Options for controlling the target. More...

#include "clang/Basic/TargetOptions.h"

Public Types
enum	CodeObjectVersionKind { COV_None , COV_2 = 200 , COV_3 = 300 , COV_4 = 400 , COV_5 = 500 }
	Enumeration value for AMDGPU code object version, which is the code object version times 100. More...

enum class	AMDGPUPrintfKind { Hostcall = 0 , Buffered = 1 }
	Enumeration values for AMDGPU printf lowering scheme. More...

Public Attributes
std::string	Triple
	The name of the target triple to compile for.

std::string	HostTriple
	When compiling for the device side, contains the triple used to compile for the host.

std::string	CPU
	If given, the name of the target CPU to generate code for.

std::string	TuneCPU
	If given, the name of the target CPU to tune code for.

std::string	FPMath
	If given, the unit to use for floating point math.

std::string	ABI
	If given, the name of the target ABI to use.

llvm::EABI	EABIVersion = llvm::EABI::Default
	The EABI version to use.

std::string	LinkerVersion
	If given, the version string of the linker in use.

std::vector< std::string >	FeaturesAsWritten
	The list of target specific features to enable or disable, as written on the command line.

std::vector< std::string >	Features
	The list of target specific features to enable or disable – this should be a list of strings starting with by '+' or '-'.

llvm::StringMap< bool >	FeatureMap
	The map of which features have been enabled disabled based on the command line.

llvm::StringMap< bool >	OpenCLFeaturesMap
	Supported OpenCL extensions and optional core features.

std::vector< std::string >	OpenCLExtensionsAsWritten
	The list of OpenCL extensions to enable or disable, as written on the command line.

bool	ForceEnableInt128 = false
	If given, enables support for __int128_t and __uint128_t types.

bool	NVPTXUseShortPointers = false
	If enabled, use 32-bit pointers for accessing const/local/shared address space.

bool	AllowAMDGPUUnsafeFPAtomics = false
	If enabled, allow AMDGPU unsafe floating point atomics.

CodeObjectVersionKind	CodeObjectVersion = CodeObjectVersionKind::COV_None
	Code object version for AMDGPU.

AMDGPUPrintfKind	AMDGPUPrintfKindVal = AMDGPUPrintfKind::Hostcall
	AMDGPU Printf lowering scheme.

std::string	CodeModel

uint64_t	LargeDataThreshold

llvm::VersionTuple	SDKVersion
	The version of the SDK which was used during the compilation.

std::string	DarwinTargetVariantTriple
	The name of the darwin target- ariant triple to compile for.

llvm::VersionTuple	DarwinTargetVariantSDKVersion
	The version of the darwin target variant SDK which was used during the compilation.

std::string	DxilValidatorVersion
	The validator version for dxil.

std::string	HLSLEntry
	The entry point name for HLSL shader being compiled as specified by -E.

Detailed Description

Options for controlling the target.

Definition at line 26 of file TargetOptions.h.

Member Enumeration Documentation

◆ AMDGPUPrintfKind

enum class clang::TargetOptions::AMDGPUPrintfKind

strong

Enumeration values for AMDGPU printf lowering scheme.

Enumerator
Hostcall	printf lowering scheme involving hostcalls, currently used by HIP programs by default
Buffered	printf lowering scheme involving implicit printf buffers,

Definition at line 94 of file TargetOptions.h.

◆ CodeObjectVersionKind

enum clang::TargetOptions::CodeObjectVersionKind

Enumeration value for AMDGPU code object version, which is the code object version times 100.

Enumerator
COV_None
COV_2
COV_3
COV_4
COV_5

Definition at line 83 of file TargetOptions.h.

Member Data Documentation

◆ ABI

std::string clang::TargetOptions::ABI

If given, the name of the target ABI to use.

Definition at line 45 of file TargetOptions.h.

Referenced by initTargetOptions().

◆ AllowAMDGPUUnsafeFPAtomics

bool clang::TargetOptions::AllowAMDGPUUnsafeFPAtomics = false

If enabled, allow AMDGPU unsafe floating point atomics.

Definition at line 79 of file TargetOptions.h.

Referenced by clang::targets::AMDGPUTargetInfo::AMDGPUTargetInfo().

◆ AMDGPUPrintfKindVal

AMDGPUPrintfKind clang::TargetOptions::AMDGPUPrintfKindVal = AMDGPUPrintfKind::Hostcall

AMDGPU Printf lowering scheme.

Definition at line 104 of file TargetOptions.h.

◆ CodeModel

std::string clang::TargetOptions::CodeModel

Definition at line 110 of file TargetOptions.h.

Referenced by FixupInvocation(), clang::targets::AArch64TargetInfo::getTargetDefines(), clang::targets::RISCVTargetInfo::getTargetDefines(), and clang::targets::X86TargetInfo::getTargetDefines().

◆ CodeObjectVersion

CodeObjectVersionKind clang::TargetOptions::CodeObjectVersion = CodeObjectVersionKind::COV_None

Code object version for AMDGPU.

Definition at line 91 of file TargetOptions.h.

◆ CPU

std::string clang::TargetOptions::CPU

If given, the name of the target CPU to generate code for.

Definition at line 36 of file TargetOptions.h.

Referenced by clang::CompilerInstance::ExecuteAction(), clang::ASTContext::getFunctionFeatureMap(), clang::targets::BPFTargetInfo::getTargetDefines(), clang::CodeGen::mergeDefaultFunctionDefinitionAttributes(), and runThinLTOBackend().

◆ DarwinTargetVariantSDKVersion

llvm::VersionTuple clang::TargetOptions::DarwinTargetVariantSDKVersion

The version of the darwin target variant SDK which was used during the compilation.

Definition at line 129 of file TargetOptions.h.

Referenced by GenerateTargetArgs(), and ParseTargetArgs().

◆ DarwinTargetVariantTriple

std::string clang::TargetOptions::DarwinTargetVariantTriple

The name of the darwin target- ariant triple to compile for.

Definition at line 125 of file TargetOptions.h.

◆ DxilValidatorVersion

std::string clang::TargetOptions::DxilValidatorVersion

The validator version for dxil.

Definition at line 132 of file TargetOptions.h.

◆ EABIVersion

llvm::EABI clang::TargetOptions::EABIVersion = llvm::EABI::Default

The EABI version to use.

Definition at line 48 of file TargetOptions.h.

Referenced by clang::targets::AArch64TargetInfo::AArch64TargetInfo(), clang::targets::ARMTargetInfo::ARMTargetInfo(), and initTargetOptions().

◆ FeatureMap

llvm::StringMap<bool> clang::TargetOptions::FeatureMap

The map of which features have been enabled disabled based on the command line.

Definition at line 62 of file TargetOptions.h.

Referenced by getCudaArch(), clang::ASTContext::getFunctionFeatureMap(), clang::ASTContext::getOpenMPDefaultSimdAlign(), and clang::targets::M68kTargetInfo::getTargetDefines().

◆ Features

std::vector<std::string> clang::TargetOptions::Features

The list of target specific features to enable or disable – this should be a list of strings starting with by '+' or '-'.

Definition at line 58 of file TargetOptions.h.

Referenced by clang::ASTContext::getFunctionFeatureMap(), clang::IncrementalExecutor::IncrementalExecutor(), overrideFunctionFeaturesWithTargetFeatures(), and runThinLTOBackend().

◆ FeaturesAsWritten

std::vector<std::string> clang::TargetOptions::FeaturesAsWritten

The list of target specific features to enable or disable, as written on the command line.

Definition at line 54 of file TargetOptions.h.

Referenced by clang::targets::ARMTargetInfo::ARMTargetInfo(), checkTargetOptions(), clang::ASTContext::getFunctionFeatureMap(), and clang::targets::NVPTXTargetInfo::NVPTXTargetInfo().

◆ ForceEnableInt128

bool clang::TargetOptions::ForceEnableInt128 = false

If given, enables support for __int128_t and __uint128_t types.

Definition at line 72 of file TargetOptions.h.

◆ FPMath

std::string clang::TargetOptions::FPMath

If given, the unit to use for floating point math.

Definition at line 42 of file TargetOptions.h.

◆ HLSLEntry

std::string clang::TargetOptions::HLSLEntry

The entry point name for HLSL shader being compiled as specified by -E.

Definition at line 135 of file TargetOptions.h.

Referenced by clang::Sema::ActOnHLSLTopLevelFunction().

◆ HostTriple

std::string clang::TargetOptions::HostTriple

When compiling for the device side, contains the triple used to compile for the host.

Definition at line 33 of file TargetOptions.h.

Referenced by clang::targets::BaseSPIRTargetInfo::BaseSPIRTargetInfo(), and clang::targets::NVPTXTargetInfo::NVPTXTargetInfo().

◆ LargeDataThreshold

uint64_t clang::TargetOptions::LargeDataThreshold

Definition at line 114 of file TargetOptions.h.

Referenced by FixupInvocation().

◆ LinkerVersion

std::string clang::TargetOptions::LinkerVersion

If given, the version string of the linker in use.

Definition at line 51 of file TargetOptions.h.

◆ NVPTXUseShortPointers

bool clang::TargetOptions::NVPTXUseShortPointers = false

If enabled, use 32-bit pointers for accessing const/local/shared address space.

Definition at line 76 of file TargetOptions.h.

Referenced by clang::targets::NVPTXTargetInfo::NVPTXTargetInfo().

◆ OpenCLExtensionsAsWritten

std::vector<std::string> clang::TargetOptions::OpenCLExtensionsAsWritten

The list of OpenCL extensions to enable or disable, as written on the command line.

Definition at line 69 of file TargetOptions.h.

◆ OpenCLFeaturesMap

llvm::StringMap<bool> clang::TargetOptions::OpenCLFeaturesMap

Supported OpenCL extensions and optional core features.

Definition at line 65 of file TargetOptions.h.

◆ SDKVersion

llvm::VersionTuple clang::TargetOptions::SDKVersion

The version of the SDK which was used during the compilation.

The option is used for two different purposes:

on darwin the version is propagated to LLVM where it's used to support SDK Version metadata (See D55673).
CUDA compilation uses it to control parts of CUDA compilation in clang that depend on specific version of the CUDA SDK.

Definition at line 122 of file TargetOptions.h.