This question already has an answer here:
- What is the native keyword in Java for? 11 answers
If we look at the Java Object class then we can find some of the methods like:
public native int hashCode()
protected native Object clone()
What are these natives and how do these methods work?
Native methods in Java are implemented using the 'Java Native Interface', known as JNI.
Native methods are implemented mostly in C and compiled to native code which runs directly on the machine. This is in contrast to normal methods, which are implemented in Java and compiled to Java byte code, which is executed by the Java Virtual Machine (JVM).
To interface to these methods from Java you need to use the Java Native Interface (JNI).
Native code is mostly needed for accessing low-level stuff. In the case of hashCode this is the address of the object in memory. My guess for clone is that it copies the raw memory from a give object to the cloned one. Other uses of native code are for access to OS features or hardware.
The drawback of using native code is that you lose the safety and security of the JVM, i.e. your program might crash or have security holes due to bugs in the native code.
Minimal example to make things clearer:
Main.java:
Main.c:
Compile and run:
Output:
Tested on Ubuntu 14.04. Also worked with Oracle JDK 1.8.0_45.
Example on GitHub for you to play with.
Interpretation:
It allows you to:
This could be used to:
with the tradeoff of lower portability.
It is also possible for you to call Java from C, but you must first create a JVM in C: How to call Java functions from C++?
Example in the OpenJDK 8
Let's find find where
Object#clone
is defined in jdk8u60-b27.First we find:
which leads us to jdk/src/share/classes/java/lang/Object.java#l212:
Now comes the hard part, finding where clone is amidst all the indirection. The query that helped me was:
which would find either C or C++ files that might implement Object's native methods. It leads us to jdk/share/native/java/lang/Object.c#l47:
which leads us to the
JVM_Clone
symbol:which leads us to hotspot/src/share/vm/prims/jvm.cpp#l580:
After expanding a bunch of macros, we come to the conclusion that this is the definition point.
These methods are either Intrinsic or written outside Java in "native" code, that is, specific to the given machine.
The ones you mention are Intrinsic and part of the JDK but you can also write native methods yourself using the Java Native Interface (JNI). This would normally use C to write the methods, but a lot of other languages, such as python allow you to write methods this way fairly easily. Code is written this way either for performance, or because it needs to access platform specific infrastructure which cannot be done in plain java.
In the case of
hashcode()
, this is implemented by the JVM. This is because often the hashcode will be related to something only the JVM knows. On early JVMs this was related to the object's location in memory - on other JVMs the Object may move in memory, and so a more complicated (but still very fast) scheme may be used.Most native methods are implemented using JNI as mentioned in other answers.
However, performance critical methods such as
Object.hashCode
are typically implemented as intrinsics. When the byte code is compiled into machine code, the Java compiler recognises the method call and inlines appropriate code directly. This is obviously going to be much faster than going through JNI for a trivial method.Many people will claim that
Object.hashCode
will return the address of the object representation in memory. In modern implementations objects actually move within memory. Instead an area of the object header is used to store the value, which may be lazily derived from the memory address at the time that the value is first requested.