tick/sched: Ensure quiet_vmstat() is called when the idle tick was stopped too

This patch incorporates an idea from Marcelo's patch [1] where a CPU-specific variable namely vmstat_dirty is used to indicate if a vmstat imbalance is present for a given CPU. Therefore, at the appropriate time, we can fold all the remaining differentials.

[1]: https://lore.kernel.org/lkml/[email protected]/

Signed-off-by Aaron Tomlin <[email protected]>

---
 mm/vmstat.c | 46 +++++++++++++++-------------------------------
 1 file changed, 15 insertions(+), 31 deletions(-)

mm/vmstat.c +10 -0

index 373d2730fcf2..51564b7c85fe 100644
@@ -195,6 +195,12 @@ void fold_vm_numa_events(void)
 #endif
 
 #ifdef CONFIG_SMP
+static DEFINE_PER_CPU_ALIGNED(bool, vmstat_dirty);
+
+static inline void mark_vmstat_dirty(void)
+{
+	this_cpu_write(vmstat_dirty, true);
+}
 
 int calculate_pressure_threshold(struct zone *zone)
 {

@@ -367,6 +373,7 @@ void __mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
 		x = 0;
 	}
 	__this_cpu_write(*p, x);
+	mark_vmstat_dirty();
 
 	if (IS_ENABLED(CONFIG_PREEMPT_RT))
 		preempt_enable();

@@ -405,6 +412,7 @@ void __mod_node_page_state(struct pglist_data *pgdat, enum node_stat_item item,
 		x = 0;
 	}
 	__this_cpu_write(*p, x);
+	mark_vmstat_dirty();
 
 	if (IS_ENABLED(CONFIG_PREEMPT_RT))
 		preempt_enable();

@@ -603,6 +611,7 @@ static inline void mod_zone_state(struct zone *zone,
 
 	if (z)
 		zone_page_state_add(z, zone, item);
+	mark_vmstat_dirty();
 }
 
 void mod_zone_page_state(struct zone *zone, enum zone_stat_item item,

@@ -671,6 +680,7 @@ static inline void mod_node_state(struct pglist_data *pgdat,
 
 	if (z)
 		node_page_state_add(z, pgdat, item);
+	mark_vmstat_dirty();
 }
 
 void mod_node_page_state(struct pglist_data *pgdat, enum node_stat_item item,

mm/vmstat.c +1 -30

 		/*
 		 * Counters were updated so we expect more updates
 		 * to occur in the future. Keep on running the

@@ -1948,35 +1960,6 @@ static void vmstat_update(struct work_struct *w)
 	}
 }
 
-/*
- * Check if the diffs for a certain cpu indicate that
- * an update is needed.
- */
-static bool need_update(int cpu)
-{
-	pg_data_t *last_pgdat = NULL;
-	struct zone *zone;
-
-	for_each_populated_zone(zone) {
-		struct per_cpu_zonestat *pzstats = per_cpu_ptr(zone->per_cpu_zonestats, cpu);
-		struct per_cpu_nodestat *n;
-
-		/*
-		 * The fast way of checking if there are any vmstat diffs.
-		 */
-		if (memchr_inv(pzstats->vm_stat_diff, 0, sizeof(pzstats->vm_stat_diff)))
-			return true;
-
-		if (last_pgdat == zone->zone_pgdat)
-			continue;
-		last_pgdat = zone->zone_pgdat;
-		n = per_cpu_ptr(zone->zone_pgdat->per_cpu_nodestats, cpu);
-		if (memchr_inv(n->vm_node_stat_diff, 0, sizeof(n->vm_node_stat_diff)))
-			return true;
-	}
-	return false;
-}
-
 /*
  * Switch off vmstat processing and then fold all the remaining differentials
  * until the diffs stay at zero. The function is used by NOHZ and can only be

@@ -1990,7 +1973,7 @@ void quiet_vmstat(void)
 	if (!delayed_work_pending(this_cpu_ptr(&vmstat_work)))
 		return;
 
-	if (!need_update(smp_processor_id()))
+	if (!__this_cpu_read(vmstat_dirty))
 		return;
 
 	/*

mm/vmstat.c +1 -1

 /*

@@ -2021,7 +2005,7 @@ static void vmstat_shepherd(struct work_struct *w)
 	for_each_online_cpu(cpu) {
 		struct delayed_work *dw = &per_cpu(vmstat_work, cpu);
 
-		if (!delayed_work_pending(dw) && need_update(cpu))
+		if (!delayed_work_pending(dw) && per_cpu(vmstat_dirty, cpu))
 			queue_delayed_work_on(cpu, mm_percpu_wq, dw, 0);
 
 		cond_resched();

Signature · 1 line

2.37.1

In the context of the idle task and an adaptive-tick mode/or a nohz_full CPU, quiet_vmstat() can be called: before stopping the idle tick, entering an idle state and on exit. In particular, for the latter case, when the idle task is required to reschedule, the idle tick can remain stopped and the timer expiration time endless i.e., KTIME_MAX. Now, indeed before a nohz_full CPU enters an idle state, CPU-specific vmstat counters should be processed to ensure the respective values have been reset and folded into the zone specific 'vm_stat[]'. That being said, it can only occur when: the idle tick was previously stopped, and reprogramming of the timer is not required.

A customer provided some evidence which indicates that the idle tick was stopped; albeit, CPU-specific vmstat counters still remained populated. Thus one can only assume quiet_vmstat() was not invoked on return to the idle loop.

If I understand correctly, I suspect this divergence might erroneously prevent a reclaim attempt by kswapd. If the number of zone specific free pages are below their per-cpu drift value then zone_page_state_snapshot() is used to compute a more accurate view of the aforementioned statistic. Thus any task blocked on the NUMA node specific pfmemalloc_wait queue will be unable to make significant progress via direct reclaim unless it is killed after being woken up by kswapd (see throttle_direct_reclaim()).

Consider the following theoretical scenario:

1. CPU Y migrated running task A to CPU X that was
in an idle state i.e. waiting for an IRQ - not polling; marked the current task on CPU X to need/or require a reschedule i.e., set TIF_NEED_RESCHED and invoked a reschedule IPI to CPU X (see sched_move_task())

2. CPU X acknowledged the reschedule IPI from CPU Y;
generic idle loop code noticed the TIF_NEED_RESCHED flag against the idle task and attempts to exit of the loop and calls the main scheduler function i.e. __schedule().

Since the idle tick was previously stopped no scheduling-clock tick would occur. So, no deferred timers would be handled

3. Post transition to kernel execution Task A
running on CPU Y, indirectly released a few pages (e.g. see __free_one_page()); CPU Y's 'vm_stat_diff[NR_FREE_PAGES]' was updated and zone specific 'vm_stat[]' update was deferred as per the CPU-specific stat threshold

4. Task A does invoke exit(2) and the kernel does
remove the task from the run-queue; the idle task was selected to execute next since there are no other runnable tasks assigned to the given CPU (see pick_next_task() and pick_next_task_idle())

5. On return to the idle loop since the idle tick
was already stopped and can remain so (see [1] below) e.g. no pending soft IRQs, no attempt is made to zero and fold CPU Y's vmstat counters since reprogramming of the scheduling-clock tick is not required/or needed (see [2])

...
do_idle
{

__current_set_polling()
tick_nohz_idle_enter()

while (!need_resched()) {

local_irq_disable()

...

/* No polling or broadcast event */ cpuidle_idle_call()
{

if (cpuidle_not_available(drv, dev)) {
tick_nohz_idle_stop_tick()
__tick_nohz_idle_stop_tick(this_cpu_ptr(&tick_cpu_sched)) {
int cpu = smp_processor_id()

if (ts->timer_expires_base)
expires = ts->timer_expires
else if (can_stop_idle_tick(cpu, ts))
(1) -------> expires = tick_nohz_next_event(ts, cpu)
else
return

ts->idle_calls++

if (expires > 0LL) {

tick_nohz_stop_tick(ts, cpu) {

if (ts->tick_stopped && (expires == ts->next_tick)) {
(2) -------> if (tick == KTIME_MAX || ts->next_tick ==
hrtimer_get_expires(&ts->sched_timer)) return
}
...
}

So the idea of with this patch is to ensure refresh_cpu_vm_stats(false) is called, when it is appropriate, on return to the idle loop when the idle tick was previously stopped too. Additionally, in the context of nohz_full, when the scheduling-tick is stopped and before exiting to user-mode, ensure no CPU-specific vmstat differentials remain.

Signed-off-by Aaron Tomlin <[email protected]>

---
 include/linux/tick.h     |  5 +++--
 kernel/time/tick-sched.c | 19 ++++++++++++++++++-
 2 files changed, 21 insertions(+), 3 deletions(-)

include/linux/tick.h +3 -2

index bfd571f18cfd..a2bbd6d32e33 100644
@@ -11,7 +11,6 @@
 #include <linux/context_tracking_state.h>
 #include <linux/cpumask.h>
 #include <linux/sched.h>
-#include <linux/rcupdate.h>
 
 #ifdef CONFIG_GENERIC_CLOCKEVENTS
 extern void __init tick_init(void);

@@ -272,6 +271,7 @@ static inline void tick_dep_clear_signal(struct signal_struct *signal,
 
 extern void tick_nohz_full_kick_cpu(int cpu);
 extern void __tick_nohz_task_switch(void);
+void __tick_nohz_user_enter_prepare(void);
 extern void __init tick_nohz_full_setup(cpumask_var_t cpumask);
 #else
 static inline bool tick_nohz_full_enabled(void) { return false; }

@@ -296,6 +296,7 @@ static inline void tick_dep_clear_signal(struct signal_struct *signal,
 
 static inline void tick_nohz_full_kick_cpu(int cpu) { }
 static inline void __tick_nohz_task_switch(void) { }
+static inline void __tick_nohz_user_enter_prepare(void) { }
 static inline void tick_nohz_full_setup(cpumask_var_t cpumask) { }
 #endif
 

@@ -308,7 +309,7 @@ static inline void tick_nohz_task_switch(void)
 static inline void tick_nohz_user_enter_prepare(void)
 {
 	if (tick_nohz_full_cpu(smp_processor_id()))
-		rcu_nocb_flush_deferred_wakeup();
+		__tick_nohz_user_enter_prepare();
 }
 
 #endif

kernel/time/tick-sched.c +18 -1

index 30049580cd62..7f75fc1bba46 100644
@@ -26,6 +26,7 @@
 #include <linux/posix-timers.h>
 #include <linux/context_tracking.h>
 #include <linux/mm.h>
+#include <linux/rcupdate.h>
 
 #include <asm/irq_regs.h>
 

@@ -519,6 +520,20 @@ void __tick_nohz_task_switch(void)
 	}
 }
 
+void __tick_nohz_user_enter_prepare(void)
+{
+	struct tick_sched *ts;
+
+	if (tick_nohz_full_cpu(smp_processor_id())) {
+		ts = this_cpu_ptr(&tick_cpu_sched);
+
+		if (ts->tick_stopped)
+			quiet_vmstat();
+		rcu_nocb_flush_deferred_wakeup();
+	}
+}
+EXPORT_SYMBOL_GPL(__tick_nohz_user_enter_prepare);
+
 /* Get the boot-time nohz CPU list from the kernel parameters. */
 void __init tick_nohz_full_setup(cpumask_var_t cpumask)
 {

@@ -890,6 +905,9 @@ static void tick_nohz_stop_tick(struct tick_sched *ts, int cpu)
 		ts->do_timer_last = 0;
 	}
 
+	/* Attempt to fold when the idle tick is stopped or not */
+	quiet_vmstat();
+
 	/* Skip reprogram of event if its not changed */
 	if (ts->tick_stopped && (expires == ts->next_tick)) {
 		/* Sanity check: make sure clockevent is actually programmed */

@@ -911,7 +929,6 @@ static void tick_nohz_stop_tick(struct tick_sched *ts, int cpu)
 	 */
 	if (!ts->tick_stopped) {
 		calc_load_nohz_start();
-		quiet_vmstat();
 
 		ts->last_tick = hrtimer_get_expires(&ts->sched_timer);
 		ts->tick_stopped = 1;

Signature · 1 line

2.37.1

Aaron,

It is also necessary to remove the delayed_work_pending check from quiet_vmstat:

(quoted diff) +1 -4

@@ -1987,10 +1970,7 @@ void quiet_vmstat(void)
        if (system_state != SYSTEM_RUNNING)
                return;
 
-       if (!delayed_work_pending(this_cpu_ptr(&vmstat_work)))
-               return;
-
-       if (!need_update(smp_processor_id()))
+       if (!__this_cpu_read(vmstat_dirty))
                return;
 
 

Because, consider the following situation:

vmstat_shepherd CPU nohz_full CPU

vmstat_dirty = false enter nohz_full return to userspace
check vmstat_dirty, unset,
delayed work cancelled.
dirties per-CPU vmstats vmstat_dirty = true return to userspace (quiet_vmstat will skip
clearing stats because delayed work was not queued).