public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Frederic Weisbecker <frederic@kernel.org>
To: "Paul E . McKenney" <paulmck@kernel.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Frederic Weisbecker <frederic@kernel.org>,
	rcu <rcu@vger.kernel.org>, Uladzislau Rezki <urezki@gmail.com>,
	Neeraj Upadhyay <quic_neeraju@quicinc.com>,
	Boqun Feng <boqun.feng@gmail.com>,
	Joel Fernandes <joel@joelfernandes.org>
Subject: [PATCH 04/10] rcu/nocb: Remove needless full barrier after callback advancing
Date: Fri,  8 Sep 2023 22:35:57 +0200	[thread overview]
Message-ID: <20230908203603.5865-5-frederic@kernel.org> (raw)
In-Reply-To: <20230908203603.5865-1-frederic@kernel.org>

A full barrier is issued from nocb_gp_wait() upon callbacks advancing
to order grace period completion with callbacks execution.

However these two events are already ordered by the
smp_mb__after_unlock_lock() barrier within the call to
raw_spin_lock_rcu_node() that is necessary for callbacks advancing to
happen.

The following litmus test shows the kind of guarantee that this barrier
provides:

	C smp_mb__after_unlock_lock

	{}

	// rcu_gp_cleanup()
	P0(spinlock_t *rnp_lock, int *gpnum)
	{
		// Grace period cleanup increase gp sequence number
		spin_lock(rnp_lock);
		WRITE_ONCE(*gpnum, 1);
		spin_unlock(rnp_lock);
	}

	// nocb_gp_wait()
	P1(spinlock_t *rnp_lock, spinlock_t *nocb_lock, int *gpnum, int *cb_ready)
	{
		int r1;

		// Call rcu_advance_cbs() from nocb_gp_wait()
		spin_lock(nocb_lock);
		spin_lock(rnp_lock);
		smp_mb__after_unlock_lock();
		r1 = READ_ONCE(*gpnum);
		WRITE_ONCE(*cb_ready, 1);
		spin_unlock(rnp_lock);
		spin_unlock(nocb_lock);
	}

	// nocb_cb_wait()
	P2(spinlock_t *nocb_lock, int *cb_ready, int *cb_executed)
	{
		int r2;

		// rcu_do_batch() -> rcu_segcblist_extract_done_cbs()
		spin_lock(nocb_lock);
		r2 = READ_ONCE(*cb_ready);
		spin_unlock(nocb_lock);

		// Actual callback execution
		WRITE_ONCE(*cb_executed, 1);
	}

	P3(int *cb_executed, int *gpnum)
	{
		int r3;

		WRITE_ONCE(*cb_executed, 2);
		smp_mb();
		r3 = READ_ONCE(*gpnum);
	}

	exists (1:r1=1 /\ 2:r2=1 /\ cb_executed=2 /\ 3:r3=0) (* Bad outcome. *)

Here the bad outcome only occurs if the smp_mb__after_unlock_lock() is
removed. This barrier orders the grace period completion against
callbacks advancing and even later callbacks invocation, thanks to the
opportunistic propagation via the ->nocb_lock to nocb_cb_wait().

Therefore the smp_mb() placed after callbacks advancing can be safely
removed.

Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
---
 kernel/rcu/tree_nocb.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/kernel/rcu/tree_nocb.h b/kernel/rcu/tree_nocb.h
index 6e63ba4788e1..2dc76f5e6e78 100644
--- a/kernel/rcu/tree_nocb.h
+++ b/kernel/rcu/tree_nocb.h
@@ -779,7 +779,6 @@ static void nocb_gp_wait(struct rcu_data *my_rdp)
 		if (rcu_segcblist_ready_cbs(&rdp->cblist)) {
 			needwake = rdp->nocb_cb_sleep;
 			WRITE_ONCE(rdp->nocb_cb_sleep, false);
-			smp_mb(); /* CB invocation -after- GP end. */
 		} else {
 			needwake = false;
 		}
-- 
2.41.0


  parent reply	other threads:[~2023-09-08 20:36 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-08 20:35 [PATCH 00/10] rcu cleanups Frederic Weisbecker
2023-09-08 20:35 ` [PATCH 01/10] rcu: Use rcu_segcblist_segempty() instead of open coding it Frederic Weisbecker
2023-10-02 15:38   ` Paul E. McKenney
2023-09-08 20:35 ` [PATCH 02/10] rcu: Rename jiffies_till_flush to jiffies_lazy_flush Frederic Weisbecker
2023-09-09  1:07   ` Joel Fernandes
2023-09-10 19:48     ` Frederic Weisbecker
2023-09-08 20:35 ` [PATCH 03/10] rcu/nocb: Remove needless LOAD-ACQUIRE Frederic Weisbecker
2023-09-09  1:48   ` Joel Fernandes
2023-09-09  1:50     ` Joel Fernandes
2023-09-10 21:17     ` Frederic Weisbecker
2023-09-08 20:35 ` Frederic Weisbecker [this message]
2023-09-09  4:31   ` [PATCH 04/10] rcu/nocb: Remove needless full barrier after callback advancing Joel Fernandes
2023-09-09 18:22     ` Boqun Feng
2023-09-10  4:09       ` Joel Fernandes
2023-09-10 10:22         ` Paul E. McKenney
2023-09-10 20:17         ` Frederic Weisbecker
2023-09-10 20:29           ` Frederic Weisbecker
2023-09-08 20:35 ` [PATCH 05/10] rcu: Assume IRQS disabled from rcu_report_dead() Frederic Weisbecker
2023-10-02 15:41   ` Paul E. McKenney
2023-09-08 20:35 ` [PATCH 06/10] rcu: Assume rcu_report_dead() is always called locally Frederic Weisbecker
2023-10-02 15:45   ` Paul E. McKenney
2023-09-08 20:36 ` [PATCH 07/10] rcu: Conditionally build CPU-hotplug teardown callbacks Frederic Weisbecker
2023-10-04 16:57   ` Paul E. McKenney
2023-09-08 20:36 ` [PATCH 08/10] rcu: Standardize explicit CPU-hotplug calls Frederic Weisbecker
2023-10-02 15:47   ` Paul E. McKenney
2023-09-08 20:36 ` [PATCH 09/10] rcu: Remove references to rcu_migrate_callbacks() from diagrams Frederic Weisbecker
2023-10-02 15:52   ` Paul E. McKenney
2023-09-08 20:36 ` [PATCH 10/10] rcu: Comment why callbacks migration can't wait for CPUHP_RCUTREE_PREP Frederic Weisbecker
2023-10-02 15:48   ` Paul E. McKenney

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=20230908203603.5865-5-frederic@kernel.org \
    --to=frederic@kernel.org \
    --cc=boqun.feng@gmail.com \
    --cc=joel@joelfernandes.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=paulmck@kernel.org \
    --cc=quic_neeraju@quicinc.com \
    --cc=rcu@vger.kernel.org \
    --cc=urezki@gmail.com \
    /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