Merge tag 'for-f2fs-3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs

Pull f2fs updates from Jaegeuk Kim:
 "In this round, a couple of sysfs entries were introduced to tune the
  f2fs at runtime.

  In addition, f2fs starts to support inline_data and improves the
  read/write performance in some workloads by refactoring bio-related
  flows.

  This patch-set includes the following major enhancement patches.
   - support inline_data
   - refactor bio operations such as merge operations and rw type
     assignment
   - enhance the direct IO path
   - enhance bio operations
   - truncate a node page when it becomes obsolete
   - add sysfs entries: small_discards, max_victim_search, and
     in-place-update
   - add a sysfs entry to control max_victim_search

  The other bug fixes are as follows.
   - fix a bug in truncate_partial_nodes
   - avoid warnings during sparse and build process
   - fix error handling flows
   - fix potential bit overflows

  And, there are a bunch of cleanups"

* tag 'for-f2fs-3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (95 commits)
  f2fs: drop obsolete node page when it is truncated
  f2fs: introduce NODE_MAPPING for code consistency
  f2fs: remove the orphan block page array
  f2fs: add help function META_MAPPING
  f2fs: move a branch for code redability
  f2fs: call mark_inode_dirty to flush dirty pages
  f2fs: clean checkpatch warnings
  f2fs: missing REQ_META and REQ_PRIO when sync_meta_pages(META_FLUSH)
  f2fs: avoid f2fs_balance_fs call during pageout
  f2fs: add delimiter to seperate name and value in debug phrase
  f2fs: use spinlock rather than mutex for better speed
  f2fs: move alloc new orphan node out of lock protection region
  f2fs: move grabing orphan pages out of protection region
  f2fs: remove the needless parameter of f2fs_wait_on_page_writeback
  f2fs: update documents and a MAINTAINERS entry
  f2fs: add a sysfs entry to control max_victim_search
  f2fs: improve write performance under frequent fsync calls
  f2fs: avoid to read inline data except first page
  f2fs: avoid to left uninitialized data in page when read inline data
  f2fs: fix truncate_partial_nodes bug
  ...

Documentation/ABI/testing/sysfs-fs-f2fs

@@ -24,3 +24,34 @@ Date:		July 2013
 Contact:	"Namjae Jeon" <namjae.jeon@samsung.com>
 Description:
 		 Controls the victim selection policy for garbage collection.
+
+What:		/sys/fs/f2fs/<disk>/reclaim_segments
+Date:		October 2013
+Contact:	"Jaegeuk Kim" <jaegeuk.kim@samsung.com>
+Description:
+		 Controls the issue rate of segment discard commands.
+
+What:		/sys/fs/f2fs/<disk>/ipu_policy
+Date:		November 2013
+Contact:	"Jaegeuk Kim" <jaegeuk.kim@samsung.com>
+Description:
+		 Controls the in-place-update policy.
+
+What:		/sys/fs/f2fs/<disk>/min_ipu_util
+Date:		November 2013
+Contact:	"Jaegeuk Kim" <jaegeuk.kim@samsung.com>
+Description:
+		 Controls the FS utilization condition for the in-place-update
+		 policies.
+
+What:		/sys/fs/f2fs/<disk>/max_small_discards
+Date:		November 2013
+Contact:	"Jaegeuk Kim" <jaegeuk.kim@samsung.com>
+Description:
+		 Controls the issue rate of small discard commands.
+
+What:		/sys/fs/f2fs/<disk>/max_victim_search
+Date:		January 2014
+Contact:	"Jaegeuk Kim" <jaegeuk.kim@samsung.com>
+Description:
+		 Controls the number of trials to find a victim segment.

Documentation/filesystems/f2fs.txt

@@ -120,6 +120,8 @@ active_logs=%u         Support configuring the number of active logs. In the
 disable_ext_identify   Disable the extension list configured by mkfs, so f2fs
                        does not aware of cold files such as media files.
 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.
 
 ================================================================================
 DEBUGFS ENTRIES
@@ -171,6 +173,28 @@ Files in /sys/fs/f2fs/<devname>
 			      conduct checkpoint to reclaim the prefree segments
 			      to free segments. By default, 100 segments, 200MB.
 
+ max_small_discards	      This parameter controls the number of discard
+			      commands that consist small blocks less than 2MB.
+			      The candidates to be discarded are cached until
+			      checkpoint is triggered, and issued during the
+			      checkpoint. By default, it is disabled with 0.
+
+ ipu_policy                   This parameter controls the policy of in-place
+                              updates in f2fs. There are five policies:
+                               0: F2FS_IPU_FORCE, 1: F2FS_IPU_SSR,
+                               2: F2FS_IPU_UTIL,  3: F2FS_IPU_SSR_UTIL,
+                               4: F2FS_IPU_DISABLE.
+
+ min_ipu_util                 This parameter controls the threshold to trigger
+                              in-place-updates. The number indicates percentage
+                              of the filesystem utilization, and used by
+                              F2FS_IPU_UTIL and F2FS_IPU_SSR_UTIL policies.
+
+ max_victim_search	      This parameter controls the number of trials to
+			      find a victim segment when conducting SSR and
+			      cleaning operations. The default value is 4096
+			      which covers 8GB block address range.
+
 ================================================================================
 USAGE
 ================================================================================

MAINTAINERS

@@ -3634,6 +3634,7 @@ W:	http://en.wikipedia.org/wiki/F2FS
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git
 S:	Maintained
 F:	Documentation/filesystems/f2fs.txt
+F:	Documentation/ABI/testing/sysfs-fs-f2fs
 F:	fs/f2fs/
 F:	include/linux/f2fs_fs.h
 

fs/f2fs/Makefile

@@ -1,6 +1,6 @@
 obj-$(CONFIG_F2FS_FS) += f2fs.o
 
-f2fs-y		:= dir.o file.o inode.o namei.o hash.o super.o
+f2fs-y		:= dir.o file.o inode.o namei.o hash.o super.o inline.o
 f2fs-y		+= checkpoint.o gc.o data.o node.o segment.o recovery.o
 f2fs-$(CONFIG_F2FS_STAT_FS) += debug.o
 f2fs-$(CONFIG_F2FS_FS_XATTR) += xattr.o

fs/f2fs/checkpoint.c

@@ -30,7 +30,7 @@ static struct kmem_cache *inode_entry_slab;
  */
 struct page *grab_meta_page(struct f2fs_sb_info *sbi, pgoff_t index)
 {
-	struct address_space *mapping = sbi->meta_inode->i_mapping;
+	struct address_space *mapping = META_MAPPING(sbi);
 	struct page *page = NULL;
 repeat:
 	page = grab_cache_page(mapping, index);
@@ -50,7 +50,7 @@ repeat:
  */
 struct page *get_meta_page(struct f2fs_sb_info *sbi, pgoff_t index)
 {
-	struct address_space *mapping = sbi->meta_inode->i_mapping;
+	struct address_space *mapping = META_MAPPING(sbi);
 	struct page *page;
 repeat:
 	page = grab_cache_page(mapping, index);
@@ -61,11 +61,12 @@ repeat:
 	if (PageUptodate(page))
 		goto out;
 
-	if (f2fs_readpage(sbi, page, index, READ_SYNC))
+	if (f2fs_submit_page_bio(sbi, page, index,
+				READ_SYNC | REQ_META | REQ_PRIO))
 		goto repeat;
 
 	lock_page(page);
-	if (page->mapping != mapping) {
+	if (unlikely(page->mapping != mapping)) {
 		f2fs_put_page(page, 1);
 		goto repeat;
 	}
@@ -81,13 +82,12 @@ static int f2fs_write_meta_page(struct page *page,
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 
 	/* Should not write any meta pages, if any IO error was occurred */
-	if (wbc->for_reclaim || sbi->por_doing ||
-			is_set_ckpt_flags(F2FS_CKPT(sbi), CP_ERROR_FLAG)) {
-		dec_page_count(sbi, F2FS_DIRTY_META);
-		wbc->pages_skipped++;
-		set_page_dirty(page);
-		return AOP_WRITEPAGE_ACTIVATE;
-	}
+	if (unlikely(sbi->por_doing ||
+			is_set_ckpt_flags(F2FS_CKPT(sbi), CP_ERROR_FLAG)))
+		goto redirty_out;
+
+	if (wbc->for_reclaim)
+		goto redirty_out;
 
 	wait_on_page_writeback(page);
 
@@ -95,24 +95,31 @@ static int f2fs_write_meta_page(struct page *page,
 	dec_page_count(sbi, F2FS_DIRTY_META);
 	unlock_page(page);
 	return 0;
+
+redirty_out:
+	dec_page_count(sbi, F2FS_DIRTY_META);
+	wbc->pages_skipped++;
+	set_page_dirty(page);
+	return AOP_WRITEPAGE_ACTIVATE;
 }
 
 static int f2fs_write_meta_pages(struct address_space *mapping,
 				struct writeback_control *wbc)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(mapping->host->i_sb);
-	struct block_device *bdev = sbi->sb->s_bdev;
+	int nrpages = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
 	long written;
 
 	if (wbc->for_kupdate)
 		return 0;
 
-	if (get_pages(sbi, F2FS_DIRTY_META) == 0)
+	/* collect a number of dirty meta pages and write together */
+	if (get_pages(sbi, F2FS_DIRTY_META) < nrpages)
 		return 0;
 
 	/* if mounting is failed, skip writing node pages */
 	mutex_lock(&sbi->cp_mutex);
-	written = sync_meta_pages(sbi, META, bio_get_nr_vecs(bdev));
+	written = sync_meta_pages(sbi, META, nrpages);
 	mutex_unlock(&sbi->cp_mutex);
 	wbc->nr_to_write -= written;
 	return 0;
@@ -121,7 +128,7 @@ static int f2fs_write_meta_pages(struct address_space *mapping,
 long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
 						long nr_to_write)
 {
-	struct address_space *mapping = sbi->meta_inode->i_mapping;
+	struct address_space *mapping = META_MAPPING(sbi);
 	pgoff_t index = 0, end = LONG_MAX;
 	struct pagevec pvec;
 	long nwritten = 0;
@@ -136,7 +143,7 @@ long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
 		nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
 				PAGECACHE_TAG_DIRTY,
 				min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
-		if (nr_pages == 0)
+		if (unlikely(nr_pages == 0))
 			break;
 
 		for (i = 0; i < nr_pages; i++) {
@@ -149,7 +156,8 @@ long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
 				unlock_page(page);
 				break;
 			}
-			if (nwritten++ >= nr_to_write)
+			nwritten++;
+			if (unlikely(nwritten >= nr_to_write))
 				break;
 		}
 		pagevec_release(&pvec);
@@ -157,7 +165,7 @@ long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
 	}
 
 	if (nwritten)
-		f2fs_submit_bio(sbi, type, nr_to_write == LONG_MAX);
+		f2fs_submit_merged_bio(sbi, type, WRITE);
 
 	return nwritten;
 }
@@ -186,31 +194,24 @@ const struct address_space_operations f2fs_meta_aops = {
 
 int acquire_orphan_inode(struct f2fs_sb_info *sbi)
 {
-	unsigned int max_orphans;
 	int err = 0;
 
-	/*
-	 * considering 512 blocks in a segment 5 blocks are needed for cp
-	 * and log segment summaries. Remaining blocks are used to keep
-	 * orphan entries with the limitation one reserved segment
-	 * for cp pack we can have max 1020*507 orphan entries
-	 */
-	max_orphans = (sbi->blocks_per_seg - 5) * F2FS_ORPHANS_PER_BLOCK;
-	mutex_lock(&sbi->orphan_inode_mutex);
-	if (sbi->n_orphans >= max_orphans)
+	spin_lock(&sbi->orphan_inode_lock);
+	if (unlikely(sbi->n_orphans >= sbi->max_orphans))
 		err = -ENOSPC;
 	else
 		sbi->n_orphans++;
-	mutex_unlock(&sbi->orphan_inode_mutex);
+	spin_unlock(&sbi->orphan_inode_lock);
+
 	return err;
 }
 
 void release_orphan_inode(struct f2fs_sb_info *sbi)
 {
-	mutex_lock(&sbi->orphan_inode_mutex);
+	spin_lock(&sbi->orphan_inode_lock);
 	f2fs_bug_on(sbi->n_orphans == 0);
 	sbi->n_orphans--;
-	mutex_unlock(&sbi->orphan_inode_mutex);
+	spin_unlock(&sbi->orphan_inode_lock);
 }
 
 void add_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
@@ -218,27 +219,30 @@ void add_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 	struct list_head *head, *this;
 	struct orphan_inode_entry *new = NULL, *orphan = NULL;
 
-	mutex_lock(&sbi->orphan_inode_mutex);
+	new = f2fs_kmem_cache_alloc(orphan_entry_slab, GFP_ATOMIC);
+	new->ino = ino;
+
+	spin_lock(&sbi->orphan_inode_lock);
 	head = &sbi->orphan_inode_list;
 	list_for_each(this, head) {
 		orphan = list_entry(this, struct orphan_inode_entry, list);
-		if (orphan->ino == ino)
-			goto out;
+		if (orphan->ino == ino) {
+			spin_unlock(&sbi->orphan_inode_lock);
+			kmem_cache_free(orphan_entry_slab, new);
+			return;
+		}
+
 		if (orphan->ino > ino)
 			break;
 		orphan = NULL;
 	}
 
-	new = f2fs_kmem_cache_alloc(orphan_entry_slab, GFP_ATOMIC);
-	new->ino = ino;
-
 	/* add new_oentry into list which is sorted by inode number */
 	if (orphan)
 		list_add(&new->list, this->prev);
 	else
 		list_add_tail(&new->list, head);
-out:
-	mutex_unlock(&sbi->orphan_inode_mutex);
+	spin_unlock(&sbi->orphan_inode_lock);
 }
 
 void remove_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
@@ -246,7 +250,7 @@ void remove_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 	struct list_head *head;
 	struct orphan_inode_entry *orphan;
 
-	mutex_lock(&sbi->orphan_inode_mutex);
+	spin_lock(&sbi->orphan_inode_lock);
 	head = &sbi->orphan_inode_list;
 	list_for_each_entry(orphan, head, list) {
 		if (orphan->ino == ino) {
@@ -257,7 +261,7 @@ void remove_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 			break;
 		}
 	}
-	mutex_unlock(&sbi->orphan_inode_mutex);
+	spin_unlock(&sbi->orphan_inode_lock);
 }
 
 static void recover_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
@@ -270,12 +274,12 @@ static void recover_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 	iput(inode);
 }
 
-int recover_orphan_inodes(struct f2fs_sb_info *sbi)
+void recover_orphan_inodes(struct f2fs_sb_info *sbi)
 {
 	block_t start_blk, orphan_blkaddr, i, j;
 
 	if (!is_set_ckpt_flags(F2FS_CKPT(sbi), CP_ORPHAN_PRESENT_FLAG))
-		return 0;
+		return;
 
 	sbi->por_doing = true;
 	start_blk = __start_cp_addr(sbi) + 1;
@@ -295,29 +299,39 @@ int recover_orphan_inodes(struct f2fs_sb_info *sbi)
 	/* clear Orphan Flag */
 	clear_ckpt_flags(F2FS_CKPT(sbi), CP_ORPHAN_PRESENT_FLAG);
 	sbi->por_doing = false;
-	return 0;
+	return;
 }
 
 static void write_orphan_inodes(struct f2fs_sb_info *sbi, block_t start_blk)
 {
-	struct list_head *head, *this, *next;
+	struct list_head *head;
 	struct f2fs_orphan_block *orphan_blk = NULL;
-	struct page *page = NULL;
 	unsigned int nentries = 0;
-	unsigned short index = 1;
-	unsigned short orphan_blocks;
-
-	orphan_blocks = (unsigned short)((sbi->n_orphans +
+	unsigned short index;
+	unsigned short orphan_blocks = (unsigned short)((sbi->n_orphans +
 		(F2FS_ORPHANS_PER_BLOCK - 1)) / F2FS_ORPHANS_PER_BLOCK);
+	struct page *page = NULL;
+	struct orphan_inode_entry *orphan = NULL;
+
+	for (index = 0; index < orphan_blocks; index++)
+		grab_meta_page(sbi, start_blk + index);
 
-	mutex_lock(&sbi->orphan_inode_mutex);
+	index = 1;
+	spin_lock(&sbi->orphan_inode_lock);
 	head = &sbi->orphan_inode_list;
 
 	/* loop for each orphan inode entry and write them in Jornal block */
-	list_for_each_safe(this, next, head) {
-		struct orphan_inode_entry *orphan;
+	list_for_each_entry(orphan, head, list) {
+		if (!page) {
+			page = find_get_page(META_MAPPING(sbi), start_blk++);
+			f2fs_bug_on(!page);
+			orphan_blk =
+				(struct f2fs_orphan_block *)page_address(page);
+			memset(orphan_blk, 0, sizeof(*orphan_blk));
+			f2fs_put_page(page, 0);
+		}
 
-		orphan = list_entry(this, struct orphan_inode_entry, list);
+		orphan_blk->ino[nentries++] = cpu_to_le32(orphan->ino);
 
 		if (nentries == F2FS_ORPHANS_PER_BLOCK) {
 			/*
@@ -331,29 +345,20 @@ static void write_orphan_inodes(struct f2fs_sb_info *sbi, block_t start_blk)
 			set_page_dirty(page);
 			f2fs_put_page(page, 1);
 			index++;
-			start_blk++;
 			nentries = 0;
 			page = NULL;
 		}
-		if (page)
-			goto page_exist;
+	}
 
-		page = grab_meta_page(sbi, start_blk);
-		orphan_blk = (struct f2fs_orphan_block *)page_address(page);
-		memset(orphan_blk, 0, sizeof(*orphan_blk));
-page_exist:
-		orphan_blk->ino[nentries++] = cpu_to_le32(orphan->ino);
+	if (page) {
+		orphan_blk->blk_addr = cpu_to_le16(index);
+		orphan_blk->blk_count = cpu_to_le16(orphan_blocks);
+		orphan_blk->entry_count = cpu_to_le32(nentries);
+		set_page_dirty(page);
+		f2fs_put_page(page, 1);
 	}
-	if (!page)
-		goto end;
 
-	orphan_blk->blk_addr = cpu_to_le16(index);
-	orphan_blk->blk_count = cpu_to_le16(orphan_blocks);
-	orphan_blk->entry_count = cpu_to_le32(nentries);
-	set_page_dirty(page);
-	f2fs_put_page(page, 1);
-end:
-	mutex_unlock(&sbi->orphan_inode_mutex);
+	spin_unlock(&sbi->orphan_inode_lock);
 }
 
 static struct page *validate_checkpoint(struct f2fs_sb_info *sbi,
@@ -428,7 +433,8 @@ int get_valid_checkpoint(struct f2fs_sb_info *sbi)
 	cp1 = validate_checkpoint(sbi, cp_start_blk_no, &cp1_version);
 
 	/* The second checkpoint pack should start at the next segment */
-	cp_start_blk_no += 1 << le32_to_cpu(fsb->log_blocks_per_seg);
+	cp_start_blk_no += ((unsigned long long)1) <<
+				le32_to_cpu(fsb->log_blocks_per_seg);
 	cp2 = validate_checkpoint(sbi, cp_start_blk_no, &cp2_version);
 
 	if (cp1 && cp2) {
@@ -465,7 +471,7 @@ static int __add_dirty_inode(struct inode *inode, struct dir_inode_entry *new)
 	list_for_each(this, head) {
 		struct dir_inode_entry *entry;
 		entry = list_entry(this, struct dir_inode_entry, list);
-		if (entry->inode == inode)
+		if (unlikely(entry->inode == inode))
 			return -EEXIST;
 	}
 	list_add_tail(&new->list, head);
@@ -513,8 +519,8 @@ void add_dirty_dir_inode(struct inode *inode)
 void remove_dirty_dir_inode(struct inode *inode)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
-	struct list_head *head = &sbi->dir_inode_list;
-	struct list_head *this;
+
+	struct list_head *this, *head;
 
 	if (!S_ISDIR(inode->i_mode))
 		return;
@@ -525,6 +531,7 @@ void remove_dirty_dir_inode(struct inode *inode)
 		return;
 	}
 
+	head = &sbi->dir_inode_list;
 	list_for_each(this, head) {
 		struct dir_inode_entry *entry;
 		entry = list_entry(this, struct dir_inode_entry, list);
@@ -546,11 +553,13 @@ void remove_dirty_dir_inode(struct inode *inode)
 
 struct inode *check_dirty_dir_inode(struct f2fs_sb_info *sbi, nid_t ino)
 {
-	struct list_head *head = &sbi->dir_inode_list;
-	struct list_head *this;
+
+	struct list_head *this, *head;
 	struct inode *inode = NULL;
 
 	spin_lock(&sbi->dir_inode_lock);
+
+	head = &sbi->dir_inode_list;
 	list_for_each(this, head) {
 		struct dir_inode_entry *entry;
 		entry = list_entry(this, struct dir_inode_entry, list);
@@ -565,11 +574,13 @@ struct inode *check_dirty_dir_inode(struct f2fs_sb_info *sbi, nid_t ino)
 
 void sync_dirty_dir_inodes(struct f2fs_sb_info *sbi)
 {
-	struct list_head *head = &sbi->dir_inode_list;
+	struct list_head *head;
 	struct dir_inode_entry *entry;
 	struct inode *inode;
 retry:
 	spin_lock(&sbi->dir_inode_lock);
+
+	head = &sbi->dir_inode_list;
 	if (list_empty(head)) {
 		spin_unlock(&sbi->dir_inode_lock);
 		return;
@@ -585,7 +596,7 @@ retry:
 		 * We should submit bio, since it exists several
 		 * wribacking dentry pages in the freeing inode.
 		 */
-		f2fs_submit_bio(sbi, DATA, true);
+		f2fs_submit_merged_bio(sbi, DATA, WRITE);
 	}
 	goto retry;
 }
@@ -760,8 +771,8 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, bool is_umount)
 	/* wait for previous submitted node/meta pages writeback */
 	wait_on_all_pages_writeback(sbi);
 
-	filemap_fdatawait_range(sbi->node_inode->i_mapping, 0, LONG_MAX);
-	filemap_fdatawait_range(sbi->meta_inode->i_mapping, 0, LONG_MAX);
+	filemap_fdatawait_range(NODE_MAPPING(sbi), 0, LONG_MAX);
+	filemap_fdatawait_range(META_MAPPING(sbi), 0, LONG_MAX);
 
 	/* update user_block_counts */
 	sbi->last_valid_block_count = sbi->total_valid_block_count;
@@ -770,7 +781,7 @@ static void do_checkpoint(struct f2fs_sb_info *sbi, bool is_umount)
 	/* Here, we only have one bio having CP pack */
 	sync_meta_pages(sbi, META_FLUSH, LONG_MAX);
 
-	if (!is_set_ckpt_flags(ckpt, CP_ERROR_FLAG)) {
+	if (unlikely(!is_set_ckpt_flags(ckpt, CP_ERROR_FLAG))) {
 		clear_prefree_segments(sbi);
 		F2FS_RESET_SB_DIRT(sbi);
 	}
@@ -791,9 +802,9 @@ void write_checkpoint(struct f2fs_sb_info *sbi, bool is_umount)
 
 	trace_f2fs_write_checkpoint(sbi->sb, is_umount, "finish block_ops");
 
-	f2fs_submit_bio(sbi, DATA, true);
-	f2fs_submit_bio(sbi, NODE, true);
-	f2fs_submit_bio(sbi, META, true);
+	f2fs_submit_merged_bio(sbi, DATA, WRITE);
+	f2fs_submit_merged_bio(sbi, NODE, WRITE);
+	f2fs_submit_merged_bio(sbi, META, WRITE);
 
 	/*
 	 * update checkpoint pack index
@@ -818,20 +829,28 @@ void write_checkpoint(struct f2fs_sb_info *sbi, bool is_umount)
 
 void init_orphan_info(struct f2fs_sb_info *sbi)
 {
-	mutex_init(&sbi->orphan_inode_mutex);
+	spin_lock_init(&sbi->orphan_inode_lock);
 	INIT_LIST_HEAD(&sbi->orphan_inode_list);
 	sbi->n_orphans = 0;
+	/*
+	 * considering 512 blocks in a segment 8 blocks are needed for cp
+	 * and log segment summaries. Remaining blocks are used to keep
+	 * orphan entries with the limitation one reserved segment
+	 * for cp pack we can have max 1020*504 orphan entries
+	 */
+	sbi->max_orphans = (sbi->blocks_per_seg - 2 - NR_CURSEG_TYPE)
+				* F2FS_ORPHANS_PER_BLOCK;
 }
 
 int __init create_checkpoint_caches(void)
 {
 	orphan_entry_slab = f2fs_kmem_cache_create("f2fs_orphan_entry",
 			sizeof(struct orphan_inode_entry), NULL);
-	if (unlikely(!orphan_entry_slab))
+	if (!orphan_entry_slab)
 		return -ENOMEM;
 	inode_entry_slab = f2fs_kmem_cache_create("f2fs_dirty_dir_entry",
 			sizeof(struct dir_inode_entry), NULL);
-	if (unlikely(!inode_entry_slab)) {
+	if (!inode_entry_slab) {
 		kmem_cache_destroy(orphan_entry_slab);
 		return -ENOMEM;
 	}

fs/f2fs/data.c

@@ -24,6 +24,195 @@
 #include "segment.h"
 #include <trace/events/f2fs.h>
 
+static void f2fs_read_end_io(struct bio *bio, int err)
+{
+	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+
+	do {
+		struct page *page = bvec->bv_page;
+
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+
+		if (unlikely(!uptodate)) {
+			ClearPageUptodate(page);
+			SetPageError(page);
+		} else {
+			SetPageUptodate(page);
+		}
+		unlock_page(page);
+	} while (bvec >= bio->bi_io_vec);
+
+	bio_put(bio);
+}
+
+static void f2fs_write_end_io(struct bio *bio, int err)
+{
+	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct f2fs_sb_info *sbi = F2FS_SB(bvec->bv_page->mapping->host->i_sb);
+
+	do {
+		struct page *page = bvec->bv_page;
+
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+
+		if (unlikely(!uptodate)) {
+			SetPageError(page);
+			set_bit(AS_EIO, &page->mapping->flags);
+			set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG);
+			sbi->sb->s_flags |= MS_RDONLY;
+		}
+		end_page_writeback(page);
+		dec_page_count(sbi, F2FS_WRITEBACK);
+	} while (bvec >= bio->bi_io_vec);
+
+	if (bio->bi_private)
+		complete(bio->bi_private);
+
+	if (!get_pages(sbi, F2FS_WRITEBACK) &&
+			!list_empty(&sbi->cp_wait.task_list))
+		wake_up(&sbi->cp_wait);
+
+	bio_put(bio);
+}
+
+/*
+ * Low-level block read/write IO operations.
+ */
+static struct bio *__bio_alloc(struct f2fs_sb_info *sbi, block_t blk_addr,
+				int npages, bool is_read)
+{
+	struct bio *bio;
+
+	/* No failure on bio allocation */
+	bio = bio_alloc(GFP_NOIO, npages);
+
+	bio->bi_bdev = sbi->sb->s_bdev;
+	bio->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
+	bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;
+
+	return bio;
+}
+
+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(rw)) {
+		trace_f2fs_submit_read_bio(io->sbi->sb, rw,
+						fio->type, io->bio);
+		submit_bio(rw, 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->bio->bi_private = &wait;
+			submit_bio(rw, io->bio);
+			wait_for_completion(&wait);
+		} else {
+			submit_bio(rw, io->bio);
+		}
+	}
+
+	io->bio = NULL;
+}
+
+void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
+				enum page_type type, int rw)
+{
+	enum page_type btype = PAGE_TYPE_OF_BIO(type);
+	struct f2fs_bio_info *io;
+
+	io = is_read_io(rw) ? &sbi->read_io : &sbi->write_io[btype];
+
+	mutex_lock(&io->io_mutex);
+
+	/* change META to META_FLUSH in the checkpoint procedure */
+	if (type >= META_FLUSH) {
+		io->fio.type = META_FLUSH;
+		io->fio.rw = WRITE_FLUSH_FUA | REQ_META | REQ_PRIO;
+	}
+	__submit_merged_bio(io);
+	mutex_unlock(&io->io_mutex);
+}
+
+/*
+ * Fill the locked page with data located in the block address.
+ * Return unlocked page.
+ */
+int f2fs_submit_page_bio(struct f2fs_sb_info *sbi, struct page *page,
+					block_t blk_addr, int rw)
+{
+	struct bio *bio;
+
+	trace_f2fs_submit_page_bio(page, blk_addr, rw);
+
+	/* Allocate a new bio */
+	bio = __bio_alloc(sbi, blk_addr, 1, is_read_io(rw));
+
+	if (bio_add_page(bio, page, PAGE_CACHE_SIZE, 0) < PAGE_CACHE_SIZE) {
+		bio_put(bio);
+		f2fs_put_page(page, 1);
+		return -EFAULT;
+	}
+
+	submit_bio(rw, bio);
+	return 0;
+}
+
+void f2fs_submit_page_mbio(struct f2fs_sb_info *sbi, struct page *page,
+			block_t blk_addr, struct f2fs_io_info *fio)
+{
+	enum page_type btype = PAGE_TYPE_OF_BIO(fio->type);
+	struct f2fs_bio_info *io;
+	bool is_read = is_read_io(fio->rw);
+
+	io = is_read ? &sbi->read_io : &sbi->write_io[btype];
+
+	verify_block_addr(sbi, blk_addr);
+
+	mutex_lock(&io->io_mutex);
+
+	if (!is_read)
+		inc_page_count(sbi, F2FS_WRITEBACK);
+
+	if (io->bio && (io->last_block_in_bio != blk_addr - 1 ||
+						io->fio.rw != fio->rw))
+		__submit_merged_bio(io);
+alloc_new:
+	if (io->bio == NULL) {
+		int bio_blocks = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
+
+		io->bio = __bio_alloc(sbi, blk_addr, bio_blocks, is_read);
+		io->fio = *fio;
+	}
+
+	if (bio_add_page(io->bio, page, PAGE_CACHE_SIZE, 0) <
+							PAGE_CACHE_SIZE) {
+		__submit_merged_bio(io);
+		goto alloc_new;
+	}
+
+	io->last_block_in_bio = blk_addr;
+
+	mutex_unlock(&io->io_mutex);
+	trace_f2fs_submit_page_mbio(page, fio->rw, fio->type, blk_addr);
+}
+
 /*
  * Lock ordering for the change of data block address:
  * ->data_page
@@ -37,7 +226,7 @@ static void __set_data_blkaddr(struct dnode_of_data *dn, block_t new_addr)
 	struct page *node_page = dn->node_page;
 	unsigned int ofs_in_node = dn->ofs_in_node;
 
-	f2fs_wait_on_page_writeback(node_page, NODE, false);
+	f2fs_wait_on_page_writeback(node_page, NODE);
 
 	rn = F2FS_NODE(node_page);
 
@@ -51,19 +240,39 @@ int reserve_new_block(struct dnode_of_data *dn)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(dn->inode->i_sb);
 
-	if (is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC))
+	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
 		return -EPERM;
-	if (!inc_valid_block_count(sbi, dn->inode, 1))
+	if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1)))
 		return -ENOSPC;
 
 	trace_f2fs_reserve_new_block(dn->inode, dn->nid, dn->ofs_in_node);
 
 	__set_data_blkaddr(dn, NEW_ADDR);
 	dn->data_blkaddr = NEW_ADDR;
+	mark_inode_dirty(dn->inode);
 	sync_inode_page(dn);
 	return 0;
 }
 
+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(!need_put && index);
+
+	err = get_dnode_of_data(dn, index, ALLOC_NODE);
+	if (err)
+		return err;
+
+	if (dn->data_blkaddr == NULL_ADDR)
+		err = reserve_new_block(dn);
+	if (err || need_put)
+		f2fs_put_dnode(dn);
+	return err;
+}
+
 static int check_extent_cache(struct inode *inode, pgoff_t pgofs,
 					struct buffer_head *bh_result)
 {
@@ -71,6 +280,9 @@ static int check_extent_cache(struct inode *inode, pgoff_t pgofs,
 	pgoff_t start_fofs, end_fofs;
 	block_t start_blkaddr;
 
+	if (is_inode_flag_set(fi, FI_NO_EXTENT))
+		return 0;
+
 	read_lock(&fi->ext.ext_lock);
 	if (fi->ext.len == 0) {
 		read_unlock(&fi->ext.ext_lock);
@@ -109,6 +321,7 @@ void update_extent_cache(block_t blk_addr, struct dnode_of_data *dn)
 	struct f2fs_inode_info *fi = F2FS_I(dn->inode);
 	pgoff_t fofs, start_fofs, end_fofs;
 	block_t start_blkaddr, end_blkaddr;
+	int need_update = true;
 
 	f2fs_bug_on(blk_addr == NEW_ADDR);
 	fofs = start_bidx_of_node(ofs_of_node(dn->node_page), fi) +
@@ -117,6 +330,9 @@ void update_extent_cache(block_t blk_addr, struct dnode_of_data *dn)
 	/* Update the page address in the parent node */
 	__set_data_blkaddr(dn, blk_addr);
 
+	if (is_inode_flag_set(fi, FI_NO_EXTENT))
+		return;
+
 	write_lock(&fi->ext.ext_lock);
 
 	start_fofs = fi->ext.fofs;
@@ -163,14 +379,21 @@ void update_extent_cache(block_t blk_addr, struct dnode_of_data *dn)
 					fofs - start_fofs + 1;
 			fi->ext.len -= fofs - start_fofs + 1;
 		}
-		goto end_update;
+	} else {
+		need_update = false;
 	}
-	write_unlock(&fi->ext.ext_lock);
-	return;
 
+	/* Finally, if the extent is very fragmented, let's drop the cache. */
+	if (fi->ext.len < F2FS_MIN_EXTENT_LEN) {
+		fi->ext.len = 0;
+		set_inode_flag(fi, FI_NO_EXTENT);
+		need_update = true;
+	}
 end_update:
 	write_unlock(&fi->ext.ext_lock);
-	sync_inode_page(dn);
+	if (need_update)
+		sync_inode_page(dn);
+	return;
 }
 
 struct page *find_data_page(struct inode *inode, pgoff_t index, bool sync)
@@ -196,7 +419,7 @@ struct page *find_data_page(struct inode *inode, pgoff_t index, bool sync)
 		return ERR_PTR(-ENOENT);
 
 	/* By fallocate(), there is no cached page, but with NEW_ADDR */
-	if (dn.data_blkaddr == NEW_ADDR)
+	if (unlikely(dn.data_blkaddr == NEW_ADDR))
 		return ERR_PTR(-EINVAL);
 
 	page = grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS);
@@ -208,11 +431,14 @@ struct page *find_data_page(struct inode *inode, pgoff_t index, bool sync)
 		return page;
 	}
 
-	err = f2fs_readpage(sbi, page, dn.data_blkaddr,
+	err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
 					sync ? READ_SYNC : READA);
+	if (err)
+		return ERR_PTR(err);
+
 	if (sync) {
 		wait_on_page_locked(page);
-		if (!PageUptodate(page)) {
+		if (unlikely(!PageUptodate(page))) {
 			f2fs_put_page(page, 0);
 			return ERR_PTR(-EIO);
 		}
@@ -246,7 +472,7 @@ repeat:
 	}
 	f2fs_put_dnode(&dn);
 
-	if (dn.data_blkaddr == NULL_ADDR) {
+	if (unlikely(dn.data_blkaddr == NULL_ADDR)) {
 		f2fs_put_page(page, 1);
 		return ERR_PTR(-ENOENT);
 	}
@@ -266,16 +492,16 @@ repeat:
 		return page;
 	}
 
-	err = f2fs_readpage(sbi, page, dn.data_blkaddr, READ_SYNC);
+	err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr, READ_SYNC);
 	if (err)
 		return ERR_PTR(err);
 
 	lock_page(page);
-	if (!PageUptodate(page)) {
+	if (unlikely(!PageUptodate(page))) {
 		f2fs_put_page(page, 1);
 		return ERR_PTR(-EIO);
 	}
-	if (page->mapping != mapping) {
+	if (unlikely(page->mapping != mapping)) {
 		f2fs_put_page(page, 1);
 		goto repeat;
 	}
@@ -286,12 +512,12 @@ repeat:
  * Caller ensures that this data page is never allocated.
  * A new zero-filled data page is allocated in the page cache.
  *
- * Also, caller should grab and release a mutex by calling mutex_lock_op() and
- * mutex_unlock_op().
- * Note that, npage is set only by make_empty_dir.
+ * Also, caller should grab and release a rwsem by calling f2fs_lock_op() and
+ * f2fs_unlock_op().
+ * Note that, ipage is set only by make_empty_dir.
  */
 struct page *get_new_data_page(struct inode *inode,
-		struct page *npage, pgoff_t index, bool new_i_size)
+		struct page *ipage, pgoff_t index, bool new_i_size)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	struct address_space *mapping = inode->i_mapping;
@@ -299,24 +525,16 @@ struct page *get_new_data_page(struct inode *inode,
 	struct dnode_of_data dn;
 	int err;
 
-	set_new_dnode(&dn, inode, npage, npage, 0);
-	err = get_dnode_of_data(&dn, index, ALLOC_NODE);
+	set_new_dnode(&dn, inode, ipage, NULL, 0);
+	err = f2fs_reserve_block(&dn, index);
 	if (err)
 		return ERR_PTR(err);
-
-	if (dn.data_blkaddr == NULL_ADDR) {
-		if (reserve_new_block(&dn)) {
-			if (!npage)
-				f2fs_put_dnode(&dn);
-			return ERR_PTR(-ENOSPC);
-		}
-	}
-	if (!npage)
-		f2fs_put_dnode(&dn);
 repeat:
 	page = grab_cache_page(mapping, index);
-	if (!page)
-		return ERR_PTR(-ENOMEM);
+	if (!page) {
+		err = -ENOMEM;
+		goto put_err;
+	}
 
 	if (PageUptodate(page))
 		return page;
@@ -325,15 +543,18 @@ repeat:
 		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
 		SetPageUptodate(page);
 	} else {
-		err = f2fs_readpage(sbi, page, dn.data_blkaddr, READ_SYNC);
+		err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
+								READ_SYNC);
 		if (err)
-			return ERR_PTR(err);
+			goto put_err;
+
 		lock_page(page);
-		if (!PageUptodate(page)) {
+		if (unlikely(!PageUptodate(page))) {
 			f2fs_put_page(page, 1);
-			return ERR_PTR(-EIO);
+			err = -EIO;
+			goto put_err;
 		}
-		if (page->mapping != mapping) {
+		if (unlikely(page->mapping != mapping)) {
 			f2fs_put_page(page, 1);
 			goto repeat;
 		}
@@ -344,140 +565,187 @@ repeat:
 		i_size_write(inode, ((index + 1) << PAGE_CACHE_SHIFT));
 		/* Only the directory inode sets new_i_size */
 		set_inode_flag(F2FS_I(inode), FI_UPDATE_DIR);
-		mark_inode_dirty_sync(inode);
 	}
 	return page;
-}
-
-static void read_end_io(struct bio *bio, int err)
-{
-	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
-	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
 
-	do {
-		struct page *page = bvec->bv_page;
-
-		if (--bvec >= bio->bi_io_vec)
-			prefetchw(&bvec->bv_page->flags);
-
-		if (uptodate) {
-			SetPageUptodate(page);
-		} else {
-			ClearPageUptodate(page);
-			SetPageError(page);
-		}
-		unlock_page(page);
-	} while (bvec >= bio->bi_io_vec);
-	bio_put(bio);
+put_err:
+	f2fs_put_dnode(&dn);
+	return ERR_PTR(err);
 }
 
-/*
- * Fill the locked page with data located in the block address.
- * Return unlocked page.
- */
-int f2fs_readpage(struct f2fs_sb_info *sbi, struct page *page,
-					block_t blk_addr, int type)
+static int __allocate_data_block(struct dnode_of_data *dn)
 {
-	struct block_device *bdev = sbi->sb->s_bdev;
-	struct bio *bio;
+	struct f2fs_sb_info *sbi = F2FS_SB(dn->inode->i_sb);
+	struct f2fs_summary sum;
+	block_t new_blkaddr;
+	struct node_info ni;
+	int type;
 
-	trace_f2fs_readpage(page, blk_addr, type);
+	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
+		return -EPERM;
+	if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1)))
+		return -ENOSPC;
 
-	down_read(&sbi->bio_sem);
+	__set_data_blkaddr(dn, NEW_ADDR);
+	dn->data_blkaddr = NEW_ADDR;
 
-	/* Allocate a new bio */
-	bio = f2fs_bio_alloc(bdev, 1);
+	get_node_info(sbi, dn->nid, &ni);
+	set_summary(&sum, dn->nid, dn->ofs_in_node, ni.version);
 
-	/* Initialize the bio */
-	bio->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
-	bio->bi_end_io = read_end_io;
+	type = CURSEG_WARM_DATA;
 
-	if (bio_add_page(bio, page, PAGE_CACHE_SIZE, 0) < PAGE_CACHE_SIZE) {
-		bio_put(bio);
-		up_read(&sbi->bio_sem);
-		f2fs_put_page(page, 1);
-		return -EFAULT;
-	}
+	allocate_data_block(sbi, NULL, NULL_ADDR, &new_blkaddr, &sum, type);
 
-	submit_bio(type, bio);
-	up_read(&sbi->bio_sem);
+	/* direct IO doesn't use extent cache to maximize the performance */
+	set_inode_flag(F2FS_I(dn->inode), FI_NO_EXTENT);
+	update_extent_cache(new_blkaddr, dn);
+	clear_inode_flag(F2FS_I(dn->inode), FI_NO_EXTENT);
+
+	dn->data_blkaddr = new_blkaddr;
 	return 0;
 }
 
 /*
- * This function should be used by the data read flow only where it
- * does not check the "create" flag that indicates block allocation.
- * The reason for this special functionality is to exploit VFS readahead
- * mechanism.
+ * get_data_block() now supported readahead/bmap/rw direct_IO with mapped bh.
+ * If original data blocks are allocated, then give them to blockdev.
+ * Otherwise,
+ *     a. preallocate requested block addresses
+ *     b. do not use extent cache for better performance
+ *     c. give the block addresses to blockdev
  */
-static int get_data_block_ro(struct inode *inode, sector_t iblock,
+static int get_data_block(struct inode *inode, sector_t iblock,
 			struct buffer_head *bh_result, int create)
 {
+	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	unsigned int blkbits = inode->i_sb->s_blocksize_bits;
 	unsigned maxblocks = bh_result->b_size >> blkbits;
 	struct dnode_of_data dn;
-	pgoff_t pgofs;
-	int err;
+	int mode = create ? ALLOC_NODE : LOOKUP_NODE_RA;
+	pgoff_t pgofs, end_offset;
+	int err = 0, ofs = 1;
+	bool allocated = false;
 
 	/* Get the page offset from the block offset(iblock) */
 	pgofs =	(pgoff_t)(iblock >> (PAGE_CACHE_SHIFT - blkbits));
 
-	if (check_extent_cache(inode, pgofs, bh_result)) {
-		trace_f2fs_get_data_block(inode, iblock, bh_result, 0);
-		return 0;
-	}
+	if (check_extent_cache(inode, pgofs, bh_result))
+		goto out;
+
+	if (create)
+		f2fs_lock_op(sbi);
 
 	/* When reading holes, we need its node page */
 	set_new_dnode(&dn, inode, NULL, NULL, 0);
-	err = get_dnode_of_data(&dn, pgofs, LOOKUP_NODE_RA);
+	err = get_dnode_of_data(&dn, pgofs, mode);
 	if (err) {
-		trace_f2fs_get_data_block(inode, iblock, bh_result, err);
-		return (err == -ENOENT) ? 0 : err;
+		if (err == -ENOENT)
+			err = 0;
+		goto unlock_out;
+	}
+	if (dn.data_blkaddr == NEW_ADDR)
+		goto put_out;
+
+	if (dn.data_blkaddr != NULL_ADDR) {
+		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
+	} else if (create) {
+		err = __allocate_data_block(&dn);
+		if (err)
+			goto put_out;
+		allocated = true;
+		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
+	} else {
+		goto put_out;
 	}
 
-	/* It does not support data allocation */
-	f2fs_bug_on(create);
+	end_offset = IS_INODE(dn.node_page) ?
+			ADDRS_PER_INODE(F2FS_I(inode)) : ADDRS_PER_BLOCK;
+	bh_result->b_size = (((size_t)1) << blkbits);
+	dn.ofs_in_node++;
+	pgofs++;
+
+get_next:
+	if (dn.ofs_in_node >= end_offset) {
+		if (allocated)
+			sync_inode_page(&dn);
+		allocated = false;
+		f2fs_put_dnode(&dn);
 
-	if (dn.data_blkaddr != NEW_ADDR && dn.data_blkaddr != NULL_ADDR) {
-		int i;
-		unsigned int end_offset;
+		set_new_dnode(&dn, inode, NULL, NULL, 0);
+		err = get_dnode_of_data(&dn, pgofs, mode);
+		if (err) {
+			if (err == -ENOENT)
+				err = 0;
+			goto unlock_out;
+		}
+		if (dn.data_blkaddr == NEW_ADDR)
+			goto put_out;
 
 		end_offset = IS_INODE(dn.node_page) ?
-				ADDRS_PER_INODE(F2FS_I(inode)) :
-				ADDRS_PER_BLOCK;
-
-		clear_buffer_new(bh_result);
+			ADDRS_PER_INODE(F2FS_I(inode)) : ADDRS_PER_BLOCK;
+	}
 
+	if (maxblocks > (bh_result->b_size >> blkbits)) {
+		block_t blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
+		if (blkaddr == NULL_ADDR && create) {
+			err = __allocate_data_block(&dn);
+			if (err)
+				goto sync_out;
+			allocated = true;
+			blkaddr = dn.data_blkaddr;
+		}
 		/* Give more consecutive addresses for the read ahead */
-		for (i = 0; i < end_offset - dn.ofs_in_node; i++)
-			if (((datablock_addr(dn.node_page,
-							dn.ofs_in_node + i))
-				!= (dn.data_blkaddr + i)) || maxblocks == i)
-				break;
-		map_bh(bh_result, inode->i_sb, dn.data_blkaddr);
-		bh_result->b_size = (i << blkbits);
+		if (blkaddr == (bh_result->b_blocknr + ofs)) {
+			ofs++;
+			dn.ofs_in_node++;
+			pgofs++;
+			bh_result->b_size += (((size_t)1) << blkbits);
+			goto get_next;
+		}
 	}
+sync_out:
+	if (allocated)
+		sync_inode_page(&dn);
+put_out:
 	f2fs_put_dnode(&dn);
-	trace_f2fs_get_data_block(inode, iblock, bh_result, 0);
-	return 0;
+unlock_out:
+	if (create)
+		f2fs_unlock_op(sbi);
+out:
+	trace_f2fs_get_data_block(inode, iblock, bh_result, err);
+	return err;
 }
 
 static int f2fs_read_data_page(struct file *file, struct page *page)
 {
-	return mpage_readpage(page, get_data_block_ro);
+	struct inode *inode = page->mapping->host;
+	int ret;
+
+	/* If the file has inline data, try to read it directlly */
+	if (f2fs_has_inline_data(inode))
+		ret = f2fs_read_inline_data(inode, page);
+	else
+		ret = mpage_readpage(page, get_data_block);
+
+	return ret;
 }
 
 static int f2fs_read_data_pages(struct file *file,
 			struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, get_data_block_ro);
+	struct inode *inode = file->f_mapping->host;
+
+	/* If the file has inline data, skip readpages */
+	if (f2fs_has_inline_data(inode))
+		return 0;
+
+	return mpage_readpages(mapping, pages, nr_pages, get_data_block);
 }
 
-int do_write_data_page(struct page *page)
+int do_write_data_page(struct page *page, struct f2fs_io_info *fio)
 {
 	struct inode *inode = page->mapping->host;
-	block_t old_blk_addr, new_blk_addr;
+	block_t old_blkaddr, new_blkaddr;
 	struct dnode_of_data dn;
 	int err = 0;
 
@@ -486,10 +754,10 @@ int do_write_data_page(struct page *page)
 	if (err)
 		return err;
 
-	old_blk_addr = dn.data_blkaddr;
+	old_blkaddr = dn.data_blkaddr;
 
