RegionUtils.cpp

Go to the documentation of this file.

//===- RegionUtils.cpp - Region-related transformation utilities ----------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
 
#include "mlir/Transforms/RegionUtils.h"
 
#include "mlir/Analysis/SliceAnalysis.h"
#include "mlir/Analysis/TopologicalSortUtils.h"
#include "mlir/IR/Block.h"
#include "mlir/IR/Dominance.h"
#include "mlir/IR/IRMapping.h"
#include "mlir/IR/Operation.h"
#include "mlir/IR/PatternMatch.h"
#include "mlir/IR/Value.h"
#include "mlir/Interfaces/ControlFlowInterfaces.h"
#include "mlir/Interfaces/SideEffectInterfaces.h"
#include "mlir/Support/LogicalResult.h"
#include "llvm/ADT/DenseSet.h"
#include "llvm/ADT/DepthFirstIterator.h"
#include "llvm/ADT/PostOrderIterator.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallVectorExtras.h"
#include "llvm/Support/DebugLog.h"
 
#include <deque>
#include <iterator>
 
using namespace mlir;
 
#define DEBUG_TYPE "region-utils"
 
void mlir::replaceAllUsesInRegionWith(Value orig, Value replacement,
                                      Region &region) {
  for (auto &use : llvm::make_early_inc_range(orig.getUses())) {
    if (region.isAncestor(use.getOwner()->getParentRegion()))
      use.set(replacement);
  }
}
 
void mlir::visitUsedValuesDefinedAbove(
    Region &region, Region &limit, function_ref<void(OpOperand *)> callback) {
  assert(limit.isAncestor(&region) &&
         "expected isolation limit to be an ancestor of the given region");
 
  // Collect proper ancestors of `limit` upfront to avoid traversing the region
  // tree for every value.
  SmallPtrSet<Region *, 4> properAncestors;
  for (auto *reg = limit.getParentRegion(); reg != nullptr;
       reg = reg->getParentRegion()) {
    properAncestors.insert(reg);
  }
 
  region.walk([callback, &properAncestors](Operation *op) {
    for (OpOperand &operand : op->getOpOperands())
      // Callback on values defined in a proper ancestor of region.
      if (properAncestors.count(operand.get().getParentRegion()))
        callback(&operand);
  });
}
 
void mlir::visitUsedValuesDefinedAbove(
    MutableArrayRef<Region> regions, function_ref<void(OpOperand *)> callback) {
  for (Region &region : regions)
    visitUsedValuesDefinedAbove(region, region, callback);
}
 
void mlir::getUsedValuesDefinedAbove(Region &region, Region &limit,
                                     SetVector<Value> &values) {
  visitUsedValuesDefinedAbove(region, limit, [&](OpOperand *operand) {
    values.insert(operand->get());
  });
}
 
void mlir::getUsedValuesDefinedAbove(MutableArrayRef<Region> regions,
                                     SetVector<Value> &values) {
  for (Region &region : regions)
    getUsedValuesDefinedAbove(region, region, values);
}
 
//===----------------------------------------------------------------------===//
// Make block isolated from above.
//===----------------------------------------------------------------------===//
 
SmallVector<Value> mlir::makeRegionIsolatedFromAbove(
    RewriterBase &rewriter, Region &region,
    llvm::function_ref<bool(Operation *)> cloneOperationIntoRegion) {
 
  // Get initial list of values used within region but defined above.
  llvm::SetVector<Value> initialCapturedValues;
  mlir::getUsedValuesDefinedAbove(region, initialCapturedValues);
 
  std::deque<Value> worklist(initialCapturedValues.begin(),
                             initialCapturedValues.end());
  llvm::DenseSet<Value> visited;
  llvm::DenseSet<Operation *> visitedOps;
 
  llvm::SetVector<Value> finalCapturedValues;
  SmallVector<Operation *> clonedOperations;
  while (!worklist.empty()) {
    Value currValue = worklist.front();
    worklist.pop_front();
    if (visited.count(currValue))
      continue;
    visited.insert(currValue);
 
    Operation *definingOp = currValue.getDefiningOp();
    if (!definingOp || visitedOps.count(definingOp)) {
      finalCapturedValues.insert(currValue);
      continue;
    }
    visitedOps.insert(definingOp);
 
    if (!cloneOperationIntoRegion(definingOp)) {
      // Defining operation isnt cloned, so add the current value to final
      // captured values list.
      finalCapturedValues.insert(currValue);
      continue;
    }
 
    // Add all operands of the operation to the worklist and mark the op as to
    // be cloned.
    for (Value operand : definingOp->getOperands()) {
      if (visited.count(operand))
        continue;
      worklist.push_back(operand);
    }
    clonedOperations.push_back(definingOp);
  }
 
  // The operations to be cloned need to be ordered in topological order
  // so that they can be cloned into the region without violating use-def
  // chains.
  mlir::computeTopologicalSorting(clonedOperations);
 
  OpBuilder::InsertionGuard g(rewriter);
  // Collect types of existing block
  Block *entryBlock = &region.front();
  SmallVector<Type> newArgTypes =
      llvm::to_vector(entryBlock->getArgumentTypes());
  SmallVector<Location> newArgLocs = llvm::map_to_vector(
      entryBlock->getArguments(), [](BlockArgument b) { return b.getLoc(); });
 
  // Append the types of the captured values.
  for (auto value : finalCapturedValues) {
    newArgTypes.push_back(value.getType());
    newArgLocs.push_back(value.getLoc());
  }
 
  // Create a new entry block.
  Block *newEntryBlock =
      rewriter.createBlock(&region, region.begin(), newArgTypes, newArgLocs);
  auto newEntryBlockArgs = newEntryBlock->getArguments();
 
  // Create a mapping between the captured values and the new arguments added.
  IRMapping map;
  auto replaceIfFn = [&](OpOperand &use) {
    return region.isAncestor(use.getOwner()->getParentRegion());
  };
 
  for (auto [arg, capturedVal] :
       llvm::zip(newEntryBlockArgs.take_back(finalCapturedValues.size()),
                 finalCapturedValues)) {
    map.map(capturedVal, arg);
    rewriter.replaceUsesWithIf(capturedVal, arg, replaceIfFn);
  }
  rewriter.setInsertionPointToStart(newEntryBlock);
  for (auto *clonedOp : clonedOperations) {
    Operation *newOp = rewriter.clone(*clonedOp, map);
    rewriter.replaceOpUsesWithIf(clonedOp, newOp->getResults(), replaceIfFn);
  }
  rewriter.mergeBlocks(
      entryBlock, newEntryBlock,
      newEntryBlock->getArguments().take_front(entryBlock->getNumArguments()));
  return llvm::to_vector(finalCapturedValues);
}
 
//===----------------------------------------------------------------------===//
// Unreachable Block Elimination
//===----------------------------------------------------------------------===//
 
/// Erase the unreachable blocks within the provided regions. Returns success
/// if any blocks were erased, failure otherwise.
// TODO: We could likely merge this with the DCE algorithm below.
LogicalResult mlir::eraseUnreachableBlocks(RewriterBase &rewriter,
                                           MutableArrayRef<Region> regions,
                                           bool recurse) {
  LDBG() << "Starting eraseUnreachableBlocks with " << regions.size()
         << " regions";
 
  // Set of blocks found to be reachable within a given region.
  llvm::df_iterator_default_set<Block *, 16> reachable;
  // If any blocks were found to be dead.
  int erasedDeadBlocks = 0;
 
