1 | //===- ASTWriter.h - AST File Writer ----------------------------*- 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 ASTWriter class, which writes an AST file |
10 | // containing a serialized representation of a translation unit. |
11 | // |
12 | //===----------------------------------------------------------------------===// |
13 | |
14 | #ifndef LLVM_CLANG_SERIALIZATION_ASTWRITER_H |
15 | #define LLVM_CLANG_SERIALIZATION_ASTWRITER_H |
16 | |
17 | #include "clang/AST/ASTMutationListener.h" |
18 | #include "clang/AST/Decl.h" |
19 | #include "clang/AST/Type.h" |
20 | #include "clang/Basic/LLVM.h" |
21 | #include "clang/Basic/Module.h" |
22 | #include "clang/Basic/SourceLocation.h" |
23 | #include "clang/Sema/Sema.h" |
24 | #include "clang/Sema/SemaConsumer.h" |
25 | #include "clang/Serialization/ASTBitCodes.h" |
26 | #include "clang/Serialization/ASTDeserializationListener.h" |
27 | #include "clang/Serialization/PCHContainerOperations.h" |
28 | #include "clang/Serialization/SourceLocationEncoding.h" |
29 | #include "llvm/ADT/ArrayRef.h" |
30 | #include "llvm/ADT/DenseMap.h" |
31 | #include "llvm/ADT/DenseSet.h" |
32 | #include "llvm/ADT/MapVector.h" |
33 | #include "llvm/ADT/STLExtras.h" |
34 | #include "llvm/ADT/SetVector.h" |
35 | #include "llvm/ADT/SmallVector.h" |
36 | #include "llvm/ADT/StringRef.h" |
37 | #include "llvm/Bitstream/BitstreamWriter.h" |
38 | #include <cassert> |
39 | #include <cstddef> |
40 | #include <cstdint> |
41 | #include <ctime> |
42 | #include <memory> |
43 | #include <queue> |
44 | #include <string> |
45 | #include <utility> |
46 | #include <vector> |
47 | |
48 | namespace clang { |
49 | |
50 | class ASTContext; |
51 | class ASTReader; |
52 | class Attr; |
53 | class CXXRecordDecl; |
54 | class FileEntry; |
55 | class FPOptionsOverride; |
56 | class FunctionDecl; |
57 | class ; |
58 | class ; |
59 | class IdentifierResolver; |
60 | class LangOptions; |
61 | class MacroDefinitionRecord; |
62 | class MacroInfo; |
63 | class Module; |
64 | class InMemoryModuleCache; |
65 | class ModuleFileExtension; |
66 | class ModuleFileExtensionWriter; |
67 | class NamedDecl; |
68 | class ObjCInterfaceDecl; |
69 | class PreprocessingRecord; |
70 | class Preprocessor; |
71 | class RecordDecl; |
72 | class Sema; |
73 | class SourceManager; |
74 | class Stmt; |
75 | class StoredDeclsList; |
76 | class SwitchCase; |
77 | class Token; |
78 | |
79 | /// Writes an AST file containing the contents of a translation unit. |
80 | /// |
81 | /// The ASTWriter class produces a bitstream containing the serialized |
82 | /// representation of a given abstract syntax tree and its supporting |
83 | /// data structures. This bitstream can be de-serialized via an |
84 | /// instance of the ASTReader class. |
85 | class ASTWriter : public ASTDeserializationListener, |
86 | public ASTMutationListener { |
87 | public: |
88 | friend class ASTDeclWriter; |
89 | friend class ASTRecordWriter; |
90 | |
91 | using RecordData = SmallVector<uint64_t, 64>; |
92 | using RecordDataImpl = SmallVectorImpl<uint64_t>; |
93 | using RecordDataRef = ArrayRef<uint64_t>; |
94 | |
95 | private: |
96 | /// Map that provides the ID numbers of each type within the |
97 | /// output stream, plus those deserialized from a chained PCH. |
98 | /// |
99 | /// The ID numbers of types are consecutive (in order of discovery) |
100 | /// and start at 1. 0 is reserved for NULL. When types are actually |
101 | /// stored in the stream, the ID number is shifted by 2 bits to |
102 | /// allow for the const/volatile qualifiers. |
103 | /// |
104 | /// Keys in the map never have const/volatile qualifiers. |
105 | using TypeIdxMap = llvm::DenseMap<QualType, serialization::TypeIdx, |
106 | serialization::UnsafeQualTypeDenseMapInfo>; |
107 | |
108 | using LocSeq = SourceLocationSequence; |
109 | |
110 | /// The bitstream writer used to emit this precompiled header. |
111 | llvm::BitstreamWriter &Stream; |
112 | |
113 | /// The buffer associated with the bitstream. |
114 | const SmallVectorImpl<char> &Buffer; |
115 | |
116 | /// The PCM manager which manages memory buffers for pcm files. |
117 | InMemoryModuleCache &ModuleCache; |
118 | |
119 | /// The ASTContext we're writing. |
120 | ASTContext *Context = nullptr; |
121 | |
122 | /// The preprocessor we're writing. |
123 | Preprocessor *PP = nullptr; |
124 | |
125 | /// The reader of existing AST files, if we're chaining. |
126 | ASTReader *Chain = nullptr; |
127 | |
128 | /// The module we're currently writing, if any. |
129 | Module *WritingModule = nullptr; |
130 | |
131 | /// The byte range representing all the UNHASHED_CONTROL_BLOCK. |
132 | std::pair<uint64_t, uint64_t> UnhashedControlBlockRange; |
133 | /// The bit offset of the AST block hash blob. |
134 | uint64_t ASTBlockHashOffset = 0; |
135 | /// The bit offset of the signature blob. |
136 | uint64_t SignatureOffset = 0; |
137 | |
138 | /// The bit offset of the first bit inside the AST_BLOCK. |
139 | uint64_t ASTBlockStartOffset = 0; |
140 | |
141 | /// The byte range representing all the AST_BLOCK. |
142 | std::pair<uint64_t, uint64_t> ASTBlockRange; |
143 | |
144 | /// The base directory for any relative paths we emit. |
145 | std::string BaseDirectory; |
146 | |
147 | /// Indicates whether timestamps should be written to the produced |
148 | /// module file. This is the case for files implicitly written to the |
