public inbox for io-uring@vger.kernel.org
 help / color / mirror / Atom feed
From: Fengnan Chang <fengnanchang@gmail.com>
To: Jens Axboe <axboe@kernel.dk>,
	asml.silence@gmail.com, io-uring@vger.kernel.org
Cc: Fengnan Chang <changfengnan@bytedance.com>,
	Diangang Li <lidiangang@bytedance.com>
Subject: Re: [RFC PATCH 2/2] io_uring: fix io may accumulation in poll mode
Date: Fri, 12 Dec 2025 10:12:15 +0800	[thread overview]
Message-ID: <f987df2c-f9a7-4656-b725-7a30651b4d86@gmail.com> (raw)
In-Reply-To: <1d8a4c67-0c30-449e-a4e3-24363de0fcfa@kernel.dk>



在 2025/12/12 09:53, Jens Axboe 写道:
> On 12/11/25 6:41 PM, Fengnan Chang wrote:
>> Oh, we can't add nr_events == iob.nr_reqs check, if
>> blk_mq_add_to_batch add failed, completed IO will not add into iob,
>> iob.nr_reqs will be 0, this may cause io hang.
> Indeed, won't work as-is.
>
> I do think we're probably making a bigger deal out of the full loop than
> necessary. At least I'd be perfectly happy with just the current patch,
> performance should be better there than we currently have it. Ideally
> we'd have just one loop for polling and catching the completed items,
> but that's a bit tricky with the batch completions.
Yes, ideally one loop would be enough, but given that there are also 
multi_queue ctx, that
doesn't seem to be possible.
>


  reply	other threads:[~2025-12-12  2:12 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-12-10  8:54 [RFC PATCH 0/2] io_uring: fix io may accumulation in poll mode Fengnan Chang
2025-12-10  8:55 ` [RFC PATCH 1/2] blk-mq: delete task running check in blk_hctx_poll Fengnan Chang
2025-12-10  9:19   ` Jens Axboe
2025-12-10  9:53   ` Jens Axboe
2025-12-10  8:55 ` [RFC PATCH 2/2] io_uring: fix io may accumulation in poll mode Fengnan Chang
2025-12-11  2:15   ` Jens Axboe
2025-12-11  4:10     ` Jens Axboe
2025-12-11  7:38       ` Fengnan
2025-12-11 10:22         ` Jens Axboe
2025-12-11 10:33           ` Jens Axboe
2025-12-11 11:13             ` Fengnan Chang
2025-12-11 11:19               ` Jens Axboe
2025-12-12  1:41             ` Fengnan Chang
2025-12-12  1:53               ` Jens Axboe
2025-12-12  2:12                 ` Fengnan Chang [this message]
2025-12-12  5:11                   ` Jens Axboe
2025-12-12  8:58                     ` Jens Axboe
2025-12-12  9:49                       ` Fengnan Chang
2025-12-12 20:22                         ` Jens Axboe
2025-12-12 13:32                     ` Diangang Li
2025-12-12 20:09                       ` Jens Axboe
2025-12-10  9:53 ` (subset) [RFC PATCH 0/2] " 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=f987df2c-f9a7-4656-b725-7a30651b4d86@gmail.com \
    --to=fengnanchang@gmail.com \
    --cc=asml.silence@gmail.com \
    --cc=axboe@kernel.dk \
    --cc=changfengnan@bytedance.com \
    --cc=io-uring@vger.kernel.org \
    --cc=lidiangang@bytedance.com \
    /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