From 90b1e56641bbab801e22141c56aa79dc095a3764 Mon Sep 17 00:00:00 2001 From: Chengming Zhou Date: Tue, 23 Jan 2024 09:33:29 +0000 Subject: mm/slub: directly load freelist from cpu partial slab in the likely case The likely case is that we get a usable slab from the cpu partial list, we can directly load freelist from it and return back, instead of going the other way that need more work, like reenable interrupt and recheck. But we need to remove the "VM_BUG_ON(!new.frozen)" in get_freelist() for reusing it, since cpu partial slab is not frozen. It seems acceptable since it's only for debug purpose. And get_freelist() also assumes it can return NULL if the freelist is empty, which is not possible for the cpu partial slab case, so we add "VM_BUG_ON(!freelist)" after get_freelist() to make it explicit. There is some small performance improvement too, which shows by: perf bench sched messaging -g 5 -t -l 100000 mm-stable slub-optimize Total time 7.473 7.209 Signed-off-by: Chengming Zhou Reviewed-by: Vlastimil Babka Signed-off-by: Vlastimil Babka --- mm/slub.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) (limited to 'mm') diff --git a/mm/slub.c b/mm/slub.c index 2ef88bbf56a3..fda402b2d649 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -3326,7 +3326,6 @@ static inline void *get_freelist(struct kmem_cache *s, struct slab *slab) counters = slab->counters; new.counters = counters; - VM_BUG_ON(!new.frozen); new.inuse = slab->objects; new.frozen = freelist != NULL; @@ -3498,18 +3497,20 @@ new_slab: slab = slub_percpu_partial(c); slub_set_percpu_partial(c, slab); - local_unlock_irqrestore(&s->cpu_slab->lock, flags); - stat(s, CPU_PARTIAL_ALLOC); - if (unlikely(!node_match(slab, node) || - !pfmemalloc_match(slab, gfpflags))) { - slab->next = NULL; - __put_partials(s, slab); - continue; + if (likely(node_match(slab, node) && + pfmemalloc_match(slab, gfpflags))) { + c->slab = slab; + freelist = get_freelist(s, slab); + VM_BUG_ON(!freelist); + stat(s, CPU_PARTIAL_ALLOC); + goto load_freelist; } - freelist = freeze_slab(s, slab); - goto retry_load_slab; + local_unlock_irqrestore(&s->cpu_slab->lock, flags); + + slab->next = NULL; + __put_partials(s, slab); } #endif -- cgit From a6def11b6dcde5d8f1fcc9e2c0ae71399432b62e Mon Sep 17 00:00:00 2001 From: Chengming Zhou Date: Tue, 23 Jan 2024 09:33:30 +0000 Subject: mm/slub: remove full list manipulation for non-debug slab Since debug slab is processed by free_to_partial_list(), and only debug slab which has SLAB_STORE_USER flag would care about the full list, we can remove these unrelated full list manipulations from __slab_free(). Acked-by: Christoph Lameter (Ampere) Reviewed-by: Vlastimil Babka Signed-off-by: Chengming Zhou Signed-off-by: Vlastimil Babka --- mm/slub.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'mm') diff --git a/mm/slub.c b/mm/slub.c index fda402b2d649..5c6fbeef05a8 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -4188,7 +4188,6 @@ static void __slab_free(struct kmem_cache *s, struct slab *slab, * then add it. */ if (!kmem_cache_has_cpu_partial(s) && unlikely(!prior)) { - remove_full(s, n, slab); add_partial(n, slab, DEACTIVATE_TO_TAIL); stat(s, FREE_ADD_PARTIAL); } @@ -4202,9 +4201,6 @@ slab_empty: */ remove_partial(n, slab); stat(s, FREE_REMOVE_PARTIAL); - } else { - /* Slab must be on the full list */ - remove_full(s, n, slab); } spin_unlock_irqrestore(&n->list_lock, flags); -- cgit From c63349fc4a2d10f5d1b5ee805cb639ee88fd6e4a Mon Sep 17 00:00:00 2001 From: Chengming Zhou Date: Tue, 23 Jan 2024 09:33:31 +0000 Subject: mm/slub: remove unused parameter in next_freelist_entry() The parameter "struct slab *slab" is unused in next_freelist_entry(), so just remove it. Acked-by: Christoph Lameter (Ampere) Reviewed-by: Vlastimil Babka Signed-off-by: Chengming Zhou Signed-off-by: Vlastimil Babka --- mm/slub.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'mm') diff --git a/mm/slub.c b/mm/slub.c index 5c6fbeef05a8..7f235fa6592d 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -2243,7 +2243,7 @@ static void __init init_freelist_randomization(void) } /* Get the next entry on the pre-computed freelist randomized */ -static void *next_freelist_entry(struct kmem_cache *s, struct slab *slab, +static void *next_freelist_entry(struct kmem_cache *s, unsigned long *pos, void *start, unsigned long page_limit, unsigned long freelist_count) @@ -2282,13 +2282,12 @@ static bool shuffle_freelist(struct kmem_cache *s, struct slab *slab) start = fixup_red_left(s, slab_address(slab)); /* First entry is used as the base of the freelist */ - cur = next_freelist_entry(s, slab, &pos, start, page_limit, - freelist_count); + cur = next_freelist_entry(s, &pos, start, page_limit, freelist_count); cur = setup_object(s, cur); slab->freelist = cur; for (idx = 1; idx < slab->objects; idx++) { - next = next_freelist_entry(s, slab, &pos, start, page_limit, + next = next_freelist_entry(s, &pos, start, page_limit, freelist_count); next = setup_object(s, next); set_freepointer(s, cur, next); -- cgit