149 | /// module cache, where we need the timestamps to determine if the module |
150 | /// file is up to date, but not otherwise. |
151 | bool IncludeTimestamps; |
152 | |
153 | /// Indicates whether the AST file being written is an implicit module. |
154 | /// If that's the case, we may be able to skip writing some information that |
155 | /// are guaranteed to be the same in the importer by the context hash. |
156 | bool BuildingImplicitModule = false; |
157 | |
158 | /// Indicates when the AST writing is actively performing |
159 | /// serialization, rather than just queueing updates. |
160 | bool WritingAST = false; |
161 | |
162 | /// Indicates that we are done serializing the collection of decls |
163 | /// and types to emit. |
164 | bool DoneWritingDeclsAndTypes = false; |
165 | |
166 | /// Indicates that the AST contained compiler errors. |
167 | bool ASTHasCompilerErrors = false; |
168 | |
169 | /// Indicates that we're going to generate the reduced BMI for C++20 |
170 | /// named modules. |
171 | bool GeneratingReducedBMI = false; |
172 | |
173 | /// Mapping from input file entries to the index into the |
174 | /// offset table where information about that input file is stored. |
175 | llvm::DenseMap<const FileEntry *, uint32_t> InputFileIDs; |
176 | |
177 | /// Stores a declaration or a type to be written to the AST file. |
178 | class DeclOrType { |
179 | public: |
180 | DeclOrType(Decl *D) : Stored(D), IsType(false) {} |
181 | DeclOrType(QualType T) : Stored(T.getAsOpaquePtr()), IsType(true) {} |
182 | |
183 | bool isType() const { return IsType; } |
184 | bool isDecl() const { return !IsType; } |
185 | |
186 | QualType getType() const { |
187 | assert(isType() && "Not a type!" ); |
188 | return QualType::getFromOpaquePtr(Ptr: Stored); |
189 | } |
190 | |
191 | Decl *getDecl() const { |
192 | assert(isDecl() && "Not a decl!" ); |
193 | return static_cast<Decl *>(Stored); |
194 | } |
195 | |
196 | private: |
197 | void *Stored; |
198 | bool IsType; |
199 | }; |
200 | |
201 | /// The declarations and types to emit. |
202 | std::queue<DeclOrType> DeclTypesToEmit; |
203 | |
204 | /// The delayed namespace to emit. Only meaningful for reduced BMI. |
205 | /// |
206 | /// In reduced BMI, we want to elide the unreachable declarations in |
207 | /// the global module fragment. However, in ASTWriterDecl, when we see |
208 | /// a namespace, all the declarations in the namespace would be emitted. |
209 | /// So the optimization become meaningless. To solve the issue, we |
210 | /// delay recording all the declarations until we emit all the declarations. |
211 | /// Then we can safely record the reached declarations only. |
212 | llvm::SmallVector<NamespaceDecl *, 16> DelayedNamespace; |
213 | |
214 | /// The first ID number we can use for our own declarations. |
215 | serialization::DeclID FirstDeclID = serialization::NUM_PREDEF_DECL_IDS; |
216 | |
217 | /// The decl ID that will be assigned to the next new decl. |
218 | serialization::DeclID NextDeclID = FirstDeclID; |
219 | |
220 | /// Map that provides the ID numbers of each declaration within |
221 | /// the output stream, as well as those deserialized from a chained PCH. |
222 | /// |
223 | /// The ID numbers of declarations are consecutive (in order of |
224 | /// discovery) and start at 2. 1 is reserved for the translation |
225 | /// unit, while 0 is reserved for NULL. |
226 | llvm::DenseMap<const Decl *, serialization::DeclID> DeclIDs; |
227 | |
228 | /// Offset of each declaration in the bitstream, indexed by |
229 | /// the declaration's ID. |
230 | std::vector<serialization::DeclOffset> DeclOffsets; |
231 | |
232 | /// The offset of the DECLTYPES_BLOCK. The offsets in DeclOffsets |
233 | /// are relative to this value. |
234 | uint64_t DeclTypesBlockStartOffset = 0; |
235 | |
236 | /// Sorted (by file offset) vector of pairs of file offset/DeclID. |
237 | using LocDeclIDsTy = |
238 | SmallVector<std::pair<unsigned, serialization::DeclID>, 64>; |
239 | struct DeclIDInFileInfo { |
240 | LocDeclIDsTy DeclIDs; |
241 | |
242 | /// Set when the DeclIDs vectors from all files are joined, this |
243 | /// indicates the index that this particular vector has in the global one. |
244 | unsigned FirstDeclIndex; |
245 | }; |
246 | using FileDeclIDsTy = |
247 | llvm::DenseMap<FileID, std::unique_ptr<DeclIDInFileInfo>>; |
248 | |
249 | /// Map from file SLocEntries to info about the file-level declarations |
250 | /// that it contains. |
251 | FileDeclIDsTy FileDeclIDs; |
252 | |
253 | void associateDeclWithFile(const Decl *D, serialization::DeclID); |
254 | |
255 | /// The first ID number we can use for our own types. |
256 | serialization::TypeID FirstTypeID = serialization::NUM_PREDEF_TYPE_IDS; |
257 | |
258 | /// The type ID that will be assigned to the next new type. |
259 | serialization::TypeID NextTypeID = FirstTypeID; |
260 | |
261 | /// Map that provides the ID numbers of each type within the |
262 | /// output stream, plus those deserialized from a chained PCH. |
263 | /// |
264 | /// The ID numbers of types are consecutive (in order of discovery) |
265 | /// and start at 1. 0 is reserved for NULL. When types are actually |
266 | /// stored in the stream, the ID number is shifted by 2 bits to |
267 | /// allow for the const/volatile qualifiers. |
268 | /// |
269 | /// Keys in the map never have const/volatile qualifiers. |
270 | TypeIdxMap TypeIdxs; |
271 | |
