pete created this revision.
pete added reviewers: ahatanak, rjmccall, erik.pilkington.
Herald added a subscriber: cfe-commits.
It is faster to directly call the ObjC runtime for methods such as
alloc/allocWithZone instead of sending a message to those functions.
This patch adds support for converting messages to alloc/allocWithZone to their
equivalent runtime calls.
Tests included for the positive case of applying this transformation, negative
tests that we ensure we only convert "alloc" to objc_alloc, not "alloc2", and
also a driver test to ensure we enable this only for supported runtime versions.
Repository:
rC Clang
https://reviews.llvm.org/D55349
Files:
include/clang/Basic/ObjCRuntime.h
include/clang/Driver/Options.td
include/clang/Frontend/CodeGenOptions.def
lib/CodeGen/CGObjC.cpp
lib/CodeGen/CodeGenFunction.h
lib/CodeGen/CodeGenModule.h
lib/Driver/ToolChains/Clang.cpp
lib/Frontend/CompilerInvocation.cpp
test/CodeGenObjC/convert-messages-to-runtime-calls.m
test/Driver/objc-convert-messages-to-runtime-calls.m
Index: test/Driver/objc-convert-messages-to-runtime-calls.m
===================================================================
--- /dev/null
+++ test/Driver/objc-convert-messages-to-runtime-calls.m
@@ -0,0 +1,7 @@
+// RUN: %clang %s -### -o %t.o 2>&1 -fsyntax-only -fobjc-convert-messages-to-runtime-calls -fno-objc-convert-messages-to-runtime-calls -target x86_64-apple-macosx10.10.0 | FileCheck %s --check-prefix=DISABLE
+// RUN: %clang %s -### -o %t.o 2>&1 -fsyntax-only -fno-objc-convert-messages-to-runtime-calls -fobjc-convert-messages-to-runtime-calls -target x86_64-apple-macosx10.10.0 | FileCheck %s --check-prefix=ENABLE
+
+// Check that we pass fobjc-convert-messages-to-runtime-calls only when supported, and not explicitly disabled.
+
+// DISABLE: "-fno-objc-convert-messages-to-runtime-calls"
+// ENABLE-NOT: "-fno-objc-convert-messages-to-runtime-calls"
Index: test/CodeGenObjC/convert-messages-to-runtime-calls.m
===================================================================
--- /dev/null
+++ test/CodeGenObjC/convert-messages-to-runtime-calls.m
@@ -0,0 +1,66 @@
+// RUN: %clang_cc1 -fobjc-runtime=macosx-10.10.0 -emit-llvm -o - %s -fno-objc-convert-messages-to-runtime-calls | FileCheck %s --check-prefix=MSGS
+// RUN: %clang_cc1 -fobjc-runtime=macosx-10.10.0 -emit-llvm -o - %s | FileCheck %s --check-prefix=CALLS
+// RUN: %clang_cc1 -fobjc-runtime=macosx-10.9.0 -emit-llvm -o - %s | FileCheck %s --check-prefix=MSGS
+// RUN: %clang_cc1 -fobjc-runtime=macosx-fragile-10.10.0 -emit-llvm -o - %s | FileCheck %s --check-prefix=MSGS
+// RUN: %clang_cc1 -fobjc-runtime=ios-8.0 -emit-llvm -o - %s | FileCheck %s --check-prefix=CALLS
+// RUN: %clang_cc1 -fobjc-runtime=ios-7.0 -emit-llvm -o - %s | FileCheck %s --check-prefix=MSGS
+// Note: This line below is for tvos for which the driver passes through to use the ios9.0 runtime.
