linux-kernel - Re: [PATCH] sched: move access of avg_rt and avg

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <85230ed0-26a0-4f08-aab0-f0a6ce03abe8@linux.vnet.ibm.com>
Date: Wed, 20 Dec 2023 20:18:44 +0530
From: Shrikanth Hegde <sshegde@...ux.vnet.ibm.com>
To: Vincent Guittot <vincent.guittot@...aro.org>
Cc: mingo@...nel.org, peterz@...radead.org, dietmar.eggemann@....com,
        linux-kernel@...r.kernel.org, srikar@...ux.vnet.ibm.com,
        yu.c.chen@...el.com, tim.c.chen@...ux.intel.com
Subject: Re: [PATCH] sched: move access of avg_rt and avg_dl into existing
 helper functions



On 12/20/23 7:29 PM, Vincent Guittot wrote:

Hi Vincent, thanks for taking a look.

> On Wed, 20 Dec 2023 at 07:55, Shrikanth Hegde
> <sshegde@...ux.vnet.ibm.com> wrote:
>>
>> This is a minor code simplification. There are helper functions called
>> cpu_util_dl and cpu_util_rt which gives the average utilization of DL
>> and RT respectively. But there are few places in code where these
>> variables are used directly.
>>
>> Instead use the helper function so that code becomes simpler and easy to
>> maintain later on.
>>
>> Signed-off-by: Shrikanth Hegde <sshegde@...ux.vnet.ibm.com>
>> ---
>>  kernel/sched/fair.c | 12 +++++-------
>>  1 file changed, 5 insertions(+), 7 deletions(-)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index bcea3d55d95d..02631060ca7e 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -9212,19 +9212,17 @@ static inline bool cfs_rq_has_blocked(struct cfs_rq *cfs_rq)
>>
>>  static inline bool others_have_blocked(struct rq *rq)
>>  {
>> -       if (READ_ONCE(rq->avg_rt.util_avg))
>> +       if (cpu_util_rt(rq))
>>                 return true;
>>
>> -       if (READ_ONCE(rq->avg_dl.util_avg))
>> +       if (cpu_util_dl(rq))
>>                 return true;
>>
>>         if (thermal_load_avg(rq))
>>                 return true;
>>
>> -#ifdef CONFIG_HAVE_SCHED_AVG_IRQ
>> -       if (READ_ONCE(rq->avg_irq.util_avg))
>> +       if (cpu_util_irq(rq))
> 
> cpu_util_irq doesn't call READ_ONCE()
> 


I see. Actually it would be right if cpu_util_irq does call READ_ONCE no? 

Sorry i havent yet understood the memory barriers in details. Please correct me 
if i am wrong here, 
since ___update_load_avg(&rq->avg_irq, 1) does use WRITE_ONCE and reading out this 
value using cpu_util_irq on a different CPU should use READ_ONCE no? 

> 
>>                 return true;
>> -#endif
>>
>>         return false;
>>  }
>> @@ -9481,8 +9479,8 @@ static unsigned long scale_rt_capacity(int cpu)
>>          * avg_thermal.load_avg tracks thermal pressure and the weighted
>>          * average uses the actual delta max capacity(load).
>>          */
>> -       used = READ_ONCE(rq->avg_rt.util_avg);
>> -       used += READ_ONCE(rq->avg_dl.util_avg);
>> +       used = cpu_util_rt(rq);
>> +       used += cpu_util_dl(rq);
>>         used += thermal_load_avg(rq);
>>
>>         if (unlikely(used >= max))
>> --
>> 2.39.3
>>