netty5/transport/src/main/java/io/netty/channel/ChannelOutboundBuffer.java

868 lines
30 KiB
Java
Raw Normal View History

Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
/*
* Copyright 2013 The Netty Project
*
* The Netty Project licenses this file to you under the Apache License,
* version 2.0 (the "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
* License for the specific language governing permissions and limitations
* under the License.
*/
package io.netty.channel;
import io.netty.buffer.ByteBuf;
import io.netty.buffer.ByteBufHolder;
2014-02-07 20:52:37 +01:00
import io.netty.buffer.Unpooled;
import io.netty.channel.socket.nio.NioSocketChannel;
2013-07-18 16:26:45 +02:00
import io.netty.util.Recycler;
import io.netty.util.Recycler.Handle;
import io.netty.util.ReferenceCountUtil;
import io.netty.util.concurrent.FastThreadLocal;
import io.netty.util.internal.InternalThreadLocalMap;
import io.netty.util.internal.OneTimeTask;
import io.netty.util.internal.PlatformDependent;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
import io.netty.util.internal.logging.InternalLogger;
import io.netty.util.internal.logging.InternalLoggerFactory;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.PrintStream;
import java.nio.ByteBuffer;
import java.nio.channels.ClosedChannelException;
import java.util.Arrays;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
import java.util.concurrent.atomic.AtomicIntegerFieldUpdater;
import java.util.concurrent.atomic.AtomicLongFieldUpdater;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
/**
* (Transport implementors only) an internal data structure used by {@link AbstractChannel} to store its pending
* outbound write requests.
* <p>
* All methods must be called by a transport implementation from an I/O thread, except the following ones:
* <ul>
* <li>{@link #size()} and {@link #isEmpty()}</li>
* <li>{@link #isWritable()}</li>
* <li>{@link #getUserDefinedWritability(int)} and {@link #setUserDefinedWritability(int, boolean)}</li>
* </ul>
* </p>
*/
public final class ChannelOutboundBuffer {
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
private static final InternalLogger logger = InternalLoggerFactory.getInstance(ChannelOutboundBuffer.class);
private static final FastThreadLocal<ByteBuffer[]> NIO_BUFFERS = new FastThreadLocal<ByteBuffer[]>() {
2013-07-18 16:26:45 +02:00
@Override
protected ByteBuffer[] initialValue() throws Exception {
return new ByteBuffer[1024];
2013-07-18 16:26:45 +02:00
}
};
private final Channel channel;
// Entry(flushedEntry) --> ... Entry(unflushedEntry) --> ... Entry(tailEntry)
//
// The Entry that is the first in the linked-list structure that was flushed
private Entry flushedEntry;
// The Entry which is the first unflushed in the linked-list structure
private Entry unflushedEntry;
// The Entry which represents the tail of the buffer
private Entry tailEntry;
// The number of flushed entries that are not written yet
private int flushed;
private int nioBufferCount;
private long nioBufferSize;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
private boolean inFail;
private static final AtomicLongFieldUpdater<ChannelOutboundBuffer> TOTAL_PENDING_SIZE_UPDATER;
@SuppressWarnings("UnusedDeclaration")
private volatile long totalPendingSize;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
private static final AtomicIntegerFieldUpdater<ChannelOutboundBuffer> UNWRITABLE_UPDATER;
@SuppressWarnings("UnusedDeclaration")
private volatile int unwritable;
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
private final Runnable fireChannelWritabilityChangedTask;
static {
AtomicIntegerFieldUpdater<ChannelOutboundBuffer> unwritableUpdater =
PlatformDependent.newAtomicIntegerFieldUpdater(ChannelOutboundBuffer.class, "unwritable");
if (unwritableUpdater == null) {
unwritableUpdater = AtomicIntegerFieldUpdater.newUpdater(ChannelOutboundBuffer.class, "unwritable");
}
UNWRITABLE_UPDATER = unwritableUpdater;
AtomicLongFieldUpdater<ChannelOutboundBuffer> pendingSizeUpdater =
PlatformDependent.newAtomicLongFieldUpdater(ChannelOutboundBuffer.class, "totalPendingSize");
if (pendingSizeUpdater == null) {
pendingSizeUpdater = AtomicLongFieldUpdater.newUpdater(ChannelOutboundBuffer.class, "totalPendingSize");
}
TOTAL_PENDING_SIZE_UPDATER = pendingSizeUpdater;
}
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
ChannelOutboundBuffer(final AbstractChannel channel) {
this.channel = channel;
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
fireChannelWritabilityChangedTask = new ChannelWritabilityChangedTask(channel);
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
/**
* Add given message to this {@link ChannelOutboundBuffer}. The given {@link ChannelPromise} will be notified once
* the message was written.
