Jens Axboe b688f11e86 io_uring: utilize the io batching infrastructure for more efficient polled IO
Wire up using an io_comp_batch for f_op->iopoll(). If the lower stack
supports it, we can handle high rates of polled IO more efficiently.

This raises the single core efficiency on my system from ~6.1M IOPS to
~6.6M IOPS running a random read workload at depth 128 on two gen2
Optane drives.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2021-10-18 14:40:46 -06:00
..
2021-10-07 11:20:08 -07:00
2021-08-19 09:02:55 +09:00
2021-04-12 15:04:29 +02:00
2021-07-06 11:06:04 -07:00
2021-09-02 12:32:12 -07:00
2021-09-07 12:18:29 -07:00
2021-08-26 22:28:02 +02:00
2021-10-07 10:18:36 -05:00
2021-09-22 09:21:02 -07:00
2021-09-04 10:25:26 -07:00
2021-09-10 09:46:48 -07:00
2021-06-30 12:21:16 -07:00
2021-09-03 15:33:47 -07:00
2021-08-16 10:50:32 -06:00
2021-08-19 09:02:55 +09:00
2021-10-18 06:17:36 -06:00
2021-08-31 11:13:35 -07:00
2021-10-18 06:17:35 -06:00
2021-09-09 13:25:49 -07:00
2021-09-11 14:48:42 -07:00
2021-08-31 11:06:32 -07:00
2021-06-29 10:53:48 -07:00
2021-08-18 22:08:24 +02:00
2021-08-10 17:57:22 +02:00