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>
Motivation:
File.createTempFile(String, String)` will create a temporary file in the system temporary directory if the 'java.io.tmpdir'. The permissions on that file utilize the umask. In a majority of cases, this means that the file that java creates has the permissions: `-rw-r--r--`, thus, any other local user on that system can read the contents of that file.
This can be a security concern if any sensitive data is stored in this file.
This was reported by Jonathan Leitschuh <jonathan.leitschuh@gmail.com> as a security problem.
Modifications:
Use Files.createTempFile(...) which will use safe-defaults when running on java 7 and later. If running on java 6 there isnt much we can do, which is fair enough as java 6 shouldnt be considered "safe" anyway.
Result:
Create temporary files with sane permissions by default.
Motiviation:
We need to ensure we only register the methods for unix-native-common once as otherwise it may have strange side-effects.
Modifications:
- Add extra method that should be called to signal that we need to register the methods. The registration will only happen once.
- Adjust code to make use of it.
Result:
No more problems due incorrect registration of these methods.
Motivation:
This reverts commit 7fb62a93b8 as it broke native loading in some cases due maven dependencies.
Modification:
Revert the commit.
Result:
Native loading works again
Motivation:
Android seems to use a different field name so we should also try to access it with the name used by android.
Modifications:
Try first fd and if this fails try descriptor as field name
Result:
Workaround for android.
Motivation:
We need to ensure we only register native methods once as otherwise we may end up in an "invalid" state. The problem here was that before it was basically the responsibility the user of transport-native-unix-common to register the methods. This is error prone as there may be multiple users of these on the classpath at the same time.
Modifications:
- Provide a way to init native lib without register the native methods of the provided classes. This is needed to be able to re-use functionality which is exposed to our internal native code
- Use flatten plugin to correctly resolve classifier and so have the correct dependency
- Call Unix.* method to ensure we register the methods correctly once
- Include native lib as well in the native jars of unix-common
Result:
Be able to have multiple artifacts of the classpath that depends on the unix-common. Related to https://github.com/netty/netty-incubator-transport-io_uring/issues/15
Motivation:
We should use aarch_64 in our classifier / jni libname on aarch64 as os.detected.arch uses the name. Being non consistent (especially across our different projects) already gave us a lot of trouble in the past.
Let's fix this once for all.
Modifications:
Use aarch_64
Result:
More consistent classifier usage on aarch64
Motivation:
In some enviroments sun.misc.Unsafe is not present. We should support these as well.
Modifications:
Fallback to JNI if we can't directly access the memoryAddress of the buffer.
Result:
Fixes https://github.com/netty/netty/issues/10813
Motivation:
https in xmlns URIs does not work and will let the maven release plugin fail:
```
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 1.779 s
[INFO] Finished at: 2020-11-10T07:45:21Z
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-release-plugin:2.5.3:prepare (default-cli) on project netty-parent: Execution default-cli of goal org.apache.maven.plugins:maven-release-plugin:2.5.3:prepare failed: The namespace xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" could not be added as a namespace to "project": The namespace prefix "xsi" collides with an additional namespace declared by the element -> [Help 1]
[ERROR]
```
See also https://issues.apache.org/jira/browse/HBASE-24014.
Modifications:
Use http for xmlns
Result:
Be able to use maven release plugin
Motivation:
We had a lot of duplication in our jni code which was mostly due macros but also related to how we support shading. By using netty-jni-util we can share all the code between netty and netty-tcnative ( and possible other jni based netty projects in the future).
Modifications:
- Use netty-jni-util and re-use its macros / functions
- Remove duplicated code
- Adjust build files
Result:
Less code duplication for JNI
Motivation:
It's important to unload all previous registered native methods when there is a failure during loading the native lib. Failing to do so may lead to an "invalid state" and so may segfault the JVM when trying to call a native method that was previous loaded.
This was observed when two versions of netty-tcnative were on the classpath which had different requirements in terms of linking.
Something like this was reported in he hs log:
```
Instructions: (pc=0x0000000116413bf0)
0x0000000116413bd0:
[error occurred during error reporting (printing registers, top of stack, instructions near pc), id 0xb]
Register to memory mapping:
RAX=0x0000000116413bf0 is an unknown value
RBX={method} {0x000000011422e708} 'aprMajorVersion' '()I' in 'io/netty/internal/tcnative/Library'
RCX=0x000000000000000a is an unknown value
RDX=0x000000000000000a is an unknown value
```
Modifications:
- Unregister previous registered native methods on failure
- Unregister previous registered native methods on on unload of the native lib
Result:
No more segfault caused by invalid state when loading of the native lib fails in between. In this case the user will receive an error now like:
Motivation:
Thread.stop() works by producing a ThreadDeath error in the target thread. EventLoops swallow all Throwables, which makes them effectively unkillable. This is effectively a memory leak, for our application. Beside this we should also just regrow all `Error` as there is almost no way to recover.
Modification:
Edit the EventLoops that swallow Throwables to instead rethrow Error.
Result:
`EventLoop` can crash if `Error` is thrown
Motivation:
HTTP is a plaintext protocol which means that someone may be able
to eavesdrop the data. To prevent this, HTTPS should be used whenever
possible. However, maintaining using https:// in all URLs may be
difficult. The nohttp tool can help here. The tool scans all the files
in a repository and reports where http:// is used.
Modifications:
- Added nohttp (via checkstyle) into the build process.
- Suppressed findings for the websites
that don't support HTTPS or that are not reachable
Result:
- Prevent using HTTP in the future.
- Encourage users to use HTTPS when they follow the links they found in
the code.
Motivation:
DuplexChannel allow for half-closure, we should have a special config interface for it as well.
Modifications:
Add DuplexChannelConfig which allows to configure half-closure.
Result:
More consistent types
Motivation:
LGTM reports multiple issues. They need to be triaged,
and real ones should be fixed.
Modifications:
- Fixed multiple issues reported by LGTM, such as redundant conditions,
resource leaks, typos, possible integer overflows.
- Suppressed false-positives.
- Added a few testcases.
Result:
Fixed several possible issues, get rid of false alarms in the LGTM report.
Motivation:
All scheduled executors should behave in accordance to their API.
The bug here is that scheduled tasks were not run more than once because we executed the runnables directly, instead of through the provided runnable future.
Modification:
We now run tasks through the provided future, so that when each run completes, the internal state of the task is reset and the ScheduledThreadPoolExecutor is informed of the completion.
This allows the executor to prepare the next run.
Result:
The UnorderedThreadPoolEventExecutor is now able to run scheduled tasks more than once.
Which is what one would expect from the API.
Motivation:
Creating exceptions is expensive so we should only do so if really needed.
Modifications:
Only create the ConnectTimeoutException if we really need it.
Result:
Less overhead
Motivation:
When we try to parse the kernel version we need to be careful what to
expect. Especially when a custom kernel is used we may get extra chars
in the version numbers. For example I had this one fail because of my
custom kernel that I built for io_uring:
5.8.7ioring-fixes+
Modifications:
- Try to be a bit more lenient when parsing
- If we cant parse the kernel version just use 0.0.0
Result:
Tests are more robust
only
Motivation:
4b7dba1 introduced a change which was not 100 % complete and so
introduce a regression when a user specified to use
InetProtocolFamily.IPv4 and trying to bind to a port (without specify
the ip).
Modifications:
- Fix regression by respect the InetProtocolFamily
- Add unit test
Result:
Fix regression when binding to port explicit
Motivation:
When we were using the netty http protocol, OOM occurred, this problem has been in 4.1.51.Final Fix [# 10424](https://github.com/netty/netty/issues/10424), even if OOM is up, the service will still receive new connection events, will occur again OOM and eventually cause the connection not to be released.
code `byteBuf = allocHandle.allocate(allocator);`
Modification:
I fail to create buffer when I try to receive new data, i determine if it is OOM then the close read event releases the connection.
```java
if (close || cause instanceof OutOfMemoryError || cause instanceof IOException) {
closeOnRead(pipeline);
}
```
Result:
Fixes # [10434](https://github.com/netty/netty/issues/10434).
Motivation:
Even if the system does not support ipv6 we should try to use it if the user explicit pass an Inet6Address. This way we ensure we fail and not try to convert this to an ipv4 address internally.
This incorrect behavior was introduced by 70731bfa7e
Modifications:
If the user explicit passed an Inet6Address we force the usage of ipv6
Result:
Fixes https://github.com/netty/netty/issues/10402
Motivation:
`transport-native-epoll` doesn't have ARM release package.
Modification:
This PR added cross compile profile for epoll. Then we can easily build aarch64 package on X86 machine.
Result:
Fixes#8279
Motivation:
netty_epoll_linuxsocket_JNI_OnLoad(...) may produce a deadlock with another thread that will load IOUtil in a static block. This seems to be a JDK bug which is not yet fixed. To workaround this we force IOUtil to be loaded from without java code before init the JNI code
Modifications:
Use Selector.open() as a workaround to load IOUtil
Result:
Fixes https://github.com/netty/netty/issues/10187
Motivation:
8dc6ad5 introduced IPV6-mapped-IPV4 address support but
copied the addresses incorrectly. It copied the first
4 bytes of the ipv6 address to the address byte array
at offset 12, instead of the other way around.
7a547aa implemented this correctly in netty_unix_socket.c
but it seems the change should've been applied to
netty_epoll_native.c as well.
The current behaviour will always set the address to
`0.0.0.0`.
Modifications:
Copy the correct bytes from the ipv6 mapped ipv4 address.
I.e. copy 4 bytes at offset 12 from the native address
to the byte array `addr` at offset 0.
Result:
When using recvmmsg with IPV6-mapped-IPV4 addresses,
the address will be correctly copied to the byte array
`addr` in the NativeDatagramPacket instance.
Motivation:
https://github.com/netty/netty/pull/9797 changed the code for recvmmsg and sendmmsg to use the syscalls directly to remvove the dependency on newer GLIBC versions. Unfortunally it made the assumption that the syscall numbers are the same for different architectures, which is not the case.
Thanks to @jayv for pointing it out
Modifications:
Add #if, #elif and #else declarations to ensure we pick the correct syscall number (or not support if if the architecture is not supported atm).
Result:
Pick the correct syscall number depending on the architecture.
Motivation:
Due a bug we did not correctly set the writerIndex of the ByteBuf when a
user specified EpollChannelOption.MAX_DATAGRAM_PAYLOAD_SIZE but we ended
up with a non scattering read.
Modifications:
- Set writerIndex to the correct value
- Add unit tests
Result:
Fixes https://github.com/netty/netty/issues/9788
Motivation
The event loop implementations had become somewhat tangled over time and
work was done recently to streamline EpollEventLoop. NioEventLoop would
benefit from the same treatment and it is more straighforward now that
we can follow the same structure as was done for epoll.
Modifications
Untangle NioEventLoop logic and mirror what's now done in EpollEventLoop
w.r.t. the volatile selector wake-up guard and scheduled task deadline
handling.
Some common refinements to EpollEventLoop have also been included - to
use constants for the "special" deadline/wakeup volatile values and to
avoid some unnecessary calls to System.nanoTime() on task-only
iterations.
Result
Hopefully cleaner, more efficient and less fragile NIO transport
implementation.
Motivation:
394a1b3485 introduced a hard dependency on GLIBC 2.12 which was not the case before. This had the effect of not be able to use the native epoll transports on platforms which ship with earlier versions of GLIBC.
To make things a backward compatible as possible we should not introduce such changes in a bugfix release.
Special thanks to @weissi with all the help to fix this.
Modifications:
- Use syscalls directly to remove dependency on GLIBC 2.12
- Make code consistent that needs newer GLIBC versions
- Adjust scattering read test to only run if recvmmsg syscall is supported
- Cleanup pom.xml as some stuff is not needed anymore after using syscalls.
Result:
Fixes https://github.com/netty/netty/issues/9758.
Motivation:
In most cases, we want to use MultithreadEventLoopGroup such as NioEventLoopGroup without setting thread numbers but thread name only. So we need to use followed code:
NioEventLoopGroup boss = new NioEventLoopGroup(0, new DefaultThreadFactory("boss"));
It looks a bit confuse or strange for the number 0 due to we only want to set thread name. So it will be better to add new constructor for this case.
Modifications:
add new constructor into all event loop groups, for example: public NioEventLoopGroup(ThreadFactory threadFactory)
Result:
User can only set thread factory without setting the thread number to 0:
NioEventLoopGroup boss = new NioEventLoopGroup(new DefaultThreadFactory("boss"));
Motivation
The recently-introduced event loop scheduling hooks can be exploited by
the epoll transport to avoid waking the event loop when scheduling
future tasks if there is a timer already set to wake up sooner.
There is also a "default" timeout which will wake the event
loop after 1 second if there are no pending future tasks. The
performance impact of these wakeups themselves is likely negligible but
there's significant overhead in having to re-arm the timer every time
the event loop goes to sleep (see #7816). It's not 100% clear why this
timeout was there originally but we're sure it's no longer needed.
Modification
Combine the existing volatile wakenUp and non-volatile prevDeadlineNanos
fields into a single AtomicLong that stores the next scheduled wakeup
time while the event loop is in epoll_wait, and is -1 while it is awake.
Use this as a guard to debounce wakeups from both immediate scheduled
tasks and future scheduled tasks, the latter using the new
before/afterScheduledTaskSubmitted overrides and based on whether the
new deadline occurs prior to an already-scheduled timer.
A similar optimization was already added to NioEventLoop, but it still
uses two separate volatiles. We should consider similar streamlining of
that in a future update.
Result
Fewer event loop wakeups when scheduling future tasks, greatly reduced
overhead when no future tasks are scheduled.
Motivation:
There is a goto statement above the current position of initialize dynamicMethods, and dynamicMethods is used after the goto which might cause undefined behavior.
Modifications:
Initialize dynamicMehtods at the top.
Result:
No more undefined behavior.
Motivation
This is a vestige that was removed in the original PR #9535 before it
was reverted, but we missed it when re-applying in #9586.
It means there is a possible race condition because a wakeup event could
be missed while shutting down, but the consequences aren't serious since
there's a 1 second safeguard timeout when waiting for it.
Modification
Remove call to epollWaitNow() in EpollEventLoop#closeAll()
Result
Cleanup redundant code, avoid shutdown delay race condition
Motivation
The current event loop shutdown logic is quite fragile and in the
epoll/NIO cases relies on the default 1 second wait/select timeout that
applies when there are no scheduled tasks. Without this default timeout
the shutdown would hang indefinitely.
The timeout only takes effect in this case because queued scheduled
tasks are first cancelled in
SingleThreadEventExecutor#confirmShutdown(), but I _think_ even this
isn't robust, since the main task queue is subsequently serviced which
could result in some new scheduled task being queued with much later
deadline.
It also means shutdowns are unnecessarily delayed by up to 1 second.
Modifications
- Add/extend unit tests to expose the issue
- Adjust SingleThreadEventExecutor shutdown and confirmShutdown methods
to explicitly add no-op tasks to the taskQueue so that the subsequent
event loop iteration doesn't enter blocking wait (as looks like was
originally intended)
Results
Faster and more robust shutdown of event loops, allows removal of the
default wait timeout
Motivation:
The build script for the module Netty Transport Native Epoll can cause
intermittent build break due to broken pipe by /usr/bin/ldd. This issue
likely to occur on a build environment with multiple processors.
Modifications:
The root cause is that the consumer head command finishes earlier the
producer ldd command. Buffering the outputs of the ldd command by an
intermediate tail command avoids the broken pipe.
Result:
A build on multiple processors can finish successfully.
Signed-off-by: Tatsushi Inagaki <e29253@jp.ibm.com>
Motivation:
At the moment we not consistently (and also not correctly) free allocated native memory in all cases during loading the JNI library. This can lead to native memory leaks in the unlikely case of failure while trying to load the library.
Beside this we also not always correctly handle the case when a new java object can not be created in native code because of out of memory.
Modification:
- Copy some macros from netty-tcnative to be able to handle errors in a more easy fashion
- Correctly account for New* functions to return NULL
- Share code
Result:
More robust and clean JNI code
Motivation
This is another iteration of #9476.
Modifications
Instead of maintaining a count of all writes performed and then using
reads during shutdown to ensure all are accounted for, just set a flag
after each write and don't reset it until the corresponding event has
been returned from epoll_wait.
This requires that while a write is still pending we don't reset
wakenUp, i.e. continue to block writes from the wakeup() method.
Result
Race condition eliminated. Fixes#9362
Co-authored-by: Norman Maurer <norman_maurer@apple.com>
Motivation:
Running tests with a `KQueueDomainSocketChannel` showed worse performance than an `NioSocketChannel`. It turns out that the default send buffer size for Nio sockets is 64k while for KQueue sockets it's 8k. I verified that manually setting the socket's send buffer size improved perf to expected levels.
Modification:
Plumb the `SO_SNDBUF` and `SO_RCVBUF` options into the `*DomainSocketChannelConfig`.
Result:
Can now configure send and receive buffer sizes for domain sockets.
Motivation
Currently an epoll_ctl syscall is made every time there is a change to
the event interest flags (EPOLLIN, EPOLLOUT, etc) of a channel. These
are only done in the event loop so can be aggregated into 0 or 1 such
calls per channel prior to the next call to epoll_wait.
Modifications
I think further streamlining/simplification is possible but for now I've
tried to minimize structural changes and added the aggregation beneath
the existing flag manipulation logic.
A new AbstractChannel#activeFlags field records the flags last set on
the epoll fd for that channel. Calls to setFlag/clearFlag update the
flags field as before but instead of calling epoll_ctl immediately, just
set or clear a bit for the channel in a new bitset in the associated
EpollEventLoop to reflect whether there's any change to the last set
value.
Prior to calling epoll_wait the event loop makes the appropriate
epoll_ctl(EPOLL_CTL_MOD) call once for each channel who's bit is set.
Result
Fewer syscalls, particularly in some auto-read=false cases. Simplified
error handling from centralization of these calls.
Motivation:
We should correctly reset the cached local and remote address when a Channel.disconnect() is called and the channel has a notion of disconnect vs close (for example DatagramChannel implementations).
Modifications:
- Correctly reset cached kicak abd remote address
- Update testcase to cover it and so ensure all transports work in a consistent way
Result:
Correctly handle disconnect()
Motivation:
Changes that were done to the EpollEventLoop to optimize some things did break some testsuite and caused timeouts. We need to investigate to see why this is the case but for
now we should just revert so we can do a release.
Modifivations:
- Partly revert 1fa7a5e697 and a22d4ba859
Result:
Testsuites pass again.
Motivation:
291f80733a introduced a change to use a byte[] to construct the InetAddress when receiving datagram messages to reduce the overhead. Unfortunally it introduced a regression when handling IPv6-mapped-IPv4 addresses and so produced an IndexOutOfBoundsException when trying to fill the byte[] in native code.
Modifications:
- Correctly use the offset on the pointer of the address.
- Add testcase
- Make tests more robust and include more details when the test fails
Result:
No more IndexOutOfBoundsException
Motivation:
394a1b3485 added support for recvmmsg(...) for unconnected datagram channels, this change also allows to use recvmmsg(...) with connected datagram channels.
Modifications:
- Always try to use recvmmsg(...) if configured to do so
- Adjust unit test to cover it
Result:
Less syscalls when reading datagram packets
Motivation:
We should also use sendmmsg on connected channels whenever possible to reduce the overhead of syscalls.
Modifications:
No matter if the channel is connected or not try to use sendmmsg when supported to reduce the overhead of syscalls
Result:
Better performance on connected UDP channels due less syscalls
Motivation:
394a1b3485 introduced the possibility to use recvmmsg(...) but did not correctly handle ipv6 mapped ip4 addresses to make it consistent with other transports.
Modifications:
- Correctly handle ipv6 mapped ipv4 addresses by only copy over the relevant bytes
- Small improvement on how to detect ipv6 mapped ipv4 addresses by using memcmp and not byte by byte compare
- Adjust test to cover this bug
Result:
Correctly handle ipv6 mapped ipv4 addresses
Motivation
This is another iteration of #9476.
Modifications
Instead of maintaining a count of all writes performed and then using
reads during shutdown to ensure all are accounted for, just set a flag
after each write and don't reset it until the corresponding event has
been returned from epoll_wait.
This requires that while a write is still pending we don't reset
wakenUp, i.e. continue to block writes from the wakeup() method.
Result
Race condition eliminated. Fixes#9362
Motivation:
When using datagram sockets which need to handle a lot of packets it makes sense to use recvmmsg to be able to read multiple datagram packets with one syscall.
Modifications:
- Add support for recvmmsg on linux
- Add new EpollChannelOption.MAX_DATAGRAM_PACKET_SIZE
- Add tests
Result:
Fixes https://github.com/netty/netty/issues/8446.
Motivation:
There are some extra log level checks (logger.isWarnEnabled()).
Modification:
Remove log level checks (logger.isWarnEnabled()) from io.netty.channel.epoll.AbstractEpollStreamChannel, io.netty.channel.DefaultFileRegion, io.netty.channel.nio.AbstractNioChannel, io.netty.util.HashedWheelTimer, io.netty.handler.stream.ChunkedWriteHandler and io.netty.channel.udt.nio.NioUdtMessageConnectorChannel
Result:
Fixes#9456
Motivation:
We should not only include the java source files but also the c source file in our source jars.
Modifications:
Add files from src/main/c as well
Result:
Fixes https://github.com/netty/netty/issues/9494
Motivation
The epoll transport was updated in #7834 to decouple setting of the
timerFd from the event loop, so that scheduling delayed tasks does not
require waking up epoll_wait. To achieve this, new overridable hooks
were added in the AbstractScheduledEventExecutor and
SingleThreadEventExecutor superclasses.
However, the minimumDelayScheduledTaskRemoved hook has no current
purpose and I can't envisage a _practical_ need for it. Removing
it would reduce complexity and avoid supporting this specific
API indefinitely. We can add something similar later if needed
but the opposite is not true.
There also isn't a _nice_ way to use the abstractions for
wakeup-avoidance optimizations in other EventLoops that don't have a
decoupled timer.
This PR replaces executeScheduledRunnable and
wakesUpForScheduledRunnable
with two new methods before/afterFutureTaskScheduled that have slightly
different semantics:
- They only apply to additions; given the current internals there's no
practical use for removals
- They allow per-submission wakeup decisions via a boolean return val,
which makes them easier to exploit from other existing EL impls (e.g.
NIO/KQueue)
- They are subjectively "cleaner", taking just the deadline parameter
and not exposing Runnables
- For current EL/queue impls, only the "after" hook is really needed,
but specialized blocking queue impls can conditionally wake on task
submission (I have one lined up)
Also included are further optimization/simplification/fixes to the
timerFd manipulation logic.
Modifications
- Remove AbstractScheduledEventExecutor#minimumDelayScheduledTaskRemoved()
and supporting methods
- Uplift NonWakeupRunnable and corresponding default wakesUpForTask()
impl from SingleThreadEventLoop to SingleThreadEventExecutor
- Change executeScheduledRunnable() to be package-private, and have a
final impl in SingleThreadEventExecutor which triggers new overridable
hooks before/afterFutureTaskScheduled()
- Remove unnecessary use of bookend tasks while draining the task queue
- Use new hooks to add simpler wake-up avoidance optimization to
NioEventLoop (primarily to demonstrate utility/simplicity)
- Reinstate removed EpollTest class
In EpollEventLoop:
- Refactor to use only the new afterFutureTaskScheduled() hook for
updating timerFd
- Fix setTimerFd race condition using a monitor
- Set nextDeadlineNanos to a negative value while the EL is awake and
use this to block timer changes from outside the EL. Restore the
known-set value prior to sleeping, updating timerFd first if necessary
- Don't read from timerFd when processing expiry event
Result
- Cleaner API for integrating with different EL/queue timing impls
- Fixed race condition to avoid missing scheduled wakeups
- Eliminate unnecessary timerFd updates while EL is awake, and
unnecessary expired timerFd reads
- Avoid unnecessary scheduled-task wakeups when using NIO transport
I did not yet further explore the suggestion of using
TFD_TIMER_ABSTIME for the timerFd.
Motivation
Currently an epoll_ctl syscall is made every time there is a change to
the event interest flags (EPOLLIN, EPOLLOUT, etc) of a channel. These
are only done in the event loop so can be aggregated into 0 or 1 such
calls per channel prior to the next call to epoll_wait.
Modifications
I think further streamlining/simplification is possible but for now I've
tried to minimize structural changes and added the aggregation beneath
the existing flag manipulation logic.
A new AbstractChannel#activeFlags field records the flags last set on
the epoll fd for that channel. Calls to setFlag/clearFlag update the
flags field as before but instead of calling epoll_ctl immediately, just
set or clear a bit for the channel in a new bitset in the associated
EpollEventLoop to reflect whether there's any change to the last set
value.
Prior to calling epoll_wait the event loop makes the appropriate
epoll_ctl(EPOLL_CTL_MOD) call once for each channel who's bit is set.
Result
Fewer syscalls, particularly in some auto-read=false cases. Simplified
error handling from centralization of these calls.
