All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Byungchul Park <byungchul.park@lge.com>,
	LKML <linux-kernel@vger.kernel.org>,
	Chris Metcalf <cmetcalf@ezchip.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Luiz Capitulino <lcapitulino@redhat.com>,
	Christoph Lameter <cl@linux.com>,
	"Paul E . McKenney" <paulmck@linux.vnet.ibm.com>,
	Mike Galbraith <efault@gmx.de>, Rik van Riel <riel@redhat.com>
Subject: Re: [PATCH 1/4] sched: Don't account tickless CPU load on tick
Date: Fri, 29 Jan 2016 10:50:16 +0100	[thread overview]
Message-ID: <20160129095016.GZ6356@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <20160128160125.GB25866@lerouge>

On Thu, Jan 28, 2016 at 05:01:26PM +0100, Frederic Weisbecker wrote:
> So lets check all the things we call on scheduler_tick():
> 
> _ sched_clock_tick(): maybe it doesn't need to be called when idle. I'm not sure.
>   Some code in idle (timers, irqs, ...) might need to call sched_clock().

Only needed if you've got a shady TSC.

> _ update_rq_clock(), task_tick(): task_tick is empty for idle class, so we probably
>   don't need an updated rq either.

Right, for regular NOHZ we'll be running the idle task, and the idle
tick handler is empty. So for NOHZ you can ignore this.

For NOHZ_FULL you'll not be running the idle task and this gets
'interesting'.

The most important part would be tracking the task runtime, which is
used for a number of user visible things. This should be doable
remotely.

> _ update_cpu_load_active(): I was about to fix the issue properly and make it account
>   correctly on idle ticks but we might as well want to spare it.

Right, we've gone over this one in detail in other emails I think.

> _ calc_global_load_tick(): no idea

Can easily be done remote. However, this only records deltas of
nr_active (:= nr_running + nr_uninterruptible) and for NOHZ and
NOHZ_FULL this should not change, therefore the delta _should_ be 0 and
you can skip this.

> _ perf_event_task_tick(): needed, some freq CPU events can trigger in idle and need
>   adjustments

Right, this is a tricky one. Maybe I should look into moving this into a
hrtimer, but that too has 'fun' problems IIRC. I'll put it on the TODO
list somewhere.

> _ trigger_load_balance(): maybe needed, I see it triggers the softirq after some
>   rebalance delay, regardless of the current CPU idleness.

We already have NOHZ remote balancing, we could (and should) probably do
the same for NOHZ_FULL. Then again, I would expect the NOHZ_FULL cpus to
not actually be part of a balance domain, so we could probably detect
and short-circuit this.

> _ rq_last_tick_reset(): not needed in idle

Right, part of the NOHZ_FULL 'hack', once you fix all the remote
accounting stuff this could go away entirely think.

  reply	other threads:[~2016-01-29  9:50 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-01-13 16:01 [RFC PATCH 0/4] sched: Improve cpu load accounting with nohz Frederic Weisbecker
2016-01-13 16:01 ` [PATCH 1/4] sched: Don't account tickless CPU load on tick Frederic Weisbecker
2016-01-19 13:08   ` Peter Zijlstra
2016-01-19 16:22     ` Frederic Weisbecker
2016-01-19 18:56       ` Peter Zijlstra
2016-01-19 22:33         ` Frederic Weisbecker
2016-01-20  5:43           ` Byungchul Park
2016-01-20 10:26             ` Byungchul Park
2016-01-28 16:01               ` Frederic Weisbecker
2016-01-29  9:50                 ` Peter Zijlstra [this message]
2016-02-01 10:05                   ` Byungchul Park
2016-02-01 10:09                     ` [PATCH] sched: calculate sched_clock_cpu without tick handling during nohz Byungchul Park
2016-02-01 10:34                     ` [PATCH 1/4] sched: Don't account tickless CPU load on tick Peter Zijlstra
2016-02-01 23:51                       ` Byungchul Park
2016-02-02  0:50                       ` Byungchul Park
2016-02-01  6:33               ` Byungchul Park
2016-01-20  8:42     ` Thomas Gleixner
2016-01-20 17:36       ` Frederic Weisbecker
2016-01-22  8:40       ` Byungchul Park
2016-01-13 16:01 ` [PATCH 2/4] sched: Consolidate nohz CPU load update code Frederic Weisbecker
2016-01-14  2:30   ` Byungchul Park
2016-01-19 13:13     ` Peter Zijlstra
2016-01-20  0:51       ` Byungchul Park
2016-01-14  5:18   ` Byungchul Park
2016-01-19 13:13     ` Peter Zijlstra
2016-01-19 16:49     ` Frederic Weisbecker
2016-01-20  1:41       ` Byungchul Park
2016-02-29 11:14   ` [tip:sched/core] sched/fair: " tip-bot for Frederic Weisbecker
2016-01-13 16:01 ` [RFC PATCH 3/4] sched: Move cpu load stats functions above fair queue callbacks Frederic Weisbecker
2016-01-13 16:01 ` [RFC PATCH 4/4] sched: Upload nohz full CPU load on task enqueue/dequeue Frederic Weisbecker
2016-01-19 13:17   ` Peter Zijlstra
2016-01-19 17:03     ` Frederic Weisbecker
2016-01-20  9:09       ` Peter Zijlstra
2016-01-20 14:54         ` Frederic Weisbecker
2016-01-20 15:11           ` Thomas Gleixner
2016-01-20 15:19             ` Christoph Lameter
2016-01-20 16:52               ` Frederic Weisbecker
2016-01-20 16:45             ` Frederic Weisbecker
2016-01-20 16:56           ` Peter Zijlstra
2016-01-20 17:21             ` Frederic Weisbecker
2016-01-20 18:25               ` Peter Zijlstra
2016-01-21 13:25                 ` Frederic Weisbecker
2016-01-20  9:03   ` Thomas Gleixner
2016-01-20 14:31     ` Frederic Weisbecker
2016-01-20 14:43       ` Thomas Gleixner
2016-01-20 16:40         ` Frederic Weisbecker
2016-01-20 16:42           ` Christoph Lameter
2016-01-20 16:47             ` Frederic Weisbecker
2016-01-14 21:19 ` [RFC PATCH 0/4] sched: Improve cpu load accounting with nohz Dietmar Eggemann
2016-01-14 21:27   ` Peter Zijlstra
2016-01-14 22:23     ` Dietmar Eggemann
2016-01-15  7:07       ` Byungchul Park
2016-01-15 16:56         ` Dietmar Eggemann
2016-01-18  0:23           ` Byungchul Park
2016-01-19 13:04           ` Peter Zijlstra
2016-01-20  0:48             ` Byungchul Park
2016-01-20 13:04               ` Dietmar Eggemann
2016-02-29 11:14         ` [tip:sched/core] sched/fair: Avoid using decay_load_missed() with a negative value tip-bot for Byungchul Park

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=20160129095016.GZ6356@twins.programming.kicks-ass.net \
    --to=peterz@infradead.org \
    --cc=byungchul.park@lge.com \
    --cc=cl@linux.com \
    --cc=cmetcalf@ezchip.com \
    --cc=efault@gmx.de \
    --cc=fweisbec@gmail.com \
    --cc=lcapitulino@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=riel@redhat.com \
    --cc=tglx@linutronix.de \
    /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.