Mastering Java Bytecode - JAX.de 2012

Mastering Java Bytecode
Anton Arhipov | ZeroTurnaround

whoami
• Anton Arhipov
• Java Dev / Product Lead
• ZeroTurnaround, JRebel

• @antonarhipov @javarebel

Today – Java Core
• Mastering Java Bytecode

• The Future of Java on Multi-
Core, Lambdas, Spliterators and
Methods

• OpenJDK JVM Internals

• Invokedynamic Deep Dive

Why Bytecode?
• Know your platform!
• Create your own compiler?
• Programming model (AOP, ORM)
• Awesome tools (JRebel )

• … just bored?

1+2

12+ PUSH 1
PUSH 2
2
1

1+2

12+ PUSH 1
PUSH 2
3
ADD

1+2

12+ ICONST_1
ICONST_2
3
IADD

Bytecode
• One-byte instructions
• 256 possible opcodes
• 200+ in use

TYPE OPERATION

• <TYPE> ::= b, s, c, i, l, f, d, a

TYPE OPERATION

• <TYPE> ::= b, s, c, i, l, f, d, a
• constant values (ldc, iconst_1)

TYPE OPERATION

• <TYPE> ::= b, s, c, i, l, f, d, a
• constant values (ldc, iconst_1)
• Local variables and stack interaction (load/store)
• Array operations (aload, astore)
• Math (add, sub, mul, div)
• Boolean/bitwise operations (iand, ixor)
• Comparisons (cmpg, cmpl, ifne, ifeq)
• Conversions (l2d, i2l)

Bytecode Taxonomy

Stack
Manipulation

Bytecode Taxonomy

Stack Flow
Manipulation Control

Bytecode Taxonomy

Stack Flow

Object
Model

Bytecode Taxonomy

Stack Flow

Object
Arithmetics
Model

Bytecode Taxonomy

Stack Flow

monitorenter
monitorexit

Object
Arithmetics
Model

javap
• Java class file disassembler
• Used with no options shows class structure
only
– Methods, superclass, interfaces, etc
• -c shows the bytecode
• -private shows all methods and members
• -s prints internal signatures
• -l prints line numbers and local variable
tables

C:workgeeconclasses>javap Hello -c

Compiled from "Hello.java"
public class Hello extends java.lang.Object{
public Hello();
Code:
0: aload_0
1: invokespecial #1; //Method java/lang/Object."<init>":()V
4: return

public Hello();
Code: the default constructor
0: aload_0
4: return

public Hello();
Code: push this to stack
0: aload_0
4: return

public Hello();
Code:
0: aload_0
4: return

invoke <init> on this

public Hello();
Code:
0: aload_0
4: return

public static void main(java.lang.String[]);
Code:
0: getstatic #2; //Field java/lang/System.out:Ljava/io/PrintStream;
3: ldc #3; //String Hello, World!
5: invokevirtual #4; //Method java/io/PrintStream.println:(Ljava/lang/String;)V

public Hello();
Code:
0: aload_0
4: return
get static field
Code:

public Hello();
Code:
0: aload_0
4: return

Code:

load string to the stack

public Hello();
Code:
0: aload_0
4: return

Code:

invoke method with parameter

What’s #1,#2, etc ?
public Hello();
Code:
0: aload_0
4: return

Code:

C:workgeeconclasses>javap Hello -c -verbose

Compiled from "Hello.java“
public class Hello extends java.lang.Object
SourceFile: "Hello.java"
minor version: 0
major version: 50
Constant pool:
const #1 = Method #6.#20; // java/lang/Object."<init>":()V
const #2 = Field #21.#22; // java/lang/System.out:Ljava/io/PrintStream;
const #3 = String #23; // Hello, World!
const #4 = Method #24.#25; // java/io/PrintStream.println:(Ljava/lang/String;)V
const #5 = class #26; // Hello
const #6 = class #27; // java/lang/Object
const #7 = Asciz <init>;
const #8 = Asciz ()V;

…
public Hello();
Code:
Stack=1, Locals=1, Args_size=1
0: aload_0
4: return
LineNumberTable:
line 1: 0

LocalVariableTable:
Start Length Slot Name Signature
0 5 0 this LHello;

…
Code:
8: return
LineNumberTable:
line 4: 0
line 5: 8
LocalVariableTable:
0 9 0 args [Ljava/lang/String;

Stack Machine
• JVM is a stack-based machine

Stack Machine
• Each thread has a stack

Stack Machine
• Stack stores frames

Stack Machine
• Frame is created on method invocation

Stack Machine
• Frame is created on method invocation
• Frame consists of:
– Operand stack
– Array of local variables

The Frame
Local variables
0 1 2 … N
Operand stack

#1
Constant
Pool

public java.lang.String getName();
Code:
0: aload_0
1: getfield #2; //Field name:Ljava/lang/String;
4: areturn
LocalVariableTable:
0 5 0 this LGet;

0 1 2 3 4

aload_0 getfield 00 02 areturn

Code:
0: aload_0
4: areturn
LocalVariableTable:
0 5 0 this LGet;

