[clang][modules] Virtualize module cache pruning

[jansvoboda11]

This PR virtualizes module cache pruning via the new ModuleCache interface. Currently this is an NFC, but I left a FIXME in InProcessModuleCache to make this more efficient for the dependency scanner.

[llvmbot]

@llvm/pr-subscribers-clang

@llvm/pr-subscribers-clang-modules

Author: Jan Svoboda (jansvoboda11)

Changes

This PR virtualizes module cache pruning via the new ModuleCache interface. Currently this is an NFC, but I left a FIXME in InProcessModuleCache to make this more efficient for the dependency scanner.

---

Full diff: https://github.com/llvm/llvm-project/pull/149113.diff

4 Files Affected:

• (modified) clang/include/clang/Serialization/ModuleCache.h (+8-1)
• (modified) clang/lib/Frontend/CompilerInstance.cpp (+4-89)
• (modified) clang/lib/Serialization/ModuleCache.cpp (+86)
• (modified) clang/lib/Tooling/DependencyScanning/InProcessModuleCache.cpp (+7)

diff --git a/clang/include/clang/Serialization/ModuleCache.h b/clang/include/clang/Serialization/ModuleCache.h
index 3117d954a09cc..ec052c5c18e0a 100644
--- a/clang/include/clang/Serialization/ModuleCache.h
+++ b/clang/include/clang/Serialization/ModuleCache.h
@@ -45,11 +45,15 @@ class ModuleCache : public RefCountedBase<ModuleCache> {
   /// were validated.
   virtual void updateModuleTimestamp(StringRef ModuleFilename) = 0;
 
+  /// Prune module files that haven't been accessed in a long time.
+  virtual void maybePrune(StringRef Path, time_t PruneInterval,
+                          time_t PruneAfter) = 0;
+
   /// Returns this process's view of the module cache.
   virtual InMemoryModuleCache &getInMemoryModuleCache() = 0;
   virtual const InMemoryModuleCache &getInMemoryModuleCache() const = 0;
 
-  // TODO: Virtualize writing/reading PCM files, pruning, etc.
+  // TODO: Virtualize writing/reading PCM files, etc.
 
   virtual ~ModuleCache() = default;
 };
@@ -59,6 +63,9 @@ class ModuleCache : public RefCountedBase<ModuleCache> {
 /// \c CompilerInstance instances participating in building modules for single
 /// translation unit in order to share the same \c InMemoryModuleCache.
 IntrusiveRefCntPtr<ModuleCache> createCrossProcessModuleCache();
+
+/// Shared implementation of `ModuleCache::maybePrune()`.
+void maybePruneImpl(StringRef Path, time_t PruneInterval, time_t PruneAfter);
 } // namespace clang
 
 #endif
diff --git a/clang/lib/Frontend/CompilerInstance.cpp b/clang/lib/Frontend/CompilerInstance.cpp
index c7b82db292a14..2f5cd84bb53e0 100644
--- a/clang/lib/Frontend/CompilerInstance.cpp
+++ b/clang/lib/Frontend/CompilerInstance.cpp
@@ -1636,90 +1636,6 @@ static void checkConfigMacros(Preprocessor &PP, Module *M,
   }
 }
 
