mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Chris Metcalf <cmetcalf@ezchip.com>
To: Gilad Ben Yossef <giladb@ezchip.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	Ingo Molnar <mingo@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	"Rik van Riel" <riel@redhat.com>, Tejun Heo <tj@kernel.org>,
	Frederic Weisbecker <fweisbec@gmail.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
	Christoph Lameter <cl@linux.com>,
	Viresh Kumar <viresh.kumar@linaro.org>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will.deacon@arm.com>,
	Andy Lutomirski <luto@amacapital.net>,
	<linux-kernel@vger.kernel.org>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Subject: [PATCH v7 06/11] nohz: task_isolation: allow tick to be fully disabled
Date: Mon, 28 Sep 2015 11:17:21 -0400	[thread overview]
Message-ID: <1443453446-7827-7-git-send-email-cmetcalf@ezchip.com> (raw)
In-Reply-To: <1443453446-7827-1-git-send-email-cmetcalf@ezchip.com>

While the current fallback to 1-second tick is still helpful for
maintaining completely correct kernel semantics, processes using
prctl(PR_SET_TASK_ISOLATION) semantics place a higher priority on
running completely tickless, so don't bound the time_delta for such
processes.  In addition, due to the way such processes quiesce by
waiting for the timer tick to stop prior to returning to userspace,
without this commit it won't be possible to use the task_isolation
mode at all.

Removing the 1-second cap was previously discussed (see link
below) and Thomas Gleixner observed that vruntime, load balancing
data, load accounting, and other things might be impacted.
Frederic Weisbecker similarly observed that allowing the tick to
be indefinitely deferred just meant that no one would ever fix the
underlying bugs.  However it's at least true that the mode proposed
in this patch can only be enabled on a nohz_full core by a process
requesting task_isolation mode, which may limit how important it is
to maintain scheduler data correctly, for example.

Paul McKenney observed that if provide a mode where the 1Hz fallback
timer is removed, this will provide an environment where new code
that relies on that tick will get punished, and we won't forgive
such assumptions silently, so it may also be worth it from that
perspective.

Finally, it's worth observing that the tile architecture has been
using similar code for its Zero-Overhead Linux for many years
(starting in 2008) and customers are very enthusiastic about the
resulting bare-metal performance on cores that are available to
run full Linux semantics on demand (crash, logging, shutdown, etc).
So this semantics is very useful if we can convince ourselves that
doing this is safe.

Link: https://lkml.kernel.org/r/alpine.DEB.2.11.1410311058500.32582@gentwo.org
Signed-off-by: Chris Metcalf <cmetcalf@ezchip.com>
---
 kernel/time/tick-sched.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 3319e16f31e5..4504c0b95d0d 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -24,6 +24,7 @@
 #include <linux/posix-timers.h>
 #include <linux/perf_event.h>
 #include <linux/context_tracking.h>
+#include <linux/isolation.h>
 
 #include <asm/irq_regs.h>
 
@@ -634,7 +635,7 @@ static ktime_t tick_nohz_stop_sched_tick(struct tick_sched *ts,
 
 #ifdef CONFIG_NO_HZ_FULL
 	/* Limit the tick delta to the maximum scheduler deferment */
-	if (!ts->inidle)
+	if (!ts->inidle && !task_isolation_enabled())
 		delta = min(delta, scheduler_tick_max_deferment());
 #endif
 
-- 
2.1.2


  parent reply	other threads:[~2015-09-28 15:20 UTC|newest]

