com.oracle.graal.graph.spi

Interface Canonicalizable

All Known Subinterfaces:

Canonicalizable.Binary<T>, Canonicalizable.BinaryCommutative<T>, Canonicalizable.Unary<T>

All Known Implementing Classes:

AbsNode, AddNode, AMD64CountLeadingZerosNode, AMD64CountTrailingZerosNode, AMD64FloatConvertNode, AMD64MathIntrinsicNode, AndNode, ArrayCopyCallNode, ArrayEqualsNode, ArrayLengthNode, AssertionNode, BinaryArithmeticNode, BinaryNode, BinaryOpLogicNode, BitCountNode, BitScanForwardNode, BitScanReverseNode, BoxNode, CallSiteTargetNode, CheckCastDynamicNode, CheckCastNode, ClassCastNode, ClassGetHubNode, ClassIsAssignableFromNode, CompareNode, CompressionNode, ConditionalNode, ConditionAnchorNode, DeferredPiNode, DivNode, DynamicDeoptimizeNode, DynamicNewArrayNode, DynamicNewInstanceNode, FixedBinaryNode, FloatConvertNode, FloatEqualsNode, FloatingReadNode, FloatLessThanNode, GetClassNode, GuardedValueNode, GuardNode, GuardProxyNode, HubGetClassNode, IdentityHashCodeNode, InstanceOfDynamicNode, InstanceOfNode, IntegerAddExactNode, IntegerBelowNode, IntegerConvertNode, IntegerDivNode, IntegerEqualsNode, IntegerLessThanNode, IntegerMulExactNode, IntegerMulHighNode, IntegerRemNode, IntegerSubExactNode, IntegerTestNode, IsCompilationConstantNode, IsNullNode, JavaReadNode, KlassLayoutHelperNode, LeftShiftNode, LoadFieldNode, LoadHubNode, LoadIndexedNode, LoadIndexedPointerNode, LoadMethodNode, LoadSnippetVarargParameterNode, LogicNegationNode, MemoryAnchorNode, MulNode, NarrowNode, NegateNode, NewFrameNode, NormalizeCompareNode, NotNode, ObjectEqualsNode, OffsetAddressNode, OrNode, PiArrayNode, PiNode, PointerEqualsNode, PureFunctionMacroNode, ReadNode, ReflectionGetCallerClassNode, RegisterFinalizerNode, ReinterpretNode, RemNode, ReverseBytesNode, RightShiftNode, ShiftNode, ShortCircuitOrNode, SignExtendNode, SnippetLocationProxyNode, SqrtNode, SubNode, TypeCheckNode, UnaryArithmeticNode, UnaryNode, UnaryOpLogicNode, UnboxNode, UnsafeAccessNode, UnsafeCastNode, UnsafeLoadNode, UnsafeStoreNode, UnsignedDivNode, UnsignedMulHighNode, UnsignedRemNode, UnsignedRightShiftNode, ValueProxyNode, WordCastNode, XorNode, ZeroExtendNode

public interface Canonicalizable

Nodes can implement Canonicalizable or one of the two sub-interfaces Canonicalizable.Unary and Canonicalizable.Binary to provide local optimizations like constant folding and strength reduction. Implementations should return a replacement that is always semantically correct for the given inputs, or "this" if they do not see an opportunity for improvement.

Implementations of canonical(CanonicalizerTool) or the equivalent methods of the two sub-interfaces must not have any side effects.
They are not allowed to change inputs, successors or properties of any node (including the current one) and they also cannot add new nodes to the graph.

In addition to pre-existing nodes they can return newly created nodes, which will be added to the graph automatically if (and only if) the effects of the canonicalization are committed. Non-cyclic graphs (DAGs) of newly created nodes (i.e., one newly created node with an input to another newly created node) will be handled correctly.

Nested Class Summary

Nested Classes

Modifier and Type	Interface and Description
static interface	Canonicalizable.Binary<T extends Node> This sub-interface of Canonicalizable is intended for nodes that have exactly two inputs.
static interface	Canonicalizable.BinaryCommutative<T extends Node> This sub-interface of Canonicalizable.Binary is for nodes with two inputs where the operation is commutative.
static interface	Canonicalizable.Unary<T extends Node> This sub-interface of Canonicalizable is intended for nodes that have exactly one input.

Method Summary

Modifier and Type	Method and Description
Node	canonical(CanonicalizerTool tool) Implementations of this method can provide local optimizations like constant folding and strength reduction.

Method Detail

canonical

Node canonical(CanonicalizerTool tool)

Implementations of this method can provide local optimizations like constant folding and strength reduction. Implementations should look at the properties and inputs of the current node and determine if there is a more optimal and always semantically correct replacement.
The return value determines the effect that the canonicalization will have:

• Returning an pre-existing node will replace the current node with the given one.
• Returning a newly created node (that was not yet added to the graph) will replace the current node with the given one, after adding it to the graph. If both the replacement and the replacee are anchored in control flow (fixed nodes), the replacement will be added to the control flow. It is invalid to replace a non-fixed node with a newly created fixed node (because its placement in the control flow cannot be determined without scheduling).
• Returning null will delete the current node and replace it with null at all usages. Note that it is not necessary to delete floating nodes that have no more usages this way - they will be deleted automatically.

Parameters:

tool - provides access to runtime interfaces like MetaAccessProvider