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], [email protected]
Subject: [PATCH 06/11] io_uring/rsrc: kill rsrc_ref_lock
Date: Thu, 30 Mar 2023 15:53:24 +0100	[thread overview]
Message-ID: <80e76722fb1317d5315528b8b57de559ca5aea33.1680187408.git.asml.silence@gmail.com> (raw)
In-Reply-To: <[email protected]>

We use ->rsrc_ref_lock spinlock to protect ->rsrc_ref_list in
io_rsrc_node_ref_zero(). Now we removed pcpu refcounting, which means
io_rsrc_node_ref_zero() is not executed from the irq context as an RCU
callback anymore, and we also put it under ->uring_lock.
io_rsrc_node_switch(), which queues up nodes into the list, is also
protected by ->uring_lock, so we can safely get rid of ->rsrc_ref_lock.

Signed-off-by: Pavel Begunkov <[email protected]>
---
 include/linux/io_uring_types.h | 2 +-
 io_uring/io_uring.c            | 1 -
 io_uring/rsrc.c                | 5 -----
 3 files changed, 1 insertion(+), 7 deletions(-)

diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
index a0a5b5964d3a..9492889f00c0 100644
--- a/include/linux/io_uring_types.h
+++ b/include/linux/io_uring_types.h
@@ -333,8 +333,8 @@ struct io_ring_ctx {
 	struct delayed_work		rsrc_put_work;
 	struct callback_head		rsrc_put_tw;
 	struct llist_head		rsrc_put_llist;
+	/* protected by ->uring_lock */
 	struct list_head		rsrc_ref_list;
-	spinlock_t			rsrc_ref_lock;
 
 	struct list_head		io_buffers_pages;
 
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index e55c48d91209..e94780c0a024 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -325,7 +325,6 @@ static __cold struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
 	INIT_LIST_HEAD(&ctx->defer_list);
 	INIT_LIST_HEAD(&ctx->timeout_list);
 	INIT_LIST_HEAD(&ctx->ltimeout_list);
-	spin_lock_init(&ctx->rsrc_ref_lock);
 	INIT_LIST_HEAD(&ctx->rsrc_ref_list);
 	INIT_DELAYED_WORK(&ctx->rsrc_put_work, io_rsrc_put_work);
 	init_task_work(&ctx->rsrc_put_tw, io_rsrc_put_tw);
diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c
index 1237fc77c250..e122b6e5f9c5 100644
--- a/io_uring/rsrc.c
+++ b/io_uring/rsrc.c
@@ -209,11 +209,9 @@ void io_rsrc_node_ref_zero(struct io_rsrc_node *node)
 	__must_hold(&node->rsrc_data->ctx->uring_lock)
 {
 	struct io_ring_ctx *ctx = node->rsrc_data->ctx;
-	unsigned long flags;
 	bool first_add = false;
 	unsigned long delay = HZ;
 
-	spin_lock_irqsave(&ctx->rsrc_ref_lock, flags);
 	node->done = true;
 
 	/* if we are mid-quiesce then do not delay */
@@ -229,7 +227,6 @@ void io_rsrc_node_ref_zero(struct io_rsrc_node *node)
 		list_del(&node->node);
 		first_add |= llist_add(&node->llist, &ctx->rsrc_put_llist);
 	}
-	spin_unlock_irqrestore(&ctx->rsrc_ref_lock, flags);
 
 	if (!first_add)
 		return;
@@ -268,9 +265,7 @@ void io_rsrc_node_switch(struct io_ring_ctx *ctx,
 		struct io_rsrc_node *rsrc_node = ctx->rsrc_node;
 
 		rsrc_node->rsrc_data = data_to_kill;
-		spin_lock_irq(&ctx->rsrc_ref_lock);
 		list_add_tail(&rsrc_node->node, &ctx->rsrc_ref_list);
-		spin_unlock_irq(&ctx->rsrc_ref_lock);
 
 		atomic_inc(&data_to_kill->refs);
 		/* put master ref */
-- 
2.39.1


  parent reply	other threads:[~2023-03-30 14:56 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-30 14:53 [RFC 00/11] optimise registered buffer/file updates Pavel Begunkov
2023-03-30 14:53 ` [PATCH 01/11] io_uring/rsrc: use non-pcpu refcounts for nodes Pavel Begunkov
2023-03-30 14:53 ` [PATCH 02/11] io_uring/rsrc: keep cached refs per node Pavel Begunkov
2023-03-30 14:53 ` [PATCH 03/11] io_uring: don't put nodes under spinlocks Pavel Begunkov
2023-03-30 14:53 ` [PATCH 04/11] io_uring: io_free_req() via tw Pavel Begunkov
2023-03-30 14:53 ` [PATCH 05/11] io_uring/rsrc: protect node refs with uring_lock Pavel Begunkov
2023-03-30 14:53 ` Pavel Begunkov [this message]
2023-03-30 14:53 ` [PATCH 07/11] io_uring/rsrc: rename rsrc_list Pavel Begunkov
2023-03-30 14:53 ` [PATCH 08/11] io_uring/rsrc: optimise io_rsrc_put allocation Pavel Begunkov
2023-03-30 14:53 ` [PATCH 09/11] io_uring/rsrc: don't offload node free Pavel Begunkov
2023-03-30 14:53 ` [PATCH 10/11] io_uring/rsrc: cache struct io_rsrc_node Pavel Begunkov
2023-03-31 14:09   ` Gabriel Krisman Bertazi
2023-03-31 16:27     ` Pavel Begunkov
2023-04-01  0:04       ` Gabriel Krisman Bertazi
2023-04-04 13:21         ` Pavel Begunkov
2023-04-04 15:48           ` Gabriel Krisman Bertazi
2023-04-04 15:52             ` Jens Axboe
2023-04-04 16:53               ` Gabriel Krisman Bertazi
2023-04-04 18:26                 ` Pavel Begunkov
2023-03-30 14:53 ` [PATCH 11/11] io_uring/rsrc: add lockdep sanity checks Pavel Begunkov
2023-03-31 13:35 ` [RFC 00/11] optimise registered buffer/file updates Gabriel Krisman Bertazi
2023-03-31 16:21   ` Pavel Begunkov
2023-03-31 15:18 ` 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=80e76722fb1317d5315528b8b57de559ca5aea33.1680187408.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