From patchwork Mon Nov 17 21:55:58 2008 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Pekka Enberg X-Patchwork-Id: 9247 Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from vger.kernel.org (vger.kernel.org [209.132.176.167]) by ozlabs.org (Postfix) with ESMTP id F1971DDDE7 for ; Tue, 18 Nov 2008 08:56:23 +1100 (EST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754603AbYKQV4B (ORCPT ); Mon, 17 Nov 2008 16:56:01 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753071AbYKQV4A (ORCPT ); Mon, 17 Nov 2008 16:56:00 -0500 Received: from courier.cs.helsinki.fi ([128.214.9.1]:42483 "EHLO mail.cs.helsinki.fi" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752997AbYKQVz7 (ORCPT ); Mon, 17 Nov 2008 16:55:59 -0500 Received: from melkki.cs.helsinki.fi (melkki.cs.helsinki.fi [128.214.9.98]) (AUTH: PLAIN cs-relay, TLS: TLSv1/SSLv3,256bits,AES256-SHA) by mail.cs.helsinki.fi with esmtp; Mon, 17 Nov 2008 23:55:58 +0200 id 00097E3C.4921E86E.000042BA Received: by melkki.cs.helsinki.fi (Postfix, from userid 54704) id C7A5917F68; Mon, 17 Nov 2008 23:55:58 +0200 (EET) Received: from localhost (localhost [127.0.0.1]) by melkki.cs.helsinki.fi (Postfix) with ESMTP id C6E3617F48; Mon, 17 Nov 2008 23:55:58 +0200 (EET) Date: Mon, 17 Nov 2008 23:55:58 +0200 (EET) From: Pekka J Enberg To: akpm@linux-foundation.org cc: adilger@sun.com, linux-kernel@vger.kernel.org, linux-ext4@vger.kernel.org, cl@linux-foundation.org, mpm@selenic.com, eduard.munteanu@linux360.ro Subject: [PATCH 4/4] ext4: allocate ->s_blockgroup_lock separately Message-ID: Mime-Version: 1.0 Sender: linux-ext4-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-ext4@vger.kernel.org From: Pekka Enberg As spotted by kmemtrace, struct ext4_sb_info is 17664 bytes on 64-bit which makes it a very bad fit for SLAB allocators. The culprit of the wasted memory is ->s_blockgroup_lock which can be as big as 16 KB when NR_CPUS >= 32. To fix that, allocate ->s_blockgroup_lock, which fits nicely in a order 2 page in the worst case, separately. This shinks down struct ext4_sb_info enough to fit a 2 KB slab cache so now we allocate 16 KB + 2 KB instead of 32 KB saving 14 KB of memory. Cc: Andreas Dilger Signed-off-by: Pekka Enberg --- fs/ext4/ext4_sb.h | 4 ++-- fs/ext4/super.c | 10 +++++++++- 2 files changed, 11 insertions(+), 3 deletions(-) diff --git a/fs/ext4/ext4_sb.h b/fs/ext4/ext4_sb.h index b21f167..ed3bc9c 100644 --- a/fs/ext4/ext4_sb.h +++ b/fs/ext4/ext4_sb.h @@ -61,7 +61,7 @@ struct ext4_sb_info { struct percpu_counter s_freeinodes_counter; struct percpu_counter s_dirs_counter; struct percpu_counter s_dirtyblocks_counter; - struct blockgroup_lock s_blockgroup_lock; + struct blockgroup_lock *s_blockgroup_lock; struct proc_dir_entry *s_proc; /* root of the per fs reservation window tree */ @@ -149,7 +149,7 @@ struct ext4_sb_info { static inline spinlock_t * sb_bgl_lock(struct ext4_sb_info *sbi, unsigned int block_group) { - return bgl_lock_ptr(&sbi->s_blockgroup_lock, block_group); + return bgl_lock_ptr(sbi->s_blockgroup_lock, block_group); } #endif /* _EXT4_SB */ diff --git a/fs/ext4/super.c b/fs/ext4/super.c index e4a241c..1add517 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -497,6 +497,7 @@ static void ext4_put_super(struct super_block *sb) ext4_blkdev_remove(sbi); } sb->s_fs_info = NULL; + kfree(sbi->s_blockgroup_lock); kfree(sbi); return; } @@ -1883,6 +1884,13 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); if (!sbi) return -ENOMEM; + + sbi->s_blockgroup_lock = + kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); + if (!sbi->s_blockgroup_lock) { + kfree(sbi); + return -ENOMEM; + } sb->s_fs_info = sbi; sbi->s_mount_opt = 0; sbi->s_resuid = EXT4_DEF_RESUID; @@ -2179,7 +2187,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) &sbi->s_inode_readahead_blks); #endif - bgl_lock_init(&sbi->s_blockgroup_lock); + bgl_lock_init(sbi->s_blockgroup_lock); for (i = 0; i < db_count; i++) { block = descriptor_loc(sb, logical_sb_block, i);