Re: [PATCH] security, keys: Optimize barrier usage for Rmw atomic bitops

[Davidlohr Bueso]

ping?

On Wed, 29 Jan 2020, Davidlohr Bueso wrote:

> For both set and clear_bit, we can avoid the unnecessary barriers
> on non LL/SC architectures, such as x86. Instead, use the
> smp_mb__{before,after}_atomic() calls.
>
> Signed-off-by: Davidlohr Bueso <dbueso@xxxxxxx>
> ---
> security/keys/gc.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/security/keys/gc.c b/security/keys/gc.c
> index 671dd730ecfc..ce7b4c22e3c4 100644
> --- a/security/keys/gc.c
> +++ b/security/keys/gc.c
> @@ -102,7 +102,7 @@ void key_gc_keytype(struct key_type *ktype)
>
> 	key_gc_dead_keytype = ktype;
> 	set_bit(KEY_GC_REAPING_KEYTYPE, &key_gc_flags);
> -	smp_mb();
> +	smp_mb__after_atomic();
> 	set_bit(KEY_GC_REAP_KEYTYPE, &key_gc_flags);
>
> 	kdebug("schedule");
> @@ -308,7 +308,7 @@ static void key_garbage_collector(struct work_struct *work)
>
> 	if (unlikely(gc_state & KEY_GC_REAPING_DEAD_3)) {
> 		kdebug("dead wake");
> -		smp_mb();
> +		smp_mb__before_atomic();
> 		clear_bit(KEY_GC_REAPING_KEYTYPE, &key_gc_flags);
> 		wake_up_bit(&key_gc_flags, KEY_GC_REAPING_KEYTYPE);
> 	}
> --
> 2.16.4