mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Eric Dumazet <eric.dumazet@gmail.com>
To: Michel Lespinasse <walken@google.com>
Cc: Rik van Riel <riel@redhat.com>,
	linux-kernel@vger.kernel.org, aquini@redhat.com,
	lwoodman@redhat.com, jeremy@goop.org,
	Jan Beulich <JBeulich@novell.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	knoel@redhat.com
Subject: Re: [RFC PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address
Date: Thu, 03 Jan 2013 05:05:43 -0800	[thread overview]
Message-ID: <1357218343.21409.24347.camel@edumazet-glaptop> (raw)
In-Reply-To: <CANN689FQK0c4VCPvT-Zg0vFLApOCMed7BwEEaR_ixFjO3gwBGQ@mail.gmail.com>

On Thu, 2013-01-03 at 04:48 -0800, Michel Lespinasse wrote:
> On Wed, Jan 2, 2013 at 9:24 PM, Rik van Riel <riel@redhat.com> wrote:
> > From: Eric Dumazet <eric.dumazet@gmail.com>
> >
> > Eric Dumazet found a regression with the spinlock backoff code,
> > in workloads where multiple spinlocks were contended, each having
> > a different wait time.
> 
> I think you should really clarify that the regression was observed
> with version 1 of your proposal. At that time,
> 
> 1- the autotune code tended to use too long delays for long held locks, and
> 
> 2- there was no exponential backoff, which meant that sharing stats
> between a long held and a short held spinlock could really hurt
> throughput on the short held spinlock
> 
> 
> I believe that with autotune v2, this really shouldnt be a problem and
> stats sharing should result in just using a delay that's appropriate
> for the shorter of the two lock hold times - which is not the optimal
> value, but is actually close enough performance wise and, most
> importantly, should not cause any regression when compared to current
> mainline.
> 
> (it's important to point that out because otherwise, you might trick
> Linus into thinking your patches are risky, which I think they
> shouldn't be after you implemented exponential backoff)
> 
> >  void ticket_spin_lock_wait(arch_spinlock_t *lock, struct __raw_tickets inc)
> >  {
> >         __ticket_t head = inc.head, ticket = inc.tail;
> >         __ticket_t waiters_ahead;
> > -       int delay = __this_cpu_read(spinlock_delay);
> > +       u32 hash = hash32_ptr(lock);
> > +       u32 slot = hash_32(hash, DELAY_HASH_SHIFT);
> > +       struct delay_entry *ent = &__get_cpu_var(spinlock_delay[slot]);
> > +       u32 delay = (ent->hash == hash) ? ent->delay : MIN_SPINLOCK_DELAY;
> 
> IMO we want to avoid MIN_SPINLOCK_DELAY here. The exponential backoff
> autotune should make us resilient to collisions (if they happen we'll
> just end up with something very close to the min of the delays that
> would have been appropriate for either locks), so it should be better
> to just let collisions happen rather than force the use of
> MIN_SPINLOCK_DELAY.
> 

exponential backoff wont help, I tried this idea last week and found
that its better to detect hash collision and safely use
MIN_SPINLOCK_DELAY in this case.

Its better to not overestimate the delay and spin much longer than
needed.

On a hash collision, we dont know at all the contention history of this
lock, unless we store the EWMA delay inside the lock.

(On x86 and NR_CPUS <= 256, we have a 16 bit hole in the spinlock that
 we could use for this)




  reply	other threads:[~2013-01-03 13:05 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-01-03  5:15 [RFC PATCH 0/5] x86,smp: make ticket spinlock proportional backoff w/ auto tuning Rik van Riel
2013-01-03  5:18 ` [RFC PATCH 1/5] x86,smp: move waiting on contended ticket lock out of line Rik van Riel
2013-01-03 10:47   ` Michel Lespinasse
2013-01-03  5:22 ` [RFC PATCH 2/5] x86,smp: proportional backoff for ticket spinlocks Rik van Riel
2013-01-03 11:35   ` Raghavendra KT
2013-01-03 11:42     ` Michel Lespinasse
2013-01-03 18:19       ` Raghavendra K T
2013-01-03  5:23 ` [RFC PATCH 3/5] x86,smp: auto tune spinlock backoff delay factor Rik van Riel
2013-01-03 12:31   ` Michel Lespinasse
2013-01-03 17:17     ` Rik van Riel
2013-01-05  0:45       ` Rik van Riel
2013-01-03  5:24 ` [RFC PATCH 4/5] x86,smp: keep spinlock delay values per hashed spinlock address Rik van Riel
2013-01-03 12:48   ` Michel Lespinasse
2013-01-03 13:05     ` Eric Dumazet [this message]
2013-01-03  5:25 ` [RFC PATCH 5/5] x86,smp: add debugging code to track spinlock delay value Rik van Riel
2013-01-03 10:46 ` [RFC PATCH 0/5] x86,smp: make ticket spinlock proportional backoff w/ auto tuning Michel Lespinasse
2013-01-03 11:29 ` Raghavendra KT

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=1357218343.21409.24347.camel@edumazet-glaptop \
    --to=eric.dumazet@gmail.com \
    --cc=JBeulich@novell.com \
    --cc=aquini@redhat.com \
    --cc=jeremy@goop.org \
    --cc=knoel@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lwoodman@redhat.com \
    --cc=riel@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=walken@google.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

all inboxes | Powered by JetHome®