Merge commit 'block/for-next'

fs/btrfs/disk-io.c

@@ -1600,6 +1600,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	sb->s_blocksize = 4096;
 	sb->s_blocksize_bits = blksize_bits(4096);
+	sb->s_bdi = &fs_info->bdi;
 
 	/*
 	 * we set the i_size on the btree inode to the max possible int.

fs/fs-writeback.c

@@ -75,13 +75,6 @@ static inline void bdi_work_init(struct bdi_work *work,
 	work->state = WS_USED;
 }
 
-static inline void bdi_work_init_on_stack(struct bdi_work *work,
-					  struct writeback_control *wbc)
-{
-	bdi_work_init(work, wbc);
-	work->state |= WS_ONSTACK;
-}
-
 /**
  * writeback_in_progress - determine whether there is writeback in progress
  * @bdi: the device's backing_dev_info structure.
@@ -146,7 +139,6 @@ static void wb_clear_pending(struct bdi_writeback *wb, struct bdi_work *work)
 
 static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work)
 {
-	if (work) {
 	work->seen = bdi->wb_mask;
 	BUG_ON(!work->seen);
 	atomic_set(&work->pending, bdi->wb_cnt);
@@ -160,7 +152,6 @@ static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work)
 	spin_lock(&bdi->wb_lock);
 	list_add_tail_rcu(&work->list, &bdi->work_list);
 	spin_unlock(&bdi->wb_lock);
-	}
 
 	/*
 	 * If the default thread isn't there, make sure we add it. When
@@ -172,14 +163,12 @@ static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work)
 		struct bdi_writeback *wb = &bdi->wb;
 
 		/*
-		 * If we failed allocating the bdi work item, wake up the wb
-		 * thread always. As a safety precaution, it'll flush out
-		 * everything
+		 * End work now if this wb has no dirty IO pending. Otherwise
+		 * wakeup the handling thread
 		 */
-		if (!wb_has_dirty_io(wb)) {
-			if (work)
+		if (!wb_has_dirty_io(wb))
 			wb_clear_pending(wb, work);
-		} else if (wb->task)
+		else if (wb->task)
 			wake_up_process(wb->task);
 	}
 }
@@ -194,48 +183,63 @@ static void bdi_wait_on_work_clear(struct bdi_work *work)
 		    TASK_UNINTERRUPTIBLE);
 }
 
-static struct bdi_work *bdi_alloc_work(struct writeback_control *wbc)
+static void bdi_alloc_queue_work(struct backing_dev_info *bdi,
+				 struct writeback_control *wbc)
 {
 	struct bdi_work *work;
 
+	/*
+	 * This is WB_SYNC_NONE writeback, so if allocation fails just
+	 * wakeup the thread for old dirty data writeback
+	 */
 	work = kmalloc(sizeof(*work), GFP_ATOMIC);
-	if (work)
+	if (work) {
 		bdi_work_init(work, wbc);
+		bdi_queue_work(bdi, work);
+	} else {
+		struct bdi_writeback *wb = &bdi->wb;
 
-	return work;
+		if (wb->task)
+			wake_up_process(wb->task);
+	}
 }
 
-void bdi_start_writeback(struct writeback_control *wbc)
+/**
+ * bdi_sync_writeback - start and wait for writeback
+ * @wbc: writeback parameters
+ *
+ * Description:
+ *   This does WB_SYNC_ALL data integrity writeback and waits for the
+ *   IO to complete. Callers must hold the sb s_umount semaphore for
+ *   reading, to avoid having the super disappear before we are done.
+ */
+static void bdi_sync_writeback(struct writeback_control *wbc)
 {
-	const bool must_wait = wbc->sync_mode == WB_SYNC_ALL;
-	struct bdi_work work_stack, *work = NULL;
+	struct bdi_work work;
 
-	if (!must_wait)
-		work = bdi_alloc_work(wbc);
+	wbc->sync_mode = WB_SYNC_ALL;
 
-	if (!work) {
-		work = &work_stack;
-		bdi_work_init_on_stack(work, wbc);
-	}
+	bdi_work_init(&work, wbc);
+	work.state |= WS_ONSTACK;
 
