From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751938AbbASM3c (ORCPT ); Mon, 19 Jan 2015 07:29:32 -0500 Received: from mx2.parallels.com ([199.115.105.18]:53554 "EHLO mx2.parallels.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751517AbbASM3b (ORCPT ); Mon, 19 Jan 2015 07:29:31 -0500 From: Vladimir Davydov To: Alexander Viro CC: , Subject: [PATCH] fs: super: do not cacheline align s_{inode,dentry}_lru Date: Mon, 19 Jan 2015 15:29:23 +0300 Message-ID: <1421670563-28228-1-git-send-email-vdavydov@parallels.com> X-Mailer: git-send-email 1.7.10.4 MIME-Version: 1.0 Content-Type: text/plain Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org super_block->s_{inode,dentry}_lru were made cacheline aligned, because the list_lru struct had a spinlock and a list embedded to it and therefore could be actively modified. Today, struct list_lru is, in fact, read-only: the lists and locks live in list_lru_node's, which list_lru points to. The list_lru_node struct is already cacheline aligned by definition. So let us drop the cacheline aligned annotation in order not to waste space in struct super_block. Signed-off-by: Vladimir Davydov --- include/linux/fs.h | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/include/linux/fs.h b/include/linux/fs.h index 0bd9669ce91d..790e534adbc7 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1309,12 +1309,9 @@ struct super_block { struct workqueue_struct *s_dio_done_wq; struct hlist_head s_pins; - /* - * Keep the lru lists last in the structure so they always sit on their - * own individual cachelines. - */ - struct list_lru s_dentry_lru ____cacheline_aligned_in_smp; - struct list_lru s_inode_lru ____cacheline_aligned_in_smp; + struct list_lru s_dentry_lru; + struct list_lru s_inode_lru; + struct rcu_head rcu; /* -- 1.7.10.4