Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760126Ab0GQTut (ORCPT ); Sat, 17 Jul 2010 15:50:49 -0400 Received: from 0122700014.0.fullrate.dk ([95.166.99.235]:44308 "EHLO kernel.dk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751487Ab0GQTus (ORCPT ); Sat, 17 Jul 2010 15:50:48 -0400 Message-ID: <4C42099C.3070708@kernel.dk> Date: Sat, 17 Jul 2010 13:50:52 -0600 From: Jens Axboe MIME-Version: 1.0 To: Dan Carpenter CC: "Rafael J. Wysocki" , Martin Pirker , LKML , Andrew Morton , Miklos Szeredi , Wu Fengguang , Richard Kennedy , mjg@redhat.com Subject: Re: 2.6.35-rc5 inconsistent lock state References: <201007142304.48660.rjw@sisk.pl> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4906 Lines: 141 On 07/17/2010 01:04 PM, Dan Carpenter wrote: > This is from: > > commit 31373d09da5b7fe21fe6f781e92bd534a3495f00 > Author: Matthew Garrett > Date: Tue Apr 6 14:25:14 2010 +0200 > > laptop-mode: Make flushes per-device > > One of the features of laptop-mode is that it forces a writeout of dirty > pages if something else triggers a physical read or write from a device. > The current implementation flushes pages on all devices, rather than only > the one that triggered the flush. This patch alters the behaviour so that > only the recently accessed block device is flushed, preventing other > disks being spun up for no terribly good reason. > > One way to fix it might be to change all the places that call > spin_lock(&bdi->wb_lock); to spin_lock_bh(&bdi->wb_lock); but I'm not > sure that's the right way. > > I don't think Matthew Garrett has a bugzilla account? I posted a patch for this the other day, but I just now notice that it was a private list of CC addresses. Can anyone try this completed untested patch? diff --git a/block/blk-core.c b/block/blk-core.c index 5ab3ac2..a108b8e 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -451,7 +451,7 @@ void blk_cleanup_queue(struct request_queue *q) */ blk_sync_queue(q); - del_timer_sync(&q->backing_dev_info.laptop_mode_wb_timer); + cancel_delayed_work_sync(&q->backing_dev_info.laptop_mode_work); mutex_lock(&q->sysfs_lock); queue_flag_set_unlocked(QUEUE_FLAG_DEAD, q); mutex_unlock(&q->sysfs_lock); @@ -515,8 +515,9 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id) return NULL; } - setup_timer(&q->backing_dev_info.laptop_mode_wb_timer, - laptop_mode_timer_fn, (unsigned long) q); + + INIT_DELAYED_WORK(&q->backing_dev_info.laptop_mode_work, + laptop_mode_work_fn); init_timer(&q->unplug_timer); setup_timer(&q->timeout, blk_rq_timed_out_timer, (unsigned long) q); INIT_LIST_HEAD(&q->timeout_list); diff --git a/include/linux/backing-dev.h b/include/linux/backing-dev.h index e536f3a..d51acff 100644 --- a/include/linux/backing-dev.h +++ b/include/linux/backing-dev.h @@ -84,7 +84,7 @@ struct backing_dev_info { struct device *dev; - struct timer_list laptop_mode_wb_timer; + struct delayed_work laptop_mode_work; #ifdef CONFIG_DEBUG_FS struct dentry *debug_dir; diff --git a/include/linux/writeback.h b/include/linux/writeback.h index c24eca7..936ef5a 100644 --- a/include/linux/writeback.h +++ b/include/linux/writeback.h @@ -87,8 +87,7 @@ static inline void inode_sync_wait(struct inode *inode) #ifdef CONFIG_BLOCK void laptop_io_completion(struct backing_dev_info *info); void laptop_sync_completion(void); -void laptop_mode_sync(struct work_struct *work); -void laptop_mode_timer_fn(unsigned long data); +void laptop_mode_work_fn(struct work_struct *); #else static inline void laptop_sync_completion(void) { } #endif diff --git a/mm/page-writeback.c b/mm/page-writeback.c index 3d2111a..9978e8e 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -698,18 +698,20 @@ int dirty_writeback_centisecs_handler(ctl_table *table, int write, } #ifdef CONFIG_BLOCK -void laptop_mode_timer_fn(unsigned long data) +void laptop_mode_work_fn(struct work_struct *work) { - struct request_queue *q = (struct request_queue *)data; + struct backing_dev_info *bdi; int nr_pages = global_page_state(NR_FILE_DIRTY) + global_page_state(NR_UNSTABLE_NFS); + bdi = container_of(work, struct backing_dev_info, laptop_mode_work.work); + /* * We want to write everything out, not just down to the dirty * threshold */ - if (bdi_has_dirty_io(&q->backing_dev_info)) - bdi_start_writeback(&q->backing_dev_info, nr_pages); + if (bdi_has_dirty_io(bdi)) + bdi_start_writeback(bdi, nr_pages); } /* @@ -717,9 +719,12 @@ void laptop_mode_timer_fn(unsigned long data) * of all dirty data a few seconds from now. If the flush is already scheduled * then push it back - the user is still using the disk. */ -void laptop_io_completion(struct backing_dev_info *info) +void laptop_io_completion(struct backing_dev_info *bdi) { - mod_timer(&info->laptop_mode_wb_timer, jiffies + laptop_mode); + if (work_pending(&bdi->laptop_mode_work.work)) + mod_timer(&bdi->laptop_mode_work.timer, jiffies + laptop_mode); + else + schedule_delayed_work(&bdi->laptop_mode_work, laptop_mode); } /* @@ -734,7 +739,7 @@ void laptop_sync_completion(void) rcu_read_lock(); list_for_each_entry_rcu(bdi, &bdi_list, bdi_list) - del_timer(&bdi->laptop_mode_wb_timer); + __cancel_delayed_work(&bdi->laptop_mode_work); rcu_read_unlock(); } -- Jens Axboe -- 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/