[PATCH v2] tick/nohz: Avoid unused timekeeping_max_deferment() calls

[PATCH v2] tick/nohz: Avoid unused timekeeping_max_deferment() calls

[Usama Arif]

tick_nohz_next_event() limits a CPU's sleep interval to the maximum
deferment supported by the current clocksource when that CPU owns the
do_timer() duty. If the duty is unassigned, the limit also applies when
the CPU's TS_FLAG_DO_TIMER_LAST flag is set.

After the early timer checks, the function currently reads the maximum
deferment unconditionally. It then replaces the result with KTIME_MAX
unless one of the two conditions above applies.

timekeeping_max_deferment() performs a seqcount-protected read of the
shared timekeeper and follows its clocksource pointer. Check the do_timer
state first and avoid this work when the result would be discarded. This
leaves the resulting expiry unchanged and reduces accesses to timekeeper
data that is modified regularly.

On x86-64 this removes 18-20 dynamically executed instructions, including
the call, from the common non-owner path when the seqcount does not retry.

Signed-off-by: Usama Arif <usama.arif@linux.dev>
---
v1 -> v2:
- Remove the unnecessary comment and delta variable (Frederic Weisbecker).
---
 kernel/time/tick-sched.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index c8f2c4a503b08..a7893a079a83f 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -816,7 +816,7 @@ u64 get_jiffies_update(unsigned long *basej)
  */
 static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
 {
-	u64 basemono, next_tick, delta, expires;
+	u64 basemono, next_tick, expires;
 	unsigned long basejiff;
 	int tick_cpu;
 
@@ -856,8 +856,7 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
 	 * If the tick is due in the next period, keep it ticking or
 	 * force prod the timer.
 	 */
-	delta = next_tick - basemono;
-	if (delta <= (u64)TICK_NSEC) {
+	if (next_tick - basemono <= (u64)TICK_NSEC) {
 		/*
 		 * We've not stopped the tick yet, and there's a timer in the
 		 * next period, so no point in stopping it either, bail.
@@ -873,17 +872,19 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
 	 * the sleep time to the timekeeping 'max_deferment' value.
 	 * Otherwise we can sleep as long as we want.
 	 */
-	delta = timekeeping_max_deferment();
 	tick_cpu = READ_ONCE(tick_do_timer_cpu);
 	if (tick_cpu != cpu &&
-	    (tick_cpu != TICK_DO_TIMER_NONE || !tick_sched_flag_test(ts, TS_FLAG_DO_TIMER_LAST)))
-		delta = KTIME_MAX;
-
-	/* Calculate the next expiry time */
-	if (delta < (KTIME_MAX - basemono))
-		expires = basemono + delta;
-	else
+	    (tick_cpu != TICK_DO_TIMER_NONE || !tick_sched_flag_test(ts, TS_FLAG_DO_TIMER_LAST))) {
 		expires = KTIME_MAX;
+	} else {
+		expires = timekeeping_max_deferment();
+
+		/* Calculate the next expiry time */
+		if (expires < (KTIME_MAX - basemono))
+			expires += basemono;
+		else
+			expires = KTIME_MAX;
+	}
 
 	ts->timer_expires = min_t(u64, expires, next_tick);
 
-- 
2.53.0-Meta

Re: [PATCH v2] tick/nohz: Avoid unused timekeeping_max_deferment() calls

[Usama Arif]

On Fri, 18 Sep 2026 06:34:08 -0700 Usama Arif <usama.arif@linux.dev> wrote:

> tick_nohz_next_event() limits a CPU's sleep interval to the maximum
> deferment supported by the current clocksource when that CPU owns the
> do_timer() duty. If the duty is unassigned, the limit also applies when
> the CPU's TS_FLAG_DO_TIMER_LAST flag is set.
> 
> After the early timer checks, the function currently reads the maximum
> deferment unconditionally. It then replaces the result with KTIME_MAX
> unless one of the two conditions above applies.
> 
> timekeeping_max_deferment() performs a seqcount-protected read of the
> shared timekeeper and follows its clocksource pointer. Check the do_timer
> state first and avoid this work when the result would be discarded. This
> leaves the resulting expiry unchanged and reduces accesses to timekeeper
> data that is modified regularly.
> 
> On x86-64 this removes 18-20 dynamically executed instructions, including
> the call, from the common non-owner path when the seqcount does not retry.
> 
> Signed-off-by: Usama Arif <usama.arif@linux.dev>
> ---
> v1 -> v2:
> - Remove the unnecessary comment and delta variable (Frederic Weisbecker).
> ---
>  kernel/time/tick-sched.c | 23 ++++++++++++-----------
>  1 file changed, 12 insertions(+), 11 deletions(-)

Hi,

Just wanted to check if there are any comments or reviews for this patch?

Thanks!
Usama

> 
> diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
> index c8f2c4a503b08..a7893a079a83f 100644
> --- a/kernel/time/tick-sched.c
> +++ b/kernel/time/tick-sched.c
> @@ -816,7 +816,7 @@ u64 get_jiffies_update(unsigned long *basej)
>   */
>  static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
>  {
> -	u64 basemono, next_tick, delta, expires;
> +	u64 basemono, next_tick, expires;
>  	unsigned long basejiff;
>  	int tick_cpu;
>  
> @@ -856,8 +856,7 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
>  	 * If the tick is due in the next period, keep it ticking or
>  	 * force prod the timer.
>  	 */
> -	delta = next_tick - basemono;
> -	if (delta <= (u64)TICK_NSEC) {
> +	if (next_tick - basemono <= (u64)TICK_NSEC) {
>  		/*
>  		 * We've not stopped the tick yet, and there's a timer in the
>  		 * next period, so no point in stopping it either, bail.
> @@ -873,17 +872,19 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
>  	 * the sleep time to the timekeeping 'max_deferment' value.
>  	 * Otherwise we can sleep as long as we want.
>  	 */
> -	delta = timekeeping_max_deferment();
>  	tick_cpu = READ_ONCE(tick_do_timer_cpu);
>  	if (tick_cpu != cpu &&
> -	    (tick_cpu != TICK_DO_TIMER_NONE || !tick_sched_flag_test(ts, TS_FLAG_DO_TIMER_LAST)))
> -		delta = KTIME_MAX;
> -
> -	/* Calculate the next expiry time */
> -	if (delta < (KTIME_MAX - basemono))
> -		expires = basemono + delta;
> -	else
> +	    (tick_cpu != TICK_DO_TIMER_NONE || !tick_sched_flag_test(ts, TS_FLAG_DO_TIMER_LAST))) {
>  		expires = KTIME_MAX;
> +	} else {
> +		expires = timekeeping_max_deferment();
> +
> +		/* Calculate the next expiry time */
> +		if (expires < (KTIME_MAX - basemono))
> +			expires += basemono;
> +		else
> +			expires = KTIME_MAX;
> +	}
>  
>  	ts->timer_expires = min_t(u64, expires, next_tick);
>  
> -- 
> 2.53.0-Meta
> 
> 

Re: [PATCH v2] tick/nohz: Avoid unused timekeeping_max_deferment() calls

[Rik van Riel]

On Fri, 2026-09-18 at 06:34 -0700, Usama Arif wrote:
> tick_nohz_next_event() limits a CPU's sleep interval to the maximum
> deferment supported by the current clocksource when that CPU owns the
> do_timer() duty. If the duty is unassigned, the limit also applies
> when
> the CPU's TS_FLAG_DO_TIMER_LAST flag is set.
> 
> After the early timer checks, the function currently reads the
> maximum
> deferment unconditionally. It then replaces the result with KTIME_MAX
> unless one of the two conditions above applies.
> 
> timekeeping_max_deferment() performs a seqcount-protected read of the
> shared timekeeper and follows its clocksource pointer. Check the
> do_timer
> state first and avoid this work when the result would be discarded.
> This
> leaves the resulting expiry unchanged and reduces accesses to
> timekeeper
> data that is modified regularly.
> 
> On x86-64 this removes 18-20 dynamically executed instructions,
> including
> the call, from the common non-owner path when the seqcount does not
> retry.
> 
> Signed-off-by: Usama Arif <usama.arif@linux.dev>

The changelog confused me a little at first, but after
reading through the code it all made sense.

Reviewed-by: Rik van Riel <riel@surriel.com>

-- 
All Rights Reversed.

[tip: timers/nohz] tick/nohz: Avoid unused timekeeping_max_deferment() calls

[tip-bot2 for Usama Arif]

The following commit has been merged into the timers/nohz branch of tip:

Commit-ID:     d305927765cf6025bc11db9d96c06ea663230ddd
Gitweb:        https://git.kernel.org/tip/d305927765cf6025bc11db9d96c06ea663230ddd
Author:        Usama Arif <usama.arif@linux.dev>
AuthorDate:    Fri, 18 Sep 2026 06:34:08 -07:00
Committer:     Thomas Gleixner <tglx@kernel.org>
CommitterDate: Tue, 29 Sep 2026 21:00:08 +02:00

tick/nohz: Avoid unused timekeeping_max_deferment() calls

tick_nohz_next_event() limits a CPU's sleep interval to the maximum
deferment supported by the current clocksource when that CPU owns the
do_timer() duty. If the duty is unassigned, the limit also applies when
the CPU's TS_FLAG_DO_TIMER_LAST flag is set.

After the early timer checks, the function currently reads the maximum
deferment unconditionally. It then replaces the result with KTIME_MAX
unless one of the two conditions above applies.

timekeeping_max_deferment() performs a seqcount-protected read of the
shared timekeeper and follows its clocksource pointer. Check the do_timer
state first and avoid this work when the result would be discarded. This
leaves the resulting expiry unchanged and reduces accesses to timekeeper
data that is modified regularly.

On x86-64 this removes 18-20 dynamically executed instructions, including
the call, from the common non-owner path when the seqcount does not retry.

Signed-off-by: Usama Arif <usama.arif@linux.dev>
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Reviewed-by: Rik van Riel <riel@surriel.com>
Link: https://patch.msgid.link/20260918133408.2834751-1-usama.arif@linux.dev
---
 kernel/time/tick-sched.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index c8f2c4a..a7893a0 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -816,7 +816,7 @@ u64 get_jiffies_update(unsigned long *basej)
  */
 static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
 {
-	u64 basemono, next_tick, delta, expires;
+	u64 basemono, next_tick, expires;
 	unsigned long basejiff;
 	int tick_cpu;
 
@@ -856,8 +856,7 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
 	 * If the tick is due in the next period, keep it ticking or
 	 * force prod the timer.
 	 */
-	delta = next_tick - basemono;
-	if (delta <= (u64)TICK_NSEC) {
+	if (next_tick - basemono <= (u64)TICK_NSEC) {
 		/*
 		 * We've not stopped the tick yet, and there's a timer in the
 		 * next period, so no point in stopping it either, bail.
@@ -873,17 +872,19 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
 	 * the sleep time to the timekeeping 'max_deferment' value.
 	 * Otherwise we can sleep as long as we want.
 	 */
-	delta = timekeeping_max_deferment();
 	tick_cpu = READ_ONCE(tick_do_timer_cpu);
 	if (tick_cpu != cpu &&
-	    (tick_cpu != TICK_DO_TIMER_NONE || !tick_sched_flag_test(ts, TS_FLAG_DO_TIMER_LAST)))
-		delta = KTIME_MAX;
-
-	/* Calculate the next expiry time */
-	if (delta < (KTIME_MAX - basemono))
-		expires = basemono + delta;
-	else
+	    (tick_cpu != TICK_DO_TIMER_NONE || !tick_sched_flag_test(ts, TS_FLAG_DO_TIMER_LAST))) {
 		expires = KTIME_MAX;
+	} else {
+		expires = timekeeping_max_deferment();
+
+		/* Calculate the next expiry time */
+		if (expires < (KTIME_MAX - basemono))
+			expires += basemono;
+		else
+			expires = KTIME_MAX;
+	}
 
 	ts->timer_expires = min_t(u64, expires, next_tick);