[PATCH 0/3] sched/numa: Update numa_balancing stats in /proc

[PATCH 0/3] sched/numa: Update numa_balancing stats in /proc

[Srikar Dronamraju]

Recent changes to numa balancing like adding group faults, merging
numa_faults and numa_mem_faults have not been reflected in /proc.

Hence some of the numa_balancing information is not exposed or displays
wrong info. This patchset tries to correct it.

Before change: /proc/pid/sched
clock-delta                                  :                   12
mm->numa_scan_seq                            :                   34
numa_migrations, 1
numa_faults_memory, 0, 0, 1, 0, -1
numa_faults_memory, 1, 0, 0, 0, -1
numa_faults_memory, 0, 1, 0, 0, -1
numa_faults_memory, 1, 1, 0, 0, -1
numa_faults_memory, 0, 2, 0, 0, -1
numa_faults_memory, 1, 2, 0, 0, -1
numa_faults_memory, 0, 3, 0, 0, -1
numa_faults_memory, 1, 3, 0, 0, -1

After change: /proc/pid/sched
clock-delta                                  :                   12
mm->numa_scan_seq                            :                   34
numa_pages_migrated                          :                    1
numa_preferred_nid                           :                    0
total_numa_faults                            :                 2627
current_node=0, numa_group_id=6958
numa_faults node=0 task_private=1344 task_shared=0 group_private=1344 group_shared=0
numa_faults node=1 task_private=641 task_shared=65 group_private=641 group_shared=65
numa_faults node=2 task_private=512 task_shared=0 group_private=512 group_shared=0
numa_faults node=3 task_private=64 task_shared=1 group_private=64 group_shared=1


Srikar Dronamraju (3):
  sched: Move print_cfs_rq declaration to kernel/sched/sched.h
  sched/numa: Show numa_group id in sched_debug task listings
  sched/numa: Fix numabalancing stats in /proc/pid/sched

 include/linux/sched.h |  2 --
 kernel/sched/debug.c  | 40 ++++++++++++++++++----------------------
 kernel/sched/fair.c   | 22 +++++++++++++++++++++-
 kernel/sched/sched.h  | 13 +++++++++++++
 4 files changed, 52 insertions(+), 25 deletions(-)

-- 
1.8.3.1

[PATCH 1/3] sched: Move print_cfs_rq declaration to kernel/sched/sched.h

[Srikar Dronamraju]

Currently print_cfs_rq() is declared in include/linux/sched.h.
However its not used outside kernel/sched. Hence move the declaration to
kernel/sched/sched.h

Also some functions are only available for CONFIG_SCHED_DEBUG. Hence move
the declarations within #ifdef.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
---
 include/linux/sched.h | 2 --
 kernel/sched/sched.h  | 5 +++++
 2 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 6633e83..eb5d5ff 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -191,8 +191,6 @@ struct task_group;
 #ifdef CONFIG_SCHED_DEBUG
 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
 extern void proc_sched_set_task(struct task_struct *p);
-extern void
-print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
 #endif
 
 /*
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 8858891..c901c9c 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1689,9 +1689,14 @@ static inline void double_rq_unlock(struct rq *rq1, struct rq *rq2)
 
 extern struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq);
 extern struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq);
+
+#ifdef	CONFIG_SCHED_DEBUG
 extern void print_cfs_stats(struct seq_file *m, int cpu);
 extern void print_rt_stats(struct seq_file *m, int cpu);
 extern void print_dl_stats(struct seq_file *m, int cpu);
+extern void
+print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
+#endif
 
 extern void init_cfs_rq(struct cfs_rq *cfs_rq);
 extern void init_rt_rq(struct rt_rq *rt_rq);
-- 
1.8.3.1

Re: [PATCH 1/3] sched: Move print_cfs_rq declaration to kernel/sched/sched.h

[Rik van Riel]

On 06/25/2015 01:21 PM, Srikar Dronamraju wrote:
> Currently print_cfs_rq() is declared in include/linux/sched.h.
> However its not used outside kernel/sched. Hence move the declaration to
> kernel/sched/sched.h
> 
> Also some functions are only available for CONFIG_SCHED_DEBUG. Hence move
> the declarations within #ifdef.
> 
> Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>

Acked-by: Rik van Riel <riel@redhat.com>

[tip:sched/urgent] sched/debug: Move print_cfs_rq() declaration to kernel/sched/sched.h

[tip-bot for Srikar Dronamraju]

Commit-ID:  6e3ecb4945f06ebbf1cf00ec316511c249492b08
Gitweb:     http://git.kernel.org/tip/6e3ecb4945f06ebbf1cf00ec316511c249492b08
Author:     Srikar Dronamraju <srikar@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 22:51:41 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 1 Jul 2015 10:31:27 +0200

sched/debug: Move print_cfs_rq() declaration to kernel/sched/sched.h

Currently print_cfs_rq() is declared in include/linux/sched.h.
However it's not used outside kernel/sched. Hence move the
declaration to kernel/sched/sched.h

Also some functions are only available for CONFIG_SCHED_DEBUG=y.
Hence move the declarations to within the #ifdef.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-2-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 include/linux/sched.h | 2 --
 kernel/sched/sched.h  | 5 +++++
 2 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 9bf4bc0..3505352 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -191,8 +191,6 @@ struct task_group;
 #ifdef CONFIG_SCHED_DEBUG
 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
 extern void proc_sched_set_task(struct task_struct *p);
-extern void
-print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
 #endif
 
 /*
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index aea7c1f..7d58952 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1668,9 +1668,14 @@ static inline void double_rq_unlock(struct rq *rq1, struct rq *rq2)
 
 extern struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq);
 extern struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq);
+
+#ifdef	CONFIG_SCHED_DEBUG
 extern void print_cfs_stats(struct seq_file *m, int cpu);
 extern void print_rt_stats(struct seq_file *m, int cpu);
 extern void print_dl_stats(struct seq_file *m, int cpu);
+extern void
+print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
+#endif
 
 extern void init_cfs_rq(struct cfs_rq *cfs_rq);
 extern void init_rt_rq(struct rt_rq *rt_rq);

[tip:sched/urgent] sched/debug: Move print_cfs_rq() declaration to kernel/sched/sched.h

[tip-bot for Srikar Dronamraju]

Commit-ID:  6b55c9654fccf69ae7ace23ca101dc37b903181b
Gitweb:     http://git.kernel.org/tip/6b55c9654fccf69ae7ace23ca101dc37b903181b
Author:     Srikar Dronamraju <srikar@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 22:51:41 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 4 Jul 2015 10:04:31 +0200

sched/debug: Move print_cfs_rq() declaration to kernel/sched/sched.h

Currently print_cfs_rq() is declared in include/linux/sched.h.
However it's not used outside kernel/sched. Hence move the
declaration to kernel/sched/sched.h

Also some functions are only available for CONFIG_SCHED_DEBUG=y.
Hence move the declarations to within the #ifdef.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-2-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 include/linux/sched.h | 2 --
 kernel/sched/sched.h  | 5 +++++
 2 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 9bf4bc0..3505352 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -191,8 +191,6 @@ struct task_group;
 #ifdef CONFIG_SCHED_DEBUG
 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
 extern void proc_sched_set_task(struct task_struct *p);
-extern void
-print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
 #endif
 
 /*
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index aea7c1f..7d58952 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1668,9 +1668,14 @@ static inline void double_rq_unlock(struct rq *rq1, struct rq *rq2)
 
 extern struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq);
 extern struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq);
+
+#ifdef	CONFIG_SCHED_DEBUG
 extern void print_cfs_stats(struct seq_file *m, int cpu);
 extern void print_rt_stats(struct seq_file *m, int cpu);
 extern void print_dl_stats(struct seq_file *m, int cpu);
+extern void
+print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
+#endif
 
 extern void init_cfs_rq(struct cfs_rq *cfs_rq);
 extern void init_rt_rq(struct rt_rq *rt_rq);

[PATCH 2/3] sched/numa: Show numa_group id in sched_debug task listings

[Srikar Dronamraju]

Having the numa group id in /proc/sched_debug helps to see how the numa
groups have spread across the system.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
---
 kernel/sched/debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 315c68e..f1dcd1d 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -142,7 +142,7 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p)
 		0LL, 0L);
 #endif
 #ifdef CONFIG_NUMA_BALANCING
-	SEQ_printf(m, " %d", task_node(p));
+	SEQ_printf(m, " %d %d", task_node(p), task_numa_group_id(p));
 #endif
 #ifdef CONFIG_CGROUP_SCHED
 	SEQ_printf(m, " %s", task_group_path(task_group(p)));
-- 
1.8.3.1

Re: [PATCH 2/3] sched/numa: Show numa_group id in sched_debug task listings

[Rik van Riel]

On 06/25/2015 01:21 PM, Srikar Dronamraju wrote:
> Having the numa group id in /proc/sched_debug helps to see how the numa
> groups have spread across the system.
> 
> Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>

Acked-by: Rik van Riel <riel@redhat.com>

[tip:sched/urgent] sched/numa: Show numa_group ID in /proc/ sched_debug task listings

[tip-bot for Srikar Dronamraju]

Commit-ID:  478fd9ada06ff17ae13a5b7f775346562036127e
Gitweb:     http://git.kernel.org/tip/478fd9ada06ff17ae13a5b7f775346562036127e
Author:     Srikar Dronamraju <srikar@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 22:51:42 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 1 Jul 2015 10:31:27 +0200

sched/numa: Show numa_group ID in /proc/sched_debug task listings

Having the numa group ID in /proc/sched_debug helps to see how
the numa groups have spread across the system.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-3-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 315c68e..f1dcd1d 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -142,7 +142,7 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p)
 		0LL, 0L);
 #endif
 #ifdef CONFIG_NUMA_BALANCING
-	SEQ_printf(m, " %d", task_node(p));
+	SEQ_printf(m, " %d %d", task_node(p), task_numa_group_id(p));
 #endif
 #ifdef CONFIG_CGROUP_SCHED
 	SEQ_printf(m, " %s", task_group_path(task_group(p)));

[tip:sched/urgent] sched/numa: Show numa_group ID in /proc/ sched_debug task listings

[tip-bot for Srikar Dronamraju]

Commit-ID:  e3d24d0a6048a826de5562d75dedb664d3a2a1b2
Gitweb:     http://git.kernel.org/tip/e3d24d0a6048a826de5562d75dedb664d3a2a1b2
Author:     Srikar Dronamraju <srikar@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 22:51:42 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 4 Jul 2015 10:04:32 +0200

sched/numa: Show numa_group ID in /proc/sched_debug task listings

Having the numa group ID in /proc/sched_debug helps to see how
the numa groups have spread across the system.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-3-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 315c68e..f1dcd1d 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -142,7 +142,7 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p)
 		0LL, 0L);
 #endif
 #ifdef CONFIG_NUMA_BALANCING
-	SEQ_printf(m, " %d", task_node(p));
+	SEQ_printf(m, " %d %d", task_node(p), task_numa_group_id(p));
 #endif
 #ifdef CONFIG_CGROUP_SCHED
 	SEQ_printf(m, " %s", task_group_path(task_group(p)));

[PATCH 3/3] sched/numa: Fix numabalancing stats in /proc/pid/sched

[Srikar Dronamraju]

Commit 44dba3d5d6a1 (sched: Refactor task_struct to use numa_faults
instead of numa_* pointers) modified the way tsk->numa_faults stats are
accounted. However that commit never touched show_numa_stats() that
displays /proc/pid/sched. Now the numbers displayed in /proc/pid/sched
dont match the actual numbers.

Fix it by making sure that /proc/pid/sched reflects the task fault numbers
Also add group fault stats too.

Also couple of more modifications are added here.
1. Format change.
- Previously we would list two entries per node, one for private and one for
  shared. Also the home node info was listed in each entry.

- Now preferred node, total_faults and current node are displayed separately.
- Now there is one entry per node, that lists private,shared task and group
  faults.

2. p->numa_pages_migrated was getting reset after every read of
   /proc/pid/sched. Its more useful to have absolute numbers since
   differential migrations between two accesses can be easily calculated.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
---
 kernel/sched/debug.c | 38 +++++++++++++++++---------------------
 kernel/sched/fair.c  | 22 +++++++++++++++++++++-
 kernel/sched/sched.h | 10 +++++++++-
 3 files changed, 47 insertions(+), 23 deletions(-)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index f1dcd1d..8b974ed 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -517,11 +517,21 @@ __initcall(init_sched_debug_procfs);
 	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F))
 
 
+#ifdef CONFIG_NUMA_BALANCING
+void print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
+		unsigned long tpf, unsigned long gsf, unsigned long gpf)
+{
+	SEQ_printf(m, "numa_faults node=%d ", node);
+	SEQ_printf(m, "task_private=%lu task_shared=%lu ", tsf, tpf);
+	SEQ_printf(m, "group_private=%lu group_shared=%lu\n", gsf, gpf);
+}
+#endif
+
+
 static void sched_show_numa(struct task_struct *p, struct seq_file *m)
 {
 #ifdef CONFIG_NUMA_BALANCING
 	struct mempolicy *pol;
-	int node, i;
 
 	if (p->mm)
 		P(mm->numa_scan_seq);
@@ -533,26 +543,12 @@ static void sched_show_numa(struct task_struct *p, struct seq_file *m)
 	mpol_get(pol);
 	task_unlock(p);
 
-	SEQ_printf(m, "numa_migrations, %ld\n", xchg(&p->numa_pages_migrated, 0));
-
-	for_each_online_node(node) {
-		for (i = 0; i < 2; i++) {
-			unsigned long nr_faults = -1;
-			int cpu_current, home_node;
-
-			if (p->numa_faults)
-				nr_faults = p->numa_faults[2*node + i];
-
-			cpu_current = !i ? (task_node(p) == node) :
-				(pol && node_isset(node, pol->v.nodes));
-
-			home_node = (p->numa_preferred_nid == node);
-
-			SEQ_printf(m, "numa_faults_memory, %d, %d, %d, %d, %ld\n",
-				i, node, cpu_current, home_node, nr_faults);
-		}
-	}
-
+	P(numa_pages_migrated);
+	P(numa_preferred_nid);
+	P(total_numa_faults);
+	SEQ_printf(m, "current_node=%d, numa_group_id=%d\n",
+			task_node(p), task_numa_group_id(p));
+	show_numa_stats(p, m);
 	mpol_put(pol);
 #endif
 }
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 3d57cc0..e31ece1 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -8473,7 +8473,27 @@ void print_cfs_stats(struct seq_file *m, int cpu)
 		print_cfs_rq(m, cpu, cfs_rq);
 	rcu_read_unlock();
 }
-#endif
+
+#ifdef CONFIG_NUMA_BALANCING
+void show_numa_stats(struct task_struct *p, struct seq_file *m)
+{
+	int node;
+	unsigned long tsf = 0, tpf = 0, gsf = 0, gpf = 0;
+
+	for_each_online_node(node) {
+		if (p->numa_faults) {
+			tsf = p->numa_faults[task_faults_idx(NUMA_MEM, node, 0)];
+			tpf = p->numa_faults[task_faults_idx(NUMA_MEM, node, 1)];
+		}
+		if (p->numa_group) {
+			gsf = p->numa_group->faults[task_faults_idx(NUMA_MEM, node, 0)],
+			gpf = p->numa_group->faults[task_faults_idx(NUMA_MEM, node, 1)];
+		}
+		print_numa_stats(m, node, tsf, tpf, gsf, gpf);
+	}
+}
+#endif /* CONFIG_NUMA_BALANCING */
+#endif /* CONFIG_SCHED_DEBUG */
 
 __init void init_sched_fair_class(void)
 {
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index c901c9c..84d4879 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1696,7 +1696,15 @@ extern void print_rt_stats(struct seq_file *m, int cpu);
 extern void print_dl_stats(struct seq_file *m, int cpu);
 extern void
 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
-#endif
+
+#ifdef CONFIG_NUMA_BALANCING
+extern void
+show_numa_stats(struct task_struct *p, struct seq_file *m);
+extern void
+print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
+	unsigned long tpf, unsigned long gsf, unsigned long gpf);
+#endif /* CONFIG_NUMA_BALANCING */
+#endif /* CONFIG_SCHED_DEBUG */
 
 extern void init_cfs_rq(struct cfs_rq *cfs_rq);
 extern void init_rt_rq(struct rt_rq *rt_rq);
-- 
1.8.3.1

Re: [PATCH 3/3] sched/numa: Fix numabalancing stats in /proc/pid/sched

[Rik van Riel]

On 06/25/2015 01:21 PM, Srikar Dronamraju wrote:
> Commit 44dba3d5d6a1 (sched: Refactor task_struct to use numa_faults
> instead of numa_* pointers) modified the way tsk->numa_faults stats are
> accounted. However that commit never touched show_numa_stats() that
> displays /proc/pid/sched. Now the numbers displayed in /proc/pid/sched
> dont match the actual numbers.
> 
> Fix it by making sure that /proc/pid/sched reflects the task fault numbers
> Also add group fault stats too.
> 
> Also couple of more modifications are added here.
> 1. Format change.
> - Previously we would list two entries per node, one for private and one for
>   shared. Also the home node info was listed in each entry.
> 
> - Now preferred node, total_faults and current node are displayed separately.
> - Now there is one entry per node, that lists private,shared task and group
>   faults.
> 
> 2. p->numa_pages_migrated was getting reset after every read of
>    /proc/pid/sched. Its more useful to have absolute numbers since
>    differential migrations between two accesses can be easily calculated.
> 
> Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>

Acked-by: Rik van Riel <riel@redhat.com>


Thank you for updating the statistics to match what
the code does nowadays.

[tip:sched/urgent] sched/numa: Fix numa balancing stats in /proc/ pid/sched

[tip-bot for Srikar Dronamraju]

Commit-ID:  46fb6b9762d13fdeeefb3a9e9c317ebcd344f578
Gitweb:     http://git.kernel.org/tip/46fb6b9762d13fdeeefb3a9e9c317ebcd344f578
Author:     Srikar Dronamraju <srikar@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 22:51:43 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 1 Jul 2015 10:31:28 +0200

sched/numa: Fix numa balancing stats in /proc/pid/sched

Commit 44dba3d5d6a1 ("sched: Refactor task_struct to use
numa_faults instead of numa_* pointers") modified the way
tsk->numa_faults stats are accounted.

However that commit never touched show_numa_stats() that is displayed
in /proc/pid/sched and thus the numbers displayed in /proc/pid/sched
don't match the actual numbers.

Fix it by making sure that /proc/pid/sched reflects the task
fault numbers. Also add group fault stats too.

Also couple of more modifications are added here:

1. Format changes:

  - Previously we would list two entries per node, one for private
    and one for shared. Also the home node info was listed in each entry.

  - Now preferred node, total_faults and current node are
    displayed separately.

  - Now there is one entry per node, that lists private,shared task and
    group faults.

2. Unit changes:

  - p->numa_pages_migrated was getting reset after every read of
    /proc/pid/sched. It's more useful to have absolute numbers since
    differential migrations between two accesses can be more easily
    calculated.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-4-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/debug.c | 38 +++++++++++++++++---------------------
 kernel/sched/fair.c  | 22 +++++++++++++++++++++-
 kernel/sched/sched.h | 10 +++++++++-
 3 files changed, 47 insertions(+), 23 deletions(-)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index f1dcd1d..4222ec5 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -517,11 +517,21 @@ __initcall(init_sched_debug_procfs);
 	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F))
 
 
+#ifdef CONFIG_NUMA_BALANCING
+void print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
+		unsigned long tpf, unsigned long gsf, unsigned long gpf)
+{
+	SEQ_printf(m, "numa_faults node=%d ", node);
+	SEQ_printf(m, "task_private=%lu task_shared=%lu ", tsf, tpf);
+	SEQ_printf(m, "group_private=%lu group_shared=%lu\n", gsf, gpf);
+}
+#endif
+
+
 static void sched_show_numa(struct task_struct *p, struct seq_file *m)
 {
 #ifdef CONFIG_NUMA_BALANCING
 	struct mempolicy *pol;
-	int node, i;
 
 	if (p->mm)
 		P(mm->numa_scan_seq);
@@ -533,26 +543,12 @@ static void sched_show_numa(struct task_struct *p, struct seq_file *m)
 	mpol_get(pol);
 	task_unlock(p);
 
-	SEQ_printf(m, "numa_migrations, %ld\n", xchg(&p->numa_pages_migrated, 0));
-
-	for_each_online_node(node) {
-		for (i = 0; i < 2; i++) {
-			unsigned long nr_faults = -1;
-			int cpu_current, home_node;
-
-			if (p->numa_faults)
-				nr_faults = p->numa_faults[2*node + i];
-
-			cpu_current = !i ? (task_node(p) == node) :
-				(pol && node_isset(node, pol->v.nodes));
-
-			home_node = (p->numa_preferred_nid == node);
-
-			SEQ_printf(m, "numa_faults_memory, %d, %d, %d, %d, %ld\n",
-				i, node, cpu_current, home_node, nr_faults);
-		}
-	}
-
+	P(numa_pages_migrated);
+	P(numa_preferred_nid);
+	P(total_numa_faults);
+	SEQ_printf(m, "current_node=%d, numa_group_id=%d\n",
+			task_node(p), task_numa_group_id(p));
+	show_numa_stats(p, m);
 	mpol_put(pol);
 #endif
 }
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 40a7fcb..7245039 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -8468,7 +8468,27 @@ void print_cfs_stats(struct seq_file *m, int cpu)
 		print_cfs_rq(m, cpu, cfs_rq);
 	rcu_read_unlock();
 }
-#endif
+
+#ifdef CONFIG_NUMA_BALANCING
+void show_numa_stats(struct task_struct *p, struct seq_file *m)
+{
+	int node;
+	unsigned long tsf = 0, tpf = 0, gsf = 0, gpf = 0;
+
+	for_each_online_node(node) {
+		if (p->numa_faults) {
+			tsf = p->numa_faults[task_faults_idx(NUMA_MEM, node, 0)];
+			tpf = p->numa_faults[task_faults_idx(NUMA_MEM, node, 1)];
+		}
+		if (p->numa_group) {
+			gsf = p->numa_group->faults[task_faults_idx(NUMA_MEM, node, 0)],
+			gpf = p->numa_group->faults[task_faults_idx(NUMA_MEM, node, 1)];
+		}
+		print_numa_stats(m, node, tsf, tpf, gsf, gpf);
+	}
+}
+#endif /* CONFIG_NUMA_BALANCING */
+#endif /* CONFIG_SCHED_DEBUG */
 
 __init void init_sched_fair_class(void)
 {
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 7d58952..7ef5968 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1675,7 +1675,15 @@ extern void print_rt_stats(struct seq_file *m, int cpu);
 extern void print_dl_stats(struct seq_file *m, int cpu);
 extern void
 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
-#endif
+
+#ifdef CONFIG_NUMA_BALANCING
+extern void
+show_numa_stats(struct task_struct *p, struct seq_file *m);
+extern void
+print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
+	unsigned long tpf, unsigned long gsf, unsigned long gpf);
+#endif /* CONFIG_NUMA_BALANCING */
+#endif /* CONFIG_SCHED_DEBUG */
 
 extern void init_cfs_rq(struct cfs_rq *cfs_rq);
 extern void init_rt_rq(struct rt_rq *rt_rq);

[tip:sched/urgent] sched/numa: Fix numa balancing stats in /proc/ pid/sched

[tip-bot for Srikar Dronamraju]

Commit-ID:  397f2378f136128623fc237746157aa2564d1082
Gitweb:     http://git.kernel.org/tip/397f2378f136128623fc237746157aa2564d1082
Author:     Srikar Dronamraju <srikar@linux.vnet.ibm.com>
AuthorDate: Thu, 25 Jun 2015 22:51:43 +0530
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 4 Jul 2015 10:04:33 +0200

sched/numa: Fix numa balancing stats in /proc/pid/sched

Commit 44dba3d5d6a1 ("sched: Refactor task_struct to use
numa_faults instead of numa_* pointers") modified the way
tsk->numa_faults stats are accounted.

However that commit never touched show_numa_stats() that is displayed
in /proc/pid/sched and thus the numbers displayed in /proc/pid/sched
don't match the actual numbers.

Fix it by making sure that /proc/pid/sched reflects the task
fault numbers. Also add group fault stats too.

Also couple of more modifications are added here:

1. Format changes:

  - Previously we would list two entries per node, one for private
    and one for shared. Also the home node info was listed in each entry.

  - Now preferred node, total_faults and current node are
    displayed separately.

  - Now there is one entry per node, that lists private,shared task and
    group faults.

