lkml.org 
[lkml]   [2022]   [May]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH] sche/fair: static cpumasks for load balance
From


On 5/26/22 01:04, Peter Zijlstra wrote:
> On Mon, May 23, 2022 at 05:42:16PM +0800, Bing Huang wrote:
>> The both cpu mask load_balance_mask and select_idle_mask just only used
>> in fair.c, but allocation in core.c in CONFIG_CPUMASK_OFFSTACK=y case,
>> and global via declare per cpu variations. More or less, it looks wired.
>>
>> Signed-off-by: Bing Huang <huangbing@kylinos.cn>
>> ---
>> kernel/sched/core.c | 13 +++----------
>> kernel/sched/fair.c | 18 ++++++++++++++++--
>> kernel/sched/sched.h | 4 ++++
>> 3 files changed, 23 insertions(+), 12 deletions(-)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index d58c0389eb23..64934c5e236d 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -9448,9 +9448,6 @@ LIST_HEAD(task_groups);
>> static struct kmem_cache *task_group_cache __read_mostly;
>> #endif
>>
>> -DECLARE_PER_CPU(cpumask_var_t, load_balance_mask);
>> -DECLARE_PER_CPU(cpumask_var_t, select_idle_mask);
>> -
>> void __init sched_init(void)
>> {
>> unsigned long ptr = 0;
>> @@ -9494,14 +9491,10 @@ void __init sched_init(void)
>>
>> #endif /* CONFIG_RT_GROUP_SCHED */
>> }
>> +
>> #ifdef CONFIG_CPUMASK_OFFSTACK
>> - for_each_possible_cpu(i) {
>> - per_cpu(load_balance_mask, i) = (cpumask_var_t)kzalloc_node(
>> - cpumask_size(), GFP_KERNEL, cpu_to_node(i));
>> - per_cpu(select_idle_mask, i) = (cpumask_var_t)kzalloc_node(
>> - cpumask_size(), GFP_KERNEL, cpu_to_node(i));
>> - }
>> -#endif /* CONFIG_CPUMASK_OFFSTACK */
>> + init_working_cpumask();
>> +#endif
> We already have init_sched_fair_class(), can't you use that?

OK, thanks for your review. I got it.

In addition, I notice that load_balance_mask and select_idle_mask are
defined
in CONFIG_SMP=y case in fair.c, so could be they moved inside
CONFIG_SMP, or just outside ?

>
>>
>> init_rt_bandwidth(&def_rt_bandwidth, global_rt_period(), global_rt_runtime());
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index a68482d66535..3a4b0dd1beb5 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -5831,8 +5831,8 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags)
>> #ifdef CONFIG_SMP
>>
>> /* Working cpumask for: load_balance, load_balance_newidle. */
>> -DEFINE_PER_CPU(cpumask_var_t, load_balance_mask);
>> -DEFINE_PER_CPU(cpumask_var_t, select_idle_mask);
>> +static DEFINE_PER_CPU(cpumask_var_t, load_balance_mask);
>> +static DEFINE_PER_CPU(cpumask_var_t, select_idle_mask);
>>
>> #ifdef CONFIG_NO_HZ_COMMON
>>
>> @@ -7052,6 +7052,20 @@ balance_fair(struct rq *rq, struct task_struct *prev, struct rq_flags *rf)
>> }
>> #endif /* CONFIG_SMP */
>>
>> +#ifdef CONFIG_CPUMASK_OFFSTACK
>> +void __init init_working_cpumask(void)
>> +{
>> + int i;
>> +
>> + for_each_possible_cpu(i) {
>> + per_cpu(load_balance_mask, i) = (cpumask_var_t)kzalloc_node(
>> + cpumask_size(), GFP_KERNEL, cpu_to_node(i));
>> + per_cpu(select_idle_mask, i) = (cpumask_var_t)kzalloc_node(
>> + cpumask_size(), GFP_KERNEL, cpu_to_node(i));
>> + }
>> +}
>> +#endif
> If you stick that #ifdef inside the function, you can remove it
> everwhere else. Less #ifdef is more better.
I will remove it, move them inside init_sched_fair_class().

\
 
 \ /
  Last update: 2022-05-26 11:05    [W:0.039 / U:1.172 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site