llvm::IRBuilderBase

Synopsis

Declared in <llvm/IR/IRBuilder.h>

class IRBuilderBase;

Types

Name	Description
`FastMathFlagGuard`	RAII guard that restores fast-math builder settings.
`InsertPoint`	InsertPoint - A saved insertion point.
`InsertPointGuard`	RAII guard that restores the builder insert point.
`OperandBundlesGuard`	RAII guard that restores default operand bundles.

Member Functions

Name	Description
`IRBuilderBase` [constructor]	Construct an IRBuilderBase.
`ClearInsertionPoint`	Clear the insertion point: created instructions will not be inserted into a block.
`CreateAShr`	Create an a shr instruction.
`CreateAdd`	Create an add instruction.
`CreateAddReduce`	Create a vector int add reduction intrinsic of the source vector.
`CreateAddrSpaceCast`	Create an addr space cast.
`CreateAggregateCast`	Cast between aggregates with matching structure.
`CreateAggregateRet`	Create an aggregate return from multiple values.
`CreateAlignedLoad`	Create an aligned load.
`CreateAlignedStore`	Create an aligned store.
`CreateAlignmentAssumption`	Create an assume intrinsic call that represents an alignment assumption on the provided pointer.
`CreateAlloca`	Create an alloca instruction.
`CreateAllocationSize`	Get allocation size of an alloca as a runtime Value* (handles both static and dynamic allocas and vscale factor).
`CreateAnd`	Create an and instruction.
`CreateAndReduce`	Create a vector int AND reduction intrinsic of the source vector.
`CreateArithmeticFence`	Create a call to the arithmetic_fence intrinsic.
`CreateAssumption`	`CreateAssumption` overloads
`CreateAtomicCmpXchg`	Create an atomic cmpxchg instruction.
`CreateAtomicRMW`	Create an atomicrmw instruction.
`CreateBinOp`	Create a bin op.
`CreateBinOpFMF`	Create a binop with fast-math flags.
`CreateBinaryIntrinsic`	Create a call to intrinsic `ID` with 2 operands which is mangled on the first type.
`CreateBitCast`	Create a bit cast instruction.
`CreateBitOrPointerCast`	Create a bit or pointer cast.
`CreateBitPreservingCastChain`	Create casts that preserve the bit pattern of a value.
`CreateBr`	Create an unconditional 'br label X' instruction.
`CreateCall`	Create a call instruction.
`CreateCallBr`	`CreateCallBr` overloads
`CreateCast`	Create a cast instruction.
`CreateCatchPad`	Create a catch pad.
`CreateCatchRet`	Create a catch ret.
`CreateCatchSwitch`	Create a catch switch.
`CreateCleanupPad`	Create a cleanup pad.
`CreateCleanupRet`	Create a cleanup ret.
`CreateCmp`	Create a cmp.
`CreateCondBr`	`CreateCondBr` overloads
`CreateConstGEP1_32`	Create a const gep1_32.
`CreateConstGEP1_64`	Create a const gep1_64.
`CreateConstGEP2_32`	Create a const gep2_32.
`CreateConstGEP2_64`	Create a const gep2_64.
`CreateConstInBoundsGEP1_32`	Create a const in bounds gep1_32.
`CreateConstInBoundsGEP1_64`	Create a const in bounds gep1_64.
`CreateConstInBoundsGEP2_32`	Create a const in bounds gep2_32.
`CreateConstInBoundsGEP2_64`	Create a const in bounds gep2_64.
`CreateConstrainedFPBinOp`	Create a constrained floating-point binop.
`CreateConstrainedFPCall`	Create a constrained fp call.
`CreateConstrainedFPCast`	Create a constrained fp cast.
`CreateConstrainedFPCmp`	Create a constrained fp cmp.
`CreateConstrainedFPIntrinsic`	Create a constrained floating-point intrinsic call.
`CreateConstrainedFPUnroundedBinOp`	Create a constrained FP binop without rounding metadata.
`CreateCopySign`	Create call to the copysign intrinsic.
`CreateCountTrailingZeroElems`	Create a call to llvm.experimental_cttz_elts
`CreateDereferenceableAssumption`	Create an assume intrinsic call that represents a dereferencable assumption on the provided pointer.
`CreateDisjointOr`	Create a disjoint or.
`CreateElementCount`	Create an expression which evaluates to the number of elements in `EC` at runtime. This can result in poison if type `Ty` is not big enough to hold the value.
`CreateElementUnorderedAtomicMemCpy`	Create and insert an element unordered-atomic memcpy between the specified pointers.
`CreateElementUnorderedAtomicMemMove`	Create and insert an element unordered-atomic memmove between the specified pointers.
`CreateElementUnorderedAtomicMemSet`	`CreateElementUnorderedAtomicMemSet` overloads
`CreateExactBinOp`	Create an exact binop instruction.
`CreateExactSDiv`	Create an exact sdiv instruction.
`CreateExactUDiv`	Create an exact udiv instruction.
`CreateExtractElement`	Create an extract element.
`CreateExtractValue`	Create an extract value.
`CreateExtractVector`	Create a call to the vector.extract intrinsic.
`CreateFAbs`	Create call to the fabs intrinsic.
`CreateFAdd`	Create a f add instruction.
`CreateFAddFMF`	Create a fadd with fast-math flags.
`CreateFAddReduce`	Create a sequential vector fadd reduction intrinsic.
`CreateFCmp`	Create a f cmp.
`CreateFCmpFMF`	Create an fcmp fmf instruction.
`CreateFCmpOEQ`	Create an fcmp oeq instruction.
`CreateFCmpOGE`	Create an fcmp oge instruction.
`CreateFCmpOGT`	Create an fcmp ogt instruction.
`CreateFCmpOLE`	Create an fcmp ole instruction.
`CreateFCmpOLT`	Create an fcmp olt instruction.
`CreateFCmpONE`	Create an fcmp one instruction.
`CreateFCmpORD`	Create an fcmp ord instruction.
`CreateFCmpS`	Create an fcmp s instruction.
`CreateFCmpUEQ`	Create an fcmp ueq instruction.
`CreateFCmpUGE`	Create an fcmp uge instruction.
`CreateFCmpUGT`	Create an fcmp ugt instruction.
`CreateFCmpULE`	Create an fcmp ule instruction.
`CreateFCmpULT`	Create an fcmp ult instruction.
`CreateFCmpUNE`	Create an fcmp une instruction.
`CreateFCmpUNO`	Create an fcmp uno instruction.
`CreateFDiv`	Create a f div instruction.
`CreateFDivFMF`	Create a fdiv with fast-math flags.
`CreateFMA`	Create call to the fma intrinsic.
`CreateFMul`	Create a f mul instruction.
`CreateFMulFMF`	Create a fmul with fast-math flags.
`CreateFMulReduce`	Create a sequential vector fmul reduction intrinsic.
`CreateFNeg`	Create a f neg.
`CreateFNegFMF`	Create a fneg with fast-math flags.
`CreateFPCast`	Create a fp cast.
`CreateFPExt`	Create a fp ext.
`CreateFPExtFMF`	Create a fpext with fast-math flags.
`CreateFPMaxReduce`	Create a vector float max reduction intrinsic of the source vector.
`CreateFPMaximumNumReduce`	Create a vector float maximum reduction intrinsic of the source vector. This variant follows the NaN and signed zero semantic of llvm.maximumnum intrinsic.
`CreateFPMaximumReduce`	Create a vector float maximum reduction intrinsic of the source vector. This variant follows the NaN and signed zero semantic of llvm.maximum intrinsic.
`CreateFPMinReduce`	Create a vector float min reduction intrinsic of the source vector.
`CreateFPMinimumNumReduce`	Create a vector float minimum reduction intrinsic of the source vector. This variant follows the NaN and signed zero semantic of llvm.minimumnum intrinsic.
`CreateFPMinimumReduce`	Create a vector float minimum reduction intrinsic of the source vector. This variant follows the NaN and signed zero semantic of llvm.minimum intrinsic.
`CreateFPToSI`	Create a fp to si.
`CreateFPToUI`	Create a fp to ui.
`CreateFPTrunc`	Create a fp trunc.
`CreateFPTruncFMF`	Create a fptrunc with fast-math flags.
`CreateFRem`	Create a f rem instruction.
`CreateFRemFMF`	Create a frem with fast-math flags.
`CreateFSub`	Create a f sub instruction.
`CreateFSubFMF`	Create a fsub with fast-math flags.
`CreateFence`	Create a fence instruction.
`CreateFree`	Generate the IR for a call to the builtin free function.
`CreateFreeze`	Create a freeze instruction.
`CreateGCGetPointerBase`	Create a call to the experimental.gc.pointer.base intrinsic to get the base pointer for the specified derived pointer.
`CreateGCGetPointerOffset`	Create a call to the experimental.gc.get.pointer.offset intrinsic to get the offset of the specified derived pointer from its base.
`CreateGCRelocate`	Create a call to the experimental.gc.relocate intrinsics to project the relocated value of one pointer from the statepoint.
`CreateGCResult`	Create a call to the experimental.gc.result intrinsic to extract the result from a call wrapped in a statepoint.
`CreateGCStatepointCall`	`CreateGCStatepointCall` overloads
`CreateGCStatepointInvoke`	`CreateGCStatepointInvoke` overloads
`CreateGEP`	Create a gep instruction.
`CreateGlobalString`	Make a new global variable with initializer type i8*
`CreateICmp`	Create an i cmp.
`CreateICmpEQ`	Create an icmp eq instruction.
`CreateICmpNE`	Create an icmp ne instruction.
`CreateICmpSGE`	Create an icmp sge instruction.
`CreateICmpSGT`	Create an icmp sgt instruction.
`CreateICmpSLE`	Create an icmp sle instruction.
`CreateICmpSLT`	Create an icmp slt instruction.
`CreateICmpUGE`	Create an icmp uge instruction.
`CreateICmpUGT`	Create an icmp ugt instruction.
`CreateICmpULE`	Create an icmp ule instruction.
`CreateICmpULT`	Create an icmp ult instruction.
`CreateInBoundsGEP`	Create an in bounds gep.
`CreateInBoundsPtrAdd`	Create an in bounds ptr add.
`CreateIndirectBr`	Create an indirect branch instruction.
`CreateInsertElement`	Create an insert element.
`CreateInsertValue`	Create an insert value.
`CreateInsertVector`	`CreateInsertVector` overloads
`CreateIntCast`	Create an int cast.
`CreateIntMaxReduce`	Create a vector integer max reduction intrinsic of the source vector.
`CreateIntMinReduce`	Create a vector integer min reduction intrinsic of the source vector.
`CreateIntToPtr`	Create an int to ptr.
`CreateIntrinsic`	Create a possibly constant-folded intrinsic call.
`CreateIntrinsicWithoutFolding`	Create an intrinsic call that is never constant-folded.
`CreateInvariantStart`	Create a call to invariant.start intrinsic.
`CreateInvoke`	Create an invoke instruction.
`CreateIsNeg`	Return a boolean value testing if `Arg` < 0.
`CreateIsNotNeg`	Return a boolean value testing if `Arg` > -1.
`CreateIsNotNull`	Return a boolean value testing if `Arg` != 0.
`CreateIsNull`	Return a boolean value testing if `Arg` == 0.
`CreateLShr`	Create an l shr instruction.
`CreateLandingPad`	Create a landing pad.
`CreateLaunderInvariantGroup`	Create a launder.invariant.group intrinsic call.
`CreateLdexp`	Create call to the ldexp intrinsic.
`CreateLifetimeEnd`	Create a lifetime.end intrinsic.
`CreateLifetimeStart`	Create a lifetime.start intrinsic.
`CreateLoad`	`CreateLoad` overloads
`CreateLogicalAnd`	Create a logical and.
`CreateLogicalOp`	Create a logical op.
`CreateLogicalOr`	Create a logical or.
`CreateMalloc`	Generate IR for a call to malloc.
`CreateMaskedCompressStore`	Create a call to Masked Compress Store intrinsic
`CreateMaskedExpandLoad`	Create a call to Masked Expand Load intrinsic
`CreateMaskedGather`	Create a call to Masked Gather intrinsic
`CreateMaskedLoad`	Create a call to Masked Load intrinsic
`CreateMaskedScatter`	Create a call to Masked Scatter intrinsic
`CreateMaskedStore`	Create a call to Masked Store intrinsic
`CreateMaxNum`	Create call to the maxnum intrinsic.
`CreateMaximum`	Create call to the maximum intrinsic.
`CreateMaximumNum`	Create call to the maximum intrinsic.
`CreateMemCpy`	`CreateMemCpy` overloads
`CreateMemCpyInline`	Create a mem cpy inline.
`CreateMemMove`	Create a memmove intrinsic.
`CreateMemSet`	`CreateMemSet` overloads
`CreateMemSetInline`	Create an inline memset intrinsic.
`CreateMemTransferInst`	Create a memory-transfer intrinsic.
`CreateMinNum`	Create call to the minnum intrinsic.
`CreateMinimum`	Create call to the minimum intrinsic.
`CreateMinimumNum`	Create call to the minimumnum intrinsic.
`CreateMul`	Create a mul instruction.
`CreateMulReduce`	Create a vector int mul reduction intrinsic of the source vector.
`CreateNAryOp`	Create either a UnaryOperator or BinaryOperator depending on `Opc.` Correct number of operands must be passed accordingly.
`CreateNSWAdd`	Create an nsw add instruction.
`CreateNSWMul`	Create an nsw mul instruction.
`CreateNSWNeg`	Create an nsw neg instruction.
`CreateNSWSub`	Create an nsw sub instruction.
`CreateNUWAdd`	Create an nuw add instruction.