0 1 2 3 4

2A B4 00 02 B0

Code:
0: aload_0
4: areturn
LocalVariableTable:
0 5 0 this LGet;

dup A
pop B
swap
dup_x1
dup2_x1

dup A
pop A
swap B
dup_x1
dup2_x1

dup B
pop A
swap
dup_x1
dup2_x1

dup B
pop A
swap B
dup_x1
dup2_x1

dup B
pop A
swap B
dup_x1 B
dup2_x1 A

dup2_x2

How do you
swap doubles?

dup2_x2
dconst_0
0.0

dup2_x2
dconst_0
dconst_1 1.0

0.0

dup2_x2
dconst_0
dconst_1 1.0
swap
0.0

dup2_x2
dconst_0
dconst_1 1.0
swap not allowed!

0.0

dup2_x2
dconst_0
dconst_1 1.0
swap2
0.0

dup2_x2
dconst_0
dconst_1 1.0
doesn’t
swap2 exist
0.0

dup2_x2
dconst_0
dconst_1 1.0
dup2_x2
0.0

1.0

dup2_x2
dconst_0
dconst_1 0.0
dup2_x2
pop2 1.0

dup2_x2
dconst_0
dconst_1 0.0
dup2_x2
pop2 1.0

profit! 

Local Variables

public int calculate(int);
Code:
…

LocalVariableTable:
0 5 0 this LLocalVariables;
0 5 1 value I

Local Variables

The table
Code: maps
Stack=2, Locals=2, Args_size=2 numbers to
…
names
LocalVariableTable:
0 5 1 value I

Local Variables

Code:
Stack=2, Locals=2, Args_size=2 Sized explicitly
…

LocalVariableTable:
0 5 1 value I

Local Variables Stack

var value depth value
ldc "Hello"
0 0
astore_0
1 iconst_1 1
astore_1
2 aload_0 2

3 3

4 4


ldc "Hello"
0 0 "Hello"
astore_0
1 iconst_1 1
astore_1
2 aload_0 2

3 3

4 4


ldc "Hello"
0 "Hello" 0
astore_0
1 iconst_1 1
astore_1
2 aload_0 2

3 3

4 4


ldc "Hello"
0 "Hello" 0 1
astore_0
1 iconst_1 1
astore_1
2 aload_0 2

3 3

4 4


ldc "Hello"
0 "Hello" 0
astore_0
1 1 iconst_1 1
astore_1
2 aload_0 2

3 3

4 4


ldc "Hello"
0 "Hello" 0 "Hello"
astore_0
1 1 iconst_1 1
astore_1
2 aload_0 2

3 3

4 4

load
Local
Variables Stack
Table

store

Object Initialization
new
0xBB

<init>
Instance initialization method

<clinit>
Class and interface
initialization method

Object Initialization: static {}

static {};
Code:
0: iconst_1
1: putstatic #2; //Field a:I
4: iconst_2
5: putstatic #3; //Field b:I
8: return

Object Initialization: static {}

<clinit>

static {};
Code:
0: iconst_1
1: putstatic #2; //Field a:I
4: iconst_2
5: putstatic #3; //Field b:I
8: return

Object Initialization: new

public Initializer();
Code:


Code:
0: aload_0


Code:
0: aload_0
4: aload_0


Code:
0: aload_0
4: aload_0
5: new #2; //class java/lang/Object
8: dup


Code:
0: aload_0
4: aload_0
8: dup
12: putfield #3; //Field o:Ljava/lang/Object;


Code:
0: aload_0
4: aload_0
8: dup
12: putfield #3; //Field o:Ljava/lang/Object;
15: return

Object Initialization: {}
public Initializer(int);
Code:
0:aload_0
1:invokespecial #1; // ..<init>
4:aload_0
5:iconst_1
6:putfield #2; //Field a:I
9:aload_0
10:iconst_2
11: putfield #3; //Field c:I
14:aload_0
15:iload_1
16:putfield #4; //Field b:I
19:return

invokeXXX
• invokestatic
• invokespecial
• invokevirtual
• invokeinterface
• invokedynamic

invokestatic
• invokestatic
• invokespecial
• invokevirtual Integer.valueOf(“42”)
• invokeinterface
• invokedynamic

invokespecial
• invokestatic <init>
• invokespecial
• invokevirtual private void foo();
• invokeinterface
• invokedynamic super.method();

invokevirtual
class A
• invokestatic A/method1
A/method2
• invokespecial
• invokevirtual
• invokeinterface
• invokedynamic

invokevirtual
class A
A/method2
• invokespecial
class B
• invokevirtual
• invokeinterface
• invokedynamic

invokevirtual
class A
A/method2
• invokespecial
class B
• invokevirtual A/method1
B/method2
• invokeinterface B/method3

• invokedynamic

invokeinterface
class A
A/method2
• invokespecial
class B impl X
B/method2
• invokeinterface B/method3
X/methodX
• invokedynamic

invokeinterface
class A
A/method2
• invokespecial
class B impl X

• invokeinterface
B/method2
B/method3
D impl X
X/methodX D/method1
• invokedynamic X/methodX

invokeinterface
class A
A/method2
• invokespecial
class B impl X

• invokeinterface
B/method2
B/method3
D impl X
X/methodX D/method1
• invokedynamic X/methodX

Efficient Implementation of Java Interfaces: Invokeinterface Considered
Harmless, Bowen Alpern, Anthony Cocchi, Stephen Fink, David Grove, and
Derek Lieber, OOPSLA’01

Method Invocation
obj.method(param1, param2);

Method Invocation

push obj
push param1
push param2
call method

Method Invocation
obj
push obj
push param1
push param2
call method

Method Invocation
param1
push obj
obj
push param1
push param2
call method

Method Invocation
param2
push obj param1
push param1
push param2 obj
call method

Method Invocation
obj?
push obj
push param1
push param2
call method

Method Invocation
this.add(1, 2);
0: aload_0
1: iconst_1
2: iconst_2
3: invokevirtual #2; //Method add:(II)I

Inner Classes

class Car$Engine extends j.l.Object{
final Car this$0;

Car$Engine(Car);

public void start();
Code:
0: aload_0
1: getfield #1; //Field this$0:LCar;
4: invokestatic #3; // Car.access$000:(LCar;)V
7: return

}

Inner Classes
public class Car extends j.l.Object{
public Car();
private void move();
class Car$Engine extends j.l.Object{
final Car this$0; static void access$000(Car);
Code:
Car$Engine(Car); 0: aload_0
1: invokespecial #1; // move: ()V;
public void start(); 4: return
Code: }
0: aload_0
1: getfield #1; //Field this$0:LCar;
4: invokestatic #3; // Car.access$000:(LCar;)V
7: return

}

object Singleton {
def test={}
}

object Singleton {
def test={}
}

$> scalac Singleton.scala

Singleton.class Singleton$.class

public final class Singleton extends java.lang.Object {
public static final void test();
Code:
0: getstatic #11; //Field Singleton$.MODULE$:LSingleton$;
3: invokevirtual #13; //Method Singleton$.test:()V
6: return
}

Code:
6: return
}
public final class Singleton$ extends java.lang.Object implements
scala.ScalaObject {
public static final Singleton$ MODULE$;

public static {};
Code:
0: new #9; //class Singleton$
3: invokespecial #12; //Method "<init>":()V
6: return

public void test();
private Singleton$();
}

Code:
6: return
}
public final class Singleton$ extends java.lang.Object implements
scala.ScalaObject {
public static final Singleton$ MODULE$;

public static {};
public void test();
private Singleton$();
Code:
0: aload_0
4: aload_0
5: putstatic #19; //Field MODULE$:LSingleton$;
8: return

public class Singleton {
public void test(){
Singleton$.MODULE$.test();
}
}

object Singleton {
def test={}
}

public class Singleton {
public void test(){
Singleton$.MODULE$.test();
}
}

public final class Singleton$
object Singleton { implements scala.ScalaObject {
def test={} public static final Singleton$ MODULE$;
} static { new Singleton$(); }

private Singleton$(){
MODULE$ = this;
}

public void test() {
}
}

class Groovy { }
$> groovyc Groovy.groovy
$> javap –c –p Groovy

class Groovy { }
$> groovyc Groovy.groovy
$> javap –c –p Groovy
public class Test extends java.lang.Object implements groovy.lang.GroovyObject{
private static org.codehaus.groovy.reflection.ClassInfo $staticClassInfo;
private transient groovy.lang.MetaClass metaClass;
public static java.lang.Long __timeStamp;
public static java.lang.Long __timeStamp__239_neverHappen1304807931117;
private static java.lang.ref.SoftReference $callSiteArray;
private static java.lang.Class $class$groovy$lang$MetaClass;
private static java.lang.Class $class$Test;
private static java.lang.Class $class$java$lang$String;
public java.lang.Object this$dist$invoke$2(java.lang.String, java.lang.Object);
public void this$dist$set$2(java.lang.String, java.lang.Object);
public java.lang.Object this$dist$get$2(java.lang.String);
protected groovy.lang.MetaClass $getStaticMetaClass();
public groovy.lang.MetaClass getMetaClass();
public void setMetaClass(groovy.lang.MetaClass);
public java.lang.Object invokeMethod(java.lang.String, java.lang.Object);
public java.lang.Object getProperty(java.lang.String);
public void setProperty(java.lang.String, java.lang.Object);

SLIDES
GOTO IDE
SLIDES
IDE: DEMO

https://github.com/antonarhipov

@antonarhipov

anton@zeroturnaround.com

Mastering Java Bytecode - JAX.de 2012

More Related Content

What's hot

Viewers also liked

Similar to Mastering Java Bytecode - JAX.de 2012

More from Anton Arhipov

Recently uploaded

Mastering Java Bytecode - JAX.de 2012