public inbox for [email protected]
 help / color / mirror / Atom feed
From: Pavel Begunkov <[email protected]>
To: Jens Axboe <[email protected]>, [email protected]
Cc: [email protected]
Subject: [PATCH 1/4] io_uring: dont kill fasync under completion_lock
Date: Sat,  2 Jan 2021 16:06:51 +0000	[thread overview]
Message-ID: <e51028690f7415a018403d3607739693188b5f7b.1609600704.git.asml.silence@gmail.com> (raw)
In-Reply-To: <[email protected]>

      CPU0                    CPU1
       ----                    ----
  lock(&new->fa_lock);
                               local_irq_disable();
                               lock(&ctx->completion_lock);
                               lock(&new->fa_lock);
  <Interrupt>
    lock(&ctx->completion_lock);

 *** DEADLOCK ***

Move kill_fasync() out of io_commit_cqring() to io_cqring_ev_posted(),
so it doesn't hold completion_lock while doing it. That saves from the
reported deadlock, and it's just nice to shorten the locking time and
untangle nested locks (compl_lock -> wq_head::lock).

Reported-by: [email protected]
Signed-off-by: Pavel Begunkov <[email protected]>
---
 fs/io_uring.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index ca46f314640b..00dd85acd039 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -1342,11 +1342,6 @@ static void __io_commit_cqring(struct io_ring_ctx *ctx)
 
 	/* order cqe stores with ring update */
 	smp_store_release(&rings->cq.tail, ctx->cached_cq_tail);
-
-	if (wq_has_sleeper(&ctx->cq_wait)) {
-		wake_up_interruptible(&ctx->cq_wait);
-		kill_fasync(&ctx->cq_fasync, SIGIO, POLL_IN);
-	}
 }
 
 static void io_put_identity(struct io_uring_task *tctx, struct io_kiocb *req)
@@ -1710,6 +1705,10 @@ static void io_cqring_ev_posted(struct io_ring_ctx *ctx)
 		wake_up(&ctx->sq_data->wait);
 	if (io_should_trigger_evfd(ctx))
 		eventfd_signal(ctx->cq_ev_fd, 1);
+	if (wq_has_sleeper(&ctx->cq_wait)) {
+		wake_up_interruptible(&ctx->cq_wait);
+		kill_fasync(&ctx->cq_fasync, SIGIO, POLL_IN);
+	}
 }
 
 /* Returns true if there are no backlogged entries after the flush */
-- 
2.24.0


  reply	other threads:[~2021-01-02 16:12 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-02 16:06 [PATCH 0/4] bunch of random fixes Pavel Begunkov
2021-01-02 16:06 ` Pavel Begunkov [this message]
2021-01-03 11:58   ` [PATCH 1/4] io_uring: dont kill fasync under completion_lock Pavel Begunkov
2021-01-02 16:06 ` [PATCH 2/4] io_uring: patch up IOPOLL overflow_flush sync Pavel Begunkov
2021-01-03 15:12   ` Jens Axboe
2021-01-03 16:29     ` Pavel Begunkov
2021-01-03 22:05       ` Jens Axboe
2021-01-02 16:06 ` [PATCH 3/4] io_uring: drop file refs after task cancel Pavel Begunkov
2021-01-02 16:06 ` [PATCH 4/4] io_uring: cancel more aggressively in exit_work Pavel Begunkov

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=e51028690f7415a018403d3607739693188b5f7b.1609600704.git.asml.silence@gmail.com \
    [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