/build/source/llvm/lib/Transforms/Scalar/SROA.cpp

Bug Summary

File:	build/source/llvm/lib/Transforms/Scalar/SROA.cpp
Warning:	line 3070, column 54 Called C++ object pointer is null
Annotated Source Code

Press '?' to see keyboard shortcuts
Show analyzer invocation
clang -cc1 -cc1 -triple x86_64-pc-linux-gnu -analyze -disable-free -clear-ast-before-backend -disable-llvm-verifier -discard-value-names -main-file-name SROA.cpp -analyzer-checker=core -analyzer-checker=apiModeling -analyzer-checker=unix -analyzer-checker=deadcode -analyzer-checker=cplusplus -analyzer-checker=security.insecureAPI.UncheckedReturn -analyzer-checker=security.insecureAPI.getpw -analyzer-checker=security.insecureAPI.gets -analyzer-checker=security.insecureAPI.mktemp -analyzer-checker=security.insecureAPI.mkstemp -analyzer-checker=security.insecureAPI.vfork -analyzer-checker=nullability.NullPassedToNonnull -analyzer-checker=nullability.NullReturnedFromNonnull -analyzer-output plist -w -setup-static-analyzer -analyzer-config-compatibility-mode=true -mrelocation-model pic -pic-level 2 -mframe-pointer=none -fmath-errno -ffp-contract=on -fno-rounding-math -mconstructor-aliases -funwind-tables=2 -target-cpu x86-64 -tune-cpu generic -debugger-tuning=gdb -ffunction-sections -fdata-sections -fcoverage-compilation-dir=/build/source/build-llvm -resource-dir /usr/lib/llvm-16/lib/clang/16 -I lib/Transforms/Scalar -I /build/source/llvm/lib/Transforms/Scalar -I include -I /build/source/llvm/include -D _DEBUG -D _GNU_SOURCE -D __STDC_CONSTANT_MACROS -D __STDC_FORMAT_MACROS -D __STDC_LIMIT_MACROS -D _FORTIFY_SOURCE=2 -D NDEBUG -U NDEBUG -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/x86_64-linux-gnu/c++/10 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/backward -internal-isystem /usr/lib/llvm-16/lib/clang/16/include -internal-isystem /usr/local/include -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/10/../../../../x86_64-linux-gnu/include -internal-externc-isystem /usr/include/x86_64-linux-gnu -internal-externc-isystem /include -internal-externc-isystem /usr/include -fmacro-prefix-map=/build/source/build-llvm=build-llvm -fmacro-prefix-map=/build/source/= -fcoverage-prefix-map=/build/source/build-llvm=build-llvm -fcoverage-prefix-map=/build/source/= -source-date-epoch 1670930111 -O3 -Wno-unused-command-line-argument -Wno-unused-parameter -Wwrite-strings -Wno-missing-field-initializers -Wno-long-long -Wno-maybe-uninitialized -Wno-class-memaccess -Wno-redundant-move -Wno-pessimizing-move -Wno-noexcept-type -Wno-comment -Wno-misleading-indentation -std=c++17 -fdeprecated-macro -fdebug-compilation-dir=/build/source/build-llvm -fdebug-prefix-map=/build/source/build-llvm=build-llvm -fdebug-prefix-map=/build/source/= -fdebug-prefix-map=/build/source/build-llvm=build-llvm -fdebug-prefix-map=/build/source/= -ferror-limit 19 -fvisibility-inlines-hidden -stack-protector 2 -fgnuc-version=4.2.1 -fcolor-diagnostics -vectorize-loops -vectorize-slp -analyzer-output=html -analyzer-config stable-report-filename=true -faddrsig -D__GCC_HAVE_DWARF2_CFI_ASM=1 -o /tmp/scan-build-2022-12-13-140357-15994-1 -x c++ /build/source/llvm/lib/Transforms/Scalar/SROA.cpp
1//===- SROA.cpp - Scalar Replacement Of Aggregates ------------------------===//
2//
3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4// See https://llvm.org/LICENSE.txt for license information.
5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6//
7//===----------------------------------------------------------------------===//
8/// \file
9/// This transformation implements the well known scalar replacement of
10/// aggregates transformation. It tries to identify promotable elements of an
11/// aggregate alloca, and promote them to registers. It will also try to
12/// convert uses of an element (or set of elements) of an alloca into a vector
13/// or bitfield-style integer scalar if appropriate.
14///
15/// It works to do this with minimal slicing of the alloca so that regions
16/// which are merely transferred in and out of external memory remain unchanged
17/// and are not decomposed to scalar code.
18///
19/// Because this also performs alloca promotion, it can be thought of as also
20/// serving the purpose of SSA formation. The algorithm iterates on the
21/// function until all opportunities for promotion have been realized.
22///
23//===----------------------------------------------------------------------===//

25#include "llvm/Transforms/Scalar/SROA.h"
26#include "llvm/ADT/APInt.h"
27#include "llvm/ADT/ArrayRef.h"
28#include "llvm/ADT/DenseMap.h"
29#include "llvm/ADT/PointerIntPair.h"
30#include "llvm/ADT/STLExtras.h"
31#include "llvm/ADT/SetVector.h"
32#include "llvm/ADT/SmallBitVector.h"
33#include "llvm/ADT/SmallPtrSet.h"
34#include "llvm/ADT/SmallVector.h"
35#include "llvm/ADT/Statistic.h"
36#include "llvm/ADT/StringRef.h"
37#include "llvm/ADT/Twine.h"
38#include "llvm/ADT/iterator.h"
39#include "llvm/ADT/iterator_range.h"
40#include "llvm/Analysis/AssumptionCache.h"
41#include "llvm/Analysis/DomTreeUpdater.h"
42#include "llvm/Analysis/GlobalsModRef.h"
43#include "llvm/Analysis/Loads.h"
44#include "llvm/Analysis/PtrUseVisitor.h"
45#include "llvm/Config/llvm-config.h"
46#include "llvm/IR/BasicBlock.h"
47#include "llvm/IR/Constant.h"
48#include "llvm/IR/ConstantFolder.h"
49#include "llvm/IR/Constants.h"
50#include "llvm/IR/DIBuilder.h"
51#include "llvm/IR/DataLayout.h"
52#include "llvm/IR/DebugInfo.h"
53#include "llvm/IR/DebugInfoMetadata.h"
54#include "llvm/IR/DerivedTypes.h"
55#include "llvm/IR/Dominators.h"
56#include "llvm/IR/Function.h"
57#include "llvm/IR/GetElementPtrTypeIterator.h"
58#include "llvm/IR/GlobalAlias.h"
59#include "llvm/IR/IRBuilder.h"
60#include "llvm/IR/InstVisitor.h"
61#include "llvm/IR/Instruction.h"
62#include "llvm/IR/Instructions.h"
63#include "llvm/IR/IntrinsicInst.h"
64#include "llvm/IR/LLVMContext.h"
65#include "llvm/IR/Metadata.h"
66#include "llvm/IR/Module.h"
67#include "llvm/IR/Operator.h"
68#include "llvm/IR/PassManager.h"
69#include "llvm/IR/Type.h"
70#include "llvm/IR/Use.h"
71#include "llvm/IR/User.h"
72#include "llvm/IR/Value.h"
73#include "llvm/InitializePasses.h"
74#include "llvm/Pass.h"
75#include "llvm/Support/Casting.h"
76#include "llvm/Support/CommandLine.h"
77#include "llvm/Support/Compiler.h"
78#include "llvm/Support/Debug.h"
79#include "llvm/Support/ErrorHandling.h"
80#include "llvm/Support/raw_ostream.h"
81#include "llvm/Transforms/Scalar.h"
82#include "llvm/Transforms/Utils/BasicBlockUtils.h"
83#include "llvm/Transforms/Utils/Local.h"
84#include "llvm/Transforms/Utils/PromoteMemToReg.h"
85#include <algorithm>
86#include <cassert>
87#include <cstddef>
88#include <cstdint>
89#include <cstring>
90#include <iterator>
91#include <string>
92#include <tuple>
93#include <utility>
94#include <vector>

96using namespace llvm;
97using namespace llvm::sroa;

99#define DEBUG_TYPE"sroa" "sroa"

101STATISTIC(NumAllocasAnalyzed, "Number of allocas analyzed for replacement")static llvm::Statistic NumAllocasAnalyzed = {"sroa", "NumAllocasAnalyzed"
, "Number of allocas analyzed for replacement"};
102STATISTIC(NumAllocaPartitions, "Number of alloca partitions formed")static llvm::Statistic NumAllocaPartitions = {"sroa", "NumAllocaPartitions"
, "Number of alloca partitions formed"};
103STATISTIC(MaxPartitionsPerAlloca, "Maximum number of partitions per alloca")static llvm::Statistic MaxPartitionsPerAlloca = {"sroa", "MaxPartitionsPerAlloca"
, "Maximum number of partitions per alloca"};
104STATISTIC(NumAllocaPartitionUses, "Number of alloca partition uses rewritten")static llvm::Statistic NumAllocaPartitionUses = {"sroa", "NumAllocaPartitionUses"
, "Number of alloca partition uses rewritten"};
105STATISTIC(MaxUsesPerAllocaPartition, "Maximum number of uses of a partition")static llvm::Statistic MaxUsesPerAllocaPartition = {"sroa", "MaxUsesPerAllocaPartition"
, "Maximum number of uses of a partition"};
106STATISTIC(NumNewAllocas, "Number of new, smaller allocas introduced")static llvm::Statistic NumNewAllocas = {"sroa", "NumNewAllocas"
, "Number of new, smaller allocas introduced"};
107STATISTIC(NumPromoted, "Number of allocas promoted to SSA values")static llvm::Statistic NumPromoted = {"sroa", "NumPromoted", "Number of allocas promoted to SSA values"
};
108STATISTIC(NumLoadsSpeculated, "Number of loads speculated to allow promotion")static llvm::Statistic NumLoadsSpeculated = {"sroa", "NumLoadsSpeculated"
, "Number of loads speculated to allow promotion"};
109STATISTIC(NumLoadsPredicated,static llvm::Statistic NumLoadsPredicated = {"sroa", "NumLoadsPredicated"
, "Number of loads rewritten into predicated loads to allow promotion"
}
        "Number of loads rewritten into predicated loads to allow promotion")static llvm::Statistic NumLoadsPredicated = {"sroa", "NumLoadsPredicated"
, "Number of loads rewritten into predicated loads to allow promotion"
};
111STATISTIC(static llvm::Statistic NumStoresPredicated = {"sroa", "NumStoresPredicated"
, "Number of stores rewritten into predicated loads to allow promotion"
}
  NumStoresPredicated,static llvm::Statistic NumStoresPredicated = {"sroa", "NumStoresPredicated"
, "Number of stores rewritten into predicated loads to allow promotion"
}
  "Number of stores rewritten into predicated loads to allow promotion")static llvm::Statistic NumStoresPredicated = {"sroa", "NumStoresPredicated"
, "Number of stores rewritten into predicated loads to allow promotion"
};
114STATISTIC(NumDeleted, "Number of instructions deleted")static llvm::Statistic NumDeleted = {"sroa", "NumDeleted", "Number of instructions deleted"
};
115STATISTIC(NumVectorized, "Number of vectorized aggregates")static llvm::Statistic NumVectorized = {"sroa", "NumVectorized"
, "Number of vectorized aggregates"};

117/// Hidden option to experiment with completely strict handling of inbounds
118/// GEPs.
119static cl::opt<bool> SROAStrictInbounds("sroa-strict-inbounds", cl::init(false),
                                      cl::Hidden);
121namespace {

123/// A custom IRBuilder inserter which prefixes all names, but only in
124/// Assert builds.
125class IRBuilderPrefixedInserter final : public IRBuilderDefaultInserter {
std::string Prefix;

Twine getNameWithPrefix(const Twine &Name) const {
  return Name.isTriviallyEmpty() ? Name : Prefix + Name;
}

132public:
void SetNamePrefix(const Twine &P) { Prefix = P.str(); }

void InsertHelper(Instruction *I, const Twine &Name, BasicBlock *BB,
                  BasicBlock::iterator InsertPt) const override {
  IRBuilderDefaultInserter::InsertHelper(I, getNameWithPrefix(Name), BB,
                                         InsertPt);
}
140};

142/// Provide a type for IRBuilder that drops names in release builds.
143using IRBuilderTy = IRBuilder<ConstantFolder, IRBuilderPrefixedInserter>;

145/// A used slice of an alloca.
146///
147/// This structure represents a slice of an alloca used by some instruction. It
148/// stores both the begin and end offsets of this use, a pointer to the use
149/// itself, and a flag indicating whether we can classify the use as splittable
150/// or not when forming partitions of the alloca.
151class Slice {
/// The beginning offset of the range.
uint64_t BeginOffset = 0;

/// The ending offset, not included in the range.
uint64_t EndOffset = 0;

/// Storage for both the use of this slice and whether it can be
/// split.
PointerIntPair<Use *, 1, bool> UseAndIsSplittable;

162public:
Slice() = default;

Slice(uint64_t BeginOffset, uint64_t EndOffset, Use *U, bool IsSplittable)
    : BeginOffset(BeginOffset), EndOffset(EndOffset),
      UseAndIsSplittable(U, IsSplittable) {}

uint64_t beginOffset() const { return BeginOffset; }
uint64_t endOffset() const { return EndOffset; }

bool isSplittable() const { return UseAndIsSplittable.getInt(); }
void makeUnsplittable() { UseAndIsSplittable.setInt(false); }

Use *getUse() const { return UseAndIsSplittable.getPointer(); }

bool isDead() const { return getUse() == nullptr; }
void kill() { UseAndIsSplittable.setPointer(nullptr); }

/// Support for ordering ranges.
///
/// This provides an ordering over ranges such that start offsets are
/// always increasing, and within equal start offsets, the end offsets are
/// decreasing. Thus the spanning range comes first in a cluster with the
/// same start position.
bool operator<(const Slice &RHS) const {
  if (beginOffset() < RHS.beginOffset())
    return true;
  if (beginOffset() > RHS.beginOffset())
    return false;
  if (isSplittable() != RHS.isSplittable())
    return !isSplittable();
  if (endOffset() > RHS.endOffset())
    return true;
  return false;
}

/// Support comparison with a single offset to allow binary searches.
friend LLVM_ATTRIBUTE_UNUSED__attribute__((__unused__)) bool operator<(const Slice &LHS,
                                            uint64_t RHSOffset) {
  return LHS.beginOffset() < RHSOffset;
}
friend LLVM_ATTRIBUTE_UNUSED__attribute__((__unused__)) bool operator<(uint64_t LHSOffset,
                                            const Slice &RHS) {
  return LHSOffset < RHS.beginOffset();
}

bool operator==(const Slice &RHS) const {
  return isSplittable() == RHS.isSplittable() &&
         beginOffset() == RHS.beginOffset() && endOffset() == RHS.endOffset();
}
bool operator!=(const Slice &RHS) const { return !operator==(RHS); }
213};

215} // end anonymous namespace

217/// Representation of the alloca slices.
218///
219/// This class represents the slices of an alloca which are formed by its
220/// various uses. If a pointer escapes, we can't fully build a representation
221/// for the slices used and we reflect that in this structure. The uses are
222/// stored, sorted by increasing beginning offset and with unsplittable slices
223/// starting at a particular offset before splittable slices.
224class llvm::sroa::AllocaSlices {
225public:
/// Construct the slices of a particular alloca.
AllocaSlices(const DataLayout &DL, AllocaInst &AI);

/// Test whether a pointer to the allocation escapes our analysis.
///
/// If this is true, the slices are never fully built and should be
/// ignored.
bool isEscaped() const { return PointerEscapingInstr; }

/// Support for iterating over the slices.
/// @{
using iterator = SmallVectorImpl<Slice>::iterator;
using range = iterator_range<iterator>;

iterator begin() { return Slices.begin(); }
iterator end() { return Slices.end(); }

using const_iterator = SmallVectorImpl<Slice>::const_iterator;
using const_range = iterator_range<const_iterator>;

const_iterator begin() const { return Slices.begin(); }
const_iterator end() const { return Slices.end(); }
/// @}

/// Erase a range of slices.
void erase(iterator Start, iterator Stop) { Slices.erase(Start, Stop); }

/// Insert new slices for this alloca.
///
/// This moves the slices into the alloca's slices collection, and re-sorts
/// everything so that the usual ordering properties of the alloca's slices
/// hold.
void insert(ArrayRef<Slice> NewSlices) {
  int OldSize = Slices.size();
  Slices.append(NewSlices.begin(), NewSlices.end());
  auto SliceI = Slices.begin() + OldSize;
  llvm::sort(SliceI, Slices.end());
  std::inplace_merge(Slices.begin(), SliceI, Slices.end());
}

// Forward declare the iterator and range accessor for walking the
// partitions.
class partition_iterator;
iterator_range<partition_iterator> partitions();

/// Access the dead users for this alloca.
ArrayRef<Instruction *> getDeadUsers() const { return DeadUsers; }

/// Access Uses that should be dropped if the alloca is promotable.
ArrayRef<Use *> getDeadUsesIfPromotable() const {
  return DeadUseIfPromotable;
}

/// Access the dead operands referring to this alloca.
///
/// These are operands which have cannot actually be used to refer to the
/// alloca as they are outside its range and the user doesn't correct for
/// that. These mostly consist of PHI node inputs and the like which we just
/// need to replace with undef.
ArrayRef<Use *> getDeadOperands() const { return DeadOperands; }

287#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
void print(raw_ostream &OS, const_iterator I, StringRef Indent = "  ") const;
void printSlice(raw_ostream &OS, const_iterator I,
                StringRef Indent = "  ") const;
void printUse(raw_ostream &OS, const_iterator I,
              StringRef Indent = "  ") const;
void print(raw_ostream &OS) const;
void dump(const_iterator I) const;
void dump() const;
296#endif

298private:
template <typename DerivedT, typename RetT = void> class BuilderBase;
class SliceBuilder;

friend class AllocaSlices::SliceBuilder;

304#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
/// Handle to alloca instruction to simplify method interfaces.
AllocaInst &AI;
307#endif

/// The instruction responsible for this alloca not having a known set
/// of slices.
///
/// When an instruction (potentially) escapes the pointer to the alloca, we
/// store a pointer to that here and abort trying to form slices of the
/// alloca. This will be null if the alloca slices are analyzed successfully.
Instruction *PointerEscapingInstr;

/// The slices of the alloca.
///
/// We store a vector of the slices formed by uses of the alloca here. This
/// vector is sorted by increasing begin offset, and then the unsplittable
/// slices before the splittable ones. See the Slice inner class for more
/// details.
SmallVector<Slice, 8> Slices;

/// Instructions which will become dead if we rewrite the alloca.
///
/// Note that these are not separated by slice. This is because we expect an
/// alloca to be completely rewritten or not rewritten at all. If rewritten,
/// all these instructions can simply be removed and replaced with poison as
/// they come from outside of the allocated space.
SmallVector<Instruction *, 8> DeadUsers;

/// Uses which will become dead if can promote the alloca.
SmallVector<Use *, 8> DeadUseIfPromotable;

/// Operands which will become dead if we rewrite the alloca.
///
/// These are operands that in their particular use can be replaced with
/// poison when we rewrite the alloca. These show up in out-of-bounds inputs
/// to PHI nodes and the like. They aren't entirely dead (there might be
/// a GEP back into the bounds using it elsewhere) and nor is the PHI, but we
/// want to swap this particular input for poison to simplify the use lists of
/// the alloca.
SmallVector<Use *, 8> DeadOperands;
345};

347/// A partition of the slices.
348///
349/// An ephemeral representation for a range of slices which can be viewed as
350/// a partition of the alloca. This range represents a span of the alloca's
351/// memory which cannot be split, and provides access to all of the slices
352/// overlapping some part of the partition.
353///
354/// Objects of this type are produced by traversing the alloca's slices, but
355/// are only ephemeral and not persistent.
356class llvm::sroa::Partition {
357private:
friend class AllocaSlices;
friend class AllocaSlices::partition_iterator;

using iterator = AllocaSlices::iterator;

/// The beginning and ending offsets of the alloca for this
/// partition.
uint64_t BeginOffset = 0, EndOffset = 0;

/// The start and end iterators of this partition.
iterator SI, SJ;

/// A collection of split slice tails overlapping the partition.
SmallVector<Slice *, 4> SplitTails;

/// Raw constructor builds an empty partition starting and ending at
/// the given iterator.
Partition(iterator SI) : SI(SI), SJ(SI) {}

377public:
/// The start offset of this partition.
///
/// All of the contained slices start at or after this offset.
uint64_t beginOffset() const { return BeginOffset; }

/// The end offset of this partition.
///
/// All of the contained slices end at or before this offset.
uint64_t endOffset() const { return EndOffset; }

/// The size of the partition.
///
/// Note that this can never be zero.
uint64_t size() const {
  assert(BeginOffset < EndOffset && "Partitions must span some bytes!")(static_cast <bool> (BeginOffset < EndOffset &&
 "Partitions must span some bytes!") ? void (0) : __assert_fail
 ("BeginOffset < EndOffset && \"Partitions must span some bytes!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 392, __extension__ __PRETTY_FUNCTION__
));
  return EndOffset - BeginOffset;
}

/// Test whether this partition contains no slices, and merely spans
/// a region occupied by split slices.
bool empty() const { return SI == SJ; }

/// \name Iterate slices that start within the partition.
/// These may be splittable or unsplittable. They have a begin offset >= the
/// partition begin offset.
/// @{
// FIXME: We should probably define a "concat_iterator" helper and use that
// to stitch together pointee_iterators over the split tails and the
// contiguous iterators of the partition. That would give a much nicer
// interface here. We could then additionally expose filtered iterators for
// split, unsplit, and unsplittable splices based on the usage patterns.
iterator begin() const { return SI; }
iterator end() const { return SJ; }
/// @}

/// Get the sequence of split slice tails.
///
/// These tails are of slices which start before this partition but are
/// split and overlap into the partition. We accumulate these while forming
/// partitions.
ArrayRef<Slice *> splitSliceTails() const { return SplitTails; }
419};

421/// An iterator over partitions of the alloca's slices.
422///
423/// This iterator implements the core algorithm for partitioning the alloca's
424/// slices. It is a forward iterator as we don't support backtracking for
425/// efficiency reasons, and re-use a single storage area to maintain the
426/// current set of split slices.
427///
428/// It is templated on the slice iterator type to use so that it can operate
429/// with either const or non-const slice iterators.
430class AllocaSlices::partition_iterator
  : public iterator_facade_base<partition_iterator, std::forward_iterator_tag,
                                Partition> {
friend class AllocaSlices;

/// Most of the state for walking the partitions is held in a class
/// with a nice interface for examining them.
Partition P;

/// We need to keep the end of the slices to know when to stop.
AllocaSlices::iterator SE;

/// We also need to keep track of the maximum split end offset seen.
/// FIXME: Do we really?
uint64_t MaxSplitSliceEndOffset = 0;

/// Sets the partition to be empty at given iterator, and sets the
/// end iterator.
partition_iterator(AllocaSlices::iterator SI, AllocaSlices::iterator SE)
    : P(SI), SE(SE) {
  // If not already at the end, advance our state to form the initial
  // partition.
  if (SI != SE)
    advance();
}

/// Advance the iterator to the next partition.
///
/// Requires that the iterator not be at the end of the slices.
void advance() {
  assert((P.SI != SE || !P.SplitTails.empty()) &&(static_cast <bool> ((P.SI != SE || !P.SplitTails.empty
()) && "Cannot advance past the end of the slices!") ?
 void (0) : __assert_fail ("(P.SI != SE || !P.SplitTails.empty()) && \"Cannot advance past the end of the slices!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 461, __extension__ __PRETTY_FUNCTION__
))
         "Cannot advance past the end of the slices!")(static_cast <bool> ((P.SI != SE || !P.SplitTails.empty
()) && "Cannot advance past the end of the slices!") ?
 void (0) : __assert_fail ("(P.SI != SE || !P.SplitTails.empty()) && \"Cannot advance past the end of the slices!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 461, __extension__ __PRETTY_FUNCTION__
));

  // Clear out any split uses which have ended.
  if (!P.SplitTails.empty()) {
    if (P.EndOffset >= MaxSplitSliceEndOffset) {
      // If we've finished all splits, this is easy.
      P.SplitTails.clear();
      MaxSplitSliceEndOffset = 0;
    } else {
      // Remove the uses which have ended in the prior partition. This
      // cannot change the max split slice end because we just checked that
      // the prior partition ended prior to that max.
      llvm::erase_if(P.SplitTails,
                     [&](Slice *S) { return S->endOffset() <= P.EndOffset; });
      assert(llvm::any_of(P.SplitTails,(static_cast <bool> (llvm::any_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset
; }) && "Could not find the current max split slice offset!"
) ? void (0) : __assert_fail ("llvm::any_of(P.SplitTails, [&](Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset; }) && \"Could not find the current max split slice offset!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 479, __extension__ __PRETTY_FUNCTION__
))
                          [&](Slice *S) {(static_cast <bool> (llvm::any_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset
; }) && "Could not find the current max split slice offset!"
) ? void (0) : __assert_fail ("llvm::any_of(P.SplitTails, [&](Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset; }) && \"Could not find the current max split slice offset!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 479, __extension__ __PRETTY_FUNCTION__
))
                            return S->endOffset() == MaxSplitSliceEndOffset;(static_cast <bool> (llvm::any_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset
; }) && "Could not find the current max split slice offset!"
) ? void (0) : __assert_fail ("llvm::any_of(P.SplitTails, [&](Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset; }) && \"Could not find the current max split slice offset!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 479, __extension__ __PRETTY_FUNCTION__
))
                          }) &&(static_cast <bool> (llvm::any_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset
; }) && "Could not find the current max split slice offset!"
) ? void (0) : __assert_fail ("llvm::any_of(P.SplitTails, [&](Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset; }) && \"Could not find the current max split slice offset!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 479, __extension__ __PRETTY_FUNCTION__
))
             "Could not find the current max split slice offset!")(static_cast <bool> (llvm::any_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset
; }) && "Could not find the current max split slice offset!"
) ? void (0) : __assert_fail ("llvm::any_of(P.SplitTails, [&](Slice *S) { return S->endOffset() == MaxSplitSliceEndOffset; }) && \"Could not find the current max split slice offset!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 479, __extension__ __PRETTY_FUNCTION__
));
      assert(llvm::all_of(P.SplitTails,(static_cast <bool> (llvm::all_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset
; }) && "Max split slice end offset is not actually the max!"
) ? void (0) : __assert_fail ("llvm::all_of(P.SplitTails, [&](Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset; }) && \"Max split slice end offset is not actually the max!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 484, __extension__ __PRETTY_FUNCTION__
))
                          [&](Slice *S) {(static_cast <bool> (llvm::all_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset
; }) && "Max split slice end offset is not actually the max!"
) ? void (0) : __assert_fail ("llvm::all_of(P.SplitTails, [&](Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset; }) && \"Max split slice end offset is not actually the max!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 484, __extension__ __PRETTY_FUNCTION__
))
                            return S->endOffset() <= MaxSplitSliceEndOffset;(static_cast <bool> (llvm::all_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset
; }) && "Max split slice end offset is not actually the max!"
) ? void (0) : __assert_fail ("llvm::all_of(P.SplitTails, [&](Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset; }) && \"Max split slice end offset is not actually the max!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 484, __extension__ __PRETTY_FUNCTION__
))
                          }) &&(static_cast <bool> (llvm::all_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset
; }) && "Max split slice end offset is not actually the max!"
) ? void (0) : __assert_fail ("llvm::all_of(P.SplitTails, [&](Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset; }) && \"Max split slice end offset is not actually the max!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 484, __extension__ __PRETTY_FUNCTION__
))
             "Max split slice end offset is not actually the max!")(static_cast <bool> (llvm::all_of(P.SplitTails, [&]
(Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset
; }) && "Max split slice end offset is not actually the max!"
) ? void (0) : __assert_fail ("llvm::all_of(P.SplitTails, [&](Slice *S) { return S->endOffset() <= MaxSplitSliceEndOffset; }) && \"Max split slice end offset is not actually the max!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 484, __extension__ __PRETTY_FUNCTION__
));
    }
  }

  // If P.SI is already at the end, then we've cleared the split tail and
  // now have an end iterator.
  if (P.SI == SE) {
    assert(P.SplitTails.empty() && "Failed to clear the split slices!")(static_cast <bool> (P.SplitTails.empty() && "Failed to clear the split slices!"
) ? void (0) : __assert_fail ("P.SplitTails.empty() && \"Failed to clear the split slices!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 491, __extension__ __PRETTY_FUNCTION__
));
    return;
  }

  // If we had a non-empty partition previously, set up the state for
  // subsequent partitions.
  if (P.SI != P.SJ) {
    // Accumulate all the splittable slices which started in the old
    // partition into the split list.
    for (Slice &S : P)
      if (S.isSplittable() && S.endOffset() > P.EndOffset) {
        P.SplitTails.push_back(&S);
        MaxSplitSliceEndOffset =
            std::max(S.endOffset(), MaxSplitSliceEndOffset);
      }

    // Start from the end of the previous partition.
    P.SI = P.SJ;

    // If P.SI is now at the end, we at most have a tail of split slices.
    if (P.SI == SE) {
      P.BeginOffset = P.EndOffset;
      P.EndOffset = MaxSplitSliceEndOffset;
      return;
    }

    // If the we have split slices and the next slice is after a gap and is
    // not splittable immediately form an empty partition for the split
    // slices up until the next slice begins.
    if (!P.SplitTails.empty() && P.SI->beginOffset() != P.EndOffset &&
        !P.SI->isSplittable()) {
      P.BeginOffset = P.EndOffset;
      P.EndOffset = P.SI->beginOffset();
      return;
    }
  }

  // OK, we need to consume new slices. Set the end offset based on the
  // current slice, and step SJ past it. The beginning offset of the
  // partition is the beginning offset of the next slice unless we have
  // pre-existing split slices that are continuing, in which case we begin
  // at the prior end offset.
  P.BeginOffset = P.SplitTails.empty() ? P.SI->beginOffset() : P.EndOffset;
  P.EndOffset = P.SI->endOffset();
  ++P.SJ;

  // There are two strategies to form a partition based on whether the
  // partition starts with an unsplittable slice or a splittable slice.
  if (!P.SI->isSplittable()) {
    // When we're forming an unsplittable region, it must always start at
    // the first slice and will extend through its end.
    assert(P.BeginOffset == P.SI->beginOffset())(static_cast <bool> (P.BeginOffset == P.SI->beginOffset
()) ? void (0) : __assert_fail ("P.BeginOffset == P.SI->beginOffset()"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 542, __extension__ __PRETTY_FUNCTION__
));

    // Form a partition including all of the overlapping slices with this
    // unsplittable slice.
    while (P.SJ != SE && P.SJ->beginOffset() < P.EndOffset) {
      if (!P.SJ->isSplittable())
        P.EndOffset = std::max(P.EndOffset, P.SJ->endOffset());
      ++P.SJ;
    }

    // We have a partition across a set of overlapping unsplittable
    // partitions.
    return;
  }

  // If we're starting with a splittable slice, then we need to form
  // a synthetic partition spanning it and any other overlapping splittable
  // splices.
  assert(P.SI->isSplittable() && "Forming a splittable partition!")(static_cast <bool> (P.SI->isSplittable() &&
 "Forming a splittable partition!") ? void (0) : __assert_fail
 ("P.SI->isSplittable() && \"Forming a splittable partition!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 560, __extension__ __PRETTY_FUNCTION__
));

  // Collect all of the overlapping splittable slices.
  while (P.SJ != SE && P.SJ->beginOffset() < P.EndOffset &&
         P.SJ->isSplittable()) {
    P.EndOffset = std::max(P.EndOffset, P.SJ->endOffset());
    ++P.SJ;
  }

  // Back upiP.EndOffset if we ended the span early when encountering an
  // unsplittable slice. This synthesizes the early end offset of
  // a partition spanning only splittable slices.
  if (P.SJ != SE && P.SJ->beginOffset() < P.EndOffset) {
    assert(!P.SJ->isSplittable())(static_cast <bool> (!P.SJ->isSplittable()) ? void (
0) : __assert_fail ("!P.SJ->isSplittable()", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 573, __extension__ __PRETTY_FUNCTION__));
    P.EndOffset = P.SJ->beginOffset();
  }
}

578public:
bool operator==(const partition_iterator &RHS) const {
  assert(SE == RHS.SE &&(static_cast <bool> (SE == RHS.SE && "End iterators don't match between compared partition iterators!"
) ? void (0) : __assert_fail ("SE == RHS.SE && \"End iterators don't match between compared partition iterators!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 581, __extension__ __PRETTY_FUNCTION__
))
         "End iterators don't match between compared partition iterators!")(static_cast <bool> (SE == RHS.SE && "End iterators don't match between compared partition iterators!"
) ? void (0) : __assert_fail ("SE == RHS.SE && \"End iterators don't match between compared partition iterators!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 581, __extension__ __PRETTY_FUNCTION__
));

  // The observed positions of partitions is marked by the P.SI iterator and
  // the emptiness of the split slices. The latter is only relevant when
  // P.SI == SE, as the end iterator will additionally have an empty split
  // slices list, but the prior may have the same P.SI and a tail of split
  // slices.
  if (P.SI == RHS.P.SI && P.SplitTails.empty() == RHS.P.SplitTails.empty()) {
    assert(P.SJ == RHS.P.SJ &&(static_cast <bool> (P.SJ == RHS.P.SJ && "Same set of slices formed two different sized partitions!"
) ? void (0) : __assert_fail ("P.SJ == RHS.P.SJ && \"Same set of slices formed two different sized partitions!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 590, __extension__ __PRETTY_FUNCTION__
))
           "Same set of slices formed two different sized partitions!")(static_cast <bool> (P.SJ == RHS.P.SJ && "Same set of slices formed two different sized partitions!"
) ? void (0) : __assert_fail ("P.SJ == RHS.P.SJ && \"Same set of slices formed two different sized partitions!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 590, __extension__ __PRETTY_FUNCTION__
));
    assert(P.SplitTails.size() == RHS.P.SplitTails.size() &&(static_cast <bool> (P.SplitTails.size() == RHS.P.SplitTails
.size() && "Same slice position with differently sized non-empty split "
 "slice tails!") ? void (0) : __assert_fail ("P.SplitTails.size() == RHS.P.SplitTails.size() && \"Same slice position with differently sized non-empty split \" \"slice tails!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 593, __extension__ __PRETTY_FUNCTION__
))
           "Same slice position with differently sized non-empty split "(static_cast <bool> (P.SplitTails.size() == RHS.P.SplitTails
.size() && "Same slice position with differently sized non-empty split "
 "slice tails!") ? void (0) : __assert_fail ("P.SplitTails.size() == RHS.P.SplitTails.size() && \"Same slice position with differently sized non-empty split \" \"slice tails!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 593, __extension__ __PRETTY_FUNCTION__
))
           "slice tails!")(static_cast <bool> (P.SplitTails.size() == RHS.P.SplitTails
.size() && "Same slice position with differently sized non-empty split "
 "slice tails!") ? void (0) : __assert_fail ("P.SplitTails.size() == RHS.P.SplitTails.size() && \"Same slice position with differently sized non-empty split \" \"slice tails!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 593, __extension__ __PRETTY_FUNCTION__
));
    return true;
  }
  return false;
}

partition_iterator &operator++() {
  advance();
  return *this;
}

Partition &operator*() { return P; }
605};

607/// A forward range over the partitions of the alloca's slices.
608///
609/// This accesses an iterator range over the partitions of the alloca's
610/// slices. It computes these partitions on the fly based on the overlapping
611/// offsets of the slices and the ability to split them. It will visit "empty"
612/// partitions to cover regions of the alloca only accessed via split
613/// slices.
614iterator_range<AllocaSlices::partition_iterator> AllocaSlices::partitions() {
return make_range(partition_iterator(begin(), end()),
                  partition_iterator(end(), end()));
617}

619static Value *foldSelectInst(SelectInst &SI) {
// If the condition being selected on is a constant or the same value is
// being selected between, fold the select. Yes this does (rarely) happen
// early on.
if (ConstantInt *CI = dyn_cast<ConstantInt>(SI.getCondition()))
  return SI.getOperand(1 + CI->isZero());
if (SI.getOperand(1) == SI.getOperand(2))
  return SI.getOperand(1);

return nullptr;
629}

631/// A helper that folds a PHI node or a select.
632static Value *foldPHINodeOrSelectInst(Instruction &I) {
if (PHINode *PN = dyn_cast<PHINode>(&I)) {
  // If PN merges together the same value, return that value.
  return PN->hasConstantValue();
}
return foldSelectInst(cast<SelectInst>(I));
638}

640/// Builder for the alloca slices.
641///
642/// This class builds a set of alloca slices by recursively visiting the uses
643/// of an alloca and making a slice for each load and store at each offset.
644class AllocaSlices::SliceBuilder : public PtrUseVisitor<SliceBuilder> {
friend class PtrUseVisitor<SliceBuilder>;
friend class InstVisitor<SliceBuilder>;

using Base = PtrUseVisitor<SliceBuilder>;

const uint64_t AllocSize;
AllocaSlices &AS;

SmallDenseMap<Instruction *, unsigned> MemTransferSliceMap;
SmallDenseMap<Instruction *, uint64_t> PHIOrSelectSizes;

/// Set to de-duplicate dead instructions found in the use walk.
SmallPtrSet<Instruction *, 4> VisitedDeadInsts;

659public:
SliceBuilder(const DataLayout &DL, AllocaInst &AI, AllocaSlices &AS)
    : PtrUseVisitor<SliceBuilder>(DL),
      AllocSize(DL.getTypeAllocSize(AI.getAllocatedType()).getFixedSize()),
      AS(AS) {}

665private:
void markAsDead(Instruction &I) {
  if (VisitedDeadInsts.insert(&I).second)
    AS.DeadUsers.push_back(&I);
}

void insertUse(Instruction &I, const APInt &Offset, uint64_t Size,
               bool IsSplittable = false) {
  // Completely skip uses which have a zero size or start either before or
  // past the end of the allocation.
  if (Size == 0 || Offset.uge(AllocSize)) {
    LLVM_DEBUG(dbgs() << "WARNING: Ignoring " << Size << " byte use @"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte use @" << Offset << " which has zero size or starts outside of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << Offsetdo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte use @" << Offset << " which has zero size or starts outside of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << " which has zero size or starts outside of the "do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte use @" << Offset << " which has zero size or starts outside of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << AllocSize << " byte alloca:\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte use @" << Offset << " which has zero size or starts outside of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << "    alloca: " << AS.AI << "\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte use @" << Offset << " which has zero size or starts outside of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << "       use: " << I << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte use @" << Offset << " which has zero size or starts outside of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false);
    return markAsDead(I);
  }

  uint64_t BeginOffset = Offset.getZExtValue();
  uint64_t EndOffset = BeginOffset + Size;

  // Clamp the end offset to the end of the allocation. Note that this is
  // formulated to handle even the case where "BeginOffset + Size" overflows.
  // This may appear superficially to be something we could ignore entirely,
  // but that is not so! There may be widened loads or PHI-node uses where
  // some instructions are dead but not others. We can't completely ignore
  // them, and so have to record at least the information here.
  assert(AllocSize >= BeginOffset)(static_cast <bool> (AllocSize >= BeginOffset) ? void
 (0) : __assert_fail ("AllocSize >= BeginOffset", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 694, __extension__ __PRETTY_FUNCTION__)); // Established above.
  if (Size > AllocSize - BeginOffset) {
    LLVM_DEBUG(dbgs() << "WARNING: Clamping a " << Size << " byte use @"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Clamping a " << Size
 << " byte use @" << Offset << " to remain within the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << Offset << " to remain within the " << AllocSizedo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Clamping a " << Size
 << " byte use @" << Offset << " to remain within the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << " byte alloca:\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Clamping a " << Size
 << " byte use @" << Offset << " to remain within the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << "    alloca: " << AS.AI << "\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Clamping a " << Size
 << " byte use @" << Offset << " to remain within the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false)
                      << "       use: " << I << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Clamping a " << Size
 << " byte use @" << Offset << " to remain within the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 I << "\n"; } } while (false);
    EndOffset = AllocSize;
  }

  AS.Slices.push_back(Slice(BeginOffset, EndOffset, U, IsSplittable));
}

void visitBitCastInst(BitCastInst &BC) {
  if (BC.use_empty())
    return markAsDead(BC);

  return Base::visitBitCastInst(BC);
}

void visitAddrSpaceCastInst(AddrSpaceCastInst &ASC) {
  if (ASC.use_empty())
    return markAsDead(ASC);

  return Base::visitAddrSpaceCastInst(ASC);
}

void visitGetElementPtrInst(GetElementPtrInst &GEPI) {
  if (GEPI.use_empty())
    return markAsDead(GEPI);

  if (SROAStrictInbounds && GEPI.isInBounds()) {
    // FIXME: This is a manually un-factored variant of the basic code inside
    // of GEPs with checking of the inbounds invariant specified in the
    // langref in a very strict sense. If we ever want to enable
    // SROAStrictInbounds, this code should be factored cleanly into
    // PtrUseVisitor, but it is easier to experiment with SROAStrictInbounds
    // by writing out the code here where we have the underlying allocation
    // size readily available.
    APInt GEPOffset = Offset;
    const DataLayout &DL = GEPI.getModule()->getDataLayout();
    for (gep_type_iterator GTI = gep_type_begin(GEPI),
                           GTE = gep_type_end(GEPI);
         GTI != GTE; ++GTI) {
      ConstantInt *OpC = dyn_cast<ConstantInt>(GTI.getOperand());
      if (!OpC)
        break;

      // Handle a struct index, which adds its field offset to the pointer.
      if (StructType *STy = GTI.getStructTypeOrNull()) {
        unsigned ElementIdx = OpC->getZExtValue();
        const StructLayout *SL = DL.getStructLayout(STy);
        GEPOffset +=
            APInt(Offset.getBitWidth(), SL->getElementOffset(ElementIdx));
      } else {
        // For array or vector indices, scale the index by the size of the
        // type.
        APInt Index = OpC->getValue().sextOrTrunc(Offset.getBitWidth());
        GEPOffset +=
            Index *
            APInt(Offset.getBitWidth(),
                  DL.getTypeAllocSize(GTI.getIndexedType()).getFixedSize());
      }

      // If this index has computed an intermediate pointer which is not
      // inbounds, then the result of the GEP is a poison value and we can
      // delete it and all uses.
      if (GEPOffset.ugt(AllocSize))
        return markAsDead(GEPI);
    }
  }

  return Base::visitGetElementPtrInst(GEPI);
}

void handleLoadOrStore(Type *Ty, Instruction &I, const APInt &Offset,
                       uint64_t Size, bool IsVolatile) {
  // We allow splitting of non-volatile loads and stores where the type is an
  // integer type. These may be used to implement 'memcpy' or other "transfer
  // of bits" patterns.
  bool IsSplittable =
      Ty->isIntegerTy() && !IsVolatile && DL.typeSizeEqualsStoreSize(Ty);

  insertUse(I, Offset, Size, IsSplittable);
}

void visitLoadInst(LoadInst &LI) {
  assert((!LI.isSimple() || LI.getType()->isSingleValueType()) &&(static_cast <bool> ((!LI.isSimple() || LI.getType()->
isSingleValueType()) && "All simple FCA loads should have been pre-split"
) ? void (0) : __assert_fail ("(!LI.isSimple() || LI.getType()->isSingleValueType()) && \"All simple FCA loads should have been pre-split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 782, __extension__ __PRETTY_FUNCTION__
))
         "All simple FCA loads should have been pre-split")(static_cast <bool> ((!LI.isSimple() || LI.getType()->
isSingleValueType()) && "All simple FCA loads should have been pre-split"
) ? void (0) : __assert_fail ("(!LI.isSimple() || LI.getType()->isSingleValueType()) && \"All simple FCA loads should have been pre-split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 782, __extension__ __PRETTY_FUNCTION__
));

  if (!IsOffsetKnown)
    return PI.setAborted(&LI);

  if (isa<ScalableVectorType>(LI.getType()))
    return PI.setAborted(&LI);

  uint64_t Size = DL.getTypeStoreSize(LI.getType()).getFixedSize();
  return handleLoadOrStore(LI.getType(), LI, Offset, Size, LI.isVolatile());
}

void visitStoreInst(StoreInst &SI) {
  Value *ValOp = SI.getValueOperand();
  if (ValOp == *U)
    return PI.setEscapedAndAborted(&SI);
  if (!IsOffsetKnown)
    return PI.setAborted(&SI);

  if (isa<ScalableVectorType>(ValOp->getType()))
    return PI.setAborted(&SI);

  uint64_t Size = DL.getTypeStoreSize(ValOp->getType()).getFixedSize();

  // If this memory access can be shown to *statically* extend outside the
  // bounds of the allocation, it's behavior is undefined, so simply
  // ignore it. Note that this is more strict than the generic clamping
  // behavior of insertUse. We also try to handle cases which might run the
  // risk of overflow.
  // FIXME: We should instead consider the pointer to have escaped if this
  // function is being instrumented for addressing bugs or race conditions.
  if (Size > AllocSize || Offset.ugt(AllocSize - Size)) {
    LLVM_DEBUG(dbgs() << "WARNING: Ignoring " << Size << " byte store @"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte store @" << Offset << " which extends past the end of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 SI << "\n"; } } while (false)
                      << Offset << " which extends past the end of the "do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte store @" << Offset << " which extends past the end of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 SI << "\n"; } } while (false)
                      << AllocSize << " byte alloca:\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte store @" << Offset << " which extends past the end of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 SI << "\n"; } } while (false)
                      << "    alloca: " << AS.AI << "\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte store @" << Offset << " which extends past the end of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 SI << "\n"; } } while (false)
                      << "       use: " << SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "WARNING: Ignoring " << Size
 << " byte store @" << Offset << " which extends past the end of the "
 << AllocSize << " byte alloca:\n" << "    alloca: "
 << AS.AI << "\n" << "       use: " <<
 SI << "\n"; } } while (false);
    return markAsDead(SI);
  }

  assert((!SI.isSimple() || ValOp->getType()->isSingleValueType()) &&(static_cast <bool> ((!SI.isSimple() || ValOp->getType
()->isSingleValueType()) && "All simple FCA stores should have been pre-split"
) ? void (0) : __assert_fail ("(!SI.isSimple() || ValOp->getType()->isSingleValueType()) && \"All simple FCA stores should have been pre-split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 823, __extension__ __PRETTY_FUNCTION__
))
         "All simple FCA stores should have been pre-split")(static_cast <bool> ((!SI.isSimple() || ValOp->getType
()->isSingleValueType()) && "All simple FCA stores should have been pre-split"
) ? void (0) : __assert_fail ("(!SI.isSimple() || ValOp->getType()->isSingleValueType()) && \"All simple FCA stores should have been pre-split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 823, __extension__ __PRETTY_FUNCTION__
));
  handleLoadOrStore(ValOp->getType(), SI, Offset, Size, SI.isVolatile());
}

void visitMemSetInst(MemSetInst &II) {
  assert(II.getRawDest() == *U && "Pointer use is not the destination?")(static_cast <bool> (II.getRawDest() == *U && "Pointer use is not the destination?"
) ? void (0) : __assert_fail ("II.getRawDest() == *U && \"Pointer use is not the destination?\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 828, __extension__ __PRETTY_FUNCTION__
));
  ConstantInt *Length = dyn_cast<ConstantInt>(II.getLength());
  if ((Length && Length->getValue() == 0) ||
      (IsOffsetKnown && Offset.uge(AllocSize)))
    // Zero-length mem transfer intrinsics can be ignored entirely.
    return markAsDead(II);

  if (!IsOffsetKnown)
    return PI.setAborted(&II);

  insertUse(II, Offset, Length ? Length->getLimitedValue()
                               : AllocSize - Offset.getLimitedValue(),
            (bool)Length);
}

void visitMemTransferInst(MemTransferInst &II) {
  ConstantInt *Length = dyn_cast<ConstantInt>(II.getLength());
  if (Length && Length->getValue() == 0)
    // Zero-length mem transfer intrinsics can be ignored entirely.
    return markAsDead(II);

  // Because we can visit these intrinsics twice, also check to see if the
  // first time marked this instruction as dead. If so, skip it.
  if (VisitedDeadInsts.count(&II))
    return;

  if (!IsOffsetKnown)
    return PI.setAborted(&II);

  // This side of the transfer is completely out-of-bounds, and so we can
  // nuke the entire transfer. However, we also need to nuke the other side
  // if already added to our partitions.
  // FIXME: Yet another place we really should bypass this when
  // instrumenting for ASan.
  if (Offset.uge(AllocSize)) {
    SmallDenseMap<Instruction *, unsigned>::iterator MTPI =
        MemTransferSliceMap.find(&II);
    if (MTPI != MemTransferSliceMap.end())
      AS.Slices[MTPI->second].kill();
    return markAsDead(II);
  }

  uint64_t RawOffset = Offset.getLimitedValue();
  uint64_t Size = Length ? Length->getLimitedValue() : AllocSize - RawOffset;

  // Check for the special case where the same exact value is used for both
  // source and dest.
  if (*U == II.getRawDest() && *U == II.getRawSource()) {
    // For non-volatile transfers this is a no-op.
    if (!II.isVolatile())
      return markAsDead(II);

    return insertUse(II, Offset, Size, /*IsSplittable=*/false);
  }

  // If we have seen both source and destination for a mem transfer, then
  // they both point to the same alloca.
  bool Inserted;
  SmallDenseMap<Instruction *, unsigned>::iterator MTPI;
  std::tie(MTPI, Inserted) =
      MemTransferSliceMap.insert(std::make_pair(&II, AS.Slices.size()));
  unsigned PrevIdx = MTPI->second;
  if (!Inserted) {
    Slice &PrevP = AS.Slices[PrevIdx];

    // Check if the begin offsets match and this is a non-volatile transfer.
    // In that case, we can completely elide the transfer.
    if (!II.isVolatile() && PrevP.beginOffset() == RawOffset) {
      PrevP.kill();
      return markAsDead(II);
    }

    // Otherwise we have an offset transfer within the same alloca. We can't
    // split those.
    PrevP.makeUnsplittable();
  }

  // Insert the use now that we've fixed up the splittable nature.
  insertUse(II, Offset, Size, /*IsSplittable=*/Inserted && Length);

  // Check that we ended up with a valid index in the map.
  assert(AS.Slices[PrevIdx].getUse()->getUser() == &II &&(static_cast <bool> (AS.Slices[PrevIdx].getUse()->getUser
() == &II && "Map index doesn't point back to a slice with this user."
) ? void (0) : __assert_fail ("AS.Slices[PrevIdx].getUse()->getUser() == &II && \"Map index doesn't point back to a slice with this user.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 910, __extension__ __PRETTY_FUNCTION__
))
         "Map index doesn't point back to a slice with this user.")(static_cast <bool> (AS.Slices[PrevIdx].getUse()->getUser
() == &II && "Map index doesn't point back to a slice with this user."
) ? void (0) : __assert_fail ("AS.Slices[PrevIdx].getUse()->getUser() == &II && \"Map index doesn't point back to a slice with this user.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 910, __extension__ __PRETTY_FUNCTION__
));
}

// Disable SRoA for any intrinsics except for lifetime invariants and
// invariant group.
// FIXME: What about debug intrinsics? This matches old behavior, but
// doesn't make sense.
void visitIntrinsicInst(IntrinsicInst &II) {
  if (II.isDroppable()) {
    AS.DeadUseIfPromotable.push_back(U);
    return;
  }

  if (!IsOffsetKnown)
    return PI.setAborted(&II);

  if (II.isLifetimeStartOrEnd()) {
    ConstantInt *Length = cast<ConstantInt>(II.getArgOperand(0));
    uint64_t Size = std::min(AllocSize - Offset.getLimitedValue(),
                             Length->getLimitedValue());
    insertUse(II, Offset, Size, true);
    return;
  }

  if (II.isLaunderOrStripInvariantGroup()) {
    enqueueUsers(II);
    return;
  }

  Base::visitIntrinsicInst(II);
}

Instruction *hasUnsafePHIOrSelectUse(Instruction *Root, uint64_t &Size) {
  // We consider any PHI or select that results in a direct load or store of
  // the same offset to be a viable use for slicing purposes. These uses
  // are considered unsplittable and the size is the maximum loaded or stored
  // size.
  SmallPtrSet<Instruction *, 4> Visited;
  SmallVector<std::pair<Instruction *, Instruction *>, 4> Uses;
  Visited.insert(Root);
  Uses.push_back(std::make_pair(cast<Instruction>(*U), Root));
  const DataLayout &DL = Root->getModule()->getDataLayout();
  // If there are no loads or stores, the access is dead. We mark that as
  // a size zero access.
  Size = 0;
  do {
    Instruction *I, *UsedI;
    std::tie(UsedI, I) = Uses.pop_back_val();

    if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
      Size = std::max(Size,
                      DL.getTypeStoreSize(LI->getType()).getFixedSize());
      continue;
    }
    if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
      Value *Op = SI->getOperand(0);
      if (Op == UsedI)
        return SI;
      Size = std::max(Size,
                      DL.getTypeStoreSize(Op->getType()).getFixedSize());
      continue;
    }

    if (GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(I)) {
      if (!GEP->hasAllZeroIndices())
        return GEP;
    } else if (!isa<BitCastInst>(I) && !isa<PHINode>(I) &&
               !isa<SelectInst>(I) && !isa<AddrSpaceCastInst>(I)) {
      return I;
    }

    for (User *U : I->users())
      if (Visited.insert(cast<Instruction>(U)).second)
        Uses.push_back(std::make_pair(I, cast<Instruction>(U)));
  } while (!Uses.empty());

  return nullptr;
}

void visitPHINodeOrSelectInst(Instruction &I) {
  assert(isa<PHINode>(I) || isa<SelectInst>(I))(static_cast <bool> (isa<PHINode>(I) || isa<SelectInst
>(I)) ? void (0) : __assert_fail ("isa<PHINode>(I) || isa<SelectInst>(I)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 990, __extension__ __PRETTY_FUNCTION__
));
  if (I.use_empty())
    return markAsDead(I);

  // If this is a PHI node before a catchswitch, we cannot insert any non-PHI
  // instructions in this BB, which may be required during rewriting. Bail out
  // on these cases.
  if (isa<PHINode>(I) &&
      I.getParent()->getFirstInsertionPt() == I.getParent()->end())
    return PI.setAborted(&I);

  // TODO: We could use simplifyInstruction here to fold PHINodes and
  // SelectInsts. However, doing so requires to change the current
  // dead-operand-tracking mechanism. For instance, suppose neither loading
  // from %U nor %other traps. Then "load (select undef, %U, %other)" does not
  // trap either.  However, if we simply replace %U with undef using the
  // current dead-operand-tracking mechanism, "load (select undef, undef,
  // %other)" may trap because the select may return the first operand
  // "undef".
  if (Value *Result = foldPHINodeOrSelectInst(I)) {
    if (Result == *U)
      // If the result of the constant fold will be the pointer, recurse
      // through the PHI/select as if we had RAUW'ed it.
      enqueueUsers(I);
    else
      // Otherwise the operand to the PHI/select is dead, and we can replace
      // it with poison.
      AS.DeadOperands.push_back(U);

    return;
  }

  if (!IsOffsetKnown)
    return PI.setAborted(&I);

  // See if we already have computed info on this node.
  uint64_t &Size = PHIOrSelectSizes[&I];
  if (!Size) {
    // This is a new PHI/Select, check for an unsafe use of it.
    if (Instruction *UnsafeI = hasUnsafePHIOrSelectUse(&I, Size))
      return PI.setAborted(UnsafeI);
  }

  // For PHI and select operands outside the alloca, we can't nuke the entire
  // phi or select -- the other side might still be relevant, so we special
  // case them here and use a separate structure to track the operands
  // themselves which should be replaced with poison.
  // FIXME: This should instead be escaped in the event we're instrumenting
  // for address sanitization.
  if (Offset.uge(AllocSize)) {
    AS.DeadOperands.push_back(U);
    return;
  }

  insertUse(I, Offset, Size);
}

void visitPHINode(PHINode &PN) { visitPHINodeOrSelectInst(PN); }

void visitSelectInst(SelectInst &SI) { visitPHINodeOrSelectInst(SI); }

/// Disable SROA entirely if there are unhandled users of the alloca.
void visitInstruction(Instruction &I) { PI.setAborted(&I); }
1053};

1055AllocaSlices::AllocaSlices(const DataLayout &DL, AllocaInst &AI)
  :
1057#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
    AI(AI),
1059#endif
    PointerEscapingInstr(nullptr) {
SliceBuilder PB(DL, AI, *this);
SliceBuilder::PtrInfo PtrI = PB.visitPtr(AI);
if (PtrI.isEscaped() || PtrI.isAborted()) {
  // FIXME: We should sink the escape vs. abort info into the caller nicely,
  // possibly by just storing the PtrInfo in the AllocaSlices.
  PointerEscapingInstr = PtrI.getEscapingInst() ? PtrI.getEscapingInst()
                                                : PtrI.getAbortingInst();
  assert(PointerEscapingInstr && "Did not track a bad instruction")(static_cast <bool> (PointerEscapingInstr && "Did not track a bad instruction"
) ? void (0) : __assert_fail ("PointerEscapingInstr && \"Did not track a bad instruction\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1068, __extension__ __PRETTY_FUNCTION__
));
  return;
}

llvm::erase_if(Slices, [](const Slice &S) { return S.isDead(); });

// Sort the uses. This arranges for the offsets to be in ascending order,
// and the sizes to be in descending order.
llvm::stable_sort(Slices);
1077}

1079#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)

1081void AllocaSlices::print(raw_ostream &OS, const_iterator I,
                       StringRef Indent) const {
printSlice(OS, I, Indent);
OS << "\n";
printUse(OS, I, Indent);
1086}

1088void AllocaSlices::printSlice(raw_ostream &OS, const_iterator I,
                            StringRef Indent) const {
OS << Indent << "[" << I->beginOffset() << "," << I->endOffset() << ")"
   << " slice #" << (I - begin())
   << (I->isSplittable() ? " (splittable)" : "");
1093}

1095void AllocaSlices::printUse(raw_ostream &OS, const_iterator I,
                          StringRef Indent) const {
OS << Indent << "  used by: " << *I->getUse()->getUser() << "\n";
1098}

1100void AllocaSlices::print(raw_ostream &OS) const {
if (PointerEscapingInstr) {
  OS << "Can't analyze slices for alloca: " << AI << "\n"
     << "  A pointer to this alloca escaped by:\n"
     << "  " << *PointerEscapingInstr << "\n";
  return;
}

OS << "Slices of alloca: " << AI << "\n";
for (const_iterator I = begin(), E = end(); I != E; ++I)
  print(OS, I);
1111}

1113LLVM_DUMP_METHOD__attribute__((noinline)) __attribute__((__used__)) void AllocaSlices::dump(const_iterator I) const {
print(dbgs(), I);
1115}
1116LLVM_DUMP_METHOD__attribute__((noinline)) __attribute__((__used__)) void AllocaSlices::dump() const { print(dbgs()); }

1118#endif // !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)

1120/// Walk the range of a partitioning looking for a common type to cover this
1121/// sequence of slices.
1122static std::pair<Type *, IntegerType *>
1123findCommonType(AllocaSlices::const_iterator B, AllocaSlices::const_iterator E,
             uint64_t EndOffset) {
Type *Ty = nullptr;
bool TyIsCommon = true;
IntegerType *ITy = nullptr;

// Note that we need to look at *every* alloca slice's Use to ensure we
// always get consistent results regardless of the order of slices.
for (AllocaSlices::const_iterator I = B; I != E; ++I) {
  Use *U = I->getUse();
  if (isa<IntrinsicInst>(*U->getUser()))
    continue;
  if (I->beginOffset() != B->beginOffset() || I->endOffset() != EndOffset)
    continue;

  Type *UserTy = nullptr;
  if (LoadInst *LI = dyn_cast<LoadInst>(U->getUser())) {
    UserTy = LI->getType();
  } else if (StoreInst *SI = dyn_cast<StoreInst>(U->getUser())) {
    UserTy = SI->getValueOperand()->getType();
  }

  if (IntegerType *UserITy = dyn_cast_or_null<IntegerType>(UserTy)) {
    // If the type is larger than the partition, skip it. We only encounter
    // this for split integer operations where we want to use the type of the
    // entity causing the split. Also skip if the type is not a byte width
    // multiple.
    if (UserITy->getBitWidth() % 8 != 0 ||
        UserITy->getBitWidth() / 8 > (EndOffset - B->beginOffset()))
      continue;

    // Track the largest bitwidth integer type used in this way in case there
    // is no common type.
    if (!ITy || ITy->getBitWidth() < UserITy->getBitWidth())
      ITy = UserITy;
  }

  // To avoid depending on the order of slices, Ty and TyIsCommon must not
  // depend on types skipped above.
  if (!UserTy || (Ty && Ty != UserTy))
    TyIsCommon = false; // Give up on anything but an iN type.
  else
    Ty = UserTy;
}

return {TyIsCommon ? Ty : nullptr, ITy};
1169}

1171/// PHI instructions that use an alloca and are subsequently loaded can be
1172/// rewritten to load both input pointers in the pred blocks and then PHI the
1173/// results, allowing the load of the alloca to be promoted.
1174/// From this:
1175///   %P2 = phi [i32* %Alloca, i32* %Other]
1176///   %V = load i32* %P2
1177/// to:
1178///   %V1 = load i32* %Alloca      -> will be mem2reg'd
1179///   ...
1180///   %V2 = load i32* %Other
1181///   ...
1182///   %V = phi [i32 %V1, i32 %V2]
1183///
1184/// We can do this to a select if its only uses are loads and if the operands
1185/// to the select can be loaded unconditionally.
1186///
1187/// FIXME: This should be hoisted into a generic utility, likely in
1188/// Transforms/Util/Local.h
1189static bool isSafePHIToSpeculate(PHINode &PN) {
const DataLayout &DL = PN.getModule()->getDataLayout();

// For now, we can only do this promotion if the load is in the same block
// as the PHI, and if there are no stores between the phi and load.
// TODO: Allow recursive phi users.
// TODO: Allow stores.
BasicBlock *BB = PN.getParent();
Align MaxAlign;
uint64_t APWidth = DL.getIndexTypeSizeInBits(PN.getType());
Type *LoadType = nullptr;
for (User *U : PN.users()) {
  LoadInst *LI = dyn_cast<LoadInst>(U);
  if (!LI || !LI->isSimple())
    return false;

  // For now we only allow loads in the same block as the PHI.  This is
  // a common case that happens when instcombine merges two loads through
  // a PHI.
  if (LI->getParent() != BB)
    return false;

  if (LoadType) {
    if (LoadType != LI->getType())
      return false;
  } else {
    LoadType = LI->getType();
  }

  // Ensure that there are no instructions between the PHI and the load that
  // could store.
  for (BasicBlock::iterator BBI(PN); &*BBI != LI; ++BBI)
    if (BBI->mayWriteToMemory())
      return false;

  MaxAlign = std::max(MaxAlign, LI->getAlign());
}

if (!LoadType)
  return false;

APInt LoadSize = APInt(APWidth, DL.getTypeStoreSize(LoadType).getFixedSize());

// We can only transform this if it is safe to push the loads into the
// predecessor blocks. The only thing to watch out for is that we can't put
// a possibly trapping load in the predecessor if it is a critical edge.
for (unsigned Idx = 0, Num = PN.getNumIncomingValues(); Idx != Num; ++Idx) {
  Instruction *TI = PN.getIncomingBlock(Idx)->getTerminator();
  Value *InVal = PN.getIncomingValue(Idx);

  // If the value is produced by the terminator of the predecessor (an
  // invoke) or it has side-effects, there is no valid place to put a load
  // in the predecessor.
  if (TI == InVal || TI->mayHaveSideEffects())
    return false;

  // If the predecessor has a single successor, then the edge isn't
  // critical.
  if (TI->getNumSuccessors() == 1)
    continue;

  // If this pointer is always safe to load, or if we can prove that there
  // is already a load in the block, then we can move the load to the pred
  // block.
  if (isSafeToLoadUnconditionally(InVal, MaxAlign, LoadSize, DL, TI))
    continue;

  return false;
}

return true;
1260}

1262static void speculatePHINodeLoads(IRBuilderTy &IRB, PHINode &PN) {
LLVM_DEBUG(dbgs() << "    original: " << PN << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << PN <<
 "\n"; } } while (false);

LoadInst *SomeLoad = cast<LoadInst>(PN.user_back());
Type *LoadTy = SomeLoad->getType();
IRB.SetInsertPoint(&PN);
PHINode *NewPN = IRB.CreatePHI(LoadTy, PN.getNumIncomingValues(),
                               PN.getName() + ".sroa.speculated");

// Get the AA tags and alignment to use from one of the loads. It does not
// matter which one we get and if any differ.
AAMDNodes AATags = SomeLoad->getAAMetadata();
Align Alignment = SomeLoad->getAlign();

// Rewrite all loads of the PN to use the new PHI.
while (!PN.use_empty()) {
  LoadInst *LI = cast<LoadInst>(PN.user_back());
  LI->replaceAllUsesWith(NewPN);
  LI->eraseFromParent();
}

// Inject loads into all of the pred blocks.
DenseMap<BasicBlock*, Value*> InjectedLoads;
for (unsigned Idx = 0, Num = PN.getNumIncomingValues(); Idx != Num; ++Idx) {
  BasicBlock *Pred = PN.getIncomingBlock(Idx);
  Value *InVal = PN.getIncomingValue(Idx);

  // A PHI node is allowed to have multiple (duplicated) entries for the same
  // basic block, as long as the value is the same. So if we already injected
  // a load in the predecessor, then we should reuse the same load for all
  // duplicated entries.
  if (Value* V = InjectedLoads.lookup(Pred)) {
    NewPN->addIncoming(V, Pred);
    continue;
  }

  Instruction *TI = Pred->getTerminator();
  IRB.SetInsertPoint(TI);

  LoadInst *Load = IRB.CreateAlignedLoad(
      LoadTy, InVal, Alignment,
      (PN.getName() + ".sroa.speculate.load." + Pred->getName()));
  ++NumLoadsSpeculated;
  if (AATags)
    Load->setAAMetadata(AATags);
  NewPN->addIncoming(Load, Pred);
  InjectedLoads[Pred] = Load;
}

LLVM_DEBUG(dbgs() << "          speculated to: " << *NewPN << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          speculated to: " <<
 *NewPN << "\n"; } } while (false);
PN.eraseFromParent();
1313}

1315sroa::SelectHandSpeculativity &
1316sroa::SelectHandSpeculativity::setAsSpeculatable(bool isTrueVal) {
if (isTrueVal)
  Bitfield::set<sroa::SelectHandSpeculativity::TrueVal>(Storage, true);
else
  Bitfield::set<sroa::SelectHandSpeculativity::FalseVal>(Storage, true);
return *this;
1322}

1324bool sroa::SelectHandSpeculativity::isSpeculatable(bool isTrueVal) const {
return isTrueVal
           ? Bitfield::get<sroa::SelectHandSpeculativity::TrueVal>(Storage)
           : Bitfield::get<sroa::SelectHandSpeculativity::FalseVal>(Storage);
1328}

1330bool sroa::SelectHandSpeculativity::areAllSpeculatable() const {
return isSpeculatable(/*isTrueVal=*/true) &&
       isSpeculatable(/*isTrueVal=*/false);
1333}

1335bool sroa::SelectHandSpeculativity::areAnySpeculatable() const {
return isSpeculatable(/*isTrueVal=*/true) ||
       isSpeculatable(/*isTrueVal=*/false);
1338}
1339bool sroa::SelectHandSpeculativity::areNoneSpeculatable() const {
return !areAnySpeculatable();
1341}

1343static sroa::SelectHandSpeculativity
1344isSafeLoadOfSelectToSpeculate(LoadInst &LI, SelectInst &SI, bool PreserveCFG) {
assert(LI.isSimple() && "Only for simple loads")(static_cast <bool> (LI.isSimple() && "Only for simple loads"
) ? void (0) : __assert_fail ("LI.isSimple() && \"Only for simple loads\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1345, __extension__ __PRETTY_FUNCTION__
));
sroa::SelectHandSpeculativity Spec;

const DataLayout &DL = SI.getModule()->getDataLayout();
for (Value *Value : {SI.getTrueValue(), SI.getFalseValue()})
  if (isSafeToLoadUnconditionally(Value, LI.getType(), LI.getAlign(), DL,
                                  &LI))
    Spec.setAsSpeculatable(/*isTrueVal=*/Value == SI.getTrueValue());
  else if (PreserveCFG)
    return Spec;

return Spec;
1357}

1359std::optional<sroa::RewriteableMemOps>
1360SROAPass::isSafeSelectToSpeculate(SelectInst &SI, bool PreserveCFG) {
RewriteableMemOps Ops;

for (User *U : SI.users()) {
  if (auto *BC = dyn_cast<BitCastInst>(U); BC && BC->hasOneUse())
    U = *BC->user_begin();

  if (auto *Store = dyn_cast<StoreInst>(U)) {
    // Note that atomic stores can be transformed; atomic semantics do not
    // have any meaning for a local alloca. Stores are not speculatable,
    // however, so if we can't turn it into a predicated store, we are done.
    if (Store->isVolatile() || PreserveCFG)
      return {}; // Give up on this `select`.
    Ops.emplace_back(Store);
    continue;
  }

  auto *LI = dyn_cast<LoadInst>(U);

  // Note that atomic loads can be transformed;
  // atomic semantics do not have any meaning for a local alloca.
  if (!LI || LI->isVolatile())
    return {}; // Give up on this `select`.

  PossiblySpeculatableLoad Load(LI);
  if (!LI->isSimple()) {
    // If the `load` is not simple, we can't speculatively execute it,
    // but we could handle this via a CFG modification. But can we?
    if (PreserveCFG)
      return {}; // Give up on this `select`.
    Ops.emplace_back(Load);
    continue;
  }

  sroa::SelectHandSpeculativity Spec =
      isSafeLoadOfSelectToSpeculate(*LI, SI, PreserveCFG);
  if (PreserveCFG && !Spec.areAllSpeculatable())
    return {}; // Give up on this `select`.

  Load.setInt(Spec);
  Ops.emplace_back(Load);
}

return Ops;
1404}

1406static void speculateSelectInstLoads(SelectInst &SI, LoadInst &LI,
                                   IRBuilderTy &IRB) {
LLVM_DEBUG(dbgs() << "    original load: " << SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original load: " << SI
 << "\n"; } } while (false);

IRB.SetInsertPoint(&SI);
Value *TV = SI.getTrueValue();
Value *FV = SI.getFalseValue();
// Replace the given load of the select with a select of two loads.

assert(LI.isSimple() && "We only speculate simple loads")(static_cast <bool> (LI.isSimple() && "We only speculate simple loads"
) ? void (0) : __assert_fail ("LI.isSimple() && \"We only speculate simple loads\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1415, __extension__ __PRETTY_FUNCTION__
));

IRB.SetInsertPoint(&LI);
LoadInst *TL =
    IRB.CreateAlignedLoad(LI.getType(), TV, LI.getAlign(),
                          LI.getName() + ".sroa.speculate.load.true");
LoadInst *FL =
    IRB.CreateAlignedLoad(LI.getType(), FV, LI.getAlign(),
                          LI.getName() + ".sroa.speculate.load.false");
NumLoadsSpeculated += 2;

// Transfer alignment and AA info if present.
TL->setAlignment(LI.getAlign());
FL->setAlignment(LI.getAlign());

AAMDNodes Tags = LI.getAAMetadata();
if (Tags) {
  TL->setAAMetadata(Tags);
  FL->setAAMetadata(Tags);
}

Value *V = IRB.CreateSelect(SI.getCondition(), TL, FL,
                            LI.getName() + ".sroa.speculated");

LLVM_DEBUG(dbgs() << "          speculated to: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          speculated to: " <<
 *V << "\n"; } } while (false);
LI.replaceAllUsesWith(V);
1441}

1443template <typename T>
1444static void rewriteMemOpOfSelect(SelectInst &SI, T &I,
                               sroa::SelectHandSpeculativity Spec,
                               DomTreeUpdater &DTU) {
assert((isa<LoadInst>(I) || isa<StoreInst>(I)) && "Only for load and store!")(static_cast <bool> ((isa<LoadInst>(I) || isa<
StoreInst>(I)) && "Only for load and store!") ? void
 (0) : __assert_fail ("(isa<LoadInst>(I) || isa<StoreInst>(I)) && \"Only for load and store!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1447, __extension__ __PRETTY_FUNCTION__
));
LLVM_DEBUG(dbgs() << "    original mem op: " << I << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original mem op: " << I
 << "\n"; } } while (false);
BasicBlock *Head = I.getParent();
Instruction *ThenTerm = nullptr;
Instruction *ElseTerm = nullptr;
if (Spec.areNoneSpeculatable())
  SplitBlockAndInsertIfThenElse(SI.getCondition(), &I, &ThenTerm, &ElseTerm,
                                SI.getMetadata(LLVMContext::MD_prof), &DTU);
else {
  SplitBlockAndInsertIfThen(SI.getCondition(), &I, /*Unreachable=*/false,
                            SI.getMetadata(LLVMContext::MD_prof), &DTU,
                            /*LI=*/nullptr, /*ThenBlock=*/nullptr);
  if (Spec.isSpeculatable(/*isTrueVal=*/true))
    cast<BranchInst>(Head->getTerminator())->swapSuccessors();
}
auto *HeadBI = cast<BranchInst>(Head->getTerminator());
Spec = {}; // Do not use `Spec` beyond this point.
BasicBlock *Tail = I.getParent();
Tail->setName(Head->getName() + ".cont");
PHINode *PN;
if (isa<LoadInst>(I))
  PN = PHINode::Create(I.getType(), 2, "", &I);
for (BasicBlock *SuccBB : successors(Head)) {
  bool IsThen = SuccBB == HeadBI->getSuccessor(0);
  int SuccIdx = IsThen ? 0 : 1;
  auto *NewMemOpBB = SuccBB == Tail ? Head : SuccBB;
  if (NewMemOpBB != Head) {
    NewMemOpBB->setName(Head->getName() + (IsThen ? ".then" : ".else"));
    if (isa<LoadInst>(I))
      ++NumLoadsPredicated;
    else
      ++NumStoresPredicated;
  } else
    ++NumLoadsSpeculated;
  auto &CondMemOp = cast<T>(*I.clone());
  CondMemOp.insertBefore(NewMemOpBB->getTerminator());
  Value *Ptr = SI.getOperand(1 + SuccIdx);
  if (auto *PtrTy = Ptr->getType();
      !PtrTy->isOpaquePointerTy() &&
      PtrTy != CondMemOp.getPointerOperandType())
    Ptr = BitCastInst::CreatePointerBitCastOrAddrSpaceCast(
        Ptr, CondMemOp.getPointerOperandType(), "", &CondMemOp);
  CondMemOp.setOperand(I.getPointerOperandIndex(), Ptr);
  if (isa<LoadInst>(I)) {
    CondMemOp.setName(I.getName() + (IsThen ? ".then" : ".else") + ".val");
    PN->addIncoming(&CondMemOp, NewMemOpBB);
  } else
    LLVM_DEBUG(dbgs() << "                 to: " << CondMemOp << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "                 to: " << CondMemOp
 << "\n"; } } while (false);
}
if (isa<LoadInst>(I)) {
  PN->takeName(&I);
  LLVM_DEBUG(dbgs() << "          to: " << *PN << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *PN <<
 "\n"; } } while (false);
  I.replaceAllUsesWith(PN);
}
1501}

1503static void rewriteMemOpOfSelect(SelectInst &SelInst, Instruction &I,
                               sroa::SelectHandSpeculativity Spec,
                               DomTreeUpdater &DTU) {
if (auto *LI = dyn_cast<LoadInst>(&I))
  rewriteMemOpOfSelect(SelInst, *LI, Spec, DTU);
else if (auto *SI = dyn_cast<StoreInst>(&I))
  rewriteMemOpOfSelect(SelInst, *SI, Spec, DTU);
else
  llvm_unreachable_internal("Only for load and store.");
1512}

1514static bool rewriteSelectInstMemOps(SelectInst &SI,
                                  const sroa::RewriteableMemOps &Ops,
                                  IRBuilderTy &IRB, DomTreeUpdater *DTU) {
bool CFGChanged = false;
LLVM_DEBUG(dbgs() << "    original select: " << SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original select: " << SI
 << "\n"; } } while (false);

for (const RewriteableMemOp &Op : Ops) {
  sroa::SelectHandSpeculativity Spec;
  Instruction *I;
  if (auto *const *US = std::get_if<UnspeculatableStore>(&Op)) {
    I = *US;
  } else {
    auto PSL = std::get<PossiblySpeculatableLoad>(Op);
    I = PSL.getPointer();
    Spec = PSL.getInt();
  }
  if (Spec.areAllSpeculatable()) {
    speculateSelectInstLoads(SI, cast<LoadInst>(*I), IRB);
  } else {
    assert(DTU && "Should not get here when not allowed to modify the CFG!")(static_cast <bool> (DTU && "Should not get here when not allowed to modify the CFG!"
) ? void (0) : __assert_fail ("DTU && \"Should not get here when not allowed to modify the CFG!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1533, __extension__ __PRETTY_FUNCTION__
));
    rewriteMemOpOfSelect(SI, *I, Spec, *DTU);
    CFGChanged = true;
  }
  I->eraseFromParent();
}

for (User *U : make_early_inc_range(SI.users()))
  cast<BitCastInst>(U)->eraseFromParent();
SI.eraseFromParent();
return CFGChanged;
1544}

1546/// Build a GEP out of a base pointer and indices.
1547///
1548/// This will return the BasePtr if that is valid, or build a new GEP
1549/// instruction using the IRBuilder if GEP-ing is needed.
1550static Value *buildGEP(IRBuilderTy &IRB, Value *BasePtr,
                     SmallVectorImpl<Value *> &Indices,
                     const Twine &NamePrefix) {
if (Indices.empty())
  return BasePtr;

// A single zero index is a no-op, so check for this and avoid building a GEP
// in that case.
if (Indices.size() == 1 && cast<ConstantInt>(Indices.back())->isZero())
  return BasePtr;

// buildGEP() is only called for non-opaque pointers.
return IRB.CreateInBoundsGEP(
    BasePtr->getType()->getNonOpaquePointerElementType(), BasePtr, Indices,
    NamePrefix + "sroa_idx");
1565}

1567/// Get a natural GEP off of the BasePtr walking through Ty toward
1568/// TargetTy without changing the offset of the pointer.
1569///
1570/// This routine assumes we've already established a properly offset GEP with
1571/// Indices, and arrived at the Ty type. The goal is to continue to GEP with
1572/// zero-indices down through type layers until we find one the same as
1573/// TargetTy. If we can't find one with the same type, we at least try to use
1574/// one with the same size. If none of that works, we just produce the GEP as
1575/// indicated by Indices to have the correct offset.
1576static Value *getNaturalGEPWithType(IRBuilderTy &IRB, const DataLayout &DL,
                                  Value *BasePtr, Type *Ty, Type *TargetTy,
                                  SmallVectorImpl<Value *> &Indices,
                                  const Twine &NamePrefix) {
if (Ty == TargetTy)
  return buildGEP(IRB, BasePtr, Indices, NamePrefix);

// Offset size to use for the indices.
unsigned OffsetSize = DL.getIndexTypeSizeInBits(BasePtr->getType());

// See if we can descend into a struct and locate a field with the correct
// type.
unsigned NumLayers = 0;
Type *ElementTy = Ty;
do {
  if (ElementTy->isPointerTy())
    break;

  if (ArrayType *ArrayTy = dyn_cast<ArrayType>(ElementTy)) {
    ElementTy = ArrayTy->getElementType();
    Indices.push_back(IRB.getIntN(OffsetSize, 0));
  } else if (VectorType *VectorTy = dyn_cast<VectorType>(ElementTy)) {
    ElementTy = VectorTy->getElementType();
    Indices.push_back(IRB.getInt32(0));
  } else if (StructType *STy = dyn_cast<StructType>(ElementTy)) {
    if (STy->element_begin() == STy->element_end())
      break; // Nothing left to descend into.
    ElementTy = *STy->element_begin();
    Indices.push_back(IRB.getInt32(0));
  } else {
    break;
  }
  ++NumLayers;
} while (ElementTy != TargetTy);
if (ElementTy != TargetTy)
  Indices.erase(Indices.end() - NumLayers, Indices.end());

return buildGEP(IRB, BasePtr, Indices, NamePrefix);
1614}

1616/// Get a natural GEP from a base pointer to a particular offset and
1617/// resulting in a particular type.
1618///
1619/// The goal is to produce a "natural" looking GEP that works with the existing
1620/// composite types to arrive at the appropriate offset and element type for
1621/// a pointer. TargetTy is the element type the returned GEP should point-to if
1622/// possible. We recurse by decreasing Offset, adding the appropriate index to
1623/// Indices, and setting Ty to the result subtype.
1624///
1625/// If no natural GEP can be constructed, this function returns null.
1626static Value *getNaturalGEPWithOffset(IRBuilderTy &IRB, const DataLayout &DL,
                                    Value *Ptr, APInt Offset, Type *TargetTy,
                                    SmallVectorImpl<Value *> &Indices,
                                    const Twine &NamePrefix) {
PointerType *Ty = cast<PointerType>(Ptr->getType());

// Don't consider any GEPs through an i8* as natural unless the TargetTy is
// an i8.
if (Ty == IRB.getInt8PtrTy(Ty->getAddressSpace()) && TargetTy->isIntegerTy(8))
  return nullptr;

Type *ElementTy = Ty->getNonOpaquePointerElementType();
if (!ElementTy->isSized())
  return nullptr; // We can't GEP through an unsized element.

SmallVector<APInt> IntIndices = DL.getGEPIndicesForOffset(ElementTy, Offset);
if (Offset != 0)
  return nullptr;

for (const APInt &Index : IntIndices)
  Indices.push_back(IRB.getInt(Index));
return getNaturalGEPWithType(IRB, DL, Ptr, ElementTy, TargetTy, Indices,
                             NamePrefix);
1649}

1651/// Compute an adjusted pointer from Ptr by Offset bytes where the
1652/// resulting pointer has PointerTy.
1653///
1654/// This tries very hard to compute a "natural" GEP which arrives at the offset
1655/// and produces the pointer type desired. Where it cannot, it will try to use
1656/// the natural GEP to arrive at the offset and bitcast to the type. Where that
1657/// fails, it will try to use an existing i8* and GEP to the byte offset and
1658/// bitcast to the type.
1659///
1660/// The strategy for finding the more natural GEPs is to peel off layers of the
1661/// pointer, walking back through bit casts and GEPs, searching for a base
1662/// pointer from which we can compute a natural GEP with the desired
1663/// properties. The algorithm tries to fold as many constant indices into
1664/// a single GEP as possible, thus making each GEP more independent of the
1665/// surrounding code.
1666static Value *getAdjustedPtr(IRBuilderTy &IRB, const DataLayout &DL, Value *Ptr,
                           APInt Offset, Type *PointerTy,
                           const Twine &NamePrefix) {
// Create i8 GEP for opaque pointers.
if (Ptr->getType()->isOpaquePointerTy()) {
  if (Offset != 0)
    Ptr = IRB.CreateInBoundsGEP(IRB.getInt8Ty(), Ptr, IRB.getInt(Offset),
                                NamePrefix + "sroa_idx");
  return IRB.CreatePointerBitCastOrAddrSpaceCast(Ptr, PointerTy,
                                                 NamePrefix + "sroa_cast");
}

// Even though we don't look through PHI nodes, we could be called on an
// instruction in an unreachable block, which may be on a cycle.
SmallPtrSet<Value *, 4> Visited;
Visited.insert(Ptr);
SmallVector<Value *, 4> Indices;

// We may end up computing an offset pointer that has the wrong type. If we
// never are able to compute one directly that has the correct type, we'll
// fall back to it, so keep it and the base it was computed from around here.
Value *OffsetPtr = nullptr;
Value *OffsetBasePtr;

// Remember any i8 pointer we come across to re-use if we need to do a raw
// byte offset.
Value *Int8Ptr = nullptr;
APInt Int8PtrOffset(Offset.getBitWidth(), 0);

PointerType *TargetPtrTy = cast<PointerType>(PointerTy);
Type *TargetTy = TargetPtrTy->getNonOpaquePointerElementType();

// As `addrspacecast` is , `Ptr` (the storage pointer) may have different
// address space from the expected `PointerTy` (the pointer to be used).
// Adjust the pointer type based the original storage pointer.
auto AS = cast<PointerType>(Ptr->getType())->getAddressSpace();
PointerTy = TargetTy->getPointerTo(AS);

do {
  // First fold any existing GEPs into the offset.
  while (GEPOperator *GEP = dyn_cast<GEPOperator>(Ptr)) {
    APInt GEPOffset(Offset.getBitWidth(), 0);
    if (!GEP->accumulateConstantOffset(DL, GEPOffset))
      break;
    Offset += GEPOffset;
    Ptr = GEP->getPointerOperand();
    if (!Visited.insert(Ptr).second)
      break;
  }

  // See if we can perform a natural GEP here.
  Indices.clear();
  if (Value *P = getNaturalGEPWithOffset(IRB, DL, Ptr, Offset, TargetTy,
                                         Indices, NamePrefix)) {
    // If we have a new natural pointer at the offset, clear out any old
    // offset pointer we computed. Unless it is the base pointer or
    // a non-instruction, we built a GEP we don't need. Zap it.
    if (OffsetPtr && OffsetPtr != OffsetBasePtr)
      if (Instruction *I = dyn_cast<Instruction>(OffsetPtr)) {
        assert(I->use_empty() && "Built a GEP with uses some how!")(static_cast <bool> (I->use_empty() && "Built a GEP with uses some how!"
) ? void (0) : __assert_fail ("I->use_empty() && \"Built a GEP with uses some how!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1725, __extension__ __PRETTY_FUNCTION__
));
        I->eraseFromParent();
      }
    OffsetPtr = P;
    OffsetBasePtr = Ptr;
    // If we also found a pointer of the right type, we're done.
    if (P->getType() == PointerTy)
      break;
  }

  // Stash this pointer if we've found an i8*.
  if (Ptr->getType()->isIntegerTy(8)) {
    Int8Ptr = Ptr;
    Int8PtrOffset = Offset;
  }

  // Peel off a layer of the pointer and update the offset appropriately.
  if (Operator::getOpcode(Ptr) == Instruction::BitCast) {
    Ptr = cast<Operator>(Ptr)->getOperand(0);
  } else if (GlobalAlias *GA = dyn_cast<GlobalAlias>(Ptr)) {
    if (GA->isInterposable())
      break;
    Ptr = GA->getAliasee();
  } else {
    break;
  }
  assert(Ptr->getType()->isPointerTy() && "Unexpected operand type!")(static_cast <bool> (Ptr->getType()->isPointerTy(
) && "Unexpected operand type!") ? void (0) : __assert_fail
 ("Ptr->getType()->isPointerTy() && \"Unexpected operand type!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1751, __extension__ __PRETTY_FUNCTION__
));
} while (Visited.insert(Ptr).second);

if (!OffsetPtr) {
  if (!Int8Ptr) {
    Int8Ptr = IRB.CreateBitCast(
        Ptr, IRB.getInt8PtrTy(PointerTy->getPointerAddressSpace()),
        NamePrefix + "sroa_raw_cast");
    Int8PtrOffset = Offset;
  }

  OffsetPtr = Int8PtrOffset == 0
                  ? Int8Ptr
                  : IRB.CreateInBoundsGEP(IRB.getInt8Ty(), Int8Ptr,
                                          IRB.getInt(Int8PtrOffset),
                                          NamePrefix + "sroa_raw_idx");
}
Ptr = OffsetPtr;

// On the off chance we were targeting i8*, guard the bitcast here.
if (cast<PointerType>(Ptr->getType()) != TargetPtrTy) {
  Ptr = IRB.CreatePointerBitCastOrAddrSpaceCast(Ptr,
                                                TargetPtrTy,
                                                NamePrefix + "sroa_cast");
}

return Ptr;
1778}

1780/// Compute the adjusted alignment for a load or store from an offset.
1781static Align getAdjustedAlignment(Instruction *I, uint64_t Offset) {
return commonAlignment(getLoadStoreAlignment(I), Offset);
1783}

1785/// Test whether we can convert a value from the old to the new type.
1786///
1787/// This predicate should be used to guard calls to convertValue in order to
1788/// ensure that we only try to convert viable values. The strategy is that we
1789/// will peel off single element struct and array wrappings to get to an
1790/// underlying value, and convert that value.
1791static bool canConvertValue(const DataLayout &DL, Type *OldTy, Type *NewTy) {
if (OldTy == NewTy)
  return true;

// For integer types, we can't handle any bit-width differences. This would
// break both vector conversions with extension and introduce endianness
// issues when in conjunction with loads and stores.
if (isa<IntegerType>(OldTy) && isa<IntegerType>(NewTy)) {
  assert(cast<IntegerType>(OldTy)->getBitWidth() !=(static_cast <bool> (cast<IntegerType>(OldTy)->
getBitWidth() != cast<IntegerType>(NewTy)->getBitWidth
() && "We can't have the same bitwidth for different int types"
) ? void (0) : __assert_fail ("cast<IntegerType>(OldTy)->getBitWidth() != cast<IntegerType>(NewTy)->getBitWidth() && \"We can't have the same bitwidth for different int types\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1801, __extension__ __PRETTY_FUNCTION__
))
             cast<IntegerType>(NewTy)->getBitWidth() &&(static_cast <bool> (cast<IntegerType>(OldTy)->
getBitWidth() != cast<IntegerType>(NewTy)->getBitWidth
() && "We can't have the same bitwidth for different int types"
) ? void (0) : __assert_fail ("cast<IntegerType>(OldTy)->getBitWidth() != cast<IntegerType>(NewTy)->getBitWidth() && \"We can't have the same bitwidth for different int types\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1801, __extension__ __PRETTY_FUNCTION__
))
         "We can't have the same bitwidth for different int types")(static_cast <bool> (cast<IntegerType>(OldTy)->
getBitWidth() != cast<IntegerType>(NewTy)->getBitWidth
() && "We can't have the same bitwidth for different int types"
) ? void (0) : __assert_fail ("cast<IntegerType>(OldTy)->getBitWidth() != cast<IntegerType>(NewTy)->getBitWidth() && \"We can't have the same bitwidth for different int types\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1801, __extension__ __PRETTY_FUNCTION__
));
  return false;
}

if (DL.getTypeSizeInBits(NewTy).getFixedSize() !=
    DL.getTypeSizeInBits(OldTy).getFixedSize())
  return false;
if (!NewTy->isSingleValueType() || !OldTy->isSingleValueType())
  return false;

// We can convert pointers to integers and vice-versa. Same for vectors
// of pointers and integers.
OldTy = OldTy->getScalarType();
NewTy = NewTy->getScalarType();
if (NewTy->isPointerTy() || OldTy->isPointerTy()) {
  if (NewTy->isPointerTy() && OldTy->isPointerTy()) {
    unsigned OldAS = OldTy->getPointerAddressSpace();
    unsigned NewAS = NewTy->getPointerAddressSpace();
    // Convert pointers if they are pointers from the same address space or
    // different integral (not non-integral) address spaces with the same
    // pointer size.
    return OldAS == NewAS ||
           (!DL.isNonIntegralAddressSpace(OldAS) &&
            !DL.isNonIntegralAddressSpace(NewAS) &&
            DL.getPointerSize(OldAS) == DL.getPointerSize(NewAS));
  }

  // We can convert integers to integral pointers, but not to non-integral
  // pointers.
  if (OldTy->isIntegerTy())
    return !DL.isNonIntegralPointerType(NewTy);

  // We can convert integral pointers to integers, but non-integral pointers
  // need to remain pointers.
  if (!DL.isNonIntegralPointerType(OldTy))
    return NewTy->isIntegerTy();

  return false;
}

return true;
1842}

1844/// Generic routine to convert an SSA value to a value of a different
1845/// type.
1846///
1847/// This will try various different casting techniques, such as bitcasts,
1848/// inttoptr, and ptrtoint casts. Use the \c canConvertValue predicate to test
1849/// two types for viability with this routine.
1850static Value *convertValue(const DataLayout &DL, IRBuilderTy &IRB, Value *V,
                         Type *NewTy) {
Type *OldTy = V->getType();
assert(canConvertValue(DL, OldTy, NewTy) && "Value not convertable to type")(static_cast <bool> (canConvertValue(DL, OldTy, NewTy) &&
 "Value not convertable to type") ? void (0) : __assert_fail (
"canConvertValue(DL, OldTy, NewTy) && \"Value not convertable to type\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1853, __extension__ __PRETTY_FUNCTION__
));

if (OldTy == NewTy)
  return V;

assert(!(isa<IntegerType>(OldTy) && isa<IntegerType>(NewTy)) &&(static_cast <bool> (!(isa<IntegerType>(OldTy) &&
 isa<IntegerType>(NewTy)) && "Integer types must be the exact same to convert."
) ? void (0) : __assert_fail ("!(isa<IntegerType>(OldTy) && isa<IntegerType>(NewTy)) && \"Integer types must be the exact same to convert.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1859, __extension__ __PRETTY_FUNCTION__
))
       "Integer types must be the exact same to convert.")(static_cast <bool> (!(isa<IntegerType>(OldTy) &&
 isa<IntegerType>(NewTy)) && "Integer types must be the exact same to convert."
) ? void (0) : __assert_fail ("!(isa<IntegerType>(OldTy) && isa<IntegerType>(NewTy)) && \"Integer types must be the exact same to convert.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1859, __extension__ __PRETTY_FUNCTION__
));

// See if we need inttoptr for this type pair. May require additional bitcast.
if (OldTy->isIntOrIntVectorTy() && NewTy->isPtrOrPtrVectorTy()) {
  // Expand <2 x i32> to i8* --> <2 x i32> to i64 to i8*
  // Expand i128 to <2 x i8*> --> i128 to <2 x i64> to <2 x i8*>
  // Expand <4 x i32> to <2 x i8*> --> <4 x i32> to <2 x i64> to <2 x i8*>
  // Directly handle i64 to i8*
  return IRB.CreateIntToPtr(IRB.CreateBitCast(V, DL.getIntPtrType(NewTy)),
                            NewTy);
}

// See if we need ptrtoint for this type pair. May require additional bitcast.
if (OldTy->isPtrOrPtrVectorTy() && NewTy->isIntOrIntVectorTy()) {
  // Expand <2 x i8*> to i128 --> <2 x i8*> to <2 x i64> to i128
  // Expand i8* to <2 x i32> --> i8* to i64 to <2 x i32>
  // Expand <2 x i8*> to <4 x i32> --> <2 x i8*> to <2 x i64> to <4 x i32>
  // Expand i8* to i64 --> i8* to i64 to i64
  return IRB.CreateBitCast(IRB.CreatePtrToInt(V, DL.getIntPtrType(OldTy)),
                           NewTy);
}

if (OldTy->isPtrOrPtrVectorTy() && NewTy->isPtrOrPtrVectorTy()) {
  unsigned OldAS = OldTy->getPointerAddressSpace();
  unsigned NewAS = NewTy->getPointerAddressSpace();
  // To convert pointers with different address spaces (they are already
  // checked convertible, i.e. they have the same pointer size), so far we
  // cannot use `bitcast` (which has restrict on the same address space) or
  // `addrspacecast` (which is not always no-op casting). Instead, use a pair
  // of no-op `ptrtoint`/`inttoptr` casts through an integer with the same bit
  // size.
  if (OldAS != NewAS) {
    assert(DL.getPointerSize(OldAS) == DL.getPointerSize(NewAS))(static_cast <bool> (DL.getPointerSize(OldAS) == DL.getPointerSize
(NewAS)) ? void (0) : __assert_fail ("DL.getPointerSize(OldAS) == DL.getPointerSize(NewAS)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1891, __extension__ __PRETTY_FUNCTION__
));
    return IRB.CreateIntToPtr(IRB.CreatePtrToInt(V, DL.getIntPtrType(OldTy)),
                              NewTy);
  }
}

return IRB.CreateBitCast(V, NewTy);
1898}

1900/// Test whether the given slice use can be promoted to a vector.
1901///
1902/// This function is called to test each entry in a partition which is slated
1903/// for a single slice.
1904static bool isVectorPromotionViableForSlice(Partition &P, const Slice &S,
                                          VectorType *Ty,
                                          uint64_t ElementSize,
                                          const DataLayout &DL) {
// First validate the slice offsets.
uint64_t BeginOffset =
    std::max(S.beginOffset(), P.beginOffset()) - P.beginOffset();
uint64_t BeginIndex = BeginOffset / ElementSize;
if (BeginIndex * ElementSize != BeginOffset ||
    BeginIndex >= cast<FixedVectorType>(Ty)->getNumElements())
  return false;
uint64_t EndOffset =
    std::min(S.endOffset(), P.endOffset()) - P.beginOffset();
uint64_t EndIndex = EndOffset / ElementSize;
if (EndIndex * ElementSize != EndOffset ||
    EndIndex > cast<FixedVectorType>(Ty)->getNumElements())
  return false;

assert(EndIndex > BeginIndex && "Empty vector!")(static_cast <bool> (EndIndex > BeginIndex &&
 "Empty vector!") ? void (0) : __assert_fail ("EndIndex > BeginIndex && \"Empty vector!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1922, __extension__ __PRETTY_FUNCTION__
));
uint64_t NumElements = EndIndex - BeginIndex;
Type *SliceTy = (NumElements == 1)
                    ? Ty->getElementType()
                    : FixedVectorType::get(Ty->getElementType(), NumElements);

Type *SplitIntTy =
    Type::getIntNTy(Ty->getContext(), NumElements * ElementSize * 8);

Use *U = S.getUse();

if (MemIntrinsic *MI = dyn_cast<MemIntrinsic>(U->getUser())) {
  if (MI->isVolatile())
    return false;
  if (!S.isSplittable())
    return false; // Skip any unsplittable intrinsics.
} else if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(U->getUser())) {
  if (!II->isLifetimeStartOrEnd() && !II->isDroppable())
    return false;
} else if (LoadInst *LI = dyn_cast<LoadInst>(U->getUser())) {
  if (LI->isVolatile())
    return false;
  Type *LTy = LI->getType();
  // Disable vector promotion when there are loads or stores of an FCA.
  if (LTy->isStructTy())
    return false;
  if (P.beginOffset() > S.beginOffset() || P.endOffset() < S.endOffset()) {
    assert(LTy->isIntegerTy())(static_cast <bool> (LTy->isIntegerTy()) ? void (0) :
 __assert_fail ("LTy->isIntegerTy()", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 1949, __extension__ __PRETTY_FUNCTION__));
    LTy = SplitIntTy;
  }
  if (!canConvertValue(DL, SliceTy, LTy))
    return false;
} else if (StoreInst *SI = dyn_cast<StoreInst>(U->getUser())) {
  if (SI->isVolatile())
    return false;
  Type *STy = SI->getValueOperand()->getType();
  // Disable vector promotion when there are loads or stores of an FCA.
  if (STy->isStructTy())
    return false;
  if (P.beginOffset() > S.beginOffset() || P.endOffset() < S.endOffset()) {
    assert(STy->isIntegerTy())(static_cast <bool> (STy->isIntegerTy()) ? void (0) :
 __assert_fail ("STy->isIntegerTy()", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 1962, __extension__ __PRETTY_FUNCTION__));
    STy = SplitIntTy;
  }
  if (!canConvertValue(DL, STy, SliceTy))
    return false;
} else {
  return false;
}

return true;
1972}

1974/// Test whether a vector type is viable for promotion.
1975///
1976/// This implements the necessary checking for \c isVectorPromotionViable over
1977/// all slices of the alloca for the given VectorType.
1978static bool checkVectorTypeForPromotion(Partition &P, VectorType *VTy,
                                      const DataLayout &DL) {
uint64_t ElementSize =
    DL.getTypeSizeInBits(VTy->getElementType()).getFixedSize();

// While the definition of LLVM vectors is bitpacked, we don't support sizes
// that aren't byte sized.
if (ElementSize % 8)
  return false;
assert((DL.getTypeSizeInBits(VTy).getFixedSize() % 8) == 0 &&(static_cast <bool> ((DL.getTypeSizeInBits(VTy).getFixedSize
() % 8) == 0 && "vector size not a multiple of element size?"
) ? void (0) : __assert_fail ("(DL.getTypeSizeInBits(VTy).getFixedSize() % 8) == 0 && \"vector size not a multiple of element size?\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1988, __extension__ __PRETTY_FUNCTION__
))
       "vector size not a multiple of element size?")(static_cast <bool> ((DL.getTypeSizeInBits(VTy).getFixedSize
() % 8) == 0 && "vector size not a multiple of element size?"
) ? void (0) : __assert_fail ("(DL.getTypeSizeInBits(VTy).getFixedSize() % 8) == 0 && \"vector size not a multiple of element size?\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 1988, __extension__ __PRETTY_FUNCTION__
));
ElementSize /= 8;

for (const Slice &S : P)
  if (!isVectorPromotionViableForSlice(P, S, VTy, ElementSize, DL))
    return false;

for (const Slice *S : P.splitSliceTails())
  if (!isVectorPromotionViableForSlice(P, *S, VTy, ElementSize, DL))
    return false;

return true;
2000}

2002/// Test whether the given alloca partitioning and range of slices can be
2003/// promoted to a vector.
2004///
2005/// This is a quick test to check whether we can rewrite a particular alloca
2006/// partition (and its newly formed alloca) into a vector alloca with only
2007/// whole-vector loads and stores such that it could be promoted to a vector
2008/// SSA value. We only can ensure this for a limited set of operations, and we
2009/// don't want to do the rewrites unless we are confident that the result will
2010/// be promotable, so we have an early test here.
2011static VectorType *isVectorPromotionViable(Partition &P, const DataLayout &DL) {
// Collect the candidate types for vector-based promotion. Also track whether
// we have different element types.
SmallVector<VectorType *, 4> CandidateTys;
Type *CommonEltTy = nullptr;
VectorType *CommonVecPtrTy = nullptr;
bool HaveVecPtrTy = false;
bool HaveCommonEltTy = true;
bool HaveCommonVecPtrTy = true;
auto CheckCandidateType = [&](Type *Ty) {
  if (auto *VTy = dyn_cast<VectorType>(Ty)) {
    // Return if bitcast to vectors is different for total size in bits.
    if (!CandidateTys.empty()) {
      VectorType *V = CandidateTys[0];
      if (DL.getTypeSizeInBits(VTy).getFixedSize() !=
          DL.getTypeSizeInBits(V).getFixedSize()) {
        CandidateTys.clear();
        return;
      }
    }
    CandidateTys.push_back(VTy);
    Type *EltTy = VTy->getElementType();

    if (!CommonEltTy)
      CommonEltTy = EltTy;
    else if (CommonEltTy != EltTy)
      HaveCommonEltTy = false;

    if (EltTy->isPointerTy()) {
      HaveVecPtrTy = true;
      if (!CommonVecPtrTy)
        CommonVecPtrTy = VTy;
      else if (CommonVecPtrTy != VTy)
        HaveCommonVecPtrTy = false;
    }
  }
};
// Consider any loads or stores that are the exact size of the slice.
for (const Slice &S : P)
  if (S.beginOffset() == P.beginOffset() &&
      S.endOffset() == P.endOffset()) {
    if (auto *LI = dyn_cast<LoadInst>(S.getUse()->getUser()))
      CheckCandidateType(LI->getType());
    else if (auto *SI = dyn_cast<StoreInst>(S.getUse()->getUser()))
      CheckCandidateType(SI->getValueOperand()->getType());
  }

// If we didn't find a vector type, nothing to do here.
if (CandidateTys.empty())
  return nullptr;

// Pointer-ness is sticky, if we had a vector-of-pointers candidate type,
// then we should choose it, not some other alternative.
// But, we can't perform a no-op pointer address space change via bitcast,
// so if we didn't have a common pointer element type, bail.
if (HaveVecPtrTy && !HaveCommonVecPtrTy)
  return nullptr;

