Btrfs: don't wait for the completion of all the ordered extents

It is very likely that there are lots of ordered extents in the filesytem,
if we wait for the completion of all of them when we want to reclaim some
space for the metadata space reservation, we would be blocked for a long
time. The performance would drop down suddenly for a long time.

Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>

8 files changed, 31 insertions, 18 deletions

diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index cb94310..3d2495e 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -391,7 +391,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
 	btrfs_dev_replace_unlock(dev_replace);
 
-	btrfs_wait_all_ordered_extents(root->fs_info);
+	btrfs_wait_ordered_roots(root->fs_info, -1);
 
 	/* force writing the updated state information to disk */
 	trans = btrfs_start_transaction(root, 0);
@@ -466,7 +466,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 		return ret;
 	}
-	btrfs_wait_all_ordered_extents(root->fs_info);
+	btrfs_wait_ordered_roots(root->fs_info, -1);
 
 	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 83bffbe..a21bbf83 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -4018,7 +4018,7 @@ static void btrfs_writeback_inodes_sb_nr(struct btrfs_root *root,
 		 */
 		btrfs_start_all_delalloc_inodes(root->fs_info, 0);
 		if (!current->journal_info)
-			btrfs_wait_all_ordered_extents(root->fs_info);
+			btrfs_wait_ordered_roots(root->fs_info, -1);
 	}
 }
 
@@ -4050,11 +4050,12 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig,
 	long time_left;
 	unsigned long nr_pages;
 	int loops;
+	int items;
 	enum btrfs_reserve_flush_enum flush;
 
 	/* Calc the number of the pages we need flush for space reservation */
-	to_reclaim = calc_reclaim_items_nr(root, to_reclaim);
-	to_reclaim *= EXTENT_SIZE_PER_ITEM;
+	items = calc_reclaim_items_nr(root, to_reclaim);
+	to_reclaim = items * EXTENT_SIZE_PER_ITEM;
 
 	trans = (struct btrfs_trans_handle *)current->journal_info;
 	block_rsv = &root->fs_info->delalloc_block_rsv;
@@ -4066,7 +4067,7 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig,
 		if (trans)
 			return;
 		if (wait_ordered)
-			btrfs_wait_all_ordered_extents(root->fs_info);
+			btrfs_wait_ordered_roots(root->fs_info, items);
 		return;
 	}
 
@@ -4105,7 +4106,7 @@ skip_async:
 
 		loops++;
 		if (wait_ordered && !trans) {
-			btrfs_wait_all_ordered_extents(root->fs_info);
+			btrfs_wait_ordered_roots(root->fs_info, items);
 		} else {
 			time_left = schedule_timeout_killable(1);
 			if (time_left)
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 6523108..d4f2861 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -572,7 +572,7 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 	if (ret)
 		return ret;
 
-	btrfs_wait_ordered_extents(root);
+	btrfs_wait_ordered_extents(root, -1);
 
 	pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_NOFS);
 	if (!pending_snapshot)
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index 8a5eff3..25a8f38 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -565,10 +565,11 @@ static void btrfs_run_ordered_extent_work(struct btrfs_work *work)
  * wait for all the ordered extents in a root.  This is done when balancing
  * space between drives.
  */
-void btrfs_wait_ordered_extents(struct btrfs_root *root)
+int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr)
 {
 	struct list_head splice, works;
 	struct btrfs_ordered_extent *ordered, *next;
+	int count = 0;
 
 	INIT_LIST_HEAD(&splice);
 	INIT_LIST_HEAD(&works);
@@ -576,7 +577,7 @@ void btrfs_wait_ordered_extents(struct btrfs_root *root)
 	mutex_lock(&root->fs_info->ordered_operations_mutex);
 	spin_lock(&root->ordered_extent_lock);
 	list_splice_init(&root->ordered_extents, &splice);
-	while (!list_empty(&splice)) {
+	while (!list_empty(&splice) && nr) {
 		ordered = list_first_entry(&splice, struct btrfs_ordered_extent,
 					   root_extent_list);
 		list_move_tail(&ordered->root_extent_list,
@@ -591,7 +592,11 @@ void btrfs_wait_ordered_extents(struct btrfs_root *root)
 
 		cond_resched();
 		spin_lock(&root->ordered_extent_lock);
+		if (nr != -1)
+			nr--;
+		count++;
 	}
+	list_splice_tail(&splice, &root->ordered_extents);
 	spin_unlock(&root->ordered_extent_lock);
 
 	list_for_each_entry_safe(ordered, next, &works, work_list) {
@@ -601,18 +606,21 @@ void btrfs_wait_ordered_extents(struct btrfs_root *root)
 		cond_resched();
 	}
 	mutex_unlock(&root->fs_info->ordered_operations_mutex);
+
+	return count;
 }
 
-void btrfs_wait_all_ordered_extents(struct btrfs_fs_info *fs_info)
+void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr)
 {
 	struct btrfs_root *root;
 	struct list_head splice;
+	int done;
 
 	INIT_LIST_HEAD(&splice);
 
 	spin_lock(&fs_info->ordered_root_lock);
 	list_splice_init(&fs_info->ordered_roots, &splice);
-	while (!list_empty(&splice)) {
+	while (!list_empty(&splice) && nr) {
 		root = list_first_entry(&splice, struct btrfs_root,
 					ordered_root);
 		root = btrfs_grab_fs_root(root);
@@ -621,10 +629,14 @@ void btrfs_wait_all_ordered_extents(struct btrfs_fs_info *fs_info)
 			       &fs_info->ordered_roots);
 		spin_unlock(&fs_info->ordered_root_lock);
 
-		btrfs_wait_ordered_extents(root);
+		done = btrfs_wait_ordered_extents(root, nr);
 		btrfs_put_fs_root(root);
 
 		spin_lock(&fs_info->ordered_root_lock);
+		if (nr != -1) {
+			nr -= done;
+			WARN_ON(nr < 0);
+		}
 	}
 	spin_unlock(&fs_info->ordered_root_lock);
 }
diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h
index 3982db1..9b0450f 100644
--- a/fs/btrfs/ordered-data.h
+++ b/fs/btrfs/ordered-data.h
@@ -195,8 +195,8 @@ int btrfs_run_ordered_operations(struct btrfs_trans_handle *trans,
 void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
 				 struct btrfs_root *root,
 				 struct inode *inode);
-void btrfs_wait_ordered_extents(struct btrfs_root *root);
-void btrfs_wait_all_ordered_extents(struct btrfs_fs_info *fs_info);
+int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr);
+void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr);
 void btrfs_get_logged_extents(struct btrfs_root *log, struct inode *inode);
 void btrfs_wait_logged_extents(struct btrfs_root *log, u64 transid);
 void btrfs_free_logged_extents(struct btrfs_root *log, u64 transid);
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 70eca79..e1b3c2c 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -4227,7 +4227,7 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 		err = ret;
 		goto out;
 	}
-	btrfs_wait_all_ordered_extents(fs_info);
+	btrfs_wait_ordered_roots(fs_info, -1);
 
 	while (1) {
 		mutex_lock(&fs_info->cleaner_mutex);
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 1de6d4d..2d8ac1b 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -920,7 +920,7 @@ int btrfs_sync_fs(struct super_block *sb, int wait)
 		return 0;
 	}
 
-	btrfs_wait_all_ordered_extents(fs_info);
+	btrfs_wait_ordered_roots(fs_info, -1);
 
 	trans = btrfs_attach_transaction_barrier(root);
 	if (IS_ERR(trans)) {
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 277fe81..32c100b 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -1636,7 +1636,7 @@ static inline int btrfs_start_delalloc_flush(struct btrfs_fs_info *fs_info)
 static inline void btrfs_wait_delalloc_flush(struct btrfs_fs_info *fs_info)
 {
 	if (btrfs_test_opt(fs_info->tree_root, FLUSHONCOMMIT))
-		btrfs_wait_all_ordered_extents(fs_info);
+		btrfs_wait_ordered_roots(fs_info, -1);
 }
 
 int btrfs_commit_transaction(struct btrfs_trans_handle *trans,