Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 837d678d authored by Joonsoo Kim's avatar Joonsoo Kim Committed by Pekka Enberg
Browse files

slub: remove one code path and reduce lock contention in __slab_free()



When we try to free object, there is some of case that we need
to take a node lock. This is the necessary step for preventing a race.
After taking a lock, then we try to cmpxchg_double_slab().
But, there is a possible scenario that cmpxchg_double_slab() is failed
with taking a lock. Following example explains it.

CPU A               CPU B
need lock
...                 need lock
...                 lock!!
lock..but spin      free success
spin...             unlock
lock!!
free fail

In this case, retry with taking a lock is occured in CPU A.
I think that in this case for CPU A,
"release a lock first, and re-take a lock if necessary" is preferable way.

There are two reasons for this.

First, this makes __slab_free()'s logic somehow simple.
With this patch, 'was_frozen = 1' is "always" handled without taking a lock.
So we can remove one code path.

Second, it may reduce lock contention.
When we do retrying, status of slab is already changed,
so we don't need a lock anymore in almost every case.
"release a lock first, and re-take a lock if necessary" policy is
helpful to this.

Signed-off-by: default avatarJoonsoo Kim <js1304@gmail.com>
Acked-by: default avatarChristoph Lameter <cl@linux.com>
Signed-off-by: default avatarPekka Enberg <penberg@kernel.org>
parent ddffeb8c
Loading
Loading
Loading
Loading
+14 −20
Original line number Original line Diff line number Diff line
@@ -2459,7 +2459,6 @@ static void __slab_free(struct kmem_cache *s, struct page *page,
	void *prior;
	void *prior;
	void **object = (void *)x;
	void **object = (void *)x;
	int was_frozen;
	int was_frozen;
	int inuse;
	struct page new;
	struct page new;
	unsigned long counters;
	unsigned long counters;
	struct kmem_cache_node *n = NULL;
	struct kmem_cache_node *n = NULL;
@@ -2472,13 +2471,17 @@ static void __slab_free(struct kmem_cache *s, struct page *page,
		return;
		return;


	do {
	do {
		if (unlikely(n)) {
			spin_unlock_irqrestore(&n->list_lock, flags);
			n = NULL;
		}
		prior = page->freelist;
		prior = page->freelist;
		counters = page->counters;
		counters = page->counters;
		set_freepointer(s, object, prior);
		set_freepointer(s, object, prior);
		new.counters = counters;
		new.counters = counters;
		was_frozen = new.frozen;
		was_frozen = new.frozen;
		new.inuse--;
		new.inuse--;
		if ((!new.inuse || !prior) && !was_frozen && !n) {
		if ((!new.inuse || !prior) && !was_frozen) {


			if (!kmem_cache_debug(s) && !prior)
			if (!kmem_cache_debug(s) && !prior)


@@ -2503,7 +2506,6 @@ static void __slab_free(struct kmem_cache *s, struct page *page,


			}
			}
		}
		}
		inuse = new.inuse;


	} while (!cmpxchg_double_slab(s, page,
	} while (!cmpxchg_double_slab(s, page,
		prior, counters,
		prior, counters,
@@ -2529,26 +2531,18 @@ static void __slab_free(struct kmem_cache *s, struct page *page,
                return;
                return;
        }
        }


	/*
	if (unlikely(!new.inuse && n->nr_partial > s->min_partial))
	 * was_frozen may have been set after we acquired the list_lock in
	 * an earlier loop. So we need to check it here again.
	 */
	if (was_frozen)
		stat(s, FREE_FROZEN);
	else {
		if (unlikely(!inuse && n->nr_partial > s->min_partial))
		goto slab_empty;
		goto slab_empty;


	/*
	/*
	 * Objects left in the slab. If it was not on the partial list before
	 * Objects left in the slab. If it was not on the partial list before
	 * then add it.
	 * then add it.
	 */
	 */
		if (unlikely(!prior)) {
	if (kmem_cache_debug(s) && unlikely(!prior)) {
		remove_full(s, page);
		remove_full(s, page);
		add_partial(n, page, DEACTIVATE_TO_TAIL);
		add_partial(n, page, DEACTIVATE_TO_TAIL);
		stat(s, FREE_ADD_PARTIAL);
		stat(s, FREE_ADD_PARTIAL);
	}
	}
	}
	spin_unlock_irqrestore(&n->list_lock, flags);
	spin_unlock_irqrestore(&n->list_lock, flags);
	return;
	return;