   /// allow resolving headers even after headers+PCH was moved to a new path.
152   std::string OriginalDir;
153 
154   std::string ModuleMapPath;
155 
156   /// Whether this precompiled header is a relocatable PCH file.
157   bool RelocatablePCH = false;
158 
159   /// Whether timestamps are included in this module file.
160   bool HasTimestamps = false;
161 
162   /// Whether the top-level module has been read from the AST file.
163   bool DidReadTopLevelSubmodule = false;
164 
165   /// The file entry for the module file.
166   OptionalFileEntryRefDegradesToFileEntryPtr File;
167 
168   /// The signature of the module file, which may be used instead of the size
169   /// and modification time to identify this particular file.
170   ASTFileSignature Signature;
171 
172   /// The signature of the AST block of the module file, this can be used to
173   /// unique module files based on AST contents.
174   ASTFileSignature ASTBlockHash;
175 
176   /// Whether this module has been directly imported by the
177   /// user.
178   bool DirectlyImported = false;
179 
180   /// The generation of which this module file is a part.
181   unsigned Generation;
182 
183   /// The memory buffer that stores the data associated with
184   /// this AST file, owned by the InMemoryModuleCache.
185   llvm::MemoryBuffer *Buffer;
186 
187   /// The size of this file, in bits.
188   uint64_t SizeInBits = 0;
189 
190   /// The global bit offset (or base) of this module
191   uint64_t GlobalBitOffset = 0;
192 
193   /// The bit offset of the AST block of this module.
194   uint64_t ASTBlockStartOffset = 0;
195 
196   /// The serialized bitstream data for this file.
197   StringRef Data;
198 
199   /// The main bitstream cursor for the main block.
200   llvm::BitstreamCursor Stream;
201 
202   /// The source location where the module was explicitly or implicitly
203   /// imported in the local translation unit.
204   ///
205   /// If module A depends on and imports module B, both modules will have the
206   /// same DirectImportLoc, but different ImportLoc (B's ImportLoc will be a
207   /// source location inside module A).
208   ///
209   /// WARNING: This is largely useless. It doesn't tell you when a module was
210   /// made visible, just when the first submodule of that module was imported.
211   SourceLocation DirectImportLoc;
212 
213   /// The source location where this module was first imported.
214   SourceLocation ImportLoc;
215 
216   /// The first source location in this module.
217   SourceLocation FirstLoc;
218 
219   /// The list of extension readers that are attached to this module
220   /// file.
221   std::vector<std::unique_ptr<ModuleFileExtensionReader>> ExtensionReaders;
222 
223   /// The module offset map data for this file. If non-empty, the various
224   /// ContinuousRangeMaps described below have not yet been populated.
225   StringRef ModuleOffsetMap;
226 
227   // === Input Files ===
228 
229   /// The cursor to the start of the input-files block.
230   llvm::BitstreamCursor InputFilesCursor;
231 
232   /// Offsets for all of the input file entries in the AST file.
233   const llvm::support::unaligned_uint64_t *InputFileOffsets = nullptr;
234 
235   /// The input files that have been loaded from this AST file.
236   std::vector<InputFile> InputFilesLoaded;
237 
238   // All user input files reside at the index range [0, NumUserInputFiles), and
239   // system input files reside at [NumUserInputFiles, InputFilesLoaded.size()).
240   unsigned NumUserInputFiles = 0;
241 
242   /// If non-zero, specifies the time when we last validated input
243   /// files.  Zero means we never validated them.
244   ///
245   /// The time is specified in seconds since the start of the Epoch.
246   uint64_t InputFilesValidationTimestamp = 0;
247 
248   // === Source Locations ===
249 
250   /// Cursor used to read source location entries.
251   llvm::BitstreamCursor SLocEntryCursor;
252 
253   /// The bit offset to the start of the SOURCE_MANAGER_BLOCK.
254   uint64_t SourceManagerBlockStartOffset = 0;
255 
256   /// The number of source location entries in this AST file.
257   unsigned LocalNumSLocEntries = 0;
258 
259   /// The base ID in the source manager's view of this module.
260   int SLocEntryBaseID = 0;
261 
262   /// The base offset in the source manager's view of this module.
263   unsigned SLocEntryBaseOffset = 0;
264 
265   /// Base file offset for the offsets in SLocEntryOffsets. Real file offset
266   /// for the entry is SLocEntryOffsetsBase + SLocEntryOffsets[i].
267   uint64_t SLocEntryOffsetsBase = 0;
268 
269   /// Offsets for all of the source location entries in the
270   /// AST file.
271   const uint32_t *SLocEntryOffsets = nullptr;
272 
273   /// SLocEntries that we're going to preload.
