Disruptor 创建过程

为什么80%的码农都做不了架构师？>>>   

1 Disruptor disruptor = new Disruptor<ValueEvent>(ValueEvent.EVENT_FACTORY, ringBufferSize, exec,
                        ProducerType.MULTI, waitStrategy);

public Disruptor(final EventFactory<T> eventFactory,
                     final int ringBufferSize,
                     final Executor executor,
                     final ProducerType producerType,
                     final WaitStrategy waitStrategy)
    {
        this(RingBuffer.create(producerType, eventFactory, ringBufferSize, waitStrategy),
             executor);
    }
　　

public static <E> RingBuffer<E> create(ProducerType    producerType,
                                           EventFactory<E> factory,
                                           int             bufferSize,
                                           WaitStrategy    waitStrategy)
    {
        switch (producerType)
        {
        case SINGLE:
            return createSingleProducer(factory, bufferSize, waitStrategy);
        case MULTI:
            return createMultiProducer(factory, bufferSize, waitStrategy);
        default:
            throw new IllegalStateException(producerType.toString());
        }
    }

　　

 createMultiProducer：

public static <E> RingBuffer<E> createMultiProducer(EventFactory<E> factory,
                                                        int             bufferSize,
                                                        WaitStrategy    waitStrategy)
    {
        MultiProducerSequencer sequencer = new MultiProducerSequencer(bufferSize, waitStrategy);//Sequencer
　　　　//public final class MultiProducerSequencer extends AbstractSequencer
　　　　//public abstract class AbstractSequencer implements Sequencer

        return new RingBuffer<E>(factory, sequencer);
    }
　　

RingBuffer(EventFactory<E> eventFactory,
               Sequencer       sequencer)
    {
        this.sequencer    = sequencer;
        this.bufferSize   = sequencer.getBufferSize();

        if (bufferSize < 1)
        {
            throw new IllegalArgumentException("bufferSize must not be less than 1");
        }
        if (Integer.bitCount(bufferSize) != 1)
        {
            throw new IllegalArgumentException("bufferSize must be a power of 2");
        }

        this.indexMask = bufferSize - 1;
        this.entries   = new Object[sequencer.getBufferSize()];
        fill(eventFactory);
    }
　　　　

private void fill(EventFactory<E> eventFactory)
    {
        for (int i = 0; i < entries.length; i++)
        {
            entries[i] = eventFactory.newInstance();
        }
    }

 

 

2 disruptor.handleEventsWith(eventHandlers);

　　

public EventHandlerGroup<T> handleEventsWith(final EventHandler<T>... handlers)
    {
        return createEventProcessors(new Sequence[0], handlers);
    }

EventHandlerGroup<T> createEventProcessors(final Sequence[] barrierSequences,
                                               final EventHandler<T>[] eventHandlers)
    {
        checkNotStarted();

        final Sequence[] processorSequences = new Sequence[eventHandlers.length];
        final SequenceBarrier barrier = ringBuffer.newBarrier(barrierSequences);//
//ProcessingSequenceBarrier  ProcessingSequenceBarrier implements SequenceBarrier

        for (int i = 0, eventHandlersLength = eventHandlers.length; i < eventHandlersLength; i++)
        {
            final EventHandler<T> eventHandler = eventHandlers[i];

            final BatchEventProcessor<T> batchEventProcessor = new BatchEventProcessor<T>(ringBuffer, barrier, eventHandler);

            if (exceptionHandler != null)
            {
                batchEventProcessor.setExceptionHandler(exceptionHandler);
            }

            consumerRepository.add(batchEventProcessor, eventHandler, barrier);
            processorSequences[i] = batchEventProcessor.getSequence();
        }

        if (processorSequences.length > 0)
        {
            consumerRepository.unMarkEventProcessorsAsEndOfChain(barrierSequences);
        }

        return new EventHandlerGroup<T>(this, consumerRepository, processorSequences);
    }
consumerRepository 与 batchEventProcessor eventHandler SequenceBarrier 关联起来
EventHandlerGroup 与disruptor consumerRepository Sequence 关联起来

 

3 disruptor.start();

public RingBuffer<T> start()
    {
        Sequence[] gatingSequences = consumerRepository.getLastSequenceInChain(true);
        ringBuffer.addGatingSequences(gatingSequences);

        checkOnlyStartedOnce();
        for (ConsumerInfo consumerInfo : consumerRepository)
        {
            consumerInfo.start(executor);//自定义的线程池执行EventProcessorInfo
        }

        return ringBuffer;
    }

class EventProcessorInfo<T> implements ConsumerInfo
    @Override
    public void start(final Executor executor)
    {
        executor.execute(eventprocessor);//batchEventProcessor 
    }

BatchEventProcessor RUN方法
  @Override
    public void run()
    {
        if (!running.compareAndSet(false, true))
        {
            throw new IllegalStateException("Thread is already running");
        }
        sequenceBarrier.clearAlert();

        notifyStart();

