1//===--- FileManager.cpp - File System Probing and Caching ----------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10//  This file implements the FileManager interface.
11//
12//===----------------------------------------------------------------------===//
13//
14// TODO: This should index all interesting directories with dirent calls.
15//  getdirentries ?
16//  opendir/readdir_r/closedir ?
17//
18//===----------------------------------------------------------------------===//
19
20#include "clang/Basic/FileManager.h"
21#include "clang/Basic/FileSystemStatCache.h"
22#include "llvm/ADT/SmallString.h"
23#include "llvm/Config/llvm-config.h"
24#include "llvm/Support/FileSystem.h"
25#include "llvm/Support/MemoryBuffer.h"
26#include "llvm/Support/Path.h"
27#include "llvm/Support/raw_ostream.h"
28#include <map>
29#include <set>
30#include <string>
31#include <system_error>
32
33using namespace clang;
34
35/// NON_EXISTENT_DIR - A special value distinct from null that is used to
36/// represent a dir name that doesn't exist on the disk.
37#define NON_EXISTENT_DIR reinterpret_cast<DirectoryEntry*>((intptr_t)-1)
38
39/// NON_EXISTENT_FILE - A special value distinct from null that is used to
40/// represent a filename that doesn't exist on the disk.
41#define NON_EXISTENT_FILE reinterpret_cast<FileEntry*>((intptr_t)-1)
42
43//===----------------------------------------------------------------------===//
44// Common logic.
45//===----------------------------------------------------------------------===//
46
47FileManager::FileManager(const FileSystemOptions &FSO,
48                         IntrusiveRefCntPtr<vfs::FileSystem> FS)
49  : FS(FS), FileSystemOpts(FSO),
50    SeenDirEntries(64), SeenFileEntries(64), NextFileUID(0) {
51  NumDirLookups = NumFileLookups = 0;
52  NumDirCacheMisses = NumFileCacheMisses = 0;
53
54  // If the caller doesn't provide a virtual file system, just grab the real
55  // file system.
56  if (!FS)
57    this->FS = vfs::getRealFileSystem();
58}
59
60FileManager::~FileManager() {
61  for (unsigned i = 0, e = VirtualFileEntries.size(); i != e; ++i)
62    delete VirtualFileEntries[i];
63  for (unsigned i = 0, e = VirtualDirectoryEntries.size(); i != e; ++i)
64    delete VirtualDirectoryEntries[i];
65}
66
67void FileManager::addStatCache(FileSystemStatCache *statCache,
68                               bool AtBeginning) {
69  assert(statCache && "No stat cache provided?");
70  if (AtBeginning || !StatCache.get()) {
71    statCache->setNextStatCache(StatCache.release());
72    StatCache.reset(statCache);
73    return;
74  }
75
76  FileSystemStatCache *LastCache = StatCache.get();
77  while (LastCache->getNextStatCache())
78    LastCache = LastCache->getNextStatCache();
79
80  LastCache->setNextStatCache(statCache);
81}
82
83void FileManager::removeStatCache(FileSystemStatCache *statCache) {
84  if (!statCache)
85    return;
86
87  if (StatCache.get() == statCache) {
88    // This is the first stat cache.
89    StatCache.reset(StatCache->takeNextStatCache());
90    return;
91  }
92
93  // Find the stat cache in the list.
94  FileSystemStatCache *PrevCache = StatCache.get();
95  while (PrevCache && PrevCache->getNextStatCache() != statCache)
96    PrevCache = PrevCache->getNextStatCache();
97
98  assert(PrevCache && "Stat cache not found for removal");
99  PrevCache->setNextStatCache(statCache->getNextStatCache());
100}
101
102void FileManager::clearStatCaches() {
103  StatCache.reset(nullptr);
104}
105
106/// \brief Retrieve the directory that the given file name resides in.
107/// Filename can point to either a real file or a virtual file.
108static const DirectoryEntry *getDirectoryFromFile(FileManager &FileMgr,
109                                                  StringRef Filename,
110                                                  bool CacheFailure) {
111  if (Filename.empty())
112    return nullptr;
113
114  if (llvm::sys::path::is_separator(Filename[Filename.size() - 1]))
115    return nullptr; // If Filename is a directory.
116
117  StringRef DirName = llvm::sys::path::parent_path(Filename);
118  // Use the current directory if file has no path component.
119  if (DirName.empty())
120    DirName = ".";
121
122  return FileMgr.getDirectory(DirName, CacheFailure);
123}
124
125/// Add all ancestors of the given path (pointing to either a file or
126/// a directory) as virtual directories.
127void FileManager::addAncestorsAsVirtualDirs(StringRef Path) {
128  StringRef DirName = llvm::sys::path::parent_path(Path);
129  if (DirName.empty())
130    return;
131
132  llvm::StringMapEntry<DirectoryEntry *> &NamedDirEnt =
133    SeenDirEntries.GetOrCreateValue(DirName);
134
135  // When caching a virtual directory, we always cache its ancestors
136  // at the same time.  Therefore, if DirName is already in the cache,
137  // we don't need to recurse as its ancestors must also already be in
138  // the cache.
139  if (NamedDirEnt.getValue())
140    return;
141
142  // Add the virtual directory to the cache.
143  DirectoryEntry *UDE = new DirectoryEntry;
144  UDE->Name = NamedDirEnt.getKeyData();
145  NamedDirEnt.setValue(UDE);
146  VirtualDirectoryEntries.push_back(UDE);
147
148  // Recursively add the other ancestors.
149  addAncestorsAsVirtualDirs(DirName);
150}
151
152const DirectoryEntry *FileManager::getDirectory(StringRef DirName,
153                                                bool CacheFailure) {
154  // stat doesn't like trailing separators except for root directory.
155  // At least, on Win32 MSVCRT, stat() cannot strip trailing '/'.
156  // (though it can strip '\\')
157  if (DirName.size() > 1 &&
158      DirName != llvm::sys::path::root_path(DirName) &&
159      llvm::sys::path::is_separator(DirName.back()))
160    DirName = DirName.substr(0, DirName.size()-1);
161#ifdef LLVM_ON_WIN32
162  // Fixing a problem with "clang C:test.c" on Windows.
163  // Stat("C:") does not recognize "C:" as a valid directory
164  std::string DirNameStr;
165  if (DirName.size() > 1 && DirName.back() == ':' &&
166      DirName.equals_lower(llvm::sys::path::root_name(DirName))) {
167    DirNameStr = DirName.str() + '.';
168    DirName = DirNameStr;
169  }
170#endif
171
172  ++NumDirLookups;
173  llvm::StringMapEntry<DirectoryEntry *> &NamedDirEnt =
174    SeenDirEntries.GetOrCreateValue(DirName);
175
176  // See if there was already an entry in the map.  Note that the map
177  // contains both virtual and real directories.
178  if (NamedDirEnt.getValue())
179    return NamedDirEnt.getValue() == NON_EXISTENT_DIR ? nullptr
180                                                      : NamedDirEnt.getValue();
181
182  ++NumDirCacheMisses;
183
184  // By default, initialize it to invalid.
185  NamedDirEnt.setValue(NON_EXISTENT_DIR);
186
187  // Get the null-terminated directory name as stored as the key of the
188  // SeenDirEntries map.
189  const char *InterndDirName = NamedDirEnt.getKeyData();
190
191  // Check to see if the directory exists.
192  FileData Data;
193  if (getStatValue(InterndDirName, Data, false, nullptr /*directory lookup*/)) {
194    // There's no real directory at the given path.
195    if (!CacheFailure)
196      SeenDirEntries.erase(DirName);
197    return nullptr;
198  }
199
200  // It exists.  See if we have already opened a directory with the
201  // same inode (this occurs on Unix-like systems when one dir is
202  // symlinked to another, for example) or the same path (on
203  // Windows).
204  DirectoryEntry &UDE = UniqueRealDirs[Data.UniqueID];
205
206  NamedDirEnt.setValue(&UDE);
207  if (!UDE.getName()) {
208    // We don't have this directory yet, add it.  We use the string
209    // key from the SeenDirEntries map as the string.
210    UDE.Name  = InterndDirName;
211  }
212
213  return &UDE;
214}
215
216const FileEntry *FileManager::getFile(StringRef Filename, bool openFile,
217                                      bool CacheFailure) {
218  ++NumFileLookups;
219
220  // See if there is already an entry in the map.
221  llvm::StringMapEntry<FileEntry *> &NamedFileEnt =
222    SeenFileEntries.GetOrCreateValue(Filename);
223
224  // See if there is already an entry in the map.
225  if (NamedFileEnt.getValue())
226    return NamedFileEnt.getValue() == NON_EXISTENT_FILE
227                 ? nullptr : NamedFileEnt.getValue();
228
229  ++NumFileCacheMisses;
230
231  // By default, initialize it to invalid.
232  NamedFileEnt.setValue(NON_EXISTENT_FILE);
233
234  // Get the null-terminated file name as stored as the key of the
235  // SeenFileEntries map.
236  const char *InterndFileName = NamedFileEnt.getKeyData();
237
238  // Look up the directory for the file.  When looking up something like
239  // sys/foo.h we'll discover all of the search directories that have a 'sys'
240  // subdirectory.  This will let us avoid having to waste time on known-to-fail
241  // searches when we go to find sys/bar.h, because all the search directories
242  // without a 'sys' subdir will get a cached failure result.
243  const DirectoryEntry *DirInfo = getDirectoryFromFile(*this, Filename,
244                                                       CacheFailure);
245  if (DirInfo == nullptr) { // Directory doesn't exist, file can't exist.
246    if (!CacheFailure)
247      SeenFileEntries.erase(Filename);
248
249    return nullptr;
250  }
251
252  // FIXME: Use the directory info to prune this, before doing the stat syscall.
253  // FIXME: This will reduce the # syscalls.
254
255  // Nope, there isn't.  Check to see if the file exists.
256  std::unique_ptr<vfs::File> F;
257  FileData Data;
258  if (getStatValue(InterndFileName, Data, true, openFile ? &F : nullptr)) {
259    // There's no real file at the given path.
260    if (!CacheFailure)
261      SeenFileEntries.erase(Filename);
262
263    return nullptr;
264  }
265
266  assert((openFile || !F) && "undesired open file");
267
268  // It exists.  See if we have already opened a file with the same inode.
269  // This occurs when one dir is symlinked to another, for example.
270  FileEntry &UFE = UniqueRealFiles[Data.UniqueID];
271
272  NamedFileEnt.setValue(&UFE);
273  if (UFE.isValid()) { // Already have an entry with this inode, return it.
274
275    // FIXME: this hack ensures that if we look up a file by a virtual path in
276    // the VFS that the getDir() will have the virtual path, even if we found
277    // the file by a 'real' path first. This is required in order to find a
278    // module's structure when its headers/module map are mapped in the VFS.
279    // We should remove this as soon as we can properly support a file having
280    // multiple names.
281    if (DirInfo != UFE.Dir && Data.IsVFSMapped)
282      UFE.Dir = DirInfo;
283
284    return &UFE;
285  }
286
287  // Otherwise, we don't have this file yet, add it.
288  UFE.Name    = Data.Name;
289  UFE.Size = Data.Size;
290  UFE.ModTime = Data.ModTime;
291  UFE.Dir     = DirInfo;
292  UFE.UID     = NextFileUID++;
293  UFE.UniqueID = Data.UniqueID;
294  UFE.IsNamedPipe = Data.IsNamedPipe;
295  UFE.InPCH = Data.InPCH;
296  UFE.File = std::move(F);
297  UFE.IsValid = true;
298  return &UFE;
299}
300
301const FileEntry *
302FileManager::getVirtualFile(StringRef Filename, off_t Size,
303                            time_t ModificationTime) {
304  ++NumFileLookups;
305
306  // See if there is already an entry in the map.
307  llvm::StringMapEntry<FileEntry *> &NamedFileEnt =
308    SeenFileEntries.GetOrCreateValue(Filename);
309
310  // See if there is already an entry in the map.
311  if (NamedFileEnt.getValue() && NamedFileEnt.getValue() != NON_EXISTENT_FILE)
312    return NamedFileEnt.getValue();
313
314  ++NumFileCacheMisses;
315
316  // By default, initialize it to invalid.
317  NamedFileEnt.setValue(NON_EXISTENT_FILE);
318
319  addAncestorsAsVirtualDirs(Filename);
320  FileEntry *UFE = nullptr;
321
322  // Now that all ancestors of Filename are in the cache, the
323  // following call is guaranteed to find the DirectoryEntry from the
324  // cache.
325  const DirectoryEntry *DirInfo = getDirectoryFromFile(*this, Filename,
326                                                       /*CacheFailure=*/true);
327  assert(DirInfo &&
328         "The directory of a virtual file should already be in the cache.");
329
330  // Check to see if the file exists. If so, drop the virtual file
331  FileData Data;
332  const char *InterndFileName = NamedFileEnt.getKeyData();
333  if (getStatValue(InterndFileName, Data, true, nullptr) == 0) {
334    Data.Size = Size;
335    Data.ModTime = ModificationTime;
336    UFE = &UniqueRealFiles[Data.UniqueID];
337
338    NamedFileEnt.setValue(UFE);
339
340    // If we had already opened this file, close it now so we don't
341    // leak the descriptor. We're not going to use the file
342    // descriptor anyway, since this is a virtual file.
343    if (UFE->File)
344      UFE->closeFile();
345
346    // If we already have an entry with this inode, return it.
347    if (UFE->isValid())
348      return UFE;
349
350    UFE->UniqueID = Data.UniqueID;
351    UFE->IsNamedPipe = Data.IsNamedPipe;
352    UFE->InPCH = Data.InPCH;
353  }
354
355  if (!UFE) {
356    UFE = new FileEntry();
357    VirtualFileEntries.push_back(UFE);
358    NamedFileEnt.setValue(UFE);
359  }
360
361  UFE->Name    = InterndFileName;
362  UFE->Size    = Size;
363  UFE->ModTime = ModificationTime;
364  UFE->Dir     = DirInfo;
365  UFE->UID     = NextFileUID++;
366  UFE->File.reset();
367  return UFE;
368}
369
370void FileManager::FixupRelativePath(SmallVectorImpl<char> &path) const {
371  StringRef pathRef(path.data(), path.size());
372
373  if (FileSystemOpts.WorkingDir.empty()
374      || llvm::sys::path::is_absolute(pathRef))
375    return;
376
377  SmallString<128> NewPath(FileSystemOpts.WorkingDir);
378  llvm::sys::path::append(NewPath, pathRef);
379  path = NewPath;
380}
381
382llvm::MemoryBuffer *FileManager::
383getBufferForFile(const FileEntry *Entry, std::string *ErrorStr,
384                 bool isVolatile, bool ShouldCloseOpenFile) {
385  std::unique_ptr<llvm::MemoryBuffer> Result;
386  std::error_code ec;
387
388  uint64_t FileSize = Entry->getSize();
389  // If there's a high enough chance that the file have changed since we
390  // got its size, force a stat before opening it.
391  if (isVolatile)
392    FileSize = -1;
393
394  const char *Filename = Entry->getName();
395  // If the file is already open, use the open file descriptor.
396  if (Entry->File) {
397    ec = Entry->File->getBuffer(Filename, Result, FileSize,
398                                /*RequiresNullTerminator=*/true, isVolatile);
399    if (ErrorStr)
400      *ErrorStr = ec.message();
401    // FIXME: we need a set of APIs that can make guarantees about whether a
402    // FileEntry is open or not.
403    if (ShouldCloseOpenFile)
404      Entry->closeFile();
405    return Result.release();
406  }
407
408  // Otherwise, open the file.
409
410  if (FileSystemOpts.WorkingDir.empty()) {
411    ec = FS->getBufferForFile(Filename, Result, FileSize,
412                              /*RequiresNullTerminator=*/true, isVolatile);
413    if (ec && ErrorStr)
414      *ErrorStr = ec.message();
415    return Result.release();
416  }
417
418  SmallString<128> FilePath(Entry->getName());
419  FixupRelativePath(FilePath);
420  ec = FS->getBufferForFile(FilePath.str(), Result, FileSize,
421                            /*RequiresNullTerminator=*/true, isVolatile);
422  if (ec && ErrorStr)
423    *ErrorStr = ec.message();
424  return Result.release();
425}
426
427llvm::MemoryBuffer *FileManager::
428getBufferForFile(StringRef Filename, std::string *ErrorStr) {
429  std::unique_ptr<llvm::MemoryBuffer> Result;
430  std::error_code ec;
431  if (FileSystemOpts.WorkingDir.empty()) {
432    ec = FS->getBufferForFile(Filename, Result);
433    if (ec && ErrorStr)
434      *ErrorStr = ec.message();
435    return Result.release();
436  }
437
438  SmallString<128> FilePath(Filename);
439  FixupRelativePath(FilePath);
440  ec = FS->getBufferForFile(FilePath.c_str(), Result);
441  if (ec && ErrorStr)
442    *ErrorStr = ec.message();
443  return Result.release();
444}
445
446/// getStatValue - Get the 'stat' information for the specified path,
447/// using the cache to accelerate it if possible.  This returns true
448/// if the path points to a virtual file or does not exist, or returns
449/// false if it's an existent real file.  If FileDescriptor is NULL,
450/// do directory look-up instead of file look-up.
451bool FileManager::getStatValue(const char *Path, FileData &Data, bool isFile,
452                               std::unique_ptr<vfs::File> *F) {
453  // FIXME: FileSystemOpts shouldn't be passed in here, all paths should be
454  // absolute!
