LLVM: llvm::AMDGPUMachineFunction Class Reference (original) (raw)

#include "[Target/AMDGPU/AMDGPUMachineFunction.h](AMDGPUMachineFunction%5F8h%5Fsource.html)"

Public Member Functions
AMDGPUMachineFunction (const Function &F, const AMDGPUSubtarget &ST)
uint64_t getExplicitKernArgSize () const
Align getMaxKernArgAlign () const
uint32_t getLDSSize () const
uint32_t getGDSSize () const
bool isEntryFunction () const
bool isModuleEntryFunction () const
bool isChainFunction () const
bool isBottomOfStack () const
bool hasNoSignedZerosFPMath () const
bool isMemoryBound () const
bool needsWaveLimiter () const
bool hasInitWholeWave () const
void setInitWholeWave ()
unsigned allocateLDSGlobal (const DataLayout &DL, const GlobalVariable &GV)
unsigned allocateLDSGlobal (const DataLayout &DL, const GlobalVariable &GV, Align Trailing)
Align getDynLDSAlign () const
void setDynLDSAlign (const Function &F, const GlobalVariable &GV)
void setUsesDynamicLDS (bool DynLDS)
bool isDynamicLDSUsed () const
- Public Member Functions inherited from llvm::MachineFunctionInfo
virtual ~MachineFunctionInfo ()
virtual MachineFunctionInfo * clone (BumpPtrAllocator &Allocator, MachineFunction &DestMF, const DenseMap< MachineBasicBlock *, MachineBasicBlock * > &Src2DstMBB) const
Make a functionally equivalent copy of this MachineFunctionInfo in MF.
Static Public Member Functions
static std::optional< uint32_t > getLDSKernelIdMetadata (const Function &F)
static std::optional< uint32_t > getLDSAbsoluteAddress (const GlobalValue &GV)
- Static Public Member Functions inherited from llvm::MachineFunctionInfo
template<typename FuncInfoTy , typename SubtargetTy = TargetSubtargetInfo>
static FuncInfoTy * create (BumpPtrAllocator &Allocator, const Function &F, const SubtargetTy *STI)
Factory function: default behavior is to call new using the supplied allocator.
template
static Ty * create (BumpPtrAllocator &Allocator, const Ty &MFI)
Protected Attributes
uint64_t ExplicitKernArgSize = 0
Align MaxKernArgAlign
uint32_t LDSSize = 0
Number of bytes in the LDS that are being used.
uint32_t GDSSize = 0
uint32_t StaticLDSSize = 0
Number of bytes in the LDS allocated statically.
uint32_t StaticGDSSize = 0
Align DynLDSAlign
Align for dynamic shared memory if any.
bool UsesDynamicLDS = false
bool IsEntryFunction = false
bool IsModuleEntryFunction = false
bool IsChainFunction = false
bool NoSignedZerosFPMath = false
bool MemoryBound = false
bool WaveLimiter = false
bool HasInitWholeWave = false

Definition at line 24 of file AMDGPUMachineFunction.h.

Definition at line 42 of file AMDGPUMachineFunction.cpp.

References llvm::CallingConv::AMDGPU_KERNEL, CC, llvm::StringRef::consumeInteger(), llvm::StringRef::empty(), ExplicitKernArgSize, F, GDSSize, llvm::AMDGPU::getIntegerPairAttribute(), getKernelDynLDSGlobalFromFunction(), llvm::Attribute::getValueAsBool(), llvm::Attribute::getValueAsString(), hasLDSKernelArgument(), llvm::Attribute::isStringAttribute(), LDSSize, MaxKernArgAlign, MemoryBound, NoSignedZerosFPMath, llvm::CallingConv::SPIR_KERNEL, StaticGDSSize, StaticLDSSize, UsesDynamicLDS, and WaveLimiter.

◆ allocateLDSGlobal() [1/2]

◆ allocateLDSGlobal() [2/2]

TODO: We should sort these to minimize wasted space due to alignment padding. Currently the padding is decided by the first encountered use during lowering.

Definition at line 93 of file AMDGPUMachineFunction.cpp.

References llvm::alignTo(), assert(), DL, GDSSize, llvm::GlobalValue::getAddressSpace(), llvm::GlobalObject::getAlign(), getLDSAbsoluteAddress(), llvm::GlobalValue::getValueType(), isModuleEntryFunction(), llvm::AMDGPU::isNamedBarrier(), LDSSize, llvm_unreachable, llvm::AMDGPUAS::LOCAL_ADDRESS, llvm::Offset, llvm::AMDGPUAS::REGION_ADDRESS, llvm::report_fatal_error(), StaticGDSSize, and StaticLDSSize.

