linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: LKML <linux-kernel@vger.kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@kernel.org>,
	Juri Lelli <juri.lelli@redhat.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	Daniel Bristot de Oliveira <bristot@redhat.com>,
	Will Deacon <will@kernel.org>, Waiman Long <longman@redhat.com>,
	Boqun Feng <boqun.feng@gmail.com>,
	Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
	Davidlohr Bueso <dave@stgolabs.net>
Subject: [patch 50/50] locking/rtmutex: Add adaptive spinwait mechanism
Date: Tue, 13 Jul 2021 17:11:44 +0200	[thread overview]
Message-ID: <20210713160750.848393199@linutronix.de> (raw)
In-Reply-To: 20210713151054.700719949@linutronix.de

From: Steven Rostedt <rostedt@goodmis.org>

Going to sleep when a spinlock or rwlock is contended can be quite
inefficient when the contention time is short and the lock owner is running
on a different CPU. The MCS mechanism is not applicable to rtmutex based
locks, so provide a simple adaptive spinwait mechanism for the RT specific
spin/rwlock implementations.

[ tglx: Provide a contemporary changelog ]

Originally-by: Gregory Haskins <ghaskins@novell.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 kernel/locking/rtmutex.c |   50 ++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 49 insertions(+), 1 deletion(-)
---
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -8,6 +8,11 @@
  *  Copyright (C) 2005-2006 Timesys Corp., Thomas Gleixner <tglx@timesys.com>
  *  Copyright (C) 2005 Kihon Technologies Inc., Steven Rostedt
  *  Copyright (C) 2006 Esben Nielsen
+ * Adaptive Spinlocks:
+ *  Copyright (C) 2008 Novell, Inc., Gregory Haskins, Sven Dietrich,
+ *				     and Peter Morreale,
+ * Adaptive Spinlocks simplification:
+ *  Copyright (C) 2008 Red Hat, Inc., Steven Rostedt <srostedt@redhat.com>
  *
  *  See Documentation/locking/rt-mutex-design.rst for details.
  */
@@ -1433,6 +1438,43 @@ static __always_inline int __rt_mutex_lo
  * Functions required for spin/rw_lock substitution on RT kernels
  */
 
+#ifdef CONFIG_SMP
+/*
+ * Note that owner is a speculative pointer and dereferencing relies
+ * on rcu_read_lock() and the check against the lock owner.
+ */
+static bool rtlock_adaptive_spinwait(struct rt_mutex *lock,
+				     struct task_struct *owner)
+{
+	bool res = true;
+
+	rcu_read_lock();
+	for (;;) {
+		/* Owner changed. Trylock again */
+		if (owner != rt_mutex_owner(lock))
+			break;
+		/*
+		 * Ensure that owner->on_cpu is dereferenced _after_
+		 * checking the above to be valid.
+		 */
+		barrier();
+		if (!owner->on_cpu) {
+			res = false;
+			break;
+		}
+		cpu_relax();
+	}
+	rcu_read_unlock();
+	return res;
+}
+#else
+static bool rtlock_adaptive_spinwait(struct rt_mutex *lock,
+				     struct task_struct *owner)
+{
+	return false;
+}
+#endif
+
 /**
  * rtlock_slowlock_locked - Slow path lock acquisition for RT locks
  * @lock:	The underlying rt mutex
@@ -1440,6 +1482,7 @@ static __always_inline int __rt_mutex_lo
 static void __sched rtlock_slowlock_locked(struct rt_mutex *lock)
 {
 	struct rt_mutex_waiter waiter;
+	struct task_struct *owner;
 
 	lockdep_assert_held(&lock->wait_lock);
 
@@ -1458,9 +1501,14 @@ static void __sched rtlock_slowlock_lock
 		if (try_to_take_rt_mutex(lock, current, &waiter))
 			break;
 
+		if (&waiter == rt_mutex_top_waiter(lock))
+			owner = rt_mutex_owner(lock);
+		else
+			owner = NULL;
 		raw_spin_unlock_irq(&lock->wait_lock);
 
-		schedule_rtlock();
+		if (!owner || !rtlock_adaptive_spinwait(lock, owner))
+			schedule_rtlock();
 
 		raw_spin_lock_irq(&lock->wait_lock);
 		set_current_state(TASK_RTLOCK_WAIT);


  parent reply	other threads:[~2021-07-13 16:15 UTC|newest]

Thread overview: 77+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-13 15:10 [patch 00/50] locking, sched: The PREEMPT-RT locking infrastructure Thomas Gleixner
2021-07-13 15:10 ` [patch 01/50] sched: Split out the wakeup state check Thomas Gleixner
2021-07-13 15:10 ` [patch 02/50] sched: Introduce TASK_RTLOCK_WAIT Thomas Gleixner
2021-07-13 15:10 ` [patch 03/50] sched: Prepare for RT sleeping spin/rwlocks Thomas Gleixner
2021-07-13 19:52   ` Waiman Long
2021-07-14 23:20   ` Valentin Schneider
2021-07-15  9:27     ` Peter Zijlstra
2021-07-15 14:08       ` Valentin Schneider
2021-07-13 15:10 ` [patch 04/50] sched: Rework the __schedule() preempt argument Thomas Gleixner
2021-07-13 20:04   ` Waiman Long
2021-07-13 15:10 ` [patch 05/50] sched: Provide schedule point for RT locks Thomas Gleixner
2021-07-14  8:28   ` Peter Zijlstra
2021-07-14  9:49     ` Thomas Gleixner
2021-07-14 10:17       ` Peter Zijlstra
2021-07-14 11:32         ` Thomas Gleixner
2021-07-13 15:11 ` [patch 06/50] sched/wake_q: Provide WAKE_Q_HEAD_INITIALIZER Thomas Gleixner
2021-07-13 15:11 ` [patch 07/50] rtmutex: Convert macros to inlines Thomas Gleixner
2021-07-13 15:11 ` [patch 08/50] rtmutex: Switch to try_cmpxchg() Thomas Gleixner
2021-07-13 15:11 ` [patch 09/50] rtmutex: Split API and implementation Thomas Gleixner
2021-07-13 15:11 ` [patch 10/50] locking/rtmutex: Provide rt_mutex_slowlock_locked() Thomas Gleixner
2021-07-27 17:20   ` Valentin Schneider
2021-07-27 19:02     ` Thomas Gleixner
2021-07-13 15:11 ` [patch 11/50] locking/rtmutex: Provide lockdep less variants of rtmutex interfaces Thomas Gleixner
2021-07-13 15:11 ` [patch 12/50] locking: Add base code for RT rw_semaphore and rwlock Thomas Gleixner
2021-07-13 15:11 ` [patch 13/50] locking/rwsem: Add rtmutex based R/W semaphore implementation Thomas Gleixner
2021-07-13 15:11 ` [patch 14/50] locking/rtmutex: Add wake_state to rt_mutex_waiter Thomas Gleixner
2021-07-14 10:18   ` Peter Zijlstra
2021-07-14 11:33     ` Thomas Gleixner
2021-07-13 15:11 ` [patch 15/50] locking/rtmutex: Provide rt_mutex_wake_q and helpers Thomas Gleixner
2021-07-13 15:11 ` [patch 16/50] locking/rtmutex: Use rt_mutex_wake_q_head Thomas Gleixner
2021-07-14  8:55   ` Peter Zijlstra
2021-07-14  9:51     ` Thomas Gleixner
2021-07-13 15:11 ` [patch 17/50] locking/rtmutex: Prepare RT rt_mutex_wake_q for RT locks Thomas Gleixner
2021-07-14  8:55   ` Peter Zijlstra
2021-07-14  9:52     ` Thomas Gleixner
2021-07-13 15:11 ` [patch 18/50] locking/rtmutex: Guard regular sleeping locks specific functions Thomas Gleixner
2021-07-13 15:11 ` [patch 19/50] locking/spinlock: Split the lock types header Thomas Gleixner
2021-07-13 15:11 ` [patch 20/50] locking/rtmutex: Prevent future include recursion hell Thomas Gleixner
2021-07-13 15:11 ` [patch 21/50] locking/lockdep: Reduce includes in debug_locks.h Thomas Gleixner
2021-07-13 15:11 ` [patch 22/50] rbtree: Split out the rbtree type definitions Thomas Gleixner
2021-07-14  9:24   ` Peter Zijlstra
2021-07-14  9:31     ` Arnaldo
2021-07-13 15:11 ` [patch 23/50] locking/rtmutex: Include only rbtree types Thomas Gleixner
2021-07-13 15:11 ` [patch 24/50] locking/spinlock: Provide RT specific spinlock type Thomas Gleixner
2021-07-14  9:54   ` Peter Zijlstra
2021-07-14 10:07     ` [PATCH] media/atomisp: Use lockdep instead of *mutex_is_locked() Peter Zijlstra
2021-07-13 15:11 ` [patch 25/50] locking/spinlock: Provide RT variant header Thomas Gleixner
2021-07-13 15:11 ` [patch 26/50] locking/rtmutex: Provide the spin/rwlock core lock function Thomas Gleixner
2021-07-14 10:59   ` Peter Zijlstra
2021-07-13 15:11 ` [patch 27/50] locking/spinlock: Provide RT variant Thomas Gleixner
2021-07-14 11:25   ` Peter Zijlstra
2021-07-13 15:11 ` [patch 28/50] locking/rwlock: " Thomas Gleixner
2021-07-13 15:11 ` [patch 29/50] locking/mutex: Consolidate core headers Thomas Gleixner
2021-07-13 15:11 ` [patch 30/50] locking/mutex: Move waiter to core header Thomas Gleixner
2021-07-13 15:11 ` [patch 31/50] locking/ww_mutex: Move ww_mutex declarations into ww_mutex.h Thomas Gleixner
2021-07-13 15:11 ` [patch 32/50] locking/mutex: Make mutex::wait_lock raw Thomas Gleixner
2021-07-13 15:11 ` [patch 33/50] locking/mutex: Introduce _mutex_t Thomas Gleixner
2021-07-15 17:31   ` Peter Zijlstra
2021-07-13 15:11 ` [patch 34/50] locking/mutex: Rename the ww_mutex relevant functions Thomas Gleixner
2021-07-13 15:11 ` [patch 35/50] locking/ww_mutex: Switch to _mutex_t Thomas Gleixner
2021-07-13 15:11 ` [patch 36/50] locking/mutex: Replace struct mutex in core code Thomas Gleixner
2021-07-13 15:11 ` [patch 37/50] locking/mutex: Rearrange items in mutex.h Thomas Gleixner
2021-07-13 15:11 ` [patch 38/50] locking/mutex: Exclude non-ww_mutex API for RT Thomas Gleixner
2021-07-13 15:11 ` [patch 39/50] locking/rtmutex: Add mutex variant " Thomas Gleixner
2021-07-13 15:11 ` [patch 40/50] lib/test_lockup: Adapt to changed variables Thomas Gleixner
2021-07-13 15:11 ` [patch 41/50] futex: Validate waiter correctly in futex_proxy_trylock_atomic() Thomas Gleixner
2021-07-13 15:11 ` [patch 42/50] futex: Cleanup stale comments Thomas Gleixner
2021-07-28 23:28   ` André Almeida
2021-07-13 15:11 ` [patch 43/50] futex: Correct the number of requeued waiters for PI Thomas Gleixner
2021-07-13 15:11 ` [patch 44/50] futex: Restructure futex_requeue() Thomas Gleixner
2021-07-13 15:11 ` [patch 45/50] futex: Clarify comment in futex_requeue() Thomas Gleixner
2021-07-13 15:11 ` [patch 46/50] futex: Prevent requeue_pi() lock nesting issue on RT Thomas Gleixner
2021-07-13 15:11 ` [patch 47/50] rtmutex: Prevent lockdep false positive with PI futexes Thomas Gleixner
2021-07-13 15:11 ` [patch 48/50] preempt: Adjust PREEMPT_LOCK_OFFSET for RT Thomas Gleixner
2021-07-13 15:11 ` [patch 49/50] locking/rtmutex: Implement equal priority lock stealing Thomas Gleixner
2021-07-13 15:11 ` Thomas Gleixner [this message]
2021-07-27 17:19 ` [patch 00/50] locking, sched: The PREEMPT-RT locking infrastructure Valentin Schneider

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=20210713160750.848393199@linutronix.de \
    --to=tglx@linutronix.de \
    --cc=bigeasy@linutronix.de \
    --cc=boqun.feng@gmail.com \
    --cc=bristot@redhat.com \
    --cc=dave@stgolabs.net \
    --cc=juri.lelli@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=longman@redhat.com \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    --cc=will@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;
as well as URLs for NNTP newsgroup(s).