-/// Write a new timestamp file with the given path.
-static void writeTimestampFile(StringRef TimestampFile) {
-  std::error_code EC;
-  llvm::raw_fd_ostream Out(TimestampFile.str(), EC, llvm::sys::fs::OF_None);
-}
-
-/// Prune the module cache of modules that haven't been accessed in
-/// a long time.
-static void pruneModuleCache(const HeaderSearchOptions &HSOpts) {
-  llvm::sys::fs::file_status StatBuf;
-  llvm::SmallString<128> TimestampFile;
-  TimestampFile = HSOpts.ModuleCachePath;
-  assert(!TimestampFile.empty());
-  llvm::sys::path::append(TimestampFile, "modules.timestamp");
-
-  // Try to stat() the timestamp file.
-  if (std::error_code EC = llvm::sys::fs::status(TimestampFile, StatBuf)) {
-    // If the timestamp file wasn't there, create one now.
-    if (EC == std::errc::no_such_file_or_directory) {
-      writeTimestampFile(TimestampFile);
-    }
-    return;
-  }
-
-  // Check whether the time stamp is older than our pruning interval.
-  // If not, do nothing.
-  time_t TimeStampModTime =
-      llvm::sys::toTimeT(StatBuf.getLastModificationTime());
-  time_t CurrentTime = time(nullptr);
-  if (CurrentTime - TimeStampModTime <= time_t(HSOpts.ModuleCachePruneInterval))
-    return;
-
-  // Write a new timestamp file so that nobody else attempts to prune.
-  // There is a benign race condition here, if two Clang instances happen to
-  // notice at the same time that the timestamp is out-of-date.
-  writeTimestampFile(TimestampFile);
-
-  // Walk the entire module cache, looking for unused module files and module
-  // indices.
-  std::error_code EC;
-  for (llvm::sys::fs::directory_iterator Dir(HSOpts.ModuleCachePath, EC),
-       DirEnd;
-       Dir != DirEnd && !EC; Dir.increment(EC)) {
-    // If we don't have a directory, there's nothing to look into.
-    if (!llvm::sys::fs::is_directory(Dir->path()))
-      continue;
-
-    // Walk all of the files within this directory.
-    for (llvm::sys::fs::directory_iterator File(Dir->path(), EC), FileEnd;
-         File != FileEnd && !EC; File.increment(EC)) {
-      // We only care about module and global module index files.
-      StringRef Extension = llvm::sys::path::extension(File->path());
-      if (Extension != ".pcm" && Extension != ".timestamp" &&
-          llvm::sys::path::filename(File->path()) != "modules.idx")
-        continue;
-
-      // Look at this file. If we can't stat it, there's nothing interesting
-      // there.
-      if (llvm::sys::fs::status(File->path(), StatBuf))
-        continue;
-
-      // If the file has been used recently enough, leave it there.
-      time_t FileAccessTime = llvm::sys::toTimeT(StatBuf.getLastAccessedTime());
-      if (CurrentTime - FileAccessTime <=
-              time_t(HSOpts.ModuleCachePruneAfter)) {
-        continue;
-      }
-
-      // Remove the file.
-      llvm::sys::fs::remove(File->path());
-
-      // Remove the timestamp file.
-      std::string TimpestampFilename = File->path() + ".timestamp";
-      llvm::sys::fs::remove(TimpestampFilename);
-    }
-
-    // If we removed all of the files in the directory, remove the directory
-    // itself.
-    if (llvm::sys::fs::directory_iterator(Dir->path(), EC) ==
-            llvm::sys::fs::directory_iterator() && !EC)
-      llvm::sys::fs::remove(Dir->path());
-  }
-}
-
 void CompilerInstance::createASTReader() {
   if (TheASTReader)
     return;
@@ -1730,11 +1646,10 @@ void CompilerInstance::createASTReader() {
   // If we're implicitly building modules but not currently recursively
   // building a module, check whether we need to prune the module cache.
   if (getSourceManager().getModuleBuildStack().empty() &&
-      !getPreprocessor().getHeaderSearchInfo().getModuleCachePath().empty() &&
-      getHeaderSearchOpts().ModuleCachePruneInterval > 0 &&
-      getHeaderSearchOpts().ModuleCachePruneAfter > 0) {
-    pruneModuleCache(getHeaderSearchOpts());
-  }
+      !getPreprocessor().getHeaderSearchInfo().getModuleCachePath().empty())
+    ModCache->maybePrune(getHeaderSearchOpts().ModuleCachePath,
+                         getHeaderSearchOpts().ModuleCachePruneInterval,
+                         getHeaderSearchOpts().ModuleCachePruneAfter);
 
   HeaderSearchOptions &HSOpts = getHeaderSearchOpts();
   std::string Sysroot = HSOpts.Sysroot;
diff --git a/clang/lib/Serialization/ModuleCache.cpp b/clang/lib/Serialization/ModuleCache.cpp
index f42bdc16d815d..96687277ebafd 100644
--- a/clang/lib/Serialization/ModuleCache.cpp
+++ b/clang/lib/Serialization/ModuleCache.cpp
@@ -16,6 +16,87 @@
 
 using namespace clang;
 
+/// Write a new timestamp file with the given path.
+static void writeTimestampFile(StringRef TimestampFile) {
+  std::error_code EC;
+  llvm::raw_fd_ostream Out(TimestampFile.str(), EC, llvm::sys::fs::OF_None);
+}
+
+void clang::maybePruneImpl(StringRef Path, time_t PruneInterval,
+                           time_t PruneAfter) {
+  if (PruneInterval <= 0 || PruneAfter <= 0)
+    return;
+
+  llvm::SmallString<128> TimestampFile(Path);
+  llvm::sys::path::append(TimestampFile, "modules.timestamp");
+
+  // Try to stat() the timestamp file.
+  llvm::sys::fs::file_status StatBuf;
+  if (std::error_code EC = llvm::sys::fs::status(TimestampFile, StatBuf)) {
+    // If the timestamp file wasn't there, create one now.
+    if (EC == std::errc::no_such_file_or_directory)
+      writeTimestampFile(TimestampFile);
+    return;
+  }
+
+  // Check whether the time stamp is older than our pruning interval.
+  // If not, do nothing.
+  time_t TimestampModTime =
+      llvm::sys::toTimeT(StatBuf.getLastModificationTime());
+  time_t CurrentTime = time(nullptr);
+  if (CurrentTime - TimestampModTime <= PruneInterval)
+    return;
+
+  // Write a new timestamp file so that nobody else attempts to prune.
+  // There is a benign race condition here, if two Clang instances happen to
+  // notice at the same time that the timestamp is out-of-date.
+  writeTimestampFile(TimestampFile);
+
+  // Walk the entire module cache, looking for unused module files and module
+  // indices.
+  std::error_code EC;
+  for (llvm::sys::fs::directory_iterator Dir(Path, EC), DirEnd;
+       Dir != DirEnd && !EC; Dir.increment(EC)) {
+    // If we don't have a directory, there's nothing to look into.
+    if (!llvm::sys::fs::is_directory(Dir->path()))
+      continue;
+
+    // Walk all the files within this directory.
+    for (llvm::sys::fs::directory_iterator File(Dir->path(), EC), FileEnd;
+         File != FileEnd && !EC; File.increment(EC)) {
+      // We only care about module and global module index files.
+      StringRef Extension = llvm::sys::path::extension(File->path());
+      if (Extension != ".pcm" && Extension != ".timestamp" &&
+          llvm::sys::path::filename(File->path()) != "modules.idx")
+        continue;
+
+      // Look at this file. If we can't stat it, there's nothing interesting
+      // there.
+      if (llvm::sys::fs::status(File->path(), StatBuf))
+        continue;
+
+      // If the file has been used recently enough, leave it there.
+      time_t FileAccessTime = llvm::sys::toTimeT(StatBuf.getLastAccessedTime());
+      if (CurrentTime - FileAccessTime <= PruneAfter)
+        continue;
+
+      // Remove the file.
+      llvm::sys::fs::remove(File->path());
+
+      // Remove the timestamp file.
+      std::string TimpestampFilename = File->path() + ".timestamp";
+      llvm::sys::fs::remove(TimpestampFilename);
+    }
+
+    // If we removed all the files in the directory, remove the directory
+    // itself.
+    if (llvm::sys::fs::directory_iterator(Dir->path(), EC) ==
+            llvm::sys::fs::directory_iterator() &&
+        !EC)
+      llvm::sys::fs::remove(Dir->path());
+  }
+}
+
 namespace {
 class CrossProcessModuleCache : public ModuleCache {
   InMemoryModuleCache InMemory;
@@ -53,6 +134,11 @@ class CrossProcessModuleCache : public ModuleCache {
     OS.clear_error(); // Avoid triggering a fatal error.
   }
 
+  void maybePrune(StringRef Path, time_t PruneInterval,
+                  time_t PruneAfter) override {
+    maybePruneImpl(Path, PruneInterval, PruneAfter);
+  }
+
   InMemoryModuleCache &getInMemoryModuleCache() override { return InMemory; }
   const InMemoryModuleCache &getInMemoryModuleCache() const override {
     return InMemory;
diff --git a/clang/lib/Tooling/DependencyScanning/InProcessModuleCache.cpp b/clang/lib/Tooling/DependencyScanning/InProcessModuleCache.cpp
index 80db2d47d940e..d1e543b438225 100644
--- a/clang/lib/Tooling/DependencyScanning/InProcessModuleCache.cpp
+++ b/clang/lib/Tooling/DependencyScanning/InProcessModuleCache.cpp
@@ -100,6 +100,13 @@ class InProcessModuleCache : public ModuleCache {
     Timestamp.store(llvm::sys::toTimeT(std::chrono::system_clock::now()));
   }
 
+  void maybePrune(StringRef Path, time_t PruneInterval,
+                  time_t PruneAfter) override {
+    // FIXME: This only needs to be ran once per build, not in every
+    // compilation. Call it once per service.
+    maybePruneImpl(Path, PruneInterval, PruneAfter);
+  }
+
   InMemoryModuleCache &getInMemoryModuleCache() override { return InMemory; }
   const InMemoryModuleCache &getInMemoryModuleCache() const override {
     return InMemory;