netty5/transport-native-io_uring/src/main/java/io/netty/channel/uring/IOUringEventLoop.java
Norman Maurer 70b7621963
Implement batching of reading and writing when using datagram with io_uring. (#10606)
Motivation:

io_uring does not support recvmmsg / sendmmsg directly and so we need to
"emulate" it by submitting multiple IORING_IO_RECVMSG /
IORING_IO_SENDMSG calls.

Modifications:

- Allow to issue multiple write / read calls at once no matter what
  concrete AbstractIOUringChannel subclass it is
- Add support for batching recvmsg / sendmsg when using
IOUringDatagramChannel

Result:

Better performance
2020-09-29 16:58:46 +02:00

357 lines
14 KiB
Java

/*
* Copyright 2020 The Netty Project
*
* The Netty Project licenses this file to you under the Apache License,
* version 2.0 (the "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
* License for the specific language governing permissions and limitations
* under the License.
*/
package io.netty.channel.uring;
import io.netty.channel.EventLoopTaskQueueFactory;
import io.netty.channel.SingleThreadEventLoop;
import io.netty.channel.unix.Errors;
import io.netty.channel.unix.FileDescriptor;
import io.netty.channel.unix.IovArray;
import io.netty.util.collection.IntObjectHashMap;
import io.netty.util.collection.IntObjectMap;
import io.netty.util.concurrent.RejectedExecutionHandler;
import io.netty.util.internal.PlatformDependent;
import io.netty.util.internal.logging.InternalLogger;
import io.netty.util.internal.logging.InternalLoggerFactory;
import java.io.IOException;
import java.util.Queue;
import java.util.concurrent.Executor;
import java.util.concurrent.atomic.AtomicLong;
final class IOUringEventLoop extends SingleThreadEventLoop implements IOUringCompletionQueueCallback {
private static final InternalLogger logger = InternalLoggerFactory.getInstance(IOUringEventLoop.class);
private final long eventfdReadBuf = PlatformDependent.allocateMemory(8);
private final IntObjectMap<AbstractIOUringChannel> channels = new IntObjectHashMap<AbstractIOUringChannel>(4096);
private final RingBuffer ringBuffer;
private static final long AWAKE = -1L;
private static final long NONE = Long.MAX_VALUE;
// nextWakeupNanos is:
// AWAKE when EL is awake
// NONE when EL is waiting with no wakeup scheduled
// other value T when EL is waiting with wakeup scheduled at time T
private final AtomicLong nextWakeupNanos = new AtomicLong(AWAKE);
private final FileDescriptor eventfd;
private final IovArrays iovArrays;
// The maximum number of bytes for an InetAddress / Inet6Address
private final byte[] inet4AddressArray = new byte[4];
private final byte[] inet6AddressArray = new byte[16];
private long prevDeadlineNanos = NONE;
private boolean pendingWakeup;
IOUringEventLoop(IOUringEventLoopGroup parent, Executor executor, int ringSize, boolean ioseqAsync,
RejectedExecutionHandler rejectedExecutionHandler, EventLoopTaskQueueFactory queueFactory) {
super(parent, executor, false, newTaskQueue(queueFactory), newTaskQueue(queueFactory),
rejectedExecutionHandler);
// Ensure that we load all native bits as otherwise it may fail when try to use native methods in IovArray
IOUring.ensureAvailability();
// TODO: Let's hard code this to 8 IovArrays to keep the memory overhead kind of small. We may want to consider
// allow to change this in the future.
iovArrays = new IovArrays(8);
ringBuffer = Native.createRingBuffer(ringSize, ioseqAsync, new Runnable() {
@Override
public void run() {
// Once we submitted its safe to clear the IovArrays and so be able to re-use these.
iovArrays.clear();
}
});
eventfd = Native.newBlockingEventFd();
logger.trace("New EventLoop: {}", this.toString());
}
private static Queue<Runnable> newTaskQueue(
EventLoopTaskQueueFactory queueFactory) {
if (queueFactory == null) {
return newTaskQueue0(DEFAULT_MAX_PENDING_TASKS);
}
return queueFactory.newTaskQueue(DEFAULT_MAX_PENDING_TASKS);
}
@Override
protected Queue<Runnable> newTaskQueue(int maxPendingTasks) {
return newTaskQueue0(maxPendingTasks);
}
private static Queue<Runnable> newTaskQueue0(int maxPendingTasks) {
// This event loop never calls takeTask()
return maxPendingTasks == Integer.MAX_VALUE? PlatformDependent.<Runnable>newMpscQueue()
: PlatformDependent.<Runnable>newMpscQueue(maxPendingTasks);
}
void add(AbstractIOUringChannel ch) {
logger.trace("Add Channel: {} ", ch.socket.intValue());
int fd = ch.socket.intValue();
channels.put(fd, ch);
}
void remove(AbstractIOUringChannel ch) {
logger.trace("Remove Channel: {}", ch.socket.intValue());
int fd = ch.socket.intValue();
AbstractIOUringChannel old = channels.remove(fd);
if (old != null && old != ch) {
// The Channel mapping was already replaced due FD reuse, put back the stored Channel.
channels.put(fd, old);
// If we found another Channel in the map that is mapped to the same FD the given Channel MUST be closed.
assert !ch.isOpen();
}
}
private void closeAll() {
logger.trace("CloseAll IOUringEvenloop");
// Using the intermediate collection to prevent ConcurrentModificationException.
// In the `close()` method, the channel is deleted from `channels` map.
AbstractIOUringChannel[] localChannels = channels.values().toArray(new AbstractIOUringChannel[0]);
for (AbstractIOUringChannel ch : localChannels) {
ch.unsafe().close(ch.unsafe().voidPromise());
}
}
@Override
protected void run() {
final IOUringCompletionQueue completionQueue = ringBuffer.ioUringCompletionQueue();
final IOUringSubmissionQueue submissionQueue = ringBuffer.ioUringSubmissionQueue();
// Lets add the eventfd related events before starting to do any real work.
addEventFdRead(submissionQueue);
for (;;) {
try {
logger.trace("Run IOUringEventLoop {}", this);
// Prepare to block wait
long curDeadlineNanos = nextScheduledTaskDeadlineNanos();
if (curDeadlineNanos == -1L) {
curDeadlineNanos = NONE; // nothing on the calendar
}
nextWakeupNanos.set(curDeadlineNanos);
// Only submit a timeout if there are no tasks to process and do a blocking operation
// on the completionQueue.
try {
if (!hasTasks()) {
if (curDeadlineNanos != prevDeadlineNanos) {
prevDeadlineNanos = curDeadlineNanos;
submissionQueue.addTimeout(deadlineToDelayNanos(curDeadlineNanos), (short) 0);
}
// Check there were any completion events to process
if (!completionQueue.hasCompletions()) {
// Block if there is nothing to process after this try again to call process(....)
logger.trace("submitAndWait {}", this);
submissionQueue.submitAndWait();
}
}
} finally {
if (nextWakeupNanos.get() == AWAKE || nextWakeupNanos.getAndSet(AWAKE) == AWAKE) {
pendingWakeup = true;
}
}
} catch (Throwable t) {
handleLoopException(t);
}
// Avoid blocking for as long as possible - loop until available work exhausted
boolean maybeMoreWork = true;
do {
try {
// CQE processing can produce tasks, and new CQEs could arrive while
// processing tasks. So run both on every iteration and break when
// they both report that nothing was done (| means always run both).
maybeMoreWork = completionQueue.process(this) != 0 | runAllTasks();
} catch (Throwable t) {
handleLoopException(t);
}
// Always handle shutdown even if the loop processing threw an exception
try {
if (isShuttingDown()) {
closeAll();
if (confirmShutdown()) {
return;
}
if (!maybeMoreWork) {
maybeMoreWork = hasTasks() || completionQueue.hasCompletions();
}
}
} catch (Throwable t) {
handleLoopException(t);
}
} while (maybeMoreWork);
}
}
/**
* Visible only for testing!
*/
void handleLoopException(Throwable t) {
logger.warn("Unexpected exception in the io_uring event loop", t);
// Prevent possible consecutive immediate failures that lead to
// excessive CPU consumption.
try {
Thread.sleep(1000);
} catch (InterruptedException e) {
// Ignore.
}
}
@Override
public void handle(int fd, int res, int flags, int op, short data) {
if (op == Native.IORING_OP_READ && eventfd.intValue() == fd) {
pendingWakeup = false;
addEventFdRead(ringBuffer.ioUringSubmissionQueue());
} else if (op == Native.IORING_OP_TIMEOUT) {
if (res == Native.ERRNO_ETIME_NEGATIVE) {
prevDeadlineNanos = NONE;
}
} else {
// Remaining events should be channel-specific
final AbstractIOUringChannel channel = channels.get(fd);
if (channel == null) {
return;
}
if (op == Native.IORING_OP_READ || op == Native.IORING_OP_ACCEPT || op == Native.IORING_OP_RECVMSG) {
handleRead(channel, res, data);
} else if (op == Native.IORING_OP_WRITEV ||
op == Native.IORING_OP_WRITE || op == Native.IORING_OP_SENDMSG) {
handleWrite(channel, res, data);
} else if (op == Native.IORING_OP_POLL_ADD) {
handlePollAdd(channel, res, data);
} else if (op == Native.IORING_OP_POLL_REMOVE) {
if (res == Errors.ERRNO_ENOENT_NEGATIVE) {
logger.trace("IORING_POLL_REMOVE not successful");
} else if (res == 0) {
logger.trace("IORING_POLL_REMOVE successful");
}
if (!channel.ioScheduled()) {
// We cancelled the POLL ops which means we are done and should remove the mapping.
remove(channel);
return;
}
} else if (op == Native.IORING_OP_CONNECT) {
handleConnect(channel, res);
}
channel.ioUringUnsafe().processDelayedClose();
}
}
private void handleRead(AbstractIOUringChannel channel, int res, int data) {
channel.ioUringUnsafe().readComplete(res, data);
}
private void handleWrite(AbstractIOUringChannel channel, int res, int data) {
channel.ioUringUnsafe().writeComplete(res, data);
}
private void handlePollAdd(AbstractIOUringChannel channel, int res, int pollMask) {
if ((pollMask & Native.POLLOUT) != 0) {
channel.ioUringUnsafe().pollOut(res);
}
if ((pollMask & Native.POLLIN) != 0) {
channel.ioUringUnsafe().pollIn(res);
}
if ((pollMask & Native.POLLRDHUP) != 0) {
channel.ioUringUnsafe().pollRdHup(res);
}
}
private void addEventFdRead(IOUringSubmissionQueue submissionQueue) {
submissionQueue.addRead(eventfd.intValue(), eventfdReadBuf, 0, 8, (short) 0);
}
private void handleConnect(AbstractIOUringChannel channel, int res) {
channel.ioUringUnsafe().connectComplete(res);
}
@Override
protected void cleanup() {
if (pendingWakeup) {
// Another thread is in the process of writing to the eventFd. We must wait to
// receive the corresponding CQE before closing it or else the fd int may be
// reassigned by the kernel in the meantime.
IOUringCompletionQueue completionQueue = ringBuffer.ioUringCompletionQueue();
IOUringCompletionQueueCallback callback = new IOUringCompletionQueueCallback() {
@Override
public void handle(int fd, int res, int flags, int op, short data) {
if (op == Native.IORING_OP_READ && eventfd.intValue() == fd) {
pendingWakeup = false;
}
}
};
completionQueue.process(callback);
while (pendingWakeup) {
completionQueue.ioUringWaitCqe();
completionQueue.process(callback);
}
}
try {
eventfd.close();
} catch (IOException e) {
logger.warn("Failed to close the event fd.", e);
}
ringBuffer.close();
iovArrays.release();
PlatformDependent.freeMemory(eventfdReadBuf);
}
RingBuffer getRingBuffer() {
return ringBuffer;
}
@Override
protected void wakeup(boolean inEventLoop) {
if (!inEventLoop && nextWakeupNanos.getAndSet(AWAKE) != AWAKE) {
// write to the evfd which will then wake-up epoll_wait(...)
Native.eventFdWrite(eventfd.intValue(), 1L);
}
}
IovArray iovArray() {
IovArray iovArray = iovArrays.next();
if (iovArray == null) {
ringBuffer.ioUringSubmissionQueue().submit();
iovArray = iovArrays.next();
assert iovArray != null;
}
return iovArray;
}
/**
* {@code byte[]} that can be used as temporary storage to encode the ipv4 address
*/
byte[] inet4AddressArray() {
return inet4AddressArray;
}
/**
* {@code byte[]} that can be used as temporary storage to encode the ipv6 address
*/
byte[] inet6AddressArray() {
return inet6AddressArray;
}
}