  SmallVector<Region *, 1> worklist;
  worklist.reserve(regions.size());
  for (Region &region : regions)
    worklist.push_back(&region);
 
  LDBG(2) << "Initial worklist size: " << worklist.size();
 
  while (!worklist.empty()) {
    Region *region = worklist.pop_back_val();
    if (region->empty()) {
      LDBG(2) << "Skipping empty region";
      continue;
    }
 
    LDBG(2) << "Processing region with " << region->getBlocks().size()
            << " blocks";
    if (region->getParentOp())
      LDBG(2) << " -> for operation:  "
              << OpWithFlags(region->getParentOp(),
                             OpPrintingFlags().skipRegions());
 
    // If this is a single block region, just collect the nested regions.
    if (region->hasOneBlock()) {
      if (recurse)
        for (Operation &op : region->front())
          for (Region &region : op.getRegions())
            worklist.push_back(&region);
      continue;
    }
 
    // Mark all reachable blocks.
    reachable.clear();
    for (Block *block : depth_first_ext(&region->front(), reachable))
      (void)block /* Mark all reachable blocks */;
 
    LDBG(2) << "Found " << reachable.size() << " reachable blocks out of "
            << region->getBlocks().size() << " total blocks";
 
    // Collect all of the dead blocks and push the live regions onto the
    // worklist.
    for (Block &block : llvm::make_early_inc_range(*region)) {
      if (!reachable.count(&block)) {
        LDBG() << "Erasing unreachable block: " << &block;
        block.dropAllDefinedValueUses();
        rewriter.eraseBlock(&block);
        ++erasedDeadBlocks;
        continue;
      }
 
      // Walk any regions within this block.
      if (recurse)
        for (Operation &op : block)
          for (Region &region : op.getRegions())
            worklist.push_back(&region);
    }
  }
 
  LDBG() << "Finished eraseUnreachableBlocks, erased " << erasedDeadBlocks
         << " dead blocks";
 
  return success(erasedDeadBlocks > 0);
}
 
//===----------------------------------------------------------------------===//
// Dead Code Elimination
//===----------------------------------------------------------------------===//
 
namespace {
/// Data structure used to track which values have already been proved live.
///
/// Because Operation's can have multiple results, this data structure tracks
/// liveness for both Value's and Operation's to avoid having to look through
/// all Operation results when analyzing a use.
///
/// This data structure essentially tracks the dataflow lattice.
/// The set of values/ops proved live increases monotonically to a fixed-point.
class LiveMap {
public:
  /// Value methods.
  bool wasProvenLive(Value value) {
    // TODO: For results that are removable, e.g. for region based control flow,
    // we could allow for these values to be tracked independently.
    if (OpResult result = dyn_cast<OpResult>(value))
      return wasProvenLive(result.getOwner());
    return wasProvenLive(cast<BlockArgument>(value));
  }
  bool wasProvenLive(BlockArgument arg) { return liveValues.count(arg); }
  void setProvedLive(Value value) {
    // TODO: For results that are removable, e.g. for region based control flow,
    // we could allow for these values to be tracked independently.
    if (OpResult result = dyn_cast<OpResult>(value))
      return setProvedLive(result.getOwner());
    setProvedLive(cast<BlockArgument>(value));
  }
  void setProvedLive(BlockArgument arg) {
    changed |= liveValues.insert(arg).second;
  }
 
  /// Operation methods.
  bool wasProvenLive(Operation *op) { return liveOps.count(op); }
  void setProvedLive(Operation *op) { changed |= liveOps.insert(op).second; }
 
  /// Methods for tracking if we have reached a fixed-point.
  void resetChanged() { changed = false; }
  bool hasChanged() { return changed; }
 
private:
  bool changed = false;
  DenseSet<Value> liveValues;
  DenseSet<Operation *> liveOps;
};
} // namespace
 
static bool isUseSpeciallyKnownDead(OpOperand &use, LiveMap &liveMap) {
  Operation *owner = use.getOwner();
  unsigned operandIndex = use.getOperandNumber();
  // This pass generally treats all uses of an op as live if the op itself is
  // considered live. However, for successor operands to terminators we need a
  // finer-grained notion where we deduce liveness for operands individually.
  // The reason for this is easiest to think about in terms of a classical phi
  // node based SSA IR, where each successor operand is really an operand to a
  // *separate* phi node, rather than all operands to the branch itself as with
  // the block argument representation that MLIR uses.
  //
  // And similarly, because each successor operand is really an operand to a phi
  // node, rather than to the terminator op itself, a terminator op can't e.g.
  // "print" the value of a successor operand.
  if (owner->hasTrait<OpTrait::IsTerminator>()) {
    if (BranchOpInterface branchInterface = dyn_cast<BranchOpInterface>(owner))
      if (auto arg = branchInterface.getSuccessorBlockArgument(operandIndex))
        return !liveMap.wasProvenLive(*arg);
    return false;
  }
  return false;
}
 
static void processValue(Value value, LiveMap &liveMap) {
  bool provedLive = llvm::any_of(value.getUses(), [&](OpOperand &use) {
    if (isUseSpeciallyKnownDead(use, liveMap))
      return false;
    return liveMap.wasProvenLive(use.getOwner());
  });
  if (provedLive)
    liveMap.setProvedLive(value);
}
 
static void propagateLiveness(Region &region, LiveMap &liveMap);
 
static void propagateTerminatorLiveness(Operation *op, LiveMap &liveMap) {
  // Terminators are always live.
  liveMap.setProvedLive(op);
 
  // Check to see if we can reason about the successor operands and mutate them.
  BranchOpInterface branchInterface = dyn_cast<BranchOpInterface>(op);
  if (!branchInterface) {
    for (Block *successor : op->getSuccessors())
      for (BlockArgument arg : successor->getArguments())
        liveMap.setProvedLive(arg);
    return;
  }
 
  // If we can't reason about the operand to a successor, conservatively mark
  // it as live.
  for (unsigned i = 0, e = op->getNumSuccessors(); i != e; ++i) {
    SuccessorOperands successorOperands =
        branchInterface.getSuccessorOperands(i);
    for (unsigned opI = 0, opE = successorOperands.getProducedOperandCount();
         opI != opE; ++opI)
      liveMap.setProvedLive(op->getSuccessor(i)->getArgument(opI));
  }
}
 
static void propagateLiveness(Operation *op, LiveMap &liveMap) {
  // Recurse on any regions the op has.
  for (Region &region : op->getRegions())
    propagateLiveness(region, liveMap);
 
  // Process terminator operations.
  if (op->hasTrait<OpTrait::IsTerminator>())
    return propagateTerminatorLiveness(op, liveMap);
 
  // Don't reprocess live operations.
  if (liveMap.wasProvenLive(op))
    return;
 
  // Process the op itself.
  if (!wouldOpBeTriviallyDead(op))
    return liveMap.setProvedLive(op);
 
  // If the op isn't intrinsically alive, check it's results.
  for (Value value : op->getResults())
    processValue(value, liveMap);
}
 
static void propagateLiveness(Region &region, LiveMap &liveMap) {
  if (region.empty())
    return;
 
  for (Block *block : llvm::post_order(&region.front())) {
    // We process block arguments after the ops in the block, to promote
    // faster convergence to a fixed point (we try to visit uses before defs).
    for (Operation &op : llvm::reverse(block->getOperations()))
      propagateLiveness(&op, liveMap);
 
