Merge tag 'for-f2fs-3.15' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs

Pull f2fs updates from Jaegeuk Kim:
 "This patch-set includes the following major enhancement patches.
   - introduce large directory support
   - introduce f2fs_issue_flush to merge redundant flush commands
   - merge write IOs as much as possible aligned to the segment
   - add sysfs entries to tune the f2fs configuration
   - use radix_tree for the free_nid_list to reduce in-memory operations
   - remove costly bit operations in f2fs_find_entry
   - enhance the readahead flow for CP/NAT/SIT/SSA blocks

  The other bug fixes are as follows:
   - recover xattr node blocks correctly after sudden-power-cut
   - fix to calculate the maximum number of node ids
   - enhance to handle many error cases

  And, there are a bunch of cleanups"

* tag 'for-f2fs-3.15' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (62 commits)
  f2fs: fix wrong statistics of inline data
  f2fs: check the acl's validity before setting
  f2fs: introduce f2fs_issue_flush to avoid redundant flush issue
  f2fs: fix to cover io->bio with io_rwsem
  f2fs: fix error path when fail to read inline data
  f2fs: use list_for_each_entry{_safe} for simplyfying code
  f2fs: avoid free slab cache under spinlock
  f2fs: avoid unneeded lookup when xattr name length is too long
  f2fs: avoid unnecessary bio submit when wait page writeback
  f2fs: return -EIO when node id is not matched
  f2fs: avoid RECLAIM_FS-ON-W warning
  f2fs: skip unnecessary node writes during fsync
  f2fs: introduce fi->i_sem to protect fi's info
  f2fs: change reclaim rate in percentage
  f2fs: add missing documentation for dir_level
  f2fs: remove unnecessary threshold
  f2fs: throttle the memory footprint with a sysfs entry
  f2fs: avoid to drop nat entries due to the negative nr_shrink
  f2fs: call f2fs_wait_on_page_writeback instead of native function
  f2fs: introduce nr_pages_to_write for segment alignment
  ...

18 files changed, 902 insertions, 506 deletions

diff --git a/fs/f2fs/acl.c b/fs/f2fs/acl.c
index fa8da4cb8c4b..e93e4ec7d165 100644
--- a/fs/f2fs/acl.c
+++ b/fs/f2fs/acl.c
@@ -174,7 +174,7 @@ struct posix_acl *f2fs_get_acl(struct inode *inode, int type)
 
 	retval = f2fs_getxattr(inode, name_index, "", NULL, 0);
 	if (retval > 0) {
-		value = kmalloc(retval, GFP_KERNEL);
+		value = kmalloc(retval, GFP_F2FS_ZERO);
 		if (!value)
 			return ERR_PTR(-ENOMEM);
 		retval = f2fs_getxattr(inode, name_index, "", value, retval);
@@ -203,6 +203,12 @@ static int __f2fs_set_acl(struct inode *inode, int type,
 	size_t size = 0;
 	int error;
 
+	if (acl) {
+		error = posix_acl_valid(acl);
+		if (error < 0)
+			return error;
+	}
+
 	switch (type) {
 	case ACL_TYPE_ACCESS:
 		name_index = F2FS_XATTR_INDEX_POSIX_ACL_ACCESS;
diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
index 293d0486a40f..4aa521aa9bc3 100644
--- a/fs/f2fs/checkpoint.c
+++ b/fs/f2fs/checkpoint.c
@@ -33,14 +33,12 @@ struct page *grab_meta_page(struct f2fs_sb_info *sbi, pgoff_t index)
 	struct address_space *mapping = META_MAPPING(sbi);
 	struct page *page = NULL;
 repeat:
-	page = grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS);
 	if (!page) {
 		cond_resched();
 		goto repeat;
 	}
 
-	/* We wait writeback only inside grab_meta_page() */
-	wait_on_page_writeback(page);
 	SetPageUptodate(page);
 	return page;
 }
@@ -75,23 +73,102 @@ out:
 	return page;
 }
 
+inline int get_max_meta_blks(struct f2fs_sb_info *sbi, int type)
+{
+	switch (type) {
+	case META_NAT:
+		return NM_I(sbi)->max_nid / NAT_ENTRY_PER_BLOCK;
+	case META_SIT:
+		return SIT_BLK_CNT(sbi);
+	case META_SSA:
+	case META_CP:
+		return 0;
+	default:
+		BUG();
+	}
+}
+
+/*
+ * Readahead CP/NAT/SIT/SSA pages
+ */
+int ra_meta_pages(struct f2fs_sb_info *sbi, int start, int nrpages, int type)
+{
+	block_t prev_blk_addr = 0;
+	struct page *page;
+	int blkno = start;
+	int max_blks = get_max_meta_blks(sbi, type);
+
+	struct f2fs_io_info fio = {
+		.type = META,
+		.rw = READ_SYNC | REQ_META | REQ_PRIO
+	};
+
+	for (; nrpages-- > 0; blkno++) {
+		block_t blk_addr;
+
+		switch (type) {
+		case META_NAT:
+			/* get nat block addr */
+			if (unlikely(blkno >= max_blks))
+				blkno = 0;
+			blk_addr = current_nat_addr(sbi,
+					blkno * NAT_ENTRY_PER_BLOCK);
+			break;
+		case META_SIT:
+			/* get sit block addr */
+			if (unlikely(blkno >= max_blks))
+				goto out;
+			blk_addr = current_sit_addr(sbi,
+					blkno * SIT_ENTRY_PER_BLOCK);
+			if (blkno != start && prev_blk_addr + 1 != blk_addr)
+				goto out;
+			prev_blk_addr = blk_addr;
+			break;
+		case META_SSA:
+		case META_CP:
+			/* get ssa/cp block addr */
+			blk_addr = blkno;
+			break;
+		default:
+			BUG();
+		}
+
+		page = grab_cache_page(META_MAPPING(sbi), blk_addr);
+		if (!page)
+			continue;
+		if (PageUptodate(page)) {
+			mark_page_accessed(page);
+			f2fs_put_page(page, 1);
+			continue;
+		}
+
+		f2fs_submit_page_mbio(sbi, page, blk_addr, &fio);
+		mark_page_accessed(page);
+		f2fs_put_page(page, 0);
+	}
+out:
+	f2fs_submit_merged_bio(sbi, META, READ);
+	return blkno - start;
+}
+
 static int f2fs_write_meta_page(struct page *page,
 				struct writeback_control *wbc)
 {
 	struct inode *inode = page->mapping->host;
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 
-	/* Should not write any meta pages, if any IO error was occurred */
-	if (unlikely(sbi->por_doing ||
-			is_set_ckpt_flags(F2FS_CKPT(sbi), CP_ERROR_FLAG)))
+	if (unlikely(sbi->por_doing))
 		goto redirty_out;
-
 	if (wbc->for_reclaim)
 		goto redirty_out;
 
-	wait_on_page_writeback(page);
+	/* Should not write any meta pages, if any IO error was occurred */
+	if (unlikely(is_set_ckpt_flags(F2FS_CKPT(sbi), CP_ERROR_FLAG)))
+		goto no_write;
 
+	f2fs_wait_on_page_writeback(page, META);
 	write_meta_page(sbi, page);
+no_write:
 	dec_page_count(sbi, F2FS_DIRTY_META);
 	unlock_page(page);
 	return 0;
@@ -99,6 +176,7 @@ static int f2fs_write_meta_page(struct page *page,
 redirty_out:
 	dec_page_count(sbi, F2FS_DIRTY_META);
 	wbc->pages_skipped++;
+	account_page_redirty(page);
 	set_page_dirty(page);
 	return AOP_WRITEPAGE_ACTIVATE;
 }
@@ -107,21 +185,23 @@ static int f2fs_write_meta_pages(struct address_space *mapping,
 				struct writeback_control *wbc)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(mapping->host->i_sb);
-	int nrpages = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
-	long written;
-
-	if (wbc->for_kupdate)
-		return 0;
+	long diff, written;
 
 	/* collect a number of dirty meta pages and write together */
-	if (get_pages(sbi, F2FS_DIRTY_META) < nrpages)
-		return 0;
+	if (wbc->for_kupdate ||
+		get_pages(sbi, F2FS_DIRTY_META) < nr_pages_to_skip(sbi, META))
+		goto skip_write;
 
 	/* if mounting is failed, skip writing node pages */
 	mutex_lock(&sbi->cp_mutex);
-	written = sync_meta_pages(sbi, META, nrpages);
+	diff = nr_pages_to_write(sbi, META, wbc);
+	written = sync_meta_pages(sbi, META, wbc->nr_to_write);
 	mutex_unlock(&sbi->cp_mutex);
-	wbc->nr_to_write -= written;
+	wbc->nr_to_write = max((long)0, wbc->nr_to_write - written - diff);
+	return 0;
+
+skip_write:
+	wbc->pages_skipped += get_pages(sbi, F2FS_DIRTY_META);
 	return 0;
 }
 
@@ -148,10 +228,22 @@ long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
 
 		for (i = 0; i < nr_pages; i++) {
 			struct page *page = pvec.pages[i];
+
 			lock_page(page);
-			f2fs_bug_on(page->mapping != mapping);
-			f2fs_bug_on(!PageDirty(page));
-			clear_page_dirty_for_io(page);
+
+			if (unlikely(page->mapping != mapping)) {
+continue_unlock:
+				unlock_page(page);
+				continue;
+			}
+			if (!PageDirty(page)) {
+				/* someone wrote it for us */
+				goto continue_unlock;
+			}
+
+			if (!clear_page_dirty_for_io(page))
+				goto continue_unlock;
+
 			if (f2fs_write_meta_page(page, &wbc)) {
 				unlock_page(page);
 				break;
@@ -216,16 +308,15 @@ void release_orphan_inode(struct f2fs_sb_info *sbi)
 
 void add_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 {
-	struct list_head *head, *this;
-	struct orphan_inode_entry *new = NULL, *orphan = NULL;
+	struct list_head *head;
+	struct orphan_inode_entry *new, *orphan;
 
 	new = f2fs_kmem_cache_alloc(orphan_entry_slab, GFP_ATOMIC);
 	new->ino = ino;
 
 	spin_lock(&sbi->orphan_inode_lock);
 	head = &sbi->orphan_inode_list;
-	list_for_each(this, head) {
-		orphan = list_entry(this, struct orphan_inode_entry, list);
+	list_for_each_entry(orphan, head, list) {
 		if (orphan->ino == ino) {
 			spin_unlock(&sbi->orphan_inode_lock);
 			kmem_cache_free(orphan_entry_slab, new);
@@ -234,14 +325,10 @@ void add_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 
 		if (orphan->ino > ino)
 			break;
-		orphan = NULL;
 	}
 
-	/* add new_oentry into list which is sorted by inode number */
-	if (orphan)
-		list_add(&new->list, this->prev);
-	else
-		list_add_tail(&new->list, head);
+	/* add new orphan entry into list which is sorted by inode number */
+	list_add_tail(&new->list, &orphan->list);
 	spin_unlock(&sbi->orphan_inode_lock);
 }
 
@@ -255,10 +342,11 @@ void remove_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 	list_for_each_entry(orphan, head, list) {
 		if (orphan->ino == ino) {
 			list_del(&orphan->list);
-			kmem_cache_free(orphan_entry_slab, orphan);
 			f2fs_bug_on(sbi->n_orphans == 0);
 			sbi->n_orphans--;
-			break;
+			spin_unlock(&sbi->orphan_inode_lock);
+			kmem_cache_free(orphan_entry_slab, orphan);
+			return;
 		}
 	}
 	spin_unlock(&sbi->orphan_inode_lock);
@@ -285,6 +373,8 @@ void recover_orphan_inodes(struct f2fs_sb_info *sbi)
 	start_blk = __start_cp_addr(sbi) + 1;
 	orphan_blkaddr = __start_sum_addr(sbi) - 1;
 
+	ra_meta_pages(sbi, start_blk, orphan_blkaddr, META_CP);
+
 	for (i = 0; i < orphan_blkaddr; i++) {
 		struct page *page = get_meta_page(sbi, start_blk + i);
 		struct f2fs_orphan_block *orphan_blk;
@@ -466,14 +556,12 @@ static int __add_dirty_inode(struct inode *inode, struct dir_inode_entry *new)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	struct list_head *head = &sbi->dir_inode_list;
-	struct list_head *this;
+	struct dir_inode_entry *entry;
 
-	list_for_each(this, head) {
-		struct dir_inode_entry *entry;
-		entry = list_entry(this, struct dir_inode_entry, list);
+	list_for_each_entry(entry, head, list)
 		if (unlikely(entry->inode == inode))
 			return -EEXIST;
-	}
+
 	list_add_tail(&new->list, head);
 	stat_inc_dirty_dir(sbi);
 	return 0;
@@ -483,6 +571,7 @@ void set_dirty_dir_page(struct inode *inode, struct page *page)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	struct dir_inode_entry *new;
+	int ret = 0;
 
 	if (!S_ISDIR(inode->i_mode))
 		return;
@@ -492,13 +581,13 @@ void set_dirty_dir_page(struct inode *inode, struct page *page)
 	INIT_LIST_HEAD(&new->list);
 
 	spin_lock(&sbi->dir_inode_lock);
-	if (__add_dirty_inode(inode, new))
-		kmem_cache_free(inode_entry_slab, new);
-
-	inc_page_count(sbi, F2FS_DIRTY_DENTS);
+	ret = __add_dirty_inode(inode, new);
 	inode_inc_dirty_dents(inode);
 	SetPagePrivate(page);
 	spin_unlock(&sbi->dir_inode_lock);
+
+	if (ret)
+		kmem_cache_free(inode_entry_slab, new);
 }
 
 void add_dirty_dir_inode(struct inode *inode)
@@ -506,44 +595,47 @@ void add_dirty_dir_inode(struct inode *inode)
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	struct dir_inode_entry *new =
 			f2fs_kmem_cache_alloc(inode_entry_slab, GFP_NOFS);
+	int ret = 0;
 
 	new->inode = inode;
 	INIT_LIST_HEAD(&new->list);
 
 	spin_lock(&sbi->dir_inode_lock);
-	if (__add_dirty_inode(inode, new))
-		kmem_cache_free(inode_entry_slab, new);
+	ret = __add_dirty_inode(inode, new);
 	spin_unlock(&sbi->dir_inode_lock);
+
+	if (ret)
+		kmem_cache_free(inode_entry_slab, new);
 }
 
 void remove_dirty_dir_inode(struct inode *inode)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
-
-	struct list_head *this, *head;
+	struct list_head *head;
+	struct dir_inode_entry *entry;
 
 	if (!S_ISDIR(inode->i_mode))
 		return;
 
 	spin_lock(&sbi->dir_inode_lock);
-	if (atomic_read(&F2FS_I(inode)->dirty_dents)) {
+	if (get_dirty_dents(inode)) {
 		spin_unlock(&sbi->dir_inode_lock);
 		return;
 	}
 
 	head = &sbi->dir_inode_list;
-	list_for_each(this, head) {
-		struct dir_inode_entry *entry;
-		entry = list_entry(this, struct dir_inode_entry, list);
+	list_for_each_entry(entry, head, list) {
 		if (entry->inode == inode) {
 			list_del(&entry->list);
-			kmem_cache_free(inode_entry_slab, entry);
 			stat_dec_dirty_dir(sbi);
-			break;
+			spin_unlock(&sbi->dir_inode_lock);
+			kmem_cache_free(inode_entry_slab, entry);
+			goto done;
 		}
 	}
 	spin_unlock(&sbi->dir_inode_lock);
 
+done:
 	/* Only from the recovery routine */
 	if (is_inode_flag_set(F2FS_I(inode), FI_DELAY_IPUT)) {
 		clear_inode_flag(F2FS_I(inode), FI_DELAY_IPUT);
@@ -554,15 +646,14 @@ void remove_dirty_dir_inode(struct inode *inode)
 struct inode *check_dirty_dir_inode(struct f2fs_sb_info *sbi, nid_t ino)
 {
 
-	struct list_head *this, *head;
+	struct list_head *head;
 	struct inode *inode = NULL;
+	struct dir_inode_entry *entry;
 
 	spin_lock(&sbi->dir_inode_lock);
 
 	head = &sbi->dir_inode_list;
-	list_for_each(this, head) {
-		struct dir_inode_entry *entry;
-		entry = list_entry(this, struct dir_inode_entry, list);
+	list_for_each_entry(entry, head, list) {
 		if (entry->inode->i_ino == ino) {
 			inode = entry->inode;
 			break;
@@ -589,7 +680,7 @@ retry:
 	inode = igrab(entry->inode);
 	spin_unlock(&sbi->dir_inode_lock);
 	if (inode) {
-		filemap_flush(inode->i_mapping);
+		filemap_fdatawrite(inode->i_mapping);
 		iput(inode);
 	} else {
 		/*
@@ -824,6 +915,7 @@ void write_checkpoint(struct f2fs_sb_info *sbi, bool is_umount)
 	unblock_operations(sbi);
 	mutex_unlock(&sbi->cp_mutex);
 
+	stat_inc_cp_count(sbi->stat_info);
 	trace_f2fs_write_checkpoint(sbi->sb, is_umount, "finish checkpoint");
 }
 
@@ -845,11 +937,11 @@ void init_orphan_info(struct f2fs_sb_info *sbi)
 int __init create_checkpoint_caches(void)
 {
 	orphan_entry_slab = f2fs_kmem_cache_create("f2fs_orphan_entry",
-			sizeof(struct orphan_inode_entry), NULL);
+			sizeof(struct orphan_inode_entry));
 	if (!orphan_entry_slab)
 		return -ENOMEM;
 	inode_entry_slab = f2fs_kmem_cache_create("f2fs_dirty_dir_entry",
-			sizeof(struct dir_inode_entry), NULL);
+			sizeof(struct dir_inode_entry));
 	if (!inode_entry_slab) {
 		kmem_cache_destroy(orphan_entry_slab);
 		return -ENOMEM;
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 2261ccdd0b5f..45abd60e2bff 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -45,7 +45,7 @@ static void f2fs_read_end_io(struct bio *bio, int err)
 
 static void f2fs_write_end_io(struct bio *bio, int err)
 {
-	struct f2fs_sb_info *sbi = F2FS_SB(bio->bi_io_vec->bv_page->mapping->host->i_sb);
+	struct f2fs_sb_info *sbi = bio->bi_private;
 	struct bio_vec *bvec;
 	int i;
 
@@ -55,15 +55,16 @@ static void f2fs_write_end_io(struct bio *bio, int err)
 		if (unlikely(err)) {
 			SetPageError(page);
 			set_bit(AS_EIO, &page->mapping->flags);
-			set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG);
-			sbi->sb->s_flags |= MS_RDONLY;
+			f2fs_stop_checkpoint(sbi);
 		}
 		end_page_writeback(page);
 		dec_page_count(sbi, F2FS_WRITEBACK);
 	}
 
-	if (bio->bi_private)
-		complete(bio->bi_private);
+	if (sbi->wait_io) {
+		complete(sbi->wait_io);
+		sbi->wait_io = NULL;
+	}
 
 	if (!get_pages(sbi, F2FS_WRITEBACK) &&
 			!list_empty(&sbi->cp_wait.task_list))
@@ -86,6 +87,7 @@ static struct bio *__bio_alloc(struct f2fs_sb_info *sbi, block_t blk_addr,
 	bio->bi_bdev = sbi->sb->s_bdev;
 	bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
 	bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;
+	bio->bi_private = sbi;
 
 	return bio;
 }
@@ -113,7 +115,7 @@ static void __submit_merged_bio(struct f2fs_bio_info *io)
 		 */
 		if (fio->type == META_FLUSH) {
 			DECLARE_COMPLETION_ONSTACK(wait);
-			io->bio->bi_private = &wait;
+			io->sbi->wait_io = &wait;
 			submit_bio(rw, io->bio);
 			wait_for_completion(&wait);
 		} else {
@@ -132,7 +134,7 @@ void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
 
 	io = is_read_io(rw) ? &sbi->read_io : &sbi->write_io[btype];
 
-	mutex_lock(&io->io_mutex);
+	down_write(&io->io_rwsem);
 
 	/* change META to META_FLUSH in the checkpoint procedure */
 	if (type >= META_FLUSH) {
@@ -140,7 +142,7 @@ void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
 		io->fio.rw = WRITE_FLUSH_FUA | REQ_META | REQ_PRIO;
 	}
 	__submit_merged_bio(io);
-	mutex_unlock(&io->io_mutex);
+	up_write(&io->io_rwsem);
 }
 
 /*
@@ -178,7 +180,7 @@ void f2fs_submit_page_mbio(struct f2fs_sb_info *sbi, struct page *page,
 
 	verify_block_addr(sbi, blk_addr);
 
-	mutex_lock(&io->io_mutex);
+	down_write(&io->io_rwsem);
 
 	if (!is_read)
 		inc_page_count(sbi, F2FS_WRITEBACK);
@@ -202,7 +204,7 @@ alloc_new:
 
 	io->last_block_in_bio = blk_addr;
 
-	mutex_unlock(&io->io_mutex);
+	up_write(&io->io_rwsem);
 	trace_f2fs_submit_page_mbio(page, fio->rw, fio->type, blk_addr);
 }
 
@@ -797,48 +799,36 @@ static int f2fs_write_data_page(struct page *page,
 	 */
 	offset = i_size & (PAGE_CACHE_SIZE - 1);
 	if ((page->index >= end_index + 1) || !offset) {
-		if (S_ISDIR(inode->i_mode)) {
-			dec_page_count(sbi, F2FS_DIRTY_DENTS);
-			inode_dec_dirty_dents(inode);
-		}
+		inode_dec_dirty_dents(inode);
 		goto out;
 	}
 
 	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
 write:
-	if (unlikely(sbi->por_doing)) {
-		err = AOP_WRITEPAGE_ACTIVATE;
+	if (unlikely(sbi->por_doing))
 		goto redirty_out;
-	}
 
 	/* Dentry blocks are controlled by checkpoint */
 	if (S_ISDIR(inode->i_mode)) {
-		dec_page_count(sbi, F2FS_DIRTY_DENTS);
 		inode_dec_dirty_dents(inode);
 		err = do_write_data_page(page, &fio);
-	} else {
-		f2fs_lock_op(sbi);
-
-		if (f2fs_has_inline_data(inode) || f2fs_may_inline(inode)) {
-			err = f2fs_write_inline_data(inode, page, offset);
-			f2fs_unlock_op(sbi);
-			goto out;
-		} else {
-			err = do_write_data_page(page, &fio);
-		}
+		goto done;
+	}
 
-		f2fs_unlock_op(sbi);
+	if (!wbc->for_reclaim)
 		need_balance_fs = true;
-	}
-	if (err == -ENOENT)
-		goto out;
-	else if (err)
+	else if (has_not_enough_free_secs(sbi, 0))
 		goto redirty_out;
 
-	if (wbc->for_reclaim) {
-		f2fs_submit_merged_bio(sbi, DATA, WRITE);
-		need_balance_fs = false;
-	}
+	f2fs_lock_op(sbi);
+	if (f2fs_has_inline_data(inode) || f2fs_may_inline(inode))
+		err = f2fs_write_inline_data(inode, page, offset);
+	else
+		err = do_write_data_page(page, &fio);
+	f2fs_unlock_op(sbi);
+done:
+	if (err && err != -ENOENT)
+		goto redirty_out;
 
 	clear_cold_data(page);
 out:
@@ -849,12 +839,11 @@ out:
 
 redirty_out:
 	wbc->pages_skipped++;
+	account_page_redirty(page);
 	set_page_dirty(page);
-	return err;
+	return AOP_WRITEPAGE_ACTIVATE;
 }
 
-#define MAX_DESIRED_PAGES_WP	4096
-
 static int __f2fs_writepage(struct page *page, struct writeback_control *wbc,
 			void *data)
 {
@@ -871,17 +860,17 @@ static int f2fs_write_data_pages(struct address_space *mapping,
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	bool locked = false;
 	int ret;
-	long excess_nrtw = 0, desired_nrtw;
+	long diff;
 
 	/* deal with chardevs and other special file */
 	if (!mapping->a_ops->writepage)
 		return 0;
 
-	if (wbc->nr_to_write < MAX_DESIRED_PAGES_WP) {
-		desired_nrtw = MAX_DESIRED_PAGES_WP;
-		excess_nrtw = desired_nrtw - wbc->nr_to_write;
-		wbc->nr_to_write = desired_nrtw;
-	}
+	if (S_ISDIR(inode->i_mode) && wbc->sync_mode == WB_SYNC_NONE &&
+			get_dirty_dents(inode) < nr_pages_to_skip(sbi, DATA))
+		goto skip_write;
+
+	diff = nr_pages_to_write(sbi, DATA, wbc);
 
 	if (!S_ISDIR(inode->i_mode)) {
 		mutex_lock(&sbi->writepages);
@@ -895,8 +884,12 @@ static int f2fs_write_data_pages(struct address_space *mapping,
 
 	remove_dirty_dir_inode(inode);
 
-	wbc->nr_to_write -= excess_nrtw;
+	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);
 	return ret;
+
+skip_write:
+	wbc->pages_skipped += get_dirty_dents(inode);
+	return 0;
 }
 
 static int f2fs_write_begin(struct file *file, struct address_space *mapping,
@@ -949,13 +942,19 @@ inline_data:
 	if (dn.data_blkaddr == NEW_ADDR) {
 		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
 	} else {
-		if (f2fs_has_inline_data(inode))
+		if (f2fs_has_inline_data(inode)) {
 			err = f2fs_read_inline_data(inode, page);
-		else
+			if (err) {
+				page_cache_release(page);
+				return err;
+			}
+		} else {
 			err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
 							READ_SYNC);
-		if (err)
-			return err;
+			if (err)
+				return err;
+		}
+
 		lock_page(page);
 		if (unlikely(!PageUptodate(page))) {
 			f2fs_put_page(page, 1);
@@ -1031,11 +1030,8 @@ static void f2fs_invalidate_data_page(struct page *page, unsigned int offset,
 				      unsigned int length)
 {
 	struct inode *inode = page->mapping->host;
-	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
-	if (S_ISDIR(inode->i_mode) && PageDirty(page)) {
-		dec_page_count(sbi, F2FS_DIRTY_DENTS);
+	if (PageDirty(page))
 		inode_dec_dirty_dents(inode);
-	}
 	ClearPagePrivate(page);
 }
 
diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index 3de9d20d0c14..b52c12cf5873 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -86,7 +86,6 @@ static void update_sit_info(struct f2fs_sb_info *sbi)
 {
 	struct f2fs_stat_info *si = F2FS_STAT(sbi);
 	unsigned int blks_per_sec, hblks_per_sec, total_vblocks, bimodal, dist;
-	struct sit_info *sit_i = SIT_I(sbi);
 	unsigned int segno, vblocks;
 	int ndirty = 0;
 
@@ -94,7 +93,6 @@ static void update_sit_info(struct f2fs_sb_info *sbi)
 	total_vblocks = 0;
 	blks_per_sec = sbi->segs_per_sec * (1 << sbi->log_blocks_per_seg);
 	hblks_per_sec = blks_per_sec / 2;
-	mutex_lock(&sit_i->sentry_lock);
 	for (segno = 0; segno < TOTAL_SEGS(sbi); segno += sbi->segs_per_sec) {
 		vblocks = get_valid_blocks(sbi, segno, sbi->segs_per_sec);
 		dist = abs(vblocks - hblks_per_sec);
@@ -105,7 +103,6 @@ static void update_sit_info(struct f2fs_sb_info *sbi)
 			ndirty++;
 		}
 	}
-	mutex_unlock(&sit_i->sentry_lock);
 	dist = TOTAL_SECS(sbi) * hblks_per_sec * hblks_per_sec / 100;
 	si->bimodal = bimodal / dist;
 	if (si->dirty_count)
@@ -236,6 +233,7 @@ static int stat_show(struct seq_file *s, void *v)
 			   si->dirty_count);
 		seq_printf(s, "  - Prefree: %d\n  - Free: %d (%d)\n\n",
 			   si->prefree_count, si->free_segs, si->free_secs);
+		seq_printf(s, "CP calls: %d\n", si->cp_count);
 		seq_printf(s, "GC calls: %d (BG: %d)\n",
 			   si->call_count, si->bg_gc);
 		seq_printf(s, "  - data segments : %d\n", si->data_segs);
@@ -252,10 +250,10 @@ static int stat_show(struct seq_file *s, void *v)
 			   si->ndirty_dent, si->ndirty_dirs);
 		seq_printf(s, "  - meta: %4d in %4d\n",
 			   si->ndirty_meta, si->meta_pages);
-		seq_printf(s, "  - NATs: %5d > %lu\n",
-			   si->nats, NM_WOUT_THRESHOLD);
-		seq_printf(s, "  - SITs: %5d\n  - free_nids: %5d\n",
-			   si->sits, si->fnids);
+		seq_printf(s, "  - NATs: %9d\n  - SITs: %9d\n",
+			   si->nats, si->sits);
+		seq_printf(s, "  - free_nids: %9d\n",
+			   si->fnids);
 		seq_puts(s, "\nDistribution of User Blocks:");
 		seq_puts(s, " [ valid | invalid | free ]\n");
 		seq_puts(s, "  [");
diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c
index 2b7c255bcbdf..972fd0ef230f 100644
--- a/fs/f2fs/dir.c
+++ b/fs/f2fs/dir.c
@@ -21,12 +21,12 @@ static unsigned long dir_blocks(struct inode *inode)
 							>> PAGE_CACHE_SHIFT;
 }
 
-static unsigned int dir_buckets(unsigned int level)
+static unsigned int dir_buckets(unsigned int level, int dir_level)
 {
 	if (level < MAX_DIR_HASH_DEPTH / 2)
-		return 1 << level;
+		return 1 << (level + dir_level);
 	else
-		return 1 << ((MAX_DIR_HASH_DEPTH / 2) - 1);
+		return 1 << ((MAX_DIR_HASH_DEPTH / 2 + dir_level) - 1);
 }
 
 static unsigned int bucket_blocks(unsigned int level)
@@ -65,13 +65,14 @@ static void set_de_type(struct f2fs_dir_entry *de, struct inode *inode)
 	de->file_type = f2fs_type_by_mode[(mode & S_IFMT) >> S_SHIFT];
 }
 
-static unsigned long dir_block_index(unsigned int level, unsigned int idx)
+static unsigned long dir_block_index(unsigned int level,
+				int dir_level, unsigned int idx)
 {
 	unsigned long i;
 	unsigned long bidx = 0;
 
 	for (i = 0; i < level; i++)
-		bidx += dir_buckets(i) * bucket_blocks(i);
+		bidx += dir_buckets(i, dir_level) * bucket_blocks(i);
 	bidx += idx * bucket_blocks(level);
 	return bidx;
 }
@@ -93,16 +94,21 @@ static struct f2fs_dir_entry *find_in_block(struct page *dentry_page,
 			f2fs_hash_t namehash, struct page **res_page)
 {
 	struct f2fs_dir_entry *de;
-	unsigned long bit_pos, end_pos, next_pos;
+	unsigned long bit_pos = 0;
 	struct f2fs_dentry_block *dentry_blk = kmap(dentry_page);
-	int slots;
+	const void *dentry_bits = &dentry_blk->dentry_bitmap;
+	int max_len = 0;
 
-	bit_pos = find_next_bit_le(&dentry_blk->dentry_bitmap,
-					NR_DENTRY_IN_BLOCK, 0);
 	while (bit_pos < NR_DENTRY_IN_BLOCK) {
+		if (!test_bit_le(bit_pos, dentry_bits)) {
+			if (bit_pos == 0)
+				max_len = 1;
+			else if (!test_bit_le(bit_pos - 1, dentry_bits))
+				max_len++;
+			bit_pos++;
+			continue;
+		}
 		de = &dentry_blk->dentry[bit_pos];
-		slots = GET_DENTRY_SLOTS(le16_to_cpu(de->name_len));
-
 		if (early_match_name(name, namelen, namehash, de)) {
 			if (!memcmp(dentry_blk->filename[bit_pos],
 							name, namelen)) {
@@ -110,20 +116,18 @@ static struct f2fs_dir_entry *find_in_block(struct page *dentry_page,
 				goto found;
 			}
 		}
-		next_pos = bit_pos + slots;
-		bit_pos = find_next_bit_le(&dentry_blk->dentry_bitmap,
-				NR_DENTRY_IN_BLOCK, next_pos);
-		if (bit_pos >= NR_DENTRY_IN_BLOCK)
-			end_pos = NR_DENTRY_IN_BLOCK;
-		else
-			end_pos = bit_pos;
-		if (*max_slots < end_pos - next_pos)
-			*max_slots = end_pos - next_pos;
+		if (max_len > *max_slots) {
+			*max_slots = max_len;
+			max_len = 0;
+		}
+		bit_pos += GET_DENTRY_SLOTS(le16_to_cpu(de->name_len));
 	}
 
 	de = NULL;
 	kunmap(dentry_page);
 found:
+	if (max_len > *max_slots)
+		*max_slots = max_len;
 	return de;
 }
 
@@ -141,10 +145,11 @@ static struct f2fs_dir_entry *find_in_level(struct inode *dir,
 
 	f2fs_bug_on(level > MAX_DIR_HASH_DEPTH);
 
-	nbucket = dir_buckets(level);
+	nbucket = dir_buckets(level, F2FS_I(dir)->i_dir_level);
 	nblock = bucket_blocks(level);
 
-	bidx = dir_block_index(level, le32_to_cpu(namehash) % nbucket);
+	bidx = dir_block_index(level, F2FS_I(dir)->i_dir_level,
+					le32_to_cpu(namehash) % nbucket);
 	end_block = bidx + nblock;
 
 	for (; bidx < end_block; bidx++) {
@@ -248,7 +253,7 @@ void f2fs_set_link(struct inode *dir, struct f2fs_dir_entry *de,
 		struct page *page, struct inode *inode)
 {
 	lock_page(page);
-	wait_on_page_writeback(page);
+	f2fs_wait_on_page_writeback(page, DATA);
 	de->ino = cpu_to_le32(inode->i_ino);
 	set_de_type(de, inode);
 	kunmap(page);
@@ -347,14 +352,11 @@ static struct page *init_inode_metadata(struct inode *inode,
 		err = f2fs_init_security(inode, dir, name, page);
 		if (err)
 			goto put_error;
-
-		wait_on_page_writeback(page);
 	} else {
 		page = get_node_page(F2FS_SB(dir->i_sb), inode->i_ino);
 		if (IS_ERR(page))
 			return page;
 
-		wait_on_page_writeback(page);
 		set_cold_node(inode, page);
 	}
 
@@ -372,6 +374,10 @@ static struct page *init_inode_metadata(struct inode *inode,
 
 put_error:
 	f2fs_put_page(page, 1);
+	/* once the failed inode becomes a bad inode, i_mode is S_IFREG */
+	truncate_inode_pages(&inode->i_data, 0);
+	truncate_blocks(inode, 0);
+	remove_dirty_dir_inode(inode);
 error:
 	remove_inode_page(inode);
 	return ERR_PTR(err);
@@ -395,9 +401,6 @@ static void update_parent_metadata(struct inode *dir, struct inode *inode,
 		set_inode_flag(F2FS_I(dir), FI_UPDATE_DIR);
 	}
 
-	if (is_inode_flag_set(F2FS_I(dir), FI_UPDATE_DIR))
-		update_inode_page(dir);
-
 	if (is_inode_flag_set(F2FS_I(inode), FI_INC_LINK))
 		clear_inode_flag(F2FS_I(inode), FI_INC_LINK);
 }
@@ -464,10 +467,11 @@ start:
 	if (level == current_depth)
 		++current_depth;
 
-	nbucket = dir_buckets(level);
+	nbucket = dir_buckets(level, F2FS_I(dir)->i_dir_level);
 	nblock = bucket_blocks(level);
 
-	bidx = dir_block_index(level, (le32_to_cpu(dentry_hash) % nbucket));
+	bidx = dir_block_index(level, F2FS_I(dir)->i_dir_level,
+				(le32_to_cpu(dentry_hash) % nbucket));
 
 	for (block = bidx; block <= (bidx + nblock - 1); block++) {
 		dentry_page = get_new_data_page(dir, NULL, block, true);
@@ -487,8 +491,9 @@ start:
 	++level;
 	goto start;
 add_dentry:
-	wait_on_page_writeback(dentry_page);
+	f2fs_wait_on_page_writeback(dentry_page, DATA);
 
+	down_write(&F2FS_I(inode)->i_sem);
 	page = init_inode_metadata(inode, dir, name);
 	if (IS_ERR(page)) {
 		err = PTR_ERR(page);
@@ -511,7 +516,12 @@ add_dentry:
 
 	update_parent_metadata(dir, inode, current_depth);
 fail:
-	clear_inode_flag(F2FS_I(dir), FI_UPDATE_DIR);
+	up_write(&F2FS_I(inode)->i_sem);
+
+	if (is_inode_flag_set(F2FS_I(dir), FI_UPDATE_DIR)) {
+		update_inode_page(dir);
+		clear_inode_flag(F2FS_I(dir), FI_UPDATE_DIR);
+	}
 	kunmap(dentry_page);
 	f2fs_put_page(dentry_page, 1);
 	return err;
@@ -528,13 +538,12 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
 	unsigned int bit_pos;
 	struct address_space *mapping = page->mapping;
 	struct inode *dir = mapping->host;
-	struct f2fs_sb_info *sbi = F2FS_SB(dir->i_sb);
 	int slots = GET_DENTRY_SLOTS(le16_to_cpu(dentry->name_len));
 	void *kaddr = page_address(page);
 	int i;
 
 	lock_page(page);
-	wait_on_page_writeback(page);
+	f2fs_wait_on_page_writeback(page, DATA);
 
 	dentry_blk = (struct f2fs_dentry_block *)kaddr;
 	bit_pos = dentry - (struct f2fs_dir_entry *)dentry_blk->dentry;
@@ -551,6 +560,10 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
 	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
 
 	if (inode) {
+		struct f2fs_sb_info *sbi = F2FS_SB(dir->i_sb);
+
+		down_write(&F2FS_I(inode)->i_sem);
+
 		if (S_ISDIR(inode->i_mode)) {
 			drop_nlink(dir);
 			update_inode_page(dir);
@@ -561,6 +574,7 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
 			drop_nlink(inode);
 			i_size_write(inode, 0);
 		}