Re: [PATCH] cpu/hotplug: Fix acquire of st->should_run in cpuhp_thread_fun()

[Peter Zijlstra]

On Wed, Sep 05, 2018 at 11:22:07AM +0530, Neeraj Upadhyay wrote:
> The smp_mb() in cpuhp_thread_fun() appears to be misplaced, and
> need to be after the load of st->should_run, to prevent
> reordering of the later load/stores w.r.t. the load of
> st->should_run.

I'm tempted to say this changelog needs improvement, but given the
comment that goes with the smp_mb(), yes this is obviously right.

Acked-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxx>

> Signed-off-by: Neeraj Upadhyay <neeraju@xxxxxxxxxxxxxx>
> ---
>  kernel/cpu.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/cpu.c b/kernel/cpu.c
> index aa7fe85..eb4041f 100644
> --- a/kernel/cpu.c
> +++ b/kernel/cpu.c
> @@ -607,15 +607,15 @@ static void cpuhp_thread_fun(unsigned int cpu)
>  	bool bringup = st->bringup;
>  	enum cpuhp_state state;
>  
> +	if (WARN_ON_ONCE(!st->should_run))
> +		return;
> +
>  	/*
>  	 * ACQUIRE for the cpuhp_should_run() load of ->should_run. Ensures
>  	 * that if we see ->should_run we also see the rest of the state.
>  	 */
>  	smp_mb();
>  
> -	if (WARN_ON_ONCE(!st->should_run))
> -		return;
> -
>  	cpuhp_lock_acquire(bringup);
>  
>  	if (st->single) {