All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: Viresh Kumar <viresh.kumar@linaro.org>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Christoph Lameter <cl@linux.com>,
	Linaro Kernel Mailman List <linaro-kernel@lists.linaro.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	vinmenon@codeaurora.org, shashim@codeaurora.org,
	Michal Hocko <mhocko@suse.cz>, Mel Gorman <mgorman@suse.de>,
	dave@stgolabs.net, Konstantin Khlebnikov <koct9i@gmail.com>,
	Linux Memory Management List <linux-mm@kvack.org>,
	Suresh Siddha <suresh.b.siddha@intel.com>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: Re: [RFC] vmstat: Avoid waking up idle-cpu to service shepherd work
Date: Mon, 30 Mar 2015 18:25:19 +0200	[thread overview]
Message-ID: <20150330162519.GB23123@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <CAKohponEAivnev-fcWdjD0OcwQaXHN58tESCfqbZ_-W+_N+DvA@mail.gmail.com>

On Mon, Mar 30, 2015 at 09:47:01PM +0530, Viresh Kumar wrote:
> And all I get it is 8256 bytes, with or without the change.

Duh, rounded up to cacheline boundary ;-)

Trades two 4 byte holes at the start for a bigger 'hole' at the end.

struct tvec_base {
        spinlock_t                 lock;                 /*     0     2 */

        /* XXX 6 bytes hole, try to pack */

        struct timer_list *        running_timer;        /*     8     8 */
        long unsigned int          timer_jiffies;        /*    16     8 */
        long unsigned int          next_timer;           /*    24     8 */
        long unsigned int          active_timers;        /*    32     8 */
        long unsigned int          all_timers;           /*    40     8 */
        int                        cpu;                  /*    48     4 */

        /* XXX 4 bytes hole, try to pack */

        struct tvec_root           tv1;                  /*    56  4096 */
        /* --- cacheline 64 boundary (4096 bytes) was 56 bytes ago --- */
        struct tvec                tv2;                  /*  4152  1024 */
        /* --- cacheline 80 boundary (5120 bytes) was 56 bytes ago --- */
        struct tvec                tv3;                  /*  5176  1024 */
        /* --- cacheline 96 boundary (6144 bytes) was 56 bytes ago --- */
        struct tvec                tv4;                  /*  6200  1024 */
        /* --- cacheline 112 boundary (7168 bytes) was 56 bytes ago --- */
        struct tvec                tv5;                  /*  7224  1024 */
        /* --- cacheline 128 boundary (8192 bytes) was 56 bytes ago --- */

        /* size: 8256, cachelines: 129, members: 12 */
        /* sum members: 8238, holes: 2, sum holes: 10 */
        /* padding: 8 */
};

vs

struct tvec_base {
	spinlock_t                 lock;                 /*     0     2 */

	/* XXX 2 bytes hole, try to pack */

	int                        cpu;                  /*     4     4 */
	struct timer_list *        running_timer;        /*     8     8 */
	long unsigned int          timer_jiffies;        /*    16     8 */
	long unsigned int          next_timer;           /*    24     8 */
	long unsigned int          active_timers;        /*    32     8 */
	long unsigned int          all_timers;           /*    40     8 */
	struct tvec_root           tv1;                  /*    48  4096 */
	/* --- cacheline 64 boundary (4096 bytes) was 48 bytes ago --- */
	struct tvec                tv2;                  /*  4144  1024 */
	/* --- cacheline 80 boundary (5120 bytes) was 48 bytes ago --- */
	struct tvec                tv3;                  /*  5168  1024 */
	/* --- cacheline 96 boundary (6144 bytes) was 48 bytes ago --- */
	struct tvec                tv4;                  /*  6192  1024 */
	/* --- cacheline 112 boundary (7168 bytes) was 48 bytes ago --- */
	struct tvec                tv5;                  /*  7216  1024 */
	/* --- cacheline 128 boundary (8192 bytes) was 48 bytes ago --- */

	/* size: 8256, cachelines: 129, members: 12 */
	/* sum members: 8238, holes: 1, sum holes: 2 */
	/* padding: 16 */
};

WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <peterz@infradead.org>
To: Viresh Kumar <viresh.kumar@linaro.org>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Christoph Lameter <cl@linux.com>,
	Linaro Kernel Mailman List <linaro-kernel@lists.linaro.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	vinmenon@codeaurora.org, shashim@codeaurora.org,
	Michal Hocko <mhocko@suse.cz>, Mel Gorman <mgorman@suse.de>,
	dave@stgolabs.net, Konstantin Khlebnikov <koct9i@gmail.com>,
	Linux Memory Management List <linux-mm@kvack.org>,
	Suresh Siddha <suresh.b.siddha@intel.com>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: Re: [RFC] vmstat: Avoid waking up idle-cpu to service shepherd work
Date: Mon, 30 Mar 2015 18:25:19 +0200	[thread overview]
Message-ID: <20150330162519.GB23123@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <CAKohponEAivnev-fcWdjD0OcwQaXHN58tESCfqbZ_-W+_N+DvA@mail.gmail.com>

On Mon, Mar 30, 2015 at 09:47:01PM +0530, Viresh Kumar wrote:
> And all I get it is 8256 bytes, with or without the change.

Duh, rounded up to cacheline boundary ;-)

Trades two 4 byte holes at the start for a bigger 'hole' at the end.

struct tvec_base {
        spinlock_t                 lock;                 /*     0     2 */

        /* XXX 6 bytes hole, try to pack */

        struct timer_list *        running_timer;        /*     8     8 */
        long unsigned int          timer_jiffies;        /*    16     8 */
        long unsigned int          next_timer;           /*    24     8 */
        long unsigned int          active_timers;        /*    32     8 */
        long unsigned int          all_timers;           /*    40     8 */
        int                        cpu;                  /*    48     4 */

        /* XXX 4 bytes hole, try to pack */

        struct tvec_root           tv1;                  /*    56  4096 */
        /* --- cacheline 64 boundary (4096 bytes) was 56 bytes ago --- */
        struct tvec                tv2;                  /*  4152  1024 */
        /* --- cacheline 80 boundary (5120 bytes) was 56 bytes ago --- */
        struct tvec                tv3;                  /*  5176  1024 */
        /* --- cacheline 96 boundary (6144 bytes) was 56 bytes ago --- */
        struct tvec                tv4;                  /*  6200  1024 */
        /* --- cacheline 112 boundary (7168 bytes) was 56 bytes ago --- */
        struct tvec                tv5;                  /*  7224  1024 */
        /* --- cacheline 128 boundary (8192 bytes) was 56 bytes ago --- */

        /* size: 8256, cachelines: 129, members: 12 */
        /* sum members: 8238, holes: 2, sum holes: 10 */
        /* padding: 8 */
};

vs

struct tvec_base {
	spinlock_t                 lock;                 /*     0     2 */

	/* XXX 2 bytes hole, try to pack */

	int                        cpu;                  /*     4     4 */
	struct timer_list *        running_timer;        /*     8     8 */
	long unsigned int          timer_jiffies;        /*    16     8 */
	long unsigned int          next_timer;           /*    24     8 */
	long unsigned int          active_timers;        /*    32     8 */
	long unsigned int          all_timers;           /*    40     8 */
	struct tvec_root           tv1;                  /*    48  4096 */
	/* --- cacheline 64 boundary (4096 bytes) was 48 bytes ago --- */
	struct tvec                tv2;                  /*  4144  1024 */
	/* --- cacheline 80 boundary (5120 bytes) was 48 bytes ago --- */
	struct tvec                tv3;                  /*  5168  1024 */
	/* --- cacheline 96 boundary (6144 bytes) was 48 bytes ago --- */
	struct tvec                tv4;                  /*  6192  1024 */
	/* --- cacheline 112 boundary (7168 bytes) was 48 bytes ago --- */
	struct tvec                tv5;                  /*  7216  1024 */
	/* --- cacheline 128 boundary (8192 bytes) was 48 bytes ago --- */

	/* size: 8256, cachelines: 129, members: 12 */
	/* sum members: 8238, holes: 1, sum holes: 2 */
	/* padding: 16 */
};

