[PATCH v2 0/2] Clean up schedstat and task delay acct config options

[PATCH v2 0/2] Clean up schedstat and task delay acct config options

[Naveen N. Rao]

Further to the earlier discussion [1], this patch set introduces CONFIG_SCHED_INFO for the common code associated with CONFIG_SCHEDSTATS and CONFIG_TASK_DELAY_ACCT. Further, /proc/<pid>/schedstat is now exposed if either of these is enabled.

[1] http://thread.gmane.org/gmane.linux.kernel/1961707

- Naveen


Naveen N. Rao (2):
  schedstat: Simplify sched_info accounting dependency
  schedstat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO is set

 fs/proc/base.c        | 11 +++++++----
 include/linux/sched.h |  6 +++---
 init/Kconfig          |  1 +
 kernel/sched/core.c   |  2 +-
 kernel/sched/stats.h  |  4 ++--
 lib/Kconfig.debug     |  5 +++++
 6 files changed, 19 insertions(+), 10 deletions(-)

-- 
2.4.0

[PATCH v2 1/2] schedstat: Simplify sched_info accounting dependency

[Naveen N. Rao]

Both CONFIG_SCHEDSTATS and CONFIG_TASK_DELAY_ACCT track task sched_info.
Simplify by introducing a common CONFIG_SCHED_INFO selected by both.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
---
 include/linux/sched.h | 6 +++---
 init/Kconfig          | 1 +
 kernel/sched/core.c   | 2 +-
 kernel/sched/stats.h  | 4 ++--
 lib/Kconfig.debug     | 5 +++++
 5 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 6633e83..9bf4bc0 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -849,7 +849,7 @@ extern struct user_struct root_user;
 struct backing_dev_info;
 struct reclaim_state;
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 struct sched_info {
 	/* cumulative counters */
 	unsigned long pcount;	      /* # of times run on this cpu */
@@ -859,7 +859,7 @@ struct sched_info {
 	unsigned long long last_arrival,/* when we last ran on a cpu */
 			   last_queued;	/* when we were last queued to run */
 };
-#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
+#endif /* CONFIG_SCHED_INFO */
 
 #ifdef CONFIG_TASK_DELAY_ACCT
 struct task_delay_info {
@@ -1408,7 +1408,7 @@ struct task_struct {
 	int rcu_tasks_idle_cpu;
 #endif /* #ifdef CONFIG_TASKS_RCU */
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 	struct sched_info sched_info;
 #endif
 
diff --git a/init/Kconfig b/init/Kconfig
index b999fa3..12cb556 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -435,6 +435,7 @@ config TASKSTATS
 config TASK_DELAY_ACCT
 	bool "Enable per-task delay accounting"
 	depends on TASKSTATS
+	select SCHED_INFO
 	help
 	  Collect information on time spent by a task waiting for system
 	  resources like cpu, synchronous block I/O completion and swapping
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index b803e1b..5caa029 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -2164,7 +2164,7 @@ int sched_fork(unsigned long clone_flags, struct task_struct *p)
 	set_task_cpu(p, cpu);
 	raw_spin_unlock_irqrestore(&p->pi_lock, flags);
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 	if (likely(sched_info_on()))
 		memset(&p->sched_info, 0, sizeof(p->sched_info));
 #endif
diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h
index 077ebbd..b0fbc76 100644
--- a/kernel/sched/stats.h
+++ b/kernel/sched/stats.h
@@ -47,7 +47,7 @@ rq_sched_info_depart(struct rq *rq, unsigned long long delta)
 # define schedstat_set(var, val)	do { } while (0)
 #endif
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 static inline void sched_info_reset_dequeued(struct task_struct *t)
 {
 	t->sched_info.last_queued = 0;
@@ -156,7 +156,7 @@ sched_info_switch(struct rq *rq,
 #define sched_info_depart(rq, t)		do { } while (0)
 #define sched_info_arrive(rq, next)		do { } while (0)
 #define sched_info_switch(rq, t, next)		do { } while (0)
-#endif /* CONFIG_SCHEDSTATS || CONFIG_TASK_DELAY_ACCT */
+#endif /* CONFIG_SCHED_INFO */
 
 /*
  * The following are functions that support scheduler-internal time accounting.
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index b908048..e2894b2 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -841,9 +841,14 @@ config SCHED_DEBUG
 	  that can help debug the scheduler. The runtime overhead of this
 	  option is minimal.
 
+config SCHED_INFO
+	bool
+	default n
+
 config SCHEDSTATS
 	bool "Collect scheduler statistics"
 	depends on DEBUG_KERNEL && PROC_FS
+	select SCHED_INFO
 	help
 	  If you say Y here, additional code will be inserted into the
 	  scheduler and related routines to collect statistics about
-- 
2.4.0

[PATCH v2 2/2] schedstat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO is set

[Naveen N. Rao]

In the unlikely scenario that the kernel is booted with nodelayacct, we
dump all zeroes so as to make it clear that scheduler statistics are not
available.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
---
 fs/proc/base.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index 286a422..a299131 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -304,14 +304,17 @@ static int proc_pid_stack(struct seq_file *m, struct pid_namespace *ns,
 }
 #endif
 
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 /*
  * Provides /proc/PID/schedstat
  */
 static int proc_pid_schedstat(struct seq_file *m, struct pid_namespace *ns,
 			      struct pid *pid, struct task_struct *task)
 {
-	seq_printf(m, "%llu %llu %lu\n",
+	if (unlikely(!sched_info_on()))
+		seq_printf(m, "0 0 0\n");
+	else
+		seq_printf(m, "%llu %lld %ld\n",
 		   (unsigned long long)task->se.sum_exec_runtime,
 		   (unsigned long long)task->sched_info.run_delay,
 		   task->sched_info.pcount);
@@ -2600,7 +2603,7 @@ static const struct pid_entry tgid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat",  S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
@@ -2948,7 +2951,7 @@ static const struct pid_entry tid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat", S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
-- 
2.4.0

Re: [PATCH v2 2/2] schedstat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO is set

[Srikar Dronamraju]

* Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> [2015-06-25 23:53:38]:

>  /*
>   * Provides /proc/PID/schedstat
>   */
>  static int proc_pid_schedstat(struct seq_file *m, struct pid_namespace *ns,
>  			      struct pid *pid, struct task_struct *task)
>  {
> -	seq_printf(m, "%llu %llu %lu\n",
> +	if (unlikely(!sched_info_on()))
> +		seq_printf(m, "0 0 0\n");
> +	else
> +		seq_printf(m, "%llu %lld %ld\n",

Should the above be seq_printf(m, "%llu %llu %lu\n", ?

>  		   (unsigned long long)task->se.sum_exec_runtime,
>  		   (unsigned long long)task->sched_info.run_delay,
>  		   task->sched_info.pcount);
> -- 
> 2.4.0
> 

-- 
Thanks and Regards
Srikar Dronamraju

Re: [PATCH v2 2/2] schedstat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO is set

[Naveen N. Rao]

On 2015/06/30 11:25AM, Srikar Dronamraju wrote:
> * Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> [2015-06-25 23:53:38]:
> 
> >  /*
> >   * Provides /proc/PID/schedstat
> >   */
> >  static int proc_pid_schedstat(struct seq_file *m, struct pid_namespace *ns,
> >  			      struct pid *pid, struct task_struct *task)
> >  {
> > -	seq_printf(m, "%llu %llu %lu\n",
> > +	if (unlikely(!sched_info_on()))
> > +		seq_printf(m, "0 0 0\n");
> > +	else
> > +		seq_printf(m, "%llu %lld %ld\n",
> 
> Should the above be seq_printf(m, "%llu %llu %lu\n", ?

Yikes! Nice catch, thanks! This looks to be left over from my tests with 
returning -1. Fixed patch on the way...

Regards,
Naveen

[PATCH v3 2/2] schedstat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO is set

[Naveen N. Rao]

In the unlikely scenario that the kernel is booted with nodelayacct, we
dump all zeroes so as to make it clear that scheduler statistics are not
available.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
---
Changes since v2: Revert to unsigned long in printf


 fs/proc/base.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index 286a422..ad63fa3 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -304,14 +304,17 @@ static int proc_pid_stack(struct seq_file *m, struct pid_namespace *ns,
 }
 #endif
 
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 /*
  * Provides /proc/PID/schedstat
  */
 static int proc_pid_schedstat(struct seq_file *m, struct pid_namespace *ns,
 			      struct pid *pid, struct task_struct *task)
 {
-	seq_printf(m, "%llu %llu %lu\n",
+	if (unlikely(!sched_info_on()))
+		seq_printf(m, "0 0 0\n");
+	else
+		seq_printf(m, "%llu %llu %lu\n",
 		   (unsigned long long)task->se.sum_exec_runtime,
 		   (unsigned long long)task->sched_info.run_delay,
 		   task->sched_info.pcount);
@@ -2600,7 +2603,7 @@ static const struct pid_entry tgid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat",  S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
@@ -2948,7 +2951,7 @@ static const struct pid_entry tid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat", S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
-- 
2.4.0

Re: [PATCH v3 2/2] schedstat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO is set

[Srikar Dronamraju]

* Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> [2015-06-30 14:36:03]:

> In the unlikely scenario that the kernel is booted with nodelayacct, we
> dump all zeroes so as to make it clear that scheduler statistics are not
> available.
> 
> Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
> ---
> Changes since v2: Revert to unsigned long in printf
> 

Looks good to me.

Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
-- 
Thanks and Regards
Srikar Dronamraju

Re: [PATCH v2 1/2] schedstat: Simplify sched_info accounting dependency

[Srikar Dronamraju]

* Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> [2015-06-25 23:53:37]:

> Both CONFIG_SCHEDSTATS and CONFIG_TASK_DELAY_ACCT track task sched_info.
> Simplify by introducing a common CONFIG_SCHED_INFO selected by both.
> 
> Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
> ---

Looks good to me.

Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
-- 
Thanks and Regards
Srikar Dronamraju

Re: [PATCH v3 2/2] schedstat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO is set

[Balbir Singh]

On Tue, Jun 30, 2015 at 6:59 PM, Srikar Dronamraju
<srikar@linux.vnet.ibm.com> wrote:
> * Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> [2015-06-30 14:36:03]:
>
>> In the unlikely scenario that the kernel is booted with nodelayacct, we
>> dump all zeroes so as to make it clear that scheduler statistics are not
>> available.
>>
>> Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
>> ---
>> Changes since v2: Revert to unsigned long in printf
>>
>
> Looks good to me.
>
> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>

Acked-by: Balbir Singh <bsingharora@gmail.com>

Re: [PATCH v2 1/2] schedstat: Simplify sched_info accounting dependency

[Balbir Singh]

On Tue, Jun 30, 2015 at 7:00 PM, Srikar Dronamraju
<srikar@linux.vnet.ibm.com> wrote:
> * Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> [2015-06-25 23:53:37]:
>
>> Both CONFIG_SCHEDSTATS and CONFIG_TASK_DELAY_ACCT track task sched_info.
>> Simplify by introducing a common CONFIG_SCHED_INFO selected by both.
>>
>> Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
>> ---
>
> Looks good to me.
>
> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>


Acked-by: Balbir Singh <bsingharora@gmail.com>

[tip:sched/urgent] sched/stat: Simplify the sched_info accounting dependency

[tip-bot for Naveen N. Rao]

Commit-ID:  640032ed8d55bf6f09b0701e7464aab7e24e0361
Gitweb:     http://git.kernel.org/tip/640032ed8d55bf6f09b0701e7464aab7e24e0361
Author:     Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 23:53:37 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 1 Jul 2015 10:02:58 +0200

sched/stat: Simplify the sched_info accounting dependency

Both CONFIG_SCHEDSTATS=y and CONFIG_TASK_DELAY_ACCT=y track task
sched_info, which results in ugly #if clauses.

Simplify the code by introducing a synthethic CONFIG_SCHED_INFO
switch, selected by both.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: a.p.zijlstra@chello.nl
Cc: ricklind@us.ibm.com
Link: http://lkml.kernel.org/r/8d19eef800811a94b0f91bcbeb27430a884d7433.1435255405.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 include/linux/sched.h | 6 +++---
 init/Kconfig          | 1 +
 kernel/sched/core.c   | 2 +-
 kernel/sched/stats.h  | 4 ++--
 lib/Kconfig.debug     | 5 +++++
 5 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 6633e83..9bf4bc0 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -849,7 +849,7 @@ extern struct user_struct root_user;
 struct backing_dev_info;
 struct reclaim_state;
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 struct sched_info {
 	/* cumulative counters */
 	unsigned long pcount;	      /* # of times run on this cpu */
@@ -859,7 +859,7 @@ struct sched_info {
 	unsigned long long last_arrival,/* when we last ran on a cpu */
 			   last_queued;	/* when we were last queued to run */
 };
-#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
+#endif /* CONFIG_SCHED_INFO */
 
 #ifdef CONFIG_TASK_DELAY_ACCT
 struct task_delay_info {
@@ -1408,7 +1408,7 @@ struct task_struct {
 	int rcu_tasks_idle_cpu;
 #endif /* #ifdef CONFIG_TASKS_RCU */
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 	struct sched_info sched_info;
 #endif
 
diff --git a/init/Kconfig b/init/Kconfig
index b999fa3..12cb556 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -435,6 +435,7 @@ config TASKSTATS
 config TASK_DELAY_ACCT
 	bool "Enable per-task delay accounting"
 	depends on TASKSTATS
+	select SCHED_INFO
 	help
 	  Collect information on time spent by a task waiting for system
 	  resources like cpu, synchronous block I/O completion and swapping
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 5bcf926..435d8e4 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1975,7 +1975,7 @@ int sched_fork(unsigned long clone_flags, struct task_struct *p)
 	set_task_cpu(p, cpu);
 	raw_spin_unlock_irqrestore(&p->pi_lock, flags);
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 	if (likely(sched_info_on()))
 		memset(&p->sched_info, 0, sizeof(p->sched_info));
 #endif
diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h
index 077ebbd..b0fbc76 100644
--- a/kernel/sched/stats.h
+++ b/kernel/sched/stats.h
@@ -47,7 +47,7 @@ rq_sched_info_depart(struct rq *rq, unsigned long long delta)
 # define schedstat_set(var, val)	do { } while (0)
 #endif
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 static inline void sched_info_reset_dequeued(struct task_struct *t)
 {
 	t->sched_info.last_queued = 0;
@@ -156,7 +156,7 @@ sched_info_switch(struct rq *rq,
 #define sched_info_depart(rq, t)		do { } while (0)
 #define sched_info_arrive(rq, next)		do { } while (0)
 #define sched_info_switch(rq, t, next)		do { } while (0)
-#endif /* CONFIG_SCHEDSTATS || CONFIG_TASK_DELAY_ACCT */
+#endif /* CONFIG_SCHED_INFO */
 
 /*
  * The following are functions that support scheduler-internal time accounting.
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index b908048..e2894b2 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -841,9 +841,14 @@ config SCHED_DEBUG
 	  that can help debug the scheduler. The runtime overhead of this
 	  option is minimal.
 
+config SCHED_INFO
+	bool
+	default n
+
 config SCHEDSTATS
 	bool "Collect scheduler statistics"
 	depends on DEBUG_KERNEL && PROC_FS
+	select SCHED_INFO
 	help
 	  If you say Y here, additional code will be inserted into the
 	  scheduler and related routines to collect statistics about

[tip:sched/urgent] sched/stat: Expose /proc/pid/ schedstat if CONFIG_SCHED_INFO=y

[tip-bot for Naveen N. Rao]

Commit-ID:  74ed568374f0e74d80a63ab5496bf32db641ec60
Gitweb:     http://git.kernel.org/tip/74ed568374f0e74d80a63ab5496bf32db641ec60
Author:     Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
AuthorDate: Tue, 30 Jun 2015 14:36:03 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 1 Jul 2015 10:02:58 +0200

sched/stat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO=y

Expand /proc/pid/schedstat output:

 - enable it on CONFIG_TASK_DELAY_ACCT=y && !CONFIG_SCHEDSTATS kernels.

 - dump all zeroes on kernels that are booted with the 'nodelayacct'
   option, which boot option disables delay accounting on
   CONFIG_TASK_DELAY_ACCT=y kernels.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: a.p.zijlstra@chello.nl
Cc: ricklind@us.ibm.com
Link: http://lkml.kernel.org/r/5ccbef17d4bc841084ea6e6421d4e4a23b7b806f.1435654789.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 fs/proc/base.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index 286a422..ad63fa3 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -304,14 +304,17 @@ static int proc_pid_stack(struct seq_file *m, struct pid_namespace *ns,
 }
 #endif
 
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 /*
  * Provides /proc/PID/schedstat
  */
 static int proc_pid_schedstat(struct seq_file *m, struct pid_namespace *ns,
 			      struct pid *pid, struct task_struct *task)
 {
-	seq_printf(m, "%llu %llu %lu\n",
+	if (unlikely(!sched_info_on()))
+		seq_printf(m, "0 0 0\n");
+	else
+		seq_printf(m, "%llu %llu %lu\n",
 		   (unsigned long long)task->se.sum_exec_runtime,
 		   (unsigned long long)task->sched_info.run_delay,
 		   task->sched_info.pcount);
@@ -2600,7 +2603,7 @@ static const struct pid_entry tgid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat",  S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
@@ -2948,7 +2951,7 @@ static const struct pid_entry tid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat", S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP

[tip:sched/urgent] sched/stat: Simplify the sched_info accounting dependency

[tip-bot for Naveen N. Rao]

Commit-ID:  f6db8347993256b58bd4746b0c4c5b935c32210d
Gitweb:     http://git.kernel.org/tip/f6db8347993256b58bd4746b0c4c5b935c32210d
Author:     Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 23:53:37 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 4 Jul 2015 10:04:30 +0200

sched/stat: Simplify the sched_info accounting dependency

Both CONFIG_SCHEDSTATS=y and CONFIG_TASK_DELAY_ACCT=y track task
sched_info, which results in ugly #if clauses.

Simplify the code by introducing a synthethic CONFIG_SCHED_INFO
switch, selected by both.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: a.p.zijlstra@chello.nl
Cc: ricklind@us.ibm.com
Link: http://lkml.kernel.org/r/8d19eef800811a94b0f91bcbeb27430a884d7433.1435255405.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 include/linux/sched.h | 6 +++---
 init/Kconfig          | 1 +
 kernel/sched/core.c   | 2 +-
 kernel/sched/stats.h  | 4 ++--
 lib/Kconfig.debug     | 5 +++++
 5 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 6633e83..9bf4bc0 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -849,7 +849,7 @@ extern struct user_struct root_user;
 struct backing_dev_info;
 struct reclaim_state;
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 struct sched_info {
 	/* cumulative counters */
 	unsigned long pcount;	      /* # of times run on this cpu */
@@ -859,7 +859,7 @@ struct sched_info {
 	unsigned long long last_arrival,/* when we last ran on a cpu */
 			   last_queued;	/* when we were last queued to run */
 };
-#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
+#endif /* CONFIG_SCHED_INFO */
 
 #ifdef CONFIG_TASK_DELAY_ACCT
 struct task_delay_info {
@@ -1408,7 +1408,7 @@ struct task_struct {
 	int rcu_tasks_idle_cpu;
 #endif /* #ifdef CONFIG_TASKS_RCU */
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 	struct sched_info sched_info;
 #endif
 
diff --git a/init/Kconfig b/init/Kconfig
index b999fa3..12cb556 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -435,6 +435,7 @@ config TASKSTATS
 config TASK_DELAY_ACCT
 	bool "Enable per-task delay accounting"
 	depends on TASKSTATS
+	select SCHED_INFO
 	help
 	  Collect information on time spent by a task waiting for system
 	  resources like cpu, synchronous block I/O completion and swapping
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index c86935a..abb8785 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1975,7 +1975,7 @@ int sched_fork(unsigned long clone_flags, struct task_struct *p)
 	set_task_cpu(p, cpu);
 	raw_spin_unlock_irqrestore(&p->pi_lock, flags);
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 	if (likely(sched_info_on()))
 		memset(&p->sched_info, 0, sizeof(p->sched_info));
 #endif
diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h
index 077ebbd..b0fbc76 100644
--- a/kernel/sched/stats.h
+++ b/kernel/sched/stats.h
@@ -47,7 +47,7 @@ rq_sched_info_depart(struct rq *rq, unsigned long long delta)
 # define schedstat_set(var, val)	do { } while (0)
 #endif
 
-#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
+#ifdef CONFIG_SCHED_INFO
 static inline void sched_info_reset_dequeued(struct task_struct *t)
 {
 	t->sched_info.last_queued = 0;
@@ -156,7 +156,7 @@ sched_info_switch(struct rq *rq,
 #define sched_info_depart(rq, t)		do { } while (0)
 #define sched_info_arrive(rq, next)		do { } while (0)
 #define sched_info_switch(rq, t, next)		do { } while (0)
-#endif /* CONFIG_SCHEDSTATS || CONFIG_TASK_DELAY_ACCT */
+#endif /* CONFIG_SCHED_INFO */
 
 /*
  * The following are functions that support scheduler-internal time accounting.
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index b908048..e2894b2 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -841,9 +841,14 @@ config SCHED_DEBUG
 	  that can help debug the scheduler. The runtime overhead of this
 	  option is minimal.
 
+config SCHED_INFO
+	bool
+	default n
+
 config SCHEDSTATS
 	bool "Collect scheduler statistics"
 	depends on DEBUG_KERNEL && PROC_FS
+	select SCHED_INFO
 	help
 	  If you say Y here, additional code will be inserted into the
 	  scheduler and related routines to collect statistics about

[tip:sched/urgent] sched/stat: Expose /proc/pid/ schedstat if CONFIG_SCHED_INFO=y

[tip-bot for Naveen N. Rao]

Commit-ID:  5968cecedd7a09f23e9fcb5f9fb4e893712f35ba
Gitweb:     http://git.kernel.org/tip/5968cecedd7a09f23e9fcb5f9fb4e893712f35ba
Author:     Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
AuthorDate: Tue, 30 Jun 2015 14:36:03 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 4 Jul 2015 10:04:31 +0200

sched/stat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO=y

Expand /proc/pid/schedstat output:

 - enable it on CONFIG_TASK_DELAY_ACCT=y && !CONFIG_SCHEDSTATS kernels.

 - dump all zeroes on kernels that are booted with the 'nodelayacct'
   option, which boot option disables delay accounting on
   CONFIG_TASK_DELAY_ACCT=y kernels.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: a.p.zijlstra@chello.nl
Cc: ricklind@us.ibm.com
Link: http://lkml.kernel.org/r/5ccbef17d4bc841084ea6e6421d4e4a23b7b806f.1435654789.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 fs/proc/base.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index 286a422..ad63fa3 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -304,14 +304,17 @@ static int proc_pid_stack(struct seq_file *m, struct pid_namespace *ns,
 }
 #endif
 
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 /*
  * Provides /proc/PID/schedstat
  */
 static int proc_pid_schedstat(struct seq_file *m, struct pid_namespace *ns,
 			      struct pid *pid, struct task_struct *task)
 {
-	seq_printf(m, "%llu %llu %lu\n",
+	if (unlikely(!sched_info_on()))
+		seq_printf(m, "0 0 0\n");
+	else
+		seq_printf(m, "%llu %llu %lu\n",
 		   (unsigned long long)task->se.sum_exec_runtime,
 		   (unsigned long long)task->sched_info.run_delay,
 		   task->sched_info.pcount);
@@ -2600,7 +2603,7 @@ static const struct pid_entry tgid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat",  S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
@@ -2948,7 +2951,7 @@ static const struct pid_entry tid_base_stuff[] = {
 #ifdef CONFIG_STACKTRACE
 	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
-#ifdef CONFIG_SCHEDSTATS
+#ifdef CONFIG_SCHED_INFO
 	ONE("schedstat", S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP