1 //===- ModuleManager.cpp - Module Manager -----------------------*- 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 ModuleManager class, which manages a set of loaded
10 //  modules for the ASTReader.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #ifndef LLVM_CLANG_SERIALIZATION_MODULEMANAGER_H
15 #define LLVM_CLANG_SERIALIZATION_MODULEMANAGER_H
16 
17 #include "clang/Basic/LLVM.h"
18 #include "clang/Basic/Module.h"
19 #include "clang/Basic/SourceLocation.h"
20 #include "clang/Serialization/ModuleFile.h"
21 #include "llvm/ADT/DenseMap.h"
22 #include "llvm/ADT/IntrusiveRefCntPtr.h"
23 #include "llvm/ADT/STLExtras.h"
24 #include "llvm/ADT/SmallPtrSet.h"
25 #include "llvm/ADT/SmallVector.h"
26 #include "llvm/ADT/StringRef.h"
27 #include "llvm/ADT/iterator.h"
28 #include "llvm/ADT/iterator_range.h"
29 #include <cstdint>
30 #include <ctime>
31 #include <memory>
32 #include <string>
33 #include <utility>
34 
35 namespace clang {
36 
37 class FileEntry;
38 class FileManager;
39 class GlobalModuleIndex;
40 class HeaderSearch;
41 class InMemoryModuleCache;
42 class PCHContainerReader;
43 
44 namespace serialization {
45 
46 /// Manages the set of modules loaded by an AST reader.
47 class ModuleManager {
48   /// The chain of AST files, in the order in which we started to load
49   /// them (this order isn't really useful for anything).
50   SmallVector<std::unique_ptr<ModuleFile>, 2> Chain;
51 
52   /// The chain of non-module PCH files. The first entry is the one named
53   /// by the user, the last one is the one that doesn't depend on anything
54   /// further.
55   SmallVector<ModuleFile *, 2> PCHChain;
56 
57   // The roots of the dependency DAG of AST files. This is used
58   // to implement short-circuiting logic when running DFS over the dependencies.
59   SmallVector<ModuleFile *, 2> Roots;
60 
61   /// All loaded modules, indexed by name.
62   llvm::DenseMap<const FileEntry *, ModuleFile *> Modules;
63 
64   /// FileManager that handles translating between filenames and
65   /// FileEntry *.
66   FileManager &FileMgr;
67 
68   /// Cache of PCM files.
69   IntrusiveRefCntPtr<InMemoryModuleCache> ModuleCache;
70 
71   /// Knows how to unwrap module containers.
72   const PCHContainerReader &PCHContainerRdr;
73 
74   /// Preprocessor's HeaderSearchInfo containing the module map.
75   const HeaderSearch &HeaderSearchInfo;
76 
77   /// A lookup of in-memory (virtual file) buffers
78   llvm::DenseMap<const FileEntry *, std::unique_ptr<llvm::MemoryBuffer>>
79       InMemoryBuffers;
80 
81   /// The visitation order.
82   SmallVector<ModuleFile *, 4> VisitOrder;
83 
84   /// The list of module files that both we and the global module index
85   /// know about.
86   ///
87   /// Either the global index or the module manager may have modules that the
88   /// other does not know about, because the global index can be out-of-date
89   /// (in which case the module manager could have modules it does not) and
90   /// this particular translation unit might not have loaded all of the modules
91   /// known to the global index.
92   SmallVector<ModuleFile *, 4> ModulesInCommonWithGlobalIndex;
93 
94   /// The global module index, if one is attached.
95   ///
96   /// The global module index will actually be owned by the ASTReader; this is
97   /// just an non-owning pointer.
98   GlobalModuleIndex *GlobalIndex = nullptr;
99 
100   /// State used by the "visit" operation to avoid malloc traffic in
101   /// calls to visit().
102   struct VisitState {
103     explicit VisitState(unsigned N) : VisitNumber(N, 0) {
104       Stack.reserve(N);
105     }
106 
107     /// The stack used when marking the imports of a particular module
108     /// as not-to-be-visited.
109     SmallVector<ModuleFile *, 4> Stack;
110 
111     /// The visit number of each module file, which indicates when
112     /// this module file was last visited.
