Java bytecode manipulation and analysis framework
Java bytecode manipulation and analysis framework
ASM is an all purpose Java bytecode manipulation and analysis framework. It can be used to modify existing classes or dynamically generate classes, directly in binary form. This demo will guide you to have basic understand how ASM works and what will ASM do.
The ASM bytecode manipulation framework is written in Java. Java Runtime Environment is required.
You can download the latest binary file from the ObjectWeb Forge. This demo uses the version 5.2.
Bytecode Outline plugin for Eclipse shows disassembled bytecode of current Java editor or class file. The Bytecode Outline plugin can be installed from the Eclipse Update Manager with the ObjectWeb Eclipse Update Site http://download.forge.objectweb.org/eclipse-update/
Here is a quick review in case you are not familiar with Java Bytecode. Java Bytecode is an intermediate code between Java source code and assembly code. Java source code .java
file can be compiled into Bytecode .class
file and run on where any computers have a Java Runtime Environment.
For example, consider the following block of code:
public class Test
{
public static void main(String[] args) {
printOne();
printOne();
printTwo();
}
public static void printOne() {
System.out.println("Hello World");
}
public static void printTwo() {
printOne();
printOne();
}
}
By using javac
to compile and then javap -c
to disassemble it, here is what we get:
public class Test {
public Test();
Code:
0: aload_0
1: invokespecial #1 // Method java/lang/Object."":()V
4: return
public static void main(java.lang.String[]);
Code:
0: invokestatic #2 // Method printOne:()V
3: invokestatic #2 // Method printOne:()V
6: invokestatic #3 // Method printTwo:()V
9: return
public static void printOne();
Code:
0: getstatic #4 // Field java/lang/System.out:Ljava/io/PrintStream;
3: ldc #5 // String Hello World
5: invokevirtual #6 // Method java/io/PrintStream.println:(Ljava/lang/String;)V
8: return
public static void printTwo();
Code:
0: invokestatic #2 // Method printOne:()V
3: invokestatic #2 // Method printOne:()V
6: return
}
As mentioned before, ASM framework includes tools to help you translate between those codes. Bytecode Outline shows disassembled bytecode of current Java editor or class file. Unlike javap
, ASMifier on compiled classes allows you to see how any given bytecode could be generated with ASM.
Reflection means the ability for a program to examine, introspect, and modify its own structure and behavior at runtime.[1] However refelction is not sufficient in many cases such as source in non-Java language. ASM framework uses a visitor-based approach to generate bytecode and drive transformations of existing classes.
ASM utilizes Visitor Pattern to accomplish dynamic dispatch on object and its behavior.
The Core package can be logically divided into two major parts:
Bytecode producers, such as a ClassReader or a custom class that can fire the proper sequence of calls to the methods of the above visitor classes.
Bytecode consumers, such as writers (ClassWriter, FieldWriter, MethodWriter, and AnnotationWriter), or any other classes implementing the above visitor interfaces.
ASM ClassReader will call accept()
to allow visitor to walk through itself. We can define our own visitor to override any methods in order to manipulate bytecode we desire to chanage.
We will use Java agent to monitor the main process and use ASM to modify the bytecode at running time.
Let us say we are particularly interested in certain methods in main
BankTransactions bank = new BankTransactions();
for (int i = 0; i < 100; i++) {
String accountId = "account" + i;
bank.login("password", accountId, "Ashley");
bank.unimportantProcessing(accountId);
bank.withdraw(accountId, Double.valueOf(i));
}
We want to keep track of those important behaviors such as login and withdraw. We can use Java annotation to mark those methods for later use.
By setting the premain flag in manifest file, our program will now start from premain function. The premain method acts as a setup hook for the agent. It allows the agent to register a class transformer. When a class transformer is registered with the JVM, that transformer will receive the bytes of every class prior to the class being loaded in the JVM.
ClassReader reader = new ClassReader(classfileBuffer);
ClassWriter writer = new ClassWriter(reader, ClassWriter.COMPUTE_FRAMES);
ClassVisitor visitor = new LogMethodClassVisitor(writer, className);
reader.accept(visitor, 0);
return writer.toByteArray();
ASM palys role here, when visitor visits any methods with annotation @Important
, we record the field related to the method and modify any bytecode as we wish. Here we simply print any important methods’ index of parameter that we care:
System.out.println(methodName);
if (isImportant) {
mv.visitFieldInsn(Opcodes.GETSTATIC, "java/lang/System", "err", "Ljava/io/PrintStream;");
mv.visitLdcInsn(parameterIndexes.toString());
mv.visitMethodInsn(Opcodes.INVOKEVIRTUAL, "java/io/PrintStream", "println", "(Ljava/lang/String;)V", false);
}
How do we know what to put in the visitor methods? As we mentioned above the ASMifier can let us know what ASM code needed to generate the target code.
Notice that the process only happens when we first time load the method, so that the method name will only print once while the parameter index will print many times, since we have already modified its bytecode in the method, whenever the method is invoked the important parameter index will also be printed. That will be a basic instrumentation by using ASM.
$ java -cp workspace/asm-all-5.2.jar -javaagent:ASM.jar BankTransactions
Starting the agent
<init>
main
unimportantProcessing
login
withdraw
[1, 2]
[0, 1]
[1, 2]
[0, 1]
...