Skip to contentSkip to conversation

timers: Clear pending_map on CPU hotplug

PATCH Superseded Liang Hao 3 participants 6 messages Aug 23 – Oct 20 High-resolution timers, timer wheel, clockevents

lore
Share & tools
lore

Liang Hao

author
#0 lore raw

PATCH 1/1 timers: Clear pending_map on CPU hotplug

CPU hotplug migration leaves pending_map bits set on the outgoing CPU while its wheel vectors are empty. After re-online, a later next_expiry recalc can treat those empty buckets as pending and temporarily pull next_expiry forward until collect drains the bits.

Clear pending_map under the base lock after migration in timers_dead_cpu(), and in timers_prepare_cpu() for a full base reinit.

---
 kernel/time/timer.c | 3 +++
 1 file changed, 3 insertions(+)
kernel/time/timer.c +3 -0
index ae9abf14688e..e550489c73b8 100644
@@ -44,6 +44,7 @@
 #include <linux/compat.h>
 #include <linux/random.h>
 #include <linux/sysctl.h>
+#include <linux/bitmap.h>
 
 #include <linux/uaccess.h>
 #include <asm/unistd.h>
@@ -2509,6 +2510,7 @@ int timers_prepare_cpu(unsigned int cpu)
 		base->next_expiry_recalc = false;
 		base->timers_pending = false;
 		base->is_idle = false;
+		bitmap_zero(base->pending_map, WHEEL_SIZE);
 	}
 	return 0;
 }
@@ -2541,6 +2543,7 @@ int timers_dead_cpu(unsigned int cpu)
 		for (i = 0; i < WHEEL_SIZE; i++)
 			migrate_timer_list(new_base, old_base->vectors + i);
 
+		bitmap_zero(old_base->pending_map, WHEEL_SIZE);
 		raw_spin_unlock(&old_base->lock);
 		raw_spin_unlock_irq(&new_base->lock);
 		put_cpu_ptr(&timer_bases);
Signature · 1 line
2.50.1 (Apple Git-155)
author
#1 lore raw

Hi,

Additional measurements on a local arm64 Android board.

Method (two sessions per build):
5 runs per session, 5 offline/online pairs per run.

Per pair:
taskset 0x8 sh -c 'n=0; while [ $n -lt 100 ]; do sleep 0.01; n=$((n+1)); done' & sleep 0.5
echo 0 > /sys/devices/system/cpu/cpu3/online
echo 1 > /sys/devices/system/cpu/cpu3/online
wait

  # then the same sleep loop again

Temporary debug dmesg counts per run:
empty_bucket - collect_expired_timers() saw pending_map bit set
with an empty hlist (next_expiry had been pulled to a stale empty bucket after re-online)
stale_pending - timers_dead_cpu() saw non-empty pending_map after
migrate (same probe on both builds; Build A then clears, Build B does not)

Build A: clear pending_map after migrate
Build B: do not clear

Build A session 1 / 2 (empty_bucket, stale_pending):

  0,6  0,6  0,6  0,6  0,5
  0,7  0,5  0,5  0,5  0,5

Build B session 1 / 2:

  33,10  18,10  10,10  12,10  6,10
  32,10  14,10  17,10  13,10  10,10

stale_pending > 0 on both builds is expected: migrate leaves bits set while vectors are empty. Build A clears them; Build B leaves them. The discriminator is empty_bucket: without clearing it is > 0 every run; with clearing it stays 0.

Thanks,
Liang Hao

maintainer
#2 lore raw
In reply to Liang Hao
Quoted 34 lines · Liang Hao

CPU hotplug migration leaves pending_map bits set on the outgoing CPU while its wheel vectors are empty. After re-online, a later next_expiry recalc can treat those empty buckets as pending and temporarily pull next_expiry forward until collect drains the bits.

Clear pending_map under the base lock after migration in timers_dead_cpu(), and in timers_prepare_cpu() for a full base reinit.

---
 kernel/time/timer.c | 3 +++
 1 file changed, 3 insertions(+)
kernel/time/timer.c +2 -0
index ae9abf14688e..e550489c73b8 100644
@@ -44,6 +44,7 @@
 #include <linux/compat.h>
 #include <linux/random.h>
 #include <linux/sysctl.h>
+#include <linux/bitmap.h>
 
 #include <linux/uaccess.h>
 #include <asm/unistd.h>
@@ -2509,6 +2510,7 @@ int timers_prepare_cpu(unsigned int cpu)
 		base->next_expiry_recalc = false;
 		base->timers_pending = false;
 		base->is_idle = false;
+		bitmap_zero(base->pending_map, WHEEL_SIZE);
 	}
 	return 0;
 }

kernel/time/timer.c int timers_dead_cpu(unsigned int cpu) · line 2546

Quoted 4 lines
kernel/time/timer.c +1 -0
@@ -2541,6 +2543,7 @@ int timers_dead_cpu(unsigned int cpu)
 		for (i = 0; i < WHEEL_SIZE; i++)
 			migrate_timer_list(new_base, old_base->vectors + i);
 
+		bitmap_zero(old_base->pending_map, WHEEL_SIZE);

Why do you need to zero the map twice - once when the CPU goes out and then again when it comes back. That does not make sense.

Thanks

tglx

author
#3 lore raw
Quoted 2 lines · Thomas Gleixner

Why do you need to zero the map twice - once when the CPU goes out and then again when it comes back. That does not make sense.

You're right - clearing it twice is redundant. The stale bits come from migration in timers_dead_cpu(); clearing pending_map there after migration is sufficient. I'll remove the timers_prepare_cpu() part in v2.

Thanks,

Liang Hao

author
#4 lore raw

PATCH v2 1/1 timers: Clear pending_map on CPU hotplug

In reply to Liang Hao

CPU hotplug migration leaves pending_map bits set on the outgoing CPU while its wheel vectors are empty. After re-online, a later next_expiry recalc can treat those empty buckets as pending and temporarily pull next_expiry forward until collect drains the bits.

Clear pending_map under the base lock after migration in timers_dead_cpu().

---
v1 -> v2:
- remove redundant clear in timers_prepare_cpu()

 kernel/time/timer.c | 2 ++
 1 file changed, 2 insertions(+)
kernel/time/timer.c +2 -0
index 42afdcb229d8..a3c4e20a4e36 100644
@@ -44,6 +44,7 @@
 #include <linux/compat.h>
 #include <linux/random.h>
 #include <linux/sysctl.h>
+#include <linux/bitmap.h>
 
 #include <linux/uaccess.h>
 #include <asm/unistd.h>
@@ -2541,6 +2542,7 @@ int timers_dead_cpu(unsigned int cpu)
 		for (i = 0; i < WHEEL_SIZE; i++)
 			migrate_timer_list(new_base, old_base->vectors + i);
 
+		bitmap_zero(old_base->pending_map, WHEEL_SIZE);
 		raw_spin_unlock(&old_base->lock);
 		raw_spin_unlock_irq(&new_base->lock);
 		put_cpu_ptr(&timer_bases);
Signature · 1 line
2.54.0 (Apple Git-157)
merge bot Applied
#5 lore raw
The following commit has been merged into the timers/urgent branch of tip:

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

Commit-ID:     e7cad586d73276cbd4466de81b2092287d7475c0
Gitweb:        https://git.kernel.org/tip/e7cad586d73276cbd4466de81b2092287d7475c0
Author:        Liang Hao <[email protected]>
AuthorDate:    Wed, 30 Sep 2026 14:36:21 +08:00
Committer:     Thomas Gleixner <[email protected]>
CommitterDate: Tue, 20 Oct 2026 15:28:14 +02:00

timers: Clear pending_map on CPU hotplug

CPU hotplug migration leaves pending_map bits set on the outgoing CPU while its wheel vectors are empty. After re-online, a later next_expiry recalc can treat those empty buckets as pending and temporarily pull next_expiry forward until collect drains the bits.

Clear pending_map under the base lock after migration in timers_dead_cpu().

Fixes 500462a9de65 ("timers: Switch to a non-cascading wheel")
---
 kernel/time/timer.c | 2 ++
 1 file changed, 2 insertions(+)
kernel/time/timer.c +2 -0
index ae9abf1..fc37444 100644
@@ -44,6 +44,7 @@
 #include <linux/compat.h>
 #include <linux/random.h>
 #include <linux/sysctl.h>
+#include <linux/bitmap.h>
 
 #include <linux/uaccess.h>
 #include <asm/unistd.h>
@@ -2541,6 +2542,7 @@ int timers_dead_cpu(unsigned int cpu)
 		for (i = 0; i < WHEEL_SIZE; i++)
 			migrate_timer_list(new_base, old_base->vectors + i);
 
+		bitmap_zero(old_base->pending_map, WHEEL_SIZE);
 		raw_spin_unlock(&old_base->lock);
 		raw_spin_unlock_irq(&new_base->lock);
 		put_cpu_ptr(&timer_bases);