Merge tag 'for-f2fs-3.19' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs

Pull f2fs updates from Jaegeuk Kim:
 "This patch-set includes lots of bug fixes based on clean-ups and
  refactored codes.  And inline_dir was introduced and two minor mount
  options were added.  Details from signed tag:

  This series includes the following enhancement with refactored flows.
   - fix inmemory page operations
   - fix wrong inline_data & inline_dir logics
   - enhance memory and IO control under memory pressure
   - consider preemption on radix_tree operation
   - fix memory leaks and deadlocks

  But also, there are a couple of new features:
   - support inline_dir to store dentries inside inode page
   - add -o fastboot to reduce booting time
   - implement -o dirsync

  And a lot of clean-ups and minor bug fixes as well"

* tag 'for-f2fs-3.19' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (88 commits)
  f2fs: avoid to ra unneeded blocks in recover flow
  f2fs: introduce is_valid_blkaddr to cleanup codes in ra_meta_pages
  f2fs: fix to enable readahead for SSA/CP blocks
  f2fs: use atomic for counting inode with inline_{dir,inode} flag
  f2fs: cleanup path to need cp at fsync
  f2fs: check if inode state is dirty at fsync
  f2fs: count the number of inmemory pages
  f2fs: release inmemory pages when the file was closed
  f2fs: set page private for inmemory pages for truncation
  f2fs: count inline_xx in do_read_inode
  f2fs: do retry operations with cond_resched
  f2fs: call radix_tree_preload before radix_tree_insert
  f2fs: use rw_semaphore for nat entry lock
  f2fs: fix missing kmem_cache_free
  f2fs: more fast lookup for gc_inode list
  f2fs: cleanup redundant macro
  f2fs: fix to return correct error number in f2fs_write_begin
  f2fs: cleanup if-statement of phase in gc_data_segment
  f2fs: fix to recover converted inline_data
  f2fs: make clean the page before writing
  ...

Documentation/filesystems/f2fs.txt

@@ -122,6 +122,10 @@ disable_ext_identify   Disable the extension list configured by mkfs, so f2fs
 inline_xattr           Enable the inline xattrs feature.
 inline_data            Enable the inline data feature: New created small(<~3.4k)
                        files can be written into inode block.
+inline_dentry          Enable the inline dir feature: data in new created
+                       directory entries can be written into inode block. The
+                       space of inode block which is used to store inline
+                       dentries is limited to ~3.4k.
 flush_merge	       Merge concurrent cache_flush commands as much as possible
                        to eliminate redundant command issues. If the underlying
 		       device handles the cache_flush command relatively slowly,
@@ -131,6 +135,9 @@ nobarrier              This option can be used if underlying storage guarantees
 		       If this option is set, no cache_flush commands are issued
 		       but f2fs still guarantees the write ordering of all the
 		       data writes.
+fastboot               This option is used when a system wants to reduce mount
+                       time as much as possible, even though normal performance
+		       can be sacrificed.
 
 ================================================================================
 DEBUGFS ENTRIES

fs/f2fs/acl.c

@@ -162,7 +162,8 @@ fail:
 	return ERR_PTR(-EINVAL);
 }
 
-struct posix_acl *f2fs_get_acl(struct inode *inode, int type)
+static struct posix_acl *__f2fs_get_acl(struct inode *inode, int type,
+						struct page *dpage)
 {
 	int name_index = F2FS_XATTR_INDEX_POSIX_ACL_DEFAULT;
 	void *value = NULL;
@@ -172,12 +173,13 @@ struct posix_acl *f2fs_get_acl(struct inode *inode, int type)
 	if (type == ACL_TYPE_ACCESS)
 		name_index = F2FS_XATTR_INDEX_POSIX_ACL_ACCESS;
 
-	retval = f2fs_getxattr(inode, name_index, "", NULL, 0);
+	retval = f2fs_getxattr(inode, name_index, "", NULL, 0, dpage);
 	if (retval > 0) {
 		value = kmalloc(retval, GFP_F2FS_ZERO);
 		if (!value)
 			return ERR_PTR(-ENOMEM);
-		retval = f2fs_getxattr(inode, name_index, "", value, retval);
+		retval = f2fs_getxattr(inode, name_index, "", value,
+							retval, dpage);
 	}
 
 	if (retval > 0)
@@ -194,6 +196,11 @@ struct posix_acl *f2fs_get_acl(struct inode *inode, int type)
 	return acl;
 }
 
+struct posix_acl *f2fs_get_acl(struct inode *inode, int type)
+{
+	return __f2fs_get_acl(inode, type, NULL);
+}
+
 static int __f2fs_set_acl(struct inode *inode, int type,
 			struct posix_acl *acl, struct page *ipage)
 {
@@ -229,7 +236,7 @@ static int __f2fs_set_acl(struct inode *inode, int type,
 	if (acl) {
 		value = f2fs_acl_to_disk(acl, &size);
 		if (IS_ERR(value)) {
-			cond_clear_inode_flag(fi, FI_ACL_MODE);
+			clear_inode_flag(fi, FI_ACL_MODE);
 			return (int)PTR_ERR(value);
 		}
 	}
@@ -240,7 +247,7 @@ static int __f2fs_set_acl(struct inode *inode, int type,
 	if (!error)
 		set_cached_acl(inode, type, acl);
 
-	cond_clear_inode_flag(fi, FI_ACL_MODE);
+	clear_inode_flag(fi, FI_ACL_MODE);
 	return error;
 }
 
@@ -249,12 +256,137 @@ int f2fs_set_acl(struct inode *inode, struct posix_acl *acl, int type)
 	return __f2fs_set_acl(inode, type, acl, NULL);
 }
 
-int f2fs_init_acl(struct inode *inode, struct inode *dir, struct page *ipage)
+/*
+ * Most part of f2fs_acl_clone, f2fs_acl_create_masq, f2fs_acl_create
+ * are copied from posix_acl.c
+ */
+static struct posix_acl *f2fs_acl_clone(const struct posix_acl *acl,
+							gfp_t flags)
 {
-	struct posix_acl *default_acl, *acl;
+	struct posix_acl *clone = NULL;
+
+	if (acl) {
+		int size = sizeof(struct posix_acl) + acl->a_count *
+				sizeof(struct posix_acl_entry);
+		clone = kmemdup(acl, size, flags);
+		if (clone)
+			atomic_set(&clone->a_refcount, 1);
+	}
+	return clone;
+}
+
+static int f2fs_acl_create_masq(struct posix_acl *acl, umode_t *mode_p)
+{
+	struct posix_acl_entry *pa, *pe;
+	struct posix_acl_entry *group_obj = NULL, *mask_obj = NULL;
+	umode_t mode = *mode_p;
+	int not_equiv = 0;
+
+	/* assert(atomic_read(acl->a_refcount) == 1); */
+
+	FOREACH_ACL_ENTRY(pa, acl, pe) {
+		switch(pa->e_tag) {
+		case ACL_USER_OBJ:
+			pa->e_perm &= (mode >> 6) | ~S_IRWXO;
+			mode &= (pa->e_perm << 6) | ~S_IRWXU;
+			break;
+
+		case ACL_USER:
+		case ACL_GROUP:
+			not_equiv = 1;
+			break;
+
+		case ACL_GROUP_OBJ:
+			group_obj = pa;
+			break;
+
+		case ACL_OTHER:
+			pa->e_perm &= mode | ~S_IRWXO;
+			mode &= pa->e_perm | ~S_IRWXO;
+			break;
+
+		case ACL_MASK:
+			mask_obj = pa;
+			not_equiv = 1;
+			break;
+
+		default:
+			return -EIO;
+		}
+	}
+
+	if (mask_obj) {
+		mask_obj->e_perm &= (mode >> 3) | ~S_IRWXO;
+		mode &= (mask_obj->e_perm << 3) | ~S_IRWXG;
+	} else {
+		if (!group_obj)
+			return -EIO;
+		group_obj->e_perm &= (mode >> 3) | ~S_IRWXO;
+		mode &= (group_obj->e_perm << 3) | ~S_IRWXG;
+	}
+
+	*mode_p = (*mode_p & ~S_IRWXUGO) | mode;
+        return not_equiv;
+}
+
+static int f2fs_acl_create(struct inode *dir, umode_t *mode,
+		struct posix_acl **default_acl, struct posix_acl **acl,
+		struct page *dpage)
+{
+	struct posix_acl *p;
+	int ret;
+
+	if (S_ISLNK(*mode) || !IS_POSIXACL(dir))
+		goto no_acl;
+
+	p = __f2fs_get_acl(dir, ACL_TYPE_DEFAULT, dpage);
+	if (IS_ERR(p)) {
+		if (p == ERR_PTR(-EOPNOTSUPP))
+			goto apply_umask;
+		return PTR_ERR(p);
+	}
+
+	if (!p)
+		goto apply_umask;
+
+	*acl = f2fs_acl_clone(p, GFP_NOFS);
+	if (!*acl)
+		return -ENOMEM;
+
+	ret = f2fs_acl_create_masq(*acl, mode);
+	if (ret < 0) {
+		posix_acl_release(*acl);
+		return -ENOMEM;
+	}
+
+	if (ret == 0) {
+		posix_acl_release(*acl);
+		*acl = NULL;
+	}
+
+	if (!S_ISDIR(*mode)) {
+		posix_acl_release(p);
+		*default_acl = NULL;
+	} else {
+		*default_acl = p;
+	}
+	return 0;
+
+apply_umask:
+	*mode &= ~current_umask();
+no_acl:
+	*default_acl = NULL;
+	*acl = NULL;
+	return 0;
+}
+
+int f2fs_init_acl(struct inode *inode, struct inode *dir, struct page *ipage,
+							struct page *dpage)
+{
+	struct posix_acl *default_acl = NULL, *acl = NULL;
 	int error = 0;
 
-	error = posix_acl_create(dir, &inode->i_mode, &default_acl, &acl);
+	error = f2fs_acl_create(dir, &inode->i_mode, &default_acl, &acl, dpage);
 	if (error)
 		return error;
 

fs/f2fs/acl.h

@@ -38,14 +38,15 @@ struct f2fs_acl_header {
 
 extern struct posix_acl *f2fs_get_acl(struct inode *, int);
 extern int f2fs_set_acl(struct inode *inode, struct posix_acl *acl, int type);
-extern int f2fs_init_acl(struct inode *, struct inode *, struct page *);
+extern int f2fs_init_acl(struct inode *, struct inode *, struct page *,
+							struct page *);
 #else
 #define f2fs_check_acl	NULL
 #define f2fs_get_acl	NULL
 #define f2fs_set_acl	NULL
 
 static inline int f2fs_init_acl(struct inode *inode, struct inode *dir,
-							struct page *page)
+				struct page *ipage, struct page *dpage)
 {
 	return 0;
 }

fs/f2fs/checkpoint.c

@@ -72,36 +72,36 @@ out:
 	return page;
 }
 
-struct page *get_meta_page_ra(struct f2fs_sb_info *sbi, pgoff_t index)
+static inline bool is_valid_blkaddr(struct f2fs_sb_info *sbi,
-{
+						block_t blkaddr, int type)
-	bool readahead = false;
-	struct page *page;
-
-	page = find_get_page(META_MAPPING(sbi), index);
-	if (!page || (page && !PageUptodate(page)))
-		readahead = true;
-	f2fs_put_page(page, 0);
-
-	if (readahead)
-		ra_meta_pages(sbi, index, MAX_BIO_BLOCKS(sbi), META_POR);
-	return get_meta_page(sbi, index);
-}
-
-static inline block_t 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;
+		break;
 	case META_SIT:
-		return SIT_BLK_CNT(sbi);
+		if (unlikely(blkaddr >= SIT_BLK_CNT(sbi)))
+			return false;
+		break;
 	case META_SSA:
+		if (unlikely(blkaddr >= MAIN_BLKADDR(sbi) ||
+			blkaddr < SM_I(sbi)->ssa_blkaddr))
+			return false;
+		break;
 	case META_CP:
-		return 0;
+		if (unlikely(blkaddr >= SIT_I(sbi)->sit_base_addr ||
+			blkaddr < __start_cp_addr(sbi)))
+			return false;
+		break;
 	case META_POR:
-		return MAX_BLKADDR(sbi);
+		if (unlikely(blkaddr >= MAX_BLKADDR(sbi) ||
+			blkaddr < MAIN_BLKADDR(sbi)))
+			return false;
+		break;
 	default:
 		BUG();
 	}
+
+	return true;
 }
 
 /*
@@ -112,7 +112,6 @@ int ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages, int type
 	block_t prev_blk_addr = 0;
 	struct page *page;
 	block_t blkno = start;
-	block_t max_blks = get_max_meta_blks(sbi, type);
 
 	struct f2fs_io_info fio = {
 		.type = META,
@@ -122,18 +121,20 @@ int ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages, int type
 	for (; nrpages-- > 0; blkno++) {
 		block_t blk_addr;
 
+		if (!is_valid_blkaddr(sbi, blkno, type))
+			goto out;
+
 		switch (type) {
 		case META_NAT:
-			/* get nat block addr */
+			if (unlikely(blkno >=
-			if (unlikely(blkno >= max_blks))
+					NAT_BLOCK_OFFSET(NM_I(sbi)->max_nid)))
 				blkno = 0;
+			/* get nat block addr */
 			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)
@@ -143,10 +144,6 @@ int ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages, int type
 		case META_SSA:
 		case META_CP:
 		case META_POR:
-			if (unlikely(blkno >= max_blks))
-				goto out;
-			if (unlikely(blkno < SEG0_BLKADDR(sbi)))
-				goto out;
 			blk_addr = blkno;
 			break;
 		default:
@@ -169,6 +166,20 @@ out:
 	return blkno - start;
 }
 
+void ra_meta_pages_cond(struct f2fs_sb_info *sbi, pgoff_t index)
+{
+	struct page *page;
+	bool readahead = false;
+
+	page = find_get_page(META_MAPPING(sbi), index);
+	if (!page || (page && !PageUptodate(page)))
+		readahead = true;
+	f2fs_put_page(page, 0);
+
+	if (readahead)
+		ra_meta_pages(sbi, index, MAX_BIO_BLOCKS(sbi), META_POR);
+}
+
 static int f2fs_write_meta_page(struct page *page,
 				struct writeback_control *wbc)
 {
@@ -178,7 +189,7 @@ static int f2fs_write_meta_page(struct page *page,
 
 	if (unlikely(sbi->por_doing))
 		goto redirty_out;
-	if (wbc->for_reclaim)
+	if (wbc->for_reclaim && page->index < GET_SUM_BLOCK(sbi, 0))
 		goto redirty_out;
 	if (unlikely(f2fs_cp_error(sbi)))
 		goto redirty_out;
@@ -187,6 +198,9 @@ static int f2fs_write_meta_page(struct page *page,
 	write_meta_page(sbi, page);
 	dec_page_count(sbi, F2FS_DIRTY_META);
 	unlock_page(page);
+
+	if (wbc->for_reclaim)
+		f2fs_submit_merged_bio(sbi, META, WRITE);
 	return 0;
 
 redirty_out:
@@ -298,46 +312,57 @@ const struct address_space_operations f2fs_meta_aops = {
 
 static void __add_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type)
 {
+	struct inode_management *im = &sbi->im[type];
 	struct ino_entry *e;
 retry:
-	spin_lock(&sbi->ino_lock[type]);
+	if (radix_tree_preload(GFP_NOFS)) {
+		cond_resched();
+		goto retry;
+	}
 
-	e = radix_tree_lookup(&sbi->ino_root[type], ino);
+	spin_lock(&im->ino_lock);
+
+	e = radix_tree_lookup(&im->ino_root, ino);
 	if (!e) {
 		e = kmem_cache_alloc(ino_entry_slab, GFP_ATOMIC);
 		if (!e) {
-			spin_unlock(&sbi->ino_lock[type]);
+			spin_unlock(&im->ino_lock);
+			radix_tree_preload_end();
 			goto retry;
 		}
-		if (radix_tree_insert(&sbi->ino_root[type], ino, e)) {
+		if (radix_tree_insert(&im->ino_root, ino, e)) {
-			spin_unlock(&sbi->ino_lock[type]);
+			spin_unlock(&im->ino_lock);
 			kmem_cache_free(ino_entry_slab, e);
+			radix_tree_preload_end();
 			goto retry;
 		}
 		memset(e, 0, sizeof(struct ino_entry));
 		e->ino = ino;
 
-		list_add_tail(&e->list, &sbi->ino_list[type]);
+		list_add_tail(&e->list, &im->ino_list);
+		if (type != ORPHAN_INO)
+			im->ino_num++;
 	}
-	spin_unlock(&sbi->ino_lock[type]);
+	spin_unlock(&im->ino_lock);
+	radix_tree_preload_end();
 }
 
 static void __remove_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type)
 {
+	struct inode_management *im = &sbi->im[type];
 	struct ino_entry *e;
 
-	spin_lock(&sbi->ino_lock[type]);
+	spin_lock(&im->ino_lock);
-	e = radix_tree_lookup(&sbi->ino_root[type], ino);
+	e = radix_tree_lookup(&im->ino_root, ino);
 	if (e) {
 		list_del(&e->list);
-		radix_tree_delete(&sbi->ino_root[type], ino);
+		radix_tree_delete(&im->ino_root, ino);
-		if (type == ORPHAN_INO)
+		im->ino_num--;
-			sbi->n_orphans--;
+		spin_unlock(&im->ino_lock);
-		spin_unlock(&sbi->ino_lock[type]);
 		kmem_cache_free(ino_entry_slab, e);
 		return;
 	}
-	spin_unlock(&sbi->ino_lock[type]);
+	spin_unlock(&im->ino_lock);
 }
 
 void add_dirty_inode(struct f2fs_sb_info *sbi, nid_t ino, int type)
@@ -355,10 +380,12 @@ void remove_dirty_inode(struct f2fs_sb_info *sbi, nid_t ino, int type)
 /* mode should be APPEND_INO or UPDATE_INO */
 bool exist_written_data(struct f2fs_sb_info *sbi, nid_t ino, int mode)
 {
+	struct inode_management *im = &sbi->im[mode];
 	struct ino_entry *e;
-	spin_lock(&sbi->ino_lock[mode]);
+
-	e = radix_tree_lookup(&sbi->ino_root[mode], ino);
+	spin_lock(&im->ino_lock);
-	spin_unlock(&sbi->ino_lock[mode]);
+	e = radix_tree_lookup(&im->ino_root, ino);
+	spin_unlock(&im->ino_lock);
 	return e ? true : false;
 }
 
@@ -368,36 +395,42 @@ void release_dirty_inode(struct f2fs_sb_info *sbi)
 	int i;
 
 	for (i = APPEND_INO; i <= UPDATE_INO; i++) {
-		spin_lock(&sbi->ino_lock[i]);
+		struct inode_management *im = &sbi->im[i];
-		list_for_each_entry_safe(e, tmp, &sbi->ino_list[i], list) {
+
+		spin_lock(&im->ino_lock);
+		list_for_each_entry_safe(e, tmp, &im->ino_list, list) {
 			list_del(&e->list);
-			radix_tree_delete(&sbi->ino_root[i], e->ino);
+			radix_tree_delete(&im->ino_root, e->ino);
 			kmem_cache_free(ino_entry_slab, e);
+			im->ino_num--;
 		}
-		spin_unlock(&sbi->ino_lock[i]);
+		spin_unlock(&im->ino_lock);
 	}
 }
 
 int acquire_orphan_inode(struct f2fs_sb_info *sbi)
 {
+	struct inode_management *im = &sbi->im[ORPHAN_INO];
 	int err = 0;
 
-	spin_lock(&sbi->ino_lock[ORPHAN_INO]);
+	spin_lock(&im->ino_lock);
-	if (unlikely(sbi->n_orphans >= sbi->max_orphans))
+	if (unlikely(im->ino_num >= sbi->max_orphans))
 		err = -ENOSPC;
 	else
-		sbi->n_orphans++;
+		im->ino_num++;
-	spin_unlock(&sbi->ino_lock[ORPHAN_INO]);
+	spin_unlock(&im->ino_lock);
 
 	return err;
 }
 
 void release_orphan_inode(struct f2fs_sb_info *sbi)
 {
-	spin_lock(&sbi->ino_lock[ORPHAN_INO]);
+	struct inode_management *im = &sbi->im[ORPHAN_INO];
-	f2fs_bug_on(sbi, sbi->n_orphans == 0);
+
-	sbi->n_orphans--;
+	spin_lock(&im->ino_lock);
-	spin_unlock(&sbi->ino_lock[ORPHAN_INO]);
+	f2fs_bug_on(sbi, im->ino_num == 0);
+	im->ino_num--;
+	spin_unlock(&im->ino_lock);
 }
 
 void add_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
@@ -460,17 +493,19 @@ static void write_orphan_inodes(struct f2fs_sb_info *sbi, block_t start_blk)
 	struct f2fs_orphan_block *orphan_blk = NULL;
 	unsigned int nentries = 0;
 	unsigned short index;
-	unsigned short orphan_blocks =
+	unsigned short orphan_blocks;
-			(unsigned short)GET_ORPHAN_BLOCKS(sbi->n_orphans);
 	struct page *page = NULL;
 	struct ino_entry *orphan = NULL;
+	struct inode_management *im = &sbi->im[ORPHAN_INO];
+
+	orphan_blocks = GET_ORPHAN_BLOCKS(im->ino_num);
 
 	for (index = 0; index < orphan_blocks; index++)
 		grab_meta_page(sbi, start_blk + index);
 
 	index = 1;
-	spin_lock(&sbi->ino_lock[ORPHAN_INO]);
+	spin_lock(&im->ino_lock);
-	head = &sbi->ino_list[ORPHAN_INO];
+	head = &im->ino_list;
 
 	/* loop for each orphan inode entry and write them in Jornal block */
 	list_for_each_entry(orphan, head, list) {
@@ -510,7 +545,7 @@ static void write_orphan_inodes(struct f2fs_sb_info *sbi, block_t start_blk)
 		f2fs_put_page(page, 1);
 	}
 
-	spin_unlock(&sbi->ino_lock[ORPHAN_INO]);
+	spin_unlock(&im->ino_lock);
 }
 
 static struct page *validate_checkpoint(struct f2fs_sb_info *sbi,
@@ -731,6 +766,9 @@ void sync_dirty_dir_inodes(struct f2fs_sb_info *sbi)
 	struct dir_inode_entry *entry;
 	struct inode *inode;
 retry:
+	if (unlikely(f2fs_cp_error(sbi)))
+		return;
+
 	spin_lock(&sbi->dir_inode_lock);
 
 	head = &sbi->dir_inode_list;
@@ -830,6 +868,7 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 	struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
 	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
+	unsigned long orphan_num = sbi->im[ORPHAN_INO].ino_num;
 	nid_t last_nid = nm_i->next_scan_nid;
 	block_t start_blk;
 	struct page *cp_page;
@@ -889,7 +928,7 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 	else
 		clear_ckpt_flags(ckpt, CP_COMPACT_SUM_FLAG);
 
-	orphan_blocks = GET_ORPHAN_BLOCKS(sbi->n_orphans);
+	orphan_blocks = GET_ORPHAN_BLOCKS(orphan_num);
 	ckpt->cp_pack_start_sum = cpu_to_le32(1 + cp_payload_blks +
 			orphan_blocks);
 
@@ -905,7 +944,7 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 				orphan_blocks);
 	}
 
-	if (sbi->n_orphans)
+	if (orphan_num)
 		set_ckpt_flags(ckpt, CP_ORPHAN_PRESENT_FLAG);
 	else
 		clear_ckpt_flags(ckpt, CP_ORPHAN_PRESENT_FLAG);
@@ -940,7 +979,7 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 		f2fs_put_page(cp_page, 1);
 	}
 
-	if (sbi->n_orphans) {
+	if (orphan_num) {
 		write_orphan_inodes(sbi, start_blk);
 		start_blk += orphan_blocks;
 	}
@@ -975,6 +1014,9 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 	/* Here, we only have one bio having CP pack */
 	sync_meta_pages(sbi, META_FLUSH, LONG_MAX);
 
+	/* wait for previous submitted meta pages writeback */
+	wait_on_all_pages_writeback(sbi);
+
 	release_dirty_inode(sbi);
 
 	if (unlikely(f2fs_cp_error(sbi)))
@@ -1036,9 +1078,12 @@ void init_ino_entry_info(struct f2fs_sb_info *sbi)
 	int i;
 
 	for (i = 0; i < MAX_INO_ENTRY; i++) {
-		INIT_RADIX_TREE(&sbi->ino_root[i], GFP_ATOMIC);
+		struct inode_management *im = &sbi->im[i];
-		spin_lock_init(&sbi->ino_lock[i]);
+
-		INIT_LIST_HEAD(&sbi->ino_list[i]);
+		INIT_RADIX_TREE(&im->ino_root, GFP_ATOMIC);
+		spin_lock_init(&im->ino_lock);
+		INIT_LIST_HEAD(&im->ino_list);
+		im->ino_num = 0;
 	}
 
 	/*
@@ -1047,7 +1092,6 @@ void init_ino_entry_info(struct f2fs_sb_info *sbi)
 	 * orphan entries with the limitation one reserved segment
 	 * for cp pack we can have max 1020*504 orphan entries
 	 */
-	sbi->n_orphans = 0;
 	sbi->max_orphans = (sbi->blocks_per_seg - F2FS_CP_PACKS -
 			NR_CURSEG_TYPE) * F2FS_ORPHANS_PER_BLOCK;
 }

fs/f2fs/data.c

@@ -61,11 +61,6 @@ static void f2fs_write_end_io(struct bio *bio, int err)
 		dec_page_count(sbi, F2FS_WRITEBACK);
 	}
 
-	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))
 		wake_up(&sbi->cp_wait);
@@ -95,34 +90,18 @@ static struct bio *__bio_alloc(struct f2fs_sb_info *sbi, block_t blk_addr,
 static void __submit_merged_bio(struct f2fs_bio_info *io)
 {
 	struct f2fs_io_info *fio = &io->fio;
-	int rw;
 
 	if (!io->bio)
 		return;
 
-	rw = fio->rw;
+	if (is_read_io(fio->rw))
-
+		trace_f2fs_submit_read_bio(io->sbi->sb, fio->rw,
-	if (is_read_io(rw)) {
+							fio->type, io->bio);
-		trace_f2fs_submit_read_bio(io->sbi->sb, rw,
+	else
-						fio->type, io->bio);
+		trace_f2fs_submit_write_bio(io->sbi->sb, fio->rw,
-		submit_bio(rw, io->bio);
+							fio->type, io->bio);
-	} else {
-		trace_f2fs_submit_write_bio(io->sbi->sb, rw,
-						fio->type, io->bio);
-		/*
-		 * META_FLUSH is only from the checkpoint procedure, and we
-		 * should wait this metadata bio for FS consistency.
-		 */
-		if (fio->type == META_FLUSH) {
-			DECLARE_COMPLETION_ONSTACK(wait);
-			io->sbi->wait_io = &wait;
-			submit_bio(rw, io->bio);
-			wait_for_completion(&wait);
-		} else {
-			submit_bio(rw, io->bio);
-		}
-	}
 
+	submit_bio(fio->rw, io->bio);
 	io->bio = NULL;
 }
 
@@ -257,9 +236,6 @@ int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index)
 	bool need_put = dn->inode_page ? false : true;
 	int err;
 
-	/* if inode_page exists, index should be zero */
-	f2fs_bug_on(F2FS_I_SB(dn->inode), !need_put && index);
-
 	err = get_dnode_of_data(dn, index, ALLOC_NODE);
 	if (err)
 		return err;
@@ -740,14 +716,14 @@ int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 static int f2fs_read_data_page(struct file *file, struct page *page)
 {
 	struct inode *inode = page->mapping->host;
-	int ret;
+	int ret = -EAGAIN;
 
 	trace_f2fs_readpage(page, DATA);
 
 	/* If the file has inline data, try to read it directly */
 	if (f2fs_has_inline_data(inode))
 		ret = f2fs_read_inline_data(inode, page);
-	else
+	if (ret == -EAGAIN)
 		ret = mpage_readpage(page, get_data_block);
 
 	return ret;
@@ -859,10 +835,11 @@ write:
 	else if (has_not_enough_free_secs(sbi, 0))
 		goto redirty_out;
 
+	err = -EAGAIN;
 	f2fs_lock_op(sbi);
-	if (f2fs_has_inline_data(inode) || f2fs_may_inline(inode))
+	if (f2fs_has_inline_data(inode))
-		err = f2fs_write_inline_data(inode, page, offset);
+		err = f2fs_write_inline_data(inode, page);
-	else
+	if (err == -EAGAIN)
 		err = do_write_data_page(page, &fio);
 	f2fs_unlock_op(sbi);
 done:
@@ -951,7 +928,7 @@ static int f2fs_write_begin(struct file *file, struct address_space *mapping,
 {
 	struct inode *inode = mapping->host;
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
-	struct page *page;
+	struct page *page, *ipage;
 	pgoff_t index = ((unsigned long long) pos) >> PAGE_CACHE_SHIFT;
 	struct dnode_of_data dn;
 	int err = 0;
@@ -959,45 +936,60 @@ static int f2fs_write_begin(struct file *file, struct address_space *mapping,
 	trace_f2fs_write_begin(inode, pos, len, flags);
 
 	f2fs_balance_fs(sbi);
-repeat:
-	err = f2fs_convert_inline_data(inode, pos + len, NULL);
-	if (err)
-		goto fail;
 
+	/*
+	 * We should check this at this moment to avoid deadlock on inode page
+	 * and #0 page. The locking rule for inline_data conversion should be:
+	 * lock_page(page #0) -> lock_page(inode_page)
+	 */
+	if (index != 0) {
+		err = f2fs_convert_inline_inode(inode);
+		if (err)
+			goto fail;
+	}
+repeat:
 	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page) {
 		err = -ENOMEM;
 		goto fail;
 	}
 
-	/* to avoid latency during memory pressure */
-	unlock_page(page);
-
 	*pagep = page;
 
-	if (f2fs_has_inline_data(inode) && (pos + len) <= MAX_INLINE_DATA)
-		goto inline_data;
-
 	f2fs_lock_op(sbi);
-	set_new_dnode(&dn, inode, NULL, NULL, 0);
+
-	err = f2fs_reserve_block(&dn, index);
+	/* check inline_data */
-	f2fs_unlock_op(sbi);
+	ipage = get_node_page(sbi, inode->i_ino);
-	if (err) {
+	if (IS_ERR(ipage)) {
-		f2fs_put_page(page, 0);
+		err = PTR_ERR(ipage);
-		goto fail;
+		goto unlock_fail;
-	}
-inline_data:
-	lock_page(page);
-	if (unlikely(page->mapping != mapping)) {
-		f2fs_put_page(page, 1);
-		goto repeat;
 	}
 
-	f2fs_wait_on_page_writeback(page, DATA);
+	set_new_dnode(&dn, inode, ipage, ipage, 0);
+
+	if (f2fs_has_inline_data(inode)) {
+		if (pos + len <= MAX_INLINE_DATA) {
+			read_inline_data(page, ipage);
+			set_inode_flag(F2FS_I(inode), FI_DATA_EXIST);
+			sync_inode_page(&dn);
+			goto put_next;
+		}
+		err = f2fs_convert_inline_page(&dn, page);
+		if (err)
+			goto put_fail;
+	}
+	err = f2fs_reserve_block(&dn, index);
+	if (err)
+		goto put_fail;
+put_next:
+	f2fs_put_dnode(&dn);
+	f2fs_unlock_op(sbi);
 
 	if ((len == PAGE_CACHE_SIZE) || PageUptodate(page))
 		return 0;
 
+	f2fs_wait_on_page_writeback(page, DATA);
+
 	if ((pos & PAGE_CACHE_MASK) >= i_size_read(inode)) {
 		unsigned start = pos & (PAGE_CACHE_SIZE - 1);
 		unsigned end = start + len;
@@ -1010,18 +1002,10 @@ inline_data:
 	if (dn.data_blkaddr == NEW_ADDR) {
 		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
 	} else {
-		if (f2fs_has_inline_data(inode)) {
+		err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
-			err = f2fs_read_inline_data(inode, page);
+					   READ_SYNC);
-			if (err) {
+		if (err)
-				page_cache_release(page);
+			goto fail;
-				goto fail;
-			}
-		} else {
-			err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
-							READ_SYNC);
-			if (err)
-				goto fail;
-		}
 
 		lock_page(page);
 		if (unlikely(!PageUptodate(page))) {
@@ -1038,6 +1022,12 @@ out:
 	SetPageUptodate(page);
 	clear_cold_data(page);
 	return 0;
+
+put_fail:
+	f2fs_put_dnode(&dn);
+unlock_fail:
+	f2fs_unlock_op(sbi);
+	f2fs_put_page(page, 1);
 fail:
 	f2fs_write_failed(mapping, pos + len);
 	return err;
@@ -1052,10 +1042,7 @@ static int f2fs_write_end(struct file *file,
 
 	trace_f2fs_write_end(inode, pos, len, copied);
 
-	if (f2fs_is_atomic_file(inode) || f2fs_is_volatile_file(inode))
+	set_page_dirty(page);
-		register_inmem_page(inode, page);
-	else
-		set_page_dirty(page);
 
 	if (pos + copied > i_size_read(inode)) {
 		i_size_write(inode, pos + copied);
@@ -1093,9 +1080,12 @@ static ssize_t f2fs_direct_IO(int rw, struct kiocb *iocb,
 	size_t count = iov_iter_count(iter);
 	int err;
 
-	/* Let buffer I/O handle the inline data case. */
+	/* we don't need to use inline_data strictly */
-	if (f2fs_has_inline_data(inode))
+	if (f2fs_has_inline_data(inode)) {
-		return 0;
+		err = f2fs_convert_inline_inode(inode);
+		if (err)
+			return err;
+	}
 
 	if (check_direct_IO(inode, rw, iter, offset))
 		return 0;
@@ -1119,6 +1109,9 @@ static void f2fs_invalidate_data_page(struct page *page, unsigned int offset,
 	if (offset % PAGE_CACHE_SIZE || length != PAGE_CACHE_SIZE)
 		return;
 
+	if (f2fs_is_atomic_file(inode) || f2fs_is_volatile_file(inode))
+		invalidate_inmem_page(inode, page);
+
 	if (PageDirty(page))
 		inode_dec_dirty_pages(inode);
 	ClearPagePrivate(page);
@@ -1138,6 +1131,12 @@ static int f2fs_set_data_page_dirty(struct page *page)
 	trace_f2fs_set_page_dirty(page, DATA);
 
 	SetPageUptodate(page);
+
+	if (f2fs_is_atomic_file(inode) || f2fs_is_volatile_file(inode)) {
+		register_inmem_page(inode, page);
+		return 1;
+	}
+
 	mark_inode_dirty(inode);
 
 	if (!PageDirty(page)) {
@@ -1152,9 +1151,12 @@ static sector_t f2fs_bmap(struct address_space *mapping, sector_t block)
 {
 	struct inode *inode = mapping->host;
 
-	if (f2fs_has_inline_data(inode))
+	/* we don't need to use inline_data strictly */
-		return 0;
+	if (f2fs_has_inline_data(inode)) {
-
+		int err = f2fs_convert_inline_inode(inode);
+		if (err)
+			return err;
+	}
 	return generic_block_bmap(mapping, block, get_data_block);
 }
 

fs/f2fs/debug.c

@@ -39,13 +39,15 @@ static void update_general_status(struct f2fs_sb_info *sbi)
 	si->ndirty_dent = get_pages(sbi, F2FS_DIRTY_DENTS);
 	si->ndirty_dirs = sbi->n_dirty_dirs;
 	si->ndirty_meta = get_pages(sbi, F2FS_DIRTY_META);
+	si->inmem_pages = get_pages(sbi, F2FS_INMEM_PAGES);
 	si->total_count = (int)sbi->user_block_count / sbi->blocks_per_seg;
 	si->rsvd_segs = reserved_segments(sbi);
 	si->overp_segs = overprovision_segments(sbi);
 	si->valid_count = valid_user_blocks(sbi);
 	si->valid_node_count = valid_node_count(sbi);
 	si->valid_inode_count = valid_inode_count(sbi);
-	si->inline_inode = sbi->inline_inode;
+	si->inline_inode = atomic_read(&sbi->inline_inode);
+	si->inline_dir = atomic_read(&sbi->inline_dir);
 	si->utilization = utilization(sbi);
 
 	si->free_segs = free_segments(sbi);
@@ -118,6 +120,7 @@ static void update_mem_info(struct f2fs_sb_info *sbi)
 {
 	struct f2fs_stat_info *si = F2FS_STAT(sbi);
 	unsigned npages;
+	int i;
 
 	if (si->base_mem)
 		goto get_cache;
@@ -167,8 +170,9 @@ get_cache:
 	si->cache_mem += npages << PAGE_CACHE_SHIFT;
 	npages = META_MAPPING(sbi)->nrpages;
 	si->cache_mem += npages << PAGE_CACHE_SHIFT;
-	si->cache_mem += sbi->n_orphans * sizeof(struct ino_entry);
 	si->cache_mem += sbi->n_dirty_dirs * sizeof(struct dir_inode_entry);
+	for (i = 0; i <= UPDATE_INO; i++)
+		si->cache_mem += sbi->im[i].ino_num * sizeof(struct ino_entry);
 }
 
 static int stat_show(struct seq_file *s, void *v)
@@ -200,6 +204,8 @@ static int stat_show(struct seq_file *s, void *v)
 			   si->valid_count - si->valid_node_count);
 		seq_printf(s, "  - Inline_data Inode: %u\n",
 			   si->inline_inode);
+		seq_printf(s, "  - Inline_dentry Inode: %u\n",
+			   si->inline_dir);
 		seq_printf(s, "\nMain area: %d segs, %d secs %d zones\n",
 			   si->main_area_segs, si->main_area_sections,
 			   si->main_area_zones);
@@ -244,6 +250,8 @@ static int stat_show(struct seq_file *s, void *v)
 		seq_printf(s, "\nExtent Hit Ratio: %d / %d\n",
 			   si->hit_ext, si->total_ext);
 		seq_puts(s, "\nBalancing F2FS Async:\n");
+		seq_printf(s, "  - inmem: %4d\n",
+			   si->inmem_pages);
 		seq_printf(s, "  - nodes: %4d in %4d\n",
 			   si->ndirty_node, si->node_pages);
 		seq_printf(s, "  - dents: %4d in dirs:%4d\n",
@@ -321,6 +329,9 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
 	si->sbi = sbi;
 	sbi->stat_info = si;
 
+	atomic_set(&sbi->inline_inode, 0);
+	atomic_set(&sbi->inline_dir, 0);
+
 	mutex_lock(&f2fs_stat_mutex);
 	list_add_tail(&si->stat_list, &f2fs_stat_list);
 	mutex_unlock(&f2fs_stat_mutex);

fs/f2fs/dir.c

@@ -37,7 +37,7 @@ static unsigned int bucket_blocks(unsigned int level)
 		return 4;
 }
 
-static unsigned char f2fs_filetype_table[F2FS_FT_MAX] = {
+unsigned char f2fs_filetype_table[F2FS_FT_MAX] = {
 	[F2FS_FT_UNKNOWN]	= DT_UNKNOWN,
 	[F2FS_FT_REG_FILE]	= DT_REG,
 	[F2FS_FT_DIR]		= DT_DIR,
@@ -59,7 +59,7 @@ static unsigned char f2fs_type_by_mode[S_IFMT >> S_SHIFT] = {
 	[S_IFLNK >> S_SHIFT]	= F2FS_FT_SYMLINK,
 };
 
-static void set_de_type(struct f2fs_dir_entry *de, struct inode *inode)
+void set_de_type(struct f2fs_dir_entry *de, struct inode *inode)
 {
 	umode_t mode = inode->i_mode;
 	de->file_type = f2fs_type_by_mode[(mode & S_IFMT) >> S_SHIFT];
@@ -90,51 +90,70 @@ static bool early_match_name(size_t namelen, f2fs_hash_t namehash,
 }
 
 static struct f2fs_dir_entry *find_in_block(struct page *dentry_page,
-			struct qstr *name, int *max_slots,
+				struct qstr *name, int *max_slots,
-			f2fs_hash_t namehash, struct page **res_page)
+				struct page **res_page)
+{
+	struct f2fs_dentry_block *dentry_blk;
+	struct f2fs_dir_entry *de;
+	struct f2fs_dentry_ptr d;
+
+	dentry_blk = (struct f2fs_dentry_block *)kmap(dentry_page);
+
+	make_dentry_ptr(&d, (void *)dentry_blk, 1);
+	de = find_target_dentry(name, max_slots, &d);
+
+	if (de)
+		*res_page = dentry_page;
+	else
+		kunmap(dentry_page);
+
+	/*
+	 * For the most part, it should be a bug when name_len is zero.
+	 * We stop here for figuring out where the bugs has occurred.
+	 */
+	f2fs_bug_on(F2FS_P_SB(dentry_page), d.max < 0);
+	return de;
+}
+
+struct f2fs_dir_entry *find_target_dentry(struct qstr *name, int *max_slots,
+						struct f2fs_dentry_ptr *d)
 {
 	struct f2fs_dir_entry *de;
 	unsigned long bit_pos = 0;
-	struct f2fs_dentry_block *dentry_blk = kmap(dentry_page);
+	f2fs_hash_t namehash = f2fs_dentry_hash(name);
-	const void *dentry_bits = &dentry_blk->dentry_bitmap;
 	int max_len = 0;
 
-	while (bit_pos < NR_DENTRY_IN_BLOCK) {
+	if (max_slots)
-		if (!test_bit_le(bit_pos, dentry_bits)) {
+		*max_slots = 0;
+	while (bit_pos < d->max) {
+		if (!test_bit_le(bit_pos, d->bitmap)) {
 			if (bit_pos == 0)
 				max_len = 1;
-			else if (!test_bit_le(bit_pos - 1, dentry_bits))
+			else if (!test_bit_le(bit_pos - 1, d->bitmap))
 				max_len++;
 			bit_pos++;
 			continue;
 		}
-		de = &dentry_blk->dentry[bit_pos];
+		de = &d->dentry[bit_pos];
-		if (early_match_name(name->len, namehash, de)) {
+		if (early_match_name(name->len, namehash, de) &&
-			if (!memcmp(dentry_blk->filename[bit_pos],
+			!memcmp(d->filename[bit_pos], name->name, name->len))
-							name->name,
+			goto found;
-							name->len)) {
+
-				*res_page = dentry_page;
+		if (max_slots && *max_slots >= 0 && max_len > *max_slots) {
-				goto found;
-			}
-		}
-		if (max_len > *max_slots) {
 			*max_slots = max_len;
 			max_len = 0;
 		}
 
-		/*
+		/* remain bug on condition */
-		 * For the most part, it should be a bug when name_len is zero.
+		if (unlikely(!de->name_len))
-		 * We stop here for figuring out where the bugs has occurred.
+			d->max = -1;
-		 */
-		f2fs_bug_on(F2FS_P_SB(dentry_page), !de->name_len);
 
 		bit_pos += GET_DENTRY_SLOTS(le16_to_cpu(de->name_len));
 	}
 
 	de = NULL;
-	kunmap(dentry_page);
 found:
-	if (max_len > *max_slots)
+	if (max_slots && max_len > *max_slots)
 		*max_slots = max_len;
 	return de;
 }
@@ -149,7 +168,7 @@ static struct f2fs_dir_entry *find_in_level(struct inode *dir,
 	struct page *dentry_page;
 	struct f2fs_dir_entry *de = NULL;
 	bool room = false;
-	int max_slots = 0;
+	int max_slots;
 
 	f2fs_bug_on(F2FS_I_SB(dir), level > MAX_DIR_HASH_DEPTH);
 
@@ -168,8 +187,7 @@ static struct f2fs_dir_entry *find_in_level(struct inode *dir,
 			continue;
 		}
 
-		de = find_in_block(dentry_page, name, &max_slots,
+		de = find_in_block(dentry_page, name, &max_slots, res_page);
-					namehash, res_page);
 		if (de)
 			break;
 
@@ -201,6 +219,9 @@ struct f2fs_dir_entry *f2fs_find_entry(struct inode *dir,
 	unsigned int max_depth;
 	unsigned int level;
 
+	if (f2fs_has_inline_dentry(dir))
+		return find_in_inline_dir(dir, child, res_page);
+
 	if (npages == 0)
 		return NULL;
 
@@ -227,6 +248,9 @@ struct f2fs_dir_entry *f2fs_parent_dir(struct inode *dir, struct page **p)
 	struct f2fs_dir_entry *de;
 	struct f2fs_dentry_block *dentry_blk;
 
+	if (f2fs_has_inline_dentry(dir))
+		return f2fs_parent_inline_dir(dir, p);
+
 	page = get_lock_data_page(dir, 0);
 	if (IS_ERR(page))
 		return NULL;
@@ -247,7 +271,7 @@ ino_t f2fs_inode_by_name(struct inode *dir, struct qstr *qstr)
 	de = f2fs_find_entry(dir, qstr, &page);
 	if (de) {
 		res = le32_to_cpu(de->ino);
-		kunmap(page);
+		f2fs_dentry_kunmap(dir, page);
 		f2fs_put_page(page, 0);
 	}
 
@@ -257,11 +281,13 @@ ino_t f2fs_inode_by_name(struct inode *dir, struct qstr *qstr)
 void f2fs_set_link(struct inode *dir, struct f2fs_dir_entry *de,
 		struct page *page, struct inode *inode)
 {
+	enum page_type type = f2fs_has_inline_dentry(dir) ? NODE : DATA;
 	lock_page(page);
-	f2fs_wait_on_page_writeback(page, DATA);
+	f2fs_wait_on_page_writeback(page, type);
 	de->ino = cpu_to_le32(inode->i_ino);
 	set_de_type(de, inode);
-	kunmap(page);
+	if (!f2fs_has_inline_dentry(dir))
+		kunmap(page);
 	set_page_dirty(page);
 	dir->i_mtime = dir->i_ctime = CURRENT_TIME;
 	mark_inode_dirty(dir);
@@ -296,36 +322,48 @@ int update_dent_inode(struct inode *inode, const struct qstr *name)
 	return 0;
 }
 
+void do_make_empty_dir(struct inode *inode, struct inode *parent,
+					struct f2fs_dentry_ptr *d)
+{
+	struct f2fs_dir_entry *de;
+
+	de = &d->dentry[0];
+	de->name_len = cpu_to_le16(1);
+	de->hash_code = 0;
+	de->ino = cpu_to_le32(inode->i_ino);
+	memcpy(d->filename[0], ".", 1);
+	set_de_type(de, inode);
+
+	de = &d->dentry[1];
+	de->hash_code = 0;
+	de->name_len = cpu_to_le16(2);
+	de->ino = cpu_to_le32(parent->i_ino);
+	memcpy(d->filename[1], "..", 2);
+	set_de_type(de, inode);
+
+	test_and_set_bit_le(0, (void *)d->bitmap);
+	test_and_set_bit_le(1, (void *)d->bitmap);
+}
+
 static int make_empty_dir(struct inode *inode,
 		struct inode *parent, struct page *page)
 {
 	struct page *dentry_page;
 	struct f2fs_dentry_block *dentry_blk;
-	struct f2fs_dir_entry *de;
+	struct f2fs_dentry_ptr d;
+
+	if (f2fs_has_inline_dentry(inode))
+		return make_empty_inline_dir(inode, parent, page);
 
 	dentry_page = get_new_data_page(inode, page, 0, true);
 	if (IS_ERR(dentry_page))
 		return PTR_ERR(dentry_page);
 
-
 	dentry_blk = kmap_atomic(dentry_page);
 
-	de = &dentry_blk->dentry[0];
+	make_dentry_ptr(&d, (void *)dentry_blk, 1);
-	de->name_len = cpu_to_le16(1);
+	do_make_empty_dir(inode, parent, &d);
-	de->hash_code = 0;
-	de->ino = cpu_to_le32(inode->i_ino);
-	memcpy(dentry_blk->filename[0], ".", 1);
-	set_de_type(de, inode);
 
-	de = &dentry_blk->dentry[1];
-	de->hash_code = 0;
-	de->name_len = cpu_to_le16(2);
-	de->ino = cpu_to_le32(parent->i_ino);
-	memcpy(dentry_blk->filename[1], "..", 2);
-	set_de_type(de, inode);
-
-	test_and_set_bit_le(0, &dentry_blk->dentry_bitmap);
-	test_and_set_bit_le(1, &dentry_blk->dentry_bitmap);
 	kunmap_atomic(dentry_blk);
 
 	set_page_dirty(dentry_page);
@@ -333,8 +371,8 @@ static int make_empty_dir(struct inode *inode,
 	return 0;
 }
 
-static struct page *init_inode_metadata(struct inode *inode,
+struct page *init_inode_metadata(struct inode *inode, struct inode *dir,
-		struct inode *dir, const struct qstr *name)
+			const struct qstr *name, struct page *dpage)
 {
 	struct page *page;
 	int err;
@@ -350,7 +388,7 @@ static struct page *init_inode_metadata(struct inode *inode,
 				goto error;
 		}
 
-		err = f2fs_init_acl(inode, dir, page);
+		err = f2fs_init_acl(inode, dir, page, dpage);
 		if (err)
 			goto put_error;
 
@@ -395,7 +433,7 @@ error:
 	return ERR_PTR(err);
 }
 
-static void update_parent_metadata(struct inode *dir, struct inode *inode,
+void update_parent_metadata(struct inode *dir, struct inode *inode,
 						unsigned int current_depth)
 {
 	if (is_inode_flag_set(F2FS_I(inode), FI_NEW_INODE)) {
@@ -417,27 +455,23 @@ static void update_parent_metadata(struct inode *dir, struct inode *inode,
 		clear_inode_flag(F2FS_I(inode), FI_INC_LINK);
 }
 
-static int room_for_filename(struct f2fs_dentry_block *dentry_blk, int slots)
+int room_for_filename(const void *bitmap, int slots, int max_slots)
 {
 	int bit_start = 0;
 	int zero_start, zero_end;
 next:
-	zero_start = find_next_zero_bit_le(&dentry_blk->dentry_bitmap,
+	zero_start = find_next_zero_bit_le(bitmap, max_slots, bit_start);
-						NR_DENTRY_IN_BLOCK,
+	if (zero_start >= max_slots)
-						bit_start);
+		return max_slots;
-	if (zero_start >= NR_DENTRY_IN_BLOCK)
-		return NR_DENTRY_IN_BLOCK;
 
-	zero_end = find_next_bit_le(&dentry_blk->dentry_bitmap,
+	zero_end = find_next_bit_le(bitmap, max_slots, zero_start);
-						NR_DENTRY_IN_BLOCK,
-						zero_start);
 	if (zero_end - zero_start >= slots)
 		return zero_start;
 
 	bit_start = zero_end + 1;
 
-	if (zero_end + 1 >= NR_DENTRY_IN_BLOCK)
+	if (zero_end + 1 >= max_slots)
-		return NR_DENTRY_IN_BLOCK;
+		return max_slots;
 	goto next;
 }
 
@@ -463,6 +497,14 @@ int __f2fs_add_link(struct inode *dir, const struct qstr *name,
 	int err = 0;
 	int i;
 
+	if (f2fs_has_inline_dentry(dir)) {
+		err = f2fs_add_inline_entry(dir, name, inode);
+		if (!err || err != -EAGAIN)
+			return err;
+		else
+			err = 0;
+	}
+
 	dentry_hash = f2fs_dentry_hash(name);
 	level = 0;
 	current_depth = F2FS_I(dir)->i_current_depth;
@@ -491,7 +533,8 @@ start:
 			return PTR_ERR(dentry_page);
 
 		dentry_blk = kmap(dentry_page);
-		bit_pos = room_for_filename(dentry_blk, slots);
+		bit_pos = room_for_filename(&dentry_blk->dentry_bitmap,
+						slots, NR_DENTRY_IN_BLOCK);
 		if (bit_pos < NR_DENTRY_IN_BLOCK)
 			goto add_dentry;
 
@@ -506,7 +549,7 @@ add_dentry:
 	f2fs_wait_on_page_writeback(dentry_page, DATA);
 
 	down_write(&F2FS_I(inode)->i_sem);
-	page = init_inode_metadata(inode, dir, name);
+	page = init_inode_metadata(inode, dir, name, NULL);
 	if (IS_ERR(page)) {
 		err = PTR_ERR(page);
 		goto fail;
@@ -545,7 +588,7 @@ int f2fs_do_tmpfile(struct inode *inode, struct inode *dir)
 	int err = 0;
 
 	down_write(&F2FS_I(inode)->i_sem);
-	page = init_inode_metadata(inode, dir, NULL);
+	page = init_inode_metadata(inode, dir, NULL, NULL);
 	if (IS_ERR(page)) {
 		err = PTR_ERR(page);
 		goto fail;
@@ -560,26 +603,57 @@ fail:
 	return err;
 }
 
+void f2fs_drop_nlink(struct inode *dir, struct inode *inode, struct page *page)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
+
+	down_write(&F2FS_I(inode)->i_sem);
+
+	if (S_ISDIR(inode->i_mode)) {
+		drop_nlink(dir);
+		if (page)
+			update_inode(dir, page);
+		else
+			update_inode_page(dir);
+	}
+	inode->i_ctime = CURRENT_TIME;
+
+	drop_nlink(inode);
+	if (S_ISDIR(inode->i_mode)) {
+		drop_nlink(inode);
+		i_size_write(inode, 0);
+	}
+	up_write(&F2FS_I(inode)->i_sem);
+	update_inode_page(inode);
+
+	if (inode->i_nlink == 0)
+		add_orphan_inode(sbi, inode->i_ino);
+	else
+		release_orphan_inode(sbi);
+}
+
 /*
  * It only removes the dentry from the dentry page, corresponding name
  * entry in name page does not need to be touched during deletion.
  */
 void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
-						struct inode *inode)
+					struct inode *dir, struct inode *inode)
 {
 	struct	f2fs_dentry_block *dentry_blk;
 	unsigned int bit_pos;
-	struct inode *dir = page->mapping->host;
 	int slots = GET_DENTRY_SLOTS(le16_to_cpu(dentry->name_len));
 	int i;
 
+	if (f2fs_has_inline_dentry(dir))
+		return f2fs_delete_inline_entry(dentry, page, dir, inode);
+
 	lock_page(page);
 	f2fs_wait_on_page_writeback(page, DATA);
 
 	dentry_blk = page_address(page);
 	bit_pos = dentry - dentry_blk->dentry;
 	for (i = 0; i < slots; i++)
-		test_and_clear_bit_le(bit_pos + i, &dentry_blk->dentry_bitmap);
+		clear_bit_le(bit_pos + i, &dentry_blk->dentry_bitmap);
 
 	/* Let's check and deallocate this dentry page */
 	bit_pos = find_next_bit_le(&dentry_blk->dentry_bitmap,
@@ -590,29 +664,8 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
 
 	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
 
-	if (inode) {
+	if (inode)
-		struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
+		f2fs_drop_nlink(dir, inode, NULL);
-
-		down_write(&F2FS_I(inode)->i_sem);
-
-		if (S_ISDIR(inode->i_mode)) {
-			drop_nlink(dir);
-			update_inode_page(dir);
-		}
-		inode->i_ctime = CURRENT_TIME;
-		drop_nlink(inode);
-		if (S_ISDIR(inode->i_mode)) {
-			drop_nlink(inode);
-			i_size_write(inode, 0);
-		}
-		up_write(&F2FS_I(inode)->i_sem);
-		update_inode_page(inode);
-
-		if (inode->i_nlink == 0)
-			add_orphan_inode(sbi, inode->i_ino);
-		else
-			release_orphan_inode(sbi);
-	}
 
 	if (bit_pos == NR_DENTRY_IN_BLOCK) {
 		truncate_hole(dir, page->index, page->index + 1);
@@ -628,9 +681,12 @@ bool f2fs_empty_dir(struct inode *dir)
 	unsigned long bidx;
 	struct page *dentry_page;
 	unsigned int bit_pos;
-	struct	f2fs_dentry_block *dentry_blk;
+	struct f2fs_dentry_block *dentry_blk;
 	unsigned long nblock = dir_blocks(dir);
 
+	if (f2fs_has_inline_dentry(dir))
+		return f2fs_empty_inline_dir(dir);
+
 	for (bidx = 0; bidx < nblock; bidx++) {
 		dentry_page = get_lock_data_page(dir, bidx);
 		if (IS_ERR(dentry_page)) {
@@ -640,7 +696,6 @@ bool f2fs_empty_dir(struct inode *dir)
 				return false;
 		}
 
-
 		dentry_blk = kmap_atomic(dentry_page);
 		if (bidx == 0)
 			bit_pos = 2;
@@ -659,19 +714,48 @@ bool f2fs_empty_dir(struct inode *dir)
 	return true;
 }
 
+bool f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d,
+						unsigned int start_pos)
+{
+	unsigned char d_type = DT_UNKNOWN;
+	unsigned int bit_pos;
+	struct f2fs_dir_entry *de = NULL;
+
+	bit_pos = ((unsigned long)ctx->pos % d->max);
+
+	while (bit_pos < d->max) {
+		bit_pos = find_next_bit_le(d->bitmap, d->max, bit_pos);
+		if (bit_pos >= d->max)
+			break;
+
+		de = &d->dentry[bit_pos];
+		if (de->file_type < F2FS_FT_MAX)
+			d_type = f2fs_filetype_table[de->file_type];
+		else
+			d_type = DT_UNKNOWN;
+		if (!dir_emit(ctx, d->filename[bit_pos],
+					le16_to_cpu(de->name_len),
+					le32_to_cpu(de->ino), d_type))
+			return true;
+
+		bit_pos += GET_DENTRY_SLOTS(le16_to_cpu(de->name_len));
+		ctx->pos = start_pos + bit_pos;
+	}
+	return false;
+}
+
 static int f2fs_readdir(struct file *file, struct dir_context *ctx)
 {
 	struct inode *inode = file_inode(file);
 	unsigned long npages = dir_blocks(inode);
-	unsigned int bit_pos = 0;
 	struct f2fs_dentry_block *dentry_blk = NULL;
-	struct f2fs_dir_entry *de = NULL;
 	struct page *dentry_page = NULL;
 	struct file_ra_state *ra = &file->f_ra;
 	unsigned int n = ((unsigned long)ctx->pos / NR_DENTRY_IN_BLOCK);
-	unsigned char d_type = DT_UNKNOWN;
+	struct f2fs_dentry_ptr d;
 
-	bit_pos = ((unsigned long)ctx->pos % NR_DENTRY_IN_BLOCK);
+	if (f2fs_has_inline_dentry(inode))
+		return f2fs_read_inline_dir(file, ctx);
 
 	/* readahead for multi pages of dir */
 	if (npages - n > 1 && !ra_has_index(ra, n))
@@ -684,28 +768,12 @@ static int f2fs_readdir(struct file *file, struct dir_context *ctx)
 			continue;
 
 		dentry_blk = kmap(dentry_page);
-		while (bit_pos < NR_DENTRY_IN_BLOCK) {
-			bit_pos = find_next_bit_le(&dentry_blk->dentry_bitmap,
-							NR_DENTRY_IN_BLOCK,
-							bit_pos);
-			if (bit_pos >= NR_DENTRY_IN_BLOCK)
-				break;
 
-			de = &dentry_blk->dentry[bit_pos];
+		make_dentry_ptr(&d, (void *)dentry_blk, 1);
-			if (de->file_type < F2FS_FT_MAX)
+
-				d_type = f2fs_filetype_table[de->file_type];
+		if (f2fs_fill_dentries(ctx, &d, n * NR_DENTRY_IN_BLOCK))
-			else
+			goto stop;
-				d_type = DT_UNKNOWN;
-			if (!dir_emit(ctx,
-					dentry_blk->filename[bit_pos],
-					le16_to_cpu(de->name_len),
-					le32_to_cpu(de->ino), d_type))
-				goto stop;
 
-			bit_pos += GET_DENTRY_SLOTS(le16_to_cpu(de->name_len));
-			ctx->pos = n * NR_DENTRY_IN_BLOCK + bit_pos;
-		}
-		bit_pos = 0;
 		ctx->pos = (n + 1) * NR_DENTRY_IN_BLOCK;
 		kunmap(dentry_page);
 		f2fs_put_page(dentry_page, 1);

fs/f2fs/f2fs.h

@@ -46,8 +46,10 @@
 #define F2FS_MOUNT_DISABLE_EXT_IDENTIFY	0x00000040
 #define F2FS_MOUNT_INLINE_XATTR		0x00000080
 #define F2FS_MOUNT_INLINE_DATA		0x00000100
-#define F2FS_MOUNT_FLUSH_MERGE		0x00000200
+#define F2FS_MOUNT_INLINE_DENTRY	0x00000200
-#define F2FS_MOUNT_NOBARRIER		0x00000400
+#define F2FS_MOUNT_FLUSH_MERGE		0x00000400
+#define F2FS_MOUNT_NOBARRIER		0x00000800
+#define F2FS_MOUNT_FASTBOOT		0x00001000
 
 #define clear_opt(sbi, option)	(sbi->mount_opt.opt &= ~F2FS_MOUNT_##option)
 #define set_opt(sbi, option)	(sbi->mount_opt.opt |= F2FS_MOUNT_##option)
@@ -211,6 +213,32 @@ static inline bool __has_cursum_space(struct f2fs_summary_block *sum, int size,
 /*
  * For INODE and NODE manager
  */
+/* for directory operations */
+struct f2fs_dentry_ptr {
+	const void *bitmap;
+	struct f2fs_dir_entry *dentry;
+	__u8 (*filename)[F2FS_SLOT_LEN];
+	int max;
+};
+
+static inline void make_dentry_ptr(struct f2fs_dentry_ptr *d,
+					void *src, int type)
+{
+	if (type == 1) {
+		struct f2fs_dentry_block *t = (struct f2fs_dentry_block *)src;
+		d->max = NR_DENTRY_IN_BLOCK;
+		d->bitmap = &t->dentry_bitmap;
+		d->dentry = t->dentry;
+		d->filename = t->filename;
+	} else {
+		struct f2fs_inline_dentry *t = (struct f2fs_inline_dentry *)src;
+		d->max = NR_INLINE_DENTRY;
+		d->bitmap = &t->dentry_bitmap;
+		d->dentry = t->dentry;
+		d->filename = t->filename;
+	}
+}
+
 /*
  * XATTR_NODE_OFFSET stores xattrs to one node block per file keeping -1
  * as its node offset to distinguish from index node blocks.
@@ -269,6 +297,7 @@ struct f2fs_inode_info {
 	struct extent_info ext;		/* in-memory extent cache entry */
 	struct dir_inode_entry *dirty_dir;	/* the pointer of dirty dir */
 
+	struct radix_tree_root inmem_root;	/* radix tree for inmem pages */
 	struct list_head inmem_pages;	/* inmemory pages managed by f2fs */
 	struct mutex inmem_lock;	/* lock for inmemory pages */
 };
@@ -303,7 +332,7 @@ struct f2fs_nm_info {
 	/* NAT cache management */
 	struct radix_tree_root nat_root;/* root of the nat entry cache */
 	struct radix_tree_root nat_set_root;/* root of the nat set cache */
-	rwlock_t nat_tree_lock;		/* protect nat_tree_lock */
+	struct rw_semaphore nat_tree_lock;	/* protect nat_tree_lock */
 	struct list_head nat_entries;	/* cached nat entry list (clean) */
 	unsigned int nat_cnt;		/* the # of cached nat entries */
 	unsigned int dirty_nat_cnt;	/* total num of nat entries in set */
@@ -433,6 +462,7 @@ enum count_type {
 	F2FS_DIRTY_DENTS,
 	F2FS_DIRTY_NODES,
 	F2FS_DIRTY_META,
+	F2FS_INMEM_PAGES,
 	NR_COUNT_TYPE,
 };
 
@@ -470,6 +500,14 @@ struct f2fs_bio_info {
 	struct rw_semaphore io_rwsem;	/* blocking op for bio */
 };
 
+/* for inner inode cache management */
+struct inode_management {
+	struct radix_tree_root ino_root;	/* ino entry array */
+	spinlock_t ino_lock;			/* for ino entry lock */
+	struct list_head ino_list;		/* inode list head */
+	unsigned long ino_num;			/* number of entries */
+};
+
 struct f2fs_sb_info {
 	struct super_block *sb;			/* pointer to VFS super block */
 	struct proc_dir_entry *s_proc;		/* proc entry */
@@ -488,7 +526,6 @@ struct f2fs_sb_info {
 	/* for bio operations */
 	struct f2fs_bio_info read_io;			/* for read bios */
 	struct f2fs_bio_info write_io[NR_PAGE_TYPE];	/* for write bios */
-	struct completion *wait_io;		/* for completion bios */
 
 	/* for checkpoint */
 	struct f2fs_checkpoint *ckpt;		/* raw checkpoint pointer */
@@ -500,13 +537,9 @@ struct f2fs_sb_info {
 	bool por_doing;				/* recovery is doing or not */
 	wait_queue_head_t cp_wait;
 
-	/* for inode management */
+	struct inode_management im[MAX_INO_ENTRY];      /* manage inode cache */
-	struct radix_tree_root ino_root[MAX_INO_ENTRY];	/* ino entry array */
-	spinlock_t ino_lock[MAX_INO_ENTRY];		/* for ino entry lock */
-	struct list_head ino_list[MAX_INO_ENTRY];	/* inode list head */
 
 	/* for orphan inode, use 0'th array */
-	unsigned int n_orphans;			/* # of orphan inodes */
 	unsigned int max_orphans;		/* max orphan inodes */
 
 	/* for directory inode management */
@@ -557,7 +590,8 @@ struct f2fs_sb_info {
 	unsigned int segment_count[2];		/* # of allocated segments */
 	unsigned int block_count[2];		/* # of allocated blocks */
 	int total_hit_ext, read_hit_ext;	/* extent cache hit ratio */
-	int inline_inode;			/* # of inline_data inodes */
+	atomic_t inline_inode;			/* # of inline_data inodes */
+	atomic_t inline_dir;			/* # of inline_dentry inodes */
 	int bg_gc;				/* background gc calls */
 	unsigned int n_dirty_dirs;		/* # of dir inodes */
 #endif
@@ -988,6 +1022,13 @@ retry:
 	return entry;
 }
 
+static inline void f2fs_radix_tree_insert(struct radix_tree_root *root,
+				unsigned long index, void *item)
+{
+	while (radix_tree_insert(root, index, item))
+		cond_resched();
+}
+
 #define RAW_IS_INODE(p)	((p)->footer.nid == (p)->footer.ino)
 
 static inline bool IS_INODE(struct page *page)
@@ -1020,7 +1061,7 @@ static inline int f2fs_test_bit(unsigned int nr, char *addr)
 	return mask & *addr;
 }
 
-static inline int f2fs_set_bit(unsigned int nr, char *addr)
+static inline int f2fs_test_and_set_bit(unsigned int nr, char *addr)
 {
 	int mask;
 	int ret;
@@ -1032,7 +1073,7 @@ static inline int f2fs_set_bit(unsigned int nr, char *addr)
 	return ret;
 }
 
-static inline int f2fs_clear_bit(unsigned int nr, char *addr)
+static inline int f2fs_test_and_clear_bit(unsigned int nr, char *addr)
 {
 	int mask;
 	int ret;
@@ -1044,6 +1085,15 @@ static inline int f2fs_clear_bit(unsigned int nr, char *addr)
 	return ret;
 }
 
+static inline void f2fs_change_bit(unsigned int nr, char *addr)
+{
+	int mask;
+
+	addr += (nr >> 3);
+	mask = 1 << (7 - (nr & 0x07));
+	*addr ^= mask;
+}
+
 /* used for f2fs_inode_info->flags */
 enum {
 	FI_NEW_INODE,		/* indicate newly allocated inode */
@@ -1057,11 +1107,13 @@ enum {
 	FI_NO_EXTENT,		/* not to use the extent cache */
 	FI_INLINE_XATTR,	/* used for inline xattr */
 	FI_INLINE_DATA,		/* used for inline data*/
+	FI_INLINE_DENTRY,	/* used for inline dentry */
 	FI_APPEND_WRITE,	/* inode has appended data */
 	FI_UPDATE_WRITE,	/* inode has in-place-update data */
 	FI_NEED_IPU,		/* used for ipu per file */
 	FI_ATOMIC_FILE,		/* indicate atomic file */
 	FI_VOLATILE_FILE,	/* indicate volatile file */
+	FI_DATA_EXIST,		/* indicate data exists */
 };
 
 static inline void set_inode_flag(struct f2fs_inode_info *fi, int flag)
@@ -1087,15 +1139,6 @@ static inline void set_acl_inode(struct f2fs_inode_info *fi, umode_t mode)
 	set_inode_flag(fi, FI_ACL_MODE);
 }
 
-static inline int cond_clear_inode_flag(struct f2fs_inode_info *fi, int flag)
-{
-	if (is_inode_flag_set(fi, FI_ACL_MODE)) {
-		clear_inode_flag(fi, FI_ACL_MODE);
-		return 1;
-	}
-	return 0;
-}
-
 static inline void get_inline_info(struct f2fs_inode_info *fi,
 					struct f2fs_inode *ri)
 {
@@ -1103,6 +1146,10 @@ static inline void get_inline_info(struct f2fs_inode_info *fi,
 		set_inode_flag(fi, FI_INLINE_XATTR);
 	if (ri->i_inline & F2FS_INLINE_DATA)
 		set_inode_flag(fi, FI_INLINE_DATA);
+	if (ri->i_inline & F2FS_INLINE_DENTRY)
+		set_inode_flag(fi, FI_INLINE_DENTRY);
+	if (ri->i_inline & F2FS_DATA_EXIST)
+		set_inode_flag(fi, FI_DATA_EXIST);
 }
 
 static inline void set_raw_inline(struct f2fs_inode_info *fi,
@@ -1114,6 +1161,10 @@ static inline void set_raw_inline(struct f2fs_inode_info *fi,
 		ri->i_inline |= F2FS_INLINE_XATTR;
 	if (is_inode_flag_set(fi, FI_INLINE_DATA))
 		ri->i_inline |= F2FS_INLINE_DATA;
+	if (is_inode_flag_set(fi, FI_INLINE_DENTRY))
+		ri->i_inline |= F2FS_INLINE_DENTRY;
+	if (is_inode_flag_set(fi, FI_DATA_EXIST))
+		ri->i_inline |= F2FS_DATA_EXIST;
 }
 
 static inline int f2fs_has_inline_xattr(struct inode *inode)
@@ -1148,6 +1199,17 @@ static inline int f2fs_has_inline_data(struct inode *inode)
 	return is_inode_flag_set(F2FS_I(inode), FI_INLINE_DATA);
 }
 
+static inline void f2fs_clear_inline_inode(struct inode *inode)
+{
+	clear_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
+	clear_inode_flag(F2FS_I(inode), FI_DATA_EXIST);
+}
+
+static inline int f2fs_exist_data(struct inode *inode)
+{
+	return is_inode_flag_set(F2FS_I(inode), FI_DATA_EXIST);
+}
+
 static inline bool f2fs_is_atomic_file(struct inode *inode)
 {
 	return is_inode_flag_set(F2FS_I(inode), FI_ATOMIC_FILE);
@@ -1164,6 +1226,23 @@ static inline void *inline_data_addr(struct page *page)
 	return (void *)&(ri->i_addr[1]);
 }
 
+static inline int f2fs_has_inline_dentry(struct inode *inode)
+{
+	return is_inode_flag_set(F2FS_I(inode), FI_INLINE_DENTRY);
+}
+
+static inline void *inline_dentry_addr(struct page *page)
+{
+	struct f2fs_inode *ri = F2FS_INODE(page);
+	return (void *)&(ri->i_addr[1]);
+}
+
+static inline void f2fs_dentry_kunmap(struct inode *dir, struct page *page)
+{
+	if (!f2fs_has_inline_dentry(dir))
+		kunmap(page);
+}
+
 static inline int f2fs_readonly(struct super_block *sb)
 {
 	return sb->s_flags & MS_RDONLY;
@@ -1224,6 +1303,19 @@ struct dentry *f2fs_get_parent(struct dentry *child);
 /*
  * dir.c
  */
+extern unsigned char f2fs_filetype_table[F2FS_FT_MAX];
+void set_de_type(struct f2fs_dir_entry *, struct inode *);
+struct f2fs_dir_entry *find_target_dentry(struct qstr *, int *,
+			struct f2fs_dentry_ptr *);
+bool f2fs_fill_dentries(struct dir_context *, struct f2fs_dentry_ptr *,
+			unsigned int);
+void do_make_empty_dir(struct inode *, struct inode *,
+			struct f2fs_dentry_ptr *);
+struct page *init_inode_metadata(struct inode *, struct inode *,
+			const struct qstr *, struct page *);
+void update_parent_metadata(struct inode *, struct inode *, unsigned int);
+int room_for_filename(const void *, int, int);
+void f2fs_drop_nlink(struct inode *, struct inode *, struct page *);
 struct f2fs_dir_entry *f2fs_find_entry(struct inode *, struct qstr *,
 							struct page **);
 struct f2fs_dir_entry *f2fs_parent_dir(struct inode *, struct page **);
@@ -1232,7 +1324,8 @@ void f2fs_set_link(struct inode *, struct f2fs_dir_entry *,
 				struct page *, struct inode *);
 int update_dent_inode(struct inode *, const struct qstr *);
 int __f2fs_add_link(struct inode *, const struct qstr *, struct inode *);
-void f2fs_delete_entry(struct f2fs_dir_entry *, struct page *, struct inode *);
+void f2fs_delete_entry(struct f2fs_dir_entry *, struct page *, struct inode *,
+							struct inode *);
 int f2fs_do_tmpfile(struct inode *, struct inode *);
 int f2fs_make_empty(struct inode *, struct inode *);
 bool f2fs_empty_dir(struct inode *);
@@ -1296,6 +1389,7 @@ void destroy_node_manager_caches(void);
  * segment.c
  */
 void register_inmem_page(struct inode *, struct page *);
+void invalidate_inmem_page(struct inode *, struct page *);
 void commit_inmem_pages(struct inode *, bool);
 void f2fs_balance_fs(struct f2fs_sb_info *);
 void f2fs_balance_fs_bg(struct f2fs_sb_info *);
@@ -1337,8 +1431,8 @@ void destroy_segment_manager_caches(void);
  */
 struct page *grab_meta_page(struct f2fs_sb_info *, pgoff_t);
 struct page *get_meta_page(struct f2fs_sb_info *, pgoff_t);
-struct page *get_meta_page_ra(struct f2fs_sb_info *, pgoff_t);
 int ra_meta_pages(struct f2fs_sb_info *, block_t, int, int);
+void ra_meta_pages_cond(struct f2fs_sb_info *, pgoff_t);
 long sync_meta_pages(struct f2fs_sb_info *, enum page_type, long);
 void add_dirty_inode(struct f2fs_sb_info *, nid_t, int type);
 void remove_dirty_inode(struct f2fs_sb_info *, nid_t, int type);
@@ -1405,7 +1499,7 @@ struct f2fs_stat_info {
 	int ndirty_node, ndirty_dent, ndirty_dirs, ndirty_meta;
 	int nats, sits, fnids;
 	int total_count, utilization;
-	int bg_gc, inline_inode;
+	int bg_gc, inline_inode, inline_dir, inmem_pages;
 	unsigned int valid_count, valid_node_count, valid_inode_count;
 	unsigned int bimodal, avg_vblocks;
 	int util_free, util_valid, util_invalid;
@@ -1438,14 +1532,23 @@ static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi)
 #define stat_inc_inline_inode(inode)					\
 	do {								\
 		if (f2fs_has_inline_data(inode))			\
-			((F2FS_I_SB(inode))->inline_inode++);		\
+			(atomic_inc(&F2FS_I_SB(inode)->inline_inode));	\
 	} while (0)
 #define stat_dec_inline_inode(inode)					\
 	do {								\
 		if (f2fs_has_inline_data(inode))			\
-			((F2FS_I_SB(inode))->inline_inode--);		\
+			(atomic_dec(&F2FS_I_SB(inode)->inline_inode));	\
+	} while (0)
+#define stat_inc_inline_dir(inode)					\
+	do {								\
+		if (f2fs_has_inline_dentry(inode))			\
+			(atomic_inc(&F2FS_I_SB(inode)->inline_dir));	\
+	} while (0)
+#define stat_dec_inline_dir(inode)					\
+	do {								\
+		if (f2fs_has_inline_dentry(inode))			\
+			(atomic_dec(&F2FS_I_SB(inode)->inline_dir));	\
 	} while (0)
-
 #define stat_inc_seg_type(sbi, curseg)					\
 		((sbi)->segment_count[(curseg)->alloc_type]++)
 #define stat_inc_block_count(sbi, curseg)				\
@@ -1492,6 +1595,8 @@ void f2fs_destroy_root_stats(void);
 #define stat_inc_read_hit(sb)
 #define stat_inc_inline_inode(inode)
 #define stat_dec_inline_inode(inode)
+#define stat_inc_inline_dir(inode)
+#define stat_dec_inline_dir(inode)
 #define stat_inc_seg_type(sbi, curseg)
 #define stat_inc_block_count(sbi, curseg)
 #define stat_inc_seg_count(si, type)
@@ -1519,9 +1624,20 @@ extern const struct inode_operations f2fs_special_inode_operations;
  * inline.c
  */
 bool f2fs_may_inline(struct inode *);
+void read_inline_data(struct page *, struct page *);
 int f2fs_read_inline_data(struct inode *, struct page *);
-int f2fs_convert_inline_data(struct inode *, pgoff_t, struct page *);
+int f2fs_convert_inline_page(struct dnode_of_data *, struct page *);
-int f2fs_write_inline_data(struct inode *, struct page *, unsigned int);
+int f2fs_convert_inline_inode(struct inode *);
-void truncate_inline_data(struct inode *, u64);
+int f2fs_write_inline_data(struct inode *, struct page *);
+void truncate_inline_data(struct page *, u64);
 bool recover_inline_data(struct inode *, struct page *);
+struct f2fs_dir_entry *find_in_inline_dir(struct inode *, struct qstr *,
+							struct page **);
+struct f2fs_dir_entry *f2fs_parent_inline_dir(struct inode *, struct page **);
+int make_empty_inline_dir(struct inode *inode, struct inode *, struct page *);
+int f2fs_add_inline_entry(struct inode *, const struct qstr *, struct inode *);
+void f2fs_delete_inline_entry(struct f2fs_dir_entry *, struct page *,
+						struct inode *, struct inode *);
+bool f2fs_empty_inline_dir(struct inode *);
+int f2fs_read_inline_dir(struct file *, struct dir_context *);
 #endif

fs/f2fs/file.c

@@ -41,18 +41,18 @@ static int f2fs_vm_page_mkwrite(struct vm_area_struct *vma,
 
 	sb_start_pagefault(inode->i_sb);
 
-	/* force to convert with normal data indices */
+	f2fs_bug_on(sbi, f2fs_has_inline_data(inode));
-	err = f2fs_convert_inline_data(inode, MAX_INLINE_DATA + 1, page);
-	if (err)
-		goto out;
 
 	/* block allocation */
 	f2fs_lock_op(sbi);
 	set_new_dnode(&dn, inode, NULL, NULL, 0);
 	err = f2fs_reserve_block(&dn, page->index);
-	f2fs_unlock_op(sbi);
+	if (err) {
-	if (err)
+		f2fs_unlock_op(sbi);
 		goto out;
+	}
+	f2fs_put_dnode(&dn);
+	f2fs_unlock_op(sbi);
 
 	file_update_time(vma->vm_file);
 	lock_page(page);
@@ -130,10 +130,45 @@ static inline bool need_do_checkpoint(struct inode *inode)
 		need_cp = true;
 	else if (F2FS_I(inode)->xattr_ver == cur_cp_version(F2FS_CKPT(sbi)))
 		need_cp = true;
+	else if (test_opt(sbi, FASTBOOT))
+		need_cp = true;
+	else if (sbi->active_logs == 2)
+		need_cp = true;
 
 	return need_cp;
 }
 
+static bool need_inode_page_update(struct f2fs_sb_info *sbi, nid_t ino)
+{
+	struct page *i = find_get_page(NODE_MAPPING(sbi), ino);
+	bool ret = false;
+	/* But we need to avoid that there are some inode updates */
+	if ((i && PageDirty(i)) || need_inode_block_update(sbi, ino))
+		ret = true;
+	f2fs_put_page(i, 0);
+	return ret;
+}
+
+static void try_to_fix_pino(struct inode *inode)
+{
+	struct f2fs_inode_info *fi = F2FS_I(inode);
+	nid_t pino;
+
+	down_write(&fi->i_sem);
+	fi->xattr_ver = 0;
+	if (file_wrong_pino(inode) && inode->i_nlink == 1 &&
+			get_parent_ino(inode, &pino)) {
+		fi->i_pino = pino;
+		file_got_pino(inode);
+		up_write(&fi->i_sem);
+
+		mark_inode_dirty_sync(inode);
+		f2fs_write_inode(inode, NULL);
+	} else {
+		up_write(&fi->i_sem);
+	}
+}
+
 int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 {
 	struct inode *inode = file->f_mapping->host;
@@ -164,19 +199,21 @@ int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 		return ret;
 	}
 
+	/* if the inode is dirty, let's recover all the time */
+	if (!datasync && is_inode_flag_set(fi, FI_DIRTY_INODE)) {
+		update_inode_page(inode);
+		goto go_write;
+	}
+
 	/*
 	 * if there is no written data, don't waste time to write recovery info.
 	 */
 	if (!is_inode_flag_set(fi, FI_APPEND_WRITE) &&
 			!exist_written_data(sbi, ino, APPEND_INO)) {
-		struct page *i = find_get_page(NODE_MAPPING(sbi), ino);
 
-		/* But we need to avoid that there are some inode updates */
+		/* it may call write_inode just prior to fsync */
-		if ((i && PageDirty(i)) || need_inode_block_update(sbi, ino)) {
+		if (need_inode_page_update(sbi, ino))
-			f2fs_put_page(i, 0);
 			goto go_write;
-		}
-		f2fs_put_page(i, 0);
 
 		if (is_inode_flag_set(fi, FI_UPDATE_WRITE) ||
 				exist_written_data(sbi, ino, UPDATE_INO))
@@ -196,49 +233,36 @@ go_write:
 	up_read(&fi->i_sem);
 
 	if (need_cp) {
-		nid_t pino;
-
 		/* all the dirty node pages should be flushed for POR */
 		ret = f2fs_sync_fs(inode->i_sb, 1);
 
-		down_write(&fi->i_sem);
+		/*
-		F2FS_I(inode)->xattr_ver = 0;
+		 * We've secured consistency through sync_fs. Following pino
-		if (file_wrong_pino(inode) && inode->i_nlink == 1 &&
+		 * will be used only for fsynced inodes after checkpoint.
-					get_parent_ino(inode, &pino)) {
+		 */
-			F2FS_I(inode)->i_pino = pino;
+		try_to_fix_pino(inode);
-			file_got_pino(inode);
+		goto out;
-			up_write(&fi->i_sem);
-			mark_inode_dirty_sync(inode);
-			ret = f2fs_write_inode(inode, NULL);
-			if (ret)
-				goto out;
-		} else {
-			up_write(&fi->i_sem);
-		}
-	} else {
-sync_nodes:
-		sync_node_pages(sbi, ino, &wbc);
-
-		if (need_inode_block_update(sbi, ino)) {
-			mark_inode_dirty_sync(inode);
-			ret = f2fs_write_inode(inode, NULL);
-			if (ret)
-				goto out;
-			goto sync_nodes;
-		}
-
-		ret = wait_on_node_pages_writeback(sbi, ino);
-		if (ret)
-			goto out;
-
-		/* once recovery info is written, don't need to tack this */
-		remove_dirty_inode(sbi, ino, APPEND_INO);
-		clear_inode_flag(fi, FI_APPEND_WRITE);
-flush_out:
-		remove_dirty_inode(sbi, ino, UPDATE_INO);
-		clear_inode_flag(fi, FI_UPDATE_WRITE);
-		ret = f2fs_issue_flush(F2FS_I_SB(inode));
 	}
+sync_nodes:
+	sync_node_pages(sbi, ino, &wbc);
+
+	if (need_inode_block_update(sbi, ino)) {
+		mark_inode_dirty_sync(inode);
+		f2fs_write_inode(inode, NULL);
+		goto sync_nodes;
+	}
+
+	ret = wait_on_node_pages_writeback(sbi, ino);
+	if (ret)
+		goto out;
+
+	/* once recovery info is written, don't need to tack this */
+	remove_dirty_inode(sbi, ino, APPEND_INO);
+	clear_inode_flag(fi, FI_APPEND_WRITE);
+flush_out:
+	remove_dirty_inode(sbi, ino, UPDATE_INO);
+	clear_inode_flag(fi, FI_UPDATE_WRITE);
+	ret = f2fs_issue_flush(sbi);
 out:
 	trace_f2fs_sync_file_exit(inode, need_cp, datasync, ret);
 	return ret;
@@ -296,7 +320,7 @@ static loff_t f2fs_seek_block(struct file *file, loff_t offset, int whence)
 		goto fail;
 
 	/* handle inline data case */
-	if (f2fs_has_inline_data(inode)) {
+	if (f2fs_has_inline_data(inode) || f2fs_has_inline_dentry(inode)) {
 		if (whence == SEEK_HOLE)
 			data_ofs = isize;
 		goto found;
@@ -374,6 +398,15 @@ static loff_t f2fs_llseek(struct file *file, loff_t offset, int whence)
 
 static int f2fs_file_mmap(struct file *file, struct vm_area_struct *vma)
 {
+	struct inode *inode = file_inode(file);
+
+	/* we don't need to use inline_data strictly */
+	if (f2fs_has_inline_data(inode)) {
+		int err = f2fs_convert_inline_inode(inode);
+		if (err)
+			return err;
+	}
+
 	file_accessed(file);
 	vma->vm_ops = &f2fs_file_vm_ops;
 	return 0;
@@ -415,20 +448,17 @@ void truncate_data_blocks(struct dnode_of_data *dn)
 	truncate_data_blocks_range(dn, ADDRS_PER_BLOCK);
 }
 
-static void truncate_partial_data_page(struct inode *inode, u64 from)
+static int truncate_partial_data_page(struct inode *inode, u64 from)
 {
 	unsigned offset = from & (PAGE_CACHE_SIZE - 1);
 	struct page *page;
 
-	if (f2fs_has_inline_data(inode))
-		return truncate_inline_data(inode, from);
-
 	if (!offset)
-		return;
+		return 0;
 
 	page = find_data_page(inode, from >> PAGE_CACHE_SHIFT, false);
 	if (IS_ERR(page))
-		return;
+		return 0;
 
 	lock_page(page);
 	if (unlikely(!PageUptodate(page) ||
@@ -438,9 +468,9 @@ static void truncate_partial_data_page(struct inode *inode, u64 from)
 	f2fs_wait_on_page_writeback(page, DATA);
 	zero_user(page, offset, PAGE_CACHE_SIZE - offset);
 	set_page_dirty(page);
-
 out:
 	f2fs_put_page(page, 1);
+	return 0;
 }
 
 int truncate_blocks(struct inode *inode, u64 from, bool lock)
@@ -450,27 +480,33 @@ int truncate_blocks(struct inode *inode, u64 from, bool lock)
 	struct dnode_of_data dn;
 	pgoff_t free_from;
 	int count = 0, err = 0;
+	struct page *ipage;
 
 	trace_f2fs_truncate_blocks_enter(inode, from);
 
-	if (f2fs_has_inline_data(inode))
-		goto done;
-
 	free_from = (pgoff_t)
-			((from + blocksize - 1) >> (sbi->log_blocksize));
+		((from + blocksize - 1) >> (sbi->log_blocksize));
 
 	if (lock)
 		f2fs_lock_op(sbi);
 
-	set_new_dnode(&dn, inode, NULL, NULL, 0);
+	ipage = get_node_page(sbi, inode->i_ino);
+	if (IS_ERR(ipage)) {
+		err = PTR_ERR(ipage);
+		goto out;
+	}
+
+	if (f2fs_has_inline_data(inode)) {
+		f2fs_put_page(ipage, 1);
+		goto out;
+	}
+
+	set_new_dnode(&dn, inode, ipage, NULL, 0);
 	err = get_dnode_of_data(&dn, free_from, LOOKUP_NODE);
 	if (err) {
 		if (err == -ENOENT)
 			goto free_next;
-		if (lock)
+		goto out;
-			f2fs_unlock_op(sbi);
-		trace_f2fs_truncate_blocks_exit(inode, err);
-		return err;
 	}
 
 	count = ADDRS_PER_PAGE(dn.node_page, F2FS_I(inode));
@@ -486,11 +522,13 @@ int truncate_blocks(struct inode *inode, u64 from, bool lock)
 	f2fs_put_dnode(&dn);
 free_next:
 	err = truncate_inode_blocks(inode, free_from);
+out:
 	if (lock)
 		f2fs_unlock_op(sbi);
-done:
+
 	/* lastly zero out the first data page */
-	truncate_partial_data_page(inode, from);
+	if (!err)
+		err = truncate_partial_data_page(inode, from);
 
 	trace_f2fs_truncate_blocks_exit(inode, err);
 	return err;
@@ -504,6 +542,12 @@ void f2fs_truncate(struct inode *inode)
 
 	trace_f2fs_truncate(inode);
 
+	/* we should check inline_data size */
+	if (f2fs_has_inline_data(inode) && !f2fs_may_inline(inode)) {
+		if (f2fs_convert_inline_inode(inode))
+			return;
+	}
+
 	if (!truncate_blocks(inode, i_size_read(inode), true)) {
 		inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 		mark_inode_dirty(inode);
@@ -561,10 +605,6 @@ int f2fs_setattr(struct dentry *dentry, struct iattr *attr)
 		return err;
 
 	if (attr->ia_valid & ATTR_SIZE) {
-		err = f2fs_convert_inline_data(inode, attr->ia_size, NULL);
-		if (err)
-			return err;
-
 		if (attr->ia_size != i_size_read(inode)) {
 			truncate_setsize(inode, attr->ia_size);
 			f2fs_truncate(inode);
@@ -665,9 +705,11 @@ static int punch_hole(struct inode *inode, loff_t offset, loff_t len)
 	if (offset >= inode->i_size)
 		return ret;
 
-	ret = f2fs_convert_inline_data(inode, MAX_INLINE_DATA + 1, NULL);
+	if (f2fs_has_inline_data(inode)) {
-	if (ret)
+		ret = f2fs_convert_inline_inode(inode);
-		return ret;
+		if (ret)
+			return ret;
+	}
 
 	pg_start = ((unsigned long long) offset) >> PAGE_CACHE_SHIFT;
 	pg_end = ((unsigned long long) offset + len) >> PAGE_CACHE_SHIFT;
@@ -721,9 +763,11 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
 	if (ret)
 		return ret;
 
-	ret = f2fs_convert_inline_data(inode, offset + len, NULL);
+	if (f2fs_has_inline_data(inode)) {
-	if (ret)
+		ret = f2fs_convert_inline_inode(inode);
-		return ret;
+		if (ret)
+			return ret;
+	}
 
 	pg_start = ((unsigned long long) offset) >> PAGE_CACHE_SHIFT;
 	pg_end = ((unsigned long long) offset + len) >> PAGE_CACHE_SHIFT;
@@ -874,7 +918,15 @@ static int f2fs_ioc_start_atomic_write(struct file *filp)
 
 	set_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE);
 
-	return f2fs_convert_inline_data(inode, MAX_INLINE_DATA + 1, NULL);
+	return f2fs_convert_inline_inode(inode);
+}
+
+static int f2fs_release_file(struct inode *inode, struct file *filp)
+{
+	/* some remained atomic pages should discarded */
+	if (f2fs_is_atomic_file(inode) || f2fs_is_volatile_file(inode))
+		commit_inmem_pages(inode, true);
+	return 0;
 }
 
 static int f2fs_ioc_commit_atomic_write(struct file *filp)
@@ -908,7 +960,8 @@ static int f2fs_ioc_start_volatile_write(struct file *filp)
 		return -EACCES;
 
 	set_inode_flag(F2FS_I(inode), FI_VOLATILE_FILE);
-	return 0;
+
+	return f2fs_convert_inline_inode(inode);
 }
 
 static int f2fs_ioc_fitrim(struct file *filp, unsigned long arg)
@@ -985,6 +1038,7 @@ const struct file_operations f2fs_file_operations = {
 	.read_iter	= generic_file_read_iter,
 	.write_iter	= generic_file_write_iter,
 	.open		= generic_file_open,
+	.release	= f2fs_release_file,
 	.mmap		= f2fs_file_mmap,
 	.fsync		= f2fs_sync_file,
 	.fallocate	= f2fs_fallocate,

fs/f2fs/gc.c

@@ -96,8 +96,6 @@ int start_gc_thread(struct f2fs_sb_info *sbi)
 	dev_t dev = sbi->sb->s_bdev->bd_dev;
 	int err = 0;
 
-	if (!test_opt(sbi, BG_GC))
-		goto out;
 	gc_th = kmalloc(sizeof(struct f2fs_gc_kthread), GFP_KERNEL);
 	if (!gc_th) {
 		err = -ENOMEM;
@@ -340,34 +338,39 @@ static const struct victim_selection default_v_ops = {
 	.get_victim = get_victim_by_default,
 };
 
-static struct inode *find_gc_inode(nid_t ino, struct list_head *ilist)
+static struct inode *find_gc_inode(struct gc_inode_list *gc_list, nid_t ino)
 {
 	struct inode_entry *ie;
 
-	list_for_each_entry(ie, ilist, list)
+	ie = radix_tree_lookup(&gc_list->iroot, ino);
-		if (ie->inode->i_ino == ino)
+	if (ie)
-			return ie->inode;
+		return ie->inode;
 	return NULL;
 }
 
-static void add_gc_inode(struct inode *inode, struct list_head *ilist)
+static void add_gc_inode(struct gc_inode_list *gc_list, struct inode *inode)
 {
 	struct inode_entry *new_ie;
 
-	if (inode == find_gc_inode(inode->i_ino, ilist)) {
+	if (inode == find_gc_inode(gc_list, inode->i_ino)) {
 		iput(inode);
 		return;
 	}
-
 	new_ie = f2fs_kmem_cache_alloc(winode_slab, GFP_NOFS);
 	new_ie->inode = inode;
-	list_add_tail(&new_ie->list, ilist);
+retry:
+	if (radix_tree_insert(&gc_list->iroot, inode->i_ino, new_ie)) {
+		cond_resched();
+		goto retry;
+	}
+	list_add_tail(&new_ie->list, &gc_list->ilist);
 }
 
-static void put_gc_inode(struct list_head *ilist)
+static void put_gc_inode(struct gc_inode_list *gc_list)
 {
 	struct inode_entry *ie, *next_ie;
-	list_for_each_entry_safe(ie, next_ie, ilist, list) {
+	list_for_each_entry_safe(ie, next_ie, &gc_list->ilist, list) {
+		radix_tree_delete(&gc_list->iroot, ie->inode->i_ino);
 		iput(ie->inode);
 		list_del(&ie->list);
 		kmem_cache_free(winode_slab, ie);
@@ -553,7 +556,7 @@ out:
  * the victim data block is ignored.
  */
 static void gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
-		struct list_head *ilist, unsigned int segno, int gc_type)
+		struct gc_inode_list *gc_list, unsigned int segno, int gc_type)
 {
 	struct super_block *sb = sbi->sb;
 	struct f2fs_summary *entry;
@@ -605,27 +608,27 @@ next_step:
 
 			data_page = find_data_page(inode,
 					start_bidx + ofs_in_node, false);
-			if (IS_ERR(data_page))
+			if (IS_ERR(data_page)) {
-				goto next_iput;
+				iput(inode);
+				continue;
+			}
 
 			f2fs_put_page(data_page, 0);
-			add_gc_inode(inode, ilist);
+			add_gc_inode(gc_list, inode);
-		} else {
+			continue;
-			inode = find_gc_inode(dni.ino, ilist);
+		}
-			if (inode) {
+
-				start_bidx = start_bidx_of_node(nofs,
+		/* phase 3 */
-								F2FS_I(inode));
+		inode = find_gc_inode(gc_list, dni.ino);
-				data_page = get_lock_data_page(inode,
+		if (inode) {
-						start_bidx + ofs_in_node);
+			start_bidx = start_bidx_of_node(nofs, F2FS_I(inode));
-				if (IS_ERR(data_page))
+			data_page = get_lock_data_page(inode,
-					continue;
+						start_bidx + ofs_in_node);
-				move_data_page(inode, data_page, gc_type);
+			if (IS_ERR(data_page))
-				stat_inc_data_blk_count(sbi, 1);
+				continue;
-			}
+			move_data_page(inode, data_page, gc_type);
+			stat_inc_data_blk_count(sbi, 1);
 		}
-		continue;
-next_iput:
-		iput(inode);
 	}
 
 	if (++phase < 4)
@@ -646,18 +649,20 @@ next_iput:
 }
 
 static int __get_victim(struct f2fs_sb_info *sbi, unsigned int *victim,
-						int gc_type, int type)
+			int gc_type)
 {
 	struct sit_info *sit_i = SIT_I(sbi);
 	int ret;
+
 	mutex_lock(&sit_i->sentry_lock);
-	ret = DIRTY_I(sbi)->v_ops->get_victim(sbi, victim, gc_type, type, LFS);
+	ret = DIRTY_I(sbi)->v_ops->get_victim(sbi, victim, gc_type,
+					      NO_CHECK_TYPE, LFS);
 	mutex_unlock(&sit_i->sentry_lock);
 	return ret;
 }
 
 static void do_garbage_collect(struct f2fs_sb_info *sbi, unsigned int segno,
-				struct list_head *ilist, int gc_type)
+				struct gc_inode_list *gc_list, int gc_type)
 {
 	struct page *sum_page;
 	struct f2fs_summary_block *sum;
@@ -675,7 +680,7 @@ static void do_garbage_collect(struct f2fs_sb_info *sbi, unsigned int segno,
 		gc_node_segment(sbi, sum->entries, segno, gc_type);
 		break;
 	case SUM_TYPE_DATA:
-		gc_data_segment(sbi, sum->entries, ilist, segno, gc_type);
+		gc_data_segment(sbi, sum->entries, gc_list, segno, gc_type);
 		break;
 	}
 	blk_finish_plug(&plug);
@@ -688,16 +693,18 @@ static void do_garbage_collect(struct f2fs_sb_info *sbi, unsigned int segno,
 
 int f2fs_gc(struct f2fs_sb_info *sbi)
 {
-	struct list_head ilist;
 	unsigned int segno, i;
 	int gc_type = BG_GC;
 	int nfree = 0;
 	int ret = -1;
-	struct cp_control cpc = {
+	struct cp_control cpc;
-		.reason = CP_SYNC,
+	struct gc_inode_list gc_list = {
+		.ilist = LIST_HEAD_INIT(gc_list.ilist),
+		.iroot = RADIX_TREE_INIT(GFP_NOFS),
 	};
 
-	INIT_LIST_HEAD(&ilist);
+	cpc.reason = test_opt(sbi, FASTBOOT) ? CP_UMOUNT : CP_SYNC;
+
 gc_more:
 	if (unlikely(!(sbi->sb->s_flags & MS_ACTIVE)))
 		goto stop;
@@ -709,7 +716,7 @@ gc_more:
 		write_checkpoint(sbi, &cpc);
 	}
 
-	if (!__get_victim(sbi, &segno, gc_type, NO_CHECK_TYPE))
+	if (!__get_victim(sbi, &segno, gc_type))
 		goto stop;
 	ret = 0;
 
@@ -719,7 +726,7 @@ gc_more:
 								META_SSA);
 
 	for (i = 0; i < sbi->segs_per_sec; i++)
-		do_garbage_collect(sbi, segno + i, &ilist, gc_type);
+		do_garbage_collect(sbi, segno + i, &gc_list, gc_type);
 
 	if (gc_type == FG_GC) {
 		sbi->cur_victim_sec = NULL_SEGNO;
@@ -735,7 +742,7 @@ gc_more:
 stop:
 	mutex_unlock(&sbi->gc_mutex);
 
-	put_gc_inode(&ilist);
+	put_gc_inode(&gc_list);
 	return ret;
 }
 

fs/f2fs/gc.h

@@ -40,6 +40,11 @@ struct inode_entry {
 	struct inode *inode;
 };
 
+struct gc_inode_list {
+	struct list_head ilist;
+	struct radix_tree_root iroot;
+};
+
 /*
  * inline functions
  */

fs/f2fs/inline.c

@@ -15,35 +15,44 @@
 
 bool f2fs_may_inline(struct inode *inode)
 {
-	block_t nr_blocks;
-	loff_t i_size;
-
 	if (!test_opt(F2FS_I_SB(inode), INLINE_DATA))
 		return false;
 
 	if (f2fs_is_atomic_file(inode))
 		return false;
 
-	nr_blocks = F2FS_I(inode)->i_xattr_nid ? 3 : 2;
+	if (!S_ISREG(inode->i_mode))
-	if (inode->i_blocks > nr_blocks)
 		return false;
 
-	i_size = i_size_read(inode);
+	if (i_size_read(inode) > MAX_INLINE_DATA)
-	if (i_size > MAX_INLINE_DATA)
 		return false;
 
 	return true;
 }
 
+void read_inline_data(struct page *page, struct page *ipage)
+{
+	void *src_addr, *dst_addr;
+
+	if (PageUptodate(page))
+		return;
+
+	f2fs_bug_on(F2FS_P_SB(page), page->index);
+
+	zero_user_segment(page, MAX_INLINE_DATA, PAGE_CACHE_SIZE);
+
+	/* Copy the whole inline data block */
+	src_addr = inline_data_addr(ipage);
+	dst_addr = kmap_atomic(page);
+	memcpy(dst_addr, src_addr, MAX_INLINE_DATA);
+	flush_dcache_page(page);
+	kunmap_atomic(dst_addr);
+	SetPageUptodate(page);
+}
+
 int f2fs_read_inline_data(struct inode *inode, struct page *page)
 {
 	struct page *ipage;
-	void *src_addr, *dst_addr;
-
-	if (page->index) {
-		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
-		goto out;
-	}
 
 	ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino);
 	if (IS_ERR(ipage)) {
@@ -51,112 +60,116 @@ int f2fs_read_inline_data(struct inode *inode, struct page *page)
 		return PTR_ERR(ipage);
 	}
 
-	zero_user_segment(page, MAX_INLINE_DATA, PAGE_CACHE_SIZE);
+	if (!f2fs_has_inline_data(inode)) {
+		f2fs_put_page(ipage, 1);
+		return -EAGAIN;
+	}
 
-	/* Copy the whole inline data block */
+	if (page->index)
-	src_addr = inline_data_addr(ipage);
+		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
-	dst_addr = kmap(page);
+	else
-	memcpy(dst_addr, src_addr, MAX_INLINE_DATA);
+		read_inline_data(page, ipage);
-	kunmap(page);
-	f2fs_put_page(ipage, 1);
 
-out:
 	SetPageUptodate(page);
+	f2fs_put_page(ipage, 1);
 	unlock_page(page);
-
 	return 0;
 }
 
-static int __f2fs_convert_inline_data(struct inode *inode, struct page *page)
+int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page)
 {
-	int err = 0;
-	struct page *ipage;
-	struct dnode_of_data dn;
 	void *src_addr, *dst_addr;
 	block_t new_blk_addr;
-	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct f2fs_io_info fio = {
 		.type = DATA,
 		.rw = WRITE_SYNC | REQ_PRIO,
 	};
+	int dirty, err;
+
+	f2fs_bug_on(F2FS_I_SB(dn->inode), page->index);
+
+	if (!f2fs_exist_data(dn->inode))
+		goto clear_out;
+
+	err = f2fs_reserve_block(dn, 0);
+	if (err)
+		return err;
+
+	f2fs_wait_on_page_writeback(page, DATA);
+
+	if (PageUptodate(page))
+		goto no_update;
+
+	zero_user_segment(page, MAX_INLINE_DATA, PAGE_CACHE_SIZE);
+
+	/* Copy the whole inline data block */
+	src_addr = inline_data_addr(dn->inode_page);
+	dst_addr = kmap_atomic(page);
+	memcpy(dst_addr, src_addr, MAX_INLINE_DATA);
+	flush_dcache_page(page);
+	kunmap_atomic(dst_addr);
+	SetPageUptodate(page);
+no_update:
+	/* clear dirty state */
+	dirty = clear_page_dirty_for_io(page);
+
+	/* write data page to try to make data consistent */
+	set_page_writeback(page);
+
+	write_data_page(page, dn, &new_blk_addr, &fio);
+	update_extent_cache(new_blk_addr, dn);
+	f2fs_wait_on_page_writeback(page, DATA);
+	if (dirty)
+		inode_dec_dirty_pages(dn->inode);
+
+	/* this converted inline_data should be recovered. */
+	set_inode_flag(F2FS_I(dn->inode), FI_APPEND_WRITE);
+
+	/* clear inline data and flag after data writeback */
+	truncate_inline_data(dn->inode_page, 0);
+clear_out:
+	stat_dec_inline_inode(dn->inode);
+	f2fs_clear_inline_inode(dn->inode);
+	sync_inode_page(dn);
+	f2fs_put_dnode(dn);
+	return 0;
+}
+
+int f2fs_convert_inline_inode(struct inode *inode)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct dnode_of_data dn;
+	struct page *ipage, *page;
+	int err = 0;
+
+	page = grab_cache_page(inode->i_mapping, 0);
+	if (!page)
+		return -ENOMEM;
 
 	f2fs_lock_op(sbi);
+
 	ipage = get_node_page(sbi, inode->i_ino);
 	if (IS_ERR(ipage)) {
 		err = PTR_ERR(ipage);
 		goto out;
 	}
 
-	/* someone else converted inline_data already */
+	set_new_dnode(&dn, inode, ipage, ipage, 0);
-	if (!f2fs_has_inline_data(inode))
-		goto out;
 
-	/*
+	if (f2fs_has_inline_data(inode))
-	 * i_addr[0] is not used for inline data,
+		err = f2fs_convert_inline_page(&dn, page);
-	 * so reserving new block will not destroy inline data
-	 */
-	set_new_dnode(&dn, inode, ipage, NULL, 0);
-	err = f2fs_reserve_block(&dn, 0);
-	if (err)
-		goto out;
 
-	f2fs_wait_on_page_writeback(page, DATA);
-	zero_user_segment(page, MAX_INLINE_DATA, PAGE_CACHE_SIZE);
-
-	/* Copy the whole inline data block */
-	src_addr = inline_data_addr(ipage);
-	dst_addr = kmap(page);
-	memcpy(dst_addr, src_addr, MAX_INLINE_DATA);
-	kunmap(page);
-	SetPageUptodate(page);
-
-	/* write data page to try to make data consistent */
-	set_page_writeback(page);
-	write_data_page(page, &dn, &new_blk_addr, &fio);
-	update_extent_cache(new_blk_addr, &dn);
-	f2fs_wait_on_page_writeback(page, DATA);
-
-	/* clear inline data and flag after data writeback */
-	zero_user_segment(ipage, INLINE_DATA_OFFSET,
-				 INLINE_DATA_OFFSET + MAX_INLINE_DATA);
-	clear_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
-	stat_dec_inline_inode(inode);
-
-	sync_inode_page(&dn);
 	f2fs_put_dnode(&dn);
 out:
 	f2fs_unlock_op(sbi);
+
+	f2fs_put_page(page, 1);
 	return err;
 }
 
-int f2fs_convert_inline_data(struct inode *inode, pgoff_t to_size,
+int f2fs_write_inline_data(struct inode *inode, struct page *page)
-						struct page *page)
-{
-	struct page *new_page = page;
-	int err;
-
-	if (!f2fs_has_inline_data(inode))
-		return 0;
-	else if (to_size <= MAX_INLINE_DATA)
-		return 0;
-
-	if (!page || page->index != 0) {
-		new_page = grab_cache_page(inode->i_mapping, 0);
-		if (!new_page)
-			return -ENOMEM;
-	}
-
-	err = __f2fs_convert_inline_data(inode, new_page);
-	if (!page || page->index != 0)
-		f2fs_put_page(new_page, 1);
-	return err;
-}
-
-int f2fs_write_inline_data(struct inode *inode,
-				struct page *page, unsigned size)
 {
 	void *src_addr, *dst_addr;
-	struct page *ipage;
 	struct dnode_of_data dn;
 	int err;
 
@@ -164,47 +177,39 @@ int f2fs_write_inline_data(struct inode *inode,
 	err = get_dnode_of_data(&dn, 0, LOOKUP_NODE);
 	if (err)
 		return err;
-	ipage = dn.inode_page;
 
-	f2fs_wait_on_page_writeback(ipage, NODE);
-	zero_user_segment(ipage, INLINE_DATA_OFFSET,
-				 INLINE_DATA_OFFSET + MAX_INLINE_DATA);
-	src_addr = kmap(page);
-	dst_addr = inline_data_addr(ipage);
-	memcpy(dst_addr, src_addr, size);
-	kunmap(page);
-
-	/* Release the first data block if it is allocated */
 	if (!f2fs_has_inline_data(inode)) {
-		truncate_data_blocks_range(&dn, 1);
+		f2fs_put_dnode(&dn);
-		set_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
+		return -EAGAIN;
-		stat_inc_inline_inode(inode);
 	}
 
+	f2fs_bug_on(F2FS_I_SB(inode), page->index);
+
+	f2fs_wait_on_page_writeback(dn.inode_page, NODE);
+	src_addr = kmap_atomic(page);
+	dst_addr = inline_data_addr(dn.inode_page);
+	memcpy(dst_addr, src_addr, MAX_INLINE_DATA);
+	kunmap_atomic(src_addr);
+
 	set_inode_flag(F2FS_I(inode), FI_APPEND_WRITE);
+	set_inode_flag(F2FS_I(inode), FI_DATA_EXIST);
+
 	sync_inode_page(&dn);
 	f2fs_put_dnode(&dn);
-
 	return 0;
 }
 
-void truncate_inline_data(struct inode *inode, u64 from)
+void truncate_inline_data(struct page *ipage, u64 from)
 {
-	struct page *ipage;
+	void *addr;
 
 	if (from >= MAX_INLINE_DATA)
 		return;
 
-	ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino);
-	if (IS_ERR(ipage))
-		return;
-
 	f2fs_wait_on_page_writeback(ipage, NODE);
 
-	zero_user_segment(ipage, INLINE_DATA_OFFSET + from,
+	addr = inline_data_addr(ipage);
-				INLINE_DATA_OFFSET + MAX_INLINE_DATA);
+	memset(addr + from, 0, MAX_INLINE_DATA - from);
-	set_page_dirty(ipage);
-	f2fs_put_page(ipage, 1);
 }
 
 bool recover_inline_data(struct inode *inode, struct page *npage)
@@ -236,6 +241,10 @@ process_inline:
 		src_addr = inline_data_addr(npage);
 		dst_addr = inline_data_addr(ipage);
 		memcpy(dst_addr, src_addr, MAX_INLINE_DATA);
+
+		set_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
+		set_inode_flag(F2FS_I(inode), FI_DATA_EXIST);
+
 		update_inode(inode, ipage);
 		f2fs_put_page(ipage, 1);
 		return true;
@@ -244,16 +253,279 @@ process_inline:
 	if (f2fs_has_inline_data(inode)) {
 		ipage = get_node_page(sbi, inode->i_ino);
 		f2fs_bug_on(sbi, IS_ERR(ipage));
-		f2fs_wait_on_page_writeback(ipage, NODE);
+		truncate_inline_data(ipage, 0);
-		zero_user_segment(ipage, INLINE_DATA_OFFSET,
+		f2fs_clear_inline_inode(inode);
-				 INLINE_DATA_OFFSET + MAX_INLINE_DATA);
-		clear_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
 		update_inode(inode, ipage);
 		f2fs_put_page(ipage, 1);
 	} else if (ri && (ri->i_inline & F2FS_INLINE_DATA)) {
 		truncate_blocks(inode, 0, false);
-		set_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
 		goto process_inline;
 	}
 	return false;
 }
+
+struct f2fs_dir_entry *find_in_inline_dir(struct inode *dir,
+				struct qstr *name, struct page **res_page)
+{
+	struct f2fs_sb_info *sbi = F2FS_SB(dir->i_sb);
+	struct f2fs_inline_dentry *inline_dentry;
+	struct f2fs_dir_entry *de;
+	struct f2fs_dentry_ptr d;
+	struct page *ipage;
+
+	ipage = get_node_page(sbi, dir->i_ino);
+	if (IS_ERR(ipage))
+		return NULL;
+
+	inline_dentry = inline_data_addr(ipage);
+
+	make_dentry_ptr(&d, (void *)inline_dentry, 2);
+	de = find_target_dentry(name, NULL, &d);
+
+	unlock_page(ipage);
+	if (de)
+		*res_page = ipage;
+	else
+		f2fs_put_page(ipage, 0);
+
+	/*
+	 * For the most part, it should be a bug when name_len is zero.
+	 * We stop here for figuring out where the bugs has occurred.
+	 */
+	f2fs_bug_on(sbi, d.max < 0);
+	return de;
+}
+
+struct f2fs_dir_entry *f2fs_parent_inline_dir(struct inode *dir,
+							struct page **p)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
+	struct page *ipage;
+	struct f2fs_dir_entry *de;
+	struct f2fs_inline_dentry *dentry_blk;
+
+	ipage = get_node_page(sbi, dir->i_ino);
+	if (IS_ERR(ipage))
+		return NULL;
+
+	dentry_blk = inline_data_addr(ipage);
+	de = &dentry_blk->dentry[1];
+	*p = ipage;
+	unlock_page(ipage);
+	return de;
+}
+
+int make_empty_inline_dir(struct inode *inode, struct inode *parent,
+							struct page *ipage)
+{
+	struct f2fs_inline_dentry *dentry_blk;
+	struct f2fs_dentry_ptr d;
+
+	dentry_blk = inline_data_addr(ipage);
+
+	make_dentry_ptr(&d, (void *)dentry_blk, 2);
+	do_make_empty_dir(inode, parent, &d);
+
+	set_page_dirty(ipage);
+
+	/* update i_size to MAX_INLINE_DATA */
+	if (i_size_read(inode) < MAX_INLINE_DATA) {
+		i_size_write(inode, MAX_INLINE_DATA);
+		set_inode_flag(F2FS_I(inode), FI_UPDATE_DIR);
+	}
+	return 0;
+}
+
+static int f2fs_convert_inline_dir(struct inode *dir, struct page *ipage,
+				struct f2fs_inline_dentry *inline_dentry)
+{
+	struct page *page;
+	struct dnode_of_data dn;
+	struct f2fs_dentry_block *dentry_blk;
+	int err;
+
+	page = grab_cache_page(dir->i_mapping, 0);
+	if (!page)
+		return -ENOMEM;
+
+	set_new_dnode(&dn, dir, ipage, NULL, 0);
+	err = f2fs_reserve_block(&dn, 0);
+	if (err)
+		goto out;
+
+	f2fs_wait_on_page_writeback(page, DATA);
+	zero_user_segment(page, 0, PAGE_CACHE_SIZE);
+
+	dentry_blk = kmap_atomic(page);
+
+	/* copy data from inline dentry block to new dentry block */
+	memcpy(dentry_blk->dentry_bitmap, inline_dentry->dentry_bitmap,
+					INLINE_DENTRY_BITMAP_SIZE);
+	memcpy(dentry_blk->dentry, inline_dentry->dentry,
+			sizeof(struct f2fs_dir_entry) * NR_INLINE_DENTRY);
+	memcpy(dentry_blk->filename, inline_dentry->filename,
+					NR_INLINE_DENTRY * F2FS_SLOT_LEN);
+
+	kunmap_atomic(dentry_blk);
+	SetPageUptodate(page);
+	set_page_dirty(page);
+
+	/* clear inline dir and flag after data writeback */
+	truncate_inline_data(ipage, 0);
+
+	stat_dec_inline_dir(dir);
+	clear_inode_flag(F2FS_I(dir), FI_INLINE_DENTRY);
+
+	if (i_size_read(dir) < PAGE_CACHE_SIZE) {
+		i_size_write(dir, PAGE_CACHE_SIZE);
+		set_inode_flag(F2FS_I(dir), FI_UPDATE_DIR);
+	}
+
+	sync_inode_page(&dn);
+out:
+	f2fs_put_page(page, 1);
+	return err;
+}
+
+int f2fs_add_inline_entry(struct inode *dir, const struct qstr *name,
+						struct inode *inode)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
+	struct page *ipage;
+	unsigned int bit_pos;
+	f2fs_hash_t name_hash;
+	struct f2fs_dir_entry *de;
+	size_t namelen = name->len;
+	struct f2fs_inline_dentry *dentry_blk = NULL;
+	int slots = GET_DENTRY_SLOTS(namelen);
+	struct page *page;
+	int err = 0;
+	int i;
+
+	name_hash = f2fs_dentry_hash(name);
+
+	ipage = get_node_page(sbi, dir->i_ino);
+	if (IS_ERR(ipage))
+		return PTR_ERR(ipage);
+
+	dentry_blk = inline_data_addr(ipage);
+	bit_pos = room_for_filename(&dentry_blk->dentry_bitmap,
+						slots, NR_INLINE_DENTRY);
+	if (bit_pos >= NR_INLINE_DENTRY) {
+		err = f2fs_convert_inline_dir(dir, ipage, dentry_blk);
+		if (!err)
+			err = -EAGAIN;
+		goto out;
+	}
+
+	down_write(&F2FS_I(inode)->i_sem);
+	page = init_inode_metadata(inode, dir, name, ipage);
+	if (IS_ERR(page)) {
+		err = PTR_ERR(page);
+		goto fail;
+	}
+
+	f2fs_wait_on_page_writeback(ipage, NODE);
+	de = &dentry_blk->dentry[bit_pos];
+	de->hash_code = name_hash;
+	de->name_len = cpu_to_le16(namelen);
+	memcpy(dentry_blk->filename[bit_pos], name->name, name->len);
+	de->ino = cpu_to_le32(inode->i_ino);
+	set_de_type(de, inode);
+	for (i = 0; i < slots; i++)
+		test_and_set_bit_le(bit_pos + i, &dentry_blk->dentry_bitmap);
+	set_page_dirty(ipage);
+
+	/* we don't need to mark_inode_dirty now */
+	F2FS_I(inode)->i_pino = dir->i_ino;
+	update_inode(inode, page);
+	f2fs_put_page(page, 1);
+
+	update_parent_metadata(dir, inode, 0);
+fail:
+	up_write(&F2FS_I(inode)->i_sem);
+
+	if (is_inode_flag_set(F2FS_I(dir), FI_UPDATE_DIR)) {
+		update_inode(dir, ipage);
+		clear_inode_flag(F2FS_I(dir), FI_UPDATE_DIR);
+	}
+out:
+	f2fs_put_page(ipage, 1);
+	return err;
+}
+
+void f2fs_delete_inline_entry(struct f2fs_dir_entry *dentry, struct page *page,
+					struct inode *dir, struct inode *inode)
+{
+	struct f2fs_inline_dentry *inline_dentry;
+	int slots = GET_DENTRY_SLOTS(le16_to_cpu(dentry->name_len));
+	unsigned int bit_pos;
+	int i;
+
+	lock_page(page);
+	f2fs_wait_on_page_writeback(page, NODE);
+
+	inline_dentry = inline_data_addr(page);
+	bit_pos = dentry - inline_dentry->dentry;
+	for (i = 0; i < slots; i++)
+		test_and_clear_bit_le(bit_pos + i,
+				&inline_dentry->dentry_bitmap);
+
+	set_page_dirty(page);
+
+	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
+
+	if (inode)
+		f2fs_drop_nlink(dir, inode, page);
+
+	f2fs_put_page(page, 1);
+}
+
+bool f2fs_empty_inline_dir(struct inode *dir)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
+	struct page *ipage;
+	unsigned int bit_pos = 2;
+	struct f2fs_inline_dentry *dentry_blk;
+
+	ipage = get_node_page(sbi, dir->i_ino);
+	if (IS_ERR(ipage))
+		return false;
+
+	dentry_blk = inline_data_addr(ipage);
+	bit_pos = find_next_bit_le(&dentry_blk->dentry_bitmap,
+					NR_INLINE_DENTRY,
+					bit_pos);
+
+	f2fs_put_page(ipage, 1);
+
+	if (bit_pos < NR_INLINE_DENTRY)
+		return false;
+
+	return true;
+}
+
+int f2fs_read_inline_dir(struct file *file, struct dir_context *ctx)
+{
+	struct inode *inode = file_inode(file);
+	struct f2fs_inline_dentry *inline_dentry = NULL;
+	struct page *ipage = NULL;
+	struct f2fs_dentry_ptr d;
+
+	if (ctx->pos == NR_INLINE_DENTRY)
+		return 0;
+
+	ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino);
+	if (IS_ERR(ipage))
+		return PTR_ERR(ipage);
+
+	inline_dentry = inline_data_addr(ipage);
+
+	make_dentry_ptr(&d, (void *)inline_dentry, 2);
+
+	if (!f2fs_fill_dentries(ctx, &d, 0))
+		ctx->pos = NR_INLINE_DENTRY;
+
+	f2fs_put_page(ipage, 1);
+	return 0;
+}

fs/f2fs/inode.c

@@ -67,12 +67,38 @@ static void __set_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
 	}
 }
 
+static int __recover_inline_status(struct inode *inode, struct page *ipage)
+{
+	void *inline_data = inline_data_addr(ipage);
+	struct f2fs_inode *ri;
+	void *zbuf;
+
+	zbuf = kzalloc(MAX_INLINE_DATA, GFP_NOFS);
+	if (!zbuf)
+		return -ENOMEM;
+
+	if (!memcmp(zbuf, inline_data, MAX_INLINE_DATA)) {
+		kfree(zbuf);
+		return 0;
+	}
+	kfree(zbuf);
+
+	f2fs_wait_on_page_writeback(ipage, NODE);
+	set_inode_flag(F2FS_I(inode), FI_DATA_EXIST);
+
+	ri = F2FS_INODE(ipage);
+	set_raw_inline(F2FS_I(inode), ri);
+	set_page_dirty(ipage);
+	return 0;
+}
+
 static int do_read_inode(struct inode *inode)
 {
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct f2fs_inode_info *fi = F2FS_I(inode);
 	struct page *node_page;
 	struct f2fs_inode *ri;
+	int err = 0;
 
 	/* Check if ino is within scope */
 	if (check_nid_range(sbi, inode->i_ino)) {
@@ -114,11 +140,19 @@ static int do_read_inode(struct inode *inode)
 	get_extent_info(&fi->ext, ri->i_ext);
 	get_inline_info(fi, ri);
 
+	/* check data exist */
+	if (f2fs_has_inline_data(inode) && !f2fs_exist_data(inode))
+		err = __recover_inline_status(inode, node_page);
+
 	/* get rdev by using inline_info */
 	__get_inode_rdev(inode, ri);
 
 	f2fs_put_page(node_page, 1);
-	return 0;
+
+	stat_inc_inline_inode(inode);
+	stat_inc_inline_dir(inode);
+
+	return err;
 }
 
 struct inode *f2fs_iget(struct super_block *sb, unsigned long ino)
@@ -156,7 +190,7 @@ make_now:
 		inode->i_op = &f2fs_dir_inode_operations;
 		inode->i_fop = &f2fs_dir_operations;
 		inode->i_mapping->a_ops = &f2fs_dblock_aops;
-		mapping_set_gfp_mask(inode->i_mapping, GFP_F2FS_ZERO);
+		mapping_set_gfp_mask(inode->i_mapping, GFP_F2FS_HIGH_ZERO);
 	} else if (S_ISLNK(inode->i_mode)) {
 		inode->i_op = &f2fs_symlink_inode_operations;
 		inode->i_mapping->a_ops = &f2fs_dblock_aops;
@@ -295,11 +329,12 @@ void f2fs_evict_inode(struct inode *inode)
 
 	f2fs_lock_op(sbi);
 	remove_inode_page(inode);
-	stat_dec_inline_inode(inode);
 	f2fs_unlock_op(sbi);
 
 	sb_end_intwrite(inode->i_sb);
 no_delete:
+	stat_dec_inline_dir(inode);
+	stat_dec_inline_inode(inode);
 	invalidate_mapping_pages(NODE_MAPPING(sbi), inode->i_ino, inode->i_ino);
 	if (xnid)
 		invalidate_mapping_pages(NODE_MAPPING(sbi), xnid, xnid);
@@ -325,8 +360,9 @@ void handle_failed_inode(struct inode *inode)
 		f2fs_truncate(inode);
 
 	remove_inode_page(inode);
-	stat_dec_inline_inode(inode);
 
+	clear_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
+	clear_inode_flag(F2FS_I(inode), FI_INLINE_DENTRY);
 	alloc_nid_failed(sbi, inode->i_ino);
 	f2fs_unlock_op(sbi);
 

fs/f2fs/namei.c

@@ -54,6 +54,12 @@ static struct inode *f2fs_new_inode(struct inode *dir, umode_t mode)
 		nid_free = true;
 		goto out;
 	}
+
+	if (f2fs_may_inline(inode))
+		set_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
+	if (test_opt(sbi, INLINE_DENTRY) && S_ISDIR(inode->i_mode))
+		set_inode_flag(F2FS_I(inode), FI_INLINE_DENTRY);
+
 	trace_f2fs_new_inode(inode, 0);
 	mark_inode_dirty(inode);
 	return inode;
@@ -129,8 +135,12 @@ static int f2fs_create(struct inode *dir, struct dentry *dentry, umode_t mode,
 
 	alloc_nid_done(sbi, ino);
 
+	stat_inc_inline_inode(inode);
 	d_instantiate(dentry, inode);
 	unlock_new_inode(inode);
+
+	if (IS_DIRSYNC(dir))
+		f2fs_sync_fs(sbi->sb, 1);
 	return 0;
 out:
 	handle_failed_inode(inode);
@@ -157,6 +167,9 @@ static int f2fs_link(struct dentry *old_dentry, struct inode *dir,
 	f2fs_unlock_op(sbi);
 
 	d_instantiate(dentry, inode);
+
+	if (IS_DIRSYNC(dir))
+		f2fs_sync_fs(sbi->sb, 1);
 	return 0;
 out:
 	clear_inode_flag(F2FS_I(inode), FI_INC_LINK);
@@ -187,14 +200,12 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry,
 	de = f2fs_find_entry(dir, &dentry->d_name, &page);
 	if (de) {
 		nid_t ino = le32_to_cpu(de->ino);
-		kunmap(page);
+		f2fs_dentry_kunmap(dir, page);
 		f2fs_put_page(page, 0);
 
 		inode = f2fs_iget(dir->i_sb, ino);
 		if (IS_ERR(inode))
 			return ERR_CAST(inode);
-
-		stat_inc_inline_inode(inode);
 	}
 
 	return d_splice_alias(inode, dentry);
@@ -219,15 +230,18 @@ static int f2fs_unlink(struct inode *dir, struct dentry *dentry)
 	err = acquire_orphan_inode(sbi);
 	if (err) {
 		f2fs_unlock_op(sbi);
-		kunmap(page);
+		f2fs_dentry_kunmap(dir, page);
 		f2fs_put_page(page, 0);
 		goto fail;
 	}
-	f2fs_delete_entry(de, page, inode);
+	f2fs_delete_entry(de, page, dir, inode);
 	f2fs_unlock_op(sbi);
 
 	/* In order to evict this inode, we set it dirty */
 	mark_inode_dirty(inode);
+
+	if (IS_DIRSYNC(dir))
+		f2fs_sync_fs(sbi->sb, 1);
 fail:
 	trace_f2fs_unlink_exit(inode, err);
 	return err;
@@ -261,6 +275,9 @@ static int f2fs_symlink(struct inode *dir, struct dentry *dentry,
 
 	d_instantiate(dentry, inode);
 	unlock_new_inode(inode);
+
+	if (IS_DIRSYNC(dir))
+		f2fs_sync_fs(sbi->sb, 1);
 	return err;
 out:
 	handle_failed_inode(inode);
@@ -291,11 +308,14 @@ static int f2fs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
 		goto out_fail;
 	f2fs_unlock_op(sbi);
 
+	stat_inc_inline_dir(inode);
 	alloc_nid_done(sbi, inode->i_ino);
 
 	d_instantiate(dentry, inode);
 	unlock_new_inode(inode);
 
+	if (IS_DIRSYNC(dir))
+		f2fs_sync_fs(sbi->sb, 1);
 	return 0;
 
 out_fail:
@@ -338,8 +358,12 @@ static int f2fs_mknod(struct inode *dir, struct dentry *dentry,
 	f2fs_unlock_op(sbi);
 
 	alloc_nid_done(sbi, inode->i_ino);
+
 	d_instantiate(dentry, inode);
 	unlock_new_inode(inode);
+
+	if (IS_DIRSYNC(dir))
+		f2fs_sync_fs(sbi->sb, 1);
 	return 0;
 out:
 	handle_failed_inode(inode);
@@ -435,7 +459,7 @@ static int f2fs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	old_inode->i_ctime = CURRENT_TIME;
 	mark_inode_dirty(old_inode);
 
-	f2fs_delete_entry(old_entry, old_page, NULL);
+	f2fs_delete_entry(old_entry, old_page, old_dir, NULL);
 
 	if (old_dir_entry) {
 		if (old_dir != new_dir) {
@@ -443,7 +467,7 @@ static int f2fs_rename(struct inode *old_dir, struct dentry *old_dentry,
 						old_dir_page, new_dir);
 			update_inode_page(old_inode);
 		} else {
-			kunmap(old_dir_page);
+			f2fs_dentry_kunmap(old_inode, old_dir_page);
 			f2fs_put_page(old_dir_page, 0);
 		}
 		drop_nlink(old_dir);
@@ -452,19 +476,22 @@ static int f2fs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	}
 
 	f2fs_unlock_op(sbi);
+
+	if (IS_DIRSYNC(old_dir) || IS_DIRSYNC(new_dir))
+		f2fs_sync_fs(sbi->sb, 1);
 	return 0;
 
 put_out_dir:
 	f2fs_unlock_op(sbi);
-	kunmap(new_page);
+	f2fs_dentry_kunmap(new_dir, new_page);
 	f2fs_put_page(new_page, 0);
 out_dir:
 	if (old_dir_entry) {
-		kunmap(old_dir_page);
+		f2fs_dentry_kunmap(old_inode, old_dir_page);
 		f2fs_put_page(old_dir_page, 0);
 	}
 out_old:
-	kunmap(old_page);
+	f2fs_dentry_kunmap(old_dir, old_page);
 	f2fs_put_page(old_page, 0);
 out:
 	return err;
@@ -588,6 +615,9 @@ static int f2fs_cross_rename(struct inode *old_dir, struct dentry *old_dentry,
 	update_inode_page(new_dir);
 
 	f2fs_unlock_op(sbi);
+
+	if (IS_DIRSYNC(old_dir) || IS_DIRSYNC(new_dir))
+		f2fs_sync_fs(sbi->sb, 1);
 	return 0;
 out_undo:
 	/* Still we may fail to recover name info of f2fs_inode here */
@@ -596,19 +626,19 @@ out_unlock:
 	f2fs_unlock_op(sbi);
 out_new_dir:
 	if (new_dir_entry) {
-		kunmap(new_dir_page);
+		f2fs_dentry_kunmap(new_inode, new_dir_page);
 		f2fs_put_page(new_dir_page, 0);
 	}
 out_old_dir:
 	if (old_dir_entry) {
-		kunmap(old_dir_page);
+		f2fs_dentry_kunmap(old_inode, old_dir_page);
 		f2fs_put_page(old_dir_page, 0);
 	}
 out_new:
-	kunmap(new_page);
+	f2fs_dentry_kunmap(new_dir, new_page);
 	f2fs_put_page(new_page, 0);
 out_old:
-	kunmap(old_page);
+	f2fs_dentry_kunmap(old_dir, old_page);
 	f2fs_put_page(old_page, 0);
 out:
 	return err;

fs/f2fs/node.c

@@ -31,22 +31,38 @@ bool available_free_memory(struct f2fs_sb_info *sbi, int type)
 {
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
 	struct sysinfo val;
+	unsigned long avail_ram;
 	unsigned long mem_size = 0;
 	bool res = false;
 
 	si_meminfo(&val);
-	/* give 25%, 25%, 50% memory for each components respectively */
+
+	/* only uses low memory */
+	avail_ram = val.totalram - val.totalhigh;
+
+	/* give 25%, 25%, 50%, 50% memory for each components respectively */
 	if (type == FREE_NIDS) {
-		mem_size = (nm_i->fcnt * sizeof(struct free_nid)) >> 12;
+		mem_size = (nm_i->fcnt * sizeof(struct free_nid)) >>
-		res = mem_size < ((val.totalram * nm_i->ram_thresh / 100) >> 2);
+							PAGE_CACHE_SHIFT;
+		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
 	} else if (type == NAT_ENTRIES) {
-		mem_size = (nm_i->nat_cnt * sizeof(struct nat_entry)) >> 12;
+		mem_size = (nm_i->nat_cnt * sizeof(struct nat_entry)) >>
-		res = mem_size < ((val.totalram * nm_i->ram_thresh / 100) >> 2);
+							PAGE_CACHE_SHIFT;
+		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
 	} else if (type == DIRTY_DENTS) {
 		if (sbi->sb->s_bdi->dirty_exceeded)
 			return false;
 		mem_size = get_pages(sbi, F2FS_DIRTY_DENTS);
-		res = mem_size < ((val.totalram * nm_i->ram_thresh / 100) >> 1);
+		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
+	} else if (type == INO_ENTRIES) {
+		int i;
+
+		if (sbi->sb->s_bdi->dirty_exceeded)
+			return false;
+		for (i = 0; i <= UPDATE_INO; i++)
+			mem_size += (sbi->im[i].ino_num *
+				sizeof(struct ino_entry)) >> PAGE_CACHE_SHIFT;
+		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
 	}
 	return res;
 }
@@ -131,7 +147,7 @@ static void __set_nat_cache_dirty(struct f2fs_nm_info *nm_i,
 
 	if (get_nat_flag(ne, IS_DIRTY))
 		return;
-retry:
+
 	head = radix_tree_lookup(&nm_i->nat_set_root, set);
 	if (!head) {
 		head = f2fs_kmem_cache_alloc(nat_entry_set_slab, GFP_ATOMIC);
@@ -140,11 +156,7 @@ retry:
 		INIT_LIST_HEAD(&head->set_list);
 		head->set = set;
 		head->entry_cnt = 0;
-
+		f2fs_radix_tree_insert(&nm_i->nat_set_root, set, head);
-		if (radix_tree_insert(&nm_i->nat_set_root, set, head)) {
-			cond_resched();
-			goto retry;
-		}
 	}
 	list_move_tail(&ne->list, &head->entry_list);
 	nm_i->dirty_nat_cnt++;
@@ -155,7 +167,7 @@ retry:
 static void __clear_nat_cache_dirty(struct f2fs_nm_info *nm_i,
 						struct nat_entry *ne)
 {
-	nid_t set = ne->ni.nid / NAT_ENTRY_PER_BLOCK;
+	nid_t set = NAT_BLOCK_OFFSET(ne->ni.nid);
 	struct nat_entry_set *head;
 
 	head = radix_tree_lookup(&nm_i->nat_set_root, set);
@@ -180,11 +192,11 @@ bool is_checkpointed_node(struct f2fs_sb_info *sbi, nid_t nid)
 	struct nat_entry *e;
 	bool is_cp = true;
 
-	read_lock(&nm_i->nat_tree_lock);
+	down_read(&nm_i->nat_tree_lock);
 	e = __lookup_nat_cache(nm_i, nid);
 	if (e && !get_nat_flag(e, IS_CHECKPOINTED))
 		is_cp = false;
-	read_unlock(&nm_i->nat_tree_lock);
+	up_read(&nm_i->nat_tree_lock);
 	return is_cp;
 }
 
@@ -194,11 +206,11 @@ bool has_fsynced_inode(struct f2fs_sb_info *sbi, nid_t ino)
 	struct nat_entry *e;
 	bool fsynced = false;
 
-	read_lock(&nm_i->nat_tree_lock);
+	down_read(&nm_i->nat_tree_lock);
 	e = __lookup_nat_cache(nm_i, ino);
 	if (e && get_nat_flag(e, HAS_FSYNCED_INODE))
 		fsynced = true;
-	read_unlock(&nm_i->nat_tree_lock);
+	up_read(&nm_i->nat_tree_lock);
 	return fsynced;
 }
 
@@ -208,13 +220,13 @@ bool need_inode_block_update(struct f2fs_sb_info *sbi, nid_t ino)
 	struct nat_entry *e;
 	bool need_update = true;
 
-	read_lock(&nm_i->nat_tree_lock);
+	down_read(&nm_i->nat_tree_lock);
 	e = __lookup_nat_cache(nm_i, ino);
 	if (e && get_nat_flag(e, HAS_LAST_FSYNC) &&
 			(get_nat_flag(e, IS_CHECKPOINTED) ||
 			 get_nat_flag(e, HAS_FSYNCED_INODE)))
 		need_update = false;
-	read_unlock(&nm_i->nat_tree_lock);
+	up_read(&nm_i->nat_tree_lock);
 	return need_update;
 }
 
@@ -222,13 +234,8 @@ static struct nat_entry *grab_nat_entry(struct f2fs_nm_info *nm_i, nid_t nid)
 {
 	struct nat_entry *new;
 
-	new = kmem_cache_alloc(nat_entry_slab, GFP_ATOMIC);
+	new = f2fs_kmem_cache_alloc(nat_entry_slab, GFP_ATOMIC);
-	if (!new)
+	f2fs_radix_tree_insert(&nm_i->nat_root, nid, new);
-		return NULL;
-	if (radix_tree_insert(&nm_i->nat_root, nid, new)) {
-		kmem_cache_free(nat_entry_slab, new);
-		return NULL;
-	}
 	memset(new, 0, sizeof(struct nat_entry));
 	nat_set_nid(new, nid);
 	nat_reset_flag(new);
@@ -241,18 +248,14 @@ static void cache_nat_entry(struct f2fs_nm_info *nm_i, nid_t nid,
 						struct f2fs_nat_entry *ne)
 {
 	struct nat_entry *e;
-retry:
+
-	write_lock(&nm_i->nat_tree_lock);
+	down_write(&nm_i->nat_tree_lock);
 	e = __lookup_nat_cache(nm_i, nid);
 	if (!e) {
 		e = grab_nat_entry(nm_i, nid);
-		if (!e) {
-			write_unlock(&nm_i->nat_tree_lock);
-			goto retry;
-		}
 		node_info_from_raw_nat(&e->ni, ne);
 	}
-	write_unlock(&nm_i->nat_tree_lock);
+	up_write(&nm_i->nat_tree_lock);
 }
 
 static void set_node_addr(struct f2fs_sb_info *sbi, struct node_info *ni,
@@ -260,15 +263,11 @@ static void set_node_addr(struct f2fs_sb_info *sbi, struct node_info *ni,
 {
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
 	struct nat_entry *e;
-retry:
+
-	write_lock(&nm_i->nat_tree_lock);
+	down_write(&nm_i->nat_tree_lock);
 	e = __lookup_nat_cache(nm_i, ni->nid);
 	if (!e) {
 		e = grab_nat_entry(nm_i, ni->nid);
-		if (!e) {
-			write_unlock(&nm_i->nat_tree_lock);
-			goto retry;
-		}
 		e->ni = *ni;
 		f2fs_bug_on(sbi, ni->blk_addr == NEW_ADDR);
 	} else if (new_blkaddr == NEW_ADDR) {
@@ -310,7 +309,7 @@ retry:
 			set_nat_flag(e, HAS_FSYNCED_INODE, true);
 		set_nat_flag(e, HAS_LAST_FSYNC, fsync_done);
 	}
-	write_unlock(&nm_i->nat_tree_lock);
+	up_write(&nm_i->nat_tree_lock);
 }
 
 int try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
@@ -320,7 +319,7 @@ int try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
 	if (available_free_memory(sbi, NAT_ENTRIES))
 		return 0;
 
-	write_lock(&nm_i->nat_tree_lock);
+	down_write(&nm_i->nat_tree_lock);
 	while (nr_shrink && !list_empty(&nm_i->nat_entries)) {
 		struct nat_entry *ne;
 		ne = list_first_entry(&nm_i->nat_entries,
@@ -328,7 +327,7 @@ int try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
 		__del_from_nat_cache(nm_i, ne);
 		nr_shrink--;
 	}
-	write_unlock(&nm_i->nat_tree_lock);
+	up_write(&nm_i->nat_tree_lock);
 	return nr_shrink;
 }
 
@@ -351,14 +350,14 @@ void get_node_info(struct f2fs_sb_info *sbi, nid_t nid, struct node_info *ni)
 	ni->nid = nid;
 
 	/* Check nat cache */
-	read_lock(&nm_i->nat_tree_lock);
+	down_read(&nm_i->nat_tree_lock);
 	e = __lookup_nat_cache(nm_i, nid);
 	if (e) {
 		ni->ino = nat_get_ino(e);
 		ni->blk_addr = nat_get_blkaddr(e);
 		ni->version = nat_get_version(e);
 	}
-	read_unlock(&nm_i->nat_tree_lock);
+	up_read(&nm_i->nat_tree_lock);
 	if (e)
 		return;
 
@@ -1298,16 +1297,22 @@ static int f2fs_write_node_page(struct page *page,
 		return 0;
 	}
 
-	if (wbc->for_reclaim)
+	if (wbc->for_reclaim) {
-		goto redirty_out;
+		if (!down_read_trylock(&sbi->node_write))
-
+			goto redirty_out;
-	down_read(&sbi->node_write);
+	} else {
+		down_read(&sbi->node_write);
+	}
 	set_page_writeback(page);
 	write_node_page(sbi, page, &fio, nid, ni.blk_addr, &new_addr);
 	set_node_addr(sbi, &ni, new_addr, is_fsync_dnode(page));
 	dec_page_count(sbi, F2FS_DIRTY_NODES);
 	up_read(&sbi->node_write);
 	unlock_page(page);
+
+	if (wbc->for_reclaim)
+		f2fs_submit_merged_bio(sbi, NODE, WRITE);
+
 	return 0;
 
 redirty_out:
@@ -1410,13 +1415,13 @@ static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build)
 
 	if (build) {
 		/* do not add allocated nids */
-		read_lock(&nm_i->nat_tree_lock);
+		down_read(&nm_i->nat_tree_lock);
 		ne = __lookup_nat_cache(nm_i, nid);
 		if (ne &&
 			(!get_nat_flag(ne, IS_CHECKPOINTED) ||
 				nat_get_blkaddr(ne) != NULL_ADDR))
 			allocated = true;
-		read_unlock(&nm_i->nat_tree_lock);
+		up_read(&nm_i->nat_tree_lock);
 		if (allocated)
 			return 0;
 	}
@@ -1425,15 +1430,22 @@ static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build)
 	i->nid = nid;
 	i->state = NID_NEW;
 
+	if (radix_tree_preload(GFP_NOFS)) {
+		kmem_cache_free(free_nid_slab, i);
+		return 0;
+	}
+
 	spin_lock(&nm_i->free_nid_list_lock);
 	if (radix_tree_insert(&nm_i->free_nid_root, i->nid, i)) {
 		spin_unlock(&nm_i->free_nid_list_lock);
+		radix_tree_preload_end();
 		kmem_cache_free(free_nid_slab, i);
 		return 0;
 	}
 	list_add_tail(&i->list, &nm_i->free_nid_list);
 	nm_i->fcnt++;
 	spin_unlock(&nm_i->free_nid_list_lock);
+	radix_tree_preload_end();
 	return 1;
 }
 
@@ -1804,21 +1816,15 @@ static void remove_nats_in_journal(struct f2fs_sb_info *sbi)
 		nid_t nid = le32_to_cpu(nid_in_journal(sum, i));
 
 		raw_ne = nat_in_journal(sum, i);
-retry:
-		write_lock(&nm_i->nat_tree_lock);
-		ne = __lookup_nat_cache(nm_i, nid);
-		if (ne)
-			goto found;
 
-		ne = grab_nat_entry(nm_i, nid);
+		down_write(&nm_i->nat_tree_lock);
+		ne = __lookup_nat_cache(nm_i, nid);
 		if (!ne) {
-			write_unlock(&nm_i->nat_tree_lock);
+			ne = grab_nat_entry(nm_i, nid);
-			goto retry;
+			node_info_from_raw_nat(&ne->ni, &raw_ne);
 		}
-		node_info_from_raw_nat(&ne->ni, &raw_ne);
-found:
 		__set_nat_cache_dirty(nm_i, ne);
-		write_unlock(&nm_i->nat_tree_lock);
+		up_write(&nm_i->nat_tree_lock);
 	}
 	update_nats_in_cursum(sum, -i);
 	mutex_unlock(&curseg->curseg_mutex);
@@ -1889,10 +1895,10 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi,
 		}
 		raw_nat_from_node_info(raw_ne, &ne->ni);
 
-		write_lock(&NM_I(sbi)->nat_tree_lock);
+		down_write(&NM_I(sbi)->nat_tree_lock);
 		nat_reset_flag(ne);
 		__clear_nat_cache_dirty(NM_I(sbi), ne);
-		write_unlock(&NM_I(sbi)->nat_tree_lock);
+		up_write(&NM_I(sbi)->nat_tree_lock);
 
 		if (nat_get_blkaddr(ne) == NULL_ADDR)
 			add_free_nid(sbi, nid, false);
@@ -1903,10 +1909,10 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi,
 	else
 		f2fs_put_page(page, 1);
 
-	if (!set->entry_cnt) {
+	f2fs_bug_on(sbi, set->entry_cnt);
-		radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set);
+
-		kmem_cache_free(nat_entry_set_slab, set);
+	radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set);
-	}
+	kmem_cache_free(nat_entry_set_slab, set);
 }
 
 /*
@@ -1923,6 +1929,8 @@ void flush_nat_entries(struct f2fs_sb_info *sbi)
 	nid_t set_idx = 0;
 	LIST_HEAD(sets);
 
+	if (!nm_i->dirty_nat_cnt)
+		return;
 	/*
 	 * if there are no enough space in journal to store dirty nat
 	 * entries, remove all entries from journal and merge them
@@ -1931,9 +1939,6 @@ void flush_nat_entries(struct f2fs_sb_info *sbi)
 	if (!__has_cursum_space(sum, nm_i->dirty_nat_cnt, NAT_JOURNAL))
 		remove_nats_in_journal(sbi);
 
-	if (!nm_i->dirty_nat_cnt)
-		return;
-
 	while ((found = __gang_lookup_nat_set(nm_i,
 					set_idx, NATVEC_SIZE, setvec))) {
 		unsigned idx;
@@ -1973,13 +1978,13 @@ static int init_node_manager(struct f2fs_sb_info *sbi)
 
 	INIT_RADIX_TREE(&nm_i->free_nid_root, GFP_ATOMIC);
 	INIT_LIST_HEAD(&nm_i->free_nid_list);
-	INIT_RADIX_TREE(&nm_i->nat_root, GFP_ATOMIC);
+	INIT_RADIX_TREE(&nm_i->nat_root, GFP_NOIO);
-	INIT_RADIX_TREE(&nm_i->nat_set_root, GFP_ATOMIC);
+	INIT_RADIX_TREE(&nm_i->nat_set_root, GFP_NOIO);
 	INIT_LIST_HEAD(&nm_i->nat_entries);
 
 	mutex_init(&nm_i->build_lock);
 	spin_lock_init(&nm_i->free_nid_list_lock);
-	rwlock_init(&nm_i->nat_tree_lock);
+	init_rwsem(&nm_i->nat_tree_lock);
 
 	nm_i->next_scan_nid = le32_to_cpu(sbi->ckpt->next_free_nid);
 	nm_i->bitmap_size = __bitmap_size(sbi, NAT_BITMAP);
@@ -2035,7 +2040,7 @@ void destroy_node_manager(struct f2fs_sb_info *sbi)
 	spin_unlock(&nm_i->free_nid_list_lock);
 
 	/* destroy nat cache */
-	write_lock(&nm_i->nat_tree_lock);
+	down_write(&nm_i->nat_tree_lock);
 	while ((found = __gang_lookup_nat_cache(nm_i,
 					nid, NATVEC_SIZE, natvec))) {
 		unsigned idx;
@@ -2044,7 +2049,7 @@ void destroy_node_manager(struct f2fs_sb_info *sbi)
 			__del_from_nat_cache(nm_i, natvec[idx]);
 	}
 	f2fs_bug_on(sbi, nm_i->nat_cnt);
-	write_unlock(&nm_i->nat_tree_lock);
+	up_write(&nm_i->nat_tree_lock);
 
 	kfree(nm_i->nat_bitmap);
 	sbi->nm_info = NULL;
@@ -2061,17 +2066,17 @@ int __init create_node_manager_caches(void)
 	free_nid_slab = f2fs_kmem_cache_create("free_nid",
 			sizeof(struct free_nid));
 	if (!free_nid_slab)
-		goto destory_nat_entry;
+		goto destroy_nat_entry;
 
 	nat_entry_set_slab = f2fs_kmem_cache_create("nat_entry_set",
 			sizeof(struct nat_entry_set));
 	if (!nat_entry_set_slab)
-		goto destory_free_nid;
+		goto destroy_free_nid;
 	return 0;
 
-destory_free_nid:
+destroy_free_nid:
 	kmem_cache_destroy(free_nid_slab);
-destory_nat_entry:
+destroy_nat_entry:
 	kmem_cache_destroy(nat_entry_slab);
 fail:
 	return -ENOMEM;

fs/f2fs/node.h

@@ -106,7 +106,8 @@ static inline void raw_nat_from_node_info(struct f2fs_nat_entry *raw_ne,
 enum mem_type {
 	FREE_NIDS,	/* indicates the free nid list */
 	NAT_ENTRIES,	/* indicates the cached nat entry */
-	DIRTY_DENTS	/* indicates dirty dentry pages */
+	DIRTY_DENTS,	/* indicates dirty dentry pages */
+	INO_ENTRIES,	/* indicates inode entries */
 };
 
 struct nat_entry_set {
@@ -192,10 +193,7 @@ static inline void set_to_next_nat(struct f2fs_nm_info *nm_i, nid_t start_nid)
 {
 	unsigned int block_off = NAT_BLOCK_OFFSET(start_nid);
 
-	if (f2fs_test_bit(block_off, nm_i->nat_bitmap))
+	f2fs_change_bit(block_off, nm_i->nat_bitmap);
-		f2fs_clear_bit(block_off, nm_i->nat_bitmap);
-	else
-		f2fs_set_bit(block_off, nm_i->nat_bitmap);
 }
 
 static inline void fill_node_footer(struct page *page, nid_t nid,

fs/f2fs/recovery.c

@@ -111,7 +111,7 @@ retry:
 			iput(einode);
 			goto out_unmap_put;
 		}
-		f2fs_delete_entry(de, page, einode);
+		f2fs_delete_entry(de, page, dir, einode);
 		iput(einode);
 		goto retry;
 	}
@@ -129,7 +129,7 @@ retry:
 	goto out;
 
 out_unmap_put:
-	kunmap(page);
+	f2fs_dentry_kunmap(dir, page);
 	f2fs_put_page(page, 0);
 out_err:
 	iput(dir);
@@ -170,13 +170,15 @@ static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head)
 	curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
 	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
 
+	ra_meta_pages(sbi, blkaddr, 1, META_POR);
+
 	while (1) {
 		struct fsync_inode_entry *entry;
 
 		if (blkaddr < MAIN_BLKADDR(sbi) || blkaddr >= MAX_BLKADDR(sbi))
 			return 0;
 
-		page = get_meta_page_ra(sbi, blkaddr);
+		page = get_meta_page(sbi, blkaddr);
 
 		if (cp_ver != cpver_of_node(page))
 			break;
@@ -227,6 +229,8 @@ next:
 		/* check next segment */
 		blkaddr = next_blkaddr_of_node(page);
 		f2fs_put_page(page, 1);
+
+		ra_meta_pages_cond(sbi, blkaddr);
 	}
 	f2fs_put_page(page, 1);
 	return err;
@@ -436,7 +440,9 @@ static int recover_data(struct f2fs_sb_info *sbi,
 		if (blkaddr < MAIN_BLKADDR(sbi) || blkaddr >= MAX_BLKADDR(sbi))
 			break;
 
-		page = get_meta_page_ra(sbi, blkaddr);
+		ra_meta_pages_cond(sbi, blkaddr);
+
+		page = get_meta_page(sbi, blkaddr);
 
 		if (cp_ver != cpver_of_node(page)) {
 			f2fs_put_page(page, 1);

fs/f2fs/segment.c

@@ -178,17 +178,47 @@ void register_inmem_page(struct inode *inode, struct page *page)
 {
 	struct f2fs_inode_info *fi = F2FS_I(inode);
 	struct inmem_pages *new;
+	int err;
+
+	SetPagePrivate(page);
 
 	new = f2fs_kmem_cache_alloc(inmem_entry_slab, GFP_NOFS);
 
 	/* add atomic page indices to the list */
 	new->page = page;
 	INIT_LIST_HEAD(&new->list);
-
+retry:
 	/* increase reference count with clean state */
 	mutex_lock(&fi->inmem_lock);
+	err = radix_tree_insert(&fi->inmem_root, page->index, new);
+	if (err == -EEXIST) {
+		mutex_unlock(&fi->inmem_lock);
+		kmem_cache_free(inmem_entry_slab, new);
+		return;
+	} else if (err) {
+		mutex_unlock(&fi->inmem_lock);
+		goto retry;
+	}
 	get_page(page);
 	list_add_tail(&new->list, &fi->inmem_pages);
+	inc_page_count(F2FS_I_SB(inode), F2FS_INMEM_PAGES);
+	mutex_unlock(&fi->inmem_lock);
+}
+
+void invalidate_inmem_page(struct inode *inode, struct page *page)
+{
+	struct f2fs_inode_info *fi = F2FS_I(inode);
+	struct inmem_pages *cur;
+
+	mutex_lock(&fi->inmem_lock);
+	cur = radix_tree_lookup(&fi->inmem_root, page->index);
+	if (cur) {
+		radix_tree_delete(&fi->inmem_root, cur->page->index);
+		f2fs_put_page(cur->page, 0);
+		list_del(&cur->list);
+		kmem_cache_free(inmem_entry_slab, cur);
+		dec_page_count(F2FS_I_SB(inode), F2FS_INMEM_PAGES);
+	}
 	mutex_unlock(&fi->inmem_lock);
 }
 
@@ -203,7 +233,16 @@ void commit_inmem_pages(struct inode *inode, bool abort)
 		.rw = WRITE_SYNC,
 	};
 
-	f2fs_balance_fs(sbi);
+	/*
+	 * The abort is true only when f2fs_evict_inode is called.
+	 * Basically, the f2fs_evict_inode doesn't produce any data writes, so
+	 * that we don't need to call f2fs_balance_fs.
+	 * Otherwise, f2fs_gc in f2fs_balance_fs can wait forever until this
+	 * inode becomes free by iget_locked in f2fs_iget.
+	 */
+	if (!abort)
+		f2fs_balance_fs(sbi);
+
 	f2fs_lock_op(sbi);
 
 	mutex_lock(&fi->inmem_lock);
@@ -216,9 +255,11 @@ void commit_inmem_pages(struct inode *inode, bool abort)
 			do_write_data_page(cur->page, &fio);
 			submit_bio = true;
 		}
+		radix_tree_delete(&fi->inmem_root, cur->page->index);
 		f2fs_put_page(cur->page, 1);
 		list_del(&cur->list);
 		kmem_cache_free(inmem_entry_slab, cur);
+		dec_page_count(F2FS_I_SB(inode), F2FS_INMEM_PAGES);
 	}
 	if (submit_bio)
 		f2fs_submit_merged_bio(sbi, DATA, WRITE);
@@ -248,7 +289,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi)
 {
 	/* check the # of cached NAT entries and prefree segments */
 	if (try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK) ||
-				excess_prefree_segs(sbi))
+			excess_prefree_segs(sbi) ||
+			available_free_memory(sbi, INO_ENTRIES))
 		f2fs_sync_fs(sbi->sb, true);
 }
 
@@ -441,10 +483,33 @@ void discard_next_dnode(struct f2fs_sb_info *sbi, block_t blkaddr)
 	}
 }
 
-static void add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc)
+static void __add_discard_entry(struct f2fs_sb_info *sbi,
+		struct cp_control *cpc, unsigned int start, unsigned int end)
 {
 	struct list_head *head = &SM_I(sbi)->discard_list;
-	struct discard_entry *new;
+	struct discard_entry *new, *last;
+
+	if (!list_empty(head)) {
+		last = list_last_entry(head, struct discard_entry, list);
+		if (START_BLOCK(sbi, cpc->trim_start) + start ==
+						last->blkaddr + last->len) {
+			last->len += end - start;
+			goto done;
+		}
+	}
+
+	new = f2fs_kmem_cache_alloc(discard_entry_slab, GFP_NOFS);
+	INIT_LIST_HEAD(&new->list);
+	new->blkaddr = START_BLOCK(sbi, cpc->trim_start) + start;
+	new->len = end - start;
+	list_add_tail(&new->list, head);
+done:
+	SM_I(sbi)->nr_discards += end - start;
+	cpc->trimmed += end - start;
+}
+
+static void add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc)
+{
 	int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
 	int max_blocks = sbi->blocks_per_seg;
 	struct seg_entry *se = get_seg_entry(sbi, cpc->trim_start);
@@ -473,13 +538,7 @@ static void add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 		}
 		mutex_unlock(&dirty_i->seglist_lock);
 
-		new = f2fs_kmem_cache_alloc(discard_entry_slab, GFP_NOFS);
+		__add_discard_entry(sbi, cpc, 0, sbi->blocks_per_seg);
-		INIT_LIST_HEAD(&new->list);
-		new->blkaddr = START_BLOCK(sbi, cpc->trim_start);
-		new->len = sbi->blocks_per_seg;
-		list_add_tail(&new->list, head);
-		SM_I(sbi)->nr_discards += sbi->blocks_per_seg;
-		cpc->trimmed += sbi->blocks_per_seg;
 		return;
 	}
 
@@ -489,7 +548,7 @@ static void add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 
 	/* SIT_VBLOCK_MAP_SIZE should be multiple of sizeof(unsigned long) */
 	for (i = 0; i < entries; i++)
-		dmap[i] = (cur_map[i] ^ ckpt_map[i]) & ckpt_map[i];
+		dmap[i] = ~(cur_map[i] | ckpt_map[i]);
 
 	while (force || SM_I(sbi)->nr_discards <= SM_I(sbi)->max_discards) {
 		start = __find_rev_next_bit(dmap, max_blocks, end + 1);
@@ -501,14 +560,7 @@ static void add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 		if (end - start < cpc->trim_minlen)
 			continue;
 
-		new = f2fs_kmem_cache_alloc(discard_entry_slab, GFP_NOFS);
+		__add_discard_entry(sbi, cpc, start, end);
-		INIT_LIST_HEAD(&new->list);
-		new->blkaddr = START_BLOCK(sbi, cpc->trim_start) + start;
-		new->len = end - start;
-		cpc->trimmed += end - start;
-
-		list_add_tail(&new->list, head);
-		SM_I(sbi)->nr_discards += end - start;
 	}
 }
 
@@ -620,10 +672,10 @@ static void update_sit_entry(struct f2fs_sb_info *sbi, block_t blkaddr, int del)
 
 	/* Update valid block bitmap */
 	if (del > 0) {
-		if (f2fs_set_bit(offset, se->cur_valid_map))
+		if (f2fs_test_and_set_bit(offset, se->cur_valid_map))
 			f2fs_bug_on(sbi, 1);
 	} else {
-		if (!f2fs_clear_bit(offset, se->cur_valid_map))
+		if (!f2fs_test_and_clear_bit(offset, se->cur_valid_map))
 			f2fs_bug_on(sbi, 1);
 	}
 	if (!f2fs_test_bit(offset, se->ckpt_valid_map))
@@ -1004,6 +1056,7 @@ int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range)
 						range->len < sbi->blocksize)
 		return -EINVAL;
 
+	cpc.trimmed = 0;
 	if (end <= MAIN_BLKADDR(sbi))
 		goto out;
 
@@ -1015,10 +1068,11 @@ int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range)
 	cpc.trim_start = start_segno;
 	cpc.trim_end = end_segno;
 	cpc.trim_minlen = range->minlen >> sbi->log_blocksize;
-	cpc.trimmed = 0;
 
 	/* do checkpoint to issue discard commands safely */
+	mutex_lock(&sbi->gc_mutex);
 	write_checkpoint(sbi, &cpc);
+	mutex_unlock(&sbi->gc_mutex);
 out:
 	range->len = cpc.trimmed << sbi->log_blocksize;
 	return 0;
@@ -1050,8 +1104,8 @@ static int __get_segment_type_4(struct page *page, enum page_type p_type)
 		else
 			return CURSEG_COLD_DATA;
 	} else {
-		if (IS_DNODE(page) && !is_cold_node(page))
+		if (IS_DNODE(page) && is_cold_node(page))
-			return CURSEG_HOT_NODE;
+			return CURSEG_WARM_NODE;
 		else
 			return CURSEG_COLD_NODE;
 	}
@@ -1524,17 +1578,7 @@ int lookup_journal_in_cursum(struct f2fs_summary_block *sum, int type,
 static struct page *get_current_sit_page(struct f2fs_sb_info *sbi,
 					unsigned int segno)
 {
-	struct sit_info *sit_i = SIT_I(sbi);
+	return get_meta_page(sbi, current_sit_addr(sbi, segno));
-	unsigned int offset = SIT_BLOCK_OFFSET(segno);
-	block_t blk_addr = sit_i->sit_base_addr + offset;
-
-	check_seg_range(sbi, segno);
-
-	/* calculate sit block address */
-	if (f2fs_test_bit(offset, sit_i->sit_bitmap))
-		blk_addr += sit_i->sit_blocks;
-
-	return get_meta_page(sbi, blk_addr);
 }
 
 static struct page *get_next_sit_page(struct f2fs_sb_info *sbi,
@@ -1687,7 +1731,7 @@ void flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 	 * #2, flush sit entries to sit page.
 	 */
 	list_for_each_entry_safe(ses, tmp, head, set_list) {
-		struct page *page;
+		struct page *page = NULL;
 		struct f2fs_sit_block *raw_sit = NULL;
 		unsigned int start_segno = ses->start_segno;
 		unsigned int end = min(start_segno + SIT_ENTRY_PER_BLOCK,
@@ -2200,7 +2244,7 @@ int __init create_segment_manager_caches(void)
 		goto fail;
 
 	sit_entry_set_slab = f2fs_kmem_cache_create("sit_entry_set",
-			sizeof(struct nat_entry_set));
+			sizeof(struct sit_entry_set));
 	if (!sit_entry_set_slab)
 		goto destory_discard_entry;
 

fs/f2fs/segment.h

@@ -657,10 +657,7 @@ static inline void set_to_next_sit(struct sit_info *sit_i, unsigned int start)
 {
 	unsigned int block_off = SIT_BLOCK_OFFSET(start);
 
-	if (f2fs_test_bit(block_off, sit_i->sit_bitmap))
+	f2fs_change_bit(block_off, sit_i->sit_bitmap);
-		f2fs_clear_bit(block_off, sit_i->sit_bitmap);
-	else
-		f2fs_set_bit(block_off, sit_i->sit_bitmap);
 }
 
 static inline unsigned long long get_mtime(struct f2fs_sb_info *sbi)
@@ -714,6 +711,9 @@ static inline unsigned int max_hw_blocks(struct f2fs_sb_info *sbi)
  */
 static inline int nr_pages_to_skip(struct f2fs_sb_info *sbi, int type)
 {
+	if (sbi->sb->s_bdi->dirty_exceeded)
+		return 0;
+
 	if (type == DATA)
 		return sbi->blocks_per_seg;
 	else if (type == NODE)

fs/f2fs/super.c

@@ -51,8 +51,10 @@ enum {
 	Opt_disable_ext_identify,
 	Opt_inline_xattr,
 	Opt_inline_data,
+	Opt_inline_dentry,
 	Opt_flush_merge,
 	Opt_nobarrier,
+	Opt_fastboot,
 	Opt_err,
 };
 
@@ -69,8 +71,10 @@ static match_table_t f2fs_tokens = {
 	{Opt_disable_ext_identify, "disable_ext_identify"},
 	{Opt_inline_xattr, "inline_xattr"},
 	{Opt_inline_data, "inline_data"},
+	{Opt_inline_dentry, "inline_dentry"},
 	{Opt_flush_merge, "flush_merge"},
 	{Opt_nobarrier, "nobarrier"},
+	{Opt_fastboot, "fastboot"},
 	{Opt_err, NULL},
 };
 
@@ -340,12 +344,18 @@ static int parse_options(struct super_block *sb, char *options)
 		case Opt_inline_data:
 			set_opt(sbi, INLINE_DATA);
 			break;
+		case Opt_inline_dentry:
+			set_opt(sbi, INLINE_DENTRY);
+			break;
 		case Opt_flush_merge:
 			set_opt(sbi, FLUSH_MERGE);
 			break;
 		case Opt_nobarrier:
 			set_opt(sbi, NOBARRIER);
 			break;
+		case Opt_fastboot:
+			set_opt(sbi, FASTBOOT);
+			break;
 		default:
 			f2fs_msg(sb, KERN_ERR,
 				"Unrecognized mount option \"%s\" or missing value",
@@ -373,6 +383,7 @@ static struct inode *f2fs_alloc_inode(struct super_block *sb)
 	fi->i_advise = 0;
 	rwlock_init(&fi->ext.ext_lock);
 	init_rwsem(&fi->i_sem);
+	INIT_RADIX_TREE(&fi->inmem_root, GFP_NOFS);
 	INIT_LIST_HEAD(&fi->inmem_pages);
 	mutex_init(&fi->inmem_lock);
 
@@ -473,9 +484,9 @@ int f2fs_sync_fs(struct super_block *sb, int sync)
 	trace_f2fs_sync_fs(sb, sync);
 
 	if (sync) {
-		struct cp_control cpc = {
+		struct cp_control cpc;
-			.reason = CP_SYNC,
+
-		};
+		cpc.reason = test_opt(sbi, FASTBOOT) ? CP_UMOUNT : CP_SYNC;
 		mutex_lock(&sbi->gc_mutex);
 		write_checkpoint(sbi, &cpc);
 		mutex_unlock(&sbi->gc_mutex);
@@ -562,10 +573,14 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
 		seq_puts(seq, ",disable_ext_identify");
 	if (test_opt(sbi, INLINE_DATA))
 		seq_puts(seq, ",inline_data");
+	if (test_opt(sbi, INLINE_DENTRY))
+		seq_puts(seq, ",inline_dentry");
 	if (!f2fs_readonly(sbi->sb) && test_opt(sbi, FLUSH_MERGE))
 		seq_puts(seq, ",flush_merge");
 	if (test_opt(sbi, NOBARRIER))
 		seq_puts(seq, ",nobarrier");
+	if (test_opt(sbi, FASTBOOT))
+		seq_puts(seq, ",fastboot");
 	seq_printf(seq, ",active_logs=%u", sbi->active_logs);
 
 	return 0;
@@ -654,7 +669,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 			f2fs_sync_fs(sb, 1);
 			need_restart_gc = true;
 		}
-	} else if (test_opt(sbi, BG_GC) && !sbi->gc_thread) {
+	} else if (!sbi->gc_thread) {
 		err = start_gc_thread(sbi);
 		if (err)
 			goto restore_opts;
@@ -667,7 +682,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 	 */
 	if ((*flags & MS_RDONLY) || !test_opt(sbi, FLUSH_MERGE)) {
 		destroy_flush_cmd_control(sbi);
-	} else if (test_opt(sbi, FLUSH_MERGE) && !SM_I(sbi)->cmd_control_info) {
+	} else if (!SM_I(sbi)->cmd_control_info) {
 		err = create_flush_cmd_control(sbi);
 		if (err)
 			goto restore_gc;
@@ -922,7 +937,7 @@ retry:
 static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 {
 	struct f2fs_sb_info *sbi;
-	struct f2fs_super_block *raw_super;
+	struct f2fs_super_block *raw_super = NULL;
 	struct buffer_head *raw_super_buf;
 	struct inode *root;
 	long err = -EINVAL;
@@ -1123,7 +1138,7 @@ try_onemore:
 	 * If filesystem is not mounted as read-only then
 	 * do start the gc_thread.
 	 */
-	if (!f2fs_readonly(sb)) {
+	if (test_opt(sbi, BG_GC) && !f2fs_readonly(sb)) {
 		/* After POR, we can run background GC thread.*/
 		err = start_gc_thread(sbi);
 		if (err)

fs/f2fs/xattr.c

@@ -83,7 +83,7 @@ static int f2fs_xattr_generic_get(struct dentry *dentry, const char *name,
 	}
 	if (strcmp(name, "") == 0)
 		return -EINVAL;
-	return f2fs_getxattr(dentry->d_inode, type, name, buffer, size);
+	return f2fs_getxattr(dentry->d_inode, type, name, buffer, size, NULL);
 }
 
 static int f2fs_xattr_generic_set(struct dentry *dentry, const char *name,
@@ -398,7 +398,7 @@ static inline int write_all_xattrs(struct inode *inode, __u32 hsize,
 }
 
 int f2fs_getxattr(struct inode *inode, int index, const char *name,
-		void *buffer, size_t buffer_size)
+		void *buffer, size_t buffer_size, struct page *ipage)
 {
 	struct f2fs_xattr_entry *entry;
 	void *base_addr;
@@ -412,7 +412,7 @@ int f2fs_getxattr(struct inode *inode, int index, const char *name,
 	if (len > F2FS_NAME_LEN)
 		return -ERANGE;
 
-	base_addr = read_all_xattrs(inode, NULL);
+	base_addr = read_all_xattrs(inode, ipage);
 	if (!base_addr)
 		return -ENOMEM;
 

fs/f2fs/xattr.h

@@ -115,7 +115,8 @@ extern const struct xattr_handler *f2fs_xattr_handlers[];
 
 extern int f2fs_setxattr(struct inode *, int, const char *,
 				const void *, size_t, struct page *, int);
-extern int f2fs_getxattr(struct inode *, int, const char *, void *, size_t);
+extern int f2fs_getxattr(struct inode *, int, const char *, void *,
+						size_t, struct page *);
 extern ssize_t f2fs_listxattr(struct dentry *, char *, size_t);
 #else
 
@@ -126,7 +127,8 @@ static inline int f2fs_setxattr(struct inode *inode, int index,
 	return -EOPNOTSUPP;
 }
 static inline int f2fs_getxattr(struct inode *inode, int index,
-		const char *name, void *buffer, size_t buffer_size)
+			const char *name, void *buffer,
+			size_t buffer_size, struct page *dpage)
 {
 	return -EOPNOTSUPP;
 }

include/linux/f2fs_fs.h

@@ -33,7 +33,8 @@
 #define F2FS_META_INO(sbi)	(sbi->meta_ino_num)
 
 /* This flag is used by node and meta inodes, and by recovery */
-#define GFP_F2FS_ZERO	(GFP_NOFS | __GFP_ZERO)
+#define GFP_F2FS_ZERO		(GFP_NOFS | __GFP_ZERO)
+#define GFP_F2FS_HIGH_ZERO	(GFP_NOFS | __GFP_ZERO | __GFP_HIGHMEM)
 
 /*
  * For further optimization on multi-head logs, on-disk layout supports maximum
@@ -170,14 +171,12 @@ struct f2fs_extent {
 
 #define F2FS_INLINE_XATTR	0x01	/* file inline xattr flag */
 #define F2FS_INLINE_DATA	0x02	/* file inline data flag */
+#define F2FS_INLINE_DENTRY	0x04	/* file inline dentry flag */
+#define F2FS_DATA_EXIST		0x08	/* file inline data exist flag */
 
 #define MAX_INLINE_DATA		(sizeof(__le32) * (DEF_ADDRS_PER_INODE - \
 						F2FS_INLINE_XATTR_ADDRS - 1))
 
-#define INLINE_DATA_OFFSET	(PAGE_CACHE_SIZE - sizeof(struct node_footer) -\
-				sizeof(__le32) * (DEF_ADDRS_PER_INODE + \
-				DEF_NIDS_PER_INODE - 1))
-
 struct f2fs_inode {
 	__le16 i_mode;			/* file mode */
 	__u8 i_advise;			/* file hints */
@@ -435,6 +434,24 @@ struct f2fs_dentry_block {
 	__u8 filename[NR_DENTRY_IN_BLOCK][F2FS_SLOT_LEN];
 } __packed;
 
+/* for inline dir */
+#define NR_INLINE_DENTRY	(MAX_INLINE_DATA * BITS_PER_BYTE / \
+				((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
+				BITS_PER_BYTE + 1))
+#define INLINE_DENTRY_BITMAP_SIZE	((NR_INLINE_DENTRY + \
+					BITS_PER_BYTE - 1) / BITS_PER_BYTE)
+#define INLINE_RESERVED_SIZE	(MAX_INLINE_DATA - \
+				((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
+				NR_INLINE_DENTRY + INLINE_DENTRY_BITMAP_SIZE))
+
+/* inline directory entry structure */
+struct f2fs_inline_dentry {
+	__u8 dentry_bitmap[INLINE_DENTRY_BITMAP_SIZE];
+	__u8 reserved[INLINE_RESERVED_SIZE];
+	struct f2fs_dir_entry dentry[NR_INLINE_DENTRY];
+	__u8 filename[NR_INLINE_DENTRY][F2FS_SLOT_LEN];
+} __packed;
+
 /* file types used in inode_info->flags */
 enum {
 	F2FS_FT_UNKNOWN,