blob: bf868642f396c9fbe86e16a38d7ed95c5266932d [file] [log] [blame]
/*
* Copyright (c) 2011, 2014, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*/
package org.graalvm.compiler.graph.spi;
import org.graalvm.compiler.graph.Graph;
import org.graalvm.compiler.graph.Node;
import jdk.vm.ci.meta.MetaAccessProvider;
/**
* Nodes can implement {@link Canonicalizable} or one of the two sub-interfaces {@link Unary} and
* {@link 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.<br/>
* <br/>
* <b>Implementations of {@link Canonicalizable#canonical(CanonicalizerTool)} or the equivalent
* methods of the two sub-interfaces must not have any side effects.</b><br/>
* 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.<br/>
* <br/>
* 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.
*/
public interface Canonicalizable {
/**
* 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.
* <br/>
* The return value determines the effect that the canonicalization will have:
* <ul>
* <li>Returning an pre-existing node will replace the current node with the given one.</li>
* <li>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).</li>
* <li>Returning {@code null} will delete the current node and replace it with {@code 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.</li>
* </ul>
*
* @param tool provides access to runtime interfaces like {@link MetaAccessProvider}
*/
Node canonical(CanonicalizerTool tool);
/**
* This sub-interface of {@link Canonicalizable} is intended for nodes that have exactly one
* input. It has an additional {@link #canonical(CanonicalizerTool, Node)} method that looks at
* the given input instead of the current input of the node - which can be used to ask "what if
* this input is changed to this node" - questions.
*
* @param <T> the common supertype of all inputs of this node
*/
public interface Unary<T extends Node> extends Canonicalizable {
/**
* Similar to {@link Canonicalizable#canonical(CanonicalizerTool)}, except that
* implementations should act as if the current input of the node was the given one, i.e.,
* they should never look at the inputs via the this pointer.
*/
Node canonical(CanonicalizerTool tool, T forValue);
/**
* Gets the current value of the input, so that calling
* {@link #canonical(CanonicalizerTool, Node)} with the value returned from this method
* should behave exactly like {@link Canonicalizable#canonical(CanonicalizerTool)}.
*/
T getValue();
@SuppressWarnings("unchecked")
@Override
default T canonical(CanonicalizerTool tool) {
return (T) canonical(tool, getValue());
}
}
/**
* This sub-interface of {@link Canonicalizable} is intended for nodes that have exactly two
* inputs. It has an additional {@link #canonical(CanonicalizerTool, Node, Node)} method that
* looks at the given inputs instead of the current inputs of the node - which can be used to
* ask "what if this input is changed to this node" - questions.
*
* @param <T> the common supertype of all inputs of this node
*/
public interface Binary<T extends Node> extends Canonicalizable {
/**
* Similar to {@link Canonicalizable#canonical(CanonicalizerTool)}, except that
* implementations should act as if the current input of the node was the given one, i.e.,
* they should never look at the inputs via the this pointer.
*/
Node canonical(CanonicalizerTool tool, T forX, T forY);
/**
* Gets the current value of the input, so that calling
* {@link #canonical(CanonicalizerTool, Node, Node)} with the value returned from this
* method should behave exactly like {@link Canonicalizable#canonical(CanonicalizerTool)}.
*/
T getX();
/**
* Gets the current value of the input, so that calling
* {@link #canonical(CanonicalizerTool, Node, Node)} with the value returned from this
* method should behave exactly like {@link Canonicalizable#canonical(CanonicalizerTool)}.
*/
T getY();
@SuppressWarnings("unchecked")
@Override
default T canonical(CanonicalizerTool tool) {
return (T) canonical(tool, getX(), getY());
}
}
/**
* This sub-interface of {@link Canonicalizable.Binary} is for nodes with two inputs where the
* operation is commutative. It is used to improve GVN by trying to merge nodes with the same
* inputs in different order.
*/
public interface BinaryCommutative<T extends Node> extends Binary<T> {
/**
* Ensure a canonical ordering of inputs for commutative nodes to improve GVN results. Order
* the inputs by increasing {@link Node#id} and call {@link Graph#findDuplicate(Node)} on
* the node if it's currently in a graph. It's assumed that if there was a constant on the
* left it's been moved to the right by other code and that ordering is left alone.
*
* @return the original node or another node with the same input ordering
*/
Node maybeCommuteInputs();
}
}