netty

网络编程Netty之ByteBuf详解

2022-11-29  本文已影响0人  virtual灬zzZ

Netty中的ByteBuf优势

NIO使用的ByteBuffer有哪些缺点

1: 无法动态扩容,ByteBuffer的长度是固定的,是初始指定的值,不能够再进行扩容了,当写入的内容大于ByteBuffer的容量时,会报越界异常

2.: API使用复杂,当要读取数据时,需要调用buffer.flip()方法,转换为读取模式,如果稍微不注意就可能出现错误,读取不到数据或者读取的数据是错误的

ByteBuf的优势和做了哪些增强

1: API操作起来更加的方便,可以直接写或者直接读

2:支持动态扩容,当写入的数据大于ByteBuf的容量时,会动态扩容,不会报错

3:提供了多种ByteBuf的实现,可以更加灵活的使用

4:提供了高效的零拷贝机制

5:ByteBuf可以内存复用

ByteBuf操作示例
ByteBuf操作

ByteBuf中有三个重要的属性:
1:capacity容量,初始指定的ByteBuf的大小

2:readIndex读取位置,顺序读的时候,记录读取数据的索引值

3:writeIndex写入位置,顺序写的时候,记录写入数据的索引值

ByteBuf常用的方法:
1:getByte和setByte,获取指定索引处的数据,是随机获取的,不会改变readIndex和writeIndex的值

2:read*,顺序读,会改变readIndex的值

3:write*,顺序写,会改变writeIndex的值

4:discardReadBytes,清除读过的内容

5:clear,清除缓冲区

6:搜索操作

7:标记和重置

8:引用计数和释放

简单的Demo示例
/**
 * ByteBuf的使用示例
 */
public class ByteBufDemo {

    public static void main(String[] args) {
        //分配非池化,10个字节的ByteBuf
        ByteBuf buf = Unpooled.buffer(10);

        //看下ByteBuf
        System.out.println("------------------------原始的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //写入内容到ByteBuf
        byte[] bytes = {1, 2, 3, 4, 5};
        buf.writeBytes(bytes);
        System.out.println("------------------------写入内容后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //从ByteBuf中读取内容
        buf.readByte();
        buf.readByte();
        System.out.println("------------------------读取一些内容后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //清除读过的内容
        //把读过的数据清除后,readIndex变为0,writeIndex变为3
        //后面尚未读取的内容,会复制到前面去,把原来的值覆盖掉
        //再次写入时,3,4,5后面的4,5会被覆盖掉
        buf.discardReadBytes();
        System.out.println("------------------------清除读过的数据后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //再次写入内容到ByteBuf
        byte[] bytesO = {6};
        buf.writeBytes(bytesO);
        System.out.println("------------------------再次写入内容后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //清空读和写的索引值
        //readIndex和writeIndex会重置为0,ByteBuf中的内容并不会重置
        buf.clear();
        System.out.println("------------------------清空读和写的索引值后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //再次写入内容到ByteBuf
        byte[] bytes2 = {1, 2, 3};
        buf.writeBytes(bytes2);
        System.out.println("------------------------再次写入内容后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //清空ByteBuf的内容
        //不会重置readIndex和writeIndex
        buf.setZero(0, buf.capacity());
        System.out.println("------------------------清空ByteBuf的内容后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");

        //再次写入超出指定容量的数据到ByteBuf
        //会进行扩容
        byte[] bytes3 = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12};
        buf.writeBytes(bytes3);
        System.out.println("------------------------再次写入超出指定容量的数据后的ByteBuf-------------------------------");
        System.out.println("ByteBuf参数:" + buf.toString());
        System.out.println("ByteBuf中的内容:" + Arrays.toString(buf.array()) + "\n");
    }
}

输出结果:



上面的例子是使用堆内的ByteBuf,下面看下堆外的ByteBuf例子:

 //分配非池化,10个字节的directBuffer
ByteBuf buf = Unpooled.directBuffer(10);

//看下ByteBuf
System.out.println("------------------------原始的ByteBuf-------------------------------");
System.out.println("ByteBuf参数:" + buf.toString());

directBuffer不能够使用array方法,否则会报错:java.lang.UnsupportedOperationException: direct buffer;而且使用ByteBuf是用它底层的分配器分配的,不是new一个出来,下面会具体说下。


上图中,可以看到,readIndex和writeIndex把缓冲区分成了三块,readIndex会小于或者等于writeIndex,这个应该好理解,还没有写到那里,就去读取了,能读取到什么呢。

