public inbox for [email protected]
 help / color / mirror / Atom feed
From: Jeff Moyer <[email protected]>
To: Pavel Begunkov <[email protected]>
Cc: Jens Axboe <[email protected]>, [email protected]
Subject: Re: [PATCH v2] io_uring: remove req cancel in ->flush()
Date: Thu, 22 Oct 2020 13:52:30 -0400	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <6cffe73a8a44084289ac792e7b152e01498ea1ef.1603380957.git.asml.silence@gmail.com> (Pavel Begunkov's message of "Thu, 22 Oct 2020 16:38:27 +0100")

Pavel Begunkov <[email protected]> writes:

> Every close(io_uring) causes cancellation of all inflight requests
> carrying ->files. That's not nice but was neccessary up until recently.
> Now task->files removal is handled in the core code, so that part of
> flush can be removed.

I don't understand the motivation for this patch.  Why would an
application close the io_uring fd with outstanding requests?

-Jeff

>
> Signed-off-by: Pavel Begunkov <[email protected]>
> ---
> v2: move exiting checks into io_uring_attempt_task_drop() (Jens)
>     remove not needed __io_uring_attempt_task_drop()
>
>  fs/io_uring.c | 28 +++++-----------------------
>  1 file changed, 5 insertions(+), 23 deletions(-)
>
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index 754363ff3ad6..29170bbdd708 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -8668,19 +8668,11 @@ static void io_uring_del_task_file(struct file *file)
>  		fput(file);
>  }
>  
> -static void __io_uring_attempt_task_drop(struct file *file)
> -{
> -	struct file *old = xa_load(&current->io_uring->xa, (unsigned long)file);
> -
> -	if (old == file)
> -		io_uring_del_task_file(file);
> -}
> -
>  /*
>   * Drop task note for this file if we're the only ones that hold it after
>   * pending fput()
>   */
> -static void io_uring_attempt_task_drop(struct file *file, bool exiting)
> +static void io_uring_attempt_task_drop(struct file *file)
>  {
>  	if (!current->io_uring)
>  		return;
> @@ -8688,10 +8680,9 @@ static void io_uring_attempt_task_drop(struct file *file, bool exiting)
>  	 * fput() is pending, will be 2 if the only other ref is our potential
>  	 * task file note. If the task is exiting, drop regardless of count.
>  	 */
> -	if (!exiting && atomic_long_read(&file->f_count) != 2)
> -		return;
> -
> -	__io_uring_attempt_task_drop(file);
> +	if (fatal_signal_pending(current) || (current->flags & PF_EXITING) ||
> +	    atomic_long_read(&file->f_count) == 2)
> +		io_uring_del_task_file(file);
>  }
>  
>  void __io_uring_files_cancel(struct files_struct *files)
> @@ -8749,16 +8740,7 @@ void __io_uring_task_cancel(void)
>  
>  static int io_uring_flush(struct file *file, void *data)
>  {
> -	struct io_ring_ctx *ctx = file->private_data;
> -
> -	/*
> -	 * If the task is going away, cancel work it may have pending
> -	 */
> -	if (fatal_signal_pending(current) || (current->flags & PF_EXITING))
> -		data = NULL;
> -
> -	io_uring_cancel_task_requests(ctx, data);
> -	io_uring_attempt_task_drop(file, !data);
> +	io_uring_attempt_task_drop(file);
>  	return 0;
>  }


  reply	other threads:[~2020-10-22 17:52 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-22 15:38 [PATCH v2] io_uring: remove req cancel in ->flush() Pavel Begunkov
2020-10-22 17:52 ` Jeff Moyer [this message]
2020-10-22 18:01   ` Jens Axboe
2020-10-22 18:49     ` Jeff Moyer
2020-10-22 20:39 ` 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] \
    /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