◆ getDynLDSAlign()

Align llvm::AMDGPUMachineFunction::getDynLDSAlign ( ) const inline

◆ getExplicitKernArgSize()

uint64_t llvm::AMDGPUMachineFunction::getExplicitKernArgSize ( ) const inline

◆ getGDSSize()

uint32_t llvm::AMDGPUMachineFunction::getGDSSize ( ) const inline

◆ getLDSAbsoluteAddress()

◆ getLDSKernelIdMetadata()

◆ getLDSSize()

uint32_t llvm::AMDGPUMachineFunction::getLDSSize ( ) const inline

◆ getMaxKernArgAlign()

Align llvm::AMDGPUMachineFunction::getMaxKernArgAlign ( ) const inline

◆ hasInitWholeWave()

bool llvm::AMDGPUMachineFunction::hasInitWholeWave ( ) const inline

◆ hasNoSignedZerosFPMath()

bool llvm::AMDGPUMachineFunction::hasNoSignedZerosFPMath ( ) const inline

◆ isBottomOfStack()

bool llvm::AMDGPUMachineFunction::isBottomOfStack ( ) const inline

◆ isChainFunction()

bool llvm::AMDGPUMachineFunction::isChainFunction ( ) const inline

◆ isDynamicLDSUsed()

bool AMDGPUMachineFunction::isDynamicLDSUsed ( ) const

◆ isEntryFunction()

bool llvm::AMDGPUMachineFunction::isEntryFunction ( ) const inline

Definition at line 89 of file AMDGPUMachineFunction.h.

References IsEntryFunction.

Referenced by llvm::SIMachineFunctionInfo::allocateWWMSpill(), llvm::SIFrameLowering::determineCalleeSaves(), llvm::SIFrameLowering::determineCalleeSavesSGPR(), llvm::SIFrameLowering::emitEntryFunctionPrologue(), llvm::SIFrameLowering::emitEpilogue(), llvm::AMDGPUAsmPrinter::emitFunctionBodyEnd(), llvm::AMDGPUAsmPrinter::emitFunctionBodyStart(), llvm::AMDGPUAsmPrinter::emitFunctionEntryLabel(), llvm::SIFrameLowering::emitPrologue(), llvm::SIFrameLowering::hasFPImpl(), isBottomOfStack(), llvm::AMDGPULegalizerInfo::legalizeImplicitArgPtr(), llvm::AMDGPULegalizerInfo::legalizeLDSKernelId(), llvm::SIFrameLowering::requiresStackPointerReference(), llvm::AMDGPUAsmPrinter::runOnMachineFunction(), and llvm::SIMachineFunctionInfo::SIMachineFunctionInfo().

◆ isMemoryBound()

bool llvm::AMDGPUMachineFunction::isMemoryBound ( ) const inline

◆ isModuleEntryFunction()

bool llvm::AMDGPUMachineFunction::isModuleEntryFunction ( ) const inline

◆ needsWaveLimiter()

bool llvm::AMDGPUMachineFunction::needsWaveLimiter ( ) const inline

◆ setDynLDSAlign()

Definition at line 207 of file AMDGPUMachineFunction.cpp.

References llvm::alignTo(), assert(), DL, DynLDSAlign, F, llvm::GlobalObject::getAlign(), getKernelDynLDSGlobalFromFunction(), getLDSAbsoluteAddress(), llvm::GlobalValue::getValueType(), LDSSize, llvm::Offset, llvm::report_fatal_error(), and StaticLDSSize.

Referenced by llvm::AMDGPULegalizerInfo::legalizeGlobalValue().

◆ setInitWholeWave()

void llvm::AMDGPUMachineFunction::setInitWholeWave ( ) inline

◆ setUsesDynamicLDS()

void AMDGPUMachineFunction::setUsesDynamicLDS ( bool DynLDS )

◆ DynLDSAlign

Align llvm::AMDGPUMachineFunction::DynLDSAlign protected

◆ ExplicitKernArgSize

uint64_t llvm::AMDGPUMachineFunction::ExplicitKernArgSize = 0 protected

◆ GDSSize

uint32_t llvm::AMDGPUMachineFunction::GDSSize = 0 protected

◆ HasInitWholeWave

bool llvm::AMDGPUMachineFunction::HasInitWholeWave = false protected

◆ IsChainFunction

bool llvm::AMDGPUMachineFunction::IsChainFunction = false protected

◆ IsEntryFunction

bool llvm::AMDGPUMachineFunction::IsEntryFunction = false protected

◆ IsModuleEntryFunction

bool llvm::AMDGPUMachineFunction::IsModuleEntryFunction = false protected

◆ LDSSize

uint32_t llvm::AMDGPUMachineFunction::LDSSize = 0 protected

◆ MaxKernArgAlign

Align llvm::AMDGPUMachineFunction::MaxKernArgAlign protected

◆ MemoryBound

bool llvm::AMDGPUMachineFunction::MemoryBound = false protected

◆ NoSignedZerosFPMath

bool llvm::AMDGPUMachineFunction::NoSignedZerosFPMath = false protected

◆ StaticGDSSize

uint32_t llvm::AMDGPUMachineFunction::StaticGDSSize = 0 protected

◆ StaticLDSSize

uint32_t llvm::AMDGPUMachineFunction::StaticLDSSize = 0 protected

◆ UsesDynamicLDS

bool llvm::AMDGPUMachineFunction::UsesDynamicLDS = false protected

◆ WaveLimiter

bool llvm::AMDGPUMachineFunction::WaveLimiter = false protected

The documentation for this class was generated from the following files: