Commit Graph

1734 Commits

Author SHA1 Message Date
Norman Maurer
8a5e42ad2e Correct fail write with NotYetConnectedException when OioDatagramChannel is not connected yet.
Motivation:

NioDatagramChannel fails a write with NotYetConnectedException when the DatagramChannel was not yet connected and a ByteBuf is written. The same should be done for OioDatagramChannel as well.

Modifications:

Make OioDatagramChannel consistent with NioDatagramChannel

Result:

Correct and consistent implementations of DatagramChannel
2017-02-06 11:06:00 +01:00
周岑
48f6541cb3 delete no useful intermediate variables
delete no useful intermediate variables
2017-02-06 07:55:29 +01:00
Norman Maurer
735d6dd636 [maven-release-plugin] prepare for next development iteration 2017-01-30 15:14:02 +01:00
Norman Maurer
76e22e63f3 [maven-release-plugin] prepare release netty-4.1.8.Final 2017-01-30 15:12:36 +01:00
Tim Brooks
3344cd21ac Wrap operations requiring SocketPermission with doPrivileged blocks
Motivation:

Currently Netty does not wrap socket connect, bind, or accept
operations in doPrivileged blocks. Nor does it wrap cases where a dns
lookup might happen.

This prevents an application utilizing the SecurityManager from
isolating SocketPermissions to Netty.

Modifications:

I have introduced a class (SocketUtils) that wraps operations
requiring SocketPermissions in doPrivileged blocks.

Result:

A user of Netty can grant SocketPermissions explicitly to the Netty
jar, without granting it to the rest of their application.
2017-01-19 21:12:52 +01:00
Norman Maurer
cd9008f95b Remove unnecessray for loop missed by fac0ca8319 2017-01-19 19:13:57 +01:00
Scott Mitchell
9a4aa617f4 PlatformDependent#getClassLoader fails in restrictive classloader environment
Motivation:
https://github.com/netty/netty/pull/6042 only addressed PlatformDependent#getSystemClassLoader but getClassLoader is also called in an optional manner in some common code paths but fails to catch a general enough exception to continue working.

Modifications:
- Calls to getClassLoader which can continue if results fail should catch Throwable

Result:
More resilient code in the presense of restrictive class loaders.
Fixes https://github.com/netty/netty/issues/6246.
2017-01-19 09:01:36 -08:00
周岑
86020a2858 Typo error: Method invoker() no longer exists
Method invoker() no longer exists
2017-01-19 12:16:31 +01:00
Norman Maurer
fac0ca8319 Warn about not-supported ChannelOption when bootstrap Channels.
Motivation:

We not warned about not-supported ChannelOptions when set the options for the ServerChannel.

Modifications:

- Share code for setting ChannelOptions during bootstrap

Result:

Warning is logged when a ChannelOption is used that is not supported during bootstrap a Channel. See also [#6192]
2017-01-19 08:27:19 +01:00
Norman Maurer
7f01da8d0f [maven-release-plugin] prepare for next development iteration 2017-01-12 11:36:51 +01:00
Norman Maurer
7a21eb1178 [maven-release-plugin] prepare release netty-4.1.7.Final 2017-01-12 11:35:58 +01:00
Norman Maurer
c1830c8b4e Fix missleading comment in AbstractChannelHandlerContext.invokeHandler()
Motivation:

The comment on AbstractChannelHandlerContext.invokeHandler() is incorrect and missleading. See [#6177]

Modifications:

Change true to false to correct the comment.

Result:

Fix missleading and incorrect comment.
2017-01-10 12:05:24 +01:00
Jon Chambers
074075de7e Expose channel pool configuration to subclasses.
Motivation:

`SimpleChannelPool` subclasses are likely to override the `connectChannel` method, and are likely to clobber the cloned `Bootstrap` handler in the process. To allow subclasses to properly notify the pool listener of new connections, we should expose (at least) the `handler` property of the pool to subclasses.

Modifications:

Expose `SimpleChannelPool` properties to subclasses via `protected` getters.

Result:

Subclasses can now use the bootstrap, handler, health checker, and health-check-on-release preoperties from their superclass.
2016-12-21 20:45:01 +01:00
Scott Mitchell
3d11334151 Fix DefaultChannelId MAC address parsing bug
Motivation:
DefaultChannelId provides a regular expression which validates if a user provided MAC address is valid. This regular expression may allow invalid MAC addresses and also not allow valid MAC addresses.

Modifications:
- Introduce a MacAddressUtil#parseMac method which can parse and validate the MAC address at the same time. The regular expression check before hand is additional overhead if we have to parse the MAC address.

Result:
Fixes https://github.com/netty/netty/issues/6132.
2016-12-20 17:06:27 -08:00
Norman Maurer
cfd8fb10db [#6134] Do not limit the PID to be <= 4194304
Motivation:

On some platforms the PID my be bigger then 4194304 so we should not limit it to 4194304.

Modifications:

Only check that the PID is a valid Integer

Result:

No more warnings on systems where the PID is bigger then 4194304.
2016-12-20 10:31:16 +01:00
Norman Maurer
89e93968ac Remove usage of own Atomic*FieldUpdater in favor of JDKs
Motivation:

In later Java8 versions our Atomic*FieldUpdater are slower then the JDK implementations so we should not use ours anymore. Even worse the JDK implementations provide for example an optimized version of addAndGet(...) using intrinsics which makes it a lot faster for this use-case.

Modifications:

- Remove methods that return our own Atomic*FieldUpdaters.
- Use the JDK implementations everywhere.

Result:

Faster code.
2016-12-15 08:09:06 +00:00
Scott Mitchell
002c99e751 NIO ServerChannel shouldn't close because of Exception
Motivation:
e102a008b6 changed a conditional where previously the NIO ServerChannel would not be closed in the event of an exception.

Modifications:
- Restore the logic prior to e102a008b6 which does not automatically close ServerChannels for IOExceptions

Result:
NIO ServerChannel doesn't close automatically for an IOException.
2016-12-05 20:51:05 -08:00
Norman Maurer
eed6791f8e Cleanup after commit fc1cdc991e 2016-12-05 12:18:35 +01:00
Norman Maurer
fc1cdc991e [#6095] Remove catching of ConcurrentModificationException as this can not happen.
Motivation:

We should not catch ConcurrentModificationException as this can never happen because things are executed on the EventLoop thread.

Modifications:

Remove try / catch

Result:

Cleaner code.
2016-12-04 18:59:10 +01:00
Derbylock
f6b37a38ba Removed final keyword from FixedChannelPool 2016-11-23 13:37:37 +01:00
Scott Mitchell
a25101dd0b Now that LocalChannel#releaseInboundBuffers is only called from the EventLoop (eb4d317b9d) it should clear readInProgress and drain/release the queue. Otherwise if a read event is pending (doBeginRead) was called we may later call channelRead or channelReadComplete after we have closed the channel.
Modifications:
LocalChannel#releaseInboundBuffers should always clear/release the queue and set readInProgress to false

Result:
LocalChannel queue is more reliably cleaned up.
2016-11-21 11:11:43 -08:00
Scott Mitchell
eb4d317b9d Fix LocalChannel close sequence
Motivation:
LocalChannel attempts to close its peer socket when ever it is closed. However if the channels are on different EventLoops we may attempt to process events for the peer channel on the wrong EventLoop.

Modifications:
- Ensure the close process ensures we are on the correct thread before accessing data

Result:
More correct LocalChannel close code.
2016-11-21 10:33:55 -08:00
Scott Mitchell
a043cf4a98 Catch exceptions from PlatformDependent#getSystemClassLoader
Motivation:
PlatformDependent#getSystemClassLoader may throw a wide variety of exceptions based upon the environment. We should handle all exceptions and continue initializing the slow path if an exception occurs.

Modifications:
- Catch Throwable in cases where PlatformDependent#getSystemClassLoader is used

Result:
Fixes https://github.com/netty/netty/issues/6038
2016-11-19 09:23:25 -08:00
Scott Mitchell
c1932a8537 ByteBuf Input Stream Reference Count Ownership
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.
2016-11-14 16:29:55 -08:00
Norman Maurer
e3cb9935c0 Take memory overhead of ChannelOutboundBuffer / PendingWriteQueue into account
Motivation:

To guard against the case that a user will enqueue a lot of empty or small buffers and so raise an OOME we need to also take the overhead of the ChannelOutboundBuffer / PendingWriteQueue into account when detect if a Channel is writable or not. This is related to #5856.

Modifications:

When calculate the memory for an message that is enqueued also add some extra bytes depending on the implementation.

Result:

Better guard against OOME.
2016-11-03 15:54:00 +01:00
Roger Kapsi
b2379e62f4 Allow customization of LocalChannel instances that are being created by LocalServerChannel.
Motivation

It's possible to extend LocalChannel as well as LocalServerChannel but the LocalServerChannel's serve(peer) method is hardcoded to create only instances of LocalChannel.

Modifications

Add a protected factory method that returns by default new LocalChannel(...) but users may override it to customize it.

Result

It's possible to customize the LocalChannel instance on either end of the virtual connection.
2016-10-30 08:32:49 +01:00
Norman Maurer
5f533b7358 [maven-release-plugin] prepare for next development iteration 2016-10-14 13:20:41 +02:00
Norman Maurer
35fb0babe2 [maven-release-plugin] prepare release netty-4.1.6.Final 2016-10-14 12:47:19 +02:00
Scott Mitchell
eb1dfb8f5e SingleThreadEventLoopTest failures
Motivation:
Some unit tests in SingleThreadEventLoopTest rely upon Thread.sleep for sequencing events between threads. This can be unreliable and result in spurious test failures if thread scheduling does not occur in a fair predictable manner.

Modifications:
- Reduce the reliance on Thread.sleep in SingleThreadEventLoopTest

Result:
Fixes https://github.com/netty/netty/issues/5851
2016-10-11 09:09:53 +02:00
Norman Maurer
6c8cd023bf [#5770] Use heapbuffers by default when using LocalChannel and LocalServerChannel.
Motivation:

The local transport is used to communicate in the same JVM so we should use heap buffers.

Modifications:

Use heapbuffers by default if not requested otherwise.

Result:

No allocating of direct buffers by default when using local transport
2016-10-10 11:16:17 +02:00
Norman Maurer
e102a008b6 [#5893] Ensure we not close NioDatagramChannel when SocketException is received.
Motivation:

When using java.nio.DatagramChannel we should not close the channel when a SocketException was thrown as we can still use the channel.

Modifications:

Not close the Channel when SocketException is thrown

Result:

More robust and correct handling of exceptions when using NioDatagramChannel.
2016-10-10 10:24:28 +02:00
Norman Maurer
a09e56850e [#5882] Ensure we even process tasks if processing of ready channels throws an Exception in event loop.
Motivation:

If an exception is thrown while processing the ready channels in the EventLoop we should still run all tasks as this may allow to recover. For example a OutOfMemoryError may be thrown and runAllTasks() will free up memory again. Beside this we should also ensure we always allow to shutdown even if an exception was thrown.

Modifications:

- Call runAllTasks() in a finally block
- Ensure shutdown is always handles.

Result:

More robust EventLoop implementations for NIO and Epoll.
2016-10-10 07:49:57 +02:00
Norman Maurer
3cf7ccbd3c Process OP_WRITE before OP_READ to free memory faster
Motivation:

We should better first process OP_WRITE before OP_READ as this may allow us to free memory in a faster fashion for previous queued writes.

Modifications:

Process OP_WRITE before OP_READ

Result:

Free memory faster for queued writes.
2016-10-10 07:42:39 +02:00
radai-rosenblatt
15ac6c4a1f Clean-up unused imports
Motivation:

the build doesnt seem to enforce this, so they piled up

Modifications:

removed unused import lines

Result:

less unused imports

Signed-off-by: radai-rosenblatt <radai.rosenblatt@gmail.com>
2016-09-30 09:08:50 +02:00
Roger Kapsi
149916d052 Adding ability omit the implicit #flush() call in EmbeddedChannel#writeOutbound() and
the implicit #fireChannelReadComplete() in EmbeddedChannel#writeInbound().

Motivation

We use EmbeddedChannels to implement a ProxyChannel of some sorts that shovels
messages between a source and a destination Channel. The latter are real network
channels (such as Epoll) and they may or may not be managed in a ChannelPool. We
could fuse both ends directly together but the EmbeddedChannel provides a nice
disposable section of a ChannelPipeline that can be used to instrument the messages
that are passing through the proxy portion.

The ideal flow looks abount like this:

source#channelRead() -> proxy#writeOutbound() -> destination#write()
source#channelReadComplete() -> proxy#flushOutbound() -> destination#flush()

destination#channelRead() -> proxy#writeInbound() -> source#write()
destination#channelReadComplete() -> proxy#flushInbound() -> source#flush()

The problem is that #writeOutbound() and #writeInbound() emit surplus #flush()
and #fireChannelReadComplete() events which in turn yield to surplus #flush()
calls on both ends of the pipeline.

Modifications

Introduce a new set of write methods that reain the same sematics as the #write()
method and #flushOutbound() and #flushInbound().

Result

It's possible to implement the above ideal flow.

Fix for EmbeddedChannel#ensureOpen() and Unit Tests for it

Some PR stuff.
2016-09-24 12:33:04 -07:00
ChuntaoLu
9008e72c2b Fix javadoc
Removes unmatched brace
2016-09-12 06:43:10 -07:00
Norman Maurer
6bbf32134a Log more details if notification of promise fails in PromiseNotifier and AbstractChannelHandlerContext
Motivation:

To make it easier to debug why notification of a promise failed we should log extra info and make it consistent.

Modifications:

- Create a new PromiseNotificationUtil that has static methods that can be used to try notify a promise and log.
- Reuse this in AbstractChannelHandlerContext, ChannelOutboundBuffer and PromiseNotifier

Result:

Easier to debug why a promise could not be notified.
2016-09-07 06:55:38 +02:00
Norman Maurer
dfa3bbbf00 Add support for Client Subnet in DNS Queries (RFC7871)
Motivation:

RFC7871 defines an extension which allows to request responses for a given subset.

Modifications:

- Add DnsOptPseudoRrRecord which can act as base class for extensions based on EDNS(0) as defined in RFC6891
- Add DnsOptEcsRecord to support the Client Subnet in DNS Queries extension
- Add tests

Result:

Client Subnet in DNS Queries extension is now supported.
2016-09-06 07:16:57 +02:00
buchgr
4af6855695 Remove @Deprecated for primitive WriteWaterMark getters and setters
Motivation:

For use cases that demand frequent updates of the write watermarks, an
API that requires immutable WriteWaterMark objects is not ideal, as it
implies a lot of object allocation.

For example, the HTTP/2 child channel API uses write watermarks for outbound
flow control and updates the write watermarks on every DATA frame write.

Modifications:

Remote @Deprecated tag from primitive getters and setters, however the corresponding
channel options remain deprecated.

Result:

Primitive getters and setters for write watermarks are no longer marked @Deprecated.
2016-09-05 10:26:05 +02:00
Norman Maurer
30fe2e868f Call finishConnect() before try to call read(...) / write(...) when using NIO
Motivation:

The JDK implementation of SocketChannel has an internal state that is tracked for its operations. Because of this we need to ensure we call finishConnect() before try to call read(...) / write(...) as otherwise it may produce a NotYetConnectedException.

Modifications:

First process OP_CONNECT flag.

Result:

No more possibility of NotYetConnectedException because OP_CONNECT is handled not early enough when processing interestedOps for a Channel.
2016-09-01 08:55:02 +02:00
Norman Maurer
54b1a100f4 [maven-release-plugin] prepare for next development iteration 2016-08-26 10:06:32 +02:00
Norman Maurer
6fd8bb8c63 [#5763] DefaultEventLoopGroup doesn't expose ctor variant that accepts custom Executor
Motivation:

The DefaultEventLoopGroup class extends MultithreadEventExecutorGroup but doesn't expose the ctor variants that accept a custom Executor like NioEventLoopGroup and EpollEventLoopGroup do.

Modifications:

Add missing constructor.

Result:

Be able to use custom Executor with DefaultEventLoopGroup.
2016-09-01 08:20:05 +02:00
Jason Tedor
00c0664ef8 Avoid inaccessible object exception replacing set
Motivation:

When attempting to set the selectedKeys fields on the selector
implementation, JDK 9 can throw an inaccessible object exception.

Modications:

Catch and log this exception as an possible course of action if the
sun.nio.ch package is not exported from java.base.

Result:

The selector replacement will fail gracefully as an expected course of
action if the sun.nio.ch package is not exported from java.base.
2016-08-31 13:59:48 +02:00
Norman Maurer
1208b90f57 [maven-release-plugin] prepare release netty-4.1.5.Final 2016-08-26 04:59:35 +02:00
Norman Maurer
d3cb95ef00 Make NIO and EPOLL transport connect errors more consistent with the JDK
Motivation:

The NIO transport used an IllegalStateException if a user tried to issue another connect(...) while the connect was still in process. For this case the JDK specified a ConnectPendingException which we should use. The same issues exists in the EPOLL transport. Beside this the EPOLL transport also does not throw the right exceptions for ENETUNREACH and EISCONN errno codes.

Modifications:

- Replace IllegalStateException with ConnectPendingException in NIO and EPOLL transport
- throw correct exceptions for ENETUNREACH and EISCONN in EPOLL transport
- Add test case

Result:

More correct error handling for connect attempts when using NIO and EPOLL transport
2016-08-27 20:57:36 +02:00
buchgr
4accd300e5 Fix write watermarks comparison to use less than and greater than.
Motivation:

The API documentation in ChannelConfig states that a a channel is writable,
if the number of pending bytes is below the low watermark and a
channel is not writable, if the number of pending bytes exceeds the high
watermark.

Therefore, we should use < operators instead of <= as well as > instead of >=.

Using <= and >= is also problematic, if the low watermark is equal to the high watermark,
as then a channel could be both writable and unwritable with the same number of pending
bytes (depending on whether remove() or addMessage() is called first).

The use of <= and >= was introduced in PR https://github.com/netty/netty/pull/3036, but
I don't understand why, as there doesn't seem to have been any discussion around that.

Modifications:

Use < and > operators instead of <= and >=.

Result:

High and low watermarks are treated as stated in the API docs.
2016-08-24 15:58:02 +02:00
Norman Maurer
5e148d5670 [#5639] Ensure fireChannelActive() is also called if Channel is closed in connect promise.
Motivation:

We need to ensure we also call fireChannelActive() if the Channel is directly closed in a ChannelFutureListener that is belongs to the promise for the connect. Otherwise we will see missing active events.

Modifications:

Ensure we always call fireChannelActive() if the Channel was active.

Result:

No missing events.
2016-08-24 08:47:49 +02:00
Norman Maurer
43626ac6ea Use NIO methods when using Java7+ in the NIO transport
Motivation:

We use often javachannel().socket().* in NIO as these methods exists in java6. The problem is that these will throw often very general Exceptions (Like SocketException) while it is more expected to throw the Exceptions listed in the nio interfaces. When possible we should use the new methods available in java7+ which throw the correct exceptions.

Modifications:

Check for java version and depending on it using the socket or the javachannel.

Result:

Throw expected Exceptions.
2016-08-24 07:36:14 +02:00
Norman Maurer
8ce7e73e78 Prevent extra peformance hit by fillInStackTrace() when create a new annotated connect exception.
Motivation:

To make it easier to debug connect exceptions we create new exceptions which also contain the remote address. For this we basically created a new instance and call setStackTrace(...). When doing this we pay an extra penality because it calls fillInStackTrace() when calling the super constructor.

Modifications:

Create special sub-classes of Exceptions that override the fillInStackTrace() method and so eliminate the overhead.

Result:

Less overhead when "annotate" connect exceptions.
2016-08-24 07:34:43 +02:00
buchgr
8d1e46ffd1 Remove reference to 5.0 release.
Motivation:

Comments stating that AUTO_CLOSE will be removed in Netty 5.0 are wrong,
as there is no Netty 5.0.

Modifications:

Removed comment.

Result:

No more references to Netty 5.0
2016-08-23 09:46:57 +02:00