1 //===- DependencyScanningFilesystem.cpp - clang-scan-deps fs --------------===//
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 #include "clang/Tooling/DependencyScanning/DependencyScanningFilesystem.h"
10 #include "clang/Lex/DependencyDirectivesSourceMinimizer.h"
11 #include "llvm/Support/MemoryBuffer.h"
12 #include "llvm/Support/Threading.h"
13 
14 using namespace clang;
15 using namespace tooling;
16 using namespace dependencies;
17 
18 CachedFileSystemEntry CachedFileSystemEntry::createFileEntry(
19     StringRef Filename, llvm::vfs::FileSystem &FS, bool Minimize) {
20   // Load the file and its content from the file system.
21   llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>> MaybeFile =
22       FS.openFileForRead(Filename);
23   if (!MaybeFile)
24     return MaybeFile.getError();
25   llvm::ErrorOr<llvm::vfs::Status> Stat = (*MaybeFile)->status();
26   if (!Stat)
27     return Stat.getError();
28 
29   llvm::vfs::File &F = **MaybeFile;
30   llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> MaybeBuffer =
31       F.getBuffer(Stat->getName());
32   if (!MaybeBuffer)
33     return MaybeBuffer.getError();
34 
35   llvm::SmallString<1024> MinimizedFileContents;
36   // Minimize the file down to directives that might affect the dependencies.
37   const auto &Buffer = *MaybeBuffer;
38   SmallVector<minimize_source_to_dependency_directives::Token, 64> Tokens;
39   if (!Minimize || minimizeSourceToDependencyDirectives(
40                        Buffer->getBuffer(), MinimizedFileContents, Tokens)) {
41     // Use the original file unless requested otherwise, or
42     // if the minimization failed.
43     // FIXME: Propage the diagnostic if desired by the client.
44     CachedFileSystemEntry Result;
45     Result.MaybeStat = std::move(*Stat);
46     Result.Contents.reserve(Buffer->getBufferSize() + 1);
47     Result.Contents.append(Buffer->getBufferStart(), Buffer->getBufferEnd());
48     // Implicitly null terminate the contents for Clang's lexer.
49     Result.Contents.push_back('\0');
50     Result.Contents.pop_back();
51     return Result;
52   }
53 
54   CachedFileSystemEntry Result;
55   size_t Size = MinimizedFileContents.size();
56   Result.MaybeStat = llvm::vfs::Status(Stat->getName(), Stat->getUniqueID(),
57                                        Stat->getLastModificationTime(),
58                                        Stat->getUser(), Stat->getGroup(), Size,
59                                        Stat->getType(), Stat->getPermissions());
60   // The contents produced by the minimizer must be null terminated.
61   assert(MinimizedFileContents.data()[MinimizedFileContents.size()] == '\0' &&
62          "not null terminated contents");
63   // Even though there's an implicit null terminator in the minimized contents,
64   // we want to temporarily make it explicit. This will ensure that the
65   // std::move will preserve it even if it needs to do a copy if the
66   // SmallString still has the small capacity.
67   MinimizedFileContents.push_back('\0');
68   Result.Contents = std::move(MinimizedFileContents);
69   // Now make the null terminator implicit again, so that Clang's lexer can find
70   // it right where the buffer ends.
71   Result.Contents.pop_back();
72 
73   // Compute the skipped PP ranges that speedup skipping over inactive
74   // preprocessor blocks.
75   llvm::SmallVector<minimize_source_to_dependency_directives::SkippedRange, 32>
76       SkippedRanges;
77   minimize_source_to_dependency_directives::computeSkippedRanges(Tokens,
78                                                                  SkippedRanges);
79   PreprocessorSkippedRangeMapping Mapping;
80   for (const auto &Range : SkippedRanges) {
81     if (Range.Length < 16) {
82       // Ignore small ranges as non-profitable.
83       // FIXME: This is a heuristic, its worth investigating the tradeoffs
84       // when it should be applied.
85       continue;
86     }
87     Mapping[Range.Offset] = Range.Length;
88   }
89   Result.PPSkippedRangeMapping = std::move(Mapping);
90 
91   return Result;
92 }
93 
94 CachedFileSystemEntry
95 CachedFileSystemEntry::createDirectoryEntry(llvm::vfs::Status &&Stat) {
96   assert(Stat.isDirectory() && "not a directory!");
97   auto Result = CachedFileSystemEntry();
98   Result.MaybeStat = std::move(Stat);
99   return Result;
100 }
101 
102 DependencyScanningFilesystemSharedCache::
103     DependencyScanningFilesystemSharedCache() {
104   // This heuristic was chosen using a empirical testing on a
105   // reasonably high core machine (iMacPro 18 cores / 36 threads). The cache
106   // sharding gives a performance edge by reducing the lock contention.
107   // FIXME: A better heuristic might also consider the OS to account for
108   // the different cost of lock contention on different OSes.
109   NumShards =
110       std::max(2u, llvm::hardware_concurrency().compute_thread_count() / 4);
111   CacheShards = std::make_unique<CacheShard[]>(NumShards);
112 }
113 
114 /// Returns a cache entry for the corresponding key.
115 ///
116 /// A new cache entry is created if the key is not in the cache. This is a
117 /// thread safe call.
118 DependencyScanningFilesystemSharedCache::SharedFileSystemEntry &
119 DependencyScanningFilesystemSharedCache::get(StringRef Key) {
120   CacheShard &Shard = CacheShards[llvm::hash_value(Key) % NumShards];
121   std::unique_lock<std::mutex> LockGuard(Shard.CacheLock);
122   auto It = Shard.Cache.try_emplace(Key);
123   return It.first->getValue();
124 }
125 
126 /// Whitelist file extensions that should be minimized, treating no extension as
127 /// a source file that should be minimized.
128 ///
129 /// This is kinda hacky, it would be better if we knew what kind of file Clang
130 /// was expecting instead.
131 static bool shouldMinimize(StringRef Filename) {
132   StringRef Ext = llvm::sys::path::extension(Filename);
133   if (Ext.empty())
134     return true; // C++ standard library
135   return llvm::StringSwitch<bool>(Ext)
136     .CasesLower(".c", ".cc", ".cpp", ".c++", ".cxx", true)
137     .CasesLower(".h", ".hh", ".hpp", ".h++", ".hxx", true)
138     .CasesLower(".m", ".mm", true)
139     .CasesLower(".i", ".ii", ".mi", ".mmi", true)
140     .CasesLower(".def", ".inc", true)
141     .Default(false);
142 }
143 
144 
145 static bool shouldCacheStatFailures(StringRef Filename) {
146   StringRef Ext = llvm::sys::path::extension(Filename);
147   if (Ext.empty())
148     return false; // This may be the module cache directory.
149   return shouldMinimize(Filename); // Only cache stat failures on source files.
150 }
151 
152 llvm::ErrorOr<const CachedFileSystemEntry *>
153 DependencyScanningWorkerFilesystem::getOrCreateFileSystemEntry(
154     const StringRef Filename) {
155   if (const CachedFileSystemEntry *Entry = getCachedEntry(Filename)) {
156     return Entry;
157   }
158 
159   // FIXME: Handle PCM/PCH files.
160   // FIXME: Handle module map files.
161 
162   bool KeepOriginalSource = IgnoredFiles.count(Filename) ||
163                             !shouldMinimize(Filename);
164   DependencyScanningFilesystemSharedCache::SharedFileSystemEntry
165       &SharedCacheEntry = SharedCache.get(Filename);
166   const CachedFileSystemEntry *Result;
167   {
168     std::unique_lock<std::mutex> LockGuard(SharedCacheEntry.ValueLock);
169     CachedFileSystemEntry &CacheEntry = SharedCacheEntry.Value;
170 
171     if (!CacheEntry.isValid()) {
172       llvm::vfs::FileSystem &FS = getUnderlyingFS();
173       auto MaybeStatus = FS.status(Filename);
174       if (!MaybeStatus) {
175         if (!shouldCacheStatFailures(Filename))
176           // HACK: We need to always restat non source files if the stat fails.
177           //   This is because Clang first looks up the module cache and module
178           //   files before building them, and then looks for them again. If we
179           //   cache the stat failure, it won't see them the second time.
180           return MaybeStatus.getError();
181         else
182           CacheEntry = CachedFileSystemEntry(MaybeStatus.getError());
183       } else if (MaybeStatus->isDirectory())
184         CacheEntry = CachedFileSystemEntry::createDirectoryEntry(
185             std::move(*MaybeStatus));
186       else
187         CacheEntry = CachedFileSystemEntry::createFileEntry(
188             Filename, FS, !KeepOriginalSource);
189     }
190 
191     Result = &CacheEntry;
192   }
193 
194   // Store the result in the local cache.
195   setCachedEntry(Filename, Result);
196   return Result;
197 }
198 
199 llvm::ErrorOr<llvm::vfs::Status>
200 DependencyScanningWorkerFilesystem::status(const Twine &Path) {
201   SmallString<256> OwnedFilename;
202   StringRef Filename = Path.toStringRef(OwnedFilename);
203   const llvm::ErrorOr<const CachedFileSystemEntry *> Result =
204       getOrCreateFileSystemEntry(Filename);
205   if (!Result)
206     return Result.getError();
207   return (*Result)->getStatus();
208 }
209 
210 namespace {
211 
212 /// The VFS that is used by clang consumes the \c CachedFileSystemEntry using
213 /// this subclass.
214 class MinimizedVFSFile final : public llvm::vfs::File {
215 public:
216   MinimizedVFSFile(std::unique_ptr<llvm::MemoryBuffer> Buffer,
217                    llvm::vfs::Status Stat)
218       : Buffer(std::move(Buffer)), Stat(std::move(Stat)) {}
219 
220   static llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>
221   create(const CachedFileSystemEntry *Entry,
222          ExcludedPreprocessorDirectiveSkipMapping *PPSkipMappings);
223 
224   llvm::ErrorOr<llvm::vfs::Status> status() override { return Stat; }
225 
226   llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>>
227   getBuffer(const Twine &Name, int64_t FileSize, bool RequiresNullTerminator,
228             bool IsVolatile) override {
229     return std::move(Buffer);
230   }
231 
232   std::error_code close() override { return {}; }
233 
234 private:
235   std::unique_ptr<llvm::MemoryBuffer> Buffer;
236   llvm::vfs::Status Stat;
237 };
238 
239 } // end anonymous namespace
240 
241 llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>> MinimizedVFSFile::create(
242     const CachedFileSystemEntry *Entry,
243     ExcludedPreprocessorDirectiveSkipMapping *PPSkipMappings) {
244   if (Entry->isDirectory())
245     return llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>(
246         std::make_error_code(std::errc::is_a_directory));
247   llvm::ErrorOr<StringRef> Contents = Entry->getContents();
248   if (!Contents)
249     return Contents.getError();
250   auto Result = std::make_unique<MinimizedVFSFile>(
251       llvm::MemoryBuffer::getMemBuffer(*Contents, Entry->getName(),
252                                        /*RequiresNullTerminator=*/false),
253       *Entry->getStatus());
254   if (!Entry->getPPSkippedRangeMapping().empty() && PPSkipMappings)
255     (*PPSkipMappings)[Result->Buffer->getBufferStart()] =
256         &Entry->getPPSkippedRangeMapping();
257   return llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>(
258       std::unique_ptr<llvm::vfs::File>(std::move(Result)));
259 }
260 
261 llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>
262 DependencyScanningWorkerFilesystem::openFileForRead(const Twine &Path) {
263   SmallString<256> OwnedFilename;
264   StringRef Filename = Path.toStringRef(OwnedFilename);
265 
266   const llvm::ErrorOr<const CachedFileSystemEntry *> Result =
267       getOrCreateFileSystemEntry(Filename);
268   if (!Result)
269     return Result.getError();
270   return MinimizedVFSFile::create(Result.get(), PPSkipMappings);
271 }
272