Skip to contentSkip to conversation

timer: Forward timer base before migrating timers

PATCH New Lingutla Chandrasekhar 2 participants 2 messages Jan 18, 2018 – Feb 28, 2018

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 timer: Forward timer base before migrating timers

In case when timers are migrated to a CPU, after it exits idle, but before timer base is forwarded, either from run_timer_softirq()/mod_timer()/add_timer_on(), it's possible that migrated timers are queued, based on older clock value. This can cause delays in handling those timers.

For example, consider below sequence of events:

- CPU0 timer1 expires = 59969 and base->clk = 59131. So, timer is queued at level 2, with next expiry for this timer = 60032 (due to granularity addition).
- CPU1 enters idle @60007, with next timer expiry @60020.
- CPU1 exits idle.
- CPU0 is hotplugged at 60009, and timers are migrated to CPU1, with new base->clk = 60007. timer1 is queued, based on 60007 at level 0, for immediate handling (in next timer softirq handling).
- CPU1's base->clk is forwarded to 60009, so, in next sched timer interrupt, timer1 is not handled.

The issue happens as timer wheel collects expired timers starting from the current clk's index onwards, but migrated timers, if enqueued, based on older clk value can result in their index being less than clk's current index. This can only happen if new base->clk is ahead of timer->expires, resulting in timer being queued at new base->clk's current index.

kernel/time/timer.c +6 -0
index 89a9e1b4264a..f66c7ad55d7a 100644
@@ -1886,6 +1886,12 @@ int timers_dead_cpu(unsigned int cpu)
 		raw_spin_lock_irq(&new_base->lock);
 		raw_spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
 
+		/*
+		 * Before migrating timers, update new base clk to avoid
+		 * queueing timers based on older clock value.
+		 */
+		forward_timer_base(new_base);
+
 		BUG_ON(old_base->running_timer);
 
 		for (i = 0; i < WHEEL_SIZE; i++)
Signature · 2 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
 a Linux Foundation Collaborative Project.
#1 lore raw

[tip:timers/urgent] timers: Forward timer base before migrating timers

timers: Forward timer base before migrating timers
Commit-ID:  c52232a49e203a65a6e1a670cd5262f59e9364a0
Gitweb:     https://git.kernel.org/tip/c52232a49e203a65a6e1a670cd5262f59e9364a0
Author:     Lingutla Chandrasekhar <[email protected]>
AuthorDate: Thu, 18 Jan 2018 17:20:22 +0530
Committer:  Thomas Gleixner <[email protected]>
CommitDate: Wed, 28 Feb 2018 23:34:33 +0100

timers: Forward timer base before migrating timers

On CPU hotunplug the enqueued timers of the unplugged CPU are migrated to a live CPU. This happens from the control thread which initiated the unplug.

If the CPU on which the control thread runs came out from a longer idle period then the base clock of that CPU might be stale because the control thread runs prior to any event which forwards the clock.

In such a case the timers from the unplugged CPU are queued on the live CPU based on the stale clock which can cause large delays due to increased granularity of the outer timer wheels which are far away from base:;clock.

But there is a worse problem than that. The following sequence of events illustrates it:

- CPU0 timer1 is queued expires = 59969 and base->clk = 59131.

The timer is queued at wheel level 2, with resulting expiry time = 60032 (due to level granularity).

- CPU1 enters idle @60007, with next timer expiry @60020.

- CPU0 is hotplugged at @60009

- CPU1 exits idle and runs the control thread which migrates the timers from CPU0

timer1 is now queued in level 0 for immediate handling in the next softirq because the requested expiry time 59969 is before CPU1 base->clk 60007

- CPU1 runs code which forwards the base clock which succeeds because the next expiring timer. which was collected at idle entry time is still set to 60020.

So it forwards beyond 60007 and therefore misses to expire the migrated timer1. That timer gets expired when the wheel wraps around again, which takes between 63 and 630ms depending on the HZ setting.

Address both problems by invoking forward_timer_base() for the control CPUs timer base. All other places, which might run into a similar problem (mod_timer()/add_timer_on()) already invoke forward_timer_base() to avoid that.

[ tglx: Massaged comment and changelog ]

---
 kernel/time/timer.c | 6 ++++++
 1 file changed, 6 insertions(+)
kernel/time/timer.c +6 -0
index 48150ab42de9..4a4fd567fb26 100644
@@ -1894,6 +1894,12 @@ int timers_dead_cpu(unsigned int cpu)
 		raw_spin_lock_irq(&new_base->lock);
 		raw_spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
 
+		/*
+		 * The current CPUs base clock might be stale. Update it
+		 * before moving the timers over.
+		 */
+		forward_timer_base(new_base);
+
 		BUG_ON(old_base->running_timer);
 
 		for (i = 0; i < WHEEL_SIZE; i++)