On Wed, 22 Feb 2023 14:36:47 +0000, Pavel Begunkov wrote:
> Improve support for registered buffers consisting of huge pages by
> keeping them as a single element bvec instead of chunking them into
> 4K pages. It improves performance quite a bit cutting CPU cycles on
> dma-mapping and promoting a more efficient use of hardware.
>
> With a custom fio/t/io_uring 64K reads benchmark with multiple Optanes
> I've got 671 -> 736 KIOPS improvement (~10%).
>
> [...]
Applied, thanks!
[1/4] io_uring/rsrc: disallow multi-source reg buffers
commit: edd478269640b360c6f301f2baa04abdda563ef3
[3/4] io_uring/rsrc: optimise single entry advance
commit: b000ae0ec2d709046ac1a3c5722fea417f8a067e
[4/4] io_uring/rsrc: optimise registered huge pages
commit: 57bebf807e2abcf87d96b9de1266104ee2d8fc2f
Best regards,
--
Jens Axboe