*/
public void addMessage(Object msg, int size, ChannelPromise promise) {
Entry entry = Entry.newInstance(msg, size, total(msg), promise);
if (tailEntry == null) {
flushedEntry = null;
tailEntry = entry;
} else {
Entry tail = tailEntry;
tail.next = entry;
tailEntry = entry;
}
if (unflushedEntry == null) {
unflushedEntry = entry;
}
// increment pending bytes after adding message to the unflushed arrays.
// See https://github.com/netty/netty/issues/1619
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
incrementPendingOutboundBytes(size, true);
}
/**
* Add a flush to this {@link ChannelOutboundBuffer}. This means all previous added messages are marked as flushed
* and so you will be able to handle them.
*/
public void addFlush() {
// There is no need to process all entries if there was already a flush before and no new messages
// where added in the meantime.
//
// See https://github.com/netty/netty/issues/2577
Entry entry = unflushedEntry;
if (entry != null) {
if (flushedEntry == null) {
// there is no flushedEntry yet, so start with the entry
flushedEntry = entry;
}
do {
flushed ++;
if (!entry.promise.setUncancellable()) {
// Was cancelled so make sure we free up memory and notify about the freed bytes
int pending = entry.cancel();
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
decrementPendingOutboundBytes(pending, true);
}
entry = entry.next;
} while (entry != null);
// All flushed so reset unflushedEntry
unflushedEntry = null;
}
}
/**
* Increment the pending bytes which will be written at some point.
* This method is thread-safe!
*/
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
void incrementPendingOutboundBytes(long size, boolean notifyWritability) {
if (size == 0) {
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
return;
}
long newWriteBufferSize = TOTAL_PENDING_SIZE_UPDATER.addAndGet(this, size);
if (newWriteBufferSize >= channel.config().getWriteBufferHighWaterMark()) {
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
setUnwritable(notifyWritability);
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
}
/**
* Decrement the pending bytes which will be written at some point.
* This method is thread-safe!
*/
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
void decrementPendingOutboundBytes(long size, boolean notifyWritability) {
if (size == 0) {
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
return;
}
long newWriteBufferSize = TOTAL_PENDING_SIZE_UPDATER.addAndGet(this, -size);
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
if (newWriteBufferSize == 0
|| newWriteBufferSize <= channel.config().getWriteBufferLowWaterMark()) {
setWritable(notifyWritability);
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
}
private static long total(Object msg) {
if (msg instanceof ByteBuf) {
return ((ByteBuf) msg).readableBytes();
}
if (msg instanceof FileRegion) {
return ((FileRegion) msg).count();
}
if (msg instanceof ByteBufHolder) {
return ((ByteBufHolder) msg).content().readableBytes();
}
return -1;
}
/**
* Return the current message to write or {@code null} if nothing was flushed before and so is ready to be written.
*/
public Object current() {
Entry entry = flushedEntry;
if (entry == null) {
return null;
}
return entry.msg;
}
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
/**
* Notify the {@link ChannelPromise} of the current message about writing progress.
*/
public void progress(long amount) {
Entry e = flushedEntry;
assert e != null;
ChannelPromise p = e.promise;
if (p instanceof ChannelProgressivePromise) {
long progress = e.progress + amount;
e.progress = progress;
((ChannelProgressivePromise) p).tryProgress(progress, e.total);
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
}
/**
* Will remove the current message, mark its {@link ChannelPromise} as success and return {@code true}. If no
* flushed message exists at the time this method is called it will return {@code false} to signal that no more
* messages are ready to be handled.
