ModuleManager.cpp revision 360784 
1//===- ModuleManager.cpp - Module Manager ---------------------------------===//
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#include "clang/Serialization/ModuleManager.h"
15#include "clang/Basic/FileManager.h"
16#include "clang/Basic/LLVM.h"
17#include "clang/Lex/HeaderSearch.h"
18#include "clang/Lex/ModuleMap.h"
19#include "clang/Serialization/GlobalModuleIndex.h"
20#include "clang/Serialization/InMemoryModuleCache.h"
21#include "clang/Serialization/ModuleFile.h"
22#include "clang/Serialization/PCHContainerOperations.h"
23#include "llvm/ADT/STLExtras.h"
24#include "llvm/ADT/SetVector.h"
25#include "llvm/ADT/SmallPtrSet.h"
26#include "llvm/ADT/SmallVector.h"
27#include "llvm/ADT/StringRef.h"
28#include "llvm/ADT/iterator.h"
29#include "llvm/Support/Chrono.h"
30#include "llvm/Support/DOTGraphTraits.h"
31#include "llvm/Support/ErrorOr.h"
32#include "llvm/Support/GraphWriter.h"
33#include "llvm/Support/MemoryBuffer.h"
34#include "llvm/Support/VirtualFileSystem.h"
35#include <algorithm>
36#include <cassert>
37#include <memory>
38#include <string>
39#include <system_error>
40
41using namespace clang;
42using namespace serialization;
43
44ModuleFile *ModuleManager::lookupByFileName(StringRef Name) const {
45  auto Entry = FileMgr.getFile(Name, /*OpenFile=*/false,
46                               /*CacheFailure=*/false);
47  if (Entry)
48    return lookup(*Entry);
49
50  return nullptr;
51}
52
53ModuleFile *ModuleManager::lookupByModuleName(StringRef Name) const {
54  if (const Module *Mod = HeaderSearchInfo.getModuleMap().findModule(Name))
55    if (const FileEntry *File = Mod->getASTFile())
56      return lookup(File);
57
58  return nullptr;
59}
60
61ModuleFile *ModuleManager::lookup(const FileEntry *File) const {
62  auto Known = Modules.find(File);
63  if (Known == Modules.end())
64    return nullptr;
65
66  return Known->second;
67}
68
69std::unique_ptr<llvm::MemoryBuffer>
70ModuleManager::lookupBuffer(StringRef Name) {
71  auto Entry = FileMgr.getFile(Name, /*OpenFile=*/false,
72                               /*CacheFailure=*/false);
73  if (!Entry)
74    return nullptr;
75  return std::move(InMemoryBuffers[*Entry]);
76}
77
78static bool checkSignature(ASTFileSignature Signature,
79                           ASTFileSignature ExpectedSignature,
80                           std::string &ErrorStr) {
81  if (!ExpectedSignature || Signature == ExpectedSignature)
82    return false;
83
84  ErrorStr =
85      Signature ? "signature mismatch" : "could not read module signature";
86  return true;
87}
88
89static void updateModuleImports(ModuleFile &MF, ModuleFile *ImportedBy,
90                                SourceLocation ImportLoc) {
91  if (ImportedBy) {
92    MF.ImportedBy.insert(ImportedBy);
93    ImportedBy->Imports.insert(&MF);
94  } else {
95    if (!MF.DirectlyImported)
96      MF.ImportLoc = ImportLoc;
97
98    MF.DirectlyImported = true;
99  }
100}
101
102ModuleManager::AddModuleResult
103ModuleManager::addModule(StringRef FileName, ModuleKind Type,
104                         SourceLocation ImportLoc, ModuleFile *ImportedBy,
105                         unsigned Generation,
106                         off_t ExpectedSize, time_t ExpectedModTime,
107                         ASTFileSignature ExpectedSignature,
108                         ASTFileSignatureReader ReadSignature,
109                         ModuleFile *&Module,
110                         std::string &ErrorStr) {
111  Module = nullptr;
112
113  // Look for the file entry. This only fails if the expected size or
114  // modification time differ.
