Maybe not quite there yet?

https://lwn.net/Articles/724198/ (May 30, 2017)

In truth, an AIO operation can block in the kernel for a number of reasons, making AIO difficult to use in situations where the calling thread truly cannot afford to block. A longstanding patch set aiming to improve this situation would appear to be nearing completion, but it is more of a step in the right direction than a true solution to the problem.

In other words, AIO seems likely to remain useful only for the handful of applications that perform direct I/O to files.

In my limited knowledge the only app I've seen use aio is oracle database.

If it works in the sense that the kernel supports the API and implements aoi internally using async mechanisms, then that would be awesome. Is that the case?