Commit Graph

5710 Commits

Author SHA1 Message Date
Norman Maurer
d3ffa1b02b [#1259] Add optimized queue for SCMP pattern and use it in NIO and native transport
This queue also produces less GC then CLQ when make use of OneTimeTask
2014-02-27 13:28:37 +01:00
Trustin Lee
ac5592fc05 Fix checkstyle 2014-02-26 16:18:19 -08:00
Trustin Lee
24fc1b9b37 Fix a bug where HttpObjectDecoder generates two LastHttpContent consecutively
Motivation:
When an HttpResponseDecoder decodes an invalid chunk, a LastHttpContent instance is produced and the decoder enters the 'BAD_MESSAGE' state, which is not supposed to produce a message any further.  However, because HttpObjectDecoder.invalidChunk() did not clear this.message out to null, decodeLast() will produce another LastHttpContent message on a certain situation.

Modification:
Do not forget to null out HttpObjectDecoder.message in invalidChunk(), and add a test case for it.

Result:
No more consecutive LastHttpContent messages produced by HttpObjectDecoder.
2014-02-26 15:53:53 -08:00
Michael Nitschinger
cd6b5def8d Always upstream full memcache messages.
This changeset is related to #2182, which exposes the failure in
the http codec, but the memcache codec works very similar. In addition,
better failure handling in the decoder has been added.
2014-02-24 10:42:17 +01:00
Michael Nitschinger
7d62594cc6 Upgrade JMH to 0.4.1 and make use of @Params. 2014-02-23 16:39:39 +01:00
cmcmaugh
cc4198971e Apply receive timeout to commPort when using RxtxChannel. Part of [#1390] 2014-02-23 16:31:34 +01:00
Norman Maurer
8c68c5a3ae Fix check to clear READ_OP and EPOLLIN. Part of [#2254] 2014-02-22 20:06:52 +01:00
Norman Maurer
b32316b33c [#2254] Correctly handle Channel.read() and ChannelHandlerContext.read()
This includes also when it is called from channelRead(...) and channelReadComplete(...) methods.
2014-02-22 18:48:47 +01:00
Norman Maurer
47fab2bfe8 Directly use memory addresses for gathering writes to reduce gc pressure. Part of [#2239]
This also does factor out some logic of ChannelOutboundBuffer. Mainly we not need nioBuffers() for many
transports and also not need to copy from heap to direct buffer. So this functionality was moved to
NioSocketChannelOutboundBuffer. Also introduce a EpollChannelOutboundBuffer which makes use of
memory addresses for all the writes to reduce GC pressure
2014-02-21 13:37:33 +01:00
Norman Maurer
78db65d0fb [#2254] Fix regression in handling autoRead and Channel.read()
This regression was introduced by e0b39159657c9eb711047bc32367537c4870d467
2014-02-21 09:19:36 +01:00
Trustin Lee
1884a5697c Avoid unnecessary IllegalStateException in ChunkedWriteHandler
Motivation:
ChunkedWriteHandler can sometimes fail to write the last chunk of a ChunkedInput due to an I/O error.  Subsequently, the ChunkedInput's associated promise is marked as failure and the connection is closed.  When the connection is closed, ChunkedWriteHandler attempts to clean up its message queue and to mark their promises as success or failure.  However, because the promise of the ChunkedInput, which was consumed completely yet failed to be written, is already marked as failure, the attempt to mark it as success fails, leading a WARN level log.

Modification:
Use trySuccess() instead of setSuccess() so that the attempt to mark a ChunkedInput as success does not raise an exception even if the promise is already done.

Result:
Fixes #2249
2014-02-20 17:14:25 -08:00
Jeff Pinner
b02531f0aa SPDY: remove SPDY/3 support 2014-02-20 15:00:32 -08:00
Trustin Lee
738d3f77b9 Fix another leak in HttpObjectAggregatorTest by removing redundant code 2014-02-20 14:30:25 -08:00
Trustin Lee
874a66c8fb Fix a memory leak in HttpObjectAggregatorTest 2014-02-20 13:49:35 -08:00
Trustin Lee
2c4aff13c7 Add an operation that resets the state of HttpObjectDecoder
Motivation:

Currently, it is impossible to give a user the full control over what to do in response to the request with 'Expect: 100-continue' header.  Currently, a user have to do one of the following:

- Accept the request and respond with 100 Continue, or
- Send the reject response and close the connection.

.. which means it is impossible to send the reject response and keep the connection alive so that the client sends additional requests.

Modification:

Added a public method called 'reset()' to HttpObjectDecoder so that a user can reset the state of the decoder easily.  Once called, the decoder will assume the next input will be the beginning of a new request.

HttpObjectAggregator now calls `reset()`right after calling 'handleOversizedMessage()' so that the decoder can continue to decode the subsequent request even after the request with 'Expect: 100-continue' header is rejected.

Added relevant unit tests / Minor clean-up

Result:

This commit completes the fix of #2211
2014-02-20 13:41:54 -08:00
Trustin Lee
e278b57489 Added a test for oversized HTTP responses 2014-02-20 11:46:41 -08:00
Trustin Lee
fcc41a62bd Overall clean-up of HttpObjectAggregator / Handle oversized response differently
- Related: #2211
2014-02-20 11:36:56 -08:00
Chris Mowforth
91376263d7 Expose a callback in HttpObjectAggregator to handle oversized messages
- Related: #2211
2014-02-20 11:36:56 -08:00
Norman Maurer
dbb2198839 Fix a regression which could lead to GenericFutureListeners never been notifed. Part of [#2186].
This regression was introduced by commit c97f2d2de00ad74835067cb6f5a62cd4651d1161
2014-02-20 15:13:25 +01:00
Trustin Lee
ddb6e3672c Fix a bug where HttpObjectAggregator doesn't always produce FullHttpMessage
- Fixes #2182
- Always convert an unfull invalid message to a full message
2014-02-19 15:31:53 -08:00
Trustin Lee
396a81c811 Ensure that the last bad chunk is LastHttpChunk 2014-02-19 15:19:35 -08:00
Trustin Lee
b1f27d9aa8 HttpObjectDecoder.invalidChunk() should return LastHttpChunk
because the decoder will enter 'BAD_MESSAGE' state and will not produce any chunk since then.
2014-02-19 14:59:18 -08:00
Norman Maurer
eb143d6d6e Correctly handle remove from epoll and EPOLLRDHUP. Also fix a bug with removing EPOLLIN and EPOLLOUT 2014-02-19 16:34:57 +01:00
Norman Maurer
23374a1854 Ensure native transport is included in all and tarball 2014-02-17 22:46:07 +01:00
Norman Maurer
6bef9c8489 Use optimized write and read calls if memoryAddress is present. Part of [#2239] 2014-02-17 17:28:37 +01:00
Norman Maurer
1c9c797e82 Move marking ChannelPromise for writes uncancellable to addFlush for keep things simple 2014-02-17 16:16:08 +01:00
Trustin Lee
5612472ae6 Fix an inspector warning 2014-02-17 05:25:17 -08:00
Trustin Lee
91da8e228b Overall clean-up in EpollSocketChannel
- Extract writev part from doWrite() for simplicity
- Clearer comments
2014-02-17 05:21:10 -08:00
Norman Maurer
8f3c09ba6b [#2238] Fix bug which caused a SIGSEGV when calling remoteAddress() or localAddress() 2014-02-17 12:00:56 +01:00
Norman Maurer
5a369e60b1 Correctly update the ByteBuffers position on write and writev 2014-02-17 07:30:25 +01:00
Trustin Lee
a0378af850 Fix resource leaks in ByteArrayEncoderTest 2014-02-16 11:50:09 -08:00
Norman Maurer
728417abd5 Use correct writerIndex when read for channel 2014-02-16 20:43:36 +01:00
Norman Maurer
9330172f80 Introduce a native transport for linux using epoll ET
This transport use JNI (C) to directly make use of epoll in Edge-Triggered mode for maximal performance on Linux. Beside this it also support using TCP_CORK and produce less GC then the NIO transport using JDK NIO.
It only builds on linux and skip the build if linux is not used. The transport produce a jar which contains all needed .so files for 32bit and 64 bit. The user only need to include the jar as dependency as usually
to make use of it and use the correct classes.

This includes also some cleanup of @trustin
2014-02-15 22:44:56 +01:00
Trustin Lee
a0e74ff984 Use thread local direct buffer for I/O when the current allocator is unpooled
- Allocating and deallocating a direct buffer for I/O is an expensive
  operation, so we have to at least have a pool of direct buffers if the
  current allocator is not pooled
2014-02-15 11:24:01 -08:00
Trustin Lee
99beb188c4 Fix inefficient loop in QueryStringDecoder.decodeComponent()
- Fixes #2228
- There's no point of looping until the end of the component if at least one '%' or '+' is found.
2014-02-14 13:26:43 -08:00
Michael Nitschinger
33197c7696 Update JMH to 0.3.2 2014-02-14 13:16:13 -08:00
Trustin Lee
fbd4385506 Determine the default allocator from system property
- Add ByteBufAllocator.DEFAULT
- The default allocator is now 'pooled'
2014-02-14 13:04:12 -08:00
Trustin Lee
f0127ec991 Do not warn about Unsafe in Android 2014-02-14 12:06:24 -08:00
Trustin Lee
bb145c0057 Fix wiki link 2014-02-14 12:04:12 -08:00
Trustin Lee
0e4c80b953 Fix IllegalStateException caused by ChannelHandlerAppender 2014-02-13 19:44:40 -08:00
Trustin Lee
abcb39b638 Do not use String.format() for log message generation
- It's slow.
2014-02-13 19:31:17 -08:00
Trustin Lee
c671103289 Deprecate CombinedChannelDuplexHandler in favor of ChannelHandlerAppender
- Fixes #2233
- Add ChannelHandlerAppender
2014-02-13 19:20:56 -08:00
Trustin Lee
7e1db4feb7 Deprecate CombinedChannelDuplexHandler in favor of ChannelHandlerAppender
- Fixes #2233
2014-02-13 19:20:25 -08:00
Trustin Lee
dea5c688fd Fix checkstyle 2014-02-13 18:51:32 -08:00
Trustin Lee
ac70dc4546 Update the version to 4.1.0.Alpha1-SNAPSHOT 2014-02-13 18:32:26 -08:00
Trustin Lee
5226f75cfa Clean up HttpObjectAggregator 2014-02-13 18:25:15 -08:00
Trustin Lee
51349352e2 Fix a bug that CompositeByteBuf.touch() does nothing 2014-02-13 18:24:36 -08:00
Trustin Lee
8837afddf8 Enable a user specify an arbitrary information with ReferenceCounted.touch()
- Related: #2163
- Add ResourceLeakHint to allow a user to provide a meaningful information about the leak when touching it
- DefaultChannelHandlerContext now implements ResourceLeakHint to tell where the message is going.
- Cleaner resource leak report by excluding noisy stack trace elements
2014-02-13 18:16:25 -08:00
Trustin Lee
0235244e55 Touch a ReferenceCounted while it traverses across a pipeline 2014-02-13 18:11:54 -08:00
Trustin Lee
45e70d9935 Add ReferenceCounted.touch() / Add missing retain() overrides
- Fixes #2163
- Inspector warnings
2014-02-13 18:10:11 -08:00