Linux io_uring

Discussions center on io_uring, Linux's asynchronous I/O interface, debating its performance advantages, comparisons to alternatives like IOCP or kqueue, and suitability for high-throughput applications.

➡️ Stable 0.9x Open Source
1,784
Comments
17
Years Active
5
Top Authors
#8869
Topic ID

Activity Over Time

2010
1
2011
1
2012
3
2013
5
2014
2
2015
8
2016
7
2017
16
2018
10
2019
69
2020
259
2021
267
2022
176
2023
279
2024
276
2025
372
2026
33

Keywords

OP phoronix.com NT FWIW SQE CQE IOCP PG kernel.dk lwn.net io_uring io thread linux fsync userspace completion priority async syscalls

Sample Comments

saagarjha • Jan 1, 2025 • View on HN

You do realize io_uring exists for a reason, right?

hinkley • Feb 10, 2024 • View on HN

If it's not solved by io_uring, then what is io_uring for?

hderms • Sep 26, 2023 • View on HN

but isn't the whole point of io_uring that it's more performant?

yxhuvud • Sep 3, 2025 • View on HN

Uh, io_uring does that just fine.

tlamponi • Nov 27, 2020 • View on HN

Care to shed some lights on the points of io_uring the OP misses? (honestly interested)

the8472 • Sep 8, 2019 • View on HN

Does io_uring change the equation a little?

lathiat • Feb 12, 2021 • View on HN

I'm not sure about io_uring not reimplementing everything.. it seems to be gaining more and more scope.

zozbot234 • Jan 28, 2020 • View on HN

With io_uring this can be done on Linux as well.

jepler • Jul 11, 2024 • View on HN

did you consider using io_uring? if not, was there a reason other than portability?

bryanwb • Apr 12, 2019 • View on HN

can you elaborate on how io_uring bypasses the kernel?