Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext4 updates from Ted Ts'o:
 "Lots of bugs fixes, including Zheng and Jan's extent status shrinker
  fixes, which should improve CPU utilization and potential soft lockups
  under heavy memory pressure, and Eric Whitney's bigalloc fixes"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (26 commits)
  ext4: ext4_da_convert_inline_data_to_extent drop locked page after error
  ext4: fix suboptimal seek_{data,hole} extents traversial
  ext4: ext4_inline_data_fiemap should respect callers argument
  ext4: prevent fsreentrance deadlock for inline_data
  ext4: forbid journal_async_commit in data=ordered mode
  jbd2: remove unnecessary NULL check before iput()
  ext4: Remove an unnecessary check for NULL before iput()
  ext4: remove unneeded code in ext4_unlink
  ext4: don't count external journal blocks as overhead
  ext4: remove never taken branch from ext4_ext_shift_path_extents()
  ext4: create nojournal_checksum mount option
  ext4: update comments regarding ext4_delete_inode()
  ext4: cleanup GFP flags inside resize path
  ext4: introduce aging to extent status tree
  ext4: cleanup flag definitions for extent status tree
  ext4: limit number of scanned extents in status tree shrinker
  ext4: move handling of list of shrinkable inodes into extent status code
  ext4: change LRU to round-robin in extent status tree shrinker
  ext4: cache extent hole in extent status tree for ext4_da_map_blocks()
  ext4: fix block reservation for bigalloc filesystems
  ...

fs/ext4/ext4.h

@@ -158,17 +158,8 @@ struct ext4_allocation_request {
 #define EXT4_MAP_MAPPED		(1 << BH_Mapped)
 #define EXT4_MAP_UNWRITTEN	(1 << BH_Unwritten)
 #define EXT4_MAP_BOUNDARY	(1 << BH_Boundary)
-/* Sometimes (in the bigalloc case, from ext4_da_get_block_prep) the caller of
- * ext4_map_blocks wants to know whether or not the underlying cluster has
- * already been accounted for. EXT4_MAP_FROM_CLUSTER conveys to the caller that
- * the requested mapping was from previously mapped (or delayed allocated)
- * cluster. We use BH_AllocFromCluster only for this flag. BH_AllocFromCluster
- * should never appear on buffer_head's state flags.
- */
-#define EXT4_MAP_FROM_CLUSTER	(1 << BH_AllocFromCluster)
 #define EXT4_MAP_FLAGS		(EXT4_MAP_NEW | EXT4_MAP_MAPPED |\
-				 EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY |\
-				 EXT4_MAP_FROM_CLUSTER)
+				 EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY)
 
 struct ext4_map_blocks {
 	ext4_fsblk_t m_pblk;
@@ -565,10 +556,8 @@ enum {
 #define EXT4_GET_BLOCKS_KEEP_SIZE		0x0080
 	/* Do not take i_data_sem locking in ext4_map_blocks */
 #define EXT4_GET_BLOCKS_NO_LOCK			0x0100
-	/* Do not put hole in extent cache */
-#define EXT4_GET_BLOCKS_NO_PUT_HOLE		0x0200
 	/* Convert written extents to unwritten */
-#define EXT4_GET_BLOCKS_CONVERT_UNWRITTEN	0x0400
+#define EXT4_GET_BLOCKS_CONVERT_UNWRITTEN	0x0200
 
 /*
  * The bit position of these flags must not overlap with any of the
@@ -889,10 +878,12 @@ struct ext4_inode_info {
 	/* extents status tree */
 	struct ext4_es_tree i_es_tree;
 	rwlock_t i_es_lock;
-	struct list_head i_es_lru;
+	struct list_head i_es_list;
 	unsigned int i_es_all_nr;	/* protected by i_es_lock */
-	unsigned int i_es_lru_nr;	/* protected by i_es_lock */
-	unsigned long i_touch_when;	/* jiffies of last accessing */
+	unsigned int i_es_shk_nr;	/* protected by i_es_lock */
+	ext4_lblk_t i_es_shrink_lblk;	/* Offset where we start searching for
+					   extents to shrink. Protected by
+					   i_es_lock  */
 
 	/* ialloc */
 	ext4_group_t	i_last_alloc_group;
@@ -1337,10 +1328,11 @@ struct ext4_sb_info {
 
 	/* Reclaim extents from extent status tree */
 	struct shrinker s_es_shrinker;
-	struct list_head s_es_lru;
+	struct list_head s_es_list;	/* List of inodes with reclaimable extents */
+	long s_es_nr_inode;
 	struct ext4_es_stats s_es_stats;
 	struct mb_cache *s_mb_cache;
-	spinlock_t s_es_lru_lock ____cacheline_aligned_in_smp;
+	spinlock_t s_es_lock ____cacheline_aligned_in_smp;
 
 	/* Ratelimit ext4 messages. */
 	struct ratelimit_state s_err_ratelimit_state;
@@ -2196,7 +2188,6 @@ extern int ext4_calculate_overhead(struct super_block *sb);
 extern void ext4_superblock_csum_set(struct super_block *sb);
 extern void *ext4_kvmalloc(size_t size, gfp_t flags);
 extern void *ext4_kvzalloc(size_t size, gfp_t flags);
-extern void ext4_kvfree(void *ptr);
 extern int ext4_alloc_flex_bg_array(struct super_block *sb,
 				    ext4_group_t ngroup);
 extern const char *ext4_decode_error(struct super_block *sb, int errno,
@@ -2647,7 +2638,7 @@ extern struct buffer_head *ext4_get_first_inline_block(struct inode *inode,
 					int *retval);
 extern int ext4_inline_data_fiemap(struct inode *inode,
 				   struct fiemap_extent_info *fieinfo,
-				   int *has_inline);
+				   int *has_inline, __u64 start, __u64 len);
 extern int ext4_try_to_evict_inline_data(handle_t *handle,
 					 struct inode *inode,
 					 int needed);
@@ -2794,16 +2785,6 @@ extern int ext4_bio_write_page(struct ext4_io_submit *io,
 /* mmp.c */
 extern int ext4_multi_mount_protect(struct super_block *, ext4_fsblk_t);
 
