Commit Graph

9366 Commits

Author SHA1 Message Date
Aleksey Yeschenko
db1e662933 Fix ByteBufChecksum optimisation for CRC32 and Adler32 (#9242)
Motivation:

Because of a simple bug in ByteBufChecksum#updateByteBuffer(Checksum),
ReflectiveByteBufChecksum is never used for CRC32 and Adler32, resulting
in direct ByteBuffers being checksummed byte by byte, which is
undesriable.

Modification:

Fix ByteBufChecksum#updateByteBuffer(Checksum) method to pass the
correct argument to Method#invoke(Checksum, ByteBuffer).

Result:

ReflectiveByteBufChecksum will now be used for Adler32 and CRC32 on
Java8+ and direct ByteBuffers will no longer be checksummed on slow
byte-by-byte basis.
2019-06-16 07:37:52 +02:00
Norman Maurer
acd13e1cf7 Allow to set parent Channel when constructing EmbeddedChannel (#9230)
Motivation:

Sometimes it is beneficial to be able to set a parent Channel in EmbeddedChannel if the handler that should be tested depend on the parent.

Modifications:

- Add another constructor which allows to specify a parent
- Add unit tests

Result:

Fixes https://github.com/netty/netty/issues/9228.
2019-06-08 09:20:19 -07:00
Stephane Landelle
576b1768f4 Introduce WebSocketClientHandshaker::absoluteUpgradeUrl, close #9205 (#9206)
Motivation:

When connecting through an HTTP proxy over clear HTTP, user agents must send requests with an absolute url. This hold true for WebSocket Upgrade request.

WebSocketClientHandshaker and subclasses currently always send requests with a relative url, which causes proxies to crash as request is malformed.

Modification:

Introduce a new parameter `absoluteUpgradeUrl` and expose it in constructors and WebSocketClientHandshakerFactory.

Result:

It's now possible to configure WebSocketClientHandshaker so it works properly with HTTP proxies over clear HTTP.
2019-06-07 16:01:49 -07:00
yipulash
68f2242ac4 delete Other "Content-" MIME Header Fields exception (#9122)
delete Other "Content-" MIME Header Fields exception

Motivation:

RFC7578 4.8. Other "Content-" Header Fields

The multipart/form-data media type does not support any MIME header
fields in parts other than Content-Type, Content-Disposition, and (in
limited circumstances) Content-Transfer-Encoding. Other header
fields MUST NOT be included and MUST be ignored.

Modification:

Ignore other Content types.

Result: 

Other "Content-" Header Fields should be ignored no exception
2019-06-07 13:51:50 -07:00
Norman Maurer
9d5420987a Add support for loopbackmode and accessing the configured interface when using epoll native transport with multicast (#9218)
Motivation:

We did not have support for enable / disable loopback mode in our native epoll transport and also missed the implemention to access the configured interface.

Modifications:

Add implementation and adjust test to cover it

Result:

More complete multicast support with native epoll transport
2019-06-07 13:45:45 -07:00
Carl Mastrangelo
c27fbb5ff2 Handle missing methods on ChannelHandlerMask (#9221)
Motivation:

When Netty is run through ProGuard, seemingly unused methods are removed.  This breaks reflection, making the Handler skipping throw a reflective error.

Modification:

If a method is seemingly absent, just disable the optimization.

Result:

Dealing with ProGuard sucks infinitesimally less.
2019-06-07 13:42:56 -07:00
Scott Mitchell
e0bfc4f91f HTTP/2 avoid closing connection when writing GOAWAY (#9227)
Motivation:
b4e3c12b8e introduced code to avoid coupling
close() to graceful close. It also added some code which attempted to infer when
a graceful close was being done in writing of a GOAWAY to preserve the
"connection is closed when all streams are closed behavior" for the child
channel API. However the implementation was too overzealous and may preemptively
close the connection if there are not currently any open streams (and close if
there are any frames which create streams in flight).

Modifications:
- Decouple writing a GOAWAY from trying to infer if a graceful close is being
  done and closing the connection. Even if we could enhance this logic (e.g.
wait to close until the second GOAWAY with no error) it is possible the user
doesn't want the connection to be closed yet. We can add a means for the codec
to orchestrate the graceful close in the future (e.g. write some special "close
the connection when all streams are closed") but for now we can just let the
application handle this.

