Bufferization/IR/BufferizableOpInterface.cpp

//===- BufferizableOpInterface.cpp - Bufferizable Ops  ---=----------------===//
//
// 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/Dialect/Bufferization/IR/BufferizableOpInterface.h"
#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
#include "mlir/Dialect/MemRef/IR/MemRef.h"
#include "mlir/IR/AsmState.h"
#include "mlir/IR/BlockAndValueMapping.h"
#include "mlir/IR/BuiltinOps.h"
#include "mlir/IR/Operation.h"
#include "mlir/IR/TypeUtilities.h"
#include "mlir/IR/Value.h"
#include "llvm/Support/Debug.h"

namespace mlir {
namespace bufferization {

#include "mlir/Dialect/Bufferization/IR/BufferizableOpInterface.cpp.inc"

} // namespace bufferization
} // namespace mlir

#define DEBUG_TYPE "bufferizable-op-interface"
#define DBGS() (llvm::dbgs() << '[' << DEBUG_TYPE << "] ")
#define LDBG(X) LLVM_DEBUG(DBGS() << (X))

using namespace mlir;
using namespace bufferization;

/// Attribute name used to mark the bufferization layout for region
/// arguments during linalg comprehensive bufferization.
constexpr const ::llvm::StringLiteral
    bufferization::BufferizableOpInterface::kBufferLayoutAttrName;

/// Attribute name used to mark region arguments that can be bufferized
/// in-place during linalg comprehensive bufferization.
constexpr const ::llvm::StringLiteral
    bufferization::BufferizableOpInterface::kInplaceableAttrName;

//===----------------------------------------------------------------------===//
// BufferizationOptions
//===----------------------------------------------------------------------===//

// Default constructor for BufferizationOptions.
BufferizationOptions::BufferizationOptions() {}

BufferizableOpInterface
BufferizationOptions::dynCastBufferizableOp(Operation *op) const {
  if (isOpAllowed(op))
    return dyn_cast<BufferizableOpInterface>(op);
  return nullptr;
}

BufferizableOpInterface
BufferizationOptions::dynCastBufferizableOp(Value value) const {
  if (auto bufferizableOp = value.getDefiningOp<BufferizableOpInterface>())
    if (isOpAllowed(bufferizableOp.getOperation()))
      return bufferizableOp;
  return nullptr;
}

//===----------------------------------------------------------------------===//
// Helper functions for BufferizableOpInterface
//===----------------------------------------------------------------------===//

static void setInsertionPointAfter(OpBuilder &b, Value value) {
  if (auto bbArg = value.dyn_cast<BlockArgument>()) {
    b.setInsertionPointToStart(bbArg.getOwner());
  } else {
    b.setInsertionPointAfter(value.getDefiningOp());
  }
}

/// Determine which OpOperand* will alias with `result` if the op is bufferized
/// in place. Return an empty vector if the op is not bufferizable.
SmallVector<OpOperand *>
BufferizationState::getAliasingOpOperand(OpResult result) const {
  if (Operation *op = result.getDefiningOp())
    if (auto bufferizableOp = dyn_cast<BufferizableOpInterface>(op))
      return bufferizableOp.getAliasingOpOperand(result, *this);
  return {};
}

/// Determine which OpResult will alias with `opOperand` if the op is bufferized
/// in place. Return an empty OpResult if the op is not bufferizable.
OpResult BufferizationState::getAliasingOpResult(OpOperand &opOperand) const {
  if (auto bufferizableOp =
          dyn_cast<BufferizableOpInterface>(opOperand.getOwner()))
    return bufferizableOp.getAliasingOpResult(opOperand, *this);
  return OpResult();
}

/// Return true if `opOperand` bufferizes to a memory read. Return `true` if the
/// op is not bufferizable.
bool BufferizationState::bufferizesToMemoryRead(OpOperand &opOperand) const {
  if (auto bufferizableOp =
          dyn_cast<BufferizableOpInterface>(opOperand.getOwner()))
    return bufferizableOp.bufferizesToMemoryRead(opOperand, *this);