-/*
- * Note that these flags will never ever appear in a buffer_head's state flag.
- * See EXT4_MAP_... to see where this is used.
- */
-enum ext4_state_bits {
-	BH_AllocFromCluster	/* allocated blocks were part of already
-				 * allocated cluster. */
-	= BH_JBDPrivateStart
-};
-
 /*
  * Add new method to test whether block and inode bitmaps are properly
  * initialized. With uninit_bg reading the block from disk is not enough

fs/ext4/extents_status.h

@@ -29,25 +29,28 @@
 /*
  * These flags live in the high bits of extent_status.es_pblk
  */
-#define ES_SHIFT	60
+enum {
+	ES_WRITTEN_B,
+	ES_UNWRITTEN_B,
+	ES_DELAYED_B,
+	ES_HOLE_B,
+	ES_REFERENCED_B,
+	ES_FLAGS
+};
+
+#define ES_SHIFT (sizeof(ext4_fsblk_t)*8 - ES_FLAGS)
+#define ES_MASK (~((ext4_fsblk_t)0) << ES_SHIFT)
 
-#define EXTENT_STATUS_WRITTEN	(1 << 3)
-#define EXTENT_STATUS_UNWRITTEN (1 << 2)
-#define EXTENT_STATUS_DELAYED	(1 << 1)
-#define EXTENT_STATUS_HOLE	(1 << 0)
+#define EXTENT_STATUS_WRITTEN	(1 << ES_WRITTEN_B)
+#define EXTENT_STATUS_UNWRITTEN (1 << ES_UNWRITTEN_B)
+#define EXTENT_STATUS_DELAYED	(1 << ES_DELAYED_B)
+#define EXTENT_STATUS_HOLE	(1 << ES_HOLE_B)
+#define EXTENT_STATUS_REFERENCED	(1 << ES_REFERENCED_B)
 
-#define EXTENT_STATUS_FLAGS	(EXTENT_STATUS_WRITTEN | \
+#define ES_TYPE_MASK	((ext4_fsblk_t)(EXTENT_STATUS_WRITTEN | \
 			  EXTENT_STATUS_UNWRITTEN | \
 			  EXTENT_STATUS_DELAYED | \
-				 EXTENT_STATUS_HOLE)
-
-#define ES_WRITTEN		(1ULL << 63)
-#define ES_UNWRITTEN		(1ULL << 62)
-#define ES_DELAYED		(1ULL << 61)
-#define ES_HOLE			(1ULL << 60)
-
-#define ES_MASK			(ES_WRITTEN | ES_UNWRITTEN | \
-				 ES_DELAYED | ES_HOLE)
+			  EXTENT_STATUS_HOLE) << ES_SHIFT)
 
 struct ext4_sb_info;
 struct ext4_extent;
@@ -65,14 +68,13 @@ struct ext4_es_tree {
 };
 
 struct ext4_es_stats {
-	unsigned long es_stats_last_sorted;
 	unsigned long es_stats_shrunk;
 	unsigned long es_stats_cache_hits;
 	unsigned long es_stats_cache_misses;
 	u64 es_stats_scan_time;
 	u64 es_stats_max_scan_time;
 	struct percpu_counter es_stats_all_cnt;
-	struct percpu_counter es_stats_lru_cnt;
+	struct percpu_counter es_stats_shk_cnt;
 };
 
 extern int __init ext4_init_es(void);
@@ -93,29 +95,49 @@ extern void ext4_es_find_delayed_extent_range(struct inode *inode,
 extern int ext4_es_lookup_extent(struct inode *inode, ext4_lblk_t lblk,
 				 struct extent_status *es);
 
+static inline unsigned int ext4_es_status(struct extent_status *es)
+{
+	return es->es_pblk >> ES_SHIFT;
+}
+
+static inline unsigned int ext4_es_type(struct extent_status *es)
+{
+	return (es->es_pblk & ES_TYPE_MASK) >> ES_SHIFT;
+}
+
 static inline int ext4_es_is_written(struct extent_status *es)
 {
-	return (es->es_pblk & ES_WRITTEN) != 0;
+	return (ext4_es_type(es) & EXTENT_STATUS_WRITTEN) != 0;
 }
 
 static inline int ext4_es_is_unwritten(struct extent_status *es)
 {
-	return (es->es_pblk & ES_UNWRITTEN) != 0;
+	return (ext4_es_type(es) & EXTENT_STATUS_UNWRITTEN) != 0;
 }
 
 static inline int ext4_es_is_delayed(struct extent_status *es)
 {
-	return (es->es_pblk & ES_DELAYED) != 0;
+	return (ext4_es_type(es) & EXTENT_STATUS_DELAYED) != 0;
 }
 
 static inline int ext4_es_is_hole(struct extent_status *es)
 {
-	return (es->es_pblk & ES_HOLE) != 0;
+	return (ext4_es_type(es) & EXTENT_STATUS_HOLE) != 0;
 }
 
-static inline unsigned int ext4_es_status(struct extent_status *es)
+static inline void ext4_es_set_referenced(struct extent_status *es)
 {
-	return es->es_pblk >> ES_SHIFT;
+	es->es_pblk |= ((ext4_fsblk_t)EXTENT_STATUS_REFERENCED) << ES_SHIFT;
+}
+
+static inline void ext4_es_clear_referenced(struct extent_status *es)
+{
+	es->es_pblk &= ~(((ext4_fsblk_t)EXTENT_STATUS_REFERENCED) << ES_SHIFT);
+}
+
+static inline int ext4_es_is_referenced(struct extent_status *es)
+{
+	return (ext4_es_status(es) & EXTENT_STATUS_REFERENCED) != 0;
 }
 
 static inline ext4_fsblk_t ext4_es_pblock(struct extent_status *es)
@@ -135,23 +157,19 @@ static inline void ext4_es_store_pblock(struct extent_status *es,
 static inline void ext4_es_store_status(struct extent_status *es,
 					unsigned int status)
 {
-	es->es_pblk = (((ext4_fsblk_t)
-			(status & EXTENT_STATUS_FLAGS) << ES_SHIFT) |
-		       (es->es_pblk & ~ES_MASK));
+	es->es_pblk = (((ext4_fsblk_t)status << ES_SHIFT) & ES_MASK) |
+		      (es->es_pblk & ~ES_MASK);
 }
 
