LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Olivier Langlois <olivier@trillion01.com>
To: Nadav Amit <nadav.amit@gmail.com>, Jens Axboe <axboe@kernel.dk>
Cc: io-uring@vger.kernel.org, linux-kernel@vger.kernel.org,
	Nadav Amit <namit@vmware.com>,
	Pavel Begunkov <asml.silence@gmail.com>
Subject: Re: [PATCH 1/2] io_uring: clear TIF_NOTIFY_SIGNAL when running task work
Date: Mon, 09 Aug 2021 17:48:38 -0400	[thread overview]
Message-ID: <fdd54421f4d4e825152192e327c838d035352945.camel@trillion01.com> (raw)
In-Reply-To: <20210808001342.964634-2-namit@vmware.com>

On Sat, 2021-08-07 at 17:13 -0700, Nadav Amit wrote:
> From: Nadav Amit <namit@vmware.com>
> 
> When using SQPOLL, the submission queue polling thread calls
> task_work_run() to run queued work. However, when work is added with
> TWA_SIGNAL - as done by io_uring itself - the TIF_NOTIFY_SIGNAL remains
> set afterwards and is never cleared.
> 
> Consequently, when the submission queue polling thread checks whether
> signal_pending(), it may always find a pending signal, if
> task_work_add() was ever called before.
> 
> The impact of this bug might be different on different kernel versions.
> It appears that on 5.14 it would only cause unnecessary calculation and
> prevent the polling thread from sleeping. On 5.13, where the bug was
> found, it stops the polling thread from finding newly submitted work.
> 
> Instead of task_work_run(), use tracehook_notify_signal() that clears
> TIF_NOTIFY_SIGNAL. Test for TIF_NOTIFY_SIGNAL in addition to
> current->task_works to avoid a race in which task_works is cleared but
> the TIF_NOTIFY_SIGNAL is set.
> 
> Fixes: 685fe7feedb96 ("io-wq: eliminate the need for a manager thread")
> Cc: Jens Axboe <axboe@kernel.dk>
> Cc: Pavel Begunkov <asml.silence@gmail.com>
> Signed-off-by: Nadav Amit <namit@vmware.com>
> ---
>  fs/io_uring.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index 5a0fd6bcd318..f39244d35f90 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -78,6 +78,7 @@
>  #include <linux/task_work.h>
>  #include <linux/pagemap.h>
>  #include <linux/io_uring.h>
> +#include <linux/tracehook.h>
>  
>  #define CREATE_TRACE_POINTS
>  #include <trace/events/io_uring.h>
> @@ -2203,9 +2204,9 @@ static inline unsigned int io_put_rw_kbuf(struct
> io_kiocb *req)
>  
>  static inline bool io_run_task_work(void)
>  {
> -       if (current->task_works) {
> +       if (test_thread_flag(TIF_NOTIFY_SIGNAL) || current->task_works)
> {
>                 __set_current_state(TASK_RUNNING);
> -               task_work_run();
> +               tracehook_notify_signal();
>                 return true;
>         }
>  

thx a lot for this patch!

This explains what I am seeing here:
https://lore.kernel.org/io-uring/4d93d0600e4a9590a48d320c5a7dd4c54d66f095.camel@trillion01.com/

I was under the impression that task_work_run() was clearing
TIF_NOTIFY_SIGNAL.

your patch made me realize that it does not...



  parent reply	other threads:[~2021-08-09 21:48 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-08  0:13 [PATCH 0/2] io_uring: bug fixes Nadav Amit
2021-08-08  0:13 ` [PATCH 1/2] io_uring: clear TIF_NOTIFY_SIGNAL when running task work Nadav Amit
2021-08-08 12:55   ` Pavel Begunkov
2021-08-08 17:31     ` Nadav Amit
2021-08-09  4:07       ` Hao Xu
2021-08-09  4:50         ` Nadav Amit
2021-08-09 10:35           ` Pavel Begunkov
2021-08-09 10:18       ` Pavel Begunkov
2021-08-09 21:48   ` Olivier Langlois [this message]
2021-08-10  8:28     ` Nadav Amit
2021-08-10 13:33       ` Olivier Langlois
2021-08-10 21:32       ` Pavel Begunkov
2021-08-11  2:33         ` Nadav Amit
2021-08-11  2:51           ` Jens Axboe
2021-08-11  5:40             ` I/O cancellation in io-uring (was: io_uring: clear TIF_NOTIFY_SIGNAL ...) Nadav Amit
2021-08-08  0:13 ` [PATCH 2/2] io_uring: Use WRITE_ONCE() when writing to sq_flags Nadav Amit
2021-08-09 13:53 ` [PATCH 0/2] io_uring: bug fixes 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=fdd54421f4d4e825152192e327c838d035352945.camel@trillion01.com \
    --to=olivier@trillion01.com \
    --cc=asml.silence@gmail.com \
    --cc=axboe@kernel.dk \
    --cc=io-uring@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=nadav.amit@gmail.com \
    --cc=namit@vmware.com \
    --subject='Re: [PATCH 1/2] io_uring: clear TIF_NOTIFY_SIGNAL when running task work' \
    /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

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).