From: Davidlohr Bueso <dave@stgolabs.net>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
Steven Rostedt <rostedt@goodmis.org>,
Mike Galbraith <umgwanakikbuti@gmail.com>,
"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH -rfc 4/4] locking/rtmutex: Support spin on owner (osq)
Date: Mon, 08 Jun 2015 21:41:42 -0700 [thread overview]
Message-ID: <1433824902.3165.61.camel@stgolabs.net> (raw)
In-Reply-To: <alpine.DEB.2.11.1506051443250.7723@nanos>
On Fri, 2015-06-05 at 15:59 +0200, Thomas Gleixner wrote:
> On Tue, 19 May 2015, Davidlohr Bueso wrote:
> >
> > +/*
> > + * Lockless alternative to rt_mutex_has_waiters() as we do not need the
> > + * wait_lock to check if we are in, for instance, a transitional state
> > + * after calling mark_rt_mutex_waiters().
>
> Before I get into a state of brain melt, could you please explain that
> in an understandable way?
With that I meant that we could check the owner to see if the
RT_MUTEX_HAS_WAITERS bit was set without taking the wait_lock and no
owner.
>
> rt_mutex_has_waiters() looks at the root pointer of the rbtree head
> whether that's empty. You can do a lockless check of that as well,
> right? So what's the FAST part of that function and how is that
> related to a point after we called mark_rt_mutex_waiters()?
You're right, we could use rt_mutex_has_waiters(). When I thought of
this originally, I was considering something like:
if (rt_mutex_has_waiters(lock)) {
if (current->prio >= rt_mutex_top_waiter(lock)->prio)
...
Which obviously requires the wait_lock, but I did not consider just
using the tree. However, the consequence I see in doing this is that we
would miss scenarios where mark_rt_mutex_waiters() is called (under nil
owner, for example), so we would force tasks to block only when there
are truly waiters.
> > + */
> > +static inline bool rt_mutex_has_waiters_fast(struct rt_mutex *lock)
> > +{
> > + unsigned long val = (unsigned long)lock->owner;
> > +
> > + if (!val)
> > + return false;
> > + return val & RT_MUTEX_HAS_WAITERS;
> > +}
> > +
>
> > +/*
> > + * Initial check for entering the mutex spinning loop
> > + */
> > +static inline bool rt_mutex_can_spin_on_owner(struct rt_mutex *lock)
> > +{
> > + struct task_struct *owner;
> > + /* default return to spin: if no owner, the lock is free */
>
>
> Rather than having a comment in the middle of the variable declaration
> section, I'd prefer a comment explaing the whole logic of this
> function.
Ok.
> > + int ret = true;
>
> > +static bool rt_mutex_optimistic_spin(struct rt_mutex *lock)
> > +{
> > + bool taken = false;
> > +
> > + preempt_disable();
> > +
> > + if (!rt_mutex_can_spin_on_owner(lock))
> > + goto done;
> > + /*
> > + * In order to avoid a stampede of mutex spinners trying to
> > + * acquire the mutex all at once, the spinners need to take a
> > + * MCS (queued) lock first before spinning on the owner field.
> > + */
> > + if (!osq_lock(&lock->osq))
> > + goto done;
>
> Hmm. The queue lock is serializing potential spinners, right?
Yes.
>
> So that's going to lead to a potential priority ordering problem
> because if a lower prio task wins the racing to the ocq_lock queue,
> then the higher prio waiter will be queued behind and blocked from
> taking the lock first.
Hmm yes, ocq is a fair lock. However I believe this is mitigated by (a)
the conservative spinning approach, and (b) by osq_lock's need_resched()
check, so at least a spinner will abort if a higher prio task comes in.
But of course, this only deals with spinners, and we cannot account for
a lower prio owner task.
So if this is not acceptable, I guess we'll have to do without the mcs
like properties.
Thanks,
Davidlohr
next prev parent reply other threads:[~2015-06-09 4:42 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-19 17:24 [PATCH -tip 0/4] rtmutex: Spin on owner Davidlohr Bueso
2015-05-19 17:24 ` [PATCH 1/4] locking/rtmutex: Implement lockless top-waiter wakeup Davidlohr Bueso
2015-06-05 12:35 ` Thomas Gleixner
2015-06-16 19:29 ` [PATCH] futex: lower the lock contention on the HB lock during wake up Sebastian Andrzej Siewior
2015-06-16 19:50 ` Davidlohr Bueso
2015-06-17 8:33 ` [PATCH v2] " Sebastian Andrzej Siewior
2015-06-17 14:17 ` Mike Galbraith
2015-06-17 14:28 ` Sebastian Andrzej Siewior
2015-06-17 14:31 ` Mike Galbraith
2015-06-21 4:35 ` Mike Galbraith
2015-06-18 20:30 ` [tip:sched/core] futex: Lower " tip-bot for Sebastian Andrzej Siewior
2015-06-19 17:51 ` [PATCH v2] futex: lower " Kevin Hilman
2015-06-19 18:54 ` Thomas Gleixner
2015-06-19 19:32 ` Kevin Hilman
2015-06-19 19:33 ` [tip:sched/locking] futex: Lower " tip-bot for Sebastian Andrzej Siewior
2015-06-18 20:30 ` [tip:sched/core] locking/rtmutex: Implement lockless top-waiter wakeup tip-bot for Davidlohr Bueso
2015-05-19 17:24 ` [PATCH 2/4] locking/rtmutex: Use cmp-cmpxchg Davidlohr Bueso
2015-06-05 12:38 ` Thomas Gleixner
2015-06-06 15:27 ` Davidlohr Bueso
2015-06-15 18:34 ` Jason Low
2015-06-15 19:37 ` Davidlohr Bueso
2015-06-16 1:00 ` Jason Low
2015-05-19 17:24 ` [PATCH 3/4] locking/rtmutex: Update stale plist comments Davidlohr Bueso
2015-06-05 12:39 ` Thomas Gleixner
2015-06-18 20:57 ` [tip:sched/core] " tip-bot for Davidlohr Bueso
2015-06-19 19:33 ` [tip:sched/locking] " tip-bot for Davidlohr Bueso
2015-05-19 17:24 ` [PATCH -rfc 4/4] locking/rtmutex: Support spin on owner (osq) Davidlohr Bueso
2015-05-20 7:11 ` Paul Bolle
2015-05-25 20:35 ` Davidlohr Bueso
2015-05-29 15:19 ` Davidlohr Bueso
2015-05-29 18:01 ` Davidlohr Bueso
2015-06-05 13:59 ` Thomas Gleixner
2015-06-09 4:41 ` Davidlohr Bueso [this message]
2015-06-09 9:29 ` Thomas Gleixner
2015-06-09 11:21 ` Peter Zijlstra
2015-06-09 12:53 ` Thomas Gleixner
2015-05-25 20:35 ` [PATCH -tip 0/4] rtmutex: Spin on owner Davidlohr Bueso
2015-05-26 19:05 ` Thomas Gleixner
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=1433824902.3165.61.camel@stgolabs.net \
--to=dave@stgolabs.net \
--cc=bigeasy@linutronix.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=paulmck@linux.vnet.ibm.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=umgwanakikbuti@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;
as well as URLs for NNTP newsgroup(s).