[PATCH] sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

[PATCH] sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

[Kirill Tkhai]

This patch touches RT group scheduling case.

Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,
while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because
changing of priority on a child level does not guarantee that the priority is
the highest all over the rq. So, this leak makes RT balancing unusable.

The short example: the task having the highest priority among all rq's RT tasks
(no one other task has the same priority) are waking on a throttle rt_rq.
The rq's cpupri is set to the task's priority equivalent, but real
rq->rt.highest_prio.curr is less.

The patch below fixes the problem.

It looks like all version have this bug, so I CC'ed stable mailing list.

Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
CC: Ingo Molnar <mingo@redhat.com>
CC: Peter Zijlstra <peterz@infradead.org>
CC: Steven Rostedt <rostedt@goodmis.org>
CC: stable@vger.kernel.org
---
 kernel/sched/rt.c |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 7d57275..1c40655 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -901,6 +901,13 @@ inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
 {
 	struct rq *rq = rq_of_rt_rq(rt_rq);
 
+#ifdef CONFIG_RT_GROUP_SCHED
+	/*
+	 * Change rq's cpupri only if rt_rq is the top queue.
+	 */
+	if (&rq->rt != rt_rq)
+		return;
+#endif
 	if (rq->online && prio < prev_prio)
 		cpupri_set(&rq->rd->cpupri, rq->cpu, prio);
 }
@@ -910,6 +917,13 @@ dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
 {
 	struct rq *rq = rq_of_rt_rq(rt_rq);
 
+#ifdef CONFIG_RT_GROUP_SCHED
+	/*
+	 * Change rq's cpupri only if rt_rq is the top queue.
+	 */
+	if (&rq->rt != rt_rq)
+		return;
+#endif
 	if (rq->online && rt_rq->highest_prio.curr != prev_prio)
 		cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr);
 }

Re: [PATCH] sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

[Kirill Tkhai]

Ping!

27.11.2013, 19:59, "Kirill Tkhai" <tkhai@yandex.ru>:
> This patch touches RT group scheduling case.
>
> Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,
> while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because
> changing of priority on a child level does not guarantee that the priority is
> the highest all over the rq. So, this leak makes RT balancing unusable.
>
> The short example: the task having the highest priority among all rq's RT tasks
> (no one other task has the same priority) are waking on a throttle rt_rq.
> The rq's cpupri is set to the task's priority equivalent, but real
> rq->rt.highest_prio.curr is less.
>
> The patch below fixes the problem.
>
> It looks like all version have this bug, so I CC'ed stable mailing list.
>
> Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
> CC: Ingo Molnar <mingo@redhat.com>
> CC: Peter Zijlstra <peterz@infradead.org>
> CC: Steven Rostedt <rostedt@goodmis.org>
> CC: stable@vger.kernel.org
>---
> kernel/sched/rt.c |   14 ++++++++++++++
> 1 files changed, 14 insertions(+), 0 deletions(-)
>diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
>index 7d57275..1c40655 100644
>--- a/kernel/sched/rt.c
>+++ b/kernel/sched/rt.c
>@@ -901,6 +901,13 @@ inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
> {
> 	struct rq *rq = rq_of_rt_rq(rt_rq);
> 
>+#ifdef CONFIG_RT_GROUP_SCHED
>+	/*
>+	 * Change rq's cpupri only if rt_rq is the top queue.
>+	 */
>+	if (&rq->rt != rt_rq)
>+		return;
>+#endif
> 	if (rq->online && prio < prev_prio)
> 		cpupri_set(&rq->rd->cpupri, rq->cpu, prio);
> }
>@@ -910,6 +917,13 @@ dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
> {
> 	struct rq *rq = rq_of_rt_rq(rt_rq);
> 
>+#ifdef CONFIG_RT_GROUP_SCHED
>+	/*
>+	 * Change rq's cpupri only if rt_rq is the top queue.
>+	 */
>+	if (&rq->rt != rt_rq)
>+		return;
>+#endif
> 	if (rq->online && rt_rq->highest_prio.curr != prev_prio)
> 		cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr);
> }

Re: [PATCH] sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

[Peter Zijlstra]

On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote:
> This patch touches RT group scheduling case.
> 
> Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,
> while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because
> changing of priority on a child level does not guarantee that the priority is
> the highest all over the rq. So, this leak makes RT balancing unusable.
> 
> The short example: the task having the highest priority among all rq's RT tasks
> (no one other task has the same priority) are waking on a throttle rt_rq.
> The rq's cpupri is set to the task's priority equivalent, but real
> rq->rt.highest_prio.curr is less.
> 
> The patch below fixes the problem.
> 
> It looks like all version have this bug, so I CC'ed stable mailing list.

Yeah, I think this is right.

cpupri stuff should indeed only be changed for the top level group.

Re: [PATCH] sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

[Kirill Tkhai]

13.12.2013, 19:42, "Peter Zijlstra" <peterz@infradead.org>:
> On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote:
>
>>  This patch touches RT group scheduling case.
>>
>>  Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,
>>  while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because
>>  changing of priority on a child level does not guarantee that the priority is
>>  the highest all over the rq. So, this leak makes RT balancing unusable.
>>
>>  The short example: the task having the highest priority among all rq's RT tasks
>>  (no one other task has the same priority) are waking on a throttle rt_rq.
>>  The rq's cpupri is set to the task's priority equivalent, but real
>>  rq->rt.highest_prio.curr is less.
>>
>>  The patch below fixes the problem.
>>
>>  It looks like all version have this bug, so I CC'ed stable mailing list.
>
> Yeah, I think this is right.
>
> cpupri stuff should indeed only be changed for the top level group.

Ingo, are you going to apply this patch? Or will you give any comments?

Re: [PATCH] sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

[Peter Zijlstra]

On Tue, Dec 17, 2013 at 04:02:58PM +0400, Kirill Tkhai wrote:
> 
> 
> 13.12.2013, 19:42, "Peter Zijlstra" <peterz@infradead.org>:
> > On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote:
> >
> >>  This patch touches RT group scheduling case.
> >>
> >>  Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,
> >>  while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because
> >>  changing of priority on a child level does not guarantee that the priority is
> >>  the highest all over the rq. So, this leak makes RT balancing unusable.
> >>
> >>  The short example: the task having the highest priority among all rq's RT tasks
> >>  (no one other task has the same priority) are waking on a throttle rt_rq.
> >>  The rq's cpupri is set to the task's priority equivalent, but real
> >>  rq->rt.highest_prio.curr is less.
> >>
> >>  The patch below fixes the problem.
> >>
> >>  It looks like all version have this bug, so I CC'ed stable mailing list.
> >
> > Yeah, I think this is right.
> >
> > cpupri stuff should indeed only be changed for the top level group.
> 
> Ingo, are you going to apply this patch? Or will you give any comments?

I queued it, Ingo should get it through me somewhere today if all things
go well.

Thanks

Re: [PATCH] sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

[Kirill Tkhai]

17.12.2013, 16:47, "Peter Zijlstra" <peterz@infradead.org>:
> On Tue, Dec 17, 2013 at 04:02:58PM +0400, Kirill Tkhai wrote:
>
>>  13.12.2013, 19:42, "Peter Zijlstra" <peterz@infradead.org>:
>>>  On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote:
>>>>   This patch touches RT group scheduling case.
>>>>
>>>>   Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,
>>>>   while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because
>>>>   changing of priority on a child level does not guarantee that the priority is
>>>>   the highest all over the rq. So, this leak makes RT balancing unusable.
>>>>
>>>>   The short example: the task having the highest priority among all rq's RT tasks
>>>>   (no one other task has the same priority) are waking on a throttle rt_rq.
>>>>   The rq's cpupri is set to the task's priority equivalent, but real
>>>>   rq->rt.highest_prio.curr is less.
>>>>
>>>>   The patch below fixes the problem.
>>>>
>>>>   It looks like all version have this bug, so I CC'ed stable mailing list.
>>>  Yeah, I think this is right.
>>>
>>>  cpupri stuff should indeed only be changed for the top level group.
>>  Ingo, are you going to apply this patch? Or will you give any comments?
>
> I queued it, Ingo should get it through me somewhere today if all things
> go well.
>
> Thanks

Thanks, Peter

[tip:sched/core] sched/rt: Fix rq's cpupri leak while enqueue/ dequeue child RT entities

[tip-bot for Kirill Tkhai]

Commit-ID:  757dfcaa41844595964f1220f1d33182dae49976
Gitweb:     http://git.kernel.org/tip/757dfcaa41844595964f1220f1d33182dae49976
Author:     Kirill Tkhai <tkhai@yandex.ru>
AuthorDate: Wed, 27 Nov 2013 19:59:13 +0400
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Tue, 17 Dec 2013 15:08:44 +0100

sched/rt: Fix rq's cpupri leak while enqueue/dequeue child RT entities

This patch touches the RT group scheduling case.

Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's
priority, while rt_rq passed to them may be not the top-level rt_rq.
This is wrong, because changing of priority on a child level does not
guarantee that the priority is the highest all over the rq. So, this
leak makes RT balancing unusable.

The short example: the task having the highest priority among all rq's
RT tasks (no one other task has the same priority) are waking on a
throttle rt_rq.  The rq's cpupri is set to the task's priority
equivalent, but real rq->rt.highest_prio.curr is less.

The patch below fixes the problem.

Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
CC: Steven Rostedt <rostedt@goodmis.org>
CC: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/49231385567953@web4m.yandex.ru
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/rt.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 7d57275..1c40655 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -901,6 +901,13 @@ inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
 {
 	struct rq *rq = rq_of_rt_rq(rt_rq);
 
+#ifdef CONFIG_RT_GROUP_SCHED
+	/*
+	 * Change rq's cpupri only if rt_rq is the top queue.
+	 */
+	if (&rq->rt != rt_rq)
+		return;
+#endif
 	if (rq->online && prio < prev_prio)
 		cpupri_set(&rq->rd->cpupri, rq->cpu, prio);
 }
@@ -910,6 +917,13 @@ dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
 {
 	struct rq *rq = rq_of_rt_rq(rt_rq);
 
+#ifdef CONFIG_RT_GROUP_SCHED
+	/*
+	 * Change rq's cpupri only if rt_rq is the top queue.
+	 */
+	if (&rq->rt != rt_rq)
+		return;
+#endif
 	if (rq->online && rt_rq->highest_prio.curr != prev_prio)
 		cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr);
 }