113     SmallVector<unsigned, 4> VisitNumber;
114 
115     /// The next visit number to use to mark visited module files.
116     unsigned NextVisitNumber = 1;
117 
118     /// The next visit state.
119     std::unique_ptr<VisitState> NextState;
120   };
121 
122   /// The first visit() state in the chain.
123   std::unique_ptr<VisitState> FirstVisitState;
124 
125   std::unique_ptr<VisitState> allocateVisitState();
126   void returnVisitState(std::unique_ptr<VisitState> State);
127 
128 public:
129   using ModuleIterator = llvm::pointee_iterator<
130       SmallVectorImpl<std::unique_ptr<ModuleFile>>::iterator>;
131   using ModuleConstIterator = llvm::pointee_iterator<
132       SmallVectorImpl<std::unique_ptr<ModuleFile>>::const_iterator>;
133   using ModuleReverseIterator = llvm::pointee_iterator<
134       SmallVectorImpl<std::unique_ptr<ModuleFile>>::reverse_iterator>;
135   using ModuleOffset = std::pair<uint32_t, StringRef>;
136 
137   explicit ModuleManager(FileManager &FileMgr, InMemoryModuleCache &ModuleCache,
138                          const PCHContainerReader &PCHContainerRdr,
139                          const HeaderSearch &HeaderSearchInfo);
140 
141   /// Forward iterator to traverse all loaded modules.
142   ModuleIterator begin() { return Chain.begin(); }
143 
144   /// Forward iterator end-point to traverse all loaded modules
145   ModuleIterator end() { return Chain.end(); }
146 
147   /// Const forward iterator to traverse all loaded modules.
148   ModuleConstIterator begin() const { return Chain.begin(); }
149 
150   /// Const forward iterator end-point to traverse all loaded modules
151   ModuleConstIterator end() const { return Chain.end(); }
152 
153   /// Reverse iterator to traverse all loaded modules.
154   ModuleReverseIterator rbegin() { return Chain.rbegin(); }
155 
156   /// Reverse iterator end-point to traverse all loaded modules.
157   ModuleReverseIterator rend() { return Chain.rend(); }
158 
159   /// A range covering the PCH and preamble module files loaded.
160   llvm::iterator_range<SmallVectorImpl<ModuleFile *>::const_iterator>
161   pch_modules() const {
162     return llvm::make_range(PCHChain.begin(), PCHChain.end());
163   }
164 
165   /// Returns the primary module associated with the manager, that is,
166   /// the first module loaded
167   ModuleFile &getPrimaryModule() { return *Chain[0]; }
168 
169   /// Returns the primary module associated with the manager, that is,
170   /// the first module loaded.
171   ModuleFile &getPrimaryModule() const { return *Chain[0]; }
172 
173   /// Returns the module associated with the given index
174   ModuleFile &operator[](unsigned Index) const { return *Chain[Index]; }
175 
176   /// Returns the module associated with the given file name.
177   ModuleFile *lookupByFileName(StringRef FileName) const;
178 
179   /// Returns the module associated with the given module name.
180   ModuleFile *lookupByModuleName(StringRef ModName) const;
181 
182   /// Returns the module associated with the given module file.
183   ModuleFile *lookup(const FileEntry *File) const;
184 
185   /// Returns the in-memory (virtual file) buffer with the given name
186   std::unique_ptr<llvm::MemoryBuffer> lookupBuffer(StringRef Name);
187 
188   /// Number of modules loaded
189   unsigned size() const { return Chain.size(); }
190 
191   /// The result of attempting to add a new module.
192   enum AddModuleResult {
193     /// The module file had already been loaded.
194     AlreadyLoaded,
195 
196     /// The module file was just loaded in response to this call.
197     NewlyLoaded,
198 
199     /// The module file is missing.
200     Missing,
201 
202     /// The module file is out-of-date.
203     OutOfDate
204   };
205 
206   using ASTFileSignatureReader = ASTFileSignature (*)(StringRef);
207 
208   /// Attempts to create a new module and add it to the list of known
209   /// modules.
210   ///
211   /// \param FileName The file name of the module to be loaded.
212   ///
213   /// \param Type The kind of module being loaded.
214   ///
215   /// \param ImportLoc The location at which the module is imported.
216   ///
217   /// \param ImportedBy The module that is importing this module, or NULL if
218   /// this module is imported directly by the user.
219   ///
220   /// \param Generation The generation in which this module was loaded.
221   ///
222   /// \param ExpectedSize The expected size of the module file, used for
223   /// validation. This will be zero if unknown.
224   ///
225   /// \param ExpectedModTime The expected modification time of the module
226   /// file, used for validation. This will be zero if unknown.
227   ///
228   /// \param ExpectedSignature The expected signature of the module file, used
229   /// for validation. This will be zero if unknown.
230   ///
231   /// \param ReadSignature Reads the signature from an AST file without actually
232   /// loading it.
233   ///
234   /// \param Module A pointer to the module file if the module was successfully
235   /// loaded.
236   ///
237   /// \param ErrorStr Will be set to a non-empty string if any errors occurred
238   /// while trying to load the module.
239   ///
240   /// \return A pointer to the module that corresponds to this file name,
241   /// and a value indicating whether the module was loaded.
242   AddModuleResult addModule(StringRef FileName, ModuleKind Type,
243                             SourceLocation ImportLoc,
244                             ModuleFile *ImportedBy, unsigned Generation,
245                             off_t ExpectedSize, time_t ExpectedModTime,
246                             ASTFileSignature ExpectedSignature,
247                             ASTFileSignatureReader ReadSignature,
248                             ModuleFile *&Module,
249                             std::string &ErrorStr);
250 
251   /// Remove the modules starting from First (to the end).
252   void removeModules(ModuleIterator First);
253 
254   /// Add an in-memory buffer the list of known buffers
255   void addInMemoryBuffer(StringRef FileName,
256                          std::unique_ptr<llvm::MemoryBuffer> Buffer);
257 
258   /// Set the global module index.
259   void setGlobalIndex(GlobalModuleIndex *Index);
260 
261   /// Notification from the AST reader that the given module file
262   /// has been "accepted", and will not (can not) be unloaded.
263   void moduleFileAccepted(ModuleFile *MF);
264 
265   /// Visit each of the modules.
266   ///
267   /// This routine visits each of the modules, starting with the
268   /// "root" modules that no other loaded modules depend on, and
269   /// proceeding to the leaf modules, visiting each module only once
270   /// during the traversal.
271   ///
272   /// This traversal is intended to support various "lookup"
273   /// operations that can find data in any of the loaded modules.
274   ///
275   /// \param Visitor A visitor function that will be invoked with each
276   /// module. The return value must be convertible to bool; when false, the
277   /// visitation continues to modules that the current module depends on. When
278   /// true, the visitation skips any modules that the current module depends on.
279   ///
280   /// \param ModuleFilesHit If non-NULL, contains the set of module files
281   /// that we know we need to visit because the global module index told us to.
282   /// Any module that is known to both the global module index and the module
283   /// manager that is *not* in this set can be skipped.
284   void visit(llvm::function_ref<bool(ModuleFile &M)> Visitor,
285              llvm::SmallPtrSetImpl<ModuleFile *> *ModuleFilesHit = nullptr);
286 
287   /// Attempt to resolve the given module file name to a file entry.
288   ///
289   /// \param FileName The name of the module file.
290   ///
291   /// \param ExpectedSize The size that the module file is expected to have.
292   /// If the actual size differs, the resolver should return \c true.
293   ///
294   /// \param ExpectedModTime The modification time that the module file is
295   /// expected to have. If the actual modification time differs, the resolver
296   /// should return \c true.
297   ///
298   /// \param File Will be set to the file if there is one, or null
299   /// otherwise.
300   ///
301   /// \returns True if a file exists but does not meet the size/
302   /// modification time criteria, false if the file is either available and
303   /// suitable, or is missing.
304   bool lookupModuleFile(StringRef FileName, off_t ExpectedSize,
305                         time_t ExpectedModTime, OptionalFileEntryRef &File);
306 
307   /// View the graphviz representation of the module graph.
308   void viewGraph();
309 
310   InMemoryModuleCache &getModuleCache() const { return *ModuleCache; }
311 };
312 
313 } // namespace serialization
314 
315 } // namespace clang
316 
317 #endif // LLVM_CLANG_SERIALIZATION_MODULEMANAGER_H
318