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 = std::max(2u, llvm::hardware_concurrency() / 4);
110   CacheShards = std::make_unique<CacheShard[]>(NumShards);
111 }
112 
113 /// Returns a cache entry for the corresponding key.
114 ///
115 /// A new cache entry is created if the key is not in the cache. This is a
116 /// thread safe call.
117 DependencyScanningFilesystemSharedCache::SharedFileSystemEntry &
118 DependencyScanningFilesystemSharedCache::get(StringRef Key) {
119   CacheShard &Shard = CacheShards[llvm::hash_value(Key) % NumShards];
120   std::unique_lock<std::mutex> LockGuard(Shard.CacheLock);
121   auto It = Shard.Cache.try_emplace(Key);
122   return It.first->getValue();
123 }
124 
125 llvm::ErrorOr<llvm::vfs::Status>
126 DependencyScanningWorkerFilesystem::status(const Twine &Path) {
127   SmallString<256> OwnedFilename;
128   StringRef Filename = Path.toStringRef(OwnedFilename);
129 
130   // Check the local cache first.
131   if (const CachedFileSystemEntry *Entry = getCachedEntry(Filename))
132     return Entry->getStatus();
133 
134   // FIXME: Handle PCM/PCH files.
135   // FIXME: Handle module map files.
136 
137   bool KeepOriginalSource = IgnoredFiles.count(Filename);
138   DependencyScanningFilesystemSharedCache::SharedFileSystemEntry
139       &SharedCacheEntry = SharedCache.get(Filename);
140   const CachedFileSystemEntry *Result;
141   {
142     std::unique_lock<std::mutex> LockGuard(SharedCacheEntry.ValueLock);
143     CachedFileSystemEntry &CacheEntry = SharedCacheEntry.Value;
144 
145     if (!CacheEntry.isValid()) {
146       llvm::vfs::FileSystem &FS = getUnderlyingFS();
147       auto MaybeStatus = FS.status(Filename);
148       if (!MaybeStatus)
149         CacheEntry = CachedFileSystemEntry(MaybeStatus.getError());
150       else if (MaybeStatus->isDirectory())
151         CacheEntry = CachedFileSystemEntry::createDirectoryEntry(
152             std::move(*MaybeStatus));
153       else
154         CacheEntry = CachedFileSystemEntry::createFileEntry(
155             Filename, FS, !KeepOriginalSource);
156     }
157 
158     Result = &CacheEntry;
159   }
160 
161   // Store the result in the local cache.
162   setCachedEntry(Filename, Result);
163   return Result->getStatus();
164 }
165 
166 namespace {
167 
168 /// The VFS that is used by clang consumes the \c CachedFileSystemEntry using
169 /// this subclass.
170 class MinimizedVFSFile final : public llvm::vfs::File {
171 public:
172   MinimizedVFSFile(std::unique_ptr<llvm::MemoryBuffer> Buffer,
173                    llvm::vfs::Status Stat)
174       : Buffer(std::move(Buffer)), Stat(std::move(Stat)) {}
175 
176   llvm::ErrorOr<llvm::vfs::Status> status() override { return Stat; }
177 
178   const llvm::MemoryBuffer *getBufferPtr() const { return Buffer.get(); }
179 
180   llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>>
181   getBuffer(const Twine &Name, int64_t FileSize, bool RequiresNullTerminator,
182             bool IsVolatile) override {
183     return std::move(Buffer);
184   }
185 
186   std::error_code close() override { return {}; }
187 
188 private:
189   std::unique_ptr<llvm::MemoryBuffer> Buffer;
190   llvm::vfs::Status Stat;
191 };
192 
193 llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>
194 createFile(const CachedFileSystemEntry *Entry,
195            ExcludedPreprocessorDirectiveSkipMapping *PPSkipMappings) {
196   if (Entry->isDirectory())
197     return llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>(
198         std::make_error_code(std::errc::is_a_directory));
199   llvm::ErrorOr<StringRef> Contents = Entry->getContents();
200   if (!Contents)
201     return Contents.getError();
202   auto Result = std::make_unique<MinimizedVFSFile>(
203       llvm::MemoryBuffer::getMemBuffer(*Contents, Entry->getName(),
204                                        /*RequiresNullTerminator=*/false),
205       *Entry->getStatus());
206   if (!Entry->getPPSkippedRangeMapping().empty() && PPSkipMappings)
207     (*PPSkipMappings)[Result->getBufferPtr()] =
208         &Entry->getPPSkippedRangeMapping();
209   return llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>(
210       std::unique_ptr<llvm::vfs::File>(std::move(Result)));
211 }
212 
213 } // end anonymous namespace
214 
215 llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>
216 DependencyScanningWorkerFilesystem::openFileForRead(const Twine &Path) {
217   SmallString<256> OwnedFilename;
218   StringRef Filename = Path.toStringRef(OwnedFilename);
219 
220   // Check the local cache first.
221   if (const CachedFileSystemEntry *Entry = getCachedEntry(Filename))
222     return createFile(Entry, PPSkipMappings);
223 
224   // FIXME: Handle PCM/PCH files.
225   // FIXME: Handle module map files.
226 
227   bool KeepOriginalSource = IgnoredFiles.count(Filename);
228   DependencyScanningFilesystemSharedCache::SharedFileSystemEntry
229       &SharedCacheEntry = SharedCache.get(Filename);
230   const CachedFileSystemEntry *Result;
231   {
232     std::unique_lock<std::mutex> LockGuard(SharedCacheEntry.ValueLock);
233     CachedFileSystemEntry &CacheEntry = SharedCacheEntry.Value;
234 
235     if (!CacheEntry.isValid()) {
236       CacheEntry = CachedFileSystemEntry::createFileEntry(
237           Filename, getUnderlyingFS(), !KeepOriginalSource);
238     }
239 
240     Result = &CacheEntry;
241   }
242 
243   // Store the result in the local cache.
244   setCachedEntry(Filename, Result);
245   return createFile(Result, PPSkipMappings);
246 }
247