Re: [PATCH] rcu: don't use __this_cpu_* from preemptible context

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Sun, Dec 28, 2014 at 11:49:40AM -0500, Sasha Levin wrote:
> Commit "rcu: Make cond_resched_rcu_qs() apply to normal RCU flavors" has added
> debug code which tries to read from __this_cpu in a preemptible code, which
> doesn't work too well.
> 
> Signed-off-by: Sasha Levin <sasha.levin@oracle.com>

Good catch, queued for 3.20 (same as the offending commit).

							Thanx, Paul

> ---
>  kernel/rcu/tree_trace.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/rcu/tree_trace.c b/kernel/rcu/tree_trace.c
> index a9b79ca..fbb6240 100644
> --- a/kernel/rcu/tree_trace.c
> +++ b/kernel/rcu/tree_trace.c
> @@ -122,7 +122,7 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp)
>  		   cpu_is_offline(rdp->cpu) ? '!' : ' ',
>  		   ulong2long(rdp->completed), ulong2long(rdp->gpnum),
>  		   rdp->passed_quiesce,
> -		   rdp->rcu_qs_ctr_snap == __this_cpu_read(rcu_qs_ctr),
> +		   rdp->rcu_qs_ctr_snap == per_cpu(rcu_qs_ctr, rdp->cpu),
>  		   rdp->qs_pending);
>  	seq_printf(m, " dt=%d/%llx/%d df=%lu",
>  		   atomic_read(&rdp->dynticks->dynticks),
> -- 
> 1.7.10.4
>