timers/nohz: Protect idle/iowait sleep time under seqcount

Reading idle/IO sleep time (eg: from /proc/stat) can race with idle exit
updates because the state machine handling the stats is not atomic and
requires a coherent read batch.

As a result reading the sleep time may report irrelevant or backward
values.

Fix this with protecting the simple state machine within a seqcount.
This is expected to be cheap enough not to add measurable performance
impact on the idle path.

Note this only fixes reader VS writer condition partitially. A race
remains that involves remote updates of the CPU iowait task counter. It
can hardly be fixed.

Reported-by: Yu Liao <liaoyu15@huawei.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20230222144649.624380-4-frederic@kernel.org

Author: Frederic Weisbecker

kernel/time/tick-sched.c

@@ -646,28 +646,34 @@ static void tick_nohz_stop_idle(struct tick_sched *ts, ktime_t now)
 
 	delta = ktime_sub(now, ts->idle_entrytime);
 
+	write_seqcount_begin(&ts->idle_sleeptime_seq);
 	if (nr_iowait_cpu(smp_processor_id()) > 0)
 		ts->iowait_sleeptime = ktime_add(ts->iowait_sleeptime, delta);
 	else
 		ts->idle_sleeptime = ktime_add(ts->idle_sleeptime, delta);
 
 	ts->idle_entrytime = now;
 	ts->idle_active = 0;
+	write_seqcount_end(&ts->idle_sleeptime_seq);
 
 	sched_clock_idle_wakeup_event();
 }
 
 static void tick_nohz_start_idle(struct tick_sched *ts)
 {
+	write_seqcount_begin(&ts->idle_sleeptime_seq);
 	ts->idle_entrytime = ktime_get();
 	ts->idle_active = 1;
+	write_seqcount_end(&ts->idle_sleeptime_seq);
+
 	sched_clock_idle_sleep_event();
 }
 
 static u64 get_cpu_sleep_time_us(struct tick_sched *ts, ktime_t *sleeptime,
 				 bool compute_delta, u64 *last_update_time)
 {
 	ktime_t now, idle;
+	unsigned int seq;
 
 	if (!tick_nohz_active)
 		return -1;
@@ -676,13 +682,17 @@ static u64 get_cpu_sleep_time_us(struct tick_sched *ts, ktime_t *sleeptime,
 	if (last_update_time)
 		*last_update_time = ktime_to_us(now);
 
-	if (ts->idle_active && compute_delta) {
-		ktime_t delta = ktime_sub(now, ts->idle_entrytime);
+	do {
+		seq = read_seqcount_begin(&ts->idle_sleeptime_seq);
 
-		idle = ktime_add(*sleeptime, delta);
-	} else {
-		idle = *sleeptime;
-	}
+		if (ts->idle_active && compute_delta) {
+			ktime_t delta = ktime_sub(now, ts->idle_entrytime);
+
+			idle = ktime_add(*sleeptime, delta);
+		} else {
+			idle = *sleeptime;
+		}
+	} while (read_seqcount_retry(&ts->idle_sleeptime_seq, seq));
 
 	return ktime_to_us(idle);
 

kernel/time/tick-sched.h

@@ -75,6 +75,7 @@ struct tick_sched {
 	ktime_t				idle_waketime;
 
 	/* Idle entry */
+	seqcount_t			idle_sleeptime_seq;
 	ktime_t				idle_entrytime;
 
 	/* Tick stop */