LLVM API Documentation
#include "llvm/Transforms/Scalar.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/Dominators.h"
#include "llvm/Analysis/Loads.h"
#include "llvm/Analysis/PtrUseVisitor.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/DIBuilder.h"
#include "llvm/DebugInfo.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/Operator.h"
#include "llvm/InstVisitor.h"
#include "llvm/Pass.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Compiler.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Transforms/Utils/PromoteMemToReg.h"
#include "llvm/Transforms/Utils/SSAUpdater.h"
Go to the source code of this file.
Classes | |
struct | llvm::isPodLike< T > |
struct | llvm::isPodLike< Slice > |
class | AllocaSlices::SliceBuilder |
Builder for the alloca slices. More... | |
Namespaces | |
llvm | |
List of target independent CodeGen pass IDs. | |
Macros | |
#define | DEBUG_TYPE "sroa" |
Functions | |
STATISTIC (NumAllocasAnalyzed,"Number of allocas analyzed for replacement") | |
STATISTIC (NumAllocaPartitions,"Number of alloca partitions formed") | |
STATISTIC (MaxPartitionsPerAlloca,"Maximum number of partitions per alloca") | |
STATISTIC (NumAllocaPartitionUses,"Number of alloca partition uses rewritten") | |
STATISTIC (MaxUsesPerAllocaPartition,"Maximum number of uses of a partition") | |
STATISTIC (NumNewAllocas,"Number of new, smaller allocas introduced") | |
STATISTIC (NumPromoted,"Number of allocas promoted to SSA values") | |
STATISTIC (NumLoadsSpeculated,"Number of loads speculated to allow promotion") | |
STATISTIC (NumDeleted,"Number of instructions deleted") | |
STATISTIC (NumVectorized,"Number of vectorized aggregates") | |
static Value * | foldSelectInst (SelectInst &SI) |
INITIALIZE_PASS_BEGIN (SROA,"sroa","Scalar Replacement Of Aggregates", false, false) INITIALIZE_PASS_END(SROA | |
Scalar Replacement Of static false Type * | findCommonType (AllocaSlices::const_iterator B, AllocaSlices::const_iterator E, uint64_t EndOffset) |
static bool | isSafePHIToSpeculate (PHINode &PN, const DataLayout *DL=0) |
static void | speculatePHINodeLoads (PHINode &PN) |
static bool | isSafeSelectToSpeculate (SelectInst &SI, const DataLayout *DL=0) |
static void | speculateSelectInstLoads (SelectInst &SI) |
static Value * | buildGEP (IRBuilderTy &IRB, Value *BasePtr, SmallVectorImpl< Value * > &Indices) |
Build a GEP out of a base pointer and indices. More... | |
static Value * | getNaturalGEPWithType (IRBuilderTy &IRB, const DataLayout &DL, Value *BasePtr, Type *Ty, Type *TargetTy, SmallVectorImpl< Value * > &Indices) |
Get a natural GEP off of the BasePtr walking through Ty toward TargetTy without changing the offset of the pointer. More... | |
static Value * | getNaturalGEPRecursively (IRBuilderTy &IRB, const DataLayout &DL, Value *Ptr, Type *Ty, APInt &Offset, Type *TargetTy, SmallVectorImpl< Value * > &Indices) |
Recursively compute indices for a natural GEP. More... | |
static Value * | getNaturalGEPWithOffset (IRBuilderTy &IRB, const DataLayout &DL, Value *Ptr, APInt Offset, Type *TargetTy, SmallVectorImpl< Value * > &Indices) |
Get a natural GEP from a base pointer to a particular offset and resulting in a particular type. More... | |
static Value * | getAdjustedPtr (IRBuilderTy &IRB, const DataLayout &DL, Value *Ptr, APInt Offset, Type *PointerTy) |
Compute an adjusted pointer from Ptr by Offset bytes where the resulting pointer has PointerTy. More... | |
static bool | canConvertValue (const DataLayout &DL, Type *OldTy, Type *NewTy) |
Test whether we can convert a value from the old to the new type. More... | |
static Value * | convertValue (const DataLayout &DL, IRBuilderTy &IRB, Value *V, Type *NewTy) |
Generic routine to convert an SSA value to a value of a different type. More... | |
static bool | isVectorPromotionViableForSlice (const DataLayout &DL, AllocaSlices &S, uint64_t SliceBeginOffset, uint64_t SliceEndOffset, VectorType *Ty, uint64_t ElementSize, AllocaSlices::const_iterator I) |
Test whether the given slice use can be promoted to a vector. More... | |
static bool | isVectorPromotionViable (const DataLayout &DL, Type *AllocaTy, AllocaSlices &S, uint64_t SliceBeginOffset, uint64_t SliceEndOffset, AllocaSlices::const_iterator I, AllocaSlices::const_iterator E, ArrayRef< AllocaSlices::iterator > SplitUses) |
Test whether the given alloca partitioning and range of slices can be promoted to a vector. More... | |
static bool | isIntegerWideningViableForSlice (const DataLayout &DL, Type *AllocaTy, uint64_t AllocBeginOffset, uint64_t Size, AllocaSlices &S, AllocaSlices::const_iterator I, bool &WholeAllocaOp) |
Test whether a slice of an alloca is valid for integer widening. More... | |
static bool | isIntegerWideningViable (const DataLayout &DL, Type *AllocaTy, uint64_t AllocBeginOffset, AllocaSlices &S, AllocaSlices::const_iterator I, AllocaSlices::const_iterator E, ArrayRef< AllocaSlices::iterator > SplitUses) |
Test whether the given alloca partition's integer operations can be widened to promotable ones. More... | |
static Value * | extractInteger (const DataLayout &DL, IRBuilderTy &IRB, Value *V, IntegerType *Ty, uint64_t Offset, const Twine &Name) |
static Value * | insertInteger (const DataLayout &DL, IRBuilderTy &IRB, Value *Old, Value *V, uint64_t Offset, const Twine &Name) |
static Value * | extractVector (IRBuilderTy &IRB, Value *V, unsigned BeginIndex, unsigned EndIndex, const Twine &Name) |
static Value * | insertVector (IRBuilderTy &IRB, Value *Old, Value *V, unsigned BeginIndex, const Twine &Name) |
static Type * | stripAggregateTypeWrapping (const DataLayout &DL, Type *Ty) |
Strip aggregate type wrapping. More... | |
static Type * | getTypePartition (const DataLayout &DL, Type *Ty, uint64_t Offset, uint64_t Size) |
Try to find a partition of the aggregate type passed in for a given offset and size. More... | |
static void | removeFinishedSplitUses (SmallVectorImpl< AllocaSlices::iterator > &SplitUses, uint64_t &MaxSplitUseEndOffset, uint64_t Offset) |
static void | enqueueUsersInWorklist (Instruction &I, SmallVectorImpl< Instruction * > &Worklist, SmallPtrSet< Instruction *, 8 > &Visited) |
Variables | |
static cl::opt< bool > | ForceSSAUpdater ("force-ssa-updater", cl::init(false), cl::Hidden) |
sroa | |
Scalar Replacement Of | Aggregates |
Scalar Replacement Of | false |
This transformation implements the well known scalar replacement of aggregates transformation. It tries to identify promotable elements of an aggregate alloca, and promote them to registers. It will also try to convert uses of an element (or set of elements) of an alloca into a vector or bitfield-style integer scalar if appropriate.
It works to do this with minimal slicing of the alloca so that regions which are merely transferred in and out of external memory remain unchanged and are not decomposed to scalar code.
Because this also performs alloca promotion, it can be thought of as also serving the purpose of SSA formation. The algorithm iterates on the function until all opportunities for promotion have been realized.
Definition in file SROA.cpp.
|
static |
Build a GEP out of a base pointer and indices.
This will return the BasePtr if that is valid, or build a new GEP instruction using the IRBuilder if GEP-ing is needed.
Definition at line 1190 of file SROA.cpp.
References llvm::SmallVectorTemplateCommon< T >::back(), llvm::SmallVectorBase::empty(), isZero(), and llvm::SmallVectorTemplateCommon< T >::size().
Referenced by getNaturalGEPWithType().
|
static |
Test whether we can convert a value from the old to the new type.
This predicate should be used to guard calls to convertValue in order to ensure that we only try to convert viable values. The strategy is that we will peel off single element struct and array wrappings to get to an underlying value, and convert that value.
Definition at line 1452 of file SROA.cpp.
References llvm::Type::getScalarType(), llvm::DataLayout::getTypeSizeInBits(), llvm::Type::isIntegerTy(), llvm::Type::isPointerTy(), and llvm::Type::isSingleValueType().
Referenced by convertValue(), isIntegerWideningViable(), isIntegerWideningViableForSlice(), and isVectorPromotionViableForSlice().
|
static |
Generic routine to convert an SSA value to a value of a different type.
This will try various different casting techniques, such as bitcasts, inttoptr, and ptrtoint casts. Use the canConvertValue
predicate to test two types for viability with this routine.
Definition at line 1485 of file SROA.cpp.
References canConvertValue(), llvm::DataLayout::getIntPtrType(), llvm::Type::getScalarType(), llvm::Value::getType(), llvm::Type::isIntegerTy(), llvm::Type::isPointerTy(), and llvm::Type::isVectorTy().
|
static |
Definition at line 3432 of file SROA.cpp.
References llvm::SmallPtrSet< PtrType, SmallSize >::insert(), llvm::SmallVectorTemplateBase< T, isPodLike< T >::value >::push_back(), llvm::Value::use_begin(), and llvm::Value::use_end().
|
static |
Definition at line 1766 of file SROA.cpp.
References llvm::dbgs(), DEBUG, llvm::IntegerType::getBitWidth(), llvm::Value::getType(), llvm::DataLayout::getTypeStoreSize(), and llvm::DataLayout::isBigEndian().
|
static |
Definition at line 1820 of file SROA.cpp.
References llvm::dbgs(), DEBUG, llvm::VectorType::getNumElements(), llvm::Value::getType(), llvm::SmallVectorTemplateBase< T, isPodLike< T >::value >::push_back(), and llvm::SmallVectorImpl< T >::reserve().
|
static |
Walk the range of a partitioning looking for a common type to cover this sequence of slices.
Definition at line 937 of file SROA.cpp.
References llvm::Value::getType(), and I.
|
static |
Definition at line 297 of file SROA.cpp.
References llvm::SelectInst::getCondition(), and llvm::User::getOperand().
|
static |
Compute an adjusted pointer from Ptr by Offset bytes where the resulting pointer has PointerTy.
This tries very hard to compute a "natural" GEP which arrives at the offset and produces the pointer type desired. Where it cannot, it will try to use the natural GEP to arrive at the offset and bitcast to the type. Where that fails, it will try to use an existing i8* and GEP to the byte offset and bitcast to the type.
The strategy for finding the more natural GEPs is to peel off layers of the pointer, walking back through bit casts and GEPs, searching for a base pointer from which we can compute a natural GEP with the desired properties. The algorithm tries to fold as many constant indices into a single GEP as possible, thus making each GEP more independent of the surrounding code.
Definition at line 1359 of file SROA.cpp.
References llvm::SmallVectorImpl< T >::clear(), llvm::Instruction::eraseFromParent(), llvm::APInt::getBitWidth(), getNaturalGEPWithOffset(), llvm::Type::getPointerElementType(), llvm::Value::getType(), llvm::SmallPtrSet< PtrType, SmallSize >::insert(), llvm::Type::isIntegerTy(), llvm::Type::isPointerTy(), P, and llvm::Value::use_empty().
|
static |
Recursively compute indices for a natural GEP.
This is the recursive step for getNaturalGEPWithOffset that walks down the element types adding appropriate indices for the GEP.
Definition at line 1250 of file SROA.cpp.
References llvm::dyn_cast(), llvm::APInt::getBitWidth(), llvm::StructLayout::getElementContainingOffset(), llvm::StructLayout::getElementOffset(), llvm::StructType::getElementType(), getNaturalGEPWithType(), llvm::StructLayout::getSizeInBytes(), llvm::DataLayout::getStructLayout(), llvm::DataLayout::getTypeAllocSize(), llvm::DataLayout::getTypeSizeInBits(), llvm::APInt::getZExtValue(), llvm::Type::isPointerTy(), llvm::SmallVectorTemplateBase< T, isPodLike< T >::value >::push_back(), llvm::APInt::sdiv(), and llvm::APInt::uge().
Referenced by getNaturalGEPWithOffset().
|
static |
Get a natural GEP from a base pointer to a particular offset and resulting in a particular type.
The goal is to produce a "natural" looking GEP that works with the existing composite types to arrive at the appropriate offset and element type for a pointer. TargetTy is the element type the returned GEP should point-to if possible. We recurse by decreasing Offset, adding the appropriate index to Indices, and setting Ty to the result subtype.
If no natural GEP can be constructed, this function returns null.
Definition at line 1320 of file SROA.cpp.
References llvm::APInt::getBitWidth(), llvm::SequentialType::getElementType(), llvm::Type::getInt8PtrTy(), getNaturalGEPRecursively(), llvm::Value::getType(), llvm::DataLayout::getTypeAllocSize(), llvm::Type::isIntegerTy(), llvm::Type::isSized(), llvm::SmallVectorTemplateBase< T, isPodLike< T >::value >::push_back(), and llvm::APInt::sdiv().
Referenced by getAdjustedPtr().
|
static |
Get a natural GEP off of the BasePtr walking through Ty toward TargetTy without changing the offset of the pointer.
This routine assumes we've already established a properly offset GEP with Indices, and arrived at the Ty type. The goal is to continue to GEP with zero-indices down through type layers until we find one the same as TargetTy. If we can't find one with the same type, we at least try to use one with the same size. If none of that works, we just produce the GEP as indicated by Indices to have the correct offset.
Definition at line 1212 of file SROA.cpp.
References buildGEP(), llvm::SmallVectorTemplateCommon< T >::end(), llvm::SmallVectorImpl< T >::erase(), llvm::DataLayout::getPointerSizeInBits(), llvm::Type::isPointerTy(), and llvm::SmallVectorTemplateBase< T, isPodLike< T >::value >::push_back().
Referenced by getNaturalGEPRecursively().
|
static |
Try to find a partition of the aggregate type passed in for a given offset and size.
This recurses through the aggregate type and tries to compute a subtype based on the offset and size. When the offset and size span a sub-section of an array, it will even compute a new array type for that sub-section, and the same for structs.
Note that this routine is very strict and tries to find a partition of the type which produces the exact right offset and size. It is not forgiving when the size or offset cause either end of type-based partition to be off. Also, this is a best-effort routine. It is reasonable to give up and not return a type if necessary.
Definition at line 2939 of file SROA.cpp.
References llvm::dyn_cast(), llvm::StructType::element_begin(), llvm::StructType::element_end(), llvm::Type::getContext(), llvm::StructLayout::getElementContainingOffset(), llvm::StructLayout::getElementOffset(), llvm::StructType::getElementType(), llvm::StructLayout::getSizeInBytes(), llvm::DataLayout::getStructLayout(), llvm::DataLayout::getTypeAllocSize(), llvm::StructType::isPacked(), llvm::makeArrayRef(), and stripAggregateTypeWrapping().
INITIALIZE_PASS_BEGIN | ( | SROA | , |
"sroa" | , | ||
"Scalar Replacement Of Aggregates" | , | ||
false | , | ||
false | |||
) |
|
static |
Definition at line 1789 of file SROA.cpp.
References llvm::dbgs(), DEBUG, llvm::IntegerType::getBitWidth(), llvm::Value::getType(), llvm::DataLayout::getTypeStoreSize(), llvm::DataLayout::isBigEndian(), llvm::APIntOps::shl(), and llvm::APInt::zext().
|
static |
Definition at line 1848 of file SROA.cpp.
References llvm::SmallVectorImpl< T >::clear(), llvm::dbgs(), DEBUG, llvm::dyn_cast(), llvm::VectorType::getNumElements(), llvm::Value::getType(), llvm::SmallVectorTemplateBase< T, isPodLike< T >::value >::push_back(), and llvm::SmallVectorImpl< T >::reserve().
|
static |
Test whether the given alloca partition's integer operations can be widened to promotable ones.
