linux-block.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jens Axboe <axboe@kernel.dk>
To: Pavel Begunkov <asml.silence@gmail.com>, io-uring@vger.kernel.org
Cc: linux-block@vger.kernel.org, jannh@google.com
Subject: Re: [PATCH 2/3] io_uring: pass in io_kiocb to fill/add CQ handlers
Date: Thu, 7 Nov 2019 06:16:36 -0700	[thread overview]
Message-ID: <2a723bbc-f031-6d2a-ef8b-96f9a5cfc70e@kernel.dk> (raw)
In-Reply-To: <df4352ab-2670-e69f-cc92-5e72f1cd6229@gmail.com>

On 11/7/19 2:53 AM, Pavel Begunkov wrote:
>> @@ -804,8 +803,10 @@ static void io_fail_links(struct io_kiocb *req)
>>   		    link->submit.sqe->opcode == IORING_OP_LINK_TIMEOUT) {
>>   			io_link_cancel_timeout(ctx, link);
>>   		} else {
>> -			io_cqring_fill_event(ctx, link->user_data, -ECANCELED);
>> -			__io_free_req(link);
>> +			io_cqring_fill_event(link, -ECANCELED);
>> +			/* drop both submit and complete references */
>> +			io_put_req(link, NULL);
>> +			io_put_req(link, NULL);
> 
> io_put_req() -> ... -> io_free_req() -> io_fail_links() -> io_put_req()
> 
> It shouldn't recurse further, but probably it would be better to avoid
> it at all.

Not sure how to improve that. We could do something ala:

if (refcount_sub_and_test(2, &link->refs))
	__io_free_req(link);

to make it clear and more resistant against recursion.

I also think we need to put that link path out-of-line in io_free_req().
I'll make those two changes.

-- 
Jens Axboe


  reply	other threads:[~2019-11-07 13:16 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-06 23:53 [PATCHSET v2 0/3] io_uring CQ ring backpressure Jens Axboe
2019-11-06 23:53 ` [PATCH 1/3] io_uring: make io_cqring_events() take 'ctx' as argument Jens Axboe
2019-11-06 23:53 ` [PATCH 2/3] io_uring: pass in io_kiocb to fill/add CQ handlers Jens Axboe
2019-11-07  9:53   ` Pavel Begunkov
2019-11-07 13:16     ` Jens Axboe [this message]
2019-11-06 23:53 ` [PATCH 3/3] io_uring: add support for backlogged CQ ring Jens Axboe
2019-11-07 16:00 [PATCHSET v3 0/3] io_uring CQ ring backpressure Jens Axboe
2019-11-07 16:00 ` [PATCH 2/3] io_uring: pass in io_kiocb to fill/add CQ handlers 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=2a723bbc-f031-6d2a-ef8b-96f9a5cfc70e@kernel.dk \
    --to=axboe@kernel.dk \
    --cc=asml.silence@gmail.com \
    --cc=io-uring@vger.kernel.org \
    --cc=jannh@google.com \
    --cc=linux-block@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).