272 | /// Offset of each type in the bitstream, indexed by |
273 | /// the type's ID. |
274 | std::vector<serialization::UnderalignedInt64> TypeOffsets; |
275 | |
276 | /// The first ID number we can use for our own identifiers. |
277 | serialization::IdentID FirstIdentID = serialization::NUM_PREDEF_IDENT_IDS; |
278 | |
279 | /// The identifier ID that will be assigned to the next new identifier. |
280 | serialization::IdentID NextIdentID = FirstIdentID; |
281 | |
282 | /// Map that provides the ID numbers of each identifier in |
283 | /// the output stream. |
284 | /// |
285 | /// The ID numbers for identifiers are consecutive (in order of |
286 | /// discovery), starting at 1. An ID of zero refers to a NULL |
287 | /// IdentifierInfo. |
288 | llvm::MapVector<const IdentifierInfo *, serialization::IdentID> IdentifierIDs; |
289 | |
290 | /// The first ID number we can use for our own macros. |
291 | serialization::MacroID FirstMacroID = serialization::NUM_PREDEF_MACRO_IDS; |
292 | |
293 | /// The identifier ID that will be assigned to the next new identifier. |
294 | serialization::MacroID NextMacroID = FirstMacroID; |
295 | |
296 | /// Map that provides the ID numbers of each macro. |
297 | llvm::DenseMap<MacroInfo *, serialization::MacroID> MacroIDs; |
298 | |
299 | struct MacroInfoToEmitData { |
300 | const IdentifierInfo *Name; |
301 | MacroInfo *MI; |
302 | serialization::MacroID ID; |
303 | }; |
304 | |
305 | /// The macro infos to emit. |
306 | std::vector<MacroInfoToEmitData> MacroInfosToEmit; |
307 | |
308 | llvm::DenseMap<const IdentifierInfo *, uint32_t> |
309 | IdentMacroDirectivesOffsetMap; |
310 | |
311 | /// @name FlushStmt Caches |
312 | /// @{ |
313 | |
314 | /// Set of parent Stmts for the currently serializing sub-stmt. |
315 | llvm::DenseSet<Stmt *> ParentStmts; |
316 | |
317 | /// Offsets of sub-stmts already serialized. The offset points |
318 | /// just after the stmt record. |
319 | llvm::DenseMap<Stmt *, uint64_t> SubStmtEntries; |
320 | |
321 | /// @} |
322 | |
323 | /// Offsets of each of the identifier IDs into the identifier |
324 | /// table. |
325 | std::vector<uint32_t> IdentifierOffsets; |
326 | |
327 | /// The first ID number we can use for our own submodules. |
328 | serialization::SubmoduleID FirstSubmoduleID = |
329 | serialization::NUM_PREDEF_SUBMODULE_IDS; |
330 | |
331 | /// The submodule ID that will be assigned to the next new submodule. |
332 | serialization::SubmoduleID NextSubmoduleID = FirstSubmoduleID; |
333 | |
334 | /// The first ID number we can use for our own selectors. |
335 | serialization::SelectorID FirstSelectorID = |
336 | serialization::NUM_PREDEF_SELECTOR_IDS; |
337 | |
338 | /// The selector ID that will be assigned to the next new selector. |
339 | serialization::SelectorID NextSelectorID = FirstSelectorID; |
340 | |
341 | /// Map that provides the ID numbers of each Selector. |
342 | llvm::MapVector<Selector, serialization::SelectorID> SelectorIDs; |
343 | |
344 | /// Offset of each selector within the method pool/selector |
345 | /// table, indexed by the Selector ID (-1). |
346 | std::vector<uint32_t> SelectorOffsets; |
347 | |
348 | /// Mapping from macro definitions (as they occur in the preprocessing |
349 | /// record) to the macro IDs. |
350 | llvm::DenseMap<const MacroDefinitionRecord *, |
351 | serialization::PreprocessedEntityID> MacroDefinitions; |
352 | |
353 | /// Cache of indices of anonymous declarations within their lexical |
354 | /// contexts. |
355 | llvm::DenseMap<const Decl *, unsigned> AnonymousDeclarationNumbers; |
356 | |
357 | /// An update to a Decl. |
358 | class DeclUpdate { |
359 | /// A DeclUpdateKind. |
360 | unsigned Kind; |
361 | union { |
362 | const Decl *Dcl; |
363 | void *Type; |
364 | SourceLocation::UIntTy Loc; |
365 | unsigned Val; |
366 | Module *Mod; |
367 | const Attr *Attribute; |
368 | }; |
369 | |
370 | public: |
371 | DeclUpdate(unsigned Kind) : Kind(Kind), Dcl(nullptr) {} |
372 | DeclUpdate(unsigned Kind, const Decl *Dcl) : Kind(Kind), Dcl(Dcl) {} |
373 | DeclUpdate(unsigned Kind, QualType Type) |
374 | : Kind(Kind), Type(Type.getAsOpaquePtr()) {} |
375 | DeclUpdate(unsigned Kind, SourceLocation Loc) |
376 | : Kind(Kind), Loc(Loc.getRawEncoding()) {} |
377 | DeclUpdate(unsigned Kind, unsigned Val) : Kind(Kind), Val(Val) {} |
378 | DeclUpdate(unsigned Kind, Module *M) : Kind(Kind), Mod(M) {} |
379 | DeclUpdate(unsigned Kind, const Attr *Attribute) |
380 | : Kind(Kind), Attribute(Attribute) {} |
381 | |
382 | unsigned getKind() const { return Kind; } |
383 | const Decl *getDecl() const { return Dcl; } |
384 | QualType getType() const { return QualType::getFromOpaquePtr(Ptr: Type); } |
385 | |
386 | SourceLocation getLoc() const { |
387 | return SourceLocation::getFromRawEncoding(Encoding: Loc); |
388 | } |
389 | |
390 | unsigned getNumber() const { return Val; } |
391 | Module *getModule() const { return Mod; } |
392 | const Attr *getAttr() const { return Attribute; } |
393 | }; |
394 | |
395 | using UpdateRecord = SmallVector<DeclUpdate, 1>; |
396 | using DeclUpdateMap = llvm::MapVector<const Decl *, UpdateRecord>; |
397 | |
398 | /// Mapping from declarations that came from a chained PCH to the |
399 | /// record containing modifications to them. |
400 | DeclUpdateMap DeclUpdates; |
401 | |
402 | /// DeclUpdates added during parsing the GMF. We split these from |
403 | /// DeclUpdates since we want to add these updates in GMF on need. |
404 | /// Only meaningful for reduced BMI. |
405 | DeclUpdateMap DeclUpdatesFromGMF; |
406 | |
407 | using FirstLatestDeclMap = llvm::DenseMap<Decl *, Decl *>; |
408 | |
409 | /// Map of first declarations from a chained PCH that point to the |
410 | /// most recent declarations in another PCH. |
411 | FirstLatestDeclMap FirstLatestDecls; |
412 | |
413 | /// Declarations encountered that might be external |
414 | /// definitions. |
415 | /// |
416 | /// We keep track of external definitions and other 'interesting' declarations |
417 | /// as we are emitting declarations to the AST file. The AST file contains a |
418 | /// separate record for these declarations, which are provided to the AST |
419 | /// consumer by the AST reader. This is behavior is required to properly cope with, |
420 | /// e.g., tentative variable definitions that occur within |
421 | /// headers. The declarations themselves are stored as declaration |
422 | /// IDs, since they will be written out to an EAGERLY_DESERIALIZED_DECLS |
423 | /// record. |
424 | SmallVector<serialization::DeclID, 16> EagerlyDeserializedDecls; |
425 | SmallVector<serialization::DeclID, 16> ModularCodegenDecls; |
426 | |
427 | /// DeclContexts that have received extensions since their serialized |
428 | /// form. |
429 | /// |
430 | /// For namespaces, when we're chaining and encountering a namespace, we check |
431 | /// if its primary namespace comes from the chain. If it does, we add the |
432 | /// primary to this set, so that we can write out lexical content updates for |
433 | /// it. |
434 | llvm::SmallSetVector<const DeclContext *, 16> UpdatedDeclContexts; |
435 | |
436 | /// Keeps track of declarations that we must emit, even though we're |
437 | /// not guaranteed to be able to find them by walking the AST starting at the |
438 | /// translation unit. |
439 | SmallVector<const Decl *, 16> DeclsToEmitEvenIfUnreferenced; |
440 | |
441 | /// The set of Objective-C class that have categories we |
442 | /// should serialize. |
443 | llvm::SetVector<ObjCInterfaceDecl *> ObjCClassesWithCategories; |
444 | |
445 | /// The set of declarations that may have redeclaration chains that |
446 | /// need to be serialized. |
447 | llvm::SmallVector<const Decl *, 16> Redeclarations; |
448 | |
449 | /// A cache of the first local declaration for "interesting" |
450 | /// redeclaration chains. |
451 | llvm::DenseMap<const Decl *, const Decl *> FirstLocalDeclCache; |
452 | |
453 | /// Mapping from SwitchCase statements to IDs. |
454 | llvm::DenseMap<SwitchCase *, unsigned> SwitchCaseIDs; |
455 | |
456 | /// The number of statements written to the AST file. |
457 | unsigned NumStatements = 0; |
458 | |
459 | /// The number of macros written to the AST file. |
460 | unsigned NumMacros = 0; |
461 | |
462 | /// The number of lexical declcontexts written to the AST |
463 | /// file. |
464 | unsigned NumLexicalDeclContexts = 0; |
465 | |
466 | /// The number of visible declcontexts written to the AST |
467 | /// file. |
468 | unsigned NumVisibleDeclContexts = 0; |
469 | |
470 | /// A mapping from each known submodule to its ID number, which will |
471 | /// be a positive integer. |
472 | llvm::DenseMap<const Module *, unsigned> SubmoduleIDs; |
473 | |
474 | /// A list of the module file extension writers. |
475 | std::vector<std::unique_ptr<ModuleFileExtensionWriter>> |
476 | ModuleFileExtensionWriters; |
477 | |
478 | /// Mapping from a source location entry to whether it is affecting or not. |
479 | llvm::BitVector IsSLocAffecting; |
480 | |
481 | /// Mapping from \c FileID to an index into the FileID adjustment table. |
482 | std::vector<FileID> NonAffectingFileIDs; |
483 | std::vector<unsigned> NonAffectingFileIDAdjustments; |
484 | |
485 | /// Mapping from an offset to an index into the offset adjustment table. |
486 | std::vector<SourceRange> NonAffectingRanges; |
487 | std::vector<SourceLocation::UIntTy> NonAffectingOffsetAdjustments; |
488 | |
489 | /// Computes input files that didn't affect compilation of the current module, |
490 | /// and initializes data structures necessary for leaving those files out |
491 | /// during \c SourceManager serialization. |
492 | void computeNonAffectingInputFiles(); |
493 | |
494 | /// Returns an adjusted \c FileID, accounting for any non-affecting input |
495 | /// files. |
496 | FileID getAdjustedFileID(FileID FID) const; |
497 | /// Returns an adjusted number of \c FileIDs created within the specified \c |
498 | /// FileID, accounting for any non-affecting input files. |
499 | unsigned getAdjustedNumCreatedFIDs(FileID FID) const; |
500 | /// Returns an adjusted \c SourceLocation, accounting for any non-affecting |
501 | /// input files. |
502 | SourceLocation getAdjustedLocation(SourceLocation Loc) const; |
503 | /// Returns an adjusted \c SourceRange, accounting for any non-affecting input |
504 | /// files. |
505 | SourceRange getAdjustedRange(SourceRange Range) const; |
506 | /// Returns an adjusted \c SourceLocation offset, accounting for any |
507 | /// non-affecting input files. |
508 | SourceLocation::UIntTy getAdjustedOffset(SourceLocation::UIntTy Offset) const; |
509 | /// Returns an adjustment for offset into SourceManager, accounting for any |
510 | /// non-affecting input files. |
511 | SourceLocation::UIntTy getAdjustment(SourceLocation::UIntTy Offset) const; |
512 | |
