On Tue, Jan 07, 2025 at 09:18:48AM +0800, Huacai Chen wrote:
> Hi, all, I like simple fixes, so is this acceptable (based on an early
> version from Koichiro Den)?
No not at all. This is bizarre - in the mail you are replying to Koichiro
agrees with me that the approach of his code that you've sent here (I don't
know why) is fundamentally broken and suggest another.
I am at a loss as to why you've sent this? Perhaps a miscommunication
somewhere? :)
In any case, please don't send '[PATCH] xxx' mails that aren't intended to
be patches, a better form of this would be to say 'oh can we just do...'
then to put this code in the mail underneath, without any '[PATCH]' prefix.
But please do review the discussion - the below is insufficient as simple
as it is (sadly) because the boot CPU's delayed work will never be
executed.
I will take a look at Koichiro's new approach as soon as I am able.
Cheers!
> ---
> mm/vmstat.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/mm/vmstat.c b/mm/vmstat.c
> index 0889b75cef14..1badc24a21ff 100644
> --- a/mm/vmstat.c
> +++ b/mm/vmstat.c
> @@ -2122,10 +2122,15 @@ static void __init start_shepherd_timer(void)
> {
> int cpu;
>
> - for_each_possible_cpu(cpu)
> + for_each_possible_cpu(cpu) {
> INIT_DEFERRABLE_WORK(per_cpu_ptr(&vmstat_work, cpu),
> vmstat_update);
>
> + /* Will be enabled on vmstat_cpu_online() */
> + if (!cpu_online(cpu))
> + disable_delayed_work_sync(&per_cpu(vmstat_work, cpu));
> + }
> +
> schedule_delayed_work(&shepherd,
> round_jiffies_relative(sysctl_stat_interval));
> }
> --
> 2.43.5
>