// Try to pick the "best" element type out of the choices.
if (!HaveCommonEltTy && HaveVecPtrTy) {
  // If there was a pointer element type, there's really only one choice.
  CandidateTys.clear();
  CandidateTys.push_back(CommonVecPtrTy);
} else if (!HaveCommonEltTy && !HaveVecPtrTy) {
  // Integer-ify vector types.
  for (VectorType *&VTy : CandidateTys) {
    if (!VTy->getElementType()->isIntegerTy())
      VTy = cast<VectorType>(VTy->getWithNewType(IntegerType::getIntNTy(
          VTy->getContext(), VTy->getScalarSizeInBits())));
  }

  // Rank the remaining candidate vector types. This is easy because we know
  // they're all integer vectors. We sort by ascending number of elements.
  auto RankVectorTypes = [&DL](VectorType *RHSTy, VectorType *LHSTy) {
    (void)DL;
    assert(DL.getTypeSizeInBits(RHSTy).getFixedSize() ==(static_cast <bool> (DL.getTypeSizeInBits(RHSTy).getFixedSize
() == DL.getTypeSizeInBits(LHSTy).getFixedSize() && "Cannot have vector types of different sizes!"
) ? void (0) : __assert_fail ("DL.getTypeSizeInBits(RHSTy).getFixedSize() == DL.getTypeSizeInBits(LHSTy).getFixedSize() && \"Cannot have vector types of different sizes!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2088, __extension__ __PRETTY_FUNCTION__
))
               DL.getTypeSizeInBits(LHSTy).getFixedSize() &&(static_cast <bool> (DL.getTypeSizeInBits(RHSTy).getFixedSize
() == DL.getTypeSizeInBits(LHSTy).getFixedSize() && "Cannot have vector types of different sizes!"
) ? void (0) : __assert_fail ("DL.getTypeSizeInBits(RHSTy).getFixedSize() == DL.getTypeSizeInBits(LHSTy).getFixedSize() && \"Cannot have vector types of different sizes!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2088, __extension__ __PRETTY_FUNCTION__
))
           "Cannot have vector types of different sizes!")(static_cast <bool> (DL.getTypeSizeInBits(RHSTy).getFixedSize
() == DL.getTypeSizeInBits(LHSTy).getFixedSize() && "Cannot have vector types of different sizes!"
) ? void (0) : __assert_fail ("DL.getTypeSizeInBits(RHSTy).getFixedSize() == DL.getTypeSizeInBits(LHSTy).getFixedSize() && \"Cannot have vector types of different sizes!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2088, __extension__ __PRETTY_FUNCTION__
));
    assert(RHSTy->getElementType()->isIntegerTy() &&(static_cast <bool> (RHSTy->getElementType()->isIntegerTy
() && "All non-integer types eliminated!") ? void (0)
 : __assert_fail ("RHSTy->getElementType()->isIntegerTy() && \"All non-integer types eliminated!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2090, __extension__ __PRETTY_FUNCTION__
))
           "All non-integer types eliminated!")(static_cast <bool> (RHSTy->getElementType()->isIntegerTy
() && "All non-integer types eliminated!") ? void (0)
 : __assert_fail ("RHSTy->getElementType()->isIntegerTy() && \"All non-integer types eliminated!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2090, __extension__ __PRETTY_FUNCTION__
));
    assert(LHSTy->getElementType()->isIntegerTy() &&(static_cast <bool> (LHSTy->getElementType()->isIntegerTy
() && "All non-integer types eliminated!") ? void (0)
 : __assert_fail ("LHSTy->getElementType()->isIntegerTy() && \"All non-integer types eliminated!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2092, __extension__ __PRETTY_FUNCTION__
))
           "All non-integer types eliminated!")(static_cast <bool> (LHSTy->getElementType()->isIntegerTy
() && "All non-integer types eliminated!") ? void (0)
 : __assert_fail ("LHSTy->getElementType()->isIntegerTy() && \"All non-integer types eliminated!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2092, __extension__ __PRETTY_FUNCTION__
));
    return cast<FixedVectorType>(RHSTy)->getNumElements() <
           cast<FixedVectorType>(LHSTy)->getNumElements();
  };
  llvm::sort(CandidateTys, RankVectorTypes);
  CandidateTys.erase(
      std::unique(CandidateTys.begin(), CandidateTys.end(), RankVectorTypes),
      CandidateTys.end());
} else {
2101// The only way to have the same element type in every vector type is to
2102// have the same vector type. Check that and remove all but one.
2103#ifndef NDEBUG
  for (VectorType *VTy : CandidateTys) {
    assert(VTy->getElementType() == CommonEltTy &&(static_cast <bool> (VTy->getElementType() == CommonEltTy
 && "Unaccounted for element type!") ? void (0) : __assert_fail
 ("VTy->getElementType() == CommonEltTy && \"Unaccounted for element type!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2106, __extension__ __PRETTY_FUNCTION__
))
           "Unaccounted for element type!")(static_cast <bool> (VTy->getElementType() == CommonEltTy
 && "Unaccounted for element type!") ? void (0) : __assert_fail
 ("VTy->getElementType() == CommonEltTy && \"Unaccounted for element type!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2106, __extension__ __PRETTY_FUNCTION__
));
    assert(VTy == CandidateTys[0] &&(static_cast <bool> (VTy == CandidateTys[0] && "Different vector types with the same element type!"
) ? void (0) : __assert_fail ("VTy == CandidateTys[0] && \"Different vector types with the same element type!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2108, __extension__ __PRETTY_FUNCTION__
))
           "Different vector types with the same element type!")(static_cast <bool> (VTy == CandidateTys[0] && "Different vector types with the same element type!"
) ? void (0) : __assert_fail ("VTy == CandidateTys[0] && \"Different vector types with the same element type!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2108, __extension__ __PRETTY_FUNCTION__
));
  }
2110#endif
  CandidateTys.resize(1);
}

// FIXME: hack. Do we have a named constant for this?
// SDAG SDNode can't have more than 65535 operands.
llvm::erase_if(CandidateTys, [](VectorType *VTy) {
  return cast<FixedVectorType>(VTy)->getNumElements() >
         std::numeric_limits<unsigned short>::max();
});

for (VectorType *VTy : CandidateTys)
  if (checkVectorTypeForPromotion(P, VTy, DL))
    return VTy;

return nullptr;
2126}

2128/// Test whether a slice of an alloca is valid for integer widening.
2129///
2130/// This implements the necessary checking for the \c isIntegerWideningViable
2131/// test below on a single slice of the alloca.
2132static bool isIntegerWideningViableForSlice(const Slice &S,
                                          uint64_t AllocBeginOffset,
                                          Type *AllocaTy,
                                          const DataLayout &DL,
                                          bool &WholeAllocaOp) {
uint64_t Size = DL.getTypeStoreSize(AllocaTy).getFixedSize();

uint64_t RelBegin = S.beginOffset() - AllocBeginOffset;
uint64_t RelEnd = S.endOffset() - AllocBeginOffset;

Use *U = S.getUse();

// Lifetime intrinsics operate over the whole alloca whose sizes are usually
// larger than other load/store slices (RelEnd > Size). But lifetime are
// always promotable and should not impact other slices' promotability of the
// partition.
if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(U->getUser())) {
  if (II->isLifetimeStartOrEnd() || II->isDroppable())
    return true;
}

// We can't reasonably handle cases where the load or store extends past
// the end of the alloca's type and into its padding.
if (RelEnd > Size)
  return false;

if (LoadInst *LI = dyn_cast<LoadInst>(U->getUser())) {
  if (LI->isVolatile())
    return false;
  // We can't handle loads that extend past the allocated memory.
  if (DL.getTypeStoreSize(LI->getType()).getFixedSize() > Size)
    return false;
  // So far, AllocaSliceRewriter does not support widening split slice tails
  // in rewriteIntegerLoad.
  if (S.beginOffset() < AllocBeginOffset)
    return false;
  // Note that we don't count vector loads or stores as whole-alloca
  // operations which enable integer widening because we would prefer to use
  // vector widening instead.
  if (!isa<VectorType>(LI->getType()) && RelBegin == 0 && RelEnd == Size)
    WholeAllocaOp = true;
  if (IntegerType *ITy = dyn_cast<IntegerType>(LI->getType())) {
    if (ITy->getBitWidth() < DL.getTypeStoreSizeInBits(ITy).getFixedSize())
      return false;
  } else if (RelBegin != 0 || RelEnd != Size ||
             !canConvertValue(DL, AllocaTy, LI->getType())) {
    // Non-integer loads need to be convertible from the alloca type so that
    // they are promotable.
    return false;
  }
} else if (StoreInst *SI = dyn_cast<StoreInst>(U->getUser())) {
  Type *ValueTy = SI->getValueOperand()->getType();
  if (SI->isVolatile())
    return false;
  // We can't handle stores that extend past the allocated memory.
  if (DL.getTypeStoreSize(ValueTy).getFixedSize() > Size)
    return false;
  // So far, AllocaSliceRewriter does not support widening split slice tails
  // in rewriteIntegerStore.
  if (S.beginOffset() < AllocBeginOffset)
    return false;
  // Note that we don't count vector loads or stores as whole-alloca
  // operations which enable integer widening because we would prefer to use
  // vector widening instead.
  if (!isa<VectorType>(ValueTy) && RelBegin == 0 && RelEnd == Size)
    WholeAllocaOp = true;
  if (IntegerType *ITy = dyn_cast<IntegerType>(ValueTy)) {
    if (ITy->getBitWidth() < DL.getTypeStoreSizeInBits(ITy).getFixedSize())
      return false;
  } else if (RelBegin != 0 || RelEnd != Size ||
             !canConvertValue(DL, ValueTy, AllocaTy)) {
    // Non-integer stores need to be convertible to the alloca type so that
    // they are promotable.
    return false;
  }
} else if (MemIntrinsic *MI = dyn_cast<MemIntrinsic>(U->getUser())) {
  if (MI->isVolatile() || !isa<Constant>(MI->getLength()))
    return false;
  if (!S.isSplittable())
    return false; // Skip any unsplittable intrinsics.
} else {
  return false;
}

return true;
2217}

2219/// Test whether the given alloca partition's integer operations can be
2220/// widened to promotable ones.
2221///
2222/// This is a quick test to check whether we can rewrite the integer loads and
2223/// stores to a particular alloca into wider loads and stores and be able to
2224/// promote the resulting alloca.
2225static bool isIntegerWideningViable(Partition &P, Type *AllocaTy,
                                  const DataLayout &DL) {
uint64_t SizeInBits = DL.getTypeSizeInBits(AllocaTy).getFixedSize();
// Don't create integer types larger than the maximum bitwidth.
if (SizeInBits > IntegerType::MAX_INT_BITS)
  return false;

// Don't try to handle allocas with bit-padding.
if (SizeInBits != DL.getTypeStoreSizeInBits(AllocaTy).getFixedSize())
  return false;

// We need to ensure that an integer type with the appropriate bitwidth can
// be converted to the alloca type, whatever that is. We don't want to force
// the alloca itself to have an integer type if there is a more suitable one.
Type *IntTy = Type::getIntNTy(AllocaTy->getContext(), SizeInBits);
if (!canConvertValue(DL, AllocaTy, IntTy) ||
    !canConvertValue(DL, IntTy, AllocaTy))
  return false;

// While examining uses, we ensure that the alloca has a covering load or
// store. We don't want to widen the integer operations only to fail to
// promote due to some other unsplittable entry (which we may make splittable
// later). However, if there are only splittable uses, go ahead and assume
// that we cover the alloca.
// FIXME: We shouldn't consider split slices that happen to start in the
// partition here...
bool WholeAllocaOp = P.empty() && DL.isLegalInteger(SizeInBits);

for (const Slice &S : P)
  if (!isIntegerWideningViableForSlice(S, P.beginOffset(), AllocaTy, DL,
                                       WholeAllocaOp))
    return false;

for (const Slice *S : P.splitSliceTails())
  if (!isIntegerWideningViableForSlice(*S, P.beginOffset(), AllocaTy, DL,
                                       WholeAllocaOp))
    return false;

return WholeAllocaOp;
2264}

2266static Value *extractInteger(const DataLayout &DL, IRBuilderTy &IRB, Value *V,
                           IntegerType *Ty, uint64_t Offset,
                           const Twine &Name) {
LLVM_DEBUG(dbgs() << "       start: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "       start: " << *V <<
 "\n"; } } while (false);
IntegerType *IntTy = cast<IntegerType>(V->getType());
assert(DL.getTypeStoreSize(Ty).getFixedSize() + Offset <=(static_cast <bool> (DL.getTypeStoreSize(Ty).getFixedSize
() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() &&
 "Element extends past full value") ? void (0) : __assert_fail
 ("DL.getTypeStoreSize(Ty).getFixedSize() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() && \"Element extends past full value\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2273, __extension__ __PRETTY_FUNCTION__
))
           DL.getTypeStoreSize(IntTy).getFixedSize() &&(static_cast <bool> (DL.getTypeStoreSize(Ty).getFixedSize
() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() &&
 "Element extends past full value") ? void (0) : __assert_fail
 ("DL.getTypeStoreSize(Ty).getFixedSize() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() && \"Element extends past full value\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2273, __extension__ __PRETTY_FUNCTION__
))
       "Element extends past full value")(static_cast <bool> (DL.getTypeStoreSize(Ty).getFixedSize
() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() &&
 "Element extends past full value") ? void (0) : __assert_fail
 ("DL.getTypeStoreSize(Ty).getFixedSize() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() && \"Element extends past full value\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2273, __extension__ __PRETTY_FUNCTION__
));
uint64_t ShAmt = 8 * Offset;
if (DL.isBigEndian())
  ShAmt = 8 * (DL.getTypeStoreSize(IntTy).getFixedSize() -
               DL.getTypeStoreSize(Ty).getFixedSize() - Offset);
if (ShAmt) {
  V = IRB.CreateLShr(V, ShAmt, Name + ".shift");
  LLVM_DEBUG(dbgs() << "     shifted: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "     shifted: " << *V <<
 "\n"; } } while (false);
}
assert(Ty->getBitWidth() <= IntTy->getBitWidth() &&(static_cast <bool> (Ty->getBitWidth() <= IntTy->
getBitWidth() && "Cannot extract to a larger integer!"
) ? void (0) : __assert_fail ("Ty->getBitWidth() <= IntTy->getBitWidth() && \"Cannot extract to a larger integer!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2283, __extension__ __PRETTY_FUNCTION__
))
       "Cannot extract to a larger integer!")(static_cast <bool> (Ty->getBitWidth() <= IntTy->
getBitWidth() && "Cannot extract to a larger integer!"
) ? void (0) : __assert_fail ("Ty->getBitWidth() <= IntTy->getBitWidth() && \"Cannot extract to a larger integer!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2283, __extension__ __PRETTY_FUNCTION__
));
if (Ty != IntTy) {
  V = IRB.CreateTrunc(V, Ty, Name + ".trunc");
  LLVM_DEBUG(dbgs() << "     trunced: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "     trunced: " << *V <<
 "\n"; } } while (false);
}
return V;
2289}

2291static Value *insertInteger(const DataLayout &DL, IRBuilderTy &IRB, Value *Old,
                          Value *V, uint64_t Offset, const Twine &Name) {
IntegerType *IntTy = cast<IntegerType>(Old->getType());
IntegerType *Ty = cast<IntegerType>(V->getType());
assert(Ty->getBitWidth() <= IntTy->getBitWidth() &&(static_cast <bool> (Ty->getBitWidth() <= IntTy->
getBitWidth() && "Cannot insert a larger integer!") ?
 void (0) : __assert_fail ("Ty->getBitWidth() <= IntTy->getBitWidth() && \"Cannot insert a larger integer!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2296, __extension__ __PRETTY_FUNCTION__
))
       "Cannot insert a larger integer!")(static_cast <bool> (Ty->getBitWidth() <= IntTy->
getBitWidth() && "Cannot insert a larger integer!") ?
 void (0) : __assert_fail ("Ty->getBitWidth() <= IntTy->getBitWidth() && \"Cannot insert a larger integer!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2296, __extension__ __PRETTY_FUNCTION__
));
LLVM_DEBUG(dbgs() << "       start: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "       start: " << *V <<
 "\n"; } } while (false);
if (Ty != IntTy) {
  V = IRB.CreateZExt(V, IntTy, Name + ".ext");
  LLVM_DEBUG(dbgs() << "    extended: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    extended: " << *V <<
 "\n"; } } while (false);
}
assert(DL.getTypeStoreSize(Ty).getFixedSize() + Offset <=(static_cast <bool> (DL.getTypeStoreSize(Ty).getFixedSize
() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() &&
 "Element store outside of alloca store") ? void (0) : __assert_fail
 ("DL.getTypeStoreSize(Ty).getFixedSize() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() && \"Element store outside of alloca store\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2304, __extension__ __PRETTY_FUNCTION__
))
           DL.getTypeStoreSize(IntTy).getFixedSize() &&(static_cast <bool> (DL.getTypeStoreSize(Ty).getFixedSize
() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() &&
 "Element store outside of alloca store") ? void (0) : __assert_fail
 ("DL.getTypeStoreSize(Ty).getFixedSize() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() && \"Element store outside of alloca store\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2304, __extension__ __PRETTY_FUNCTION__
))
       "Element store outside of alloca store")(static_cast <bool> (DL.getTypeStoreSize(Ty).getFixedSize
() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() &&
 "Element store outside of alloca store") ? void (0) : __assert_fail
 ("DL.getTypeStoreSize(Ty).getFixedSize() + Offset <= DL.getTypeStoreSize(IntTy).getFixedSize() && \"Element store outside of alloca store\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2304, __extension__ __PRETTY_FUNCTION__
));
uint64_t ShAmt = 8 * Offset;
if (DL.isBigEndian())
  ShAmt = 8 * (DL.getTypeStoreSize(IntTy).getFixedSize() -
               DL.getTypeStoreSize(Ty).getFixedSize() - Offset);
if (ShAmt) {
  V = IRB.CreateShl(V, ShAmt, Name + ".shift");
  LLVM_DEBUG(dbgs() << "     shifted: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "     shifted: " << *V <<
 "\n"; } } while (false);
}

if (ShAmt || Ty->getBitWidth() < IntTy->getBitWidth()) {
  APInt Mask = ~Ty->getMask().zext(IntTy->getBitWidth()).shl(ShAmt);
  Old = IRB.CreateAnd(Old, Mask, Name + ".mask");
  LLVM_DEBUG(dbgs() << "      masked: " << *Old << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "      masked: " << *Old <<
 "\n"; } } while (false);
  V = IRB.CreateOr(Old, V, Name + ".insert");
  LLVM_DEBUG(dbgs() << "    inserted: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    inserted: " << *V <<
 "\n"; } } while (false);
}
return V;
2322}

2324static Value *extractVector(IRBuilderTy &IRB, Value *V, unsigned BeginIndex,
                          unsigned EndIndex, const Twine &Name) {
auto *VecTy = cast<FixedVectorType>(V->getType());
unsigned NumElements = EndIndex - BeginIndex;
assert(NumElements <= VecTy->getNumElements() && "Too many elements!")(static_cast <bool> (NumElements <= VecTy->getNumElements
() && "Too many elements!") ? void (0) : __assert_fail
 ("NumElements <= VecTy->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2328, __extension__ __PRETTY_FUNCTION__
));

if (NumElements == VecTy->getNumElements())
  return V;

if (NumElements == 1) {
  V = IRB.CreateExtractElement(V, IRB.getInt32(BeginIndex),
                               Name + ".extract");
  LLVM_DEBUG(dbgs() << "     extract: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "     extract: " << *V <<
 "\n"; } } while (false);
  return V;
}

auto Mask = llvm::to_vector<8>(llvm::seq<int>(BeginIndex, EndIndex));
V = IRB.CreateShuffleVector(V, Mask, Name + ".extract");
LLVM_DEBUG(dbgs() << "     shuffle: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "     shuffle: " << *V <<
 "\n"; } } while (false);
return V;
2344}

2346static Value *insertVector(IRBuilderTy &IRB, Value *Old, Value *V,
                         unsigned BeginIndex, const Twine &Name) {
VectorType *VecTy = cast<VectorType>(Old->getType());
assert(VecTy && "Can only insert a vector into a vector")(static_cast <bool> (VecTy && "Can only insert a vector into a vector"
) ? void (0) : __assert_fail ("VecTy && \"Can only insert a vector into a vector\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2349, __extension__ __PRETTY_FUNCTION__
));

VectorType *Ty = dyn_cast<VectorType>(V->getType());
if (!Ty) {
  // Single element to insert.
  V = IRB.CreateInsertElement(Old, V, IRB.getInt32(BeginIndex),
                              Name + ".insert");
  LLVM_DEBUG(dbgs() << "     insert: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "     insert: " << *V <<
 "\n"; } } while (false);
  return V;
}

assert(cast<FixedVectorType>(Ty)->getNumElements() <=(static_cast <bool> (cast<FixedVectorType>(Ty)->
getNumElements() <= cast<FixedVectorType>(VecTy)->
getNumElements() && "Too many elements!") ? void (0) :
 __assert_fail ("cast<FixedVectorType>(Ty)->getNumElements() <= cast<FixedVectorType>(VecTy)->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2362, __extension__ __PRETTY_FUNCTION__
))
           cast<FixedVectorType>(VecTy)->getNumElements() &&(static_cast <bool> (cast<FixedVectorType>(Ty)->
getNumElements() <= cast<FixedVectorType>(VecTy)->
getNumElements() && "Too many elements!") ? void (0) :
 __assert_fail ("cast<FixedVectorType>(Ty)->getNumElements() <= cast<FixedVectorType>(VecTy)->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2362, __extension__ __PRETTY_FUNCTION__
))
       "Too many elements!")(static_cast <bool> (cast<FixedVectorType>(Ty)->
getNumElements() <= cast<FixedVectorType>(VecTy)->
getNumElements() && "Too many elements!") ? void (0) :
 __assert_fail ("cast<FixedVectorType>(Ty)->getNumElements() <= cast<FixedVectorType>(VecTy)->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2362, __extension__ __PRETTY_FUNCTION__
));
if (cast<FixedVectorType>(Ty)->getNumElements() ==
    cast<FixedVectorType>(VecTy)->getNumElements()) {
  assert(V->getType() == VecTy && "Vector type mismatch")(static_cast <bool> (V->getType() == VecTy &&
 "Vector type mismatch") ? void (0) : __assert_fail ("V->getType() == VecTy && \"Vector type mismatch\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2365, __extension__ __PRETTY_FUNCTION__
));
  return V;
}
unsigned EndIndex = BeginIndex + cast<FixedVectorType>(Ty)->getNumElements();

// When inserting a smaller vector into the larger to store, we first
// use a shuffle vector to widen it with undef elements, and then
// a second shuffle vector to select between the loaded vector and the
// incoming vector.
SmallVector<int, 8> Mask;
Mask.reserve(cast<FixedVectorType>(VecTy)->getNumElements());
for (unsigned i = 0; i != cast<FixedVectorType>(VecTy)->getNumElements(); ++i)
  if (i >= BeginIndex && i < EndIndex)
    Mask.push_back(i - BeginIndex);
  else
    Mask.push_back(-1);
V = IRB.CreateShuffleVector(V, Mask, Name + ".expand");
LLVM_DEBUG(dbgs() << "    shuffle: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    shuffle: " << *V <<
 "\n"; } } while (false);

SmallVector<Constant *, 8> Mask2;
Mask2.reserve(cast<FixedVectorType>(VecTy)->getNumElements());
for (unsigned i = 0; i != cast<FixedVectorType>(VecTy)->getNumElements(); ++i)
  Mask2.push_back(IRB.getInt1(i >= BeginIndex && i < EndIndex));

V = IRB.CreateSelect(ConstantVector::get(Mask2), V, Old, Name + "blend");

LLVM_DEBUG(dbgs() << "    blend: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    blend: " << *V <<
 "\n"; } } while (false);
return V;
2393}

2395/// Visitor to rewrite instructions using p particular slice of an alloca
2396/// to use a new alloca.
2397///
2398/// Also implements the rewriting to vector-based accesses when the partition
2399/// passes the isVectorPromotionViable predicate. Most of the rewriting logic
2400/// lives here.
2401class llvm::sroa::AllocaSliceRewriter
  : public InstVisitor<AllocaSliceRewriter, bool> {
// Befriend the base class so it can delegate to private visit methods.
friend class InstVisitor<AllocaSliceRewriter, bool>;

using Base = InstVisitor<AllocaSliceRewriter, bool>;

const DataLayout &DL;
AllocaSlices &AS;
SROAPass &Pass;
AllocaInst &OldAI, &NewAI;
const uint64_t NewAllocaBeginOffset, NewAllocaEndOffset;
Type *NewAllocaTy;

// This is a convenience and flag variable that will be null unless the new
// alloca's integer operations should be widened to this integer type due to
// passing isIntegerWideningViable above. If it is non-null, the desired
// integer type will be stored here for easy access during rewriting.
IntegerType *IntTy;

// If we are rewriting an alloca partition which can be written as pure
// vector operations, we stash extra information here. When VecTy is
// non-null, we have some strict guarantees about the rewritten alloca:
//   - The new alloca is exactly the size of the vector type here.
//   - The accesses all either map to the entire vector or to a single
//     element.
//   - The set of accessing instructions is only one of those handled above
//     in isVectorPromotionViable. Generally these are the same access kinds
//     which are promotable via mem2reg.
VectorType *VecTy;
Type *ElementTy;
uint64_t ElementSize;

// The original offset of the slice currently being rewritten relative to
// the original alloca.
uint64_t BeginOffset = 0;
uint64_t EndOffset = 0;

// The new offsets of the slice currently being rewritten relative to the
// original alloca.
uint64_t NewBeginOffset = 0, NewEndOffset = 0;

uint64_t SliceSize = 0;
bool IsSplittable = false;
bool IsSplit = false;
Use *OldUse = nullptr;
Instruction *OldPtr = nullptr;

// Track post-rewrite users which are PHI nodes and Selects.
SmallSetVector<PHINode *, 8> &PHIUsers;
SmallSetVector<SelectInst *, 8> &SelectUsers;

// Utility IR builder, whose name prefix is setup for each visited use, and
// the insertion point is set to point to the user.
IRBuilderTy IRB;

// Return the new alloca, addrspacecasted if required to avoid changing the
// addrspace of a volatile access.
Value *getPtrToNewAI(unsigned AddrSpace, bool IsVolatile) {
  if (!IsVolatile || AddrSpace == NewAI.getType()->getPointerAddressSpace())
    return &NewAI;

  Type *AccessTy = NewAI.getAllocatedType()->getPointerTo(AddrSpace);
  return IRB.CreateAddrSpaceCast(&NewAI, AccessTy);
}

2467public:
AllocaSliceRewriter(const DataLayout &DL, AllocaSlices &AS, SROAPass &Pass,
                    AllocaInst &OldAI, AllocaInst &NewAI,
                    uint64_t NewAllocaBeginOffset,
                    uint64_t NewAllocaEndOffset, bool IsIntegerPromotable,
                    VectorType *PromotableVecTy,
                    SmallSetVector<PHINode *, 8> &PHIUsers,
                    SmallSetVector<SelectInst *, 8> &SelectUsers)
    : DL(DL), AS(AS), Pass(Pass), OldAI(OldAI), NewAI(NewAI),
      NewAllocaBeginOffset(NewAllocaBeginOffset),
      NewAllocaEndOffset(NewAllocaEndOffset),
      NewAllocaTy(NewAI.getAllocatedType()),
      IntTy(
          IsIntegerPromotable
              ? Type::getIntNTy(NewAI.getContext(),
                                DL.getTypeSizeInBits(NewAI.getAllocatedType())
                                    .getFixedSize())
              : nullptr),
      VecTy(PromotableVecTy),
      ElementTy(VecTy ? VecTy->getElementType() : nullptr),
      ElementSize(VecTy ? DL.getTypeSizeInBits(ElementTy).getFixedSize() / 8
                        : 0),
      PHIUsers(PHIUsers), SelectUsers(SelectUsers),
      IRB(NewAI.getContext(), ConstantFolder()) {
  if (VecTy) {
    assert((DL.getTypeSizeInBits(ElementTy).getFixedSize() % 8) == 0 &&(static_cast <bool> ((DL.getTypeSizeInBits(ElementTy).getFixedSize
() % 8) == 0 && "Only multiple-of-8 sized vector elements are viable"
) ? void (0) : __assert_fail ("(DL.getTypeSizeInBits(ElementTy).getFixedSize() % 8) == 0 && \"Only multiple-of-8 sized vector elements are viable\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2493, __extension__ __PRETTY_FUNCTION__
))
           "Only multiple-of-8 sized vector elements are viable")(static_cast <bool> ((DL.getTypeSizeInBits(ElementTy).getFixedSize
() % 8) == 0 && "Only multiple-of-8 sized vector elements are viable"
) ? void (0) : __assert_fail ("(DL.getTypeSizeInBits(ElementTy).getFixedSize() % 8) == 0 && \"Only multiple-of-8 sized vector elements are viable\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2493, __extension__ __PRETTY_FUNCTION__
));
    ++NumVectorized;
  }
  assert((!IntTy && !VecTy) || (IntTy && !VecTy) || (!IntTy && VecTy))(static_cast <bool> ((!IntTy && !VecTy) || (IntTy
 && !VecTy) || (!IntTy && VecTy)) ? void (0) :
 __assert_fail ("(!IntTy && !VecTy) || (IntTy && !VecTy) || (!IntTy && VecTy)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2496, __extension__ __PRETTY_FUNCTION__
));
}

bool visit(AllocaSlices::const_iterator I) {
  bool CanSROA = true;
  BeginOffset = I->beginOffset();
  EndOffset = I->endOffset();
  IsSplittable = I->isSplittable();
  IsSplit =
      BeginOffset < NewAllocaBeginOffset || EndOffset > NewAllocaEndOffset;
  LLVM_DEBUG(dbgs() << "  rewriting " << (IsSplit ? "split " : ""))do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  rewriting " << (IsSplit ?
 "split " : ""); } } while (false);
  LLVM_DEBUG(AS.printSlice(dbgs(), I, ""))do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { AS.printSlice(dbgs(), I, ""); } } while (false);
  LLVM_DEBUG(dbgs() << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "\n"; } } while (false);

  // Compute the intersecting offset range.
  assert(BeginOffset < NewAllocaEndOffset)(static_cast <bool> (BeginOffset < NewAllocaEndOffset
) ? void (0) : __assert_fail ("BeginOffset < NewAllocaEndOffset"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2511, __extension__ __PRETTY_FUNCTION__
));
  assert(EndOffset > NewAllocaBeginOffset)(static_cast <bool> (EndOffset > NewAllocaBeginOffset
) ? void (0) : __assert_fail ("EndOffset > NewAllocaBeginOffset"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2512, __extension__ __PRETTY_FUNCTION__
));
  NewBeginOffset = std::max(BeginOffset, NewAllocaBeginOffset);
  NewEndOffset = std::min(EndOffset, NewAllocaEndOffset);

  SliceSize = NewEndOffset - NewBeginOffset;

  OldUse = I->getUse();
  OldPtr = cast<Instruction>(OldUse->get());

  Instruction *OldUserI = cast<Instruction>(OldUse->getUser());
  IRB.SetInsertPoint(OldUserI);
  IRB.SetCurrentDebugLocation(OldUserI->getDebugLoc());
  IRB.getInserter().SetNamePrefix(
      Twine(NewAI.getName()) + "." + Twine(BeginOffset) + ".");

  CanSROA &= visit(cast<Instruction>(OldUse->getUser()));
  if (VecTy || IntTy)
    assert(CanSROA)(static_cast <bool> (CanSROA) ? void (0) : __assert_fail
 ("CanSROA", "llvm/lib/Transforms/Scalar/SROA.cpp", 2529, __extension__
 __PRETTY_FUNCTION__));
  return CanSROA;
}

2533private:
// Make sure the other visit overloads are visible.
using Base::visit;

// Every instruction which can end up as a user must have a rewrite rule.
bool visitInstruction(Instruction &I) {
  LLVM_DEBUG(dbgs() << "    !!!! Cannot rewrite: " << I << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    !!!! Cannot rewrite: " <<
 I << "\n"; } } while (false);
  llvm_unreachable("No rewrite rule for this instruction!")::llvm::llvm_unreachable_internal("No rewrite rule for this instruction!"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2540);
}

Value *getNewAllocaSlicePtr(IRBuilderTy &IRB, Type *PointerTy) {
  // Note that the offset computation can use BeginOffset or NewBeginOffset
  // interchangeably for unsplit slices.
  assert(IsSplit || BeginOffset == NewBeginOffset)(static_cast <bool> (IsSplit || BeginOffset == NewBeginOffset
) ? void (0) : __assert_fail ("IsSplit || BeginOffset == NewBeginOffset"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2546, __extension__ __PRETTY_FUNCTION__
));
  uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;

2549#ifndef NDEBUG
  StringRef OldName = OldPtr->getName();
  // Skip through the last '.sroa.' component of the name.
  size_t LastSROAPrefix = OldName.rfind(".sroa.");
  if (LastSROAPrefix != StringRef::npos) {
    OldName = OldName.substr(LastSROAPrefix + strlen(".sroa."));
    // Look for an SROA slice index.
    size_t IndexEnd = OldName.find_first_not_of("0123456789");
    if (IndexEnd != StringRef::npos && OldName[IndexEnd] == '.') {
      // Strip the index and look for the offset.
      OldName = OldName.substr(IndexEnd + 1);
      size_t OffsetEnd = OldName.find_first_not_of("0123456789");
      if (OffsetEnd != StringRef::npos && OldName[OffsetEnd] == '.')
        // Strip the offset.
        OldName = OldName.substr(OffsetEnd + 1);
    }
  }
  // Strip any SROA suffixes as well.
  OldName = OldName.substr(0, OldName.find(".sroa_"));
2568#endif

  return getAdjustedPtr(IRB, DL, &NewAI,
                        APInt(DL.getIndexTypeSizeInBits(PointerTy), Offset),
                        PointerTy,
2573#ifndef NDEBUG
                        Twine(OldName) + "."
2575#else
                        Twine()
2577#endif
                        );
}

/// Compute suitable alignment to access this slice of the *new*
/// alloca.
///
/// You can optionally pass a type to this routine and if that type's ABI
/// alignment is itself suitable, this will return zero.
Align getSliceAlign() {
  return commonAlignment(NewAI.getAlign(),
                         NewBeginOffset - NewAllocaBeginOffset);
}

unsigned getIndex(uint64_t Offset) {
  assert(VecTy && "Can only call getIndex when rewriting a vector")(static_cast <bool> (VecTy && "Can only call getIndex when rewriting a vector"
) ? void (0) : __assert_fail ("VecTy && \"Can only call getIndex when rewriting a vector\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2592, __extension__ __PRETTY_FUNCTION__
));
  uint64_t RelOffset = Offset - NewAllocaBeginOffset;
  assert(RelOffset / ElementSize < UINT32_MAX && "Index out of bounds")(static_cast <bool> (RelOffset / ElementSize < (4294967295U
) && "Index out of bounds") ? void (0) : __assert_fail
 ("RelOffset / ElementSize < UINT32_MAX && \"Index out of bounds\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2594, __extension__ __PRETTY_FUNCTION__
));
  uint32_t Index = RelOffset / ElementSize;
  assert(Index * ElementSize == RelOffset)(static_cast <bool> (Index * ElementSize == RelOffset) ?
 void (0) : __assert_fail ("Index * ElementSize == RelOffset"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2596, __extension__ __PRETTY_FUNCTION__
));
  return Index;
}

void deleteIfTriviallyDead(Value *V) {
  Instruction *I = cast<Instruction>(V);
  if (isInstructionTriviallyDead(I))
    Pass.DeadInsts.push_back(I);
}

Value *rewriteVectorizedLoadInst(LoadInst &LI) {
  unsigned BeginIndex = getIndex(NewBeginOffset);
  unsigned EndIndex = getIndex(NewEndOffset);
  assert(EndIndex > BeginIndex && "Empty vector!")(static_cast <bool> (EndIndex > BeginIndex &&
 "Empty vector!") ? void (0) : __assert_fail ("EndIndex > BeginIndex && \"Empty vector!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2609, __extension__ __PRETTY_FUNCTION__
));

  LoadInst *Load = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                         NewAI.getAlign(), "load");

  Load->copyMetadata(LI, {LLVMContext::MD_mem_parallel_loop_access,
                          LLVMContext::MD_access_group});
  return extractVector(IRB, Load, BeginIndex, EndIndex, "vec");
}

Value *rewriteIntegerLoad(LoadInst &LI) {
  assert(IntTy && "We cannot insert an integer to the alloca")(static_cast <bool> (IntTy && "We cannot insert an integer to the alloca"
) ? void (0) : __assert_fail ("IntTy && \"We cannot insert an integer to the alloca\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2620, __extension__ __PRETTY_FUNCTION__
));
  assert(!LI.isVolatile())(static_cast <bool> (!LI.isVolatile()) ? void (0) : __assert_fail
 ("!LI.isVolatile()", "llvm/lib/Transforms/Scalar/SROA.cpp", 2621
, __extension__ __PRETTY_FUNCTION__));
  Value *V = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                   NewAI.getAlign(), "load");
  V = convertValue(DL, IRB, V, IntTy);
  assert(NewBeginOffset >= NewAllocaBeginOffset && "Out of bounds offset")(static_cast <bool> (NewBeginOffset >= NewAllocaBeginOffset
 && "Out of bounds offset") ? void (0) : __assert_fail
 ("NewBeginOffset >= NewAllocaBeginOffset && \"Out of bounds offset\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2625, __extension__ __PRETTY_FUNCTION__
));
  uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
  if (Offset > 0 || NewEndOffset < NewAllocaEndOffset) {
    IntegerType *ExtractTy = Type::getIntNTy(LI.getContext(), SliceSize * 8);
    V = extractInteger(DL, IRB, V, ExtractTy, Offset, "extract");
  }
  // It is possible that the extracted type is not the load type. This
  // happens if there is a load past the end of the alloca, and as
  // a consequence the slice is narrower but still a candidate for integer
  // lowering. To handle this case, we just zero extend the extracted
  // integer.
  assert(cast<IntegerType>(LI.getType())->getBitWidth() >= SliceSize * 8 &&(static_cast <bool> (cast<IntegerType>(LI.getType
())->getBitWidth() >= SliceSize * 8 && "Can only handle an extract for an overly wide load"
) ? void (0) : __assert_fail ("cast<IntegerType>(LI.getType())->getBitWidth() >= SliceSize * 8 && \"Can only handle an extract for an overly wide load\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2637, __extension__ __PRETTY_FUNCTION__
))
         "Can only handle an extract for an overly wide load")(static_cast <bool> (cast<IntegerType>(LI.getType
())->getBitWidth() >= SliceSize * 8 && "Can only handle an extract for an overly wide load"
) ? void (0) : __assert_fail ("cast<IntegerType>(LI.getType())->getBitWidth() >= SliceSize * 8 && \"Can only handle an extract for an overly wide load\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2637, __extension__ __PRETTY_FUNCTION__
));
  if (cast<IntegerType>(LI.getType())->getBitWidth() > SliceSize * 8)
    V = IRB.CreateZExt(V, LI.getType());
  return V;
}

bool visitLoadInst(LoadInst &LI) {
  LLVM_DEBUG(dbgs() << "    original: " << LI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << LI <<
 "\n"; } } while (false);
  Value *OldOp = LI.getOperand(0);
  assert(OldOp == OldPtr)(static_cast <bool> (OldOp == OldPtr) ? void (0) : __assert_fail
 ("OldOp == OldPtr", "llvm/lib/Transforms/Scalar/SROA.cpp", 2646
, __extension__ __PRETTY_FUNCTION__));

  AAMDNodes AATags = LI.getAAMetadata();

  unsigned AS = LI.getPointerAddressSpace();

  Type *TargetTy = IsSplit ? Type::getIntNTy(LI.getContext(), SliceSize * 8)
                           : LI.getType();
  const bool IsLoadPastEnd =
      DL.getTypeStoreSize(TargetTy).getFixedSize() > SliceSize;
  bool IsPtrAdjusted = false;
  Value *V;
  if (VecTy) {
    V = rewriteVectorizedLoadInst(LI);
  } else if (IntTy && LI.getType()->isIntegerTy()) {
    V = rewriteIntegerLoad(LI);
  } else if (NewBeginOffset == NewAllocaBeginOffset &&
             NewEndOffset == NewAllocaEndOffset &&
             (canConvertValue(DL, NewAllocaTy, TargetTy) ||
              (IsLoadPastEnd && NewAllocaTy->isIntegerTy() &&
               TargetTy->isIntegerTy()))) {
    Value *NewPtr =
        getPtrToNewAI(LI.getPointerAddressSpace(), LI.isVolatile());
    LoadInst *NewLI = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), NewPtr,
                                            NewAI.getAlign(), LI.isVolatile(),
                                            LI.getName());
    if (AATags)
      NewLI->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
    if (LI.isVolatile())
      NewLI->setAtomic(LI.getOrdering(), LI.getSyncScopeID());
    if (NewLI->isAtomic())
      NewLI->setAlignment(LI.getAlign());

    // Any !nonnull metadata or !range metadata on the old load is also valid
    // on the new load. This is even true in some cases even when the loads
    // are different types, for example by mapping !nonnull metadata to
    // !range metadata by modeling the null pointer constant converted to the
    // integer type.
    // FIXME: Add support for range metadata here. Currently the utilities
    // for this don't propagate range metadata in trivial cases from one
    // integer load to another, don't handle non-addrspace-0 null pointers
    // correctly, and don't have any support for mapping ranges as the
    // integer type becomes winder or narrower.
    if (MDNode *N = LI.getMetadata(LLVMContext::MD_nonnull))
      copyNonnullMetadata(LI, N, *NewLI);

    // Try to preserve nonnull metadata
    V = NewLI;

    // If this is an integer load past the end of the slice (which means the
    // bytes outside the slice are undef or this load is dead) just forcibly
    // fix the integer size with correct handling of endianness.
    if (auto *AITy = dyn_cast<IntegerType>(NewAllocaTy))
      if (auto *TITy = dyn_cast<IntegerType>(TargetTy))
        if (AITy->getBitWidth() < TITy->getBitWidth()) {
          V = IRB.CreateZExt(V, TITy, "load.ext");
          if (DL.isBigEndian())
            V = IRB.CreateShl(V, TITy->getBitWidth() - AITy->getBitWidth(),
                              "endian_shift");
        }
  } else {
    Type *LTy = TargetTy->getPointerTo(AS);
    LoadInst *NewLI =
        IRB.CreateAlignedLoad(TargetTy, getNewAllocaSlicePtr(IRB, LTy),
                              getSliceAlign(), LI.isVolatile(), LI.getName());
    if (AATags)
      NewLI->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
    if (LI.isVolatile())
      NewLI->setAtomic(LI.getOrdering(), LI.getSyncScopeID());
    NewLI->copyMetadata(LI, {LLVMContext::MD_mem_parallel_loop_access,
                             LLVMContext::MD_access_group});

    V = NewLI;
    IsPtrAdjusted = true;
  }
  V = convertValue(DL, IRB, V, TargetTy);

  if (IsSplit) {
    assert(!LI.isVolatile())(static_cast <bool> (!LI.isVolatile()) ? void (0) : __assert_fail
 ("!LI.isVolatile()", "llvm/lib/Transforms/Scalar/SROA.cpp", 2724
, __extension__ __PRETTY_FUNCTION__));
    assert(LI.getType()->isIntegerTy() &&(static_cast <bool> (LI.getType()->isIntegerTy() &&
 "Only integer type loads and stores are split") ? void (0) :
 __assert_fail ("LI.getType()->isIntegerTy() && \"Only integer type loads and stores are split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2726, __extension__ __PRETTY_FUNCTION__
))
           "Only integer type loads and stores are split")(static_cast <bool> (LI.getType()->isIntegerTy() &&
 "Only integer type loads and stores are split") ? void (0) :
 __assert_fail ("LI.getType()->isIntegerTy() && \"Only integer type loads and stores are split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2726, __extension__ __PRETTY_FUNCTION__
));
    assert(SliceSize < DL.getTypeStoreSize(LI.getType()).getFixedSize() &&(static_cast <bool> (SliceSize < DL.getTypeStoreSize
(LI.getType()).getFixedSize() && "Split load isn't smaller than original load"
) ? void (0) : __assert_fail ("SliceSize < DL.getTypeStoreSize(LI.getType()).getFixedSize() && \"Split load isn't smaller than original load\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2728, __extension__ __PRETTY_FUNCTION__
))
           "Split load isn't smaller than original load")(static_cast <bool> (SliceSize < DL.getTypeStoreSize
(LI.getType()).getFixedSize() && "Split load isn't smaller than original load"
) ? void (0) : __assert_fail ("SliceSize < DL.getTypeStoreSize(LI.getType()).getFixedSize() && \"Split load isn't smaller than original load\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2728, __extension__ __PRETTY_FUNCTION__
));
    assert(DL.typeSizeEqualsStoreSize(LI.getType()) &&(static_cast <bool> (DL.typeSizeEqualsStoreSize(LI.getType
()) && "Non-byte-multiple bit width") ? void (0) : __assert_fail
 ("DL.typeSizeEqualsStoreSize(LI.getType()) && \"Non-byte-multiple bit width\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2730, __extension__ __PRETTY_FUNCTION__
))
           "Non-byte-multiple bit width")(static_cast <bool> (DL.typeSizeEqualsStoreSize(LI.getType
()) && "Non-byte-multiple bit width") ? void (0) : __assert_fail
 ("DL.typeSizeEqualsStoreSize(LI.getType()) && \"Non-byte-multiple bit width\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2730, __extension__ __PRETTY_FUNCTION__
));
    // Move the insertion point just past the load so that we can refer to it.
    IRB.SetInsertPoint(&*std::next(BasicBlock::iterator(&LI)));
    // Create a placeholder value with the same type as LI to use as the
    // basis for the new value. This allows us to replace the uses of LI with
    // the computed value, and then replace the placeholder with LI, leaving
    // LI only used for this computation.
    Value *Placeholder = new LoadInst(
        LI.getType(), PoisonValue::get(LI.getType()->getPointerTo(AS)), "",
        false, Align(1));
    V = insertInteger(DL, IRB, Placeholder, V, NewBeginOffset - BeginOffset,
                      "insert");
    LI.replaceAllUsesWith(V);
    Placeholder->replaceAllUsesWith(&LI);
    Placeholder->deleteValue();
  } else {
    LI.replaceAllUsesWith(V);
  }

  Pass.DeadInsts.push_back(&LI);
  deleteIfTriviallyDead(OldOp);
  LLVM_DEBUG(dbgs() << "          to: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *V <<
 "\n"; } } while (false);
  return !LI.isVolatile() && !IsPtrAdjusted;
}

bool rewriteVectorizedStoreInst(Value *V, StoreInst &SI, Value *OldOp,
                                AAMDNodes AATags) {
  if (V->getType() != VecTy) {
    unsigned BeginIndex = getIndex(NewBeginOffset);
    unsigned EndIndex = getIndex(NewEndOffset);
    assert(EndIndex > BeginIndex && "Empty vector!")(static_cast <bool> (EndIndex > BeginIndex &&
 "Empty vector!") ? void (0) : __assert_fail ("EndIndex > BeginIndex && \"Empty vector!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2760, __extension__ __PRETTY_FUNCTION__
));
    unsigned NumElements = EndIndex - BeginIndex;
    assert(NumElements <= cast<FixedVectorType>(VecTy)->getNumElements() &&(static_cast <bool> (NumElements <= cast<FixedVectorType
>(VecTy)->getNumElements() && "Too many elements!"
) ? void (0) : __assert_fail ("NumElements <= cast<FixedVectorType>(VecTy)->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2763, __extension__ __PRETTY_FUNCTION__
))
           "Too many elements!")(static_cast <bool> (NumElements <= cast<FixedVectorType
>(VecTy)->getNumElements() && "Too many elements!"
) ? void (0) : __assert_fail ("NumElements <= cast<FixedVectorType>(VecTy)->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2763, __extension__ __PRETTY_FUNCTION__
));
    Type *SliceTy = (NumElements == 1)
                        ? ElementTy
                        : FixedVectorType::get(ElementTy, NumElements);
    if (V->getType() != SliceTy)
      V = convertValue(DL, IRB, V, SliceTy);

    // Mix in the existing elements.
    Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                       NewAI.getAlign(), "load");
    V = insertVector(IRB, Old, V, BeginIndex, "vec");
  }
  StoreInst *Store = IRB.CreateAlignedStore(V, &NewAI, NewAI.getAlign());
  Store->copyMetadata(SI, {LLVMContext::MD_mem_parallel_loop_access,
                           LLVMContext::MD_access_group});
  if (AATags)
    Store->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
  Pass.DeadInsts.push_back(&SI);

  LLVM_DEBUG(dbgs() << "          to: " << *Store << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *Store <<
 "\n"; } } while (false);
  return true;
}

bool rewriteIntegerStore(Value *V, StoreInst &SI, AAMDNodes AATags) {
  assert(IntTy && "We cannot extract an integer from the alloca")(static_cast <bool> (IntTy && "We cannot extract an integer from the alloca"
) ? void (0) : __assert_fail ("IntTy && \"We cannot extract an integer from the alloca\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2787, __extension__ __PRETTY_FUNCTION__
));
  assert(!SI.isVolatile())(static_cast <bool> (!SI.isVolatile()) ? void (0) : __assert_fail
 ("!SI.isVolatile()", "llvm/lib/Transforms/Scalar/SROA.cpp", 2788
, __extension__ __PRETTY_FUNCTION__));
  if (DL.getTypeSizeInBits(V->getType()).getFixedSize() !=
      IntTy->getBitWidth()) {
    Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                       NewAI.getAlign(), "oldload");
    Old = convertValue(DL, IRB, Old, IntTy);
    assert(BeginOffset >= NewAllocaBeginOffset && "Out of bounds offset")(static_cast <bool> (BeginOffset >= NewAllocaBeginOffset
 && "Out of bounds offset") ? void (0) : __assert_fail
 ("BeginOffset >= NewAllocaBeginOffset && \"Out of bounds offset\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2794, __extension__ __PRETTY_FUNCTION__
));
    uint64_t Offset = BeginOffset - NewAllocaBeginOffset;
    V = insertInteger(DL, IRB, Old, SI.getValueOperand(), Offset, "insert");
  }
  V = convertValue(DL, IRB, V, NewAllocaTy);
  StoreInst *Store = IRB.CreateAlignedStore(V, &NewAI, NewAI.getAlign());
  Store->copyMetadata(SI, {LLVMContext::MD_mem_parallel_loop_access,
                           LLVMContext::MD_access_group});
  if (AATags)
    Store->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
  Pass.DeadInsts.push_back(&SI);
  LLVM_DEBUG(dbgs() << "          to: " << *Store << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *Store <<
 "\n"; } } while (false);
  return true;
}

bool visitStoreInst(StoreInst &SI) {
  LLVM_DEBUG(dbgs() << "    original: " << SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << SI <<
 "\n"; } } while (false);
  Value *OldOp = SI.getOperand(1);
  assert(OldOp == OldPtr)(static_cast <bool> (OldOp == OldPtr) ? void (0) : __assert_fail
 ("OldOp == OldPtr", "llvm/lib/Transforms/Scalar/SROA.cpp", 2812
, __extension__ __PRETTY_FUNCTION__));

  AAMDNodes AATags = SI.getAAMetadata();
  Value *V = SI.getValueOperand();

  // Strip all inbounds GEPs and pointer casts to try to dig out any root
  // alloca that should be re-examined after promoting this alloca.
  if (V->getType()->isPointerTy())
    if (AllocaInst *AI = dyn_cast<AllocaInst>(V->stripInBoundsOffsets()))
      Pass.PostPromotionWorklist.insert(AI);

  if (SliceSize < DL.getTypeStoreSize(V->getType()).getFixedSize()) {
    assert(!SI.isVolatile())(static_cast <bool> (!SI.isVolatile()) ? void (0) : __assert_fail
 ("!SI.isVolatile()", "llvm/lib/Transforms/Scalar/SROA.cpp", 2824
, __extension__ __PRETTY_FUNCTION__));
    assert(V->getType()->isIntegerTy() &&(static_cast <bool> (V->getType()->isIntegerTy() &&
 "Only integer type loads and stores are split") ? void (0) :
 __assert_fail ("V->getType()->isIntegerTy() && \"Only integer type loads and stores are split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2826, __extension__ __PRETTY_FUNCTION__
))
           "Only integer type loads and stores are split")(static_cast <bool> (V->getType()->isIntegerTy() &&
 "Only integer type loads and stores are split") ? void (0) :
 __assert_fail ("V->getType()->isIntegerTy() && \"Only integer type loads and stores are split\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2826, __extension__ __PRETTY_FUNCTION__
));
    assert(DL.typeSizeEqualsStoreSize(V->getType()) &&(static_cast <bool> (DL.typeSizeEqualsStoreSize(V->getType
()) && "Non-byte-multiple bit width") ? void (0) : __assert_fail
 ("DL.typeSizeEqualsStoreSize(V->getType()) && \"Non-byte-multiple bit width\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2828, __extension__ __PRETTY_FUNCTION__
))
           "Non-byte-multiple bit width")(static_cast <bool> (DL.typeSizeEqualsStoreSize(V->getType
()) && "Non-byte-multiple bit width") ? void (0) : __assert_fail
 ("DL.typeSizeEqualsStoreSize(V->getType()) && \"Non-byte-multiple bit width\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2828, __extension__ __PRETTY_FUNCTION__
));
    IntegerType *NarrowTy = Type::getIntNTy(SI.getContext(), SliceSize * 8);
    V = extractInteger(DL, IRB, V, NarrowTy, NewBeginOffset - BeginOffset,
                       "extract");
  }

  if (VecTy)
    return rewriteVectorizedStoreInst(V, SI, OldOp, AATags);
  if (IntTy && V->getType()->isIntegerTy())
    return rewriteIntegerStore(V, SI, AATags);

  const bool IsStorePastEnd =
      DL.getTypeStoreSize(V->getType()).getFixedSize() > SliceSize;
  StoreInst *NewSI;
  if (NewBeginOffset == NewAllocaBeginOffset &&
      NewEndOffset == NewAllocaEndOffset &&
      (canConvertValue(DL, V->getType(), NewAllocaTy) ||
       (IsStorePastEnd && NewAllocaTy->isIntegerTy() &&
        V->getType()->isIntegerTy()))) {
    // If this is an integer store past the end of slice (and thus the bytes
    // past that point are irrelevant or this is unreachable), truncate the
    // value prior to storing.
    if (auto *VITy = dyn_cast<IntegerType>(V->getType()))
      if (auto *AITy = dyn_cast<IntegerType>(NewAllocaTy))
        if (VITy->getBitWidth() > AITy->getBitWidth()) {
          if (DL.isBigEndian())
            V = IRB.CreateLShr(V, VITy->getBitWidth() - AITy->getBitWidth(),
                               "endian_shift");
          V = IRB.CreateTrunc(V, AITy, "load.trunc");
        }

    V = convertValue(DL, IRB, V, NewAllocaTy);
    Value *NewPtr =
        getPtrToNewAI(SI.getPointerAddressSpace(), SI.isVolatile());

    NewSI =
        IRB.CreateAlignedStore(V, NewPtr, NewAI.getAlign(), SI.isVolatile());
  } else {
    unsigned AS = SI.getPointerAddressSpace();
    Value *NewPtr = getNewAllocaSlicePtr(IRB, V->getType()->getPointerTo(AS));
    NewSI =
        IRB.CreateAlignedStore(V, NewPtr, getSliceAlign(), SI.isVolatile());
  }
  NewSI->copyMetadata(SI, {LLVMContext::MD_mem_parallel_loop_access,
                           LLVMContext::MD_access_group});
  if (AATags)
    NewSI->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
  if (SI.isVolatile())
    NewSI->setAtomic(SI.getOrdering(), SI.getSyncScopeID());
  if (NewSI->isAtomic())
    NewSI->setAlignment(SI.getAlign());
  Pass.DeadInsts.push_back(&SI);
  deleteIfTriviallyDead(OldOp);

  LLVM_DEBUG(dbgs() << "          to: " << *NewSI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *NewSI <<
 "\n"; } } while (false);
  return NewSI->getPointerOperand() == &NewAI &&
         NewSI->getValueOperand()->getType() == NewAllocaTy &&
         !SI.isVolatile();
}

/// Compute an integer value from splatting an i8 across the given
/// number of bytes.
///
/// Note that this routine assumes an i8 is a byte. If that isn't true, don't
/// call this routine.
/// FIXME: Heed the advice above.
///
/// \param V The i8 value to splat.
/// \param Size The number of bytes in the output (assuming i8 is one byte)
Value *getIntegerSplat(Value *V, unsigned Size) {
  assert(Size > 0 && "Expected a positive number of bytes.")(static_cast <bool> (Size > 0 && "Expected a positive number of bytes."
) ? void (0) : __assert_fail ("Size > 0 && \"Expected a positive number of bytes.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2898, __extension__ __PRETTY_FUNCTION__
));
  IntegerType *VTy = cast<IntegerType>(V->getType());
  assert(VTy->getBitWidth() == 8 && "Expected an i8 value for the byte")(static_cast <bool> (VTy->getBitWidth() == 8 &&
 "Expected an i8 value for the byte") ? void (0) : __assert_fail
 ("VTy->getBitWidth() == 8 && \"Expected an i8 value for the byte\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2900, __extension__ __PRETTY_FUNCTION__
));
  if (Size == 1)
    return V;

  Type *SplatIntTy = Type::getIntNTy(VTy->getContext(), Size * 8);
  V = IRB.CreateMul(
      IRB.CreateZExt(V, SplatIntTy, "zext"),
      IRB.CreateUDiv(Constant::getAllOnesValue(SplatIntTy),
                     IRB.CreateZExt(Constant::getAllOnesValue(V->getType()),
                                    SplatIntTy)),
      "isplat");
  return V;
}

/// Compute a vector splat for a given element value.
Value *getVectorSplat(Value *V, unsigned NumElements) {
  V = IRB.CreateVectorSplat(NumElements, V, "vsplat");
  LLVM_DEBUG(dbgs() << "       splat: " << *V << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "       splat: " << *V <<
 "\n"; } } while (false);
  return V;
}

bool visitMemSetInst(MemSetInst &II) {
  LLVM_DEBUG(dbgs() << "    original: " << II << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << II <<
 "\n"; } } while (false);
  assert(II.getRawDest() == OldPtr)(static_cast <bool> (II.getRawDest() == OldPtr) ? void (
0) : __assert_fail ("II.getRawDest() == OldPtr", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 2923, __extension__ __PRETTY_FUNCTION__));

  AAMDNodes AATags = II.getAAMetadata();

  // If the memset has a variable size, it cannot be split, just adjust the
  // pointer to the new alloca.
  if (!isa<ConstantInt>(II.getLength())) {
    assert(!IsSplit)(static_cast <bool> (!IsSplit) ? void (0) : __assert_fail
 ("!IsSplit", "llvm/lib/Transforms/Scalar/SROA.cpp", 2930, __extension__
 __PRETTY_FUNCTION__));
    assert(NewBeginOffset == BeginOffset)(static_cast <bool> (NewBeginOffset == BeginOffset) ? void
 (0) : __assert_fail ("NewBeginOffset == BeginOffset", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 2931, __extension__ __PRETTY_FUNCTION__));
    II.setDest(getNewAllocaSlicePtr(IRB, OldPtr->getType()));
    II.setDestAlignment(getSliceAlign());

    deleteIfTriviallyDead(OldPtr);
    return false;
  }

  // Record this instruction for deletion.
  Pass.DeadInsts.push_back(&II);

  Type *AllocaTy = NewAI.getAllocatedType();
  Type *ScalarTy = AllocaTy->getScalarType();

  const bool CanContinue = [&]() {
    if (VecTy || IntTy)
      return true;
    if (BeginOffset > NewAllocaBeginOffset ||
        EndOffset < NewAllocaEndOffset)
      return false;
    // Length must be in range for FixedVectorType.
    auto *C = cast<ConstantInt>(II.getLength());
    const uint64_t Len = C->getLimitedValue();
    if (Len > std::numeric_limits<unsigned>::max())
      return false;
    auto *Int8Ty = IntegerType::getInt8Ty(NewAI.getContext());
    auto *SrcTy = FixedVectorType::get(Int8Ty, Len);
    return canConvertValue(DL, SrcTy, AllocaTy) &&
           DL.isLegalInteger(DL.getTypeSizeInBits(ScalarTy).getFixedSize());
  }();

  // If this doesn't map cleanly onto the alloca type, and that type isn't
  // a single value type, just emit a memset.
  if (!CanContinue) {
    Type *SizeTy = II.getLength()->getType();
    Constant *Size = ConstantInt::get(SizeTy, NewEndOffset - NewBeginOffset);
    CallInst *New = IRB.CreateMemSet(
        getNewAllocaSlicePtr(IRB, OldPtr->getType()), II.getValue(), Size,
        MaybeAlign(getSliceAlign()), II.isVolatile());
    if (AATags)
      New->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
    LLVM_DEBUG(dbgs() << "          to: " << *New << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *New <<
 "\n"; } } while (false);
    return false;
  }

  // If we can represent this as a simple value, we have to build the actual
  // value to store, which requires expanding the byte present in memset to
  // a sensible representation for the alloca type. This is essentially
  // splatting the byte to a sufficiently wide integer, splatting it across
  // any desired vector width, and bitcasting to the final type.
  Value *V;

  if (VecTy) {
    // If this is a memset of a vectorized alloca, insert it.
    assert(ElementTy == ScalarTy)(static_cast <bool> (ElementTy == ScalarTy) ? void (0) :
 __assert_fail ("ElementTy == ScalarTy", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 2985, __extension__ __PRETTY_FUNCTION__));

    unsigned BeginIndex = getIndex(NewBeginOffset);
    unsigned EndIndex = getIndex(NewEndOffset);
    assert(EndIndex > BeginIndex && "Empty vector!")(static_cast <bool> (EndIndex > BeginIndex &&
 "Empty vector!") ? void (0) : __assert_fail ("EndIndex > BeginIndex && \"Empty vector!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2989, __extension__ __PRETTY_FUNCTION__
));
    unsigned NumElements = EndIndex - BeginIndex;
    assert(NumElements <= cast<FixedVectorType>(VecTy)->getNumElements() &&(static_cast <bool> (NumElements <= cast<FixedVectorType
>(VecTy)->getNumElements() && "Too many elements!"
) ? void (0) : __assert_fail ("NumElements <= cast<FixedVectorType>(VecTy)->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2992, __extension__ __PRETTY_FUNCTION__
))
           "Too many elements!")(static_cast <bool> (NumElements <= cast<FixedVectorType
>(VecTy)->getNumElements() && "Too many elements!"
) ? void (0) : __assert_fail ("NumElements <= cast<FixedVectorType>(VecTy)->getNumElements() && \"Too many elements!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 2992, __extension__ __PRETTY_FUNCTION__
));

    Value *Splat = getIntegerSplat(
        II.getValue(), DL.getTypeSizeInBits(ElementTy).getFixedSize() / 8);
    Splat = convertValue(DL, IRB, Splat, ElementTy);
    if (NumElements > 1)
      Splat = getVectorSplat(Splat, NumElements);

    Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                       NewAI.getAlign(), "oldload");
    V = insertVector(IRB, Old, Splat, BeginIndex, "vec");
  } else if (IntTy) {
    // If this is a memset on an alloca where we can widen stores, insert the
    // set integer.
    assert(!II.isVolatile())(static_cast <bool> (!II.isVolatile()) ? void (0) : __assert_fail
 ("!II.isVolatile()", "llvm/lib/Transforms/Scalar/SROA.cpp", 3006
, __extension__ __PRETTY_FUNCTION__));

    uint64_t Size = NewEndOffset - NewBeginOffset;
    V = getIntegerSplat(II.getValue(), Size);

    if (IntTy && (BeginOffset != NewAllocaBeginOffset ||
                  EndOffset != NewAllocaBeginOffset)) {
      Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                         NewAI.getAlign(), "oldload");
      Old = convertValue(DL, IRB, Old, IntTy);
      uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
      V = insertInteger(DL, IRB, Old, V, Offset, "insert");
    } else {
      assert(V->getType() == IntTy &&(static_cast <bool> (V->getType() == IntTy &&
 "Wrong type for an alloca wide integer!") ? void (0) : __assert_fail
 ("V->getType() == IntTy && \"Wrong type for an alloca wide integer!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3020, __extension__ __PRETTY_FUNCTION__
))
             "Wrong type for an alloca wide integer!")(static_cast <bool> (V->getType() == IntTy &&
 "Wrong type for an alloca wide integer!") ? void (0) : __assert_fail
 ("V->getType() == IntTy && \"Wrong type for an alloca wide integer!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3020, __extension__ __PRETTY_FUNCTION__
));
    }
    V = convertValue(DL, IRB, V, AllocaTy);
  } else {
    // Established these invariants above.
    assert(NewBeginOffset == NewAllocaBeginOffset)(static_cast <bool> (NewBeginOffset == NewAllocaBeginOffset
) ? void (0) : __assert_fail ("NewBeginOffset == NewAllocaBeginOffset"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3025, __extension__ __PRETTY_FUNCTION__
));
    assert(NewEndOffset == NewAllocaEndOffset)(static_cast <bool> (NewEndOffset == NewAllocaEndOffset
) ? void (0) : __assert_fail ("NewEndOffset == NewAllocaEndOffset"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3026, __extension__ __PRETTY_FUNCTION__
));

    V = getIntegerSplat(II.getValue(),
                        DL.getTypeSizeInBits(ScalarTy).getFixedSize() / 8);
    if (VectorType *AllocaVecTy = dyn_cast<VectorType>(AllocaTy))
      V = getVectorSplat(
          V, cast<FixedVectorType>(AllocaVecTy)->getNumElements());

    V = convertValue(DL, IRB, V, AllocaTy);
  }

  Value *NewPtr = getPtrToNewAI(II.getDestAddressSpace(), II.isVolatile());
  StoreInst *New =
      IRB.CreateAlignedStore(V, NewPtr, NewAI.getAlign(), II.isVolatile());
  New->copyMetadata(II, {LLVMContext::MD_mem_parallel_loop_access,
                         LLVMContext::MD_access_group});
  if (AATags)
    New->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
  LLVM_DEBUG(dbgs() << "          to: " << *New << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *New <<
 "\n"; } } while (false);
  return !II.isVolatile();
}

bool visitMemTransferInst(MemTransferInst &II) {
  // Rewriting of memory transfer instructions can be a bit tricky. We break
  // them into two categories: split intrinsics and unsplit intrinsics.

  LLVM_DEBUG(dbgs() << "    original: " << II << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << II <<
 "\n"; } } while (false);
1
Assuming 'DebugFlag' is false→
2
←
Loop condition is false.  Exiting loop→

  AAMDNodes AATags = II.getAAMetadata();

  bool IsDest = &II.getRawDestUse() == OldUse;
  assert((IsDest && II.getRawDest() == OldPtr) ||(static_cast <bool> ((IsDest && II.getRawDest()
 == OldPtr) || (!IsDest && II.getRawSource() == OldPtr
)) ? void (0) : __assert_fail ("(IsDest && II.getRawDest() == OldPtr) || (!IsDest && II.getRawSource() == OldPtr)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3058, __extension__ __PRETTY_FUNCTION__
))
3
←
Assuming 'IsDest' is false→
4
←
Assuming pointer value is null→
5
←
'?' condition is true→
         (!IsDest && II.getRawSource() == OldPtr))(static_cast <bool> ((IsDest && II.getRawDest()
 == OldPtr) || (!IsDest && II.getRawSource() == OldPtr
)) ? void (0) : __assert_fail ("(IsDest && II.getRawDest() == OldPtr) || (!IsDest && II.getRawSource() == OldPtr)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3058, __extension__ __PRETTY_FUNCTION__
));

  Align SliceAlign = getSliceAlign();

  // For unsplit intrinsics, we simply modify the source and destination
  // pointers in place. This isn't just an optimization, it is a matter of
  // correctness. With unsplit intrinsics we may be dealing with transfers
  // within a single alloca before SROA ran, or with transfers that have
  // a variable length. We may also be dealing with memmove instead of
  // memcpy, and so simply updating the pointers is the necessary for us to
  // update both source and dest of a single call.
  if (!IsSplittable) {
6
←
Assuming field 'IsSplittable' is false→
7
←
Taking true branch→
    Value *AdjustedPtr = getNewAllocaSlicePtr(IRB, OldPtr->getType());
8
←
Called C++ object pointer is null
    if (IsDest) {
      II.setDest(AdjustedPtr);
      II.setDestAlignment(SliceAlign);
    }
    else {
      II.setSource(AdjustedPtr);
      II.setSourceAlignment(SliceAlign);
    }

    LLVM_DEBUG(dbgs() << "          to: " << II << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << II <<
 "\n"; } } while (false);
    deleteIfTriviallyDead(OldPtr);
    return false;
  }
  // For split transfer intrinsics we have an incredibly useful assurance:
  // the source and destination do not reside within the same alloca, and at
  // least one of them does not escape. This means that we can replace
  // memmove with memcpy, and we don't need to worry about all manner of
  // downsides to splitting and transforming the operations.