    // We currently do not remove entry block arguments, so there is no need to
    // track their liveness.
    // TODO: We could track these and enable removing dead operands/arguments
    // from region control flow operations.
    if (block->isEntryBlock())
      continue;
 
    for (Value value : block->getArguments()) {
      if (!liveMap.wasProvenLive(value))
        processValue(value, liveMap);
    }
  }
}
 
static void eraseTerminatorSuccessorOperands(Operation *terminator,
                                             LiveMap &liveMap) {
  BranchOpInterface branchOp = dyn_cast<BranchOpInterface>(terminator);
  if (!branchOp)
    return;
 
  for (unsigned succI = 0, succE = terminator->getNumSuccessors();
       succI < succE; succI++) {
    // Iterating successors in reverse is not strictly needed, since we
    // aren't erasing any successors. But it is slightly more efficient
    // since it will promote later operands of the terminator being erased
    // first, reducing the quadratic-ness.
    unsigned succ = succE - succI - 1;
    SuccessorOperands succOperands = branchOp.getSuccessorOperands(succ);
    Block *successor = terminator->getSuccessor(succ);
 
    for (unsigned argI = 0, argE = succOperands.size(); argI < argE; ++argI) {
      // Iterating args in reverse is needed for correctness, to avoid
      // shifting later args when earlier args are erased.
      unsigned arg = argE - argI - 1;
      if (!liveMap.wasProvenLive(successor->getArgument(arg)))
        succOperands.erase(arg);
    }
  }
}
 
static LogicalResult deleteDeadness(RewriterBase &rewriter,
                                    MutableArrayRef<Region> regions,
                                    LiveMap &liveMap) {
  bool erasedAnything = false;
  for (Region &region : regions) {
    if (region.empty())
      continue;
    bool hasSingleBlock = region.hasOneBlock();
 
    // Delete every operation that is not live. Graph regions may have cycles
    // in the use-def graph, so we must explicitly dropAllUses() from each
    // operation as we erase it. Visiting the operations in post-order
    // guarantees that in SSA CFG regions value uses are removed before defs,
    // which makes dropAllUses() a no-op.
    for (Block *block : llvm::post_order(&region.front())) {
      if (!hasSingleBlock)
        eraseTerminatorSuccessorOperands(block->getTerminator(), liveMap);
      for (Operation &childOp :
           llvm::make_early_inc_range(llvm::reverse(block->getOperations()))) {
        if (!liveMap.wasProvenLive(&childOp)) {
          erasedAnything = true;
          childOp.dropAllUses();
          rewriter.eraseOp(&childOp);
        } else {
          erasedAnything |= succeeded(
              deleteDeadness(rewriter, childOp.getRegions(), liveMap));
        }
      }
    }
    // Delete block arguments.
    // The entry block has an unknown contract with their enclosing block, so
    // skip it.
    for (Block &block : llvm::drop_begin(region.getBlocks(), 1)) {
      block.eraseArguments(
          [&](BlockArgument arg) { return !liveMap.wasProvenLive(arg); });
    }
  }
  return success(erasedAnything);
}
 
// This function performs a simple dead code elimination algorithm over the
// given regions.
//
// The overall goal is to prove that Values are dead, which allows deleting ops
// and block arguments.
//
// This uses an optimistic algorithm that assumes everything is dead until
// proved otherwise, allowing it to delete recursively dead cycles.
//
// This is a simple fixed-point dataflow analysis algorithm on a lattice
// {Dead,Alive}. Because liveness flows backward, we generally try to
// iterate everything backward to speed up convergence to the fixed-point. This
// allows for being able to delete recursively dead cycles of the use-def graph,
// including block arguments.
//
// This function returns success if any operations or arguments were deleted,
// failure otherwise.
LogicalResult mlir::runRegionDCE(RewriterBase &rewriter,
                                 MutableArrayRef<Region> regions) {
  LiveMap liveMap;
  do {
    liveMap.resetChanged();
 
    for (Region &region : regions)
      propagateLiveness(region, liveMap);
  } while (liveMap.hasChanged());
 
  return deleteDeadness(rewriter, regions, liveMap);
}
 
bool mlir::eliminateTriviallyDeadOps(RewriterBase &rewriter, Region &region,
                                     bool includeNestedRegions) {
  LDBG() << "Starting eliminateTriviallyDeadOps with "
         << region.getBlocks().size()
         << " blocks, includeNestedRegions=" << includeNestedRegions;
  if (Operation *parentOp = region.getParentOp())
    LDBG(2) << " -> parent operation: "
            << OpWithFlags(parentOp, OpPrintingFlags().skipRegions());
 
  bool changed = false;
  unsigned erasedOps = 0;
  unsigned seededOps = 0;
  unsigned enqueuedDefs = 0;
 
  // Step 1: walk each op in reverse program order. If the op is already
  // trivially dead, erase it outright — there's no point recursing into
  // regions that will be destroyed with it. Otherwise, if
  // `includeNestedRegions` is set, recurse into its nested regions so values
  // defined in `region` may lose their last user and show up as dead in
  // step 2's seed. Reverse iteration lets dead chains propagate within this
  // single pass.
  for (Block &block : llvm::reverse(region)) {
    LDBG(2) << "Scanning block " << &block << " with "
            << block.getOperations().size() << " operations";
    for (Operation &op :
         llvm::make_early_inc_range(llvm::reverse(block.getOperations()))) {
      LDBG(3) << "Visiting operation: "
              << OpWithFlags(&op, OpPrintingFlags().skipRegions());
      if (isOpTriviallyDead(&op)) {
        LDBG() << "Erasing trivially dead operation: "
               << OpWithFlags(&op, OpPrintingFlags().skipRegions());
        rewriter.eraseOp(&op);
        changed = true;
        ++erasedOps;
        continue;
      }
      if (includeNestedRegions) {
        unsigned regionIdx = 0;
        for (Region &nested : op.getRegions()) {
          LDBG(2) << "Recursing into nested region #" << regionIdx
                  << " of operation " << op.getName();
          bool nestedChanged =
              eliminateTriviallyDeadOps(rewriter, nested, includeNestedRegions);
          LDBG(2) << "Finished nested region #" << regionIdx << " of operation "
                  << op.getName() << ", changed=" << nestedChanged;
          changed |= nestedChanged;
          ++regionIdx;
        }
      }
    }
  }
 
  // Step 2: worklist over ops in this region only.
  //
  // Worklist invariant: an op is pushed only once we have verified it is
  // trivially dead. No speculative enqueues: every op on the worklist will
  // be erased when popped. Two things enforce this:
  //   - the initial seed below calls isOpTriviallyDead before enqueueing,
  //   - the propagation inside the loop drops the erasing op's use of
  //     `defOp` *before* re-checking isOpTriviallyDead(defOp), so the check
  //     sees the post-erase use count and only enqueues when actually dead.
  // Deadness is monotonic within this pass (we never add users, only remove
  // them), so an op that was dead at enqueue time is still dead at pop time.
  SmallVector<Operation *> worklist;
 
