Modules: Cache PCMs in memory and avoid a use-after-free
Clang's internal build system for implicit modules uses lock files to
ensure that after a process writes a PCM it will read the same one back
in (without contention from other -cc1 commands). Since PCMs are read
from disk repeatedly while invalidating, building, and importing, the
lock is not released quickly. Furthermore, the LockFileManager is not
robust in every environment. Other -cc1 commands can stall until
timeout (after about eight minutes).
This commit changes the lock file from being necessary for correctness
to a (possibly dubious) performance hack. The remaining benefit is to
reduce duplicate work in competing -cc1 commands which depend on the
same module. Follow-up commits will change the internal build system to
continue after a timeout, and reduce the timeout. Perhaps we should
reconsider blocking at all.
This also fixes a use-after-free, when one part of a compilation
validates a PCM and starts using it, and another tries to swap out the
PCM for something new.
The PCMCache is a new type called MemoryBufferCache, which saves memory
buffers based on their filename. Its ownership is shared by the
CompilerInstance and ModuleManager.
- The ModuleManager stores PCMs there that it loads from disk, never
touching the disk if the cache is hot.
- When modules fail to validate, they're removed from the cache.
- When a CompilerInstance is spawned to build a new module, each
already-loaded PCM is assumed to be valid, and is frozen to avoid
the use-after-free.
- Any newly-built module is written directly to the cache to avoid the
round-trip to the filesystem, making lock files unnecessary for
correctness.
Original patch by Manman Ren; most testcases by Adrian Prantl!
llvm-svn: 298165
diff --git a/clang/lib/Serialization/ASTReader.cpp b/clang/lib/Serialization/ASTReader.cpp
index e4eddae..eac0854 100644
--- a/clang/lib/Serialization/ASTReader.cpp
+++ b/clang/lib/Serialization/ASTReader.cpp
@@ -37,6 +37,7 @@
#include "clang/Basic/FileManager.h"
#include "clang/Basic/FileSystemOptions.h"
#include "clang/Basic/LangOptions.h"
+#include "clang/Basic/MemoryBufferCache.h"
#include "clang/Basic/ObjCRuntime.h"
#include "clang/Basic/OperatorKinds.h"
#include "clang/Basic/Sanitizers.h"
@@ -463,6 +464,30 @@
return checkDiagnosticGroupMappings(StoredDiags, Diags, Complain);
}
+/// Return the top import module if it is implicit, nullptr otherwise.
+static Module *getTopImportImplicitModule(ModuleManager &ModuleMgr,
+ Preprocessor &PP) {
+ // If the original import came from a file explicitly generated by the user,
+ // don't check the diagnostic mappings.
+ // FIXME: currently this is approximated by checking whether this is not a
+ // module import of an implicitly-loaded module file.
+ // Note: ModuleMgr.rbegin() may not be the current module, but it must be in
+ // the transitive closure of its imports, since unrelated modules cannot be
+ // imported until after this module finishes validation.
+ ModuleFile *TopImport = &*ModuleMgr.rbegin();
+ while (!TopImport->ImportedBy.empty())
+ TopImport = TopImport->ImportedBy[0];
+ if (TopImport->Kind != MK_ImplicitModule)
+ return nullptr;
+
+ StringRef ModuleName = TopImport->ModuleName;
+ assert(!ModuleName.empty() && "diagnostic options read before module name");
+
+ Module *M = PP.getHeaderSearchInfo().lookupModule(ModuleName);
+ assert(M && "missing module");
+ return M;
+}
+
bool PCHValidator::ReadDiagnosticOptions(
IntrusiveRefCntPtr<DiagnosticOptions> DiagOpts, bool Complain) {
DiagnosticsEngine &ExistingDiags = PP.getDiagnostics();
@@ -476,28 +501,14 @@
ModuleManager &ModuleMgr = Reader.getModuleManager();
assert(ModuleMgr.size() >= 1 && "what ASTFile is this then");
- // If the original import came from a file explicitly generated by the user,
- // don't check the diagnostic mappings.
- // FIXME: currently this is approximated by checking whether this is not a
- // module import of an implicitly-loaded module file.
- // Note: ModuleMgr.rbegin() may not be the current module, but it must be in
- // the transitive closure of its imports, since unrelated modules cannot be
- // imported until after this module finishes validation.
- ModuleFile *TopImport = &*ModuleMgr.rbegin();
- while (!TopImport->ImportedBy.empty())
- TopImport = TopImport->ImportedBy[0];
- if (TopImport->Kind != MK_ImplicitModule)
+ Module *TopM = getTopImportImplicitModule(ModuleMgr, PP);
+ if (!TopM)
return false;
- StringRef ModuleName = TopImport->ModuleName;
- assert(!ModuleName.empty() && "diagnostic options read before module name");
-
- Module *M = PP.getHeaderSearchInfo().lookupModule(ModuleName);
- assert(M && "missing module");
-
// FIXME: if the diagnostics are incompatible, save a DiagnosticOptions that
// contains the union of their flags.
