padata: use __this_cpu_read per-cpu helper

For bottom halves off, __this_cpu_read is better.

Signed-off-by: Shan Wei <davidshan@tencent.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This commit is contained in:
Shan Wei 2012-12-06 17:16:23 +08:00 committed by Herbert Xu
parent a465348ff5
commit f0fcf2002b

View file

@ -171,7 +171,7 @@ static struct padata_priv *padata_get_next(struct parallel_data *pd)
{
int cpu, num_cpus;
unsigned int next_nr, next_index;
struct padata_parallel_queue *queue, *next_queue;
struct padata_parallel_queue *next_queue;
struct padata_priv *padata;
struct padata_list *reorder;
@ -204,8 +204,7 @@ static struct padata_priv *padata_get_next(struct parallel_data *pd)
goto out;
}
queue = per_cpu_ptr(pd->pqueue, smp_processor_id());
if (queue->cpu_index == next_queue->cpu_index) {
if (__this_cpu_read(pd->pqueue->cpu_index) == next_queue->cpu_index) {
padata = ERR_PTR(-ENODATA);
goto out;
}