堆内和堆外内存

socket是操作系统底层提供给上层应用使用的网络通信API,当要去读取或者写入的数据在JVM的堆中,那么就先需要把JVM堆中需要读取的数据拷贝一份到操作系统中,然后socket再去读取,而直接内存的好处是socket可以直接读取,少了拷贝这一步操作。

ByteBuf动态扩容
下面以堆内的ByteBuf为例,查看源码,分析ByteBuf的动态扩容:
动态扩容肯定是写入数据的时候,ByteBuf的容量不够了,才去扩容的,所以需要跟踪下面的代码:

buf.writeBytes(bytes);

跟踪上面的writeBytes,首先进入了ByteBuf这个抽象类中,进入了下面这个抽象方法:

public abstract ByteBuf writeBytes(byte[] src);

它的实现类如下:


进入第一个AbstractByteBuf的方法:

 @Override
 public ByteBuf writeBytes(byte[] src) {
      writeBytes(src, 0, src.length);
      return this;
  }

再次调用了下面的方法:

 @Override
 public ByteBuf writeBytes(byte[] src, int srcIndex, int length) {
      //检查是否可以写入
      ensureWritable(length);
      setBytes(writerIndex, src, srcIndex, length);
      //把当前的写入位置加上写入数据的长度
      writerIndex += length;
      return this;
  }

src是需要写入的数据,length是写入数据的长度
然后会进入ensureWritable方法,传入的参数是:写入数据的长度

@Override
public ByteBuf ensureWritable(int minWritableBytes) {
    //参数校验
    checkPositiveOrZero(minWritableBytes, "minWritableBytes");
    //检查容量是否可以写入这么多数据
    ensureWritable0(minWritableBytes);
    return this;
}

//检查参数是否小于0
public static int checkPositiveOrZero(int i, String name) {
    if (i < 0) {
         throw new IllegalArgumentException(name + ": " + i + " (expected: >= 0)");
     }
     return i;
 }

参数校验完成后会进入ensureWritable0方法:

final void ensureWritable0(int minWritableBytes) {
        //确保缓冲区可以访问
        ensureAccessible();
        //如果写入的数据长度小于等于剩余可写数据的容量,就直接返回
        //就是说,容量足够写入,不需要扩容
        if (minWritableBytes <= writableBytes()) {
            return;
        }
        if (checkBounds) {
            //maxCapacity是int的最大值
            //检查写入的数据长度是否比可以写入的最大容量还要大
            //是的话就抛异常
            if (minWritableBytes > maxCapacity - writerIndex) {
                throw new IndexOutOfBoundsException(String.format(
                        "writerIndex(%d) + minWritableBytes(%d) exceeds maxCapacity(%d): %s",
                        writerIndex, minWritableBytes, maxCapacity, this));
            }
        }

        //正式的扩容方法
        int newCapacity = alloc().calculateNewCapacity(writerIndex + minWritableBytes, maxCapacity);

        //把扩容后的新容量设置进去
        capacity(newCapacity);
}

进入AbstractByteBufAllocator类的扩容方法:

//常量 4M
static final int CALCULATE_THRESHOLD = 1048576 * 4; // 4 MiB page

 @Override
 public int calculateNewCapacity(int minNewCapacity, int maxCapacity) {
        //校验参数
        checkPositiveOrZero(minNewCapacity, "minNewCapacity");
        //minNewCapacity = writerIndex + minWritableBytes
        //已经写入的数据索引加上当前写入的数据长度,就是需要的最小的容量
        //判断是否比最大容量还大,是的话就抛异常
        if (minNewCapacity > maxCapacity) {
            throw new IllegalArgumentException(String.format(
                    "minNewCapacity: %d (expected: not greater than maxCapacity(%d)",
                    minNewCapacity, maxCapacity));
        }
        final int threshold = CALCULATE_THRESHOLD; // 4 MiB page
        //如果需要的最小容量等于4M,就直接返回4M,作为扩容后的容量
        if (minNewCapacity == threshold) {
            return threshold;
        }

        //如果需要的最小容量大于4M,就按照下面的扩容方式扩容
        if (minNewCapacity > threshold) {
            //newCapacity = 15 / 4194304 * 4194304 
            int newCapacity = minNewCapacity / threshold * threshold;
            //如果计算出的容量大于最大容量减去4M,就把最大容量赋值给新的容量
            if (newCapacity > maxCapacity - threshold) {
                newCapacity = maxCapacity;
            } else {
                newCapacity += threshold;
            }
            return newCapacity;
        }

