fs/fs-writeback.c: add a new writeback list for sync

wait_sb_inodes() currently does a walk of all inodes in the filesystem
to find dirty one to wait on during sync.  This is highly inefficient
and wastes a lot of CPU when there are lots of clean cached inodes that
we don't need to wait on.

To avoid this "all inode" walk, we need to track inodes that are
currently under writeback that we need to wait for.  We do this by
adding inodes to a writeback list on the sb when the mapping is first
tagged as having pages under writeback.  wait_sb_inodes() can then walk
this list of "inodes under IO" and wait specifically just for the inodes
that the current sync(2) needs to wait for.

Define a couple helpers to add/remove an inode from the writeback list
and call them when the overall mapping is tagged for or cleared from
writeback.  Update wait_sb_inodes() to walk only the inodes under
writeback due to the sync.

With this change, filesystem sync times are significantly reduced for
fs' with largely populated inode caches and otherwise no other work to
do.  For example, on a 16xcpu 2GHz x86-64 server, 10TB XFS filesystem
with a ~10m entry inode cache, sync times are reduced from ~7.3s to less
than 0.1s when the filesystem is fully clean.

Link: http://lkml.kernel.org/r/1466594593-6757-2-git-send-email-bfoster@redhat.com
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Tested-by: Holger Hoffstätte <holger.hoffstaette@applied-asynchrony.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

3 files changed, 85 insertions, 25 deletions

diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index fe7e83a45eff..1fcce8345da3 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -981,6 +981,37 @@ void inode_io_list_del(struct inode *inode)
 }
 
 /*
+ * mark an inode as under writeback on the sb
+ */
+void sb_mark_inode_writeback(struct inode *inode)
+{
+	struct super_block *sb = inode->i_sb;
+	unsigned long flags;
+
+	if (list_empty(&inode->i_wb_list)) {
+		spin_lock_irqsave(&sb->s_inode_wblist_lock, flags);
+		if (list_empty(&inode->i_wb_list))
+			list_add_tail(&inode->i_wb_list, &sb->s_inodes_wb);
+		spin_unlock_irqrestore(&sb->s_inode_wblist_lock, flags);
+	}
+}
+
+/*
+ * clear an inode as under writeback on the sb
+ */
+void sb_clear_inode_writeback(struct inode *inode)
+{
+	struct super_block *sb = inode->i_sb;
+	unsigned long flags;
+
+	if (!list_empty(&inode->i_wb_list)) {
+		spin_lock_irqsave(&sb->s_inode_wblist_lock, flags);
+		list_del_init(&inode->i_wb_list);
+		spin_unlock_irqrestore(&sb->s_inode_wblist_lock, flags);
+	}
+}
+
+/*
  * Redirty an inode: set its when-it-was dirtied timestamp and move it to the
  * furthest end of its superblock's dirty-inode list.
  *
@@ -2154,7 +2185,7 @@ EXPORT_SYMBOL(__mark_inode_dirty);
  */
 static void wait_sb_inodes(struct super_block *sb)
 {
-	struct inode *inode, *old_inode = NULL;
+	LIST_HEAD(sync_list);
 
 	/*
 	 * We need to be protected against the filesystem going from
@@ -2163,38 +2194,60 @@ static void wait_sb_inodes(struct super_block *sb)
 	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
 	mutex_lock(&sb->s_sync_lock);
-	spin_lock(&sb->s_inode_list_lock);
 
 	/*
-	 * Data integrity sync. Must wait for all pages under writeback,
-	 * because there may have been pages dirtied before our sync
-	 * call, but which had writeout started before we write it out.
-	 * In which case, the inode may not be on the dirty list, but
-	 * we still have to wait for that writeout.
+	 * Splice the writeback list onto a temporary list to avoid waiting on
+	 * inodes that have started writeback after this point.
+	 *
+	 * Use rcu_read_lock() to keep the inodes around until we have a
+	 * reference. s_inode_wblist_lock protects sb->s_inodes_wb as well as
+	 * the local list because inodes can be dropped from either by writeback
+	 * completion.
+	 */
+	rcu_read_lock();
+	spin_lock_irq(&sb->s_inode_wblist_lock);
+	list_splice_init(&sb->s_inodes_wb, &sync_list);
+
+	/*
+	 * Data integrity sync. Must wait for all pages under writeback, because
+	 * there may have been pages dirtied before our sync call, but which had
+	 * writeout started before we write it out.  In which case, the inode
+	 * may not be on the dirty list, but we still have to wait for that
+	 * writeout.
 	 */
-	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
+	while (!list_empty(&sync_list)) {
+		struct inode *inode = list_first_entry(&sync_list, struct inode,
+						       i_wb_list);
 		struct address_space *mapping = inode->i_mapping;
 
+		/*
+		 * Move each inode back to the wb list before we drop the lock
+		 * to preserve consistency between i_wb_list and the mapping
+		 * writeback tag. Writeback completion is responsible to remove
+		 * the inode from either list once the writeback tag is cleared.
+		 */
+		list_move_tail(&inode->i_wb_list, &sb->s_inodes_wb);
+
+		/*
+		 * The mapping can appear untagged while still on-list since we
+		 * do not have the mapping lock. Skip it here, wb completion
+		 * will remove it.
+		 */
+		if (!mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK))
+			continue;
+
+		spin_unlock_irq(&sb->s_inode_wblist_lock);
+
 		spin_lock(&inode->i_lock);
-		if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) ||
-		    (mapping->nrpages == 0)) {
+		if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) {
 			spin_unlock(&inode->i_lock);
+
+			spin_lock_irq(&sb->s_inode_wblist_lock);
 			continue;
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&sb->s_inode_list_lock);
-
-		/*
-		 * We hold a reference to 'inode' so it couldn't have been
-		 * removed from s_inodes list while we dropped the
-		 * s_inode_list_lock.  We cannot iput the inode now as we can
-		 * be holding the last reference and we cannot iput it under
-		 * s_inode_list_lock. So we keep the reference and iput it
-		 * later.
-		 */
-		iput(old_inode);
-		old_inode = inode;
+		rcu_read_unlock();
 
 		/*
 		 * We keep the error status of individual mapping so that
@@ -2205,10 +2258,13 @@ static void wait_sb_inodes(struct super_block *sb)
 
 		cond_resched();
 
-		spin_lock(&sb->s_inode_list_lock);
+		iput(inode);
+
+		rcu_read_lock();
+		spin_lock_irq(&sb->s_inode_wblist_lock);
 	}
-	spin_unlock(&sb->s_inode_list_lock);
-	iput(old_inode);
+	spin_unlock_irq(&sb->s_inode_wblist_lock);
+	rcu_read_unlock();
 	mutex_unlock(&sb->s_sync_lock);
 }
 
diff --git a/fs/inode.c b/fs/inode.c
index 4ccbc21b30ce..e171f7b5f9e4 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -365,6 +365,7 @@ void inode_init_once(struct inode *inode)
 	INIT_HLIST_NODE(&inode->i_hash);
 	INIT_LIST_HEAD(&inode->i_devices);
 	INIT_LIST_HEAD(&inode->i_io_list);
+	INIT_LIST_HEAD(&inode->i_wb_list);
 	INIT_LIST_HEAD(&inode->i_lru);
 	address_space_init_once(&inode->i_data);
 	i_size_ordered_init(inode);
@@ -507,6 +508,7 @@ void clear_inode(struct inode *inode)
 	BUG_ON(!list_empty(&inode->i_data.private_list));
 	BUG_ON(!(inode->i_state & I_FREEING));
 	BUG_ON(inode->i_state & I_CLEAR);
+	BUG_ON(!list_empty(&inode->i_wb_list));
 	/* don't need i_lock here, no concurrent mods to i_state */
 	inode->i_state = I_FREEING | I_CLEAR;
 }
diff --git a/fs/super.c b/fs/super.c
index d78b9847e6cb..5806ffd45563 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -206,6 +206,8 @@ static struct super_block *alloc_super(struct file_system_type *type, int flags)
 	mutex_init(&s->s_sync_lock);
 	INIT_LIST_HEAD(&s->s_inodes);
 	spin_lock_init(&s->s_inode_list_lock);
+	INIT_LIST_HEAD(&s->s_inodes_wb);
+	spin_lock_init(&s->s_inode_wblist_lock);
 
 	if (list_lru_init_memcg(&s->s_dentry_lru))
 		goto fail;