  LDBG(2) << "Stage 2: Seeding trivially dead operation worklist";
  for (Operation &op : region.getOps()) {
    if (isOpTriviallyDead(&op)) {
      LDBG(2) << "Seeded worklist with operation: "
              << OpWithFlags(&op, OpPrintingFlags().skipRegions());
      worklist.push_back(&op);
      changed = true;
      ++seededOps;
    }
  }
  LDBG(2) << "Initial worklist size: " << worklist.size();
 
  while (!worklist.empty()) {
    Operation *op = worklist.pop_back_val();
    LDBG(2) << "Popped operation from worklist: "
            << OpWithFlags(op, OpPrintingFlags().skipRegions());
    /// Erase each operand to drop its use count before checking its defining
    /// op: by the time we call isOpTriviallyDead on defOp, the
    /// about-to-be-erased `op` is no longer counted as a user. Only
    /// actually-dead ops enter the worklist.
    ///
    /// Walk nested operations as well because erasing `op` also implicitly
    /// erases every operation nested under it and therefore drops their operand
    /// uses.
    op->walk([&](Operation *erasedOp) {
      LDBG(3) << "Processing operands of operation erased: "
              << OpWithFlags(erasedOp, OpPrintingFlags().skipRegions());
      for (OpOperand &opOperand : erasedOp->getOpOperands()) {
        Operation *defOp = opOperand.get().getDefiningOp();
        if (!defOp) {
          LDBG(4) << "Skipping operand #" << opOperand.getOperandNumber()
                  << ": value has no defining operation";
          continue;
        }
        if (defOp->getParentRegion() != &region) {
          LDBG(4) << "Skipping operand #" << opOperand.getOperandNumber()
                  << ": defining operation is outside the current region";
          continue;
        }
        LDBG(4) << "Dropping operand #" << opOperand.getOperandNumber()
                << " from defining operation: "
                << OpWithFlags(defOp, OpPrintingFlags().skipRegions());
        opOperand.drop();
        if (isOpTriviallyDead(defOp)) {
          LDBG(2) << "Enqueued newly trivially dead defining operation: "
                  << OpWithFlags(defOp, OpPrintingFlags().skipRegions());
          worklist.push_back(defOp);
          ++enqueuedDefs;
        } else {
          LDBG(4) << "Defining operation is still not trivially dead: "
                  << OpWithFlags(defOp, OpPrintingFlags().skipRegions());
        }
      }
    });
    LDBG() << "Erasing trivially dead worklist operation: "
           << OpWithFlags(op, OpPrintingFlags().skipRegions());
    rewriter.eraseOp(op);
    ++erasedOps;
  }
  LDBG() << "Finished eliminateTriviallyDeadOps, erased " << erasedOps
         << " operations, seeded " << seededOps << " operations, enqueued "
         << enqueuedDefs << " defining operations, changed=" << changed;
  return changed;
}
 
//===----------------------------------------------------------------------===//
// Block Merging
//===----------------------------------------------------------------------===//
 
//===----------------------------------------------------------------------===//
// BlockEquivalenceData
//===----------------------------------------------------------------------===//
 
namespace {
/// This class contains the information for comparing the equivalencies of two
/// blocks. Blocks are considered equivalent if they contain the same operations
/// in the same order. The only allowed divergence is for operands that come
/// from sources outside of the parent block, i.e. the uses of values produced
/// within the block must be equivalent.
///   e.g.,
/// Equivalent:
///  ^bb1(%arg0: i32)
///    return %arg0, %foo : i32, i32
///  ^bb2(%arg1: i32)
///    return %arg1, %bar : i32, i32
/// Not Equivalent:
///  ^bb1(%arg0: i32)
///    return %foo, %arg0 : i32, i32
///  ^bb2(%arg1: i32)
///    return %arg1, %bar : i32, i32
struct BlockEquivalenceData {
  BlockEquivalenceData(Block *block);
 
  /// Return the order index for the given value that is within the block of
  /// this data.
  unsigned getOrderOf(Value value) const;
 
  /// The block this data refers to.
  Block *block;
  /// A hash value for this block.
  llvm::hash_code hash;
  /// A map of result producing operations to their relative orders within this
  /// block. The order of an operation is the number of defined values that are
  /// produced within the block before this operation.
  DenseMap<Operation *, unsigned> opOrderIndex;
};
} // namespace
 
BlockEquivalenceData::BlockEquivalenceData(Block *block)
    : block(block), hash(0) {
  unsigned orderIt = block->getNumArguments();
  for (Operation &op : *block) {
    if (unsigned numResults = op.getNumResults()) {
      opOrderIndex.try_emplace(&op, orderIt);
      orderIt += numResults;
    }
    auto opHash = OperationEquivalence::computeHash(
        &op, OperationEquivalence::ignoreHashValue,
        OperationEquivalence::ignoreHashValue,
        OperationEquivalence::IgnoreLocations);
    hash = llvm::hash_combine(hash, opHash);
  }
}
 
unsigned BlockEquivalenceData::getOrderOf(Value value) const {
  assert(value.getParentBlock() == block && "expected value of this block");
 
  // Arguments use the argument number as the order index.
  if (BlockArgument arg = dyn_cast<BlockArgument>(value))
    return arg.getArgNumber();
 
  // Otherwise, the result order is offset from the parent op's order.
  OpResult result = cast<OpResult>(value);
  auto opOrderIt = opOrderIndex.find(result.getDefiningOp());
  assert(opOrderIt != opOrderIndex.end() && "expected op to have an order");
  return opOrderIt->second + result.getResultNumber();
}
 
//===----------------------------------------------------------------------===//
// BlockMergeCluster
//===----------------------------------------------------------------------===//
 
namespace {
/// This class represents a cluster of blocks to be merged together.
class BlockMergeCluster {
public:
  BlockMergeCluster(BlockEquivalenceData &&leaderData)
      : leaderData(std::move(leaderData)) {}
 
  /// Attempt to add the given block to this cluster. Returns success if the
  /// block was merged, failure otherwise.
  LogicalResult addToCluster(BlockEquivalenceData &blockData);
 
  /// Try to merge all of the blocks within this cluster into the leader block.
  LogicalResult merge(RewriterBase &rewriter);
 
private:
  /// The equivalence data for the leader of the cluster.
  BlockEquivalenceData leaderData;
 
  /// The set of blocks that can be merged into the leader.
  llvm::SmallSetVector<Block *, 1> blocksToMerge;
 
  /// A set of operand+index pairs that correspond to operands that need to be
  /// replaced by arguments when the cluster gets merged.
  std::set<std::pair<int, int>> operandsToMerge;
};
} // namespace
 
LogicalResult BlockMergeCluster::addToCluster(BlockEquivalenceData &blockData) {
  if (leaderData.hash != blockData.hash)
    return failure();
  Block *leaderBlock = leaderData.block, *mergeBlock = blockData.block;
  if (leaderBlock->getArgumentTypes() != mergeBlock->getArgumentTypes())
    return failure();
 
  // A set of operands that mismatch between the leader and the new block.
  SmallVector<std::pair<int, int>, 8> mismatchedOperands;
  auto lhsIt = leaderBlock->begin(), lhsE = leaderBlock->end();
  auto rhsIt = blockData.block->begin(), rhsE = blockData.block->end();
  for (int opI = 0; lhsIt != lhsE && rhsIt != rhsE; ++lhsIt, ++rhsIt, ++opI) {
    // Check that the operations are equivalent.
    if (!OperationEquivalence::isEquivalentTo(
            &*lhsIt, &*rhsIt, OperationEquivalence::ignoreValueEquivalence,
            /*markEquivalent=*/nullptr,
            OperationEquivalence::Flags::IgnoreLocations))
      return failure();
 
    // Compare the operands of the two operations. If the operand is within
    // the block, it must refer to the same operation.
    auto lhsOperands = lhsIt->getOperands(), rhsOperands = rhsIt->getOperands();
    for (int operand : llvm::seq<int>(0, lhsIt->getNumOperands())) {
      Value lhsOperand = lhsOperands[operand];
      Value rhsOperand = rhsOperands[operand];
      if (lhsOperand == rhsOperand)
        continue;
      // Check that the types of the operands match.
      if (lhsOperand.getType() != rhsOperand.getType())
        return failure();
 
      // Check that these uses are both external, or both internal.
      bool lhsIsInBlock = lhsOperand.getParentBlock() == leaderBlock;
      bool rhsIsInBlock = rhsOperand.getParentBlock() == mergeBlock;
      if (lhsIsInBlock != rhsIsInBlock)
        return failure();
      // Let the operands differ if they are defined in a different block. These
      // will become new arguments if the blocks get merged.
      if (!lhsIsInBlock) {
 
        // Check whether the operands aren't the result of an immediate
        // predecessors terminator. In that case we are not able to use it as a
        // successor operand when branching to the merged block as it does not
        // dominate its producing operation.
        auto isValidSuccessorArg = [](Block *block, Value operand) {
          if (operand.getDefiningOp() !=
              operand.getParentBlock()->getTerminator())
            return true;
          return !llvm::is_contained(block->getPredecessors(),
                                     operand.getParentBlock());
        };
 
        if (!isValidSuccessorArg(leaderBlock, lhsOperand) ||
            !isValidSuccessorArg(mergeBlock, rhsOperand))
          return failure();
 
        mismatchedOperands.emplace_back(opI, operand);
        continue;
      }
 
      // Otherwise, these operands must have the same logical order within the
      // parent block.
      if (leaderData.getOrderOf(lhsOperand) != blockData.getOrderOf(rhsOperand))
        return failure();
    }
 
    // If the lhs or rhs has external uses, the blocks cannot be merged as the
    // merged version of this operation will not be either the lhs or rhs
    // alone (thus semantically incorrect), but some mix dependending on which
    // block preceeded this.
    // TODO allow merging of operations when one block does not dominate the
    // other
    if (rhsIt->isUsedOutsideOfBlock(mergeBlock) ||
        lhsIt->isUsedOutsideOfBlock(leaderBlock)) {
      return failure();
    }
  }
  // Make sure that the block sizes are equivalent.
  if (lhsIt != lhsE || rhsIt != rhsE)
    return failure();
 
  // If we get here, the blocks are equivalent and can be merged.
  operandsToMerge.insert(mismatchedOperands.begin(), mismatchedOperands.end());
  blocksToMerge.insert(blockData.block);
  return success();
}
 
/// Returns true if the predecessor terminators of the given block can have
/// their operands updated by appending values of the given types: each must
/// implement BranchOpInterface and be willing to forward every one of the
/// types to the block.
static bool ableToUpdatePredOperands(Block *block, ArrayRef<Type> types) {
  for (auto it = block->pred_begin(), e = block->pred_end(); it != e; ++it) {
    auto branch = dyn_cast<BranchOpInterface>((*it)->getTerminator());
    if (!branch)
      return false;
    unsigned succIndex = it.getSuccessorIndex();
    for (Type type : types)
      if (!branch.mayForwardTypeToSuccessor(succIndex, type))
        return false;
  }
  return true;
}
 
/// Prunes the redundant list of new arguments. E.g., if we are passing an
/// argument list like [x, y, z, x] this would return [x, y, z] and it would
/// update the `block` (to whom the argument are passed to) accordingly. The new
/// arguments are passed as arguments at the back of the block, hence we need to
/// know how many `numOldArguments` were before, in order to correctly replace
/// the new arguments in the block
static SmallVector<SmallVector<Value, 8>, 2> pruneRedundantArguments(
    const SmallVector<SmallVector<Value, 8>, 2> &newArguments,
    RewriterBase &rewriter, unsigned numOldArguments, Block *block) {
 
  SmallVector<SmallVector<Value, 8>, 2> newArgumentsPruned(
      newArguments.size(), SmallVector<Value, 8>());
 
  if (newArguments.empty())
    return newArguments;
 
  // `newArguments` is a 2D array of size `numLists` x `numArgs`
  unsigned numLists = newArguments.size();
  unsigned numArgs = newArguments[0].size();
 
  // Map that for each arg index contains the index that we can use in place of
  // the original index. E.g., if we have newArgs = [x, y, z, x], we will have
  // idxToReplacement[3] = 0
  llvm::DenseMap<unsigned, unsigned> idxToReplacement;
 
  // This is a useful data structure to track the first appearance of a Value
  // on a given list of arguments
  DenseMap<Value, unsigned> firstValueToIdx;
  for (unsigned j = 0; j < numArgs; ++j) {
    Value newArg = newArguments[0][j];
    firstValueToIdx.try_emplace(newArg, j);
  }
 
  // Go through the first list of arguments (list 0).
  for (unsigned j = 0; j < numArgs; ++j) {
    // Look back to see if there are possible redundancies in list 0. Please
    // note that we are using a map to annotate when an argument was seen first
    // to avoid a O(N^2) algorithm. This has the drawback that if we have two
    // lists like:
    // list0: [%a, %a, %a]
    // list1: [%c, %b, %b]
    // We cannot simplify it, because firstValueToIdx[%a] = 0, but we cannot
    // point list1[1](==%b) or list1[2](==%b) to list1[0](==%c).  However, since
    // the number of arguments can be potentially unbounded we cannot afford a
    // O(N^2) algorithm (to search to all the possible pairs) and we need to
    // accept the trade-off.
    unsigned k = firstValueToIdx[newArguments[0][j]];
    if (k == j)
      continue;
 
    bool shouldReplaceJ = true;
    unsigned replacement = k;
    // If a possible redundancy is found, then scan the other lists: we
    // can prune the arguments if and only if they are redundant in every
    // list.
    for (unsigned i = 1; i < numLists; ++i)
      shouldReplaceJ =
          shouldReplaceJ && (newArguments[i][k] == newArguments[i][j]);
    // Save the replacement.
    if (shouldReplaceJ)
      idxToReplacement[j] = replacement;
  }
 
  // Populate the pruned argument list.
  for (unsigned i = 0; i < numLists; ++i)
    for (unsigned j = 0; j < numArgs; ++j)
      if (!idxToReplacement.contains(j))
        newArgumentsPruned[i].push_back(newArguments[i][j]);
 
  // Replace the block's redundant arguments.
  SmallVector<unsigned> toErase;
  for (auto [idx, arg] : llvm::enumerate(block->getArguments())) {
    if (idxToReplacement.contains(idx)) {
      Value oldArg = block->getArgument(numOldArguments + idx);
      Value newArg =
          block->getArgument(numOldArguments + idxToReplacement[idx]);
      rewriter.replaceAllUsesWith(oldArg, newArg);
      toErase.push_back(numOldArguments + idx);
    }
  }
 
  // Erase the block's redundant arguments.
  for (unsigned idxToErase : llvm::reverse(toErase))
    block->eraseArgument(idxToErase);
  return newArgumentsPruned;
}
 
LogicalResult BlockMergeCluster::merge(RewriterBase &rewriter) {
  // Don't consider clusters that don't have blocks to merge.
  if (blocksToMerge.empty())
    return failure();
 
  Block *leaderBlock = leaderData.block;
  if (!operandsToMerge.empty()) {
    // If the cluster has operands to merge, verify that the predecessor
    // terminators of each of the blocks can have their successor operands
    // updated: merging threads the mismatched values through them as new
    // successor operands, so each terminator must be able to forward values
    // of those types. The types are read off the leader block; addToCluster
    // already required every block's mismatched operand types to match.
    // TODO: We could try and sub-partition this cluster if only some blocks
    // cause the mismatch.
    SmallVector<Type> operandTypes;
    operandTypes.reserve(operandsToMerge.size());
    {
      unsigned curOpIndex = 0;
      Block::iterator opIt = leaderBlock->begin();
      for (const auto &it : operandsToMerge) {
        std::advance(opIt, it.first - curOpIndex);
        curOpIndex = it.first;
        operandTypes.push_back(opIt->getOperand(it.second).getType());
      }
    }
    if (!ableToUpdatePredOperands(leaderBlock, operandTypes) ||
        !llvm::all_of(blocksToMerge, [&](Block *block) {
          return ableToUpdatePredOperands(block, operandTypes);
        }))
      return failure();
 
    // Collect the iterators for each of the blocks to merge. We will walk all
    // of the iterators at once to avoid operand index invalidation.
    SmallVector<Block::iterator, 2> blockIterators;
    blockIterators.reserve(blocksToMerge.size() + 1);
    blockIterators.push_back(leaderBlock->begin());
    for (Block *mergeBlock : blocksToMerge)
      blockIterators.push_back(mergeBlock->begin());
 
    // Update each of the predecessor terminators with the new arguments.
    SmallVector<SmallVector<Value, 8>, 2> newArguments(
        1 + blocksToMerge.size(),
        SmallVector<Value, 8>(operandsToMerge.size()));
    unsigned curOpIndex = 0;
    unsigned numOldArguments = leaderBlock->getNumArguments();
    for (const auto &it : llvm::enumerate(operandsToMerge)) {
      unsigned nextOpOffset = it.value().first - curOpIndex;
      curOpIndex = it.value().first;
 
      // Process the operand for each of the block iterators.
      for (unsigned i = 0, e = blockIterators.size(); i != e; ++i) {
        Block::iterator &blockIter = blockIterators[i];
        std::advance(blockIter, nextOpOffset);
        auto &operand = blockIter->getOpOperand(it.value().second);
        newArguments[i][it.index()] = operand.get();
 
        // Update the operand and insert an argument if this is the leader.
        if (i == 0) {
          Value operandVal = operand.get();
          operand.set(leaderBlock->addArgument(operandVal.getType(),
                                               operandVal.getLoc()));
        }
      }
    }
 
    // Prune redundant arguments and update the leader block argument list
    newArguments = pruneRedundantArguments(newArguments, rewriter,
                                           numOldArguments, leaderBlock);
 
    // Update the predecessors for each of the blocks.
    auto updatePredecessors = [&](Block *block, unsigned clusterIndex) {
      for (auto predIt = block->pred_begin(), predE = block->pred_end();
           predIt != predE; ++predIt) {
        auto branch = cast<BranchOpInterface>((*predIt)->getTerminator());
        unsigned succIndex = predIt.getSuccessorIndex();
        branch.getSuccessorOperands(succIndex).append(
            newArguments[clusterIndex]);
      }
    };
    updatePredecessors(leaderBlock, /*clusterIndex=*/0);
    for (unsigned i = 0, e = blocksToMerge.size(); i != e; ++i)
      updatePredecessors(blocksToMerge[i], /*clusterIndex=*/i + 1);
  }
 
  // Replace all uses of the merged blocks with the leader and erase them.
  for (Block *block : blocksToMerge) {
    block->replaceAllUsesWith(leaderBlock);
    rewriter.eraseBlock(block);
  }
  return success();
}
 
/// Identify identical blocks within the given region and merge them, inserting
/// new block arguments as necessary. Returns success if any blocks were merged,
/// failure otherwise.
static LogicalResult mergeIdenticalBlocks(RewriterBase &rewriter,
                                          Region &region) {
  if (region.empty() || region.hasOneBlock())
    return failure();
 
  // Identify sets of blocks, other than the entry block, that branch to the
  // same successors. We will use these groups to create clusters of equivalent
  // blocks.
  DenseMap<SuccessorRange, SmallVector<Block *, 1>> matchingSuccessors;
  for (Block &block : llvm::drop_begin(region, 1))
    matchingSuccessors[block.getSuccessors()].push_back(&block);
 
  bool mergedAnyBlocks = false;
  for (ArrayRef<Block *> blocks : llvm::make_second_range(matchingSuccessors)) {
    if (blocks.size() == 1)
      continue;
 
    SmallVector<BlockMergeCluster, 1> clusters;
    for (Block *block : blocks) {
      BlockEquivalenceData data(block);
 
      // Don't allow merging if this block has any regions.
      // TODO: Add support for regions if necessary.
      bool hasNonEmptyRegion = llvm::any_of(*block, [](Operation &op) {
        return llvm::any_of(op.getRegions(),
                            [](Region &region) { return !region.empty(); });
      });
      if (hasNonEmptyRegion)
        continue;
 
      // Don't allow merging if this block's arguments are used outside of the
      // original block.
      bool argHasExternalUsers = llvm::any_of(
          block->getArguments(), [block](mlir::BlockArgument &arg) {
            return arg.isUsedOutsideOfBlock(block);
          });
      if (argHasExternalUsers)
        continue;
 
      // Try to add this block to an existing cluster.
      bool addedToCluster = false;
      for (auto &cluster : clusters)
        if ((addedToCluster = succeeded(cluster.addToCluster(data))))
          break;
      if (!addedToCluster)
        clusters.emplace_back(std::move(data));
    }
    for (auto &cluster : clusters)
      mergedAnyBlocks |= succeeded(cluster.merge(rewriter));
  }
 
  return success(mergedAnyBlocks);
}
 
/// Identify identical blocks within the given regions and merge them, inserting
/// new block arguments as necessary.
static LogicalResult mergeIdenticalBlocks(RewriterBase &rewriter,
                                          MutableArrayRef<Region> regions) {
  llvm::SmallSetVector<Region *, 1> worklist;
  for (auto &region : regions)
    worklist.insert(&region);
  bool anyChanged = false;
  while (!worklist.empty()) {
    Region *region = worklist.pop_back_val();
    if (succeeded(mergeIdenticalBlocks(rewriter, *region))) {
      worklist.insert(region);
      anyChanged = true;
    }
 
    // Add any nested regions to the worklist.
    for (Block &block : *region)
      for (auto &op : block)
        for (auto &nestedRegion : op.getRegions())
          worklist.insert(&nestedRegion);
  }
 
  return success(anyChanged);
}
 
/// If a block's argument is always the same across different invocations, then
/// drop the argument and use the value directly inside the block
static LogicalResult dropRedundantArguments(RewriterBase &rewriter,
                                            Block &block) {
  SmallVector<size_t> argsToErase;
 
  // Go through the arguments of the block.
  for (auto [argIdx, blockOperand] : llvm::enumerate(block.getArguments())) {
    bool sameArg = true;
    Value commonValue;
 
    // Go through the block predecessor and flag if they pass to the block
    // different values for the same argument.
    for (Block::pred_iterator predIt = block.pred_begin(),
                              predE = block.pred_end();
         predIt != predE; ++predIt) {
      auto branch = dyn_cast<BranchOpInterface>((*predIt)->getTerminator());
      if (!branch) {
        sameArg = false;
        break;
      }
      unsigned succIndex = predIt.getSuccessorIndex();
      SuccessorOperands succOperands = branch.getSuccessorOperands(succIndex);
 
      // Produced operands are generated by the terminator operation itself
      // (e.g., results of an async call) and cannot be forwarded or dropped.
      if (succOperands.isOperandProduced(argIdx)) {
        sameArg = false;
        break;
      }
 
      // Get the forwarded operand value using operator[] which correctly
      // adjusts for the produced operand offset.
      Value operandValue = succOperands[argIdx];
      if (!commonValue) {
        commonValue = operandValue;
        continue;
      }
      if (operandValue != commonValue) {
        sameArg = false;
        break;
      }
    }
 
    // If they are passing the same value, drop the argument.
    if (commonValue && sameArg) {
      argsToErase.push_back(argIdx);
 
      // Remove the argument from the block.
      rewriter.replaceAllUsesWith(blockOperand, commonValue);
    }
  }
 
  // Remove the arguments.
  for (size_t argIdx : llvm::reverse(argsToErase)) {
    block.eraseArgument(argIdx);
 
    // Remove the argument from the branch ops.
    for (auto predIt = block.pred_begin(), predE = block.pred_end();
         predIt != predE; ++predIt) {
      auto branch = cast<BranchOpInterface>((*predIt)->getTerminator());
      unsigned succIndex = predIt.getSuccessorIndex();
      SuccessorOperands succOperands = branch.getSuccessorOperands(succIndex);
      succOperands.erase(argIdx);
    }
  }
  return success(!argsToErase.empty());
}
 
/// This optimization drops redundant argument to blocks. I.e., if a given
/// argument to a block receives the same value from each of the block
/// predecessors, we can remove the argument from the block and use directly the
/// original value. This is a simple example:
///
/// %cond = llvm.call @rand() : () -> i1
/// %val0 = llvm.mlir.constant(1 : i64) : i64
/// %val1 = llvm.mlir.constant(2 : i64) : i64
/// %val2 = llvm.mlir.constant(3 : i64) : i64
/// llvm.cond_br %cond, ^bb1(%val0 : i64, %val1 : i64), ^bb2(%val0 : i64, %val2
/// : i64)
///
/// ^bb1(%arg0 : i64, %arg1 : i64):
///    llvm.call @foo(%arg0, %arg1)
///
/// The previous IR can be rewritten as:
/// %cond = llvm.call @rand() : () -> i1
/// %val0 = llvm.mlir.constant(1 : i64) : i64
/// %val1 = llvm.mlir.constant(2 : i64) : i64
/// %val2 = llvm.mlir.constant(3 : i64) : i64
/// llvm.cond_br %cond, ^bb1(%val1 : i64), ^bb2(%val2 : i64)
///
/// ^bb1(%arg0 : i64):
///    llvm.call @foo(%val0, %arg0)
///
static LogicalResult dropRedundantArguments(RewriterBase &rewriter,
                                            MutableArrayRef<Region> regions) {
  llvm::SmallSetVector<Region *, 1> worklist;
  for (Region &region : regions)
    worklist.insert(&region);
  bool anyChanged = false;
  while (!worklist.empty()) {
    Region *region = worklist.pop_back_val();
 
    // Add any nested regions to the worklist.
    for (Block &block : *region) {
      anyChanged =
          succeeded(dropRedundantArguments(rewriter, block)) || anyChanged;
 
      for (Operation &op : block)
        for (Region &nestedRegion : op.getRegions())
          worklist.insert(&nestedRegion);
    }
  }
  return success(anyChanged);
}
 
//===----------------------------------------------------------------------===//
// Region Simplification
//===----------------------------------------------------------------------===//
 
/// Run a set of structural simplifications over the given regions. This
/// includes transformations like unreachable block elimination, dead argument
/// elimination, as well as some other DCE. This function returns success if any
/// of the regions were simplified, failure otherwise.
LogicalResult mlir::simplifyRegions(RewriterBase &rewriter,
                                    MutableArrayRef<Region> regions,
                                    bool mergeBlocks) {
  bool eliminatedBlocks = succeeded(eraseUnreachableBlocks(rewriter, regions));
  bool eliminatedOpsOrArgs = succeeded(runRegionDCE(rewriter, regions));
  bool mergedIdenticalBlocks = false;
  bool droppedRedundantArguments = false;
  if (mergeBlocks) {
    mergedIdenticalBlocks = succeeded(mergeIdenticalBlocks(rewriter, regions));
    droppedRedundantArguments =
        succeeded(dropRedundantArguments(rewriter, regions));
  }
  return success(eliminatedBlocks || eliminatedOpsOrArgs ||
                 mergedIdenticalBlocks || droppedRedundantArguments);
}
 
//===---------------------------------------------------------------------===//
// Move operation dependencies
//===---------------------------------------------------------------------===//
 
/// Check if moving operations in the slice before `insertionPoint` would break
/// dominance due to block argument operands. Returns true if all block args
/// dominate the insertion point (no issue), false otherwise. If `failingOp` is
/// provided, it will be set to the first problematic op.
///
/// For operands defined by ops: either the defining op is in the slice (so
/// dominance preserved), or it already dominates insertionPoint (otherwise it
/// would be in the slice). So we only need to check block argument operands,
/// both as direct operands and as values captured inside regions.
static bool blockArgsDominateInsertionPoint(
    const llvm::SetVector<Operation *> &slice, Operation *insertionPoint,
    DominanceInfo &dominance, Operation **failingOp = nullptr) {
  Block *insertionBlock = insertionPoint->getBlock();
 
  // Returns true if the block arg dominates, false otherwise. Sets failingOp
  // on failure.
  auto argDominates = [&](BlockArgument arg, Operation *op) {
    Block *argBlock = arg.getOwner();
    bool dominates = argBlock == insertionBlock ||
                     dominance.dominates(argBlock, insertionBlock);
    if (!dominates && failingOp)
      *failingOp = op;
    return dominates;
  };
 
  for (Operation *op : slice) {
    // Check direct operands.
    for (Value operand : op->getOperands()) {
      auto arg = dyn_cast<BlockArgument>(operand);
      if (!arg)
        continue;
      if (!argDominates(arg, op))
        return false;
    }
 
    // Check block arguments captured inside regions. Process one region at a
    // time to enable early exit without collecting values from all regions.
    for (Region &region : op->getRegions()) {
      SetVector<Value> capturedValues;
      getUsedValuesDefinedAbove(region, region, capturedValues);
      for (Value val : capturedValues) {
        auto arg = dyn_cast<BlockArgument>(val);
        if (!arg)
          continue;
        if (!argDominates(arg, op))
          return false;
      }
    }
  }
  return true;
}
 
/// Check if any region between an operation and an ancestor block is
/// isolated from above. If so, moving the operation out would break
/// the isolation semantics.
static bool hasIsolatedRegionBetween(Operation *op, Block *ancestorBlock) {
  Region *ancestorRegion = ancestorBlock->getParent();
 
  // Walk up from the op's region to find if there's an isolated region
  // between the op and the ancestor.
  Region *region = op->getParentRegion();
  while (region && region != ancestorRegion) {
    Operation *parentOp = region->getParentOp();
    if (!parentOp)
      break;
 
    if (parentOp->hasTrait<OpTrait::IsIsolatedFromAbove>())
      return true;
 
    region = parentOp->getParentRegion();
  }
  return false;
}
 
LogicalResult mlir::moveOperationDependencies(RewriterBase &rewriter,
                                              Operation *op,
                                              Operation *insertionPoint,
                                              DominanceInfo &dominance) {
  Block *insertionBlock = insertionPoint->getBlock();
 
  // If `insertionPoint` does not dominate `op`, do nothing.
  if (!dominance.properlyDominates(insertionPoint, op)) {
    return rewriter.notifyMatchFailure(op,
                                       "insertion point does not dominate op");
  }
 
  // Verify we're not crossing an isolated region.
  if (hasIsolatedRegionBetween(op, insertionBlock)) {
    return rewriter.notifyMatchFailure(
        op, "cannot move operation across isolated-from-above region");
  }
 
  // Find the backward slice of operation for each `Value` the operation
  // depends on. Prune the slice to only include operations not already
  // dominated by the `insertionPoint`.
  BackwardSliceOptions options;
  options.inclusive = false;
  options.omitUsesFromAbove = false;
  // Block arguments cannot be moved; dominance check handles this case.
  options.omitBlockArguments = true;
  bool dependsOnSideEffectingOp = false;
  options.filter = [&](Operation *sliceBoundaryOp) {
    // Skip the root op - we're moving its dependencies, not the op itself.
    // The root op is filtered out by options.inclusive = false anyway.
    if (sliceBoundaryOp == op)
      return true;
    bool dominated =
        dominance.properlyDominates(sliceBoundaryOp, insertionPoint);
    // Op is already before insertion point, no need to include in slice.
    if (dominated)
      return false;
    // Op needs to move but is side-effecting - stop traversal early.
    if (!isPure(sliceBoundaryOp)) {
      dependsOnSideEffectingOp = true;
      return false;
    }
    return true;
  };
  llvm::SetVector<Operation *> slice;
  LogicalResult result = getBackwardSlice(op, &slice, options);
  assert(result.succeeded() && "expected a backward slice");
  (void)result;
 
  // Check if any operation in the slice is side-effecting.
  if (dependsOnSideEffectingOp) {
    return rewriter.notifyMatchFailure(
        op, "cannot move operation with side-effecting dependencies");
  }
 
  // If the slice contains `insertionPoint` cannot move the dependencies.
  if (slice.contains(insertionPoint)) {
    return rewriter.notifyMatchFailure(
        op,
        "cannot move dependencies before operation in backward slice of op");
  }
 
  // Verify no operation in the slice uses a block argument that wouldn't
  // dominate at the new location.
  Operation *badOp = nullptr;
  if (!blockArgsDominateInsertionPoint(slice, insertionPoint, dominance,
                                       &badOp)) {
    return rewriter.notifyMatchFailure(
        badOp, "moving op would break dominance for block argument operand");
  }
 
  // We should move the slice in topological order, but `getBackwardSlice`
  // already does that. So no need to sort again.
  for (Operation *op : slice) {
    rewriter.moveOpBefore(op, insertionPoint);
  }
  return success();
}
 
LogicalResult mlir::moveOperationDependencies(RewriterBase &rewriter,
                                              Operation *op,
                                              Operation *insertionPoint) {
  DominanceInfo dominance(op);
  return moveOperationDependencies(rewriter, op, insertionPoint, dominance);
}
 
LogicalResult mlir::moveValueDefinitions(RewriterBase &rewriter,
                                         ValueRange values,
                                         Operation *insertionPoint,
                                         DominanceInfo &dominance) {
  // Remove the values that already dominate the insertion point.
  SmallVector<Value> prunedValues;
  for (auto value : values) {
    if (dominance.properlyDominates(value, insertionPoint))
      continue;
    // Block arguments are not supported.
    if (isa<BlockArgument>(value)) {
      return rewriter.notifyMatchFailure(
          insertionPoint,
          "unsupported case of moving block argument before insertion point");
    }
 
    Block *insertionBlock = insertionPoint->getBlock();
    Operation *definingOp = value.getDefiningOp();
    Block *definingBlock = definingOp->getBlock();
 
    // Verify we're not crossing an isolated region.
    if (hasIsolatedRegionBetween(definingOp, insertionBlock)) {
      return rewriter.notifyMatchFailure(
          insertionPoint,
          "cannot move value definition across isolated-from-above region");
    }
 
    // Verify the insertion point's block dominates the defining block,
    // otherwise we're trying to move "backwards" in the CFG which doesn't
    // make sense.
    if (!dominance.dominates(insertionBlock, definingBlock)) {
      return rewriter.notifyMatchFailure(
          insertionPoint,
          "insertion point block does not dominate the value's defining "
          "block");
    }
    prunedValues.push_back(value);
  }
 
  // Find the backward slice of operation for each `Value` the operation
  // depends on. Prune the slice to only include operations not already
  // dominated by the `insertionPoint`
  BackwardSliceOptions options;
  options.inclusive = true;
  options.omitUsesFromAbove = false;
  // Block arguments cannot be moved, so we stop the slice computation there.
  // If an op uses a block argument that wouldn't dominate at the new location,
  // the dominance check will catch it.
  options.omitBlockArguments = true;
  bool dependsOnSideEffectingOp = false;
  options.filter = [&](Operation *sliceBoundaryOp) {
    bool dominated =
        dominance.properlyDominates(sliceBoundaryOp, insertionPoint);
    // Op is already before insertion point, no need to include in slice.
    if (dominated)
      return false;
    // Op needs to move but is side-effecting - stop traversal early.
    if (!isPure(sliceBoundaryOp)) {
      dependsOnSideEffectingOp = true;
      return false;
    }
    return true;
  };
  llvm::SetVector<Operation *> slice;
  for (auto value : prunedValues) {
    LogicalResult result = getBackwardSlice(value, &slice, options);
    assert(result.succeeded() && "expected a backward slice");
    (void)result;
  }
 
  // Check if any operation in the slice is side-effecting.
  if (dependsOnSideEffectingOp) {
    return rewriter.notifyMatchFailure(
        insertionPoint, "cannot move value definitions with side-effecting "
                        "operations in the slice");
  }
 
  // If the slice contains `insertionPoint` cannot move the dependencies.
  if (slice.contains(insertionPoint)) {
    return rewriter.notifyMatchFailure(
        insertionPoint,
        "cannot move dependencies before operation in backward slice of op");
  }
 
  // Sort operations topologically. This is needed because we call
  // getBackwardSlice multiple times (once per value), and the combined slice
  // may not be in topological order when independent subgraphs interleave.
  mlir::topologicalSort(slice);
 
  // Verify no operation in the slice uses a block argument that wouldn't
  // dominate at the new location.
  Operation *badOp = nullptr;
  if (!blockArgsDominateInsertionPoint(slice, insertionPoint, dominance,
                                       &badOp)) {
    return rewriter.notifyMatchFailure(
        badOp, "moving op would break dominance for block argument operand");
  }
 
  for (Operation *op : slice)
    rewriter.moveOpBefore(op, insertionPoint);
  return success();
}
 
LogicalResult mlir::moveValueDefinitions(RewriterBase &rewriter,
                                         ValueRange values,
                                         Operation *insertionPoint) {
  DominanceInfo dominance(insertionPoint);
  return moveValueDefinitions(rewriter, values, insertionPoint, dominance);
}