Re: [PATCH v2 1/3] sched/fair: Disable runtime_enabled on dying rq

[Kirill Tkhai <ktkhai@parallels.com>]

В Вт, 24/06/2014 в 23:26 +0400, Kirill Tkhai пишет:
> On 24.06.2014 23:13, bsegall@google.com wrote:
> > Kirill Tkhai <tkhai@yandex.ru> writes:
> > 
> >> On 24.06.2014 21:03, bsegall@google.com wrote:
> >>> Kirill Tkhai <ktkhai@parallels.com> writes:
> >>>
> >>>> We kill rq->rd on the CPU_DOWN_PREPARE stage:
> >>>>
> >>>> 	cpuset_cpu_inactive -> cpuset_update_active_cpus -> partition_sched_domains ->
> >>>> 	-> cpu_attach_domain -> rq_attach_root -> set_rq_offline
> >>>>
> >>>> This unthrottles all throttled cfs_rqs.
> >>>>
> >>>> But the cpu is still able to call schedule() till
> >>>>
> >>>> 	take_cpu_down->__cpu_disable()
> >>>>
> >>>> is called from stop_machine.
> >>>>
> >>>> This case the tasks from just unthrottled cfs_rqs are pickable
> >>>> in a standard scheduler way, and they are picked by dying cpu.
> >>>> The cfs_rqs becomes throttled again, and migrate_tasks()
> >>>> in migration_call skips their tasks (one more unthrottle
> >>>> in migrate_tasks()->CPU_DYING does not happen, because rq->rd
> >>>> is already NULL).
> >>>>
> >>>> Patch sets runtime_enabled to zero. This guarantees, the runtime
> >>>> is not accounted, and the cfs_rqs won't exceed given
> >>>> cfs_rq->runtime_remaining = 1, and tasks will be pickable
> >>>> in migrate_tasks(). runtime_enabled is recalculated again
> >>>> when rq becomes online again.
> >>>>
> >>>> Ben Segall also noticed, we always enable runtime in
> >>>> tg_set_cfs_bandwidth(). Actually, we should do that for online
> >>>> cpus only. To fix that, we check if a cpu is online when
> >>>> its rq is locked. This guarantees we do not have races with
> >>>> set_rq_offline(), which also requires rq->lock.
> >>>>
> >>>> v2: Fix race with tg_set_cfs_bandwidth().
> >>>>     Move cfs_rq->runtime_enabled=0 above unthrottle_cfs_rq().
> >>>>
> >>>> Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
> >>>> CC: Konstantin Khorenko <khorenko@parallels.com>
> >>>> CC: Ben Segall <bsegall@google.com>
> >>>> CC: Paul Turner <pjt@google.com>
> >>>> CC: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
> >>>> CC: Mike Galbraith <umgwanakikbuti@gmail.com>
> >>>> CC: Peter Zijlstra <peterz@infradead.org>
> >>>> CC: Ingo Molnar <mingo@kernel.org>
> >>>> ---
> >>>>  kernel/sched/core.c |   15 +++++++++++----
> >>>>  kernel/sched/fair.c |   22 ++++++++++++++++++++++
> >>>>  2 files changed, 33 insertions(+), 4 deletions(-)
> >>>>
> >>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> >>>> index 7f3063c..707a3c5 100644
> >>>> --- a/kernel/sched/core.c
> >>>> +++ b/kernel/sched/core.c
> >>>> @@ -7842,11 +7842,18 @@ static int tg_set_cfs_bandwidth(struct task_group *tg, u64 period, u64 quota)
> >>>>  		struct rq *rq = cfs_rq->rq;
> >>>>  
> >>>>  		raw_spin_lock_irq(&rq->lock);
> >>>> -		cfs_rq->runtime_enabled = runtime_enabled;
> >>>> -		cfs_rq->runtime_remaining = 0;
> >>>> +		/*
> >>>> +		 * Do not enable runtime on offline runqueues. We specially
> >>>> +		 * make it disabled in unthrottle_offline_cfs_rqs().
> >>>> +		 */
> >>>> +		if (cpu_online(i)) {
> >>>> +			cfs_rq->runtime_enabled = runtime_enabled;
> >>>> +			cfs_rq->runtime_remaining = 0;
> >>>> +
> >>>> +			if (cfs_rq->throttled)
> >>>> +				unthrottle_cfs_rq(cfs_rq);
> >>>> +		}
> >>>
> >>> We can just do for_each_online_cpu, yes? Also we probably need
> >>> get_online_cpus/put_online_cpus, and/or want cpu_active_mask instead
> >>> right?
> >>>
> >>
> >> Yes, we can use for_each_online_cpu/for_each_active_cpu with
> >> get_online_cpus() taken. But it adds one more lock dependence.
> >> This looks worse for me.
> > 
> > I mean, you need get_online_cpus anyway - cpu_online is just a test
> > against the same mask that for_each_online_cpu uses, and without taking
> > the lock you can still race with offlining and reset runtime_enabled.
> > 
> 
> Oh, I see. Thanks.

But we can check for rq->online, don't we? How about this?

    sched/fair: Disable runtime_enabled on dying rq
    
    We kill rq->rd on the CPU_DOWN_PREPARE stage:
    
    	cpuset_cpu_inactive -> cpuset_update_active_cpus -> partition_sched_domains ->
    	-> cpu_attach_domain -> rq_attach_root -> set_rq_offline
    
    This unthrottles all throttled cfs_rqs.
    
    But the cpu is still able to call schedule() till
    
    	take_cpu_down->__cpu_disable()
    
    is called from stop_machine.
    
    This case the tasks from just unthrottled cfs_rqs are pickable
    in a standard scheduler way, and they are picked by dying cpu.
    The cfs_rqs becomes throttled again, and migrate_tasks()
    in migration_call skips their tasks (one more unthrottle
    in migrate_tasks()->CPU_DYING does not happen, because rq->rd
    is already NULL).
    
    Patch sets runtime_enabled to zero. This guarantees, the runtime
    is not accounted, and the cfs_rqs won't exceed given
    cfs_rq->runtime_remaining = 1, and tasks will be pickable
    in migrate_tasks(). runtime_enabled is recalculated again
    when rq becomes online again.
    
    Ben Segall also noticed, we always enable runtime in
    tg_set_cfs_bandwidth(). Actually, we should do that for online
    cpus only. To fix that, we check if a cpu is online when
    its rq is locked. This guarantees we do not have races with
    set_rq_offline(), which also requires rq->lock.
    
    v2: Fix race with tg_set_cfs_bandwidth().
        Move cfs_rq->runtime_enabled=0 above unthrottle_cfs_rq().
    v3: Check for rq->online instead of cpu_active.
    
    Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index ceea8d0..4c163c9 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7837,11 +7837,17 @@ static int tg_set_cfs_bandwidth(struct task_group *tg, u64 period, u64 quota)
 		struct rq *rq = cfs_rq->rq;
 
 		raw_spin_lock_irq(&rq->lock);
-		cfs_rq->runtime_enabled = runtime_enabled;
-		cfs_rq->runtime_remaining = 0;
+		/*
+		 * Do not enable runtime on offline runqueues. We specially
+		 * disable it in unthrottle_offline_cfs_rqs().
+		 */
+		if (rq->online) {
+			cfs_rq->runtime_enabled = runtime_enabled;
+			cfs_rq->runtime_remaining = 0;
 
-		if (cfs_rq->throttled)
-			unthrottle_cfs_rq(cfs_rq);
+			if (cfs_rq->throttled)
+				unthrottle_cfs_rq(cfs_rq);
+		}
 		raw_spin_unlock_irq(&rq->lock);
 	}
 	if (runtime_was_enabled && !runtime_enabled)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 1f9c457..5616d23 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3776,6 +3776,19 @@ static void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
 	hrtimer_cancel(&cfs_b->slack_timer);
 }
 
+static void __maybe_unused update_runtime_enabled(struct rq *rq)
+{
+	struct cfs_rq *cfs_rq;
+
+	for_each_leaf_cfs_rq(rq, cfs_rq) {
+		struct cfs_bandwidth *cfs_b = &cfs_rq->tg->cfs_bandwidth;
+
+		raw_spin_lock(&cfs_b->lock);
+		cfs_rq->runtime_enabled = cfs_b->quota != RUNTIME_INF;
+		raw_spin_unlock(&cfs_b->lock);
+	}
+}
+
 static void __maybe_unused unthrottle_offline_cfs_rqs(struct rq *rq)
 {
 	struct cfs_rq *cfs_rq;
@@ -3789,6 +3802,12 @@ static void __maybe_unused unthrottle_offline_cfs_rqs(struct rq *rq)
 		 * there's some valid quota amount
 		 */
 		cfs_rq->runtime_remaining = 1;
+		/*
+		 * Offline rq is schedulable till cpu is completely disabled
+		 * in take_cpu_down(), so we prevent new cfs throttling here.
+		 */
+		cfs_rq->runtime_enabled = 0;
+
 		if (cfs_rq_throttled(cfs_rq))
 			unthrottle_cfs_rq(cfs_rq);
 	}
@@ -3832,6 +3851,7 @@ static inline struct cfs_bandwidth *tg_cfs_bandwidth(struct task_group *tg)
 	return NULL;
 }
 static inline void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
+static inline void update_runtime_enabled(struct rq *rq) {}
 static inline void unthrottle_offline_cfs_rqs(struct rq *rq) {}
 
 #endif /* CONFIG_CFS_BANDWIDTH */
@@ -7325,6 +7345,8 @@ void trigger_load_balance(struct rq *rq)
 static void rq_online_fair(struct rq *rq)
 {
 	update_sysctl();
+
+	update_runtime_enabled(rq);
 }
 
 static void rq_offline_fair(struct rq *rq)