From: peterz@infradead.org (Peter Zijlstra)
To: linux-arm-kernel@lists.infradead.org
Subject: [RFC PATCH v3 3/6] sched: pack small tasks
Date: Wed, 27 Mar 2013 09:46:20 +0100 [thread overview]
Message-ID: <1364373980.5053.57.camel@laptop> (raw)
In-Reply-To: <5151BEF2.9090100@linux.intel.com>
On Tue, 2013-03-26 at 08:29 -0700, Arjan van de Ven wrote:
> > Isn't this basically related to picking the NO_HZ cpu; if the system
> > isn't fully symmetric with its power gates you want the NO_HZ cpu to be
> > the 'special' cpu. If it is symmetric we really don't care which core
> > is left 'running' and we can even select a new pack cpu from the idle
> > cores once the old one is fully utilized.
>
> you don't really care much sure, but there's some advantages for sorting "all the way left",
> e.g. to linux cpu 0.
> Some tasks only run there, and interrupts tend to be favored to that cpu as well on x86.
Right, and I suspect all the big-little nonsense will have the little
cores on low numbers as well (is this architected or can a creative
licensee screw us over?)
So find_new_ilb() already does cpumask_first(), so it has a strong
leftmost preference. We just need to make sure it indeed does the right
thing and doesn't have some unintended side effect.
WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <peterz@infradead.org>
To: Arjan van de Ven <arjan@linux.intel.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>,
linux-kernel@vger.kernel.org,
linux-arm-kernel@lists.infradead.org,
linaro-kernel@lists.linaro.org, mingo@kernel.org,
linux@arm.linux.org.uk, pjt@google.com, santosh.shilimkar@ti.com,
morten.rasmussen@arm.com, chander.kashyap@linaro.org,
cmetcalf@tilera.com, tony.luck@intel.com, alex.shi@intel.com,
preeti@linux.vnet.ibm.com, paulmck@linux.vnet.ibm.com,
tglx@linutronix.de, len.brown@intel.com,
amit.kucheria@linaro.org, corbet@lwn.net
Subject: Re: [RFC PATCH v3 3/6] sched: pack small tasks
Date: Wed, 27 Mar 2013 09:46:20 +0100 [thread overview]
Message-ID: <1364373980.5053.57.camel@laptop> (raw)
In-Reply-To: <5151BEF2.9090100@linux.intel.com>
On Tue, 2013-03-26 at 08:29 -0700, Arjan van de Ven wrote:
> > Isn't this basically related to picking the NO_HZ cpu; if the system
> > isn't fully symmetric with its power gates you want the NO_HZ cpu to be
> > the 'special' cpu. If it is symmetric we really don't care which core
> > is left 'running' and we can even select a new pack cpu from the idle
> > cores once the old one is fully utilized.
>
> you don't really care much sure, but there's some advantages for sorting "all the way left",
> e.g. to linux cpu 0.
> Some tasks only run there, and interrupts tend to be favored to that cpu as well on x86.
Right, and I suspect all the big-little nonsense will have the little
cores on low numbers as well (is this architected or can a creative
licensee screw us over?)
So find_new_ilb() already does cpumask_first(), so it has a strong
leftmost preference. We just need to make sure it indeed does the right
thing and doesn't have some unintended side effect.
next prev parent reply other threads:[~2013-03-27 8:46 UTC|newest]
Thread overview: 103+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-03-22 12:25 [RFC PATCH v3 0/6] sched: packing small tasks Vincent Guittot
2013-03-22 12:25 ` Vincent Guittot
2013-03-22 12:25 ` [RFC PATCH v3 1/6] Revert "sched: Introduce temporary FAIR_GROUP_SCHED dependency for load-tracking" Vincent Guittot
2013-03-22 12:25 ` Vincent Guittot
2013-03-22 12:25 ` [RFC PATCH v3 2/6] sched: add a new SD_SHARE_POWERDOMAIN flag for sched_domain Vincent Guittot
2013-03-22 12:25 ` Vincent Guittot
2013-03-22 12:25 ` [RFC PATCH v3 3/6] sched: pack small tasks Vincent Guittot
2013-03-22 12:25 ` Vincent Guittot
2013-03-26 12:26 ` Peter Zijlstra
2013-03-26 12:26 ` Peter Zijlstra
2013-03-27 10:21 ` Preeti U Murthy
2013-03-27 10:21 ` Preeti U Murthy
2013-03-27 11:00 ` Vincent Guittot
2013-03-27 11:00 ` Vincent Guittot
2013-04-26 10:30 ` Peter Zijlstra
2013-04-26 10:30 ` Peter Zijlstra
2013-04-26 11:34 ` Vincent Guittot
2013-04-26 11:34 ` Vincent Guittot
2013-04-26 10:18 ` Peter Zijlstra
2013-04-26 10:18 ` Peter Zijlstra
2013-04-26 10:32 ` Preeti U Murthy
2013-04-26 10:32 ` Preeti U Murthy
2013-03-26 12:37 ` Peter Zijlstra
2013-03-26 12:37 ` Peter Zijlstra
2013-03-26 13:00 ` Vincent Guittot
2013-03-26 13:00 ` Vincent Guittot
2013-03-27 4:33 ` Preeti U Murthy
2013-03-27 4:33 ` Preeti U Murthy
2013-03-27 4:48 ` Alex Shi
2013-03-27 4:48 ` Alex Shi
2013-03-27 8:51 ` Peter Zijlstra
2013-03-27 8:51 ` Peter Zijlstra
2013-03-26 12:46 ` Peter Zijlstra
2013-03-26 12:46 ` Peter Zijlstra
2013-03-26 13:53 ` Vincent Guittot
2013-03-26 13:53 ` Vincent Guittot
2013-03-26 15:29 ` Arjan van de Ven
2013-03-26 15:29 ` Arjan van de Ven
2013-03-27 8:46 ` Peter Zijlstra [this message]
2013-03-27 8:46 ` Peter Zijlstra
2013-03-27 8:54 ` Vincent Guittot
2013-03-27 8:54 ` Vincent Guittot
2013-03-27 9:00 ` Peter Zijlstra
2013-03-27 9:00 ` Peter Zijlstra
2013-03-27 11:18 ` Catalin Marinas
2013-03-27 14:13 ` Peter Zijlstra
2013-03-27 16:36 ` Catalin Marinas
2013-03-27 17:18 ` Nicolas Pitre
2013-03-27 17:18 ` Nicolas Pitre
2013-03-27 17:37 ` Catalin Marinas
2013-03-27 17:37 ` Catalin Marinas
2013-03-27 17:20 ` Vincent Guittot
2013-03-27 18:01 ` Catalin Marinas
2013-03-27 15:37 ` Nicolas Pitre
2013-03-27 15:37 ` Nicolas Pitre
2013-03-22 12:25 ` [RFC PATCH v3 4/6] sched: secure access to other CPU statistics Vincent Guittot
2013-03-22 12:25 ` Vincent Guittot
2013-03-26 12:50 ` Peter Zijlstra
2013-03-26 12:50 ` Peter Zijlstra
2013-03-26 13:06 ` Vincent Guittot
2013-03-26 13:06 ` Vincent Guittot
2013-03-22 12:25 ` [RFC PATCH v3 5/6] sched: pack the idle load balance Vincent Guittot
2013-03-22 12:25 ` Vincent Guittot
2013-03-26 12:52 ` Peter Zijlstra
2013-03-26 12:52 ` Peter Zijlstra
2013-03-26 14:03 ` Vincent Guittot
2013-03-26 14:03 ` Vincent Guittot
2013-03-26 14:42 ` Peter Zijlstra
2013-03-26 14:42 ` Peter Zijlstra
2013-03-26 15:55 ` Vincent Guittot
2013-03-26 15:55 ` Vincent Guittot
2013-03-27 4:56 ` Alex Shi
2013-03-27 4:56 ` Alex Shi
2013-03-27 8:05 ` Vincent Guittot
2013-03-27 8:05 ` Vincent Guittot
2013-03-27 8:47 ` Alex Shi
2013-03-27 8:47 ` Alex Shi
2013-03-27 10:30 ` Vincent Guittot
2013-03-27 10:30 ` Vincent Guittot
2013-03-27 13:32 ` Alex Shi
2013-03-27 13:32 ` Alex Shi
2013-03-27 8:49 ` Peter Zijlstra
2013-03-27 8:49 ` Peter Zijlstra
2013-04-05 11:08 ` Vincent Guittot
2013-04-05 11:08 ` Vincent Guittot
2013-04-22 5:45 ` Preeti U Murthy
2013-04-22 5:45 ` Preeti U Murthy
[not found] ` <CAKfTPtCCCifC=c+xjjnAH_HSqkR80PiQoddQKXPHuZwZawbvcA@mail.gmail.com>
2013-04-23 2:23 ` Alex Shi
2013-04-23 2:23 ` Alex Shi
2013-04-23 4:57 ` Preeti U Murthy
2013-04-23 4:57 ` Preeti U Murthy
2013-04-23 15:30 ` Arjan van de Ven
2013-04-23 15:30 ` Arjan van de Ven
2013-04-26 10:54 ` Peter Zijlstra
2013-04-26 10:54 ` Peter Zijlstra
2013-04-23 4:36 ` Preeti U Murthy
2013-04-23 4:36 ` Preeti U Murthy
2013-03-22 12:25 ` [RFC PATCH v3 6/6] ARM: sched: clear SD_SHARE_POWERDOMAIN Vincent Guittot
2013-03-22 12:25 ` Vincent Guittot
2013-03-23 11:55 ` [RFC PATCH v3 0/6] sched: packing small tasks Preeti U Murthy
2013-03-23 11:55 ` Preeti U Murthy
2013-03-25 9:58 ` Vincent Guittot
2013-03-25 9:58 ` Vincent Guittot
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=1364373980.5053.57.camel@laptop \
--to=peterz@infradead.org \
--cc=linux-arm-kernel@lists.infradead.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.