513 | /// Retrieve or create a submodule ID for this module. |
514 | unsigned getSubmoduleID(Module *Mod); |
515 | |
516 | /// Write the given subexpression to the bitstream. |
517 | void WriteSubStmt(Stmt *S); |
518 | |
519 | void WriteBlockInfoBlock(); |
520 | void WriteControlBlock(Preprocessor &PP, ASTContext &Context, |
521 | StringRef isysroot); |
522 | |
523 | /// Write out the signature and diagnostic options, and return the signature. |
524 | void writeUnhashedControlBlock(Preprocessor &PP, ASTContext &Context); |
525 | ASTFileSignature backpatchSignature(); |
526 | |
527 | /// Calculate hash of the pcm content. |
528 | std::pair<ASTFileSignature, ASTFileSignature> createSignature() const; |
529 | |
530 | void (SourceManager &SourceMgr, HeaderSearchOptions &HSOpts); |
531 | void WriteSourceManagerBlock(SourceManager &SourceMgr, |
532 | const Preprocessor &PP); |
533 | void WritePreprocessor(const Preprocessor &PP, bool IsModule); |
534 | void (const HeaderSearch &HS); |
535 | void WritePreprocessorDetail(PreprocessingRecord &PPRec, |
536 | uint64_t MacroOffsetsBase); |
537 | void WriteSubmodules(Module *WritingModule); |
538 | |
539 | void WritePragmaDiagnosticMappings(const DiagnosticsEngine &Diag, |
540 | bool isModule); |
541 | |
542 | unsigned TypeExtQualAbbrev = 0; |
543 | void WriteTypeAbbrevs(); |
544 | void WriteType(QualType T); |
545 | |
546 | bool isLookupResultExternal(StoredDeclsList &Result, DeclContext *DC); |
547 | bool isLookupResultEntirelyExternalOrUnreachable(StoredDeclsList &Result, |
548 | DeclContext *DC); |
549 | |
550 | void GenerateNameLookupTable(const DeclContext *DC, |
551 | llvm::SmallVectorImpl<char> &LookupTable); |
552 | uint64_t WriteDeclContextLexicalBlock(ASTContext &Context, DeclContext *DC); |
553 | uint64_t WriteDeclContextVisibleBlock(ASTContext &Context, DeclContext *DC); |
554 | void WriteTypeDeclOffsets(); |
555 | void WriteFileDeclIDsMap(); |
556 | void (); |
557 | void WriteSelectors(Sema &SemaRef); |
558 | void WriteReferencedSelectorsPool(Sema &SemaRef); |
559 | void WriteIdentifierTable(Preprocessor &PP, IdentifierResolver &IdResolver, |
560 | bool IsModule); |
561 | void WriteDeclAndTypes(ASTContext &Context); |
562 | void PrepareWritingSpecialDecls(Sema &SemaRef); |
563 | void WriteSpecialDeclRecords(Sema &SemaRef); |
564 | void WriteDeclUpdatesBlocks(RecordDataImpl &OffsetsRecord); |
565 | void WriteDeclContextVisibleUpdate(const DeclContext *DC); |
566 | void WriteFPPragmaOptions(const FPOptionsOverride &Opts); |
567 | void WriteOpenCLExtensions(Sema &SemaRef); |
568 | void WriteCUDAPragmas(Sema &SemaRef); |
569 | void WriteObjCCategories(); |
570 | void WriteLateParsedTemplates(Sema &SemaRef); |
571 | void WriteOptimizePragmaOptions(Sema &SemaRef); |
572 | void WriteMSStructPragmaOptions(Sema &SemaRef); |
573 | void WriteMSPointersToMembersPragmaOptions(Sema &SemaRef); |
574 | void WritePackPragmaOptions(Sema &SemaRef); |
575 | void WriteFloatControlPragmaOptions(Sema &SemaRef); |
576 | void WriteModuleFileExtension(Sema &SemaRef, |
577 | ModuleFileExtensionWriter &Writer); |
578 | |
579 | unsigned DeclParmVarAbbrev = 0; |
580 | unsigned DeclContextLexicalAbbrev = 0; |
581 | unsigned DeclContextVisibleLookupAbbrev = 0; |
582 | unsigned UpdateVisibleAbbrev = 0; |
583 | unsigned DeclRecordAbbrev = 0; |
584 | unsigned DeclTypedefAbbrev = 0; |
585 | unsigned DeclVarAbbrev = 0; |
586 | unsigned DeclFieldAbbrev = 0; |
587 | unsigned DeclEnumAbbrev = 0; |
588 | unsigned DeclObjCIvarAbbrev = 0; |
589 | unsigned DeclCXXMethodAbbrev = 0; |
590 | unsigned DeclDependentNonTemplateCXXMethodAbbrev = 0; |
591 | unsigned DeclTemplateCXXMethodAbbrev = 0; |
592 | unsigned DeclMemberSpecializedCXXMethodAbbrev = 0; |
593 | unsigned DeclTemplateSpecializedCXXMethodAbbrev = 0; |
594 | unsigned DeclDependentSpecializationCXXMethodAbbrev = 0; |
595 | unsigned DeclTemplateTypeParmAbbrev = 0; |
596 | unsigned DeclUsingShadowAbbrev = 0; |
597 | |
598 | unsigned DeclRefExprAbbrev = 0; |
599 | unsigned CharacterLiteralAbbrev = 0; |
600 | unsigned IntegerLiteralAbbrev = 0; |
601 | unsigned ExprImplicitCastAbbrev = 0; |
602 | unsigned BinaryOperatorAbbrev = 0; |
603 | unsigned CompoundAssignOperatorAbbrev = 0; |
604 | unsigned CallExprAbbrev = 0; |
605 | unsigned CXXOperatorCallExprAbbrev = 0; |
606 | unsigned CXXMemberCallExprAbbrev = 0; |
607 | |
608 | unsigned CompoundStmtAbbrev = 0; |
609 | |
610 | void WriteDeclAbbrevs(); |
611 | void WriteDecl(ASTContext &Context, Decl *D); |
612 | |
613 | ASTFileSignature WriteASTCore(Sema &SemaRef, StringRef isysroot, |
614 | Module *WritingModule); |
615 | |
616 | public: |
617 | /// Create a new precompiled header writer that outputs to |
618 | /// the given bitstream. |
619 | ASTWriter(llvm::BitstreamWriter &Stream, SmallVectorImpl<char> &Buffer, |
620 | InMemoryModuleCache &ModuleCache, |
621 | ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions, |
622 | bool IncludeTimestamps = true, bool BuildingImplicitModule = false, |
623 | bool GeneratingReducedBMI = false); |
624 | ~ASTWriter() override; |
625 | |
626 | ASTContext &getASTContext() const { |
627 | assert(Context && "requested AST context when not writing AST" ); |
628 | return *Context; |
629 | } |
630 | |
631 | const LangOptions &getLangOpts() const; |
632 | |
633 | /// Get a timestamp for output into the AST file. The actual timestamp |