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2015-03-30 16:25 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-03-26  5:39 [RFC] vmstat: Avoid waking up idle-cpu to service shepherd work Viresh Kumar
2015-03-26  5:39 ` Viresh Kumar
2015-03-26 20:18 ` Andrew Morton
2015-03-26 20:18   ` Andrew Morton
2015-03-27  4:49   ` Viresh Kumar
2015-03-27  4:49     ` Viresh Kumar
2015-03-27  9:16     ` Peter Zijlstra
2015-03-27  9:16       ` Peter Zijlstra
2015-03-27  9:30       ` Peter Zijlstra
2015-03-27  9:30         ` Peter Zijlstra
2015-03-27 11:11         ` Christoph Lameter
2015-03-27 11:11           ` Christoph Lameter
2015-03-27 12:02           ` Peter Zijlstra
2015-03-27 12:02             ` Peter Zijlstra
2015-03-27 19:45             ` Christoph Lameter
2015-03-27 19:45               ` Christoph Lameter
2015-03-28  4:28             ` Viresh Kumar
2015-03-28  4:28               ` Viresh Kumar
2015-03-28 11:41               ` Peter Zijlstra
2015-03-28 11:41                 ` Peter Zijlstra
2015-03-28  4:18         ` Viresh Kumar
2015-03-28  4:18           ` Viresh Kumar
2015-03-28  9:53           ` Peter Zijlstra
2015-03-28  9:53             ` Peter Zijlstra
2015-03-28 11:57             ` viresh kumar
2015-03-28 11:57               ` viresh kumar
2015-03-28 12:04               ` Viresh Kumar
2015-03-28 12:04                 ` Viresh Kumar
2015-03-28 13:44               ` Peter Zijlstra
2015-03-28 13:44                 ` Peter Zijlstra
2015-03-29 10:24                 ` Peter Zijlstra
2015-03-29 10:24                   ` Peter Zijlstra
2015-03-30 12:02                   ` Viresh Kumar
2015-03-30 12:02                     ` Viresh Kumar
2015-03-30 12:47                     ` Peter Zijlstra
2015-03-30 12:47                       ` Peter Zijlstra
2015-03-30 13:14                       ` Viresh Kumar
2015-03-30 13:14                         ` Viresh Kumar
2015-03-30 13:59                         ` Peter Zijlstra
2015-03-30 13:59                           ` Peter Zijlstra
2015-03-30 16:17                           ` Viresh Kumar
2015-03-30 16:17                             ` Viresh Kumar
2015-03-30 16:25                             ` Peter Zijlstra [this message]
2015-03-30 16:25                               ` Peter Zijlstra
2015-03-29 12:01                 ` Viresh Kumar
2015-03-29 12:01                   ` Viresh Kumar
2015-03-29 17:24                   ` Peter Zijlstra
2015-03-29 17:24                     ` Peter Zijlstra
2015-03-30 15:08             ` Michal Hocko
2015-03-30 15:08               ` Michal Hocko
2015-03-30 15:14               ` Peter Zijlstra
2015-03-30 15:14                 ` Peter Zijlstra
2015-03-30 15:42               ` Christoph Lameter
2015-03-30 15:42                 ` Christoph Lameter
2015-03-27 14:19 ` Michal Hocko
2015-03-27 14:19   ` Michal Hocko
2015-03-28  4:34   ` Viresh Kumar
2015-03-28  4:34     ` Viresh Kumar

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=20150330162519.GB23123@twins.programming.kicks-ass.net \
    --to=peterz@infradead.org \
    --cc=akpm@linux-foundation.org \
    --cc=cl@linux.com \
    --cc=dave@stgolabs.net \
    --cc=hannes@cmpxchg.org \
    --cc=koct9i@gmail.com \
    --cc=linaro-kernel@lists.linaro.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mgorman@suse.de \
    --cc=mhocko@suse.cz \
    --cc=shashim@codeaurora.org \
    --cc=suresh.b.siddha@intel.com \
    --cc=tglx@linutronix.de \
    --cc=vinmenon@codeaurora.org \
    --cc=viresh.kumar@linaro.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.