node_match() with node=NUMA_NO_NODE always returns 1.
Duplicate check by goto statement is meaningless. Remove it.

Signed-off-by: JaeSang Yoo <jsyoo5b@gmail.com>
---
 mm/slub.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/mm/slub.c b/mm/slub.c
index 9fe000fd19ca..a65e282b8238 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -2913,7 +2913,6 @@ static void *___slab_alloc(struct kmem_cache *s, gfp_t gfpflags, int node,
 		 */
 		if (!node_isset(node, slab_nodes)) {
 			node = NUMA_NO_NODE;
-			goto redo;
 		} else {
 			stat(s, ALLOC_NODE_MISMATCH);
 			goto deactivate_slab;
-- 
2.25.1

On Sat, Apr 9, 2022 at 10:43 PM JaeSang Yoo <js.yoo.5b@gmail.com> wrote:
>
> node_match() with node=NUMA_NO_NODE always returns 1.
> Duplicate check by goto statement is meaningless. Remove it.
>
> Signed-off-by: JaeSang Yoo <jsyoo5b@gmail.com>

LGTM.

Reviewed-by: Muchun Song <songmuchun@bytedance.com>

On Sat, 9 Apr 2022, JaeSang Yoo wrote:

> node_match() with node=NUMA_NO_NODE always returns 1.
> Duplicate check by goto statement is meaningless. Remove it.
> 
> Signed-off-by: JaeSang Yoo <jsyoo5b@gmail.com>

Acked-by: David Rientjes <rientjes@google.com>

On 4/9/22 16:42, JaeSang Yoo wrote:
> node_match() with node=NUMA_NO_NODE always returns 1.
> Duplicate check by goto statement is meaningless. Remove it.
> 
> Signed-off-by: JaeSang Yoo <jsyoo5b@gmail.com>

Thanks, added.

> ---
>  mm/slub.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/mm/slub.c b/mm/slub.c
> index 9fe000fd19ca..a65e282b8238 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -2913,7 +2913,6 @@ static void *___slab_alloc(struct kmem_cache *s, gfp_t gfpflags, int node,
>  		 */
>  		if (!node_isset(node, slab_nodes)) {
>  			node = NUMA_NO_NODE;
> -			goto redo;
>  		} else {
>  			stat(s, ALLOC_NODE_MISMATCH);
>  			goto deactivate_slab;

On Sat, Apr 09, 2022 at 11:42:39PM +0900, JaeSang Yoo wrote:
> node_match() with node=NUMA_NO_NODE always returns 1.
> Duplicate check by goto statement is meaningless. Remove it.

Yeah, I think node = NUMA_NO_NODE here is to remove constraints
when later calling get_partial_node() or new_slab().

So no further check is required.

> 
> Signed-off-by: JaeSang Yoo <jsyoo5b@gmail.com>
> ---
>  mm/slub.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/mm/slub.c b/mm/slub.c
> index 9fe000fd19ca..a65e282b8238 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -2913,7 +2913,6 @@ static void *___slab_alloc(struct kmem_cache *s, gfp_t gfpflags, int node,
>  		 */
>  		if (!node_isset(node, slab_nodes)) {
>  			node = NUMA_NO_NODE;
> -			goto redo;
>  		} else {
>  			stat(s, ALLOC_NODE_MISMATCH);
>  			goto deactivate_slab;
> -- 
> 2.25.1

Looks good to me.

Reviewed-by: Hyeonggon Yoo <42.hyeyoo@gmail.com>

Thanks!

> 
> 

-- 
Thanks,
Hyeonggon