-	bdi_queue_work(wbc->bdi, work);
+	bdi_queue_work(wbc->bdi, &work);
+	bdi_wait_on_work_clear(&work);
+}
 
-	/*
-	 * If the sync mode is WB_SYNC_ALL, block waiting for the work to
-	 * complete. If not, we only need to wait for the work to be started,
-	 * if we allocated it on-stack. We use the same mechanism, if the
-	 * wait bit is set in the bdi_work struct, then threads will not
-	 * clear pending until after they are done.
+/**
+ * bdi_start_writeback - start writeback
+ * @wbc: writeback parameters
+ *
+ * Description:
+ *   This does WB_SYNC_NONE opportunistic writeback. The IO is only
+ *   started when this function returns, we make no guarentees on
+ *   completion. Caller need not hold sb s_umount semaphore.
  *
-	 * Note that work == &work_stack if must_wait is true, so we don't
-	 * need to do call_rcu() here ever, since the completion path will
-	 * have done that for us.
  */
-	if (must_wait || work == &work_stack) {
-		bdi_wait_on_work_clear(work);
-		if (work != &work_stack)
-			call_rcu(&work->rcu_head, bdi_work_free);
-	}
+void bdi_start_writeback(struct writeback_control *wbc)
+{
+	wbc->sync_mode = WB_SYNC_NONE;
+	bdi_alloc_queue_work(wbc->bdi, wbc);
 }
 
 /*
@@ -858,67 +862,24 @@ int bdi_writeback_task(struct bdi_writeback *wb)
 }
 
 /*
- * Schedule writeback for all backing devices. Expensive! If this is a data
- * integrity operation, writeback will be complete when this returns. If
- * we are simply called for WB_SYNC_NONE, then writeback will merely be
- * scheduled to run.
+ * Schedule writeback for all backing devices. Can only be used for
+ * WB_SYNC_NONE writeback, WB_SYNC_ALL should use bdi_start_writeback()
+ * and pass in the superblock.
  */
 static void bdi_writeback_all(struct writeback_control *wbc)
 {
-	const bool must_wait = wbc->sync_mode == WB_SYNC_ALL;
 	struct backing_dev_info *bdi;
-	struct bdi_work *work;
-	LIST_HEAD(list);
-
-restart:
-	spin_lock(&bdi_lock);
-
-	list_for_each_entry(bdi, &bdi_list, bdi_list) {
-		struct bdi_work *work;
 
-		if (!bdi_has_dirty_io(bdi))
-			continue;
-
-		/*
-		 * If work allocation fails, do the writes inline. We drop
-		 * the lock and restart the list writeout. This should be OK,
-		 * since this happens rarely and because the writeout should
-		 * eventually make more free memory available.
-		 */
-		work = bdi_alloc_work(wbc);
-		if (!work) {
-			struct writeback_control __wbc;
-
-			/*
-			 * Not a data integrity writeout, just continue
-			 */
-			if (!must_wait)
-				continue;
+	WARN_ON(wbc->sync_mode == WB_SYNC_ALL);
 
-			spin_unlock(&bdi_lock);
-			__wbc = *wbc;
-			__wbc.bdi = bdi;
-			writeback_inodes_wbc(&__wbc);
-			goto restart;
-		}
-		if (must_wait)
-			list_add_tail(&work->wait_list, &list);
+	rcu_read_lock();
 
-		bdi_queue_work(bdi, work);
+	list_for_each_entry(bdi, &bdi_list, bdi_list) {
+		if (bdi_has_dirty_io(bdi))
+			bdi_alloc_queue_work(bdi, wbc);
 	}
 
-	spin_unlock(&bdi_lock);
-
-	/*
-	 * If this is for WB_SYNC_ALL, wait for pending work to complete
-	 * before returning.
-	 */
-	while (!list_empty(&list)) {
-		work = list_entry(list.next, struct bdi_work, wait_list);
-		list_del(&work->wait_list);
-		bdi_wait_on_work_clear(work);
-		call_rcu(&work->rcu_head, bdi_work_free);
-	}
+	rcu_read_unlock();
 }
 
 /*
@@ -1175,14 +1136,14 @@ long sync_inodes_sb(struct super_block *sb)
 {
 	struct writeback_control wbc = {
 		.sb		= sb,
-		.sync_mode	= WB_SYNC_ALL,
+		.bdi		= sb->s_bdi,
 		.range_start	= 0,
 		.range_end	= LLONG_MAX,
 	};
 	long nr_to_write = LONG_MAX; /* doesn't actually matter */
 
 	wbc.nr_to_write = nr_to_write;
-	bdi_writeback_all(&wbc);
+	bdi_sync_writeback(&wbc);
 	wait_sb_inodes(&wbc);
 	return nr_to_write - wbc.nr_to_write;
 }

fs/fuse/inode.c

@@ -908,6 +908,8 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
 	if (err)
 		goto err_put_conn;
 
+	sb->s_bdi = &fc->bdi;
+
 	/* Handle umasking inside the fuse code */
 	if (sb->s_flags & MS_POSIXACL)
 		fc->dont_mask = 1;

fs/super.c

@@ -707,6 +707,12 @@ static int set_bdev_super(struct super_block *s, void *data)
 {
 	s->s_bdev = data;
 	s->s_dev = s->s_bdev->bd_dev;
+
+	/*
+	 * We set the bdi here to the queue backing, file systems can
+	 * overwrite this in ->fill_super()
+	 */
+	s->s_bdi = &bdev_get_queue(s->s_bdev)->backing_dev_info;
 	return 0;
 }
 

fs/sync.c

@@ -27,6 +27,13 @@
  */
 static int __sync_filesystem(struct super_block *sb, int wait)
 {
+	/*
+	 * This should be safe, as we require bdi backing to actually
+	 * write out data in the first place
+	 */
+	if (!sb->s_bdi)
+		return 0;
+
 	/* Avoid doing twice syncing and cache pruning for quota sync */
 	if (!wait) {
 		writeout_quota_sb(sb, -1);
@@ -101,7 +108,7 @@ restart:
 		spin_unlock(&sb_lock);
 
 		down_read(&sb->s_umount);
-		if (!(sb->s_flags & MS_RDONLY) && sb->s_root)
+		if (!(sb->s_flags & MS_RDONLY) && sb->s_root && sb->s_bdi)
 			__sync_filesystem(sb, wait);
 		up_read(&sb->s_umount);
 

fs/ubifs/super.c

@@ -1967,6 +1967,7 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 	if (err)
 		goto out_bdi;
 
+	sb->s_bdi = &c->bdi;
 	sb->s_fs_info = c;
 	sb->s_magic = UBIFS_SUPER_MAGIC;
 	sb->s_blocksize = UBIFS_BLOCK_SIZE;

include/linux/backing-dev.h

@@ -59,6 +59,7 @@ struct bdi_writeback {
 
 struct backing_dev_info {
 	struct list_head bdi_list;
+	struct rcu_head rcu_head;
 	unsigned long ra_pages;	/* max readahead in PAGE_CACHE_SIZE units */
 	unsigned long state;	/* Always use atomic bitops on this */
 	unsigned int capabilities; /* Device capabilities */

include/linux/fs.h

@@ -1343,6 +1343,7 @@ struct super_block {
 	int			s_nr_dentry_unused;	/* # of dentry on lru */
 
 	struct block_device	*s_bdev;
+	struct backing_dev_info *s_bdi;
 	struct mtd_info		*s_mtd;
 	struct list_head	s_instances;
 	struct quota_info	s_dquot;	/* Diskquota specific options */

mm/backing-dev.c

@@ -26,6 +26,12 @@ struct backing_dev_info default_backing_dev_info = {
 EXPORT_SYMBOL_GPL(default_backing_dev_info);
 
 static struct class *bdi_class;
+
+/*
+ * bdi_lock protects updates to bdi_list and bdi_pending_list, as well as
+ * reader side protection for bdi_pending_list. bdi_list has RCU reader side
+ * locking.
+ */
 DEFINE_SPINLOCK(bdi_lock);
 LIST_HEAD(bdi_list);
 LIST_HEAD(bdi_pending_list);
@@ -284,9 +290,9 @@ static int bdi_start_fn(void *ptr)
 	/*
 	 * Add us to the active bdi_list
 	 */
-	spin_lock(&bdi_lock);
-	list_add(&bdi->bdi_list, &bdi_list);
-	spin_unlock(&bdi_lock);
+	spin_lock_bh(&bdi_lock);
+	list_add_rcu(&bdi->bdi_list, &bdi_list);
+	spin_unlock_bh(&bdi_lock);
 
 	bdi_task_init(bdi, wb);
 
@@ -389,7 +395,7 @@ static int bdi_forker_task(void *ptr)
 		if (wb_has_dirty_io(me) || !list_empty(&me->bdi->work_list))
 			wb_do_writeback(me, 0);
 
-		spin_lock(&bdi_lock);
+		spin_lock_bh(&bdi_lock);
 
 		/*
 		 * Check if any existing bdi's have dirty data without
@@ -410,7 +416,7 @@ static int bdi_forker_task(void *ptr)
 		if (list_empty(&bdi_pending_list)) {
 			unsigned long wait;
 
-			spin_unlock(&bdi_lock);
+			spin_unlock_bh(&bdi_lock);
 			wait = msecs_to_jiffies(dirty_writeback_interval * 10);
 			schedule_timeout(wait);
 			try_to_freeze();
@@ -426,7 +432,7 @@ static int bdi_forker_task(void *ptr)
 		bdi = list_entry(bdi_pending_list.next, struct backing_dev_info,
 				 bdi_list);
 		list_del_init(&bdi->bdi_list);
-		spin_unlock(&bdi_lock);
+		spin_unlock_bh(&bdi_lock);
 
 		wb = &bdi->wb;
 		wb->task = kthread_run(bdi_start_fn, wb, "flush-%s",
@@ -445,9 +451,9 @@ static int bdi_forker_task(void *ptr)
 			 * a chance to flush other bdi's to free
 			 * memory.
 			 */
-			spin_lock(&bdi_lock);
+			spin_lock_bh(&bdi_lock);
 			list_add_tail(&bdi->bdi_list, &bdi_pending_list);
-			spin_unlock(&bdi_lock);
+			spin_unlock_bh(&bdi_lock);
 
 			bdi_flush_io(bdi);
 		}
@@ -456,6 +462,24 @@ static int bdi_forker_task(void *ptr)
 	return 0;
 }
 
+static void bdi_add_to_pending(struct rcu_head *head)
+{
+	struct backing_dev_info *bdi;
+
+	bdi = container_of(head, struct backing_dev_info, rcu_head);
+	INIT_LIST_HEAD(&bdi->bdi_list);
+
+	spin_lock(&bdi_lock);
+	list_add_tail(&bdi->bdi_list, &bdi_pending_list);
+	spin_unlock(&bdi_lock);
+
+	/*
+	 * We are now on the pending list, wake up bdi_forker_task()
+	 * to finish the job and add us back to the active bdi_list
+	 */
+	wake_up_process(default_backing_dev_info.wb.task);
+}
+
 /*
  * Add the default flusher task that gets created for any bdi
  * that has dirty data pending writeout
@@ -478,16 +502,29 @@ void static bdi_add_default_flusher_task(struct backing_dev_info *bdi)
 	 * waiting for previous additions to finish.
 	 */
 	if (!test_and_set_bit(BDI_pending, &bdi->state)) {
-		list_move_tail(&bdi->bdi_list, &bdi_pending_list);
+		list_del_rcu(&bdi->bdi_list);
 
 		/*
-		 * We are now on the pending list, wake up bdi_forker_task()
-		 * to finish the job and add us back to the active bdi_list
+		 * We must wait for the current RCU period to end before
+		 * moving to the pending list. So schedule that operation
+		 * from an RCU callback.
 		 */
-		wake_up_process(default_backing_dev_info.wb.task);
+		call_rcu(&bdi->rcu_head, bdi_add_to_pending);
 	}
 }
 
+/*
+ * Remove bdi from bdi_list, and ensure that it is no longer visible
+ */
+static void bdi_remove_from_list(struct backing_dev_info *bdi)
+{
+	spin_lock_bh(&bdi_lock);
+	list_del_rcu(&bdi->bdi_list);
+	spin_unlock_bh(&bdi_lock);
+
+	synchronize_rcu();
+}
+
 int bdi_register(struct backing_dev_info *bdi, struct device *parent,
 		const char *fmt, ...)
 {
@@ -506,9 +543,9 @@ int bdi_register(struct backing_dev_info *bdi, struct device *parent,
 		goto exit;
 	}
 
-	spin_lock(&bdi_lock);
-	list_add_tail(&bdi->bdi_list, &bdi_list);
-	spin_unlock(&bdi_lock);
+	spin_lock_bh(&bdi_lock);
+	list_add_tail_rcu(&bdi->bdi_list, &bdi_list);
+	spin_unlock_bh(&bdi_lock);
 
 	bdi->dev = dev;
 
@@ -526,9 +563,7 @@ int bdi_register(struct backing_dev_info *bdi, struct device *parent,
 			wb->task = NULL;
 			ret = -ENOMEM;
 
-			spin_lock(&bdi_lock);
-			list_del(&bdi->bdi_list);
-			spin_unlock(&bdi_lock);
+			bdi_remove_from_list(bdi);
 			goto exit;
 		}
 	}
@@ -565,9 +600,7 @@ static void bdi_wb_shutdown(struct backing_dev_info *bdi)
 	/*
 	 * Make sure nobody finds us on the bdi_list anymore
 	 */
-	spin_lock(&bdi_lock);
-	list_del(&bdi->bdi_list);
-	spin_unlock(&bdi_lock);
+	bdi_remove_from_list(bdi);
 
 	/*
 	 * Finally, kill the kernel threads. We don't need to be RCU
@@ -599,6 +632,7 @@ int bdi_init(struct backing_dev_info *bdi)
 	bdi->max_ratio = 100;
 	bdi->max_prop_frac = PROP_FRAC_BASE;
 	spin_lock_init(&bdi->wb_lock);
+	INIT_RCU_HEAD(&bdi->rcu_head);
 	INIT_LIST_HEAD(&bdi->bdi_list);
 	INIT_LIST_HEAD(&bdi->wb_list);
 	INIT_LIST_HEAD(&bdi->work_list);

mm/page-writeback.c

@@ -315,7 +315,7 @@ int bdi_set_min_ratio(struct backing_dev_info *bdi, unsigned int min_ratio)
 {
 	int ret = 0;
 
-	spin_lock(&bdi_lock);
+	spin_lock_bh(&bdi_lock);
 	if (min_ratio > bdi->max_ratio) {
 		ret = -EINVAL;
 	} else {
@@ -327,7 +327,7 @@ int bdi_set_min_ratio(struct backing_dev_info *bdi, unsigned int min_ratio)
 			ret = -EINVAL;
 		}
 	}
-	spin_unlock(&bdi_lock);
+	spin_unlock_bh(&bdi_lock);
 
 	return ret;
 }
@@ -339,14 +339,14 @@ int bdi_set_max_ratio(struct backing_dev_info *bdi, unsigned max_ratio)
 	if (max_ratio > 100)
 		return -EINVAL;
 
-	spin_lock(&bdi_lock);
+	spin_lock_bh(&bdi_lock);
 	if (bdi->min_ratio > max_ratio) {
 		ret = -EINVAL;
 	} else {
 		bdi->max_ratio = max_ratio;
 		bdi->max_prop_frac = (PROP_FRAC_BASE * max_ratio) / 100;
 	}
-	spin_unlock(&bdi_lock);
+	spin_unlock_bh(&bdi_lock);
 
 	return ret;
 }
@@ -585,7 +585,6 @@ static void balance_dirty_pages(struct address_space *mapping)
 					  > background_thresh))) {
 		struct writeback_control wbc = {
 			.bdi		= bdi,
-			.sync_mode	= WB_SYNC_NONE,
 			.nr_to_write	= nr_writeback,
 		};