" to make use of cmpxchg flow-based information"),
cl::init(true));
+static cl::opt<bool>
+EnableARMLoadStoreOpt("arm-load-store-opt", cl::Hidden,
+ cl::desc("Enable ARM load/store optimization pass"),
+ cl::init(true));
+
+// FIXME: Unify control over GlobalMerge.
+static cl::opt<cl::boolOrDefault>
+EnableGlobalMerge("arm-global-merge", cl::Hidden,
+ cl::desc("Enable the global merge pass"));
+
extern "C" void LLVMInitializeARMTarget() {
// Register the target.
RegisterTargetMachine<ARMLETargetMachine> X(TheARMLETarget);
// FIXME: This is duplicated code from the front end and should be unified.
if (TT.isOSBinFormatMachO()) {
if (TT.getEnvironment() == llvm::Triple::EABI ||
- (TT.getOS() == llvm::Triple::UnknownOS &&
- TT.getObjectFormat() == llvm::Triple::MachO) ||
+ (TT.getOS() == llvm::Triple::UnknownOS && TT.isOSBinFormatMachO()) ||
CPU.startswith("cortex-m")) {
TargetABI = ARMBaseTargetMachine::ARM_ABI_AAPCS;
} else {
TargetABI = ARMBaseTargetMachine::ARM_ABI_APCS;
break;
default:
- if (TT.getOS() == llvm::Triple::NetBSD)
- TargetABI = ARMBaseTargetMachine::ARM_ABI_APCS;
+ if (TT.isOSNetBSD())
+ TargetABI = ARMBaseTargetMachine::ARM_ABI_APCS;
else
- TargetABI = ARMBaseTargetMachine::ARM_ABI_AAPCS;
+ TargetABI = ARMBaseTargetMachine::ARM_ABI_AAPCS;
break;
}
}
return TargetABI;
}
-static std::string computeDataLayout(const Triple &TT,
- ARMBaseTargetMachine::ARMABI ABI,
+static std::string computeDataLayout(const Triple &TT, StringRef CPU,
+ const TargetOptions &Options,
bool isLittle) {
+ auto ABI = computeTargetABI(TT, CPU, Options);
std::string Ret = "";
if (isLittle)
/// TargetMachine ctor - Create an ARM architecture model.
///
-ARMBaseTargetMachine::ARMBaseTargetMachine(const Target &T, StringRef TT,
+ARMBaseTargetMachine::ARMBaseTargetMachine(const Target &T, const Triple &TT,
StringRef CPU, StringRef FS,
const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OL, bool isLittle)
- : LLVMTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL),
- TargetABI(computeTargetABI(Triple(TT), CPU, Options)),
- DL(computeDataLayout(Triple(TT), TargetABI, isLittle)),
- TLOF(createTLOF(Triple(getTargetTriple()))),
+ : LLVMTargetMachine(T, computeDataLayout(TT, CPU, Options, isLittle), TT,
+ CPU, FS, Options, RM, CM, OL),
+ TargetABI(computeTargetABI(TT, CPU, Options)),
+ TLOF(createTLOF(getTargetTriple())),
Subtarget(TT, CPU, FS, *this, isLittle), isLittle(isLittle) {
// Default to triple-appropriate float ABI
// function before we can generate a subtarget. We also need to use
// it as a key for the subtarget since that can be the only difference
// between two functions.
- Attribute SFAttr = F.getFnAttribute("use-soft-float");
- bool SoftFloat = !SFAttr.hasAttribute(Attribute::None)
- ? SFAttr.getValueAsString() == "true"
- : Options.UseSoftFloat;
-
- auto &I = SubtargetMap[CPU + FS + (SoftFloat ? "use-soft-float=true"
- : "use-soft-float=false")];
+ bool SoftFloat =
+ F.hasFnAttribute("use-soft-float") &&
+ F.getFnAttribute("use-soft-float").getValueAsString() == "true";
+ // If the soft float attribute is set on the function turn on the soft float
+ // subtarget feature.
+ if (SoftFloat)
+ FS += FS.empty() ? "+soft-float" : ",+soft-float";
+
+ auto &I = SubtargetMap[CPU + FS];
if (!I) {
// This needs to be done before we create a new subtarget since any
// creation will depend on the TM and the code generation flags on the
}
TargetIRAnalysis ARMBaseTargetMachine::getTargetIRAnalysis() {
- return TargetIRAnalysis(
- [this](Function &F) { return TargetTransformInfo(ARMTTIImpl(this, F)); });
+ return TargetIRAnalysis([this](const Function &F) {
+ return TargetTransformInfo(ARMTTIImpl(this, F));
+ });
}
+void ARMTargetMachine::anchor() {}
-void ARMTargetMachine::anchor() { }
-
-ARMTargetMachine::ARMTargetMachine(const Target &T, StringRef TT, StringRef CPU,
- StringRef FS, const TargetOptions &Options,
+ARMTargetMachine::ARMTargetMachine(const Target &T, const Triple &TT,
+ StringRef CPU, StringRef FS,
+ const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OL, bool isLittle)
: ARMBaseTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL, isLittle) {
"support ARM mode execution!");
}
-void ARMLETargetMachine::anchor() { }
+void ARMLETargetMachine::anchor() {}
-ARMLETargetMachine::ARMLETargetMachine(const Target &T, StringRef TT,
+ARMLETargetMachine::ARMLETargetMachine(const Target &T, const Triple &TT,
StringRef CPU, StringRef FS,
const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OL)
: ARMTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL, true) {}
-void ARMBETargetMachine::anchor() { }
+void ARMBETargetMachine::anchor() {}
-ARMBETargetMachine::ARMBETargetMachine(const Target &T, StringRef TT,
+ARMBETargetMachine::ARMBETargetMachine(const Target &T, const Triple &TT,
StringRef CPU, StringRef FS,
const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OL)
: ARMTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL, false) {}
-void ThumbTargetMachine::anchor() { }
+void ThumbTargetMachine::anchor() {}
-ThumbTargetMachine::ThumbTargetMachine(const Target &T, StringRef TT,
+ThumbTargetMachine::ThumbTargetMachine(const Target &T, const Triple &TT,
StringRef CPU, StringRef FS,
const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OL, bool isLittle)
- : ARMBaseTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL,
- isLittle) {
+ : ARMBaseTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL, isLittle) {
initAsmInfo();
}
-void ThumbLETargetMachine::anchor() { }
+void ThumbLETargetMachine::anchor() {}
-ThumbLETargetMachine::ThumbLETargetMachine(const Target &T, StringRef TT,
+ThumbLETargetMachine::ThumbLETargetMachine(const Target &T, const Triple &TT,
StringRef CPU, StringRef FS,
const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OL)
: ThumbTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL, true) {}
-void ThumbBETargetMachine::anchor() { }
+void ThumbBETargetMachine::anchor() {}
-ThumbBETargetMachine::ThumbBETargetMachine(const Target &T, StringRef TT,
+ThumbBETargetMachine::ThumbBETargetMachine(const Target &T, const Triple &TT,
StringRef CPU, StringRef FS,
const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
return getTM<ARMBaseTargetMachine>();
}
- const ARMSubtarget &getARMSubtarget() const {
- return *getARMTargetMachine().getSubtargetImpl();
- }
-
void addIRPasses() override;
bool addPreISel() override;
bool addInstSelector() override;
// Cmpxchg instructions are often used with a subsequent comparison to
// determine whether it succeeded. We can exploit existing control-flow in
// ldrex/strex loops to simplify this, but it needs tidying up.
- const ARMSubtarget *Subtarget = &getARMSubtarget();
- if (Subtarget->hasAnyDataBarrier() && !Subtarget->isThumb1Only())
- if (TM->getOptLevel() != CodeGenOpt::None && EnableAtomicTidy)
- addPass(createCFGSimplificationPass());
+ if (TM->getOptLevel() != CodeGenOpt::None && EnableAtomicTidy)
+ addPass(createCFGSimplificationPass(-1, [this](const Function &F) {
+ const auto &ST = this->TM->getSubtarget<ARMSubtarget>(F);
+ return ST.hasAnyDataBarrier() && !ST.isThumb1Only();
+ }));
TargetPassConfig::addIRPasses();
+
+ // Match interleaved memory accesses to ldN/stN intrinsics.
+ if (TM->getOptLevel() != CodeGenOpt::None)
+ addPass(createInterleavedAccessPass(TM));
}
bool ARMPassConfig::addPreISel() {
- if (TM->getOptLevel() != CodeGenOpt::None)
- addPass(createGlobalMergePass(TM));
+ if ((TM->getOptLevel() != CodeGenOpt::None &&
+ EnableGlobalMerge == cl::BOU_UNSET) ||
+ EnableGlobalMerge == cl::BOU_TRUE) {
+ // FIXME: This is using the thumb1 only constant value for
+ // maximal global offset for merging globals. We may want
+ // to look into using the old value for non-thumb1 code of
+ // 4095 based on the TargetMachine, but this starts to become
+ // tricky when doing code gen per function.
+ bool OnlyOptimizeForSize = (TM->getOptLevel() < CodeGenOpt::Aggressive) &&
+ (EnableGlobalMerge == cl::BOU_UNSET);
+ // Merging of extern globals is enabled by default on non-Mach-O as we
+ // expect it to be generally either beneficial or harmless. On Mach-O it
+ // is disabled as we emit the .subsections_via_symbols directive which
+ // means that merging extern globals is not safe.
+ bool MergeExternalByDefault = !TM->getTargetTriple().isOSBinFormatMachO();
+ addPass(createGlobalMergePass(TM, 127, OnlyOptimizeForSize,
+ MergeExternalByDefault));
+ }
return false;
}
bool ARMPassConfig::addInstSelector() {
addPass(createARMISelDag(getARMTargetMachine(), getOptLevel()));
- const ARMSubtarget *Subtarget = &getARMSubtarget();
- if (Subtarget->isTargetELF() && !Subtarget->isThumb1Only() &&
- TM->Options.EnableFastISel)
+ if (TM->getTargetTriple().isOSBinFormatELF() && TM->Options.EnableFastISel)
addPass(createARMGlobalBaseRegPass());
return false;
}
void ARMPassConfig::addPreRegAlloc() {
- if (getOptLevel() != CodeGenOpt::None)
- addPass(createARMLoadStoreOptimizationPass(true));
- if (getOptLevel() != CodeGenOpt::None && getARMSubtarget().isCortexA9())
+ if (getOptLevel() != CodeGenOpt::None) {
addPass(createMLxExpansionPass());
- // Since the A15SDOptimizer pass can insert VDUP instructions, it can only be
- // enabled when NEON is available.
- if (getOptLevel() != CodeGenOpt::None && getARMSubtarget().isCortexA15() &&
- getARMSubtarget().hasNEON() && !DisableA15SDOptimization) {
- addPass(createA15SDOptimizerPass());
+
+ if (EnableARMLoadStoreOpt)
+ addPass(createARMLoadStoreOptimizationPass(/* pre-register alloc */ true));
+
+ if (!DisableA15SDOptimization)
+ addPass(createA15SDOptimizerPass());
}
}
void ARMPassConfig::addPreSched2() {
if (getOptLevel() != CodeGenOpt::None) {
- addPass(createARMLoadStoreOptimizationPass());
+ if (EnableARMLoadStoreOpt)
+ addPass(createARMLoadStoreOptimizationPass());
- if (getARMSubtarget().hasNEON())
- addPass(createExecutionDependencyFixPass(&ARM::DPRRegClass));
+ addPass(createExecutionDependencyFixPass(&ARM::DPRRegClass));
}
// Expand some pseudo instructions into multiple instructions to allow
addPass(createARMExpandPseudoPass());
if (getOptLevel() != CodeGenOpt::None) {
- if (!getARMSubtarget().isThumb1Only()) {
- // in v8, IfConversion depends on Thumb instruction widths
- if (getARMSubtarget().restrictIT() &&
- !getARMSubtarget().prefers32BitThumb())
- addPass(createThumb2SizeReductionPass());
- addPass(&IfConverterID);
- }
+ // in v8, IfConversion depends on Thumb instruction widths
+ addPass(createThumb2SizeReductionPass([this](const Function &F) {
+ return this->TM->getSubtarget<ARMSubtarget>(F).restrictIT();
+ }));
+
+ addPass(createIfConverter([this](const Function &F) {
+ return !this->TM->getSubtarget<ARMSubtarget>(F).isThumb1Only();
+ }));
}
- if (getARMSubtarget().isThumb2())
- addPass(createThumb2ITBlockPass());
+ addPass(createThumb2ITBlockPass());
}
void ARMPassConfig::addPreEmitPass() {
- if (getARMSubtarget().isThumb2()) {
- if (!getARMSubtarget().prefers32BitThumb())
- addPass(createThumb2SizeReductionPass());
+ addPass(createThumb2SizeReductionPass());
- // Constant island pass work on unbundled instructions.
- addPass(&UnpackMachineBundlesID);
- }
+ // Constant island pass work on unbundled instructions.
+ addPass(createUnpackMachineBundles([this](const Function &F) {
+ return this->TM->getSubtarget<ARMSubtarget>(F).isThumb2();
+ }));
+
+ // Don't optimize barriers at -O0.
+ if (getOptLevel() != CodeGenOpt::None)
+ addPass(createARMOptimizeBarriersPass());
- addPass(createARMOptimizeBarriersPass());
addPass(createARMConstantIslandPass());
}