Thread overview: 236+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-08 17:58 [PATCH 0/6] support "dataplane" mode for nohz_full Chris Metcalf
2015-05-08 17:58 ` [PATCH 1/6] nohz_full: add support for "dataplane" mode Chris Metcalf
2015-05-08 17:58 ` [PATCH 2/6] nohz: dataplane: allow tick to be fully disabled for dataplane Chris Metcalf
2015-05-12  9:26   ` Peter Zijlstra
2015-05-12 13:12     ` Paul E. McKenney
2015-05-14 20:55       ` Chris Metcalf
2015-05-08 17:58 ` [PATCH 3/6] dataplane nohz: run softirqs synchronously on user entry Chris Metcalf
2015-05-09  7:04   ` Mike Galbraith
2015-05-11 20:13     ` Chris Metcalf
2015-05-12  2:21       ` Mike Galbraith
2015-05-12  9:28       ` Peter Zijlstra
2015-05-12  9:32       ` Peter Zijlstra
2015-05-12 13:08         ` Paul E. McKenney
2015-05-08 17:58 ` [PATCH 4/6] nohz: support PR_DATAPLANE_QUIESCE Chris Metcalf
2015-05-12  9:33   ` Peter Zijlstra
2015-05-12  9:50     ` Ingo Molnar
2015-05-12 10:38       ` Peter Zijlstra
2015-05-12 12:52         ` Ingo Molnar
2015-05-13  4:35           ` Andy Lutomirski
2015-05-13 17:51             ` Paul E. McKenney
2015-05-14 20:55               ` Chris Metcalf
2015-05-14 20:54     ` Chris Metcalf
2015-05-08 17:58 ` [PATCH 5/6] nohz: support PR_DATAPLANE_STRICT mode Chris Metcalf
2015-05-09  7:28   ` Andy Lutomirski
2015-05-09 10:37     ` Gilad Ben Yossef
2015-05-11 19:13     ` Chris Metcalf
2015-05-11 22:28       ` Andy Lutomirski
2015-05-12 21:06         ` Chris Metcalf
2015-05-12 22:23           ` Andy Lutomirski
2015-05-15 21:25             ` Chris Metcalf
2015-05-12  9:38   ` Peter Zijlstra
2015-05-12 13:20     ` Paul E. McKenney
2015-05-08 17:58 ` [PATCH 6/6] nohz: add dataplane_debug boot flag Chris Metcalf
2015-05-08 21:18 ` [PATCH 0/6] support "dataplane" mode for nohz_full Andrew Morton
2015-05-08 21:22   ` Steven Rostedt
2015-05-08 23:11     ` Chris Metcalf
2015-05-08 23:19       ` Andrew Morton
2015-05-09  7:05         ` Ingo Molnar
2015-05-09  7:19           ` Andy Lutomirski
2015-05-11 19:54             ` Chris Metcalf
2015-05-11 22:15               ` Andy Lutomirski
     [not found]             ` <55510885.9070101@ezchip.com>
2015-05-12 13:18               ` Paul E. McKenney
2015-05-09  7:19           ` Mike Galbraith
2015-05-09 10:18             ` Gilad Ben Yossef
2015-05-11 12:57           ` Steven Rostedt
2015-05-11 15:36             ` Frederic Weisbecker
2015-05-11 19:19               ` Mike Galbraith
2015-05-11 19:25                 ` Chris Metcalf
2015-05-12  1:47                   ` Mike Galbraith
2015-05-12  4:35                     ` Mike Galbraith
2015-05-15 15:05                     ` Chris Metcalf
2015-05-15 18:44                       ` Mike Galbraith
2015-05-26 19:51                         ` Chris Metcalf
2015-05-27  3:28                           ` Mike Galbraith
2015-05-11 17:19             ` Paul E. McKenney
2015-05-11 17:27               ` Andrew Morton
2015-05-11 17:33                 ` Frederic Weisbecker
2015-05-11 18:00                   ` Steven Rostedt
2015-05-11 18:09                     ` Chris Metcalf
2015-05-11 18:36                       ` Steven Rostedt
2015-05-12  9:10                       ` CONFIG_ISOLATION=y (was: [PATCH 0/6] support "dataplane" mode for nohz_full) Ingo Molnar
2015-05-12 11:48                         ` Peter Zijlstra
2015-05-12 12:34                           ` Ingo Molnar
2015-05-12 12:39                             ` Peter Zijlstra
2015-05-12 12:43                               ` Ingo Molnar
2015-05-12 15:36                             ` Frederic Weisbecker
2015-05-12 21:05                         ` CONFIG_ISOLATION=y Chris Metcalf
2015-05-12 10:46             ` [PATCH 0/6] support "dataplane" mode for nohz_full Peter Zijlstra
2015-05-15 15:10               ` Chris Metcalf
2015-05-15 21:26 ` [PATCH v2 0/5] support "cpu_isolated" " Chris Metcalf
2015-05-15 21:27   ` [PATCH v2 1/5] nohz_full: add support for "cpu_isolated" mode Chris Metcalf
2015-05-15 21:27     ` [PATCH v2 2/5] nohz: support PR_CPU_ISOLATED_STRICT mode Chris Metcalf
2015-05-15 21:27     ` [PATCH v2 3/5] nohz: cpu_isolated strict mode configurable signal Chris Metcalf
