Namespaces
	analysis_impl

	pass_options

	pdl_function_builder

	storage_user_base_impl

	StorageUserTrait

Classes
struct	AffineBinaryOpExprStorage
	A binary operation appearing in an affine expression. More...

struct	AffineConstantExprStorage
	An integer constant appearing in affine expression. More...

struct	AffineDimExprStorage
	A dimensional or symbolic identifier appearing in an affine expression. More...

struct	AffineExprStorage
	Base storage class appearing in an affine expression. More...

struct	AffineMapAccessInterfaceInterfaceTraits

struct	AffineMapAccessInterfaceTrait

struct	AffineMapStorage

struct	AffineReadOpInterfaceInterfaceTraits

struct	AffineReadOpInterfaceTrait

struct	AffineWriteOpInterfaceInterfaceTraits

struct	AffineWriteOpInterfaceTrait

struct	AliasAnalysisTraits
	This class contains various internal trait classes used by the main AliasAnalysis class below. More...

struct	AnalysisConcept
	The abstract polymorphic base class representing an analysis. More...

class	AnalysisMap
	This class represents a cache of analyses for a single operation. More...

struct	AnalysisModel
	A derived analysis model used to hold a specific analysis object. More...

struct	AnyCapturedValueMatcher
	Terminal matcher, always returns true. More...

struct	AnyValueMatcher
	Terminal matcher, always returns true. More...

class	AsmParserImpl
	This class provides the implementation of the generic parser methods within AsmParser. More...

class	AsmStateImpl

struct	attr_value_binder
	The matcher that matches a certain kind of Attribute and binds the value inside the Attribute. More...

class	AttributeUniquer

class	BlockArgumentImpl
	The internal implementation of a BlockArgument. More...

class	CallbackOstream
	A simple raw ostream subclass that forwards write_impl calls to the user-supplied callback together with opaque user-supplied data. More...

struct	constant_float_op_binder
	The matcher that matches a constant scalar / vector splat / tensor splat float operation and binds the constant float value. More...

struct	constant_float_predicate_matcher
	The matcher that matches a given target constant scalar / vector splat / tensor splat float value that fulfills a predicate. More...

struct	constant_int_op_binder
	The matcher that matches a constant scalar / vector splat / tensor splat integer operation and binds the constant integer value. More...

struct	constant_int_predicate_matcher
	The matcher that matches a given target constant scalar / vector splat / tensor splat integer value that fulfills a predicate. More...

struct	constant_op_binder
	The matcher that matches operations that have the `ConstantLike` trait, and binds the folded attribute value. More...

struct	constant_op_matcher
	The matcher that matches operations that have the `ConstantLike` trait. More...

struct	ConversionPatternRewriterImpl

struct	count_if_t_impl
	Template utility that computes the number of elements within `T` that satisfy the given predicate. More...

struct	count_if_t_impl< Pred, N, T, Us... >

class	DefaultTimingManagerImpl
	Implementation details of the `DefaultTimingManager`. More...

class	DenseArrayAttrImpl
	Base class for DenseArrayAttr that is instantiated and specialized for each supported element type below. More...

class	DenseElementIndexedIteratorImpl
	Impl iterator for indexed DenseElementsAttr iterators that records a data pointer and data index that is adjusted for the case of a splat attribute. More...

struct	DenseElementsAttributeStorage
	An attribute representing a reference to a dense vector or tensor object. More...

struct	DenseIntOrFPElementsAttrStorage
	An attribute representing a reference to a dense vector or tensor object. More...

class	DenseResourceElementsAttrBase
	Base class for DenseResourceElementsAttr that is instantiated and specialized for each supported element type below. More...

struct	DenseStringElementsAttrStorage
	An attribute representing a reference to a dense vector or tensor object containing strings. More...

struct	DiagnosticEngineImpl

class	DialectInterfaceBase
	The base class used for all derived interface types. More...

class	DialectInterfaceCollectionBase
	This class is the base class for a collection of instances for a specific interface kind. More...

class	DominanceInfoBase

struct	DynamicAttrStorage
	Storage of DynamicAttr. More...

struct	DynamicTypeStorage
	Storage of DynamicType. More...

struct	ElementsAttrIndexer
	This class provides support for indexing into the element range of an ElementsAttr. More...

class	ElementsAttrIterator
	This class implements a generic iterator for ElementsAttr. More...

class	ElementsAttrRange
	This class provides iterator utilities for an ElementsAttr range. More...

struct	ExpectedDiag
	This class represents an expected output diagnostic. More...

class	FallbackTypeIDResolver
	This class provides a fallback for resolving TypeIDs. More...

struct	FunctionTypeStorage
	Function Type Storage and Uniquing. More...

struct	InlineOpResult
	This class provides the implementation for an operation result whose index can be represented "inline" in the underlying ValueImpl. More...

struct	InlineTypeIDResolver
	This class provides utilities for resolving the TypeID of a class that provides a `static TypeID resolveTypeID()` method. More...

struct	IntegerSetStorage

struct	IntegerTypeStorage
	Integer Type Storage and Uniquing. More...

class	Interface
	This class represents an abstract interface. More...

class	InterfaceMap
	This class provides an efficient mapping between a given `Interface` type, and a particular implementation of its concept. More...

class	IROperandBase
	This class is the base for IROperand, and provides all of the non-templated facilities for operand use management. More...

struct	is_complex_t
	Type trait detector that checks if a given type T is a complex type. More...

struct	is_complex_t< std::complex< T > >

struct	NestedAnalysisMap
	An analysis map that contains a map for the current operation, and a set of maps for any child operations. More...

class	op_filter_iterator
	A utility iterator that filters out operations that are not 'OpT'. More...

class	op_iterator
	This class provides iteration over the held operations of a block for a specific operation type. More...

struct	op_matcher
	The matcher that matches a certain kind of op. More...

class	OperandStorage
	This class handles the management of operation operands. More...

struct	OpOrInterfaceRewritePatternBase
	OpOrInterfaceRewritePatternBase is a wrapper around RewritePattern that allows for matching and rewriting against an instance of a derived operation class or Interface. More...

struct	OpPassManagerImpl

class	OpResultImpl
	This class provides the implementation for an operation result. More...

class	OpToOpPassAdaptor
	An adaptor pass used to run operation passes over nested operations. More...

class	OutOfLineOpResult
	This class provides the implementation for an operation result whose index cannot be represented "inline", and thus requires an additional index field. More...

struct	ParallelDiagnosticHandlerImpl

class	Parser
	This class implement support for parsing global entities like attributes and types. More...

struct	ParserState
	This class refers to all of the state maintained globally by the parser, such as the current lexer position etc. More...

class	PassCrashReproducerGenerator

struct	PassExecutionState
	The state for a single execution of a pass. More...

struct	PassInstrumentorImpl

class	PassOptions
	Base container class and manager for all pass options. More...

struct	PassPipelineCLParserImpl

struct	PatternMatcherValue
	Binds to a specific value and matches it. More...

class	PDLByteCode
	The bytecode class is also the interpreter. More...

class	PDLByteCodeMutableState
	This class contains the mutable state of a bytecode instance. More...

class	PDLByteCodePattern
	All of the data pertaining to a specific pattern within the bytecode. More...

class	PreservedAnalyses
	A utility class to represent the analyses that are known to be preserved. More...

struct	RecoveryReproducerContext
	This class contains all of the context for generating a recovery reproducer. More...

struct	RecursivePatternMatcher
	RecursivePatternMatcher that composes. More...

struct	SourceMgrDiagnosticHandlerImpl

struct	SourceMgrDiagnosticVerifierHandlerImpl

struct	StorageUniquerImpl
	This is the implementation of the StorageUniquer class. More...

class	StorageUserBase
	Utility class for implementing users of storage classes uniqued by a StorageUniquer. More...

class	StorageUserTraitBase
	Helper class for implementing traits for storage classes. More...

struct	StringAttrStorage

struct	SymbolState
	This class contains record of any parsed top-level symbols. More...

class	TimingManagerImpl
	Private implementation details of the `TimingManager`. More...

struct	TupleTypeStorage
	A type representing a collection of other types. More...

struct	TypedValue
	TypedValue is a Value with a statically know type. More...

class	TypeIDResolver
	This class provides a resolver for getting the ID for a given class T. More...

class	TypeIDResolver< T, std::enable_if_t< InlineTypeIDResolver::has_resolve_typeid< T >::value > >
	This class provides a resolver for getting the ID for a given class T, when the class provides a `static TypeID resolveTypeID()` method. More...

struct	TypeUniquer
	A utility class to get, or create, unique instances of types within an MLIRContext. More...

class	ValueImpl
	The base class for all derived Value classes. More...

struct	Vector1D

struct	Vector1D< T, Dim, false >

struct	Vector1D< T, Dim, true >

Typedefs
using	DenseIterPtrAndSplat = std::pair< const char *, bool >
	Pair of raw pointer and a boolean flag of whether the pointer holds a splat,. More...

template<typename T , typename OperationOrValue >
using	has_operation_or_value_matcher_t = decltype(std::declval< T >().match(std::declval< OperationOrValue >()))
	Trait to check whether T provides a 'match' method with type `OperationOrValue`. More...

template<typename T >
using	first_argument = decltype(first_argument_type(std::declval< T >()))
	Type definition of the first argument to the given callable 'T'. More...

template<typename FnT >
using	walkResultType = decltype(walk(nullptr, std::declval< FnT >()))
	Utility to provide the return type of a templated walk method. More...

template<template< class > class Pred, typename... Ts>
using	count_if_t = count_if_t_impl< Pred, 0, Ts... >

template<typename T >
using	is_interface_t = decltype(isInterfaceImpl(std::declval< T & >()))

template<typename T >
using	IsInterface = llvm::is_detected< is_interface_t, T >

template<typename ImplTy , typename... Args>
using	has_impltype_getkey_t = decltype(ImplTy::getKey(std::declval< Args >()...))
	Trait to check if ImplTy provides a 'getKey' method with types 'Args'. More...

template<typename ImplTy , typename T >
using	has_impltype_hash_t = decltype(ImplTy::hashKey(std::declval< T >()))
	Trait to check if ImplTy provides a 'hashKey' method for 'T'. More...

using	ByteCodeField = uint16_t
	Use generic bytecode types. More...

using	ByteCodeAddr = uint32_t

using	OwningOpRange = llvm::OwningArrayRef< Operation * >

Functions
constexpr bool	isPowerOf2 (int n)

constexpr unsigned	nextPowerOf2 (int n)

template<size_t N>
std::array< int64_t, N >	makeStrides (ArrayRef< int64_t > shape)
	Given a shape with sizes greater than 0 along all dimensions, returns the distance, in number of elements, between a slice in a dimension and the next slice in the same dimension. More...

template<int N, typename T >
std::enable_if<(N >=1), StridedMemRefType< T, N > >::type	makeStridedMemRefDescriptor (T ptr, T alignedPtr, ArrayRef< int64_t > shape, ArrayRef< int64_t > shapeAlloc)
	Build a `StridedMemRefDescriptor<T, N>` that matches the MLIR ABI. More...

template<int N, typename T >
std::enable_if<(N==0), StridedMemRefType< T, 0 > >::type	makeStridedMemRefDescriptor (T ptr, T alignedPtr, ArrayRef< int64_t > shape={}, ArrayRef< int64_t > shapeAlloc={})
	Build a `StridedMemRefDescriptor<T, 0>` that matches the MLIR ABI. More...

template<typename T >
std::pair< T , T >	allocAligned (size_t nElements, AllocFunType allocFun=&::malloc, llvm::Optional< uint64_t > alignment=llvm::Optional< uint64_t >())
	Align `nElements` of type T with an optional `alignment`. More...

Optional< BlockArgument >	getBranchSuccessorArgument (const SuccessorOperands &operands, unsigned operandIndex, Block *successor)
	Return the `BlockArgument` corresponding to operand `operandIndex` in some successor if `operandIndex` is within the range of `operands`, or None if `operandIndex` isn't a successor operand index. More...

LogicalResult	verifyBranchSuccessorOperands (Operation *op, unsigned succNo, const SuccessorOperands &operands)
	Verify that the given operands match those of the given successor block. More...

LogicalResult	verifyTypesAlongControlFlowEdges (Operation *op)
	Verify that types match along control flow edges described the given op. More...

unsigned	getDefaultTypeSize (Type type, const DataLayout &dataLayout, DataLayoutEntryListRef params)
	Default handler for the type size request. More...

unsigned	getDefaultTypeSizeInBits (Type type, const DataLayout &dataLayout, DataLayoutEntryListRef params)
	Default handler for the type size in bits request. More...

unsigned	getDefaultABIAlignment (Type type, const DataLayout &dataLayout, ArrayRef< DataLayoutEntryInterface > params)
	Default handler for the required alignemnt request. More...

unsigned	getDefaultPreferredAlignment (Type type, const DataLayout &dataLayout, ArrayRef< DataLayoutEntryInterface > params)
	Default handler for the preferred alignemnt request. More...

DataLayoutEntryList	filterEntriesForType (DataLayoutEntryListRef entries, TypeID typeID)
	Given a list of data layout entries, returns a new list containing the entries with keys having the given type ID, i.e. More...

DataLayoutEntryInterface	filterEntryForIdentifier (DataLayoutEntryListRef entries, StringAttr id)
	Given a list of data layout entries, returns the entry that has the given identifier as key, if such an entry exists in the list. More...

LogicalResult	verifyDataLayoutOp (Operation *op)
	Verifies that the operation implementing the data layout interface, or a module operation, is valid. More...

LogicalResult	verifyDataLayoutSpec (DataLayoutSpecInterface spec, Location loc)
	Verifies that a data layout spec is valid. More...

LogicalResult	inferReturnTensorTypes (function_ref< LogicalResult(MLIRContext , Optional< Location > location, ValueShapeRange operands, DictionaryAttr attributes, RegionRange regions, SmallVectorImpl< ShapedTypeComponents > &retComponents)> componentTypeFn, MLIRContext context, Optional< Location > location, ValueRange operands, DictionaryAttr attributes, RegionRange regions, SmallVectorImpl< Type > &inferredReturnTypes)

LogicalResult	verifyInferredResultTypes (Operation *op)
	Verifies that the inferred result types match the actual result types for the op. More...

LogicalResult	verifyParallelCombiningOpInterface (Operation *op)

LogicalResult	verifyShapedDimOpInterface (Operation *op)
	Verify invariants of ops that implement the ShapedDimOpInterface. More...

LogicalResult	verifyOffsetSizeAndStrideOp (OffsetSizeAndStrideOpInterface op)

bool	sameOffsetsSizesAndStrides (OffsetSizeAndStrideOpInterface a, OffsetSizeAndStrideOpInterface b, llvm::function_ref< bool(OpFoldResult, OpFoldResult)> cmp)

template<int N>
void	bindDims (MLIRContext *ctx)

template<int N, typename AffineExprTy , typename... AffineExprTy2>
void	bindDims (MLIRContext *ctx, AffineExprTy &e, AffineExprTy2 &...exprs)

template<int N>
void	bindSymbols (MLIRContext *ctx)

template<int N, typename AffineExprTy , typename... AffineExprTy2>
void	bindSymbols (MLIRContext *ctx, AffineExprTy &e, AffineExprTy2 &...exprs)

template<typename AffineExprTy >
void	bindSymbolsList (MLIRContext *ctx, SmallVectorImpl< AffineExprTy > &exprs)

LogicalResult	verifyAffineMapAsLayout (AffineMap m, ArrayRef< int64_t > shape, function_ref< InFlightDiagnostic()> emitError)

static bool	isConstantLike (Operation *op)
	Check to see if the specified operation is ConstantLike. More...

template<typename MatcherClass >
std::enable_if_t< llvm::is_detected< detail::has_operation_or_value_matcher_t, MatcherClass, Value >::value, bool >	matchOperandOrValueAtIndex (Operation *op, unsigned idx, MatcherClass &matcher)
	Statically switch to a Value matcher. More...

template<typename MatcherClass >
std::enable_if_t< llvm::is_detected< detail::has_operation_or_value_matcher_t, MatcherClass, Operation * >::value, bool >	matchOperandOrValueAtIndex (Operation *op, unsigned idx, MatcherClass &matcher)
	Statically switch to an Operation matcher. More...

template<typename TupleT , class CallbackT , std::size_t... Is>
constexpr void	enumerateImpl (TupleT &&tuple, CallbackT &&callback, std::index_sequence< Is... >)

template<typename... Tys, typename CallbackT >
constexpr void	enumerate (std::tuple< Tys... > &tuple, CallbackT &&callback)

llvm::unique_function< InFlightDiagnostic()>	getDefaultDiagnosticEmitFn (MLIRContext *ctx)
	Utility method to generate a callback that can be used to generate a diagnostic when checking the construction invariants of a storage object. More...

llvm::unique_function< InFlightDiagnostic()>	getDefaultDiagnosticEmitFn (const Location &loc)

LogicalResult	verifySymbolTable (Operation *op)

LogicalResult	verifySymbol (Operation *op)

template<typename Ret , typename Arg , typename... Rest>
Arg	first_argument_type (Ret(*)(Arg, Rest...))
	Helper templates to deduce the first argument of a callback parameter. More...

template<typename Ret , typename F , typename Arg , typename... Rest>
Arg	first_argument_type (Ret(F::*)(Arg, Rest...))

void	walk (Operation op, function_ref< void(Region )> callback, WalkOrder order)
	Walk all of the regions, blocks, or operations nested under (and including) the given operation. More...

void	walk (Operation op, function_ref< void(Block )> callback, WalkOrder order)

void	walk (Operation op, function_ref< void(Operation )> callback, WalkOrder order)

WalkResult	walk (Operation op, function_ref< WalkResult(Region )> callback, WalkOrder order)
	Walk all of the regions, blocks, or operations nested under (and including) the given operation. More...

WalkResult	walk (Operation op, function_ref< WalkResult(Block )> callback, WalkOrder order)

WalkResult	walk (Operation op, function_ref< WalkResult(Operation )> callback, WalkOrder order)

template<WalkOrder Order = WalkOrder::PostOrder, typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()(std::declval<ArgT>()))>
std::enable_if_t< llvm::is_one_of< ArgT, Operation , Region , Block * >::value, RetT >	walk (Operation *op, FuncTy &&callback)
	Walk all of the regions, blocks, or operations nested under (and including) the given operation. More...

template<WalkOrder Order = WalkOrder::PostOrder, typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()(std::declval<ArgT>()))>
std::enable_if_t< !llvm::is_one_of< ArgT, Operation , Region , Block * >::value &&std::is_same< RetT, void >::value, RetT >	walk (Operation *op, FuncTy &&callback)
	Walk all of the operations of type 'ArgT' nested under and including the given operation. More...

template<WalkOrder Order = WalkOrder::PostOrder, typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()(std::declval<ArgT>()))>
std::enable_if_t< !llvm::is_one_of< ArgT, Operation , Region , Block * >::value &&std::is_same< RetT, WalkResult >::value, RetT >	walk (Operation *op, FuncTy &&callback)
	Walk all of the operations of type 'ArgT' nested under and including the given operation. More...

void	walk (Operation op, function_ref< void(Operation , const WalkStage &stage)> callback)
	Generic walkers with stage aware callbacks. More...

WalkResult	walk (Operation op, function_ref< WalkResult(Operation , const WalkStage &stage)> callback)
	Walk all the operations nested under (and including) the given operation, with the callback being invoked on each operation N+1 times, where N is the number of regions attached to the operation. More...

template<typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()( std::declval<ArgT>(), std::declval<const WalkStage &>()))>
std::enable_if_t< std::is_same< ArgT, Operation * >::value, RetT >	walk (Operation *op, FuncTy &&callback)
	Walk all of the operations nested under and including the given operation. More...

template<typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()( std::declval<ArgT>(), std::declval<const WalkStage &>()))>
std::enable_if_t<!std::is_same< ArgT, Operation * >::value &&std::is_same< RetT, void >::value, RetT >	walk (Operation *op, FuncTy &&callback)
	Walk all of the operations of type 'ArgT' nested under and including the given operation. More...

template<typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()( std::declval<ArgT>(), std::declval<const WalkStage &>()))>
std::enable_if_t<!std::is_same< ArgT, Operation * >::value &&std::is_same< RetT, WalkResult >::value, RetT >	walk (Operation *op, FuncTy &&callback)
	Walk all of the operations of type 'ArgT' nested under and including the given operation. More...

template<typename ContainerOpT >
OwningOpRef< ContainerOpT >	constructContainerOpForParserIfNecessary (Block parsedBlock, MLIRContext context, Location sourceFileLoc)
	Given a block containing operations that have just been parsed, if the block contains a single operation of `ContainerOpT` type then remove it from the block and return it. More...

template<typename ContainerOpT , typename... ParserArgs>
OwningOpRef< ContainerOpT >	parseSourceFile (const ParserConfig &config, ParserArgs &&...args)
	The internal implementation of the templated `parseSourceFile` methods below, that simply forwards to the non-templated version. More...

template<typename ConcreteType , typename ValueT , typename Traits , typename BaseType , template< typename, template< typename > class > class BaseTrait>
void	isInterfaceImpl (Interface< ConcreteType, ValueT, Traits, BaseType, BaseTrait > &)

size_t	getDenseElementBitWidth (Type eltType)
	Return the bit width which DenseElementsAttr should use for this type. More...

bool	isSupportedMemorySpace (Attribute memorySpace)
	Checks if the memorySpace has supported Attribute type. More...

Attribute	wrapIntegerMemorySpace (unsigned memorySpace, MLIRContext *ctx)
	Wraps deprecated integer memory space to the new Attribute form. More...

Attribute	skipDefaultMemorySpace (Attribute memorySpace)
	Replaces default memorySpace (integer == `0`) with empty Attribute. More...

unsigned	getMemorySpaceAsInt (Attribute memorySpace)
	[deprecated] Returns the memory space in old raw integer representation. More...

Variables
template<typename F >
decltype(first_argument_type(&F::operator()))	first_argument_type (F)

Typedef Documentation

◆ ByteCodeAddr

using mlir::detail::ByteCodeAddr = typedef uint32_t

Definition at line 30 of file ByteCode.h.

◆ ByteCodeField

using mlir::detail::ByteCodeField = typedef uint16_t

Use generic bytecode types.

ByteCodeField refers to the actual bytecode entries. ByteCodeAddr refers to size of indices into the bytecode.

Definition at line 29 of file ByteCode.h.

◆ count_if_t

template<template< class > class Pred, typename... Ts>

using mlir::detail::count_if_t = typedef count_if_t_impl<Pred, 0, Ts...>

Definition at line 143 of file InterfaceSupport.h.

◆ DenseIterPtrAndSplat

using mlir::detail::DenseIterPtrAndSplat = typedef std::pair<const char *, bool>

Pair of raw pointer and a boolean flag of whether the pointer holds a splat,.

Definition at line 40 of file BuiltinAttributes.h.

◆ first_argument

template<typename T >

using mlir::detail::first_argument = typedef decltype(first_argument_type(std::declval<T>()))

Type definition of the first argument to the given callable 'T'.

Definition at line 112 of file Visitors.h.

◆ has_impltype_getkey_t

template<typename ImplTy , typename... Args>

using mlir::detail::has_impltype_getkey_t = typedef decltype(ImplTy::getKey(std::declval<Args>()...))

Trait to check if ImplTy provides a 'getKey' method with types 'Args'.

Definition at line 26 of file StorageUniquer.h.

◆ has_impltype_hash_t

template<typename ImplTy , typename T >

using mlir::detail::has_impltype_hash_t = typedef decltype(ImplTy::hashKey(std::declval<T>()))

Trait to check if ImplTy provides a 'hashKey' method for 'T'.

Definition at line 30 of file StorageUniquer.h.

◆ has_operation_or_value_matcher_t

template<typename T , typename OperationOrValue >

using mlir::detail::has_operation_or_value_matcher_t = typedef decltype(std::declval<T>().match(std::declval<OperationOrValue>()))

Trait to check whether T provides a 'match' method with type OperationOrValue.

Definition at line 177 of file Matchers.h.

◆ is_interface_t

template<typename T >

using mlir::detail::is_interface_t = typedef decltype(isInterfaceImpl(std::declval<T &>()))

Definition at line 278 of file InterfaceSupport.h.

◆ IsInterface

template<typename T >

using mlir::detail::IsInterface = typedef llvm::is_detected<is_interface_t, T>

Definition at line 281 of file InterfaceSupport.h.

◆ OwningOpRange

using mlir::detail::OwningOpRange = typedef llvm::OwningArrayRef<Operation *>

Definition at line 31 of file ByteCode.h.

◆ walkResultType

template<typename FnT >

using mlir::detail::walkResultType = typedef decltype(walk(nullptr, std::declval<FnT>()))

Utility to provide the return type of a templated walk method.

Definition at line 319 of file Visitors.h.

Function Documentation

◆ allocAligned()

template<typename T >

std::pair<T , T > mlir::detail::allocAligned	(	size_t	nElements,
		AllocFunType	allocFun = `&::malloc`,
		llvm::Optional< uint64_t >	alignment = `llvm::Optional<uint64_t>()`
	)

Align nElements of type T with an optional alignment.

This replaces a portable posix_memalign. alignment must be a power of 2 and greater than the size of T. By default the alignment is sizeof(T).

Definition at line 103 of file MemRefUtils.h.

References nextPowerOf2().

◆ bindDims() [1/2]

template<int N>

void mlir::detail::bindDims ( MLIRContext * ctx )

Definition at line 307 of file AffineExpr.h.

Referenced by bindDims().

◆ bindDims() [2/2]

template<int N, typename AffineExprTy , typename... AffineExprTy2>

void mlir::detail::bindDims	(	MLIRContext *	ctx,
		AffineExprTy &	e,
		AffineExprTy2 &...	exprs
	)

Definition at line 310 of file AffineExpr.h.

References bindDims(), and mlir::getAffineDimExpr().

◆ bindSymbols() [1/2]

template<int N>

void mlir::detail::bindSymbols ( MLIRContext * ctx )

Definition at line 316 of file AffineExpr.h.

Referenced by bindSymbols().

◆ bindSymbols() [2/2]

template<int N, typename AffineExprTy , typename... AffineExprTy2>

void mlir::detail::bindSymbols	(	MLIRContext *	ctx,
		AffineExprTy &	e,
		AffineExprTy2 &...	exprs
	)

Definition at line 319 of file AffineExpr.h.

References bindSymbols(), and mlir::getAffineSymbolExpr().

◆ bindSymbolsList()

template<typename AffineExprTy >

void mlir::detail::bindSymbolsList	(	MLIRContext *	ctx,
		SmallVectorImpl< AffineExprTy > &	exprs
	)

Definition at line 325 of file AffineExpr.h.

References mlir::getAffineSymbolExpr().

◆ constructContainerOpForParserIfNecessary()

template<typename ContainerOpT >

OwningOpRef<ContainerOpT> mlir::detail::constructContainerOpForParserIfNecessary	(	Block *	parsedBlock,
		MLIRContext *	context,
		Location	sourceFileLoc
	)

inline

Given a block containing operations that have just been parsed, if the block contains a single operation of ContainerOpT type then remove it from the block and return it.

If the block does not contain just that operation, create a new operation instance of ContainerOpT and move all of the operations within parsedBlock into the first block of the first region. ContainerOpT is required to have a single region containing a single block, and must implement the SingleBlockImplicitTerminator trait.

Definition at line 38 of file Parser.h.

References mlir::Block::begin(), mlir::OpBuilder::create(), mlir::emitError(), mlir::failed(), mlir::Block::front(), mlir::Block::getOperations(), mlir::parseSourceFile(), mlir::parseSourceString(), and value.

◆ enumerate()

template<typename... Tys, typename CallbackT >

constexpr void mlir::detail::enumerate	(	std::tuple< Tys... > &	tuple,
		CallbackT &&	callback
	)

Definition at line 232 of file Matchers.h.

References enumerateImpl().

Referenced by ableToUpdatePredOperands(), adjustIter(), mlir::alignAffineMapWithValues(), mlir::bufferization::allocateTensorForShapedValue(), mlir::PatternApplicator::applyCostModel(), applyPatterns(), mlir::applyPermutationToVector(), areSameBitwidthScalarType(), mlir::bufferization::func_ext::CallOpInterface::bufferize(), mlir::bufferization::func_ext::FuncOpInterface::bufferize(), buildCostGraph(), mlir::python::PyOpView::buildGeneric(), mlir::scf::buildLoopNest(), buildPredicateList(), cloneAndFuseFirstUse(), collapseGenericOpIterationDims(), composeAffineMapAndOperands(), computeMemRefRankReductionMask(), mlir::LLVM::detail::connectPHINodes(), constructResultVector(), convertForOp(), mlir::LLVMTypeConverter::convertFunctionSignature(), convertIntrinsicResult(), convertOperationImpl(), convertYieldOp(), mlir::tensor::createDimValues(), mlir::tensor::createDynamicDimValues(), createFullTiles(), createI32Constant(), createLessThanFunc(), mlir::ConvertToLLVMPattern::createMemRefDescriptor(), mlir::tensor::createPadHighOp(), createReassociationMapsForCollapse(), mlir::ValueDecomposer::decomposeValue(), deduceCanonicalResource(), defaultAllocBufferCallBack(), diagnoseInvalidOperandDominance(), mlir::linalg::GenerateLoopNest< LoopTy >::doit(), doubleBuffer(), mlir::bufferization::dropEquivalentBufferResults(), dropSwitchCasesThatMatchDefault(), mlir::ComputationSliceState::dump(), elementwiseMatchAndRewriteHelper(), encodeKernelName(), findDuplicateElement(), foldSwitch(), foldTransferInBoundsAttribute(), forwardTerminatorOperands(), fuse(), generateCollapsedIndexingRegion(), generateTileLoopNest(), mlir::sparse_tensor::genReshapeDstShape(), getBackwardSliceImpl(), getBlockIt(), getCollapsableIterationSpaceDims(), getDimMap(), mlir::linalg::getDynOperands(), getExpandedDimToCollapsedDimMap(), getExpandedOutputDimFromInputShape(), getGenericEffectsImpl(), mlir::getLinearizedDimensions(), getLspDiagnoticFromDiag(), mlir::scf::getParallelForInductionVarOwner(), getPHISourceValue(), mlir::linalg::getReassociationMapForFoldingUnitDims(), getSemiAffineExprFromFlatForm(), getShapeDefiningLoopRange(), mlir::getSinglyExecutedRegionsToSink(), getTiledSliceDims(), getTreePredicates(), getTypeNumBytes(), getZero(), mlir::LLVM::detail::handleMultidimensionalVectors(), handleMultidimensionalVectors(), mlir::linalg::hoistRedundantVectorTransfersOnTensor(), injectGpuIndexOperations(), inlinePayload(), inlineRegionImpl(), insertCopyLoops(), mlir::scf::insideMutuallyExclusiveBranches(), mlir::inversePermutation(), invertPermutationVector(), mlir::bytecode::detail::IRNumberingState::IRNumberingState(), mlir::vector::isBroadcastableTo(), isConstantOne(), isDefinedByCallTo(), isDimSequencePreserved(), isFusableWithReshapeByDimExpansion(), isMemRefNormalizable(), mlir::AffineMap::isMinorIdentityWithBroadcasting(), mlir::isOpwiseShiftValid(), mlir::AffineMap::isPermutationOfMinorIdentityWithBroadcasting(), mlir::isReassociationValid(), isTrivialSubViewOp(), mlir::linalg::LinalgDependenceGraph::LinalgDependenceGraph(), mlir::spirv::linearizeIndex(), logFailure(), lowerAsEntryFunction(), lowerEntryPointABIAttr(), mlir::makeStridedLinearLayoutMap(), mlir::GPUFuncOpLowering::matchAndRewrite(), TransferReadPermutationLowering::matchAndRewrite(), ReduceMultiDimReductionRank::matchAndRewrite(), TransposeOpLowering::matchAndRewrite(), AllocaScopeHoister::matchAndRewrite(), CanonicalizeContractAdd< AddOpType >::matchAndRewrite(), moveRegionToNewWarpOpAndAppendReturns(), outlineKernelFuncImpl(), packFunctionArguments(), padOperandToSmallestStaticBoundingBox(), parseGlobalMemrefOpTypeAndInitialValue(), parseInferType(), parseOptionalKeywordAlternative(), mlir::permuteLoops(), permuteValues(), poolingInferReturnTypes(), printDenseI64ArrayAttr(), mlir::AsmPrinter::Impl::printIntegerSet(), printResultsAsList(), printSwitchOpCases(), produceSliceErrorMsg(), mlir::linalg::promoteSubviewAsNewBuffer(), mlir::detail::ConversionPatternRewriterImpl::remapValues(), mlir::replaceAllMemRefUsesWith(), replaceUnitDimIndexOps(), mlir::reshapeLikeShapesAreCompatible(), mlir::linalg::LinalgPaddingPattern::returningMatchAndRewrite(), mlir::linalg::rewriteAsPaddedOp(), simplifySwitchFromDefaultSwitchOnSameCondition(), sliceTransferIndices(), mlir::SPIRVTypeConverter::SPIRVTypeConverter(), splatZero(), mlir::linalg::splitReduction(), mlir::scf::tileConsumerAndFuseProducerGreedilyUsingSCFForOp(), tileLinalgOpImpl(), mlir::scf::tileParallelLoop(), mlir::linalg::TileLoopNest::tileRootOp(), mlir::scf::tileUsingSCFForOp(), mlir::linalg::transformIndexOps(), mlir::sparse_tensor::translateIndicesArray(), updateDestinationOperandsForTiledOp(), updateFuncOp(), vectorizeLinalgYield(), verifyAttributions(), verifyCollapsedShape(), mlir::verifyReshapeLikeTypes(), verifyResultTypesAreInferrable(), verifyTypesAlongAllEdges(), mlir::dataflow::AbstractSparseDataFlowAnalysis::visit(), walkReferenceCountedValues(), wrapExternalFunction(), wrapForExternalCallers(), and yieldTiledValues().

◆ enumerateImpl()

template<typename TupleT , class CallbackT , std::size_t... Is>

constexpr void mlir::detail::enumerateImpl	(	TupleT &&	tuple,
		CallbackT &&	callback,
		std::index_sequence< Is... >
	)

Definition at line 224 of file Matchers.h.

Referenced by enumerate().

◆ filterEntriesForType()

DataLayoutEntryList mlir::detail::filterEntriesForType	(	DataLayoutEntryListRef	entries,
		TypeID	typeID
	)

Given a list of data layout entries, returns a new list containing the entries with keys having the given type ID, i.e.

belonging to the same type class.

Definition at line 217 of file DataLayoutInterfaces.cpp.

References mlir::Type::getTypeID().

◆ filterEntryForIdentifier()

DataLayoutEntryInterface mlir::detail::filterEntryForIdentifier	(	DataLayoutEntryListRef	entries,
		StringAttr	id
	)

Given a list of data layout entries, returns the entry that has the given identifier as key, if such an entry exists in the list.

Definition at line 227 of file DataLayoutInterfaces.cpp.

◆ first_argument_type() [1/2]

template<typename Ret , typename Arg , typename... Rest>

Arg mlir::detail::first_argument_type ( Ret(*)(Arg, Rest...) )

Helper templates to deduce the first argument of a callback parameter.

◆ first_argument_type() [2/2]

template<typename Ret , typename F , typename Arg , typename... Rest>

Arg mlir::detail::first_argument_type ( Ret(F::*)(Arg, Rest...) )

Referenced by mlir::WalkStage::getNextRegion().

◆ getBranchSuccessorArgument()

Optional< BlockArgument > mlir::detail::getBranchSuccessorArgument	(	const SuccessorOperands &	operands,
		unsigned	operandIndex,
		Block *	successor
	)

Return the BlockArgument corresponding to operand operandIndex in some successor if operandIndex is within the range of operands, or None if operandIndex isn't a successor operand index.

Returns the BlockArgument corresponding to operand operandIndex in some successor if 'operandIndex' is within the range of 'operands', or None if operandIndex isn't a successor operand index.

Definition at line 40 of file ControlFlowInterfaces.cpp.

References mlir::Block::getArgument(), mlir::OperandRange::getBeginOperandIndex(), mlir::SuccessorOperands::getForwardedOperands(), and mlir::SuccessorOperands::getProducedOperandCount().

Referenced by mlir::SuccessorOperands::getOperandIndex().

◆ getDefaultABIAlignment()

unsigned mlir::detail::getDefaultABIAlignment	(	Type	type,
		const DataLayout &	dataLayout,
		ArrayRef< DataLayoutEntryInterface >	params
	)

Default handler for the required alignemnt request.

Computes results for built-in types and dispatches to the DataLayoutTypeInterface for other types.

Definition at line 133 of file DataLayoutInterfaces.cpp.

References mlir::Type::dyn_cast(), mlir::Type::getContext(), getFloatTypeABIAlignment(), getIndexBitwidth(), getIntegerTypeABIAlignment(), mlir::DataLayout::getTypeABIAlignment(), mlir::DataLayout::getTypeSize(), mlir::Type::isa(), and reportMissingDataLayout().

Referenced by mlir::DataLayout::getTypeABIAlignment().

◆ getDefaultDiagnosticEmitFn() [1/2]

llvm::unique_function< InFlightDiagnostic()> mlir::detail::getDefaultDiagnosticEmitFn ( MLIRContext * ctx )

Utility method to generate a callback that can be used to generate a diagnostic when checking the construction invariants of a storage object.

This is defined out-of-line to avoid the need to include Location.h.

Definition at line 1069 of file MLIRContext.cpp.

References mlir::emitError().

Referenced by mlir::detail::StorageUserBase< ConcreteT, BaseT, StorageT, UniquerT, Traits >::get(), mlir::DynamicType::get(), and mlir::detail::StorageUserBase< ConcreteT, BaseT, StorageT, UniquerT, Traits >::getChecked().

◆ getDefaultDiagnosticEmitFn() [2/2]

llvm::unique_function< InFlightDiagnostic()> mlir::detail::getDefaultDiagnosticEmitFn ( const Location & loc )

Definition at line 1073 of file MLIRContext.cpp.

References mlir::emitError().

◆ getDefaultPreferredAlignment()

unsigned mlir::detail::getDefaultPreferredAlignment	(	Type	type,
		const DataLayout &	dataLayout,
		ArrayRef< DataLayoutEntryInterface >	params
	)

Default handler for the preferred alignemnt request.

Computes results for built-in types and dispatches to the DataLayoutTypeInterface for other types.

Definition at line 186 of file DataLayoutInterfaces.cpp.

References mlir::Type::dyn_cast(), mlir::Type::getContext(), getFloatTypePreferredAlignment(), getIndexBitwidth(), getIntegerTypePreferredAlignment(), mlir::DataLayout::getTypeABIAlignment(), mlir::DataLayout::getTypePreferredAlignment(), mlir::Type::isa(), and reportMissingDataLayout().

Referenced by getDefaultTypeSizeInBits(), and mlir::DataLayout::getTypePreferredAlignment().

◆ getDefaultTypeSize()

unsigned mlir::detail::getDefaultTypeSize	(	Type	type,
		const DataLayout &	dataLayout,
		DataLayoutEntryListRef	params
	)

Default handler for the type size request.

Computes results for built-in types and dispatches to the DataLayoutTypeInterface for other types.

Definition at line 45 of file DataLayoutInterfaces.cpp.

References getDefaultTypeSizeInBits().

Referenced by mlir::DataLayout::getTypeSize().

◆ getDefaultTypeSizeInBits()

unsigned mlir::detail::getDefaultTypeSizeInBits	(	Type	type,
		const DataLayout &	dataLayout,
		DataLayoutEntryListRef	params
	)

Default handler for the type size in bits request.

Computes results for built-in types and dispatches to the DataLayoutTypeInterface for other types.

Definition at line 51 of file DataLayoutInterfaces.cpp.

References mlir::Type::dyn_cast(), mlir::Type::getContext(), getDefaultPreferredAlignment(), getIndexBitwidth(), mlir::Type::getIntOrFloatBitWidth(), mlir::DataLayout::getTypeSize(), mlir::DataLayout::getTypeSizeInBits(), mlir::Type::isa(), and reportMissingDataLayout().

Referenced by getDefaultTypeSize(), and mlir::DataLayout::getTypeSizeInBits().

◆ getDenseElementBitWidth()

size_t mlir::detail::getDenseElementBitWidth ( Type eltType )

inline

Return the bit width which DenseElementsAttr should use for this type.

Definition at line 34 of file AttributeDetail.h.

References mlir::Type::dyn_cast(), mlir::Type::getIntOrFloatBitWidth(), and mlir::Type::isIndex().

Referenced by mlir::DenseElementsAttr::bitcast(), mlir::DenseElementsAttr::get(), getDenseElementStorageWidth(), mlir::detail::DenseIntOrFPElementsAttrStorage::getKey(), isValidIntOrFloat(), mappingHelper(), mlir::DenseElementsAttr::BoolElementIterator::operator*(), and mlir::DenseElementsAttr::IntElementIterator::operator*().

◆ getMemorySpaceAsInt()

unsigned mlir::detail::getMemorySpaceAsInt ( Attribute memorySpace )

[deprecated] Returns the memory space in old raw integer representation.

New Attribute getMemorySpace() method should be used instead.

Definition at line 529 of file BuiltinTypes.cpp.

References mlir::Attribute::cast(), mlir::emitError(), mlir::failed(), mlir::failure(), mlir::Type::getContext(), getElementType(), mlir::AffineMap::getMultiDimIdentityMap(), getShape(), mlir::Attribute::isa(), isSupportedMemorySpace(), mlir::BaseMemRefType::isValidElementType(), skipDefaultMemorySpace(), mlir::success(), mlir::verify(), and wrapIntegerMemorySpace().

Referenced by convertTransferReadToLoads(), mlir::AffineDmaStartOp::getDstMemorySpace(), mlir::BaseMemRefType::getMemorySpaceAsInt(), mlir::AffineDmaStartOp::getSrcMemorySpace(), and mlir::nvgpu::optimizeSharedMemoryReadsAndWrites().

◆ inferReturnTensorTypes()

LogicalResult mlir::detail::inferReturnTensorTypes	(	function_ref< LogicalResult(MLIRContext *, Optional< Location > location, ValueShapeRange operands, DictionaryAttr attributes, RegionRange regions, SmallVectorImpl< ShapedTypeComponents > &retComponents)>	componentTypeFn,
		MLIRContext *	context,
		Optional< Location >	location,
		ValueRange	operands,
		DictionaryAttr	attributes,
		RegionRange	regions,
		SmallVectorImpl< Type > &	inferredReturnTypes
	)

Definition at line 176 of file InferTypeOpInterface.cpp.

References mlir::failed(), mlir::failure(), and mlir::success().

Referenced by mlir::ValueShapeRange::getValues(), and mlir::OpTrait::InferTensorType< ConcreteType >::inferReturnTypes().

◆ isConstantLike()

static bool mlir::detail::isConstantLike ( Operation * op )

static

Check to see if the specified operation is ConstantLike.

This includes some quick filters to avoid a semi-expensive test in the common case.

Definition at line 52 of file Matchers.h.

References mlir::Operation::getNumOperands(), mlir::Operation::getNumResults(), and mlir::Operation::hasTrait().

Referenced by mlir::detail::constant_op_matcher::match(), and mlir::detail::constant_op_binder< AttrT >::match().

◆ isInterfaceImpl()

template<typename ConcreteType , typename ValueT , typename Traits , typename BaseType , template< typename, template< typename > class > class BaseTrait>

void mlir::detail::isInterfaceImpl ( Interface< ConcreteType, ValueT, Traits, BaseType, BaseTrait > & )

◆ isPowerOf2()

constexpr bool mlir::detail::isPowerOf2 ( int n )

Definition at line 47 of file CRunnerUtils.h.

Referenced by nextPowerOf2().

◆ isSupportedMemorySpace()

bool mlir::detail::isSupportedMemorySpace ( Attribute memorySpace )

Checks if the memorySpace has supported Attribute type.

Definition at line 497 of file BuiltinTypes.cpp.

References mlir::Attribute::getDialect(), and mlir::Attribute::isa().

Referenced by getMemorySpaceAsInt().

◆ makeStridedMemRefDescriptor() [1/2]

template<int N, typename T >

std::enable_if<(N >= 1), StridedMemRefType<T, N> >::type mlir::detail::makeStridedMemRefDescriptor	(	T *	ptr,
		T *	alignedPtr,
		ArrayRef< int64_t >	shape,
		ArrayRef< int64_t >	shapeAlloc
	)

Build a StridedMemRefDescriptor<T, N> that matches the MLIR ABI.

This is an implementation detail that is kept in sync with MLIR codegen conventions. Additionally takes a shapeAlloc array which is used instead of shape to allocate "more aligned" data and compute the corresponding strides.

Definition at line 65 of file MemRefUtils.h.

References StridedMemRefType< T, N >::basePtr, and copy().

◆ makeStridedMemRefDescriptor() [2/2]

template<int N, typename T >

std::enable_if<(N == 0), StridedMemRefType<T, 0> >::type mlir::detail::makeStridedMemRefDescriptor	(	T *	ptr,
		T *	alignedPtr,
		ArrayRef< int64_t >	shape = `{}`,
		ArrayRef< int64_t >	shapeAlloc = `{}`
	)

Build a StridedMemRefDescriptor<T, 0> that matches the MLIR ABI.

This is an implementation detail that is kept in sync with MLIR codegen conventions. Additionally takes a shapeAlloc array which is used instead of shape to allocate "more aligned" data and compute the corresponding strides.

Definition at line 86 of file MemRefUtils.h.

References StridedMemRefType< T, 0 >::basePtr.

◆ makeStrides()

template<size_t N>

std::array<int64_t, N> mlir::detail::makeStrides ( ArrayRef< int64_t > shape )

inline

Given a shape with sizes greater than 0 along all dimensions, returns the distance, in number of elements, between a slice in a dimension and the next slice in the same dimension.

e.g. shape[3, 4, 5] -> strides[20, 5, 1]

Definition at line 46 of file MemRefUtils.h.

◆ matchOperandOrValueAtIndex() [1/2]

template<typename MatcherClass >

std::enable_if_t< llvm::is_detected<detail::has_operation_or_value_matcher_t, MatcherClass, Value>::value, bool> mlir::detail::matchOperandOrValueAtIndex	(	Operation *	op,
		unsigned	idx,
		MatcherClass &	matcher
	)

Statically switch to a Value matcher.

Definition at line 185 of file Matchers.h.

References mlir::Operation::getOperand(), and value.

◆ matchOperandOrValueAtIndex() [2/2]

template<typename MatcherClass >

std::enable_if_t< llvm::is_detected<detail::has_operation_or_value_matcher_t, MatcherClass, Operation *>::value, bool> mlir::detail::matchOperandOrValueAtIndex	(	Operation *	op,
		unsigned	idx,
		MatcherClass &	matcher
	)

Statically switch to an Operation matcher.

Definition at line 195 of file Matchers.h.

References mlir::Value::getDefiningOp(), and mlir::Operation::getOperand().

◆ nextPowerOf2()

constexpr unsigned mlir::detail::nextPowerOf2 ( int n )

Definition at line 49 of file CRunnerUtils.h.

References isPowerOf2().

Referenced by allocAligned(), mlir::detail::Vector1D< T, Dim, false >::operator[](), mlir::OwningMemRef< T, Rank >::OwningMemRef(), mlir::detail::Vector1D< T, Dim, true >::Vector1D(), and mlir::detail::Vector1D< T, Dim, false >::Vector1D().

◆ parseSourceFile()

template<typename ContainerOpT , typename... ParserArgs>

OwningOpRef<ContainerOpT> mlir::detail::parseSourceFile	(	const ParserConfig &	config,
		ParserArgs &&...	args
	)

inline

The internal implementation of the templated parseSourceFile methods below, that simply forwards to the non-templated version.

Definition at line 134 of file Parser.h.

References mlir::failed(), mlir::ParserConfig::getContext(), and mlir::parseSourceFile().

◆ sameOffsetsSizesAndStrides()

bool mlir::detail::sameOffsetsSizesAndStrides	(	OffsetSizeAndStrideOpInterface	a,
		OffsetSizeAndStrideOpInterface	b,
		llvm::function_ref< bool(OpFoldResult, OpFoldResult)>	cmp
	)

Definition at line 129 of file ViewLikeInterface.cpp.

◆ skipDefaultMemorySpace()

Attribute mlir::detail::skipDefaultMemorySpace ( Attribute memorySpace )

Replaces default memorySpace (integer == 0) with empty Attribute.

Definition at line 521 of file BuiltinTypes.cpp.

References mlir::Attribute::dyn_cast_or_null().

Referenced by getMemorySpaceAsInt().

◆ verifyAffineMapAsLayout()

LogicalResult mlir::detail::verifyAffineMapAsLayout	(	AffineMap	m,
		ArrayRef< int64_t >	shape,
		function_ref< InFlightDiagnostic()>	emitError
	)

Definition at line 77 of file BuiltinAttributeInterfaces.cpp.

References mlir::emitError(), mlir::AffineMap::getNumDims(), and mlir::success().

Referenced by mlir::detail::ElementsAttrRange< IteratorT >::size().

◆ verifyBranchSuccessorOperands()

LogicalResult mlir::detail::verifyBranchSuccessorOperands	(	Operation *	op,
		unsigned	succNo,
		const SuccessorOperands &	operands
	)

Verify that the given operands match those of the given successor block.

Definition at line 61 of file ControlFlowInterfaces.cpp.

References mlir::Operation::emitError(), mlir::Block::getArgument(), mlir::Block::getNumArguments(), mlir::SuccessorOperands::getProducedOperandCount(), mlir::Operation::getSuccessor(), mlir::Value::getType(), mlir::SuccessorOperands::size(), and mlir::success().

Referenced by mlir::SuccessorOperands::getOperandIndex().

◆ verifyDataLayoutOp()

LogicalResult mlir::detail::verifyDataLayoutOp ( Operation * op )

Verifies that the operation implementing the data layout interface, or a module operation, is valid.

This calls the verifier of the spec attribute and checks if the layout is compatible with specs attached to the enclosing operations.

Definition at line 311 of file DataLayoutInterfaces.cpp.

References mlir::InFlightDiagnostic::attachNote(), collectParentLayouts(), diag(), mlir::Operation::emitError(), mlir::failed(), mlir::failure(), getCombinedDataLayout(), mlir::Operation::getLoc(), getSpec(), and mlir::success().

Referenced by mlir::DataLayoutSpecAttr::print().

◆ verifyDataLayoutSpec()

LogicalResult mlir::detail::verifyDataLayoutSpec	(	DataLayoutSpecInterface	spec,
		Location	loc
	)

Verifies that a data layout spec is valid.

