Motivation:
The MacOS-specific `connectx(2)` system call make it possible to establish client-side connections with TCP FastOpen.
Modification:
Add support for TCP FastOpen to the KQueue transport, and add the `connectx(2)` system call to `BsdSocket`.
Result:
It's now possible to use TCP FastOpen when initiating connections on MacOS.
Motivation:
This test is inherently flaky due to file descriptor reuse.
Even though we have taken steps to make it less flaky, it still fails sometimes.
When it does, the error message is not very helpful.
Modification:
Make use of assertThrows and assertThat to get more descriptive error messages when the tests fail.
Result:
More meaningful messages on test failures, which may help us make the tests more resilient in the future
Motivation:
There are lots of redundant variable declarations which should be inlined to make good look better.
Modification:
Made variables inlined.
Result:
Less redundant variable and more readable code.
Motivation:
TCP FastOpen is a pure optimisation, that is opportunistically applied.
There is no reason to make it specific to the epoll transport, and in the future we could add support to other transports.
Besides, the client-side equivalent, TCP_FASTOPEN_CONNECT, is already transport agnostic.
Modification:
Move the TCP_FASTOPEN channel option from EpollChannelOption to ChannelOption.
Mark the field in EpollChannelOption as deprecated.
Result:
All channel options related to TCP FastOpen are now transport agnostic.
However, we still only actually support TFO on the epoll transport.
Motivation:
We cannot control when "the system" reuses file descriptors.
This makes any test that assert on the behaviour of closed file descriptors inherently racy.
Modification:
Allow the EpollSocketChannelConfigTest socketoption tests a few tries to get the correct assertion on the behaviour of closed socket file descriptors.
Result:
The EpollSocketChannelConfigTest should now be much less flaky.
Motivation:
There are lots of imports which are unused. We should get rid of them to make the code look better,
Modification:
Removed unused imports.
Result:
No unused imports.
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
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
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
Motivation:
JUnit 5 is more expressive, extensible, and composable in many ways, and it's better able to run tests in parallel.
Modifications:
Use JUnit5 in tests
Result:
Related to https://github.com/netty/netty/issues/10757
Motivation:
We used assumeTrue(...) in some places before to detect if we could load the native library but this could lead to the sitation that we not notice if we break native loading.
Modifications:
Always fail if we cant load the native library
Result:
Ensure we not cause any regression in the native loading code in the future
Motivation:
`PlatformDependent#normalizedOs()` already caches normalized variant of
the value of `os.name` system property. Instead of inconsistently
normalizing it in every case, use the utility method.
Modifications:
- `PlatformDependent`: `isWindows0()` and `isOsx0()` use `NORMALIZED_OS`;
- `PlatformDependent#normalizeOs(String)` define `darwin` as `osx`;
- `OpenSsl#loadTcNative()` does not require `equalsIgnoreCase` bcz `os`
is already normalized;
- Epoll and KQueue: `Native#loadNativeLibrary()` use `normalizedOs()`;
- Use consistent `Locale.US` for lower case conversion of `os.name`;
- `MacOSDnsServerAddressStreamProvider#loadNativeLibrary()` uses
`PlatformDependent.isOsx()`;
Result:
Consistent approach for `os.name` parsing.
Motivation:
It turns out it is quite easy to cause a classloader deadlock in more recent java updates if you cause classloading while you are in native code. Because of this we should just workaround this issue by pre-load all the classes that needs to be accessed in the OnLoad function.
Modifications:
- Preload all classes that would otherwise be loaded by native OnLoad functions.
Result:
Workaround for https://github.com/netty/netty/issues/11209 and https://bugs.openjdk.java.net/browse/JDK-8266310
Motivation:
NullChecks resulting in a NullPointerException or IllegalArgumentException, numeric ranges (>0, >=0) checks, not empty strings/arrays checks must never be anonymous but with the parameter or variable name which is checked. They must be specific and should not be done with an "OR-Logic" (if a == null || b == null) throw new NullPointerEx.
Modifications:
* import static relevant checks
* Replace manual checks with ObjectUtil methods
Result:
All checks needed are done with ObjectUtil, some exception texts are improved.
Fixes#11170
Motivation:
While adding support for GRO (b05fdf3ff8) we broke support for IP_RECVORIGDSTADDR when using the native transport. Beside this we also didnt correctly handle IP_RECVORIGDSTADDR when recvmmsg was used.
Modifications:
- Fix support for IP_RECVORIGDSTADDR when using the native epoll transport for normal reads (recvmsg) but also for scattering reads (recvmmsg)
- Remove code from unix code-base as the support is linux specific and we not need the code there anymore
Result:
Fixes https://github.com/netty/netty/issues/11141
Motivation:
We should avoid blocking in the event loop as much as possible.
The InputStream.read() is a blocking method, and we don't need to call it if available() returns a positive number.
Modification:
Bypass calling InputStream.read() if available() returns a positive number.
Result:
Fewer blocking calls in the event loop, in general, when ChunkedStream is used.
Motivation:
As we can supported SegmentedDatagramPacket in multiple native
transports (like in epoll and io_uring) we should just move it to
unix-common so we can share code.
Modification:
- Move SegmentedDatagrampPacket to transport-native-unixu
- Mark the SegmentedDatagramPacket in epoll as deprecated
- Update code to use updated package.
Result:
Possibility of code re-use
Motivation:
UDP_GRO can improve performance when reading UDP datagrams. This patch adds support for it.
See https://lwn.net/Articles/768995/
Modifications:
- Add recvmsg(...)
- Add support for UDP_GRO in recvmsg(...) and recvmmsg(...)
- Remove usage of recvfrom(...) and just always use recvmsg(...) or recvmmsg(...) to simplify things
- Refactor some code for sharing
- Add EpollChannelOption.UDP_GRO and the getter / setter in EpollDatagramConfig
Result:
UDP_GRO is supported when the underlying system supports it.
netty-jni-util 0.0.2.Final is incompatible with static linking. Before
the netty-jni-util dependency was introduced netty-tcnative supported
static linking via NETTY_BUILD_STATIC. netty-jni-util 0.0.3.Final adds
static linking compatibility.
Modifications:
Bump netty-jni-util to version 0.0.3.Final and update to its new API
which requires the caller to manage packagePrefix.
Result:
Using latest version of netty-jni-util and restored static linking
compatibility.
Motivation:
LSE (https://mysqlonarm.github.io/ARM-LSE-and-MySQL/) can have a huge performance difference. Let's ensure we use a compiler that can support it.
Modifications:
Update to gc10 when cross-compiling as it supports LSE and enables it by default
Result:
More optimized builds for aarch64
... number of bytes when using DatagramChannels
Motivation:
In our FixedRecvByteBufAllocator we dont continue to read if the number of bytes is less then what was configured. This is correct when using it for TCP but not when using it for UDP. When using UDP the number of bytes is the maximum of what we want to support but we often end up processing smaller datagrams in general. Because of this we should use contineReading(UncheckedBooleanSupplier) to determite if we should continue reading
Modifications:
- use contineReading(UncheckedBooleanSupplier) for DatagramChannels
Result:
Read more then once in the general case for DatagramChannels with the default config
Motivation:
Allow to configure the maximum number of messages to write per eventloop run. This can be useful to ensure we read data in a timely manner and not let writes dominate the CPU time. This is especially useful in protocols like QUIC where you need to read "fast enough" as otherwise you may not read the ACKs fast enough.
Modifications:
- Add new ChannelOption / config that allows to limit the number of messages to write per eventloop run.
- Respect this setting for DatagramChannels
Result:
Reduce the risk of having WRITES block the processing of other events in a timely manner
Co-authored-by: terrarier2111 <58695553+terrarier2111@users.noreply.github.com>
Motivation:
SslHandler owns the responsibility to flush non-application data
(e.g. handshake, renegotiation, etc.) to the socket. However when
TCP Fast Open is supported but the client_hello cannot be written
in the SYN the client_hello may not always be flushed. SslHandler
may not wrap/flush previously written/flushed data in the event
it was not able to be wrapped due to NEED_UNWRAP state being
encountered in wrap (e.g. peer initiated renegotiation).
Modifications:
- SslHandler to flush in channelActive() if TFO is enabled and
the client_hello cannot be written in the SYN.
- SslHandler to wrap application data after non-application data
wrap and handshake status is FINISHED.
- SocketSslEchoTest only flushes when writes are done, and waits
for the handshake to complete before writing.
Result:
SslHandler flushes handshake data for TFO, and previously flushed
application data after peer initiated renegotiation finishes.
Motivation:
The EpollSocketConnectTest was not correctly configuring TCP Fast Open on the server socket.
It's an option, not a child option.
Modification:
EpollSocketConnectTest now correctly enables TCP Fast Open on the server side, when available, for the test that needs it.
Result:
Test covers what it was intended to.
Motivation:
c22c6b845d introduced support for
UDP_SEGMENT but did restrict it to continous buffers. This is not needed
as it is also fine to use CompositeByteBuf
Modifications:
- Allow to use CompositeByteBuf as well
- Add unit test
Result:
More flexible usage of segmented datagrams possible
Motivation:
For protocols like QUIC using UDP_SEGMENT (GSO) can help to reduce the
overhead quite a bit. We should support it.
Modifications:
- Add a SegmentedDatagramPacket which can be used to use UDP_SEGMENT
- Add unit test
Result:
Be able to make use of UDP_SEGMENT
Motivation:
#10995
when `io.netty.channel.unix.Socket` is ipv6 and join a multicast group with ipv4 address will cause `io.netty.channel.ChannelException: setsockopt() failed: Invalid argument` (at least in `Linux centos.dev 4.18.0-240.10.1.el8_3.x86_64 #1 SMP Mon Jan 18 17:05:51 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux`)
Modification:
check if target group address is ipv6 before call `io.netty.channel.epoll.LinuxSocket#joinGroup(java.net.InetAddress, java.net.NetworkInterface, java.net.InetAddress)`
I'm not sure if this modification is currect, but i checked source code of java NIO
```
Java_sun_nio_ch_Net_canJoin6WithIPv4Group0(JNIEnv* env, jclass cl)
{
#if defined(__APPLE__)
/* IPV6_ADD_MEMBERSHIP can be used to join IPv4 multicast groups */
return JNI_TRUE;
#else
/* IPV6_ADD_MEMBERSHIP cannot be used to join IPv4 multicast groups */
return JNI_FALSE;
#endif
}
```
seems ipv6 address can't join ipv4 group except osx
Result:
test on `Linux 3.10.0-514.el7.x86_64 #1 SMP Tue Nov 22 16:42:41 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux` exception ` setsockopt() failed: Invalid argument` has fixed
Fixes#10995
Support TCP Fast Open for clients and make SslHandler take advantage
Motivation:
- TCP Fast Open allow us to send a small amount of data along side the initial SYN packet when establishing a TCP connection.
- The TLS Client Hello packet is small enough to fit in there, and is also idempotent (another requirement for using TCP Fast Open), so if we can save a round-trip when establishing TLS connections when using TFO.
Modification:
- Add support for client-side TCP Fast Open for Epoll, and also lowers the Linux kernel version requirements to 3.6.
- When adding the SslHandler to a pipeline, if TCP Fast Open is enabled for the channel (and the channel is not already active) then start the handshake early by writing it to the outbound buffer.
- An important detail to note here, is that the outbound buffer is not flushed at this point, like it would for normal handshakes. The flushing happens later as part of establishing the TCP connection.
Result:
- It is now possible for clients (on epoll) to open connections with TCP Fast Open.
- The SslHandler automatically detects when this is the case, and now send its Client Hello message as part of the initial data in the TCP Fast Open flow when available, saving a round-trip when establishing TLS connections.
Co-authored-by: Colin Godsey <crgodsey@gmail.com>