Result:
Fixes https://github.com/netty/netty/issues/9207
2019-06-06 17:44:43 -07:00
Carl Mastrangelo
f01278616a Properly debounce wakeups (#9191)
Motivation:
The wakeup logic in EpollEventLoop is overly complex

Modification:
* Simplify the race to wakeup the loop
* Dont let the event loop wake up itself (it's already awake!)
* Make event loop check if there are any more tasks after preparing to
sleep.  There is small window where the non-eventloop writers can issue
eventfd writes here, but that is okay.

Result:
Cleaner wakeup logic.

Benchmarks:

```
BEFORE
Benchmark                                   Mode  Cnt       Score      Error  Units
EpollSocketChannelBenchmark.executeMulti   thrpt   20  408381.411 ± 2857.498  ops/s
EpollSocketChannelBenchmark.executeSingle  thrpt   20  157022.360 ± 1240.573  ops/s
EpollSocketChannelBenchmark.pingPong       thrpt   20   60571.704 ±  331.125  ops/s

Benchmark                                   Mode  Cnt       Score      Error  Units
EpollSocketChannelBenchmark.executeMulti   thrpt   20  440546.953 ± 1652.823  ops/s
EpollSocketChannelBenchmark.executeSingle  thrpt   20  168114.751 ± 1176.609  ops/s
EpollSocketChannelBenchmark.pingPong       thrpt   20   61231.878 ±  520.108  ops/s
```
2019-06-04 05:27:15 -07:00
EliyahuStern
5dacc5f3da Resolve the pid field in PeerCredentials of KQueueDomainSocketChannels. (#9219)
Motivation:

This resolves a TODO from the initial transport-native-kqueue implementation, supplying the user with the pid of the local peer client/server process.

Modification:

Inside netty_kqueue_bsdsocket_getPeerCredentials, Call getsockopt with LOCAL_PEERPID and pass it to PeerCredentials constructor.
Add a test case in KQueueSocketTest.

Result:

PeerCredentials now have pid field set. Fixes https://github.com/netty/netty/issues/9213
2019-06-04 05:16:11 -07:00
Jon Chambers
26275e8a3d Close delegate resolver from RoundRobinInetAddressResolver (#9214)
Motivation:

RoundRobinDnsAddressResolverGroup ultimately opens UDP
ports for DNS resolution. Callers likely expect that
RoundRobinDnsAddressResolverGroup#close() will close those
ports, but that is not currently true (see #9212).

Modifications:

Overrode RoundRobinInetAddressResolver#close() to close
the delegate name resolver, which in turn closes any UDP
ports used for name resolution.

Result:

RoundRobinDnsAddressResolverGroup#close() closes UDP ports
as expected. This fixes #9212.
2019-06-04 05:14:28 -07:00
Nick Hill
4ba75b99af De-duplicate UnpooledDirectByteBuf/UnpooledUnsafeDirectByteBuf (#9085)
Motivation

While digging around looking at something else I noticed that these
share a lot of logic and it would be nice to reduce that duplication.

Modifications

Have UnpooledUnsafeDirectByteBuf extend UnpooledDirectByteBuf and make
adjustments to ensure existing behaviour remains unchanged.

The most significant addition needed to UnpooledUnsafeDirectByteBuf was
re-overriding the getPrimitive/setPrimitive methods to revert back to
the AbstractByteBuf versions which include bounds checks
(UnpooledDirectByteBuf excludes these as an optimization, relying on
those done by underlying ByteBuffer).

Result

~200 fewer lines, less duplicate logic.
2019-06-03 13:15:43 +02:00
Norman Maurer
4b8db65b16 Allow null sender when using DatagramPacketEncoder (#9204)
Motivation:

It is valid to use null as sender so we should support it when DatagramPacketEncoder checks if it supports the message.

Modifications:

- Add null check
- Add unit test

Result:

Fixes https://github.com/netty/netty/issues/9199.
2019-06-03 08:44:48 +02:00
Norman Maurer
a63584715f ByteToMessageDecoder.handlerRemoved(...) should only call fireChannelReadComplete() if fireChannelRead(...) was called before (#9211)
Motivation:

At the moment ByteToMessageDecoder always calls fireChannelReadComplete() when the handler is removed from the pipeline and the cumulation buffer is not null. We should only call it when we also call fireChannelRead(...), which only happens if the cumulation buffer is not null and readable.