115  const FileEntry *Entry;
116  if (Type == MK_ExplicitModule || Type == MK_PrebuiltModule) {
117    // If we're not expecting to pull this file out of the module cache, it
118    // might have a different mtime due to being moved across filesystems in
119    // a distributed build. The size must still match, though. (As must the
120    // contents, but we can't check that.)
121    ExpectedModTime = 0;
122  }
123  // Note: ExpectedSize and ExpectedModTime will be 0 for MK_ImplicitModule
124  // when using an ASTFileSignature.
125  if (lookupModuleFile(FileName, ExpectedSize, ExpectedModTime, Entry)) {
126    ErrorStr = "module file out of date";
127    return OutOfDate;
128  }
129
130  if (!Entry && FileName != "-") {
131    ErrorStr = "module file not found";
132    return Missing;
133  }
134
135  // Check whether we already loaded this module, before
136  if (ModuleFile *ModuleEntry = Modules.lookup(Entry)) {
137    // Check the stored signature.
138    if (checkSignature(ModuleEntry->Signature, ExpectedSignature, ErrorStr))
139      return OutOfDate;
140
141    Module = ModuleEntry;
142    updateModuleImports(*ModuleEntry, ImportedBy, ImportLoc);
143    return AlreadyLoaded;
144  }
145
146  // Allocate a new module.
147  auto NewModule = std::make_unique<ModuleFile>(Type, Generation);
148  NewModule->Index = Chain.size();
149  NewModule->FileName = FileName.str();
150  NewModule->File = Entry;
151  NewModule->ImportLoc = ImportLoc;
152  NewModule->InputFilesValidationTimestamp = 0;
153
154  if (NewModule->Kind == MK_ImplicitModule) {
155    std::string TimestampFilename = NewModule->getTimestampFilename();
156    llvm::vfs::Status Status;
157    // A cached stat value would be fine as well.
158    if (!FileMgr.getNoncachedStatValue(TimestampFilename, Status))
159      NewModule->InputFilesValidationTimestamp =
160          llvm::sys::toTimeT(Status.getLastModificationTime());
161  }
162
163  // Load the contents of the module
164  if (std::unique_ptr<llvm::MemoryBuffer> Buffer = lookupBuffer(FileName)) {
165    // The buffer was already provided for us.
166    NewModule->Buffer = &ModuleCache->addBuiltPCM(FileName, std::move(Buffer));
167    // Since the cached buffer is reused, it is safe to close the file
168    // descriptor that was opened while stat()ing the PCM in
169    // lookupModuleFile() above, it won't be needed any longer.
170    Entry->closeFile();
171  } else if (llvm::MemoryBuffer *Buffer =
172                 getModuleCache().lookupPCM(FileName)) {
173    NewModule->Buffer = Buffer;
174    // As above, the file descriptor is no longer needed.
175    Entry->closeFile();
176  } else if (getModuleCache().shouldBuildPCM(FileName)) {
177    // Report that the module is out of date, since we tried (and failed) to
178    // import it earlier.
179    Entry->closeFile();
180    return OutOfDate;
181  } else {
182    // Open the AST file.
183    llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> Buf((std::error_code()));
184    if (FileName == "-") {
185      Buf = llvm::MemoryBuffer::getSTDIN();
186    } else {
187      // Get a buffer of the file and close the file descriptor when done.
188      Buf = FileMgr.getBufferForFile(NewModule->File, /*isVolatile=*/false);
189    }
190
191    if (!Buf) {
192      ErrorStr = Buf.getError().message();
193      return Missing;
194    }
195
196    NewModule->Buffer = &getModuleCache().addPCM(FileName, std::move(*Buf));
197  }
198
199  // Initialize the stream.
200  NewModule->Data = PCHContainerRdr.ExtractPCH(*NewModule->Buffer);
201
202  // Read the signature eagerly now so that we can check it.  Avoid calling
203  // ReadSignature unless there's something to check though.
204  if (ExpectedSignature && checkSignature(ReadSignature(NewModule->Data),
205                                          ExpectedSignature, ErrorStr))
206    return OutOfDate;
207
208  // We're keeping this module.  Store it everywhere.
209  Module = Modules[Entry] = NewModule.get();
210
211  updateModuleImports(*NewModule, ImportedBy, ImportLoc);
212
213  if (!NewModule->isModule())
214    PCHChain.push_back(NewModule.get());
215  if (!ImportedBy)
216    Roots.push_back(NewModule.get());
217
218  Chain.push_back(std::move(NewModule));
219  return NewlyLoaded;
220}
221
222void ModuleManager::removeModules(ModuleIterator First, ModuleMap *modMap) {
223  auto Last = end();
224  if (First == Last)
225    return;
226
227  // Explicitly clear VisitOrder since we might not notice it is stale.
228  VisitOrder.clear();
229
230  // Collect the set of module file pointers that we'll be removing.
231  llvm::SmallPtrSet<ModuleFile *, 4> victimSet(
232      (llvm::pointer_iterator<ModuleIterator>(First)),
233      (llvm::pointer_iterator<ModuleIterator>(Last)));
234
235  auto IsVictim = [&](ModuleFile *MF) {
236    return victimSet.count(MF);
237  };
238  // Remove any references to the now-destroyed modules.
239  for (auto I = begin(); I != First; ++I) {
240    I->Imports.remove_if(IsVictim);
241    I->ImportedBy.remove_if(IsVictim);
242  }
243  Roots.erase(std::remove_if(Roots.begin(), Roots.end(), IsVictim),
244              Roots.end());
245
246  // Remove the modules from the PCH chain.
247  for (auto I = First; I != Last; ++I) {
248    if (!I->isModule()) {
249      PCHChain.erase(llvm::find(PCHChain, &*I), PCHChain.end());
250      break;
251    }
252  }
253
254  // Delete the modules and erase them from the various structures.
255  for (ModuleIterator victim = First; victim != Last; ++victim) {
256    Modules.erase(victim->File);
257
258    if (modMap) {
259      StringRef ModuleName = victim->ModuleName;
260      if (Module *mod = modMap->findModule(ModuleName)) {
261        mod->setASTFile(nullptr);
262      }
263    }
264  }
265
266  // Delete the modules.
267  Chain.erase(Chain.begin() + (First - begin()), Chain.end());
268}
269
270void
271ModuleManager::addInMemoryBuffer(StringRef FileName,
272                                 std::unique_ptr<llvm::MemoryBuffer> Buffer) {
273  const FileEntry *Entry =
274      FileMgr.getVirtualFile(FileName, Buffer->getBufferSize(), 0);
275  InMemoryBuffers[Entry] = std::move(Buffer);
276}
277
278ModuleManager::VisitState *ModuleManager::allocateVisitState() {
279  // Fast path: if we have a cached state, use it.
280  if (FirstVisitState) {
281    VisitState *Result = FirstVisitState;
282    FirstVisitState = FirstVisitState->NextState;
283    Result->NextState = nullptr;
284    return Result;
285  }
286
287  // Allocate and return a new state.
288  return new VisitState(size());
289}
290
291void ModuleManager::returnVisitState(VisitState *State) {
292  assert(State->NextState == nullptr && "Visited state is in list?");
293  State->NextState = FirstVisitState;
294  FirstVisitState = State;
295}
296
297void ModuleManager::setGlobalIndex(GlobalModuleIndex *Index) {
298  GlobalIndex = Index;
299  if (!GlobalIndex) {
300    ModulesInCommonWithGlobalIndex.clear();
301    return;
302  }
303
304  // Notify the global module index about all of the modules we've already
305  // loaded.
306  for (ModuleFile &M : *this)
307    if (!GlobalIndex->loadedModuleFile(&M))
308      ModulesInCommonWithGlobalIndex.push_back(&M);
309}
310
311void ModuleManager::moduleFileAccepted(ModuleFile *MF) {
312  if (!GlobalIndex || GlobalIndex->loadedModuleFile(MF))
313    return;
314
315  ModulesInCommonWithGlobalIndex.push_back(MF);
316}
317
318ModuleManager::ModuleManager(FileManager &FileMgr,
319                             InMemoryModuleCache &ModuleCache,
320                             const PCHContainerReader &PCHContainerRdr,
321                             const HeaderSearch &HeaderSearchInfo)
322    : FileMgr(FileMgr), ModuleCache(&ModuleCache),
323      PCHContainerRdr(PCHContainerRdr), HeaderSearchInfo(HeaderSearchInfo) {}
324
325ModuleManager::~ModuleManager() { delete FirstVisitState; }
326
327void ModuleManager::visit(llvm::function_ref<bool(ModuleFile &M)> Visitor,
328                          llvm::SmallPtrSetImpl<ModuleFile *> *ModuleFilesHit) {
329  // If the visitation order vector is the wrong size, recompute the order.
330  if (VisitOrder.size() != Chain.size()) {
331    unsigned N = size();
332    VisitOrder.clear();
333    VisitOrder.reserve(N);
334
335    // Record the number of incoming edges for each module. When we
336    // encounter a module with no incoming edges, push it into the queue
337    // to seed the queue.
338    SmallVector<ModuleFile *, 4> Queue;
339    Queue.reserve(N);
340    llvm::SmallVector<unsigned, 4> UnusedIncomingEdges;
341    UnusedIncomingEdges.resize(size());
342    for (ModuleFile &M : llvm::reverse(*this)) {
343      unsigned Size = M.ImportedBy.size();
344      UnusedIncomingEdges[M.Index] = Size;
345      if (!Size)
346        Queue.push_back(&M);
347    }
348
349    // Traverse the graph, making sure to visit a module before visiting any
350    // of its dependencies.
351    while (!Queue.empty()) {
352      ModuleFile *CurrentModule = Queue.pop_back_val();
353      VisitOrder.push_back(CurrentModule);
354
355      // For any module that this module depends on, push it on the
356      // stack (if it hasn't already been marked as visited).
357      for (auto M = CurrentModule->Imports.rbegin(),
358                MEnd = CurrentModule->Imports.rend();
359           M != MEnd; ++M) {
360        // Remove our current module as an impediment to visiting the
361        // module we depend on. If we were the last unvisited module
362        // that depends on this particular module, push it into the
363        // queue to be visited.
364        unsigned &NumUnusedEdges = UnusedIncomingEdges[(*M)->Index];
365        if (NumUnusedEdges && (--NumUnusedEdges == 0))
366          Queue.push_back(*M);
367      }
368    }
369
370    assert(VisitOrder.size() == N && "Visitation order is wrong?");
371
372    delete FirstVisitState;
373    FirstVisitState = nullptr;
374  }
375
376  VisitState *State = allocateVisitState();
377  unsigned VisitNumber = State->NextVisitNumber++;
378
379  // If the caller has provided us with a hit-set that came from the global
380  // module index, mark every module file in common with the global module
381  // index that is *not* in that set as 'visited'.
382  if (ModuleFilesHit && !ModulesInCommonWithGlobalIndex.empty()) {
383    for (unsigned I = 0, N = ModulesInCommonWithGlobalIndex.size(); I != N; ++I)
384    {
385      ModuleFile *M = ModulesInCommonWithGlobalIndex[I];
386      if (!ModuleFilesHit->count(M))
387        State->VisitNumber[M->Index] = VisitNumber;
388    }
389  }
390
391  for (unsigned I = 0, N = VisitOrder.size(); I != N; ++I) {
392    ModuleFile *CurrentModule = VisitOrder[I];
393    // Should we skip this module file?
394    if (State->VisitNumber[CurrentModule->Index] == VisitNumber)
395      continue;
396
397    // Visit the module.
398    assert(State->VisitNumber[CurrentModule->Index] == VisitNumber - 1);
399    State->VisitNumber[CurrentModule->Index] = VisitNumber;
400    if (!Visitor(*CurrentModule))
401      continue;
402
403    // The visitor has requested that cut off visitation of any
404    // module that the current module depends on. To indicate this
405    // behavior, we mark all of the reachable modules as having been visited.
406    ModuleFile *NextModule = CurrentModule;
407    do {
408      // For any module that this module depends on, push it on the
409      // stack (if it hasn't already been marked as visited).
410      for (llvm::SetVector<ModuleFile *>::iterator
411             M = NextModule->Imports.begin(),
412             MEnd = NextModule->Imports.end();
413           M != MEnd; ++M) {
414        if (State->VisitNumber[(*M)->Index] != VisitNumber) {
415          State->Stack.push_back(*M);
416          State->VisitNumber[(*M)->Index] = VisitNumber;
417        }
418      }
419
420      if (State->Stack.empty())
421        break;
422
423      // Pop the next module off the stack.
424      NextModule = State->Stack.pop_back_val();
425    } while (true);
426  }
427
428  returnVisitState(State);
429}
430
431bool ModuleManager::lookupModuleFile(StringRef FileName,
432                                     off_t ExpectedSize,
433                                     time_t ExpectedModTime,
434                                     const FileEntry *&File) {
435  if (FileName == "-") {
436    File = nullptr;
437    return false;
438  }
439
440  // Open the file immediately to ensure there is no race between stat'ing and
441  // opening the file.
442  auto FileOrErr = FileMgr.getFile(FileName, /*OpenFile=*/true,
443                                   /*CacheFailure=*/false);
444  if (!FileOrErr) {
445    File = nullptr;
446    return false;
447  }
448  File = *FileOrErr;
449
450  if ((ExpectedSize && ExpectedSize != File->getSize()) ||
451      (ExpectedModTime && ExpectedModTime != File->getModificationTime()))
452    // Do not destroy File, as it may be referenced. If we need to rebuild it,
453    // it will be destroyed by removeModules.
454    return true;
455
456  return false;
457}
458
459#ifndef NDEBUG
460namespace llvm {
461
462  template<>
463  struct GraphTraits<ModuleManager> {
464    using NodeRef = ModuleFile *;
465    using ChildIteratorType = llvm::SetVector<ModuleFile *>::const_iterator;
466    using nodes_iterator = pointer_iterator<ModuleManager::ModuleConstIterator>;
467
468    static ChildIteratorType child_begin(NodeRef Node) {
469      return Node->Imports.begin();
470    }
471
472    static ChildIteratorType child_end(NodeRef Node) {
473      return Node->Imports.end();
474    }
475
476    static nodes_iterator nodes_begin(const ModuleManager &Manager) {
477      return nodes_iterator(Manager.begin());
478    }
479
480    static nodes_iterator nodes_end(const ModuleManager &Manager) {
481      return nodes_iterator(Manager.end());
482    }
483  };
484
485  template<>
486  struct DOTGraphTraits<ModuleManager> : public DefaultDOTGraphTraits {
487    explicit DOTGraphTraits(bool IsSimple = false)
488        : DefaultDOTGraphTraits(IsSimple) {}
489
490    static bool renderGraphFromBottomUp() { return true; }
491
492    std::string getNodeLabel(ModuleFile *M, const ModuleManager&) {
493      return M->ModuleName;
494    }
495  };
496
497} // namespace llvm
498
499void ModuleManager::viewGraph() {
500  llvm::ViewGraph(*this, "Modules");
501}
502#endif
503