*/
public boolean remove() {
Entry e = flushedEntry;
if (e == null) {
clearNioBuffers();
return false;
}
Object msg = e.msg;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
ChannelPromise promise = e.promise;
int size = e.pendingSize;
removeEntry(e);
2014-02-07 20:52:37 +01:00
if (!e.cancelled) {
// only release message, notify and decrement if it was not canceled before.
ReferenceCountUtil.safeRelease(msg);
2014-02-07 20:52:37 +01:00
safeSuccess(promise);
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
decrementPendingOutboundBytes(size, true);
2014-02-07 20:52:37 +01:00
}
// recycle the entry
e.recycle();
return true;
}
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
/**
* Will remove the current message, mark its {@link ChannelPromise} as failure using the given {@link Throwable}
* and return {@code true}. If no flushed message exists at the time this method is called it will return
* {@code false} to signal that no more messages are ready to be handled.
*/
public boolean remove(Throwable cause) {
return remove0(cause, true);
}
private boolean remove0(Throwable cause, boolean notifyWritability) {
Entry e = flushedEntry;
if (e == null) {
clearNioBuffers();
return false;
}
Object msg = e.msg;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
ChannelPromise promise = e.promise;
int size = e.pendingSize;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
removeEntry(e);
2014-02-07 20:52:37 +01:00
if (!e.cancelled) {
// only release message, fail and decrement if it was not canceled before.
ReferenceCountUtil.safeRelease(msg);
2014-02-07 20:52:37 +01:00
safeFail(promise, cause);
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
decrementPendingOutboundBytes(size, notifyWritability);
2014-02-07 20:52:37 +01:00
}
// recycle the entry
e.recycle();
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
return true;
}
private void removeEntry(Entry e) {
if (-- flushed == 0) {
// processed everything
flushedEntry = null;
if (e == tailEntry) {
tailEntry = null;
unflushedEntry = null;
}
} else {
flushedEntry = e.next;
}
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
/**
* Removes the fully written entries and update the reader index of the partially written entry.
* This operation assumes all messages in this buffer is {@link ByteBuf}.
*/
public void removeBytes(long writtenBytes) {
for (;;) {
Object msg = current();
if (!(msg instanceof ByteBuf)) {
assert writtenBytes == 0;
break;
}
final ByteBuf buf = (ByteBuf) msg;
final int readerIndex = buf.readerIndex();
final int readableBytes = buf.writerIndex() - readerIndex;
if (readableBytes <= writtenBytes) {
if (writtenBytes != 0) {
progress(readableBytes);
writtenBytes -= readableBytes;
}
remove();
} else { // readableBytes > writtenBytes
if (writtenBytes != 0) {
buf.readerIndex(readerIndex + (int) writtenBytes);
progress(writtenBytes);
}
break;
}
}
clearNioBuffers();
}
// Clear all ByteBuffer from the array so these can be GC'ed.
// See https://github.com/netty/netty/issues/3837
private void clearNioBuffers() {
int count = nioBufferCount;
if (count > 0) {
nioBufferCount = 0;
Arrays.fill(NIO_BUFFERS.get(), 0, count, null);
}
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
/**
* Returns an array of direct NIO buffers if the currently pending messages are made of {@link ByteBuf} only.
* {@link #nioBufferCount()} and {@link #nioBufferSize()} will return the number of NIO buffers in the returned
* array and the total number of readable bytes of the NIO buffers respectively.
* <p>
* Note that the returned array is reused and thus should not escape
* {@link AbstractChannel#doWrite(ChannelOutboundBuffer)}.
* Refer to {@link NioSocketChannel#doWrite(ChannelOutboundBuffer)} for an example.
* </p>
*/
public ByteBuffer[] nioBuffers() {
long nioBufferSize = 0;
int nioBufferCount = 0;
final InternalThreadLocalMap threadLocalMap = InternalThreadLocalMap.get();
ByteBuffer[] nioBuffers = NIO_BUFFERS.get(threadLocalMap);
Entry entry = flushedEntry;
while (isFlushedEntry(entry) && entry.msg instanceof ByteBuf) {
if (!entry.cancelled) {
ByteBuf buf = (ByteBuf) entry.msg;
final int readerIndex = buf.readerIndex();
final int readableBytes = buf.writerIndex() - readerIndex;
if (readableBytes > 0) {
if (Integer.MAX_VALUE - readableBytes < nioBufferSize) {
// If the nioBufferSize + readableBytes will overflow an Integer we stop populate the
// ByteBuffer array. This is done as bsd/osx don't allow to write more bytes then
// Integer.MAX_VALUE with one writev(...) call and so will return 'EINVAL', which will
// raise an IOException. On Linux it may work depending on the
// architecture and kernel but to be safe we also enforce the limit here.
// This said writing more the Integer.MAX_VALUE is not a good idea anyway.
//
// See also:
// - https://www.freebsd.org/cgi/man.cgi?query=write&sektion=2
// - http://linux.die.net/man/2/writev
break;
}
nioBufferSize += readableBytes;
int count = entry.count;
if (count == -1) {
//noinspection ConstantValueVariableUse
entry.count = count = buf.nioBufferCount();
}
int neededSpace = nioBufferCount + count;
if (neededSpace > nioBuffers.length) {
nioBuffers = expandNioBufferArray(nioBuffers, neededSpace, nioBufferCount);
NIO_BUFFERS.set(threadLocalMap, nioBuffers);
}
if (count == 1) {
ByteBuffer nioBuf = entry.buf;
if (nioBuf == null) {
// cache ByteBuffer as it may need to create a new ByteBuffer instance if its a
// derived buffer
entry.buf = nioBuf = buf.internalNioBuffer(readerIndex, readableBytes);
}
nioBuffers[nioBufferCount ++] = nioBuf;
} else {
ByteBuffer[] nioBufs = entry.bufs;
if (nioBufs == null) {
// cached ByteBuffers as they may be expensive to create in terms
// of Object allocation
entry.bufs = nioBufs = buf.nioBuffers();
}
nioBufferCount = fillBufferArray(nioBufs, nioBuffers, nioBufferCount);
}
}
}
entry = entry.next;
}
this.nioBufferCount = nioBufferCount;
this.nioBufferSize = nioBufferSize;
return nioBuffers;
}
private static int fillBufferArray(ByteBuffer[] nioBufs, ByteBuffer[] nioBuffers, int nioBufferCount) {
for (ByteBuffer nioBuf: nioBufs) {
if (nioBuf == null) {
break;
}
nioBuffers[nioBufferCount ++] = nioBuf;
}
return nioBufferCount;
}
private static ByteBuffer[] expandNioBufferArray(ByteBuffer[] array, int neededSpace, int size) {
int newCapacity = array.length;
do {
// double capacity until it is big enough
// See https://github.com/netty/netty/issues/1890
newCapacity <<= 1;
if (newCapacity < 0) {
throw new IllegalStateException();
}
} while (neededSpace > newCapacity);
ByteBuffer[] newArray = new ByteBuffer[newCapacity];
System.arraycopy(array, 0, newArray, 0, size);
return newArray;
}
/**
* Returns the number of {@link ByteBuffer} that can be written out of the {@link ByteBuffer} array that was
* obtained via {@link #nioBuffers()}. This method <strong>MUST</strong> be called after {@link #nioBuffers()}
* was called.
*/
public int nioBufferCount() {
return nioBufferCount;
}
/**
* Returns the number of bytes that can be written out of the {@link ByteBuffer} array that was
* obtained via {@link #nioBuffers()}. This method <strong>MUST</strong> be called after {@link #nioBuffers()}
* was called.
*/
public long nioBufferSize() {
return nioBufferSize;
}
/**
* Returns {@code true} if and only if {@linkplain #totalPendingWriteBytes() the total number of pending bytes} did
* not exceed the write watermark of the {@link Channel} and
* no {@linkplain #setUserDefinedWritability(int, boolean) user-defined writability flag} has been set to
* {@code false}.
*/
public boolean isWritable() {
return unwritable == 0;
}
/**
* Returns {@code true} if and only if the user-defined writability flag at the specified index is set to
* {@code true}.
*/
public boolean getUserDefinedWritability(int index) {
return (unwritable & writabilityMask(index)) == 0;
}
/**
* Sets a user-defined writability flag at the specified index.
*/
public void setUserDefinedWritability(int index, boolean writable) {
if (writable) {
setUserDefinedWritability(index);
} else {
clearUserDefinedWritability(index);
}
}
private void setUserDefinedWritability(int index) {
final int mask = ~writabilityMask(index);
for (;;) {
final int oldValue = unwritable;
final int newValue = oldValue & mask;
if (UNWRITABLE_UPDATER.compareAndSet(this, oldValue, newValue)) {
if (oldValue != 0 && newValue == 0) {
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
fireChannelWritabilityChanged();
}
break;
}
}
}
private void clearUserDefinedWritability(int index) {
final int mask = writabilityMask(index);
for (;;) {
final int oldValue = unwritable;
final int newValue = oldValue | mask;
if (UNWRITABLE_UPDATER.compareAndSet(this, oldValue, newValue)) {
if (oldValue == 0 && newValue != 0) {
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
fireChannelWritabilityChanged();
}
break;
}
}
}
private static int writabilityMask(int index) {
if (index < 1 || index > 31) {
throw new IllegalArgumentException("index: " + index + " (expected: 1~31)");
}
return 1 << index;
}
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
private void setWritable(boolean notify) {
for (;;) {
final int oldValue = unwritable;
final int newValue = oldValue & ~1;
if (UNWRITABLE_UPDATER.compareAndSet(this, oldValue, newValue)) {
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
if (notify && oldValue != 0 && newValue == 0) {
fireChannelWritabilityChanged();
}
break;
}
}
}
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
private void setUnwritable(boolean notify) {
for (;;) {
final int oldValue = unwritable;
final int newValue = oldValue | 1;
if (UNWRITABLE_UPDATER.compareAndSet(this, oldValue, newValue)) {
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
if (notify && oldValue == 0 && newValue != 0) {
fireChannelWritabilityChanged();
}
break;
}
}
}
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
private void fireChannelWritabilityChanged() {
// Always invoke it later to prevent re-entrance bug.
// See https://github.com/netty/netty/issues/5028
channel.eventLoop().execute(fireChannelWritabilityChangedTask);
}
/**
* Returns the number of flushed messages in this {@link ChannelOutboundBuffer}.
*/
public int size() {
return flushed;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
/**
* Returns {@code true} if there are flushed messages in this {@link ChannelOutboundBuffer} or {@code false}
* otherwise.
*/
public boolean isEmpty() {
return flushed == 0;
}
void failFlushed(Throwable cause, boolean notify) {
// Make sure that this method does not reenter. A listener added to the current promise can be notified by the
// current thread in the tryFailure() call of the loop below, and the listener can trigger another fail() call
// indirectly (usually by closing the channel.)
//
// See https://github.com/netty/netty/issues/1501
if (inFail) {
return;
}
try {
inFail = true;
for (;;) {
if (!remove0(cause, notify)) {
break;
}
}
} finally {
inFail = false;
}
}
void close(final ClosedChannelException cause) {
if (inFail) {
channel.eventLoop().execute(new OneTimeTask() {
@Override
public void run() {
close(cause);
}
});
return;
}
inFail = true;
if (channel.isOpen()) {
throw new IllegalStateException("close() must be invoked after the channel is closed.");
}
if (!isEmpty()) {
throw new IllegalStateException("close() must be invoked after all flushed writes are handled.");
}
// Release all unflushed messages.
try {
Entry e = unflushedEntry;
while (e != null) {
// Just decrease; do not trigger any events via decrementPendingOutboundBytes()
int size = e.pendingSize;
TOTAL_PENDING_SIZE_UPDATER.addAndGet(this, -size);
2014-02-07 20:52:37 +01:00
if (!e.cancelled) {
ReferenceCountUtil.safeRelease(e.msg);
2014-02-07 20:52:37 +01:00
safeFail(e.promise, cause);
}
e = e.recycleAndGetNext();
}
} finally {
inFail = false;
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}
clearNioBuffers();
}
2013-07-18 03:29:34 +02:00
private static void safeSuccess(ChannelPromise promise) {
if (!(promise instanceof VoidChannelPromise) && !promise.trySuccess()) {
Throwable err = promise.cause();
if (err == null) {
logger.warn("Failed to mark a promise as success because it has succeeded already: {}", promise);
} else {
logger.warn(
"Failed to mark a promise as success because it has failed already: {}, unnotified cause {}",
promise, stackTraceToString(err));
}
}
}
private static void safeFail(ChannelPromise promise, Throwable cause) {
if (!(promise instanceof VoidChannelPromise) && !promise.tryFailure(cause)) {
Throwable err = promise.cause();
if (err == null) {
logger.warn("Failed to mark a promise as failure because it has succeeded already: {}", promise, cause);
} else {
logger.warn(
"Failed to mark a promise as failure because it hass failed already: {}, unnotified cause {}",
promise, stackTraceToString(err), cause);
}
}
}
private static String stackTraceToString(Throwable cause) {
ByteArrayOutputStream out = new ByteArrayOutputStream();
PrintStream pout = new PrintStream(out);
cause.printStackTrace(pout);
pout.flush();
try {
return new String(out.toByteArray());
} finally {
try {
out.close();
} catch (IOException ignore) {
// ignore as should never happen
}
2013-07-18 03:29:34 +02:00
}
}
@Deprecated
2013-08-21 19:28:37 +02:00
public void recycle() {
// NOOP
2013-08-21 19:28:37 +02:00
}
public long totalPendingWriteBytes() {
return totalPendingSize;
}
/**
* Get how many bytes can be written until {@link #isWritable()} returns {@code false}.
* This quantity will always be non-negative. If {@link #isWritable()} is {@code false} then 0.
*/
public long bytesBeforeUnwritable() {
long bytes = channel.config().getWriteBufferHighWaterMark() - totalPendingSize;
// If bytes is negative we know we are not writable, but if bytes is non-negative we have to check writability.
// Note that totalPendingSize and isWritable() use different volatile variables that are not synchronized
// together. totalPendingSize will be updated before isWritable().
if (bytes > 0) {
return isWritable() ? bytes : 0;
}
return 0;
}
/**
* Get how many bytes must be drained from the underlying buffer until {@link #isWritable()} returns {@code true}.
* This quantity will always be non-negative. If {@link #isWritable()} is {@code true} then 0.
*/
public long bytesBeforeWritable() {
long bytes = totalPendingSize - channel.config().getWriteBufferLowWaterMark();
// If bytes is negative we know we are writable, but if bytes is non-negative we have to check writability.
// Note that totalPendingSize and isWritable() use different volatile variables that are not synchronized
// together. totalPendingSize will be updated before isWritable().
if (bytes > 0) {
return isWritable() ? 0 : bytes;
}
return 0;
}
/**
* Call {@link MessageProcessor#processMessage(Object)} for each flushed message
* in this {@link ChannelOutboundBuffer} until {@link MessageProcessor#processMessage(Object)}
* returns {@code false} or there are no more flushed messages to process.
*/
public void forEachFlushedMessage(MessageProcessor processor) throws Exception {
if (processor == null) {
throw new NullPointerException("processor");
}
Entry entry = flushedEntry;
if (entry == null) {
return;
}
do {
if (!entry.cancelled) {
if (!processor.processMessage(entry.msg)) {
return;
}
}
entry = entry.next;
} while (isFlushedEntry(entry));
}
private boolean isFlushedEntry(Entry e) {
return e != null && e != unflushedEntry;
Optimize native transport for gathering writes Motivation: While benchmarking the native transport with gathering writes I noticed that it is quite slow. This is due the fact that we need to do a lot of array copies to get the buffers into the iov array. Modification: Introduce a new class calles IovArray which allows to fill buffers directly in a iov array that can be passed over to JNI without any array copies. This gives a nice optimization in terms of speed when doing gathering writes. Result: Big performance improvement when doing gathering writes. See the included benchmark... Before: [nmaurer@xxx]~% wrk/wrk -H 'Host: localhost' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8' -H 'Connection: keep-alive' -d 120 -c 256 -t 16 --pipeline 256 http://xxx:8080/plaintext Running 2m test @ http://xxx:8080/plaintext 16 threads and 256 connections Thread Stats Avg Stdev Max +/- Stdev Latency 23.44ms 16.37ms 259.57ms 91.77% Req/Sec 181.99k 31.69k 304.60k 78.12% 346544071 requests in 2.00m, 46.48GB read Requests/sec: 2887885.09 Transfer/sec: 396.59MB With this change: [nmaurer@xxx]~% wrk/wrk -H 'Host: localhost' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8' -H 'Connection: keep-alive' -d 120 -c 256 -t 16 --pipeline 256 http://xxx:8080/plaintext Running 2m test @ http://xxx:8080/plaintext 16 threads and 256 connections Thread Stats Avg Stdev Max +/- Stdev Latency 21.93ms 16.33ms 305.73ms 92.34% Req/Sec 194.56k 33.75k 309.33k 77.04% 369617503 requests in 2.00m, 49.57GB read Requests/sec: 3080169.65 Transfer/sec: 423.00MB
2014-07-22 22:27:50 +02:00
}
public interface MessageProcessor {
/**
* Will be called for each flushed message until it either there are no more flushed messages or this
* method returns {@code false}.
*/
boolean processMessage(Object msg) throws Exception;
}
static final class Entry {
private static final Recycler<Entry> RECYCLER = new Recycler<Entry>() {
@Override
protected Entry newObject(Handle handle) {
return new Entry(handle);
}
};
private final Handle handle;
Entry next;
Object msg;
ByteBuffer[] bufs;
ByteBuffer buf;
ChannelPromise promise;
long progress;
long total;
int pendingSize;
int count = -1;
2014-02-07 20:52:37 +01:00
boolean cancelled;
private Entry(Handle handle) {
this.handle = handle;
}
static Entry newInstance(Object msg, int size, long total, ChannelPromise promise) {
Entry entry = RECYCLER.get();
entry.msg = msg;
entry.pendingSize = size;
entry.total = total;
entry.promise = promise;
return entry;
}
int cancel() {
2014-02-07 20:52:37 +01:00
if (!cancelled) {
cancelled = true;
int pSize = pendingSize;
// release message and replace with an empty buffer
ReferenceCountUtil.safeRelease(msg);
2014-02-07 20:52:37 +01:00
msg = Unpooled.EMPTY_BUFFER;
pendingSize = 0;
total = 0;
progress = 0;
bufs = null;
buf = null;
2014-02-07 20:52:37 +01:00
return pSize;
}
return 0;
}
void recycle() {
next = null;
bufs = null;
buf = null;
msg = null;
promise = null;
progress = 0;
total = 0;
pendingSize = 0;
count = -1;
2014-02-07 20:52:37 +01:00
cancelled = false;
RECYCLER.recycle(this, handle);
}
Entry recycleAndGetNext() {
Entry next = this.next;
recycle();
return next;
}
}
[#5028] Fix re-entrance issue with channelWritabilityChanged(...) and write(...) Motivation: When always triggered fireChannelWritabilityChanged() directly when the update the pending bytes in the ChannelOutboundBuffer was made from within the EventLoop. This is problematic as this can cause some re-entrance issue if the user has a custom ChannelOutboundHandler that does multiple writes from within the write(...) method and also has a handler that will intercept the channelWritabilityChanged event and trigger another write when the Channel is writable. This can also easily happen if the user just use a MessageToMessageEncoder subclass and triggers a write from channelWritabilityChanged(). Beside this we also triggered fireChannelWritabilityChanged() too often when a user did a write from outside the EventLoop. In this case we increased the pending bytes of the outboundbuffer before scheduled the actual write and decreased again before the write then takes place. Both of this may trigger a fireChannelWritabilityChanged() event which then may be re-triggered once the actual write ends again in the ChannelOutboundBuffer. The third gotcha was that a user may get multiple events even if the writability of the channel not changed. Modification: - Always invoke the fireChannelWritabilityChanged() later on the EventLoop. - Only trigger the fireChannelWritabilityChanged() if the channel is still active and if the writability of the channel changed. No need to cause events that were already triggered without a real writability change. - when write(...) is called from outside the EventLoop we only increase the pending bytes in the outbound buffer (so that Channel.isWritable() is updated directly) but not cause a fireChannelWritabilityChanged(). The fireChannelWritabilityChanged() is then triggered once the task is picked up by the EventLoop as usual. Result: No more re-entrance possible because of writes from within channelWritabilityChanged(...) method and no events without a real writability change.
2016-04-01 11:45:43 +02:00
private static final class ChannelWritabilityChangedTask implements Runnable {
private final Channel channel;
private boolean writable = true;
ChannelWritabilityChangedTask(Channel channel) {
this.channel = channel;
}
@Override
public void run() {
if (channel.isActive()) {
boolean newWritable = channel.isWritable();
if (writable != newWritable) {
writable = newWritable;
channel.pipeline().fireChannelWritabilityChanged();
}
}
}
}
Revamp the core API to reduce memory footprint and consumption The API changes made so far turned out to increase the memory footprint and consumption while our intention was actually decreasing them. Memory consumption issue: When there are many connections which does not exchange data frequently, the old Netty 4 API spent a lot more memory than 3 because it always allocates per-handler buffer for each connection unless otherwise explicitly stated by a user. In a usual real world load, a client doesn't always send requests without pausing, so the idea of having a buffer whose life cycle if bound to the life cycle of a connection didn't work as expected. Memory footprint issue: The old Netty 4 API decreased overall memory footprint by a great deal in many cases. It was mainly because the old Netty 4 API did not allocate a new buffer and event object for each read. Instead, it created a new buffer for each handler in a pipeline. This works pretty well as long as the number of handlers in a pipeline is only a few. However, for a highly modular application with many handlers which handles connections which lasts for relatively short period, it actually makes the memory footprint issue much worse. Changes: All in all, this is about retaining all the good changes we made in 4 so far such as better thread model and going back to the way how we dealt with message events in 3. To fix the memory consumption/footprint issue mentioned above, we made a hard decision to break the backward compatibility again with the following changes: - Remove MessageBuf - Merge Buf into ByteBuf - Merge ChannelInboundByte/MessageHandler and ChannelStateHandler into ChannelInboundHandler - Similar changes were made to the adapter classes - Merge ChannelOutboundByte/MessageHandler and ChannelOperationHandler into ChannelOutboundHandler - Similar changes were made to the adapter classes - Introduce MessageList which is similar to `MessageEvent` in Netty 3 - Replace inboundBufferUpdated(ctx) with messageReceived(ctx, MessageList) - Replace flush(ctx, promise) with write(ctx, MessageList, promise) - Remove ByteToByteEncoder/Decoder/Codec - Replaced by MessageToByteEncoder<ByteBuf>, ByteToMessageDecoder<ByteBuf>, and ByteMessageCodec<ByteBuf> - Merge EmbeddedByteChannel and EmbeddedMessageChannel into EmbeddedChannel - Add SimpleChannelInboundHandler which is sometimes more useful than ChannelInboundHandlerAdapter - Bring back Channel.isWritable() from Netty 3 - Add ChannelInboundHandler.channelWritabilityChanges() event - Add RecvByteBufAllocator configuration property - Similar to ReceiveBufferSizePredictor in Netty 3 - Some existing configuration properties such as DatagramChannelConfig.receivePacketSize is gone now. - Remove suspend/resumeIntermediaryDeallocation() in ByteBuf This change would have been impossible without @normanmaurer's help. He fixed, ported, and improved many parts of the changes.
2013-05-28 13:40:19 +02:00
}