linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "David Nyström" <david.c.nystrom@gmail.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Christoph Lameter <cl@linux.com>,
	linux-kernel@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Peter Zijlstra <peterz@infradead.org>,
	Clark Williams <clark.williams@gmail.com>,
	Frederic Weisbecker <fweisbec@gmail.com>,
	Li Zefan <lizf@cn.fujitsu.com>, Ingo Molnar <mingo@kernel.org>,
	"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
	Mike Galbraith <efault@gmx.de>
Subject: Re: [PATCH 00/32] [RFC] nohz/cpuset: Start discussions on nohz CPUs
Date: Fri, 02 Nov 2012 15:50:37 +0100	[thread overview]
Message-ID: <5093DDBD.4000603@gmail.com> (raw)
In-Reply-To: <1351867045.4004.141.camel@gandalf.local.home>

On 11/02/2012 03:37 PM, Steven Rostedt wrote:
> On Fri, 2012-11-02 at 14:23 +0000, Christoph Lameter wrote:
>> On Mon, 29 Oct 2012, Steven Rostedt wrote:
>>
>>> A while ago Frederic posted a series of patches to get an idea on
>>> how to implement nohz cpusets. Where you can add a task to a cpuset
>>> and mark the set to be 'nohz'. When the task runs on a CPU and is
>>> the only task scheduled (nr_running == 1), the tick will stop.
>>> The idea is to give the task the least amount of kernel interference
>>> as possible. If the task doesn't do any system calls (and possibly
>>> even if it does), no timer interrupt will bother it. By using
>>> isocpus and nohz cpuset, a task would be able to achieve true cpu
>>> isolation.
>>

One other aspect that this patch probably needs to address is the cache 
localization of irq spinlocks.

At least in 3.6, with !CONFIG_SPARSE_IRQ
--
struct  irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = {
	[0 ... NR_IRQS-1] = {
		.handle_irq	= handle_bad_irq,
		.depth		= 1,
		.lock		= __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock),
	}
};
--

You are likely to get a cache miss in the top half of your low latency 
CPU anytime some other CPU has taken a spinlock which lies within the 
same cache line.

Or is my understanding of the __cacheline_aligned_in_smp declaration wrong ?

