public inbox for [email protected]
 help / color / mirror / Atom feed
From: Jens Axboe <[email protected]>
To: [email protected], [email protected]
Cc: [email protected], [email protected], [email protected],
	[email protected], Pavel Begunkov <[email protected]>,
	[email protected], Jens Axboe <[email protected]>
Subject: [PATCH] io_uring: Use io_schedule* in cqring wait
Date: Tue, 18 Jul 2023 13:49:15 -0600	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

From: Andres Freund <[email protected]>

I observed poor performance of io_uring compared to synchronous IO. That
turns out to be caused by deeper CPU idle states entered with io_uring,
due to io_uring using plain schedule(), whereas synchronous IO uses
io_schedule().

The losses due to this are substantial. On my cascade lake workstation,
t/io_uring from the fio repository e.g. yields regressions between 20%
and 40% with the following command:
./t/io_uring -r 5 -X0 -d 1 -s 1 -c 1 -p 0 -S$use_sync -R 0 /mnt/t2/fio/write.0.0

This is repeatable with different filesystems, using raw block devices
and using different block devices.

Use io_schedule_prepare() / io_schedule_finish() in
io_cqring_wait_schedule() to address the difference.

After that using io_uring is on par or surpassing synchronous IO (using
registered files etc makes it reliably win, but arguably is a less fair
comparison).

There are other calls to schedule() in io_uring/, but none immediately
jump out to be similarly situated, so I did not touch them. Similarly,
it's possible that mutex_lock_io() should be used, but it's not clear if
there are cases where that matters.

Cc: [email protected] # 5.10+
Cc: Pavel Begunkov <[email protected]>
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Andres Freund <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
[axboe: minor style fixup]
Signed-off-by: Jens Axboe <[email protected]>
---
 io_uring/io_uring.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index e8096d502a7c..7505de2428e0 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -2489,6 +2489,8 @@ int io_run_task_work_sig(struct io_ring_ctx *ctx)
 static inline int io_cqring_wait_schedule(struct io_ring_ctx *ctx,
 					  struct io_wait_queue *iowq)
 {
+	int token, ret;
+
 	if (unlikely(READ_ONCE(ctx->check_cq)))
 		return 1;
 	if (unlikely(!llist_empty(&ctx->work_llist)))
@@ -2499,11 +2501,20 @@ static inline int io_cqring_wait_schedule(struct io_ring_ctx *ctx,
 		return -EINTR;
 	if (unlikely(io_should_wake(iowq)))
 		return 0;
+
+	/*
+	 * Use io_schedule_prepare/finish, so cpufreq can take into account
+	 * that the task is waiting for IO - turns out to be important for low
+	 * QD IO.
+	 */
+	token = io_schedule_prepare();
+	ret = 0;
 	if (iowq->timeout == KTIME_MAX)
 		schedule();
 	else if (!schedule_hrtimeout(&iowq->timeout, HRTIMER_MODE_ABS))
-		return -ETIME;
-	return 0;
+		ret = -ETIME;
+	io_schedule_finish(token);
+	return ret;
 }
 
 /*
-- 
2.40.1


  reply	other threads:[~2023-07-18 19:49 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-18 19:49 [PATCHSET v2 0/5] Improve async iomap DIO performance Jens Axboe
2023-07-18 19:49 ` Jens Axboe [this message]
2023-07-18 19:50   ` [PATCH] io_uring: Use io_schedule* in cqring wait Jens Axboe
2023-07-18 19:49 ` [PATCH 1/5] iomap: simplify logic for when a dio can get completed inline Jens Axboe
2023-07-18 22:56   ` Dave Chinner
2023-07-19 15:23     ` Jens Axboe
2023-07-18 19:49 ` [PATCH 2/5] fs: add IOCB flags related to passing back dio completions Jens Axboe
2023-07-18 19:49 ` [PATCH 3/5] io_uring/rw: add write support for IOCB_DIO_DEFER Jens Axboe
2023-07-18 19:49 ` [PATCH 4/5] iomap: add local 'iocb' variable in iomap_dio_bio_end_io() Jens Axboe
2023-07-18 19:49 ` [PATCH 5/5] iomap: support IOCB_DIO_DEFER Jens Axboe
2023-07-18 23:50   ` Dave Chinner
2023-07-19 19:55     ` Jens Axboe
  -- strict thread matches above, loose matches on Subject: below --
2023-07-24 15:35 [PATCH] io_uring: Use io_schedule* in cqring wait Phil Elwell
2023-07-24 15:48 ` Greg KH
2023-07-24 15:50   ` Jens Axboe
2023-07-24 15:58     ` Jens Axboe
2023-07-24 16:07       ` Phil Elwell
2023-07-24 16:08         ` Jens Axboe
2023-07-24 16:48           ` Phil Elwell
2023-07-24 18:22             ` Jens Axboe
2023-07-24 19:22       ` Pavel Begunkov
2023-07-24 20:27         ` Jeff Moyer
2023-07-24 15:48 ` Jens Axboe
2023-07-24 16:16   ` Andres Freund
2023-07-24 16:20     ` Jens Axboe
2023-07-24 17:24     ` Andres Freund
2023-07-24 17:44       ` Jens Axboe

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox