Re: [PATCH v3 2/4] sched/core: Avoid double calling update_rq_clock() in __balance_push_cpu_stop()

[Vincent Guittot]

On Mon, 15 May 2023 at 08:39, Hao Jia <jiahao.os@xxxxxxxxxxxxx> wrote:
>
> The WARN_DOUBLE_CLOCK warning is triggered during cpu offline.
> ------------[ cut here ]------------
> rq->clock_update_flags & RQCF_UPDATED
> WARNING: CPU: 17 PID: 138 at kernel/sched/core.c:741
> update_rq_clock+0xaf/0x180
> Call Trace:
>  <TASK>
>  __balance_push_cpu_stop+0x146/0x180
>  ? migration_cpu_stop+0x2a0/0x2a0
>  cpu_stopper_thread+0xa3/0x140
>  smpboot_thread_fn+0x14f/0x210
>  ? sort_range+0x20/0x20
>  kthread+0xe6/0x110
>  ? kthread_complete_and_exit+0x20/0x20
>  ret_from_fork+0x1f/0x30
>
> To avoid this warning, we remove update_rq_clock() from
> the __migrate_task() function. And in order to avoid
> missing rq clock update, add update_rq_clock() call before
> migration_cpu_stop() calls __migrate_task().
>
> Signed-off-by: Hao Jia <jiahao.os@xxxxxxxxxxxxx>

Reviewed-by: Vincent Guittot <vincent.guittot@xxxxxxxxxx>

> ---
>  kernel/sched/core.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 944c3ae39861..2bcf6a8e71fc 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -2398,7 +2398,6 @@ static struct rq *__migrate_task(struct rq *rq, struct rq_flags *rf,
>         if (!is_cpu_allowed(p, dest_cpu))
>                 return rq;
>
> -       update_rq_clock(rq);
>         rq = move_queued_task(rq, rf, p, dest_cpu);
>
>         return rq;
> @@ -2456,10 +2455,12 @@ static int migration_cpu_stop(void *data)
>                                 goto out;
>                 }
>
> -               if (task_on_rq_queued(p))
> +               if (task_on_rq_queued(p)) {
> +                       update_rq_clock(rq);
>                         rq = __migrate_task(rq, &rf, p, arg->dest_cpu);
> -               else
> +               } else {
>                         p->wake_cpu = arg->dest_cpu;
> +               }
>
>                 /*
>                  * XXX __migrate_task() can fail, at which point we might end
> --
> 2.37.0
>