  // Unknown op that returns a tensor. The inplace analysis does not support it.
  // Conservatively return true.
  return true;
}

/// Return true if `opOperand` bufferizes to a memory write. Return
/// `true` if the op is not bufferizable.
bool BufferizationState::bufferizesToMemoryWrite(OpOperand &opOperand) const {
  if (auto bufferizableOp =
          dyn_cast<BufferizableOpInterface>(opOperand.getOwner()))
    return bufferizableOp.bufferizesToMemoryWrite(opOperand, *this);

  // Unknown op that returns a tensor. The inplace analysis does not support it.
  // Conservatively return true.
  return true;
}

/// Return true if `opOperand` does neither read nor write but bufferizes to an
/// alias. Return false if the op is not bufferizable.
bool BufferizationState::bufferizesToAliasOnly(OpOperand &opOperand) const {
  if (auto bufferizableOp =
          dyn_cast<BufferizableOpInterface>(opOperand.getOwner()))
    return bufferizableOp.bufferizesToAliasOnly(opOperand, *this);

  // Unknown op that returns a tensor. The inplace analysis does not support it.
  // Conservatively return false.
  return false;
}

/// Return true if the given value is read by an op that bufferizes to a memory
/// read. Also takes into account ops that create an alias but do not read by
/// themselves (e.g., ExtractSliceOp).
bool BufferizationState::isValueRead(Value value) const {
  assert(value.getType().isa<TensorType>() && "expected TensorType");
  SmallVector<OpOperand *> workingSet;
  for (OpOperand &use : value.getUses())
    workingSet.push_back(&use);

  while (!workingSet.empty()) {
    OpOperand *uMaybeReading = workingSet.pop_back_val();
    // Skip over all ops that neither read nor write (but create an alias).
    if (bufferizesToAliasOnly(*uMaybeReading))
      for (OpOperand &use : getAliasingOpResult(*uMaybeReading).getUses())
        workingSet.push_back(&use);
    if (bufferizesToMemoryRead(*uMaybeReading))
      return true;
  }

  return false;
}

// Starting from `value`, follow the use-def chain in reverse, always selecting
// the aliasing OpOperands. Find and return Values for which `condition`
// evaluates to true. OpOperands of such matching Values are not traversed any
// further.
llvm::SetVector<Value> BufferizationState::findValueInReverseUseDefChain(
    Value value, llvm::function_ref<bool(Value)> condition) const {
  llvm::SetVector<Value> result, workingSet;
  workingSet.insert(value);

  while (!workingSet.empty()) {
    Value value = workingSet.pop_back_val();
    if (condition(value) || value.isa<BlockArgument>()) {
      result.insert(value);
      continue;
    }

    OpResult opResult = value.cast<OpResult>();
    SmallVector<OpOperand *> opOperands = getAliasingOpOperand(opResult);
    if (opOperands.empty() || !options.isOpAllowed(value.getDefiningOp())) {
      result.insert(value);
      continue;
    }

    for (OpOperand *o : opOperands)
      workingSet.insert(o->get());
  }

  return result;
}

// Find the Values of the last preceding write of a given Value.
llvm::SetVector<Value>
BufferizationState::findLastPrecedingWrite(Value value) const {
  return findValueInReverseUseDefChain(value, [&](Value value) {
    Operation *op = value.getDefiningOp();
    if (!op)
      return true;
    auto bufferizableOp = options.dynCastBufferizableOp(op);
    if (!bufferizableOp)
      return true;
    return bufferizableOp.isMemoryWrite(value.cast<OpResult>(), *this);
  });
}

BufferizationState::BufferizationState(const BufferizationOptions &options)
    : options(options) {}

