[C++20] [Modules] Bring Decls Hash to BMI for C++20 Module units #71627
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Close llvm#71034 This patch introduces thin BMI, which doesn't contain the definitions of functions and variables if its definitions won't contribute to the ABI. Testing is a big part of the patch. We want to make sure the thin BMI contains the same behavior with the existing and relatively stable fatBMI. This is pretty helpful for further reduction. For user interfeaces, this patch introduces `-fthinBMI-output=` arguments to specify the position of thin BMI. This should be used when compiling a single module unit. The design is helpful to use thin BMI in two phase compilations too. With thin BMI, In two phase compilations, we'll generate 2 BMIs, one thin BMI for being used by consumers, one fat BMI for compiling itself to object files. Maybe it sounds confusing to have 2 BMIs for one module unit. But only the thin BMI will be the BMI we're talking about generally and the fat BMI is only visible by the module unit itself. With one phase compilation, we may find the behavior of `-fthinBMI-output=` is pretty similar with `-fmodule-output=`, except one generating thin BMI and the other generating fat BMI. The design here is based on 2 things: (1) The serialization of C++ is pretty complex. We can't be sure we're handling every detail correctly in the every beginning. (2) The fat BMI is relatively widely used and relatively stable. So it looks not good to replace the fat BMI immediately with thin BMI. But, of course, in the end of the day, we want the consumers to use the thin BMI only. When that day comes, the `-fmodule-output=` will be an alias to `-fthinBMI-output=`. Another design choice is to reuse `-fmodule-output=` and introduce a flag `-femit-thin-BMI`. Then `-femit-thin-BMI -fmodule-output=` will have the same effect with `-fthinBMI-output=` now. The flag `-femit-thin-BMI` should be opt-in now and opt-off later and finally deprecated. The roadmap for thin BMI in my mind is: (1) In clang18, release thin BMI and mark it as experimental. Also encourage users and build systems to try this new mode. (2) In clang19 or clang20 (based on the issue feedbacks), remove the experimental mark for thin BMI and mark fat BMI as deprecated to be used by consumers. (3) In clang21 or clang22, error out if we found the users are trying to import a fat BMI.
llvmbot
added
clang
ChuanqiXu9
removed
clang
|
@llvm/pr-subscribers-clang-driver @llvm/pr-subscribers-clang Author: Chuanqi Xu (ChuanqiXu9) ChangesClose #71618 This contains #71622 and so that it is not easy to review this until we have stacked reviews. So the main purpose of current patch now is to have a feeling about what we want to do. The motivating example is: After we change the implementation of a() from return 43; to return 44;, we can avoid recompiling use.cc to use.o since the interface doesn't change. To be continued Patch is 157.49 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/71627.diff 122 Files Affected:
diff --git a/clang/include/clang/AST/ODRHash.h b/clang/include/clang/AST/ODRHash.h
index cedf644520fc320..5f5d8f99402edce 100644
--- a/clang/include/clang/AST/ODRHash.h
+++ b/clang/include/clang/AST/ODRHash.h
@@ -101,6 +101,9 @@ class ODRHash {
// Save booleans until the end to lower the size of data to process.
void AddBoolean(bool value);
+ // Add intergeers to ID.
+ void AddInteger(unsigned value);
+
static bool isSubDeclToBeProcessed(const Decl *D, const DeclContext *Parent);
private:
diff --git a/clang/include/clang/Basic/DiagnosticDriverKinds.td b/clang/include/clang/Basic/DiagnosticDriverKinds.td
index 676f1a62b49dd0d..aad67a9f4c7da01 100644
--- a/clang/include/clang/Basic/DiagnosticDriverKinds.td
+++ b/clang/include/clang/Basic/DiagnosticDriverKinds.td
@@ -158,6 +158,8 @@ def err_drv_invalid_output_with_multiple_archs : Error<
def err_drv_no_input_files : Error<"no input files">;
def err_drv_output_argument_with_multiple_files : Error<
"cannot specify -o when generating multiple output files">;
+def err_drv_thin_bmi_output_argument_with_multiple_files : Error <
+ "cannot specify -fthinBMI-output when generating multiple module files">;
def err_drv_out_file_argument_with_multiple_sources : Error<
"cannot specify '%0%1' when compiling multiple source files">;
def err_no_external_assembler : Error<
diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 36052511203f65c..1294b3ae1be29e9 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -2915,6 +2915,11 @@ def fmodule_output : Flag<["-"], "fmodule-output">, Flags<[NoXarchOption]>,
Visibility<[ClangOption, CC1Option]>,
HelpText<"Save intermediate module file results when compiling a standard C++ module unit.">;
+def fthinBMI_output_EQ : Joined<["-"], "fthinBMI-output=">, Group<f_Group>,
+ HelpText<"Specify the output path for the thin BMI for C++20 Named modules">,
+ Visibility<[ClangOption, CC1Option, CLOption, DXCOption]>,
+ MarshallingInfoString<FrontendOpts<"ThinBMIPath">>;
+
def fmodules_prune_interval : Joined<["-"], "fmodules-prune-interval=">, Group<i_Group>,
Visibility<[ClangOption, CC1Option]>, MetaVarName<"<seconds>">,
HelpText<"Specify the interval (in seconds) between attempts to prune the module cache">,
@@ -5088,6 +5093,9 @@ def muclibc : Flag<["-"], "muclibc">, Group<m_libc_Group>, Flags<[HelpHidden]>;
def module_file_info : Flag<["-"], "module-file-info">, Flags<[]>,
Visibility<[ClangOption, CC1Option]>, Group<Action_Group>,
HelpText<"Provide information about a particular module file">;
+def get_bmi_decls_hash : Flag<["-"], "get-bmi-decls-hash">, Flags<[]>,
+ Visibility<[ClangOption, CC1Option]>, Group<Action_Group>,
+ HelpText<"Get the BMI Decls hash value for a particular module file">;
def mthumb : Flag<["-"], "mthumb">, Group<m_Group>;
def mtune_EQ : Joined<["-"], "mtune=">, Group<m_Group>,
HelpText<"Only supported on AArch64, PowerPC, RISC-V, SystemZ, and X86">;
@@ -7223,7 +7231,9 @@ def ast_view : Flag<["-"], "ast-view">,
def emit_module : Flag<["-"], "emit-module">,
HelpText<"Generate pre-compiled module file from a module map">;
def emit_module_interface : Flag<["-"], "emit-module-interface">,
- HelpText<"Generate pre-compiled module file from a C++ module interface">;
+ HelpText<"Generate pre-compiled module file from a standard C++ module interface unit">;
+def emit_thin_module_interface : Flag<["-"], "emit-thin-module-interface">,
+ HelpText<"Generate reduced prebuilt module interface from a standard C++ module interface unit">;
def emit_header_unit : Flag<["-"], "emit-header-unit">,
HelpText<"Generate C++20 header units from header files">;
def emit_pch : Flag<["-"], "emit-pch">,
diff --git a/clang/include/clang/Frontend/FrontendActions.h b/clang/include/clang/Frontend/FrontendActions.h
index 3940e00eeb8dba7..d4ff4f8f6c57d91 100644
--- a/clang/include/clang/Frontend/FrontendActions.h
+++ b/clang/include/clang/Frontend/FrontendActions.h
@@ -118,6 +118,9 @@ class GenerateModuleAction : public ASTFrontendAction {
CreateOutputFile(CompilerInstance &CI, StringRef InFile) = 0;
protected:
+ std::vector<std::unique_ptr<ASTConsumer>>
+ CreateMultiplexConsumer(CompilerInstance &CI, StringRef InFile);
+
std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
StringRef InFile) override;
@@ -147,14 +150,27 @@ class GenerateModuleFromModuleMapAction : public GenerateModuleAction {
CreateOutputFile(CompilerInstance &CI, StringRef InFile) override;
};
+/// Generates fatBMI (which contains full information to generate the object
+/// files) for C++20 Named Modules. Also generates the thin BMI (only contains
+/// necessary information for importers) if `-fthinBMI-output=`.
