Re: [PATCH v3] cpuacct: VIRT_CPU_ACCOUNTING don't prevent percpu cputime count

[KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>]

> 
> * KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:
> 
> > 
> > Changelog:
> >   since v2
> >   - revert using percpu_counter_sum()
> > 
> >   since v1
> >   - use percpu_counter_sum() instead percpu_counter_read()
> > 
> > 
> > -------------------------------------
> > Subject: [PATCH v3] cpuacct: VIRT_CPU_ACCOUNTING don't prevent percpu cputime count
> > 
> > cpuacct_update_stats() is called at every tick updating. and it use percpu_counter
> > for avoiding performance degression.
> > 
> > For archs which define VIRT_CPU_ACCOUNTING, every tick would result
> > in >1000 units of cputime updates and since this is much much greater
> > than percpu_batch_counter, we end up taking spinlock on every tick.
> > 
> > This patch change batch rule. now, any cpu can store "percpu_counter_bach * jiffies"
> > cputime in per-cpu cache.
> > it mean this patch don't have behavior change if VIRT_CPU_ACCOUNTING=n.
> > 
> > Cc: Bharata B Rao <bharata@linux.vnet.ibm.com>
> > Cc: Balaji Rao <balajirrao@gmail.com>
> > Cc: Dhaval Giani <dhaval@linux.vnet.ibm.com>
> > Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
> > Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
> > Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
> > Cc: Ingo Molnar <mingo@elte.hu>
> > Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
> > Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> > ---
> >  kernel/sched.c |    6 +++++-
> >  1 file changed, 5 insertions(+), 1 deletion(-)
> > 
> > Index: b/kernel/sched.c
> > ===================================================================
> > --- a/kernel/sched.c	2009-04-30 11:37:47.000000000 +0900
> > +++ b/kernel/sched.c	2009-05-07 16:34:00.000000000 +0900
> > @@ -10221,6 +10221,7 @@ struct cpuacct {
> >  };
> >  
> >  struct cgroup_subsys cpuacct_subsys;
> > +static s32 cpuacct_batch;
> 
> should be __read_mostly i guess.
> 
> >  
> >  /* return cpu accounting group corresponding to this container */
> >  static inline struct cpuacct *cgroup_ca(struct cgroup *cgrp)
> > @@ -10250,6 +10251,9 @@ static struct cgroup_subsys_state *cpuac
> >  	if (!ca->cpuusage)
> >  		goto out_free_ca;
> >  
> > +	if (!cpuacct_batch)
> > +		cpuacct_batch = jiffies_to_cputime(percpu_counter_batch);
> 
> So this is a once per boot condition? Why not initialize it in 
> sched_init() or so? (instead of incuring this ugly check all the 
> time)

Thanks! fixed.



Changelog:
  since V3
  - rewirte patch description (thanks Bharata!)
  - append read_mostly to cpuacct_batch
  - cpuacct_batch is initialized by sched_init()

  since v2
  - revert using percpu_counter_sum()

  since v1
  - use percpu_counter_sum() instead percpu_counter_read()


---------------------------------------------------------
Subject: [PATCH v4] cpuacct: Use bigger percpu counter batch values for stats counters on archs that have VIRT_CPU_ACCOUNTING=y

percpu counters used to accumulate statistics in cpuacct controller use
the default batch value [max(2*nr_cpus, 32)] which can be too small for
archs that define VIRT_CPU_ACCOUNTING. In such archs, a tick could result in
cputime updates in the range of thousands. As a result, cpuacct_update_stats()
would end up acquiring the percpu counter spinlock on every tick which
is not good for performance.

Let those architectures to have a bigger batch threshold so that percpu counter
spinlock isn't taken on every tick. This change doesn't affect the archs which
don't define VIRT_CPU_ACCOUNTING and they continue to have the default
percpu counter batch value.

Cc: Balaji Rao <balajirrao@gmail.com>
Cc: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
---
 kernel/sched.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

Index: b/kernel/sched.c
===================================================================
--- a/kernel/sched.c	2009-05-09 16:48:08.000000000 +0900
+++ b/kernel/sched.c	2009-05-09 18:16:00.000000000 +0900
@@ -824,8 +824,12 @@ static struct file_operations sched_feat
 	.release	= single_release,
 };
 
+static __read_mostly s32 cpuacct_batch;
+
 static __init int sched_init_debug(void)
 {
+	cpuacct_batch = jiffies_to_cputime(percpu_counter_batch);
+
 	debugfs_create_file("sched_features", 0644, NULL, NULL,
 			&sched_feat_fops);
 
@@ -10457,7 +10461,8 @@ static void cpuacct_update_stats(struct 
 	ca = task_ca(tsk);
 
 	do {
-		percpu_counter_add(&ca->cpustat[idx], val);
+		__percpu_counter_add(&ca->cpustat[idx], val, cpuacct_batch);
+
 		ca = ca->parent;
 	} while (ca);
 	rcu_read_unlock();