On Fri, 2024-12-06 at 13:55 +1100, NeilBrown wrote:
> svc_thread_init_status() contains an open-coded
> store_release_wake_up(). It is cleaner to use that function directly
> rather than needing to remember the barrier.
>
> Signed-off-by: NeilBrown <neilb@xxxxxxx>
> ---
> include/linux/sunrpc/svc.h | 7 +------
> 1 file changed, 1 insertion(+), 6 deletions(-)
>
> diff --git a/include/linux/sunrpc/svc.h b/include/linux/sunrpc/svc.h
> index e68fecf6eab5..e4f09f58d58c 100644
> --- a/include/linux/sunrpc/svc.h
> +++ b/include/linux/sunrpc/svc.h
> @@ -327,12 +327,7 @@ static inline bool svc_thread_should_stop(struct svc_rqst *rqstp)
> */
> static inline void svc_thread_init_status(struct svc_rqst *rqstp, int err)
> {
> - rqstp->rq_err = err;
> - /* memory barrier ensures assignment to error above is visible before
> - * waitqueue_active() test below completes.
> - */
> - smp_mb();
> - wake_up_var(&rqstp->rq_err);
> + store_release_wake_up(&rqstp->rq_err, err);
> if (err)
> kthread_exit(1);
> }
Reviewed-by: Jeff Layton <jlayton@xxxxxxxxxx>
