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 196     // Report that the module is out of date, since we tried (and failed) to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