Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760223AbZGIHDX (ORCPT ); Thu, 9 Jul 2009 03:03:23 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1759973AbZGIHAN (ORCPT ); Thu, 9 Jul 2009 03:00:13 -0400 Received: from smtp.nokia.com ([192.100.122.233]:30107 "EHLO mgw-mx06.nokia.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758314AbZGIHAK (ORCPT ); Thu, 9 Jul 2009 03:00:10 -0400 From: Artem Bityutskiy To: Al Viro , Jens Axboe Cc: linux-fsdevel@vger.kernel.org, Artem Bityutskiy , linux-kernel@vger.kernel.org Date: Thu, 09 Jul 2009 11:50:26 +0300 Message-Id: <20090709085026.12122.11937.sendpatchset@localhost.localdomain> In-Reply-To: <20090709084822.12122.79749.sendpatchset@localhost.localdomain> References: <20090709084822.12122.79749.sendpatchset@localhost.localdomain> Subject: [PATCH v3 18/18] writeback: optimize periodic sync_supers X-OriginalArrivalTime: 09 Jul 2009 06:59:58.0196 (UTC) FILETIME=[DF059340:01CA0062] X-Nokia-AV: Clean Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3901 Lines: 122 The sync_supers thread wakes up every 5 seconds (by default) and writes back all super blocks. It keeps waking up even if there are no dirty super-blocks. This patch improves it and makes sleep if there is nothing to do. This optimization is quite important for small battery-powered devices. Signed-off-by: Artem Bityutskiy --- include/linux/fs.h | 5 +---- mm/backing-dev.c | 46 +++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 46 insertions(+), 5 deletions(-) diff --git a/include/linux/fs.h b/include/linux/fs.h index 7882a61..ae626b7 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1793,10 +1793,7 @@ int __put_super_and_need_restart(struct super_block *sb); * Note, VFS does not provide any protection for the super block clean/dirty * state. File-systems should take care of this. */ -static inline void mark_sb_dirty(struct super_block *sb) -{ - sb->s_dirty = 1; -} +void mark_sb_dirty(struct super_block *sb); static inline void mark_sb_clean(struct super_block *sb) { sb->s_dirty = 0; diff --git a/mm/backing-dev.c b/mm/backing-dev.c index 75e6c47..96f4b2a 100644 --- a/mm/backing-dev.c +++ b/mm/backing-dev.c @@ -32,6 +32,8 @@ LIST_HEAD(bdi_pending_list); static struct task_struct *sync_supers_tsk; static struct timer_list sync_supers_timer; +static int supers_timer_armed; +static DEFINE_SPINLOCK(supers_timer_lock); static int bdi_sync_supers(void *); static void sync_supers_timer_fn(unsigned long); @@ -440,6 +442,11 @@ static void bdi_flush_io(struct backing_dev_info *bdi) * or we risk deadlocking on ->s_umount. The longer term solution would be * to implement sync_supers_bdi() or similar and simply do it from the * bdi writeback tasks individually. + * + * Historically this thread woke up periodically, regardless of whether + * there was any dirty super block. However, nowadays it is optimized to + * wake up only when there is something to sync - this is better from the + * power management point of view. */ static int bdi_sync_supers(void *unused) { @@ -449,10 +456,24 @@ static int bdi_sync_supers(void *unused) set_current_state(TASK_INTERRUPTIBLE); schedule(); + spin_lock(&supers_timer_lock); + /* Indicate that 'sync_supers' is in progress */ + supers_timer_armed = -1; + spin_unlock(&supers_timer_lock); + /* * Do this periodically, like kupdated() did before. */ sync_supers(); + + spin_lock(&supers_timer_lock); + if (supers_timer_armed == 1) + /* A super block was marked as dirty meanwhile */ + arm_supers_timer(); + else + /* No more dirty super blocks - we've synced'em all */ + supers_timer_armed = 0; + spin_unlock(&supers_timer_lock); } return 0; @@ -469,9 +490,32 @@ static void arm_supers_timer(void) static void sync_supers_timer_fn(unsigned long unused) { wake_up_process(sync_supers_tsk); - arm_supers_timer(); } +void mark_sb_dirty(struct super_block *sb) +{ + sb->s_dirty = 1; + + /* + * A super block has been marked dirty - arm the 'sync_supers' kernel + * thread timer to make sure the super block is synchronized later. + */ + spin_lock(&supers_timer_lock); + if (!supers_timer_armed) { + arm_supers_timer(); + supers_timer_armed = 1; + } else if (supers_timer_armed == -1) { + /* + * The super-blocks are being synchronized at the moment, + * indicate that a new super block has been marked as dirty and + * the timer should be armed again. + */ + supers_timer_armed = 1; + } + spin_unlock(&supers_timer_lock); +} +EXPORT_SYMBOL(mark_sb_dirty); + static int bdi_forker_task(void *ptr) { struct bdi_writeback *me = ptr; -- 1.6.0.6 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/