* Re: [PATCH] sched,nohz: Change rq->nr_running always using wrappers
2014-05-08 23:00 [PATCH] sched,nohz: Change rq->nr_running always using wrappers Kirill Tkhai
@ 2014-05-09 9:12 ` Peter Zijlstra
2014-05-12 17:56 ` Frederic Weisbecker
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Peter Zijlstra @ 2014-05-09 9:12 UTC (permalink / raw)
To: Kirill Tkhai; +Cc: linux-kernel, fweisbec, mingo
[-- Attachment #1: Type: text/plain, Size: 593 bytes --]
On Fri, May 09, 2014 at 03:00:14AM +0400, Kirill Tkhai wrote:
> Sometimes nr_running may cross 2 but interrupt is not being
> sent to rq's cpu. In this case we don't reenable timer.
> Looks like, this may be a reason of rare unexpected effects,
> if nohz is enabled.
>
> Patch replaces all places of direct changing of nr_running
> and makes add_nr_running() caring about crossing border.
>
> Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
> CC: Frederic Weisbecker <fweisbec@gmail.com>
> CC: Peter Zijlstra <peterz@infradead.org>
> CC: Ingo Molnar <mingo@kernel.org>
Thanks!
[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] sched,nohz: Change rq->nr_running always using wrappers
2014-05-08 23:00 [PATCH] sched,nohz: Change rq->nr_running always using wrappers Kirill Tkhai
2014-05-09 9:12 ` Peter Zijlstra
@ 2014-05-12 17:56 ` Frederic Weisbecker
2014-05-19 13:09 ` [tip:sched/core] sched,nohz: Change rq-> nr_running " tip-bot for Kirill Tkhai
2014-05-22 12:28 ` [tip:sched/core] sched, nohz: Change rq-> nr_running to always use wrappers tip-bot for Kirill Tkhai
3 siblings, 0 replies; 5+ messages in thread
From: Frederic Weisbecker @ 2014-05-12 17:56 UTC (permalink / raw)
To: Kirill Tkhai; +Cc: linux-kernel, peterz, mingo
On Fri, May 09, 2014 at 03:00:14AM +0400, Kirill Tkhai wrote:
> Sometimes nr_running may cross 2 but interrupt is not being
> sent to rq's cpu. In this case we don't reenable timer.
> Looks like, this may be a reason of rare unexpected effects,
> if nohz is enabled.
>
> Patch replaces all places of direct changing of nr_running
> and makes add_nr_running() caring about crossing border.
>
> Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
> CC: Frederic Weisbecker <fweisbec@gmail.com>
> CC: Peter Zijlstra <peterz@infradead.org>
> CC: Ingo Molnar <mingo@kernel.org>
Right I had that issue with throttling in my TODO list.
Thanks for fixing that:
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
^ permalink raw reply [flat|nested] 5+ messages in thread
* [tip:sched/core] sched,nohz: Change rq-> nr_running always using wrappers
2014-05-08 23:00 [PATCH] sched,nohz: Change rq->nr_running always using wrappers Kirill Tkhai
2014-05-09 9:12 ` Peter Zijlstra
2014-05-12 17:56 ` Frederic Weisbecker
@ 2014-05-19 13:09 ` tip-bot for Kirill Tkhai
2014-05-22 12:28 ` [tip:sched/core] sched, nohz: Change rq-> nr_running to always use wrappers tip-bot for Kirill Tkhai
3 siblings, 0 replies; 5+ messages in thread
From: tip-bot for Kirill Tkhai @ 2014-05-19 13:09 UTC (permalink / raw)
To: linux-tip-commits; +Cc: linux-kernel, hpa, mingo, fweisbec, peterz, tkhai, tglx
Commit-ID: 6d0909086bb3ceb22a5a482f27874d7e4b1bdbd2
Gitweb: http://git.kernel.org/tip/6d0909086bb3ceb22a5a482f27874d7e4b1bdbd2
Author: Kirill Tkhai <tkhai@yandex.ru>
AuthorDate: Fri, 9 May 2014 03:00:14 +0400
Committer: Thomas Gleixner <tglx@linutronix.de>
CommitDate: Mon, 19 May 2014 22:02:41 +0900
sched,nohz: Change rq->nr_running always using wrappers
Sometimes nr_running may cross 2 but interrupt is not being
sent to rq's cpu. In this case we don't reenable timer.
Looks like, this may be a reason of rare unexpected effects,
if nohz is enabled.
Patch replaces all places of direct changing of nr_running
and makes add_nr_running() caring about crossing border.
Cc: Ingo Molnar <mingo@kernel.org>
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140508225830.2469.97461.stgit@localhost
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
kernel/sched/deadline.c | 4 ++--
kernel/sched/fair.c | 8 ++++----
kernel/sched/rt.c | 4 ++--
kernel/sched/sched.h | 12 +++++++-----
kernel/sched/stop_task.c | 4 ++--
5 files changed, 17 insertions(+), 15 deletions(-)
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 800e99b..e0a04ae 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -741,7 +741,7 @@ void inc_dl_tasks(struct sched_dl_entity *dl_se, struct dl_rq *dl_rq)
WARN_ON(!dl_prio(prio));
dl_rq->dl_nr_running++;
- inc_nr_running(rq_of_dl_rq(dl_rq));
+ add_nr_running(rq_of_dl_rq(dl_rq), 1);
inc_dl_deadline(dl_rq, deadline);
inc_dl_migration(dl_se, dl_rq);
@@ -755,7 +755,7 @@ void dec_dl_tasks(struct sched_dl_entity *dl_se, struct dl_rq *dl_rq)
WARN_ON(!dl_prio(prio));
WARN_ON(!dl_rq->dl_nr_running);
dl_rq->dl_nr_running--;
- dec_nr_running(rq_of_dl_rq(dl_rq));
+ sub_nr_running(rq_of_dl_rq(dl_rq), 1);
dec_dl_deadline(dl_rq, dl_se->deadline);
dec_dl_migration(dl_se, dl_rq);
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 26ec668..f7cac2b 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3325,7 +3325,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
}
if (!se)
- rq->nr_running -= task_delta;
+ sub_nr_running(rq, task_delta);
cfs_rq->throttled = 1;
cfs_rq->throttled_clock = rq_clock(rq);
@@ -3376,7 +3376,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
}
if (!se)
- rq->nr_running += task_delta;
+ add_nr_running(rq, task_delta);
/* determine whether we need to wake up potentially idle cpu */
if (rq->curr == rq->idle && rq->cfs.nr_running)
@@ -3908,7 +3908,7 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
if (!se) {
update_rq_runnable_avg(rq, rq->nr_running);
- inc_nr_running(rq);
+ add_nr_running(rq, 1);
}
hrtick_update(rq);
}
@@ -3968,7 +3968,7 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags)
}
if (!se) {
- dec_nr_running(rq);
+ sub_nr_running(rq, 1);
update_rq_runnable_avg(rq, 1);
}
hrtick_update(rq);
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 7795e29..0ebfd7a 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -973,7 +973,7 @@ dequeue_top_rt_rq(struct rt_rq *rt_rq)
BUG_ON(!rq->nr_running);
- rq->nr_running -= rt_rq->rt_nr_running;
+ sub_nr_running(rq, rt_rq->rt_nr_running);
rt_rq->rt_queued = 0;
}
@@ -989,7 +989,7 @@ enqueue_top_rt_rq(struct rt_rq *rt_rq)
if (rt_rq_throttled(rt_rq) || !rt_rq->rt_nr_running)
return;
- rq->nr_running += rt_rq->rt_nr_running;
+ add_nr_running(rq, rt_rq->rt_nr_running);
rt_rq->rt_queued = 1;
}
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index b2cbe81..600e229 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1206,12 +1206,14 @@ extern void update_idle_cpu_load(struct rq *this_rq);
extern void init_task_runnable_average(struct task_struct *p);
-static inline void inc_nr_running(struct rq *rq)
+static inline void add_nr_running(struct rq *rq, unsigned count)
{
- rq->nr_running++;
+ unsigned prev_nr = rq->nr_running;
+
+ rq->nr_running = prev_nr + count;
#ifdef CONFIG_NO_HZ_FULL
- if (rq->nr_running == 2) {
+ if (prev_nr < 2 && rq->nr_running >= 2) {
if (tick_nohz_full_cpu(rq->cpu)) {
/* Order rq->nr_running write against the IPI */
smp_wmb();
@@ -1221,9 +1223,9 @@ static inline void inc_nr_running(struct rq *rq)
#endif
}
-static inline void dec_nr_running(struct rq *rq)
+static inline void sub_nr_running(struct rq *rq, unsigned count)
{
- rq->nr_running--;
+ rq->nr_running -= count;
}
static inline void rq_last_tick_reset(struct rq *rq)
diff --git a/kernel/sched/stop_task.c b/kernel/sched/stop_task.c
index d6ce65d..bfe0eda 100644
--- a/kernel/sched/stop_task.c
+++ b/kernel/sched/stop_task.c
@@ -41,13 +41,13 @@ pick_next_task_stop(struct rq *rq, struct task_struct *prev)
static void
enqueue_task_stop(struct rq *rq, struct task_struct *p, int flags)
{
- inc_nr_running(rq);
+ add_nr_running(rq, 1);
}
static void
dequeue_task_stop(struct rq *rq, struct task_struct *p, int flags)
{
- dec_nr_running(rq);
+ sub_nr_running(rq, 1);
}
static void yield_task_stop(struct rq *rq)
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [tip:sched/core] sched, nohz: Change rq-> nr_running to always use wrappers
2014-05-08 23:00 [PATCH] sched,nohz: Change rq->nr_running always using wrappers Kirill Tkhai
` (2 preceding siblings ...)
2014-05-19 13:09 ` [tip:sched/core] sched,nohz: Change rq-> nr_running " tip-bot for Kirill Tkhai
@ 2014-05-22 12:28 ` tip-bot for Kirill Tkhai
3 siblings, 0 replies; 5+ messages in thread
From: tip-bot for Kirill Tkhai @ 2014-05-22 12:28 UTC (permalink / raw)
To: linux-tip-commits; +Cc: linux-kernel, hpa, mingo, fweisbec, peterz, tkhai, tglx
Commit-ID: 72465447867b9de6b5cdea5d10f9781585136270
Gitweb: http://git.kernel.org/tip/72465447867b9de6b5cdea5d10f9781585136270
Author: Kirill Tkhai <tkhai@yandex.ru>
AuthorDate: Fri, 9 May 2014 03:00:14 +0400
Committer: Ingo Molnar <mingo@kernel.org>
CommitDate: Thu, 22 May 2014 11:16:33 +0200
sched, nohz: Change rq->nr_running to always use wrappers
Sometimes ->nr_running may cross 2 but interrupt is not being
sent to rq's cpu. In this case we don't reenable the timer.
Looks like this may be the reason for rare unexpected effects,
if nohz is enabled.
Patch replaces all places of direct changing of nr_running
and makes add_nr_running() caring about crossing border.
Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140508225830.2469.97461.stgit@localhost
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
kernel/sched/deadline.c | 4 ++--
kernel/sched/fair.c | 8 ++++----
kernel/sched/rt.c | 4 ++--
kernel/sched/sched.h | 12 +++++++-----
kernel/sched/stop_task.c | 4 ++--
5 files changed, 17 insertions(+), 15 deletions(-)
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 800e99b..e0a04ae 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -741,7 +741,7 @@ void inc_dl_tasks(struct sched_dl_entity *dl_se, struct dl_rq *dl_rq)
WARN_ON(!dl_prio(prio));
dl_rq->dl_nr_running++;
- inc_nr_running(rq_of_dl_rq(dl_rq));
+ add_nr_running(rq_of_dl_rq(dl_rq), 1);
inc_dl_deadline(dl_rq, deadline);
inc_dl_migration(dl_se, dl_rq);
@@ -755,7 +755,7 @@ void dec_dl_tasks(struct sched_dl_entity *dl_se, struct dl_rq *dl_rq)
WARN_ON(!dl_prio(prio));
WARN_ON(!dl_rq->dl_nr_running);
dl_rq->dl_nr_running--;
- dec_nr_running(rq_of_dl_rq(dl_rq));
+ sub_nr_running(rq_of_dl_rq(dl_rq), 1);
dec_dl_deadline(dl_rq, dl_se->deadline);
dec_dl_migration(dl_se, dl_rq);
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 26ec668..f7cac2b 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3325,7 +3325,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
}
if (!se)
- rq->nr_running -= task_delta;
+ sub_nr_running(rq, task_delta);
cfs_rq->throttled = 1;
cfs_rq->throttled_clock = rq_clock(rq);
@@ -3376,7 +3376,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
}
if (!se)
- rq->nr_running += task_delta;
+ add_nr_running(rq, task_delta);
/* determine whether we need to wake up potentially idle cpu */
if (rq->curr == rq->idle && rq->cfs.nr_running)
@@ -3908,7 +3908,7 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
if (!se) {
update_rq_runnable_avg(rq, rq->nr_running);
- inc_nr_running(rq);
+ add_nr_running(rq, 1);
}
hrtick_update(rq);
}
@@ -3968,7 +3968,7 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags)
}
if (!se) {
- dec_nr_running(rq);
+ sub_nr_running(rq, 1);
update_rq_runnable_avg(rq, 1);
}
hrtick_update(rq);
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 7795e29..0ebfd7a 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -973,7 +973,7 @@ dequeue_top_rt_rq(struct rt_rq *rt_rq)
BUG_ON(!rq->nr_running);
- rq->nr_running -= rt_rq->rt_nr_running;
+ sub_nr_running(rq, rt_rq->rt_nr_running);
rt_rq->rt_queued = 0;
}
@@ -989,7 +989,7 @@ enqueue_top_rt_rq(struct rt_rq *rt_rq)
if (rt_rq_throttled(rt_rq) || !rt_rq->rt_nr_running)
return;
- rq->nr_running += rt_rq->rt_nr_running;
+ add_nr_running(rq, rt_rq->rt_nr_running);
rt_rq->rt_queued = 1;
}
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index b2cbe81..600e229 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1206,12 +1206,14 @@ extern void update_idle_cpu_load(struct rq *this_rq);
extern void init_task_runnable_average(struct task_struct *p);
-static inline void inc_nr_running(struct rq *rq)
+static inline void add_nr_running(struct rq *rq, unsigned count)
{
- rq->nr_running++;
+ unsigned prev_nr = rq->nr_running;
+
+ rq->nr_running = prev_nr + count;
#ifdef CONFIG_NO_HZ_FULL
- if (rq->nr_running == 2) {
+ if (prev_nr < 2 && rq->nr_running >= 2) {
if (tick_nohz_full_cpu(rq->cpu)) {
/* Order rq->nr_running write against the IPI */
smp_wmb();
@@ -1221,9 +1223,9 @@ static inline void inc_nr_running(struct rq *rq)
#endif
}
-static inline void dec_nr_running(struct rq *rq)
+static inline void sub_nr_running(struct rq *rq, unsigned count)
{
- rq->nr_running--;
+ rq->nr_running -= count;
}
static inline void rq_last_tick_reset(struct rq *rq)
diff --git a/kernel/sched/stop_task.c b/kernel/sched/stop_task.c
index d6ce65d..bfe0eda 100644
--- a/kernel/sched/stop_task.c
+++ b/kernel/sched/stop_task.c
@@ -41,13 +41,13 @@ pick_next_task_stop(struct rq *rq, struct task_struct *prev)
static void
enqueue_task_stop(struct rq *rq, struct task_struct *p, int flags)
{
- inc_nr_running(rq);
+ add_nr_running(rq, 1);
}
static void
dequeue_task_stop(struct rq *rq, struct task_struct *p, int flags)
{
- dec_nr_running(rq);
+ sub_nr_running(rq, 1);
}
static void yield_task_stop(struct rq *rq)
^ permalink raw reply related [flat|nested] 5+ messages in thread