From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758174Ab2C1LBT (ORCPT ); Wed, 28 Mar 2012 07:01:19 -0400 Received: from fgwmail5.fujitsu.co.jp ([192.51.44.35]:49803 "EHLO fgwmail5.fujitsu.co.jp" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757771Ab2C1LBR (ORCPT ); Wed, 28 Mar 2012 07:01:17 -0400 X-SecurityPolicyCheck: OK by SHieldMailChecker v1.7.4 Message-ID: <4F72EF10.3060209@jp.fujitsu.com> Date: Wed, 28 Mar 2012 19:59:28 +0900 From: KAMEZAWA Hiroyuki User-Agent: Mozilla/5.0 (Windows NT 6.0; rv:11.0) Gecko/20120312 Thunderbird/11.0 MIME-Version: 1.0 To: "linux-mm@kvack.org" CC: Linux Kernel , Michal Hocko , Johannes Weiner , Andrew Morton , Han Ying , Glauber Costa , Konstantin Khlebnikov , Suleiman Souhlal Subject: [RFC][[PATCH 4/6] memcg: remove mem_cgroup pointer from page_cgroup References: <4F72EB84.7080000@jp.fujitsu.com> In-Reply-To: <4F72EB84.7080000@jp.fujitsu.com> Content-Type: text/plain; charset=ISO-2022-JP Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org This patch removes pc->mem_cgroup and merge the pointer into flags as 63 4 0 | pointer to memcg..........|flags| After this, memory cgroup's overhead is 8(4)bytes per a page. Changelog: - added BUILD_BUG_ON() - update comments in Kconfig Signed-off-by: KAMEZAWA Hiroyuki --- include/linux/page_cgroup.h | 21 +++++++++------------ init/Kconfig | 2 +- 2 files changed, 10 insertions(+), 13 deletions(-) diff --git a/include/linux/page_cgroup.h b/include/linux/page_cgroup.h index 3f3b4ff..7e3a3c7 100644 --- a/include/linux/page_cgroup.h +++ b/include/linux/page_cgroup.h @@ -3,6 +3,10 @@ #include +/* + * These flags are encoded into low bits of unsigned long, + * ORed with a pointer to mem_cgroup. + */ enum { /* flags for mem_cgroup */ PCG_LOCK, /* Lock for pc->mem_cgroup and following bits. */ @@ -12,6 +16,8 @@ enum { __NR_PCG_FLAGS, }; +#define PCG_FLAGS_MASK ((1 << __NR_PCG_FLAGS) - 1) + #ifndef __GENERATING_BOUNDS_H #include @@ -27,7 +33,6 @@ enum { */ struct page_cgroup { unsigned long flags; - struct mem_cgroup *mem_cgroup; }; void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat); @@ -93,7 +98,7 @@ extern struct mem_cgroup* root_mem_cgroup; static inline struct mem_cgroup* pc_to_mem_cgroup(struct page_cgroup *pc) { if (likely(!PageCgroupReset(pc))) - return pc->mem_cgroup; + return (struct mem_cgroup*)(pc->flags & ~PCG_FLAGS_MASK); return root_mem_cgroup; } @@ -101,16 +106,8 @@ static inline void pc_set_mem_cgroup_and_flags(struct page_cgroup *pc, struct mem_cgroup *memcg, unsigned long flags) { - pc->mem_cgroup = memcg; - /* - * We access a page_cgroup asynchronously without lock_page_cgroup(). - * Especially when a page_cgroup is taken from a page, pc's mem_cgroup - * is accessed after testing USED bit. To make pc's mem_cgroup visible - * before USED bit, we need memory barrier here. - * See mem_cgroup_add_lru_list(), etc. - */ - smp_wmb(); - pc->flags = flags; + BUILD_BUG_ON(__NR_PCG_FLAGS > 5); /* assume 32byte alignment. */ + pc->flags = (unsigned long)memcg | flags; } #else /* CONFIG_CGROUP_MEM_RES_CTLR */ diff --git a/init/Kconfig b/init/Kconfig index 6cfd71d..e0bfe92 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -651,7 +651,7 @@ config CGROUP_MEM_RES_CTLR Note that setting this option increases fixed memory overhead associated with each page of memory in the system. By this, - 20(40)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory + 4(8)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory usage tracking struct at boot. Total amount of this is printed out at boot. -- 1.7.4.1