        //如果需要的最小容量小于4M,就按照下面的方式扩容
        int newCapacity = 64;
        while (newCapacity < minNewCapacity) {
            newCapacity <<= 1;
        }

        return Math.min(newCapacity, maxCapacity);
}

再看下capacity方法:
下面的把扩容后的容量放到ByteBuf,就是使用了arraycopy方法

 @Override
    public ByteBuf capacity(int newCapacity) {
        checkNewCapacity(newCapacity);

        int oldCapacity = array.length;
        byte[] oldArray = array;
        if (newCapacity > oldCapacity) {
            byte[] newArray = allocateArray(newCapacity);
            System.arraycopy(oldArray, 0, newArray, 0, oldArray.length);
            setArray(newArray);
            freeArray(oldArray);
        } else if (newCapacity < oldCapacity) {
            byte[] newArray = allocateArray(newCapacity);
            int readerIndex = readerIndex();
            if (readerIndex < newCapacity) {
                int writerIndex = writerIndex();
                if (writerIndex > newCapacity) {
                    writerIndex(writerIndex = newCapacity);
                }
                System.arraycopy(oldArray, readerIndex, newArray, readerIndex, writerIndex - readerIndex);
            } else {
                setIndex(newCapacity, newCapacity);
            }
            setArray(newArray);
            freeArray(oldArray);
        }
        return this;
    }

下面是跟踪的代码步骤:


总结下动态扩容机制:
1:write*方法调用的时候,会通过ensureWritable0方法检查
2:calculateNewCapacity方法是用来计算容量的方法

扩容计算方法:
1:需要的容量没有超过4M,会从64字节开始扩容,每次增加一倍,直到计算出来的容量满足需要的最小容量,假如,当前大小是256,已经写入了200字节,再次写入60字节,需要的最小容量是260字节,那么扩容后的容量是64 * 2 * 2 * 2=512
2:需要的容量超过4M,扩容计算方法为:新容量 = 新容量的最小要求 / 4M * 4M + 4M,假如当前大小是3M,已经写了2M,再写入3M,需要的最小容量是5M,那么扩容后的容量是 5 / 4 * 4 + 4 = 8M

图示1:需要的容量小于4M:


图示2:需要的容量大于4M:


ByteBuf有哪些实现
ByteBuf从3个维度,有8种实现方式:


ByteBuf类图


//堆内
ByteBuf buf = Unpooled.buffer(10);
//堆外
ByteBuf buf = Unpooled.directBuffer(10);

ByteBuf提供了Unpooled非池化的类,可以直接使用,没有提供Pool池化的类,下面追踪源码看下ByteBuf是怎样分配的:

Unpooled.buffer分配方式

首先进入Unpooled类:

private static final ByteBufAllocator ALLOC = UnpooledByteBufAllocator.DEFAULT;
//使用默认的分配器分配堆内buffer
public static ByteBuf buffer(int initialCapacity) {
  return ALLOC.heapBuffer(initialCapacity);
}

下面会进入接口类ByteBufAllocator:

//分配一个指定容量的堆内buf
ByteBuf heapBuffer(int initialCapacity);

然后进入AbstractByteBufAllocator抽象类:

//如果没有指定初始容量,默认的初始容量大小是256
static final int DEFAULT_INITIAL_CAPACITY = 256;
//最大容量,为int的最大值
static final int DEFAULT_MAX_CAPACITY = Integer.MAX_VALUE;

@Override
public ByteBuf heapBuffer(int initialCapacity) {
   return heapBuffer(initialCapacity, DEFAULT_MAX_CAPACITY);
}

 @Override
 public ByteBuf heapBuffer(int initialCapacity, int maxCapacity) {
     //如果初始化的容量是0,最大的容量也是0,就返回一个空的Buf
     if (initialCapacity == 0 && maxCapacity == 0) {
         return emptyBuf;
     }
     validate(initialCapacity, maxCapacity);
     return newHeapBuffer(initialCapacity, maxCapacity);
 }

//校验参数
private static void validate(int initialCapacity, int maxCapacity) {
   //检查参数
   checkPositiveOrZero(initialCapacity, "initialCapacity");
   //如果初始化的容量大于最大容量,就抛异常
    if (initialCapacity > maxCapacity) {
        throw new IllegalArgumentException(String.format(
                "initialCapacity: %d (expected: not greater than maxCapacity(%d)",
                initialCapacity, maxCapacity));
    }
}

然后是newHeapBuffer抽象方法:

protected abstract ByteBuf newHeapBuffer(int initialCapacity, int maxCapacity);

因为这里初始化的是非池化的,所以会进入UnpooledByteBufAllocator类:

@Override
protected ByteBuf newHeapBuffer(int initialCapacity, int maxCapacity) {
     //PlatformDependent.hasUnsafe()是检查当前操作系统是否支持unsafe操作
     //根据支持与否,进入不同的类
     return PlatformDependent.hasUnsafe() ?
             new InstrumentedUnpooledUnsafeHeapByteBuf(this, initialCapacity, maxCapacity) :
             new InstrumentedUnpooledHeapByteBuf(this, initialCapacity, maxCapacity);
 }

支持Unsafe操作的进入下面:

 InstrumentedUnpooledUnsafeHeapByteBuf(UnpooledByteBufAllocator alloc, int initialCapacity, int maxCapacity) {
         super(alloc, initialCapacity, maxCapacity);
 }

不支持Unsafe的进入下面这个:

InstrumentedUnpooledHeapByteBuf(UnpooledByteBufAllocator alloc, int initialCapacity, int maxCapacity) {
         super(alloc, initialCapacity, maxCapacity);
}

现在以支持Unsafe操作往下面走,进入UnpooledUnsafeHeapByteBuf类:

 UnpooledUnsafeHeapByteBuf(ByteBufAllocator alloc, int initialCapacity, int maxCapacity) {
        super(alloc, initialCapacity, maxCapacity);
}

再次调用了父类UnpooledHeapByteBuf:

//分配器
private final ByteBufAllocator alloc;
//byte数组,ByteBuf数据底层就是使用这个存储
byte[] array;

public UnpooledHeapByteBuf(ByteBufAllocator alloc, int initialCapacity, int maxCapacity) {
        super(maxCapacity);

        //检查分配器是否为空
        checkNotNull(alloc, "alloc");
        //如果初始化的容量大于最大容量,就抛异常
        if (initialCapacity > maxCapacity) {
            throw new IllegalArgumentException(String.format(
                    "initialCapacity(%d) > maxCapacity(%d)", initialCapacity, maxCapacity));
        }

        this.alloc = alloc;
        //设置当前的数组是分配之后的数组
        setArray(allocateArray(initialCapacity));
        //初始化readIndex和writeIndex
        setIndex(0, 0);
 }

//分配数组
protected byte[] allocateArray(int initialCapacity) {
     //返回一个具有initialCapacity容量大小的byte数组
     return new byte[initialCapacity];
}

//set数组
 private void setArray(byte[] initialArray) {
        array = initialArray;
        tmpNioBuf = null;
}

AbstractByteBuf类下的setIndex方法:

//初始化readerIndex和writerIndex
@Override
public ByteBuf setIndex(int readerIndex, int writerIndex) {
    if (checkBounds) {
        checkIndexBounds(readerIndex, writerIndex, capacity());
    }
    setIndex0(readerIndex, writerIndex);
    return this;
}

final void setIndex0(int readerIndex, int writerIndex) {
      this.readerIndex = readerIndex;
      this.writerIndex = writerIndex;
}

上面走到AbstractByteBuf后,就分配完了一个非池化、堆内的ByteBuf,下面是追踪的代码:


总结:
可以看到,分配一个非池化、堆内的ByteBuf,它的底层就是byte数组

Unpooled.directBuffer分配方式

首先进入的也是Unpooled类:

public static ByteBuf directBuffer(int initialCapacity) {
     return ALLOC.directBuffer(initialCapacity);
}

然后进入ByteBufAllocator抽象类:

ByteBuf directBuffer(int initialCapacity);

然后到AbstractByteBufAllocator类:

@Override
public ByteBuf directBuffer(int initialCapacity) {
     return directBuffer(initialCapacity, DEFAULT_MAX_CAPACITY);
 }
 
@Override
public ByteBuf directBuffer(int initialCapacity, int maxCapacity) {
     //如果初始化的容量和最大容量都是0,就返回一个空的Buf
     if (initialCapacity == 0 && maxCapacity == 0) {
         return emptyBuf;
     }
     //校验参数
     validate(initialCapacity, maxCapacity);
     return newDirectBuffer(initialCapacity, maxCapacity);
 }
 
protected abstract ByteBuf newDirectBuffer(int initialCapacity, int maxCapacity);

由于分配的也是一个非池化的,所以newDirectBuffer会进入UnpooledByteBufAllocator类中的实现类:

@Override
protected ByteBuf newDirectBuffer(int initialCapacity, int maxCapacity) {
     final ByteBuf buf;
     //同样的,会判断是否支持unsafe操作
     if (PlatformDependent.hasUnsafe()) {
         buf = noCleaner ? new InstrumentedUnpooledUnsafeNoCleanerDirectByteBuf(this, initialCapacity, maxCapacity) :
                 new InstrumentedUnpooledUnsafeDirectByteBuf(this, initialCapacity, maxCapacity);
     } else {
         buf = new InstrumentedUnpooledDirectByteBuf(this, initialCapacity, maxCapacity);
     }
     return disableLeakDetector ? buf : toLeakAwareBuffer(buf);
 }