This dispatches to individual entry verifiers, and then to the verifiers implemented by the relevant type and dialect interfaces for type and identifier keys respectively.

Definition at line 474 of file DataLayoutInterfaces.cpp.

References mlir::emitError(), mlir::failed(), mlir::failure(), mlir::success(), and value.

◆ verifyInferredResultTypes()

LogicalResult mlir::detail::verifyInferredResultTypes ( Operation * op )

Verifies that the inferred result types match the actual result types for the op.

Precondition: op implements InferTypeOpInterface.

Definition at line 203 of file InferTypeOpInterface.cpp.

References mlir::Operation::getAttrDictionary(), mlir::Operation::getContext(), mlir::Operation::getLoc(), mlir::Operation::getOperands(), mlir::Operation::getRegions(), and mlir::Operation::getResultTypes().

Referenced by mlir::ValueShapeRange::getValues().

◆ verifyOffsetSizeAndStrideOp()

LogicalResult mlir::detail::verifyOffsetSizeAndStrideOp ( OffsetSizeAndStrideOpInterface op )

Definition at line 38 of file ViewLikeInterface.cpp.

References mlir::failed(), mlir::failure(), mlir::success(), and mlir::verifyListOfOperandsOrIntegers().

◆ verifyParallelCombiningOpInterface()

LogicalResult mlir::detail::verifyParallelCombiningOpInterface ( Operation * op )

Definition at line 18 of file ParallelCombiningOpInterface.cpp.

References mlir::Operation::emitError(), mlir::Operation::getNumRegions(), mlir::Operation::getRegion(), mlir::Region::hasOneBlock(), and mlir::success().

◆ verifyShapedDimOpInterface()

LogicalResult mlir::detail::verifyShapedDimOpInterface ( Operation * op )

Verify invariants of ops that implement the ShapedDimOpInterface.

Definition at line 17 of file ShapedOpInterfaces.cpp.

References mlir::Operation::emitError(), mlir::Operation::getNumResults(), mlir::Operation::getResult(), mlir::Value::getType(), mlir::Type::isIndex(), and mlir::success().

◆ verifySymbol()

LogicalResult mlir::detail::verifySymbol ( Operation * op )

Definition at line 453 of file SymbolTable.cpp.

References mlir::Operation::emitOpError(), mlir::Operation::getAttr(), mlir::Operation::getAttrOfType(), mlir::SymbolTable::getSymbolAttrName(), mlir::SymbolTable::getVisibilityAttrName(), and mlir::success().

Referenced by mlir::SymbolUserMap::useEmpty().

◆ verifySymbolTable()

LogicalResult mlir::detail::verifySymbolTable ( Operation * op )

Definition at line 412 of file SymbolTable.cpp.

References mlir::WalkResult::advance(), mlir::InFlightDiagnostic::append(), mlir::Operation::emitError(), mlir::Operation::emitOpError(), mlir::Operation::getAttrOfType(), mlir::Operation::getLoc(), mlir::Operation::getNumRegions(), mlir::Operation::getRegion(), mlir::Operation::getRegions(), mlir::SymbolTable::getSymbolAttrName(), mlir::success(), and walkSymbolTable().

Referenced by mlir::SymbolUserMap::useEmpty(), and mlir::OpTrait::SymbolTable< ConcreteType >::verifyRegionTrait().

◆ verifyTypesAlongControlFlowEdges()

LogicalResult mlir::detail::verifyTypesAlongControlFlowEdges ( Operation * op )

Verify that types match along control flow edges described the given op.

Definition at line 151 of file ControlFlowInterfaces.cpp.

References mlir::failed(), mlir::failure(), mlir::Operation::getNumRegions(), and verifyTypesAlongAllEdges().

Referenced by mlir::SuccessorOperands::getOperandIndex().

◆ walk() [1/14]

void mlir::detail::walk	(	Operation *	op,
		function_ref< void(Region *)>	callback,
		WalkOrder	order
	)

Walk all of the regions, blocks, or operations nested under (and including) the given operation.

Walk all of the regions/blocks/operations nested under and including the given operation.

Regions, blocks and operations at the same nesting level are visited in lexicographical order. The walk order for enclosing regions, blocks and operations with respect to their nested ones is specified by 'order'. These methods are invoked for void-returning callbacks. A callback on a block or operation is allowed to erase that block or operation only if the walk is in post-order. See non-void method for pre-order erasure.

Definition at line 24 of file Visitors.cpp.

References mlir::Operation::getRegions(), mlir::PostOrder, and mlir::PreOrder.

Referenced by mlir::PatternApplicator::applyDefaultCostModel(), mlir::pdll::ast::Node::getLoc(), isMemRefNormalizable(), mlir::AffineExpr::operator!(), GpuAsyncRegionPass::SingleTokenUseCallback::operator()(), verifyConstantType(), mlir::pdll::ast::Node::walk(), walk(), mlir::Operation::walk(), JamBlockGatherer::walk(), mlir::PatternApplicator::walkAllPatterns(), and walkSymbolUses().

◆ walk() [2/14]

void mlir::detail::walk	(	Operation *	op,
		function_ref< void(Block *)>	callback,
		WalkOrder	order
	)

Definition at line 40 of file Visitors.cpp.

References mlir::Operation::getRegions(), mlir::PostOrder, mlir::PreOrder, and walk().

◆ walk() [3/14]

void mlir::detail::walk	(	Operation *	op,
		function_ref< void(Operation *)>	callback,
		WalkOrder	order
	)

Definition at line 55 of file Visitors.cpp.

References mlir::WalkStage::advance(), mlir::Operation::getRegions(), mlir::PostOrder, mlir::PreOrder, and walk().

◆ walk() [4/14]

WalkResult mlir::detail::walk	(	Operation *	op,
		function_ref< WalkResult(Region *)>	callback,
		WalkOrder	order
	)

Walk all of the regions, blocks, or operations nested under (and including) the given operation.

Walk all of the regions/blocks/operations nested under and including the given operation.

Regions, blocks and operations at the same nesting level are visited in lexicographical order. The walk order for enclosing regions, blocks and operations with respect to their nested ones is specified by 'order'. This method is invoked for skippable or interruptible callbacks. A callback on a block or operation is allowed to erase that block or operation if either:

the walk is in post-order, or
the walk is in pre-order and the walk is skipped after the erasure.

These functions walk operations until an interrupt result is returned by the callback. Walks on regions, blocks and operations may also be skipped if the callback returns a skip result. Regions, blocks and operations at the same nesting level are visited in lexicographical order. The walk order for enclosing regions, blocks and operations with respect to their nested ones is specified by 'order'. A callback on a block or operation is allowed to erase that block or operation if either:

the walk is in post-order, or
the walk is in pre-order and the walk is skipped after the erasure.

Definition at line 102 of file Visitors.cpp.

References mlir::WalkResult::advance(), mlir::Operation::getRegions(), mlir::WalkResult::interrupt(), mlir::PostOrder, mlir::PreOrder, walk(), mlir::WalkResult::wasInterrupted(), and mlir::WalkResult::wasSkipped().

◆ walk() [5/14]

WalkResult mlir::detail::walk	(	Operation *	op,
		function_ref< WalkResult(Block *)>	callback,
		WalkOrder	order
	)

Definition at line 130 of file Visitors.cpp.

References mlir::WalkResult::advance(), mlir::Operation::getRegions(), mlir::WalkResult::interrupt(), mlir::PostOrder, mlir::PreOrder, walk(), mlir::WalkResult::wasInterrupted(), and mlir::WalkResult::wasSkipped().

◆ walk() [6/14]

WalkResult mlir::detail::walk	(	Operation *	op,
		function_ref< WalkResult(Operation *)>	callback,
		WalkOrder	order
	)

Definition at line 157 of file Visitors.cpp.

References mlir::WalkResult::advance(), mlir::WalkStage::advance(), mlir::Operation::getRegions(), mlir::WalkResult::interrupt(), mlir::PostOrder, mlir::PreOrder, walk(), mlir::WalkResult::wasInterrupted(), and mlir::WalkResult::wasSkipped().

◆ walk() [7/14]

template<WalkOrder Order = WalkOrder::PostOrder, typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()(std::declval<ArgT>()))>

std::enable_if_t<llvm::is_one_of<ArgT, Operation , Region , Block *>::value, RetT> mlir::detail::walk	(	Operation *	op,
		FuncTy &&	callback
	)

Walk all of the regions, blocks, or operations nested under (and including) the given operation.

Regions, blocks and operations at the same nesting level are visited in lexicographical order. The walk order for enclosing regions, blocks and operations with respect to their nested ones is specified by 'Order' (post-order by default). A callback on a block or operation is allowed to erase that block or operation if either:

the walk is in post-order, or
the walk is in pre-order and the walk is skipped after the erasure. This method is selected for callbacks that operate on Region*, Block*, and Operation*.

Example: op->walk([](Region *r) { ... }); op->walk([](Block *b) { ... }); op->walk([](Operation *op) { ... });

Definition at line 169 of file Visitors.h.

References mlir::PostOrder, value, and walk().

◆ walk() [8/14]

template<WalkOrder Order = WalkOrder::PostOrder, typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()(std::declval<ArgT>()))>

std::enable_if_t< !llvm::is_one_of<ArgT, Operation , Region , Block *>::value && std::is_same<RetT, void>::value, RetT> mlir::detail::walk	(	Operation *	op,
		FuncTy &&	callback
	)

Walk all of the operations of type 'ArgT' nested under and including the given operation.

Regions, blocks and operations at the same nesting level are visited in lexicographical order. The walk order for enclosing regions, blocks and operations with respect to their nested ones is specified by 'order' (post-order by default). This method is selected for void-returning callbacks that operate on a specific derived operation type. A callback on an operation is allowed to erase that operation only if the walk is in post-order. See non-void method for pre-order erasure.

Example: op->walk([](ReturnOp op) { ... });

Definition at line 192 of file Visitors.h.

References mlir::PostOrder, value, and walk().

◆ walk() [9/14]

template<WalkOrder Order = WalkOrder::PostOrder, typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()(std::declval<ArgT>()))>

std::enable_if_t< !llvm::is_one_of<ArgT, Operation , Region , Block *>::value && std::is_same<RetT, WalkResult>::value, RetT> mlir::detail::walk	(	Operation *	op,
		FuncTy &&	callback
	)

Walk all of the operations of type 'ArgT' nested under and including the given operation.

Regions, blocks and operations at the same nesting level are visited in lexicographical order. The walk order for enclosing regions, blocks and operations with respect to their nested ones is specified by 'Order' (post-order by default). This method is selected for WalkReturn returning skippable or interruptible callbacks that operate on a specific derived operation type. A callback on an operation is allowed to erase that operation if either:

the walk is in post-order, or
the walk is in pre-order and the walk is skipped after the erasure.

Example: op->walk([](ReturnOp op) { if (some_invariant) return WalkResult::skip(); if (another_invariant) return WalkResult::interrupt(); return WalkResult::advance(); });

Definition at line 227 of file Visitors.h.

References mlir::WalkResult::advance(), value, walk(), and mlir::WalkResult::WalkResult().

◆ walk() [10/14]

void mlir::detail::walk	(	Operation *	op,
		function_ref< void(Operation *, const WalkStage &stage)>	callback
	)

Generic walkers with stage aware callbacks.

Walk all the operations nested under (and including) the given operation, with the callback being invoked on each operation N+1 times, where N is the number of regions attached to the operation. The stage input to the callback indicates the current walk stage. This method is invoked for void returning callbacks.

◆ walk() [11/14]

WalkResult mlir::detail::walk	(	Operation *	op,
		function_ref< WalkResult(Operation *, const WalkStage &stage)>	callback
	)

Walk all the operations nested under (and including) the given operation, with the callback being invoked on each operation N+1 times, where N is the number of regions attached to the operation.

The stage input to the callback indicates the current walk stage. This method is invoked for skippable or interruptible callbacks.

◆ walk() [12/14]

template<typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()( std::declval<ArgT>(), std::declval<const WalkStage &>()))>

std::enable_if_t<std::is_same<ArgT, Operation *>::value, RetT> mlir::detail::walk	(	Operation *	op,
		FuncTy &&	callback
	)

Walk all of the operations nested under and including the given operation.

This method is selected for stage-aware callbacks that operate on Operation*.

Example: op->walk([](Operation *op, const WalkStage &stage) { ... });

Definition at line 265 of file Visitors.h.

References value, and walk().

◆ walk() [13/14]

template<typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()( std::declval<ArgT>(), std::declval<const WalkStage &>()))>

std::enable_if_t<!std::is_same<ArgT, Operation *>::value && std::is_same<RetT, void>::value, RetT> mlir::detail::walk	(	Operation *	op,
		FuncTy &&	callback
	)

Walk all of the operations of type 'ArgT' nested under and including the given operation.

This method is selected for void returning callbacks that operate on a specific derived operation type.

Example: op->walk([](ReturnOp op, const WalkStage &stage) { ... });

Definition at line 282 of file Visitors.h.

References value, and walk().

◆ walk() [14/14]

template<typename FuncTy , typename ArgT = detail::first_argument<FuncTy>, typename RetT = decltype(std::declval<FuncTy>()( std::declval<ArgT>(), std::declval<const WalkStage &>()))>

std::enable_if_t<!std::is_same<ArgT, Operation *>::value && std::is_same<RetT, WalkResult>::value, RetT> mlir::detail::walk	(	Operation *	op,
		FuncTy &&	callback
	)

Walk all of the operations of type 'ArgT' nested under and including the given operation.

This method is selected for WalkReturn returning interruptible callbacks that operate on a specific derived operation type.

Example: op->walk(op, [](ReturnOp op, const WalkStage &stage) { if (some_invariant) return WalkResult::interrupt(); return WalkResult::advance(); });

Definition at line 307 of file Visitors.h.

References mlir::WalkResult::advance(), and walk().

Referenced by walk().

◆ wrapIntegerMemorySpace()

Attribute mlir::detail::wrapIntegerMemorySpace	(	unsigned	memorySpace,
		MLIRContext *	ctx
	)

Wraps deprecated integer memory space to the new Attribute form.

Definition at line 513 of file BuiltinTypes.cpp.

Referenced by getMemorySpaceAsInt().

Variable Documentation

◆ first_argument_type

template<typename F >

decltype(first_argument_type(&F::operator())) mlir::detail::first_argument_type(F)

Definition at line 108 of file Visitors.h.

Namespaces

Classes

Typedefs

Functions

Variables

Typedef Documentation

◆ ByteCodeAddr

◆ ByteCodeField

◆ count_if_t

◆ DenseIterPtrAndSplat

◆ first_argument

◆ has_impltype_getkey_t

◆ has_impltype_hash_t

◆ has_operation_or_value_matcher_t

◆ is_interface_t

◆ IsInterface

◆ OwningOpRange

◆ walkResultType

Function Documentation

◆ allocAligned()

◆ bindDims() [1/2]

◆ bindDims() [2/2]

◆ bindSymbols() [1/2]

◆ bindSymbols() [2/2]

◆ bindSymbolsList()

◆ constructContainerOpForParserIfNecessary()

◆ enumerate()

◆ enumerateImpl()

◆ filterEntriesForType()

◆ filterEntryForIdentifier()

◆ first_argument_type() [1/2]

◆ first_argument_type() [2/2]

◆ getBranchSuccessorArgument()

◆ getDefaultABIAlignment()

◆ getDefaultDiagnosticEmitFn() [1/2]

◆ getDefaultDiagnosticEmitFn() [2/2]

◆ getDefaultPreferredAlignment()

◆ getDefaultTypeSize()

◆ getDefaultTypeSizeInBits()

◆ getDenseElementBitWidth()

◆ getMemorySpaceAsInt()

◆ inferReturnTensorTypes()

◆ isConstantLike()

◆ isInterfaceImpl()

◆ isPowerOf2()

◆ isSupportedMemorySpace()

◆ makeStridedMemRefDescriptor() [1/2]

◆ makeStridedMemRefDescriptor() [2/2]

◆ makeStrides()

◆ matchOperandOrValueAtIndex() [1/2]

◆ matchOperandOrValueAtIndex() [2/2]

◆ nextPowerOf2()

◆ parseSourceFile()

◆ sameOffsetsSizesAndStrides()

◆ skipDefaultMemorySpace()

◆ verifyAffineMapAsLayout()

◆ verifyBranchSuccessorOperands()

◆ verifyDataLayoutOp()

◆ verifyDataLayoutSpec()

◆ verifyInferredResultTypes()

◆ verifyOffsetSizeAndStrideOp()

◆ verifyParallelCombiningOpInterface()

◆ verifyShapedDimOpInterface()

◆ verifySymbol()

◆ verifySymbolTable()

◆ verifyTypesAlongControlFlowEdges()

◆ walk() [1/14]

◆ walk() [2/14]

◆ walk() [3/14]

◆ walk() [4/14]

◆ walk() [5/14]

◆ walk() [6/14]

◆ walk() [7/14]

◆ walk() [8/14]

◆ walk() [9/14]

◆ walk() [10/14]

◆ walk() [11/14]

◆ walk() [12/14]

◆ walk() [13/14]

◆ walk() [14/14]