634 | /// of the specified file may be ignored if we have been instructed to not |
635 | /// include timestamps in the output file. |
636 | time_t getTimestampForOutput(const FileEntry *E) const; |
637 | |
638 | /// Write a precompiled header for the given semantic analysis. |
639 | /// |
640 | /// \param SemaRef a reference to the semantic analysis object that processed |
641 | /// the AST to be written into the precompiled header. |
642 | /// |
643 | /// \param WritingModule The module that we are writing. If null, we are |
644 | /// writing a precompiled header. |
645 | /// |
646 | /// \param isysroot if non-empty, write a relocatable file whose headers |
647 | /// are relative to the given system root. If we're writing a module, its |
648 | /// build directory will be used in preference to this if both are available. |
649 | /// |
650 | /// \return the module signature, which eventually will be a hash of |
651 | /// the module but currently is merely a random 32-bit number. |
652 | ASTFileSignature WriteAST(Sema &SemaRef, StringRef OutputFile, |
653 | Module *WritingModule, StringRef isysroot, |
654 | bool ShouldCacheASTInMemory = false); |
655 | |
656 | /// Emit a token. |
657 | void AddToken(const Token &Tok, RecordDataImpl &Record); |
658 | |
659 | /// Emit a AlignPackInfo. |
660 | void AddAlignPackInfo(const Sema::AlignPackInfo &Info, |
661 | RecordDataImpl &Record); |
662 | |
663 | /// Emit a FileID. |
664 | void AddFileID(FileID FID, RecordDataImpl &Record); |
665 | |
666 | /// Emit a source location. |
667 | void AddSourceLocation(SourceLocation Loc, RecordDataImpl &Record, |
668 | LocSeq *Seq = nullptr); |
669 | |
670 | /// Emit a source range. |
671 | void AddSourceRange(SourceRange Range, RecordDataImpl &Record, |
672 | LocSeq *Seq = nullptr); |
673 | |
674 | /// Emit a reference to an identifier. |
675 | void AddIdentifierRef(const IdentifierInfo *II, RecordDataImpl &Record); |
676 | |
677 | /// Get the unique number used to refer to the given selector. |
678 | serialization::SelectorID getSelectorRef(Selector Sel); |
679 | |
680 | /// Get the unique number used to refer to the given identifier. |
681 | serialization::IdentID getIdentifierRef(const IdentifierInfo *II); |
682 | |
683 | /// Get the unique number used to refer to the given macro. |
684 | serialization::MacroID getMacroRef(MacroInfo *MI, const IdentifierInfo *Name); |
685 | |
686 | /// Determine the ID of an already-emitted macro. |
687 | serialization::MacroID getMacroID(MacroInfo *MI); |
688 | |
689 | uint32_t getMacroDirectivesOffset(const IdentifierInfo *Name); |
690 | |
691 | /// Emit a reference to a type. |
692 | void AddTypeRef(QualType T, RecordDataImpl &Record); |
693 | |
694 | /// Force a type to be emitted and get its ID. |
695 | serialization::TypeID GetOrCreateTypeID(QualType T); |
696 | |
697 | /// Determine the type ID of an already-emitted type. |
698 | serialization::TypeID getTypeID(QualType T) const; |
699 | |
700 | /// Find the first local declaration of a given local redeclarable |
701 | /// decl. |
702 | const Decl *getFirstLocalDecl(const Decl *D); |
703 | |
704 | /// Is this a local declaration (that is, one that will be written to |
705 | /// our AST file)? This is the case for declarations that are neither imported |
706 | /// from another AST file nor predefined. |
707 | bool IsLocalDecl(const Decl *D) { |
708 | if (D->isFromASTFile()) |
709 | return false; |
710 | auto I = DeclIDs.find(Val: D); |
711 | return (I == DeclIDs.end() || |
712 | I->second >= serialization::NUM_PREDEF_DECL_IDS); |
713 | }; |
714 | |
715 | /// Emit a reference to a declaration. |
716 | void AddDeclRef(const Decl *D, RecordDataImpl &Record); |
717 | // Emit a reference to a declaration if the declaration was emitted. |
718 | void AddEmittedDeclRef(const Decl *D, RecordDataImpl &Record); |
719 | |
720 | /// Force a declaration to be emitted and get its ID. |
721 | serialization::DeclID GetDeclRef(const Decl *D); |
722 | |
723 | /// Determine the declaration ID of an already-emitted |
724 | /// declaration. |
725 | serialization::DeclID getDeclID(const Decl *D); |
726 | |
727 | /// Whether or not the declaration got emitted. If not, it wouldn't be |
728 | /// emitted. |
729 | /// |
730 | /// This may only be called after we've done the job to write the |
731 | /// declarations (marked by DoneWritingDeclsAndTypes). |
732 | /// |
733 | /// A declaration may only be omitted in reduced BMI. |
734 | bool wasDeclEmitted(const Decl *D) const; |
735 | |
736 | unsigned getAnonymousDeclarationNumber(const NamedDecl *D); |
737 | |
738 | /// Add a string to the given record. |
739 | void AddString(StringRef Str, RecordDataImpl &Record); |
740 | |
741 | /// Convert a path from this build process into one that is appropriate |
742 | /// for emission in the module file. |
743 | bool PreparePathForOutput(SmallVectorImpl<char> &Path); |
744 | |
745 | /// Add a path to the given record. |
746 | void AddPath(StringRef Path, RecordDataImpl &Record); |
747 | |
748 | /// Emit the current record with the given path as a blob. |
749 | void EmitRecordWithPath(unsigned Abbrev, RecordDataRef Record, |
750 | StringRef Path); |
751 | |
752 | /// Add a version tuple to the given record |
753 | void AddVersionTuple(const VersionTuple &Version, RecordDataImpl &Record); |
754 | |
755 | /// Retrieve or create a submodule ID for this module, or return 0 if |