// bufferization.to_memref is not allowed to change the rank.
static void ensureToMemrefOpIsValid(Value tensor, Type memrefType) {
#ifndef NDEBUG
  auto rankedTensorType = tensor.getType().dyn_cast<RankedTensorType>();
  assert((!rankedTensorType || memrefType.cast<MemRefType>().getRank() ==
                                   rankedTensorType.getRank()) &&
         "to_memref would be invalid: mismatching ranks");
#endif
}

static Value lookupBuffer(RewriterBase &rewriter, Value tensor,
                          const BufferizationOptions &options) {
  auto tensorType = tensor.getType().dyn_cast<TensorType>();
  assert(tensorType && "unexpected non-tensor type");

  // Replace "%t = to_tensor %m" with %m.
  if (auto toTensorOp = tensor.getDefiningOp<bufferization::ToTensorOp>())
    return toTensorOp.memref();

  // Insert to_memref op.
  OpBuilder::InsertionGuard g(rewriter);
  setInsertionPointAfter(rewriter, tensor);
  Type memrefType = getMemRefType(tensorType, options);
  ensureToMemrefOpIsValid(tensor, memrefType);
  return rewriter.create<bufferization::ToMemrefOp>(tensor.getLoc(), memrefType,
                                                    tensor);
}

/// Return the result buffer (memref) for a given OpResult (tensor). Allocate
/// a new buffer and copy over data from the existing buffer if out-of-place
/// bufferization is necessary.
FailureOr<Value> BufferizationState::getBuffer(
    RewriterBase &rewriter, OpOperand &opOperand, bool forceInPlace,
    Optional<Operation *> customCopyInsertionPoint) const {
  OpBuilder::InsertionGuard guard(rewriter);
  Operation *op = opOperand.getOwner();
  Location loc = op->getLoc();
  Value operand = opOperand.get();
  Value operandBuffer = lookupBuffer(rewriter, operand, options);

  if (forceInPlace || isInPlace(opOperand))
    return operandBuffer;

  // Bufferizing out-of-place: Allocate a new buffer.
  // Move insertion point right after `operandBuffer`. That is where the
  // allocation should be inserted (in the absence of allocation hoisting).
  setInsertionPointAfter(rewriter, operandBuffer);
  // Allocate the result buffer.
  FailureOr<Value> resultBuffer = createAlloc(rewriter, loc, operandBuffer,
                                              options.createDeallocs, options);
  if (failed(resultBuffer))
    return failure();
  // Do not copy if the last preceding writes of `operand` are ops that do
  // not write (skipping ops that merely create aliases). E.g., InitTensorOp.
  // Note: If `findLastPrecedingWrite` reaches the end of the reverse SSA
  // use-def chain, it returns that value, regardless of whether it is a
  // memory write or not.
  SetVector<Value> lastWrites = findLastPrecedingWrite(operand);
  if (llvm::none_of(lastWrites, [&](Value lastWrite) {
        if (auto bufferizableOp = options.dynCastBufferizableOp(lastWrite))
          return bufferizableOp.isMemoryWrite(lastWrite.cast<OpResult>(),
                                              *this);
        return true;
      }))
    return resultBuffer;
  // Do not copy if the copied data is never read.
  OpResult aliasingOpResult = getAliasingOpResult(opOperand);
  if (aliasingOpResult && !bufferizesToMemoryRead(opOperand) &&
      !isValueRead(aliasingOpResult))
    return resultBuffer;
  // Do not copy if this op does not read the data, but writes it.
  if (bufferizesToMemoryWrite(opOperand) && !bufferizesToMemoryRead(opOperand))
    return resultBuffer;

  if (customCopyInsertionPoint) {
    rewriter.setInsertionPoint(*customCopyInsertionPoint);
  } else {
    // The copy happens right before the op that is bufferized.
    rewriter.setInsertionPoint(op);
  }
  if (failed(
          createMemCpy(rewriter, loc, operandBuffer, *resultBuffer, options)))
    return failure();

  return resultBuffer;
}

void bufferization::replaceOpWithBufferizedValues(RewriterBase &rewriter,
                                                  Operation *op,
                                                  ValueRange values) {
  OpBuilder::InsertionGuard g(rewriter);