2015-05-15 21:27     ` [PATCH v2 4/5] nohz: add cpu_isolated_debug boot flag Chris Metcalf
2015-05-15 21:27     ` [PATCH v2 5/5] nohz: cpu_isolated: allow tick to be fully disabled Chris Metcalf
2015-05-15 22:17     ` [PATCH v2 1/5] nohz_full: add support for "cpu_isolated" mode Thomas Gleixner
2015-05-28 20:38       ` Chris Metcalf
2015-06-03 15:29   ` [PATCH v3 0/5] support "cpu_isolated" mode for nohz_full Chris Metcalf
2015-06-03 15:29     ` [PATCH v3 1/5] nohz_full: add support for "cpu_isolated" mode Chris Metcalf
2015-06-03 15:29     ` [PATCH v3 2/5] nohz: support PR_CPU_ISOLATED_STRICT mode Chris Metcalf
2015-06-03 15:29     ` [PATCH v3 3/5] nohz: cpu_isolated strict mode configurable signal Chris Metcalf
2015-06-03 15:29     ` [PATCH v3 4/5] nohz: add cpu_isolated_debug boot flag Chris Metcalf
2015-06-03 15:29     ` [PATCH v3 5/5] nohz: cpu_isolated: allow tick to be fully disabled Chris Metcalf
2015-07-13 19:57     ` [PATCH v4 0/5] support "cpu_isolated" mode for nohz_full Chris Metcalf
2015-07-13 19:57       ` [PATCH v4 1/5] nohz_full: add support for "cpu_isolated" mode Chris Metcalf
2015-07-13 20:40         ` Andy Lutomirski
2015-07-13 21:01           ` Chris Metcalf
2015-07-13 21:45             ` Andy Lutomirski
2015-07-21 19:10               ` Chris Metcalf
2015-07-21 19:26                 ` Andy Lutomirski
2015-07-21 20:36                   ` Paul E. McKenney
2015-07-22 13:57                     ` Christoph Lameter
2015-07-22 19:28                       ` Paul E. McKenney
2015-07-22 20:02                         ` Christoph Lameter
2015-07-24 20:21                           ` Chris Metcalf
2015-07-24 20:22                   ` Chris Metcalf
2015-07-24 14:03                 ` Frederic Weisbecker
2015-07-24 20:19                   ` Chris Metcalf
2015-07-24 13:27         ` Frederic Weisbecker
2015-07-24 20:21           ` Chris Metcalf
2015-07-13 19:57       ` [PATCH v4 2/5] nohz: support PR_CPU_ISOLATED_STRICT mode Chris Metcalf
2015-07-13 21:47         ` Andy Lutomirski
2015-07-21 19:34           ` Chris Metcalf
2015-07-21 19:42             ` Andy Lutomirski
2015-07-24 20:29               ` Chris Metcalf
2015-07-13 19:57       ` [PATCH v4 3/5] nohz: cpu_isolated strict mode configurable signal Chris Metcalf
2015-07-13 19:58       ` [PATCH v4 4/5] nohz: add cpu_isolated_debug boot flag Chris Metcalf
2015-07-13 19:58       ` [PATCH v4 5/5] nohz: cpu_isolated: allow tick to be fully disabled Chris Metcalf
2015-07-28 19:49       ` [PATCH v5 0/6] support "cpu_isolated" mode for nohz_full Chris Metcalf
2015-07-28 19:49         ` [PATCH v5 1/6] vmstat: provide a function to quiet down the diff processing Chris Metcalf
2015-07-28 19:49         ` [PATCH v5 2/6] cpu_isolated: add initial support Chris Metcalf
2015-08-12 16:00           ` Frederic Weisbecker
2015-08-12 18:22             ` Chris Metcalf
2015-08-26 15:26               ` Frederic Weisbecker
2015-08-26 15:55                 ` Chris Metcalf
2015-07-28 19:49         ` [PATCH v5 3/6] cpu_isolated: support PR_CPU_ISOLATED_STRICT mode Chris Metcalf
2015-07-28 19:49         ` [PATCH v5 4/6] cpu_isolated: provide strict mode configurable signal Chris Metcalf
2015-07-28 19:49         ` [PATCH v5 5/6] cpu_isolated: add debug boot flag Chris Metcalf
2015-07-28 19:49         ` [PATCH v5 6/6] nohz: cpu_isolated: allow tick to be fully disabled Chris Metcalf
2015-08-25 19:55         ` [PATCH v6 0/6] support "task_isolated" mode for nohz_full Chris Metcalf
2015-08-25 19:55           ` [PATCH v6 1/6] vmstat: provide a function to quiet down the diff processing Chris Metcalf
2015-08-25 19:55           ` [PATCH v6 2/6] task_isolation: add initial support Chris Metcalf
2015-08-25 19:55           ` [PATCH v6 3/6] task_isolation: support PR_TASK_ISOLATION_STRICT mode Chris Metcalf
2015-08-26 10:36             ` Will Deacon
2015-08-26 15:10               ` Chris Metcalf
2015-09-02 10:13                 ` Will Deacon
2015-08-28 15:31               ` [PATCH v6.1 " Chris Metcalf
2015-08-25 19:55           ` [PATCH v6 4/6] task_isolation: provide strict mode configurable signal Chris Metcalf
2015-08-28 19:22             ` Andy Lutomirski