        T event = null;
//Concurrent sequence class used for tracking the progress of
// * the ring buffer and event processors.  Support a number
 //* of concurrent operations including CAS and order writes
        long nextSequence = sequence.get() + 1L;//使用CAS取得下一个序列号
        try
        {
            while (true)
            {
                try
                {
                    final long availableSequence = sequenceBarrier.waitFor(nextSequence);//等待策略取得可用序列号

                    if (nextSequence > availableSequence)//
                    {
                        Thread.yield();
                    }

                    while (nextSequence <= availableSequence)
                    {
                        event = dataProvider.get(nextSequence);//dataProvider 就是ringbuffer
                        eventHandler.onEvent(event, nextSequence, nextSequence == availableSequence);//触发自定义事件
                        nextSequence++;
                    }

                    sequence.set(availableSequence);
                }
                catch (final TimeoutException e)
                {
                    notifyTimeout(sequence.get());
                }
                catch (final AlertException ex)
                {
                    if (!running.get())
                    {
                        break;
                    }
                }
                catch (final Throwable ex)
                {
                    exceptionHandler.handleEventException(ex, nextSequence, event);
                    sequence.set(nextSequence);
                    nextSequence++;
                }
            }
        }
        finally
        {
            notifyShutdown();
            running.set(false);
        }
    }

 4 介绍 final long availableSequence = sequenceBarrier.waitFor(nextSequence);//等待可用序列号

　　内容代码

 

    @Override
    public long waitFor(final long sequence)
        throws AlertException, InterruptedException, TimeoutException
    {
        checkAlert();

        long availableSequence = waitStrategy.waitFor(sequence, cursorSequence, dependentSequence, this);

        if (availableSequence < sequence)
        {
            return availableSequence;
        }

        return sequencer.getHighestPublishedSequence(sequence, availableSequence);
    }

 

　　6种等待策略

　　

 

//MultiProducerSequencer   Sequencer   
 @Override
    public long getHighestPublishedSequence(long lowerBound, long availableSequence)
    {
        for (long sequence = lowerBound; sequence <= availableSequence; sequence++)
        {
            if (!isAvailable(sequence))
            {
                return sequence - 1;//实际上实现了循环等待
            }
        }

        return availableSequence;
    }

    @Override
    public boolean isAvailable(long sequence)
    {
        int index = calculateIndex(sequence);
        int flag = calculateAvailabilityFlag(sequence);
        long bufferAddress = (index * SCALE) + BASE;
        return UNSAFE.getIntVolatile(availableBuffer, bufferAddress) == flag;
    }

 

 

 

5 disruptor.publishEvent(eventTranslator, msg);

public <A> void publishEvent(final EventTranslatorOneArg<T, A> eventTranslator, A arg)
    {
        ringBuffer.publishEvent(eventTranslator, arg);
    }

RingBuffer.class
public <A> void publishEvent(EventTranslatorOneArg<E, A> translator, A arg0)
    {
        final long sequence = sequencer.next();
        translateAndPublish(translator, sequence, arg0);
    }

  final long sequence = sequencer.next(); 
//MultiProducerSequencer 
    public long next()
    {
        return next(1);
    }
  public long next(int n)
    {
        if (n < 1)
        {
            throw new IllegalArgumentException("n must be > 0");
        }

        long current;
        long next;

        do
        {
            current = cursor.get();
            next = current + n;

            long wrapPoint = next - bufferSize;
            long cachedGatingSequence = gatingSequenceCache.get();

            if (wrapPoint > cachedGatingSequence || cachedGatingSequence > current)
            {
                long gatingSequence = Util.getMinimumSequence(gatingSequences, current);

                if (wrapPoint > gatingSequence)
                {
                    LockSupport.parkNanos(1); // TODO, should we spin based on the wait strategy?
                    continue;
                }

                gatingSequenceCache.set(gatingSequence);
            }
            else if (cursor.compareAndSet(current, next))
            {
                break;
            }
        }
        while (true);

        return next;
    }

private void translateAndPublish(EventTranslator<E> translator, long sequence)
    {
        try
        {
            translator.translateTo(get(sequence), sequence);// 取位置 调用自定义类设置值 
　　　　　　　　//上面是自定义设置值类    
　　　　// EventTranslatorOneArg<ValueEvent, String> eventTranslator = new EventTranslatorOneArg<ValueEvent, String>() {
      //   @Override
      //  public void translateTo(ValueEvent event, long sequence, String msg) {
      //      event.setValue(msg);
      //  }
    };
        }
        finally
        {
            sequencer.publish(sequence);
        }
    }
//   MultiProducerSequencer
 public void publish(final long sequence)
    {
        setAvailable(sequence);//设置可用
        waitStrategy.signalAllWhenBlocking();
    }
//SingleProducerSequencer
    public void publish(long sequence)
    {
        cursor.set(sequence);//cursor就是 Sequence
        waitStrategy.signalAllWhenBlocking();
    }

 

 

转载于:https://my.oschina.net/sbcagf/blog/782949