Modifications:

Only call fireChannelReadComplete() if fireChannelRead(...) is called before during removal of the handler.

Result:

More correct semantics
2019-06-03 08:44:06 +02:00
Idel Pivnitskiy
b5ff56b6ee Make UnpooledUnsafeHeapByteBuf class public (#9184)
Motivation:

1. Users will be able to use an optimized version of
`UnpooledHeapByteBuf` and override behavior of methods if required.
2. Consistency with `UnpooledDirectByteBuf`, `UnpooledHeapByteBuf`, and
`UnpooledUnsafeDirectByteBuf`.

Modifications:

- Add `public` access modifier to `UnpooledUnsafeHeapByteBuf` class and
ctor;

Result:

Public access for optimized version of `UnpooledHeapByteBuf`.
2019-05-31 07:04:52 +02:00
Norman Maurer
63f7854c1e Don't read from timerfd and eventfd on each EventLoop tick (#9192)
Motivation:

We do not need to issue a read on timerfd and eventfd when the EventLoop wakes up if we register these as Edge-Triggered. This removes the overhead of 2 syscalls and so helps to reduce latency.

Modifications:

- Ensure we register the timerfd and eventfd with EPOLLET flag
- If eventfd_write fails with EAGAIN, call eventfd_read and try eventfd_write again as we only use it as wake-up mechanism.

Result:

Less syscalls and so reducing overhead.

Co-authored-by: Carl Mastrangelo <carl@carlmastrangelo.com>
2019-05-31 07:00:44 +02:00
SplotyCode
5b654fb4f7 Fixed toString() exception in MqttSubscribePayload and MqttUnsubscribePayload (#9202)
Motivation:
The toString() methods of MqttSubscribePayload and MqttUnsubscribePayload are causing exceptions when no topics are set.

Modification:
The toString() methods will not throw Excpetions anymore.

Result:
Fixes #9197
2019-05-31 06:48:21 +02:00
Nick Hill
ed23cfae5e Simplify SingleThreadEventExecutor.awaitTermination() implementation (#9081)
Motivation

A Semaphore is currently dedicated to this purpose but a simple
CountDownLatch will do.

Modification

Remove private threadLock Semaphore from SingleThreadEventExecutor and just use a CountDownLatch.

Also eliminate use of PlatformDependent.throwException() in startThread
method, and combine some nested if clauses.

Result

Cleaner EventLoop termination notification.
2019-05-27 16:09:41 +02:00
Norman Maurer
be7c808ce1 Set the HOST header in Http2ClientInitializer when trying to start an upgrade request (#9177)
Motivation:

The io.netty.example.http2.helloworld.client.Http2Client example should work in the h2c (HTTP2 cleartext - non-TLS) mode, which is the default for this example unless you set a -Dssl VM param. As we do not set the HOST header some servers do reject the upgrade request.

Modifications:

Set the HOST header

Result:

Fixes https://github.com/netty/netty/issues/9115.
2019-05-27 16:04:04 +02:00
Nick Hill
1e575d7389 Fix redundant or missing checks and other inconsistencies in ByteBuf impls (#9119)
Motivation

There are a few minor inconsistencies / redundant operations in the
ByteBuf implementations which would be good to fix.

Modifications

- Unnecessary ByteBuffer.duplicate() performed in
CompositeByteBuf.nioBuffer(int,int)
- Add missing checkIndex(...) check to
ReadOnlyByteBufferBuf.nioBuffer(int,int)
- Remove duplicate bounds check in
ReadOnlyByteBufferBuf.getBytes(int,byte[],int,int)
- Omit redundant bounds check in
UnpooledHeapByteBuf.getBytes(int,ByteBuffer)

Result

More consistency and slightly less overhead
2019-05-27 15:32:42 +02:00
Norman Maurer
f966a0a041 Correctly detect InternetProtocolFamily when EpollDatagramChannel is created with existing FileDescriptor (#9185)
Motivation:

When EpollDatagramChannel is created with an existing FileDescriptor we should detect the correct InternetProtocolFamily.

Modifications:

Obtain the InternetProtocolFamily from the given FD

Result:

Use correct InternetProtocolFamily when EpollDatagramChannel is created via existing FileDescriptor
2019-05-26 20:30:05 +02:00
Steve Buzzard
33d1a91083 Added UDP multicast (with caveats: getInterface, getNetworkInterface, block or loopback-mode-disabled operations).
Motivation:

Provide epoll/native multicast to support high load multicast users (we are using it for a high load telecomm app at my day job).

Modification:

Added support for source specific and any source multicast for epoll transport. Some caveats: no support for disabling loop back mode, retrieval of interface and block operation, all of which tend to be less frequently used.

Result:

Provides epoll transport multicast for common use cases.

Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2019-05-25 10:06:13 +02:00
Norman Maurer
5694dc3e2b Do not use static exceptions for websocket handshake timeout (#9174)
Motivation:

f17bfd0f64 removed the usage of static exception instances to reduce the risk of OOME due addSupressed calls. We should do the same for exceptions used to signal handshake timeouts.

Modifications:

Do not use static instances

Result:

No risk of OOME due addSuppressed calls
2019-05-23 08:36:37 +02:00
noSim
2e7e6d37d5 Updated jboss-marshalling dependency to current license (#9172)
Motivation:

The mentioned license for the jboss-marshalling dependency is outdated. The license has moved from LGPL v2.1 to Apache 2.0.
The version used by Netty (1.4.11Final) is on Apache 2.0 see https://github.com/jboss-remoting/jboss-marshalling/blob/1.4.11.Final/LICENSE.txt

Modification:

Updated NOTICE file with correct license for jboss-marshalling.

Result:

NOTICE file shows correct license.
2019-05-23 07:21:57 +02:00
Nick Hill
4fa7b99da2 Introduce ByteBuf.maxFastWritableBytes() method (#9086)
Motivation

ByteBuf capacity is automatically increased as needed up to maxCapacity
when writing beyond the buffer's current capacity. However there's no
way to tell in general whether such an increase will result in a
relatively costly internal buffer re-allocation.

For unpooled buffers it always does, in pooled cases it depends on the
size of the associated chunk of allocated memory, which I don't think is
currently exposed in any way.

It would sometimes be useful to know where this limit is when making
external decisions about whether to reuse or preemptively reallocate.

It would also be advantageous to take this limit into account when
auto-increasing the capacity during writes, to defer such reallocation
until really necessary.

Modifications

Introduce new AbstractByteBuf.maxFastWritableBytes() method which will
return a value >= writableBytes() and <= maxWritableBytes().

Make use of the new method in the sizing decision made by the
AbstractByteBuf.ensureWritable(...) methods.

Result

Less reallocation/copying.
2019-05-22 20:11:56 +02:00
Vojin Jovanovic
f7a8a6f1ae Remove deprecated GraalVM native-image flags (#9118)
Motivation:

The first final version of GraalVM was released which deprecated some flags. We should use the new ones.

Modifications:

Removes the use of deprecated GraalVM native-image flags
Adds a flag to initialize netty at build time.

Result:

Do not use deprecated flags
2019-05-22 19:22:03 +02:00
Norman Maurer
9c51a8c6d4 Correctly detect that KeyManagerFactory is not supported when using OpenSSL 1.1.0+ (#9170)
Motivation:

How we tried to detect if KeyManagerFactory is supported was not good enough for OpenSSL 1.1.0+ as it partly provided the API but not all of what is required.

This then lead to failures like:

[ERROR] Tests run: 1, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 1.102 s <<< FAILURE! - in io.netty.channel.epoll.EpollDomainSocketStartTlsTest
[ERROR] initializationError(io.netty.channel.epoll.EpollDomainSocketStartTlsTest)  Time elapsed: 0.016 s  <<< ERROR!
javax.net.ssl.SSLException: failed to set certificate and key
	at io.netty.handler.ssl.ReferenceCountedOpenSslServerContext.newSessionContext(ReferenceCountedOpenSslServerContext.java:130)
	at io.netty.handler.ssl.OpenSslServerContext.<init>(OpenSslServerContext.java:353)
	at io.netty.handler.ssl.OpenSslServerContext.<init>(OpenSslServerContext.java:334)
	at io.netty.handler.ssl.SslContext.newServerContextInternal(SslContext.java:468)
	at io.netty.handler.ssl.SslContextBuilder.build(SslContextBuilder.java:457)
	at io.netty.testsuite.transport.socket.SocketStartTlsTest.data(SocketStartTlsTest.java:93)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.lang.reflect.Method.invoke(Method.java:498)
	at org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:50)
	at org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:12)
	at org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:47)
	at org.junit.runners.Parameterized.allParameters(Parameterized.java:280)
	at org.junit.runners.Parameterized.<init>(Parameterized.java:248)
	at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
	at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)
	at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
	at java.lang.reflect.Constructor.newInstance(Constructor.java:423)
	at org.junit.internal.builders.AnnotatedBuilder.buildRunner(AnnotatedBuilder.java:104)
	at org.junit.internal.builders.AnnotatedBuilder.runnerForClass(AnnotatedBuilder.java:86)
	at org.junit.runners.model.RunnerBuilder.safeRunnerForClass(RunnerBuilder.java:59)
	at org.junit.internal.builders.AllDefaultPossibilitiesBuilder.runnerForClass(AllDefaultPossibilitiesBuilder.java:26)
	at org.junit.runners.model.RunnerBuilder.safeRunnerForClass(RunnerBuilder.java:59)
	at org.junit.internal.requests.ClassRequest.getRunner(ClassRequest.java:33)
	at org.apache.maven.surefire.junit4.JUnit4Provider.execute(JUnit4Provider.java:362)
	at org.apache.maven.surefire.junit4.JUnit4Provider.executeWithRerun(JUnit4Provider.java:273)
	at org.apache.maven.surefire.junit4.JUnit4Provider.executeTestSet(JUnit4Provider.java:238)
	at org.apache.maven.surefire.junit4.JUnit4Provider.invoke(JUnit4Provider.java:159)
	at org.apache.maven.surefire.booter.ForkedBooter.invokeProviderInSameClassLoader(ForkedBooter.java:384)
	at org.apache.maven.surefire.booter.ForkedBooter.runSuitesInProcess(ForkedBooter.java:345)
	at org.apache.maven.surefire.booter.ForkedBooter.execute(ForkedBooter.java:126)
	at org.apache.maven.surefire.booter.ForkedBooter.main(ForkedBooter.java:418)
Caused by: java.lang.Exception: Requires OpenSSL 1.0.2+
	at io.netty.internal.tcnative.SSLContext.setCertificateCallback(Native Method)
	at io.netty.handler.ssl.ReferenceCountedOpenSslServerContext.newSessionContext(ReferenceCountedOpenSslServerContext.java:126)
	... 32 more

Modifications:

Also try to set the certification callback and only if this works as well mark KeyManagerFactory support as enabled.

Result:

Also correctly work when OpenSSL 1.1.0 is used.
2019-05-22 19:08:40 +02:00
Norman Maurer
f81cd7d05b
Use correct object to synchronize on in DefaultChannelPipeline (#9171)
Motivation:

8e72071d76 did adjust how synchronization is done but missed to update one block and so used synchronized (this) while it should be synchronized (handlers) .

Modifications:

Use synchronized (handlers)

Result:

Correctly synchronize
2019-05-22 18:50:28 +02:00
秦世成
1465e3ce06 Support handshake timeout in websocket handlers (#8856)
Motivation:

Support handshake timeout option in websocket handlers. It makes sense to limit the time we need to move from `HANDSHAKE_ISSUED` to `HANDSHAKE_COMPLETE` states when upgrading to WebSockets

Modification:

- Add `handshakeTimeoutMillis` option in `WebSocketClientProtocolHandshakeHandler`  and `WebSocketServerProtocolHandshakeHandler`.
- Schedule a timeout task, the task will trigger user event `HANDSHAKE_TIMEOUT` if the handshake timed out.

Result:

Fixes issue https://github.com/netty/netty/issues/8841
2019-05-22 13:56:36 +02:00
Nick Hill
23554e6997 Ensure "full" ownership of msgs passed to EmbeddedChannel.writeInbound() (#9058)
Motivation

Pipeline handlers are free to "take control" of input buffers if they have singular refcount - in particular to mutate their raw data if non-readonly via discarding of read bytes, etc.

However there are various places (primarily unit tests) where a wrapped byte-array buffer is passed in and the wrapped array is assumed not to change (used after the wrapped buffer is passed to EmbeddedChannel.writeInbound()). This invalid assumption could result in unexpected errors, such as those exposed by #8931.

Modifications

Anywhere that the data passed to writeInbound() might be used again, ensure that either:
- A copy is used rather than wrapping a shared byte array, or
- The buffer is otherwise protected from modification by making it read-only

For the tests, copying is preferred since it still allows the "mutating" optimizations to be exercised.

Results

Avoid possible errors when pipeline assumes it has full control of input buffer.
2019-05-22 12:35:03 +02:00
Carl Mastrangelo
912aa7d867 Don't double record stacktrace in Annotated*Exception (#9117)
Motivation:
When initializing the AnnotatedSocketException in AbstractChannel, both
the cause and the stack trace are set, leaving a trailing "Caused By"
that is compressed when printing the trace.

Modification:
Don't include the stack trace in the exception, but leave it in the cause.

Result:
Clearer stack trace
2019-05-22 12:07:06 +02:00
Nick Hill
d29f91540d Fix possible unsafe sharing of internal NIO buffer in CompositeByteBuf (#9169)
Motivation

A small thread-safety bug was introduced during the internal
optimizations of ComponentByteBuf made a while back in #8437. When there
is a single component which was added as a slice,
internalNioBuffer(int,int) will currently return the unwrapped slice's
un-duplicated internal NIO buffer. This is not safe since it could be
modified concurrently with other usage of that parent buffer.

Modifications

Delegate internalNioBuffer to nioBuffer in this case, which returns a
duplicate. This matches what's done in derived buffers in general
(for the same reason). Add unit test.

Result

Fixed possible thread-safety bug
2019-05-22 11:07:27 +02:00
Fabien Renaud
73dae929e3 codec-memcache: copy metadata in binary full request response (#9160)
Motivations
-----------
Calling `copy()`, `duplicate()` or `replace()` on `FullBinaryMemcacheResponse`
or `FullBinaryMemcacheRequest` instances should copy status, opCode, etc.
that are defined in `AbstractBinaryMemcacheMessage`.

Modifications
-------------
 - Modified duplicate, copy and replace methods in
DefaultFullBinaryMemcacheRequest and DefaultFullBinaryMemcacheResponse
to always copy metadata from parent classes.
 - Unit tests verifying duplicate, copy and replace methods for
DefaultFullBinaryMemcacheRequest and DefaultFullBinaryMemcacheResponse
copy buffers and metadata as expected.

Result
------
Calling copy(), duplicate() or replace() methods on
DefaultFullBinaryMemcacheRequest or DefaultFullBinaryMemcacheResponse
produces valid copies with all expected metadata.

Fixes #9159
2019-05-22 11:06:16 +02:00
Norman Maurer
1176d10ebc Skip usage of Japicmp maven plugin for now
Motivation:

We did not release any version of netty 5 yet so we need to skip the usage of the Japicmp plugin.

Modifications:

Skip plugin execution

Result:

Build works
2019-05-22 11:04:52 +02:00
Julien Viet
f85583047e KQueueEventLoop | EpollEventLoop may incorrectly update registration when FD is reused.
Motivation:

The current KQueueEventLoop implementation does not process concurrent domain socket channel registration/unregistration in the order they actual
happen since unregistration are delated by an event loop task scheduling. When a domain socket is closed, it's file descriptor might be reused
quickly and therefore trigger a new channel registration using the same descriptor.

Consequently the KQueueEventLoop#add(AbstractKQueueChannel) method will overwrite the current inactive channels having the same descriptor
and the delayed KQueueEventLoop#remove(AbstractKQueueChannel) will remove the active channel that replaced the inactive one.

As active channels are registered, events for this file descriptor won't be processed anymore and the channels will never be closed.

The same problem can also happen in EpollEventLoop. Beside this we also may never remove the AbstractEpollChannel from the internal map
when it is unregistered which will prevent it from be GC'ed

Modifications:

- Change logic of native KQueue and Epoll implementations to ensure we correctly handle the case of FD reuse
- Only try to update kevent / epoll if the Channel is still open (as otherwise it will be handled by kqueue / epoll itself)
- Correctly remove AbstractEpollChannel from internal map in all cases
- Make implementation of closeAll() consistent for Epoll and KQueueEventLoop

Result:

KQueue and Epoll native transports correctly handle FD reuse

Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2019-05-22 10:11:42 +02:00
Norman Maurer
bbb397ac5c Remove usage of io.netty.handler.ssl.openssl.useKeyManagerFactory system property
Motivation:

Usafe of io.netty.handler.ssl.openssl.useKeyManagerFactory system property was deprecated in 4.1 so let us remove it.