Br,
David


  reply	other threads:[~2012-11-02 14:50 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-10-29 20:27 [PATCH 00/32] [RFC] nohz/cpuset: Start discussions on nohz CPUs Steven Rostedt
2012-10-29 20:27 ` [PATCH 01/32] nohz: Move nohz load balancer selection into idle logic Steven Rostedt
2012-10-30  8:32   ` Charles Wang
2012-10-30 15:39     ` Steven Rostedt
2012-10-29 20:27 ` [PATCH 02/32] cpuset: Set up interface for nohz flag Steven Rostedt
2012-10-30 17:16   ` Steven Rostedt
2012-10-29 20:27 ` [PATCH 03/32] nohz: Try not to give the timekeeping duty to an adaptive tickless cpu Steven Rostedt
2012-10-30 17:33   ` Steven Rostedt
2012-10-29 20:27 ` [PATCH 04/32] x86: New cpuset nohz irq vector Steven Rostedt
2012-10-30 17:39   ` Steven Rostedt
2012-10-30 23:51     ` Frederic Weisbecker
2012-10-31  0:07       ` Steven Rostedt
2012-10-31  0:45         ` Frederic Weisbecker
2012-10-29 20:27 ` [PATCH 05/32] nohz: Adaptive tick stop and restart on nohz cpuset Steven Rostedt
2012-10-30 18:23   ` Steven Rostedt
2012-10-29 20:27 ` [PATCH 06/32] nohz/cpuset: Dont turn off the tick if rcu needs it Steven Rostedt
2012-10-30 18:30   ` Steven Rostedt
2012-10-29 20:27 ` [PATCH 07/32] nohz/cpuset: Wake up adaptive nohz CPU when a timer gets enqueued Steven Rostedt
2012-10-29 20:27 ` [PATCH 08/32] nohz/cpuset: Dont stop the tick if posix cpu timers are running Steven Rostedt
2012-10-29 20:27 ` [PATCH 09/32] nohz/cpuset: Restart tick when nohz flag is cleared on cpuset Steven Rostedt
2012-10-30 18:55   ` Steven Rostedt
2012-10-29 20:27 ` [PATCH 10/32] nohz/cpuset: Restart the tick if printk needs it Steven Rostedt
2012-10-30 19:01   ` Steven Rostedt
2012-10-30 23:54     ` Frederic Weisbecker
2012-10-29 20:27 ` [PATCH 11/32] rcu: Restart the tick on non-responding adaptive nohz CPUs Steven Rostedt
2012-10-29 20:27 ` [PATCH 12/32] rcu: Restart tick if we enqueue a callback in a nohz/cpuset CPU Steven Rostedt
2012-10-29 20:27 ` [PATCH 13/32] nohz: Generalize tickless cpu time accounting Steven Rostedt
2012-10-29 20:27 ` [PATCH 14/32] nohz/cpuset: Account user and system times in adaptive nohz mode Steven Rostedt
2012-10-29 20:27 ` [PATCH 15/32] nohz/cpuset: New API to flush cputimes on nohz cpusets Steven Rostedt
2012-10-29 20:27 ` [PATCH 16/32] nohz/cpuset: Flush cputime on threads in nohz cpusets when waiting leader Steven Rostedt
2012-10-29 20:27 ` [PATCH 17/32] nohz/cpuset: Flush cputimes on procfs stat file read Steven Rostedt
2012-10-29 20:27 ` [PATCH 18/32] nohz/cpuset: Flush cputimes for getrusage() and times() syscalls Steven Rostedt
2012-10-29 20:27 ` [PATCH 19/32] x86: Syscall hooks for nohz cpusets Steven Rostedt
2012-10-29 20:27 ` [PATCH 20/32] nohz/cpuset: enable addition&removal of cpus while in adaptive nohz mode Steven Rostedt
2012-10-29 20:27 ` [PATCH 21/32] nohz: Dont restart the tick before scheduling to idle Steven Rostedt
2012-10-29 20:27 ` [PATCH 22/32] sched: Comment on rq->clock correctness in ttwu_do_wakeup() in nohz Steven Rostedt
2012-10-29 20:27 ` [PATCH 23/32] sched: Update rq clock on nohz CPU before migrating tasks Steven Rostedt
2012-10-29 20:27 ` [PATCH 24/32] sched: Update rq clock on nohz CPU before setting fair group shares Steven Rostedt
2012-10-29 20:27 ` [PATCH 25/32] sched: Update rq clock on tickless CPUs before calling check_preempt_curr() Steven Rostedt
2012-10-29 20:27 ` [PATCH 26/32] sched: Update rq clock earlier in unthrottle_cfs_rq Steven Rostedt
2012-10-29 20:27 ` [PATCH 27/32] sched: Update clock of nohz busiest rq before balancing Steven Rostedt
2012-10-29 20:27 ` [PATCH 28/32] sched: Update rq clock before idle balancing Steven Rostedt
2012-10-29 20:27 ` [PATCH 29/32] sched: Update nohz rq clock before searching busiest group on load balancing Steven Rostedt
2012-10-29 20:27 ` [PATCH 30/32] rcu: Switch to extended quiescent state in userspace from nohz cpuset Steven Rostedt
2012-10-29 20:27 ` [PATCH 31/32] nohz/cpuset: Disable under some configs Steven Rostedt
2012-10-29 20:27 ` [PATCH 32/32] nohz, not for merge: Add tickless tracing Steven Rostedt
2012-10-30 14:02 ` [PATCH 00/32] [RFC] nohz/cpuset: Start discussions on nohz CPUs Gilad Ben-Yossef
2012-11-02 14:23 ` Christoph Lameter
2012-11-02 14:37   ` Steven Rostedt
2012-11-02 14:50     ` David Nyström [this message]
2012-11-02 15:03     ` Christoph Lameter
2012-11-02 15:14       ` Steven Rostedt
2012-11-02 18:35       ` Paul E. McKenney
2012-11-02 20:16         ` Christoph Lameter
2012-11-02 20:41           ` Paul E. McKenney
2012-11-02 20:51             ` Steven Rostedt
2012-11-03  2:08               ` Paul E. McKenney
2012-11-05 15:17                 ` Christoph Lameter
2012-11-05 22:41                   ` Frederic Weisbecker
2012-11-05 22:32   ` Frederic Weisbecker

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=5093DDBD.4000603@gmail.com \
    --to=david.c.nystrom@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=cl@linux.com \
    --cc=clark.williams@gmail.com \
    --cc=efault@gmx.de \
    --cc=fweisbec@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lizf@cn.fujitsu.com \
    --cc=mingo@kernel.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    --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 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).