274   SmallVector<uint64_t, 4> PreloadSLocEntries;
275 
276   /// Remapping table for source locations in this module.
277   ContinuousRangeMap<uint32_t, int, 2> SLocRemap;
278 
279   // === Identifiers ===
280 
281   /// The number of identifiers in this AST file.
282   unsigned LocalNumIdentifiers = 0;
283 
284   /// Offsets into the identifier table data.
285   ///
286   /// This array is indexed by the identifier ID (-1), and provides
287   /// the offset into IdentifierTableData where the string data is
288   /// stored.
289   const uint32_t *IdentifierOffsets = nullptr;
290 
291   /// Base identifier ID for identifiers local to this module.
292   serialization::IdentID BaseIdentifierID = 0;
293 
294   /// Remapping table for identifier IDs in this module.
295   ContinuousRangeMap<uint32_t, int, 2> IdentifierRemap;
296 
297   /// Actual data for the on-disk hash table of identifiers.
298   ///
299   /// This pointer points into a memory buffer, where the on-disk hash
300   /// table for identifiers actually lives.
301   const char *IdentifierTableData = nullptr;
302 
303   /// A pointer to an on-disk hash table of opaque type
304   /// IdentifierHashTable.
305   void *IdentifierLookupTable = nullptr;
306 
307   /// Offsets of identifiers that we're going to preload within
308   /// IdentifierTableData.
309   std::vector<unsigned> PreloadIdentifierOffsets;
310 
311   // === Macros ===
312 
313   /// The cursor to the start of the preprocessor block, which stores
314   /// all of the macro definitions.
315   llvm::BitstreamCursor MacroCursor;
316 
317   /// The number of macros in this AST file.
318   unsigned LocalNumMacros = 0;
319 
320   /// Base file offset for the offsets in MacroOffsets. Real file offset for
321   /// the entry is MacroOffsetsBase + MacroOffsets[i].
322   uint64_t MacroOffsetsBase = 0;
323 
324   /// Offsets of macros in the preprocessor block.
325   ///
326   /// This array is indexed by the macro ID (-1), and provides
327   /// the offset into the preprocessor block where macro definitions are
328   /// stored.
329   const uint32_t *MacroOffsets = nullptr;
330 
331   /// Base macro ID for macros local to this module.
332   serialization::MacroID BaseMacroID = 0;
333 
334   /// Remapping table for macro IDs in this module.
335   ContinuousRangeMap<uint32_t, int, 2> MacroRemap;
336 
337   /// The offset of the start of the set of defined macros.
338   uint64_t MacroStartOffset = 0;
339 
340   // === Detailed PreprocessingRecord ===
341 
342   /// The cursor to the start of the (optional) detailed preprocessing
343   /// record block.
344   llvm::BitstreamCursor PreprocessorDetailCursor;
345 
346   /// The offset of the start of the preprocessor detail cursor.
347   uint64_t PreprocessorDetailStartOffset = 0;
348 
349   /// Base preprocessed entity ID for preprocessed entities local to
350   /// this module.
351   serialization::PreprocessedEntityID BasePreprocessedEntityID = 0;
352 
353   /// Remapping table for preprocessed entity IDs in this module.
354   ContinuousRangeMap<uint32_t, int, 2> PreprocessedEntityRemap;
355 
356   const PPEntityOffset *PreprocessedEntityOffsets = nullptr;
357   unsigned NumPreprocessedEntities = 0;
358 
359   /// Base ID for preprocessed skipped ranges local to this module.
360   unsigned BasePreprocessedSkippedRangeID = 0;
361 
362   const PPSkippedRange *PreprocessedSkippedRangeOffsets = nullptr;
363   unsigned NumPreprocessedSkippedRanges = 0;
364 
365   // === Header search information ===
366 
367   /// The number of local HeaderFileInfo structures.
368   unsigned LocalNumHeaderFileInfos = 0;
369 
370   /// Actual data for the on-disk hash table of header file
371   /// information.
372   ///
373   /// This pointer points into a memory buffer, where the on-disk hash
374   /// table for header file information actually lives.
375   const char *HeaderFileInfoTableData = nullptr;
376 
377   /// The on-disk hash table that contains information about each of
378   /// the header files.
379   void *HeaderFileInfoTable = nullptr;
380 
381   // === Submodule information ===
382 
383   /// The number of submodules in this module.
384   unsigned LocalNumSubmodules = 0;
385 
386   /// Base submodule ID for submodules local to this module.
387   serialization::SubmoduleID BaseSubmoduleID = 0;
388 
389   /// Remapping table for submodule IDs in this module.
390   ContinuousRangeMap<uint32_t, int, 2> SubmoduleRemap;
391 
392   // === Selectors ===
393 
394   /// The number of selectors new to this file.
395   ///
396   /// This is the number of entries in SelectorOffsets.
397   unsigned LocalNumSelectors = 0;
398 
399   /// Offsets into the selector lookup table's data array
400   /// where each selector resides.
401   const uint32_t *SelectorOffsets = nullptr;
402 
403   /// Base selector ID for selectors local to this module.
404   serialization::SelectorID BaseSelectorID = 0;
405 
406   /// Remapping table for selector IDs in this module.
407   ContinuousRangeMap<uint32_t, int, 2> SelectorRemap;
408 
409   /// A pointer to the character data that comprises the selector table
410   ///
411   /// The SelectorOffsets table refers into this memory.
412   const unsigned char *SelectorLookupTableData = nullptr;
413 
414   /// A pointer to an on-disk hash table of opaque type
415   /// ASTSelectorLookupTable.
416   ///
417   /// This hash table provides the IDs of all selectors, and the associated
418   /// instance and factory methods.
419   void *SelectorLookupTable = nullptr;
420 
421   // === Declarations ===
422 
423   /// DeclsCursor - This is a cursor to the start of the DECLTYPES_BLOCK block.
424   /// It has read all the abbreviations at the start of the block and is ready
425   /// to jump around with these in context.
426   llvm::BitstreamCursor DeclsCursor;
427 
428   /// The offset to the start of the DECLTYPES_BLOCK block.
429   uint64_t DeclsBlockStartOffset = 0;
430 
431   /// The number of declarations in this AST file.
432   unsigned LocalNumDecls = 0;
433 
434   /// Offset of each declaration within the bitstream, indexed
435   /// by the declaration ID (-1).
436   const DeclOffset *DeclOffsets = nullptr;
437 
438   /// Base declaration ID for declarations local to this module.
439   serialization::DeclID BaseDeclID = 0;
440 
441   /// Remapping table for declaration IDs in this module.
442   ContinuousRangeMap<uint32_t, int, 2> DeclRemap;
443 
444   /// Mapping from the module files that this module file depends on
445   /// to the base declaration ID for that module as it is understood within this
446   /// module.
447   ///
448   /// This is effectively a reverse global-to-local mapping for declaration
449   /// IDs, so that we can interpret a true global ID (for this translation unit)
450   /// as a local ID (for this module file).
451   llvm::DenseMap<ModuleFile *, serialization::DeclID> GlobalToLocalDeclIDs;
452 
453   /// Array of file-level DeclIDs sorted by file.
454   const serialization::DeclID *FileSortedDecls = nullptr;
455   unsigned NumFileSortedDecls = 0;
456 
457   /// Array of category list location information within this
458   /// module file, sorted by the definition ID.
459   const serialization::ObjCCategoriesInfo *ObjCCategoriesMap = nullptr;
460 
461   /// The number of redeclaration info entries in ObjCCategoriesMap.
462   unsigned LocalNumObjCCategoriesInMap = 0;
463 
464   /// The Objective-C category lists for categories known to this
465   /// module.
466   SmallVector<uint64_t, 1> ObjCCategories;
467 
468   // === Types ===
469 
470   /// The number of types in this AST file.
471   unsigned LocalNumTypes = 0;
472 
473   /// Offset of each type within the bitstream, indexed by the
474   /// type ID, or the representation of a Type*.
475   const UnderalignedInt64 *TypeOffsets = nullptr;
476 
477   /// Base type ID for types local to this module as represented in
478   /// the global type ID space.
479   serialization::TypeID BaseTypeIndex = 0;
480 
481   /// Remapping table for type IDs in this module.
482   ContinuousRangeMap<uint32_t, int, 2> TypeRemap;
483 
484   // === Miscellaneous ===
485 
486   /// Diagnostic IDs and their mappings that the user changed.
487   SmallVector<uint64_t, 8> PragmaDiagMappings;
488 
489   /// List of modules which depend on this module
490   llvm::SetVector<ModuleFile *> ImportedBy;
491 
492   /// List of modules which this module depends on
493   llvm::SetVector<ModuleFile *> Imports;
494 
495   /// Determine whether this module was directly imported at
496   /// any point during translation.
isDirectlyImported()497   bool isDirectlyImported() const { return DirectlyImported; }
498 
499   /// Is this a module file for a module (rather than a PCH or similar).
isModule()500   bool isModule() const {
501     return Kind == MK_ImplicitModule || Kind == MK_ExplicitModule ||
502            Kind == MK_PrebuiltModule;
503   }
504 
505   /// Dump debugging output for this module.
506   void dump();
507 };
508 
509 } // namespace serialization
510 
511 } // namespace clang
512 
513 #endif // LLVM_CLANG_SERIALIZATION_MODULEFILE_H
514