public inbox for io-uring@vger.kernel.org
 help / color / mirror / Atom feed
From: Jens Axboe <axboe@kernel.dk>
To: Pavel Begunkov <asml.silence@gmail.com>, io-uring@vger.kernel.org
Subject: Re: [PATCH 1/1] io_uring: add lockdep checks for io_handle_tw_list
Date: Tue, 1 Apr 2025 10:13:27 -0600	[thread overview]
Message-ID: <65765d68-cda0-41fb-acdf-58e7b5c1243f@kernel.dk> (raw)
In-Reply-To: <ffd30102aee729e48911f595d1c05804e59b0403.1743522348.git.asml.silence@gmail.com>

On 4/1/25 9:46 AM, Pavel Begunkov wrote:
> Add a lockdep check to io_handle_tw_list() verifying that the context is
> locked and no task work drops it by accident.

I think we'd want a bit more of a "why" explanation here, but I can add
that while committing.

> diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
> index 6df996d01ccf..13e0b48d1aac 100644
> --- a/io_uring/io_uring.c
> +++ b/io_uring/io_uring.c
> @@ -1054,6 +1054,10 @@ struct llist_node *io_handle_tw_list(struct llist_node *node,
>  			mutex_lock(&ctx->uring_lock);
>  			percpu_ref_get(&ctx->refs);
>  		}
> +
> +		lockdep_assert(req->ctx == ctx);
> +		lockdep_assert_held(&ctx->uring_lock);
> +
>  		INDIRECT_CALL_2(req->io_task_work.func,
>  				io_poll_task_func, io_req_rw_complete,
>  				req, ts);

If the assumption is that some previous tw messed things up, might not
be a bad idea to include dumping of that if one of the above lockdep
asserts fail? Preferably in such a way that code generation is the same
when lockdep isn't set...

-- 
Jens Axboe

  reply	other threads:[~2025-04-01 16:13 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-04-01 15:46 [PATCH 1/1] io_uring: add lockdep checks for io_handle_tw_list Pavel Begunkov
2025-04-01 16:13 ` Jens Axboe [this message]
2025-04-01 17:19   ` Pavel Begunkov
2025-04-01 18:14     ` 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=65765d68-cda0-41fb-acdf-58e7b5c1243f@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