* Re: [PATCH v2] sched/fair: WARN and refuse to set buddy when !se->on_rq
2017-05-10 20:11 [PATCH v2] sched/fair: WARN and refuse to set buddy when !se->on_rq Daniel Axtens
@ 2017-06-19 14:05 ` Daniel Axtens
2017-06-20 7:41 ` Konstantin Khlebnikov
2017-06-20 10:27 ` [PATCH] sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well Ingo Molnar
2017-06-20 13:26 ` [tip:sched/core] sched/fair: WARN() and refuse to set buddy when !se->on_rq tip-bot for Daniel Axtens
2 siblings, 1 reply; 6+ messages in thread
From: Daniel Axtens @ 2017-06-19 14:05 UTC (permalink / raw)
To: linux-kernel, peterz, mingo, Konstantin Khlebnikov
Cc: Ben Segall, Thomas Gleixner
Hi Konstantin and Peter,
Just checking if this version was OK with you - I hadn't heard anything
and I noticed it's not in -next so I just wanted to check to see if
there were any other changes you wanted.
Regards,
Daniel
Daniel Axtens <dja@axtens.net> writes:
> If we set a next or last buddy for a se that is not on_rq, we will
> end up taking a NULL pointer dereference in wakeup_preempt_entity
> via pick_next_task_fair.
>
> Detect when we would be about to do that, throw a warning and
> then refuse to actually set it.
>
> This has been suggested at least twice[0][1]: just do it.
>
> [0] https://marc.info/?l=linux-kernel&m=146651668921468&w=2
> [1] https://lkml.org/lkml/2016/6/16/663
>
> Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
> Cc: Ben Segall <bsegall@google.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Signed-off-by: Daniel Axtens <dja@axtens.net>
>
> ---
>
> I recently had to debug a problem with these (we hadn't backported
> Konstantin's patches in this area) and this would have saved a lot
> of time/pain.
>
> v2: use SCHED_WARN_ON to restrict when the test is run. This is a
> macro for WARN_ON_ONCE, which is convenient.
> ---
> kernel/sched/fair.c | 10 ++++++++--
> 1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index d71109321841..44b94cfe02cb 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -6168,8 +6168,11 @@ static void set_last_buddy(struct sched_entity *se)
> if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
> return;
>
> - for_each_sched_entity(se)
> + for_each_sched_entity(se) {
> + if (SCHED_WARN_ON(!se->on_rq))
> + return;
> cfs_rq_of(se)->last = se;
> + }
> }
>
> static void set_next_buddy(struct sched_entity *se)
> @@ -6177,8 +6180,11 @@ static void set_next_buddy(struct sched_entity *se)
> if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
> return;
>
> - for_each_sched_entity(se)
> + for_each_sched_entity(se) {
> + if (SCHED_WARN_ON(!se->on_rq))
> + return;
> cfs_rq_of(se)->next = se;
> + }
> }
>
> static void set_skip_buddy(struct sched_entity *se)
> --
> 2.11.0
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH v2] sched/fair: WARN and refuse to set buddy when !se->on_rq
2017-06-19 14:05 ` Daniel Axtens
@ 2017-06-20 7:41 ` Konstantin Khlebnikov
0 siblings, 0 replies; 6+ messages in thread
From: Konstantin Khlebnikov @ 2017-06-20 7:41 UTC (permalink / raw)
To: Daniel Axtens, linux-kernel, peterz, mingo; +Cc: Ben Segall, Thomas Gleixner
On 19.06.2017 17:05, Daniel Axtens wrote:
> Hi Konstantin and Peter,
>
> Just checking if this version was OK with you - I hadn't heard anything
> and I noticed it's not in -next so I just wanted to check to see if
> there were any other changes you wanted.
Looks good for me. Exactly that debug helped me alot at that time.
>
> Regards,
> Daniel
>
>
> Daniel Axtens <dja@axtens.net> writes:
>
>> If we set a next or last buddy for a se that is not on_rq, we will
>> end up taking a NULL pointer dereference in wakeup_preempt_entity
>> via pick_next_task_fair.
>>
>> Detect when we would be about to do that, throw a warning and
>> then refuse to actually set it.
>>
>> This has been suggested at least twice[0][1]: just do it.
>>
>> [0] https://marc.info/?l=linux-kernel&m=146651668921468&w=2
>> [1] https://lkml.org/lkml/2016/6/16/663
>>
>> Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
>> Cc: Ben Segall <bsegall@google.com>
>> Cc: Peter Zijlstra <peterz@infradead.org>
>> Cc: Thomas Gleixner <tglx@linutronix.de>
>> Signed-off-by: Daniel Axtens <dja@axtens.net>
>>
>> ---
>>
>> I recently had to debug a problem with these (we hadn't backported
>> Konstantin's patches in this area) and this would have saved a lot
>> of time/pain.
>>
>> v2: use SCHED_WARN_ON to restrict when the test is run. This is a
>> macro for WARN_ON_ONCE, which is convenient.
>> ---
>> kernel/sched/fair.c | 10 ++++++++--
>> 1 file changed, 8 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index d71109321841..44b94cfe02cb 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -6168,8 +6168,11 @@ static void set_last_buddy(struct sched_entity *se)
>> if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
>> return;
>>
>> - for_each_sched_entity(se)
>> + for_each_sched_entity(se) {
>> + if (SCHED_WARN_ON(!se->on_rq))
>> + return;
>> cfs_rq_of(se)->last = se;
>> + }
>> }
>>
>> static void set_next_buddy(struct sched_entity *se)
>> @@ -6177,8 +6180,11 @@ static void set_next_buddy(struct sched_entity *se)
>> if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
>> return;
>>
>> - for_each_sched_entity(se)
>> + for_each_sched_entity(se) {
>> + if (SCHED_WARN_ON(!se->on_rq))
>> + return;
>> cfs_rq_of(se)->next = se;
>> + }
>> }
>>
>> static void set_skip_buddy(struct sched_entity *se)
>> --
>> 2.11.0
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH] sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well
2017-05-10 20:11 [PATCH v2] sched/fair: WARN and refuse to set buddy when !se->on_rq Daniel Axtens
2017-06-19 14:05 ` Daniel Axtens
@ 2017-06-20 10:27 ` Ingo Molnar
2017-06-20 13:52 ` Daniel Axtens
2017-06-20 13:26 ` [tip:sched/core] sched/fair: WARN() and refuse to set buddy when !se->on_rq tip-bot for Daniel Axtens
2 siblings, 1 reply; 6+ messages in thread
From: Ingo Molnar @ 2017-06-20 10:27 UTC (permalink / raw)
To: Daniel Axtens
Cc: linux-kernel, peterz, mingo, Konstantin Khlebnikov, Ben Segall,
Thomas Gleixner
* Daniel Axtens <dja@axtens.net> wrote:
> If we set a next or last buddy for a se that is not on_rq, we will
> end up taking a NULL pointer dereference in wakeup_preempt_entity
> via pick_next_task_fair.
>
> Detect when we would be about to do that, throw a warning and
> then refuse to actually set it.
>
> This has been suggested at least twice[0][1]: just do it.
>
> [0] https://marc.info/?l=linux-kernel&m=146651668921468&w=2
> [1] https://lkml.org/lkml/2016/6/16/663
>
> Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
> Cc: Ben Segall <bsegall@google.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Signed-off-by: Daniel Axtens <dja@axtens.net>
>
> ---
>
> I recently had to debug a problem with these (we hadn't backported
> Konstantin's patches in this area) and this would have saved a lot
> of time/pain.
>
> v2: use SCHED_WARN_ON to restrict when the test is run. This is a
> macro for WARN_ON_ONCE, which is convenient.
> ---
> kernel/sched/fair.c | 10 ++++++++--
> 1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index d71109321841..44b94cfe02cb 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -6168,8 +6168,11 @@ static void set_last_buddy(struct sched_entity *se)
> if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
> return;
>
> - for_each_sched_entity(se)
> + for_each_sched_entity(se) {
> + if (SCHED_WARN_ON(!se->on_rq))
> + return;
> cfs_rq_of(se)->last = se;
> + }
This won't build in the !CONFIG_SCHED_DEBUG case, because of the naive definition
in sched.h:
#define SCHED_WARN_ON(x) ((void)(x))
That should be changed to something like:
#define SCHED_WARN_ON(x) ((void)(x))
I've applied the fix below. (untested at the moment)
Thanks,
Ingo
========================>
>From 6d3aed3d8a0573d0a6eb1160ccd0a0713f4dbc2f Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 20 Jun 2017 12:24:42 +0200
Subject: [PATCH] sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well
This definition of SCHED_WARN_ON():
#define SCHED_WARN_ON(x) ((void)(x))
is not fully compatible with the 'real' WARN_ON_ONCE() primitive, as it
has no return value, so it cannot be used in conditionals.
Fix it.
Cc: Daniel Axtens <dja@axtens.net>
Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
kernel/sched/sched.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index f2ef759a4cb6..e0329d10bdb8 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -39,9 +39,9 @@
#include "cpuacct.h"
#ifdef CONFIG_SCHED_DEBUG
-#define SCHED_WARN_ON(x) WARN_ONCE(x, #x)
+# define SCHED_WARN_ON(x) WARN_ONCE(x, #x)
#else
-#define SCHED_WARN_ON(x) ((void)(x))
+# define SCHED_WARN_ON(x) ({ (void)(x), 0; })
#endif
struct rq;
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH] sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well
2017-06-20 10:27 ` [PATCH] sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well Ingo Molnar
@ 2017-06-20 13:52 ` Daniel Axtens
0 siblings, 0 replies; 6+ messages in thread
From: Daniel Axtens @ 2017-06-20 13:52 UTC (permalink / raw)
To: Ingo Molnar
Cc: linux-kernel, peterz, mingo, Konstantin Khlebnikov, Ben Segall,
Thomas Gleixner
Hi Ingo,
Good catch - thanks for picking that up and taking the patch!
Regards,
Daniel
> * Daniel Axtens <dja@axtens.net> wrote:
>
>> If we set a next or last buddy for a se that is not on_rq, we will
>> end up taking a NULL pointer dereference in wakeup_preempt_entity
>> via pick_next_task_fair.
>>
>> Detect when we would be about to do that, throw a warning and
>> then refuse to actually set it.
>>
>> This has been suggested at least twice[0][1]: just do it.
>>
>> [0] https://marc.info/?l=linux-kernel&m=146651668921468&w=2
>> [1] https://lkml.org/lkml/2016/6/16/663
>>
>> Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
>> Cc: Ben Segall <bsegall@google.com>
>> Cc: Peter Zijlstra <peterz@infradead.org>
>> Cc: Thomas Gleixner <tglx@linutronix.de>
>> Signed-off-by: Daniel Axtens <dja@axtens.net>
>>
>> ---
>>
>> I recently had to debug a problem with these (we hadn't backported
>> Konstantin's patches in this area) and this would have saved a lot
>> of time/pain.
>>
>> v2: use SCHED_WARN_ON to restrict when the test is run. This is a
>> macro for WARN_ON_ONCE, which is convenient.
>> ---
>> kernel/sched/fair.c | 10 ++++++++--
>> 1 file changed, 8 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index d71109321841..44b94cfe02cb 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -6168,8 +6168,11 @@ static void set_last_buddy(struct sched_entity *se)
>> if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
>> return;
>>
>> - for_each_sched_entity(se)
>> + for_each_sched_entity(se) {
>> + if (SCHED_WARN_ON(!se->on_rq))
>> + return;
>> cfs_rq_of(se)->last = se;
>> + }
>
> This won't build in the !CONFIG_SCHED_DEBUG case, because of the naive definition
> in sched.h:
>
> #define SCHED_WARN_ON(x) ((void)(x))
>
> That should be changed to something like:
>
> #define SCHED_WARN_ON(x) ((void)(x))
>
> I've applied the fix below. (untested at the moment)
>
> Thanks,
>
> Ingo
>
> ========================>
> From 6d3aed3d8a0573d0a6eb1160ccd0a0713f4dbc2f Mon Sep 17 00:00:00 2001
> From: Ingo Molnar <mingo@kernel.org>
> Date: Tue, 20 Jun 2017 12:24:42 +0200
> Subject: [PATCH] sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well
>
> This definition of SCHED_WARN_ON():
>
> #define SCHED_WARN_ON(x) ((void)(x))
>
> is not fully compatible with the 'real' WARN_ON_ONCE() primitive, as it
> has no return value, so it cannot be used in conditionals.
>
> Fix it.
>
> Cc: Daniel Axtens <dja@axtens.net>
> Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Mike Galbraith <efault@gmx.de>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Cc: linux-kernel@vger.kernel.org
> Signed-off-by: Ingo Molnar <mingo@kernel.org>
> ---
> kernel/sched/sched.h | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> index f2ef759a4cb6..e0329d10bdb8 100644
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -39,9 +39,9 @@
> #include "cpuacct.h"
>
> #ifdef CONFIG_SCHED_DEBUG
> -#define SCHED_WARN_ON(x) WARN_ONCE(x, #x)
> +# define SCHED_WARN_ON(x) WARN_ONCE(x, #x)
> #else
> -#define SCHED_WARN_ON(x) ((void)(x))
> +# define SCHED_WARN_ON(x) ({ (void)(x), 0; })
> #endif
>
> struct rq;
^ permalink raw reply [flat|nested] 6+ messages in thread
* [tip:sched/core] sched/fair: WARN() and refuse to set buddy when !se->on_rq
2017-05-10 20:11 [PATCH v2] sched/fair: WARN and refuse to set buddy when !se->on_rq Daniel Axtens
2017-06-19 14:05 ` Daniel Axtens
2017-06-20 10:27 ` [PATCH] sched/debug: Fix SCHED_WARN_ON() to return a value on !CONFIG_SCHED_DEBUG as well Ingo Molnar
@ 2017-06-20 13:26 ` tip-bot for Daniel Axtens
2 siblings, 0 replies; 6+ messages in thread
From: tip-bot for Daniel Axtens @ 2017-06-20 13:26 UTC (permalink / raw)
To: linux-tip-commits
Cc: peterz, khlebnikov, bsegall, mingo, efault, hpa, linux-kernel,
torvalds, dja, tglx
Commit-ID: c5ae366e12b2bd56fc7d7e9d484836bec9ddc110
Gitweb: http://git.kernel.org/tip/c5ae366e12b2bd56fc7d7e9d484836bec9ddc110
Author: Daniel Axtens <dja@axtens.net>
AuthorDate: Thu, 11 May 2017 06:11:39 +1000
Committer: Ingo Molnar <mingo@kernel.org>
CommitDate: Tue, 20 Jun 2017 12:26:52 +0200
sched/fair: WARN() and refuse to set buddy when !se->on_rq
If we set a next or last buddy for a se that is not on_rq, we will
end up taking a NULL pointer dereference in wakeup_preempt_entity
via pick_next_task_fair.
Detect when we would be about to do that, throw a warning and
then refuse to actually set it.
This has been suggested at least twice:
https://marc.info/?l=linux-kernel&m=146651668921468&w=2
https://lkml.org/lkml/2016/6/16/663
I recently had to debug a problem with these (we hadn't backported
Konstantin's patches in this area) and this would have saved a lot
of time/pain.
Just do it.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Cc: Ben Segall <bsegall@google.com>
Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170510201139.16236-1-dja@axtens.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
kernel/sched/fair.c | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 396bca9..cb3a3da 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -6164,8 +6164,11 @@ static void set_last_buddy(struct sched_entity *se)
if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
return;
- for_each_sched_entity(se)
+ for_each_sched_entity(se) {
+ if (SCHED_WARN_ON(!se->on_rq))
+ return;
cfs_rq_of(se)->last = se;
+ }
}
static void set_next_buddy(struct sched_entity *se)
@@ -6173,8 +6176,11 @@ static void set_next_buddy(struct sched_entity *se)
if (entity_is_task(se) && unlikely(task_of(se)->policy == SCHED_IDLE))
return;
- for_each_sched_entity(se)
+ for_each_sched_entity(se) {
+ if (SCHED_WARN_ON(!se->on_rq))
+ return;
cfs_rq_of(se)->next = se;
+ }
}
static void set_skip_buddy(struct sched_entity *se)
^ permalink raw reply [flat|nested] 6+ messages in thread