Commit Graph

10481 Commits

Author SHA1 Message Date
Heng Zhang
b3e0e09a82
Distribue weight to children when closing stream (#11490)
Motivation:

As suggested in [section 5.3.4 in http2 spec](https://datatracker.ietf.org/doc/html/rfc7540#section-5.3.4):

> When a stream is removed from the dependency tree, its dependencies can be moved to become dependent on the parent of the closed stream. The weights of new dependencies are recalculated by distributing the weight of the dependency of the closed stream proportionally based on the weights of its dependencies.

For example, we have stream A and B depend on connection stream with default weights (16), and stream C depends on A with max weight (256). When stream A was closed, we move stream C to become dependent on the connection stream, then we should distribute the weight of stream A to its children (only stream C), so the new weight of stream C will be 16. If we keep the weight of stream C unchanged, it will get more resource than stream B

Modification:
- distribute weight to its children when closing a stream
- add a unit test for the case above and fix other related unit tests

Result:

More spec-compliant and more appropriate stream reprioritization

Co-authored-by: Heng Zhang <zhangheng@imo.im>
2021-07-20 14:17:56 +02:00
Netty Project Bot
2a184ab932 [maven-release-plugin] prepare for next development iteration 2021-07-16 06:23:23 +00:00
Netty Project Bot
e43d0d99f1 [maven-release-plugin] prepare release netty-4.1.66.Final 2021-07-16 06:23:17 +00:00
Chris Vest
9e8b8ff53a
Fix a bug where SslHandler clients would not process Server Hello messages in a timely manner (#11472)
Motivation:
The TLS handshake must be able to finish on its own, without being driven by outside read calls.
This is currently not the case when TCP FastOpen is enabled.
Reads must be permitted and marked as pending, even when a channel is not active.

This is important because, with TCP FastOpen, the handshake processing of a TLS connection will start
before the connection has been established -- before the process of connecting has even been started.

The SslHandler on the client side will add the Client Hello message to the ChannelOutboundBuffer, then
issue a `ctx.read` call for the anticipated Server Hello response, and then flush the Client Hello
message which, in the case of TCP FastOpen, will cause the TCP connection to be established.

In this transaction, it is important that the `ctx.read` call is not ignored since, if auto-read is
turned off, this could delay or even prevent the Server Hello message from being processed, causing
the server-side handshake to time out.

Modification:
Attach a listener to the SslHandler.handshakeFuture in the EchoClient, that will call ctx.read.

Result:
The SocketSslEchoTest now tests that the SslHandler can finish handshakes on its own, without being driven by 3rd party ctx.read calls.
The various channel implementations have been updated to comply with this behaviour.
2021-07-15 09:02:03 +02:00
Norman Maurer
55957d3e75
SelfSignedCertificate should try BouncyCastle first (#11487)
Motivation:

In JDK version >= 9 the access to sun.* is not permitted anymore by default. Because of this we should better first try the BouncyCastle based implementation before falling back to the JDK based version.

Modifications:

Switch ordering of usage of BouncyCastle vs JDK internals.

Result:

Less surprising errors when using SelfSignedCertificate in Java >9
2021-07-15 08:59:20 +02:00
Norman Maurer
e236e99006
Migrate the rest of the ssl package to junit5 (#11483)
Motivation:

We should use junit5 everywhere.

Modifications:

- Refactor rest of tests to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-07-15 08:50:24 +02:00
Norman Maurer
2914134522
Disable flaky test for now (#11488)
Motivation:

JdkOpenSslEngineInteroptTest.testMutualAuthSameCerts() is flaky on the CI and so fails the PR build quite often.
Let's disable it for now until we were able to reproduce it locally and fix it.

Modifications:

Disable flaky test

Result:

More stable CI builds
2021-07-14 16:42:34 +02:00
Nitesh Kant
8337e3a973
Improve name matching in DNS answers (#11474)
__Motivation__

Upon receiving a DNS answer, we match whether the name in the question matches the name in the record. Some DNS servers we have encountered append a search domain to the record name which fails this match. eg: for question name `netty` and search domains `io` and `com`, we will do 2 queries: `netty.io.` and `netty.com.`, if the answer for `netty.io` contains `netty.com` then we ignore this record.

__Modification__

If the name in the record does not match the name in the question, append configured search domains to the question name to see if it matches the record name.

