Registering kernel-managed ring buffers

[A248]

Hi, I'm not sure if this has been requested already. I just did some reading on io_uring, and I apologize if this question seems naive.

Are kernel-managed ring buffers usable with this library? It seems like the op code IOURING_REGISTER_BUFFERS can allow the application to gain more performance if the buffers are pre-allocated with the kernel.

If this is supported, what would be an example of how to use it?

[A248]

Also IORING_REGISTER_PBUF_RING appears to be the newer, and more efficient, mechanism since Linux 5.17.

https://github.com/axboe/liburing/wiki/io_uring-and-networking-in-2023

[Berrysoft]

Are kernel-managed ring buffers usable with this library?

We provide bufring support, but no registered buffer support. The bufring is supported through read_managed* methods.

mechanism since Linux 5.17.

It's OK. We require 5.19 currently.

The biggest problem it that the features are too new to decide how to design the APIs. Do you have any ideas?