  // If this doesn't map cleanly onto the alloca type, and that type isn't
  // a single value type, just emit a memcpy.
  bool EmitMemCpy =
      !VecTy && !IntTy &&
      (BeginOffset > NewAllocaBeginOffset || EndOffset < NewAllocaEndOffset ||
       SliceSize !=
           DL.getTypeStoreSize(NewAI.getAllocatedType()).getFixedSize() ||
       !NewAI.getAllocatedType()->isSingleValueType());

  // If we're just going to emit a memcpy, the alloca hasn't changed, and the
  // size hasn't been shrunk based on analysis of the viable range, this is
  // a no-op.
  if (EmitMemCpy && &OldAI == &NewAI) {
    // Ensure the start lines up.
    assert(NewBeginOffset == BeginOffset)(static_cast <bool> (NewBeginOffset == BeginOffset) ? void
 (0) : __assert_fail ("NewBeginOffset == BeginOffset", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 3104, __extension__ __PRETTY_FUNCTION__));

    // Rewrite the size as needed.
    if (NewEndOffset != EndOffset)
      II.setLength(ConstantInt::get(II.getLength()->getType(),
                                    NewEndOffset - NewBeginOffset));
    return false;
  }
  // Record this instruction for deletion.
  Pass.DeadInsts.push_back(&II);

  // Strip all inbounds GEPs and pointer casts to try to dig out any root
  // alloca that should be re-examined after rewriting this instruction.
  Value *OtherPtr = IsDest ? II.getRawSource() : II.getRawDest();
  if (AllocaInst *AI =
          dyn_cast<AllocaInst>(OtherPtr->stripInBoundsOffsets())) {
    assert(AI != &OldAI && AI != &NewAI &&(static_cast <bool> (AI != &OldAI && AI != &
NewAI && "Splittable transfers cannot reach the same alloca on both ends."
) ? void (0) : __assert_fail ("AI != &OldAI && AI != &NewAI && \"Splittable transfers cannot reach the same alloca on both ends.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3121, __extension__ __PRETTY_FUNCTION__
))
           "Splittable transfers cannot reach the same alloca on both ends.")(static_cast <bool> (AI != &OldAI && AI != &
NewAI && "Splittable transfers cannot reach the same alloca on both ends."
) ? void (0) : __assert_fail ("AI != &OldAI && AI != &NewAI && \"Splittable transfers cannot reach the same alloca on both ends.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3121, __extension__ __PRETTY_FUNCTION__
));
    Pass.Worklist.insert(AI);
  }

  Type *OtherPtrTy = OtherPtr->getType();
  unsigned OtherAS = OtherPtrTy->getPointerAddressSpace();

  // Compute the relative offset for the other pointer within the transfer.
  unsigned OffsetWidth = DL.getIndexSizeInBits(OtherAS);
  APInt OtherOffset(OffsetWidth, NewBeginOffset - BeginOffset);
  Align OtherAlign =
      (IsDest ? II.getSourceAlign() : II.getDestAlign()).valueOrOne();
  OtherAlign =
      commonAlignment(OtherAlign, OtherOffset.zextOrTrunc(64).getZExtValue());

  if (EmitMemCpy) {
    // Compute the other pointer, folding as much as possible to produce
    // a single, simple GEP in most cases.
    OtherPtr = getAdjustedPtr(IRB, DL, OtherPtr, OtherOffset, OtherPtrTy,
                              OtherPtr->getName() + ".");

    Value *OurPtr = getNewAllocaSlicePtr(IRB, OldPtr->getType());
    Type *SizeTy = II.getLength()->getType();
    Constant *Size = ConstantInt::get(SizeTy, NewEndOffset - NewBeginOffset);

    Value *DestPtr, *SrcPtr;
    MaybeAlign DestAlign, SrcAlign;
    // Note: IsDest is true iff we're copying into the new alloca slice
    if (IsDest) {
      DestPtr = OurPtr;
      DestAlign = SliceAlign;
      SrcPtr = OtherPtr;
      SrcAlign = OtherAlign;
    } else {
      DestPtr = OtherPtr;
      DestAlign = OtherAlign;
      SrcPtr = OurPtr;
      SrcAlign = SliceAlign;
    }
    CallInst *New = IRB.CreateMemCpy(DestPtr, DestAlign, SrcPtr, SrcAlign,
                                     Size, II.isVolatile());
    if (AATags)
      New->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
    LLVM_DEBUG(dbgs() << "          to: " << *New << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *New <<
 "\n"; } } while (false);
    return false;
  }

  bool IsWholeAlloca = NewBeginOffset == NewAllocaBeginOffset &&
                       NewEndOffset == NewAllocaEndOffset;
  uint64_t Size = NewEndOffset - NewBeginOffset;
  unsigned BeginIndex = VecTy ? getIndex(NewBeginOffset) : 0;
  unsigned EndIndex = VecTy ? getIndex(NewEndOffset) : 0;
  unsigned NumElements = EndIndex - BeginIndex;
  IntegerType *SubIntTy =
      IntTy ? Type::getIntNTy(IntTy->getContext(), Size * 8) : nullptr;

  // Reset the other pointer type to match the register type we're going to
  // use, but using the address space of the original other pointer.
  Type *OtherTy;
  if (VecTy && !IsWholeAlloca) {
    if (NumElements == 1)
      OtherTy = VecTy->getElementType();
    else
      OtherTy = FixedVectorType::get(VecTy->getElementType(), NumElements);
  } else if (IntTy && !IsWholeAlloca) {
    OtherTy = SubIntTy;
  } else {
    OtherTy = NewAllocaTy;
  }
  OtherPtrTy = OtherTy->getPointerTo(OtherAS);

  Value *AdjPtr = getAdjustedPtr(IRB, DL, OtherPtr, OtherOffset, OtherPtrTy,
                                 OtherPtr->getName() + ".");
  MaybeAlign SrcAlign = OtherAlign;
  MaybeAlign DstAlign = SliceAlign;
  if (!IsDest)
    std::swap(SrcAlign, DstAlign);

  Value *SrcPtr;
  Value *DstPtr;

  if (IsDest) {
    DstPtr = getPtrToNewAI(II.getDestAddressSpace(), II.isVolatile());
    SrcPtr = AdjPtr;
  } else {
    DstPtr = AdjPtr;
    SrcPtr = getPtrToNewAI(II.getSourceAddressSpace(), II.isVolatile());
  }

  Value *Src;
  if (VecTy && !IsWholeAlloca && !IsDest) {
    Src = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                NewAI.getAlign(), "load");
    Src = extractVector(IRB, Src, BeginIndex, EndIndex, "vec");
  } else if (IntTy && !IsWholeAlloca && !IsDest) {
    Src = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                NewAI.getAlign(), "load");
    Src = convertValue(DL, IRB, Src, IntTy);
    uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
    Src = extractInteger(DL, IRB, Src, SubIntTy, Offset, "extract");
  } else {
    LoadInst *Load = IRB.CreateAlignedLoad(OtherTy, SrcPtr, SrcAlign,
                                           II.isVolatile(), "copyload");
    Load->copyMetadata(II, {LLVMContext::MD_mem_parallel_loop_access,
                            LLVMContext::MD_access_group});
    if (AATags)
      Load->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
    Src = Load;
  }

  if (VecTy && !IsWholeAlloca && IsDest) {
    Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                       NewAI.getAlign(), "oldload");
    Src = insertVector(IRB, Old, Src, BeginIndex, "vec");
  } else if (IntTy && !IsWholeAlloca && IsDest) {
    Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
                                       NewAI.getAlign(), "oldload");
    Old = convertValue(DL, IRB, Old, IntTy);
    uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
    Src = insertInteger(DL, IRB, Old, Src, Offset, "insert");
    Src = convertValue(DL, IRB, Src, NewAllocaTy);
  }

  StoreInst *Store = cast<StoreInst>(
      IRB.CreateAlignedStore(Src, DstPtr, DstAlign, II.isVolatile()));
  Store->copyMetadata(II, {LLVMContext::MD_mem_parallel_loop_access,
                           LLVMContext::MD_access_group});
  if (AATags)
    Store->setAAMetadata(AATags.shift(NewBeginOffset - BeginOffset));
  LLVM_DEBUG(dbgs() << "          to: " << *Store << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *Store <<
 "\n"; } } while (false);
  return !II.isVolatile();
}

bool visitIntrinsicInst(IntrinsicInst &II) {
  assert((II.isLifetimeStartOrEnd() || II.isDroppable()) &&(static_cast <bool> ((II.isLifetimeStartOrEnd() || II.isDroppable
()) && "Unexpected intrinsic!") ? void (0) : __assert_fail
 ("(II.isLifetimeStartOrEnd() || II.isDroppable()) && \"Unexpected intrinsic!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3256, __extension__ __PRETTY_FUNCTION__
))
         "Unexpected intrinsic!")(static_cast <bool> ((II.isLifetimeStartOrEnd() || II.isDroppable
()) && "Unexpected intrinsic!") ? void (0) : __assert_fail
 ("(II.isLifetimeStartOrEnd() || II.isDroppable()) && \"Unexpected intrinsic!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3256, __extension__ __PRETTY_FUNCTION__
));
  LLVM_DEBUG(dbgs() << "    original: " << II << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << II <<
 "\n"; } } while (false);

  // Record this instruction for deletion.
  Pass.DeadInsts.push_back(&II);

  if (II.isDroppable()) {
    assert(II.getIntrinsicID() == Intrinsic::assume && "Expected assume")(static_cast <bool> (II.getIntrinsicID() == Intrinsic::
assume && "Expected assume") ? void (0) : __assert_fail
 ("II.getIntrinsicID() == Intrinsic::assume && \"Expected assume\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3263, __extension__ __PRETTY_FUNCTION__
));
    // TODO For now we forget assumed information, this can be improved.
    OldPtr->dropDroppableUsesIn(II);
    return true;
  }

  assert(II.getArgOperand(1) == OldPtr)(static_cast <bool> (II.getArgOperand(1) == OldPtr) ? void
 (0) : __assert_fail ("II.getArgOperand(1) == OldPtr", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 3269, __extension__ __PRETTY_FUNCTION__));
  // Lifetime intrinsics are only promotable if they cover the whole alloca.
  // Therefore, we drop lifetime intrinsics which don't cover the whole
  // alloca.
  // (In theory, intrinsics which partially cover an alloca could be
  // promoted, but PromoteMemToReg doesn't handle that case.)
  // FIXME: Check whether the alloca is promotable before dropping the
  // lifetime intrinsics?
  if (NewBeginOffset != NewAllocaBeginOffset ||
      NewEndOffset != NewAllocaEndOffset)
    return true;

  ConstantInt *Size =
      ConstantInt::get(cast<IntegerType>(II.getArgOperand(0)->getType()),
                       NewEndOffset - NewBeginOffset);
  // Lifetime intrinsics always expect an i8* so directly get such a pointer
  // for the new alloca slice.
  Type *PointerTy = IRB.getInt8PtrTy(OldPtr->getType()->getPointerAddressSpace());
  Value *Ptr = getNewAllocaSlicePtr(IRB, PointerTy);
  Value *New;
  if (II.getIntrinsicID() == Intrinsic::lifetime_start)
    New = IRB.CreateLifetimeStart(Ptr, Size);
  else
    New = IRB.CreateLifetimeEnd(Ptr, Size);

  (void)New;
  LLVM_DEBUG(dbgs() << "          to: " << *New << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *New <<
 "\n"; } } while (false);

  return true;
}

void fixLoadStoreAlign(Instruction &Root) {
  // This algorithm implements the same visitor loop as
  // hasUnsafePHIOrSelectUse, and fixes the alignment of each load
  // or store found.
  SmallPtrSet<Instruction *, 4> Visited;
  SmallVector<Instruction *, 4> Uses;
  Visited.insert(&Root);
  Uses.push_back(&Root);
  do {
    Instruction *I = Uses.pop_back_val();

    if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
      LI->setAlignment(std::min(LI->getAlign(), getSliceAlign()));
      continue;
    }
    if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
      SI->setAlignment(std::min(SI->getAlign(), getSliceAlign()));
      continue;
    }

    assert(isa<BitCastInst>(I) || isa<AddrSpaceCastInst>(I) ||(static_cast <bool> (isa<BitCastInst>(I) || isa<
AddrSpaceCastInst>(I) || isa<PHINode>(I) || isa<SelectInst
>(I) || isa<GetElementPtrInst>(I)) ? void (0) : __assert_fail
 ("isa<BitCastInst>(I) || isa<AddrSpaceCastInst>(I) || isa<PHINode>(I) || isa<SelectInst>(I) || isa<GetElementPtrInst>(I)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3322, __extension__ __PRETTY_FUNCTION__
))
           isa<PHINode>(I) || isa<SelectInst>(I) ||(static_cast <bool> (isa<BitCastInst>(I) || isa<
AddrSpaceCastInst>(I) || isa<PHINode>(I) || isa<SelectInst
>(I) || isa<GetElementPtrInst>(I)) ? void (0) : __assert_fail
 ("isa<BitCastInst>(I) || isa<AddrSpaceCastInst>(I) || isa<PHINode>(I) || isa<SelectInst>(I) || isa<GetElementPtrInst>(I)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3322, __extension__ __PRETTY_FUNCTION__
))
           isa<GetElementPtrInst>(I))(static_cast <bool> (isa<BitCastInst>(I) || isa<
AddrSpaceCastInst>(I) || isa<PHINode>(I) || isa<SelectInst
>(I) || isa<GetElementPtrInst>(I)) ? void (0) : __assert_fail
 ("isa<BitCastInst>(I) || isa<AddrSpaceCastInst>(I) || isa<PHINode>(I) || isa<SelectInst>(I) || isa<GetElementPtrInst>(I)"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3322, __extension__ __PRETTY_FUNCTION__
));
    for (User *U : I->users())
      if (Visited.insert(cast<Instruction>(U)).second)
        Uses.push_back(cast<Instruction>(U));
  } while (!Uses.empty());
}

bool visitPHINode(PHINode &PN) {
  LLVM_DEBUG(dbgs() << "    original: " << PN << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << PN <<
 "\n"; } } while (false);
  assert(BeginOffset >= NewAllocaBeginOffset && "PHIs are unsplittable")(static_cast <bool> (BeginOffset >= NewAllocaBeginOffset
 && "PHIs are unsplittable") ? void (0) : __assert_fail
 ("BeginOffset >= NewAllocaBeginOffset && \"PHIs are unsplittable\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3331, __extension__ __PRETTY_FUNCTION__
));
  assert(EndOffset <= NewAllocaEndOffset && "PHIs are unsplittable")(static_cast <bool> (EndOffset <= NewAllocaEndOffset
 && "PHIs are unsplittable") ? void (0) : __assert_fail
 ("EndOffset <= NewAllocaEndOffset && \"PHIs are unsplittable\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3332, __extension__ __PRETTY_FUNCTION__
));

  // We would like to compute a new pointer in only one place, but have it be
  // as local as possible to the PHI. To do that, we re-use the location of
  // the old pointer, which necessarily must be in the right position to
  // dominate the PHI.
  IRBuilderBase::InsertPointGuard Guard(IRB);
  if (isa<PHINode>(OldPtr))
    IRB.SetInsertPoint(&*OldPtr->getParent()->getFirstInsertionPt());
  else
    IRB.SetInsertPoint(OldPtr);
  IRB.SetCurrentDebugLocation(OldPtr->getDebugLoc());

  Value *NewPtr = getNewAllocaSlicePtr(IRB, OldPtr->getType());
  // Replace the operands which were using the old pointer.
  std::replace(PN.op_begin(), PN.op_end(), cast<Value>(OldPtr), NewPtr);

  LLVM_DEBUG(dbgs() << "          to: " << PN << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << PN <<
 "\n"; } } while (false);
  deleteIfTriviallyDead(OldPtr);

  // Fix the alignment of any loads or stores using this PHI node.
  fixLoadStoreAlign(PN);

  // PHIs can't be promoted on their own, but often can be speculated. We
  // check the speculation outside of the rewriter so that we see the
  // fully-rewritten alloca.
  PHIUsers.insert(&PN);
  return true;
}

bool visitSelectInst(SelectInst &SI) {
  LLVM_DEBUG(dbgs() << "    original: " << SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << SI <<
 "\n"; } } while (false);
  assert((SI.getTrueValue() == OldPtr || SI.getFalseValue() == OldPtr) &&(static_cast <bool> ((SI.getTrueValue() == OldPtr || SI
.getFalseValue() == OldPtr) && "Pointer isn't an operand!"
) ? void (0) : __assert_fail ("(SI.getTrueValue() == OldPtr || SI.getFalseValue() == OldPtr) && \"Pointer isn't an operand!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3365, __extension__ __PRETTY_FUNCTION__
))
         "Pointer isn't an operand!")(static_cast <bool> ((SI.getTrueValue() == OldPtr || SI
.getFalseValue() == OldPtr) && "Pointer isn't an operand!"
) ? void (0) : __assert_fail ("(SI.getTrueValue() == OldPtr || SI.getFalseValue() == OldPtr) && \"Pointer isn't an operand!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3365, __extension__ __PRETTY_FUNCTION__
));
  assert(BeginOffset >= NewAllocaBeginOffset && "Selects are unsplittable")(static_cast <bool> (BeginOffset >= NewAllocaBeginOffset
 && "Selects are unsplittable") ? void (0) : __assert_fail
 ("BeginOffset >= NewAllocaBeginOffset && \"Selects are unsplittable\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3366, __extension__ __PRETTY_FUNCTION__
));
  assert(EndOffset <= NewAllocaEndOffset && "Selects are unsplittable")(static_cast <bool> (EndOffset <= NewAllocaEndOffset
 && "Selects are unsplittable") ? void (0) : __assert_fail
 ("EndOffset <= NewAllocaEndOffset && \"Selects are unsplittable\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3367, __extension__ __PRETTY_FUNCTION__
));

  Value *NewPtr = getNewAllocaSlicePtr(IRB, OldPtr->getType());
  // Replace the operands which were using the old pointer.
  if (SI.getOperand(1) == OldPtr)
    SI.setOperand(1, NewPtr);
  if (SI.getOperand(2) == OldPtr)
    SI.setOperand(2, NewPtr);

  LLVM_DEBUG(dbgs() << "          to: " << SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << SI <<
 "\n"; } } while (false);
  deleteIfTriviallyDead(OldPtr);

  // Fix the alignment of any loads or stores using this select.
  fixLoadStoreAlign(SI);

  // Selects can't be promoted on their own, but often can be speculated. We
  // check the speculation outside of the rewriter so that we see the
  // fully-rewritten alloca.
  SelectUsers.insert(&SI);
  return true;
}
3388};

3390namespace {

3392/// Visitor to rewrite aggregate loads and stores as scalar.
3393///
3394/// This pass aggressively rewrites all aggregate loads and stores on
3395/// a particular pointer (or any pointer derived from it which we can identify)
3396/// with scalar loads and stores.
3397class AggLoadStoreRewriter : public InstVisitor<AggLoadStoreRewriter, bool> {
// Befriend the base class so it can delegate to private visit methods.
friend class InstVisitor<AggLoadStoreRewriter, bool>;

/// Queue of pointer uses to analyze and potentially rewrite.
SmallVector<Use *, 8> Queue;

/// Set to prevent us from cycling with phi nodes and loops.
SmallPtrSet<User *, 8> Visited;

/// The current pointer use being rewritten. This is used to dig up the used
/// value (as opposed to the user).
Use *U = nullptr;

/// Used to calculate offsets, and hence alignment, of subobjects.
const DataLayout &DL;

IRBuilderTy &IRB;

3416public:
AggLoadStoreRewriter(const DataLayout &DL, IRBuilderTy &IRB)
    : DL(DL), IRB(IRB) {}

/// Rewrite loads and stores through a pointer and all pointers derived from
/// it.
bool rewrite(Instruction &I) {
  LLVM_DEBUG(dbgs() << "  Rewriting FCA loads and stores...\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting FCA loads and stores...\n"
; } } while (false);
  enqueueUsers(I);
  bool Changed = false;
  while (!Queue.empty()) {
    U = Queue.pop_back_val();
    Changed |= visit(cast<Instruction>(U->getUser()));
  }
  return Changed;
}

3433private:
/// Enqueue all the users of the given instruction for further processing.
/// This uses a set to de-duplicate users.
void enqueueUsers(Instruction &I) {
  for (Use &U : I.uses())
    if (Visited.insert(U.getUser()).second)
      Queue.push_back(&U);
}

// Conservative default is to not rewrite anything.
bool visitInstruction(Instruction &I) { return false; }

/// Generic recursive split emission class.
template <typename Derived> class OpSplitter {
protected:
  /// The builder used to form new instructions.
  IRBuilderTy &IRB;

  /// The indices which to be used with insert- or extractvalue to select the
  /// appropriate value within the aggregate.
  SmallVector<unsigned, 4> Indices;

  /// The indices to a GEP instruction which will move Ptr to the correct slot
  /// within the aggregate.
  SmallVector<Value *, 4> GEPIndices;

  /// The base pointer of the original op, used as a base for GEPing the
  /// split operations.
  Value *Ptr;

  /// The base pointee type being GEPed into.
  Type *BaseTy;

  /// Known alignment of the base pointer.
  Align BaseAlign;

  /// To calculate offset of each component so we can correctly deduce
  /// alignments.
  const DataLayout &DL;

  /// Initialize the splitter with an insertion point, Ptr and start with a
  /// single zero GEP index.
  OpSplitter(Instruction *InsertionPoint, Value *Ptr, Type *BaseTy,
             Align BaseAlign, const DataLayout &DL, IRBuilderTy &IRB)
      : IRB(IRB), GEPIndices(1, IRB.getInt32(0)), Ptr(Ptr), BaseTy(BaseTy),
        BaseAlign(BaseAlign), DL(DL) {
    IRB.SetInsertPoint(InsertionPoint);
  }

public:
  /// Generic recursive split emission routine.
  ///
  /// This method recursively splits an aggregate op (load or store) into
  /// scalar or vector ops. It splits recursively until it hits a single value
  /// and emits that single value operation via the template argument.
  ///
  /// The logic of this routine relies on GEPs and insertvalue and
  /// extractvalue all operating with the same fundamental index list, merely
  /// formatted differently (GEPs need actual values).
  ///
  /// \param Ty  The type being split recursively into smaller ops.
  /// \param Agg The aggregate value being built up or stored, depending on
  /// whether this is splitting a load or a store respectively.
  void emitSplitOps(Type *Ty, Value *&Agg, const Twine &Name) {
    if (Ty->isSingleValueType()) {
      unsigned Offset = DL.getIndexedOffsetInType(BaseTy, GEPIndices);
      return static_cast<Derived *>(this)->emitFunc(
          Ty, Agg, commonAlignment(BaseAlign, Offset), Name);
    }

    if (ArrayType *ATy = dyn_cast<ArrayType>(Ty)) {
      unsigned OldSize = Indices.size();
      (void)OldSize;
      for (unsigned Idx = 0, Size = ATy->getNumElements(); Idx != Size;
           ++Idx) {
        assert(Indices.size() == OldSize && "Did not return to the old size")(static_cast <bool> (Indices.size() == OldSize &&
 "Did not return to the old size") ? void (0) : __assert_fail
 ("Indices.size() == OldSize && \"Did not return to the old size\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3508, __extension__ __PRETTY_FUNCTION__
));
        Indices.push_back(Idx);
        GEPIndices.push_back(IRB.getInt32(Idx));
        emitSplitOps(ATy->getElementType(), Agg, Name + "." + Twine(Idx));
        GEPIndices.pop_back();
        Indices.pop_back();
      }
      return;
    }

    if (StructType *STy = dyn_cast<StructType>(Ty)) {
      unsigned OldSize = Indices.size();
      (void)OldSize;
      for (unsigned Idx = 0, Size = STy->getNumElements(); Idx != Size;
           ++Idx) {
        assert(Indices.size() == OldSize && "Did not return to the old size")(static_cast <bool> (Indices.size() == OldSize &&
 "Did not return to the old size") ? void (0) : __assert_fail
 ("Indices.size() == OldSize && \"Did not return to the old size\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3523, __extension__ __PRETTY_FUNCTION__
));
        Indices.push_back(Idx);
        GEPIndices.push_back(IRB.getInt32(Idx));
        emitSplitOps(STy->getElementType(Idx), Agg, Name + "." + Twine(Idx));
        GEPIndices.pop_back();
        Indices.pop_back();
      }
      return;
    }

    llvm_unreachable("Only arrays and structs are aggregate loadable types")::llvm::llvm_unreachable_internal("Only arrays and structs are aggregate loadable types"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3533);
  }
};

struct LoadOpSplitter : public OpSplitter<LoadOpSplitter> {
  AAMDNodes AATags;

  LoadOpSplitter(Instruction *InsertionPoint, Value *Ptr, Type *BaseTy,
                 AAMDNodes AATags, Align BaseAlign, const DataLayout &DL,
                 IRBuilderTy &IRB)
      : OpSplitter<LoadOpSplitter>(InsertionPoint, Ptr, BaseTy, BaseAlign, DL,
                                   IRB),
        AATags(AATags) {}

  /// Emit a leaf load of a single value. This is called at the leaves of the
  /// recursive emission to actually load values.
  void emitFunc(Type *Ty, Value *&Agg, Align Alignment, const Twine &Name) {
    assert(Ty->isSingleValueType())(static_cast <bool> (Ty->isSingleValueType()) ? void
 (0) : __assert_fail ("Ty->isSingleValueType()", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 3550, __extension__ __PRETTY_FUNCTION__));
    // Load the single value and insert it using the indices.
    Value *GEP =
        IRB.CreateInBoundsGEP(BaseTy, Ptr, GEPIndices, Name + ".gep");
    LoadInst *Load =
        IRB.CreateAlignedLoad(Ty, GEP, Alignment, Name + ".load");

    APInt Offset(
        DL.getIndexSizeInBits(Ptr->getType()->getPointerAddressSpace()), 0);
    if (AATags &&
        GEPOperator::accumulateConstantOffset(BaseTy, GEPIndices, DL, Offset))
      Load->setAAMetadata(AATags.shift(Offset.getZExtValue()));

    Agg = IRB.CreateInsertValue(Agg, Load, Indices, Name + ".insert");
    LLVM_DEBUG(dbgs() << "          to: " << *Load << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *Load <<
 "\n"; } } while (false);
  }
};

bool visitLoadInst(LoadInst &LI) {
  assert(LI.getPointerOperand() == *U)(static_cast <bool> (LI.getPointerOperand() == *U) ? void
 (0) : __assert_fail ("LI.getPointerOperand() == *U", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 3569, __extension__ __PRETTY_FUNCTION__));
  if (!LI.isSimple() || LI.getType()->isSingleValueType())
    return false;

  // We have an aggregate being loaded, split it apart.
  LLVM_DEBUG(dbgs() << "    original: " << LI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << LI <<
 "\n"; } } while (false);
  LoadOpSplitter Splitter(&LI, *U, LI.getType(), LI.getAAMetadata(),
                          getAdjustedAlignment(&LI, 0), DL, IRB);
  Value *V = PoisonValue::get(LI.getType());
  Splitter.emitSplitOps(LI.getType(), V, LI.getName() + ".fca");
  Visited.erase(&LI);
  LI.replaceAllUsesWith(V);
  LI.eraseFromParent();
  return true;
}

struct StoreOpSplitter : public OpSplitter<StoreOpSplitter> {
  StoreOpSplitter(Instruction *InsertionPoint, Value *Ptr, Type *BaseTy,
                  AAMDNodes AATags, Align BaseAlign, const DataLayout &DL,
                  IRBuilderTy &IRB)
      : OpSplitter<StoreOpSplitter>(InsertionPoint, Ptr, BaseTy, BaseAlign,
                                    DL, IRB),
        AATags(AATags) {}
  AAMDNodes AATags;
  /// Emit a leaf store of a single value. This is called at the leaves of the
  /// recursive emission to actually produce stores.
  void emitFunc(Type *Ty, Value *&Agg, Align Alignment, const Twine &Name) {
    assert(Ty->isSingleValueType())(static_cast <bool> (Ty->isSingleValueType()) ? void
 (0) : __assert_fail ("Ty->isSingleValueType()", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 3596, __extension__ __PRETTY_FUNCTION__));
    // Extract the single value and store it using the indices.
    //
    // The gep and extractvalue values are factored out of the CreateStore
    // call to make the output independent of the argument evaluation order.
    Value *ExtractValue =
        IRB.CreateExtractValue(Agg, Indices, Name + ".extract");
    Value *InBoundsGEP =
        IRB.CreateInBoundsGEP(BaseTy, Ptr, GEPIndices, Name + ".gep");
    StoreInst *Store =
        IRB.CreateAlignedStore(ExtractValue, InBoundsGEP, Alignment);

    APInt Offset(
        DL.getIndexSizeInBits(Ptr->getType()->getPointerAddressSpace()), 0);
    if (AATags &&
        GEPOperator::accumulateConstantOffset(BaseTy, GEPIndices, DL, Offset))
      Store->setAAMetadata(AATags.shift(Offset.getZExtValue()));

    LLVM_DEBUG(dbgs() << "          to: " << *Store << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          to: " << *Store <<
 "\n"; } } while (false);
  }
};

bool visitStoreInst(StoreInst &SI) {
  if (!SI.isSimple() || SI.getPointerOperand() != *U)
    return false;
  Value *V = SI.getValueOperand();
  if (V->getType()->isSingleValueType())
    return false;

  // We have an aggregate being stored, split it apart.
  LLVM_DEBUG(dbgs() << "    original: " << SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    original: " << SI <<
 "\n"; } } while (false);
  StoreOpSplitter Splitter(&SI, *U, V->getType(), SI.getAAMetadata(),
                           getAdjustedAlignment(&SI, 0), DL, IRB);
  Splitter.emitSplitOps(V->getType(), V, V->getName() + ".fca");
  Visited.erase(&SI);
  SI.eraseFromParent();
  return true;
}

bool visitBitCastInst(BitCastInst &BC) {
  enqueueUsers(BC);
  return false;
}

bool visitAddrSpaceCastInst(AddrSpaceCastInst &ASC) {
  enqueueUsers(ASC);
  return false;
}

// Fold gep (select cond, ptr1, ptr2) => select cond, gep(ptr1), gep(ptr2)
bool foldGEPSelect(GetElementPtrInst &GEPI) {
  if (!GEPI.hasAllConstantIndices())
    return false;

  SelectInst *Sel = cast<SelectInst>(GEPI.getPointerOperand());

  LLVM_DEBUG(dbgs() << "  Rewriting gep(select) -> select(gep):"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting gep(select) -> select(gep):"
 << "\n    original: " << *Sel << "\n              "
 << GEPI; } } while (false)
                    << "\n    original: " << *Seldo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting gep(select) -> select(gep):"
 << "\n    original: " << *Sel << "\n              "
 << GEPI; } } while (false)
                    << "\n              " << GEPI)do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting gep(select) -> select(gep):"
 << "\n    original: " << *Sel << "\n              "
 << GEPI; } } while (false);

  IRB.SetInsertPoint(&GEPI);
  SmallVector<Value *, 4> Index(GEPI.indices());
  bool IsInBounds = GEPI.isInBounds();

  Type *Ty = GEPI.getSourceElementType();
  Value *True = Sel->getTrueValue();
  Value *NTrue = IRB.CreateGEP(Ty, True, Index, True->getName() + ".sroa.gep",
                               IsInBounds);

  Value *False = Sel->getFalseValue();

  Value *NFalse = IRB.CreateGEP(Ty, False, Index,
                                False->getName() + ".sroa.gep", IsInBounds);

  Value *NSel = IRB.CreateSelect(Sel->getCondition(), NTrue, NFalse,
                                 Sel->getName() + ".sroa.sel");
  Visited.erase(&GEPI);
  GEPI.replaceAllUsesWith(NSel);
  GEPI.eraseFromParent();
  Instruction *NSelI = cast<Instruction>(NSel);
  Visited.insert(NSelI);
  enqueueUsers(*NSelI);

  LLVM_DEBUG(dbgs() << "\n          to: " << *NTruedo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "\n          to: " << *NTrue
 << "\n              " << *NFalse << "\n              "
 << *NSel << '\n'; } } while (false)
                    << "\n              " << *NFalsedo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "\n          to: " << *NTrue
 << "\n              " << *NFalse << "\n              "
 << *NSel << '\n'; } } while (false)
                    << "\n              " << *NSel << '\n')do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "\n          to: " << *NTrue
 << "\n              " << *NFalse << "\n              "
 << *NSel << '\n'; } } while (false);

  return true;
}

// Fold gep (phi ptr1, ptr2) => phi gep(ptr1), gep(ptr2)
bool foldGEPPhi(GetElementPtrInst &GEPI) {
  if (!GEPI.hasAllConstantIndices())
    return false;

  PHINode *PHI = cast<PHINode>(GEPI.getPointerOperand());
  if (GEPI.getParent() != PHI->getParent() ||
      llvm::any_of(PHI->incoming_values(), [](Value *In)
        { Instruction *I = dyn_cast<Instruction>(In);
          return !I || isa<GetElementPtrInst>(I) || isa<PHINode>(I) ||
                 succ_empty(I->getParent()) ||
                 !I->getParent()->isLegalToHoistInto();
        }))
    return false;

  LLVM_DEBUG(dbgs() << "  Rewriting gep(phi) -> phi(gep):"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting gep(phi) -> phi(gep):"
 << "\n    original: " << *PHI << "\n              "
 << GEPI << "\n          to: "; } } while (false)
                    << "\n    original: " << *PHIdo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting gep(phi) -> phi(gep):"
 << "\n    original: " << *PHI << "\n              "
 << GEPI << "\n          to: "; } } while (false)
                    << "\n              " << GEPIdo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting gep(phi) -> phi(gep):"
 << "\n    original: " << *PHI << "\n              "
 << GEPI << "\n          to: "; } } while (false)
                    << "\n          to: ")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting gep(phi) -> phi(gep):"
 << "\n    original: " << *PHI << "\n              "
 << GEPI << "\n          to: "; } } while (false);

  SmallVector<Value *, 4> Index(GEPI.indices());
  bool IsInBounds = GEPI.isInBounds();
  IRB.SetInsertPoint(GEPI.getParent()->getFirstNonPHI());
  PHINode *NewPN = IRB.CreatePHI(GEPI.getType(), PHI->getNumIncomingValues(),
                                 PHI->getName() + ".sroa.phi");
  for (unsigned I = 0, E = PHI->getNumIncomingValues(); I != E; ++I) {
    BasicBlock *B = PHI->getIncomingBlock(I);
    Value *NewVal = nullptr;
    int Idx = NewPN->getBasicBlockIndex(B);
    if (Idx >= 0) {
      NewVal = NewPN->getIncomingValue(Idx);
    } else {
      Instruction *In = cast<Instruction>(PHI->getIncomingValue(I));

      IRB.SetInsertPoint(In->getParent(), std::next(In->getIterator()));
      Type *Ty = GEPI.getSourceElementType();
      NewVal = IRB.CreateGEP(Ty, In, Index, In->getName() + ".sroa.gep",
                             IsInBounds);
    }
    NewPN->addIncoming(NewVal, B);
  }

  Visited.erase(&GEPI);
  GEPI.replaceAllUsesWith(NewPN);
  GEPI.eraseFromParent();
  Visited.insert(NewPN);
  enqueueUsers(*NewPN);