__Result__

Records names with appended search domains are still returned as valid answers.
2021-07-14 14:11:22 +02:00
Norman Maurer
876624a22b
SelfSignedCertificate should work in http2 tests (#11486)
Motivation:

We need to add `--add-exports java.base/sun.security.x509=ALL-UNNAMED` when running the tests for codec-http2 as some of the tests use SelfSignedCertificate.

Modifications:

- Add `--add-exports java.base/sun.security.x509=ALL-UNNAMED` when running the tests for codec-http2
- Ensure we export correct when running with JDK12, 13, 14 and 15 as well

Result:

No more tests failure due not be able to access classes
2021-07-14 07:46:17 +02:00
Aayush Atharva
3b8e7c3e60
Use Static Constants as Sample Data (#11485)
Motivation:
#11468 was merged but didn't fix tests completely. There is a fight between `LF` and `CRLF`. So to eliminate this, we should just get rid of them.

Modification:
Use a small sample dataset without `LF` and `CRLF`.

Result:
Simple and passing test.
2021-07-14 07:45:21 +02:00
Norman Maurer
043e9e309e
Remove rest of junit4 usage (#11484)
Motivation:

We did migrate all these modules to junit5 before but missed a few usages of junit4

Modifications:

Replace all junit4 imports by junit5 apis

Result:

Part of  https://github.com/netty/netty/issues/10757
2021-07-13 20:59:57 +02:00
Aayush Atharva
7ec198ddf9
Fix JavaDoc and Run Brotli tests on Windows (#11468)
Motivation:
JavaDoc of StandardCompressionOptions should point towards public methods. Also, Brotli tests were failing on Windows.

Modification:
Fixed JavaDoc and enabled Brotli tests on Windows.

Result:
Better JavaDoc and Brotli tests will run on Windows

Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2021-07-13 16:24:58 +02:00
Norman Maurer
98c446c493 Revert "Only do a partial build for a few jobs when verify PRs (#11464)"
This reverts commit e5adc1131a.
2021-07-13 15:32:42 +02:00
skyguard1
df22356a3a
Refactor HttpContentCompressor using CompressionEncoderFactory (#11480)
Motivation:

The `HttpContentCompressor.beginEncode()` method has too many if else, so consider refactoring

Modification:

Create the corresponding `CompressionEncoderFactory` according to the compression algorithm, remove the if else

Result:

The code of `HttpContentCompressor` is cleaner than the previous implementation

Signed-off-by: xingrufei <xingrufei@sogou-inc.com>
Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2021-07-13 12:25:30 +02:00
Aayush Atharva
6085cd7ab6
Update HttpContentCompressor to pass correct message to ObjectUtil (#11482)
Motivation:
In #11256, We introduced `Iterable` as a parameter but later in review, it was removed. But we forgot to change `compressionOptionsIterable` to just `compressionOptions`.

Modification:
Changed `compressionOptionsIterable` to `compressionOptions`.

Result:
Correct ObjectUtil message
2021-07-13 08:11:01 +02:00
Adrian Antkowiak
8cbdefae23
HAProxyMessage.sourceAddress() can be null (#11475)
Motivation:

If `send-proxy-v2` is used `sourceAddress()` can be `null`.

Modification:

Update docs to reflect this.

Result:

Docs are more correct.
2021-07-12 11:55:39 +02:00
skyguard1
6ce36f1909
Add zstd http content compression support (#11470)
Motivation:

netty needs to support zstd content-encoding http content compression

Modification:

Add ZstdOptions, and modify HttpContentCompressor and CompressorHttp2ConnectionEncoder to support zstd compression

Result:

netty supports zstd http content compression

Signed-off-by: xingrufei <xingrufei@sogou-inc.com>
2021-07-12 08:49:04 +02:00
Violeta Georgieva
3f9a5f50be
Add support for Unix domain datagram sockets when using native epoll/kqueue transport (#11423)
Motivation:

There are use cases when Unix domain datagram sockets are needed for communication. This PR adds such support for Epoll/KQueue.

Modification:

- Expose Channel, Config and Packet interfaces/classes for Unix domain datagram sockets. All interfaces/classes are in `transport-native-unix-common` module in order to be available for KQueue and Epoll implementations
- Add JNI code for Unix domain datagram sockets
- Refactor `DatagramUnicastTest` so that it can be used for testing also Unix domain datagram sockets
- Add Unix domain datagram sockets implementation for KQueue transport
- Add Unix domain datagram sockets implementation for Epoll transport

Result:

Fixes #6737
2021-07-09 09:45:43 +02:00
Norman Maurer
e5adc1131a
Only do a partial build for a few jobs when verify PRs (#11464)
Motivation:

Currently the PR verification takes a lot of time. We should try to only run builds for the affected modules in most cases to speed up the build.

Modifications:

Adjust docker files and workflow to only run build for affected modules for a lot of jobs while still run full builds for windows + java11 with BoringSSL

Result:

Hopefully quicker feedback loop when verify PRs
2021-07-09 06:47:06 +02:00
Ikko Ashimine
d1ef28a22f
Fix typo in ReferenceCountedOpenSslEngine (#11467)
Motivation:

There should be no typos in comments

Modifications:
```
alway -> always
```

Result:

Fixed typo.
2021-07-08 16:29:58 +02:00
Norman Maurer
40fb6026ef
Introduce OpenSslAsyncPrivateKeyMethod which allows to asynchronously sign / decrypt the private key (#11390) (#11460)
Motivation:

At the moment we only support signing / decrypting the private key in a synchronous fashion. This is quite limited as we may want to do a network call to do so on a remote system for example.

Modifications:

- Update to latest netty-tcnative which supports running tasks in an asynchronous fashion.
- Add OpenSslAsyncPrivateKeyMethod interface
- Adjust SslHandler to be able to handle asynchronous task execution
- Adjust unit tests to test that asynchronous task execution works in all cases

Result:

Be able to asynchronous do key signing operations
2021-07-08 16:19:22 +02:00
Aayush Atharva
fef761d03e
Introduce BrotliEncoder (#11256)
Motivation:
Currently, Netty only has BrotliDecoder which can decode Brotli encoded data. However, BrotliEncoder is missing which will encode normal data to Brotli encoded data.

Modification:
Added BrotliEncoder and CompressionOption

Result:
Fixes #6899.

Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2021-07-08 11:51:27 +02:00
skyguard1
7825fa8d7a
Add zstd http header value (#11463)
Motivation:

ZSTD has a wide range of uses on the Internet, so should consider adding `application/zstd` HTTP media-type and `zstd` content-encoding, see  https://tools.ietf.org/html/rfc8478

Modification:

Add `application/zstd` HTTP media-type and `zstd` content-encoding

Result:

netty provides `application/zstd` HTTP media-type and `zstd content-encoding` as http headers

Signed-off-by: xingrufei <xingrufei@sogou-inc.com>
Co-authored-by: xingrufei <xingrufei@sogou-inc.com>
2021-07-08 11:44:54 +02:00
Norman Maurer
3e350993b7
Add script which can be used to build affected modules only (#11461)
Motivation:

At the moment we always build all modules. This script can be used to only build affected modules for a given change

Modifications:

Add script that will only build modules that are affected by a change

Result:

More targeted build
2021-07-08 10:18:16 +02:00
Norman Maurer
7e39b96402
Only run one SSL task per delegation (#11462)
Motivation:

We should only run one SSL task per delegation to allow more SSLEngines to make progress in a timely manner

Modifications:

- Only run one task per delegation to the executor
- Only create new SSL task if really needed
- Only schedule if not on the EventExecutor thread

Result:

More fair usage of resources and less allocations
2021-07-08 07:56:15 +02:00
Aayush Atharva
ede7a604f1
Add SslProtocols and Cipher suites constants (#11457)
Motivation:
Protocols and Cipher suites constants to prevent typos in protocol and cipher suites names and ease of use.

Modification:
Added Protocols and Cipher suites as constants in their respective classes.

Result:
Fixes #11393
2021-07-07 21:15:43 +02:00
Norman Maurer
0b2e955aff Revert "Introduce OpenSslAsyncPrivateKeyMethod which allows to asynchronously sign / decrypt the private key (#11390)"
This reverts commit 7c57c4be17.
2021-07-07 08:26:27 +02:00
Norman Maurer
ac91eaaae8
Ensure we only try to wrap if handler was not removed yet. (#11455)
Motivation:
7c57c4be17 did add a way to async sign keys but did not guard against the handler been removed before try to wrap in cause of an error which could lead to a harmless NPE.

Modifications:

Add check

Result:

No more harmless NPE
2021-07-06 15:01:49 +02:00
skyguard1
266c987339
[Feature] Add zstd encoder (#11437)
Motivation:

As discussed in #10422, ZstdEncoder can be added separately

Modification:

Add ZstdEncoder separately

Result:

netty supports ZSTD with ZstdEncoder

Signed-off-by: xingrufei <xingrufei@sogou-inc.com>
Co-authored-by: xingrufei <xingrufei@sogou-inc.com>
2021-07-06 14:57:09 +02:00
Nitesh Kant
f2295628e9
ApplicationProtocolNegotiationHandler should drain buffer messages on channel close (#11445)
__Motivation__

`ApplicationProtocolNegotiationHandler` buffers messages which are read before SSL handshake complete event is received and drains them when the handler is removed. However, the channel may be closed (or input shutdown) before SSL handshake  event is received in which case we may fire channel read after channel closure (from `handlerRemoved()`).

__Modification__

Intercept `channelInactive()` and input closed event and drain the buffer.

__Result__

If channel is closed before SSL handshake complete event is received, we still maintain the order of message read and channel closure.

Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2021-07-06 14:01:17 +02:00
Norman Maurer
7c57c4be17
Introduce OpenSslAsyncPrivateKeyMethod which allows to asynchronously sign / decrypt the private key (#11390)
Motivation:

At the moment we only support signing / decrypting the private key in a synchronous fashion. This is quite limited as we may want to do a network call to do so on a remote system for example.

Modifications:

- Update to latest netty-tcnative which supports running tasks in an asynchronous fashion.
- Add OpenSslAsyncPrivateKeyMethod interface
- Adjust SslHandler to be able to handle asynchronous task execution
- Adjust unit tests to test that asynchronous task execution works in all cases

Result:

Be able to asynchronous do key signing operations
2021-07-06 08:25:31 +02:00
Norman Maurer
ae7944459d Fix test failure introduced by 7cc31b8653 2021-07-05 08:37:41 +02:00
Norman Maurer
7cc31b8653
Use Junit5 for handler module (#11444)
Motivation:

We should aim to use junit5 everywhere

Modifications:

Migrate most of the handler module to use junit5

Result:

Part of #10757
2021-07-02 15:06:06 +02:00
Kasimir Torri
c97981403d
Improve PerMessageDeflateClientExtensionHandler (#11413)
Motivation:

The `PerMessageDeflateClientExtensionHandler` has the following strange behaviors currently:

* The `requestedServerNoContext` parameter doesn't actually add the `server_no_context_takeover` parameter to the client offer; instead it depends on the requested server window size.
* The handshake will fail if the server responds with a `server_no_context_takeover` parameter and `requestedServerNoContext` is false. According to RFC 7692 (7.1.1.1) the server may do this, and this means that to cover both cases one needs to use two handshakers in the channel pipeline: one with `requestedServerNoContext = true` and one with `requestedServerNoContext = false`.
* The value of the `server_max_window_bits` parameter in the server response is never checked (should be between 8 and 15). And the value of `client_max_window_bits` is checked only in the branch handling the server window parameter.

Modification:

* Add the `server_no_context_takeover` parameter if `requestedServerNoContext` is true.
* Accept a server handshake response which includes the server no context takeover parameter even if we did not request it.
* Check the values of the client and server window size in their respective branches and fail the handshake if they are out of bounds.

Result:

There will be no need to use two handshakers in the pipeline to be lenient in what handshakes are accepted.
2021-07-02 14:47:59 +02:00
Norman Maurer
842e73f8d3
Migrate microbenchmark to junit5 (#11440) (#11443)
Motivation:

We should update to use junit5 in all modules.

Modifications:

Adjust microbenchmark to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-07-02 08:05:03 +02:00
Norman Maurer
7c6be99da0
Migrate rest of epoll module to junit5 (#11442)
Motivation:

We should update to use junit5 in all modules.

Modifications:

Adjust missing epoll tests to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-07-02 08:04:48 +02:00
Norman Maurer
5179e53294
Migrate codec-http to junit5 (#11440)
Motivation:

We should update to use junit5 in all modules.

Modifications:

Adjust codec-http tests to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-07-01 18:53:41 +02:00
Norman Maurer
c82c17782c
Migrate common to junit5 (#11431) (#11438)
Motivation:

We should update to use junit5 in all modules.

Modifications:

Adjust common tests to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-07-01 18:19:44 +02:00
Norman Maurer
8003ea8a03
Migrate transport-udt to junit5 (#11439)
Motivation:

We should update to use junit5 in all modules.

Modifications:

Adjust transport-udt tests to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-07-01 18:19:24 +02:00
Aleksandar Gradinac
bc206ff173
Delay initialization of io.netty.handler.codec.compression.BrotliDecoder to runtime (#11428)
Motivation:
Including codec-http in the project and building a native-image out of it using a GraalVM 21.2 nightly can result in a failure.

Modification:
By delaying the initialization of `io.netty.handler.codec.compression.BrotliDecoder` to runtime, native-image will not try to eagerly initialize the class during the image build, avoiding the build failure described in the issue.

Result:
Fixes #11427
2021-07-01 15:08:51 +02:00
Aayush Atharva
750d23583c
Add ALPN Buffering to support HTTP/2 Prior Knowledge (#11407)
Motivation:
Currently, Netty cannot handle HTTP/2 Preface messages if the client used the Prior knowledge technique. In Prior knowledge, the client sends an HTTP/2 preface message immediately after finishing TLS Handshake. But in Netty, when TLS Handshake is finished, ALPNHandler is triggered to configure the pipeline. And between these 2 operations, if an HTTP/2 preface message arrives, it gets dropped.

Modification:

Buffer messages until we are done with the ALPN handling.

Result:
Fixes #11403.

Co-authored-by: Norman Maurer <norman_maurer@apple.com>
2021-07-01 14:10:52 +02:00
Norman Maurer
126e9f873b
Revert "Fix deprecated GC logging switch (#11395)" (#11436)
This reverts commit 077e5f282c.
2021-07-01 12:09:52 +02:00
Alex Budovski
077e5f282c
Fix deprecated GC logging switch (#11395)
Motivation:

This caused test failures due to the deprecation warning and produced a
dumpstream.

Modification:

Replace deprecated flag with recommended one.

Result:
Fix deprecation and cause of test failure in codec project.
2021-07-01 11:30:30 +02:00
Maksym Ostroverkhov
2abe20a6b5
Tail tasks queue: configure separately from tasks queue (#11400)
Motivation:

IO transports (primarily epoll, but also applies to kqueue, nio) cant be configured with separate tail tasks queue factory -
instead single queue factory is used for both normal tasks and tail tasks.

Modifications:

Add constructor accepting tail EventLoopTaskQueueFactory to aforementioned transports

Result:

IO transports can be configured with separate tail tasks
2021-07-01 10:51:10 +02:00
Norman Maurer
20e4ccbd33
Skip Brotli related tests on platforms which not support Brotli (#11435)
Motivation:

The native module is not yet available on aarch64 Mac / Windows thus causing tests in codec/ to fail (specifically all the Brotli ones, since the module could not be loaded).

Modification:

Disable Brotli tests when platform is not supported

Result:

Tests under codec/ now pass under Mac/aarch64 and Windows/aarch64
2021-07-01 09:59:24 +02:00
wujimin
3226e77485
Add support for GMSSL (#11406) (#11410)
__Motivation__

Add support for GMSSL protocol to SslUtils.

__Modification__

Modify `SslUtils.getEncryptedPacketLength(ByteBuf buffer, int offset)` to get packet length when protocol is GMSSL.
Modify `SslUtils.getEncryptedPacketLength(ByteBuffer buffer)` to get packet length when protocol is GMSSL.

__Result__

`SslUtils.getEncryptedPacketLength` now supports GMSSL protocol. Fixes https://github.com/netty/netty/issues/11406
2021-07-01 08:17:45 +02:00
Stephane Landelle
801819b359
Fix HttpHeaderValue#XML_HTTP_REQUEST case (#11433)
Motivation:

HTTP header values are case sensitive. The expected value for `x-request-with` header is `XMLHttpRequest`, not `XmlHttpRequest`.

Modification:

Fix constant's case.

Result:

Correct `XMLHttpRequest` HTTP header value.
2021-07-01 08:13:10 +02:00
Tamara Braun
194a81ff4f
Fix Buffer Overflow During Lz4FrameEncoder Close (#11429)
Motivation:

We failed to account for the last header when estimating the buffer
size. If the data does not compress enough to make space for the
last header we would exceed the ByteBuf's capacity.

Modifications:

Call #ensureWritable with appropriate capacity for footer ByteBuf
befor writing footer.

Result:

If there is not enough space left in the buffer, the buffer will be
expanded.
2021-06-30 18:47:07 +02:00
Norman Maurer
8d76f402b1
Migrate transport to junit5 (#11431) (#11432)
Motivation:

We should update to use junit5 in all modules.

Modifications:

Adjust transport tests to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-06-30 16:17:57 +02:00
Norman Maurer
94a4880358
Migrate codec-mqtt to junit5 (#11431)
Motivation:

We should update to use junit5 in all modules.

Modifications:

Adjust codec-mqtt tests to use junit5

Result:

Part of https://github.com/netty/netty/issues/10757
2021-06-30 15:32:16 +02:00