public inbox for [email protected]
 help / color / mirror / Atom feed
From: Pavel Begunkov <[email protected]>
To: [email protected]
Cc: Jens Axboe <[email protected]>, [email protected]
Subject: [PATCH for-next 4/4] io_uring: consolidate hash_locked io-wq handling
Date: Thu,  7 Jul 2022 15:13:17 +0100	[thread overview]
Message-ID: <0ff0ffdfaa65b3d536131535c3dad3c63d9b7bb0.1657203020.git.asml.silence@gmail.com> (raw)
In-Reply-To: <[email protected]>

Don't duplicate code disabling REQ_F_HASH_LOCKED for IO_URING_F_UNLOCKED
(i.e. io-wq), move the handling into __io_arm_poll_handler().

Signed-off-by: Pavel Begunkov <[email protected]>
---
 io_uring/poll.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/io_uring/poll.c b/io_uring/poll.c
index 77b669b06046..76592063abe7 100644
--- a/io_uring/poll.c
+++ b/io_uring/poll.c
@@ -523,8 +523,12 @@ static int __io_arm_poll_handler(struct io_kiocb *req,
 	 * io_poll_can_finish_inline() tries to deal with that.
 	 */
 	ipt->owning = issue_flags & IO_URING_F_UNLOCKED;
-
 	atomic_set(&req->poll_refs, (int)ipt->owning);
+
+	/* io-wq doesn't hold uring_lock */
+	if (issue_flags & IO_URING_F_UNLOCKED)
+		req->flags &= ~REQ_F_HASH_LOCKED;
+
 	mask = vfs_poll(req->file, &ipt->pt) & poll->events;
 
 	if (unlikely(ipt->error || !ipt->nr_entries)) {
@@ -618,8 +622,7 @@ int io_arm_poll_handler(struct io_kiocb *req, unsigned issue_flags)
 	 * apoll requests already grab the mutex to complete in the tw handler,
 	 * so removal from the mutex-backed hash is free, use it by default.
 	 */
-	if (!(issue_flags & IO_URING_F_UNLOCKED))
-		req->flags |= REQ_F_HASH_LOCKED;
+	req->flags |= REQ_F_HASH_LOCKED;
 
 	if (!def->pollin && !def->pollout)
 		return IO_APOLL_ABORTED;
@@ -876,8 +879,7 @@ int io_poll_add(struct io_kiocb *req, unsigned int issue_flags)
 	 * If sqpoll or single issuer, there is no contention for ->uring_lock
 	 * and we'll end up holding it in tw handlers anyway.
 	 */
-	if (!(issue_flags & IO_URING_F_UNLOCKED) &&
-	    (req->ctx->flags & (IORING_SETUP_SQPOLL | IORING_SETUP_SINGLE_ISSUER)))
+	if (req->ctx->flags & (IORING_SETUP_SQPOLL|IORING_SETUP_SINGLE_ISSUER))
 		req->flags |= REQ_F_HASH_LOCKED;
 
 	ret = __io_arm_poll_handler(req, poll, &ipt, poll->events, issue_flags);
-- 
2.36.1



      parent reply	other threads:[~2022-07-07 14:14 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-07 14:13 [PATCH for-next 0/4] poll locking fixes Pavel Begunkov
2022-07-07 14:13 ` [PATCH for-next 1/4] io_uring: don't miss setting REQ_F_DOUBLE_POLL Pavel Begunkov
2022-07-07 14:13 ` [PATCH for-next 2/4] io_uring: don't race double poll setting REQ_F_ASYNC_DATA Pavel Begunkov
2022-07-07 14:13 ` [PATCH for-next 3/4] io_uring: clear REQ_F_HASH_LOCKED on hash removal Pavel Begunkov
2022-07-07 14:13 ` Pavel Begunkov [this message]

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=0ff0ffdfaa65b3d536131535c3dad3c63d9b7bb0.1657203020.git.asml.silence@gmail.com \
    [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