Modifications:

Remove io.netty.handler.ssl.openssl.useKeyManagerFactory usage.

Result:

Remove support of deprecated system property
2019-05-22 09:09:32 +02:00
Tim Brooks
e4ef8f6ff3 Prefer direct io buffers if direct buffers pooled (#9167)
Motivation

Direct buffers are normally preferred when interfacing with raw
sockets. Currently netty will only return direct io buffers (for reading
from a channel) when a platform has unsafe. However, this is
inconsistent with the write-side (filterOutboundMessage) where a direct
byte buffer will be returned if pooling is enabled. This means that
environments without unsafe (and no manual netty configurations) end up
with many pooled heap byte buffers for reading, many pooled direct byte
buffers for writing, and jdk pooled byte buffers (for reading).

Modifications

This commit modifies the AbstractByteBufAllocator to return a direct
byte buffer for io handling when the platform has unsafe or direct byte
buffers are pooled.

Result:

Use direct buffers when direct buffers are pooled for IO.
2019-05-22 07:33:06 +02:00
Norman Maurer
89e796cec6 Update to latest JDK13 EA release (#9166)
Motivation:

We should use the latest EA release when trying to compile with JDK13.

Modifications:

Update to latest release

Result:

Test with latest release on the CI
2019-05-21 20:10:34 +02:00
秦世成
a9159d28fb Format code to align unaligned code. (#9062)
Motivation:
Format code to align unaligned code.

Modification:
Reformat the code

Result:

Cleaner code
2019-05-20 12:07:45 +02:00
Norman Maurer
a65f231bc4 Only try to use reflection to access default nameservers when using Java8 and lower (#9157)
Motivation:

We should only try to use  reflection to access default nameservers when using Java8 and lower as otherwise we will produce an Illegal reflective access warning like:

WARNING: Illegal reflective access by io.netty.resolver.dns.DefaultDnsServerAddressStreamProvider

Modifications:

Add Java version check before try to use reflective access.

Result:

No more warning when Java9+ is used.
2019-05-18 08:21:59 +02:00
Norman Maurer
ed61e5f543 Only use static Exception instances when we can ensure addSuppressed … (#9152)
Motivation:

OOME is occurred by increasing suppressedExceptions because other libraries call Throwable#addSuppressed. As we have no control over what other libraries do we need to ensure this can not lead to OOME.

Modifications:

Only use static instances of the Exceptions if we can either dissable addSuppressed or we run on java6.

Result:

Not possible to OOME because of addSuppressed. Fixes https://github.com/netty/netty/issues/9151.
2019-05-17 22:42:53 +02:00
Norman Maurer
211dde4e08 Add support for TCP fallback when we receive a truncated DnsResponse (#9139)
Motivation:

Sometimes DNS responses can be very large which mean they will not fit in a UDP packet. When this is happening the DNS server will set the TC flag (truncated flag) to tell the resolver that the response was truncated. When a truncated response was received we should allow to retry via TCP and use the received response (if possible) as a replacement for the truncated one.

See https://tools.ietf.org/html/rfc7766.

Modifications:

- Add support for TCP fallback by allow to specify a socketChannelFactory / socketChannelType on the DnsNameResolverBuilder. If this is set to something different then null we will try to fallback to TCP.
- Add decoder / encoder for TCP
- Add unit tests

Result:

Support for TCP fallback as defined by https://tools.ietf.org/html/rfc7766 when using DnsNameResolver.
2019-05-17 14:47:10 +02:00
Norman Maurer
260a8a0e9e Add missing assume checks to skip tests if KeyManagerFactory can not be used (#9148)
Motivation:

Depending on what OpenSSL library version we use / system property that is set we need to skip tests that use KeyManagerFactory.

Modifications:

Add missing assume checks for tests that use KeyManagerFactory.

Result:

All tests pass even if KeyManagerFactory is not supported
2019-05-15 07:25:03 +02:00
RoganDawes
0a1786c32c Remove the Handler only after it has initialized the channel (#9132)
Motivation:

Previously, any 'relative' pipeline operations, such as
ctx.pipeline().replace(), .addBefore(), addAfter(), etc
would fail as the handler was not present in the pipeline.

Modification:

Used the pattern from ChannelInitializer when invoking configurePipeline().

Result:

Fixes #9131
2019-05-13 13:55:17 +02:00
Nick Hill
dd9e0d5a19 AsciiString.lastIndexOf(...) is implemented incorrectly (#9103)
Motivation

@xiaoheng1 reported incorrect behaviour of AsciiString.lastIndexOf in
#9099. Upon closer inspection it appears that it was never implemented
correctly and searches between the provided index and the end of the
string similar to indexOf(...), rather than between the provided index
and the beginning of the string as the javadoc states (and in line with
java.lang.String).

Modifications

Fix AsciiString.lastIndexOf implementation and corresponding unit tests
to behave the same as the equivalent String methods.

Result

Fixes #9099
2019-05-13 07:03:51 +02:00
Nick Hill
5e786e549e Fix incorrect behavior of ReadOnlyByteBufferBuf.getBytes(int,ByteBuffer) (#9125)
* Fix incorrect behavior of ReadOnlyByteBufferBuf.getBytes(int,ByteBuffer)

Motivation

It currently will succeed when the destination is larger than the source
range, but the ByteBuf javadoc states this should be a failure, as is
the case with all the other implementations.

Modifications

- Fix logic to fail the bounds check in this case
- Remove explicit null check which isn't done in any equivalent method
- Add unit test

Result

More correct/consistent behaviour
2019-05-13 07:03:10 +02:00
Norman Maurer
debc1cbc0a DnsNameResolver.resolveAll(DnsQuestion) should not try to filter duplicates (#9141)
Motivation:

https://github.com/netty/netty/pull/9021 did apply some changes to filter out duplicates InetAddress when calling resolveAll(...) to mimic JDK behaviour. Unfortunally this also introduced a regression as we should not filter duplicates when the user explicit calls resolveAll(DnsQuestion).

Modifications:

- Only filter duplicates if resolveAll(String) is used
- Add unit test

Result:

Fixes regressions introduces by https://github.com/netty/netty/pull/9021
2019-05-13 06:59:28 +02:00
SplotyCode
d3a13a0d6a Allow to specify KeyStore type in SslContext (#9003)
Motivation:

As brought up in https://github.com/netty/netty/issues/8998, JKS can be substantially faster than pkcs12, JDK's new default. Without an option to set the KeyStore type you must change the configuration of the entire JVM which is impractical.

Modification:

- Allow to specify KeyStore type
- Add test case

Result:

Fixes https://github.com/netty/netty/issues/8998.
2019-05-10 07:51:20 +02:00
Norman Maurer
83210c47ca Allow to have DnsNameResolver.resolveAll(...) notify as soon as the preferred records were resolved (#9136)
Motivation:

075cf8c02e introduced a change to allow resolve(...) to notify as soon as the preferred record was resolved. This works great but we should also allow the user to configure that we want to do the same for resolveAll(...), which means we should be able to notify as soon as all records for a preferred record were resolved.

Modifications:

- Add a new DnsNameResolverBuilder method to allow configure this (use false as default to not change default behaviour)
- Add unit test

Result:

Be able to speed up resolving.
2019-05-09 08:08:47 +02:00
Andrey Mizurov
a69ae1aced Fixed HttpHelloWorldServerHandler for handling HTTP 1.0/1.1 (#9124)
Motivation:

HttpHelloWorldServer example works incorrect for HTTP 1.1, the value of header connection is always set to close for each request.

Modification:

Correctly set header

Result:

Fixed HttpHelloWorldServerHandler for handling HTTP 1.0/1.1
2019-05-08 09:05:18 +02:00
Anuraag Agrawal
b45a1cd15b Add equality check to contentEquals instance methods. (#9130)
Motivation:

An instance is always equal to itself. It makes sense to skip processing for this case, which isn't uncommon since `AsciiString` is often memoized within an application when used as HTTP header names.

Modification:

`contentEquals` methods first check for instance equality before doing processing.

Result:

`contentEquals` will be faster when comparing an instance with itself.

I couldn't find any unit tests for these methods, only the static version. Let me know if I should add something to `AsciiStringCharacterTest`.

Came up here:
https://github.com/line/armeria/pull/1731#discussion_r280396280
2019-05-08 07:31:09 +02:00