 	/* This page is already truncated */
-	if (old_blk_addr == NULL_ADDR)
+	if (old_blkaddr == NULL_ADDR)
 		goto out_writepage;
 
 	set_page_writeback(page);
@@ -498,15 +766,13 @@ int do_write_data_page(struct page *page)
 	 * If current allocation needs SSR,
 	 * it had better in-place writes for updated data.
 	 */
-	if (unlikely(old_blk_addr != NEW_ADDR &&
+	if (unlikely(old_blkaddr != NEW_ADDR &&
 			!is_cold_data(page) &&
 			need_inplace_update(inode))) {
-		rewrite_data_page(F2FS_SB(inode->i_sb), page,
-						old_blk_addr);
+		rewrite_data_page(page, old_blkaddr, fio);
 	} else {
-		write_data_page(inode, page, &dn,
-				old_blk_addr, &new_blk_addr);
-		update_extent_cache(new_blk_addr, &dn);
+		write_data_page(page, &dn, &new_blkaddr, fio);
+		update_extent_cache(new_blkaddr, &dn);
 	}
 out_writepage:
 	f2fs_put_dnode(&dn);
@@ -521,9 +787,13 @@ static int f2fs_write_data_page(struct page *page,
 	loff_t i_size = i_size_read(inode);
 	const pgoff_t end_index = ((unsigned long long) i_size)
 							>> PAGE_CACHE_SHIFT;
-	unsigned offset;
+	unsigned offset = 0;
 	bool need_balance_fs = false;
 	int err = 0;
+	struct f2fs_io_info fio = {
+		.type = DATA,
+		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE,
+	};
 
 	if (page->index < end_index)
 		goto write;
@@ -543,7 +813,7 @@ static int f2fs_write_data_page(struct page *page,
 
 	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
 write:
-	if (sbi->por_doing) {
+	if (unlikely(sbi->por_doing)) {
 		err = AOP_WRITEPAGE_ACTIVATE;
 		goto redirty_out;
 	}
@@ -552,10 +822,18 @@ write:
 	if (S_ISDIR(inode->i_mode)) {
 		dec_page_count(sbi, F2FS_DIRTY_DENTS);
 		inode_dec_dirty_dents(inode);
-		err = do_write_data_page(page);
+		err = do_write_data_page(page, &fio);
 	} else {
 		f2fs_lock_op(sbi);
-		err = do_write_data_page(page);
+
+		if (f2fs_has_inline_data(inode) || f2fs_may_inline(inode)) {
+			err = f2fs_write_inline_data(inode, page, offset);
+			f2fs_unlock_op(sbi);
+			goto out;
+		} else {
+			err = do_write_data_page(page, &fio);
+		}
+
 		f2fs_unlock_op(sbi);
 		need_balance_fs = true;
 	}
@@ -564,8 +842,10 @@ write:
 	else if (err)
 		goto redirty_out;
 
-	if (wbc->for_reclaim)
-		f2fs_submit_bio(sbi, DATA, true);
+	if (wbc->for_reclaim) {
+		f2fs_submit_merged_bio(sbi, DATA, WRITE);
+		need_balance_fs = false;
+	}
 
 	clear_cold_data(page);
 out:
@@ -617,7 +897,8 @@ static int f2fs_write_data_pages(struct address_space *mapping,
 	ret = write_cache_pages(mapping, wbc, __f2fs_writepage, mapping);
 	if (locked)
 		mutex_unlock(&sbi->writepages);
-	f2fs_submit_bio(sbi, DATA, (wbc->sync_mode == WB_SYNC_ALL));
+
+	f2fs_submit_merged_bio(sbi, DATA, WRITE);
 
 	remove_dirty_dir_inode(inode);
 
@@ -638,27 +919,28 @@ static int f2fs_write_begin(struct file *file, struct address_space *mapping,
 
 	f2fs_balance_fs(sbi);
 repeat:
+	err = f2fs_convert_inline_data(inode, pos + len);
+	if (err)
+		return err;
+
 	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 	*pagep = page;
 
-	f2fs_lock_op(sbi);
+	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 = get_dnode_of_data(&dn, index, ALLOC_NODE);
-	if (err)
-		goto err;
-
-	if (dn.data_blkaddr == NULL_ADDR)
-		err = reserve_new_block(&dn);
-
-	f2fs_put_dnode(&dn);
-	if (err)
-		goto err;
-
+	err = f2fs_reserve_block(&dn, index);
 	f2fs_unlock_op(sbi);
 
+	if (err) {
+		f2fs_put_page(page, 1);
+		return err;
+	}
+inline_data:
 	if ((len == PAGE_CACHE_SIZE) || PageUptodate(page))
 		return 0;
 
@@ -674,15 +956,19 @@ repeat:
 	if (dn.data_blkaddr == NEW_ADDR) {
 		zero_user_segment(page, 0, PAGE_CACHE_SIZE);
 	} else {
-		err = f2fs_readpage(sbi, page, dn.data_blkaddr, READ_SYNC);
+		if (f2fs_has_inline_data(inode))
+			err = f2fs_read_inline_data(inode, page);
+		else
+			err = f2fs_submit_page_bio(sbi, page, dn.data_blkaddr,
+							READ_SYNC);
 		if (err)
 			return err;
 		lock_page(page);
-		if (!PageUptodate(page)) {
+		if (unlikely(!PageUptodate(page))) {
 			f2fs_put_page(page, 1);
 			return -EIO;
 		}
-		if (page->mapping != mapping) {
+		if (unlikely(page->mapping != mapping)) {
 			f2fs_put_page(page, 1);
 			goto repeat;
 		}
@@ -691,11 +977,6 @@ out:
 	SetPageUptodate(page);
 	clear_cold_data(page);
 	return 0;
-
-err:
-	f2fs_unlock_op(sbi);
-	f2fs_put_page(page, 1);
-	return err;
 }
 
 static int f2fs_write_end(struct file *file,
@@ -714,23 +995,43 @@ static int f2fs_write_end(struct file *file,
 		update_inode_page(inode);
 	}
 
-	unlock_page(page);
-	page_cache_release(page);
+	f2fs_put_page(page, 1);
 	return copied;
 }
 
+static int check_direct_IO(struct inode *inode, int rw,
+		const struct iovec *iov, loff_t offset, unsigned long nr_segs)
+{
+	unsigned blocksize_mask = inode->i_sb->s_blocksize - 1;
+	int i;
+
+	if (rw == READ)
+		return 0;
+
+	if (offset & blocksize_mask)
+		return -EINVAL;
+
+	for (i = 0; i < nr_segs; i++)
+		if (iov[i].iov_len & blocksize_mask)
+			return -EINVAL;
+	return 0;
+}
+
 static ssize_t f2fs_direct_IO(int rw, struct kiocb *iocb,
 		const struct iovec *iov, loff_t offset, unsigned long nr_segs)
 {
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file->f_mapping->host;
 
-	if (rw == WRITE)
+	/* Let buffer I/O handle the inline data case. */
+	if (f2fs_has_inline_data(inode))
+		return 0;
+
+	if (check_direct_IO(inode, rw, iov, offset, nr_segs))
 		return 0;
 
-	/* Needs synchronization with the cleaner */
 	return blockdev_direct_IO(rw, iocb, inode, iov, offset, nr_segs,
-						  get_data_block_ro);
+							get_data_block);
 }
 
 static void f2fs_invalidate_data_page(struct page *page, unsigned int offset,
@@ -759,6 +1060,8 @@ static int f2fs_set_data_page_dirty(struct page *page)
 	trace_f2fs_set_page_dirty(page, DATA);
 
 	SetPageUptodate(page);
+	mark_inode_dirty(inode);
+
 	if (!PageDirty(page)) {
 		__set_page_dirty_nobuffers(page);
 		set_dirty_dir_page(inode, page);
@@ -769,7 +1072,7 @@ static int f2fs_set_data_page_dirty(struct page *page)
 
 static sector_t f2fs_bmap(struct address_space *mapping, sector_t block)
 {
-	return generic_block_bmap(mapping, block, get_data_block_ro);
+	return generic_block_bmap(mapping, block, get_data_block);
 }
 
 const struct address_space_operations f2fs_dblock_aops = {

fs/f2fs/debug.c

@@ -24,7 +24,7 @@
 #include "gc.h"
 
 static LIST_HEAD(f2fs_stat_list);
-static struct dentry *debugfs_root;
+static struct dentry *f2fs_debugfs_root;
 static DEFINE_MUTEX(f2fs_stat_mutex);
 
 static void update_general_status(struct f2fs_sb_info *sbi)
@@ -45,14 +45,15 @@ static void update_general_status(struct f2fs_sb_info *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->utilization = utilization(sbi);
 
 	si->free_segs = free_segments(sbi);
 	si->free_secs = free_sections(sbi);
 	si->prefree_count = prefree_segments(sbi);
 	si->dirty_count = dirty_segments(sbi);
-	si->node_pages = sbi->node_inode->i_mapping->nrpages;
-	si->meta_pages = sbi->meta_inode->i_mapping->nrpages;
+	si->node_pages = NODE_MAPPING(sbi)->nrpages;
+	si->meta_pages = META_MAPPING(sbi)->nrpages;
 	si->nats = NM_I(sbi)->nat_cnt;
 	si->sits = SIT_I(sbi)->dirty_sentries;
 	si->fnids = NM_I(sbi)->fcnt;
@@ -165,9 +166,9 @@ get_cache:
 	/* free nids */
 	si->cache_mem = NM_I(sbi)->fcnt;
 	si->cache_mem += NM_I(sbi)->nat_cnt;
-	npages = sbi->node_inode->i_mapping->nrpages;
+	npages = NODE_MAPPING(sbi)->nrpages;
 	si->cache_mem += npages << PAGE_CACHE_SHIFT;
-	npages = sbi->meta_inode->i_mapping->nrpages;
+	npages = META_MAPPING(sbi)->nrpages;
 	si->cache_mem += npages << PAGE_CACHE_SHIFT;
 	si->cache_mem += sbi->n_orphans * sizeof(struct orphan_inode_entry);
 	si->cache_mem += sbi->n_dirty_dirs * sizeof(struct dir_inode_entry);
@@ -200,6 +201,8 @@ static int stat_show(struct seq_file *s, void *v)
 		seq_printf(s, "Other: %u)\n  - Data: %u\n",
 			   si->valid_node_count - si->valid_inode_count,
 			   si->valid_count - si->valid_node_count);
+		seq_printf(s, "  - Inline_data Inode: %u\n",
+			   si->inline_inode);
 		seq_printf(s, "\nMain area: %d segs, %d secs %d zones\n",
 			   si->main_area_segs, si->main_area_sections,
 			   si->main_area_zones);
@@ -242,14 +245,14 @@ static int stat_show(struct seq_file *s, void *v)
 		seq_printf(s, "  - node blocks : %d\n", si->node_blks);
 		seq_printf(s, "\nExtent Hit Ratio: %d / %d\n",
 			   si->hit_ext, si->total_ext);
-		seq_printf(s, "\nBalancing F2FS Async:\n");
-		seq_printf(s, "  - nodes %4d in %4d\n",
+		seq_puts(s, "\nBalancing F2FS Async:\n");
+		seq_printf(s, "  - nodes: %4d in %4d\n",
 			   si->ndirty_node, si->node_pages);
-		seq_printf(s, "  - dents %4d in dirs:%4d\n",
+		seq_printf(s, "  - dents: %4d in dirs:%4d\n",
 			   si->ndirty_dent, si->ndirty_dirs);
-		seq_printf(s, "  - meta %4d in %4d\n",
+		seq_printf(s, "  - meta: %4d in %4d\n",
 			   si->ndirty_meta, si->meta_pages);
-		seq_printf(s, "  - NATs %5d > %lu\n",
+		seq_printf(s, "  - NATs: %5d > %lu\n",
 			   si->nats, NM_WOUT_THRESHOLD);
 		seq_printf(s, "  - SITs: %5d\n  - free_nids: %5d\n",
 			   si->sits, si->fnids);
@@ -340,14 +343,32 @@ void f2fs_destroy_stats(struct f2fs_sb_info *sbi)
 
 void __init f2fs_create_root_stats(void)
 {
-	debugfs_root = debugfs_create_dir("f2fs", NULL);
-	if (debugfs_root)
-		debugfs_create_file("status", S_IRUGO, debugfs_root,
-					 NULL, &stat_fops);
+	struct dentry *file;
+
+	f2fs_debugfs_root = debugfs_create_dir("f2fs", NULL);
+	if (!f2fs_debugfs_root)
+		goto bail;
+
+	file = debugfs_create_file("status", S_IRUGO, f2fs_debugfs_root,
+			NULL, &stat_fops);
+	if (!file)
+		goto free_debugfs_dir;
+
+	return;
+
+free_debugfs_dir:
+	debugfs_remove(f2fs_debugfs_root);
+
+bail:
+	f2fs_debugfs_root = NULL;
+	return;
 }
 
 void f2fs_destroy_root_stats(void)
 {
-	debugfs_remove_recursive(debugfs_root);
-	debugfs_root = NULL;
+	if (!f2fs_debugfs_root)
+		return;
+
+	debugfs_remove_recursive(f2fs_debugfs_root);
+	f2fs_debugfs_root = NULL;
 }

fs/f2fs/dir.c

@@ -190,9 +190,6 @@ struct f2fs_dir_entry *f2fs_find_entry(struct inode *dir,
 	unsigned int max_depth;
 	unsigned int level;
 
-	if (namelen > F2FS_NAME_LEN)
-		return NULL;
-
 	if (npages == 0)
 		return NULL;
 
@@ -259,20 +256,17 @@ void f2fs_set_link(struct inode *dir, struct f2fs_dir_entry *de,
 	dir->i_mtime = dir->i_ctime = CURRENT_TIME;
 	mark_inode_dirty(dir);
 
-	/* update parent inode number before releasing dentry page */
-	F2FS_I(inode)->i_pino = dir->i_ino;
-
 	f2fs_put_page(page, 1);
 }
 
 static void init_dent_inode(const struct qstr *name, struct page *ipage)
 {
-	struct f2fs_node *rn;
+	struct f2fs_inode *ri;
 
 	/* copy name info. to this inode page */
-	rn = F2FS_NODE(ipage);
-	rn->i.i_namelen = cpu_to_le32(name->len);
-	memcpy(rn->i.i_name, name->name, name->len);
+	ri = F2FS_INODE(ipage);
+	ri->i_namelen = cpu_to_le32(name->len);
+	memcpy(ri->i_name, name->name, name->len);
 	set_page_dirty(ipage);
 }
 
@@ -348,11 +342,11 @@ static struct page *init_inode_metadata(struct inode *inode,
 
 		err = f2fs_init_acl(inode, dir, page);
 		if (err)
-			goto error;
+			goto put_error;
 
 		err = f2fs_init_security(inode, dir, name, page);
 		if (err)
-			goto error;
+			goto put_error;
 
 		wait_on_page_writeback(page);
 	} else {
@@ -376,8 +370,9 @@ static struct page *init_inode_metadata(struct inode *inode,
 	}
 	return page;
 
-error:
+put_error:
 	f2fs_put_page(page, 1);
+error:
 	remove_inode_page(inode);
 	return ERR_PTR(err);
 }
@@ -393,6 +388,8 @@ static void update_parent_metadata(struct inode *dir, struct inode *inode,
 		clear_inode_flag(F2FS_I(inode), FI_NEW_INODE);
 	}
 	dir->i_mtime = dir->i_ctime = CURRENT_TIME;
+	mark_inode_dirty(dir);
+
 	if (F2FS_I(dir)->i_current_depth != current_depth) {
 		F2FS_I(dir)->i_current_depth = current_depth;
 		set_inode_flag(F2FS_I(dir), FI_UPDATE_DIR);
@@ -400,8 +397,6 @@ static void update_parent_metadata(struct inode *dir, struct inode *inode,
 
 	if (is_inode_flag_set(F2FS_I(dir), FI_UPDATE_DIR))
 		update_inode_page(dir);
-	else
-		mark_inode_dirty(dir);
 
 	if (is_inode_flag_set(F2FS_I(inode), FI_INC_LINK))
 		clear_inode_flag(F2FS_I(inode), FI_INC_LINK);
@@ -432,10 +427,11 @@ next:
 }
 
 /*
- * Caller should grab and release a mutex by calling mutex_lock_op() and
- * mutex_unlock_op().
+ * Caller should grab and release a rwsem by calling f2fs_lock_op() and
+ * f2fs_unlock_op().
  */
-int __f2fs_add_link(struct inode *dir, const struct qstr *name, struct inode *inode)
+int __f2fs_add_link(struct inode *dir, const struct qstr *name,
+						struct inode *inode)
 {
 	unsigned int bit_pos;
 	unsigned int level;
@@ -461,7 +457,7 @@ int __f2fs_add_link(struct inode *dir, const struct qstr *name, struct inode *in
 	}
 
 start:
-	if (current_depth == MAX_DIR_HASH_DEPTH)
+	if (unlikely(current_depth == MAX_DIR_HASH_DEPTH))
 		return -ENOSPC;
 
 	/* Increase the depth, if required */
@@ -554,14 +550,11 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
 
 	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
 
-	if (inode && S_ISDIR(inode->i_mode)) {
-		drop_nlink(dir);
-		update_inode_page(dir);
-	} else {
-		mark_inode_dirty(dir);
-	}
-
 	if (inode) {
+		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)) {
@@ -636,7 +629,7 @@ static int f2fs_readdir(struct file *file, struct dir_context *ctx)
 
 	bit_pos = ((unsigned long)ctx->pos % NR_DENTRY_IN_BLOCK);
 
-	for ( ; n < npages; n++) {
+	for (; n < npages; n++) {
 		dentry_page = get_lock_data_page(inode, n);
 		if (IS_ERR(dentry_page))
 			continue;

fs/f2fs/f2fs.h

@@ -22,8 +22,10 @@
 
 #ifdef CONFIG_F2FS_CHECK_FS
 #define f2fs_bug_on(condition)	BUG_ON(condition)
+#define f2fs_down_write(x, y)	down_write_nest_lock(x, y)
 #else
 #define f2fs_bug_on(condition)
+#define f2fs_down_write(x, y)	down_write(x)
 #endif
 
 /*
@@ -37,6 +39,7 @@
 #define F2FS_MOUNT_POSIX_ACL		0x00000020
 #define F2FS_MOUNT_DISABLE_EXT_IDENTIFY	0x00000040
 #define F2FS_MOUNT_INLINE_XATTR		0x00000080
+#define F2FS_MOUNT_INLINE_DATA		0x00000100
 
 #define clear_opt(sbi, option)	(sbi->mount_opt.opt &= ~F2FS_MOUNT_##option)
 #define set_opt(sbi, option)	(sbi->mount_opt.opt |= F2FS_MOUNT_##option)
@@ -97,6 +100,13 @@ struct dir_inode_entry {
 	struct inode *inode;	/* vfs inode pointer */
 };
 
+/* for the list of blockaddresses to be discarded */
+struct discard_entry {
+	struct list_head list;	/* list head */
+	block_t blkaddr;	/* block address to be discarded */
+	int len;		/* # of consecutive blocks of the discard */
+};
+
 /* for the list of fsync inodes, used only during recovery */
 struct fsync_inode_entry {
 	struct list_head list;	/* list head */
@@ -155,13 +165,15 @@ enum {
 	LOOKUP_NODE,			/* look up a node without readahead */
 	LOOKUP_NODE_RA,			/*
 					 * look up a node with readahead called
-					 * by get_datablock_ro.
+					 * by get_data_block.
 					 */
 };
 
 #define F2FS_LINK_MAX		32000	/* maximum link count per file */
 
 /* for in-memory extent cache entry */
+#define F2FS_MIN_EXTENT_LEN	16	/* minimum extent length */
+
 struct extent_info {
 	rwlock_t ext_lock;	/* rwlock for consistency */
 	unsigned int fofs;	/* start offset in a file */
@@ -308,6 +320,14 @@ struct f2fs_sm_info {
 
 	/* a threshold to reclaim prefree segments */
 	unsigned int rec_prefree_segments;
+
+	/* for small discard management */
+	struct list_head discard_list;		/* 4KB discard list */
+	int nr_discards;			/* # of discards in the list */
+	int max_discards;			/* max. discards to be issued */
+
+	unsigned int ipu_policy;	/* in-place-update policy */
+	unsigned int min_ipu_util;	/* in-place-update threshold */
 };
 
 /*
@@ -338,6 +358,7 @@ enum count_type {
  *			with waiting the bio's completion
  * ...			Only can be used with META.
  */
+#define PAGE_TYPE_OF_BIO(type)	((type) > META ? META : (type))
 enum page_type {
 	DATA,
 	NODE,
@@ -346,6 +367,20 @@ enum page_type {
 	META_FLUSH,
 };
 
+struct f2fs_io_info {
+	enum page_type type;	/* contains DATA/NODE/META/META_FLUSH */
+	int rw;			/* contains R/RS/W/WS with REQ_META/REQ_PRIO */
+};
+
+#define is_read_io(rw)	(((rw) & 1) == READ)
+struct f2fs_bio_info {
+	struct f2fs_sb_info *sbi;	/* f2fs superblock */
+	struct bio *bio;		/* bios to merge */
+	sector_t last_block_in_bio;	/* last block number */
+	struct f2fs_io_info fio;	/* store buffered io info. */
+	struct mutex io_mutex;		/* mutex for bio */
+};
+
 struct f2fs_sb_info {
 	struct super_block *sb;			/* pointer to VFS super block */
 	struct proc_dir_entry *s_proc;		/* proc entry */
@@ -359,9 +394,10 @@ struct f2fs_sb_info {
 
 	/* for segment-related operations */
 	struct f2fs_sm_info *sm_info;		/* segment manager */
-	struct bio *bio[NR_PAGE_TYPE];		/* bios to merge */
-	sector_t last_block_in_bio[NR_PAGE_TYPE];	/* last block number */
-	struct rw_semaphore bio_sem;		/* IO semaphore */
+
+	/* for bio operations */
+	struct f2fs_bio_info read_io;			/* for read bios */
+	struct f2fs_bio_info write_io[NR_PAGE_TYPE];	/* for write bios */
 
 	/* for checkpoint */
 	struct f2fs_checkpoint *ckpt;		/* raw checkpoint pointer */
@@ -376,8 +412,9 @@ struct f2fs_sb_info {
 
 	/* for orphan inode management */
 	struct list_head orphan_inode_list;	/* orphan inode list */
-	struct mutex orphan_inode_mutex;	/* for orphan inode list */
+	spinlock_t orphan_inode_lock;		/* for orphan inode list */
 	unsigned int n_orphans;			/* # of orphan inodes */
+	unsigned int max_orphans;		/* max orphan inodes */
 
 	/* for directory inode management */
 	struct list_head dir_inode_list;	/* dir inode list */
@@ -414,6 +451,9 @@ struct f2fs_sb_info {
 	struct f2fs_gc_kthread	*gc_thread;	/* GC thread */
 	unsigned int cur_victim_sec;		/* current victim section num */
 
+	/* maximum # of trials to find a victim segment for SSR and GC */
+	unsigned int max_victim_search;
+
 	/*
 	 * for stat information.
 	 * one is for the LFS mode, and the other is for the SSR mode.
@@ -423,6 +463,7 @@ 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 */
 	int bg_gc;				/* background gc calls */
 	unsigned int n_dirty_dirs;		/* # of dir inodes */
 #endif
@@ -462,6 +503,11 @@ static inline struct f2fs_node *F2FS_NODE(struct page *page)
 	return (struct f2fs_node *)page_address(page);
 }
 
+static inline struct f2fs_inode *F2FS_INODE(struct page *page)
+{
+	return &((struct f2fs_node *)page_address(page))->i;
+}
+
 static inline struct f2fs_nm_info *NM_I(struct f2fs_sb_info *sbi)
 {
 	return (struct f2fs_nm_info *)(sbi->nm_info);
@@ -487,6 +533,16 @@ static inline struct dirty_seglist_info *DIRTY_I(struct f2fs_sb_info *sbi)
 	return (struct dirty_seglist_info *)(SM_I(sbi)->dirty_info);
 }
 
+static inline struct address_space *META_MAPPING(struct f2fs_sb_info *sbi)
+{
+	return sbi->meta_inode->i_mapping;
+}
+
+static inline struct address_space *NODE_MAPPING(struct f2fs_sb_info *sbi)
+{
+	return sbi->node_inode->i_mapping;
+}
+
 static inline void F2FS_SET_SB_DIRT(struct f2fs_sb_info *sbi)
 {
 	sbi->s_dirty = 1;
@@ -534,7 +590,7 @@ static inline void f2fs_unlock_op(struct f2fs_sb_info *sbi)
 
 static inline void f2fs_lock_all(struct f2fs_sb_info *sbi)
 {
-	down_write_nest_lock(&sbi->cp_rwsem, &sbi->cp_mutex);
+	f2fs_down_write(&sbi->cp_rwsem, &sbi->cp_mutex);
 }
 
 static inline void f2fs_unlock_all(struct f2fs_sb_info *sbi)
@@ -548,7 +604,7 @@ static inline void f2fs_unlock_all(struct f2fs_sb_info *sbi)
 static inline int check_nid_range(struct f2fs_sb_info *sbi, nid_t nid)
 {
 	WARN_ON((nid >= NM_I(sbi)->max_nid));
-	if (nid >= NM_I(sbi)->max_nid)
+	if (unlikely(nid >= NM_I(sbi)->max_nid))
 		return -EINVAL;
 	return 0;
 }
@@ -561,9 +617,9 @@ static inline int check_nid_range(struct f2fs_sb_info *sbi, nid_t nid)
 static inline int F2FS_HAS_BLOCKS(struct inode *inode)
 {
 	if (F2FS_I(inode)->i_xattr_nid)
-		return (inode->i_blocks > F2FS_DEFAULT_ALLOCATED_BLOCKS + 1);
+		return inode->i_blocks > F2FS_DEFAULT_ALLOCATED_BLOCKS + 1;
 	else
-		return (inode->i_blocks > F2FS_DEFAULT_ALLOCATED_BLOCKS);
+		return inode->i_blocks > F2FS_DEFAULT_ALLOCATED_BLOCKS;
 }
 
 static inline bool inc_valid_block_count(struct f2fs_sb_info *sbi,
@@ -574,7 +630,7 @@ static inline bool inc_valid_block_count(struct f2fs_sb_info *sbi,
 	spin_lock(&sbi->stat_lock);
 	valid_block_count =
 		sbi->total_valid_block_count + (block_t)count;
-	if (valid_block_count > sbi->user_block_count) {
+	if (unlikely(valid_block_count > sbi->user_block_count)) {
 		spin_unlock(&sbi->stat_lock);
 		return false;
 	}
@@ -585,7 +641,7 @@ static inline bool inc_valid_block_count(struct f2fs_sb_info *sbi,
 	return true;
 }
 
-static inline int dec_valid_block_count(struct f2fs_sb_info *sbi,
+static inline void dec_valid_block_count(struct f2fs_sb_info *sbi,
 						struct inode *inode,
 						blkcnt_t count)
 {
@@ -595,7 +651,6 @@ static inline int dec_valid_block_count(struct f2fs_sb_info *sbi,
 	inode->i_blocks -= count;
 	sbi->total_valid_block_count -= (block_t)count;
 	spin_unlock(&sbi->stat_lock);
-	return 0;
 }
 
 static inline void inc_page_count(struct f2fs_sb_info *sbi, int count_type)
@@ -686,50 +741,48 @@ static inline block_t __start_sum_addr(struct f2fs_sb_info *sbi)
 }
 
 static inline bool inc_valid_node_count(struct f2fs_sb_info *sbi,
-						struct inode *inode,
-						unsigned int count)
+						struct inode *inode)
 {
 	block_t	valid_block_count;
 	unsigned int valid_node_count;
 
 	spin_lock(&sbi->stat_lock);
 
-	valid_block_count = sbi->total_valid_block_count + (block_t)count;
-	sbi->alloc_valid_block_count += (block_t)count;
-	valid_node_count = sbi->total_valid_node_count + count;
-
-	if (valid_block_count > sbi->user_block_count) {
+	valid_block_count = sbi->total_valid_block_count + 1;
+	if (unlikely(valid_block_count > sbi->user_block_count)) {
 		spin_unlock(&sbi->stat_lock);
 		return false;
 	}
 
-	if (valid_node_count > sbi->total_node_count) {
+	valid_node_count = sbi->total_valid_node_count + 1;
+	if (unlikely(valid_node_count > sbi->total_node_count)) {
 		spin_unlock(&sbi->stat_lock);
 		return false;
 	}
 
 	if (inode)
-		inode->i_blocks += count;
-	sbi->total_valid_node_count = valid_node_count;
-	sbi->total_valid_block_count = valid_block_count;
+		inode->i_blocks++;
+
+	sbi->alloc_valid_block_count++;
+	sbi->total_valid_node_count++;
+	sbi->total_valid_block_count++;
 	spin_unlock(&sbi->stat_lock);
 
 	return true;
 }
 
 static inline void dec_valid_node_count(struct f2fs_sb_info *sbi,
-						struct inode *inode,
-						unsigned int count)
+						struct inode *inode)
 {
 	spin_lock(&sbi->stat_lock);
 
-	f2fs_bug_on(sbi->total_valid_block_count < count);
-	f2fs_bug_on(sbi->total_valid_node_count < count);
-	f2fs_bug_on(inode->i_blocks < count);
+	f2fs_bug_on(!sbi->total_valid_block_count);
+	f2fs_bug_on(!sbi->total_valid_node_count);
+	f2fs_bug_on(!inode->i_blocks);
 
-	inode->i_blocks -= count;
-	sbi->total_valid_node_count -= count;
-	sbi->total_valid_block_count -= (block_t)count;
+	inode->i_blocks--;
+	sbi->total_valid_node_count--;
+	sbi->total_valid_block_count--;
 
 	spin_unlock(&sbi->stat_lock);
 }
@@ -751,13 +804,12 @@ static inline void inc_valid_inode_count(struct f2fs_sb_info *sbi)
 	spin_unlock(&sbi->stat_lock);
 }
 
-static inline int dec_valid_inode_count(struct f2fs_sb_info *sbi)
+static inline void dec_valid_inode_count(struct f2fs_sb_info *sbi)
 {
 	spin_lock(&sbi->stat_lock);
 	f2fs_bug_on(!sbi->total_valid_inode_count);
 	sbi->total_valid_inode_count--;
 	spin_unlock(&sbi->stat_lock);
-	return 0;
 }
 
 static inline unsigned int valid_inode_count(struct f2fs_sb_info *sbi)
@@ -771,7 +823,7 @@ static inline unsigned int valid_inode_count(struct f2fs_sb_info *sbi)
 
 static inline void f2fs_put_page(struct page *page, int unlock)
 {
-	if (!page || IS_ERR(page))
+	if (!page)
 		return;
 
 	if (unlock) {
@@ -876,7 +928,9 @@ enum {
 	FI_NO_ALLOC,		/* should not allocate any blocks */
 	FI_UPDATE_DIR,		/* should update inode block for consistency */
 	FI_DELAY_IPUT,		/* used for the recovery */
+	FI_NO_EXTENT,		/* not to use the extent cache */
 	FI_INLINE_XATTR,	/* used for inline xattr */
+	FI_INLINE_DATA,		/* used for inline data*/
 };
 
 static inline void set_inode_flag(struct f2fs_inode_info *fi, int flag)
@@ -914,6 +968,8 @@ static inline void get_inline_info(struct f2fs_inode_info *fi,
 {
 	if (ri->i_inline & F2FS_INLINE_XATTR)
 		set_inode_flag(fi, FI_INLINE_XATTR);
+	if (ri->i_inline & F2FS_INLINE_DATA)
+		set_inode_flag(fi, FI_INLINE_DATA);
 }
 
 static inline void set_raw_inline(struct f2fs_inode_info *fi,
@@ -923,6 +979,8 @@ static inline void set_raw_inline(struct f2fs_inode_info *fi,
 
 	if (is_inode_flag_set(fi, FI_INLINE_XATTR))
 		ri->i_inline |= F2FS_INLINE_XATTR;
+	if (is_inode_flag_set(fi, FI_INLINE_DATA))
+		ri->i_inline |= F2FS_INLINE_DATA;
 }
 
 static inline unsigned int addrs_per_inode(struct f2fs_inode_info *fi)
@@ -948,6 +1006,18 @@ static inline int inline_xattr_size(struct inode *inode)
 		return 0;
 }
 
+static inline int f2fs_has_inline_data(struct inode *inode)
+{
+	return is_inode_flag_set(F2FS_I(inode), FI_INLINE_DATA);
+}
+
+static inline void *inline_data_addr(struct page *page)
+{
+	struct f2fs_inode *ri;
+	ri = (struct f2fs_inode *)page_address(page);
+	return (void *)&(ri->i_addr[1]);
+}
+
 static inline int f2fs_readonly(struct super_block *sb)
 {
 	return sb->s_flags & MS_RDONLY;
@@ -958,6 +1028,7 @@ static inline int f2fs_readonly(struct super_block *sb)
  */
 int f2fs_sync_file(struct file *, loff_t, loff_t, int);
 void truncate_data_blocks(struct dnode_of_data *);
+int truncate_blocks(struct inode *, u64);
 void f2fs_truncate(struct inode *);
 int f2fs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
 int f2fs_setattr(struct dentry *, struct iattr *);
@@ -1027,7 +1098,7 @@ int get_dnode_of_data(struct dnode_of_data *, pgoff_t, int);
 int truncate_inode_blocks(struct inode *, pgoff_t);
 int truncate_xattr_node(struct inode *, struct page *);
 int wait_on_node_pages_writeback(struct f2fs_sb_info *, nid_t);
-int remove_inode_page(struct inode *);
+void remove_inode_page(struct inode *);
 struct page *new_inode_page(struct inode *, const struct qstr *);
 struct page *new_node_page(struct dnode_of_data *, unsigned int, struct page *);
 void ra_node_page(struct f2fs_sb_info *, nid_t);
@@ -1059,19 +1130,19 @@ void clear_prefree_segments(struct f2fs_sb_info *);
 int npages_for_summary_flush(struct f2fs_sb_info *);
 void allocate_new_segments(struct f2fs_sb_info *);
 struct page *get_sum_page(struct f2fs_sb_info *, unsigned int);
-struct bio *f2fs_bio_alloc(struct block_device *, int);
-void f2fs_submit_bio(struct f2fs_sb_info *, enum page_type, bool);
-void f2fs_wait_on_page_writeback(struct page *, enum page_type, bool);
 void write_meta_page(struct f2fs_sb_info *, struct page *);
-void write_node_page(struct f2fs_sb_info *, struct page *, unsigned int,
-					block_t, block_t *);
-void write_data_page(struct inode *, struct page *, struct dnode_of_data*,
-					block_t, block_t *);
-void rewrite_data_page(struct f2fs_sb_info *, struct page *, block_t);
+void write_node_page(struct f2fs_sb_info *, struct page *,
+		struct f2fs_io_info *, unsigned int, block_t, block_t *);
+void write_data_page(struct page *, struct dnode_of_data *, block_t *,
+					struct f2fs_io_info *);
+void rewrite_data_page(struct page *, block_t, struct f2fs_io_info *);
 void recover_data_page(struct f2fs_sb_info *, struct page *,
 				struct f2fs_summary *, block_t, block_t);
 void rewrite_node_page(struct f2fs_sb_info *, struct page *,
 				struct f2fs_summary *, block_t, block_t);
+void allocate_data_block(struct f2fs_sb_info *, struct page *,
+		block_t, block_t *, struct f2fs_summary *, int);
+void f2fs_wait_on_page_writeback(struct page *, enum page_type);
 void write_data_summaries(struct f2fs_sb_info *, block_t);
 void write_node_summaries(struct f2fs_sb_info *, block_t);
 int lookup_journal_in_cursum(struct f2fs_summary_block *,
@@ -1079,6 +1150,8 @@ int lookup_journal_in_cursum(struct f2fs_summary_block *,
 void flush_sit_entries(struct f2fs_sb_info *);
 int build_segment_manager(struct f2fs_sb_info *);
 void destroy_segment_manager(struct f2fs_sb_info *);
+int __init create_segment_manager_caches(void);
+void destroy_segment_manager_caches(void);
 
 /*
  * checkpoint.c
@@ -1090,7 +1163,7 @@ int acquire_orphan_inode(struct f2fs_sb_info *);
 void release_orphan_inode(struct f2fs_sb_info *);
 void add_orphan_inode(struct f2fs_sb_info *, nid_t);
 void remove_orphan_inode(struct f2fs_sb_info *, nid_t);
-int recover_orphan_inodes(struct f2fs_sb_info *);
+void recover_orphan_inodes(struct f2fs_sb_info *);
 int get_valid_checkpoint(struct f2fs_sb_info *);
 void set_dirty_dir_page(struct inode *, struct page *);
 void add_dirty_dir_inode(struct inode *);
@@ -1105,13 +1178,17 @@ void destroy_checkpoint_caches(void);
 /*
  * data.c
  */
+void f2fs_submit_merged_bio(struct f2fs_sb_info *, enum page_type, int);
+int f2fs_submit_page_bio(struct f2fs_sb_info *, struct page *, block_t, int);
+void f2fs_submit_page_mbio(struct f2fs_sb_info *, struct page *, block_t,
+						struct f2fs_io_info *);
 int reserve_new_block(struct dnode_of_data *);
+int f2fs_reserve_block(struct dnode_of_data *, pgoff_t);
 void update_extent_cache(block_t, struct dnode_of_data *);
 struct page *find_data_page(struct inode *, pgoff_t, bool);
 struct page *get_lock_data_page(struct inode *, pgoff_t);
 struct page *get_new_data_page(struct inode *, struct page *, pgoff_t, bool);
-int f2fs_readpage(struct f2fs_sb_info *, struct page *, block_t, int);
-int do_write_data_page(struct page *);
+int do_write_data_page(struct page *, struct f2fs_io_info *);
 
 /*
  * gc.c
@@ -1144,7 +1221,7 @@ struct f2fs_stat_info {
 	int ndirty_node, ndirty_dent, ndirty_dirs, ndirty_meta;
 	int nats, sits, fnids;
 	int total_count, utilization;
-	int bg_gc;
+	int bg_gc, inline_inode;
 	unsigned int valid_count, valid_node_count, valid_inode_count;
 	unsigned int bimodal, avg_vblocks;
 	int util_free, util_valid, util_invalid;
@@ -1164,7 +1241,7 @@ struct f2fs_stat_info {
 
 static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi)
 {
-	return (struct f2fs_stat_info*)sbi->stat_info;
+	return (struct f2fs_stat_info *)sbi->stat_info;
 }
 
 #define stat_inc_call_count(si)		((si)->call_count++)
@@ -1173,6 +1250,17 @@ static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi)
 #define stat_dec_dirty_dir(sbi)		((sbi)->n_dirty_dirs--)
 #define stat_inc_total_hit(sb)		((F2FS_SB(sb))->total_hit_ext++)
 #define stat_inc_read_hit(sb)		((F2FS_SB(sb))->read_hit_ext++)
+#define stat_inc_inline_inode(inode)					\
+	do {								\
+		if (f2fs_has_inline_data(inode))			\
+			((F2FS_SB(inode->i_sb))->inline_inode++);	\
+	} while (0)
+#define stat_dec_inline_inode(inode)					\
+	do {								\
+		if (f2fs_has_inline_data(inode))			\
+			((F2FS_SB(inode->i_sb))->inline_inode--);	\
+	} while (0)
+
 #define stat_inc_seg_type(sbi, curseg)					\
 		((sbi)->segment_count[(curseg)->alloc_type]++)
 #define stat_inc_block_count(sbi, curseg)				\
@@ -1216,6 +1304,8 @@ void f2fs_destroy_root_stats(void);
 #define stat_dec_dirty_dir(sbi)
 #define stat_inc_total_hit(sb)
 #define stat_inc_read_hit(sb)
+#define stat_inc_inline_inode(inode)
+#define stat_dec_inline_inode(inode)
 #define stat_inc_seg_type(sbi, curseg)
 #define stat_inc_block_count(sbi, curseg)
 #define stat_inc_seg_count(si, type)
@@ -1238,4 +1328,13 @@ extern const struct address_space_operations f2fs_meta_aops;
 extern const struct inode_operations f2fs_dir_inode_operations;
 extern const struct inode_operations f2fs_symlink_inode_operations;
 extern const struct inode_operations f2fs_special_inode_operations;
+
+/*
+ * inline.c
+ */
+bool f2fs_may_inline(struct inode *);
+int f2fs_read_inline_data(struct inode *, struct page *);
+int f2fs_convert_inline_data(struct inode *, pgoff_t);
+int f2fs_write_inline_data(struct inode *, struct page *, unsigned int);
+int recover_inline_data(struct inode *, struct page *);
 #endif

fs/f2fs/file.c

@@ -33,7 +33,6 @@ static int f2fs_vm_page_mkwrite(struct vm_area_struct *vma,
 	struct page *page = vmf->page;
 	struct inode *inode = file_inode(vma->vm_file);
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
-	block_t old_blk_addr;
 	struct dnode_of_data dn;
 	int err;
 
@@ -44,30 +43,16 @@ static int f2fs_vm_page_mkwrite(struct vm_area_struct *vma,
 	/* block allocation */
 	f2fs_lock_op(sbi);
 	set_new_dnode(&dn, inode, NULL, NULL, 0);
-	err = get_dnode_of_data(&dn, page->index, ALLOC_NODE);
-	if (err) {
-		f2fs_unlock_op(sbi);
-		goto out;
-	}
-
-	old_blk_addr = dn.data_blkaddr;
-
-	if (old_blk_addr == NULL_ADDR) {
-		err = reserve_new_block(&dn);
-		if (err) {
-			f2fs_put_dnode(&dn);
-			f2fs_unlock_op(sbi);
-			goto out;
-		}
-	}
-	f2fs_put_dnode(&dn);
+	err = f2fs_reserve_block(&dn, page->index);
 	f2fs_unlock_op(sbi);
+	if (err)
+		goto out;
 
 	file_update_time(vma->vm_file);
 	lock_page(page);
-	if (page->mapping != inode->i_mapping ||
+	if (unlikely(page->mapping != inode->i_mapping ||
 			page_offset(page) > i_size_read(inode) ||
-			!PageUptodate(page)) {
+			!PageUptodate(page))) {
 		unlock_page(page);
 		err = -EFAULT;
 		goto out;
@@ -130,12 +115,12 @@ int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	int ret = 0;
 	bool need_cp = false;
 	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_ALL,
+		.sync_mode = WB_SYNC_NONE,
 		.nr_to_write = LONG_MAX,
 		.for_reclaim = 0,
 	};
 
-	if (f2fs_readonly(inode->i_sb))
+	if (unlikely(f2fs_readonly(inode->i_sb)))
 		return 0;
 
 	trace_f2fs_sync_file_enter(inode);
@@ -217,7 +202,7 @@ int truncate_data_blocks_range(struct dnode_of_data *dn, int count)
 	raw_node = F2FS_NODE(dn->node_page);
 	addr = blkaddr_in_node(raw_node) + ofs;
 
-	for ( ; count > 0; count--, addr++, dn->ofs_in_node++) {
+	for (; count > 0; count--, addr++, dn->ofs_in_node++) {
 		block_t blkaddr = le32_to_cpu(*addr);
 		if (blkaddr == NULL_ADDR)
 			continue;
@@ -256,7 +241,7 @@ static void truncate_partial_data_page(struct inode *inode, u64 from)
 		return;
 
 	lock_page(page);
-	if (page->mapping != inode->i_mapping) {
+	if (unlikely(page->mapping != inode->i_mapping)) {
 		f2fs_put_page(page, 1);
 		return;
 	}
@@ -266,21 +251,24 @@ static void truncate_partial_data_page(struct inode *inode, u64 from)
 	f2fs_put_page(page, 1);
 }
 
-static int truncate_blocks(struct inode *inode, u64 from)
+int truncate_blocks(struct inode *inode, u64 from)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	unsigned int blocksize = inode->i_sb->s_blocksize;
 	struct dnode_of_data dn;
 	pgoff_t free_from;
-	int count = 0;
-	int err;
+	int count = 0, err = 0;
 
 	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));
 
 	f2fs_lock_op(sbi);
+
 	set_new_dnode(&dn, inode, NULL, NULL, 0);
 	err = get_dnode_of_data(&dn, free_from, LOOKUP_NODE);
 	if (err) {
@@ -308,7 +296,7 @@ static int truncate_blocks(struct inode *inode, u64 from)
 free_next:
 	err = truncate_inode_blocks(inode, free_from);
 	f2fs_unlock_op(sbi);
-
+done:
 	/* lastly zero out the first data page */
 	truncate_partial_data_page(inode, from);
 
@@ -382,6 +370,10 @@ int f2fs_setattr(struct dentry *dentry, struct iattr *attr)
 
 	if ((attr->ia_valid & ATTR_SIZE) &&
 			attr->ia_size != i_size_read(inode)) {
+		err = f2fs_convert_inline_data(inode, attr->ia_size);
+		if (err)
+			return err;
+
 		truncate_setsize(inode, attr->ia_size);
 		f2fs_truncate(inode);
 		f2fs_balance_fs(F2FS_SB(inode->i_sb));
@@ -459,12 +451,16 @@ int truncate_hole(struct inode *inode, pgoff_t pg_start, pgoff_t pg_end)
 	return 0;
 }
 
-static int punch_hole(struct inode *inode, loff_t offset, loff_t len, int mode)
+static int punch_hole(struct inode *inode, loff_t offset, loff_t len)
 {
 	pgoff_t pg_start, pg_end;
 	loff_t off_start, off_end;
 	int ret = 0;
 
+	ret = f2fs_convert_inline_data(inode, MAX_INLINE_DATA + 1);
+	if (ret)
+		return ret;
+
 	pg_start = ((unsigned long long) offset) >> PAGE_CACHE_SHIFT;
 	pg_end = ((unsigned long long) offset + len) >> PAGE_CACHE_SHIFT;
 
@@ -499,12 +495,6 @@ static int punch_hole(struct inode *inode, loff_t offset, loff_t len, int mode)
 		}
 	}
 
-	if (!(mode & FALLOC_FL_KEEP_SIZE) &&
-		i_size_read(inode) <= (offset + len)) {
-		i_size_write(inode, offset);
-		mark_inode_dirty(inode);
-	}
-
 	return ret;
 }
 
@@ -521,6 +511,10 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
 	if (ret)
 		return ret;
 
+	ret = f2fs_convert_inline_data(inode, offset + len);
+	if (ret)
+		return ret;
+
 	pg_start = ((unsigned long long) offset) >> PAGE_CACHE_SHIFT;
 	pg_end = ((unsigned long long) offset + len) >> PAGE_CACHE_SHIFT;
 
@@ -532,22 +526,10 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
 
 		f2fs_lock_op(sbi);
 		set_new_dnode(&dn, inode, NULL, NULL, 0);
-		ret = get_dnode_of_data(&dn, index, ALLOC_NODE);
-		if (ret) {
-			f2fs_unlock_op(sbi);
-			break;
-		}
-
-		if (dn.data_blkaddr == NULL_ADDR) {
-			ret = reserve_new_block(&dn);
-			if (ret) {
-				f2fs_put_dnode(&dn);
-				f2fs_unlock_op(sbi);
-				break;
-			}
-		}
-		f2fs_put_dnode(&dn);
+		ret = f2fs_reserve_block(&dn, index);
 		f2fs_unlock_op(sbi);
+		if (ret)
+			break;
 
 		if (pg_start == pg_end)
 			new_size = offset + len;
@@ -578,7 +560,7 @@ static long f2fs_fallocate(struct file *file, int mode,
 		return -EOPNOTSUPP;
 
 	if (mode & FALLOC_FL_PUNCH_HOLE)
-		ret = punch_hole(inode, offset, len, mode);
+		ret = punch_hole(inode, offset, len);
 	else
 		ret = expand_inode_data(inode, offset, len, mode);
 

fs/f2fs/gc.c

@@ -119,7 +119,6 @@ int start_gc_thread(struct f2fs_sb_info *sbi)
 		kfree(gc_th);
 		sbi->gc_thread = NULL;
 	}
-
 out:
 	return err;
 }
@@ -164,8 +163,8 @@ static void select_policy(struct f2fs_sb_info *sbi, int gc_type,
 		p->ofs_unit = sbi->segs_per_sec;
 	}
 