+// RUN: %clang_cc1 -fobjc-runtime=ios-9.0 -emit-llvm -o - %s | FileCheck %s --check-prefix=CALLS
+// RUN: %clang_cc1 -fobjc-runtime=watchos-2.0 -emit-llvm -o - %s | FileCheck %s --check-prefix=CALLS
+
+#define nil (id)0
+
+@interface NSObject
++ (id)alloc;
++ (id)allocWithZone:(void*)zone;
++ (id)alloc2;
+@end
+
+// CHECK-LABEL: define {{.*}}void @test1
+void test1(id x) {
+ // MSGS: {{call.*@objc_msgSend}}
+ // CALLS: {{call.*@objc_alloc}}
+ [NSObject alloc];
+ [NSObject allocWithZone:nil];
+}
+
+// CHECK-LABEL: define {{.*}}void @test2
+void test2(void* x) {
+ // MSGS: {{call.*@objc_msgSend}}
+ // MSGS: {{call.*@objc_msgSend}}
+ // MSGS: {{call.*@objc_msgSend}}
+ // CALLS: {{call.*@objc_msgSend}}
+ // CALLS: {{call.*@objc_msgSend}}
+ // CALLS: {{call.*@objc_msgSend}}
+ [NSObject alloc2];
+ [NSObject allocWithZone:(void*)-1];
+ [NSObject allocWithZone:x];
+}
+
+@class A;
+@interface B
++ (A*) alloc;
++ (A*)allocWithZone:(void*)zone;
+@end
+
+// Make sure we get a bitcast on the return type as the
+// call will return i8* which we have to cast to A*
+// CHECK-LABEL: define {{.*}}void @test_alloc_class_ptr
+A* test_alloc_class_ptr() {
+ // CALLS: {{call.*@objc_alloc}}
+ // CALLS-NEXT: bitcast i8*
+ // CALLS-NEXT: ret
+ return [B alloc];
+}
+
+// Make sure we get a bitcast on the return type as the
+// call will return i8* which we have to cast to A*
+// CHECK-LABEL: define {{.*}}void @test_alloc_class_ptr
+A* test_allocWithZone_class_ptr() {
+ // CALLS: {{call.*@objc_alloc}}
+ // CALLS-NEXT: bitcast i8*
+ // CALLS-NEXT: ret
+ return [B allocWithZone:nil];
+}
+
+
Index: lib/Frontend/CompilerInvocation.cpp
===================================================================
--- lib/Frontend/CompilerInvocation.cpp
+++ lib/Frontend/CompilerInvocation.cpp
@@ -1155,6 +1155,10 @@
}
}
+
+ if (Args.hasArg(OPT_fno_objc_convert_messages_to_runtime_calls))
+ Opts.ObjCConvertMessagesToRuntimeCalls = 0;
+
if (Args.getLastArg(OPT_femulated_tls) ||
Args.getLastArg(OPT_fno_emulated_tls)) {
Opts.ExplicitEmulatedTLS = true;
Index: lib/Driver/ToolChains/Clang.cpp
===================================================================
--- lib/Driver/ToolChains/Clang.cpp
+++ lib/Driver/ToolChains/Clang.cpp
@@ -2864,6 +2864,18 @@
Args.ClaimAllArgs(options::OPT_fno_objc_arc_exceptions);
}
+ // Allow the user to control whether messages can be converted to runtime
+ // functions.
+ if (types::isObjC(Input.getType())) {
+ auto *Arg = Args.getLastArg(
+ options::OPT_fobjc_convert_messages_to_runtime_calls,
+ options::OPT_fno_objc_convert_messages_to_runtime_calls);
+ if (Arg &&
+ Arg->getOption().matches(
+ options::OPT_fno_objc_convert_messages_to_runtime_calls))
+ CmdArgs.push_back("-fno-objc-convert-messages-to-runtime-calls");
+ }
+
// -fobjc-infer-related-result-type is the default, except in the Objective-C
// rewriter.
if (InferCovariantReturns)
Index: lib/CodeGen/CodeGenModule.h
===================================================================
--- lib/CodeGen/CodeGenModule.h
+++ lib/CodeGen/CodeGenModule.h
@@ -119,7 +119,13 @@
struct ObjCEntrypoints {
ObjCEntrypoints() { memset(this, 0, sizeof(*this)); }
- /// void objc_autoreleasePoolPop(void*);
+ /// void objc_alloc(id);
+ llvm::Constant *objc_alloc;
+
+ /// void objc_allocWithZone(id);
+ llvm::Constant *objc_allocWithZone;
+
+ /// void objc_autoreleasePoolPop(void*);
llvm::Constant *objc_autoreleasePoolPop;
/// void *objc_autoreleasePoolPush(void);
Index: lib/CodeGen/CodeGenFunction.h
===================================================================
--- lib/CodeGen/CodeGenFunction.h
+++ lib/CodeGen/CodeGenFunction.h
@@ -3805,6 +3805,10 @@
std::pair<LValue,llvm::Value*>
EmitARCStoreUnsafeUnretained(const BinaryOperator *e, bool ignored);
+ llvm::Value *EmitObjCAlloc(llvm::Value *value,
+ llvm::Type *returnType);
+ llvm::Value *EmitObjCAllocWithZone(llvm::Value *value,
+ llvm::Type *returnType);
llvm::Value *EmitObjCThrowOperand(const Expr *expr);
llvm::Value *EmitObjCConsumeObject(QualType T, llvm::Value *Ptr);
llvm::Value *EmitObjCExtendObjectLifetime(QualType T, llvm::Value *Ptr);
Index: lib/CodeGen/CGObjC.cpp
===================================================================
--- lib/CodeGen/CGObjC.cpp
+++ lib/CodeGen/CGObjC.cpp
@@ -352,6 +352,58 @@
return nullptr;
}
+/// The ObjC runtime may provide entrypoints that are likely to be faster
+/// than an ordinary message send of the appropriate selector.
+///
+/// The entrypoints are guaranteed to be equivalent to just sending the
+/// corresponding message. If the entrypoint is implemented naively as just a
+/// message send, using it is a trade-off: it sacrifices a few cycles of
+/// overhead to save a small amount of code. However, it's possible for
+/// runtimes to detect and special-case classes that use "standard"
+/// behavior; if that's dynamically a large proportion of all objects, using
+/// the entrypoint will also be faster than using a message send.
+///
+/// If the runtime does support a required entrypoint, then this method will
+/// generate a call and return the resulting value. Otherwise it will return
+/// None and the caller can generate a msgSend instead.
+static Optional<llvm::Value *>
+tryGenerateSpecializedMessageSend(CodeGenFunction &CGF, QualType ResultType,
+ llvm::Value *Receiver,
+ const CallArgList& Args, Selector Sel,
+ const ObjCMethodDecl *method) {
+ auto &CGM = CGF.CGM;
+ if (!CGM.getCodeGenOpts().ObjCConvertMessagesToRuntimeCalls)
+ return None;
+
+ auto &Runtime = CGM.getLangOpts().ObjCRuntime;
+ switch (Sel.getMethodFamily()) {
+ case OMF_alloc:
+ // Make sure the name is exactly 'alloc'. All methods with that
+ // prefix are identified as OMF_alloc but we only want to call the
+ // runtime for this version.
+ if (Runtime.shouldUseRuntimeFunctionsForAlloc() &&
+ ResultType->isObjCObjectPointerType()) {
+ if (Sel.isUnarySelector() && Sel.getNameForSlot(0) == "alloc")
+ return CGF.EmitObjCAlloc(Receiver, CGF.ConvertType(ResultType));
+ // [Foo allocWithZone:nil] -> objc_allocWithZone(Foo)
+ if (Sel.isKeywordSelector() && Sel.getNumArgs() == 1 &&
+ Args.size() == 1 && Args.front().getKnownRValue().isScalar() &&
+ Sel.getNameForSlot(0) == "allocWithZone") {
+ const llvm::Value* arg = Args.front().getKnownRValue().getScalarVal();
+ if (isa<llvm::ConstantPointerNull>(arg))
+ return CGF.EmitObjCAllocWithZone(Receiver,
+ CGF.ConvertType(ResultType));
+ return None;
+ }
+ }
+ break;
+
+ default:
+ break;
+ }
+ return None;
+}
+
RValue CodeGenFunction::EmitObjCMessageExpr(const ObjCMessageExpr *E,
ReturnValueSlot Return) {
// Only the lookup mechanism and first two arguments of the method
@@ -474,10 +526,16 @@
Args,
method);
} else {
- result = Runtime.GenerateMessageSend(*this, Return, ResultType,
- E->getSelector(),
- Receiver, Args, OID,
- method);
+ // Call runtime methods directly if we can.
+ if (Optional<llvm::Value *> SpecializedResult =
+ tryGenerateSpecializedMessageSend(*this, ResultType, Receiver, Args,
+ E->getSelector(), method)) {
+ result = RValue::get(SpecializedResult.getValue());
+ } else {
+ result = Runtime.GenerateMessageSend(*this, Return, ResultType,
+ E->getSelector(), Receiver, Args,
+ OID, method);
+ }
}
// For delegate init calls in ARC, implicitly store the result of
@@ -1845,6 +1903,7 @@
/// where a null input causes a no-op and returns null.
static llvm::Value *emitARCValueOperation(CodeGenFunction &CGF,
llvm::Value *value,
+ llvm::Type *returnType,
llvm::Constant *&fn,
StringRef fnName,
bool isTailCall = false) {
@@ -1858,7 +1917,7 @@
}
// Cast the argument to 'id'.
- llvm::Type *origType = value->getType();
+ llvm::Type *origType = returnType ? returnType : value->getType();
value = CGF.Builder.CreateBitCast(value, CGF.Int8PtrTy);
// Call the function.
@@ -1964,7 +2023,7 @@
/// Retain the given object, with normal retain semantics.
/// call i8* \@objc_retain(i8* %value)
llvm::Value *CodeGenFunction::EmitARCRetainNonBlock(llvm::Value *value) {
- return emitARCValueOperation(*this, value,
+ return emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_retain,
"objc_retain");
}
@@ -1978,7 +2037,7 @@
llvm::Value *CodeGenFunction::EmitARCRetainBlock(llvm::Value *value,
bool mandatory) {
llvm::Value *result
- = emitARCValueOperation(*this, value,
+ = emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_retainBlock,
"objc_retainBlock");
@@ -2048,7 +2107,7 @@
llvm::Value *
CodeGenFunction::EmitARCRetainAutoreleasedReturnValue(llvm::Value *value) {
emitAutoreleasedReturnValueMarker(*this);
- return emitARCValueOperation(*this, value,
+ return emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_retainAutoreleasedReturnValue,
"objc_retainAutoreleasedReturnValue");
}
@@ -2063,7 +2122,7 @@
llvm::Value *
CodeGenFunction::EmitARCUnsafeClaimAutoreleasedReturnValue(llvm::Value *value) {
emitAutoreleasedReturnValueMarker(*this);
- return emitARCValueOperation(*this, value,
+ return emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_unsafeClaimAutoreleasedReturnValue,
"objc_unsafeClaimAutoreleasedReturnValue");
}
@@ -2178,7 +2237,7 @@
/// Autorelease the given object.
/// call i8* \@objc_autorelease(i8* %value)
llvm::Value *CodeGenFunction::EmitARCAutorelease(llvm::Value *value) {
- return emitARCValueOperation(*this, value,
+ return emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_autorelease,
"objc_autorelease");
}
@@ -2187,7 +2246,7 @@
/// call i8* \@objc_autoreleaseReturnValue(i8* %value)
llvm::Value *
CodeGenFunction::EmitARCAutoreleaseReturnValue(llvm::Value *value) {
- return emitARCValueOperation(*this, value,
+ return emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_autoreleaseReturnValue,
"objc_autoreleaseReturnValue",
/*isTailCall*/ true);
@@ -2197,7 +2256,7 @@
/// call i8* \@objc_retainAutoreleaseReturnValue(i8* %value)
llvm::Value *
CodeGenFunction::EmitARCRetainAutoreleaseReturnValue(llvm::Value *value) {
- return emitARCValueOperation(*this, value,
+ return emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_retainAutoreleaseReturnValue,
"objc_retainAutoreleaseReturnValue",
/*isTailCall*/ true);
@@ -2226,7 +2285,7 @@
/// call i8* \@objc_retainAutorelease(i8* %value)
llvm::Value *
CodeGenFunction::EmitARCRetainAutoreleaseNonBlock(llvm::Value *value) {
- return emitARCValueOperation(*this, value,
+ return emitARCValueOperation(*this, value, nullptr,
CGM.getObjCEntrypoints().objc_retainAutorelease,
"objc_retainAutorelease");
}
@@ -2385,6 +2444,24 @@
return InitRV.getScalarVal();
}
+/// Allocate the given objc object.
+/// call i8* \@objc_alloc(i8* %value)
+llvm::Value *CodeGenFunction::EmitObjCAlloc(llvm::Value *value,
+ llvm::Type *resultType) {
+ return emitARCValueOperation(*this, value, resultType,
+ CGM.getObjCEntrypoints().objc_alloc,
+ "objc_alloc");
+}
+
+/// Allocate the given objc object.
+/// call i8* \@objc_allocWithZone(i8* %value)
+llvm::Value *CodeGenFunction::EmitObjCAllocWithZone(llvm::Value *value,
+ llvm::Type *resultType) {
+ return emitARCValueOperation(*this, value, resultType,
+ CGM.getObjCEntrypoints().objc_allocWithZone,
+ "objc_allocWithZone");
+}
+
/// Produce the code to do a primitive release.
/// [tmp drain];
void CodeGenFunction::EmitObjCMRRAutoreleasePoolPop(llvm::Value *Arg) {
Index: include/clang/Frontend/CodeGenOptions.def
===================================================================
--- include/clang/Frontend/CodeGenOptions.def
+++ include/clang/Frontend/CodeGenOptions.def
@@ -149,6 +149,8 @@
CODEGENOPT(NoZeroInitializedInBSS , 1, 0) ///< -fno-zero-initialized-in-bss.
/// Method of Objective-C dispatch to use.
ENUM_CODEGENOPT(ObjCDispatchMethod, ObjCDispatchMethodKind, 2, Legacy)
+/// Replace certain message sends with calls to ObjC runtime entrypoints
+CODEGENOPT(ObjCConvertMessagesToRuntimeCalls , 1, 1)
CODEGENOPT(OmitLeafFramePointer , 1, 0) ///< Set when -momit-leaf-frame-pointer is
///< enabled.
Index: include/clang/Driver/Options.td
===================================================================
--- include/clang/Driver/Options.td
+++ include/clang/Driver/Options.td
@@ -1473,6 +1473,10 @@
def fobjc_arc : Flag<["-"], "fobjc-arc">, Group<f_Group>, Flags<[CC1Option]>,
HelpText<"Synthesize retain and release calls for Objective-C pointers">;
def fno_objc_arc : Flag<["-"], "fno-objc-arc">, Group<f_Group>;
+def fobjc_convert_messages_to_runtime_calls :
+ Flag<["-"], "fobjc-convert-messages-to-runtime-calls">, Group<f_Group>;
+def fno_objc_convert_messages_to_runtime_calls :
+ Flag<["-"], "fno-objc-convert-messages-to-runtime-calls">, Group<f_Group>, Flags<[CC1Option]>;
def fobjc_arc_exceptions : Flag<["-"], "fobjc-arc-exceptions">, Group<f_Group>, Flags<[CC1Option]>,
HelpText<"Use EH-safe code when synthesizing retains and releases in -fobjc-arc">;
def fno_objc_arc_exceptions : Flag<["-"], "fno-objc-arc-exceptions">, Group<f_Group>;
Index: include/clang/Basic/ObjCRuntime.h
===================================================================
--- include/clang/Basic/ObjCRuntime.h
+++ include/clang/Basic/ObjCRuntime.h
@@ -173,6 +173,42 @@
llvm_unreachable("bad kind");
}
+ /// Does this runtime provide entrypoints that are likely to be faster
+ /// than an ordinary message send of the "alloc" selector?
+ ///
+ /// The "alloc" entrypoint is guaranteed to be equivalent to just sending the
+ /// corresponding message. If the entrypoint is implemented naively as just a
+ /// message send, using it is a trade-off: it sacrifices a few cycles of
+ /// overhead to save a small amount of code. However, it's possible for
+ /// runtimes to detect and special-case classes that use "standard"
+ /// alloc behavior; if that's dynamically a large proportion of all
+ /// objects, using the entrypoint will also be faster than using a message
+ /// send.
+ ///
+ /// When this method returns true, Clang will turn non-super message sends of
+ /// certain selectors into calls to the correspond entrypoint:
+ /// alloc => objc_alloc
+ bool shouldUseRuntimeFunctionsForAlloc() const {
+ switch (getKind()) {
+ case FragileMacOSX:
+ return false;
+ case MacOSX:
+ return getVersion() >= VersionTuple(10, 10);
+ case iOS:
+ return getVersion() >= VersionTuple(8);
+ case WatchOS:
+ return true;
+
+ case GCC:
+ return false;
+ case GNUstep:
+ return false;
+ case ObjFW:
+ return false;
+ }
+ llvm_unreachable("bad kind");
+ }
+
/// Does this runtime supports optimized setter entrypoints?
bool hasOptimizedSetter() const {
switch (getKind()) {
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
