Say I have a class
public class Foo {
public void printHi() {
System.out.print("Hi");
}
}
and in some client code I do something like
public static void main() {
Foo foo = new Foo();
(new Thread(() -> {foo.printHi();})).start();
}
and take away the happens-before guarantee for calling Thread Start.
Then is it possible that the Foo reference might not be visible to that thread using it or worse, the method that belongs to that class is not visible, but the Foo reference is visible. I am not sure how method is stored in an object like fields, but this assumes that it is just something in memory belonging to that object, so it might have visibility issues, but that I am not sure of. Can someone also explain that part to me?
I ask this because Foo is immutable, and in the JCIP book Goetz says that
"Immutable objects, on the other hand, can be safely accessed even when synchronization is not used to publish the object reference. For this guarantee of initialization safety to hold, all of the requirements for immutability must be met: unmodi-fiable state, all fields are final, and proper construction" (Goetz, 3.5.2)
However, it doesn't have any final fields, so does it count as if all fields are final? Since no fields = all fields?
There are different ways to get to the same answer.
Your object is immutable, as it bears no state that can be modified.
All of its fields are final, as there is no field that is not final.
There is no possible race condition, as there are no data that could be modified while being accessed.
Even if there were some non-final fields declared in Foo, the invocation of printHi(), which does not read the object’s state, bears no potential data race. Note that this only applies to exact instances of Foo, produced by new Foo(…) expressions, as subclasses could override printHi() to access shared mutable state.
It’s important to emphasize that race condition are about the shared mutable data, not necessarily objects. So if printHi() accesses a shared static variable of a different class, it could produce a data race, even if the Foo instance is immutable and/or properly published. The code invoking foo.printHi() in another thread is safe, if printHi() does not access shared mutable state (or only using proper guarding).
As Elliott Frisch mentioned, a lambda expression behaves like an immutable object anyway, so the code would be safe even without the happens-before relationship of Thread.start() or the immutability of Foo (assuming the Foo instance is not modified afterwards).
foo must be (effectively) final to use here.
Foo foo = null; // <-- for example,
foo = new Foo();
(new Thread(() -> {
foo.printHi(); // <-- compiler error
})).start();
Related
Goetz's Java Concurrency in Practice, page 41, mentions how this reference can escape during construction. A "don't do this" example:
public class ThisEscape {
public ThisEscape(EventSource source) {
source.registerListener(
new EventListener() {
public void onEvent(Event e) {
doSomething(e);
}
});
}
}
Here this is "escaping" via the fact that doSomething(e) refers to the enclosing ThisEscape instance. The situation can be fixed by using static factory methods (first construct the plain object, then register the listener) instead of public constructors (doing all the work). The book goes on:
Publishing an object from within its constructor can publish an incompletely constructed object. This is true even if the publication is the last statement in the constructor. If the this reference escapes during construction, the object is considered not properly constructed.
I don't quite get this. If the publication is the last statement in the constructor, hasn't all the constructing work been done before that? How come is this not valid by then? Apparently there's some voodoo going on after that, but what?
The end of a constructor is a special place in terms of concurrency, with respect to final fields. From section 17.5 of the Java Language Specification:
An object is considered to be
completely initialized when its
constructor finishes. A thread that
can only see a reference to an object
after that object has been completely
initialized is guaranteed to see the
correctly initialized values for that
object's final fields.
The usage model for final fields is a
simple one. Set the final fields for
an object in that object's
constructor. Do not write a reference
to the object being constructed in a
place where another thread can see it
before the object's constructor is
finished. If this is followed, then
when the object is seen by another
thread, that thread will always see
the correctly constructed version of
that object's final fields. It will
also see versions of any object or
array referenced by those final fields
that are at least as up-to-date as the
final fields are.
In other words, your listener could end up seeing final fields with their default values if it examines the object in another thread. This wouldn't happen if listener registration happened after the constructor has completed.
In terms of what's going on, I suspect there's an implicit memory barrier at the very end of a constructor, making sure that all threads "see" the new data; without that memory barrier having been applied, there could be problems.
Another problem arises when you subclass ThisEscape, and the child class invokes this consructor. The implicit this reference in the EventListener would have an incompletely constructed object.
There is a small but finite time between the registerListener ending and the constructor returning. Another thread could use come in at that time and attempt to call doSomething(). If the runtime didn't return straight to your code at that time, the object could be in a invalid state.
I'm not sure of java really but one example I can think of is where possibly the runtime relocates the instance before returning to you.
Its a small chance I grant you.
I read somewhere that even if ConcurrentHashMap is guaranteed to be safe for using in multiple threads it should be declared as final, even private final. My questions are the following:
1) Will CocurrentHashMap still keep thread safety without declaring it as final?
2) The same question about private keyword. Probably it's better to ask more general question - do public/private keywords affect on runtime behavior? I understand their meaning in terms of visibility/usage in internal/external classes but what about meaning in the context of multithreading runtime? I believe code like public ConcurrentHashMap may be incorrect only in coding style terms not in runtime, am I right?
It might be helpful to give a more concrete example of what I was talking about in the comments. Let's say I do something like this:
public class CHMHolder {
private /*non-final*/ CHMHolder instance;
public static CHMHolder getInstance() {
if (instance == null) {
instance = new CHMHolder();
}
return instance;
}
private ConcurrentHashMap<String, String> map = new ConcurrentHashMap<>();
public ConcurrentHashMap<String, String> getMap() {
return map;
}
}
Now, this is not thread-safe for a whole bunch of reasons! But let's say that threadA sees a null value for instance and thus instantiates the CHMHolder, and then threadB, by a happy coincidence, sees that same CHMHolder instance (which is not guaranteed, since there's no synchronization). You would think that threadB sees a non-null CHMHolder.map, right? It might not, since there's no formal happens-before edge between threadA's map = new ... and threadB's return map.
What this means in practice is that something like CHMHolder.getInstance().getMap().isEmpty() could throw a NullPointerException, which would be confusing — after all, getInstance looks like it should always return a non-null CHMHolder, and CHMHolder looks like it should always have a non-null map. Ah, the joys of multithreading!
If map were marked final, then the JLS bit that user2864740 referenced applies. That means that if threadB sees the same instance that threadA sees (which, again, it might not), then it'll also see the map = new... action that threadA did -- that is, it will see the non-null CHM instance. Once it sees that, CHM's internal thread safety will be enough to ensure safe access.
final and private say nothing about the thread-safety (or lack thereof) of the object named by said variable. (They modify the variable, not the object.) Anyway ..
The variable will be consistent across threads if it is a final field:
An object is considered to be completely initialized when its constructor finishes. A thread that can only see a reference to an object after that object has been completely initialized is guaranteed to see the correctly initialized values for that object's final fields.
The actual ConcurrentHashMap object is "thread safe" insofar as the guarantees it makes. In particular, only single method calls/operations are guaranteed and as such using larger synchronization code may be required .. which is easily controlled if the CHM is only accessible from the object that created it.
Using private is normally considered good because it prevents other code from "accidently" accessing a variable (and thus the object it names) when they should not. However, the private modifier does not establish the same happens-before guarantee as the final modifier and is thus orthogonal to thread-safety.
I would like to known if each instance of a class has its own copy of the methods in that class?
Lets say, I have following class MyClass:
public MyClass {
private String s1;
private String s2;
private String method1(String s1){
...
}
private String method2(String s2){
...
}
}
So if two differents users make an instance of MyClass like:
MyClass instanceOfUser1 = new MyClass();
MyClass instanceOfUser2 = new MyClass();
Does know each user have in his thread a copy of the methods of MyClass? If yes, the instance variables are then thread-safe, as long as only the instance methods manipulate them, right?
I am asking this question because I often read that instance variables are not thread-safe. And I can not see why it should be like that, when each user gets an instance by calling the new operator?
Each object gets its own copy of the class's instance variables - it's static variables that are shared between all instances of a class. The reason that instance variables are not necessarily thread-safe is that they might be simultaneously modified by multiple threads calling unsynchronized instance methods.
class Example {
private int instanceVariable = 0;
public void increment() {
instanceVariable++;
}
}
Now if two different threads call increment at the same then you've got a data race - instanceVariable might increment by 1 or 2 at the end of the two methods returning. You could eliminate this data race by adding the synchronized keyword to increment, or using an AtomicInteger instead of an int, etc, but the point is that just because each object gets its own copy of the class's instance variables does not necessarily mean that the variables are accessed in a thread-safe manner - this depends on the class's methods. (The exception is final immutable variables, which can't be accessed in a thread-unsafe manner, short of something goofy like a serialization hack.)
Issues with multi-threading arise primarily with static variables and instances of a class being accessed at the same time.
You shouldn't worry about methods in the class but more about the fields (meaning scoped at the class level). If multiple references to an instance of a class exist, different execution paths may attempt to access the instance at the same time, causing unintended consequences such as race conditions.
A class is basically a blueprint for making an instance of an object. When the object is instantiated it receives a spot in memory that is accessed by a reference. If more than one thread has a handle to this reference it can cause occurrences where the instance is accessed simultaneously, this will cause fields to be manipulated by both threads.
'Instance Variables are not thread safe' - this statement depends on the context.
It is true, if for example you are talking about Servlets. It is because, Servlets create only one instance and multiple threads access it. So in that case Instance Variables are not thread safe.
In the above simplified case, if you are creating new instance for each thread, then your instance variables are thread safe.
Hope this answers your question
A method is nothing but a set of instructions. Whichever thread calls the method, get a copy of those instructions. After that the execution begins. The method may use local variables which are method and thread-scoped, or it may use shared resources, like static resources, shared objects or other resources, which are visible across threads.
Each instance has its own set of instance variables. How would you detect whether every instance had a distinct "copy" of the methods? Wouldn't the difference be visible only by examining the state of the instance variables?
In fact, no, there is only one copy of the method, meaning the set of instructions executed when the method is invoked. But, when executing, an instance method can refer to the instance on which it's being invoked with the reserved identifier this. The this identifier refers to the current instance. If you don't qualify an instance variable (or method) with something else, this is implied.
For example,
final class Example {
private boolean flag;
public void setFlag(boolean value) {
this.flag = value;
}
public void setAnotherFlag(Example friend) {
friend.flag = this.flag;
}
}
There's only one copy of the bytes that make up the VM instructions for the setFlag() and setAnotherFlag() methods. But when they are invoked, this is set to the instance upon which the invocation occurred. Because this is implied for an unqualified variable, you could delete all the references to this in the example, and it would still function exactly the same.
However, if a variable is qualified, like friend.flag above, the variables of another instance can be referenced. This is how you can get into trouble in a multi-threaded program. But, as long as an object doesn't "escape" from one thread to be visible to others, there's nothing to worry about.
There are many situations in which an instance may be accessible from multiple classes. For example, if your instance is a static variable in another class, then all threads would share that instance, and you can get into big trouble that way. That's just the first way that pops into my mind...
Currently I'm reading "Java concurrency in practice", which contains this sentence:
Since the action of a thread accessing a stateless object can't affect the correctness of operations on other threads, stateless objects are thread-safe.
So, what is stateless object?
Stateless object is an instance of a class without instance fields (instance variables). The class may have fields, but they are compile-time constants (static final).
A very much related term is immutable. Immutable objects may have state, but it does not change when a method is invoked (method invocations do not assign new values to fields). These objects are also thread-safe.
If the object doesn't have any instance fields, it it stateless. Also it can be stateless if it has some fields, but their values are known and don't change.
This is a stateless object:
class Stateless {
void test() {
System.out.println("Test!");
}
}
This is also a stateless object:
class Stateless {
//No static modifier because we're talking about the object itself
final String TEST = "Test!";
void test() {
System.out.println(TEST);
}
}
This object has state, so it is not stateless. However, it has its state set only once, and it doesn't change later, this type of objects is called immutable:
class Immutable {
final String testString;
Immutable(String testString) {
this.testString = testString;
}
void test() {
System.out.println(testString);
}
}
The concept of stateless object is highly coupled with concept of side effects.
Shortly, that is the object that has no fields underneath which could have different values, dependently on different order of method calls.
In simple terms state of object means value of internal variables in that object.
Stateful - state of object can be changed, means internal values off member variables of that object can be changed
How values can be changed?
By setting the value.
When can you set that value?
When the variable is not final..
So, to make the class stateless, make the variable final, so that the value of that variable can't be changed neither in setter not in another method. It can be used only for computing.
An object without state, like instance variables that can change and vary depending on what has already happened to the object
Stateless: it has no fields and references no fields from other classes.
The state for a particular computation exists solely in local variables that are stored on the thread’s stack and are accessible only to the executing thread.
One thread accessing a method/class cannot influence the result of another thread accessing the same method/class; because the two threads do not share state, it is as if they were accessing different instances.
Since the actions of a thread accessing a stateless object cannot
affect the correctness of operations in other threads, Stateless objects are threadsafe.
A stateless object is an object that doesn't have any internal state (internal variable)
Just a clarification.
You can consider your class as stateless in the way that is stated before, even when it has an instance variable as far as this variable is final AND immutable.
If the instance variable is just final but mutable, a List of Strings in example, yes the variable's reference can not be changed but the contents of the List and thus the state of the class can be changed.
If you can not change any parameter or value etc. of an object, after its creation, then that object is thread-safe.
An objects that have absolutely no state then there is no problem with reusing them
at this point the question is: if they have absolutely no state why not make all the methods static and never create one at all?
Goetz's Java Concurrency in Practice, page 41, mentions how this reference can escape during construction. A "don't do this" example:
public class ThisEscape {
public ThisEscape(EventSource source) {
source.registerListener(
new EventListener() {
public void onEvent(Event e) {
doSomething(e);
}
});
}
}
Here this is "escaping" via the fact that doSomething(e) refers to the enclosing ThisEscape instance. The situation can be fixed by using static factory methods (first construct the plain object, then register the listener) instead of public constructors (doing all the work). The book goes on:
Publishing an object from within its constructor can publish an incompletely constructed object. This is true even if the publication is the last statement in the constructor. If the this reference escapes during construction, the object is considered not properly constructed.
I don't quite get this. If the publication is the last statement in the constructor, hasn't all the constructing work been done before that? How come is this not valid by then? Apparently there's some voodoo going on after that, but what?
The end of a constructor is a special place in terms of concurrency, with respect to final fields. From section 17.5 of the Java Language Specification:
An object is considered to be
completely initialized when its
constructor finishes. A thread that
can only see a reference to an object
after that object has been completely
initialized is guaranteed to see the
correctly initialized values for that
object's final fields.
The usage model for final fields is a
simple one. Set the final fields for
an object in that object's
constructor. Do not write a reference
to the object being constructed in a
place where another thread can see it
before the object's constructor is
finished. If this is followed, then
when the object is seen by another
thread, that thread will always see
the correctly constructed version of
that object's final fields. It will
also see versions of any object or
array referenced by those final fields
that are at least as up-to-date as the
final fields are.
In other words, your listener could end up seeing final fields with their default values if it examines the object in another thread. This wouldn't happen if listener registration happened after the constructor has completed.
In terms of what's going on, I suspect there's an implicit memory barrier at the very end of a constructor, making sure that all threads "see" the new data; without that memory barrier having been applied, there could be problems.
Another problem arises when you subclass ThisEscape, and the child class invokes this consructor. The implicit this reference in the EventListener would have an incompletely constructed object.
There is a small but finite time between the registerListener ending and the constructor returning. Another thread could use come in at that time and attempt to call doSomething(). If the runtime didn't return straight to your code at that time, the object could be in a invalid state.
I'm not sure of java really but one example I can think of is where possibly the runtime relocates the instance before returning to you.
Its a small chance I grant you.