Motivation:
Thought there may be a bug so added a testcase to verify everything works as expected.
Modifications:
Added testcase
Result:
More test-coverage.
Motiviation:
We used ReferenceCountUtil.releaseLater(...) in our tests which simplifies a bit the releasing of ReferenceCounted objects. The problem with this is that while it simplifies stuff it increase memory usage a lot as memory may not be freed up in a timely manner.
Modifications:
- Deprecate releaseLater(...)
- Remove usage of releaseLater(...) in tests.
Result:
Less memory needed to build netty while running the tests.
Motivation:
Netty provides a adaptor from ByteBuf to Java's InputStream interface. The JDK Stream interfaces have an explicit lifetime because they implement the Closable interface. This lifetime may be differnt than the ByteBuf which is wrapped, and controlled by the interface which accepts the JDK Stream. However Netty's ByteBufInputStream currently does not take reference count ownership of the underlying ByteBuf. There may be no way for existing classes which only accept the InputStream interface to communicate when they are done with the stream, other than calling close(). This means that when the stream is closed it may be appropriate to release the underlying ByteBuf, as the ownership of the underlying ByteBuf resource may be transferred to the Java Stream.
Motivation:
- ByteBufInputStream.close() supports taking reference count ownership of the underyling ByteBuf
Result:
ByteBufInputStream can assume reference count ownership so the underlying ByteBuf can be cleaned up when the stream is closed.
Motivation:
ObjectOutputStream uses a Channel Attribute to cache a ObjectOutputStream which is backed by a ByteBuf that may be released after an object is encoded and the underlying buffer is written to the channel. On subsequent encode operations the cached ObjectOutputStream will be invalid and lead to a reference count exception.
Modifications:
- CompatibleObjectEncoder should not cache a ObjectOutputStream.
Result:
CompatibleObjectEncoder doesn't use a cached object backed by a released ByteBuf.
Motivation:
JCTools supports both non-unsafe, unsafe versions of queues and JDK6 which allows us to shade the library in netty-common allowing it to stay "zero dependency".
Modifications:
- Remove copy paste JCTools code and shade the library (dependencies that are shaded should be removed from the <dependencies> section of the generated POM).
- Remove usage of OneTimeTask and remove it all together.
Result:
Less code to maintain and easier to update JCTools and less GC pressure as the queue implementation nt creates so much garbage
Motivation:
We should ensure we null out the cumulation buffer before we fire it through the pipleine in handlerRemoved(...) as in theory it could be possible that another method is triggered as result of the fireChannelRead(...) or fireChannelReadComplete() that will try to access the cumulation.
Modifications:
Null out cumulation buffer early in handlerRemoved(...)
Result:
No possible to access the cumulation buffer that was already handed over.
Motivation:
At the moment the user is responsible to increase the writer index of the composite buffer when a new component is added. We should add some methods that handle this for the user as this is the most popular usage of the composite buffer.
Modifications:
Add new methods that autoamtically increase the writerIndex when buffers are added.
Result:
Easier usage of CompositeByteBuf.
Motivation:
99dfc9ea799348430a1c25776ce30a95bc10a1ff introduced some code that will more frequently try to forward messages out of the list of decoded messages to reduce latency and memory footprint. Unfortunally this has the side-effect that RecycleableArrayList.clear() will be called more often and so introduce some overhead as ArrayList will null out the array on each call.
Modifications:
- Introduce a CodecOutputList which allows to not null out the array until we recycle it and also allows to access internal array with extra range checks.
- Add benchmark that add elements to different List implementations and clear them
Result:
Less overhead when decode / encode messages.
Benchmark (elements) Mode Cnt Score Error Units
CodecOutputListBenchmark.arrayList 1 thrpt 20 24853764.609 ± 161582.376 ops/s
CodecOutputListBenchmark.arrayList 4 thrpt 20 17310636.508 ± 930517.403 ops/s
CodecOutputListBenchmark.codecOutList 1 thrpt 20 26670751.661 ± 587812.655 ops/s
CodecOutputListBenchmark.codecOutList 4 thrpt 20 25166421.089 ± 166945.599 ops/s
CodecOutputListBenchmark.recyclableArrayList 1 thrpt 20 24565992.626 ± 210017.290 ops/s
CodecOutputListBenchmark.recyclableArrayList 4 thrpt 20 18477881.775 ± 157003.777 ops/s
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 246.748 sec - in io.netty.handler.codec.CodecOutputListBenchmark
Motivation:
b112673554bafc1eccfd43913a3e8605337dd7fb added ChannelInputShutdownEvent support to ByteToMessageDecoder but missed updating the code for ReplayingDecoder. This has the effect:
- If a ChannelInputShutdownEvent is fired ByteToMessageDecoder (the super-class of ReplayingDecoder) will call the channelInputClosed(...) method which will pass the incorrect buffer to the decode method of ReplayingDecoder.
Modifications:
Share more code between ByteToMessageDEcoder and ReplayingDecoder and so also support ChannelInputShutdownEvent correctly in ReplayingDecoder
Result:
ChannelInputShutdownEvent is corrrectly handle in ReplayingDecoder as well.
We need to check if this handler was removed before continuing with decoding.
If it was removed, it is not safe to continue to operate on the buffer. This was already fixed for ByteToMessageDecoder in 4cdbe3928424b5b38695967c0cc1062dccf1a83c but missed for ReplayingDecoder.
Modifications:
Check if decoder was removed after fire messages through the pipeline.
Result:
No illegal buffer access when decoder was removed.
Motivation:
We use ByteBuf.readBytes(int) in various places where we could either remove it completely or use readSlice(int).retain().
Modifications:
- Remove ByteBuf.readBytes(int) when possible or replace by readSlice(int).retain().
Result:
Faster code.
Motivation:
If the input buffer is empty we should not have decodeLast(...) call decode(...) as the user may not expect this.
Modifications:
- Not call decode(...) in decodeLast(...) if the input buffer is empty.
- Add testcases.
Result:
decodeLast(...) will not call decode(...) if input buffer is empty.
Motivation:
Each call of ByteBuf.getByte(int) method does boundary checking. This can be eliminated by using ByteBuf.forEachByte(ByteProcessor) method and ByteProcessor.FIND_LF processor.
Modifications:
Find end of line with ByteProcessor.FIND_LF
Result:
A little better performance of LineBasedFrameDecoder.
Motivation:
ByteToMessageDecoder must ensure that read components of the CompositeByteBuf can be discard by default when discardSomeReadBytes() is called. This may not be the case before as because of the default maxNumComponents that will cause consolidation.
Modifications:
Ensure we not do any consolidation to actually be abel to discard read components
Result:
Less memory usage and allocations.
Motivation:
b714297a44ced34643fa60ca854f1880d02ba649 introduced ChannelInputShutdownEvent support for HttpObjectDecoder. However this should have been added to the super class ByteToMessageDecoder, and ByteToMessageDecoder should not propegate a channelInactive event through the pipeline in this case.
Modifications:
- Move the ChannelInputShutdownEvent handling from HttpObjectDecoder to ByteToMessageDecoder
- ByteToMessageDecoder doesn't call ctx.fireChannelInactive() on ChannelInputShutdownEvent
Result:
Half closed events are treated more generically, and don't get translated into a channelInactive pipeline event.
Motivation:
We not correctly added newlines if the src data needed to be padded. This regression was introduced by '63426fc3ed083513c07a58b45381f5c10dd47061'
Modifications:
- Correctly handling newlines
- Add unit test that proves the fix.
Result:
No more invalid base64 encoded data.
Motivation:
According to https://github.com/google/snappy/blob/master/format_description.txt#L55 , Snappy.decodeLiteral should handle the cases of 60, 61, 62 and 63. However right now it processes 64 instead of 63. I believe it's a typo since `tag >> 2 & 0x3F` must be less than 64.
Modifications:
Use the correct value 63.
Result:
Snappy.decodeLiteral handles the correct case.
Motivation:
`JdkZlibDecoder` is available since Netty 4.0.8 and works with JDK7+.
However, `io.netty.noJdkZlibDecoder` System prop evaluation always defaults to
true, causing Netty to always use JZLib when decompressing on the
client side when the property insn't explictly set to `false`.
Modifications:
Default to `false` instead of `true` when JDK7+.
Result:
JZLib optional as expected on JDK7+.
Motivation:
We need to check if this handler was removed before continuing with decoding.
If it was removed, it is not safe to continue to operate on the buffer.
Modifications:
Check if decoder was removed after fire messages through the pipeline.
Result:
No illegal buffer access when decoder was removed.
Motivation:
We should not use Unpooled to allocate buffers for performance reasons.
Modifications:
Allow to pass in ByteBufAllocate which is used to allocate buffers or use the allocate of the src buffer.
Result:
Better performance if the PooledByteBufAllocator is used.
Motivation:
We need to ensure we not add a newline if the Base64 encoded buffer ends directly on the MAX_LINE_LENGTH. If we miss to do so this produce invalid data.
Because of this bug OpenSslServerContext and OpenSslClientContext may fail to load a cert.
Modifications:
- Only add NEW_LINE if we not are on the end of the dst buffer.
- Add unit test
Result:
Correct result in all cases
Motivation:
We should use OneTimeTask where possible to reduce object creation.
Modifications:
Replace Runnable with OneTimeTask
Result:
Less object creation
Motivation:
At the moment we only forward decoded messages that were added the out List once the full decode loop was completed. This has the affect that resources may not be released as fast as possible and as an application may incounter higher latency if the user triggeres a writeAndFlush(...) as a result of the decoded messages.
Modifications:
- forward decoded messages after each decode call
Result:
Forwarding decoded messages through the pipeline in a more eager fashion.
Motivation:
If a remote peer writes fast enough it may take a long time to have fireChannelReadComplete(...) triggered. Because of this we need to take special care and ensure we try to discard some bytes if channelRead(...) is called to often in ByteToMessageDecoder.
Modifications:
- Add ByteToMessageDecoder.setDiscardAfterReads(...) which allows to set the number of reads after which we try to discard the read bytes
- Use default value of 16 for max reads.
Result:
No risk of OOME.
Motivation:
We missed to correctly implement the handlerRemoved(...) / channelInactive(...) and channelReadComplete(...) method, this leaded to multiple problems:
- Missed to forward bytes when the codec is removed from the pipeline
- Missed to call decodeLast(...) once the Channel goes in active
- No correct handling of channelReadComplete that could lead to grow of cumulation buffer.
Modifications:
- Correctly implement methods and forward to the internal ByteToMessageDecoder
- Add unit test.
Result:
Correct behaviour
Motivation:
Sometimes it is useful to detect if a ByteBuf contains a HAProxy header, for example if you want to write something like the PortUnification example.
Modifications:
- Add ProtocolDetectionResult which can be used as a return type for detecting different protocol.
- Add new method which allows to detect HA Proxy messages.
Result:
Easier to detect protocol.
Motivation:
We are currently doing a memory cop to extract the frame in LengthFieldBasedFrameDecoder which can be eliminated.
Modifications:
Use buffer.slice(...).retain() to eliminate the memory copy.
Result:
Better performance.
Motivation:
The LineBasedFrameDecoder discardedBytes counting different compare to
DelimiterBasedFrameDecoder.
Modifications:
Add plus sign
Result:
DiscardedBytes counting correctly
Motivation:
Our automatically handling of non-auto-read failed because it not detected the need of calling read again by itself if nothing was decoded. Beside this handling of non-auto-read never worked for SslHandler as it always triggered a read even if it decoded a message and auto-read was false.
This fixes [#3529] and [#3587].
Modifications:
- Implement handling of calling read when nothing was decoded (with non-auto-read) to ByteToMessageDecoder again
- Correctly respect non-auto-read by SslHandler
Result:
No more stales and correctly respecting of non-auto-read by SslHandler.
Motivation:
The ReplayingDecoderBuffer does not match the naming scheme we use for ByteBuf types.
Modifications:
Rename to ReplayingDecoderByteBuf to match naming scheme
Result:
Consistent naming
Motivation:
While the LengthFieldBasedFrameDecoder supports a byte order the LengthFieldPrepender does not.
That means that I can simply add a LengthFieldBasedFrameDecoder with ByteOrder.LITTLE_ENDIAN to my pipeline
but have to write my own Encoder to write length fields in little endian byte order.
Modifications:
Added a constructor that takes a byte order and all other parameters.
All other constructors delegate to this one with ByteOrder.BIG_ENDIAN.
LengthFieldPrepender.encode() uses this byte order to write the length field.
Result:
LengthFieldPrepender will write the length field in the defined byte order.
Motivation:
This will avoid one unncessary method invokation which will slightly improve performance.
Modifications:
Instead of calling isReadable we just check for the value of readableBytes()
Result:
Nothing functionally speaking change.
Related:
- 8b2fb2b985cd969719f23da689eb3dc67282070a
Motivation:
The commit mentioned above introduced a regression where
channelReadComplete() event is swallowed by a handler which was added
dynamically.
Modifications:
Do not suppress channelReadComplete() if the current handler's
channelRead() method was not invoked at all, so that a just-added
handler does not suppress channelReadComplete().
Result:
Regression is gone, and channelReadComplete() is invoked when necessary.
Motivation:
The JdkZlibDecoder and JZlibDecoder call isReadable and readableBytes in the same method. There is an opportunity to reduce the number of methods calls to just use readableBytes. JdkZlibDecoder reads from a ByteBuf with an absolute index instead of using readerIndex()
Modifications:
- Use readableBytes where isReadable was used
- Correct absolute ByteBuf index to be relative to readerIndex()
Result:
Less method calls duplicating work and preventing an index out of bounds exception.
Motivation:
ctx.fireChannelReadComplete() should only be called if something is produced during a channelRead(...) operation. Also we must ensure that it will be called
if channelRead(...) produced something at some point as channelRead(...) maybe called multiple times by the transport before channelReadComplete(...) is called.
Modifications:
- Ensure channelReadComplete(...) only triggers ctx.fireChannelReadComplete() when a previous channelRead(...) call produced a message
- Ensure read() is called of more data is needed
Result:
Correct semantic with channelReadComplete(...) events and also ensure no stales
Motivation:
Currently when there are bytes left in the cumulation buffer we do a byte copy to produce the input buffer for the decode method. This can put quite some overhead on the impl.
Modification:
- Use a CompositeByteBuf to eliminate the byte copy.
- Allow to specify if a CompositeBytebug should be used or not as some handlers can only act on one ByteBuffer in an efficient way (like SslHandler :( ).
Result:
Performance improvement as shown in the following benchmark.
Without this patch:
[xxx@xxx ~]$ ./wrk-benchmark
Running 5m test @ http://xxx:8080/plaintext
16 threads and 256 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 20.19ms 38.34ms 1.02s 98.70%
Req/Sec 241.10k 26.50k 303.45k 93.46%
1153994119 requests in 5.00m, 155.84GB read
Requests/sec: 3846702.44
Transfer/sec: 531.93MB
With the patch:
[xxx@xxx ~]$ ./wrk-benchmark
Running 5m test @ http://xxx:8080/plaintext
16 threads and 256 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 17.34ms 27.14ms 877.62ms 98.26%
Req/Sec 252.55k 23.77k 329.50k 87.71%
1209772221 requests in 5.00m, 163.37GB read
Requests/sec: 4032584.22
Transfer/sec: 557.64MB
Motivation:
Found performance issues via FindBugs and PMD.
Modifications:
- Removed unnecessary boxing/unboxing operations in DefaultTextHeaders.convertToInt(CharSequence) and DefaultTextHeaders.convertToLong(CharSequence). A boxed primitive is created from a string, just to extract the unboxed primitive value.
- Added a static modifier for DefaultHttp2Connection.ParentChangedEvent class. This class is an inner class, but does not use its embedded reference to the object which created it. This reference makes the instances of the class larger, and may keep the reference to the creator object alive longer than necessary.
- Added a static compiled Pattern to avoid compile it each time it is used when we need to replace some part of authority.
- Improved using of StringBuilders.
Result:
Performance improvements.
Motivation:
Currently, we only test our ZlibEncoders against our ZlibDecoders. It is
convenient to write such tests, but it does not necessarily guarantee
their correctness. For example, both encoder and decoder might be faulty
even if the tests pass.
Modifications:
Add another test that makes sure that our GZIP encoder generates the
GZIP trailer, using the fact that GZIPInputStream raises an EOFException
when GZIP trailer is missing.
Result:
More coverage for GZIP compression
Motiviation:
The HttpContentEncoder does not account for a EmptyLastHttpContent being provided as input. This is useful in situations where the client is unable to determine if the current content chunk is the last content chunk (i.e. a proxy forwarding content when transfer encoding is chunked).
Modifications:
- HttpContentEncoder should not attempt to compress empty HttpContent objects
Result:
HttpContentEncoder supports a EmptyLastHttpContent to terminate the response.
Motivation:
ExtensionRegistry is a subclass of ExtensionRegistryLite. The ProtobufDecoder
doesn't use the registry directly, it simply passes it through to the Protobuf
API. The Protobuf calls in question are themselves written in terms
ExtensionRegistryLite not ExtensionRegistry.
Modifications:
Require ExtensionRegistryLite instead of ExtensionRegistry in ProtobufDecoder.
Result:
Consumers can use ExtensionRegistryLite with ProtobufDecoder.
