Re: [PATCH RT 1/8] sched: migrate_enable: Use sleeping_lock to indicate involuntary sleep

[Steven Rostedt <rostedt@xxxxxxxxxxx>]

On Sat, 27 Jul 2019 00:56:31 -0500
Scott Wood <swood@xxxxxxxxxx> wrote:

> Without this, rcu_note_context_switch() will complain if an RCU read
> lock is held when migrate_enable() calls stop_one_cpu().
> 
> Signed-off-by: Scott Wood <swood@xxxxxxxxxx>
>

> index d3c6542b306f..c3407707e367 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -7405,7 +7405,9 @@ void migrate_enable(void)
>  			unpin_current_cpu();
>  			preempt_lazy_enable();
>  			preempt_enable();

This looks like it needs a comment to explain why this is.

-- Steve

> +			sleeping_lock_inc();
>  			stop_one_cpu(task_cpu(p), migration_cpu_stop, &arg);
> +			sleeping_lock_dec();
>  			return;
>  		}
>  	}