From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754977AbaEGGEt (ORCPT ); Wed, 7 May 2014 02:04:49 -0400 Received: from LGEMRELSE7Q.lge.com ([156.147.1.151]:42287 "EHLO lgemrelse7q.lge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753512AbaEGGEj (ORCPT ); Wed, 7 May 2014 02:04:39 -0400 X-Original-SENDERIP: 10.177.220.145 X-Original-MAILFROM: iamjoonsoo.kim@lge.com From: Joonsoo Kim To: Pekka Enberg Cc: Christoph Lameter , Andrew Morton , David Rientjes , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Joonsoo Kim , Joonsoo Kim Subject: [PATCH v2 08/10] slab: destroy a slab without holding any alien cache lock Date: Wed, 7 May 2014 15:06:18 +0900 Message-Id: <1399442780-28748-9-git-send-email-iamjoonsoo.kim@lge.com> X-Mailer: git-send-email 1.7.9.5 In-Reply-To: <1399442780-28748-1-git-send-email-iamjoonsoo.kim@lge.com> References: <1399442780-28748-1-git-send-email-iamjoonsoo.kim@lge.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org I haven't heard that this alien cache lock is contended, but to reduce chance of contention would be better generally. And with this change, we can simplify complex lockdep annotation in slab code. In the following patch, it will be implemented. Acked-by: Christoph Lameter Signed-off-by: Joonsoo Kim diff --git a/mm/slab.c b/mm/slab.c index 889957b..3bb5e11 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -997,9 +997,9 @@ static void free_alien_cache(struct alien_cache **alc_ptr) } static void __drain_alien_cache(struct kmem_cache *cachep, - struct array_cache *ac, int node) + struct array_cache *ac, int node, + struct list_head *list) { - LIST_HEAD(list); struct kmem_cache_node *n = cachep->node[node]; if (ac->avail) { @@ -1012,10 +1012,9 @@ static void __drain_alien_cache(struct kmem_cache *cachep, if (n->shared) transfer_objects(n->shared, ac, ac->limit); - free_block(cachep, ac->entry, ac->avail, node, &list); + free_block(cachep, ac->entry, ac->avail, node, list); ac->avail = 0; spin_unlock(&n->list_lock); - slabs_destroy(cachep, &list); } } @@ -1033,8 +1032,11 @@ static void reap_alien(struct kmem_cache *cachep, struct kmem_cache_node *n) if (alc) { ac = &alc->ac; if (ac->avail && spin_trylock_irq(&alc->lock)) { - __drain_alien_cache(cachep, ac, node); + LIST_HEAD(list); + + __drain_alien_cache(cachep, ac, node, &list); spin_unlock_irq(&alc->lock); + slabs_destroy(cachep, &list); } } } @@ -1051,10 +1053,13 @@ static void drain_alien_cache(struct kmem_cache *cachep, for_each_online_node(i) { alc = alien[i]; if (alc) { + LIST_HEAD(list); + ac = &alc->ac; spin_lock_irqsave(&alc->lock, flags); - __drain_alien_cache(cachep, ac, i); + __drain_alien_cache(cachep, ac, i, &list); spin_unlock_irqrestore(&alc->lock, flags); + slabs_destroy(cachep, &list); } } } @@ -1085,10 +1090,11 @@ static inline int cache_free_alien(struct kmem_cache *cachep, void *objp) spin_lock(&alien->lock); if (unlikely(ac->avail == ac->limit)) { STATS_INC_ACOVERFLOW(cachep); - __drain_alien_cache(cachep, ac, nodeid); + __drain_alien_cache(cachep, ac, nodeid, &list); } ac_put_obj(cachep, ac, objp); spin_unlock(&alien->lock); + slabs_destroy(cachep, &list); } else { spin_lock(&(cachep->node[nodeid])->list_lock); free_block(cachep, &objp, 1, nodeid, &list); -- 1.7.9.5