 static inline void ext4_es_store_pblock_status(struct extent_status *es,
 					       ext4_fsblk_t pb,
 					       unsigned int status)
 {
-	es->es_pblk = (((ext4_fsblk_t)
-			(status & EXTENT_STATUS_FLAGS) << ES_SHIFT) |
-		       (pb & ~ES_MASK));
+	es->es_pblk = (((ext4_fsblk_t)status << ES_SHIFT) & ES_MASK) |
+		      (pb & ~ES_MASK);
 }
 
 extern int ext4_es_register_shrinker(struct ext4_sb_info *sbi);
 extern void ext4_es_unregister_shrinker(struct ext4_sb_info *sbi);
-extern void ext4_es_lru_add(struct inode *inode);
-extern void ext4_es_lru_del(struct inode *inode);
 
 #endif /* _EXT4_EXTENTS_STATUS_H */

fs/ext4/file.c

@@ -273,24 +273,19 @@ static int ext4_file_open(struct inode * inode, struct file * filp)
  * we determine this extent as a data or a hole according to whether the
  * page cache has data or not.
  */
-static int ext4_find_unwritten_pgoff(struct inode *inode,
-				     int whence,
-				     struct ext4_map_blocks *map,
-				     loff_t *offset)
+static int ext4_find_unwritten_pgoff(struct inode *inode, int whence,
+				     loff_t endoff, loff_t *offset)
 {
 	struct pagevec pvec;
-	unsigned int blkbits;
 	pgoff_t index;
 	pgoff_t end;
-	loff_t endoff;
 	loff_t startoff;
 	loff_t lastoff;
 	int found = 0;
 
-	blkbits = inode->i_sb->s_blocksize_bits;
 	startoff = *offset;
 	lastoff = startoff;
-	endoff = (loff_t)(map->m_lblk + map->m_len) << blkbits;
+
 
 	index = startoff >> PAGE_CACHE_SHIFT;
 	end = endoff >> PAGE_CACHE_SHIFT;
@@ -408,147 +403,144 @@ static int ext4_find_unwritten_pgoff(struct inode *inode,
 static loff_t ext4_seek_data(struct file *file, loff_t offset, loff_t maxsize)
 {
 	struct inode *inode = file->f_mapping->host;
-	struct ext4_map_blocks map;
-	struct extent_status es;
-	ext4_lblk_t start, last, end;
-	loff_t dataoff, isize;
-	int blkbits;
-	int ret = 0;
+	struct fiemap_extent_info fie;
+	struct fiemap_extent ext[2];
+	loff_t next;
+	int i, ret = 0;
 
 	mutex_lock(&inode->i_mutex);
-
-	isize = i_size_read(inode);
-	if (offset >= isize) {
+	if (offset >= inode->i_size) {
 		mutex_unlock(&inode->i_mutex);
 		return -ENXIO;
 	}
+	fie.fi_flags = 0;
+	fie.fi_extents_max = 2;
+	fie.fi_extents_start = (struct fiemap_extent __user *) &ext;
+	while (1) {
+		mm_segment_t old_fs = get_fs();
 
-	blkbits = inode->i_sb->s_blocksize_bits;
-	start = offset >> blkbits;
-	last = start;
-	end = isize >> blkbits;
-	dataoff = offset;
+		fie.fi_extents_mapped = 0;
+		memset(ext, 0, sizeof(*ext) * fie.fi_extents_max);
 
-	do {
-		map.m_lblk = last;
-		map.m_len = end - last + 1;
-		ret = ext4_map_blocks(NULL, inode, &map, 0);
-		if (ret > 0 && !(map.m_flags & EXT4_MAP_UNWRITTEN)) {
-			if (last != start)
-				dataoff = (loff_t)last << blkbits;
+		set_fs(get_ds());
+		ret = ext4_fiemap(inode, &fie, offset, maxsize - offset);
+		set_fs(old_fs);
+		if (ret)
+			break;
+
+		/* No extents found, EOF */
+		if (!fie.fi_extents_mapped) {
+			ret = -ENXIO;
 			break;
 		}
+		for (i = 0; i < fie.fi_extents_mapped; i++) {
+			next = (loff_t)(ext[i].fe_length + ext[i].fe_logical);
 
+			if (offset < (loff_t)ext[i].fe_logical)
+				offset = (loff_t)ext[i].fe_logical;
 			/*
-		 * If there is a delay extent at this offset,
-		 * it will be as a data.
+			 * If extent is not unwritten, then it contains valid
+			 * data, mapped or delayed.
 			 */
-		ext4_es_find_delayed_extent_range(inode, last, last, &es);
-		if (es.es_len != 0 && in_range(last, es.es_lblk, es.es_len)) {
-			if (last != start)
-				dataoff = (loff_t)last << blkbits;
-			break;
-		}
+			if (!(ext[i].fe_flags & FIEMAP_EXTENT_UNWRITTEN))
+				goto out;
 
 			/*
 			 * If there is a unwritten extent at this offset,
 			 * it will be as a data or a hole according to page
 			 * cache that has data or not.
 			 */
-		if (map.m_flags & EXT4_MAP_UNWRITTEN) {
-			int unwritten;
-			unwritten = ext4_find_unwritten_pgoff(inode, SEEK_DATA,
-							      &map, &dataoff);
-			if (unwritten)
-				break;
-		}
-
-		last++;
-		dataoff = (loff_t)last << blkbits;
-	} while (last <= end);
+			if (ext4_find_unwritten_pgoff(inode, SEEK_DATA,
+						      next, &offset))
+				goto out;
 
+			if (ext[i].fe_flags & FIEMAP_EXTENT_LAST) {
+				ret = -ENXIO;
+				goto out;
+			}
+			offset = next;
+		}
+	}
+	if (offset > inode->i_size)
+		offset = inode->i_size;
+out:
 	mutex_unlock(&inode->i_mutex);
+	if (ret)
+		return ret;
 
-	if (dataoff > isize)
-		return -ENXIO;
-
-	return vfs_setpos(file, dataoff, maxsize);
+	return vfs_setpos(file, offset, maxsize);
 }
 
 /*
- * ext4_seek_hole() retrieves the offset for SEEK_HOLE.
+ * ext4_seek_hole() retrieves the offset for SEEK_HOLE
  */
 static loff_t ext4_seek_hole(struct file *file, loff_t offset, loff_t maxsize)
 {
 	struct inode *inode = file->f_mapping->host;
-	struct ext4_map_blocks map;
-	struct extent_status es;
-	ext4_lblk_t start, last, end;
-	loff_t holeoff, isize;
-	int blkbits;
-	int ret = 0;
+	struct fiemap_extent_info fie;
+	struct fiemap_extent ext[2];
+	loff_t next;
+	int i, ret = 0;
 
 	mutex_lock(&inode->i_mutex);
-
-	isize = i_size_read(inode);
-	if (offset >= isize) {
+	if (offset >= inode->i_size) {
 		mutex_unlock(&inode->i_mutex);
 		return -ENXIO;
 	}
 
-	blkbits = inode->i_sb->s_blocksize_bits;
-	start = offset >> blkbits;
-	last = start;
-	end = isize >> blkbits;
-	holeoff = offset;
+	fie.fi_flags = 0;
+	fie.fi_extents_max = 2;
+	fie.fi_extents_start = (struct fiemap_extent __user *)&ext;
+	while (1) {
+		mm_segment_t old_fs = get_fs();
 
-	do {
-		map.m_lblk = last;
-		map.m_len = end - last + 1;
-		ret = ext4_map_blocks(NULL, inode, &map, 0);
-		if (ret > 0 && !(map.m_flags & EXT4_MAP_UNWRITTEN)) {
-			last += ret;
-			holeoff = (loff_t)last << blkbits;
-			continue;
-		}
+		fie.fi_extents_mapped = 0;
+		memset(ext, 0, sizeof(*ext));
 
+		set_fs(get_ds());
+		ret = ext4_fiemap(inode, &fie, offset, maxsize - offset);
+		set_fs(old_fs);
+		if (ret)
+			break;
+
+		/* No extents found */
+		if (!fie.fi_extents_mapped)
+			break;
+
+		for (i = 0; i < fie.fi_extents_mapped; i++) {
+			next = (loff_t)(ext[i].fe_logical + ext[i].fe_length);
 			/*
-		 * If there is a delay extent at this offset,
-		 * we will skip this extent.
+			 * If extent is not unwritten, then it contains valid
+			 * data, mapped or delayed.
 			 */
-		ext4_es_find_delayed_extent_range(inode, last, last, &es);
-		if (es.es_len != 0 && in_range(last, es.es_lblk, es.es_len)) {
-			last = es.es_lblk + es.es_len;
-			holeoff = (loff_t)last << blkbits;
+			if (!(ext[i].fe_flags & FIEMAP_EXTENT_UNWRITTEN)) {
+				if (offset < (loff_t)ext[i].fe_logical)
+					goto out;
+				offset = next;
 				continue;
 			}
-
 			/*
 			 * If there is a unwritten extent at this offset,
 			 * it will be as a data or a hole according to page
 			 * cache that has data or not.
 			 */
-		if (map.m_flags & EXT4_MAP_UNWRITTEN) {
-			int unwritten;
-			unwritten = ext4_find_unwritten_pgoff(inode, SEEK_HOLE,
-							      &map, &holeoff);
-			if (!unwritten) {
-				last += ret;
-				holeoff = (loff_t)last << blkbits;
-				continue;
+			if (ext4_find_unwritten_pgoff(inode, SEEK_HOLE,
+						      next, &offset))
+				goto out;
+
+			offset = next;
+			if (ext[i].fe_flags & FIEMAP_EXTENT_LAST)
+				goto out;
 		}
 	}
-
-		/* find a hole */
-		break;
-	} while (last <= end);
-
+	if (offset > inode->i_size)
+		offset = inode->i_size;
+out:
 	mutex_unlock(&inode->i_mutex);
+	if (ret)
+		return ret;
 
-	if (holeoff > isize)
-		holeoff = isize;
-
-	return vfs_setpos(file, holeoff, maxsize);
+	return vfs_setpos(file, offset, maxsize);
 }
 
 /*

fs/ext4/inline.c

@@ -811,8 +811,11 @@ static int ext4_da_convert_inline_data_to_extent(struct address_space *mapping,
 	ret = __block_write_begin(page, 0, inline_size,
 				  ext4_da_get_block_prep);
 	if (ret) {
+		up_read(&EXT4_I(inode)->xattr_sem);
+		unlock_page(page);
+		page_cache_release(page);
 		ext4_truncate_failed_write(inode);
-		goto out;
+		return ret;
 	}
 
 	SetPageDirty(page);
@@ -870,6 +873,12 @@ int ext4_da_write_inline_data_begin(struct address_space *mapping,
 			goto out_journal;
 	}
 
+	/*
+	 * We cannot recurse into the filesystem as the transaction
+	 * is already started.
+	 */
+	flags |= AOP_FLAG_NOFS;
+
 	if (ret == -ENOSPC) {
 		ret = ext4_da_convert_inline_data_to_extent(mapping,
 							    inode,
@@ -882,11 +891,6 @@ int ext4_da_write_inline_data_begin(struct address_space *mapping,
 		goto out;
 	}
 
-	/*
-	 * We cannot recurse into the filesystem as the transaction
-	 * is already started.
-	 */
-	flags |= AOP_FLAG_NOFS;
 
 	page = grab_cache_page_write_begin(mapping, 0, flags);
 	if (!page) {
@@ -1807,11 +1811,12 @@ int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
 
 int ext4_inline_data_fiemap(struct inode *inode,
 			    struct fiemap_extent_info *fieinfo,
-			    int *has_inline)
+			    int *has_inline, __u64 start, __u64 len)
 {
 	__u64 physical = 0;
-	__u64 length;
-	__u32 flags = FIEMAP_EXTENT_DATA_INLINE | FIEMAP_EXTENT_LAST;
+	__u64 inline_len;
+	__u32 flags = FIEMAP_EXTENT_DATA_INLINE | FIEMAP_EXTENT_NOT_ALIGNED |
+		FIEMAP_EXTENT_LAST;
 	int error = 0;
 	struct ext4_iloc iloc;
 
@@ -1820,6 +1825,13 @@ int ext4_inline_data_fiemap(struct inode *inode,
 		*has_inline = 0;
 		goto out;
 	}
+	inline_len = min_t(size_t, ext4_get_inline_size(inode),
+			   i_size_read(inode));
+	if (start >= inline_len)
+		goto out;
+	if (start + len < inline_len)
+		inline_len = start + len;
+	inline_len -= start;
 
 	error = ext4_get_inode_loc(inode, &iloc);
 	if (error)
@@ -1828,11 +1840,10 @@ int ext4_inline_data_fiemap(struct inode *inode,
 	physical = (__u64)iloc.bh->b_blocknr << inode->i_sb->s_blocksize_bits;
 	physical += (char *)ext4_raw_inode(&iloc) - iloc.bh->b_data;
 	physical += offsetof(struct ext4_inode, i_block);
-	length = i_size_read(inode);
 
 	if (physical)
-		error = fiemap_fill_next_extent(fieinfo, 0, physical,
-						length, flags);
+		error = fiemap_fill_next_extent(fieinfo, start, physical,
+						inline_len, flags);
 	brelse(iloc.bh);
 out:
 	up_read(&EXT4_I(inode)->xattr_sem);

fs/ext4/inode.c

@@ -416,11 +416,6 @@ static void ext4_map_blocks_es_recheck(handle_t *handle,
 	}
 	if (!(flags & EXT4_GET_BLOCKS_NO_LOCK))
 		up_read((&EXT4_I(inode)->i_data_sem));
-	/*
-	 * Clear EXT4_MAP_FROM_CLUSTER and EXT4_MAP_BOUNDARY flag
-	 * because it shouldn't be marked in es_map->m_flags.
-	 */
-	map->m_flags &= ~(EXT4_MAP_FROM_CLUSTER | EXT4_MAP_BOUNDARY);
 
 	/*
 	 * We don't check m_len because extent will be collpased in status
@@ -491,7 +486,6 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
 
 	/* Lookup extent status tree firstly */
 	if (ext4_es_lookup_extent(inode, map->m_lblk, &es)) {
-		ext4_es_lru_add(inode);
 		if (ext4_es_is_written(&es) || ext4_es_is_unwritten(&es)) {
 			map->m_pblk = ext4_es_pblock(&es) +
 					map->m_lblk - es.es_lblk;
@@ -1393,7 +1387,6 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
 
 	/* Lookup extent status tree firstly */
 	if (ext4_es_lookup_extent(inode, iblock, &es)) {
-		ext4_es_lru_add(inode);
 		if (ext4_es_is_hole(&es)) {
 			retval = 0;
 			down_read(&EXT4_I(inode)->i_data_sem);
@@ -1434,24 +1427,12 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
 	 * file system block.
 	 */
 	down_read(&EXT4_I(inode)->i_data_sem);
-	if (ext4_has_inline_data(inode)) {
-		/*
-		 * We will soon create blocks for this page, and let
-		 * us pretend as if the blocks aren't allocated yet.
-		 * In case of clusters, we have to handle the work
-		 * of mapping from cluster so that the reserved space
-		 * is calculated properly.
-		 */
-		if ((EXT4_SB(inode->i_sb)->s_cluster_ratio > 1) &&
-		    ext4_find_delalloc_cluster(inode, map->m_lblk))
-			map->m_flags |= EXT4_MAP_FROM_CLUSTER;
+	if (ext4_has_inline_data(inode))
 		retval = 0;
-	} else if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
-		retval = ext4_ext_map_blocks(NULL, inode, map,
-					     EXT4_GET_BLOCKS_NO_PUT_HOLE);
+	else if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
+		retval = ext4_ext_map_blocks(NULL, inode, map, 0);
 	else
-		retval = ext4_ind_map_blocks(NULL, inode, map,
-					     EXT4_GET_BLOCKS_NO_PUT_HOLE);
+		retval = ext4_ind_map_blocks(NULL, inode, map, 0);
 
 add_delayed:
 	if (retval == 0) {
@@ -1465,7 +1446,8 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
 		 * then we don't need to reserve it again. However we still need
 		 * to reserve metadata for every block we're going to write.
 		 */
-		if (!(map->m_flags & EXT4_MAP_FROM_CLUSTER)) {
+		if (EXT4_SB(inode->i_sb)->s_cluster_ratio <= 1 ||
+		    !ext4_find_delalloc_cluster(inode, map->m_lblk)) {
 			ret = ext4_da_reserve_space(inode, iblock);
 			if (ret) {
 				/* not enough space to reserve */
@@ -1481,11 +1463,6 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
 			goto out_unlock;
 		}
 
-		/* Clear EXT4_MAP_FROM_CLUSTER flag since its purpose is served
-		 * and it should not appear on the bh->b_state.
-		 */
-		map->m_flags &= ~EXT4_MAP_FROM_CLUSTER;
-
 		map_bh(bh, inode->i_sb, invalid_block);
 		set_buffer_new(bh);
 		set_buffer_delay(bh);
@@ -3643,7 +3620,7 @@ void ext4_truncate(struct inode *inode)
 	 * If this was a simple ftruncate() and the file will remain alive,
 	 * then we need to clear up the orphan record which we created above.
 	 * However, if this was a real unlink then we were called by
-	 * ext4_delete_inode(), and we allow that function to clean up the
+	 * ext4_evict_inode(), and we allow that function to clean up the
 	 * orphan info for us.
 	 */
 	if (inode->i_nlink)

fs/ext4/ioctl.c

@@ -78,8 +78,6 @@ static void swap_inode_data(struct inode *inode1, struct inode *inode2)
 	memswap(&ei1->i_disksize, &ei2->i_disksize, sizeof(ei1->i_disksize));
 	ext4_es_remove_extent(inode1, 0, EXT_MAX_BLOCKS);
 	ext4_es_remove_extent(inode2, 0, EXT_MAX_BLOCKS);
-	ext4_es_lru_del(inode1);
-	ext4_es_lru_del(inode2);
 
 	isize = i_size_read(inode1);
 	i_size_write(inode1, i_size_read(inode2));

fs/ext4/mballoc.c

@@ -2358,7 +2358,7 @@ int ext4_mb_alloc_groupinfo(struct super_block *sb, ext4_group_t ngroups)
 	if (sbi->s_group_info) {
 		memcpy(new_groupinfo, sbi->s_group_info,
 		       sbi->s_group_info_size * sizeof(*sbi->s_group_info));
-		ext4_kvfree(sbi->s_group_info);
+		kvfree(sbi->s_group_info);
 	}
 	sbi->s_group_info = new_groupinfo;
 	sbi->s_group_info_size = size / sizeof(*sbi->s_group_info);
@@ -2385,7 +2385,7 @@ int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group,
 	if (group % EXT4_DESC_PER_BLOCK(sb) == 0) {
 		metalen = sizeof(*meta_group_info) <<
 			EXT4_DESC_PER_BLOCK_BITS(sb);
-		meta_group_info = kmalloc(metalen, GFP_KERNEL);
+		meta_group_info = kmalloc(metalen, GFP_NOFS);
 		if (meta_group_info == NULL) {
 			ext4_msg(sb, KERN_ERR, "can't allocate mem "
 				 "for a buddy group");
@@ -2399,7 +2399,7 @@ int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group,
 		sbi->s_group_info[group >> EXT4_DESC_PER_BLOCK_BITS(sb)];
 	i = group & (EXT4_DESC_PER_BLOCK(sb) - 1);
 
-	meta_group_info[i] = kmem_cache_zalloc(cachep, GFP_KERNEL);
+	meta_group_info[i] = kmem_cache_zalloc(cachep, GFP_NOFS);
 	if (meta_group_info[i] == NULL) {
 		ext4_msg(sb, KERN_ERR, "can't allocate buddy mem");
 		goto exit_group_info;
@@ -2428,7 +2428,7 @@ int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group,
 	{
 		struct buffer_head *bh;
 		meta_group_info[i]->bb_bitmap =
-			kmalloc(sb->s_blocksize, GFP_KERNEL);
+			kmalloc(sb->s_blocksize, GFP_NOFS);
 		BUG_ON(meta_group_info[i]->bb_bitmap == NULL);
 		bh = ext4_read_block_bitmap(sb, group);
 		BUG_ON(bh == NULL);
@@ -2495,7 +2495,7 @@ static int ext4_mb_init_backend(struct super_block *sb)
 		kfree(sbi->s_group_info[i]);
 	iput(sbi->s_buddy_cache);
 err_freesgi:
-	ext4_kvfree(sbi->s_group_info);
+	kvfree(sbi->s_group_info);
 	return -ENOMEM;
 }
 
@@ -2708,11 +2708,10 @@ int ext4_mb_release(struct super_block *sb)
 			EXT4_DESC_PER_BLOCK_BITS(sb);
 		for (i = 0; i < num_meta_group_infos; i++)
 			kfree(sbi->s_group_info[i]);
-		ext4_kvfree(sbi->s_group_info);
+		kvfree(sbi->s_group_info);
 	}
 	kfree(sbi->s_mb_offsets);
 	kfree(sbi->s_mb_maxs);
-	if (sbi->s_buddy_cache)
 	iput(sbi->s_buddy_cache);
 	if (sbi->s_mb_stats) {
 		ext4_msg(sb, KERN_INFO,

fs/ext4/migrate.c

@@ -592,7 +592,7 @@ int ext4_ext_migrate(struct inode *inode)
 
 	/*
 	 * set the  i_blocks count to zero
-	 * so that the ext4_delete_inode does the
+	 * so that the ext4_evict_inode() does the
 	 * right job
 	 *
 	 * We don't need to take the i_lock because

fs/ext4/move_extent.c

@@ -273,6 +273,7 @@ move_extent_per_page(struct file *o_filp, struct inode *donor_inode,
 	int replaced_count = 0;
 	int from = data_offset_in_page << orig_inode->i_blkbits;
 	int blocks_per_page = PAGE_CACHE_SIZE >> orig_inode->i_blkbits;
+	struct super_block *sb = orig_inode->i_sb;
 
 	/*
 	 * It needs twice the amount of ordinary journal buffers because
@@ -405,10 +406,13 @@ move_extent_per_page(struct file *o_filp, struct inode *donor_inode,
 	page_cache_release(pagep[1]);
 stop_journal:
 	ext4_journal_stop(handle);
+	if (*err == -ENOSPC &&
+	    ext4_should_retry_alloc(sb, &retries))
+		goto again;
 	/* Buffer was busy because probably is pinned to journal transaction,
 	 * force transaction commit may help to free it. */
-	if (*err == -EBUSY && ext4_should_retry_alloc(orig_inode->i_sb,
-						      &retries))
+	if (*err == -EBUSY && retries++ < 4 && EXT4_SB(sb)->s_journal &&
+	    jbd2_journal_force_commit_nested(EXT4_SB(sb)->s_journal))
 		goto again;
 	return replaced_count;
 

fs/ext4/namei.c

@@ -2814,7 +2814,6 @@ static int ext4_unlink(struct inode *dir, struct dentry *dentry)
 		ext4_orphan_add(handle, inode);
 	inode->i_ctime = ext4_current_time(inode);
 	ext4_mark_inode_dirty(handle, inode);
-	retval = 0;
 
 end_unlink:
 	brelse(bh);

fs/ext4/resize.c

@@ -856,7 +856,7 @@ static int add_new_gdb(handle_t *handle, struct inode *inode,
 	n_group_desc[gdb_num] = gdb_bh;
 	EXT4_SB(sb)->s_group_desc = n_group_desc;
 	EXT4_SB(sb)->s_gdb_count++;
-	ext4_kvfree(o_group_desc);
+	kvfree(o_group_desc);
 
 	le16_add_cpu(&es->s_reserved_gdt_blocks, -1);
 	err = ext4_handle_dirty_super(handle, sb);
@@ -866,7 +866,7 @@ static int add_new_gdb(handle_t *handle, struct inode *inode,
 	return err;
 
 exit_inode:
-	ext4_kvfree(n_group_desc);
+	kvfree(n_group_desc);
 	brelse(iloc.bh);
 exit_dind:
 	brelse(dind);
@@ -909,7 +909,7 @@ static int add_new_gdb_meta_bg(struct super_block *sb,
 	n_group_desc[gdb_num] = gdb_bh;
 	EXT4_SB(sb)->s_group_desc = n_group_desc;
 	EXT4_SB(sb)->s_gdb_count++;
-	ext4_kvfree(o_group_desc);
+	kvfree(o_group_desc);
 	BUFFER_TRACE(gdb_bh, "get_write_access");
 	err = ext4_journal_get_write_access(handle, gdb_bh);
 	if (unlikely(err))

fs/ext4/super.c

@@ -176,15 +176,6 @@ void *ext4_kvzalloc(size_t size, gfp_t flags)
 	return ret;
 }
 
-void ext4_kvfree(void *ptr)
-{
-	if (is_vmalloc_addr(ptr))
-		vfree(ptr);
-	else
-		kfree(ptr);
-
-}
-
 ext4_fsblk_t ext4_block_bitmap(struct super_block *sb,
 			       struct ext4_group_desc *bg)
 {
@@ -811,8 +802,8 @@ static void ext4_put_super(struct super_block *sb)
 
 	for (i = 0; i < sbi->s_gdb_count; i++)
 		brelse(sbi->s_group_desc[i]);
-	ext4_kvfree(sbi->s_group_desc);
-	ext4_kvfree(sbi->s_flex_groups);
+	kvfree(sbi->s_group_desc);
+	kvfree(sbi->s_flex_groups);
 	percpu_counter_destroy(&sbi->s_freeclusters_counter);
 	percpu_counter_destroy(&sbi->s_freeinodes_counter);
 	percpu_counter_destroy(&sbi->s_dirs_counter);
@@ -880,10 +871,10 @@ static struct inode *ext4_alloc_inode(struct super_block *sb)
 	spin_lock_init(&ei->i_prealloc_lock);
 	ext4_es_init_tree(&ei->i_es_tree);
 	rwlock_init(&ei->i_es_lock);
-	INIT_LIST_HEAD(&ei->i_es_lru);
+	INIT_LIST_HEAD(&ei->i_es_list);
 	ei->i_es_all_nr = 0;
-	ei->i_es_lru_nr = 0;
-	ei->i_touch_when = 0;
+	ei->i_es_shk_nr = 0;
+	ei->i_es_shrink_lblk = 0;
 	ei->i_reserved_data_blocks = 0;
 	ei->i_reserved_meta_blocks = 0;
 	ei->i_allocated_meta_blocks = 0;
@@ -973,7 +964,6 @@ void ext4_clear_inode(struct inode *inode)
 	dquot_drop(inode);
 	ext4_discard_preallocations(inode);
 	ext4_es_remove_extent(inode, 0, EXT_MAX_BLOCKS);
-	ext4_es_lru_del(inode);
 	if (EXT4_I(inode)->jinode) {
 		jbd2_journal_release_jbd_inode(EXT4_JOURNAL(inode),
 					       EXT4_I(inode)->jinode);
@@ -1153,7 +1143,7 @@ enum {
 	Opt_inode_readahead_blks, Opt_journal_ioprio,
 	Opt_dioread_nolock, Opt_dioread_lock,
 	Opt_discard, Opt_nodiscard, Opt_init_itable, Opt_noinit_itable,
-	Opt_max_dir_size_kb,
+	Opt_max_dir_size_kb, Opt_nojournal_checksum,
 };
 
 static const match_table_t tokens = {
@@ -1187,6 +1177,7 @@ static const match_table_t tokens = {
 	{Opt_journal_dev, "journal_dev=%u"},
 	{Opt_journal_path, "journal_path=%s"},
 	{Opt_journal_checksum, "journal_checksum"},
+	{Opt_nojournal_checksum, "nojournal_checksum"},
 	{Opt_journal_async_commit, "journal_async_commit"},
 	{Opt_abort, "abort"},
 	{Opt_data_journal, "data=journal"},
@@ -1368,6 +1359,8 @@ static const struct mount_opts {
 	 MOPT_EXT4_ONLY | MOPT_SET | MOPT_EXPLICIT},
 	{Opt_nodelalloc, EXT4_MOUNT_DELALLOC,
 	 MOPT_EXT4_ONLY | MOPT_CLEAR},
+	{Opt_nojournal_checksum, EXT4_MOUNT_JOURNAL_CHECKSUM,
+	 MOPT_EXT4_ONLY | MOPT_CLEAR},
 	{Opt_journal_checksum, EXT4_MOUNT_JOURNAL_CHECKSUM,
 	 MOPT_EXT4_ONLY | MOPT_SET},
 	{Opt_journal_async_commit, (EXT4_MOUNT_JOURNAL_ASYNC_COMMIT |
@@ -1709,6 +1702,12 @@ static int parse_options(char *options, struct super_block *sb,
 			return 0;
 		}
 	}
+	if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA &&
+	    test_opt(sb, JOURNAL_ASYNC_COMMIT)) {
+		ext4_msg(sb, KERN_ERR, "can't mount with journal_async_commit "
+			 "in data=ordered mode");
+		return 0;
+	}
 	return 1;
 }
 
@@ -1946,7 +1945,7 @@ int ext4_alloc_flex_bg_array(struct super_block *sb, ext4_group_t ngroup)
 		memcpy(new_groups, sbi->s_flex_groups,
 		       (sbi->s_flex_groups_allocated *
 			sizeof(struct flex_groups)));
-		ext4_kvfree(sbi->s_flex_groups);
+		kvfree(sbi->s_flex_groups);
 	}
 	sbi->s_flex_groups = new_groups;
 	sbi->s_flex_groups_allocated = size / sizeof(struct flex_groups);
@@ -3317,7 +3316,7 @@ int ext4_calculate_overhead(struct super_block *sb)
 	struct ext4_super_block *es = sbi->s_es;
 	ext4_group_t i, ngroups = ext4_get_groups_count(sb);
 	ext4_fsblk_t overhead = 0;
-	char *buf = (char *) get_zeroed_page(GFP_KERNEL);
+	char *buf = (char *) get_zeroed_page(GFP_NOFS);
 
 	if (!buf)
 		return -ENOMEM;
@@ -3345,8 +3344,8 @@ int ext4_calculate_overhead(struct super_block *sb)
 			memset(buf, 0, PAGE_SIZE);
 		cond_resched();
 	}
-	/* Add the journal blocks as well */
-	if (sbi->s_journal)
+	/* Add the internal journal blocks as well */
+	if (sbi->s_journal && !sbi->journal_bdev)
 		overhead += EXT4_NUM_B2C(sbi, sbi->s_journal->j_maxlen);
 
 	sbi->s_overhead = overhead;
@@ -4232,7 +4231,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 failed_mount6:
 	ext4_mb_release(sb);
 	if (sbi->s_flex_groups)
-		ext4_kvfree(sbi->s_flex_groups);
+		kvfree(sbi->s_flex_groups);
 	percpu_counter_destroy(&sbi->s_freeclusters_counter);
 	percpu_counter_destroy(&sbi->s_freeinodes_counter);
 	percpu_counter_destroy(&sbi->s_dirs_counter);
@@ -4261,7 +4260,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 failed_mount2:
 	for (i = 0; i < db_count; i++)
 		brelse(sbi->s_group_desc[i]);
-	ext4_kvfree(sbi->s_group_desc);
+	kvfree(sbi->s_group_desc);
 failed_mount:
 	if (sbi->s_chksum_driver)
 		crypto_free_shash(sbi->s_chksum_driver);
@@ -4862,6 +4861,14 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 		goto restore_opts;
 	}
 
+	if ((old_opts.s_mount_opt & EXT4_MOUNT_JOURNAL_CHECKSUM) ^
+	    test_opt(sb, JOURNAL_CHECKSUM)) {
+		ext4_msg(sb, KERN_ERR, "changing journal_checksum "
+			 "during remount not supported");
+		err = -EINVAL;
+		goto restore_opts;
+	}
+
 	if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) {
 		if (test_opt2(sb, EXPLICIT_DELALLOC)) {
 			ext4_msg(sb, KERN_ERR, "can't mount with "

fs/jbd2/journal.c

@@ -1714,7 +1714,6 @@ int jbd2_journal_destroy(journal_t *journal)
 
 	if (journal->j_proc_entry)
 		jbd2_stats_proc_exit(journal);
-	if (journal->j_inode)
 	iput(journal->j_inode);
 	if (journal->j_revoke)
 		jbd2_journal_destroy_revoke(journal);

include/trace/events/ext4.h

@@ -43,15 +43,13 @@ struct extent_status;
 	{ EXT4_GET_BLOCKS_METADATA_NOFAIL,	"METADATA_NOFAIL" },	\
 	{ EXT4_GET_BLOCKS_NO_NORMALIZE,		"NO_NORMALIZE" },	\
 	{ EXT4_GET_BLOCKS_KEEP_SIZE,		"KEEP_SIZE" },		\
-	{ EXT4_GET_BLOCKS_NO_LOCK,		"NO_LOCK" },		\
-	{ EXT4_GET_BLOCKS_NO_PUT_HOLE,		"NO_PUT_HOLE" })
+	{ EXT4_GET_BLOCKS_NO_LOCK,		"NO_LOCK" })
 
 #define show_mflags(flags) __print_flags(flags, "",	\
 	{ EXT4_MAP_NEW,		"N" },			\
 	{ EXT4_MAP_MAPPED,	"M" },			\
 	{ EXT4_MAP_UNWRITTEN,	"U" },			\
-	{ EXT4_MAP_BOUNDARY,	"B" },			\
-	{ EXT4_MAP_FROM_CLUSTER, "C" })
+	{ EXT4_MAP_BOUNDARY,	"B" })
 
 #define show_free_flags(flags) __print_flags(flags, "|",	\
 	{ EXT4_FREE_BLOCKS_METADATA,		"METADATA" },	\
@@ -2452,15 +2450,14 @@ TRACE_EVENT(ext4_collapse_range,
 
 TRACE_EVENT(ext4_es_shrink,
 	TP_PROTO(struct super_block *sb, int nr_shrunk, u64 scan_time,
-		 int skip_precached, int nr_skipped, int retried),
+		 int nr_skipped, int retried),
 
-	TP_ARGS(sb, nr_shrunk, scan_time, skip_precached, nr_skipped, retried),
+	TP_ARGS(sb, nr_shrunk, scan_time, nr_skipped, retried),
 
 	TP_STRUCT__entry(
 		__field(	dev_t,		dev		)
 		__field(	int,		nr_shrunk	)
 		__field(	unsigned long long, scan_time	)
-		__field(	int,		skip_precached	)
 		__field(	int,		nr_skipped	)
 		__field(	int,		retried		)
 	),
@@ -2469,16 +2466,14 @@ TRACE_EVENT(ext4_es_shrink,
 		__entry->dev		= sb->s_dev;
 		__entry->nr_shrunk	= nr_shrunk;
 		__entry->scan_time	= div_u64(scan_time, 1000);
-		__entry->skip_precached = skip_precached;
 		__entry->nr_skipped	= nr_skipped;
 		__entry->retried	= retried;
 	),
 
-	TP_printk("dev %d,%d nr_shrunk %d, scan_time %llu skip_precached %d "
+	TP_printk("dev %d,%d nr_shrunk %d, scan_time %llu "
 		  "nr_skipped %d retried %d",
 		  MAJOR(__entry->dev), MINOR(__entry->dev), __entry->nr_shrunk,
-		  __entry->scan_time, __entry->skip_precached,
-		  __entry->nr_skipped, __entry->retried)
+		  __entry->scan_time, __entry->nr_skipped, __entry->retried)
 );
 
 #endif /* _TRACE_EXT4_H */