-	if (p->max_search > MAX_VICTIM_SEARCH)
-		p->max_search = MAX_VICTIM_SEARCH;
+	if (p->max_search > sbi->max_victim_search)
+		p->max_search = sbi->max_victim_search;
 
 	p->offset = sbi->last_victim[p->gc_mode];
 }
@@ -429,7 +428,7 @@ next_step:
 
 		/* set page dirty and write it */
 		if (gc_type == FG_GC) {
-			f2fs_wait_on_page_writeback(node_page, NODE, true);
+			f2fs_wait_on_page_writeback(node_page, NODE);
 			set_page_dirty(node_page);
 		} else {
 			if (!PageWriteback(node_page))
@@ -521,6 +520,11 @@ static int check_dnode(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
 
 static void move_data_page(struct inode *inode, struct page *page, int gc_type)
 {
+	struct f2fs_io_info fio = {
+		.type = DATA,
+		.rw = WRITE_SYNC,
+	};
+
 	if (gc_type == BG_GC) {
 		if (PageWriteback(page))
 			goto out;
@@ -529,7 +533,7 @@ static void move_data_page(struct inode *inode, struct page *page, int gc_type)
 	} else {
 		struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 
-		f2fs_wait_on_page_writeback(page, DATA, true);
+		f2fs_wait_on_page_writeback(page, DATA);
 
 		if (clear_page_dirty_for_io(page) &&
 			S_ISDIR(inode->i_mode)) {
@@ -537,7 +541,7 @@ static void move_data_page(struct inode *inode, struct page *page, int gc_type)
 			inode_dec_dirty_dents(inode);
 		}
 		set_cold_data(page);
-		do_write_data_page(page);
+		do_write_data_page(page, &fio);
 		clear_cold_data(page);
 	}
 out:
@@ -631,7 +635,7 @@ next_iput:
 		goto next_step;
 
 	if (gc_type == FG_GC) {
-		f2fs_submit_bio(sbi, DATA, true);
+		f2fs_submit_merged_bio(sbi, DATA, WRITE);
 
 		/*
 		 * In the case of FG_GC, it'd be better to reclaim this victim
@@ -664,8 +668,6 @@ static void do_garbage_collect(struct f2fs_sb_info *sbi, unsigned int segno,
 
 	/* read segment summary of victim */
 	sum_page = get_sum_page(sbi, segno);
-	if (IS_ERR(sum_page))
-		return;
 
 	blk_start_plug(&plug);
 
@@ -697,7 +699,7 @@ int f2fs_gc(struct f2fs_sb_info *sbi)
 
 	INIT_LIST_HEAD(&ilist);
 gc_more:
-	if (!(sbi->sb->s_flags & MS_ACTIVE))
+	if (unlikely(!(sbi->sb->s_flags & MS_ACTIVE)))
 		goto stop;
 
 	if (gc_type == BG_GC && has_not_enough_free_secs(sbi, nfree)) {

fs/f2fs/gc.h

@@ -20,7 +20,7 @@
 #define LIMIT_FREE_BLOCK	40 /* percentage over invalid + free space */
 
 /* Search max. number of dirty segments to select a victim segment */
-#define MAX_VICTIM_SEARCH 4096 /* covers 8GB */
+#define DEF_MAX_VICTIM_SEARCH 4096 /* covers 8GB */
 
 struct f2fs_gc_kthread {
 	struct task_struct *f2fs_gc_task;

fs/f2fs/inline.c

@@ -0,0 +1,222 @@
+/*
+ * fs/f2fs/inline.c
+ * Copyright (c) 2013, Intel Corporation
+ * Authors: Huajun Li <huajun.li@intel.com>
+ *          Haicheng Li <haicheng.li@intel.com>
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/fs.h>
+#include <linux/f2fs_fs.h>
+
+#include "f2fs.h"
+
+bool f2fs_may_inline(struct inode *inode)
+{
+	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
+	block_t nr_blocks;
+	loff_t i_size;
+
+	if (!test_opt(sbi, INLINE_DATA))
+		return false;
+
+	nr_blocks = F2FS_I(inode)->i_xattr_nid ? 3 : 2;
+	if (inode->i_blocks > nr_blocks)
+		return false;
+
+	i_size = i_size_read(inode);
+	if (i_size > MAX_INLINE_DATA)
+		return false;
+
+	return true;
+}
+
+int f2fs_read_inline_data(struct inode *inode, struct page *page)
+{
+	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
+	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(sbi, inode->i_ino);
+	if (IS_ERR(ipage))
+		return PTR_ERR(ipage);
+
+	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);
+	f2fs_put_page(ipage, 1);
+
+out:
+	SetPageUptodate(page);
+	unlock_page(page);
+
+	return 0;
+}
+
+static int __f2fs_convert_inline_data(struct inode *inode, struct page *page)
+{
+	int err;
+	struct page *ipage;
+	struct dnode_of_data dn;
+	void *src_addr, *dst_addr;
+	block_t new_blk_addr;
+	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
+	struct f2fs_io_info fio = {
+		.type = DATA,
+		.rw = WRITE_SYNC | REQ_PRIO,
+	};
+
+	f2fs_lock_op(sbi);
+	ipage = get_node_page(sbi, inode->i_ino);
+	if (IS_ERR(ipage))
+		return PTR_ERR(ipage);
+
+	/*
+	 * i_addr[0] is not used for inline data,
+	 * 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) {
+		f2fs_unlock_op(sbi);
+		return err;
+	}
+
+	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);
+	f2fs_unlock_op(sbi);
+	return err;
+}
+
+int f2fs_convert_inline_data(struct inode *inode, pgoff_t to_size)
+{
+	struct page *page;
+	int err;
+
+	if (!f2fs_has_inline_data(inode))
+		return 0;
+	else if (to_size <= MAX_INLINE_DATA)
+		return 0;
+
+	page = grab_cache_page_write_begin(inode->i_mapping, 0, AOP_FLAG_NOFS);
+	if (!page)
+		return -ENOMEM;
+
+	err = __f2fs_convert_inline_data(inode, page);
+	f2fs_put_page(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;
+
+	set_new_dnode(&dn, inode, NULL, NULL, 0);
+	err = get_dnode_of_data(&dn, 0, LOOKUP_NODE);
+	if (err)
+		return err;
+	ipage = dn.inode_page;
+
+	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);
+		set_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
+		stat_inc_inline_inode(inode);
+	}
+
+	sync_inode_page(&dn);
+	f2fs_put_dnode(&dn);
+
+	return 0;
+}
+
+int recover_inline_data(struct inode *inode, struct page *npage)
+{
+	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
+	struct f2fs_inode *ri = NULL;
+	void *src_addr, *dst_addr;
+	struct page *ipage;
+
+	/*
+	 * The inline_data recovery policy is as follows.
+	 * [prev.] [next] of inline_data flag
+	 *    o       o  -> recover inline_data
+	 *    o       x  -> remove inline_data, and then recover data blocks
+	 *    x       o  -> remove inline_data, and then recover inline_data
+	 *    x       x  -> recover data blocks
+	 */
+	if (IS_INODE(npage))
+		ri = F2FS_INODE(npage);
+
+	if (f2fs_has_inline_data(inode) &&
+			ri && ri->i_inline & F2FS_INLINE_DATA) {
+process_inline:
+		ipage = get_node_page(sbi, inode->i_ino);
+		f2fs_bug_on(IS_ERR(ipage));
+
+		src_addr = inline_data_addr(npage);
+		dst_addr = inline_data_addr(ipage);
+		memcpy(dst_addr, src_addr, MAX_INLINE_DATA);
+		update_inode(inode, ipage);
+		f2fs_put_page(ipage, 1);
+		return -1;
+	}
+
+	if (f2fs_has_inline_data(inode)) {
+		ipage = get_node_page(sbi, inode->i_ino);
+		f2fs_bug_on(IS_ERR(ipage));
+		zero_user_segment(ipage, INLINE_DATA_OFFSET,
+				 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);
+		set_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
+		goto process_inline;
+	}
+	return 0;
+}

fs/f2fs/inode.c

@@ -42,9 +42,11 @@ static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
 	if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode) ||
 			S_ISFIFO(inode->i_mode) || S_ISSOCK(inode->i_mode)) {
 		if (ri->i_addr[0])
-			inode->i_rdev = old_decode_dev(le32_to_cpu(ri->i_addr[0]));
+			inode->i_rdev =
+				old_decode_dev(le32_to_cpu(ri->i_addr[0]));
 		else
-			inode->i_rdev = new_decode_dev(le32_to_cpu(ri->i_addr[1]));
+			inode->i_rdev =
+				new_decode_dev(le32_to_cpu(ri->i_addr[1]));
 	}
 }
 
@@ -52,11 +54,13 @@ static void __set_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
 {
 	if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) {
 		if (old_valid_dev(inode->i_rdev)) {
-			ri->i_addr[0] = cpu_to_le32(old_encode_dev(inode->i_rdev));
+			ri->i_addr[0] =
+				cpu_to_le32(old_encode_dev(inode->i_rdev));
 			ri->i_addr[1] = 0;
 		} else {
 			ri->i_addr[0] = 0;
-			ri->i_addr[1] = cpu_to_le32(new_encode_dev(inode->i_rdev));
+			ri->i_addr[1] =
+				cpu_to_le32(new_encode_dev(inode->i_rdev));
 			ri->i_addr[2] = 0;
 		}
 	}
@@ -67,7 +71,6 @@ static int do_read_inode(struct inode *inode)
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	struct f2fs_inode_info *fi = F2FS_I(inode);
 	struct page *node_page;
-	struct f2fs_node *rn;
 	struct f2fs_inode *ri;
 
 	/* Check if ino is within scope */
@@ -81,8 +84,7 @@ static int do_read_inode(struct inode *inode)
 	if (IS_ERR(node_page))
 		return PTR_ERR(node_page);
 
-	rn = F2FS_NODE(node_page);
-	ri = &(rn->i);
+	ri = F2FS_INODE(node_page);
 
 	inode->i_mode = le16_to_cpu(ri->i_mode);
 	i_uid_write(inode, le32_to_cpu(ri->i_uid));
@@ -175,13 +177,11 @@ bad_inode:
 
 void update_inode(struct inode *inode, struct page *node_page)
 {
-	struct f2fs_node *rn;
 	struct f2fs_inode *ri;
 
-	f2fs_wait_on_page_writeback(node_page, NODE, false);
+	f2fs_wait_on_page_writeback(node_page, NODE);
 
-	rn = F2FS_NODE(node_page);
-	ri = &(rn->i);
+	ri = F2FS_INODE(node_page);
 
 	ri->i_mode = cpu_to_le16(inode->i_mode);
 	ri->i_advise = F2FS_I(inode)->i_advise;
@@ -281,6 +281,7 @@ 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);

fs/f2fs/namei.c

@@ -424,11 +424,13 @@ static int f2fs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		}
 
 		f2fs_set_link(new_dir, new_entry, new_page, old_inode);
+		F2FS_I(old_inode)->i_pino = new_dir->i_ino;
 
 		new_inode->i_ctime = CURRENT_TIME;
 		if (old_dir_entry)
 			drop_nlink(new_inode);
 		drop_nlink(new_inode);
+		mark_inode_dirty(new_inode);
 
 		if (!new_inode->i_nlink)
 			add_orphan_inode(sbi, new_inode->i_ino);
@@ -457,11 +459,14 @@ static int f2fs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		if (old_dir != new_dir) {
 			f2fs_set_link(old_inode, old_dir_entry,
 						old_dir_page, new_dir);
+			F2FS_I(old_inode)->i_pino = new_dir->i_ino;
+			update_inode_page(old_inode);
 		} else {
 			kunmap(old_dir_page);
 			f2fs_put_page(old_dir_page, 0);
 		}
 		drop_nlink(old_dir);
+		mark_inode_dirty(old_dir);
 		update_inode_page(old_dir);
 	}
 

fs/f2fs/node.c

@@ -87,17 +87,19 @@ static struct page *get_next_nat_page(struct f2fs_sb_info *sbi, nid_t nid)
  */
 static void ra_nat_pages(struct f2fs_sb_info *sbi, int nid)
 {
-	struct address_space *mapping = sbi->meta_inode->i_mapping;
+	struct address_space *mapping = META_MAPPING(sbi);
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
-	struct blk_plug plug;
 	struct page *page;
 	pgoff_t index;
 	int i;
+	struct f2fs_io_info fio = {
+		.type = META,
+		.rw = READ_SYNC | REQ_META | REQ_PRIO
+	};
 
-	blk_start_plug(&plug);
 
 	for (i = 0; i < FREE_NID_PAGES; i++, nid += NAT_ENTRY_PER_BLOCK) {
-		if (nid >= nm_i->max_nid)
+		if (unlikely(nid >= nm_i->max_nid))
 			nid = 0;
 		index = current_nat_addr(sbi, nid);
 
@@ -105,15 +107,15 @@ static void ra_nat_pages(struct f2fs_sb_info *sbi, int nid)
 		if (!page)
 			continue;
 		if (PageUptodate(page)) {
+			mark_page_accessed(page);
 			f2fs_put_page(page, 1);
 			continue;
 		}
-		if (f2fs_readpage(sbi, page, index, READ))
-			continue;
-
+		f2fs_submit_page_mbio(sbi, page, index, &fio);
+		mark_page_accessed(page);
 		f2fs_put_page(page, 0);
 	}
-	blk_finish_plug(&plug);
+	f2fs_submit_merged_bio(sbi, META, READ);
 }
 
 static struct nat_entry *__lookup_nat_cache(struct f2fs_nm_info *nm_i, nid_t n)
@@ -391,8 +393,8 @@ got:
 
 /*
  * Caller should call f2fs_put_dnode(dn).
- * Also, it should grab and release a mutex by calling mutex_lock_op() and
- * mutex_unlock_op() only if ro is not set RDONLY_NODE.
+ * Also, it should grab and release a rwsem by calling f2fs_lock_op() and
+ * f2fs_unlock_op() only if ro is not set RDONLY_NODE.
  * In the case of RDONLY_NODE, we don't need to care about mutex.
  */
 int get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode)
@@ -502,7 +504,7 @@ static void truncate_node(struct dnode_of_data *dn)
 
 	/* Deallocate node address */
 	invalidate_blocks(sbi, ni.blk_addr);
-	dec_valid_node_count(sbi, dn->inode, 1);
+	dec_valid_node_count(sbi, dn->inode);
 	set_node_addr(sbi, &ni, NULL_ADDR);
 
 	if (dn->nid == dn->inode->i_ino) {
@@ -516,6 +518,10 @@ invalidate:
 	F2FS_SET_SB_DIRT(sbi);
 
 	f2fs_put_page(dn->node_page, 1);
+
+	invalidate_mapping_pages(NODE_MAPPING(sbi),
+			dn->node_page->index, dn->node_page->index);
+
 	dn->node_page = NULL;
 	trace_f2fs_truncate_node(dn->inode, dn->nid, ni.blk_addr);
 }
@@ -631,19 +637,19 @@ static int truncate_partial_nodes(struct dnode_of_data *dn,
 		return 0;
 
 	/* get indirect nodes in the path */
-	for (i = 0; i < depth - 1; i++) {
+	for (i = 0; i < idx + 1; i++) {
 		/* refernece count'll be increased */
 		pages[i] = get_node_page(sbi, nid[i]);
 		if (IS_ERR(pages[i])) {
-			depth = i + 1;
 			err = PTR_ERR(pages[i]);
+			idx = i - 1;
 			goto fail;
 		}
 		nid[i + 1] = get_nid(pages[i], offset[i + 1], false);
 	}
 
 	/* free direct nodes linked to a partial indirect node */
-	for (i = offset[depth - 1]; i < NIDS_PER_BLOCK; i++) {
+	for (i = offset[idx + 1]; i < NIDS_PER_BLOCK; i++) {
 		child_nid = get_nid(pages[idx], i, false);
 		if (!child_nid)
 			continue;
@@ -654,7 +660,7 @@ static int truncate_partial_nodes(struct dnode_of_data *dn,
 		set_nid(pages[idx], i, 0, false);
 	}
 
-	if (offset[depth - 1] == 0) {
+	if (offset[idx + 1] == 0) {
 		dn->node_page = pages[idx];
 		dn->nid = nid[idx];
 		truncate_node(dn);
@@ -662,9 +668,10 @@ static int truncate_partial_nodes(struct dnode_of_data *dn,
 		f2fs_put_page(pages[idx], 1);
 	}
 	offset[idx]++;
-	offset[depth - 1] = 0;
+	offset[idx + 1] = 0;
+	idx--;
 fail:
-	for (i = depth - 3; i >= 0; i--)
+	for (i = idx; i >= 0; i--)
 		f2fs_put_page(pages[i], 1);
 
 	trace_f2fs_truncate_partial_nodes(dn->inode, nid, depth, err);
@@ -678,11 +685,10 @@ fail:
 int truncate_inode_blocks(struct inode *inode, pgoff_t from)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
-	struct address_space *node_mapping = sbi->node_inode->i_mapping;
 	int err = 0, cont = 1;
 	int level, offset[4], noffset[4];
 	unsigned int nofs = 0;
-	struct f2fs_node *rn;
+	struct f2fs_inode *ri;
 	struct dnode_of_data dn;
 	struct page *page;
 
@@ -699,7 +705,7 @@ restart:
 	set_new_dnode(&dn, inode, page, NULL, 0);
 	unlock_page(page);
 
-	rn = F2FS_NODE(page);
+	ri = F2FS_INODE(page);
 	switch (level) {
 	case 0:
 	case 1:
@@ -709,7 +715,7 @@ restart:
 		nofs = noffset[1];
 		if (!offset[level - 1])
 			goto skip_partial;
-		err = truncate_partial_nodes(&dn, &rn->i, offset, level);
+		err = truncate_partial_nodes(&dn, ri, offset, level);
 		if (err < 0 && err != -ENOENT)
 			goto fail;
 		nofs += 1 + NIDS_PER_BLOCK;
@@ -718,7 +724,7 @@ restart:
 		nofs = 5 + 2 * NIDS_PER_BLOCK;
 		if (!offset[level - 1])
 			goto skip_partial;
-		err = truncate_partial_nodes(&dn, &rn->i, offset, level);
+		err = truncate_partial_nodes(&dn, ri, offset, level);
 		if (err < 0 && err != -ENOENT)
 			goto fail;
 		break;
@@ -728,7 +734,7 @@ restart:
 
 skip_partial:
 	while (cont) {
-		dn.nid = le32_to_cpu(rn->i.i_nid[offset[0] - NODE_DIR1_BLOCK]);
+		dn.nid = le32_to_cpu(ri->i_nid[offset[0] - NODE_DIR1_BLOCK]);
 		switch (offset[0]) {
 		case NODE_DIR1_BLOCK:
 		case NODE_DIR2_BLOCK:
@@ -751,14 +757,14 @@ skip_partial:
 		if (err < 0 && err != -ENOENT)
 			goto fail;
 		if (offset[1] == 0 &&
-				rn->i.i_nid[offset[0] - NODE_DIR1_BLOCK]) {
+				ri->i_nid[offset[0] - NODE_DIR1_BLOCK]) {
 			lock_page(page);
-			if (page->mapping != node_mapping) {
+			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
 				f2fs_put_page(page, 1);
 				goto restart;
 			}
 			wait_on_page_writeback(page);
-			rn->i.i_nid[offset[0] - NODE_DIR1_BLOCK] = 0;
+			ri->i_nid[offset[0] - NODE_DIR1_BLOCK] = 0;
 			set_page_dirty(page);
 			unlock_page(page);
 		}
@@ -794,38 +800,34 @@ int truncate_xattr_node(struct inode *inode, struct page *page)
 	set_new_dnode(&dn, inode, page, npage, nid);
 
 	if (page)
-		dn.inode_page_locked = 1;
+		dn.inode_page_locked = true;
 	truncate_node(&dn);
 	return 0;
 }
 
 /*
- * Caller should grab and release a mutex by calling mutex_lock_op() and
- * mutex_unlock_op().
+ * Caller should grab and release a rwsem by calling f2fs_lock_op() and
+ * f2fs_unlock_op().
  */
-int remove_inode_page(struct inode *inode)
+void remove_inode_page(struct inode *inode)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
 	struct page *page;
 	nid_t ino = inode->i_ino;
 	struct dnode_of_data dn;
-	int err;
 
 	page = get_node_page(sbi, ino);
 	if (IS_ERR(page))
-		return PTR_ERR(page);
+		return;
 
-	err = truncate_xattr_node(inode, page);
-	if (err) {
+	if (truncate_xattr_node(inode, page)) {
 		f2fs_put_page(page, 1);
-		return err;
+		return;
 	}
-
 	/* 0 is possible, after f2fs_new_inode() is failed */
 	f2fs_bug_on(inode->i_blocks != 0 && inode->i_blocks != 1);
 	set_new_dnode(&dn, inode, page, page, ino);
 	truncate_node(&dn);
-	return 0;
 }
 
 struct page *new_inode_page(struct inode *inode, const struct qstr *name)
@@ -843,19 +845,18 @@ struct page *new_node_page(struct dnode_of_data *dn,
 				unsigned int ofs, struct page *ipage)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(dn->inode->i_sb);
-	struct address_space *mapping = sbi->node_inode->i_mapping;
 	struct node_info old_ni, new_ni;
 	struct page *page;
 	int err;
 
-	if (is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC))
+	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
 		return ERR_PTR(-EPERM);
 
-	page = grab_cache_page(mapping, dn->nid);
+	page = grab_cache_page(NODE_MAPPING(sbi), dn->nid);
 	if (!page)
 		return ERR_PTR(-ENOMEM);
 
-	if (!inc_valid_node_count(sbi, dn->inode, 1)) {
+	if (unlikely(!inc_valid_node_count(sbi, dn->inode))) {
 		err = -ENOSPC;
 		goto fail;
 	}
@@ -898,14 +899,14 @@ fail:
  * LOCKED_PAGE: f2fs_put_page(page, 1)
  * error: nothing
  */
-static int read_node_page(struct page *page, int type)
+static int read_node_page(struct page *page, int rw)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(page->mapping->host->i_sb);
 	struct node_info ni;
 
 	get_node_info(sbi, page->index, &ni);
 
-	if (ni.blk_addr == NULL_ADDR) {
+	if (unlikely(ni.blk_addr == NULL_ADDR)) {
 		f2fs_put_page(page, 1);
 		return -ENOENT;
 	}
@@ -913,7 +914,7 @@ static int read_node_page(struct page *page, int type)
 	if (PageUptodate(page))
 		return LOCKED_PAGE;
 
-	return f2fs_readpage(sbi, page, ni.blk_addr, type);
+	return f2fs_submit_page_bio(sbi, page, ni.blk_addr, rw);
 }
 
 /*
@@ -921,18 +922,17 @@ static int read_node_page(struct page *page, int type)
  */
 void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid)
 {
-	struct address_space *mapping = sbi->node_inode->i_mapping;
 	struct page *apage;
 	int err;
 
-	apage = find_get_page(mapping, nid);
+	apage = find_get_page(NODE_MAPPING(sbi), nid);
 	if (apage && PageUptodate(apage)) {
 		f2fs_put_page(apage, 0);
 		return;
 	}
 	f2fs_put_page(apage, 0);
 
-	apage = grab_cache_page(mapping, nid);
+	apage = grab_cache_page(NODE_MAPPING(sbi), nid);
 	if (!apage)
 		return;
 
@@ -945,11 +945,10 @@ void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid)
 
 struct page *get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid)
 {
-	struct address_space *mapping = sbi->node_inode->i_mapping;
 	struct page *page;
 	int err;
 repeat:
-	page = grab_cache_page(mapping, nid);
+	page = grab_cache_page(NODE_MAPPING(sbi), nid);
 	if (!page)
 		return ERR_PTR(-ENOMEM);
 
@@ -960,11 +959,11 @@ repeat:
 		goto got_it;
 
 	lock_page(page);
-	if (!PageUptodate(page)) {
+	if (unlikely(!PageUptodate(page))) {
 		f2fs_put_page(page, 1);
 		return ERR_PTR(-EIO);
 	}
-	if (page->mapping != mapping) {
+	if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
 		f2fs_put_page(page, 1);
 		goto repeat;
 	}
@@ -981,7 +980,6 @@ got_it:
 struct page *get_node_page_ra(struct page *parent, int start)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(parent->mapping->host->i_sb);
-	struct address_space *mapping = sbi->node_inode->i_mapping;
 	struct blk_plug plug;
 	struct page *page;
 	int err, i, end;
@@ -992,7 +990,7 @@ struct page *get_node_page_ra(struct page *parent, int start)
 	if (!nid)
 		return ERR_PTR(-ENOENT);
 repeat:
-	page = grab_cache_page(mapping, nid);
+	page = grab_cache_page(NODE_MAPPING(sbi), nid);
 	if (!page)
 		return ERR_PTR(-ENOMEM);
 
@@ -1017,12 +1015,12 @@ repeat:
 	blk_finish_plug(&plug);
 
 	lock_page(page);
-	if (page->mapping != mapping) {
+	if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
 		f2fs_put_page(page, 1);
 		goto repeat;
 	}
 page_hit:
-	if (!PageUptodate(page)) {
+	if (unlikely(!PageUptodate(page))) {
 		f2fs_put_page(page, 1);
 		return ERR_PTR(-EIO);
 	}
@@ -1048,7 +1046,6 @@ void sync_inode_page(struct dnode_of_data *dn)
 int sync_node_pages(struct f2fs_sb_info *sbi, nid_t ino,
 					struct writeback_control *wbc)
 {
-	struct address_space *mapping = sbi->node_inode->i_mapping;
 	pgoff_t index, end;
 	struct pagevec pvec;
 	int step = ino ? 2 : 0;
@@ -1062,7 +1059,7 @@ next_step:
 
 	while (index <= end) {
 		int i, nr_pages;
-		nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
+		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
 				PAGECACHE_TAG_DIRTY,
 				min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
 		if (nr_pages == 0)
@@ -1095,7 +1092,7 @@ next_step:
 			else if (!trylock_page(page))
 				continue;
 
-			if (unlikely(page->mapping != mapping)) {
+			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
 continue_unlock:
 				unlock_page(page);
 				continue;
@@ -1122,7 +1119,7 @@ continue_unlock:
 				set_fsync_mark(page, 0);
 				set_dentry_mark(page, 0);
 			}
-			mapping->a_ops->writepage(page, wbc);
+			NODE_MAPPING(sbi)->a_ops->writepage(page, wbc);
 			wrote++;
 
 			if (--wbc->nr_to_write == 0)
@@ -1143,31 +1140,31 @@ continue_unlock:
 	}
 
 	if (wrote)
-		f2fs_submit_bio(sbi, NODE, wbc->sync_mode == WB_SYNC_ALL);
-
+		f2fs_submit_merged_bio(sbi, NODE, WRITE);
 	return nwritten;
 }
 
 int wait_on_node_pages_writeback(struct f2fs_sb_info *sbi, nid_t ino)
 {
-	struct address_space *mapping = sbi->node_inode->i_mapping;
 	pgoff_t index = 0, end = LONG_MAX;
 	struct pagevec pvec;
-	int nr_pages;
 	int ret2 = 0, ret = 0;
 
 	pagevec_init(&pvec, 0);
-	while ((index <= end) &&
-			(nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
-			PAGECACHE_TAG_WRITEBACK,
-			min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1)) != 0) {
-		unsigned i;
+
+	while (index <= end) {
+		int i, nr_pages;
+		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
+				PAGECACHE_TAG_WRITEBACK,
+				min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
+		if (nr_pages == 0)
+			break;
 
 		for (i = 0; i < nr_pages; i++) {
 			struct page *page = pvec.pages[i];
 
 			/* until radix tree lookup accepts end_index */
-			if (page->index > end)
+			if (unlikely(page->index > end))
 				continue;
 
 			if (ino && ino_of_node(page) == ino) {
@@ -1180,9 +1177,9 @@ int wait_on_node_pages_writeback(struct f2fs_sb_info *sbi, nid_t ino)
 		cond_resched();
 	}
 
-	if (test_and_clear_bit(AS_ENOSPC, &mapping->flags))
+	if (unlikely(test_and_clear_bit(AS_ENOSPC, &NODE_MAPPING(sbi)->flags)))
 		ret2 = -ENOSPC;
-	if (test_and_clear_bit(AS_EIO, &mapping->flags))
+	if (unlikely(test_and_clear_bit(AS_EIO, &NODE_MAPPING(sbi)->flags)))
 		ret2 = -EIO;
 	if (!ret)
 		ret = ret2;
@@ -1196,8 +1193,12 @@ static int f2fs_write_node_page(struct page *page,
 	nid_t nid;
 	block_t new_addr;
 	struct node_info ni;
+	struct f2fs_io_info fio = {
+		.type = NODE,
+		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE,
+	};
 
-	if (sbi->por_doing)
+	if (unlikely(sbi->por_doing))
 		goto redirty_out;
 
 	wait_on_page_writeback(page);
@@ -1209,7 +1210,7 @@ static int f2fs_write_node_page(struct page *page,
 	get_node_info(sbi, nid, &ni);
 
 	/* This page is already truncated */
-	if (ni.blk_addr == NULL_ADDR) {
+	if (unlikely(ni.blk_addr == NULL_ADDR)) {
 		dec_page_count(sbi, F2FS_DIRTY_NODES);
 		unlock_page(page);
 		return 0;
@@ -1220,7 +1221,7 @@ static int f2fs_write_node_page(struct page *page,
 
 	mutex_lock(&sbi->node_write);
 	set_page_writeback(page);
-	write_node_page(sbi, page, nid, ni.blk_addr, &new_addr);
+	write_node_page(sbi, page, &fio, nid, ni.blk_addr, &new_addr);
 	set_node_addr(sbi, &ni, new_addr);
 	dec_page_count(sbi, F2FS_DIRTY_NODES);
 	mutex_unlock(&sbi->node_write);
@@ -1255,6 +1256,7 @@ static int f2fs_write_node_pages(struct address_space *mapping,
 
 	/* if mounting is failed, skip writing node pages */
 	wbc->nr_to_write = 3 * max_hw_blocks(sbi);
+	wbc->sync_mode = WB_SYNC_NONE;
 	sync_node_pages(sbi, 0, wbc);
 	wbc->nr_to_write = nr_to_write - (3 * max_hw_blocks(sbi) -
 						wbc->nr_to_write);
@@ -1333,7 +1335,7 @@ static int add_free_nid(struct f2fs_nm_info *nm_i, nid_t nid, bool build)
 		return -1;
 
 	/* 0 nid should not be used */
-	if (nid == 0)
+	if (unlikely(nid == 0))
 		return 0;
 
 	if (build) {
@@ -1386,7 +1388,7 @@ static void scan_nat_page(struct f2fs_nm_info *nm_i,
 
 	for (; i < NAT_ENTRY_PER_BLOCK; i++, start_nid++) {
 
-		if (start_nid >= nm_i->max_nid)
+		if (unlikely(start_nid >= nm_i->max_nid))
 			break;
 
 		blk_addr = le32_to_cpu(nat_blk->entries[i].block_addr);
@@ -1420,7 +1422,7 @@ static void build_free_nids(struct f2fs_sb_info *sbi)
 		f2fs_put_page(page, 1);
 
 		nid += (NAT_ENTRY_PER_BLOCK - (nid % NAT_ENTRY_PER_BLOCK));
-		if (nid >= nm_i->max_nid)
+		if (unlikely(nid >= nm_i->max_nid))
 			nid = 0;
 
 		if (i++ == FREE_NID_PAGES)
@@ -1454,7 +1456,7 @@ bool alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid)
 	struct free_nid *i = NULL;
 	struct list_head *this;
 retry:
-	if (sbi->total_valid_node_count + 1 >= nm_i->max_nid)
+	if (unlikely(sbi->total_valid_node_count + 1 >= nm_i->max_nid))
 		return false;
 
 	spin_lock(&nm_i->free_nid_list_lock);
@@ -1535,13 +1537,12 @@ void recover_node_page(struct f2fs_sb_info *sbi, struct page *page,
 
 int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
 {
-	struct address_space *mapping = sbi->node_inode->i_mapping;
-	struct f2fs_node *src, *dst;
+	struct f2fs_inode *src, *dst;
 	nid_t ino = ino_of_node(page);
 	struct node_info old_ni, new_ni;
 	struct page *ipage;
 
-	ipage = grab_cache_page(mapping, ino);
+	ipage = grab_cache_page(NODE_MAPPING(sbi), ino);
 	if (!ipage)
 		return -ENOMEM;
 
@@ -1552,19 +1553,19 @@ int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
 	SetPageUptodate(ipage);
 	fill_node_footer(ipage, ino, ino, 0, true);
 
-	src = F2FS_NODE(page);
-	dst = F2FS_NODE(ipage);
+	src = F2FS_INODE(page);
+	dst = F2FS_INODE(ipage);
 
-	memcpy(dst, src, (unsigned long)&src->i.i_ext - (unsigned long)&src->i);
-	dst->i.i_size = 0;
-	dst->i.i_blocks = cpu_to_le64(1);
-	dst->i.i_links = cpu_to_le32(1);
-	dst->i.i_xattr_nid = 0;
+	memcpy(dst, src, (unsigned long)&src->i_ext - (unsigned long)src);
+	dst->i_size = 0;
+	dst->i_blocks = cpu_to_le64(1);
+	dst->i_links = cpu_to_le32(1);
+	dst->i_xattr_nid = 0;
 
 	new_ni = old_ni;
 	new_ni.ino = ino;
 
-	if (!inc_valid_node_count(sbi, NULL, 1))
+	if (unlikely(!inc_valid_node_count(sbi, NULL)))
 		WARN_ON(1);
 	set_node_addr(sbi, &new_ni, NEW_ADDR);
 	inc_valid_inode_count(sbi);
@@ -1572,47 +1573,88 @@ int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
 	return 0;
 }
 
+/*
+ * ra_sum_pages() merge contiguous pages into one bio and submit.
+ * these pre-readed pages are linked in pages list.
+ */
+static int ra_sum_pages(struct f2fs_sb_info *sbi, struct list_head *pages,
+				int start, int nrpages)
+{
+	struct page *page;
+	int page_idx = start;
+	struct f2fs_io_info fio = {
+		.type = META,
+		.rw = READ_SYNC | REQ_META | REQ_PRIO
+	};
+
+	for (; page_idx < start + nrpages; page_idx++) {
+		/* alloc temporal page for read node summary info*/
+		page = alloc_page(GFP_F2FS_ZERO);
+		if (!page) {
+			struct page *tmp;
+			list_for_each_entry_safe(page, tmp, pages, lru) {
+				list_del(&page->lru);
+				unlock_page(page);
+				__free_pages(page, 0);
+			}
+			return -ENOMEM;
+		}
+
+		lock_page(page);
+		page->index = page_idx;
+		list_add_tail(&page->lru, pages);
+	}
+
+	list_for_each_entry(page, pages, lru)
+		f2fs_submit_page_mbio(sbi, page, page->index, &fio);
+
+	f2fs_submit_merged_bio(sbi, META, READ);
+	return 0;
+}
+
 int restore_node_summary(struct f2fs_sb_info *sbi,
 			unsigned int segno, struct f2fs_summary_block *sum)
 {
 	struct f2fs_node *rn;
 	struct f2fs_summary *sum_entry;
-	struct page *page;
+	struct page *page, *tmp;
 	block_t addr;
-	int i, last_offset;
-
-	/* alloc temporal page for read node */
-	page = alloc_page(GFP_NOFS | __GFP_ZERO);
-	if (!page)
-		return -ENOMEM;
-	lock_page(page);
+	int bio_blocks = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
+	int i, last_offset, nrpages, err = 0;
+	LIST_HEAD(page_list);
 
 	/* scan the node segment */
 	last_offset = sbi->blocks_per_seg;
 	addr = START_BLOCK(sbi, segno);
 	sum_entry = &sum->entries[0];
 
-	for (i = 0; i < last_offset; i++, sum_entry++) {
-		/*
-		 * In order to read next node page,
-		 * we must clear PageUptodate flag.
-		 */
-		ClearPageUptodate(page);
+	for (i = 0; i < last_offset; i += nrpages, addr += nrpages) {
+		nrpages = min(last_offset - i, bio_blocks);
 
-		if (f2fs_readpage(sbi, page, addr, READ_SYNC))
-			goto out;
+		/* read ahead node pages */
+		err = ra_sum_pages(sbi, &page_list, addr, nrpages);
+		if (err)
+			return err;
 
-		lock_page(page);
-		rn = F2FS_NODE(page);
-		sum_entry->nid = rn->footer.nid;
-		sum_entry->version = 0;
-		sum_entry->ofs_in_node = 0;
-		addr++;
+		list_for_each_entry_safe(page, tmp, &page_list, lru) {
+
+			lock_page(page);
+			if (unlikely(!PageUptodate(page))) {
+				err = -EIO;
+			} else {
+				rn = F2FS_NODE(page);
+				sum_entry->nid = rn->footer.nid;
+				sum_entry->version = 0;
+				sum_entry->ofs_in_node = 0;
+				sum_entry++;
+			}
+
+			list_del(&page->lru);
+			unlock_page(page);
+			__free_pages(page, 0);
+		}
 	}
-	unlock_page(page);
-out:
-	__free_pages(page, 0);
-	return 0;
+	return err;
 }
 
 static bool flush_nats_in_journal(struct f2fs_sb_info *sbi)

fs/f2fs/node.h

@@ -224,7 +224,13 @@ static inline block_t next_blkaddr_of_node(struct page *node_page)
  *    |            `- direct node (5 + N => 5 + 2N - 1)
  *    `- double indirect node (5 + 2N)
  *                 `- indirect node (6 + 2N)
- *                       `- direct node (x(N + 1))
+ *                       `- direct node
+ *                 ......
+ *                 `- indirect node ((6 + 2N) + x(N + 1))
+ *                       `- direct node
+ *                 ......
+ *                 `- indirect node ((6 + 2N) + (N - 1)(N + 1))
+ *                       `- direct node
  */
 static inline bool IS_DNODE(struct page *node_page)
 {

fs/f2fs/recovery.c

@@ -40,8 +40,7 @@ static struct fsync_inode_entry *get_fsync_inode(struct list_head *head,
 
 static int recover_dentry(struct page *ipage, struct inode *inode)
 {
-	struct f2fs_node *raw_node = F2FS_NODE(ipage);
-	struct f2fs_inode *raw_inode = &(raw_node->i);
+	struct f2fs_inode *raw_inode = F2FS_INODE(ipage);
 	nid_t pino = le32_to_cpu(raw_inode->i_pino);
 	struct f2fs_dir_entry *de;
 	struct qstr name;
@@ -62,6 +61,12 @@ static int recover_dentry(struct page *ipage, struct inode *inode)
 
 	name.len = le32_to_cpu(raw_inode->i_namelen);
 	name.name = raw_inode->i_name;
+
+	if (unlikely(name.len > F2FS_NAME_LEN)) {
+		WARN_ON(1);
+		err = -ENAMETOOLONG;
+		goto out;
+	}
 retry:
 	de = f2fs_find_entry(dir, &name, &page);
 	if (de && inode->i_ino == le32_to_cpu(de->ino))
@@ -90,17 +95,16 @@ out_unmap_put:
 	kunmap(page);
 	f2fs_put_page(page, 0);
 out:
-	f2fs_msg(inode->i_sb, KERN_NOTICE, "recover_inode and its dentry: "
-			"ino = %x, name = %s, dir = %lx, err = %d",
-			ino_of_node(ipage), raw_inode->i_name,
+	f2fs_msg(inode->i_sb, KERN_NOTICE,
+			"%s: ino = %x, name = %s, dir = %lx, err = %d",
+			__func__, ino_of_node(ipage), raw_inode->i_name,
 			IS_ERR(dir) ? 0 : dir->i_ino, err);
 	return err;
 }
 
 static int recover_inode(struct inode *inode, struct page *node_page)
 {
-	struct f2fs_node *raw_node = F2FS_NODE(node_page);
-	struct f2fs_inode *raw_inode = &(raw_node->i);
+	struct f2fs_inode *raw_inode = F2FS_INODE(node_page);
 
 	if (!IS_INODE(node_page))
 		return 0;
@@ -143,9 +147,9 @@ static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head)
 	while (1) {
 		struct fsync_inode_entry *entry;
 
-		err = f2fs_readpage(sbi, page, blkaddr, READ_SYNC);
+		err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
 		if (err)
-			goto out;
+			return err;
 
 		lock_page(page);
 
@@ -191,9 +195,10 @@ next:
 		/* check next segment */
 		blkaddr = next_blkaddr_of_node(page);
 	}
+
 	unlock_page(page);
-out:
 	__free_pages(page, 0);
+
 	return err;
 }
 
@@ -293,6 +298,9 @@ static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode,
 	struct node_info ni;
 	int err = 0, recovered = 0;
 
+	if (recover_inline_data(inode, page))
+		goto out;
+
 	start = start_bidx_of_node(ofs_of_node(page), fi);
 	if (IS_INODE(page))
 		end = start + ADDRS_PER_INODE(fi);
@@ -300,12 +308,13 @@ static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode,
 		end = start + ADDRS_PER_BLOCK;
 
 	f2fs_lock_op(sbi);
+
 	set_new_dnode(&dn, inode, NULL, NULL, 0);
 
 	err = get_dnode_of_data(&dn, start, ALLOC_NODE);
 	if (err) {
 		f2fs_unlock_op(sbi);
-		return err;
+		goto out;
 	}
 
 	wait_on_page_writeback(dn.node_page);
@@ -356,10 +365,10 @@ static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode,
 err:
 	f2fs_put_dnode(&dn);
 	f2fs_unlock_op(sbi);
-
-	f2fs_msg(sbi->sb, KERN_NOTICE, "recover_data: ino = %lx, "
-			"recovered_data = %d blocks, err = %d",
-			inode->i_ino, recovered, err);
+out:
+	f2fs_msg(sbi->sb, KERN_NOTICE,
+		"recover_data: ino = %lx, recovered = %d blocks, err = %d",
+		inode->i_ino, recovered, err);
 	return err;
 }
 
@@ -377,7 +386,7 @@ static int recover_data(struct f2fs_sb_info *sbi,
 	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
 
 	/* read node page */
-	page = alloc_page(GFP_NOFS | __GFP_ZERO);
+	page = alloc_page(GFP_F2FS_ZERO);
 	if (!page)
 		return -ENOMEM;
 
@@ -386,9 +395,9 @@ static int recover_data(struct f2fs_sb_info *sbi,
 	while (1) {
 		struct fsync_inode_entry *entry;
 
-		err = f2fs_readpage(sbi, page, blkaddr, READ_SYNC);
+		err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
 		if (err)
-			goto out;
+			return err;
 
 		lock_page(page);
 
@@ -412,8 +421,8 @@ next:
 		/* check next segment */
 		blkaddr = next_blkaddr_of_node(page);
 	}
+
 	unlock_page(page);
-out:
 	__free_pages(page, 0);
 
 	if (!err)
@@ -429,7 +438,7 @@ int recover_fsync_data(struct f2fs_sb_info *sbi)
 
 	fsync_entry_slab = f2fs_kmem_cache_create("f2fs_fsync_inode_entry",
 			sizeof(struct fsync_inode_entry), NULL);
-	if (unlikely(!fsync_entry_slab))
+	if (!fsync_entry_slab)
 		return -ENOMEM;
 
 	INIT_LIST_HEAD(&inode_list);

fs/f2fs/segment.c

@@ -14,12 +14,163 @@
 #include <linux/blkdev.h>
 #include <linux/prefetch.h>
 #include <linux/vmalloc.h>
+#include <linux/swap.h>
 
 #include "f2fs.h"
 #include "segment.h"
 #include "node.h"
 #include <trace/events/f2fs.h>
 
+#define __reverse_ffz(x) __reverse_ffs(~(x))
+
+static struct kmem_cache *discard_entry_slab;
+
+/*
+ * __reverse_ffs is copied from include/asm-generic/bitops/__ffs.h since
+ * MSB and LSB are reversed in a byte by f2fs_set_bit.
+ */
+static inline unsigned long __reverse_ffs(unsigned long word)
+{
+	int num = 0;
+
+#if BITS_PER_LONG == 64
+	if ((word & 0xffffffff) == 0) {
+		num += 32;
+		word >>= 32;
+	}
+#endif
+	if ((word & 0xffff) == 0) {
+		num += 16;
+		word >>= 16;
+	}
+	if ((word & 0xff) == 0) {
+		num += 8;
+		word >>= 8;
+	}
+	if ((word & 0xf0) == 0)
+		num += 4;
+	else
+		word >>= 4;
+	if ((word & 0xc) == 0)
+		num += 2;
+	else
+		word >>= 2;
+	if ((word & 0x2) == 0)
+		num += 1;
+	return num;
+}
+
+/*
+ * __find_rev_next(_zero)_bit is copied from lib/find_next_bit.c becasue
+ * f2fs_set_bit makes MSB and LSB reversed in a byte.
+ * Example:
+ *                             LSB <--> MSB
+ *   f2fs_set_bit(0, bitmap) => 0000 0001
+ *   f2fs_set_bit(7, bitmap) => 1000 0000
+ */
+static unsigned long __find_rev_next_bit(const unsigned long *addr,
+			unsigned long size, unsigned long offset)
+{
+	const unsigned long *p = addr + BIT_WORD(offset);
+	unsigned long result = offset & ~(BITS_PER_LONG - 1);
+	unsigned long tmp;
+	unsigned long mask, submask;
+	unsigned long quot, rest;
+
+	if (offset >= size)
+		return size;
+
+	size -= result;
+	offset %= BITS_PER_LONG;
+	if (!offset)
+		goto aligned;
+
+	tmp = *(p++);
+	quot = (offset >> 3) << 3;
+	rest = offset & 0x7;
+	mask = ~0UL << quot;
+	submask = (unsigned char)(0xff << rest) >> rest;
+	submask <<= quot;
+	mask &= submask;
+	tmp &= mask;
+	if (size < BITS_PER_LONG)
+		goto found_first;
+	if (tmp)
+		goto found_middle;
+
+	size -= BITS_PER_LONG;
+	result += BITS_PER_LONG;
+aligned:
+	while (size & ~(BITS_PER_LONG-1)) {
+		tmp = *(p++);
+		if (tmp)
+			goto found_middle;
+		result += BITS_PER_LONG;
+		size -= BITS_PER_LONG;
+	}
+	if (!size)
+		return result;
+	tmp = *p;
+found_first:
+	tmp &= (~0UL >> (BITS_PER_LONG - size));
+	if (tmp == 0UL)		/* Are any bits set? */
+		return result + size;   /* Nope. */
+found_middle:
+	return result + __reverse_ffs(tmp);
+}
+
+static unsigned long __find_rev_next_zero_bit(const unsigned long *addr,
+			unsigned long size, unsigned long offset)
+{
+	const unsigned long *p = addr + BIT_WORD(offset);
+	unsigned long result = offset & ~(BITS_PER_LONG - 1);
+	unsigned long tmp;
+	unsigned long mask, submask;
+	unsigned long quot, rest;
+
+	if (offset >= size)
+		return size;
+
+	size -= result;
+	offset %= BITS_PER_LONG;
+	if (!offset)
+		goto aligned;
+
+	tmp = *(p++);
+	quot = (offset >> 3) << 3;
+	rest = offset & 0x7;
+	mask = ~(~0UL << quot);
+	submask = (unsigned char)~((unsigned char)(0xff << rest) >> rest);
+	submask <<= quot;
+	mask += submask;
+	tmp |= mask;
+	if (size < BITS_PER_LONG)
+		goto found_first;
+	if (~tmp)
+		goto found_middle;
+
+	size -= BITS_PER_LONG;
+	result += BITS_PER_LONG;
+aligned:
+	while (size & ~(BITS_PER_LONG - 1)) {
+		tmp = *(p++);
+		if (~tmp)
+			goto found_middle;
+		result += BITS_PER_LONG;
+		size -= BITS_PER_LONG;
+	}
+	if (!size)
+		return result;
+	tmp = *p;
+
+found_first:
+	tmp |= ~0UL << size;
+	if (tmp == ~0UL)        /* Are any bits zero? */
+		return result + size;   /* Nope. */
+found_middle:
+	return result + __reverse_ffz(tmp);
+}
+
 /*
  * This function balances dirty node and dentry pages.
  * In addition, it controls garbage collection.
@@ -116,6 +267,56 @@ static void locate_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno)
 	mutex_unlock(&dirty_i->seglist_lock);
 }
 
+static void f2fs_issue_discard(struct f2fs_sb_info *sbi,
+				block_t blkstart, block_t blklen)
+{
+	sector_t start = SECTOR_FROM_BLOCK(sbi, blkstart);
+	sector_t len = SECTOR_FROM_BLOCK(sbi, blklen);
+	blkdev_issue_discard(sbi->sb->s_bdev, start, len, GFP_NOFS, 0);
+	trace_f2fs_issue_discard(sbi->sb, blkstart, blklen);
+}
+
+static void add_discard_addrs(struct f2fs_sb_info *sbi,
+			unsigned int segno, struct seg_entry *se)
+{
+	struct list_head *head = &SM_I(sbi)->discard_list;
+	struct discard_entry *new;
+	int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
+	int max_blocks = sbi->blocks_per_seg;
+	unsigned long *cur_map = (unsigned long *)se->cur_valid_map;
+	unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map;
+	unsigned long dmap[entries];
+	unsigned int start = 0, end = -1;
+	int i;
+
+	if (!test_opt(sbi, DISCARD))
+		return;
+
+	/* zero block will be discarded through the prefree list */
+	if (!se->valid_blocks || se->valid_blocks == max_blocks)
+		return;
+
+	/* 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];
+
+	while (SM_I(sbi)->nr_discards <= SM_I(sbi)->max_discards) {
+		start = __find_rev_next_bit(dmap, max_blocks, end + 1);
+		if (start >= max_blocks)
+			break;
+
+		end = __find_rev_next_zero_bit(dmap, max_blocks, start + 1);
+
+		new = f2fs_kmem_cache_alloc(discard_entry_slab, GFP_NOFS);
+		INIT_LIST_HEAD(&new->list);
+		new->blkaddr = START_BLOCK(sbi, segno) + start;
+		new->len = end - start;
+
+		list_add_tail(&new->list, head);
+		SM_I(sbi)->nr_discards += end - start;
+	}
+}
+
 /*
  * Should call clear_prefree_segments after checkpoint is done.
  */
@@ -138,6 +339,9 @@ static void set_prefree_as_free_segments(struct f2fs_sb_info *sbi)
 
 void clear_prefree_segments(struct f2fs_sb_info *sbi)
 {
+	struct list_head *head = &(SM_I(sbi)->discard_list);
+	struct list_head *this, *next;
+	struct discard_entry *entry;
 	struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
 	unsigned long *prefree_map = dirty_i->dirty_segmap[PRE];
 	unsigned int total_segs = TOTAL_SEGS(sbi);
@@ -160,14 +364,19 @@ void clear_prefree_segments(struct f2fs_sb_info *sbi)
 		if (!test_opt(sbi, DISCARD))
 			continue;
 
-		blkdev_issue_discard(sbi->sb->s_bdev,
-				START_BLOCK(sbi, start) <<
-				sbi->log_sectors_per_block,
-				(1 << (sbi->log_sectors_per_block +
-				sbi->log_blocks_per_seg)) * (end - start),
-				GFP_NOFS, 0);
+		f2fs_issue_discard(sbi, START_BLOCK(sbi, start),
+				(end - start) << sbi->log_blocks_per_seg);
 	}
 	mutex_unlock(&dirty_i->seglist_lock);
+
+	/* send small discards */
+	list_for_each_safe(this, next, head) {
+		entry = list_entry(this, struct discard_entry, list);
+		f2fs_issue_discard(sbi, entry->blkaddr, entry->len);
+		list_del(&entry->list);
+		SM_I(sbi)->nr_discards -= entry->len;
+		kmem_cache_free(discard_entry_slab, entry);
+	}
 }
 
 static void __mark_sit_entry_dirty(struct f2fs_sb_info *sbi, unsigned int segno)
@@ -459,13 +668,18 @@ static void __next_free_blkoff(struct f2fs_sb_info *sbi,
 			struct curseg_info *seg, block_t start)
 {
 	struct seg_entry *se = get_seg_entry(sbi, seg->segno);
-	block_t ofs;
-	for (ofs = start; ofs < sbi->blocks_per_seg; ofs++) {
-		if (!f2fs_test_bit(ofs, se->ckpt_valid_map)
-			&& !f2fs_test_bit(ofs, se->cur_valid_map))
-			break;
-	}
-	seg->next_blkoff = ofs;
+	int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
+	unsigned long target_map[entries];
+	unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map;
+	unsigned long *cur_map = (unsigned long *)se->cur_valid_map;
+	int i, pos;
+
+	for (i = 0; i < entries; i++)
+		target_map[i] = ckpt_map[i] | cur_map[i];
+
+	pos = __find_rev_next_zero_bit(target_map, sbi->blocks_per_seg, start);
+
+	seg->next_blkoff = pos;
 }
 
 /*
@@ -573,148 +787,6 @@ static const struct segment_allocation default_salloc_ops = {
 	.allocate_segment = allocate_segment_by_default,
 };
 
-static void f2fs_end_io_write(struct bio *bio, int err)
-{
-	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
-	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
-	struct bio_private *p = bio->bi_private;
-
-	do {
-		struct page *page = bvec->bv_page;
-
-		if (--bvec >= bio->bi_io_vec)
-			prefetchw(&bvec->bv_page->flags);
-		if (!uptodate) {
-			SetPageError(page);
-			if (page->mapping)
-				set_bit(AS_EIO, &page->mapping->flags);
-			set_ckpt_flags(p->sbi->ckpt, CP_ERROR_FLAG);
-			p->sbi->sb->s_flags |= MS_RDONLY;
-		}
-		end_page_writeback(page);
-		dec_page_count(p->sbi, F2FS_WRITEBACK);
-	} while (bvec >= bio->bi_io_vec);
-
-	if (p->is_sync)
-		complete(p->wait);
-
-	if (!get_pages(p->sbi, F2FS_WRITEBACK) &&
-			!list_empty(&p->sbi->cp_wait.task_list))
-		wake_up(&p->sbi->cp_wait);
-
-	kfree(p);
-	bio_put(bio);
-}
-
-struct bio *f2fs_bio_alloc(struct block_device *bdev, int npages)
-{
-	struct bio *bio;
-
-	/* No failure on bio allocation */
-	bio = bio_alloc(GFP_NOIO, npages);
-	bio->bi_bdev = bdev;
-	bio->bi_private = NULL;
-
-	return bio;
-}
-
-static void do_submit_bio(struct f2fs_sb_info *sbi,
-				enum page_type type, bool sync)
-{
-	int rw = sync ? WRITE_SYNC : WRITE;
-	enum page_type btype = type > META ? META : type;
-
-	if (type >= META_FLUSH)
-		rw = WRITE_FLUSH_FUA;
-
-	if (btype == META)
-		rw |= REQ_META;
-
-	if (sbi->bio[btype]) {
-		struct bio_private *p = sbi->bio[btype]->bi_private;
-		p->sbi = sbi;
-		sbi->bio[btype]->bi_end_io = f2fs_end_io_write;
-
-		trace_f2fs_do_submit_bio(sbi->sb, btype, sync, sbi->bio[btype]);
-
-		if (type == META_FLUSH) {
-			DECLARE_COMPLETION_ONSTACK(wait);
-			p->is_sync = true;
-			p->wait = &wait;
-			submit_bio(rw, sbi->bio[btype]);
-			wait_for_completion(&wait);
-		} else {
-			p->is_sync = false;
-			submit_bio(rw, sbi->bio[btype]);
-		}
-		sbi->bio[btype] = NULL;
-	}
-}
-
-void f2fs_submit_bio(struct f2fs_sb_info *sbi, enum page_type type, bool sync)
-{
-	down_write(&sbi->bio_sem);
-	do_submit_bio(sbi, type, sync);
-	up_write(&sbi->bio_sem);
-}
-
-static void submit_write_page(struct f2fs_sb_info *sbi, struct page *page,
-				block_t blk_addr, enum page_type type)
-{
-	struct block_device *bdev = sbi->sb->s_bdev;
-	int bio_blocks;
-
-	verify_block_addr(sbi, blk_addr);
-
-	down_write(&sbi->bio_sem);
-
-	inc_page_count(sbi, F2FS_WRITEBACK);
-
-	if (sbi->bio[type] && sbi->last_block_in_bio[type] != blk_addr - 1)
-		do_submit_bio(sbi, type, false);
-alloc_new:
-	if (sbi->bio[type] == NULL) {
-		struct bio_private *priv;
-retry:
-		priv = kmalloc(sizeof(struct bio_private), GFP_NOFS);
-		if (!priv) {
-			cond_resched();
-			goto retry;
-		}
-
-		bio_blocks = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
-		sbi->bio[type] = f2fs_bio_alloc(bdev, bio_blocks);
-		sbi->bio[type]->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
-		sbi->bio[type]->bi_private = priv;
-		/*
-		 * The end_io will be assigned at the sumbission phase.
-		 * Until then, let bio_add_page() merge consecutive IOs as much
-		 * as possible.
-		 */
-	}
-
-	if (bio_add_page(sbi->bio[type], page, PAGE_CACHE_SIZE, 0) <
-							PAGE_CACHE_SIZE) {
-		do_submit_bio(sbi, type, false);
-		goto alloc_new;
-	}
-
-	sbi->last_block_in_bio[type] = blk_addr;
-
-	up_write(&sbi->bio_sem);
-	trace_f2fs_submit_write_page(page, blk_addr, type);
-}
-
-void f2fs_wait_on_page_writeback(struct page *page,
-				enum page_type type, bool sync)
-{
-	struct f2fs_sb_info *sbi = F2FS_SB(page->mapping->host->i_sb);
-	if (PageWriteback(page)) {
-		f2fs_submit_bio(sbi, type, sync);
-		wait_on_page_writeback(page);
-	}
-}
-
 static bool __has_curseg_space(struct f2fs_sb_info *sbi, int type)
 {
 	struct curseg_info *curseg = CURSEG_I(sbi, type);
@@ -782,16 +854,14 @@ static int __get_segment_type(struct page *page, enum page_type p_type)
 	return __get_segment_type_6(page, p_type);
 }
 
-static void do_write_page(struct f2fs_sb_info *sbi, struct page *page,
-			block_t old_blkaddr, block_t *new_blkaddr,
-			struct f2fs_summary *sum, enum page_type p_type)
+void allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
+		block_t old_blkaddr, block_t *new_blkaddr,
+		struct f2fs_summary *sum, int type)
 {
 	struct sit_info *sit_i = SIT_I(sbi);
 	struct curseg_info *curseg;
 	unsigned int old_cursegno;
-	int type;
 
-	type = __get_segment_type(page, p_type);
 	curseg = CURSEG_I(sbi, type);
 
 	mutex_lock(&curseg->curseg_mutex);
@@ -824,49 +894,64 @@ static void do_write_page(struct f2fs_sb_info *sbi, struct page *page,
 	locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
 	mutex_unlock(&sit_i->sentry_lock);
 
-	if (p_type == NODE)
+	if (page && IS_NODESEG(type))
 		fill_node_footer_blkaddr(page, NEXT_FREE_BLKADDR(sbi, curseg));
 
-	/* writeout dirty page into bdev */
-	submit_write_page(sbi, page, *new_blkaddr, p_type);
-
 	mutex_unlock(&curseg->curseg_mutex);
 }
 
+static void do_write_page(struct f2fs_sb_info *sbi, struct page *page,
+			block_t old_blkaddr, block_t *new_blkaddr,
+			struct f2fs_summary *sum, struct f2fs_io_info *fio)
+{
+	int type = __get_segment_type(page, fio->type);
+
+	allocate_data_block(sbi, page, old_blkaddr, new_blkaddr, sum, type);
+
+	/* writeout dirty page into bdev */
+	f2fs_submit_page_mbio(sbi, page, *new_blkaddr, fio);
+}
+
 void write_meta_page(struct f2fs_sb_info *sbi, struct page *page)
 {
+	struct f2fs_io_info fio = {
+		.type = META,
+		.rw = WRITE_SYNC | REQ_META | REQ_PRIO
+	};
+
 	set_page_writeback(page);
-	submit_write_page(sbi, page, page->index, META);
+	f2fs_submit_page_mbio(sbi, page, page->index, &fio);
 }
 
 void write_node_page(struct f2fs_sb_info *sbi, struct page *page,
+		struct f2fs_io_info *fio,
 		unsigned int nid, block_t old_blkaddr, block_t *new_blkaddr)
 {
 	struct f2fs_summary sum;
 	set_summary(&sum, nid, 0, 0);
-	do_write_page(sbi, page, old_blkaddr, new_blkaddr, &sum, NODE);
+	do_write_page(sbi, page, old_blkaddr, new_blkaddr, &sum, fio);
 }
 
-void write_data_page(struct inode *inode, struct page *page,
-		struct dnode_of_data *dn, block_t old_blkaddr,
-		block_t *new_blkaddr)
+void write_data_page(struct page *page, struct dnode_of_data *dn,
+		block_t *new_blkaddr, struct f2fs_io_info *fio)
 {
-	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
+	struct f2fs_sb_info *sbi = F2FS_SB(dn->inode->i_sb);
 	struct f2fs_summary sum;
 	struct node_info ni;
 
-	f2fs_bug_on(old_blkaddr == NULL_ADDR);
+	f2fs_bug_on(dn->data_blkaddr == NULL_ADDR);
 	get_node_info(sbi, dn->nid, &ni);
 	set_summary(&sum, dn->nid, dn->ofs_in_node, ni.version);
 
-	do_write_page(sbi, page, old_blkaddr,
-			new_blkaddr, &sum, DATA);
+	do_write_page(sbi, page, dn->data_blkaddr, new_blkaddr, &sum, fio);
 }
 
-void rewrite_data_page(struct f2fs_sb_info *sbi, struct page *page,
-					block_t old_blk_addr)
+void rewrite_data_page(struct page *page, block_t old_blkaddr,
+					struct f2fs_io_info *fio)
 {
-	submit_write_page(sbi, page, old_blk_addr, DATA);
+	struct inode *inode = page->mapping->host;
+	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
+	f2fs_submit_page_mbio(sbi, page, old_blkaddr, fio);
 }
 
 void recover_data_page(struct f2fs_sb_info *sbi,
@@ -925,6 +1010,10 @@ void rewrite_node_page(struct f2fs_sb_info *sbi,
 	unsigned int segno, old_cursegno;
 	block_t next_blkaddr = next_blkaddr_of_node(page);
 	unsigned int next_segno = GET_SEGNO(sbi, next_blkaddr);
+	struct f2fs_io_info fio = {
+		.type = NODE,
+		.rw = WRITE_SYNC,
+	};
 
 	curseg = CURSEG_I(sbi, type);
 
@@ -953,8 +1042,8 @@ void rewrite_node_page(struct f2fs_sb_info *sbi,
 
 	/* rewrite node page */
 	set_page_writeback(page);
-	submit_write_page(sbi, page, new_blkaddr, NODE);
-	f2fs_submit_bio(sbi, NODE, true);
+	f2fs_submit_page_mbio(sbi, page, new_blkaddr, &fio);
+	f2fs_submit_merged_bio(sbi, NODE, WRITE);
 	refresh_sit_entry(sbi, old_blkaddr, new_blkaddr);
 
 	locate_dirty_segment(sbi, old_cursegno);
@@ -964,6 +1053,16 @@ void rewrite_node_page(struct f2fs_sb_info *sbi,
 	mutex_unlock(&curseg->curseg_mutex);
 }
 
+void f2fs_wait_on_page_writeback(struct page *page,
+				enum page_type type)
+{
+	struct f2fs_sb_info *sbi = F2FS_SB(page->mapping->host->i_sb);
+	if (PageWriteback(page)) {
+		f2fs_submit_merged_bio(sbi, type, WRITE);
+		wait_on_page_writeback(page);
+	}
+}
+
 static int read_compacted_summaries(struct f2fs_sb_info *sbi)
 {
 	struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
@@ -1314,6 +1413,10 @@ void flush_sit_entries(struct f2fs_sb_info *sbi)
 
 		sit_offset = SIT_ENTRY_OFFSET(sit_i, segno);
 
+		/* add discard candidates */
+		if (SM_I(sbi)->nr_discards < SM_I(sbi)->max_discards)
+			add_discard_addrs(sbi, segno, se);
+
 		if (flushed)
 			goto to_sit_page;
 
@@ -1480,41 +1583,94 @@ static int build_curseg(struct f2fs_sb_info *sbi)
 	return restore_curseg_summaries(sbi);
 }
 
+static int ra_sit_pages(struct f2fs_sb_info *sbi, int start, int nrpages)
+{
+	struct address_space *mapping = META_MAPPING(sbi);
+	struct page *page;
+	block_t blk_addr, prev_blk_addr = 0;
+	int sit_blk_cnt = SIT_BLK_CNT(sbi);
+	int blkno = start;
+	struct f2fs_io_info fio = {
+		.type = META,
+		.rw = READ_SYNC | REQ_META | REQ_PRIO
+	};
+
+	for (; blkno < start + nrpages && blkno < sit_blk_cnt; blkno++) {
+
+		blk_addr = current_sit_addr(sbi, blkno * SIT_ENTRY_PER_BLOCK);
+
+		if (blkno != start && prev_blk_addr + 1 != blk_addr)
+			break;
+		prev_blk_addr = blk_addr;
+repeat:
+		page = grab_cache_page(mapping, blk_addr);
+		if (!page) {
+			cond_resched();
+			goto repeat;
+		}
+		if (PageUptodate(page)) {
+			mark_page_accessed(page);
+			f2fs_put_page(page, 1);
+			continue;
+		}
+
+		f2fs_submit_page_mbio(sbi, page, blk_addr, &fio);
+
+		mark_page_accessed(page);
+		f2fs_put_page(page, 0);
+	}
+
+	f2fs_submit_merged_bio(sbi, META, READ);
+	return blkno - start;
+}
+
 static void build_sit_entries(struct f2fs_sb_info *sbi)
 {
 	struct sit_info *sit_i = SIT_I(sbi);
 	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
 	struct f2fs_summary_block *sum = curseg->sum_blk;
-	unsigned int start;
+	int sit_blk_cnt = SIT_BLK_CNT(sbi);
+	unsigned int i, start, end;
+	unsigned int readed, start_blk = 0;
+	int nrpages = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
 
-	for (start = 0; start < TOTAL_SEGS(sbi); start++) {
-		struct seg_entry *se = &sit_i->sentries[start];
-		struct f2fs_sit_block *sit_blk;
-		struct f2fs_sit_entry sit;
-		struct page *page;
-		int i;
-
-		mutex_lock(&curseg->curseg_mutex);
-		for (i = 0; i < sits_in_cursum(sum); i++) {
-			if (le32_to_cpu(segno_in_journal(sum, i)) == start) {
-				sit = sit_in_journal(sum, i);
-				mutex_unlock(&curseg->curseg_mutex);
-				goto got_it;
+	do {
+		readed = ra_sit_pages(sbi, start_blk, nrpages);
+
+		start = start_blk * sit_i->sents_per_block;
+		end = (start_blk + readed) * sit_i->sents_per_block;
+
+		for (; start < end && start < TOTAL_SEGS(sbi); start++) {
+			struct seg_entry *se = &sit_i->sentries[start];
+			struct f2fs_sit_block *sit_blk;
+			struct f2fs_sit_entry sit;
+			struct page *page;
+
+			mutex_lock(&curseg->curseg_mutex);
+			for (i = 0; i < sits_in_cursum(sum); i++) {
+				if (le32_to_cpu(segno_in_journal(sum, i))
+								== start) {
+					sit = sit_in_journal(sum, i);
+					mutex_unlock(&curseg->curseg_mutex);
+					goto got_it;
+				}
 			}
-		}
-		mutex_unlock(&curseg->curseg_mutex);
-		page = get_current_sit_page(sbi, start);
-		sit_blk = (struct f2fs_sit_block *)page_address(page);
-		sit = sit_blk->entries[SIT_ENTRY_OFFSET(sit_i, start)];
-		f2fs_put_page(page, 1);
+			mutex_unlock(&curseg->curseg_mutex);
+
+			page = get_current_sit_page(sbi, start);
+			sit_blk = (struct f2fs_sit_block *)page_address(page);
+			sit = sit_blk->entries[SIT_ENTRY_OFFSET(sit_i, start)];
+			f2fs_put_page(page, 1);
 got_it:
-		check_block_count(sbi, start, &sit);
-		seg_info_from_raw_sit(se, &sit);
-		if (sbi->segs_per_sec > 1) {
-			struct sec_entry *e = get_sec_entry(sbi, start);
-			e->valid_blocks += se->valid_blocks;
+			check_block_count(sbi, start, &sit);
+			seg_info_from_raw_sit(se, &sit);
+			if (sbi->segs_per_sec > 1) {
+				struct sec_entry *e = get_sec_entry(sbi, start);
+				e->valid_blocks += se->valid_blocks;
+			}
 		}
-	}
+		start_blk += readed;
+	} while (start_blk < sit_blk_cnt);
 }
 
 static void init_free_segmap(struct f2fs_sb_info *sbi)
@@ -1644,6 +1800,12 @@ int build_segment_manager(struct f2fs_sb_info *sbi)
 	sm_info->main_segments = le32_to_cpu(raw_super->segment_count_main);
 	sm_info->ssa_blkaddr = le32_to_cpu(raw_super->ssa_blkaddr);
 	sm_info->rec_prefree_segments = DEF_RECLAIM_PREFREE_SEGMENTS;
+	sm_info->ipu_policy = F2FS_IPU_DISABLE;
+	sm_info->min_ipu_util = DEF_MIN_IPU_UTIL;
+
+	INIT_LIST_HEAD(&sm_info->discard_list);
+	sm_info->nr_discards = 0;
+	sm_info->max_discards = 0;
 
 	err = build_sit_info(sbi);
 	if (err)
@@ -1760,3 +1922,17 @@ void destroy_segment_manager(struct f2fs_sb_info *sbi)
 	sbi->sm_info = NULL;
 	kfree(sm_info);
 }
+
+int __init create_segment_manager_caches(void)
+{
+	discard_entry_slab = f2fs_kmem_cache_create("discard_entry",
+			sizeof(struct discard_entry), NULL);
+	if (!discard_entry_slab)
+		return -ENOMEM;
+	return 0;
+}
+
+void destroy_segment_manager_caches(void)
+{
+	kmem_cache_destroy(discard_entry_slab);
+}

fs/f2fs/segment.h

@@ -20,13 +20,8 @@
 #define GET_L2R_SEGNO(free_i, segno)	(segno - free_i->start_segno)
 #define GET_R2L_SEGNO(free_i, segno)	(segno + free_i->start_segno)
 
-#define IS_DATASEG(t)							\
-	((t == CURSEG_HOT_DATA) || (t == CURSEG_COLD_DATA) ||		\
-	(t == CURSEG_WARM_DATA))
-
-#define IS_NODESEG(t)							\
-	((t == CURSEG_HOT_NODE) || (t == CURSEG_COLD_NODE) ||		\
-	(t == CURSEG_WARM_NODE))
+#define IS_DATASEG(t)	(t <= CURSEG_COLD_DATA)
+#define IS_NODESEG(t)	(t >= CURSEG_HOT_NODE)
 
 #define IS_CURSEG(sbi, seg)						\
 	((seg == CURSEG_I(sbi, CURSEG_HOT_DATA)->segno) ||	\
@@ -83,25 +78,20 @@
 	(segno / SIT_ENTRY_PER_BLOCK)
 #define	START_SEGNO(sit_i, segno)		\
 	(SIT_BLOCK_OFFSET(sit_i, segno) * SIT_ENTRY_PER_BLOCK)
+#define SIT_BLK_CNT(sbi)			\
+	((TOTAL_SEGS(sbi) + SIT_ENTRY_PER_BLOCK - 1) / SIT_ENTRY_PER_BLOCK)
 #define f2fs_bitmap_size(nr)			\
 	(BITS_TO_LONGS(nr) * sizeof(unsigned long))
 #define TOTAL_SEGS(sbi)	(SM_I(sbi)->main_segments)
 #define TOTAL_SECS(sbi)	(sbi->total_sections)
 
 #define SECTOR_FROM_BLOCK(sbi, blk_addr)				\
-	(blk_addr << ((sbi)->log_blocksize - F2FS_LOG_SECTOR_SIZE))
+	(((sector_t)blk_addr) << (sbi)->log_sectors_per_block)
 #define SECTOR_TO_BLOCK(sbi, sectors)					\
-	(sectors >> ((sbi)->log_blocksize - F2FS_LOG_SECTOR_SIZE))
+	(sectors >> (sbi)->log_sectors_per_block)
 #define MAX_BIO_BLOCKS(max_hw_blocks)					\
 	(min((int)max_hw_blocks, BIO_MAX_PAGES))
 
-/* during checkpoint, bio_private is used to synchronize the last bio */
-struct bio_private {
-	struct f2fs_sb_info *sbi;
-	bool is_sync;
-	void *wait;
-};
-
 /*
  * indicate a block allocation direction: RIGHT and LEFT.
  * RIGHT means allocating new sections towards the end of volume.
@@ -458,8 +448,8 @@ static inline int reserved_sections(struct f2fs_sb_info *sbi)
 
 static inline bool need_SSR(struct f2fs_sb_info *sbi)
 {
-	return ((prefree_segments(sbi) / sbi->segs_per_sec)
-			+ free_sections(sbi) < overprovision_sections(sbi));
+	return (prefree_segments(sbi) / sbi->segs_per_sec)
+			+ free_sections(sbi) < overprovision_sections(sbi);
 }
 
 static inline bool has_not_enough_free_secs(struct f2fs_sb_info *sbi, int freed)
@@ -467,38 +457,71 @@ static inline bool has_not_enough_free_secs(struct f2fs_sb_info *sbi, int freed)
 	int node_secs = get_blocktype_secs(sbi, F2FS_DIRTY_NODES);
 	int dent_secs = get_blocktype_secs(sbi, F2FS_DIRTY_DENTS);
 
-	if (sbi->por_doing)
+	if (unlikely(sbi->por_doing))
 		return false;
 
-	return ((free_sections(sbi) + freed) <= (node_secs + 2 * dent_secs +
-						reserved_sections(sbi)));
+	return (free_sections(sbi) + freed) <= (node_secs + 2 * dent_secs +
+						reserved_sections(sbi));
 }
 
 static inline bool excess_prefree_segs(struct f2fs_sb_info *sbi)
 {
-	return (prefree_segments(sbi) > SM_I(sbi)->rec_prefree_segments);
+	return prefree_segments(sbi) > SM_I(sbi)->rec_prefree_segments;
 }
 
 static inline int utilization(struct f2fs_sb_info *sbi)
 {
-	return div_u64((u64)valid_user_blocks(sbi) * 100, sbi->user_block_count);
+	return div_u64((u64)valid_user_blocks(sbi) * 100,
+					sbi->user_block_count);
 }
 
 /*
  * Sometimes f2fs may be better to drop out-of-place update policy.
- * So, if fs utilization is over MIN_IPU_UTIL, then f2fs tries to write
- * data in the original place likewise other traditional file systems.
- * But, currently set 100 in percentage, which means it is disabled.
- * See below need_inplace_update().
+ * And, users can control the policy through sysfs entries.
+ * There are five policies with triggering conditions as follows.
+ * F2FS_IPU_FORCE - all the time,
+ * F2FS_IPU_SSR - if SSR mode is activated,
+ * F2FS_IPU_UTIL - if FS utilization is over threashold,
+ * F2FS_IPU_SSR_UTIL - if SSR mode is activated and FS utilization is over
+ *                     threashold,
+ * F2FS_IPUT_DISABLE - disable IPU. (=default option)
  */
-#define MIN_IPU_UTIL		100
+#define DEF_MIN_IPU_UTIL	70
+
+enum {
+	F2FS_IPU_FORCE,
+	F2FS_IPU_SSR,
+	F2FS_IPU_UTIL,
+	F2FS_IPU_SSR_UTIL,
+	F2FS_IPU_DISABLE,
+};
+
 static inline bool need_inplace_update(struct inode *inode)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
+
+	/* IPU can be done only for the user data */
 	if (S_ISDIR(inode->i_mode))
 		return false;
-	if (need_SSR(sbi) && utilization(sbi) > MIN_IPU_UTIL)
+
+	switch (SM_I(sbi)->ipu_policy) {
+	case F2FS_IPU_FORCE:
 		return true;
+	case F2FS_IPU_SSR:
+		if (need_SSR(sbi))
+			return true;
+		break;
+	case F2FS_IPU_UTIL:
+		if (utilization(sbi) > SM_I(sbi)->min_ipu_util)
+			return true;
+		break;
+	case F2FS_IPU_SSR_UTIL:
+		if (need_SSR(sbi) && utilization(sbi) > SM_I(sbi)->min_ipu_util)
+			return true;
+		break;
+	case F2FS_IPU_DISABLE:
+		break;
+	}
 	return false;
 }
 

fs/f2fs/super.c

@@ -50,6 +50,7 @@ enum {
 	Opt_active_logs,
 	Opt_disable_ext_identify,
 	Opt_inline_xattr,
+	Opt_inline_data,
 	Opt_err,
 };
 
@@ -65,6 +66,7 @@ static match_table_t f2fs_tokens = {
 	{Opt_active_logs, "active_logs=%u"},
 	{Opt_disable_ext_identify, "disable_ext_identify"},
 	{Opt_inline_xattr, "inline_xattr"},
+	{Opt_inline_data, "inline_data"},
 	{Opt_err, NULL},
 };
 
@@ -72,6 +74,7 @@ static match_table_t f2fs_tokens = {
 enum {
 	GC_THREAD,	/* struct f2fs_gc_thread */
 	SM_INFO,	/* struct f2fs_sm_info */
+	F2FS_SBI,	/* struct f2fs_sb_info */
 };
 
 struct f2fs_attr {
@@ -89,6 +92,8 @@ static unsigned char *__struct_ptr(struct f2fs_sb_info *sbi, int struct_type)
 		return (unsigned char *)sbi->gc_thread;
 	else if (struct_type == SM_INFO)
 		return (unsigned char *)SM_I(sbi);
+	else if (struct_type == F2FS_SBI)
+		return (unsigned char *)sbi;
 	return NULL;
 }
 
@@ -175,6 +180,10 @@ F2FS_RW_ATTR(GC_THREAD, f2fs_gc_kthread, gc_max_sleep_time, max_sleep_time);
 F2FS_RW_ATTR(GC_THREAD, f2fs_gc_kthread, gc_no_gc_sleep_time, no_gc_sleep_time);
 F2FS_RW_ATTR(GC_THREAD, f2fs_gc_kthread, gc_idle, gc_idle);
 F2FS_RW_ATTR(SM_INFO, f2fs_sm_info, reclaim_segments, rec_prefree_segments);
+F2FS_RW_ATTR(SM_INFO, f2fs_sm_info, max_small_discards, max_discards);
+F2FS_RW_ATTR(SM_INFO, f2fs_sm_info, ipu_policy, ipu_policy);
+F2FS_RW_ATTR(SM_INFO, f2fs_sm_info, min_ipu_util, min_ipu_util);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, max_victim_search, max_victim_search);
 
 #define ATTR_LIST(name) (&f2fs_attr_##name.attr)
 static struct attribute *f2fs_attrs[] = {
@@ -183,6 +192,10 @@ static struct attribute *f2fs_attrs[] = {
 	ATTR_LIST(gc_no_gc_sleep_time),
 	ATTR_LIST(gc_idle),
 	ATTR_LIST(reclaim_segments),
+	ATTR_LIST(max_small_discards),
+	ATTR_LIST(ipu_policy),
+	ATTR_LIST(min_ipu_util),
+	ATTR_LIST(max_victim_search),
 	NULL,
 };
 
@@ -311,6 +324,9 @@ static int parse_options(struct super_block *sb, char *options)
 		case Opt_disable_ext_identify:
 			set_opt(sbi, DISABLE_EXT_IDENTIFY);
 			break;
+		case Opt_inline_data:
+			set_opt(sbi, INLINE_DATA);
+			break;
 		default:
 			f2fs_msg(sb, KERN_ERR,
 				"Unrecognized mount option \"%s\" or missing value",
@@ -325,7 +341,7 @@ static struct inode *f2fs_alloc_inode(struct super_block *sb)
 {
 	struct f2fs_inode_info *fi;
 
-	fi = kmem_cache_alloc(f2fs_inode_cachep, GFP_NOFS | __GFP_ZERO);
+	fi = kmem_cache_alloc(f2fs_inode_cachep, GFP_F2FS_ZERO);
 	if (!fi)
 		return NULL;
 
@@ -508,7 +524,8 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
 #endif
 	if (test_opt(sbi, DISABLE_EXT_IDENTIFY))
 		seq_puts(seq, ",disable_ext_identify");
-
+	if (test_opt(sbi, INLINE_DATA))
+		seq_puts(seq, ",inline_data");
 	seq_printf(seq, ",active_logs=%u", sbi->active_logs);
 
 	return 0;
@@ -518,7 +535,8 @@ static int segment_info_seq_show(struct seq_file *seq, void *offset)
 {
 	struct super_block *sb = seq->private;
 	struct f2fs_sb_info *sbi = F2FS_SB(sb);
-	unsigned int total_segs = le32_to_cpu(sbi->raw_super->segment_count_main);
+	unsigned int total_segs =
+			le32_to_cpu(sbi->raw_super->segment_count_main);
 	int i;
 
 	for (i = 0; i < total_segs; i++) {
@@ -618,7 +636,7 @@ static struct inode *f2fs_nfs_get_inode(struct super_block *sb,
 	struct f2fs_sb_info *sbi = F2FS_SB(sb);
 	struct inode *inode;
 
-	if (ino < F2FS_ROOT_INO(sbi))
+	if (unlikely(ino < F2FS_ROOT_INO(sbi)))
 		return ERR_PTR(-ESTALE);
 
 	/*
@@ -629,7 +647,7 @@ static struct inode *f2fs_nfs_get_inode(struct super_block *sb,
 	inode = f2fs_iget(sb, ino);
 	if (IS_ERR(inode))
 		return ERR_CAST(inode);
-	if (generation && inode->i_generation != generation) {
+	if (unlikely(generation && inode->i_generation != generation)) {
 		/* we didn't find the right inode.. */
 		iput(inode);
 		return ERR_PTR(-ESTALE);
@@ -732,10 +750,10 @@ static int sanity_check_ckpt(struct f2fs_sb_info *sbi)
 	fsmeta += le32_to_cpu(ckpt->rsvd_segment_count);
 	fsmeta += le32_to_cpu(raw_super->segment_count_ssa);
 
-	if (fsmeta >= total)
+	if (unlikely(fsmeta >= total))
 		return 1;
 
-	if (is_set_ckpt_flags(ckpt, CP_ERROR_FLAG)) {
+	if (unlikely(is_set_ckpt_flags(ckpt, CP_ERROR_FLAG))) {
 		f2fs_msg(sbi->sb, KERN_ERR, "A bug case: need to run fsck");
 		return 1;
 	}
@@ -763,6 +781,7 @@ static void init_sb_info(struct f2fs_sb_info *sbi)
 	sbi->node_ino_num = le32_to_cpu(raw_super->node_ino);
 	sbi->meta_ino_num = le32_to_cpu(raw_super->meta_ino);
 	sbi->cur_victim_sec = NULL_SECNO;
+	sbi->max_victim_search = DEF_MAX_VICTIM_SEARCH;
 
 	for (i = 0; i < NR_COUNT_TYPE; i++)
 		atomic_set(&sbi->nr_pages[i], 0);
@@ -798,9 +817,10 @@ retry:
 	/* sanity checking of raw super */
 	if (sanity_check_raw_super(sb, *raw_super)) {
 		brelse(*raw_super_buf);
-		f2fs_msg(sb, KERN_ERR, "Can't find a valid F2FS filesystem "
-				"in %dth superblock", block + 1);
-		if(block == 0) {
+		f2fs_msg(sb, KERN_ERR,
+			"Can't find valid F2FS filesystem in %dth superblock",
+								block + 1);
+		if (block == 0) {
 			block++;
 			goto retry;
 		} else {
@@ -818,6 +838,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 	struct buffer_head *raw_super_buf;
 	struct inode *root;
 	long err = -EINVAL;
+	int i;
 
 	/* allocate memory for f2fs-specific super block info */
 	sbi = kzalloc(sizeof(struct f2fs_sb_info), GFP_KERNEL);
@@ -825,7 +846,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 		return -ENOMEM;
 
 	/* set a block size */
-	if (!sb_set_blocksize(sb, F2FS_BLKSIZE)) {
+	if (unlikely(!sb_set_blocksize(sb, F2FS_BLKSIZE))) {
 		f2fs_msg(sb, KERN_ERR, "unable to set blocksize");
 		goto free_sbi;
 	}
@@ -874,7 +895,16 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 	mutex_init(&sbi->node_write);
 	sbi->por_doing = false;
 	spin_lock_init(&sbi->stat_lock);
-	init_rwsem(&sbi->bio_sem);
+
+	mutex_init(&sbi->read_io.io_mutex);
+	sbi->read_io.sbi = sbi;
+	sbi->read_io.bio = NULL;
+	for (i = 0; i < NR_PAGE_TYPE; i++) {
+		mutex_init(&sbi->write_io[i].io_mutex);
+		sbi->write_io[i].sbi = sbi;
+		sbi->write_io[i].bio = NULL;
+	}
+
 	init_rwsem(&sbi->cp_rwsem);
 	init_waitqueue_head(&sbi->cp_wait);
 	init_sb_info(sbi);
@@ -939,9 +969,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 	}
 
 	/* if there are nt orphan nodes free them */
-	err = -EINVAL;
-	if (recover_orphan_inodes(sbi))
-		goto free_node_inode;
+	recover_orphan_inodes(sbi);
 
 	/* read root inode and dentry */
 	root = f2fs_iget(sb, F2FS_ROOT_INO(sbi));
@@ -950,8 +978,10 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 		err = PTR_ERR(root);
 		goto free_node_inode;
 	}
-	if (!S_ISDIR(root->i_mode) || !root->i_blocks || !root->i_size)
+	if (!S_ISDIR(root->i_mode) || !root->i_blocks || !root->i_size) {
+		err = -EINVAL;
 		goto free_root_inode;
+	}
 
 	sb->s_root = d_make_root(root); /* allocate root dentry */
 	if (!sb->s_root) {
@@ -1053,7 +1083,7 @@ static int __init init_inodecache(void)
 {
 	f2fs_inode_cachep = f2fs_kmem_cache_create("f2fs_inode_cache",
 			sizeof(struct f2fs_inode_info), NULL);
-	if (f2fs_inode_cachep == NULL)
+	if (!f2fs_inode_cachep)
 		return -ENOMEM;
 	return 0;
 }
@@ -1078,9 +1108,12 @@ static int __init init_f2fs_fs(void)
 	err = create_node_manager_caches();
 	if (err)
 		goto free_inodecache;
-	err = create_gc_caches();
+	err = create_segment_manager_caches();
 	if (err)
 		goto free_node_manager_caches;
+	err = create_gc_caches();
+	if (err)
+		goto free_segment_manager_caches;
 	err = create_checkpoint_caches();
 	if (err)
 		goto free_gc_caches;
@@ -1102,6 +1135,8 @@ free_checkpoint_caches:
 	destroy_checkpoint_caches();
 free_gc_caches:
 	destroy_gc_caches();
+free_segment_manager_caches:
+	destroy_segment_manager_caches();
 free_node_manager_caches:
 	destroy_node_manager_caches();
 free_inodecache:
@@ -1117,6 +1152,7 @@ static void __exit exit_f2fs_fs(void)
 	unregister_filesystem(&f2fs_fs_type);
 	destroy_checkpoint_caches();
 	destroy_gc_caches();
+	destroy_segment_manager_caches();
 	destroy_node_manager_caches();
 	destroy_inodecache();
 	kset_unregister(f2fs_kset);

fs/f2fs/xattr.c

@@ -522,7 +522,7 @@ static int __f2fs_setxattr(struct inode *inode, int name_index,
 		if (found)
 			free = free + ENTRY_SIZE(here);
 
-		if (free < newsize) {
+		if (unlikely(free < newsize)) {
 			error = -ENOSPC;
 			goto exit;
 		}

include/linux/f2fs_fs.h

@@ -153,6 +153,13 @@ struct f2fs_extent {
 #define	NODE_DIND_BLOCK		(DEF_ADDRS_PER_INODE + 5)
 
 #define F2FS_INLINE_XATTR	0x01	/* file inline xattr flag */
+#define F2FS_INLINE_DATA	0x02	/* file inline data 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 + 5 - 1))
 
 struct f2fs_inode {
 	__le16 i_mode;			/* file mode */

include/trace/events/f2fs.h

@@ -16,15 +16,28 @@
 		{ META,		"META" },				\
 		{ META_FLUSH,	"META_FLUSH" })
 
-#define show_bio_type(type)						\
-	__print_symbolic(type,						\
-		{ READ, 	"READ" },				\
-		{ READA, 	"READAHEAD" },				\
-		{ READ_SYNC, 	"READ_SYNC" },				\
-		{ WRITE, 	"WRITE" },				\
-		{ WRITE_SYNC, 	"WRITE_SYNC" },				\
-		{ WRITE_FLUSH,	"WRITE_FLUSH" },			\
-		{ WRITE_FUA, 	"WRITE_FUA" })
+#define F2FS_BIO_MASK(t)	(t & (READA | WRITE_FLUSH_FUA))
+#define F2FS_BIO_EXTRA_MASK(t)	(t & (REQ_META | REQ_PRIO))
+
+#define show_bio_type(type)	show_bio_base(type), show_bio_extra(type)
+
+#define show_bio_base(type)						\
+	__print_symbolic(F2FS_BIO_MASK(type),				\
+		{ READ, 		"READ" },			\
+		{ READA, 		"READAHEAD" },			\
+		{ READ_SYNC, 		"READ_SYNC" },			\
+		{ WRITE, 		"WRITE" },			\
+		{ WRITE_SYNC, 		"WRITE_SYNC" },			\
+		{ WRITE_FLUSH,		"WRITE_FLUSH" },		\
+		{ WRITE_FUA, 		"WRITE_FUA" },			\
+		{ WRITE_FLUSH_FUA,	"WRITE_FLUSH_FUA" })
+
+#define show_bio_extra(type)						\
+	__print_symbolic(F2FS_BIO_EXTRA_MASK(type),			\
+		{ REQ_META, 		"(M)" },			\
+		{ REQ_PRIO, 		"(P)" },			\
+		{ REQ_META | REQ_PRIO,	"(MP)" },			\
+		{ 0, " \b" })
 
 #define show_data_type(type)						\
 	__print_symbolic(type,						\
@@ -421,7 +434,7 @@ TRACE_EVENT(f2fs_truncate_partial_nodes,
 		__entry->err)
 );
 
-TRACE_EVENT_CONDITION(f2fs_readpage,
+TRACE_EVENT_CONDITION(f2fs_submit_page_bio,
 
 	TP_PROTO(struct page *page, sector_t blkaddr, int type),
 
@@ -446,7 +459,7 @@ TRACE_EVENT_CONDITION(f2fs_readpage,
 	),
 
 	TP_printk("dev = (%d,%d), ino = %lu, page_index = 0x%lx, "
-		"blkaddr = 0x%llx, bio_type = %s",
+		"blkaddr = 0x%llx, bio_type = %s%s",
 		show_dev_ino(__entry),
 		(unsigned long)__entry->index,
 		(unsigned long long)__entry->blkaddr,
@@ -598,36 +611,54 @@ TRACE_EVENT(f2fs_reserve_new_block,
 		__entry->ofs_in_node)
 );
 
-TRACE_EVENT(f2fs_do_submit_bio,
+DECLARE_EVENT_CLASS(f2fs__submit_bio,
 
-	TP_PROTO(struct super_block *sb, int btype, bool sync, struct bio *bio),
+	TP_PROTO(struct super_block *sb, int rw, int type, struct bio *bio),
 
-	TP_ARGS(sb, btype, sync, bio),
+	TP_ARGS(sb, rw, type, bio),
 
 	TP_STRUCT__entry(
 		__field(dev_t,	dev)
-		__field(int,	btype)
-		__field(bool,	sync)
+		__field(int,	rw)
+		__field(int,	type)
 		__field(sector_t,	sector)
 		__field(unsigned int,	size)
 	),
 
 	TP_fast_assign(
 		__entry->dev		= sb->s_dev;
-		__entry->btype		= btype;
-		__entry->sync		= sync;
+		__entry->rw		= rw;
+		__entry->type		= type;
 		__entry->sector		= bio->bi_sector;
 		__entry->size		= bio->bi_size;
 	),
 
-	TP_printk("dev = (%d,%d), type = %s, io = %s, sector = %lld, size = %u",
+	TP_printk("dev = (%d,%d), %s%s, %s, sector = %lld, size = %u",
 		show_dev(__entry),
-		show_block_type(__entry->btype),
-		__entry->sync ? "sync" : "no sync",
+		show_bio_type(__entry->rw),
+		show_block_type(__entry->type),
 		(unsigned long long)__entry->sector,
 		__entry->size)
 );
 
+DEFINE_EVENT_CONDITION(f2fs__submit_bio, f2fs_submit_write_bio,
+
+	TP_PROTO(struct super_block *sb, int rw, int type, struct bio *bio),
+
+	TP_ARGS(sb, rw, type, bio),
+
+	TP_CONDITION(bio)
+);
+
+DEFINE_EVENT_CONDITION(f2fs__submit_bio, f2fs_submit_read_bio,
+
+	TP_PROTO(struct super_block *sb, int rw, int type, struct bio *bio),
+
+	TP_ARGS(sb, rw, type, bio),
+
+	TP_CONDITION(bio)
+);
+
 DECLARE_EVENT_CLASS(f2fs__page,
 
 	TP_PROTO(struct page *page, int type),
@@ -674,15 +705,16 @@ DEFINE_EVENT(f2fs__page, f2fs_vm_page_mkwrite,
 	TP_ARGS(page, type)
 );
 
-TRACE_EVENT(f2fs_submit_write_page,
+TRACE_EVENT(f2fs_submit_page_mbio,
 
-	TP_PROTO(struct page *page, block_t blk_addr, int type),
+	TP_PROTO(struct page *page, int rw, int type, block_t blk_addr),
 
-	TP_ARGS(page, blk_addr, type),
+	TP_ARGS(page, rw, type, blk_addr),
 
 	TP_STRUCT__entry(
 		__field(dev_t,	dev)
 		__field(ino_t,	ino)
+		__field(int, rw)
 		__field(int, type)
 		__field(pgoff_t, index)
 		__field(block_t, block)
@@ -691,13 +723,15 @@ TRACE_EVENT(f2fs_submit_write_page,
 	TP_fast_assign(
 		__entry->dev	= page->mapping->host->i_sb->s_dev;
 		__entry->ino	= page->mapping->host->i_ino;
+		__entry->rw	= rw;
 		__entry->type	= type;
 		__entry->index	= page->index;
 		__entry->block	= blk_addr;
 	),
 
-	TP_printk("dev = (%d,%d), ino = %lu, %s, index = %lu, blkaddr = 0x%llx",
+	TP_printk("dev = (%d,%d), ino = %lu, %s%s, %s, index = %lu, blkaddr = 0x%llx",
 		show_dev_ino(__entry),
+		show_bio_type(__entry->rw),
 		show_block_type(__entry->type),
 		(unsigned long)__entry->index,
 		(unsigned long long)__entry->block)
@@ -727,6 +761,29 @@ TRACE_EVENT(f2fs_write_checkpoint,
 		__entry->msg)
 );
 
+TRACE_EVENT(f2fs_issue_discard,
+
+	TP_PROTO(struct super_block *sb, block_t blkstart, block_t blklen),
+
+	TP_ARGS(sb, blkstart, blklen),
+
+	TP_STRUCT__entry(
+		__field(dev_t,	dev)
+		__field(block_t, blkstart)
+		__field(block_t, blklen)
+	),
+
+	TP_fast_assign(
+		__entry->dev	= sb->s_dev;
+		__entry->blkstart = blkstart;
+		__entry->blklen = blklen;
+	),
+
+	TP_printk("dev = (%d,%d), blkstart = 0x%llx, blklen = 0x%llx",
+		show_dev(__entry),
+		(unsigned long long)__entry->blkstart,
+		(unsigned long long)__entry->blklen)
+);
 #endif /* _TRACE_F2FS_H */
 
  /* This part must be outside protection */