以InstrumentedUnpooledUnsafeNoCleanerDirectByteBuf为例,后面两个其实也相差不大,进入UnpooledUnsafeNoCleanerDirectByteBuf类的构造方法:

 UnpooledUnsafeNoCleanerDirectByteBuf(ByteBufAllocator alloc, int initialCapacity, int maxCapacity) {
        super(alloc, initialCapacity, maxCapacity);
    }

再次调用的父类UnpooledUnsafeDirectByteBuf:

ByteBuffer buffer;

public UnpooledUnsafeDirectByteBuf(ByteBufAllocator alloc, int initialCapacity, int maxCapacity) {
        super(maxCapacity);
        if (alloc == null) {
            throw new NullPointerException("alloc");
        }
        //校验参数
        checkPositiveOrZero(initialCapacity, "initialCapacity");
        checkPositiveOrZero(maxCapacity, "maxCapacity");
        if (initialCapacity > maxCapacity) {
            throw new IllegalArgumentException(String.format(
                    "initialCapacity(%d) > maxCapacity(%d)", initialCapacity, maxCapacity));
        }

        this.alloc = alloc;
        setByteBuffer(allocateDirect(initialCapacity), false);
}

//分配的是一个NIO中的ByteBuffer
protected ByteBuffer allocateDirect(int initialCapacity) {
      return ByteBuffer.allocateDirect(initialCapacity);
}

final void setByteBuffer(ByteBuffer buffer, boolean tryFree) {
        if (tryFree) {
            ByteBuffer oldBuffer = this.buffer;
            if (oldBuffer != null) {
                if (doNotFree) {
                    doNotFree = false;
                } else {
                    freeDirect(oldBuffer);
                }
            }
        }
        this.buffer = buffer;
        memoryAddress = PlatformDependent.directBufferAddress(buffer);
        tmpNioBuf = null;
        capacity = buffer.remaining();
}

ByteBuffer类下面的allocateDirect:

 public static ByteBuffer allocateDirect(int capacity) {
        return new DirectByteBuffer(capacity);
 }

代码跟踪图:


总结:
分配非池化、堆外的ByteBuf,可以看到底层是NIO的DirectByteBuffer实现的

ByteBufAllocator类图

ByteBuf内存复用

分配池化内存

在上面根据源码知道了怎么去分配非池化内存,那么池化内存要怎么分配呢?看下面的图示:


上面就是分配池化内存的步骤,接下来会根据源码具体分析

内存缓存池

jemalloc内存分配机制

1:内存池中有三大区域,分别是:tiny、small、normal
2:每个区域分了不同大小的格子,每个格子只能缓存对应大小的内存块
3:支持最大的格子内存是32kb,超过这个大小的不能被缓存,只能被释放掉
4:每个类型的格子都有对应的数量:tiny:512个,small:256个,normal:64个,例如tiny区域的每个大小的格子都有512个,如果满了就不会被回收,内存会被释放掉

回收池化内存

分配池化内存的过程
上面分析了分配非池化内存,下面看下怎么分配池化内存:

 ByteBufAllocator allocator = ByteBufAllocator.DEFAULT;
 //分配的内存最大长度为496
 ByteBuf buf1 = allocator.ioBuffer(495);
 System.out.printf("buf1: 0x%X%n", buf1.memoryAddress());
 //此时会被回收到tiny的512b格子中
 buf1.release();

 //从tiny的512b格子去取
 ByteBuf buf2 = allocator.ioBuffer(495);
 System.out.printf("buf2: 0x%X%n", buf2.memoryAddress());
 buf2.release();

先来看下ByteBufAllocator类:

//默认ByteBuf分配器,在ByteBufUtil中初始化
ByteBufAllocator DEFAULT = ByteBufUtil.DEFAULT_ALLOCATOR;

跟踪第一次的allocator.ioBuffer(495)代码,首先进入AbstractByteBufAllocator类:

@Override
public ByteBuf ioBuffer(int initialCapacity) {
    //如果支持Unsafe,就分配堆外内存
    if (PlatformDependent.hasUnsafe()) {
        return directBuffer(initialCapacity);
    }
    //不支持Unsafe,就分配堆内内存
    return heapBuffer(initialCapacity);
}

然后调用了该类下面的directBuffer方法:

@Override
public ByteBuf directBuffer(int initialCapacity) {
     return directBuffer(initialCapacity, DEFAULT_MAX_CAPACITY);
 }

 @Override
 public ByteBuf directBuffer(int initialCapacity, int maxCapacity) {
     //如果初始化的容量和最大容量等于0,就返回一个空的ByteBuf
     if (initialCapacity == 0 && maxCapacity == 0) {
         return emptyBuf;
     }
     validate(initialCapacity, maxCapacity);
     return newDirectBuffer(initialCapacity, maxCapacity);
 }
 //校验参数
 private static void validate(int initialCapacity, int maxCapacity) {
        checkPositiveOrZero(initialCapacity, "initialCapacity");
        if (initialCapacity > maxCapacity) {
            throw new IllegalArgumentException(String.format(
                    "initialCapacity: %d (expected: not greater than maxCapacity(%d)",
                    initialCapacity, maxCapacity));
        }
}

然后会进入池化的ByteBuf分配器PooledByteBufAllocator类,可以实现内存的复用:

// cache sizes  缓存默认值
DEFAULT_TINY_CACHE_SIZE = SystemPropertyUtil.getInt("io.netty.allocator.tinyCacheSize", 512);
DEFAULT_SMALL_CACHE_SIZE = SystemPropertyUtil.getInt("io.netty.allocator.smallCacheSize", 256);
DEFAULT_NORMAL_CACHE_SIZE = SystemPropertyUtil.getInt("io.netty.allocator.normalCacheSize", 64);

 @Override
 protected ByteBuf newDirectBuffer(int initialCapacity, int maxCapacity) {
     //从当前线程中获取cache对象
     PoolThreadCache cache = threadCache.get();
     //从cache中获取Arena
     //Arena可以理解为一个netty提供的实际进行buf分配和管理的工具
     PoolArena<ByteBuffer> directArena = cache.directArena;

     final ByteBuf buf;
     //如果有directArena就分配池化内存
     if (directArena != null) {
         buf = directArena.allocate(cache, initialCapacity, maxCapacity);
     } else { //如果没有directArena,就使用非池化Unpooled
         buf = PlatformDependent.hasUnsafe() ?
                 UnsafeByteBufUtil.newUnsafeDirectByteBuf(this, initialCapacity, maxCapacity) :
                 new UnpooledDirectByteBuf(this, initialCapacity, maxCapacity);
     }

     return toLeakAwareBuffer(buf);
 }

再次跟踪后进入PoolArena类:
可以看到下面有三种类型tiny、small、normal

enum SizeClass {
     Tiny,
     Small,
    Normal
}
PooledByteBuf<T> allocate(PoolThreadCache cache, int reqCapacity, int maxCapacity) {
      //获取一个ByteBuf对象
      PooledByteBuf<T> buf = newByteBuf(maxCapacity);
      //分配内存
      allocate(cache, buf, reqCapacity);
      return buf;
}

@Override
protected PooledByteBuf<ByteBuffer> newByteBuf(int maxCapacity) {
    //如果支持Unsafe,就初始化一个PooledUnsafeDirectByteBuf
    if (HAS_UNSAFE) {
        return PooledUnsafeDirectByteBuf.newInstance(maxCapacity);
    } else { //不支持Unsafe,就初始化一个PooledDirectByteBuf
        return PooledDirectByteBuf.newInstance(maxCapacity);
    }
}

下面进入PooledUnsafeDirectByteBuf类:
从线程回收栈中获取一个buf,如果栈中没有,就会创建一个新的,如果有,就会返回栈中的buf

//调用RECYCLER.get()时,线程栈中没有可以复用的时,会调用newObject方法,此时创建出来的buf是空的
 private static final Recycler<PooledUnsafeDirectByteBuf> RECYCLER = new Recycler<PooledUnsafeDirectByteBuf>() {
      @Override
       protected PooledUnsafeDirectByteBuf newObject(Handle<PooledUnsafeDirectByteBuf> handle) {
           return new PooledUnsafeDirectByteBuf(handle, 0);
       }
};

static PooledUnsafeDirectByteBuf newInstance(int maxCapacity) {
       //RECYCLER,回收机制
       PooledUnsafeDirectByteBuf buf = RECYCLER.get();
       //取出来的可能是之前的buf,使用之前清理一下
       buf.reuse(maxCapacity);
       return buf;
}

然后再次回到PoolArena类中的allocate方法,分配内存:

private void allocate(PoolThreadCache cache, PooledByteBuf<T> buf, final int reqCapacity) {
         //将需要的内存大小计算为2^n
        final int normCapacity = normalizeCapacity(reqCapacity);
        //需要分配的内存是否是tiny或者small类型
        if (isTinyOrSmall(normCapacity)) { // capacity < pageSize
            int tableIdx;
            PoolSubpage<T>[] table;
            boolean tiny = isTiny(normCapacity);
            if (tiny) { // < 512 //分配一个tiny内存
                if (cache.allocateTiny(this, buf, reqCapacity, normCapacity)) {
                    // was able to allocate out of the cache so move on
                    return;
                }
                tableIdx = tinyIdx(normCapacity);
                table = tinySubpagePools;
            } else {
                if (cache.allocateSmall(this, buf, reqCapacity, normCapacity)) {
                    // was able to allocate out of the cache so move on
                    return;
                }
                tableIdx = smallIdx(normCapacity);
                table = smallSubpagePools;
            }

            final PoolSubpage<T> head = table[tableIdx];

          
            synchronized (head) {
                final PoolSubpage<T> s = head.next;
                if (s != head) {
                    assert s.doNotDestroy && s.elemSize == normCapacity;
                    long handle = s.allocate();
                    assert handle >= 0;
                    s.chunk.initBufWithSubpage(buf, null, handle, reqCapacity);
                    incTinySmallAllocation(tiny);
                    return;
                }
            }
            synchronized (this) {
                //分配一块新的内存
                allocateNormal(buf, reqCapacity, normCapacity);
            }

            incTinySmallAllocation(tiny);
            return;
        }
        if (normCapacity <= chunkSize) {
            if (cache.allocateNormal(this, buf, reqCapacity, normCapacity)) {
                // was able to allocate out of the cache so move on
                return;
            }
            synchronized (this) {
                allocateNormal(buf, reqCapacity, normCapacity);
                ++allocationsNormal;
            }
        } else {
            // Huge allocations are never served via the cache so just call allocateHuge
            allocateHuge(buf, reqCapacity);
        }
}

PoolThreadCache类下的allocateTiny方法:

boolean allocateTiny(PoolArena<?> area, PooledByteBuf<?> buf, int reqCapacity, int normCapacity) {
    return allocate(cacheForTiny(area, normCapacity), buf, reqCapacity);
}

//从cache中获取buf
 private MemoryRegionCache<?> cacheForTiny(PoolArena<?> area, int normCapacity) {
        int idx = PoolArena.tinyIdx(normCapacity);
        if (area.isDirect()) {
            return cache(tinySubPageDirectCaches, idx);
        }
        return cache(tinySubPageHeapCaches, idx);
    }

根据需要的容量获取对应的格子,走到PoolArena类下面的tinyIdx方法:

static int tinyIdx(int normCapacity) {
        return normCapacity >>> 4;
}

PoolThreadCache类下的allocate方法,把缓存格子的内存分配到buf

private boolean allocate(MemoryRegionCache<?> cache, PooledByteBuf buf, int reqCapacity) {
      if (cache == null) {
          // no cache found so just return false here
          return false;
      }
      boolean allocated = cache.allocate(buf, reqCapacity);
      if (++ allocations >= freeSweepAllocationThreshold) {
          allocations = 0;
          trim();
      }
      return allocated;
  }

下面是具体跟踪代码的步骤图:


上面的源码是以tiny类型为例,其他两种类型类似,当第一次分配创建了一块新的内存,然后被成功回收到内存缓冲池后,再次分配对应大小的内存,会直接从内存缓冲池中取,不会再次分配一块新的内存了

内存回收的过程

接下来跟踪release()方法,看下内存回收的过程

buf1.release();

第一次进入AbstractReferenceCountedByteBuf类:
Buf的引用计数器,用于内存复用,有一个计数器refCnt,retain()计数器加一,release()计数器减一,
直到计数器为0,才调用deallocate()释放,deallocate()方法由具体的buf自己实现。

 @Override
 public boolean release() {
     return release0(1);
 }
 private boolean release0(int decrement) {
        int rawCnt = nonVolatileRawCnt(), realCnt = toLiveRealCnt(rawCnt, decrement);
        //判断当前buf有没有被引用了,没有的话就调用deallocate
        if (decrement == realCnt) {
            if (refCntUpdater.compareAndSet(this, rawCnt, 1)) {
                deallocate();
                return true;
            }
            return retryRelease0(decrement);
        }
        return releaseNonFinal0(decrement, rawCnt, realCnt);
}

进入PooledByteBuf类:

@Override
protected final void deallocate() {
    if (handle >= 0) {
        final long handle = this.handle;
        //表示当前的buf不在使用任何一块内存区域
        this.handle = -1;
        //设置memory为null
        memory = null;
        //释放buf的内存
        chunk.arena.free(chunk, tmpNioBuf, handle, maxLength, cache);
        tmpNioBuf = null;
        chunk = null;
        //把buf对象放入对象回收栈
        recycle();
    }
}

再次进入PoolArena类:

void free(PoolChunk<T> chunk, ByteBuffer nioBuffer, long handle, int normCapacity, PoolThreadCache cache) {
        //判断是否是unpooled
        if (chunk.unpooled) {
            int size = chunk.chunkSize();
            destroyChunk(chunk);
            activeBytesHuge.add(-size);
            deallocationsHuge.increment();
        } else {
            //判断是哪种类型,tiny、small、normal
            SizeClass sizeClass = sizeClass(normCapacity);
            //放入缓存
            if (cache != null && cache.add(this, chunk, nioBuffer, handle, normCapacity, sizeClass)) {
                // cached so not free it.
                return;
            }

            freeChunk(chunk, handle, sizeClass, nioBuffer);
        }
}

//计算内存区域是哪种类型
private SizeClass sizeClass(int normCapacity) {
        if (!isTinyOrSmall(normCapacity)) {
            return SizeClass.Normal;
        }
        return isTiny(normCapacity) ? SizeClass.Tiny : SizeClass.Small;
}

然后到PoolThreadCache类:

boolean add(PoolArena<?> area, PoolChunk chunk, ByteBuffer nioBuffer,
                long handle, int normCapacity, SizeClass sizeClass) {
     MemoryRegionCache<?> cache = cache(area, normCapacity, sizeClass);
     if (cache == null) {
         return false;
     }
     //加入到缓存队列
     return cache.add(chunk, nioBuffer, handle);
}

 private MemoryRegionCache<?> cache(PoolArena<?> area, int normCapacity, SizeClass sizeClass) {
        //判断是哪种类型,然后把内存回收到哪一块
        switch (sizeClass) {
        case Normal:
            return cacheForNormal(area, normCapacity);
        case Small:
            return cacheForSmall(area, normCapacity);
        case Tiny:
            return cacheForTiny(area, normCapacity);
        default:
            throw new Error();
        }
}

  private MemoryRegionCache<?> cacheForTiny(PoolArena<?> area, int normCapacity) {
        int idx = PoolArena.tinyIdx(normCapacity);
        if (area.isDirect()) {
            return cache(tinySubPageDirectCaches, idx);
        }
        return cache(tinySubPageHeapCaches, idx);
    }

上述跟踪代码步骤图:


ByteBuf零拷贝机制

Netty的零拷贝机制,是一种应用层的实现,和底层的JVM、操作系统内存机制没有过多的关联

几种示例
一:CompositeByteBuf,将多个ByteBuf合并为一个逻辑上的ByteBuf,避免了各个ByteBuf之间的拷贝
public static void test1() {
       ByteBuf buf1 = Unpooled.buffer(4);
       ByteBuf buf2 = Unpooled.buffer(3);
       byte[] bytes1 = {1,2};
       byte[] bytes2 = {3,4,5};
       buf1.writeBytes(bytes1);
       buf2.writeBytes(bytes2);
       CompositeByteBuf byteBuf = Unpooled.compositeBuffer();
       byteBuf = byteBuf.addComponents(true, buf1, buf2);
       System.out.println("byteBuf: " + byteBuf.toString());
}

上面输出结果,ridx是顺序读的读取位置,widx是顺序写的写入位置,cap是新的ByteBuf的容量,components是指新的ByteBuf是由几个ByteBuf组成


二:wrappedBuffer()方法,将byte[]数组包装成ByteBuf对象
public static void test2() {
      byte[] bytes = {1,2,3,4,5};
      ByteBuf buf = Unpooled.wrappedBuffer(bytes);
      System.out.println("buf:" + buf.toString());
}

输出结果中:ridx是顺序读的读取位置,widx是顺序写的写入位置,cap是ByteBuf的容量,新的ByteBuf里存的是数组的引用地址,实质操作的还是原来的数组


三:slice()方法,将一个ByteBuf对象切分成多个ByteBuf对象
public static void test3() {
     ByteBuf buf = Unpooled.wrappedBuffer("hello".getBytes());
     ByteBuf byteBuf = buf.slice(1,2);
     System.out.println("byteBuf:" + byteBuf.toString());
}

输出结果中,可以看到,有两个ByteBuf,其中一个是原有的,新的ByteBuf中存放了原来的ByteBuf的引用地址,另一个是分割后的ByteBuf的引用地址

上一篇下一篇

猜你喜欢

热点阅读