linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Frederic Weisbecker <frederic@kernel.org>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Frederic Weisbecker <frederic@kernel.org>,
	Anna-Maria Gleixner <anna-maria@linutronix.de>,
	Peter Zijlstra <peterz@infradead.org>,
	Juri Lelli <juri.lelli@redhat.com>
Subject: [RFC PATCH 08/10] timer: Spare timer softirq until next expiry
Date: Wed,  1 Jul 2020 03:10:28 +0200	[thread overview]
Message-ID: <20200701011030.14324-9-frederic@kernel.org> (raw)
In-Reply-To: <20200701011030.14324-1-frederic@kernel.org>

Now that the core timer infrastructure doesn't depend anymore on
periodic base->clk increments, even when the CPU is not in NO_HZ mode,
we can delay the timer softirqs until we have actual timers to expire.

Some spurious softirqs can still remain since base->next_expiry doesn't
keep track of canceled timers but we are still way ahead of the
unconditional periodic softirqs (~15 times less of them with 1000 Hz
and ~5 times less with 100 Hz).

Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Juri Lelli <juri.lelli@redhat.com>
---
 kernel/time/timer.c | 49 ++++++++-------------------------------------
 1 file changed, 8 insertions(+), 41 deletions(-)

diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 6c021be0e76f..95f51b646447 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1449,10 +1449,10 @@ static void expire_timers(struct timer_base *base, struct hlist_head *head)
 	}
 }
 
-static int __collect_expired_timers(struct timer_base *base,
-				    struct hlist_head *heads)
+static int collect_expired_timers(struct timer_base *base,
+				  struct hlist_head *heads)
 {
-	unsigned long clk = base->clk;
+	unsigned long clk = base->clk = base->next_expiry;
 	struct hlist_head *vec;
 	int i, levels = 0;
 	unsigned int idx;
@@ -1675,40 +1675,6 @@ void timer_clear_idle(void)
 	 */
 	base->is_idle = false;
 }
-
-static int collect_expired_timers(struct timer_base *base,
-				  struct hlist_head *heads)
-{
-	unsigned long now = READ_ONCE(jiffies);
-
-	/*
-	 * NOHZ optimization. After a long idle sleep we need to forward the
-	 * base to current jiffies. Avoid a loop by searching the bitfield for
-	 * the next expiring timer.
-	 */
-	if ((long)(now - base->clk) > 2) {
-		/*
-		 * If the next timer is ahead of time forward to current
-		 * jiffies, otherwise forward to the next expiry time:
-		 */
-		if (time_after(base->next_expiry, now)) {
-			/*
-			 * The call site will increment base->clk and then
-			 * terminate the expiry loop immediately.
-			 */
-			base->clk = now;
-			return 0;
-		}
-		base->clk = base->next_expiry;
-	}
-	return __collect_expired_timers(base, heads);
-}
-#else
-static inline int collect_expired_timers(struct timer_base *base,
-					 struct hlist_head *heads)
-{
-	return __collect_expired_timers(base, heads);
-}
 #endif
 
 /*
@@ -1741,7 +1707,7 @@ static inline void __run_timers(struct timer_base *base)
 	struct hlist_head heads[LVL_DEPTH];
 	int levels;
 
-	if (!time_after_eq(jiffies, base->clk))
+	if (time_before(jiffies, base->next_expiry))
 		return;
 
 	timer_base_lock_expiry(base);
@@ -1754,7 +1720,8 @@ static inline void __run_timers(struct timer_base *base)
 	 */
 	base->must_forward_clk = false;
 
-	while (time_after_eq(jiffies, base->clk)) {
+	while (time_after_eq(jiffies, base->clk) &&
+	       time_after_eq(jiffies, base->next_expiry)) {
 
 		levels = collect_expired_timers(base, heads);
 		base->clk++;
@@ -1789,12 +1756,12 @@ void run_local_timers(void)
 
 	hrtimer_run_queues();
 	/* Raise the softirq only if required. */
-	if (time_before(jiffies, base->clk)) {
+	if (time_before(jiffies, base->next_expiry)) {
 		if (!IS_ENABLED(CONFIG_NO_HZ_COMMON))
 			return;
 		/* CPU is awake, so check the deferrable base. */
 		base++;
-		if (time_before(jiffies, base->clk))
+		if (time_before(jiffies, base->next_expiry))
 			return;
 	}
 	raise_softirq(TIMER_SOFTIRQ);
-- 
2.26.2


  parent reply	other threads:[~2020-07-01  1:11 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-01  1:10 [RFC PATCH 00/10] timer: Reduce timers softirq (and other optimizations) Frederic Weisbecker
2020-07-01  1:10 ` [RFC PATCH 01/10] timer: Prevent base->clk from moving backward Frederic Weisbecker
2020-07-01 16:35   ` Juri Lelli
2020-07-01 23:20     ` Frederic Weisbecker
2020-07-02  9:59       ` Juri Lelli
2020-07-02 14:04         ` Frederic Weisbecker
2020-07-02 14:32         ` Frederic Weisbecker
2020-07-02 15:57           ` Juri Lelli
2020-07-02  9:48   ` Thomas Gleixner
2020-07-01  1:10 ` [RFC PATCH 02/10] timer: Move trigger_dyntick_cpu() to enqueue_timer() Frederic Weisbecker
2020-07-01  1:10 ` [RFC PATCH 03/10] timer: Simplify LVL_START() and calc_index() Frederic Weisbecker
2020-07-02 11:59   ` Thomas Gleixner
2020-07-02 12:27     ` Frederic Weisbecker
2020-07-01  1:10 ` [RFC PATCH 04/10] timer: Optimize _next_timer_interrupt() level iteration Frederic Weisbecker
2020-07-01  1:10 ` [RFC PATCH 05/10] timers: Always keep track of next expiry Frederic Weisbecker
2020-07-01  1:10 ` [RFC PATCH 06/10] timer: Reuse next expiry cache after nohz exit Frederic Weisbecker
2020-07-01  1:10 ` [RFC PATCH 07/10] timer: Expand clk forward logic beyond nohz Frederic Weisbecker
2020-07-01  1:10 ` Frederic Weisbecker [this message]
2020-07-01  1:10 ` [RFC PATCH 09/10] timer: Remove must_forward_clk Frederic Weisbecker
2020-07-01  1:10 ` [RFC PATCH 10/10] timer: Lower base clock forwarding threshold Frederic Weisbecker
2020-07-02 13:21   ` Thomas Gleixner
2020-07-02 13:32     ` Frederic Weisbecker
2020-07-02 15:14       ` Thomas Gleixner
2020-07-03  0:12         ` Frederic Weisbecker
2020-07-03  9:13           ` 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=20200701011030.14324-9-frederic@kernel.org \
    --to=frederic@kernel.org \
    --cc=anna-maria@linutronix.de \
    --cc=juri.lelli@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=peterz@infradead.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).