/***
* ASM: a very small and fast Java bytecode manipulation framework
* Copyright (c) 2000-2011 INRIA, France Telecom
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the copyright holders nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
* THE POSSIBILITY OF SUCH DAMAGE.
*/
package nginx.clojure.asm.tree.analysis;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import nginx.clojure.asm.Opcodes;
import nginx.clojure.asm.Type;
import nginx.clojure.asm.tree.AbstractInsnNode;
import nginx.clojure.asm.tree.IincInsnNode;
import nginx.clojure.asm.tree.InsnList;
import nginx.clojure.asm.tree.JumpInsnNode;
import nginx.clojure.asm.tree.LabelNode;
import nginx.clojure.asm.tree.LookupSwitchInsnNode;
import nginx.clojure.asm.tree.MethodNode;
import nginx.clojure.asm.tree.TableSwitchInsnNode;
import nginx.clojure.asm.tree.TryCatchBlockNode;
import nginx.clojure.asm.tree.VarInsnNode;
/**
* A semantic bytecode analyzer. <i>This class does not fully check that JSR and
* RET instructions are valid.</i>
*
* @param <V>
* type of the Value used for the analysis.
*
* @author Eric Bruneton
*/
public class Analyzer<V extends Value> implements Opcodes {
private final Interpreter<V> interpreter;
private int n;
private InsnList insns;
private List<TryCatchBlockNode>[] handlers;
private Frame<V>[] frames;
private Subroutine[] subroutines;
private boolean[] queued;
private int[] queue;
private int top;
/**
* Constructs a new {@link Analyzer}.
*
* @param interpreter
* the interpreter to be used to symbolically interpret the
* bytecode instructions.
*/
public Analyzer(final Interpreter<V> interpreter) {
this.interpreter = interpreter;
}
/**
* Analyzes the given method.
*
* @param owner
* the internal name of the class to which the method belongs.
* @param m
* the method to be analyzed.
* @return the symbolic state of the execution stack frame at each bytecode
* instruction of the method. The size of the returned array is
* equal to the number of instructions (and labels) of the method. A
* given frame is <tt>null</tt> if and only if the corresponding
* instruction cannot be reached (dead code).
* @throws AnalyzerException
* if a problem occurs during the analysis.
*/
public Frame<V>[] analyze(final String owner, final MethodNode m)
throws AnalyzerException {
if ((m.access & (ACC_ABSTRACT | ACC_NATIVE)) != 0) {
frames = (Frame<V>[]) new Frame<?>[0];
return frames;
}
n = m.instructions.size();
insns = m.instructions;
handlers = (List<TryCatchBlockNode>[]) new List<?>[n];
frames = (Frame<V>[]) new Frame<?>[n];
subroutines = new Subroutine[n];
queued = new boolean[n];
queue = new int[n];
top = 0;
// computes exception handlers for each instruction
for (int i = 0; i < m.tryCatchBlocks.size(); ++i) {
TryCatchBlockNode tcb = m.tryCatchBlocks.get(i);
int begin = insns.indexOf(tcb.start);
int end = insns.indexOf(tcb.end);
for (int j = begin; j < end; ++j) {
List<TryCatchBlockNode> insnHandlers = handlers[j];
if (insnHandlers == null) {
insnHandlers = new ArrayList<TryCatchBlockNode>();
handlers[j] = insnHandlers;
}
insnHandlers.add(tcb);
}
}
// computes the subroutine for each instruction:
Subroutine main = new Subroutine(null, m.maxLocals, null);
List<AbstractInsnNode> subroutineCalls = new ArrayList<AbstractInsnNode>();
Map<LabelNode, Subroutine> subroutineHeads = new HashMap<LabelNode, Subroutine>();
findSubroutine(0, main, subroutineCalls);
while (!subroutineCalls.isEmpty()) {
JumpInsnNode jsr = (JumpInsnNode) subroutineCalls.remove(0);
Subroutine sub = subroutineHeads.get(jsr.label);
if (sub == null) {
sub = new Subroutine(jsr.label, m.maxLocals, jsr);
subroutineHeads.put(jsr.label, sub);
findSubroutine(insns.indexOf(jsr.label), sub, subroutineCalls);
} else {
sub.callers.add(jsr);
}
}
for (int i = 0; i < n; ++i) {
if (subroutines[i] != null && subroutines[i].start == null) {
subroutines[i] = null;
}
}
// initializes the data structures for the control flow analysis
Frame<V> current = newFrame(m.maxLocals, m.maxStack);
Frame<V> handler = newFrame(m.maxLocals, m.maxStack);
current.setReturn(interpreter.newValue(Type.getReturnType(m.desc)));
Type[] args = Type.getArgumentTypes(m.desc);
int local = 0;
if ((m.access & ACC_STATIC) == 0) {
Type ctype = Type.getObjectType(owner);
current.setLocal(local++, interpreter.newValue(ctype));
}
for (int i = 0; i < args.length; ++i) {
current.setLocal(local++, interpreter.newValue(args[i]));
if (args[i].getSize() == 2) {
current.setLocal(local++, interpreter.newValue(null));
}
}
while (local < m.maxLocals) {
current.setLocal(local++, interpreter.newValue(null));
}
merge(0, current, null);
init(owner, m);
// control flow analysis
while (top > 0) {
int insn = queue[--top];
Frame<V> f = frames[insn];
Subroutine subroutine = subroutines[insn];
queued[insn] = false;
AbstractInsnNode insnNode = null;
try {
insnNode = m.instructions.get(insn);
int insnOpcode = insnNode.getOpcode();
int insnType = insnNode.getType();
if (insnType == AbstractInsnNode.LABEL
|| insnType == AbstractInsnNode.LINE
|| insnType == AbstractInsnNode.FRAME) {
merge(insn + 1, f, subroutine);
newControlFlowEdge(insn, insn + 1);
} else {
current.init(f).execute(insnNode, interpreter);
subroutine = subroutine == null ? null : subroutine.copy();
if (insnNode instanceof JumpInsnNode) {
JumpInsnNode j = (JumpInsnNode) insnNode;
if (insnOpcode != GOTO && insnOpcode != JSR) {
merge(insn + 1, current, subroutine);
newControlFlowEdge(insn, insn + 1);
}
int jump = insns.indexOf(j.label);
if (insnOpcode == JSR) {
merge(jump, current, new Subroutine(j.label,
m.maxLocals, j));
} else {
merge(jump, current, subroutine);
}
newControlFlowEdge(insn, jump);
} else if (insnNode instanceof LookupSwitchInsnNode) {
LookupSwitchInsnNode lsi = (LookupSwitchInsnNode) insnNode;
int jump = insns.indexOf(lsi.dflt);
merge(jump, current, subroutine);
newControlFlowEdge(insn, jump);
for (int j = 0; j < lsi.labels.size(); ++j) {
LabelNode label = lsi.labels.get(j);
jump = insns.indexOf(label);
merge(jump, current, subroutine);
newControlFlowEdge(insn, jump);
}
} else if (insnNode instanceof TableSwitchInsnNode) {
TableSwitchInsnNode tsi = (TableSwitchInsnNode) insnNode;
int jump = insns.indexOf(tsi.dflt);
merge(jump, current, subroutine);
newControlFlowEdge(insn, jump);
for (int j = 0; j < tsi.labels.size(); ++j) {
LabelNode label = tsi.labels.get(j);
jump = insns.indexOf(label);
merge(jump, current, subroutine);
newControlFlowEdge(insn, jump);
}
} else if (insnOpcode == RET) {
if (subroutine == null) {
throw new AnalyzerException(insnNode,
"RET instruction outside of a sub routine");
}
for (int i = 0; i < subroutine.callers.size(); ++i) {
JumpInsnNode caller = subroutine.callers.get(i);
int call = insns.indexOf(caller);
if (frames[call] != null) {
merge(call + 1, frames[call], current,
subroutines[call], subroutine.access);
newControlFlowEdge(insn, call + 1);
}
}
} else if (insnOpcode != ATHROW
&& (insnOpcode < IRETURN || insnOpcode > RETURN)) {
if (subroutine != null) {
if (insnNode instanceof VarInsnNode) {
int var = ((VarInsnNode) insnNode).var;
subroutine.access[var] = true;
if (insnOpcode == LLOAD || insnOpcode == DLOAD
|| insnOpcode == LSTORE
|| insnOpcode == DSTORE) {
subroutine.access[var + 1] = true;
}
} else if (insnNode instanceof IincInsnNode) {
int var = ((IincInsnNode) insnNode).var;
subroutine.access[var] = true;
}
}
merge(insn + 1, current, subroutine);
newControlFlowEdge(insn, insn + 1);
}
}
List<TryCatchBlockNode> insnHandlers = handlers[insn];
if (insnHandlers != null) {
for (int i = 0; i < insnHandlers.size(); ++i) {
TryCatchBlockNode tcb = insnHandlers.get(i);
Type type;
if (tcb.type == null) {
type = Type.getObjectType("java/lang/Throwable");
} else {
type = Type.getObjectType(tcb.type);
}
int jump = insns.indexOf(tcb.handler);
if (newControlFlowExceptionEdge(insn, tcb)) {
handler.init(f);
handler.clearStack();
handler.push(interpreter.newValue(type));
merge(jump, handler, subroutine);
}
}
}
} catch (AnalyzerException e) {
throw new AnalyzerException(e.node, "Error at instruction "
+ insn + ": " + e.getMessage(), e);
} catch (Exception e) {
throw new AnalyzerException(insnNode, "Error at instruction "
+ insn + ": " + e.getMessage(), e);
}
}
return frames;
}
private void findSubroutine(int insn, final Subroutine sub,
final List<AbstractInsnNode> calls) throws AnalyzerException {
while (true) {
if (insn < 0 || insn >= n) {
throw new AnalyzerException(null,
"Execution can fall off end of the code");
}
if (subroutines[insn] != null) {
return;
}
subroutines[insn] = sub.copy();
AbstractInsnNode node = insns.get(insn);
// calls findSubroutine recursively on normal successors
if (node instanceof JumpInsnNode) {
if (node.getOpcode() == JSR) {
// do not follow a JSR, it leads to another subroutine!
calls.add(node);
} else {
JumpInsnNode jnode = (JumpInsnNode) node;
findSubroutine(insns.indexOf(jnode.label), sub, calls);
}
} else if (node instanceof TableSwitchInsnNode) {
TableSwitchInsnNode tsnode = (TableSwitchInsnNode) node;
findSubroutine(insns.indexOf(tsnode.dflt), sub, calls);
for (int i = tsnode.labels.size() - 1; i >= 0; --i) {
LabelNode l = tsnode.labels.get(i);
findSubroutine(insns.indexOf(l), sub, calls);
}
} else if (node instanceof LookupSwitchInsnNode) {
LookupSwitchInsnNode lsnode = (LookupSwitchInsnNode) node;
findSubroutine(insns.indexOf(lsnode.dflt), sub, calls);
for (int i = lsnode.labels.size() - 1; i >= 0; --i) {
LabelNode l = lsnode.labels.get(i);
findSubroutine(insns.indexOf(l), sub, calls);
}
}
// calls findSubroutine recursively on exception handler successors
List<TryCatchBlockNode> insnHandlers = handlers[insn];
if (insnHandlers != null) {
for (int i = 0; i < insnHandlers.size(); ++i) {
TryCatchBlockNode tcb = insnHandlers.get(i);
findSubroutine(insns.indexOf(tcb.handler), sub, calls);
}
}
// if insn does not falls through to the next instruction, return.
switch (node.getOpcode()) {
case GOTO:
case RET:
case TABLESWITCH:
case LOOKUPSWITCH:
case IRETURN:
case LRETURN:
case FRETURN:
case DRETURN:
case ARETURN:
case RETURN:
case ATHROW:
return;
}
insn++;
}
}
/**
* Returns the symbolic stack frame for each instruction of the last
* recently analyzed method.
*
* @return the symbolic state of the execution stack frame at each bytecode
* instruction of the method. The size of the returned array is
* equal to the number of instructions (and labels) of the method. A
* given frame is <tt>null</tt> if the corresponding instruction
* cannot be reached, or if an error occured during the analysis of
* the method.
*/
public Frame<V>[] getFrames() {
return frames;
}
/**
* Returns the exception handlers for the given instruction.
*
* @param insn
* the index of an instruction of the last recently analyzed
* method.
* @return a list of {@link TryCatchBlockNode} objects.
*/
public List<TryCatchBlockNode> getHandlers(final int insn) {
return handlers[insn];
}
/**
* Initializes this analyzer. This method is called just before the
* execution of control flow analysis loop in #analyze. The default
* implementation of this method does nothing.
*
* @param owner
* the internal name of the class to which the method belongs.
* @param m
* the method to be analyzed.
* @throws AnalyzerException
* if a problem occurs.
*/
protected void init(String owner, MethodNode m) throws AnalyzerException {
}
/**
* Constructs a new frame with the given size.
*
* @param nLocals
* the maximum number of local variables of the frame.
* @param nStack
* the maximum stack size of the frame.
* @return the created frame.
*/
protected Frame<V> newFrame(final int nLocals, final int nStack) {
return new Frame<V>(nLocals, nStack);
}
/**
* Constructs a new frame that is identical to the given frame.
*
* @param src
* a frame.
* @return the created frame.
*/
protected Frame<V> newFrame(final Frame<? extends V> src) {
return new Frame<V>(src);
}
/**
* Creates a control flow graph edge. The default implementation of this
* method does nothing. It can be overriden in order to construct the
* control flow graph of a method (this method is called by the
* {@link #analyze analyze} method during its visit of the method's code).
*
* @param insn
* an instruction index.
* @param successor
* index of a successor instruction.
*/
protected void newControlFlowEdge(final int insn, final int successor) {
}
/**
* Creates a control flow graph edge corresponding to an exception handler.
* The default implementation of this method does nothing. It can be
* overridden in order to construct the control flow graph of a method (this
* method is called by the {@link #analyze analyze} method during its visit
* of the method's code).
*
* @param insn
* an instruction index.
* @param successor
* index of a successor instruction.
* @return true if this edge must be considered in the data flow analysis
* performed by this analyzer, or false otherwise. The default
* implementation of this method always returns true.
*/
protected boolean newControlFlowExceptionEdge(final int insn,
final int successor) {
return true;
}
/**
* Creates a control flow graph edge corresponding to an exception handler.
* The default implementation of this method delegates to
* {@link #newControlFlowExceptionEdge(int, int)
* newControlFlowExceptionEdge(int, int)}. It can be overridden in order to
* construct the control flow graph of a method (this method is called by
* the {@link #analyze analyze} method during its visit of the method's
* code).
*
* @param insn
* an instruction index.
* @param tcb
* TryCatchBlockNode corresponding to this edge.
* @return true if this edge must be considered in the data flow analysis
* performed by this analyzer, or false otherwise. The default
* implementation of this method delegates to
* {@link #newControlFlowExceptionEdge(int, int)
* newControlFlowExceptionEdge(int, int)}.
*/
protected boolean newControlFlowExceptionEdge(final int insn,
final TryCatchBlockNode tcb) {
return newControlFlowExceptionEdge(insn, insns.indexOf(tcb.handler));
}
// -------------------------------------------------------------------------
private void merge(final int insn, final Frame<V> frame,
final Subroutine subroutine) throws AnalyzerException {
Frame<V> oldFrame = frames[insn];
Subroutine oldSubroutine = subroutines[insn];
boolean changes;
if (oldFrame == null) {
frames[insn] = newFrame(frame);
changes = true;
} else {
changes = oldFrame.merge(frame, interpreter);
}
if (oldSubroutine == null) {
if (subroutine != null) {
subroutines[insn] = subroutine.copy();
changes = true;
}
} else {
if (subroutine != null) {
changes |= oldSubroutine.merge(subroutine);
}
}
if (changes && !queued[insn]) {
queued[insn] = true;
queue[top++] = insn;
}
}
private void merge(final int insn, final Frame<V> beforeJSR,
final Frame<V> afterRET, final Subroutine subroutineBeforeJSR,
final boolean[] access) throws AnalyzerException {
Frame<V> oldFrame = frames[insn];
Subroutine oldSubroutine = subroutines[insn];
boolean changes;
afterRET.merge(beforeJSR, access);
if (oldFrame == null) {
frames[insn] = newFrame(afterRET);
changes = true;
} else {
changes = oldFrame.merge(afterRET, interpreter);
}
if (oldSubroutine != null && subroutineBeforeJSR != null) {
changes |= oldSubroutine.merge(subroutineBeforeJSR);
}
if (changes && !queued[insn]) {
queued[insn] = true;
queue[top++] = insn;
}
}
}