Re: [rfc/patch] mm/slub: restore/expand unfreeze_partials() local exclusion scope
From: Thomas Gleixner
Date: Fri Jul 09 2021 - 15:28:16 EST
On Fri, Jul 09 2021 at 07:21, Mike Galbraith wrote:
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -2497,7 +2497,9 @@ static void put_cpu_partial(struct kmem_
> * partial array is full. Move the existing
> * set to the per node partial list.
> */
> + local_lock(&s->cpu_slab->lock);
> unfreeze_partials(s);
> + local_unlock(&s->cpu_slab->lock);
> oldpage = NULL;
> pobjects = 0;
> pages = 0;
> @@ -2579,7 +2581,9 @@ static void flush_cpu_slab(struct work_s
> if (c->page)
> flush_slab(s, c, true);
>
> + local_lock(&s->cpu_slab->lock);
> unfreeze_partials(s);
> + local_unlock(&s->cpu_slab->lock);
> }
>
> static bool has_cpu_slab(int cpu, struct kmem_cache *s)
> @@ -2632,8 +2636,11 @@ static int slub_cpu_dead(unsigned int cp
> struct kmem_cache *s;
>
> mutex_lock(&slab_mutex);
> - list_for_each_entry(s, &slab_caches, list)
> + list_for_each_entry(s, &slab_caches, list) {
> + local_lock(&s->cpu_slab->lock);
This one is odd. It locks the cpu_slab lock of the CPU which runs this
callback and then flushes the slab of the dead CPU.
Thanks,
tglx