class GenerateModuleInterfaceAction : public GenerateModuleAction {
-private:
+protected:
bool BeginSourceFileAction(CompilerInstance &CI) override;
+ std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
+ StringRef InFile) override;
+
std::unique_ptr<raw_pwrite_stream>
CreateOutputFile(CompilerInstance &CI, StringRef InFile) override;
};
+/// Only generates the thin BMI. This action is mainly used by tests.
+class GenerateThinModuleInterfaceAction : public GenerateModuleInterfaceAction {
+private:
+ std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
+ StringRef InFile) override;
+};
+
class GenerateHeaderUnitAction : public GenerateModuleAction {
private:
@@ -174,9 +190,8 @@ class SyntaxOnlyAction : public ASTFrontendAction {
bool hasCodeCompletionSupport() const override { return true; }
};
-/// Dump information about the given module file, to be used for
-/// basic debugging and discovery.
-class DumpModuleInfoAction : public ASTFrontendAction {
+// Base action for dumping module informations.
+class DumpModuleInfoActionBase : public ASTFrontendAction {
// Allow other tools (ex lldb) to direct output for their use.
std::shared_ptr<llvm::raw_ostream> OutputStream;
@@ -184,11 +199,12 @@ class DumpModuleInfoAction : public ASTFrontendAction {
std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
StringRef InFile) override;
bool BeginInvocation(CompilerInstance &CI) override;
- void ExecuteAction() override;
+ // Setup the output file.
+ llvm::raw_ostream &getOutputStream();
public:
- DumpModuleInfoAction() = default;
- explicit DumpModuleInfoAction(std::shared_ptr<llvm::raw_ostream> Out)
+ DumpModuleInfoActionBase() = default;
+ explicit DumpModuleInfoActionBase(std::shared_ptr<llvm::raw_ostream> Out)
: OutputStream(Out) {}
bool hasPCHSupport() const override { return false; }
bool hasASTFileSupport() const override { return true; }
@@ -196,6 +212,28 @@ class DumpModuleInfoAction : public ASTFrontendAction {
bool hasCodeCompletionSupport() const override { return false; }
};
+/// Dump information about the given module file, to be used for
+/// basic debugging and discovery.
+class DumpModuleInfoAction : public DumpModuleInfoActionBase {
+ void ExecuteAction() override;
+
+public:
+ DumpModuleInfoAction() = default;
+ explicit DumpModuleInfoAction(std::shared_ptr<llvm::raw_ostream> Out)
+ : DumpModuleInfoActionBase(Out) {}
+};
+
+/// Get the modules decl hash value action. The information is contained by
+/// DumpModuleInfoAction too. But this should be much faster.
+class GetModuleDeclsHashAction : public DumpModuleInfoActionBase {
+ void ExecuteAction() override;
+
+public:
+ GetModuleDeclsHashAction() = default;
+ explicit GetModuleDeclsHashAction(std::shared_ptr<llvm::raw_ostream> Out)
+ : DumpModuleInfoActionBase(Out) {}
+};
+
class VerifyPCHAction : public ASTFrontendAction {
protected:
std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
diff --git a/clang/include/clang/Frontend/FrontendOptions.h b/clang/include/clang/Frontend/FrontendOptions.h
index 53a8681cfdbba04..a4a23f3ca03f5be 100644
--- a/clang/include/clang/Frontend/FrontendOptions.h
+++ b/clang/include/clang/Frontend/FrontendOptions.h
@@ -85,9 +85,13 @@ enum ActionKind {
/// Generate pre-compiled module from a module map.
GenerateModule,
- /// Generate pre-compiled module from a C++ module interface file.
+ /// Generate pre-compiled module from a standard C++ module interface unit.
GenerateModuleInterface,
+ /// Generate reduced module interface for a standard C++ module interface
+ /// unit.
+ GenerateThinModuleInterface,
+
/// Generate a C++20 header unit module from a header file.
GenerateHeaderUnit,
@@ -103,6 +107,9 @@ enum ActionKind {
/// Dump information about a module file.
ModuleFileInfo,
+ /// Get BMI Decls Hash about a module file.
+ GetBMIDeclsHash,
+
/// Load and verify that a PCH file is usable.
VerifyPCH,
@@ -549,6 +556,9 @@ class FrontendOptions {
/// Path which stores the output files for -ftime-trace
std::string TimeTracePath;
+ /// Path to the thin BMI for -fthinbmi-output=
+ std::string ThinBMIPath;
+
public:
FrontendOptions()
: DisableFree(false), RelocatablePCH(false), ShowHelp(false),
diff --git a/clang/include/clang/Serialization/ASTBitCodes.h b/clang/include/clang/Serialization/ASTBitCodes.h
index 5c32fbc079c9a65..72a93bbb2ebd801 100644
--- a/clang/include/clang/Serialization/ASTBitCodes.h
+++ b/clang/include/clang/Serialization/ASTBitCodes.h
@@ -695,6 +695,9 @@ enum ASTRecordTypes {
/// Record code for an unterminated \#pragma clang assume_nonnull begin
/// recorded in a preamble.
