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