  // Replace all OpResults with the given values.
  for (OpResult opResult : op->getOpResults()) {
    // Skip OpResult if it has no uses.
    if (opResult.getUses().empty())
      continue;

    Value replacement = values[opResult.getResultNumber()];
    if (opResult.getType().isa<TensorType>()) {
      // The OpResult is a tensor. Such values are replaced with memrefs during
      // bufferization.
      assert((replacement.getType().isa<MemRefType>() ||
              replacement.getType().isa<UnrankedMemRefType>()) &&
             "tensor op result should be replaced with a memref value");
      // The existing uses of the OpResult still expect a tensor. Insert a
      // ToTensorOp. Throughout bufferization, this ToTensorOp will gradually
      // loose all of its users and eventually DCE away.
      rewriter.setInsertionPointAfter(op);
      replacement = rewriter.create<bufferization::ToTensorOp>(
          replacement.getLoc(), replacement);
    }
    opResult.replaceAllUsesWith(replacement);
  }

  rewriter.eraseOp(op);
}

AlwaysCopyBufferizationState::AlwaysCopyBufferizationState(
    const BufferizationOptions &options)
    : BufferizationState(options) {}

/// Return `true` if the given OpResult has been decided to bufferize inplace.
bool AlwaysCopyBufferizationState::isInPlace(OpOperand &opOperand) const {
  // OpOperands that bufferize to a memory write are out-of-place, i.e., an
  // alloc and copy is inserted.
  return !bufferizesToMemoryWrite(opOperand);
}

/// Return true if `v1` and `v2` bufferize to equivalent buffers.
bool AlwaysCopyBufferizationState::areEquivalentBufferizedValues(
    Value v1, Value v2) const {
  // There is no analysis, so we do not know if the values are equivalent. The
  // conservative answer is "false".
  return false;
}

//===----------------------------------------------------------------------===//
// Bufferization-specific scoped alloc/dealloc insertion support.
//===----------------------------------------------------------------------===//

/// Move the insertion point of the given builder to the beginning of a
/// surrounding block as much as possible, while not crossing any allocation
/// hoisting barriers.
static void moveInsertionPointToAllocationHoistingBarrier(OpBuilder &b) {
  Operation *op = b.getInsertionBlock()->getParentOp();
  while (op) {
    if (auto bufferizableOp = dyn_cast<BufferizableOpInterface>(op))
      if (bufferizableOp.isAllocationHoistingBarrier())
        break;
    op = op->getParentOp();
  }

  if (!op) {
    // No allocation hoisting barrier found. Hoist to FuncOp.
    op = b.getInsertionBlock()->getParentOp();
    if (!isa<FuncOp>(op))
      op = op->getParentOfType<FuncOp>();
    assert(op && "could not find enclosing FuncOp");
  }

  // TODO: Handle cases where allocation hoisting barrier has more than one
  // region or block.
  assert(op->getNumRegions() == 1 &&
         "allocation hoisting barriers with >1 regions not supported");
  assert(op->getRegion(0).getBlocks().size() == 1 &&
         "allocation hoisting barriers with >1 blocks not supported");
  b.setInsertionPointToStart(&(op->getRegion(0).front()));
}

/// Compute the type of the `memref` to use for allocating the buffer for
/// `shapedValue`. Also returns (by reference in `dynShape`), the value for the
/// dynamic dimensions in the returned `memref` type. The function may also set
/// the insertion point to an earlier location, where the allocation should
/// happen ("allocation hoisting").
static MemRefType getAllocationTypeAndShape(OpBuilder &b, Location loc,
                                            Value shapedValue,
                                            SmallVectorImpl<Value> &dynShape) {
  MemRefType allocMemRefType =
      getContiguousMemRefType(shapedValue.getType().cast<ShapedType>());

  // Compute the dynamic part of the shape.
  bool reifiedShapes = false;
  if (auto rankedOp = dyn_cast_or_null<ReifyRankedShapedTypeOpInterface>(
          shapedValue.getDefiningOp())) {
    ReifiedRankedShapedTypeDims resultDims;
    if (succeeded(rankedOp.reifyResultShapes(b, resultDims))) {
      reifiedShapes = true;
      OpResult resultValue = shapedValue.dyn_cast<OpResult>();
      auto &shape = resultDims[resultValue.getResultNumber()];
      for (const auto &dim : enumerate(allocMemRefType.getShape()))
        if (ShapedType::isDynamic(dim.value()))
          dynShape.push_back(shape[dim.index()]);
    }
  }

  if (!reifiedShapes) {
    for (const auto &dim : enumerate(allocMemRefType.getShape()))
      if (ShapedType::isDynamic(dim.value())) {
        assert((shapedValue.getType().isa<UnrankedMemRefType>() ||
                shapedValue.getType().isa<MemRefType>()) &&
               "expected MemRef type");
        dynShape.push_back(
            b.create<memref::DimOp>(loc, shapedValue, dim.index()));
      }
  }

  // If the buffer is statically shaped, try to hoist it to the first enclosing
  // parallel region.
  // TODO: also hoist in the dynamic case. For now this relies on subsequent
  // calls to LICM and buffer hoisting which will most likely not succeed.
  // TODO: when packing, allocate a static bounding box which will enable more
  // hoisting.
  if (dynShape.empty())
    moveInsertionPointToAllocationHoistingBarrier(b);

  return allocMemRefType;
}

/// Create an AllocOp/DeallocOp pair, where the AllocOp is after
/// `shapedValue.getDefiningOp` (or at the top of the block in case of a
/// bbArg) and the DeallocOp is at the end of the block.
FailureOr<Value>
bufferization::createAlloc(OpBuilder &b, Location loc, Value shapedValue,
                           bool deallocMemref,
                           const BufferizationOptions &options) {
  // Take a guard before anything else.
  OpBuilder::InsertionGuard g(b);

  // 1. Create memory allocation.
  assert(shapedValue.getType().isa<ShapedType>());
  MemRefType memRefType = shapedValue.getType().dyn_cast<MemRefType>();
  SmallVector<Value> dynShape;
  // Note: getAllocationTypeAndShape also sets the insertion point.
  MemRefType allocMemRefType =
      getAllocationTypeAndShape(b, loc, shapedValue, dynShape);
  FailureOr<Value> allocated =
      createAlloc(b, loc, allocMemRefType, dynShape, options);
  if (failed(allocated))
    return failure();
  Value casted = allocated.getValue();
  if (memRefType && memRefType != allocMemRefType) {
    assert(memref::CastOp::areCastCompatible(allocated.getValue().getType(),
                                             memRefType) &&
           "createAlloc: cast incompatible");
    casted = b.create<memref::CastOp>(loc, memRefType, allocated.getValue());
  }

  if (deallocMemref) {
    // 2. Create memory deallocation.
    b.setInsertionPoint(allocated.getValue().getParentBlock()->getTerminator());
    if (failed(createDealloc(b, loc, allocated.getValue(), options)))
      return failure();
  }

  return casted;
}

/// Create a memref allocation with the given type and dynamic extents.
FailureOr<Value>
bufferization::createAlloc(OpBuilder &b, Location loc, MemRefType type,
                           ValueRange dynShape,
                           const BufferizationOptions &options) {
  if (options.allocationFn)
    return (*options.allocationFn)(b, loc, type, dynShape);

  // Default bufferallocation via AllocOp.
  Value allocated = b.create<memref::AllocOp>(
      loc, type, dynShape, b.getI64IntegerAttr(kBufferAlignments));
  return allocated;
}

/// Create a memref allocation with the given type and dynamic extents. May also
/// deallocate the memref again.
FailureOr<Value>
bufferization::createAlloc(OpBuilder &b, Location loc, MemRefType type,
                           ValueRange dynShape, bool deallocMemref,
                           const BufferizationOptions &options) {
  OpBuilder::InsertionGuard g(b);

  FailureOr<Value> alloc = createAlloc(b, loc, type, dynShape, options);
  if (failed(alloc))
    return failure();

  if (deallocMemref) {
    // Dealloc at the end of the block.
    b.setInsertionPoint(alloc.getValue().getParentBlock()->getTerminator());
    if (failed(createDealloc(b, loc, *alloc, options)))
      return failure();
  }

  return alloc;
}

/// Create a memref deallocation.
LogicalResult
bufferization::createDealloc(OpBuilder &b, Location loc, Value allocatedBuffer,
                             const BufferizationOptions &options) {
  if (options.deallocationFn)
    return (*options.deallocationFn)(b, loc, allocatedBuffer);

  // Default buffer deallocation via DeallocOp.
  b.create<memref::DeallocOp>(loc, allocatedBuffer);
  return success();
}

/// Create a memory copy between two memref buffers.
LogicalResult bufferization::createMemCpy(OpBuilder &b, Location loc,
                                          Value from, Value to,
                                          const BufferizationOptions &options) {
  if (options.memCpyFn)
    return (*options.memCpyFn)(b, loc, from, to);

  b.create<memref::CopyOp>(loc, from, to);
  return success();
}

//===----------------------------------------------------------------------===//
// Bufferization-specific BlockAndValueMapping support with debugging.
//===----------------------------------------------------------------------===//

bool bufferization::isFunctionArgument(Value value) {
  auto bbArg = value.dyn_cast<BlockArgument>();
  if (!bbArg)
    return false;
  return isa<FuncOp>(bbArg.getOwner()->getParentOp());
}

MemRefType bufferization::getContiguousMemRefType(ShapedType shapedType,
                                                  Attribute memorySpace) {
  MemRefLayoutAttrInterface layout = {};
  return MemRefType::get(shapedType.getShape(), shapedType.getElementType(),
                         layout, memorySpace);
}

BaseMemRefType bufferization::getMemRefType(TensorType tensorType,
                                            const BufferizationOptions &options,
                                            MemRefLayoutAttrInterface layout,
                                            Attribute memorySpace) {
  // Case 1: Unranked memref type.
  if (auto unrankedTensorType = tensorType.dyn_cast<UnrankedTensorType>()) {
    assert(!layout && "UnrankedTensorType cannot have a layout map");
    return UnrankedMemRefType::get(unrankedTensorType.getElementType(),
                                   memorySpace);
  }

  // Case 2: Ranked memref type with specified layout. If fully dynamic layout
  // maps are not requested, generate a type with `layout`, which is empty (no
  // layout map) by default.
  auto rankedTensorType = tensorType.cast<RankedTensorType>();
  if (layout || !options.fullyDynamicLayoutMaps) {
    return MemRefType::get(rankedTensorType.getShape(),
                           rankedTensorType.getElementType(), layout,
                           memorySpace);
  }

  // Case 3: Ranked memref type with unspecified layout. Choose the most dynamic
  // one.
  // TODO: address space decisions to connect with the actual alloc.
  int64_t dynamicOffset = ShapedType::kDynamicStrideOrOffset;
  SmallVector<int64_t> dynamicStrides(rankedTensorType.getRank(),
                                      ShapedType::kDynamicStrideOrOffset);
  AffineMap stridedLayout = makeStridedLinearLayoutMap(
      dynamicStrides, dynamicOffset, rankedTensorType.getContext());
  return MemRefType::get(rankedTensorType.getShape(),
                         rankedTensorType.getElementType(), stridedLayout,
                         memorySpace);
}