455  if (FileSystemOpts.WorkingDir.empty())
456    return FileSystemStatCache::get(Path, Data, isFile, F,StatCache.get(), *FS);
457
458  SmallString<128> FilePath(Path);
459  FixupRelativePath(FilePath);
460
461  return FileSystemStatCache::get(FilePath.c_str(), Data, isFile, F,
462                                  StatCache.get(), *FS);
463}
464
465bool FileManager::getNoncachedStatValue(StringRef Path,
466                                        vfs::Status &Result) {
467  SmallString<128> FilePath(Path);
468  FixupRelativePath(FilePath);
469
470  llvm::ErrorOr<vfs::Status> S = FS->status(FilePath.c_str());
471  if (!S)
472    return true;
473  Result = *S;
474  return false;
475}
476
477void FileManager::invalidateCache(const FileEntry *Entry) {
478  assert(Entry && "Cannot invalidate a NULL FileEntry");
479
480  SeenFileEntries.erase(Entry->getName());
481
482  // FileEntry invalidation should not block future optimizations in the file
483  // caches. Possible alternatives are cache truncation (invalidate last N) or
484  // invalidation of the whole cache.
485  UniqueRealFiles.erase(Entry->getUniqueID());
486}
487
488
489void FileManager::GetUniqueIDMapping(
490                   SmallVectorImpl<const FileEntry *> &UIDToFiles) const {
491  UIDToFiles.clear();
492  UIDToFiles.resize(NextFileUID);
493
494  // Map file entries
495  for (llvm::StringMap<FileEntry*, llvm::BumpPtrAllocator>::const_iterator
496         FE = SeenFileEntries.begin(), FEEnd = SeenFileEntries.end();
497       FE != FEEnd; ++FE)
498    if (FE->getValue() && FE->getValue() != NON_EXISTENT_FILE)
499      UIDToFiles[FE->getValue()->getUID()] = FE->getValue();
500
501  // Map virtual file entries
502  for (SmallVectorImpl<FileEntry *>::const_iterator
503         VFE = VirtualFileEntries.begin(), VFEEnd = VirtualFileEntries.end();
504       VFE != VFEEnd; ++VFE)
505    if (*VFE && *VFE != NON_EXISTENT_FILE)
506      UIDToFiles[(*VFE)->getUID()] = *VFE;
507}
508
509void FileManager::modifyFileEntry(FileEntry *File,
510                                  off_t Size, time_t ModificationTime) {
511  File->Size = Size;
512  File->ModTime = ModificationTime;
513}
514
515StringRef FileManager::getCanonicalName(const DirectoryEntry *Dir) {
516  // FIXME: use llvm::sys::fs::canonical() when it gets implemented
517#ifdef LLVM_ON_UNIX
518  llvm::DenseMap<const DirectoryEntry *, llvm::StringRef>::iterator Known
519    = CanonicalDirNames.find(Dir);
520  if (Known != CanonicalDirNames.end())
521    return Known->second;
522
523  StringRef CanonicalName(Dir->getName());
524  char CanonicalNameBuf[PATH_MAX];
525  if (realpath(Dir->getName(), CanonicalNameBuf)) {
526    unsigned Len = strlen(CanonicalNameBuf);
527    char *Mem = static_cast<char *>(CanonicalNameStorage.Allocate(Len, 1));
528    memcpy(Mem, CanonicalNameBuf, Len);
529    CanonicalName = StringRef(Mem, Len);
530  }
531
532  CanonicalDirNames.insert(std::make_pair(Dir, CanonicalName));
533  return CanonicalName;
534#else
535  return StringRef(Dir->getName());
536#endif
537}
538
539void FileManager::PrintStats() const {
540  llvm::errs() << "\n*** File Manager Stats:\n";
541  llvm::errs() << UniqueRealFiles.size() << " real files found, "
542               << UniqueRealDirs.size() << " real dirs found.\n";
543  llvm::errs() << VirtualFileEntries.size() << " virtual files found, "
544               << VirtualDirectoryEntries.size() << " virtual dirs found.\n";
545  llvm::errs() << NumDirLookups << " dir lookups, "
546               << NumDirCacheMisses << " dir cache misses.\n";
547  llvm::errs() << NumFileLookups << " file lookups, "
548               << NumFileCacheMisses << " file cache misses.\n";
549
550  //llvm::errs() << PagesMapped << BytesOfPagesMapped << FSLookups;
551}
552