`CreateNUWMul`	Create an nuw mul instruction.
`CreateNUWSub`	Create an nuw sub instruction.
`CreateNeg`	Create a neg instruction.
`CreateNoAliasScopeDeclaration`	`CreateNoAliasScopeDeclaration` overloads
`CreateNoWrapBinOp`	Create a no wrap bin op.
`CreateNonnullAssumption`	Create an assume intrinsic call that represents a nonnull assumption on the provided pointer.
`CreateNot`	Create a not instruction.
`CreateOr`	Create an or instruction.
`CreateOrReduce`	Create a vector int OR reduction intrinsic of the source vector.
`CreatePHI`	Create a phi instruction.
`CreatePointerBitCastOrAddrSpaceCast`	Create a pointer bit cast or addr space cast.
`CreatePointerCast`	Create a pointer cast.
`CreatePreserveArrayAccessIndex`	Create a preserve array access index.
`CreatePreserveStructAccessIndex`	Create a preserve struct access index.
`CreatePreserveUnionAccessIndex`	Create a preserve union access index.
`CreatePtrAdd`	Create a ptr add.
`CreatePtrDiff`	`CreatePtrDiff` overloads
`CreatePtrToAddr`	Create a ptr to addr.
`CreatePtrToInt`	Create a ptr to int.
`CreateResume`	Create a resume.
`CreateRet`	Create a 'ret <val>' instruction.
`CreateRetVoid`	Create a 'ret void' instruction.
`CreateSDiv`	Create a s div instruction.
`CreateSExt`	Create a s ext instruction.
`CreateSExtOrBitCast`	Create a s ext or bit cast.
`CreateSExtOrTrunc`	Create a SExt or Trunc from the integer value V to DestTy. Return the value untouched if the type of V is already DestTy.
`CreateSIToFP`	Create a si to fp.
`CreateSRem`	Create a s rem instruction.
`CreateSelect`	Create a select instruction.
`CreateSelectFMF`	Create a select with fast-math flags.
`CreateSelectFMFWithUnknownProfile`	Create a select fmf with unknown profile.
`CreateSelectWithUnknownProfile`	Create a select with unknown profile.
`CreateShl`	Create a shl instruction.
`CreateShuffleVector`	`CreateShuffleVector` overloads
`CreateStackRestore`	Create a call to llvm.stackrestore
`CreateStackSave`	Create a call to llvm.stacksave
`CreateStepVector`	Creates a vector of type `DstType` with the linear sequence <0, 1, ...>
`CreateStore`	Create a store instruction.
`CreateStripInvariantGroup`	Create a strip.invariant.group intrinsic call.
`CreateStructGEP`	Create a struct gep.
`CreateStructuredAlloca`	Create a structured alloca.
`CreateStructuredGEP`	Create a structured gep.
`CreateSub`	Create a sub instruction.
`CreateSwitch`	Create a switch instruction with the specified value, default dest, and with a hint for the number of cases that will be added (for efficient allocation).
`CreateThreadLocalAddress`	Create a call to llvm.threadlocal.address intrinsic.
`CreateTrunc`	Create a trunc instruction.
`CreateTruncOrBitCast`	Create a trunc or bit cast.
`CreateTypeSize`	Create an expression for the runtime size of a type.
`CreateUDiv`	Create an u div instruction.
`CreateUIToFP`	Create an ui to fp.
`CreateURem`	Create an u rem instruction.
`CreateUnOp`	Create an un op.
`CreateUnaryIntrinsic`	Create a call to intrinsic `ID` with 1 operand which is mangled on its type.
`CreateUnreachable`	Create an unreachable.
`CreateVAArg`	Create a va arg instruction.
`CreateVScale`	Create a call to llvm.vscale.<Ty>().
`CreateVectorInterleave`	Create a vector interleave.
`CreateVectorReverse`	Return a vector value that contains the vector V reversed
`CreateVectorSplat`	Return a vector value that contains
`CreateVectorSpliceLeft`	Create a vector.splice.left operation.
`CreateVectorSpliceRight`	Create a vector.splice.right operation.
`CreateXor`	Create a xor instruction.
`CreateXorReduce`	Create a vector int XOR reduction intrinsic of the source vector.
`CreateZExt`	Create a z ext instruction.
`CreateZExtOrBitCast`	Create a z ext or bit cast.
`CreateZExtOrTrunc`	Create a ZExt or Trunc from the integer value V to DestTy. Return the value untouched if the type of V is already DestTy.
`GetInsertBlock`	Return the basic block of the current insert point.
`GetInsertPoint`	Return the iterator of the current insert point.
`Insert`	`Insert` overloads
`SetCurrentDebugLocation`	Set location information used by debugging information.
`SetInsertPoint`	`SetInsertPoint` overloads
`SetInsertPointPastAllocas`	Set the insert point past entry-block static allocas.
`SetInstDebugLocation`	If this builder has a current debug location, set it on the specified instruction.
`clearFastMathFlags`	Clear the fast-math flags.
`createIsFPClass`	Create an is.fpclass intrinsic call.
`getAllOnesMask`	Return an all true boolean vector (mask) with `NumElts` lanes.
`getBFloatTy`	Fetch the type representing a 16-bit brain floating point value.
`getByte128Ty`	Fetch the type representing a 128-bit byte.
`getByte16Ty`	Fetch the type representing a 16-bit byte.
`getByte32Ty`	Fetch the type representing a 32-bit byte.
`getByte64Ty`	Fetch the type representing a 64-bit byte.
`getByte8Ty`	Fetch the type representing an 8-bit byte.
`getByteNTy`	Fetch the type representing an N-bit byte.
`getBytePtrTy`	Fetch the type of a byte with size at least as big as that of a pointer in the given address space.
`getContext`	Return the LLVM context used by this builder.
`getCurrentDebugLocation`	Get location information used by debugging information.
`getCurrentFunctionReturnType`	Get the return type of the current function that we're emitting into.
`getDefaultConstrainedExcept`	Get the exception handling used with constrained floating point
`getDefaultConstrainedRounding`	Get the rounding mode handling used with constrained floating point
`getDefaultFPMathTag`	Get the floating point math metadata being used.
`getDoubleTy`	Fetch the type representing a 64-bit floating point value.
`getFalse`	Get the constant value for i1 false.
`getFastMathFlags`	`getFastMathFlags` overloads
`getFloatTy`	Fetch the type representing a 32-bit floating point value.
`getHalfTy`	Fetch the type representing a 16-bit floating point value.
`getIndexTy`	Fetch the type of an integer that should be used to index GEP operations within AddressSpace.
`getInt`	Get a constant integer value.
`getInt1`	Get a constant value representing either true or false.
`getInt128Ty`	Fetch the type representing a 128-bit integer.
`getInt16`	Get a constant 16-bit value.
`getInt16Ty`	Fetch the type representing a 16-bit integer.
`getInt1Ty`	Fetch the type representing a single bit
`getInt32`	Get a constant 32-bit value.
`getInt32Ty`	Fetch the type representing a 32-bit integer.
`getInt64`	Get a constant 64-bit value.
`getInt64Ty`	Fetch the type representing a 64-bit integer.
`getInt8`	Get a constant 8-bit value.
`getInt8Ty`	Fetch the type representing an 8-bit integer.
`getIntN`	Get a constant N-bit value, zero extended from a 64-bit value.
`getIntNTy`	Fetch the type representing an N-bit integer.
`getIntPtrTy`	Fetch the type of an integer with size at least as big as that of a pointer in the given address space.
`getIsFPConstrained`	Query for the use of constrained floating point math
`getPtrTy`	Fetch the type representing a pointer.
`getTrue`	Get the constant value for i1 true.
`getVoidTy`	Fetch the type representing void.
`restoreIP`	Sets the current insert point to a previously-saved location.
`saveAndClearIP`	Returns the current insert point, clearing it in the process.
`saveIP`	Returns the current insert point.
`setConstrainedFPCallAttr`	Add the StrictFP attribute to a call instruction.
`setConstrainedFPFunctionAttr`	Add the StrictFP attribute to the current function.
`setDefaultConstrainedExcept`	Set the exception handling to be used with constrained floating point
`setDefaultConstrainedRounding`	Set the rounding mode handling to be used with constrained floating point
`setDefaultFPMathTag`	Set the floating point math metadata to be used.
`setDefaultOperandBundles`	Set the default operand bundles for new calls.
`setFastMathFlags`	Set the fast-math flags to be used with generated fp-math operators
`setIsFPConstrained`	Enable or disable constrained floating-point math.

Protected Data Members

Name	Description
`BB`	Basic block used as the current insert point.
`Context`	LLVM context used to create IR objects.
`DefaultConstrainedExcept`	Default exception behavior for constrained FP.
`DefaultConstrainedRounding`	Default rounding mode for constrained FP.
`DefaultFPMathTag`	Default floating-point math metadata tag.
`DefaultOperandBundles`	Default operand bundles attached to new calls.
`FMF`	Fast-math flags applied to created floating-point ops.
`Folder`	Constant folder used when creating instructions.
`InsertPt`	Iterator within BB where new instructions are inserted.
`Inserter`	Helper that inserts newly created instructions.
`IsFPConstrained`	Whether CreateF* builds constrained FP intrinsics.

Non-Member Functions

Name	Description
`buildCmpXchgValue`	Emit IR to implement the given cmpxchg operation on values in registers, returning the new value.
`concatenateVectors`	Concatenate a list of vectors.
`createAnyOfReduction`	Create a reduction of the given vector `Src` for a reduction of kind RecurKind::AnyOf. The start value of the reduction is `InitVal.`
`createBitMaskForGaps`	Create a mask that filters the members of an interleave group where there are gaps.
`createMinMaxOp`	Returns a Min/Max operation corresponding to MinMaxRecurrenceKind. The Builder's fast-math-flags must be set to propagate the expected values.
`createOrderedReduction`	Create an ordered reduction intrinsic using the given recurrence kind `RdxKind.`
`createOrderedReduction`	Overloaded function to generate vector-predication intrinsics for ordered reduction.
`createSimpleReduction`	Overloaded function to generate vector-predication intrinsics for reduction.
`createSimpleReduction`	Create a vector reduction for recurrence kind `RdxKind.`
`emitGEPOffset`	Emit code to compute a GEP's offset from its base pointer.
`expandReductionViaLoop`	Expand a scalable vector reduction into an element-wise runtime loop.
`getOrderedReduction`	Generates an ordered vector reduction using extracts to reduce the value.
`getShuffleReduction`	Generates a vector reduction using shufflevectors to reduce the value. Fast-math-flags are propagated using the IRBuilder's setting.
`splitBB`	Split a basic block at an IRBuilder insertion point.
`splitBBWithSuffix`	Split a block and suffix the current block name for the new successor.

Derived Classes

Name	Description
`IRBuilder`	Uniform API for creating and inserting LLVM IR instructions.