- return checkDiagnosticMappings(*Diags, ExistingDiags, M->IsSystem, Complain);
+ return checkDiagnosticMappings(*Diags, ExistingDiags, TopM->IsSystem,
+ Complain);
}
/// \brief Collect the macro definitions provided by the given preprocessor
@@ -4064,12 +4075,41 @@
Listener.get(),
WasImportedBy ? false : HSOpts.ModulesValidateDiagnosticOptions);
+ // If F was directly imported by another module, it's implicitly validated by
+ // the importing module.
if (DisableValidation || WasImportedBy ||
(AllowConfigurationMismatch && Result == ConfigurationMismatch))
return Success;
- if (Result == Failure)
+ if (Result == Failure) {
Error("malformed block record in AST file");
+ return Failure;
+ }
+
+ if (Result == OutOfDate && F.Kind == MK_ImplicitModule) {
+ // If this module has already been finalized in the PCMCache, we're stuck
+ // with it; we can only load a single version of each module.
+ //
+ // This can happen when a module is imported in two contexts: in one, as a
+ // user module; in another, as a system module (due to an import from
+ // another module marked with the [system] flag). It usually indicates a
+ // bug in the module map: this module should also be marked with [system].
+ //
+ // If -Wno-system-headers (the default), and the first import is as a
+ // system module, then validation will fail during the as-user import,
+ // since -Werror flags won't have been validated. However, it's reasonable
+ // to treat this consistently as a system module.
+ //
+ // If -Wsystem-headers, the PCM on disk was built with
+ // -Wno-system-headers, and the first import is as a user module, then
+ // validation will fail during the as-system import since the PCM on disk
+ // doesn't guarantee that -Werror was respected. However, the -Werror
+ // flags were checked during the initial as-user import.
+ if (PCMCache.isBufferFinal(F.FileName)) {
+ Diag(diag::warn_module_system_bit_conflict) << F.FileName;
+ return Success;
+ }
+ }
return Result;
}
@@ -4122,7 +4162,7 @@
if (Listener && ValidateDiagnosticOptions &&
!AllowCompatibleConfigurationMismatch &&
ParseDiagnosticOptions(Record, Complain, *Listener))
- return OutOfDate;
+ Result = OutOfDate; // Don't return early. Read the signature.
break;
}
case DIAG_PRAGMA_MAPPINGS:
@@ -7301,7 +7341,7 @@
/// by heap-backed versus mmap'ed memory.
void ASTReader::getMemoryBufferSizes(MemoryBufferSizes &sizes) const {
for (ModuleFile &I : ModuleMgr) {
- if (llvm::MemoryBuffer *buf = I.Buffer.get()) {
+ if (llvm::MemoryBuffer *buf = I.Buffer) {
size_t bytes = buf->getBufferSize();
switch (buf->getBufferKind()) {
case llvm::MemoryBuffer::MemoryBuffer_Malloc:
@@ -9628,8 +9668,10 @@
: cast<ASTReaderListener>(new PCHValidator(PP, *this))),
SourceMgr(PP.getSourceManager()), FileMgr(PP.getFileManager()),
PCHContainerRdr(PCHContainerRdr), Diags(PP.getDiagnostics()), PP(PP),
- Context(Context), ModuleMgr(PP.getFileManager(), PCHContainerRdr),
- DummyIdResolver(PP), ReadTimer(std::move(ReadTimer)), isysroot(isysroot),
+ Context(Context),
+ ModuleMgr(PP.getFileManager(), PP.getPCMCache(), PCHContainerRdr),
+ PCMCache(PP.getPCMCache()), DummyIdResolver(PP),
+ ReadTimer(std::move(ReadTimer)), isysroot(isysroot),
DisableValidation(DisableValidation),
AllowASTWithCompilerErrors(AllowASTWithCompilerErrors),
AllowConfigurationMismatch(AllowConfigurationMismatch),
diff --git a/clang/lib/Serialization/ASTWriter.cpp b/clang/lib/Serialization/ASTWriter.cpp
index e9e64c2..524bc2d 100644
--- a/clang/lib/Serialization/ASTWriter.cpp
+++ b/clang/lib/Serialization/ASTWriter.cpp
@@ -35,6 +35,7 @@
#include "clang/Basic/FileSystemOptions.h"
#include "clang/Basic/LLVM.h"
#include "clang/Basic/LangOptions.h"
+#include "clang/Basic/MemoryBufferCache.h"
#include "clang/Basic/Module.h"
#include "clang/Basic/ObjCRuntime.h"
#include "clang/Basic/SourceManager.h"
@@ -4304,10 +4305,11 @@
}
ASTWriter::ASTWriter(llvm::BitstreamWriter &Stream,
- SmallVectorImpl<char> &Buffer,
+ SmallVectorImpl<char> &Buffer, MemoryBufferCache &PCMCache,
ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions,
bool IncludeTimestamps)
- : Stream(Stream), Buffer(Buffer), IncludeTimestamps(IncludeTimestamps) {
+ : Stream(Stream), Buffer(Buffer), PCMCache(PCMCache),
+ IncludeTimestamps(IncludeTimestamps) {
for (const auto &Ext : Extensions) {
if (auto Writer = Ext->createExtensionWriter(*this))
ModuleFileExtensionWriters.push_back(std::move(Writer));
@@ -4354,6 +4356,12 @@
this->BaseDirectory.clear();
WritingAST = false;
+ if (SemaRef.Context.getLangOpts().ImplicitModules && WritingModule) {
+ // Construct MemoryBuffer and update buffer manager.
+ PCMCache.addBuffer(OutputFile,
+ llvm::MemoryBuffer::getMemBufferCopy(
+ StringRef(Buffer.begin(), Buffer.size())));
+ }
return Signature;
}
diff --git a/clang/lib/Serialization/GeneratePCH.cpp b/clang/lib/Serialization/GeneratePCH.cpp
index 141a559..429ae8e 100644
--- a/clang/lib/Serialization/GeneratePCH.cpp
+++ b/clang/lib/Serialization/GeneratePCH.cpp
@@ -28,7 +28,8 @@
bool AllowASTWithErrors, bool IncludeTimestamps)
: PP(PP), OutputFile(OutputFile), isysroot(isysroot.str()),
SemaPtr(nullptr), Buffer(Buffer), Stream(Buffer->Data),
- Writer(Stream, Buffer->Data, Extensions, IncludeTimestamps),
+ Writer(Stream, Buffer->Data, PP.getPCMCache(), Extensions,
+ IncludeTimestamps),
AllowASTWithErrors(AllowASTWithErrors) {
Buffer->IsComplete = false;
}
diff --git a/clang/lib/Serialization/ModuleManager.cpp b/clang/lib/Serialization/ModuleManager.cpp
index f0654fc..d9c9fb7 100644
--- a/clang/lib/Serialization/ModuleManager.cpp
+++ b/clang/lib/Serialization/ModuleManager.cpp
@@ -12,6 +12,7 @@
//
//===----------------------------------------------------------------------===//
#include "clang/Serialization/ModuleManager.h"
+#include "clang/Basic/MemoryBufferCache.h"
#include "clang/Frontend/PCHContainerOperations.h"
#include "clang/Lex/HeaderSearch.h"
#include "clang/Lex/ModuleMap.h"
@@ -137,7 +138,9 @@
// Load the contents of the module
if (std::unique_ptr<llvm::MemoryBuffer> Buffer = lookupBuffer(FileName)) {
// The buffer was already provided for us.
- NewModule->Buffer = std::move(Buffer);
+ NewModule->Buffer = &PCMCache->addBuffer(FileName, std::move(Buffer));
+ } else if (llvm::MemoryBuffer *Buffer = PCMCache->lookupBuffer(FileName)) {
+ NewModule->Buffer = Buffer;
} else {
// Open the AST file.
llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> Buf((std::error_code()));
@@ -158,7 +161,7 @@
return Missing;
}
- NewModule->Buffer = std::move(*Buf);
+ NewModule->Buffer = &PCMCache->addBuffer(FileName, std::move(*Buf));
}
// Initialize the stream.
@@ -167,8 +170,13 @@
// Read the signature eagerly now so that we can check it. Avoid calling
// ReadSignature unless there's something to check though.
if (ExpectedSignature && checkSignature(ReadSignature(NewModule->Data),
- ExpectedSignature, ErrorStr))
+ ExpectedSignature, ErrorStr)) {
+ // Try to remove the buffer. If it can't be removed, then it was already
+ // validated by this process.
+ if (!PCMCache->tryToRemoveBuffer(NewModule->FileName))
+ FileMgr.invalidateCache(NewModule->File);
return OutOfDate;
+ }
// We're keeping this module. Store it everywhere.
Module = Modules[Entry] = NewModule.get();
@@ -235,7 +243,12 @@
// Files that didn't make it through ReadASTCore successfully will be
// rebuilt (or there was an error). Invalidate them so that we can load the
// new files that will be renamed over the old ones.
- if (LoadedSuccessfully.count(&*victim) == 0)
+ //
+ // The PCMCache tracks whether the module was succesfully loaded in another
+ // thread/context; in that case, it won't need to be rebuilt (and we can't
+ // safely invalidate it anyway).
+ if (LoadedSuccessfully.count(&*victim) == 0 &&
+ !PCMCache->tryToRemoveBuffer(victim->FileName))
FileMgr.invalidateCache(victim->File);
}
@@ -292,10 +305,10 @@
ModulesInCommonWithGlobalIndex.push_back(MF);
}
-ModuleManager::ModuleManager(FileManager &FileMgr,
+ModuleManager::ModuleManager(FileManager &FileMgr, MemoryBufferCache &PCMCache,
const PCHContainerReader &PCHContainerRdr)
- : FileMgr(FileMgr), PCHContainerRdr(PCHContainerRdr), GlobalIndex(),
- FirstVisitState(nullptr) {}
+ : FileMgr(FileMgr), PCMCache(&PCMCache), PCHContainerRdr(PCHContainerRdr),
+ GlobalIndex(), FirstVisitState(nullptr) {}
ModuleManager::~ModuleManager() { delete FirstVisitState; }