/*
* Copyright (c) 2009, 2011, 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 com.oracle.max.graal.compiler.alloc;
import java.util.*;
import com.oracle.max.graal.compiler.lir.*;
/**
* This class optimizes moves, particularly those that result from eliminating SSA form.
*
* When a block has more than one predecessor, and all predecessors end with
* the {@linkplain #same(LIRInstruction, LIRInstruction) same} sequence of
* {@linkplain LIROpcode#Move move} instructions, then these sequences
* can be replaced with a single copy of the sequence at the beginning of the block.
*
* Similarly, when a block has more than one successor, then same sequences of
* moves at the beginning of the successors can be placed once at the end of
* the block. But because the moves must be inserted before all branch
* instructions, this works only when there is exactly one conditional branch
* at the end of the block (because the moves must be inserted before all
* branches, but after all compares).
*
* This optimization affects all kind of moves (reg->reg, reg->stack and
* stack->reg). Because this optimization works best when a block contains only
* a few moves, it has a huge impact on the number of blocks that are totally
* empty.
*/
final class EdgeMoveOptimizer {
/**
* Optimizes moves on block edges.
*
* @param blockList a list of blocks whose moves should be optimized
*/
public static void optimize(List<LIRBlock> blockList) {
EdgeMoveOptimizer optimizer = new EdgeMoveOptimizer();
// ignore the first block in the list (index 0 is not processed)
for (int i = blockList.size() - 1; i >= 1; i--) {
LIRBlock block = blockList.get(i);
if (block.numberOfPreds() > 1) {
optimizer.optimizeMovesAtBlockEnd(block);
}
if (block.numberOfSux() == 2) {
optimizer.optimizeMovesAtBlockBegin(block);
}
}
}
private final List<List<LIRInstruction>> edgeInstructionSeqences;
private EdgeMoveOptimizer() {
edgeInstructionSeqences = new ArrayList<List<LIRInstruction>>(4);
}
/**
* Determines if two operations are both {@linkplain LIROpcode#Move moves}
* that have the same {@linkplain LIRInstruction#operand() source} and {@linkplain LIRInstruction#result() destination}
* operands and they have the same {@linkplain LIRInstruction#info debug info}.
*
* @param op1 the first instruction to compare
* @param op2 the second instruction to compare
* @return {@code true} if {@code op1} and {@code op2} are the same by the above algorithm
*/
private boolean same(LIRInstruction op1, LIRInstruction op2) {
assert op1 != null;
assert op2 != null;
if (op1.code == StandardOpcode.MOVE && op2.code == StandardOpcode.MOVE) {
if (op1.info == op2.info && op1.input(0).equals(op2.input(0)) && op1.result().equals(op2.result())) {
// these moves are exactly equal and can be optimized
return true;
}
}
return false;
}
/**
* Moves the longest {@linkplain #same common} subsequence at the end all
* predecessors of {@code block} to the start of {@code block}.
*/
private void optimizeMovesAtBlockEnd(LIRBlock block) {
if (block.isPredecessor(block)) {
// currently we can't handle this correctly.
return;
}
// clear all internal data structures
edgeInstructionSeqences.clear();
int numPreds = block.numberOfPreds();
assert numPreds > 1 : "do not call otherwise";
// setup a list with the LIR instructions of all predecessors
for (int i = 0; i < numPreds; i++) {
LIRBlock pred = block.predAt(i);
assert pred != null;
assert pred.lir() != null;
List<LIRInstruction> predInstructions = pred.lir();
if (pred.numberOfSux() != 1) {
// this can happen with switch-statements where multiple edges are between
// the same blocks.
return;
}
if (predInstructions.get(predInstructions.size() - 1).code == StandardOpcode.XIR) {
return;
}
assert pred.suxAt(0) == block : "invalid control flow";
assert predInstructions.get(predInstructions.size() - 1).code == StandardOpcode.JUMP : "block must end with unconditional jump";
if (predInstructions.get(predInstructions.size() - 1).info != null) {
// can not optimize instructions that have debug info
return;
}
// ignore the unconditional branch at the end of the block
List<LIRInstruction> seq = predInstructions.subList(0, predInstructions.size() - 1);
edgeInstructionSeqences.add(seq);
}
// process lir-instructions while all predecessors end with the same instruction
while (true) {
List<LIRInstruction> seq = edgeInstructionSeqences.get(0);
if (seq.isEmpty()) {
return;
}
LIRInstruction op = last(seq);
for (int i = 1; i < numPreds; ++i) {
List<LIRInstruction> otherSeq = edgeInstructionSeqences.get(i);
if (otherSeq.isEmpty() || !same(op, last(otherSeq))) {
return;
}
}
// insert the instruction at the beginning of the current block
block.lir().add(1, op);
// delete the instruction at the end of all predecessors
for (int i = 0; i < numPreds; i++) {
seq = edgeInstructionSeqences.get(i);
removeLast(seq);
}
}
}
/**
* Moves the longest {@linkplain #same common} subsequence at the start of all
* successors of {@code block} to the end of {@code block} just prior to the
* branch instruction ending {@code block}.
*/
private void optimizeMovesAtBlockBegin(LIRBlock block) {
edgeInstructionSeqences.clear();
int numSux = block.numberOfSux();
List<LIRInstruction> instructions = block.lir();
assert numSux == 2 : "method should not be called otherwise";
if (instructions.get(instructions.size() - 1).code == StandardOpcode.XIR) {
// cannot optimize when last instruction is Xir.
return;
}
assert instructions.get(instructions.size() - 1).code == StandardOpcode.JUMP : "block must end with unconditional jump";
if (instructions.get(instructions.size() - 1).info != null) {
// cannot optimize instructions when debug info is needed
return;
}
LIRInstruction branch = instructions.get(instructions.size() - 2);
if (!(branch instanceof LIRBranch) || branch.info != null) {
// not a valid case for optimization
// currently, only blocks that end with two branches (conditional branch followed
// by unconditional branch) are optimized
return;
}
// now it is guaranteed that the block ends with two branch instructions.
// the instructions are inserted at the end of the block before these two branches
int insertIdx = instructions.size() - 2;
// setup a list with the lir-instructions of all successors
for (int i = 0; i < numSux; i++) {
LIRBlock sux = block.suxAt(i);
List<LIRInstruction> suxInstructions = sux.lir();
assert suxInstructions.get(0).code == StandardOpcode.LABEL : "block must start with label";
if (sux.numberOfPreds() != 1) {
// this can happen with switch-statements where multiple edges are between
// the same blocks.
return;
}
assert sux.predAt(0) == block : "invalid control flow";
// ignore the label at the beginning of the block
List<LIRInstruction> seq = suxInstructions.subList(1, suxInstructions.size());
edgeInstructionSeqences.add(seq);
}
// process LIR instructions while all successors begin with the same instruction
while (true) {
List<LIRInstruction> seq = edgeInstructionSeqences.get(0);
if (seq.isEmpty()) {
return;
}
LIRInstruction op = first(seq);
for (int i = 1; i < numSux; i++) {
List<LIRInstruction> otherSeq = edgeInstructionSeqences.get(i);
if (otherSeq.isEmpty() || !same(op, first(otherSeq))) {
// these instructions are different and cannot be optimized .
// no further optimization possible
return;
}
}
// insert instruction at end of current block
block.lir().add(insertIdx, op);
insertIdx++;
// delete the instructions at the beginning of all successors
for (int i = 0; i < numSux; i++) {
seq = edgeInstructionSeqences.get(i);
removeFirst(seq);
}
}
}
/**
* Gets the first element from a LIR instruction sequence.
*/
private static LIRInstruction first(List<LIRInstruction> seq) {
return seq.get(0);
}
/**
* Gets the last element from a LIR instruction sequence.
*/
private static LIRInstruction last(List<LIRInstruction> seq) {
return seq.get(seq.size() - 1);
}
/**
* Removes the first element from a LIR instruction sequence.
*/
private static void removeFirst(List<LIRInstruction> seq) {
seq.remove(0);
}
/**
* Removes the last element from a LIR instruction sequence.
*/
private static void removeLast(List<LIRInstruction> seq) {
seq.remove(seq.size() - 1);
}
}