PP_ASSUME_NONNULL_LOC = 67,
+
+ /// Record code for the decls hash in the thin BMI.
+ BMI_DECLS_HASH = 68,
};
/// Record types used within a source manager block.
diff --git a/clang/include/clang/Serialization/ASTReader.h b/clang/include/clang/Serialization/ASTReader.h
index 7eefdca6815cdad..a46ed6a95aae96f 100644
--- a/clang/include/clang/Serialization/ASTReader.h
+++ b/clang/include/clang/Serialization/ASTReader.h
@@ -937,6 +937,9 @@ class ASTReader
/// Sema tracks these to emit deferred diags.
llvm::SmallSetVector<serialization::DeclID, 4> DeclsToCheckForDeferredDiags;
+ /// The hash value of read C++20 thin BMI.
+ std::optional<uint64_t> ReadedBMIDeclsHash;
+
private:
struct ImportedSubmodule {
serialization::SubmoduleID ID;
@@ -1794,6 +1797,13 @@ class ASTReader
StringRef ExistingModuleCachePath,
bool RequireStrictOptionMatches = false);
+ static std::optional<uint64_t> getBMIHash(StringRef Filename,
+ FileManager &FileMgr);
+
+ std::optional<uint64_t> getReadedBMIDeclsHash() const {
+ return ReadedBMIDeclsHash;
+ }
+
/// Returns the suggested contents of the predefines buffer,
/// which contains a (typically-empty) subset of the predefines
/// build prior to including the precompiled header.
diff --git a/clang/include/clang/Serialization/ASTWriter.h b/clang/include/clang/Serialization/ASTWriter.h
index 3019bbc2ddc9cc7..5dfeaf9d6c62ba0 100644
--- a/clang/include/clang/Serialization/ASTWriter.h
+++ b/clang/include/clang/Serialization/ASTWriter.h
@@ -166,6 +166,15 @@ class ASTWriter : public ASTDeserializationListener,
/// Indicates that the AST contained compiler errors.
bool ASTHasCompilerErrors = false;
+ /// Indicates that we're going to generate the reduced BMI for C++20
+ /// named modules.
+ bool GeneratingThinBMI = false;
+
+ /// The hash for recorded decls for C++20 named modules. The parts of decls
+ /// which not affecting the ABI may not be recorded. e.g.,
+ /// the function body of a non-inline function.
+ llvm::hash_code BMIDeclsHash = 0;
+
/// Mapping from input file entries to the index into the
/// offset table where information about that input file is stored.
llvm::DenseMap<const FileEntry *, uint32_t> InputFileIDs;
@@ -582,7 +591,8 @@ class ASTWriter : public ASTDeserializationListener,
ASTWriter(llvm::BitstreamWriter &Stream, SmallVectorImpl<char> &Buffer,
InMemoryModuleCache &ModuleCache,
ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions,
- bool IncludeTimestamps = true, bool BuildingImplicitModule = false);
+ bool IncludeTimestamps = true, bool BuildingImplicitModule = false,
+ bool GeneratingThinBMI = false);
~ASTWriter() override;
ASTContext &getASTContext() const {
@@ -813,6 +823,13 @@ class PCHGenerator : public SemaConsumer {
const ASTWriter &getWriter() const { return Writer; }
SmallVectorImpl<char> &getPCH() const { return Buffer->Data; }
+ bool isComplete() const { return Buffer->IsComplete; }
+ PCHBuffer *getBufferPtr() { return Buffer.get(); }
+ StringRef getOutputFile() const { return OutputFile; }
+ DiagnosticsEngine &getDiagnostics() const {
+ return SemaPtr->getDiagnostics();
+ }
+
public:
PCHGenerator(const Preprocessor &PP, InMemoryModuleCache &ModuleCache,
StringRef OutputFile, StringRef isysroot,
@@ -820,7 +837,8 @@ class PCHGenerator : public SemaConsumer {
ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions,
bool AllowASTWithErrors = false, bool IncludeTimestamps = true,
bool BuildingImplicitModule = false,
- bool ShouldCacheASTInMemory = false);
+ bool ShouldCacheASTInMemory = false,
+ bool GeneratingThinBMI = false);
~PCHGenerator() override;
void InitializeSema(Sema &S) override { SemaPtr = &S; }
@@ -830,6 +848,19 @@ class PCHGenerator : public SemaConsumer {
bool hasEmittedPCH() const { return Buffer->IsComplete; }
};
+class ThinBMIGenerator : public PCHGenerator {
+public:
+ ThinBMIGenerator(const Preprocessor &PP, InMemoryModuleCache &ModuleCache,
+ StringRef OutputFile, std::shared_ptr<PCHBuffer> Buffer,
+ bool IncludeTimestamps);
+
+ void HandleTranslationUnit(ASTContext &Ctx) override;
+};
+
+/// If the definition may impact the ABI. If yes, we're allowed to eliminate
+/// the definition of D in thin BMI.
+bool MayDefAffectABI(const Decl *D);
+
/// A simple helper class to pack several bits in order into (a) 32 bit
/// integer(s).
class BitsPacker {
diff --git a/clang/lib/AST/ODRHash.cpp b/clang/lib/AST/ODRHash.cpp
index aea1a93ae1fa828..ace24eb4d29d852 100644
--- a/clang/lib/AST/ODRHash.cpp
+++ b/clang/lib/AST/ODRHash.cpp
@@ -1249,3 +1249,5 @@ void ODRHash::AddQualType(QualType T) {
void ODRHash::AddBoolean(bool Value) {
Bools.push_back(Value);
}
+
+void ODRHash::AddInteger(unsigned Value) { ID.AddInteger(Value); }
diff --git a/clang/lib/Driver/Driver.cpp b/clang/lib/Driver/Driver.cpp
index 6f5ff8141032677..42bd241f1ff70fe 100644
--- a/clang/lib/Driver/Driver.cpp
+++ b/clang/lib/Driver/Driver.cpp
@@ -354,6 +354,7 @@ phases::ID Driver::getFinalPhase(const DerivedArgList &DAL,
} else if ((PhaseArg = DAL.getLastArg(options::OPT_fsyntax_only)) ||
(PhaseArg = DAL.getLastArg(options::OPT_print_supported_cpus)) ||
(PhaseArg = DAL.getLastArg(options::OPT_module_file_info)) ||
+ (PhaseArg = DAL.getLastArg(options::OPT_get_bmi_decls_hash)) ||
(PhaseArg = DAL.getLastArg(options::OPT_verify_pch)) ||
(PhaseArg = DAL.getLastArg(options::OPT_rewrite_objc)) ||
(PhaseArg = DAL.getLastArg(options::OPT_rewrite_legacy_objc)) ||
@@ -4086,6 +4087,13 @@ void Driver::BuildActions(Compilation &C, DerivedArgList &Args,
}
}
+ // Diagnose misuse of -fthinBMI-output. It should be an error if we specify
+ // -fthinBMI-output with multiple precompilation jobs. Here we didn't check if
+ // there are multiple module units in the inputs.