756 | /// the submodule is neither local (a submodle of the currently-written module) |
757 | /// nor from an imported module. |
758 | unsigned getLocalOrImportedSubmoduleID(const Module *Mod); |
759 | |
760 | /// Note that the identifier II occurs at the given offset |
761 | /// within the identifier table. |
762 | void SetIdentifierOffset(const IdentifierInfo *II, uint32_t Offset); |
763 | |
764 | /// Note that the selector Sel occurs at the given offset |
765 | /// within the method pool/selector table. |
766 | void SetSelectorOffset(Selector Sel, uint32_t Offset); |
767 | |
768 | /// Record an ID for the given switch-case statement. |
769 | unsigned RecordSwitchCaseID(SwitchCase *S); |
770 | |
771 | /// Retrieve the ID for the given switch-case statement. |
772 | unsigned getSwitchCaseID(SwitchCase *S); |
773 | |
774 | void ClearSwitchCaseIDs(); |
775 | |
776 | unsigned getTypeExtQualAbbrev() const { |
777 | return TypeExtQualAbbrev; |
778 | } |
779 | |
780 | unsigned getDeclParmVarAbbrev() const { return DeclParmVarAbbrev; } |
781 | unsigned getDeclRecordAbbrev() const { return DeclRecordAbbrev; } |
782 | unsigned getDeclTypedefAbbrev() const { return DeclTypedefAbbrev; } |
783 | unsigned getDeclVarAbbrev() const { return DeclVarAbbrev; } |
784 | unsigned getDeclFieldAbbrev() const { return DeclFieldAbbrev; } |
785 | unsigned getDeclEnumAbbrev() const { return DeclEnumAbbrev; } |
786 | unsigned getDeclObjCIvarAbbrev() const { return DeclObjCIvarAbbrev; } |
787 | unsigned getDeclCXXMethodAbbrev(FunctionDecl::TemplatedKind Kind) const { |
788 | switch (Kind) { |
789 | case FunctionDecl::TK_NonTemplate: |
790 | return DeclCXXMethodAbbrev; |
791 | case FunctionDecl::TK_FunctionTemplate: |
792 | return DeclTemplateCXXMethodAbbrev; |
793 | case FunctionDecl::TK_MemberSpecialization: |
794 | return DeclMemberSpecializedCXXMethodAbbrev; |
795 | case FunctionDecl::TK_FunctionTemplateSpecialization: |
796 | return DeclTemplateSpecializedCXXMethodAbbrev; |
797 | case FunctionDecl::TK_DependentNonTemplate: |
798 | return DeclDependentNonTemplateCXXMethodAbbrev; |
799 | case FunctionDecl::TK_DependentFunctionTemplateSpecialization: |
800 | return DeclDependentSpecializationCXXMethodAbbrev; |
801 | } |
802 | llvm_unreachable("Unknwon Template Kind!" ); |
803 | } |
804 | unsigned getDeclTemplateTypeParmAbbrev() const { |
805 | return DeclTemplateTypeParmAbbrev; |
806 | } |
807 | unsigned getDeclUsingShadowAbbrev() const { return DeclUsingShadowAbbrev; } |
808 | |
809 | unsigned getDeclRefExprAbbrev() const { return DeclRefExprAbbrev; } |
810 | unsigned getCharacterLiteralAbbrev() const { return CharacterLiteralAbbrev; } |
811 | unsigned getIntegerLiteralAbbrev() const { return IntegerLiteralAbbrev; } |
812 | unsigned getExprImplicitCastAbbrev() const { return ExprImplicitCastAbbrev; } |
813 | unsigned getBinaryOperatorAbbrev() const { return BinaryOperatorAbbrev; } |
814 | unsigned getCompoundAssignOperatorAbbrev() const { |
815 | return CompoundAssignOperatorAbbrev; |
816 | } |
817 | unsigned getCallExprAbbrev() const { return CallExprAbbrev; } |
818 | unsigned getCXXOperatorCallExprAbbrev() { return CXXOperatorCallExprAbbrev; } |
819 | unsigned getCXXMemberCallExprAbbrev() { return CXXMemberCallExprAbbrev; } |
820 | |
821 | unsigned getCompoundStmtAbbrev() const { return CompoundStmtAbbrev; } |
822 | |
823 | bool hasChain() const { return Chain; } |
824 | ASTReader *getChain() const { return Chain; } |
825 | |
826 | bool isWritingStdCXXNamedModules() const { |
827 | return WritingModule && WritingModule->isNamedModule(); |
828 | } |
829 | |
830 | bool isGeneratingReducedBMI() const { return GeneratingReducedBMI; } |
831 | |
832 | bool getDoneWritingDeclsAndTypes() const { return DoneWritingDeclsAndTypes; } |
833 | |
834 | private: |
835 | // ASTDeserializationListener implementation |
836 | void ReaderInitialized(ASTReader *Reader) override; |
837 | void IdentifierRead(serialization::IdentID ID, IdentifierInfo *II) override; |
838 | void MacroRead(serialization::MacroID ID, MacroInfo *MI) override; |
839 | void TypeRead(serialization::TypeIdx Idx, QualType T) override; |
840 | void SelectorRead(serialization::SelectorID ID, Selector Sel) override; |
841 | void MacroDefinitionRead(serialization::PreprocessedEntityID ID, |
842 | MacroDefinitionRecord *MD) override; |
843 | void ModuleRead(serialization::SubmoduleID ID, Module *Mod) override; |
844 | |
845 | // ASTMutationListener implementation. |
846 | void CompletedTagDefinition(const TagDecl *D) override; |
847 | void AddedVisibleDecl(const DeclContext *DC, const Decl *D) override; |
848 | void AddedCXXImplicitMember(const CXXRecordDecl *RD, const Decl *D) override; |
849 | void AddedCXXTemplateSpecialization( |
850 | const ClassTemplateDecl *TD, |
851 | const ClassTemplateSpecializationDecl *D) override; |
852 | void AddedCXXTemplateSpecialization( |
853 | const VarTemplateDecl *TD, |
854 | const VarTemplateSpecializationDecl *D) override; |
855 | void AddedCXXTemplateSpecialization(const FunctionTemplateDecl *TD, |
856 | const FunctionDecl *D) override; |
857 | void ResolvedExceptionSpec(const FunctionDecl *FD) override; |
858 | void DeducedReturnType(const FunctionDecl *FD, QualType ReturnType) override; |
859 | void ResolvedOperatorDelete(const CXXDestructorDecl *DD, |
860 | const FunctionDecl *Delete, |
861 | Expr *ThisArg) override; |
862 | void CompletedImplicitDefinition(const FunctionDecl *D) override; |
863 | void InstantiationRequested(const ValueDecl *D) override; |
864 | void VariableDefinitionInstantiated(const VarDecl *D) override; |
865 | void FunctionDefinitionInstantiated(const FunctionDecl *D) override; |
866 | void DefaultArgumentInstantiated(const ParmVarDecl *D) override; |
867 | void DefaultMemberInitializerInstantiated(const FieldDecl *D) override; |
868 | void AddedObjCCategoryToInterface(const ObjCCategoryDecl *CatD, |
869 | const ObjCInterfaceDecl *IFD) override; |
870 | void DeclarationMarkedUsed(const Decl *D) override; |
871 | void DeclarationMarkedOpenMPThreadPrivate(const Decl *D) override; |
872 | void DeclarationMarkedOpenMPDeclareTarget(const Decl *D, |
873 | const Attr *Attr) override; |
874 | void DeclarationMarkedOpenMPAllocate(const Decl *D, const Attr *A) override; |
875 | void RedefinedHiddenDefinition(const NamedDecl *D, Module *M) override; |
876 | void AddedAttributeToRecord(const Attr *Attr, |
877 | const RecordDecl *Record) override; |
878 | void EnteringModulePurview() override; |
879 | void AddedManglingNumber(const Decl *D, unsigned) override; |
880 | void AddedStaticLocalNumbers(const Decl *D, unsigned) override; |
881 | void AddedAnonymousNamespace(const TranslationUnitDecl *, |
882 | NamespaceDecl *AnonNamespace) override; |
883 | }; |
884 | |
885 | /// AST and semantic-analysis consumer that generates a |
886 | /// precompiled header from the parsed source code. |
887 | class PCHGenerator : public SemaConsumer { |
888 | Preprocessor &PP; |
889 | std::string OutputFile; |
890 | std::string isysroot; |
891 | Sema *SemaPtr; |
892 | std::shared_ptr<PCHBuffer> Buffer; |
893 | llvm::BitstreamWriter Stream; |
894 | ASTWriter Writer; |
895 | bool AllowASTWithErrors; |
896 | bool ShouldCacheASTInMemory; |
897 | |
898 | protected: |
899 | ASTWriter &getWriter() { return Writer; } |
900 | const ASTWriter &getWriter() const { return Writer; } |
901 | SmallVectorImpl<char> &getPCH() const { return Buffer->Data; } |
902 | |
903 | bool isComplete() const { return Buffer->IsComplete; } |
904 | PCHBuffer *getBufferPtr() { return Buffer.get(); } |
905 | StringRef getOutputFile() const { return OutputFile; } |
906 | DiagnosticsEngine &getDiagnostics() const { |
907 | return SemaPtr->getDiagnostics(); |
908 | } |
909 | Preprocessor &getPreprocessor() { return PP; } |
910 | |
911 | virtual Module *getEmittingModule(ASTContext &Ctx); |
912 | |
913 | public: |
914 | PCHGenerator(Preprocessor &PP, InMemoryModuleCache &ModuleCache, |
915 | StringRef OutputFile, StringRef isysroot, |
916 | std::shared_ptr<PCHBuffer> Buffer, |
917 | ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions, |
918 | bool AllowASTWithErrors = false, bool IncludeTimestamps = true, |
919 | bool BuildingImplicitModule = false, |
920 | bool ShouldCacheASTInMemory = false, |
921 | bool GeneratingReducedBMI = false); |
922 | ~PCHGenerator() override; |
923 | |
924 | void InitializeSema(Sema &S) override { SemaPtr = &S; } |
925 | void HandleTranslationUnit(ASTContext &Ctx) override; |
926 | ASTMutationListener *GetASTMutationListener() override; |
927 | ASTDeserializationListener *GetASTDeserializationListener() override; |
928 | bool hasEmittedPCH() const { return Buffer->IsComplete; } |
929 | }; |
930 | |
931 | class ReducedBMIGenerator : public PCHGenerator { |
932 | protected: |
933 | virtual Module *getEmittingModule(ASTContext &Ctx) override; |
934 | |
935 | public: |
936 | ReducedBMIGenerator(Preprocessor &PP, InMemoryModuleCache &ModuleCache, |
937 | StringRef OutputFile); |
938 | |
939 | void HandleTranslationUnit(ASTContext &Ctx) override; |
940 | }; |
941 | |
942 | /// If we can elide the definition of \param D in reduced BMI. |
943 | /// |
944 | /// Generally, we can elide the definition of a declaration if it won't affect |
945 | /// the ABI. e.g., the non-inline function bodies. |
946 | bool CanElideDeclDef(const Decl *D); |
947 | |
948 | /// A simple helper class to pack several bits in order into (a) 32 bit |
949 | /// integer(s). |
950 | class BitsPacker { |
951 | constexpr static uint32_t BitIndexUpbound = 32u; |
952 | |
953 | public: |
954 | BitsPacker() = default; |
955 | BitsPacker(const BitsPacker &) = delete; |
956 | BitsPacker(BitsPacker &&) = delete; |
957 | BitsPacker operator=(const BitsPacker &) = delete; |
958 | BitsPacker operator=(BitsPacker &&) = delete; |
959 | ~BitsPacker() = default; |
960 | |
961 | bool canWriteNextNBits(uint32_t BitsWidth) const { |
962 | return CurrentBitIndex + BitsWidth < BitIndexUpbound; |
963 | } |
964 | |
965 | void reset(uint32_t Value) { |
966 | UnderlyingValue = Value; |
967 | CurrentBitIndex = 0; |
968 | } |
969 | |
970 | void addBit(bool Value) { addBits(Value, BitsWidth: 1); } |
971 | void addBits(uint32_t Value, uint32_t BitsWidth) { |
972 | assert(BitsWidth < BitIndexUpbound); |
973 | assert((Value < (1u << BitsWidth)) && "Passing narrower bit width!" ); |
974 | assert(canWriteNextNBits(BitsWidth) && |
975 | "Inserting too much bits into a value!" ); |
976 | |
977 | UnderlyingValue |= Value << CurrentBitIndex; |
978 | CurrentBitIndex += BitsWidth; |
979 | } |
980 | |
981 | operator uint32_t() { return UnderlyingValue; } |
982 | |
983 | private: |
984 | uint32_t UnderlyingValue = 0; |
985 | uint32_t CurrentBitIndex = 0; |
986 | }; |
987 | |
988 | } // namespace clang |
989 | |
990 | #endif // LLVM_CLANG_SERIALIZATION_ASTWRITER_H |
991 | |