All of lore.kernel.org
 help / color / mirror / Atom feed
From: Davidlohr Bueso <dave@stgolabs.net>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Dave Chinner <david@fromorbit.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Waiman Long <Waiman.Long@hp.com>, Ingo Molnar <mingo@kernel.org>
Subject: Re: [4.2, Regression] Queued spinlocks cause major XFS performance regression
Date: Sun, 6 Sep 2015 17:05:46 -0700	[thread overview]
Message-ID: <20150907000546.GA27993@linux-q0g1.site> (raw)
In-Reply-To: <20150904152523.GR18673@twins.programming.kicks-ass.net>

On Fri, 04 Sep 2015, Peter Zijlstra wrote:

>-static inline bool virt_queued_spin_lock(struct qspinlock *lock)
>+static inline bool virt_spin_lock(struct qspinlock *lock)

Given that we fall back to the cmpxchg loop even when PARAVIRT is not in the
picture, I believe this function is horribly misnamed.

> {
> 	if (!static_cpu_has(X86_FEATURE_HYPERVISOR))
> 		return false;
>
>-	while (atomic_cmpxchg(&lock->val, 0, _Q_LOCKED_VAL) != 0)
>-		cpu_relax();
>+	/*
>+	 * On hypervisors without PARAVIRT_SPINLOCKS support we fall
>+	 * back to a Test-and-Set spinlock, because fair locks have
>+	 * horrible lock 'holder' preemption issues.
>+	 */
>+

This comment is also misleading... but if you tuck the whole function
under some PARAVIRT option, it obviously makes sense to just leave as is.
And let native actually _use_ qspinlocks.

>+	do {
>+		while (atomic_read(&lock->val) != 0)
>+			cpu_relax();
>+	} while (atomic_cmpxchg(&lock->val, 0, _Q_LOCKED_VAL) != 0);

CCAS to the rescue again.

Thanks,
Davidlohr

  parent reply	other threads:[~2015-09-07  0:06 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-09-04  5:48 [4.2, Regression] Queued spinlocks cause major XFS performance regression Dave Chinner
2015-09-04  6:39 ` Linus Torvalds
2015-09-04  7:11   ` Dave Chinner
2015-09-04  7:31     ` Juergen Gross
2015-09-04  7:55     ` Peter Zijlstra
2015-09-04  8:29     ` Dave Chinner
2015-09-04 15:05       ` Linus Torvalds
2015-09-04 15:14         ` Peter Zijlstra
2015-09-04 15:21           ` Linus Torvalds
2015-09-04 15:30             ` Peter Zijlstra
2015-09-04 15:54               ` Peter Zijlstra
2015-09-10  2:06                 ` Waiman Long
2015-09-04 15:58               ` Linus Torvalds
2015-09-05 17:45                 ` Peter Zijlstra
2015-09-04 15:25           ` Peter Zijlstra
2015-09-06 23:32             ` Dave Chinner
2015-09-07  0:05             ` Davidlohr Bueso [this message]
2015-09-07  6:57               ` Peter Zijlstra
2015-09-07 20:45                 ` Linus Torvalds
2015-09-08  6:37                   ` Davidlohr Bueso
2015-09-08 10:05                   ` Peter Zijlstra
2015-09-08 17:45                     ` Linus Torvalds
2015-09-13 10:55             ` [tip:locking/core] locking/qspinlock/x86: Fix performance regression under unaccelerated VMs tip-bot for Peter Zijlstra
2015-09-04  7:39   ` [4.2, Regression] Queued spinlocks cause major XFS performance regression Peter Zijlstra
2015-09-04  8:12     ` Dave Chinner
2015-09-04 11:32       ` Peter Zijlstra
2015-09-04 22:03         ` Dave Chinner
2015-09-06 23:47         ` Dave Chinner
2015-09-10  2:09           ` Waiman Long
     [not found]         ` <CAC=cRTOraeOeu3Z8C1qx6w=GMSzD_4VevrEzn0mMhrqy=7n3wQ@mail.gmail.com>
     [not found]           ` <56094F05.4090809@hpe.com>
2015-09-29  0:47             ` huang ying
2015-09-29  2:57               ` Waiman Long
2015-09-10  2:01 ` Waiman Long

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=20150907000546.GA27993@linux-q0g1.site \
    --to=dave@stgolabs.net \
    --cc=Waiman.Long@hp.com \
    --cc=david@fromorbit.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    --cc=torvalds@linux-foundation.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.