+ if (C.getArgs().getLastArg(options::OPT_fthinBMI_output_EQ) &&
+ Inputs.size() > 1)
+ Diag(clang::diag::err_drv_thin_bmi_output_argument_with_multiple_files);
+
handleArguments(C, Args, Inputs, Actions);
bool UseNewOffloadingDriver =
@@ -4729,7 +4737,8 @@ Action *Driver::ConstructPhaseAction(
return C.MakeAction<MigrateJobAction>(Input, types::TY_Remap);
if (Args.hasArg(options::OPT_emit_ast))
return C.MakeAction<CompileJobAction>(Input, types::TY_AST);
- if (Args.hasArg(options::OPT_module_file_info))
+ if (Args.hasArg(options::OPT_module_file_info) ||
+ Args.hasArg(options::OPT_get_bmi_decls_hash))
return C.MakeAction<CompileJobAction>(Input, types::TY_ModuleFile);
if (Args.hasArg(options::OPT_verify_pch))
return C.MakeAction<VerifyPCHJobAction>(Input, types::TY_Nothing);
@@ -5811,7 +5820,8 @@ const char *Driver::GetNamedOutputPath(Compilation &C, const JobAction &JA,
}
if (JA.getType() == types::TY_ModuleFile &&
- C.getArgs().getLastArg(options::OPT_module_file_info)) {
+ (C.getArgs().getLastArg(options::OPT_module_file_info) ||
+ C.getArgs().getLastArg(options::OPT_get_bmi_decls_hash))) {
return "-";
}
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 22f992166ded6c0..c091a94ea0d7805 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -3940,6 +3940,8 @@ static bool RenderModulesOptions(Compilation &C, const Driver &D,
Args.ClaimAllArgs(options::OPT_fmodule_output);
Args.ClaimAllArgs(options::OPT_fmodule_output_EQ);
+ Args.AddLastArg(CmdArgs, options::OPT_fthinBMI_output_EQ);
+
return HaveModules;
}
@@ -4958,7 +4960,10 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
} else if (JA.getType() == types::TY_AST) {
CmdArgs.push_back("-emit-pch");
} else if (JA.getType() == types::TY_ModuleFile) {
- CmdArgs.push_back("-module-file-info");
+ if (Args.hasArg(options::OPT_get_bmi_decls_hash))
+ CmdArgs.push_back("-get-bmi-decls-hash");
+ else
+ CmdArgs.push_back("-module-file-info");
} else if (JA.getType() == types::TY_RewrittenObjC) {
CmdArgs.push_back("-rewrite-objc");
rewriteKind = RK_NonFragile;
diff --git a/clang/lib/Frontend/CompilerInvocation.cpp b/clang/lib/Frontend/CompilerInvocation.cpp
index 637c6a35af6532b..177b40a98578947 100644
--- a/clang/lib/Frontend/CompilerInvocation.cpp
+++ b/clang/lib/Frontend/CompilerInvocation.cpp
@@ -2554,12 +2554,14 @@ static const auto &getFrontendActionTable() {
{frontend::GenerateModule, OPT_emit_module},
{frontend::GenerateModuleInterface, OPT_emit_module_interface},
+ {frontend::GenerateThinModuleInterface, OPT_emit_thin_module_interface},
{frontend::GenerateHeaderUnit, OPT_emit_header_unit},
{frontend::GeneratePCH, OPT_emit_pch},
{frontend::GenerateInterfaceStubs, OPT_emit_interface_stubs},
{frontend::InitOnly, OPT_init_only},
{frontend::ParseSyntaxOnly, OPT_fsyntax_only},
{frontend::ModuleFileInfo, OPT_module_file_info},
+ {frontend::GetBMIDeclsHash, OPT_get_bmi_decls_hash},
{frontend::VerifyPCH, OPT_verify_pch},
{frontend::PrintPreamble, OPT_print_preamble},
{frontend::PrintPreprocessedInput, OPT_E},
@@ -4236,11 +4238,13 @@ static bool isStrictlyPreprocessorAction(frontend::ActionKind Action) {
case frontend::FixIt:
case frontend::GenerateModule:
case frontend::GenerateModuleInterface:
+ case frontend::GenerateThinModuleInterface:
case frontend::GenerateHeaderUnit:
case frontend::GeneratePCH:
case frontend::GenerateInterfaceStubs:
case frontend::ParseSyntaxOnly:
case frontend::ModuleFileInfo:
+ case frontend::GetBMIDeclsHash:
case frontend::VerifyPCH:
case frontend::PluginAction:
case frontend::RewriteObjC:
diff --git a/clang/lib/Frontend/FrontendActions.cpp b/clang/lib/Frontend/FrontendActions.cpp
index 2afcf1cf9f68c81..06b15079887c774 100644
--- a/clang/lib/Frontend/FrontendActions.cpp
+++ b/clang/lib/Frontend/FrontendActions.cpp
@@ -184,12 +184,12 @@ bool GeneratePCHAction::BeginSourceFileAction(CompilerInstance &CI) {
return true;
}
-std::unique_ptr<ASTConsumer>
-GenerateModuleAction::CreateASTConsumer(CompilerInstance &CI,
- StringRef InFile) {
+std::vector<std::unique_ptr<ASTConsumer...
