mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: mingo@kernel.org, tglx@linutronix.de, luto@amacapital.net,
	nicolas.pitre@linaro.org, daniel.lezcano@linaro.org,
	peterz@infradead.org, umgwanakikbuti@gmail.com
Cc: linux-kernel@vger.kernel.org
Subject: [RFC][PATCH 6/8] sched,idle: Avoid spurious wakeup IPIs
Date: Fri, 11 Apr 2014 15:42:49 +0200	[thread overview]
Message-ID: <20140411135218.478299389@infradead.org> (raw)
In-Reply-To: <20140411134243.160989490@infradead.org>

[-- Attachment #1: peterz-sched-resched_cpu.patch --]
[-- Type: text/plain, Size: 2698 bytes --]

Because mwait_idle_with_hints() gets called from !idle context it must
call current_clr_polling(). This however means that resched_task() is
very likely to send an IPI even when we were polling:

  CPU0					CPU1

  if (current_set_polling_and_test())
    goto out;

  __monitor(&ti->flags);
  if (!need_resched())
    __mwait(eax, ecx);
					set_tsk_need_resched(p);
					smp_mb();
out:
  current_clr_polling();
					if (!tsk_is_polling(p))
					  smp_send_reschedule(cpu);


So while it is correct (extra IPIs aren't a problem, whereas a missed
IPI would be) it is a performance problem (for some).

Avoid this issue by using fetch_or() to atomically set NEED_RESCHED
and test if POLLING_NRFLAG is set.

Since a CPU stuck in mwait is unlikely to modify the flags word,
contention on the cmpxchg is unlikely and thus we should mostly
succeed in a single go.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
---
 kernel/sched/core.c |   41 ++++++++++++++++++++++++++++++++++++-----
 1 file changed, 36 insertions(+), 5 deletions(-)

--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -505,6 +505,39 @@ static inline void init_hrtick(void)
 #endif	/* CONFIG_SCHED_HRTICK */
 
 /*
+ * cmpxchg based fetch_or, macro so it works for different integer types
+ */
+#define fetch_or(ptr, val)						\
+({	typeof(*(ptr)) __old, __val = *(ptr);				\
+ 	for (;;) {							\
+ 		__old = cmpxchg((ptr), __val, __val | (val));		\
+ 		if (__old == __val)					\
+ 			break;						\
+ 		__val = __old;						\
+ 	}								\
+ 	__old;								\
+})
+
+#ifdef TIF_POLLING_NRFLAG
+/*
+ * Atomically set TIF_NEED_RESCHED and test for TIF_POLLING_NRFLAG,
+ * this avoids any races wrt polling state changes and thereby avoids
+ * spurious IPIs.
+ */
+static bool set_nr_and_not_polling(struct task_struct *p)
+{
+	struct thread_info *ti = task_thread_info(p);
+	return !(fetch_or(&ti->flags, _TIF_NEED_RESCHED) & _TIF_POLLING_NRFLAG);
+}
+#else
+static bool set_nr_and_not_polling(struct task_struct *p)
+{
+	set_tsk_need_resched(p);
+	return true;
+}
+#endif
+
+/*
  * resched_task - mark a task 'to be rescheduled now'.
  *
  * On UP this means the setting of the need_resched flag, on SMP it
@@ -520,17 +553,15 @@ void resched_task(struct task_struct *p)
 	if (test_tsk_need_resched(p))
 		return;
 
-	set_tsk_need_resched(p);
-
 	cpu = task_cpu(p);
+
 	if (cpu == smp_processor_id()) {
+		set_tsk_need_resched(p);
 		set_preempt_need_resched();
 		return;
 	}
 
-	/* NEED_RESCHED must be visible before we test polling */
-	smp_mb();
-	if (!tsk_is_polling(p))
+	if (set_nr_and_not_polling(p))
 		smp_send_reschedule(cpu);
 }
 



  parent reply	other threads:[~2014-04-11 13:57 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-04-11 13:42 [RFC][PATCH 0/8] sched,idle: need resched polling rework Peter Zijlstra
2014-04-11 13:42 ` [RFC][PATCH 1/8] sched,idle,alpha: Switch from TS_POLLING to TIF_POLLING_NRFLAG Peter Zijlstra
2014-04-11 14:38   ` Richard Henderson
2014-04-11 13:42 ` [RFC][PATCH 2/8] sched,idle,tile: " Peter Zijlstra
2014-04-11 15:15   ` Chris Metcalf
2014-04-11 15:30     ` Peter Zijlstra
2014-04-11 13:42 ` [RFC][PATCH 3/8] sched,idle,ia64: " Peter Zijlstra
2014-04-11 13:42 ` [RFC][PATCH 4/8] sched,idle,x86: " Peter Zijlstra
2014-04-11 13:42 ` [RFC][PATCH 5/8] sched,idle: Remove TS_POLLING support Peter Zijlstra
2014-04-11 13:42 ` Peter Zijlstra [this message]
2014-04-13 21:41   ` [RFC][PATCH 6/8] sched,idle: Avoid spurious wakeup IPIs Nicolas Pitre
2014-05-09 13:37   ` James Hogan
2014-05-09 14:15     ` Peter Zijlstra
2014-05-09 14:40       ` Catalin Marinas
2014-05-09 14:50         ` Peter Zijlstra
2014-05-09 14:57           ` Catalin Marinas
2014-05-09 17:02             ` Peter Zijlstra
2014-05-09 17:06               ` Peter Zijlstra
2014-05-09 17:09                 ` Catalin Marinas
2014-05-09 17:20                   ` Peter Zijlstra
2014-05-19 12:54                 ` [tip:sched/arch] arm64: Remove TIF_POLLING_NRFLAG tip-bot for Peter Zijlstra
2014-05-22 12:26                 ` [tip:sched/core] " tip-bot for Peter Zijlstra
2014-05-09 14:51       ` [RFC][PATCH 6/8] sched,idle: Avoid spurious wakeup IPIs James Hogan
2014-05-15  9:17         ` James Hogan
2014-05-19 12:54         ` [tip:sched/arch] metag: Remove TIF_POLLING_NRFLAG tip-bot for James Hogan
2014-05-22 12:26         ` [tip:sched/core] " tip-bot for James Hogan
2014-04-11 13:42 ` [RFC][PATCH 7/8] sched,idle: Delay clearing the polling bit Peter Zijlstra
2014-04-13 21:51   ` Nicolas Pitre
2014-04-11 13:42 ` [RFC][PATCH 8/8] sched,idle: Reflow cpuidle_idle_call() Peter Zijlstra
2014-04-13 21:36   ` Nicolas Pitre
2014-04-14  8:59     ` Peter Zijlstra
2014-04-14  9:25       ` Peter Zijlstra
2014-04-14 13:55         ` Nicolas Pitre
2014-04-11 15:00 ` [RFC][PATCH 0/8] sched,idle: need resched polling rework Andy Lutomirski
2014-04-11 15:14   ` Peter Zijlstra
2014-05-22 12:58   ` Peter Zijlstra
2014-05-22 13:09     ` Peter Zijlstra
2014-05-29  0:01       ` Andy Lutomirski
2014-05-29  6:48         ` Peter Zijlstra
2014-06-03  6:40           ` Andy Lutomirski
2014-06-03  6:51             ` Peter Zijlstra
2014-06-03 10:43             ` Peter Zijlstra
2014-06-03 14:02               ` Peter Zijlstra
2014-06-03 16:05                 ` Andy Lutomirski
2014-06-03 16:19                   ` Peter Zijlstra
2014-06-03 16:52                     ` Andy Lutomirski
2014-06-03 17:00                       ` Peter Zijlstra
2014-06-03 18:28                         ` Peter Zijlstra
2014-06-03 18:44                           ` Andy Lutomirski
2014-06-03 20:07                             ` Andy Lutomirski
2014-04-12  8:35 ` Mike Galbraith

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=20140411135218.478299389@infradead.org \
    --to=peterz@infradead.org \
    --cc=daniel.lezcano@linaro.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luto@amacapital.net \
    --cc=mingo@kernel.org \
    --cc=nicolas.pitre@linaro.org \
    --cc=tglx@linutronix.de \
    --cc=umgwanakikbuti@gmail.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

Powered by JetHome