lkml.org 
[lkml]   [2011]   [Mar]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: [patch 02/15] sched: validate CFS quota hierarchies
(2011/03/23 12:03), Paul Turner wrote:
> @@ -9251,7 +9255,13 @@ static int tg_set_cfs_bandwidth(struct t
> if (period > max_cfs_quota_period)
> return -EINVAL;
>
> - mutex_lock(&mutex);
> + mutex_lock(&cfs_constraints_mutex);
> + if (sysctl_sched_cfs_bandwidth_consistent) {
> + ret = __cfs_schedulable(tg, period, quota);

At this point:
period => scale in ns unit
quota => scale in ns unit, or RUNTIME_INF

And both are unsigned. But...

> @@ -9339,6 +9350,108 @@ static int cpu_cfs_period_write_u64(stru
> return tg_set_cfs_period(cgroup_tg(cgrp), cfs_period_us);
> }
>
> +
> +struct cfs_schedulable_data {
> + struct task_group *tg;
> + u64 period, quota;
> +};
> +
> +/*
> + * normalize group quota/period to be quota/max_period
> + * note: units are usecs
> + */
> +static u64 normalize_cfs_quota(struct task_group *tg,
> + struct cfs_schedulable_data *d)
> +{
> + u64 quota, period;
> + struct load_weight lw;
> +
> + if (tg == d->tg) {
> + period = d->period;
> + quota = d->quota;
> + } else {
> + period = tg_get_cfs_period(tg);
> + quota = tg_get_cfs_quota(tg);
> + }

... at this point:
period => scale in us unit
quota => scale in us unit, or -1
Moreover:
d->period => (scale in ns unit) / NSEC_PER_USEC
d->quota => (scale in ns unit, or RUNTIME_INF) / NSEC_PER_USEC

Therefore, ...

> +
> + if (quota == RUNTIME_INF)
> + return RUNTIME_INF;

This check doesn't work properly.

I found this problem because I could not get child group back to be
unconstrained:

[root@localhost group0]# cat cpu.cfs_*
500000
500000
[root@localhost group0]# cat sub0/cpu.cfs_*
500000
100000
[root@localhost group0]# cat sub1/cpu.cfs_*
500000
100000
[root@localhost group0]# echo -1 > sub1/cpu.cfs_quota_us
bash: echo: write error: Invalid argument

I confirmed that this write error is removed by the following
change. I'm looking forward to seeing your V6 soon.

Reviewed-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>


Thanks,
H.Seto

---
kernel/sched.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/kernel/sched.c b/kernel/sched.c
index 6d764b5..c8f9820 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -9467,8 +9467,8 @@ static u64 normalize_cfs_quota(struct task_group *tg,
period = d->period;
quota = d->quota;
} else {
- period = tg_get_cfs_period(tg);
- quota = tg_get_cfs_quota(tg);
+ period = ktime_to_ns(tg_cfs_bandwidth(tg)->period);
+ quota = tg_cfs_bandwidth(tg)->quota;
}

if (quota == RUNTIME_INF)
@@ -9515,8 +9515,8 @@ static int __cfs_schedulable(struct task_group *tg, u64 period, u64 quota)
int ret;
struct cfs_schedulable_data data = {
.tg = tg,
- .period = period / NSEC_PER_USEC,
- .quota = quota / NSEC_PER_USEC,
+ .period = period,
+ .quota = quota,
};

if (!sysctl_sched_cfs_bandwidth_consistent)
--
1.7.4


\
 
 \ /
  Last update: 2011-03-29 09:01    [W:0.304 / U:1.840 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site