[truncated]
|
|
@llvm/pr-subscribers-clang-driver Author: Chuanqi Xu (ChuanqiXu9) ChangesClose #71618 This contains #71622 and so that it is not easy to review this until we have stacked reviews. So the main purpose of current patch now is to have a feeling about what we want to do. The motivating example is: After we change the implementation of a() from return 43; to return 44;, we can avoid recompiling use.cc to use.o since the interface doesn't change. To be continued Patch is 157.49 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/71627.diff 122 Files Affected:
diff --git a/clang/include/clang/AST/ODRHash.h b/clang/include/clang/AST/ODRHash.h
index cedf644520fc320..5f5d8f99402edce 100644
--- a/clang/include/clang/AST/ODRHash.h
+++ b/clang/include/clang/AST/ODRHash.h
@@ -101,6 +101,9 @@ class ODRHash {
// Save booleans until the end to lower the size of data to process.
void AddBoolean(bool value);
+ // Add intergeers to ID.
+ void AddInteger(unsigned value);
+
static bool isSubDeclToBeProcessed(const Decl *D, const DeclContext *Parent);
private:
diff --git a/clang/include/clang/Basic/DiagnosticDriverKinds.td b/clang/include/clang/Basic/DiagnosticDriverKinds.td
index 676f1a62b49dd0d..aad67a9f4c7da01 100644
--- a/clang/include/clang/Basic/DiagnosticDriverKinds.td
+++ b/clang/include/clang/Basic/DiagnosticDriverKinds.td
@@ -158,6 +158,8 @@ def err_drv_invalid_output_with_multiple_archs : Error<
def err_drv_no_input_files : Error<"no input files">;
def err_drv_output_argument_with_multiple_files : Error<
"cannot specify -o when generating multiple output files">;
+def err_drv_thin_bmi_output_argument_with_multiple_files : Error <
+ "cannot specify -fthinBMI-output when generating multiple module files">;
def err_drv_out_file_argument_with_multiple_sources : Error<
"cannot specify '%0%1' when compiling multiple source files">;
def err_no_external_assembler : Error<
diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 36052511203f65c..1294b3ae1be29e9 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -2915,6 +2915,11 @@ def fmodule_output : Flag<["-"], "fmodule-output">, Flags<[NoXarchOption]>,
Visibility<[ClangOption, CC1Option]>,
HelpText<"Save intermediate module file results when compiling a standard C++ module unit.">;
+def fthinBMI_output_EQ : Joined<["-"], "fthinBMI-output=">, Group<f_Group>,
+ HelpText<"Specify the output path for the thin BMI for C++20 Named modules">,
+ Visibility<[ClangOption, CC1Option, CLOption, DXCOption]>,
+ MarshallingInfoString<FrontendOpts<"ThinBMIPath">>;
+
def fmodules_prune_interval : Joined<["-"], "fmodules-prune-interval=">, Group<i_Group>,
Visibility<[ClangOption, CC1Option]>, MetaVarName<"<seconds>">,
HelpText<"Specify the interval (in seconds) between attempts to prune the module cache">,
@@ -5088,6 +5093,9 @@ def muclibc : Flag<["-"], "muclibc">, Group<m_libc_Group>, Flags<[HelpHidden]>;
def module_file_info : Flag<["-"], "module-file-info">, Flags<[]>,
Visibility<[ClangOption, CC1Option]>, Group<Action_Group>,
HelpText<"Provide information about a particular module file">;
+def get_bmi_decls_hash : Flag<["-"], "get-bmi-decls-hash">, Flags<[]>,
+ Visibility<[ClangOption, CC1Option]>, Group<Action_Group>,
+ HelpText<"Get the BMI Decls hash value for a particular module file">;
def mthumb : Flag<["-"], "mthumb">, Group<m_Group>;
def mtune_EQ : Joined<["-"], "mtune=">, Group<m_Group>,
HelpText<"Only supported on AArch64, PowerPC, RISC-V, SystemZ, and X86">;
@@ -7223,7 +7231,9 @@ def ast_view : Flag<["-"], "ast-view">,
def emit_module : Flag<["-"], "emit-module">,
HelpText<"Generate pre-compiled module file from a module map">;
def emit_module_interface : Flag<["-"], "emit-module-interface">,
- HelpText<"Generate pre-compiled module file from a C++ module interface">;
+ HelpText<"Generate pre-compiled module file from a standard C++ module interface unit">;
+def emit_thin_module_interface : Flag<["-"], "emit-thin-module-interface">,
+ HelpText<"Generate reduced prebuilt module interface from a standard C++ module interface unit">;
def emit_header_unit : Flag<["-"], "emit-header-unit">,
HelpText<"Generate C++20 header units from header files">;
def emit_pch : Flag<["-"], "emit-pch">,
diff --git a/clang/include/clang/Frontend/FrontendActions.h b/clang/include/clang/Frontend/FrontendActions.h
index 3940e00eeb8dba7..d4ff4f8f6c57d91 100644
--- a/clang/include/clang/Frontend/FrontendActions.h
+++ b/clang/include/clang/Frontend/FrontendActions.h
@@ -118,6 +118,9 @@ class GenerateModuleAction : public ASTFrontendAction {
CreateOutputFile(CompilerInstance &CI, StringRef InFile) = 0;
protected:
+ std::vector<std::unique_ptr<ASTConsumer>>
+ CreateMultiplexConsumer(CompilerInstance &CI, StringRef InFile);
+
std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
StringRef InFile) override;
@@ -147,14 +150,27 @@ class GenerateModuleFromModuleMapAction : public GenerateModuleAction {
CreateOutputFile(CompilerInstance &CI, StringRef InFile) override;
};
+/// Generates fatBMI (which contains full information to generate the object
+/// files) for C++20 Named Modules. Also generates the thin BMI (only contains
+/// necessary information for importers) if `-fthinBMI-output=`.
class GenerateModuleInterfaceAction : public GenerateModuleAction {
-private:
+protected:
bool BeginSourceFileAction(CompilerInstance &CI) override;
+ std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
+ StringRef InFile) override;
+
std::unique_ptr<raw_pwrite_stream>
CreateOutputFile(CompilerInstance &CI, StringRef InFile) override;
};
+/// Only generates the thin BMI. This action is mainly used by tests.
+class GenerateThinModuleInterfaceAction : public GenerateModuleInterfaceAction {
+private:
+ std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
+ StringRef InFile) override;
+};
+
class GenerateHeaderUnitAction : public GenerateModuleAction {
private:
@@ -174,9 +190,8 @@ class SyntaxOnlyAction : public ASTFrontendAction {
bool hasCodeCompletionSupport() const override { return true; }
};
-/// Dump information about the given module file, to be used for
-/// basic debugging and discovery.