This is a quick test to check whether we can rewrite the integer loads and stores to a particular alloca into wider loads and stores and be able to promote the resulting alloca.
Definition at line 1720 of file SROA.cpp.
References llvm::ArrayRef< T >::begin(), canConvertValue(), llvm::ArrayRef< T >::end(), llvm::Type::getContext(), llvm::DataLayout::getTypeSizeInBits(), llvm::DataLayout::getTypeStoreSize(), llvm::DataLayout::getTypeStoreSizeInBits(), I, isIntegerWideningViableForSlice(), and llvm::DataLayout::isLegalInteger().
|
static |
Test whether a slice of an alloca is valid for integer widening.
This implements the necessary checking for the isIntegerWideningViable
test below on a single slice of the alloca.
Definition at line 1652 of file SROA.cpp.
References canConvertValue(), llvm::IntrinsicInst::getIntrinsicID(), llvm::Value::getType(), llvm::DataLayout::getTypeStoreSizeInBits(), llvm::LoadInst::isVolatile(), llvm::Intrinsic::lifetime_end, llvm::Intrinsic::lifetime_start, and llvm::A64CC::MI.
Referenced by isIntegerWideningViable().
|
static |
PHI instructions that use an alloca and are subsequently loaded can be rewritten to load both input pointers in the pred blocks and then PHI the results, allowing the load of the alloca to be promoted. From this: P2 = phi [i32* Alloca, i32* Other] V = load i32* P2 to: V1 = load i32* Alloca -> will be mem2reg'd ... V2 = load i32* Other ... V = phi [i32 V1, i32 V2]
We can do this to a select if its only uses are loads and if the operands to the select can be loaded unconditionally.
FIXME: This should be hoisted into a generic utility, likely in Transforms/Util/Local.h
Definition at line 1005 of file SROA.cpp.
References llvm::dyn_cast(), llvm::LoadInst::getAlignment(), llvm::PHINode::getIncomingBlock(), llvm::PHINode::getIncomingValue(), llvm::PHINode::getNumIncomingValues(), llvm::TerminatorInst::getNumSuccessors(), llvm::Instruction::getParent(), llvm::BasicBlock::getTerminator(), llvm::Value::isDereferenceablePointer(), llvm::isSafeToLoadUnconditionally(), llvm::LoadInst::isSimple(), LI, llvm::Instruction::mayHaveSideEffects(), llvm::Value::use_begin(), and llvm::Value::use_end().
|
static |
Select instructions that use an alloca and are subsequently loaded can be rewritten to load both input pointers and then select between the result, allowing the load of the alloca to be promoted. From this: P2 = select i1 cond, i32* Alloca, i32* Other V = load i32* P2 to: V1 = load i32* Alloca -> will be mem2reg'd V2 = load i32* Other V = select i1 cond, i32 V1, i32 V2
We can do this to a select if its only uses are loads and if the operand to the select can be loaded unconditionally.
Definition at line 1124 of file SROA.cpp.
References llvm::dyn_cast(), llvm::LoadInst::getAlignment(), llvm::SelectInst::getFalseValue(), llvm::SelectInst::getTrueValue(), llvm::Value::isDereferenceablePointer(), llvm::isSafeToLoadUnconditionally(), llvm::LoadInst::isSimple(), llvm::Value::use_begin(), and llvm::Value::use_end().
|
static |
Test whether the given alloca partitioning and range of slices can be promoted to a vector.
This is a quick test to check whether we can rewrite a particular alloca partition (and its newly formed alloca) into a vector alloca with only whole-vector loads and stores such that it could be promoted to a vector SSA value. We only can ensure this for a limited set of operations, and we don't want to do the rewrites unless we are confident that the result will be promotable, so we have an early test here.
Definition at line 1614 of file SROA.cpp.
References llvm::ArrayRef< T >::begin(), llvm::dyn_cast(), llvm::ArrayRef< T >::end(), llvm::Type::getScalarType(), llvm::DataLayout::getTypeSizeInBits(), I, and isVectorPromotionViableForSlice().
|
static |
Test whether the given slice use can be promoted to a vector.
This function is called to test each entry in a partioning which is slated for a single slice.
Definition at line 1539 of file SROA.cpp.
References canConvertValue(), llvm::Type::getContext(), llvm::SequentialType::getElementType(), llvm::VectorType::getNumElements(), llvm::Value::getType(), llvm::Type::isIntegerTy(), llvm::LoadInst::isVolatile(), and llvm::A64CC::MI.
Referenced by isVectorPromotionViable().
|
static |
Definition at line 3193 of file SROA.cpp.
References llvm::SmallVectorTemplateCommon< T >::begin(), llvm::SmallVectorImpl< T >::clear(), llvm::SmallVectorTemplateCommon< T >::end(), llvm::SmallVectorImpl< T >::erase(), and llvm::SmallVectorTemplateCommon< T >::size().
|
static |
Definition at line 1070 of file SROA.cpp.
References llvm::PHINode::addIncoming(), Align(), llvm::dbgs(), DEBUG, llvm::Instruction::eraseFromParent(), llvm::LoadInst::getAlignment(), llvm::PHINode::getIncomingBlock(), llvm::PHINode::getIncomingValue(), llvm::Instruction::getMetadata(), llvm::Value::getName(), llvm::PHINode::getNumIncomingValues(), llvm::BasicBlock::getTerminator(), llvm::Value::getType(), llvm::SPII::Load, llvm::Value::replaceAllUsesWith(), llvm::LoadInst::setAlignment(), llvm::Instruction::setMetadata(), llvm::Value::use_begin(), and llvm::Value::use_empty().
|
static |
Definition at line 1150 of file SROA.cpp.
References llvm::dbgs(), DEBUG, llvm::Instruction::eraseFromParent(), llvm::LoadInst::getAlignment(), llvm::SelectInst::getCondition(), llvm::SelectInst::getFalseValue(), llvm::Instruction::getMetadata(), llvm::Value::getName(), llvm::SelectInst::getTrueValue(), llvm::LoadInst::isSimple(), llvm::Value::replaceAllUsesWith(), llvm::LoadInst::setAlignment(), llvm::Instruction::setMetadata(), llvm::Value::use_begin(), and llvm::Value::use_empty().
STATISTIC | ( | NumAllocasAnalyzed | , |
"Number of allocas analyzed for replacement" | |||
) |
STATISTIC | ( | NumAllocaPartitions | , |
"Number of alloca partitions formed" | |||
) |
STATISTIC | ( | MaxPartitionsPerAlloca | , |
"Maximum number of partitions per alloca" | |||
) |
STATISTIC | ( | NumAllocaPartitionUses | , |
"Number of alloca partition uses rewritten" | |||
) |
STATISTIC | ( | MaxUsesPerAllocaPartition | , |
"Maximum number of uses of a partition" | |||
) |
STATISTIC | ( | NumNewAllocas | , |
"Number of | new, | ||
smaller allocas introduced" | |||
) |
STATISTIC | ( | NumPromoted | , |
"Number of allocas promoted to SSA values" | |||
) |
STATISTIC | ( | NumLoadsSpeculated | , |
"Number of loads speculated to allow promotion" | |||
) |
STATISTIC | ( | NumDeleted | , |
"Number of instructions deleted" | |||
) |
STATISTIC | ( | NumVectorized | , |
"Number of vectorized aggregates" | |||
) |
|
static |
Strip aggregate type wrapping.
This removes no-op aggregate types wrapping an underlying type. It will strip as many layers of types as it can without changing either the type size or the allocated size.
Definition at line 2901 of file SROA.cpp.
References llvm::StructLayout::getElementContainingOffset(), llvm::DataLayout::getStructLayout(), llvm::DataLayout::getTypeAllocSize(), llvm::DataLayout::getTypeSizeInBits(), and llvm::Type::isSingleValueType().
Referenced by getTypePartition().
Hidden option to force the pass to not use DomTree and mem2reg, instead forming SSA values through the SSAUpdater infrastructure.