From: Jens Axboe <axboe@kernel.dk>
To: Pavel Begunkov <asml.silence@gmail.com>, io-uring@vger.kernel.org
Subject: Re: [PATCH v2 13/13] io_uring/io-wq: return 2-step work swap scheme
Date: Thu, 4 Feb 2021 07:52:02 -0700 [thread overview]
Message-ID: <8acbd513-531c-0a12-ea3f-ecf0cd94c9e2@kernel.dk> (raw)
In-Reply-To: <014eff28b71c8e5da5edaa4ad9d142916317c839.1612446019.git.asml.silence@gmail.com>
On 2/4/21 6:52 AM, Pavel Begunkov wrote:
> Saving one lock/unlock for io-wq is not super important, but adds some
> ugliness in the code. More important, atomic decs not turning it to zero
> for some archs won't give the right ordering/barriers so the
> io_steal_work() may pretty easily get subtly and completely broken.
>
> Return back 2-step io-wq work exchange and clean it up.
IIRC, this wasn't done to skip the lock/unlock exchange, which I agree
doesn't matter, but to ensure that a link would not need another io-wq
punt. And that is a big deal, it's much faster to run it from that
same thread, rather than needing a new async queue and new thread grab
to get there.
Just want to make sure that's on your mind... Maybe it's still fine
as-is, didn't look too closely yet or test it.
--
Jens Axboe
next prev parent reply other threads:[~2021-02-04 14:53 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-04 13:51 [PATCH v2 5.12 00/13] a second pack of 5.12 cleanups Pavel Begunkov
2021-02-04 13:51 ` [PATCH v2 01/13] io_uring: deduplicate core cancellations sequence Pavel Begunkov
2021-02-04 13:51 ` [PATCH v2 02/13] io_uring: refactor scheduling in io_cqring_wait Pavel Begunkov
2021-02-04 13:51 ` [PATCH v2 03/13] io_uring: refactor io_cqring_wait Pavel Begunkov
2021-02-04 13:51 ` [PATCH v2 04/13] io_uring: refactor io_read for unsupported nowait Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 05/13] io_uring: further simplify do_read error parsing Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 06/13] io_uring: let io_setup_async_rw take care of iovec Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 07/13] io_uring: don't forget to adjust io_size Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 08/13] io_uring: inline io_read()'s iovec freeing Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 09/13] io_uring: highlight read-retry loop Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 10/13] io_uring: treat NONBLOCK and RWF_NOWAIT similarly Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 11/13] io_uring: io_import_iovec return type cleanup Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 12/13] io_uring: deduplicate file table slot calculation Pavel Begunkov
2021-02-04 13:52 ` [PATCH v2 13/13] io_uring/io-wq: return 2-step work swap scheme Pavel Begunkov
2021-02-04 14:52 ` Jens Axboe [this message]
2021-02-04 14:56 ` Pavel Begunkov
2021-02-04 15:05 ` Jens Axboe
2021-02-04 15:07 ` [PATCH v2 5.12 00/13] a second pack of 5.12 cleanups 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 \
--in-reply-to=8acbd513-531c-0a12-ea3f-ecf0cd94c9e2@kernel.dk \
--to=axboe@kernel.dk \
--cc=asml.silence@gmail.com \
--cc=io-uring@vger.kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).