2. Unit changes:

  - p->numa_pages_migrated was getting reset after every read of
    /proc/pid/sched. It's more useful to have absolute numbers since
    differential migrations between two accesses can be more easily
    calculated.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-4-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/debug.c | 38 +++++++++++++++++---------------------
 kernel/sched/fair.c  | 22 +++++++++++++++++++++-
 kernel/sched/sched.h | 10 +++++++++-
 3 files changed, 47 insertions(+), 23 deletions(-)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index f1dcd1d..4222ec5 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -517,11 +517,21 @@ __initcall(init_sched_debug_procfs);
 	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F))
 
 
+#ifdef CONFIG_NUMA_BALANCING
+void print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
+		unsigned long tpf, unsigned long gsf, unsigned long gpf)
+{
+	SEQ_printf(m, "numa_faults node=%d ", node);
+	SEQ_printf(m, "task_private=%lu task_shared=%lu ", tsf, tpf);
+	SEQ_printf(m, "group_private=%lu group_shared=%lu\n", gsf, gpf);
+}
+#endif
+
+
 static void sched_show_numa(struct task_struct *p, struct seq_file *m)
 {
 #ifdef CONFIG_NUMA_BALANCING
 	struct mempolicy *pol;
-	int node, i;
 
 	if (p->mm)
 		P(mm->numa_scan_seq);
@@ -533,26 +543,12 @@ static void sched_show_numa(struct task_struct *p, struct seq_file *m)
 	mpol_get(pol);
 	task_unlock(p);
 
-	SEQ_printf(m, "numa_migrations, %ld\n", xchg(&p->numa_pages_migrated, 0));
-
-	for_each_online_node(node) {
-		for (i = 0; i < 2; i++) {
-			unsigned long nr_faults = -1;
-			int cpu_current, home_node;
-
-			if (p->numa_faults)
-				nr_faults = p->numa_faults[2*node + i];
-
-			cpu_current = !i ? (task_node(p) == node) :
-				(pol && node_isset(node, pol->v.nodes));
-
-			home_node = (p->numa_preferred_nid == node);
-
-			SEQ_printf(m, "numa_faults_memory, %d, %d, %d, %d, %ld\n",
-				i, node, cpu_current, home_node, nr_faults);
-		}
-	}
-
+	P(numa_pages_migrated);
+	P(numa_preferred_nid);
+	P(total_numa_faults);
+	SEQ_printf(m, "current_node=%d, numa_group_id=%d\n",
+			task_node(p), task_numa_group_id(p));
+	show_numa_stats(p, m);
 	mpol_put(pol);
 #endif
 }
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 40a7fcb..7245039 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -8468,7 +8468,27 @@ void print_cfs_stats(struct seq_file *m, int cpu)
 		print_cfs_rq(m, cpu, cfs_rq);
 	rcu_read_unlock();
 }
-#endif
+
+#ifdef CONFIG_NUMA_BALANCING
+void show_numa_stats(struct task_struct *p, struct seq_file *m)
+{
+	int node;
+	unsigned long tsf = 0, tpf = 0, gsf = 0, gpf = 0;
+
+	for_each_online_node(node) {
+		if (p->numa_faults) {
+			tsf = p->numa_faults[task_faults_idx(NUMA_MEM, node, 0)];
+			tpf = p->numa_faults[task_faults_idx(NUMA_MEM, node, 1)];
+		}
+		if (p->numa_group) {
+			gsf = p->numa_group->faults[task_faults_idx(NUMA_MEM, node, 0)],
+			gpf = p->numa_group->faults[task_faults_idx(NUMA_MEM, node, 1)];
+		}
+		print_numa_stats(m, node, tsf, tpf, gsf, gpf);
+	}
+}
+#endif /* CONFIG_NUMA_BALANCING */
+#endif /* CONFIG_SCHED_DEBUG */
 
 __init void init_sched_fair_class(void)
 {
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 7d58952..7ef5968 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1675,7 +1675,15 @@ extern void print_rt_stats(struct seq_file *m, int cpu);
 extern void print_dl_stats(struct seq_file *m, int cpu);
 extern void
 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
-#endif
+
+#ifdef CONFIG_NUMA_BALANCING
+extern void
+show_numa_stats(struct task_struct *p, struct seq_file *m);
+extern void
+print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
+	unsigned long tpf, unsigned long gsf, unsigned long gpf);
+#endif /* CONFIG_NUMA_BALANCING */
+#endif /* CONFIG_SCHED_DEBUG */
 
 extern void init_cfs_rq(struct cfs_rq *cfs_rq);
 extern void init_rt_rq(struct rt_rq *rt_rq);