[PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Kirill Tkhai]

The race is in unlocked task_rq() access. In pair with parallel
call of sched_setaffinity() it may be a reason of corruption
of internal rq's data.

Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
CC: Juri Lelli <juri.lelli@gmail.com>
CC: Peter Zijlstra <peterz@infradead.org>
CC: Ingo Molnar <mingo@redhat.com>
Cc: <stable@vger.kernel.org> # v3.14
---
 kernel/sched/deadline.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 800e99b..ffb023a 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -513,9 +513,16 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
 						     struct sched_dl_entity,
 						     dl_timer);
 	struct task_struct *p = dl_task_of(dl_se);
-	struct rq *rq = task_rq(p);
+	struct rq *rq;
+again:
+	rq = task_rq(p);
 	raw_spin_lock(&rq->lock);
 
+	if (unlikely(rq != task_rq(p))) {
+		raw_spin_unlock(&rq->lock);
+		goto again;
+	}
+
 	/*
 	 * We need to take care of a possible races here. In fact, the
 	 * task might have changed its scheduling policy to something

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Juri Lelli]

On Sat, 17 May 2014 01:30:03 +0400
Kirill Tkhai <tkhai@yandex.ru> wrote:

> The race is in unlocked task_rq() access. In pair with parallel
> call of sched_setaffinity() it may be a reason of corruption
> of internal rq's data.
> 

Sure, the thing can happen!

> Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
> CC: Juri Lelli <juri.lelli@gmail.com>
> CC: Peter Zijlstra <peterz@infradead.org>
> CC: Ingo Molnar <mingo@redhat.com>
> Cc: <stable@vger.kernel.org> # v3.14
> ---
>  kernel/sched/deadline.c |    9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
> index 800e99b..ffb023a 100644
> --- a/kernel/sched/deadline.c
> +++ b/kernel/sched/deadline.c
> @@ -513,9 +513,16 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
>  						     struct sched_dl_entity,
>  						     dl_timer);
>  	struct task_struct *p = dl_task_of(dl_se);
> -	struct rq *rq = task_rq(p);
> +	struct rq *rq;

We could maybe add a comment here, in line with what we have below, to
document why we need this.

Thanks,

- Juri

> +again:
> +	rq = task_rq(p);
>  	raw_spin_lock(&rq->lock);
>  
> +	if (unlikely(rq != task_rq(p))) {
> +		raw_spin_unlock(&rq->lock);
> +		goto again;
> +	}
> +
>  	/*
>  	 * We need to take care of a possible races here. In fact, the
>  	 * task might have changed its scheduling policy to something
> 

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Kirill Tkhai]

19.05.2014, 17:11, "Juri Lelli" <juri.lelli@gmail.com>:
> On Sat, 17 May 2014 01:30:03 +0400
> Kirill Tkhai <tkhai@yandex.ru> wrote:
>
>>  The race is in unlocked task_rq() access. In pair with parallel
>>  call of sched_setaffinity() it may be a reason of corruption
>>  of internal rq's data.
>
> Sure, the thing can happen!

[snipped]

>>  @@ -513,9 +513,16 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
>>                                                        struct sched_dl_entity,
>>                                                        dl_timer);
>>           struct task_struct *p = dl_task_of(dl_se);
>>  - struct rq *rq = task_rq(p);
>>  + struct rq *rq;
>
> We could maybe add a comment here, in line with what we have below, to
> document why we need this.

How about this? (I added comment and rewrote changelog).

[PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

Throttled task is still on rq, and it may be moved to other cpu
if user is playing with sched_setaffinity(). Therefore, unlocked
task_rq() access makes the race.

To fix that we do the same as made in __task_rq_lock(). We do not
use __task_rq_lock() itself, because it has a useful lockdep check,
which is not correct in case of dl_task_timer(). This case is
an exception.

Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
CC: Juri Lelli <juri.lelli@gmail.com>
CC: Peter Zijlstra <peterz@infradead.org>
CC: Ingo Molnar <mingo@redhat.com>
Cc: <stable@vger.kernel.org> # v3.14

diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 800e99b..c0a6921 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
 						     struct sched_dl_entity,
 						     dl_timer);
 	struct task_struct *p = dl_task_of(dl_se);
-	struct rq *rq = task_rq(p);
+	struct rq *rq;
+again:
+	rq = task_rq(p);
 	raw_spin_lock(&rq->lock);
 
+	if (unlikely(rq != task_rq(p))) {
+		/* Task was moved, retrying. */
+		raw_spin_unlock(&rq->lock);
+		goto again;
+	}
+
 	/*
 	 * We need to take care of a possible races here. In fact, the
 	 * task might have changed its scheduling policy to something

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Peter Zijlstra]

[-- Attachment #1: Type: text/plain, Size: 577 bytes --]

On Mon, May 19, 2014 at 11:31:19PM +0400, Kirill Tkhai wrote:
> @@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
>  						     struct sched_dl_entity,
>  						     dl_timer);
>  	struct task_struct *p = dl_task_of(dl_se);
> -	struct rq *rq = task_rq(p);
> +	struct rq *rq;
> +again:
> +	rq = task_rq(p);
>  	raw_spin_lock(&rq->lock);
>  
> +	if (unlikely(rq != task_rq(p))) {
> +		/* Task was moved, retrying. */
> +		raw_spin_unlock(&rq->lock);
> +		goto again;
> +	}
> +

That thing is called: rq = __task_rq_lock(p);

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Kirill Tkhai]

20.05.2014, 04:00, "Peter Zijlstra" <peterz@infradead.org>:
> On Mon, May 19, 2014 at 11:31:19PM +0400, Kirill Tkhai wrote:
>
>>  @@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
>>                                                        struct sched_dl_entity,
>>                                                        dl_timer);
>>           struct task_struct *p = dl_task_of(dl_se);
>>  - struct rq *rq = task_rq(p);
>>  + struct rq *rq;
>>  +again:
>>  + rq = task_rq(p);
>>           raw_spin_lock(&rq->lock);
>>
>>  + if (unlikely(rq != task_rq(p))) {
>>  + /* Task was moved, retrying. */
>>  + raw_spin_unlock(&rq->lock);
>>  + goto again;
>>  + }
>>  +
>
> That thing is called: rq = __task_rq_lock(p);

But p->pi_lock is not held. The problem is __task_rq_lock() has lockdep assert.
Should we change it?

Kirill

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Kirill Tkhai]

20.05.2014, 09:08, "Kirill Tkhai" <tkhai@yandex.ru>:
> 20.05.2014, 04:00, "Peter Zijlstra" <peterz@infradead.org>:
>
>>  On Mon, May 19, 2014 at 11:31:19PM +0400, Kirill Tkhai wrote:
>>>   @@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
>>>                                                         struct sched_dl_entity,
>>>                                                         dl_timer);
>>>            struct task_struct *p = dl_task_of(dl_se);
>>>   - struct rq *rq = task_rq(p);
>>>   + struct rq *rq;
>>>   +again:
>>>   + rq = task_rq(p);
>>>            raw_spin_lock(&rq->lock);
>>>
>>>   + if (unlikely(rq != task_rq(p))) {
>>>   + /* Task was moved, retrying. */
>>>   + raw_spin_unlock(&rq->lock);
>>>   + goto again;
>>>   + }
>>>   +
>>  That thing is called: rq = __task_rq_lock(p);
>
> But p->pi_lock is not held. The problem is __task_rq_lock() has lockdep assert.
> Should we change it?

Or make something like this?

static inline struct rq *_task_rq_lock(struct task_struct *p)
{
        lockdep_assert_held(&p->pi_lock);

        return __task_rq_lock(p);
}

Thanks!

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Peter Zijlstra]

On Tue, May 20, 2014 at 09:08:53AM +0400, Kirill Tkhai wrote:
> 
> 
> 20.05.2014, 04:00, "Peter Zijlstra" <peterz@infradead.org>:
> > On Mon, May 19, 2014 at 11:31:19PM +0400, Kirill Tkhai wrote:
> >
> >>  @@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
> >>                                                        struct sched_dl_entity,
> >>                                                        dl_timer);
> >>           struct task_struct *p = dl_task_of(dl_se);
> >>  - struct rq *rq = task_rq(p);
> >>  + struct rq *rq;
> >>  +again:
> >>  + rq = task_rq(p);
> >>           raw_spin_lock(&rq->lock);
> >>
> >>  + if (unlikely(rq != task_rq(p))) {
> >>  + /* Task was moved, retrying. */
> >>  + raw_spin_unlock(&rq->lock);
> >>  + goto again;
> >>  + }
> >>  +
> >
> > That thing is called: rq = __task_rq_lock(p);
> 
> But p->pi_lock is not held. The problem is __task_rq_lock() has lockdep assert.
> Should we change it?

Ok, so now that I'm awake ;-)

So the trivial problem as described by your initial changelog isn't
right, because we cannot call sched_setaffinity() on deadline tasks, or
rather we can, but we can't actually change the affinity mask.

Now I suppose the problem can still actually happen when you change the
root domain and trigger a effective affinity change that way.

That said, no leave it as you proposed, adding a *task_rq_lock() variant
without lockdep assert in will only confuse things, as normally we
really should be also taking ->pi_lock.

The only reason we don't strictly need ->pi_lock now is because we're
guaranteed to have p->state == TASK_RUNNING here and are thus free of
ttwu races.

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Juri Lelli]

Hi,

On Tue, 20 May 2014 09:53:15 +0200
Peter Zijlstra <peterz@infradead.org> wrote:

> On Tue, May 20, 2014 at 09:08:53AM +0400, Kirill Tkhai wrote:
> > 
> > 
> > 20.05.2014, 04:00, "Peter Zijlstra" <peterz@infradead.org>:
> > > On Mon, May 19, 2014 at 11:31:19PM +0400, Kirill Tkhai wrote:
> > >
> > >>  @@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
> > >>                                                        struct sched_dl_entity,
> > >>                                                        dl_timer);
> > >>           struct task_struct *p = dl_task_of(dl_se);
> > >>  - struct rq *rq = task_rq(p);
> > >>  + struct rq *rq;
> > >>  +again:
> > >>  + rq = task_rq(p);
> > >>           raw_spin_lock(&rq->lock);
> > >>
> > >>  + if (unlikely(rq != task_rq(p))) {
> > >>  + /* Task was moved, retrying. */
> > >>  + raw_spin_unlock(&rq->lock);
> > >>  + goto again;
> > >>  + }
> > >>  +
> > >
> > > That thing is called: rq = __task_rq_lock(p);
> > 
> > But p->pi_lock is not held. The problem is __task_rq_lock() has lockdep assert.
> > Should we change it?
> 
> Ok, so now that I'm awake ;-)
> 
> So the trivial problem as described by your initial changelog isn't
> right, because we cannot call sched_setaffinity() on deadline tasks, or
> rather we can, but we can't actually change the affinity mask.
> 

Well, if we disable AC we can. And I was able to recreate that race in
that case.

> Now I suppose the problem can still actually happen when you change the
> root domain and trigger a effective affinity change that way.
> 

Yeah, I think here too.

> That said, no leave it as you proposed, adding a *task_rq_lock() variant
> without lockdep assert in will only confuse things, as normally we
> really should be also taking ->pi_lock.
> 
> The only reason we don't strictly need ->pi_lock now is because we're
> guaranteed to have p->state == TASK_RUNNING here and are thus free of
> ttwu races.

Maybe we could add this as part of the comment.

Thanks,

- Juri

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Kirill Tkhai]

20.05.2014, 12:16, "Juri Lelli" <juri.lelli@gmail.com>:
> Hi,
>
> On Tue, 20 May 2014 09:53:15 +0200
> Peter Zijlstra <peterz@infradead.org> wrote:
>
>>  On Tue, May 20, 2014 at 09:08:53AM +0400, Kirill Tkhai wrote:
>>>  20.05.2014, 04:00, "Peter Zijlstra" <peterz@infradead.org>:
>>>>  On Mon, May 19, 2014 at 11:31:19PM +0400, Kirill Tkhai wrote:
>>>>>   @@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
>>>>>                                                         struct sched_dl_entity,
>>>>>                                                         dl_timer);
>>>>>            struct task_struct *p = dl_task_of(dl_se);
>>>>>   - struct rq *rq = task_rq(p);
>>>>>   + struct rq *rq;
>>>>>   +again:
>>>>>   + rq = task_rq(p);
>>>>>            raw_spin_lock(&rq->lock);
>>>>>
>>>>>   + if (unlikely(rq != task_rq(p))) {
>>>>>   + /* Task was moved, retrying. */
>>>>>   + raw_spin_unlock(&rq->lock);
>>>>>   + goto again;
>>>>>   + }
>>>>>   +
>>>>  That thing is called: rq = __task_rq_lock(p);
>>>  But p->pi_lock is not held. The problem is __task_rq_lock() has lockdep assert.
>>>  Should we change it?
>>  Ok, so now that I'm awake ;-)
>>
>>  So the trivial problem as described by your initial changelog isn't
>>  right, because we cannot call sched_setaffinity() on deadline tasks, or
>>  rather we can, but we can't actually change the affinity mask.
>
> Well, if we disable AC we can. And I was able to recreate that race in
> that case.
>
>>  Now I suppose the problem can still actually happen when you change the
>>  root domain and trigger a effective affinity change that way.
>
> Yeah, I think here too.
>
>>  That said, no leave it as you proposed, adding a *task_rq_lock() variant
>>  without lockdep assert in will only confuse things, as normally we
>>  really should be also taking ->pi_lock.
>>
>>  The only reason we don't strictly need ->pi_lock now is because we're
>>  guaranteed to have p->state == TASK_RUNNING here and are thus free of
>>  ttwu races.
>
> Maybe we could add this as part of the comment.

Peter, Juri, thanks for comment. Hope, I understood you right :)

[PATCH] sched/dl: Fix race in dl_task_timer()

Throttled task is still on rq, and it may be moved to other cpu
if user is playing with sched_setaffinity(). Therefore, unlocked
task_rq() access makes the race.

Juri Lelli reports he got this race when dl_bandwidth_enabled()
was not set.

Other thing, pointed by Peter Zijlstra:

   "Now I suppose the problem can still actually happen when
    you change the root domain and trigger a effective affinity
    change that way".

To fix that we do the same as made in __task_rq_lock(). We do not
use __task_rq_lock() itself, because it has a useful lockdep check,
which is not correct in case of dl_task_timer(). We do not need
pi_lock locked here. This case is an exception (PeterZ):

   "The only reason we don't strictly need ->pi_lock now is because
    we're guaranteed to have p->state == TASK_RUNNING here and are
    thus free of ttwu races".

Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
CC: Juri Lelli <juri.lelli@gmail.com>
CC: Peter Zijlstra <peterz@infradead.org>
CC: Ingo Molnar <mingo@redhat.com>
Cc: <stable@vger.kernel.org> # v3.14
---
 kernel/sched/deadline.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 800e99b..14bc348 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
 						     struct sched_dl_entity,
 						     dl_timer);
 	struct task_struct *p = dl_task_of(dl_se);
-	struct rq *rq = task_rq(p);
+	struct rq *rq;
+again:
+	rq = task_rq(p);
 	raw_spin_lock(&rq->lock);
 
+	if (rq != task_rq(p)) {
+		/* Task was moved, retrying. */
+		raw_spin_unlock(&rq->lock);
+		goto again;
+	}
+
 	/*
 	 * We need to take care of a possible races here. In fact, the
 	 * task might have changed its scheduling policy to something

Re: [PATCH] sched/dl: Fix race between dl_task_timer() and sched_setaffinity()

[Peter Zijlstra]

[-- Attachment #1: Type: text/plain, Size: 1272 bytes --]

On Tue, May 20, 2014 at 01:33:42PM +0400, Kirill Tkhai wrote:
> [PATCH] sched/dl: Fix race in dl_task_timer()
> 
> Throttled task is still on rq, and it may be moved to other cpu
> if user is playing with sched_setaffinity(). Therefore, unlocked
> task_rq() access makes the race.
> 
> Juri Lelli reports he got this race when dl_bandwidth_enabled()
> was not set.
> 
> Other thing, pointed by Peter Zijlstra:
> 
>    "Now I suppose the problem can still actually happen when
>     you change the root domain and trigger a effective affinity
>     change that way".
> 
> To fix that we do the same as made in __task_rq_lock(). We do not
> use __task_rq_lock() itself, because it has a useful lockdep check,
> which is not correct in case of dl_task_timer(). We do not need
> pi_lock locked here. This case is an exception (PeterZ):
> 
>    "The only reason we don't strictly need ->pi_lock now is because
>     we're guaranteed to have p->state == TASK_RUNNING here and are
>     thus free of ttwu races".
> 
> Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
> CC: Juri Lelli <juri.lelli@gmail.com>
> CC: Peter Zijlstra <peterz@infradead.org>
> CC: Ingo Molnar <mingo@redhat.com>
> Cc: <stable@vger.kernel.org> # v3.14
> ---

thanks Kirill!

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

[tip:sched/urgent] sched/dl: Fix race in dl_task_timer()

[tip-bot for Kirill Tkhai]

Commit-ID:  0f397f2c90ce68821ee864c2c53baafe78de765d
Gitweb:     http://git.kernel.org/tip/0f397f2c90ce68821ee864c2c53baafe78de765d
Author:     Kirill Tkhai <tkhai@yandex.ru>
AuthorDate: Tue, 20 May 2014 13:33:42 +0400
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Thu, 5 Jun 2014 11:51:12 +0200

sched/dl: Fix race in dl_task_timer()

Throttled task is still on rq, and it may be moved to other cpu
if user is playing with sched_setaffinity(). Therefore, unlocked
task_rq() access makes the race.

Juri Lelli reports he got this race when dl_bandwidth_enabled()
was not set.

Other thing, pointed by Peter Zijlstra:

   "Now I suppose the problem can still actually happen when
    you change the root domain and trigger a effective affinity
    change that way".

To fix that we do the same as made in __task_rq_lock(). We do not
use __task_rq_lock() itself, because it has a useful lockdep check,
which is not correct in case of dl_task_timer(). We do not need
pi_lock locked here. This case is an exception (PeterZ):

   "The only reason we don't strictly need ->pi_lock now is because
    we're guaranteed to have p->state == TASK_RUNNING here and are
    thus free of ttwu races".

Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: <stable@vger.kernel.org> # v3.14+
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/3056991400578422@web14g.yandex.ru
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/deadline.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 800e99b..14bc348 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -513,9 +513,17 @@ static enum hrtimer_restart dl_task_timer(struct hrtimer *timer)
 						     struct sched_dl_entity,
 						     dl_timer);
 	struct task_struct *p = dl_task_of(dl_se);
-	struct rq *rq = task_rq(p);
+	struct rq *rq;
+again:
+	rq = task_rq(p);
 	raw_spin_lock(&rq->lock);
 
+	if (rq != task_rq(p)) {
+		/* Task was moved, retrying. */
+		raw_spin_unlock(&rq->lock);
+		goto again;
+	}
+
 	/*
 	 * We need to take care of a possible races here. In fact, the
 	 * task might have changed its scheduling policy to something