2015-09-02 18:38               ` [PATCH v6.2 3/6] task_isolation: support PR_TASK_ISOLATION_STRICT mode Chris Metcalf
2015-08-25 19:55           ` [PATCH v6 5/6] task_isolation: add debug boot flag Chris Metcalf
2015-08-25 19:55           ` [PATCH v6 6/6] nohz: task_isolation: allow tick to be fully disabled Chris Metcalf
2015-09-28 15:17           ` [PATCH v7 00/11] support "task_isolated" mode for nohz_full Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 01/11] vmstat: provide a function to quiet down the diff processing Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 02/11] task_isolation: add initial support Chris Metcalf
2015-10-01 12:14               ` Frederic Weisbecker
2015-10-01 12:18                 ` Thomas Gleixner
2015-10-01 12:23                   ` Frederic Weisbecker
2015-10-01 12:31                     ` Thomas Gleixner
2015-10-01 17:02                   ` Chris Metcalf
2015-10-01 21:20                     ` Thomas Gleixner
2015-10-02 17:15                       ` Chris Metcalf
2015-10-02 19:02                         ` Thomas Gleixner
2015-10-01 19:25                 ` Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 03/11] task_isolation: support PR_TASK_ISOLATION_STRICT mode Chris Metcalf
2015-09-28 20:51               ` Andy Lutomirski
2015-09-28 21:54                 ` Chris Metcalf
2015-09-28 22:38                   ` Andy Lutomirski
2015-09-29 17:35                     ` Chris Metcalf
2015-09-29 17:46                       ` Andy Lutomirski
2015-09-29 17:57                         ` Chris Metcalf
2015-09-29 18:00                           ` Andy Lutomirski
2015-10-01 19:25                             ` Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 04/11] task_isolation: provide strict mode configurable signal Chris Metcalf
2015-09-28 20:54               ` Andy Lutomirski
2015-09-28 21:54                 ` Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 05/11] task_isolation: add debug boot flag Chris Metcalf
2015-09-28 20:59               ` Andy Lutomirski
2015-09-28 21:55                 ` Chris Metcalf
2015-09-28 22:40                   ` Andy Lutomirski
2015-09-29 17:35                     ` Chris Metcalf
2015-10-05 17:07               ` Luiz Capitulino
2015-10-08  0:33                 ` Chris Metcalf
2015-10-08 20:28                   ` Luiz Capitulino
2015-09-28 15:17             ` Chris Metcalf [this message]
2015-09-28 20:40               ` [PATCH v7 06/11] nohz: task_isolation: allow tick to be fully disabled Andy Lutomirski
2015-10-01 13:07                 ` Frederic Weisbecker
2015-10-01 14:13                   ` Thomas Gleixner
2015-09-28 15:17             ` [PATCH v7 07/11] arch/x86: enable task isolation functionality Chris Metcalf
2015-09-28 20:59               ` Andy Lutomirski
2015-09-28 21:57                 ` Chris Metcalf
2015-09-28 22:43                   ` Andy Lutomirski
2015-09-29 17:42                     ` Chris Metcalf
2015-09-29 17:57                       ` Andy Lutomirski
2015-09-30 20:25                         ` Thomas Gleixner
2015-09-30 20:58                           ` Chris Metcalf
2015-09-30 22:02                             ` Thomas Gleixner
2015-09-30 22:11                               ` Andy Lutomirski
2015-10-01  8:12                                 ` Thomas Gleixner
2015-10-01  9:08                                   ` Christoph Lameter
2015-10-01 10:10                                     ` Thomas Gleixner
2015-10-01 19:25                                   ` Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 08/11] arch/arm64: adopt prepare_exit_to_usermode() model from x86 Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 09/11] arch/arm64: enable task isolation functionality Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 10/11] arch/tile: adopt prepare_exit_to_usermode() model from x86 Chris Metcalf
2015-09-28 15:17             ` [PATCH v7 11/11] arch/tile: enable task isolation functionality Chris Metcalf
2015-10-20 20:35             ` [PATCH v8 00/14] support "task_isolation" mode for nohz_full Chris Metcalf
2015-10-20 20:35               ` [PATCH v8 01/14] vmstat: provide a function to quiet down the diff processing Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 02/14] vmstat: add vmstat_idle function Chris Metcalf
2015-10-20 20:45                 ` Christoph Lameter
2015-10-20 20:36               ` [PATCH v8 03/14] lru_add_drain_all: factor out lru_add_drain_needed Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 04/14] task_isolation: add initial support Chris Metcalf
2015-10-20 20:56                 ` Andy Lutomirski
2015-10-20 21:20                   ` Chris Metcalf
2015-10-20 21:26                     ` Andy Lutomirski
2015-10-21  0:29                       ` Steven Rostedt
2015-10-26 20:19                         ` Chris Metcalf
2015-10-26 21:13                         ` Chris Metcalf
2015-10-26 20:32                       ` Chris Metcalf
2015-10-21 16:12                 ` Frederic Weisbecker
2015-10-27 16:40                   ` Chris Metcalf
2016-01-28 16:38                     ` Frederic Weisbecker
2016-02-11 19:58                       ` Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 05/14] task_isolation: support PR_TASK_ISOLATION_STRICT mode Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 06/14] task_isolation: provide strict mode configurable signal Chris Metcalf
2015-10-21  0:56                 ` Steven Rostedt
2015-10-21  1:30                   ` Chris Metcalf
2015-10-21  1:41                     ` Steven Rostedt
2015-10-21  1:42                     ` Andy Lutomirski
2015-10-21  6:41                       ` Gilad Ben Yossef
2015-10-21 18:53                         ` Andy Lutomirski
2015-10-22 20:44                           ` Chris Metcalf
2015-10-22 21:00                             ` Andy Lutomirski
2015-10-27 19:37                               ` Chris Metcalf
2015-10-24  9:16                           ` Gilad Ben Yossef
2015-10-20 20:36               ` [PATCH v8 07/14] task_isolation: add debug boot flag Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 08/14] nohz_full: allow disabling the 1Hz minimum tick at boot Chris Metcalf
2015-10-20 21:03                 ` Frederic Weisbecker
2015-10-20 21:18                   ` Chris Metcalf
2015-10-21  0:59                     ` Steven Rostedt
2015-10-21  6:56                   ` Gilad Ben Yossef
2015-10-21 14:28                   ` Christoph Lameter
2015-10-21 15:35                     ` Frederic Weisbecker
2015-10-20 20:36               ` [PATCH v8 09/14] arch/x86: enable task isolation functionality Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 10/14] arch/arm64: adopt prepare_exit_to_usermode() model from x86 Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 11/14] arch/arm64: enable task isolation functionality Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 12/14] arch/tile: adopt prepare_exit_to_usermode() model from x86 Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 13/14] arch/tile: turn off timer tick for oneshot_stopped state Chris Metcalf
2015-10-20 20:36               ` [PATCH v8 14/14] arch/tile: enable task isolation functionality Chris Metcalf
2015-10-21 12:39               ` [PATCH v8 00/14] support "task_isolation" mode for nohz_full Peter Zijlstra
2015-10-22 20:31                 ` Chris Metcalf
2015-10-23  2:33                   ` Frederic Weisbecker
2015-10-23  8:49                     ` Peter Zijlstra
2015-10-23 13:29                       ` Frederic Weisbecker
2015-10-23  9:04                   ` Peter Zijlstra
2015-10-23 11:52                     ` Theodore Ts'o

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=1443453446-7827-7-git-send-email-cmetcalf@ezchip.com \
    --to=cmetcalf@ezchip.com \
    --cc=akpm@linux-foundation.org \
    --cc=catalin.marinas@arm.com \
    --cc=cl@linux.com \
    --cc=fweisbec@gmail.com \
    --cc=giladb@ezchip.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luto@amacapital.net \
    --cc=mingo@kernel.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=peterz@infradead.org \
    --cc=riel@redhat.com \
    --cc=rostedt@goodmis.org \
    --cc=tglx@linutronix.de \
    --cc=tj@kernel.org \
    --cc=viresh.kumar@linaro.org \
    --cc=will.deacon@arm.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®