  LLVM_DEBUG(for (Value *In : NewPN->incoming_values())do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { for (Value *In : NewPN->incoming_values()) dbgs
() << "\n              " << *In; dbgs() << "\n              "
 << *NewPN << '\n'; } } while (false)
               dbgs() << "\n              " << *In;do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { for (Value *In : NewPN->incoming_values()) dbgs
() << "\n              " << *In; dbgs() << "\n              "
 << *NewPN << '\n'; } } while (false)
             dbgs() << "\n              " << *NewPN << '\n')do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { for (Value *In : NewPN->incoming_values()) dbgs
() << "\n              " << *In; dbgs() << "\n              "
 << *NewPN << '\n'; } } while (false);

  return true;
}

bool visitGetElementPtrInst(GetElementPtrInst &GEPI) {
  if (isa<SelectInst>(GEPI.getPointerOperand()) &&
      foldGEPSelect(GEPI))
    return true;

  if (isa<PHINode>(GEPI.getPointerOperand()) &&
      foldGEPPhi(GEPI))
    return true;

  enqueueUsers(GEPI);
  return false;
}

bool visitPHINode(PHINode &PN) {
  enqueueUsers(PN);
  return false;
}

bool visitSelectInst(SelectInst &SI) {
  enqueueUsers(SI);
  return false;
}
3763};

3765} // end anonymous namespace

3767/// Strip aggregate type wrapping.
3768///
3769/// This removes no-op aggregate types wrapping an underlying type. It will
3770/// strip as many layers of types as it can without changing either the type
3771/// size or the allocated size.
3772static Type *stripAggregateTypeWrapping(const DataLayout &DL, Type *Ty) {
if (Ty->isSingleValueType())
  return Ty;

uint64_t AllocSize = DL.getTypeAllocSize(Ty).getFixedSize();
uint64_t TypeSize = DL.getTypeSizeInBits(Ty).getFixedSize();

Type *InnerTy;
if (ArrayType *ArrTy = dyn_cast<ArrayType>(Ty)) {
  InnerTy = ArrTy->getElementType();
} else if (StructType *STy = dyn_cast<StructType>(Ty)) {
  const StructLayout *SL = DL.getStructLayout(STy);
  unsigned Index = SL->getElementContainingOffset(0);
  InnerTy = STy->getElementType(Index);
} else {
  return Ty;
}

if (AllocSize > DL.getTypeAllocSize(InnerTy).getFixedSize() ||
    TypeSize > DL.getTypeSizeInBits(InnerTy).getFixedSize())
  return Ty;

return stripAggregateTypeWrapping(DL, InnerTy);
3795}

3797/// Try to find a partition of the aggregate type passed in for a given
3798/// offset and size.
3799///
3800/// This recurses through the aggregate type and tries to compute a subtype
3801/// based on the offset and size. When the offset and size span a sub-section
3802/// of an array, it will even compute a new array type for that sub-section,
3803/// and the same for structs.
3804///
3805/// Note that this routine is very strict and tries to find a partition of the
3806/// type which produces the *exact* right offset and size. It is not forgiving
3807/// when the size or offset cause either end of type-based partition to be off.
3808/// Also, this is a best-effort routine. It is reasonable to give up and not
3809/// return a type if necessary.
3810static Type *getTypePartition(const DataLayout &DL, Type *Ty, uint64_t Offset,
                            uint64_t Size) {
if (Offset == 0 && DL.getTypeAllocSize(Ty).getFixedSize() == Size)
  return stripAggregateTypeWrapping(DL, Ty);
if (Offset > DL.getTypeAllocSize(Ty).getFixedSize() ||
    (DL.getTypeAllocSize(Ty).getFixedSize() - Offset) < Size)
  return nullptr;

if (isa<ArrayType>(Ty) || isa<VectorType>(Ty)) {
   Type *ElementTy;
   uint64_t TyNumElements;
   if (auto *AT = dyn_cast<ArrayType>(Ty)) {
     ElementTy = AT->getElementType();
     TyNumElements = AT->getNumElements();
   } else {
     // FIXME: This isn't right for vectors with non-byte-sized or
     // non-power-of-two sized elements.
     auto *VT = cast<FixedVectorType>(Ty);
     ElementTy = VT->getElementType();
     TyNumElements = VT->getNumElements();
  }
  uint64_t ElementSize = DL.getTypeAllocSize(ElementTy).getFixedSize();
  uint64_t NumSkippedElements = Offset / ElementSize;
  if (NumSkippedElements >= TyNumElements)
    return nullptr;
  Offset -= NumSkippedElements * ElementSize;

  // First check if we need to recurse.
  if (Offset > 0 || Size < ElementSize) {
    // Bail if the partition ends in a different array element.
    if ((Offset + Size) > ElementSize)
      return nullptr;
    // Recurse through the element type trying to peel off offset bytes.
    return getTypePartition(DL, ElementTy, Offset, Size);
  }
  assert(Offset == 0)(static_cast <bool> (Offset == 0) ? void (0) : __assert_fail
 ("Offset == 0", "llvm/lib/Transforms/Scalar/SROA.cpp", 3845,
 __extension__ __PRETTY_FUNCTION__));

  if (Size == ElementSize)
    return stripAggregateTypeWrapping(DL, ElementTy);
  assert(Size > ElementSize)(static_cast <bool> (Size > ElementSize) ? void (0) :
 __assert_fail ("Size > ElementSize", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 3849, __extension__ __PRETTY_FUNCTION__));
  uint64_t NumElements = Size / ElementSize;
  if (NumElements * ElementSize != Size)
    return nullptr;
  return ArrayType::get(ElementTy, NumElements);
}

StructType *STy = dyn_cast<StructType>(Ty);
if (!STy)
  return nullptr;

const StructLayout *SL = DL.getStructLayout(STy);
if (Offset >= SL->getSizeInBytes())
  return nullptr;
uint64_t EndOffset = Offset + Size;
if (EndOffset > SL->getSizeInBytes())
  return nullptr;

unsigned Index = SL->getElementContainingOffset(Offset);
Offset -= SL->getElementOffset(Index);

Type *ElementTy = STy->getElementType(Index);
uint64_t ElementSize = DL.getTypeAllocSize(ElementTy).getFixedSize();
if (Offset >= ElementSize)
  return nullptr; // The offset points into alignment padding.

// See if any partition must be contained by the element.
if (Offset > 0 || Size < ElementSize) {
  if ((Offset + Size) > ElementSize)
    return nullptr;
  return getTypePartition(DL, ElementTy, Offset, Size);
}
assert(Offset == 0)(static_cast <bool> (Offset == 0) ? void (0) : __assert_fail
 ("Offset == 0", "llvm/lib/Transforms/Scalar/SROA.cpp", 3881,
 __extension__ __PRETTY_FUNCTION__));

if (Size == ElementSize)
  return stripAggregateTypeWrapping(DL, ElementTy);

StructType::element_iterator EI = STy->element_begin() + Index,
                             EE = STy->element_end();
if (EndOffset < SL->getSizeInBytes()) {
  unsigned EndIndex = SL->getElementContainingOffset(EndOffset);
  if (Index == EndIndex)
    return nullptr; // Within a single element and its padding.

  // Don't try to form "natural" types if the elements don't line up with the
  // expected size.
  // FIXME: We could potentially recurse down through the last element in the
  // sub-struct to find a natural end point.
  if (SL->getElementOffset(EndIndex) != EndOffset)
    return nullptr;

  assert(Index < EndIndex)(static_cast <bool> (Index < EndIndex) ? void (0) : __assert_fail
 ("Index < EndIndex", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 3900, __extension__ __PRETTY_FUNCTION__));
  EE = STy->element_begin() + EndIndex;
}

// Try to build up a sub-structure.
StructType *SubTy =
    StructType::get(STy->getContext(), makeArrayRef(EI, EE), STy->isPacked());
const StructLayout *SubSL = DL.getStructLayout(SubTy);
if (Size != SubSL->getSizeInBytes())
  return nullptr; // The sub-struct doesn't have quite the size needed.

return SubTy;
3912}

3914/// Pre-split loads and stores to simplify rewriting.
3915///
3916/// We want to break up the splittable load+store pairs as much as
3917/// possible. This is important to do as a preprocessing step, as once we
3918/// start rewriting the accesses to partitions of the alloca we lose the
3919/// necessary information to correctly split apart paired loads and stores
3920/// which both point into this alloca. The case to consider is something like
3921/// the following:
3922///
3923///   %a = alloca [12 x i8]
3924///   %gep1 = getelementptr i8, ptr %a, i32 0
3925///   %gep2 = getelementptr i8, ptr %a, i32 4
3926///   %gep3 = getelementptr i8, ptr %a, i32 8
3927///   store float 0.0, ptr %gep1
3928///   store float 1.0, ptr %gep2
3929///   %v = load i64, ptr %gep1
3930///   store i64 %v, ptr %gep2
3931///   %f1 = load float, ptr %gep2
3932///   %f2 = load float, ptr %gep3
3933///
3934/// Here we want to form 3 partitions of the alloca, each 4 bytes large, and
3935/// promote everything so we recover the 2 SSA values that should have been
3936/// there all along.
3937///
3938/// \returns true if any changes are made.
3939bool SROAPass::presplitLoadsAndStores(AllocaInst &AI, AllocaSlices &AS) {
LLVM_DEBUG(dbgs() << "Pre-splitting loads and stores\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "Pre-splitting loads and stores\n"
; } } while (false);

// Track the loads and stores which are candidates for pre-splitting here, in
// the order they first appear during the partition scan. These give stable
// iteration order and a basis for tracking which loads and stores we
// actually split.
SmallVector<LoadInst *, 4> Loads;
SmallVector<StoreInst *, 4> Stores;

// We need to accumulate the splits required of each load or store where we
// can find them via a direct lookup. This is important to cross-check loads
// and stores against each other. We also track the slice so that we can kill
// all the slices that end up split.
struct SplitOffsets {
  Slice *S;
  std::vector<uint64_t> Splits;
};
SmallDenseMap<Instruction *, SplitOffsets, 8> SplitOffsetsMap;

// Track loads out of this alloca which cannot, for any reason, be pre-split.
// This is important as we also cannot pre-split stores of those loads!
// FIXME: This is all pretty gross. It means that we can be more aggressive
// in pre-splitting when the load feeding the store happens to come from
// a separate alloca. Put another way, the effectiveness of SROA would be
// decreased by a frontend which just concatenated all of its local allocas
// into one big flat alloca. But defeating such patterns is exactly the job
// SROA is tasked with! Sadly, to not have this discrepancy we would have
// change store pre-splitting to actually force pre-splitting of the load
// that feeds it *and all stores*. That makes pre-splitting much harder, but
// maybe it would make it more principled?
SmallPtrSet<LoadInst *, 8> UnsplittableLoads;

LLVM_DEBUG(dbgs() << "  Searching for candidate loads and stores\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Searching for candidate loads and stores\n"
; } } while (false);
for (auto &P : AS.partitions()) {
  for (Slice &S : P) {
    Instruction *I = cast<Instruction>(S.getUse()->getUser());
    if (!S.isSplittable() || S.endOffset() <= P.endOffset()) {
      // If this is a load we have to track that it can't participate in any
      // pre-splitting. If this is a store of a load we have to track that
      // that load also can't participate in any pre-splitting.
      if (auto *LI = dyn_cast<LoadInst>(I))
        UnsplittableLoads.insert(LI);
      else if (auto *SI = dyn_cast<StoreInst>(I))
        if (auto *LI = dyn_cast<LoadInst>(SI->getValueOperand()))
          UnsplittableLoads.insert(LI);
      continue;
    }
    assert(P.endOffset() > S.beginOffset() &&(static_cast <bool> (P.endOffset() > S.beginOffset()
 && "Empty or backwards partition!") ? void (0) : __assert_fail
 ("P.endOffset() > S.beginOffset() && \"Empty or backwards partition!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3988, __extension__ __PRETTY_FUNCTION__
))
           "Empty or backwards partition!")(static_cast <bool> (P.endOffset() > S.beginOffset()
 && "Empty or backwards partition!") ? void (0) : __assert_fail
 ("P.endOffset() > S.beginOffset() && \"Empty or backwards partition!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3988, __extension__ __PRETTY_FUNCTION__
));

    // Determine if this is a pre-splittable slice.
    if (auto *LI = dyn_cast<LoadInst>(I)) {
      assert(!LI->isVolatile() && "Cannot split volatile loads!")(static_cast <bool> (!LI->isVolatile() && "Cannot split volatile loads!"
) ? void (0) : __assert_fail ("!LI->isVolatile() && \"Cannot split volatile loads!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 3992, __extension__ __PRETTY_FUNCTION__
));

      // The load must be used exclusively to store into other pointers for
      // us to be able to arbitrarily pre-split it. The stores must also be
      // simple to avoid changing semantics.
      auto IsLoadSimplyStored = [](LoadInst *LI) {
        for (User *LU : LI->users()) {
          auto *SI = dyn_cast<StoreInst>(LU);
          if (!SI || !SI->isSimple())
            return false;
        }
        return true;
      };
      if (!IsLoadSimplyStored(LI)) {
        UnsplittableLoads.insert(LI);
        continue;
      }

      Loads.push_back(LI);
    } else if (auto *SI = dyn_cast<StoreInst>(I)) {
      if (S.getUse() != &SI->getOperandUse(SI->getPointerOperandIndex()))
        // Skip stores *of* pointers. FIXME: This shouldn't even be possible!
        continue;
      auto *StoredLoad = dyn_cast<LoadInst>(SI->getValueOperand());
      if (!StoredLoad || !StoredLoad->isSimple())
        continue;
      assert(!SI->isVolatile() && "Cannot split volatile stores!")(static_cast <bool> (!SI->isVolatile() && "Cannot split volatile stores!"
) ? void (0) : __assert_fail ("!SI->isVolatile() && \"Cannot split volatile stores!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4018, __extension__ __PRETTY_FUNCTION__
));

      Stores.push_back(SI);
    } else {
      // Other uses cannot be pre-split.
      continue;
    }

    // Record the initial split.
    LLVM_DEBUG(dbgs() << "    Candidate: " << *I << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    Candidate: " << *I <<
 "\n"; } } while (false);
    auto &Offsets = SplitOffsetsMap[I];
    assert(Offsets.Splits.empty() &&(static_cast <bool> (Offsets.Splits.empty() && "Should not have splits the first time we see an instruction!"
) ? void (0) : __assert_fail ("Offsets.Splits.empty() && \"Should not have splits the first time we see an instruction!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4030, __extension__ __PRETTY_FUNCTION__
))
           "Should not have splits the first time we see an instruction!")(static_cast <bool> (Offsets.Splits.empty() && "Should not have splits the first time we see an instruction!"
) ? void (0) : __assert_fail ("Offsets.Splits.empty() && \"Should not have splits the first time we see an instruction!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4030, __extension__ __PRETTY_FUNCTION__
));
    Offsets.S = &S;
    Offsets.Splits.push_back(P.endOffset() - S.beginOffset());
  }

  // Now scan the already split slices, and add a split for any of them which
  // we're going to pre-split.
  for (Slice *S : P.splitSliceTails()) {
    auto SplitOffsetsMapI =
        SplitOffsetsMap.find(cast<Instruction>(S->getUse()->getUser()));
    if (SplitOffsetsMapI == SplitOffsetsMap.end())
      continue;
    auto &Offsets = SplitOffsetsMapI->second;

    assert(Offsets.S == S && "Found a mismatched slice!")(static_cast <bool> (Offsets.S == S && "Found a mismatched slice!"
) ? void (0) : __assert_fail ("Offsets.S == S && \"Found a mismatched slice!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4044, __extension__ __PRETTY_FUNCTION__
));
    assert(!Offsets.Splits.empty() &&(static_cast <bool> (!Offsets.Splits.empty() &&
 "Cannot have an empty set of splits on the second partition!"
) ? void (0) : __assert_fail ("!Offsets.Splits.empty() && \"Cannot have an empty set of splits on the second partition!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4046, __extension__ __PRETTY_FUNCTION__
))
           "Cannot have an empty set of splits on the second partition!")(static_cast <bool> (!Offsets.Splits.empty() &&
 "Cannot have an empty set of splits on the second partition!"
) ? void (0) : __assert_fail ("!Offsets.Splits.empty() && \"Cannot have an empty set of splits on the second partition!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4046, __extension__ __PRETTY_FUNCTION__
));
    assert(Offsets.Splits.back() ==(static_cast <bool> (Offsets.Splits.back() == P.beginOffset
() - Offsets.S->beginOffset() && "Previous split does not end where this one begins!"
) ? void (0) : __assert_fail ("Offsets.Splits.back() == P.beginOffset() - Offsets.S->beginOffset() && \"Previous split does not end where this one begins!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4049, __extension__ __PRETTY_FUNCTION__
))
               P.beginOffset() - Offsets.S->beginOffset() &&(static_cast <bool> (Offsets.Splits.back() == P.beginOffset
() - Offsets.S->beginOffset() && "Previous split does not end where this one begins!"
) ? void (0) : __assert_fail ("Offsets.Splits.back() == P.beginOffset() - Offsets.S->beginOffset() && \"Previous split does not end where this one begins!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4049, __extension__ __PRETTY_FUNCTION__
))
           "Previous split does not end where this one begins!")(static_cast <bool> (Offsets.Splits.back() == P.beginOffset
() - Offsets.S->beginOffset() && "Previous split does not end where this one begins!"
) ? void (0) : __assert_fail ("Offsets.Splits.back() == P.beginOffset() - Offsets.S->beginOffset() && \"Previous split does not end where this one begins!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4049, __extension__ __PRETTY_FUNCTION__
));

    // Record each split. The last partition's end isn't needed as the size
    // of the slice dictates that.
    if (S->endOffset() > P.endOffset())
      Offsets.Splits.push_back(P.endOffset() - Offsets.S->beginOffset());
  }
}

// We may have split loads where some of their stores are split stores. For
// such loads and stores, we can only pre-split them if their splits exactly
// match relative to their starting offset. We have to verify this prior to
// any rewriting.
llvm::erase_if(Stores, [&UnsplittableLoads, &SplitOffsetsMap](StoreInst *SI) {
  // Lookup the load we are storing in our map of split
  // offsets.
  auto *LI = cast<LoadInst>(SI->getValueOperand());
  // If it was completely unsplittable, then we're done,
  // and this store can't be pre-split.
  if (UnsplittableLoads.count(LI))
    return true;

  auto LoadOffsetsI = SplitOffsetsMap.find(LI);
  if (LoadOffsetsI == SplitOffsetsMap.end())
    return false; // Unrelated loads are definitely safe.
  auto &LoadOffsets = LoadOffsetsI->second;

  // Now lookup the store's offsets.
  auto &StoreOffsets = SplitOffsetsMap[SI];

  // If the relative offsets of each split in the load and
  // store match exactly, then we can split them and we
  // don't need to remove them here.
  if (LoadOffsets.Splits == StoreOffsets.Splits)
    return false;

  LLVM_DEBUG(dbgs() << "    Mismatched splits for load and store:\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    Mismatched splits for load and store:\n"
 << "      " << *LI << "\n" << "      "
 << *SI << "\n"; } } while (false)
                    << "      " << *LI << "\n"do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    Mismatched splits for load and store:\n"
 << "      " << *LI << "\n" << "      "
 << *SI << "\n"; } } while (false)
                    << "      " << *SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    Mismatched splits for load and store:\n"
 << "      " << *LI << "\n" << "      "
 << *SI << "\n"; } } while (false);

  // We've found a store and load that we need to split
  // with mismatched relative splits. Just give up on them
  // and remove both instructions from our list of
  // candidates.
  UnsplittableLoads.insert(LI);
  return true;
});
// Now we have to go *back* through all the stores, because a later store may
// have caused an earlier store's load to become unsplittable and if it is
// unsplittable for the later store, then we can't rely on it being split in
// the earlier store either.
llvm::erase_if(Stores, [&UnsplittableLoads](StoreInst *SI) {
  auto *LI = cast<LoadInst>(SI->getValueOperand());
  return UnsplittableLoads.count(LI);
});
// Once we've established all the loads that can't be split for some reason,
// filter any that made it into our list out.
llvm::erase_if(Loads, [&UnsplittableLoads](LoadInst *LI) {
  return UnsplittableLoads.count(LI);
});

// If no loads or stores are left, there is no pre-splitting to be done for
// this alloca.
if (Loads.empty() && Stores.empty())
  return false;

// From here on, we can't fail and will be building new accesses, so rig up
// an IR builder.
IRBuilderTy IRB(&AI);

// Collect the new slices which we will merge into the alloca slices.
SmallVector<Slice, 4> NewSlices;

// Track any allocas we end up splitting loads and stores for so we iterate
// on them.
SmallPtrSet<AllocaInst *, 4> ResplitPromotableAllocas;

// At this point, we have collected all of the loads and stores we can
// pre-split, and the specific splits needed for them. We actually do the
// splitting in a specific order in order to handle when one of the loads in
// the value operand to one of the stores.
//
// First, we rewrite all of the split loads, and just accumulate each split
// load in a parallel structure. We also build the slices for them and append
// them to the alloca slices.
SmallDenseMap<LoadInst *, std::vector<LoadInst *>, 1> SplitLoadsMap;
std::vector<LoadInst *> SplitLoads;
const DataLayout &DL = AI.getModule()->getDataLayout();
for (LoadInst *LI : Loads) {
  SplitLoads.clear();

  auto &Offsets = SplitOffsetsMap[LI];
  unsigned SliceSize = Offsets.S->endOffset() - Offsets.S->beginOffset();
  assert(LI->getType()->getIntegerBitWidth() % 8 == 0 &&(static_cast <bool> (LI->getType()->getIntegerBitWidth
() % 8 == 0 && "Load must have type size equal to store size"
) ? void (0) : __assert_fail ("LI->getType()->getIntegerBitWidth() % 8 == 0 && \"Load must have type size equal to store size\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4143, __extension__ __PRETTY_FUNCTION__
))
         "Load must have type size equal to store size")(static_cast <bool> (LI->getType()->getIntegerBitWidth
() % 8 == 0 && "Load must have type size equal to store size"
) ? void (0) : __assert_fail ("LI->getType()->getIntegerBitWidth() % 8 == 0 && \"Load must have type size equal to store size\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4143, __extension__ __PRETTY_FUNCTION__
));
  assert(LI->getType()->getIntegerBitWidth() / 8 >= SliceSize &&(static_cast <bool> (LI->getType()->getIntegerBitWidth
() / 8 >= SliceSize && "Load must be >= slice size"
) ? void (0) : __assert_fail ("LI->getType()->getIntegerBitWidth() / 8 >= SliceSize && \"Load must be >= slice size\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4145, __extension__ __PRETTY_FUNCTION__
))
         "Load must be >= slice size")(static_cast <bool> (LI->getType()->getIntegerBitWidth
() / 8 >= SliceSize && "Load must be >= slice size"
) ? void (0) : __assert_fail ("LI->getType()->getIntegerBitWidth() / 8 >= SliceSize && \"Load must be >= slice size\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4145, __extension__ __PRETTY_FUNCTION__
));

  uint64_t BaseOffset = Offsets.S->beginOffset();
  assert(BaseOffset + SliceSize > BaseOffset &&(static_cast <bool> (BaseOffset + SliceSize > BaseOffset
 && "Cannot represent alloca access size using 64-bit integers!"
) ? void (0) : __assert_fail ("BaseOffset + SliceSize > BaseOffset && \"Cannot represent alloca access size using 64-bit integers!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4149, __extension__ __PRETTY_FUNCTION__
))
         "Cannot represent alloca access size using 64-bit integers!")(static_cast <bool> (BaseOffset + SliceSize > BaseOffset
 && "Cannot represent alloca access size using 64-bit integers!"
) ? void (0) : __assert_fail ("BaseOffset + SliceSize > BaseOffset && \"Cannot represent alloca access size using 64-bit integers!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4149, __extension__ __PRETTY_FUNCTION__
));

  Instruction *BasePtr = cast<Instruction>(LI->getPointerOperand());
  IRB.SetInsertPoint(LI);

  LLVM_DEBUG(dbgs() << "  Splitting load: " << *LI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Splitting load: " << *LI
 << "\n"; } } while (false);

  uint64_t PartOffset = 0, PartSize = Offsets.Splits.front();
  int Idx = 0, Size = Offsets.Splits.size();
  for (;;) {
    auto *PartTy = Type::getIntNTy(LI->getContext(), PartSize * 8);
    auto AS = LI->getPointerAddressSpace();
    auto *PartPtrTy = PartTy->getPointerTo(AS);
    LoadInst *PLoad = IRB.CreateAlignedLoad(
        PartTy,
        getAdjustedPtr(IRB, DL, BasePtr,
                       APInt(DL.getIndexSizeInBits(AS), PartOffset),
                       PartPtrTy, BasePtr->getName() + "."),
        getAdjustedAlignment(LI, PartOffset),
        /*IsVolatile*/ false, LI->getName());
    PLoad->copyMetadata(*LI, {LLVMContext::MD_mem_parallel_loop_access,
                              LLVMContext::MD_access_group});

    // Append this load onto the list of split loads so we can find it later
    // to rewrite the stores.
    SplitLoads.push_back(PLoad);

    // Now build a new slice for the alloca.
    NewSlices.push_back(
        Slice(BaseOffset + PartOffset, BaseOffset + PartOffset + PartSize,
              &PLoad->getOperandUse(PLoad->getPointerOperandIndex()),
              /*IsSplittable*/ false));
    LLVM_DEBUG(dbgs() << "    new slice [" << NewSlices.back().beginOffset()do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    new slice [" << NewSlices
.back().beginOffset() << ", " << NewSlices.back()
.endOffset() << "): " << *PLoad << "\n"; } }
 while (false)
                      << ", " << NewSlices.back().endOffset()do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    new slice [" << NewSlices
.back().beginOffset() << ", " << NewSlices.back()
.endOffset() << "): " << *PLoad << "\n"; } }
 while (false)
                      << "): " << *PLoad << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    new slice [" << NewSlices
.back().beginOffset() << ", " << NewSlices.back()
.endOffset() << "): " << *PLoad << "\n"; } }
 while (false);

    // See if we've handled all the splits.
    if (Idx >= Size)
      break;

    // Setup the next partition.
    PartOffset = Offsets.Splits[Idx];
    ++Idx;
    PartSize = (Idx < Size ? Offsets.Splits[Idx] : SliceSize) - PartOffset;
  }

  // Now that we have the split loads, do the slow walk over all uses of the
  // load and rewrite them as split stores, or save the split loads to use
  // below if the store is going to be split there anyways.
  bool DeferredStores = false;
  for (User *LU : LI->users()) {
    StoreInst *SI = cast<StoreInst>(LU);
    if (!Stores.empty() && SplitOffsetsMap.count(SI)) {
      DeferredStores = true;
      LLVM_DEBUG(dbgs() << "    Deferred splitting of store: " << *SIdo { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    Deferred splitting of store: "
 << *SI << "\n"; } } while (false)
                        << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    Deferred splitting of store: "
 << *SI << "\n"; } } while (false);
      continue;
    }

    Value *StoreBasePtr = SI->getPointerOperand();
    IRB.SetInsertPoint(SI);

    LLVM_DEBUG(dbgs() << "    Splitting store of load: " << *SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    Splitting store of load: " <<
 *SI << "\n"; } } while (false);

    for (int Idx = 0, Size = SplitLoads.size(); Idx < Size; ++Idx) {
      LoadInst *PLoad = SplitLoads[Idx];
      uint64_t PartOffset = Idx == 0 ? 0 : Offsets.Splits[Idx - 1];
      auto *PartPtrTy =
          PLoad->getType()->getPointerTo(SI->getPointerAddressSpace());

      auto AS = SI->getPointerAddressSpace();
      StoreInst *PStore = IRB.CreateAlignedStore(
          PLoad,
          getAdjustedPtr(IRB, DL, StoreBasePtr,
                         APInt(DL.getIndexSizeInBits(AS), PartOffset),
                         PartPtrTy, StoreBasePtr->getName() + "."),
          getAdjustedAlignment(SI, PartOffset),
          /*IsVolatile*/ false);
      PStore->copyMetadata(*SI, {LLVMContext::MD_mem_parallel_loop_access,
                                 LLVMContext::MD_access_group});
      LLVM_DEBUG(dbgs() << "      +" << PartOffset << ":" << *PStore << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "      +" << PartOffset <<
 ":" << *PStore << "\n"; } } while (false);
    }

    // We want to immediately iterate on any allocas impacted by splitting
    // this store, and we have to track any promotable alloca (indicated by
    // a direct store) as needing to be resplit because it is no longer
    // promotable.
    if (AllocaInst *OtherAI = dyn_cast<AllocaInst>(StoreBasePtr)) {
      ResplitPromotableAllocas.insert(OtherAI);
      Worklist.insert(OtherAI);
    } else if (AllocaInst *OtherAI = dyn_cast<AllocaInst>(
                   StoreBasePtr->stripInBoundsOffsets())) {
      Worklist.insert(OtherAI);
    }

    // Mark the original store as dead.
    DeadInsts.push_back(SI);
  }

  // Save the split loads if there are deferred stores among the users.
  if (DeferredStores)
    SplitLoadsMap.insert(std::make_pair(LI, std::move(SplitLoads)));

  // Mark the original load as dead and kill the original slice.
  DeadInsts.push_back(LI);
  Offsets.S->kill();
}

// Second, we rewrite all of the split stores. At this point, we know that
// all loads from this alloca have been split already. For stores of such
// loads, we can simply look up the pre-existing split loads. For stores of
// other loads, we split those loads first and then write split stores of
// them.
for (StoreInst *SI : Stores) {
  auto *LI = cast<LoadInst>(SI->getValueOperand());
  IntegerType *Ty = cast<IntegerType>(LI->getType());
  assert(Ty->getBitWidth() % 8 == 0)(static_cast <bool> (Ty->getBitWidth() % 8 == 0) ? void
 (0) : __assert_fail ("Ty->getBitWidth() % 8 == 0", "llvm/lib/Transforms/Scalar/SROA.cpp"
, 4265, __extension__ __PRETTY_FUNCTION__));
  uint64_t StoreSize = Ty->getBitWidth() / 8;
  assert(StoreSize > 0 && "Cannot have a zero-sized integer store!")(static_cast <bool> (StoreSize > 0 && "Cannot have a zero-sized integer store!"
) ? void (0) : __assert_fail ("StoreSize > 0 && \"Cannot have a zero-sized integer store!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4267, __extension__ __PRETTY_FUNCTION__
));

  auto &Offsets = SplitOffsetsMap[SI];
  assert(StoreSize == Offsets.S->endOffset() - Offsets.S->beginOffset() &&(static_cast <bool> (StoreSize == Offsets.S->endOffset
() - Offsets.S->beginOffset() && "Slice size should always match load size exactly!"
) ? void (0) : __assert_fail ("StoreSize == Offsets.S->endOffset() - Offsets.S->beginOffset() && \"Slice size should always match load size exactly!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4271, __extension__ __PRETTY_FUNCTION__
))
         "Slice size should always match load size exactly!")(static_cast <bool> (StoreSize == Offsets.S->endOffset
() - Offsets.S->beginOffset() && "Slice size should always match load size exactly!"
) ? void (0) : __assert_fail ("StoreSize == Offsets.S->endOffset() - Offsets.S->beginOffset() && \"Slice size should always match load size exactly!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4271, __extension__ __PRETTY_FUNCTION__
));
  uint64_t BaseOffset = Offsets.S->beginOffset();
  assert(BaseOffset + StoreSize > BaseOffset &&(static_cast <bool> (BaseOffset + StoreSize > BaseOffset
 && "Cannot represent alloca access size using 64-bit integers!"
) ? void (0) : __assert_fail ("BaseOffset + StoreSize > BaseOffset && \"Cannot represent alloca access size using 64-bit integers!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4274, __extension__ __PRETTY_FUNCTION__
))
         "Cannot represent alloca access size using 64-bit integers!")(static_cast <bool> (BaseOffset + StoreSize > BaseOffset
 && "Cannot represent alloca access size using 64-bit integers!"
) ? void (0) : __assert_fail ("BaseOffset + StoreSize > BaseOffset && \"Cannot represent alloca access size using 64-bit integers!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4274, __extension__ __PRETTY_FUNCTION__
));

  Value *LoadBasePtr = LI->getPointerOperand();
  Instruction *StoreBasePtr = cast<Instruction>(SI->getPointerOperand());

  LLVM_DEBUG(dbgs() << "  Splitting store: " << *SI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Splitting store: " << *SI
 << "\n"; } } while (false);

  // Check whether we have an already split load.
  auto SplitLoadsMapI = SplitLoadsMap.find(LI);
  std::vector<LoadInst *> *SplitLoads = nullptr;
  if (SplitLoadsMapI != SplitLoadsMap.end()) {
    SplitLoads = &SplitLoadsMapI->second;
    assert(SplitLoads->size() == Offsets.Splits.size() + 1 &&(static_cast <bool> (SplitLoads->size() == Offsets.Splits
.size() + 1 && "Too few split loads for the number of splits in the store!"
) ? void (0) : __assert_fail ("SplitLoads->size() == Offsets.Splits.size() + 1 && \"Too few split loads for the number of splits in the store!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4287, __extension__ __PRETTY_FUNCTION__
))
           "Too few split loads for the number of splits in the store!")(static_cast <bool> (SplitLoads->size() == Offsets.Splits
.size() + 1 && "Too few split loads for the number of splits in the store!"
) ? void (0) : __assert_fail ("SplitLoads->size() == Offsets.Splits.size() + 1 && \"Too few split loads for the number of splits in the store!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4287, __extension__ __PRETTY_FUNCTION__
));
  } else {
    LLVM_DEBUG(dbgs() << "          of load: " << *LI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "          of load: " << *LI
 << "\n"; } } while (false);
  }

  uint64_t PartOffset = 0, PartSize = Offsets.Splits.front();
  int Idx = 0, Size = Offsets.Splits.size();
  for (;;) {
    auto *PartTy = Type::getIntNTy(Ty->getContext(), PartSize * 8);
    auto *LoadPartPtrTy = PartTy->getPointerTo(LI->getPointerAddressSpace());
    auto *StorePartPtrTy = PartTy->getPointerTo(SI->getPointerAddressSpace());

    // Either lookup a split load or create one.
    LoadInst *PLoad;
    if (SplitLoads) {
      PLoad = (*SplitLoads)[Idx];
    } else {
      IRB.SetInsertPoint(LI);
      auto AS = LI->getPointerAddressSpace();
      PLoad = IRB.CreateAlignedLoad(
          PartTy,
          getAdjustedPtr(IRB, DL, LoadBasePtr,
                         APInt(DL.getIndexSizeInBits(AS), PartOffset),
                         LoadPartPtrTy, LoadBasePtr->getName() + "."),
          getAdjustedAlignment(LI, PartOffset),
          /*IsVolatile*/ false, LI->getName());
      PLoad->copyMetadata(*LI, {LLVMContext::MD_mem_parallel_loop_access,
                                LLVMContext::MD_access_group});
    }

    // And store this partition.
    IRB.SetInsertPoint(SI);
    auto AS = SI->getPointerAddressSpace();
    StoreInst *PStore = IRB.CreateAlignedStore(
        PLoad,
        getAdjustedPtr(IRB, DL, StoreBasePtr,
                       APInt(DL.getIndexSizeInBits(AS), PartOffset),
                       StorePartPtrTy, StoreBasePtr->getName() + "."),
        getAdjustedAlignment(SI, PartOffset),
        /*IsVolatile*/ false);
    PStore->copyMetadata(*SI, {LLVMContext::MD_mem_parallel_loop_access,
                               LLVMContext::MD_access_group});

    // Now build a new slice for the alloca.
    NewSlices.push_back(
        Slice(BaseOffset + PartOffset, BaseOffset + PartOffset + PartSize,
              &PStore->getOperandUse(PStore->getPointerOperandIndex()),
              /*IsSplittable*/ false));
    LLVM_DEBUG(dbgs() << "    new slice [" << NewSlices.back().beginOffset()do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    new slice [" << NewSlices
.back().beginOffset() << ", " << NewSlices.back()
.endOffset() << "): " << *PStore << "\n"; }
 } while (false)
                      << ", " << NewSlices.back().endOffset()do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    new slice [" << NewSlices
.back().beginOffset() << ", " << NewSlices.back()
.endOffset() << "): " << *PStore << "\n"; }
 } while (false)
                      << "): " << *PStore << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "    new slice [" << NewSlices
.back().beginOffset() << ", " << NewSlices.back()
.endOffset() << "): " << *PStore << "\n"; }
 } while (false);
    if (!SplitLoads) {
      LLVM_DEBUG(dbgs() << "      of split load: " << *PLoad << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "      of split load: " << *
PLoad << "\n"; } } while (false);
    }

    // See if we've finished all the splits.
    if (Idx >= Size)
      break;

    // Setup the next partition.
    PartOffset = Offsets.Splits[Idx];
    ++Idx;
    PartSize = (Idx < Size ? Offsets.Splits[Idx] : StoreSize) - PartOffset;
  }

  // We want to immediately iterate on any allocas impacted by splitting
  // this load, which is only relevant if it isn't a load of this alloca and
  // thus we didn't already split the loads above. We also have to keep track
  // of any promotable allocas we split loads on as they can no longer be
  // promoted.
  if (!SplitLoads) {
    if (AllocaInst *OtherAI = dyn_cast<AllocaInst>(LoadBasePtr)) {
      assert(OtherAI != &AI && "We can't re-split our own alloca!")(static_cast <bool> (OtherAI != &AI && "We can't re-split our own alloca!"
) ? void (0) : __assert_fail ("OtherAI != &AI && \"We can't re-split our own alloca!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4359, __extension__ __PRETTY_FUNCTION__
));
      ResplitPromotableAllocas.insert(OtherAI);
      Worklist.insert(OtherAI);
    } else if (AllocaInst *OtherAI = dyn_cast<AllocaInst>(
                   LoadBasePtr->stripInBoundsOffsets())) {
      assert(OtherAI != &AI && "We can't re-split our own alloca!")(static_cast <bool> (OtherAI != &AI && "We can't re-split our own alloca!"
) ? void (0) : __assert_fail ("OtherAI != &AI && \"We can't re-split our own alloca!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4364, __extension__ __PRETTY_FUNCTION__
));
      Worklist.insert(OtherAI);
    }
  }

  // Mark the original store as dead now that we've split it up and kill its
  // slice. Note that we leave the original load in place unless this store
  // was its only use. It may in turn be split up if it is an alloca load
  // for some other alloca, but it may be a normal load. This may introduce
  // redundant loads, but where those can be merged the rest of the optimizer
  // should handle the merging, and this uncovers SSA splits which is more
  // important. In practice, the original loads will almost always be fully
  // split and removed eventually, and the splits will be merged by any
  // trivial CSE, including instcombine.
  if (LI->hasOneUse()) {
    assert(*LI->user_begin() == SI && "Single use isn't this store!")(static_cast <bool> (*LI->user_begin() == SI &&
 "Single use isn't this store!") ? void (0) : __assert_fail (
"*LI->user_begin() == SI && \"Single use isn't this store!\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4379, __extension__ __PRETTY_FUNCTION__
));
    DeadInsts.push_back(LI);
  }
  DeadInsts.push_back(SI);
  Offsets.S->kill();
}

// Remove the killed slices that have ben pre-split.
llvm::erase_if(AS, [](const Slice &S) { return S.isDead(); });

// Insert our new slices. This will sort and merge them into the sorted
// sequence.
AS.insert(NewSlices);

LLVM_DEBUG(dbgs() << "  Pre-split slices:\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Pre-split slices:\n"; } } while
 (false);
4394#ifndef NDEBUG
for (auto I = AS.begin(), E = AS.end(); I != E; ++I)
  LLVM_DEBUG(AS.print(dbgs(), I, "    "))do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { AS.print(dbgs(), I, "    "); } } while (false);
4397#endif

// Finally, don't try to promote any allocas that new require re-splitting.
// They have already been added to the worklist above.
llvm::erase_if(PromotableAllocas, [&](AllocaInst *AI) {
  return ResplitPromotableAllocas.count(AI);
});

return true;
4406}

4408/// Rewrite an alloca partition's users.
4409///
4410/// This routine drives both of the rewriting goals of the SROA pass. It tries
4411/// to rewrite uses of an alloca partition to be conducive for SSA value
4412/// promotion. If the partition needs a new, more refined alloca, this will
4413/// build that new alloca, preserving as much type information as possible, and
4414/// rewrite the uses of the old alloca to point at the new one and have the
4415/// appropriate new offsets. It also evaluates how successful the rewrite was
4416/// at enabling promotion and if it was successful queues the alloca to be
4417/// promoted.
4418AllocaInst *SROAPass::rewritePartition(AllocaInst &AI, AllocaSlices &AS,
                                     Partition &P) {
// Try to compute a friendly type for this partition of the alloca. This
// won't always succeed, in which case we fall back to a legal integer type
// or an i8 array of an appropriate size.
Type *SliceTy = nullptr;
VectorType *SliceVecTy = nullptr;
const DataLayout &DL = AI.getModule()->getDataLayout();
std::pair<Type *, IntegerType *> CommonUseTy =
    findCommonType(P.begin(), P.end(), P.endOffset());
// Do all uses operate on the same type?
if (CommonUseTy.first)
  if (DL.getTypeAllocSize(CommonUseTy.first).getFixedSize() >= P.size()) {
    SliceTy = CommonUseTy.first;
    SliceVecTy = dyn_cast<VectorType>(SliceTy);
  }
// If not, can we find an appropriate subtype in the original allocated type?
if (!SliceTy)
  if (Type *TypePartitionTy = getTypePartition(DL, AI.getAllocatedType(),
                                               P.beginOffset(), P.size()))
    SliceTy = TypePartitionTy;

// If still not, can we use the largest bitwidth integer type used?
if (!SliceTy && CommonUseTy.second)
  if (DL.getTypeAllocSize(CommonUseTy.second).getFixedSize() >= P.size()) {
    SliceTy = CommonUseTy.second;
    SliceVecTy = dyn_cast<VectorType>(SliceTy);
  }
if ((!SliceTy || (SliceTy->isArrayTy() &&
                  SliceTy->getArrayElementType()->isIntegerTy())) &&
    DL.isLegalInteger(P.size() * 8)) {
  SliceTy = Type::getIntNTy(*C, P.size() * 8);
}

// If the common use types are not viable for promotion then attempt to find
// another type that is viable.
if (SliceVecTy && !checkVectorTypeForPromotion(P, SliceVecTy, DL))
  if (Type *TypePartitionTy = getTypePartition(DL, AI.getAllocatedType(),
                                               P.beginOffset(), P.size())) {
    VectorType *TypePartitionVecTy = dyn_cast<VectorType>(TypePartitionTy);
    if (TypePartitionVecTy &&
        checkVectorTypeForPromotion(P, TypePartitionVecTy, DL))
      SliceTy = TypePartitionTy;
  }

if (!SliceTy)
  SliceTy = ArrayType::get(Type::getInt8Ty(*C), P.size());
assert(DL.getTypeAllocSize(SliceTy).getFixedSize() >= P.size())(static_cast <bool> (DL.getTypeAllocSize(SliceTy).getFixedSize
() >= P.size()) ? void (0) : __assert_fail ("DL.getTypeAllocSize(SliceTy).getFixedSize() >= P.size()"
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4465, __extension__ __PRETTY_FUNCTION__
));

bool IsIntegerPromotable = isIntegerWideningViable(P, SliceTy, DL);

VectorType *VecTy =
    IsIntegerPromotable ? nullptr : isVectorPromotionViable(P, DL);
if (VecTy)
  SliceTy = VecTy;

// Check for the case where we're going to rewrite to a new alloca of the
// exact same type as the original, and with the same access offsets. In that
// case, re-use the existing alloca, but still run through the rewriter to
// perform phi and select speculation.
// P.beginOffset() can be non-zero even with the same type in a case with
// out-of-bounds access (e.g. @PR35657 function in SROA/basictest.ll).
AllocaInst *NewAI;
if (SliceTy == AI.getAllocatedType() && P.beginOffset() == 0) {
  NewAI = &AI;
  // FIXME: We should be able to bail at this point with "nothing changed".
  // FIXME: We might want to defer PHI speculation until after here.
  // FIXME: return nullptr;
} else {
  // Make sure the alignment is compatible with P.beginOffset().
  const Align Alignment = commonAlignment(AI.getAlign(), P.beginOffset());
  // If we will get at least this much alignment from the type alone, leave
  // the alloca's alignment unconstrained.
  const bool IsUnconstrained = Alignment <= DL.getABITypeAlign(SliceTy);
  NewAI = new AllocaInst(
      SliceTy, AI.getAddressSpace(), nullptr,
      IsUnconstrained ? DL.getPrefTypeAlign(SliceTy) : Alignment,
      AI.getName() + ".sroa." + Twine(P.begin() - AS.begin()), &AI);
  // Copy the old AI debug location over to the new one.
  NewAI->setDebugLoc(AI.getDebugLoc());
  ++NumNewAllocas;
}

LLVM_DEBUG(dbgs() << "Rewriting alloca partition "do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "Rewriting alloca partition " <<
 "[" << P.beginOffset() << "," << P.endOffset
() << ") to: " << *NewAI << "\n"; } } while
 (false)
                  << "[" << P.beginOffset() << "," << P.endOffset()do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "Rewriting alloca partition " <<
 "[" << P.beginOffset() << "," << P.endOffset
() << ") to: " << *NewAI << "\n"; } } while
 (false)
                  << ") to: " << *NewAI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "Rewriting alloca partition " <<
 "[" << P.beginOffset() << "," << P.endOffset
() << ") to: " << *NewAI << "\n"; } } while
 (false);

// Track the high watermark on the worklist as it is only relevant for
// promoted allocas. We will reset it to this point if the alloca is not in
// fact scheduled for promotion.
unsigned PPWOldSize = PostPromotionWorklist.size();
unsigned NumUses = 0;
SmallSetVector<PHINode *, 8> PHIUsers;
SmallSetVector<SelectInst *, 8> SelectUsers;

AllocaSliceRewriter Rewriter(DL, AS, *this, AI, *NewAI, P.beginOffset(),
                             P.endOffset(), IsIntegerPromotable, VecTy,
                             PHIUsers, SelectUsers);
bool Promotable = true;
for (Slice *S : P.splitSliceTails()) {
  Promotable &= Rewriter.visit(S);
  ++NumUses;
}
for (Slice &S : P) {
  Promotable &= Rewriter.visit(&S);
  ++NumUses;
}

NumAllocaPartitionUses += NumUses;
MaxUsesPerAllocaPartition.updateMax(NumUses);

// Now that we've processed all the slices in the new partition, check if any
// PHIs or Selects would block promotion.
for (PHINode *PHI : PHIUsers)
  if (!isSafePHIToSpeculate(*PHI)) {
    Promotable = false;
    PHIUsers.clear();
    SelectUsers.clear();
    break;
  }

SmallVector<std::pair<SelectInst *, RewriteableMemOps>, 2>
    NewSelectsToRewrite;
NewSelectsToRewrite.reserve(SelectUsers.size());
for (SelectInst *Sel : SelectUsers) {
  std::optional<RewriteableMemOps> Ops =
      isSafeSelectToSpeculate(*Sel, PreserveCFG);
  if (!Ops) {
    Promotable = false;
    PHIUsers.clear();
    SelectUsers.clear();
    NewSelectsToRewrite.clear();
    break;
  }
  NewSelectsToRewrite.emplace_back(std::make_pair(Sel, *Ops));
}

if (Promotable) {
  for (Use *U : AS.getDeadUsesIfPromotable()) {
    auto *OldInst = dyn_cast<Instruction>(U->get());
    Value::dropDroppableUse(*U);
    if (OldInst)
      if (isInstructionTriviallyDead(OldInst))
        DeadInsts.push_back(OldInst);
  }
  if (PHIUsers.empty() && SelectUsers.empty()) {
    // Promote the alloca.
    PromotableAllocas.push_back(NewAI);
  } else {
    // If we have either PHIs or Selects to speculate, add them to those
    // worklists and re-queue the new alloca so that we promote in on the
    // next iteration.
    for (PHINode *PHIUser : PHIUsers)
      SpeculatablePHIs.insert(PHIUser);
    SelectsToRewrite.reserve(SelectsToRewrite.size() +
                             NewSelectsToRewrite.size());
    for (auto &&KV : llvm::make_range(
             std::make_move_iterator(NewSelectsToRewrite.begin()),
             std::make_move_iterator(NewSelectsToRewrite.end())))
      SelectsToRewrite.insert(std::move(KV));
    Worklist.insert(NewAI);
  }
} else {
  // Drop any post-promotion work items if promotion didn't happen.
  while (PostPromotionWorklist.size() > PPWOldSize)
    PostPromotionWorklist.pop_back();

  // We couldn't promote and we didn't create a new partition, nothing
  // happened.
  if (NewAI == &AI)
    return nullptr;

  // If we can't promote the alloca, iterate on it to check for new
  // refinements exposed by splitting the current alloca. Don't iterate on an
  // alloca which didn't actually change and didn't get promoted.
  Worklist.insert(NewAI);
}

return NewAI;
4597}

4599/// Walks the slices of an alloca and form partitions based on them,
4600/// rewriting each of their uses.
4601bool SROAPass::splitAlloca(AllocaInst &AI, AllocaSlices &AS) {
if (AS.begin() == AS.end())
  return false;

unsigned NumPartitions = 0;
bool Changed = false;
const DataLayout &DL = AI.getModule()->getDataLayout();

// First try to pre-split loads and stores.
Changed |= presplitLoadsAndStores(AI, AS);

// Now that we have identified any pre-splitting opportunities,
// mark loads and stores unsplittable except for the following case.
// We leave a slice splittable if all other slices are disjoint or fully
// included in the slice, such as whole-alloca loads and stores.
// If we fail to split these during pre-splitting, we want to force them
// to be rewritten into a partition.
bool IsSorted = true;

uint64_t AllocaSize =
    DL.getTypeAllocSize(AI.getAllocatedType()).getFixedSize();
const uint64_t MaxBitVectorSize = 1024;
if (AllocaSize <= MaxBitVectorSize) {
  // If a byte boundary is included in any load or store, a slice starting or
  // ending at the boundary is not splittable.
  SmallBitVector SplittableOffset(AllocaSize + 1, true);
  for (Slice &S : AS)
    for (unsigned O = S.beginOffset() + 1;
         O < S.endOffset() && O < AllocaSize; O++)
      SplittableOffset.reset(O);

  for (Slice &S : AS) {
    if (!S.isSplittable())
      continue;

    if ((S.beginOffset() > AllocaSize || SplittableOffset[S.beginOffset()]) &&
        (S.endOffset() > AllocaSize || SplittableOffset[S.endOffset()]))
      continue;

    if (isa<LoadInst>(S.getUse()->getUser()) ||
        isa<StoreInst>(S.getUse()->getUser())) {
      S.makeUnsplittable();
      IsSorted = false;
    }
  }
}
else {
  // We only allow whole-alloca splittable loads and stores
  // for a large alloca to avoid creating too large BitVector.
  for (Slice &S : AS) {
    if (!S.isSplittable())
      continue;

    if (S.beginOffset() == 0 && S.endOffset() >= AllocaSize)
      continue;

    if (isa<LoadInst>(S.getUse()->getUser()) ||
        isa<StoreInst>(S.getUse()->getUser())) {
      S.makeUnsplittable();
      IsSorted = false;
    }
  }
}

if (!IsSorted)
  llvm::sort(AS);

/// Describes the allocas introduced by rewritePartition in order to migrate
/// the debug info.
struct Fragment {
  AllocaInst *Alloca;
  uint64_t Offset;
  uint64_t Size;
  Fragment(AllocaInst *AI, uint64_t O, uint64_t S)
    : Alloca(AI), Offset(O), Size(S) {}
};
SmallVector<Fragment, 4> Fragments;

// Rewrite each partition.
for (auto &P : AS.partitions()) {
  if (AllocaInst *NewAI = rewritePartition(AI, AS, P)) {
    Changed = true;
    if (NewAI != &AI) {
      uint64_t SizeOfByte = 8;
      uint64_t AllocaSize =
          DL.getTypeSizeInBits(NewAI->getAllocatedType()).getFixedSize();
      // Don't include any padding.
      uint64_t Size = std::min(AllocaSize, P.size() * SizeOfByte);
      Fragments.push_back(Fragment(NewAI, P.beginOffset() * SizeOfByte, Size));
    }
  }
  ++NumPartitions;
}

NumAllocaPartitions += NumPartitions;
MaxPartitionsPerAlloca.updateMax(NumPartitions);

// Migrate debug information from the old alloca to the new alloca(s)
// and the individual partitions.
TinyPtrVector<DbgVariableIntrinsic *> DbgDeclares = FindDbgAddrUses(&AI);
for (DbgVariableIntrinsic *DbgDeclare : DbgDeclares) {
  auto *Expr = DbgDeclare->getExpression();
  DIBuilder DIB(*AI.getModule(), /*AllowUnresolved*/ false);
  uint64_t AllocaSize =
      DL.getTypeSizeInBits(AI.getAllocatedType()).getFixedSize();
  for (auto Fragment : Fragments) {
    // Create a fragment expression describing the new partition or reuse AI's
    // expression if there is only one partition.
    auto *FragmentExpr = Expr;
    if (Fragment.Size < AllocaSize || Expr->isFragment()) {
      // If this alloca is already a scalar replacement of a larger aggregate,
      // Fragment.Offset describes the offset inside the scalar.
      auto ExprFragment = Expr->getFragmentInfo();
      uint64_t Offset = ExprFragment ? ExprFragment->OffsetInBits : 0;
      uint64_t Start = Offset + Fragment.Offset;
      uint64_t Size = Fragment.Size;
      if (ExprFragment) {
        uint64_t AbsEnd =
            ExprFragment->OffsetInBits + ExprFragment->SizeInBits;
        if (Start >= AbsEnd)
          // No need to describe a SROAed padding.
          continue;
        Size = std::min(Size, AbsEnd - Start);
      }
      // The new, smaller fragment is stenciled out from the old fragment.
      if (auto OrigFragment = FragmentExpr->getFragmentInfo()) {
        assert(Start >= OrigFragment->OffsetInBits &&(static_cast <bool> (Start >= OrigFragment->OffsetInBits
 && "new fragment is outside of original fragment") ?
 void (0) : __assert_fail ("Start >= OrigFragment->OffsetInBits && \"new fragment is outside of original fragment\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4728, __extension__ __PRETTY_FUNCTION__
))
               "new fragment is outside of original fragment")(static_cast <bool> (Start >= OrigFragment->OffsetInBits
 && "new fragment is outside of original fragment") ?
 void (0) : __assert_fail ("Start >= OrigFragment->OffsetInBits && \"new fragment is outside of original fragment\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4728, __extension__ __PRETTY_FUNCTION__
));
        Start -= OrigFragment->OffsetInBits;
      }

      // The alloca may be larger than the variable.
      auto VarSize = DbgDeclare->getVariable()->getSizeInBits();
      if (VarSize) {
        if (Size > *VarSize)
          Size = *VarSize;
        if (Size == 0 || Start + Size > *VarSize)
          continue;
      }

      // Avoid creating a fragment expression that covers the entire variable.
      if (!VarSize || *VarSize != Size) {
        if (auto E =
                DIExpression::createFragmentExpression(Expr, Start, Size))
          FragmentExpr = *E;
        else
          continue;
      }
    }

    // Remove any existing intrinsics on the new alloca describing
    // the variable fragment.
    for (DbgVariableIntrinsic *OldDII : FindDbgAddrUses(Fragment.Alloca)) {
      auto SameVariableFragment = [](const DbgVariableIntrinsic *LHS,
                                     const DbgVariableIntrinsic *RHS) {
        return LHS->getVariable() == RHS->getVariable() &&
               LHS->getDebugLoc()->getInlinedAt() ==
                   RHS->getDebugLoc()->getInlinedAt();
      };
      if (SameVariableFragment(OldDII, DbgDeclare))
        OldDII->eraseFromParent();
    }

    DIB.insertDeclare(Fragment.Alloca, DbgDeclare->getVariable(), FragmentExpr,
                      DbgDeclare->getDebugLoc(), &AI);
  }
}
return Changed;
4769}

4771/// Clobber a use with poison, deleting the used value if it becomes dead.
4772void SROAPass::clobberUse(Use &U) {
Value *OldV = U;
// Replace the use with an poison value.
U = PoisonValue::get(OldV->getType());

// Check for this making an instruction dead. We have to garbage collect
// all the dead instructions to ensure the uses of any alloca end up being
// minimal.
if (Instruction *OldI = dyn_cast<Instruction>(OldV))
  if (isInstructionTriviallyDead(OldI)) {
    DeadInsts.push_back(OldI);
  }
4784}

4786/// Analyze an alloca for SROA.
4787///
4788/// This analyzes the alloca to ensure we can reason about it, builds
4789/// the slices of the alloca, and then hands it off to be split and
4790/// rewritten as needed.
4791std::pair<bool /*Changed*/, bool /*CFGChanged*/>
4792SROAPass::runOnAlloca(AllocaInst &AI) {
bool Changed = false;
bool CFGChanged = false;

LLVM_DEBUG(dbgs() << "SROA alloca: " << AI << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "SROA alloca: " << AI <<
 "\n"; } } while (false);
++NumAllocasAnalyzed;

// Special case dead allocas, as they're trivial.
if (AI.use_empty()) {
  AI.eraseFromParent();
  Changed = true;
  return {Changed, CFGChanged};
}
const DataLayout &DL = AI.getModule()->getDataLayout();

// Skip alloca forms that this analysis can't handle.
auto *AT = AI.getAllocatedType();
if (AI.isArrayAllocation() || !AT->isSized() || isa<ScalableVectorType>(AT) ||
    DL.getTypeAllocSize(AT).getFixedSize() == 0)
  return {Changed, CFGChanged};

// First, split any FCA loads and stores touching this alloca to promote
// better splitting and promotion opportunities.
IRBuilderTy IRB(&AI);
AggLoadStoreRewriter AggRewriter(DL, IRB);
Changed |= AggRewriter.rewrite(AI);

// Build the slices using a recursive instruction-visiting builder.
AllocaSlices AS(DL, AI);
LLVM_DEBUG(AS.print(dbgs()))do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { AS.print(dbgs()); } } while (false);
if (AS.isEscaped())
  return {Changed, CFGChanged};

// Delete all the dead users of this alloca before splitting and rewriting it.
for (Instruction *DeadUser : AS.getDeadUsers()) {
  // Free up everything used by this instruction.
  for (Use &DeadOp : DeadUser->operands())
    clobberUse(DeadOp);

  // Now replace the uses of this instruction.
  DeadUser->replaceAllUsesWith(PoisonValue::get(DeadUser->getType()));

  // And mark it for deletion.
  DeadInsts.push_back(DeadUser);
  Changed = true;
}
for (Use *DeadOp : AS.getDeadOperands()) {
  clobberUse(*DeadOp);
  Changed = true;
}

// No slices to split. Leave the dead alloca for a later pass to clean up.
if (AS.begin() == AS.end())
  return {Changed, CFGChanged};

Changed |= splitAlloca(AI, AS);

LLVM_DEBUG(dbgs() << "  Speculating PHIs\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Speculating PHIs\n"; } } while
 (false);
while (!SpeculatablePHIs.empty())
  speculatePHINodeLoads(IRB, *SpeculatablePHIs.pop_back_val());

LLVM_DEBUG(dbgs() << "  Rewriting Selects\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "  Rewriting Selects\n"; } } while
 (false);
auto RemainingSelectsToRewrite = SelectsToRewrite.takeVector();
while (!RemainingSelectsToRewrite.empty()) {
  const auto [K, V] = RemainingSelectsToRewrite.pop_back_val();
  CFGChanged |=
      rewriteSelectInstMemOps(*K, V, IRB, PreserveCFG ? nullptr : DTU);
}

return {Changed, CFGChanged};
4862}

4864/// Delete the dead instructions accumulated in this run.
4865///
4866/// Recursively deletes the dead instructions we've accumulated. This is done
4867/// at the very end to maximize locality of the recursive delete and to
4868/// minimize the problems of invalidated instruction pointers as such pointers
4869/// are used heavily in the intermediate stages of the algorithm.
4870///
4871/// We also record the alloca instructions deleted here so that they aren't
4872/// subsequently handed to mem2reg to promote.
4873bool SROAPass::deleteDeadInstructions(
  SmallPtrSetImpl<AllocaInst *> &DeletedAllocas) {
bool Changed = false;
while (!DeadInsts.empty()) {
  Instruction *I = dyn_cast_or_null<Instruction>(DeadInsts.pop_back_val());
  if (!I)
    continue;
  LLVM_DEBUG(dbgs() << "Deleting dead instruction: " << *I << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "Deleting dead instruction: " <<
 *I << "\n"; } } while (false);

  // If the instruction is an alloca, find the possible dbg.declare connected
  // to it, and remove it too. We must do this before calling RAUW or we will
  // not be able to find it.
  if (AllocaInst *AI = dyn_cast<AllocaInst>(I)) {
    DeletedAllocas.insert(AI);
    for (DbgVariableIntrinsic *OldDII : FindDbgAddrUses(AI))
      OldDII->eraseFromParent();
  }

  I->replaceAllUsesWith(UndefValue::get(I->getType()));

  for (Use &Operand : I->operands())
    if (Instruction *U = dyn_cast<Instruction>(Operand)) {
      // Zero out the operand and see if it becomes trivially dead.
      Operand = nullptr;
      if (isInstructionTriviallyDead(U))
        DeadInsts.push_back(U);
    }

  ++NumDeleted;
  I->eraseFromParent();
  Changed = true;
}
return Changed;
4906}

4908/// Promote the allocas, using the best available technique.
4909///
4910/// This attempts to promote whatever allocas have been identified as viable in
4911/// the PromotableAllocas list. If that list is empty, there is nothing to do.
4912/// This function returns whether any promotion occurred.
4913bool SROAPass::promoteAllocas(Function &F) {
if (PromotableAllocas.empty())
  return false;

NumPromoted += PromotableAllocas.size();

LLVM_DEBUG(dbgs() << "Promoting allocas with mem2reg...\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "Promoting allocas with mem2reg...\n"
; } } while (false);
PromoteMemToReg(PromotableAllocas, DTU->getDomTree(), AC);
PromotableAllocas.clear();
return true;
4923}

4925PreservedAnalyses SROAPass::runImpl(Function &F, DomTreeUpdater &RunDTU,
                                  AssumptionCache &RunAC) {
LLVM_DEBUG(dbgs() << "SROA function: " << F.getName() << "\n")do { if (::llvm::DebugFlag && ::llvm::isCurrentDebugType
("sroa")) { dbgs() << "SROA function: " << F.getName
() << "\n"; } } while (false);
C = &F.getContext();
DTU = &RunDTU;
AC = &RunAC;

BasicBlock &EntryBB = F.getEntryBlock();
for (BasicBlock::iterator I = EntryBB.begin(), E = std::prev(EntryBB.end());
     I != E; ++I) {
  if (AllocaInst *AI = dyn_cast<AllocaInst>(I)) {
    if (isa<ScalableVectorType>(AI->getAllocatedType())) {
      if (isAllocaPromotable(AI))
        PromotableAllocas.push_back(AI);
    } else {
      Worklist.insert(AI);
    }
  }
}

bool Changed = false;
bool CFGChanged = false;
// A set of deleted alloca instruction pointers which should be removed from
// the list of promotable allocas.
SmallPtrSet<AllocaInst *, 4> DeletedAllocas;

do {
  while (!Worklist.empty()) {
    auto [IterationChanged, IterationCFGChanged] =
        runOnAlloca(*Worklist.pop_back_val());
    Changed |= IterationChanged;
    CFGChanged |= IterationCFGChanged;

    Changed |= deleteDeadInstructions(DeletedAllocas);

    // Remove the deleted allocas from various lists so that we don't try to
    // continue processing them.
    if (!DeletedAllocas.empty()) {
      auto IsInSet = [&](AllocaInst *AI) { return DeletedAllocas.count(AI); };
      Worklist.remove_if(IsInSet);
      PostPromotionWorklist.remove_if(IsInSet);
      llvm::erase_if(PromotableAllocas, IsInSet);
      DeletedAllocas.clear();
    }
  }

  Changed |= promoteAllocas(F);

  Worklist = PostPromotionWorklist;
  PostPromotionWorklist.clear();
} while (!Worklist.empty());

assert((!CFGChanged || Changed) && "Can not only modify the CFG.")(static_cast <bool> ((!CFGChanged || Changed) &&
 "Can not only modify the CFG.") ? void (0) : __assert_fail (
"(!CFGChanged || Changed) && \"Can not only modify the CFG.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4977, __extension__ __PRETTY_FUNCTION__
));
assert((!CFGChanged || !PreserveCFG) &&(static_cast <bool> ((!CFGChanged || !PreserveCFG) &&
 "Should not have modified the CFG when told to preserve it."
) ? void (0) : __assert_fail ("(!CFGChanged || !PreserveCFG) && \"Should not have modified the CFG when told to preserve it.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4979, __extension__ __PRETTY_FUNCTION__
))
       "Should not have modified the CFG when told to preserve it.")(static_cast <bool> ((!CFGChanged || !PreserveCFG) &&
 "Should not have modified the CFG when told to preserve it."
) ? void (0) : __assert_fail ("(!CFGChanged || !PreserveCFG) && \"Should not have modified the CFG when told to preserve it.\""
, "llvm/lib/Transforms/Scalar/SROA.cpp", 4979, __extension__ __PRETTY_FUNCTION__
));

if (!Changed)
  return PreservedAnalyses::all();

PreservedAnalyses PA;
if (!CFGChanged)
  PA.preserveSet<CFGAnalyses>();
PA.preserve<DominatorTreeAnalysis>();
return PA;
4989}

4991PreservedAnalyses SROAPass::runImpl(Function &F, DominatorTree &RunDT,
                                  AssumptionCache &RunAC) {
DomTreeUpdater DTU(RunDT, DomTreeUpdater::UpdateStrategy::Lazy);
return runImpl(F, DTU, RunAC);
4995}

4997PreservedAnalyses SROAPass::run(Function &F, FunctionAnalysisManager &AM) {
return runImpl(F, AM.getResult<DominatorTreeAnalysis>(F),
               AM.getResult<AssumptionAnalysis>(F));
5000}

5002void SROAPass::printPipeline(
  raw_ostream &OS, function_ref<StringRef(StringRef)> MapClassName2PassName) {
static_cast<PassInfoMixin<SROAPass> *>(this)->printPipeline(
    OS, MapClassName2PassName);
OS << (PreserveCFG ? "<preserve-cfg>" : "<modify-cfg>");
5007}

5009SROAPass::SROAPass(SROAOptions PreserveCFG_)
  : PreserveCFG(PreserveCFG_ == SROAOptions::PreserveCFG) {}

5012/// A legacy pass for the legacy pass manager that wraps the \c SROA pass.
5013///
5014/// This is in the llvm namespace purely to allow it to be a friend of the \c
5015/// SROA pass.
5016class llvm::sroa::SROALegacyPass : public FunctionPass {
/// The SROA implementation.
SROAPass Impl;

5020public:
static char ID;

SROALegacyPass(SROAOptions PreserveCFG = SROAOptions::PreserveCFG)
    : FunctionPass(ID), Impl(PreserveCFG) {
  initializeSROALegacyPassPass(*PassRegistry::getPassRegistry());
}

bool runOnFunction(Function &F) override {
  if (skipFunction(F))
    return false;

  auto PA = Impl.runImpl(
      F, getAnalysis<DominatorTreeWrapperPass>().getDomTree(),
      getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F));
  return !PA.areAllPreserved();
}

void getAnalysisUsage(AnalysisUsage &AU) const override {
  AU.addRequired<AssumptionCacheTracker>();
  AU.addRequired<DominatorTreeWrapperPass>();
  AU.addPreserved<GlobalsAAWrapperPass>();
  AU.addPreserved<DominatorTreeWrapperPass>();
}

StringRef getPassName() const override { return "SROA"; }
5046};

5048char SROALegacyPass::ID = 0;

5050FunctionPass *llvm::createSROAPass(bool PreserveCFG) {
return new SROALegacyPass(PreserveCFG ? SROAOptions::PreserveCFG
                                      : SROAOptions::ModifyCFG);
5053}

5055INITIALIZE_PASS_BEGIN(SROALegacyPass, "sroa",static void *initializeSROALegacyPassPassOnce(PassRegistry &
Registry) {
                    "Scalar Replacement Of Aggregates", false, false)static void *initializeSROALegacyPassPassOnce(PassRegistry &
Registry) {
5057INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)initializeAssumptionCacheTrackerPass(Registry);
5058INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)initializeDominatorTreeWrapperPassPass(Registry);
5059INITIALIZE_PASS_END(SROALegacyPass, "sroa", "Scalar Replacement Of Aggregates",PassInfo *PI = new PassInfo( "Scalar Replacement Of Aggregates"
, "sroa", &SROALegacyPass::ID, PassInfo::NormalCtor_t(callDefaultCtor
<SROALegacyPass>), false, false); Registry.registerPass
(*PI, true); return PI; } static llvm::once_flag InitializeSROALegacyPassPassFlag
; void llvm::initializeSROALegacyPassPass(PassRegistry &Registry
) { llvm::call_once(InitializeSROALegacyPassPassFlag, initializeSROALegacyPassPassOnce
, std::ref(Registry)); }
                  false, false)PassInfo *PI = new PassInfo( "Scalar Replacement Of Aggregates"
, "sroa", &SROALegacyPass::ID, PassInfo::NormalCtor_t(callDefaultCtor
<SROALegacyPass>), false, false); Registry.registerPass
(*PI, true); return PI; } static llvm::once_flag InitializeSROALegacyPassPassFlag
; void llvm::initializeSROALegacyPassPass(PassRegistry &Registry
) { llvm::call_once(InitializeSROALegacyPassPassFlag, initializeSROALegacyPassPassOnce
, std::ref(Registry)); }