-class DumpModuleInfoAction : public ASTFrontendAction {
+// Base action for dumping module informations.
+class DumpModuleInfoActionBase : public ASTFrontendAction {
// Allow other tools (ex lldb) to direct output for their use.
std::shared_ptr<llvm::raw_ostream> OutputStream;
@@ -184,11 +199,12 @@ class DumpModuleInfoAction : public ASTFrontendAction {
std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
StringRef InFile) override;
bool BeginInvocation(CompilerInstance &CI) override;
- void ExecuteAction() override;
+ // Setup the output file.
+ llvm::raw_ostream &getOutputStream();
public:
- DumpModuleInfoAction() = default;
- explicit DumpModuleInfoAction(std::shared_ptr<llvm::raw_ostream> Out)
+ DumpModuleInfoActionBase() = default;
+ explicit DumpModuleInfoActionBase(std::shared_ptr<llvm::raw_ostream> Out)
: OutputStream(Out) {}
bool hasPCHSupport() const override { return false; }
bool hasASTFileSupport() const override { return true; }
@@ -196,6 +212,28 @@ class DumpModuleInfoAction : public ASTFrontendAction {
bool hasCodeCompletionSupport() const override { return false; }
};
+/// Dump information about the given module file, to be used for
+/// basic debugging and discovery.
+class DumpModuleInfoAction : public DumpModuleInfoActionBase {
+ void ExecuteAction() override;
+
+public:
+ DumpModuleInfoAction() = default;
+ explicit DumpModuleInfoAction(std::shared_ptr<llvm::raw_ostream> Out)
+ : DumpModuleInfoActionBase(Out) {}
+};
+
+/// Get the modules decl hash value action. The information is contained by
+/// DumpModuleInfoAction too. But this should be much faster.
+class GetModuleDeclsHashAction : public DumpModuleInfoActionBase {
+ void ExecuteAction() override;
+
+public:
+ GetModuleDeclsHashAction() = default;
+ explicit GetModuleDeclsHashAction(std::shared_ptr<llvm::raw_ostream> Out)
+ : DumpModuleInfoActionBase(Out) {}
+};
+
class VerifyPCHAction : public ASTFrontendAction {
protected:
std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
diff --git a/clang/include/clang/Frontend/FrontendOptions.h b/clang/include/clang/Frontend/FrontendOptions.h
index 53a8681cfdbba04..a4a23f3ca03f5be 100644
--- a/clang/include/clang/Frontend/FrontendOptions.h
+++ b/clang/include/clang/Frontend/FrontendOptions.h
@@ -85,9 +85,13 @@ enum ActionKind {
/// Generate pre-compiled module from a module map.
GenerateModule,
- /// Generate pre-compiled module from a C++ module interface file.
+ /// Generate pre-compiled module from a standard C++ module interface unit.
GenerateModuleInterface,
+ /// Generate reduced module interface for a standard C++ module interface
+ /// unit.
+ GenerateThinModuleInterface,
+
/// Generate a C++20 header unit module from a header file.
GenerateHeaderUnit,
@@ -103,6 +107,9 @@ enum ActionKind {
/// Dump information about a module file.
ModuleFileInfo,
+ /// Get BMI Decls Hash about a module file.
+ GetBMIDeclsHash,
+
/// Load and verify that a PCH file is usable.
VerifyPCH,
@@ -549,6 +556,9 @@ class FrontendOptions {
/// Path which stores the output files for -ftime-trace
std::string TimeTracePath;
+ /// Path to the thin BMI for -fthinbmi-output=
+ std::string ThinBMIPath;
+
public:
FrontendOptions()
: DisableFree(false), RelocatablePCH(false), ShowHelp(false),
diff --git a/clang/include/clang/Serialization/ASTBitCodes.h b/clang/include/clang/Serialization/ASTBitCodes.h
index 5c32fbc079c9a65..72a93bbb2ebd801 100644
--- a/clang/include/clang/Serialization/ASTBitCodes.h
+++ b/clang/include/clang/Serialization/ASTBitCodes.h
@@ -695,6 +695,9 @@ enum ASTRecordTypes {
/// Record code for an unterminated \#pragma clang assume_nonnull begin
/// recorded in a preamble.
PP_ASSUME_NONNULL_LOC = 67,
+
+ /// Record code for the decls hash in the thin BMI.
+ BMI_DECLS_HASH = 68,
};
/// Record types used within a source manager block.
diff --git a/clang/include/clang/Serialization/ASTReader.h b/clang/include/clang/Serialization/ASTReader.h
index 7eefdca6815cdad..a46ed6a95aae96f 100644
--- a/clang/include/clang/Serialization/ASTReader.h
+++ b/clang/include/clang/Serialization/ASTReader.h
@@ -937,6 +937,9 @@ class ASTReader
/// Sema tracks these to emit deferred diags.
llvm::SmallSetVector<serialization::DeclID, 4> DeclsToCheckForDeferredDiags;
+ /// The hash value of read C++20 thin BMI.
+ std::optional<uint64_t> ReadedBMIDeclsHash;
+
private:
struct ImportedSubmodule {
serialization::SubmoduleID ID;
@@ -1794,6 +1797,13 @@ class ASTReader
StringRef ExistingModuleCachePath,
bool RequireStrictOptionMatches = false);
+ static std::optional<uint64_t> getBMIHash(StringRef Filename,
+ FileManager &FileMgr);
+
+ std::optional<uint64_t> getReadedBMIDeclsHash() const {
+ return ReadedBMIDeclsHash;
+ }
+
/// Returns the suggested contents of the predefines buffer,
/// which contains a (typically-empty) subset of the predefines
/// build prior to including the precompiled header.
diff --git a/clang/include/clang/Serialization/ASTWriter.h b/clang/include/clang/Serialization/ASTWriter.h
index 3019bbc2ddc9cc7..5dfeaf9d6c62ba0 100644
--- a/clang/include/clang/Serialization/ASTWriter.h
+++ b/clang/include/clang/Serialization/ASTWriter.h
@@ -166,6 +166,15 @@ class ASTWriter : public ASTDeserializationListener,
/// Indicates that the AST contained compiler errors.
bool ASTHasCompilerErrors = false;
+ /// Indicates that we're going to generate the reduced BMI for C++20
+ /// named modules.
+ bool GeneratingThinBMI = false;
+
+ /// The hash for recorded decls for C++20 named modules. The parts of decls
+ /// which not affecting the ABI may not be recorded. e.g.,
+ /// the function body of a non-inline function.
+ llvm::hash_code BMIDeclsHash = 0;
+
/// Mapping from input file entries to the index into the
/// offset table where information about that input file is stored.
llvm::DenseMap<const FileEntry *, uint32_t> InputFileIDs;
@@ -582,7 +591,8 @@ class ASTWriter : public ASTDeserializationListener,
ASTWriter(llvm::BitstreamWriter &Stream, SmallVectorImpl<char> &Buffer,
InMemoryModuleCache &ModuleCache,
ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions,
- bool IncludeTimestamps = true, bool BuildingImplicitModule = false);
+ bool IncludeTimestamps = true, bool BuildingImplicitModule = false,
+ bool GeneratingThinBMI = false);
~ASTWriter() override;
ASTContext &getASTContext() const {
@@ -813,6 +823,13 @@ class PCHGenerator : public SemaConsumer {
const ASTWriter &getWriter() const { return Writer; }
SmallVectorImpl<char> &getPCH() const { return Buffer->Data; }
+ bool isComplete() const { return Buffer->IsComplete; }
+ PCHBuffer *getBufferPtr() { return Buffer.get(); }
+ StringRef getOutputFile() const { return OutputFile; }
+ DiagnosticsEngine &getDiagnostics() const {
+ return SemaPtr->getDiagnostics();
+ }
+
public:
PCHGenerator(const Preprocessor &PP, InMemoryModuleCache &ModuleCache,
StringRef OutputFile, StringRef isysroot,
@@ -820,7 +837,8 @@ class PCHGenerator : public SemaConsumer {
ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions,
bool AllowASTWithErrors = false, bool IncludeTimestamps = true,
bool BuildingImplicitModule = false,
- bool ShouldCacheASTInMemory = false);
+ bool ShouldCacheASTInMemory = false,
+ bool GeneratingThinBMI = false);
~PCHGenerator() override;
void InitializeSema(Sema &S) override { SemaPtr = &S; }
@@ -830,6 +848,19 @@ class PCHGenerator : public SemaConsumer {
bool hasEmittedPCH() const { return Buffer->IsComplete; }
};
+class ThinBMIGenerator : public PCHGenerator {
+public:
+ ThinBMIGenerator(const Preprocessor &PP, InMemoryModuleCache &ModuleCache,
+ StringRef OutputFile, std::shared_ptr<PCHBuffer> Buffer,
+ bool IncludeTimestamps);
+
+ void HandleTranslationUnit(ASTContext &Ctx) override;
+};
+
+/// If the definition may impact the ABI. If yes, we're allowed to eliminate
+/// the definition of D in thin BMI.
+bool MayDefAffectABI(const Decl *D);
+
/// A simple helper class to pack several bits in order into (a) 32 bit
/// integer(s).
class BitsPacker {
diff --git a/clang/lib/AST/ODRHash.cpp b/clang/lib/AST/ODRHash.cpp
index aea1a93ae1fa828..ace24eb4d29d852 100644
--- a/clang/lib/AST/ODRHash.cpp
+++ b/clang/lib/AST/ODRHash.cpp
@@ -1249,3 +1249,5 @@ void ODRHash::AddQualType(QualType T) {
void ODRHash::AddBoolean(bool Value) {
Bools.push_back(Value);
}
+
+void ODRHash::AddInteger(unsigned Value) { ID.AddInteger(Value); }
diff --git a/clang/lib/Driver/Driver.cpp b/clang/lib/Driver/Driver.cpp
index 6f5ff8141032677..42bd241f1ff70fe 100644
--- a/clang/lib/Driver/Driver.cpp
+++ b/clang/lib/Driver/Driver.cpp
@@ -354,6 +354,7 @@ phases::ID Driver::getFinalPhase(const DerivedArgList &DAL,
} else if ((PhaseArg = DAL.getLastArg(options::OPT_fsyntax_only)) ||
(PhaseArg = DAL.getLastArg(options::OPT_print_supported_cpus)) ||
(PhaseArg = DAL.getLastArg(options::OPT_module_file_info)) ||
+ (PhaseArg = DAL.getLastArg(options::OPT_get_bmi_decls_hash)) ||
(PhaseArg = DAL.getLastArg(options::OPT_verify_pch)) ||
(PhaseArg = DAL.getLastArg(options::OPT_rewrite_objc)) ||
(PhaseArg = DAL.getLastArg(options::OPT_rewrite_legacy_objc)) ||
@@ -4086,6 +4087,13 @@ void Driver::BuildActions(Compilation &C, DerivedArgList &Args,
}
}
+ // Diagnose misuse of -fthinBMI-output. It should be an error if we specify
+ // -fthinBMI-output with multiple precompilation jobs. Here we didn't check if
+ // there are multiple module units in the inputs.
+ if (C.getArgs().getLastArg(options::OPT_fthinBMI_output_EQ) &&
+ Inputs.size() > 1)
+ Diag(clang::diag::err_drv_thin_bmi_output_argument_with_multiple_files);
+
handleArguments(C, Args, Inputs, Actions);
bool UseNewOffloadingDriver =
@@ -4729,7 +4737,8 @@ Action *Driver::ConstructPhaseAction(
return C.MakeAction<MigrateJobAction>(Input, types::TY_Remap);
if (Args.hasArg(options::OPT_emit_ast))
return C.MakeAction<CompileJobAction>(Input, types::TY_AST);
- if (Args.hasArg(options::OPT_module_file_info))
+ if (Args.hasArg(options::OPT_module_file_info) ||
+ Args.hasArg(options::OPT_get_bmi_decls_hash))
return C.MakeAction<CompileJobAction>(Input, types::TY_ModuleFile);
if (Args.hasArg(options::OPT_verify_pch))
return C.MakeAction<VerifyPCHJobAction>(Input, types::TY_Nothing);
@@ -5811,7 +5820,8 @@ const char *Driver::GetNamedOutputPath(Compilation &C, const JobAction &JA,
}
if (JA.getType() == types::TY_ModuleFile &&
- C.getArgs().getLastArg(options::OPT_module_file_info)) {
+ (C.getArgs().getLastArg(options::OPT_module_file_info) ||
+ C.getArgs().getLastArg(options::OPT_get_bmi_decls_hash))) {
return "-";
}
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 22f992166ded6c0..c091a94ea0d7805 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -3940,6 +3940,8 @@ static bool RenderModulesOptions(Compilation &C, const Driver &D,
Args.ClaimAllArgs(options::OPT_fmodule_output);
Args.ClaimAllArgs(options::OPT_fmodule_output_EQ);
+ Args.AddLastArg(CmdArgs, options::OPT_fthinBMI_output_EQ);
+
return HaveModules;
}
@@ -4958,7 +4960,10 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
} else if (JA.getType() == types::TY_AST) {
CmdArgs.push_back("-emit-pch");
} else if (JA.getType() == types::TY_ModuleFile) {
- CmdArgs.push_back("-module-file-info");
+ if (Args.hasArg(options::OPT_get_bmi_decls_hash))
+ CmdArgs.push_back("-get-bmi-decls-hash");
+ else
+ CmdArgs.push_back("-module-file-info");
} else if (JA.getType() == types::TY_RewrittenObjC) {
CmdArgs.push_back("-rewrite-objc");
rewriteKind = RK_NonFragile;
diff --git a/clang/lib/Frontend/CompilerInvocation.cpp b/clang/lib/Frontend/CompilerInvocation.cpp
index 637c6a35af6532b..177b40a98578947 100644
--- a/clang/lib/Frontend/CompilerInvocation.cpp
+++ b/clang/lib/Frontend/CompilerInvocation.cpp
@@ -2554,12 +2554,14 @@ static const auto &getFrontendActionTable() {
{frontend::GenerateModule, OPT_emit_module},
{frontend::GenerateModuleInterface, OPT_emit_module_interface},
+ {frontend::GenerateThinModuleInterface, OPT_emit_thin_module_interface},
{frontend::GenerateHeaderUnit, OPT_emit_header_unit},
{frontend::GeneratePCH, OPT_emit_pch},
{frontend::GenerateInterfaceStubs, OPT_emit_interface_stubs},
{frontend::InitOnly, OPT_init_only},
{frontend::ParseSyntaxOnly, OPT_fsyntax_only},
{frontend::ModuleFileInfo, OPT_module_file_info},
+ {frontend::GetBMIDeclsHash, OPT_get_bmi_decls_hash},
{frontend::VerifyPCH, OPT_verify_pch},
{frontend::PrintPreamble, OPT_print_preamble},
{frontend::PrintPreprocessedInput, OPT_E},
@@ -4236,11 +4238,13 @@ static bool isStrictlyPreprocessorAction(frontend::ActionKind Action) {
case frontend::FixIt:
case frontend::GenerateModule:
case frontend::GenerateModuleInterface:
+ case frontend::GenerateThinModuleInterface:
case frontend::GenerateHeaderUnit:
case frontend::GeneratePCH:
case frontend::GenerateInterfaceStubs:
case frontend::ParseSyntaxOnly:
case frontend::ModuleFileInfo:
+ case frontend::GetBMIDeclsHash:
case frontend::VerifyPCH:
case frontend::PluginAction:
case frontend::RewriteObjC:
diff --git a/clang/lib/Frontend/FrontendActions.cpp b/clang/lib/Frontend/FrontendActions.cpp
index 2afcf1cf9f68c81..06b15079887c774 100644
--- a/clang/lib/Frontend/FrontendActions.cpp
+++ b/clang/lib/Frontend/FrontendActions.cpp
@@ -184,12 +184,12 @@ bool GeneratePCHAction::BeginSourceFileAction(CompilerInstance &CI) {
return true;
}
-std::unique_ptr<ASTConsumer>
-GenerateModuleAction::CreateASTConsumer(CompilerInstance &CI,
- StringRef InFile) {
+std::vector<std::unique_ptr<ASTConsumer...
[truncated]
|
llvmbot
added
clang
|
Let's discuss the higher level problems in Let's discuss the higher level problems in https://discourse.llvm.org/t/rfc-c-20-modules-introduce-thin-bmi-and-decls-hash/74755. Especially this one is not good to review since we don't support stacked review now. |
|
As I summarized in https://discourse.llvm.org/t/rfc-c-20-modules-introduce-thin-bmi-and-decls-hash/74755/43, people are interested in this direction, but it may be too early for us to implement it. Let's postpone this. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Close #71618
This contains #71622 and so that it is not easy to review this until we have stacked reviews. So the main purpose of current patch now is to have a feeling about what we want to do.
Motivating Examples
The motivating example is:
After we change the implementation of a() from return 43; to return 44;, we can avoid recompiling use.cc to use.o since the interface doesn't change.
This is a pretty appealing feature.
Interfaces
The interface introduced in this patch is:
-module-file-infoflag or-get-bmi-decls-hashflag.For example,
or
The difference is that
-module-file-infomay bring more information like compiler versions and compilation flags but-module-file-infomay be slower. And-get-bmi-decls-hashfocus on the decls hash only and it should be faster.Note that the value doesn't contain information like compiler versions and compilation flags, just as its name shows. The thought is that the 2 informations should be already known by the build systems already.
Usages
The decls hash allow the build systems to use a compare and swap strategy when generating BMI. For example, the build system can generate the BMI to a temporary place, (and if the compiler version and compilation flags doesn't change), the build system can try to query and compare the decls hash value for the generated BMI and the existing BMI. And only replace the existing BMI if their hash value differs.
Then we should be able to make the motivating examples.
Why not generating 2 same BMIs directly
People may feel the process is complicated. Why not generating the same BMI at the very beginning? The answer is that we can't. We generate too many things in the BMI and we just serialize the source file as is. The most simple example is, the generated BMI is different for:
and
Just because the function
a()has different source locations.The strategies of computing decls hash
Currently the strategy is simple, we don't count the definition of an non-inline non-dependent function or variables. And we don't count the non-exported non-inline non-dependent entities completely.
For example,
and
will have the same decls hash value.
We can find more examples in https://github.com/llvm/llvm-project/pull/71627/files#diff-c8483a51f357861802c5db18942ed799c5c8d10b4571c5a49274c2c800a12042
An interesting case may be inline functions. The ABI says the inline functions will be generated in the unit that calling them. So we choose to update the BMI hash every time we find the inline functions changes. But the strategy may be too conservative. For example,
and
These two units can have the same decls hash since the inline function
func()is not reachable to the consumers. But we choose to generate different decls hash for simpler implementation. Technically, it is possible to make it by implementing a context sensitive reachable analysis. But let's leave it to future.The key point here is our strategy to treat such cases. I think: