History

Close IOUringEventLoop wakeup/eventfd close shutdown race (#10615 )

Motivation

There is a race condition when shutting down the event loop where the
eventFd write performed in the wakeup() method may actually hit a
different fd if it's closed and reassigned in the meantime.

This was already encountered and addressed in the epoll case.

Modifications

Similar to what's done for epoll, in IOUringEventLoop:
- Reinstate pendingWakeup flag which tracks when there is a wakeup
pending (CAS of nextWakeupNanos performed by other thread in the
wakeup() method)
- Add logic to the cleanup() method to wait for corresponding READ CQE
before closing the eventFd
- Remove unused fields from IOUringCompletionQueue (cleanup)

Result

No event loop shutdown race

2020-09-28 08:47:23 +02:00

src

Close IOUringEventLoop wakeup/eventfd close shutdown race (#10615 )

2020-09-28 08:47:23 +02:00

pom.xml

Explicit need to specify -Piouring-native to compile the native bits … (#10546 )

2020-09-09 09:50:36 +02:00

README.md

Update README to reflect kernel requirements for iouring transport (#10539 )

2020-09-07 12:05:49 +02:00

README.md

Netty io_uring

The new io_uring interface added to the Linux Kernel 5.1 is a high I/O performance scalable interface for fully asynchronous Linux syscalls

Requirements:

x86-64 processor
at least 5.9-rc4
to run the tests, you have to increase memlock(default 64K)

See our wiki page.