Re: [PATCH v2] cpuacct: VIRT_CPU_ACCOUNTING don't prevent percpu cputime count

[Balbir Singh <balbir@linux.vnet.ibm.com>]

* KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> [2009-04-30 15:11:15]:

> 
> Changelog:
>   since v1
>   - use percpu_counter_sum() instead percpu_counter_read()
> 
> 
> -------------------------------------
> Subject: [PATCH v2] 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 |    8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
> 
> Index: b/kernel/sched.c
> ===================================================================
> --- a/kernel/sched.c	2009-04-30 11:37:47.000000000 +0900
> +++ b/kernel/sched.c	2009-04-30 14:17:00.000000000 +0900
> @@ -10221,6 +10221,7 @@ struct cpuacct {
>  };
> 
>  struct cgroup_subsys cpuacct_subsys;
> +static s32 cpuacct_batch;
> 
>  /* 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);
> +
>  	for (i = 0; i < CPUACCT_STAT_NSTATS; i++)
>  		if (percpu_counter_init(&ca->cpustat[i], 0))
>  			goto out_free_counters;
> @@ -10376,7 +10380,7 @@ static int cpuacct_stats_show(struct cgr
>  	int i;
> 
>  	for (i = 0; i < CPUACCT_STAT_NSTATS; i++) {
> -		s64 val = percpu_counter_read(&ca->cpustat[i]);
> +		s64 val = percpu_counter_sum(&ca->cpustat[i]);
>  		val = cputime64_to_clock_t(val);
>  		cb->fill(cb, cpuacct_stat_desc[i], val);
>  	}
> @@ -10446,7 +10450,7 @@ 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();
> 
>

What do the test results look like with this? I'll see if I can find
some time to test this patch. On a patch read level this seems much better
to me, Peter?

Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>

-- 
	Balbir