clang/lib/Serialization/ModuleManager.cpp

   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
  41 using namespace clang;
  42 using namespace serialization;
  43
  44 ModuleFile *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
  53 ModuleFile *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
  61 ModuleFile *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
  69 std::unique_ptr<llvm::MemoryBuffer>
  70 ModuleManager::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
  78 static 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
  89 static 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
 102 ModuleManager::AddModuleResult
 103 ModuleManager::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   OptionalFileEntryRefDegradesToFileEntryPtr 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   // The ModuleManager's use of FileEntry nodes as the keys for its map of
 136   // loaded modules is less than ideal. Uniqueness for FileEntry nodes is
 137   // maintained by FileManager, which in turn uses inode numbers on hosts
 138   // that support that. When coupled with the module cache's proclivity for
 139   // turning over and deleting stale PCMs, this means entries for different
 140   // module files can wind up reusing the same underlying inode. When this
 141   // happens, subsequent accesses to the Modules map will disagree on the
 142   // ModuleFile associated with a given file. In general, it is not sufficient
 143   // to resolve this conundrum with a type like FileEntryRef that stores the
 144   // name of the FileEntry node on first access because of path canonicalization
 145   // issues. However, the paths constructed for implicit module builds are
 146   // fully under Clang's control. We *can*, therefore, rely on their structure
 147   // being consistent across operating systems and across subsequent accesses
 148   // to the Modules map.
 149   auto implicitModuleNamesMatch = [](ModuleKind Kind, const ModuleFile *MF,
 150                                      const FileEntry *Entry) -> bool {
 151     if (Kind != MK_ImplicitModule)
 152       return true;
 153     return Entry->getName() == MF->FileName;
 154   };
 155
 156   // Check whether we already loaded this module, before
 157   if (ModuleFile *ModuleEntry = Modules.lookup(Entry)) {
 158     if (implicitModuleNamesMatch(Type, ModuleEntry, Entry)) {
 159       // Check the stored signature.
 160       if (checkSignature(ModuleEntry->Signature, ExpectedSignature, ErrorStr))
 161         return OutOfDate;
 162
 163       Module = ModuleEntry;
 164       updateModuleImports(*ModuleEntry, ImportedBy, ImportLoc);
 165       return AlreadyLoaded;
 166     }
 167   }
 168
 169   // Allocate a new module.
 170   auto NewModule = std::make_unique<ModuleFile>(Type, Generation);
 171   NewModule->Index = Chain.size();
 172   NewModule->FileName = FileName.str();
 173   NewModule->File = Entry;
 174   NewModule->ImportLoc = ImportLoc;
 175   NewModule->InputFilesValidationTimestamp = 0;
 176
 177   if (NewModule->Kind == MK_ImplicitModule) {
 178     std::string TimestampFilename = NewModule->getTimestampFilename();
 179     llvm::vfs::Status Status;
 180     // A cached stat value would be fine as well.
 181     if (!FileMgr.getNoncachedStatValue(TimestampFilename, Status))
 182       NewModule->InputFilesValidationTimestamp =
 183           llvm::sys::toTimeT(Status.getLastModificationTime());
 184   }
 185
 186   // Load the contents of the module
 187   if (std::unique_ptr<llvm::MemoryBuffer> Buffer = lookupBuffer(FileName)) {
 188     // The buffer was already provided for us.
 189     NewModule->Buffer = &ModuleCache->addBuiltPCM(FileName, std::move(Buffer));
 190     // Since the cached buffer is reused, it is safe to close the file
 191     // descriptor that was opened while stat()ing the PCM in
 192     // lookupModuleFile() above, it won't be needed any longer.
 193     Entry->closeFile();
 194   } else if (llvm::MemoryBuffer *Buffer =
 195                  getModuleCache().lookupPCM(FileName)) {
 196     NewModule->Buffer = Buffer;
 197     // As above, the file descriptor is no longer needed.
 198     Entry->closeFile();
 199   } else if (getModuleCache().shouldBuildPCM(FileName)) {
 200     // Report that the module is out of date, since we tried (and failed) to
 201     // import it earlier.
 202     Entry->closeFile();
 203     return OutOfDate;
 204   } else {
 205     // Open the AST file.
 206     llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> Buf((std::error_code()));
 207     if (FileName == "-") {
 208       Buf = llvm::MemoryBuffer::getSTDIN();
 209     } else {
 210       // Get a buffer of the file and close the file descriptor when done.
 211       // The file is volatile because in a parallel build we expect multiple
 212       // compiler processes to use the same module file rebuilding it if needed.
 213       //
 214       // RequiresNullTerminator is false because module files don't need it, and
 215       // this allows the file to still be mmapped.
 216       Buf = FileMgr.getBufferForFile(NewModule->File,
 217                                      /*IsVolatile=*/true,
 218                                      /*RequiresNullTerminator=*/false);
 219     }
 220
 221     if (!Buf) {
 222       ErrorStr = Buf.getError().message();
 223       return Missing;
 224     }
 225
 226     NewModule->Buffer = &getModuleCache().addPCM(FileName, std::move(*Buf));
 227   }
 228
 229   // Initialize the stream.
 230   NewModule->Data = PCHContainerRdr.ExtractPCH(*NewModule->Buffer);
 231
 232   // Read the signature eagerly now so that we can check it.  Avoid calling
 233   // ReadSignature unless there's something to check though.
 234   if (ExpectedSignature && checkSignature(ReadSignature(NewModule->Data),
 235                                           ExpectedSignature, ErrorStr))
 236     return OutOfDate;
 237
 238   // We're keeping this module.  Store it everywhere.
 239   Module = Modules[Entry] = NewModule.get();
 240
 241   updateModuleImports(*NewModule, ImportedBy, ImportLoc);
 242
 243   if (!NewModule->isModule())
 244     PCHChain.push_back(NewModule.get());
 245   if (!ImportedBy)
 246     Roots.push_back(NewModule.get());
 247
 248   Chain.push_back(std::move(NewModule));
 249   return NewlyLoaded;
 250 }
 251
 252 void ModuleManager::removeModules(ModuleIterator First, ModuleMap *modMap) {
 253   auto Last = end();
 254   if (First == Last)
 255     return;
 256
 257   // Explicitly clear VisitOrder since we might not notice it is stale.
 258   VisitOrder.clear();
 259
 260   // Collect the set of module file pointers that we'll be removing.
 261   llvm::SmallPtrSet<ModuleFile *, 4> victimSet(
 262       (llvm::pointer_iterator<ModuleIterator>(First)),
 263       (llvm::pointer_iterator<ModuleIterator>(Last)));
 264
 265   auto IsVictim = [&](ModuleFile *MF) {
 266     return victimSet.count(MF);
 267   };
 268   // Remove any references to the now-destroyed modules.
 269   for (auto I = begin(); I != First; ++I) {
 270     I->Imports.remove_if(IsVictim);
 271     I->ImportedBy.remove_if(IsVictim);
 272   }
 273   llvm::erase_if(Roots, IsVictim);
 274
 275   // Remove the modules from the PCH chain.
 276   for (auto I = First; I != Last; ++I) {
 277     if (!I->isModule()) {
 278       PCHChain.erase(llvm::find(PCHChain, &*I), PCHChain.end());
 279       break;
 280     }
 281   }
 282
 283   // Delete the modules and erase them from the various structures.
 284   for (ModuleIterator victim = First; victim != Last; ++victim) {
 285     Modules.erase(victim->File);
 286
 287     if (modMap) {
 288       StringRef ModuleName = victim->ModuleName;
 289       if (Module *mod = modMap->findModule(ModuleName)) {
 290         mod->setASTFile(None);
 291       }
 292     }
 293   }
 294
 295   // Delete the modules.
 296   Chain.erase(Chain.begin() + (First - begin()), Chain.end());
 297 }
 298
 299 void
 300 ModuleManager::addInMemoryBuffer(StringRef FileName,
 301                                  std::unique_ptr<llvm::MemoryBuffer> Buffer) {
 302   const FileEntry *Entry =
 303       FileMgr.getVirtualFile(FileName, Buffer->getBufferSize(), 0);
 304   InMemoryBuffers[Entry] = std::move(Buffer);
 305 }
 306
 307 std::unique_ptr<ModuleManager::VisitState> ModuleManager::allocateVisitState() {
 308   // Fast path: if we have a cached state, use it.
 309   if (FirstVisitState) {
 310     auto Result = std::move(FirstVisitState);
 311     FirstVisitState = std::move(Result->NextState);
 312     return Result;
 313   }
 314
 315   // Allocate and return a new state.
 316   return std::make_unique<VisitState>(size());
 317 }
 318
 319 void ModuleManager::returnVisitState(std::unique_ptr<VisitState> State) {
 320   assert(State->NextState == nullptr && "Visited state is in list?");
 321   State->NextState = std::move(FirstVisitState);
 322   FirstVisitState = std::move(State);
 323 }
 324
 325 void ModuleManager::setGlobalIndex(GlobalModuleIndex *Index) {
 326   GlobalIndex = Index;
 327   if (!GlobalIndex) {
 328     ModulesInCommonWithGlobalIndex.clear();
 329     return;
 330   }
 331
 332   // Notify the global module index about all of the modules we've already
 333   // loaded.
 334   for (ModuleFile &M : *this)
 335     if (!GlobalIndex->loadedModuleFile(&M))
 336       ModulesInCommonWithGlobalIndex.push_back(&M);
 337 }
 338
 339 void ModuleManager::moduleFileAccepted(ModuleFile *MF) {
 340   if (!GlobalIndex || GlobalIndex->loadedModuleFile(MF))
 341     return;
 342
 343   ModulesInCommonWithGlobalIndex.push_back(MF);
 344 }
 345
 346 ModuleManager::ModuleManager(FileManager &FileMgr,
 347                              InMemoryModuleCache &ModuleCache,
 348                              const PCHContainerReader &PCHContainerRdr,
 349                              const HeaderSearch &HeaderSearchInfo)
 350     : FileMgr(FileMgr), ModuleCache(&ModuleCache),
 351       PCHContainerRdr(PCHContainerRdr), HeaderSearchInfo(HeaderSearchInfo) {}
 352
 353 void ModuleManager::visit(llvm::function_ref<bool(ModuleFile &M)> Visitor,
 354                           llvm::SmallPtrSetImpl<ModuleFile *> *ModuleFilesHit) {
 355   // If the visitation order vector is the wrong size, recompute the order.
 356   if (VisitOrder.size() != Chain.size()) {
 357     unsigned N = size();
 358     VisitOrder.clear();
 359     VisitOrder.reserve(N);
 360
 361     // Record the number of incoming edges for each module. When we
 362     // encounter a module with no incoming edges, push it into the queue
 363     // to seed the queue.
 364     SmallVector<ModuleFile *, 4> Queue;
 365     Queue.reserve(N);
 366     llvm::SmallVector<unsigned, 4> UnusedIncomingEdges;
 367     UnusedIncomingEdges.resize(size());
 368     for (ModuleFile &M : llvm::reverse(*this)) {
 369       unsigned Size = M.ImportedBy.size();
 370       UnusedIncomingEdges[M.Index] = Size;
 371       if (!Size)
 372         Queue.push_back(&M);
 373     }
 374
 375     // Traverse the graph, making sure to visit a module before visiting any
 376     // of its dependencies.
 377     while (!Queue.empty()) {
 378       ModuleFile *CurrentModule = Queue.pop_back_val();
 379       VisitOrder.push_back(CurrentModule);
 380
 381       // For any module that this module depends on, push it on the
 382       // stack (if it hasn't already been marked as visited).
 383       for (ModuleFile *M : llvm::reverse(CurrentModule->Imports)) {
 384         // Remove our current module as an impediment to visiting the
 385         // module we depend on. If we were the last unvisited module
 386         // that depends on this particular module, push it into the
 387         // queue to be visited.
 388         unsigned &NumUnusedEdges = UnusedIncomingEdges[M->Index];
 389         if (NumUnusedEdges && (--NumUnusedEdges == 0))
 390           Queue.push_back(M);
 391       }
 392     }
 393
 394     assert(VisitOrder.size() == N && "Visitation order is wrong?");
 395
 396     FirstVisitState = nullptr;
 397   }
 398
 399   auto State = allocateVisitState();
 400   unsigned VisitNumber = State->NextVisitNumber++;
 401
 402   // If the caller has provided us with a hit-set that came from the global
 403   // module index, mark every module file in common with the global module
 404   // index that is *not* in that set as 'visited'.
 405   if (ModuleFilesHit && !ModulesInCommonWithGlobalIndex.empty()) {
 406     for (unsigned I = 0, N = ModulesInCommonWithGlobalIndex.size(); I != N; ++I)
 407     {
 408       ModuleFile *M = ModulesInCommonWithGlobalIndex[I];
 409       if (!ModuleFilesHit->count(M))
 410         State->VisitNumber[M->Index] = VisitNumber;
 411     }
 412   }
 413
 414   for (unsigned I = 0, N = VisitOrder.size(); I != N; ++I) {
 415     ModuleFile *CurrentModule = VisitOrder[I];
 416     // Should we skip this module file?
 417     if (State->VisitNumber[CurrentModule->Index] == VisitNumber)
 418       continue;
 419
 420     // Visit the module.
 421     assert(State->VisitNumber[CurrentModule->Index] == VisitNumber - 1);
 422     State->VisitNumber[CurrentModule->Index] = VisitNumber;
 423     if (!Visitor(*CurrentModule))
 424       continue;
 425
 426     // The visitor has requested that cut off visitation of any
 427     // module that the current module depends on. To indicate this
 428     // behavior, we mark all of the reachable modules as having been visited.
 429     ModuleFile *NextModule = CurrentModule;
 430     do {
 431       // For any module that this module depends on, push it on the
 432       // stack (if it hasn't already been marked as visited).
 433       for (llvm::SetVector<ModuleFile *>::iterator
 434              M = NextModule->Imports.begin(),
 435              MEnd = NextModule->Imports.end();
 436            M != MEnd; ++M) {
 437         if (State->VisitNumber[(*M)->Index] != VisitNumber) {
 438           State->Stack.push_back(*M);
 439           State->VisitNumber[(*M)->Index] = VisitNumber;
 440         }
 441       }
 442
 443       if (State->Stack.empty())
 444         break;
 445
 446       // Pop the next module off the stack.
 447       NextModule = State->Stack.pop_back_val();
 448     } while (true);
 449   }
 450
 451   returnVisitState(std::move(State));
 452 }
 453
 454 bool ModuleManager::lookupModuleFile(StringRef FileName, off_t ExpectedSize,
 455                                      time_t ExpectedModTime,
 456                                      Optional<FileEntryRef> &File) {
 457   File = None;
 458   if (FileName == "-")
 459     return false;
 460
 461   // Open the file immediately to ensure there is no race between stat'ing and
 462   // opening the file.
 463   Optional<FileEntryRef> FileOrErr =
 464       expectedToOptional(FileMgr.getFileRef(FileName, /*OpenFile=*/true,
 465                                             /*CacheFailure=*/false));
 466   if (!FileOrErr)
 467     return false;
 468
 469   File = *FileOrErr;
 470
 471   if ((ExpectedSize && ExpectedSize != File->getSize()) ||
 472       (ExpectedModTime && ExpectedModTime != File->getModificationTime()))
 473     // Do not destroy File, as it may be referenced. If we need to rebuild it,
 474     // it will be destroyed by removeModules.
 475     return true;
 476
 477   return false;
 478 }
 479
 480 #ifndef NDEBUG
 481 namespace llvm {
 482
 483   template<>
 484   struct GraphTraits<ModuleManager> {
 485     using NodeRef = ModuleFile *;
 486     using ChildIteratorType = llvm::SetVector<ModuleFile *>::const_iterator;
 487     using nodes_iterator = pointer_iterator<ModuleManager::ModuleConstIterator>;
 488
 489     static ChildIteratorType child_begin(NodeRef Node) {
 490       return Node->Imports.begin();
 491     }
 492
 493     static ChildIteratorType child_end(NodeRef Node) {
 494       return Node->Imports.end();
 495     }
 496
 497     static nodes_iterator nodes_begin(const ModuleManager &Manager) {
 498       return nodes_iterator(Manager.begin());
 499     }
 500
 501     static nodes_iterator nodes_end(const ModuleManager &Manager) {
 502       return nodes_iterator(Manager.end());
 503     }
 504   };
 505
 506   template<>
 507   struct DOTGraphTraits<ModuleManager> : public DefaultDOTGraphTraits {
 508     explicit DOTGraphTraits(bool IsSimple = false)
 509         : DefaultDOTGraphTraits(IsSimple) {}
 510
 511     static bool renderGraphFromBottomUp() { return true; }
 512
 513     std::string getNodeLabel(ModuleFile *M, const ModuleManager&) {
 514       return M->ModuleName;
 515     }
 516   };
 517
 518 } // namespace llvm
 519
 520 void ModuleManager::viewGraph() {
 521   llvm::ViewGraph(*this, "Modules");
 522 }
 523 #endif