From: Peter Zijlstra <peterz@infradead.org>
To: Sasha Levin <sasha.levin@oracle.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>,
lenb@kernel.org, rjw@rjwysocki.net,
Eliezer Tamir <eliezer.tamir@linux.intel.com>,
rui.zhang@intel.com, jacob.jun.pan@linux.intel.com,
Mike Galbraith <bitbucket@online.de>,
Ingo Molnar <mingo@kernel.org>,
hpa@zytor.com, paulmck@linux.vnet.ibm.com,
Thomas Gleixner <tglx@linutronix.de>,
John Stultz <john.stultz@linaro.org>,
Andy Lutomirski <luto@amacapital.net>,
linux-kernel@vger.kernel.org, dyoung@redhat.com
Subject: Re: [PATCH 13/15] sched: Use a static_key for sched_clock_stable
Date: Wed, 22 Jan 2014 12:59:18 +0100 [thread overview]
Message-ID: <20140122115918.GG3694@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <20140122104532.GJ31570@twins.programming.kicks-ass.net>
On Wed, Jan 22, 2014 at 11:45:32AM +0100, Peter Zijlstra wrote:
> Ho humm.
OK, so I had me a ponder; does the below fix things for you and David?
I've only done a boot test on real proper hardware :-)
---
kernel/sched/clock.c | 42 +++++++++++++++++++++++++++++++++---------
1 file changed, 33 insertions(+), 9 deletions(-)
diff --git a/kernel/sched/clock.c b/kernel/sched/clock.c
index 6bd6a6731b21..6bbcd97f4532 100644
--- a/kernel/sched/clock.c
+++ b/kernel/sched/clock.c
@@ -77,35 +77,45 @@ __read_mostly int sched_clock_running;
#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
static struct static_key __sched_clock_stable = STATIC_KEY_INIT;
+static int __sched_clock_stable_early;
int sched_clock_stable(void)
{
- if (static_key_false(&__sched_clock_stable))
- return false;
- return true;
+ return static_key_false(&__sched_clock_stable);
}
void set_sched_clock_stable(void)
{
+ __sched_clock_stable_early = 1;
+
+ smp_mb(); /* matches sched_clock_init() */
+
+ if (!sched_clock_running)
+ return;
+
if (!sched_clock_stable())
- static_key_slow_dec(&__sched_clock_stable);
+ static_key_slow_inc(&__sched_clock_stable);
}
static void __clear_sched_clock_stable(struct work_struct *work)
{
/* XXX worry about clock continuity */
if (sched_clock_stable())
- static_key_slow_inc(&__sched_clock_stable);
+ static_key_slow_dec(&__sched_clock_stable);
}
static DECLARE_WORK(sched_clock_work, __clear_sched_clock_stable);
void clear_sched_clock_stable(void)
{
- if (keventd_up())
- schedule_work(&sched_clock_work);
- else
- __clear_sched_clock_stable(&sched_clock_work);
+ __sched_clock_stable_early = 0;
+
+ smp_mb(); /* matches sched_clock_init() */
+
+ if (!sched_clock_running)
+ return;
+
+ schedule_work(&sched_clock_work);
}
struct sched_clock_data {
@@ -140,6 +150,20 @@ void sched_clock_init(void)
}
sched_clock_running = 1;
+
+ /*
+ * Ensure that it is impossible to not do a static_key update.
+ *
+ * Either {set,clear}_sched_clock_stable() must see sched_clock_running
+ * and do the update, or we must see their __sched_clock_stable_early
+ * and do the update, or both.
+ */
+ smp_mb(); /* matches {set,clear}_sched_clock_stable() */
+
+ if (__sched_clock_stable_early)
+ set_sched_clock_stable();
+ else
+ clear_sched_clock_stable();
}
/*
next prev parent reply other threads:[~2014-01-22 11:59 UTC|newest]
Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-12-12 14:08 [PATCH 00/15] cleanups and optimizations Peter Zijlstra
2013-12-12 14:08 ` [PATCH 01/15] x86, acpi, idle: Restructure the mwait idle routines Peter Zijlstra
2013-12-19 20:09 ` [tip:x86/idle] " tip-bot for Peter Zijlstra
2013-12-19 20:13 ` H. Peter Anvin
2013-12-19 20:16 ` Peter Zijlstra
2013-12-12 14:08 ` [PATCH 02/15] sched, preempt: Fixup missed PREEMPT_NEED_RESCHED folding Peter Zijlstra
2013-12-12 14:08 ` [PATCH 03/15] preempt, locking: Rework local_bh_{dis,en}able() Peter Zijlstra
2013-12-12 14:08 ` [PATCH 04/15] locking: Optimize lock_bh functions Peter Zijlstra
2013-12-12 14:08 ` [PATCH 05/15] sched, net: Clean up preempt_enable_no_resched() abuse Peter Zijlstra
2013-12-12 14:08 ` [PATCH 06/15] sched, net: Fixup busy_loop_us_clock() Peter Zijlstra
2013-12-12 14:08 ` [PATCH 07/15] sched, thermal: Clean up preempt_enable_no_resched() abuse Peter Zijlstra
2013-12-12 14:08 ` [PATCH 08/15] preempt: Take away preempt_enable_no_resched() from modules Peter Zijlstra
2013-12-12 14:08 ` [PATCH 09/15] x86: Use mul_u64_u32_shr() for native_sched_clock() Peter Zijlstra
2013-12-12 14:08 ` [PATCH 10/15] x86: Move some code around Peter Zijlstra
2013-12-12 14:08 ` [PATCH 11/15] x86: Rewrite cyc2ns to avoid the need to disable IRQs Peter Zijlstra
2014-06-16 17:13 ` Viresh Kumar
2014-06-17 12:15 ` Peter Zijlstra
2014-06-17 12:42 ` Viresh Kumar
2014-06-17 16:32 ` Mauro
2014-06-17 12:47 ` Mauro
2013-12-12 14:08 ` [PATCH 12/15] sched: Remove local_irq_disable() from the clocks Peter Zijlstra
2013-12-12 14:08 ` [PATCH 13/15] sched: Use a static_key for sched_clock_stable Peter Zijlstra
2014-01-21 22:28 ` Sasha Levin
2014-01-22 10:45 ` Peter Zijlstra
2014-01-22 11:59 ` Peter Zijlstra [this message]
2014-01-23 1:53 ` Dave Young
2014-01-23 16:46 ` [tip:sched/urgent] sched/clock: Fixup early initialization tip-bot for Peter Zijlstra
2014-01-22 12:00 ` [PATCH 13/15] sched: Use a static_key for sched_clock_stable Markus Trippelsdorf
2014-01-22 12:07 ` Peter Zijlstra
2014-01-22 12:16 ` Peter Zijlstra
2014-01-22 12:26 ` Markus Trippelsdorf
2014-01-22 12:30 ` Peter Zijlstra
2014-01-22 13:14 ` Markus Trippelsdorf
2014-01-22 14:26 ` Sasha Levin
2014-01-22 18:35 ` Markus Trippelsdorf
2014-01-22 18:42 ` Peter Zijlstra
2014-01-22 18:45 ` Markus Trippelsdorf
2014-01-22 19:17 ` Josh Boyer
2014-01-22 19:09 ` Markus Trippelsdorf
2014-01-22 19:12 ` Markus Trippelsdorf
2014-01-22 20:16 ` Peter Zijlstra
2014-01-22 21:08 ` Peter Zijlstra
2014-01-22 21:17 ` Markus Trippelsdorf
2014-01-23 9:48 ` Peter Zijlstra
2014-01-23 10:01 ` Markus Trippelsdorf
2014-01-23 10:04 ` Peter Zijlstra
2014-01-23 13:32 ` Josh Boyer
2014-01-23 16:46 ` [tip:sched/urgent] sched/x86/tsc: Initialize multiplier to 0 tip-bot for Peter Zijlstra
2014-01-22 23:53 ` [PATCH 13/15] sched: Use a static_key for sched_clock_stable Josh Boyer
2014-01-23 1:53 ` Dave Young
2014-01-23 2:10 ` Dave Young
2014-01-23 16:56 ` Peter Zijlstra
2014-01-24 3:15 ` Dave Young
2014-01-24 7:58 ` Ingo Molnar
2014-01-22 17:14 ` Peter Zijlstra
2014-01-22 22:31 ` Sasha Levin
2013-12-12 14:08 ` [PATCH 14/15] sched, clock: Fixup clear_sched_clock_stable() Peter Zijlstra
2013-12-12 14:08 ` [PATCH 15/15] x86: Avoid a runtime condition in native_sched_clock() Peter Zijlstra
2013-12-13 3:30 ` [PATCH 00/15] cleanups and optimizations Mike Galbraith
2013-12-13 10:49 ` Eliezer Tamir
2013-12-13 13:56 ` Peter Zijlstra
2013-12-16 17:48 ` Eliezer Tamir
2013-12-17 13:32 ` Peter Zijlstra
2013-12-17 14:03 ` Eliezer Tamir
2013-12-17 15:13 ` Peter Zijlstra
2013-12-17 18:19 ` Eliezer Tamir
2013-12-17 22:11 ` Thomas Gleixner
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=20140122115918.GG3694@twins.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=arjan@linux.intel.com \
--cc=bitbucket@online.de \
--cc=dyoung@redhat.com \
--cc=eliezer.tamir@linux.intel.com \
--cc=hpa@zytor.com \
--cc=jacob.jun.pan@linux.intel.com \
--cc=john.stultz@linaro.org \
--cc=lenb@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=mingo@kernel.org \
--cc=paulmck@linux.vnet.ibm.com \
--cc=rjw@rjwysocki.net \
--cc=rui.zhang@intel.com \
--cc=sasha.levin@oracle.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome