Merge commit 'origin/master' into next

57 files changed, 1518 insertions, 1079 deletions

diff --git a/fs/bio.c b/fs/bio.c
index 5f80848c320c..24c914043532 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -25,7 +25,6 @@
 #include <linux/module.h>
 #include <linux/mempool.h>
 #include <linux/workqueue.h>
-#include <linux/blktrace_api.h>
 #include <scsi/sg.h>		/* for struct sg_iovec */
 
 #include <trace/events/block.h>
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 0d50d49d990a..d28d29c95f7c 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -42,6 +42,8 @@
 static struct extent_io_ops btree_extent_io_ops;
 static void end_workqueue_fn(struct btrfs_work *work);
 
+static atomic_t btrfs_bdi_num = ATOMIC_INIT(0);
+
 /*
  * end_io_wq structs are used to do processing in task context when an IO is
  * complete.  This is used during reads to verify checksums, and it is used
@@ -1342,12 +1344,25 @@ static void btrfs_unplug_io_fn(struct backing_dev_info *bdi, struct page *page)
 	free_extent_map(em);
 }
 
+/*
+ * If this fails, caller must call bdi_destroy() to get rid of the
+ * bdi again.
+ */
 static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
 {
-	bdi_init(bdi);
+	int err;
+
+	bdi->capabilities = BDI_CAP_MAP_COPY;
+	err = bdi_init(bdi);
+	if (err)
+		return err;
+
+	err = bdi_register(bdi, NULL, "btrfs-%d",
+				atomic_inc_return(&btrfs_bdi_num));
+	if (err)
+		return err;
+
 	bdi->ra_pages	= default_backing_dev_info.ra_pages;
-	bdi->state		= 0;
-	bdi->capabilities	= default_backing_dev_info.capabilities;
 	bdi->unplug_io_fn	= btrfs_unplug_io_fn;
 	bdi->unplug_io_data	= info;
 	bdi->congested_fn	= btrfs_congested_fn;
@@ -1569,7 +1584,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->sb = sb;
 	fs_info->max_extent = (u64)-1;
 	fs_info->max_inline = 8192 * 1024;
-	setup_bdi(fs_info, &fs_info->bdi);
+	if (setup_bdi(fs_info, &fs_info->bdi))
+		goto fail_bdi;
 	fs_info->btree_inode = new_inode(sb);
 	fs_info->btree_inode->i_ino = 1;
 	fs_info->btree_inode->i_nlink = 1;
@@ -1946,8 +1962,8 @@ fail_iput:
 
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
+fail_bdi:
 	bdi_destroy(&fs_info->bdi);
-
 fail:
 	kfree(extent_root);
 	kfree(tree_root);
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 2e177d7f4bb9..4e83457ea253 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -543,13 +543,13 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 			btrfs_free_log(trans, root);
 			btrfs_update_reloc_root(trans, root);
 
-			if (root->commit_root == root->node)
-				continue;
-
-			free_extent_buffer(root->commit_root);
-			root->commit_root = btrfs_root_node(root);
+			if (root->commit_root != root->node) {
+				free_extent_buffer(root->commit_root);
+				root->commit_root = btrfs_root_node(root);
+				btrfs_set_root_node(&root->root_item,
+						    root->node);
+			}
 
-			btrfs_set_root_node(&root->root_item, root->node);
 			err = btrfs_update_root(trans, fs_info->tree_root,
 						&root->root_key,
 						&root->root_item);
diff --git a/fs/fat/cache.c b/fs/fat/cache.c
index b42602298087..923990e4f16e 100644
--- a/fs/fat/cache.c
+++ b/fs/fat/cache.c
@@ -241,7 +241,7 @@ int fat_get_cluster(struct inode *inode, int cluster, int *fclus, int *dclus)
 	while (*fclus < cluster) {
 		/* prevent the infinite loop of cluster chain */
 		if (*fclus > limit) {
-			fat_fs_panic(sb, "%s: detected the cluster chain loop"
+			fat_fs_error(sb, "%s: detected the cluster chain loop"
 				     " (i_pos %lld)", __func__,
 				     MSDOS_I(inode)->i_pos);
 			nr = -EIO;
@@ -252,7 +252,7 @@ int fat_get_cluster(struct inode *inode, int cluster, int *fclus, int *dclus)
 		if (nr < 0)
 			goto out;
 		else if (nr == FAT_ENT_FREE) {
-			fat_fs_panic(sb, "%s: invalid cluster chain"
+			fat_fs_error(sb, "%s: invalid cluster chain"
 				     " (i_pos %lld)", __func__,
 				     MSDOS_I(inode)->i_pos);
 			nr = -EIO;
@@ -285,7 +285,7 @@ static int fat_bmap_cluster(struct inode *inode, int cluster)
 	if (ret < 0)
 		return ret;
 	else if (ret == FAT_ENT_EOF) {
-		fat_fs_panic(sb, "%s: request beyond EOF (i_pos %lld)",
+		fat_fs_error(sb, "%s: request beyond EOF (i_pos %lld)",
 			     __func__, MSDOS_I(inode)->i_pos);
 		return -EIO;
 	}
diff --git a/fs/fat/dir.c b/fs/fat/dir.c
index f3500294eec5..3b8e71b412fd 100644
--- a/fs/fat/dir.c
+++ b/fs/fat/dir.c
@@ -1334,7 +1334,7 @@ found:
 			goto error_remove;
 		}
 		if (dir->i_size & (sbi->cluster_size - 1)) {
-			fat_fs_panic(sb, "Odd directory size");
+			fat_fs_error(sb, "Odd directory size");
 			dir->i_size = (dir->i_size + sbi->cluster_size - 1)
 				& ~((loff_t)sbi->cluster_size - 1);
 		}
diff --git a/fs/fat/fat.h b/fs/fat/fat.h
index e4d88527b5dd..adb0e72a176d 100644
--- a/fs/fat/fat.h
+++ b/fs/fat/fat.h
@@ -17,6 +17,10 @@
 #define VFAT_SFN_CREATE_WIN95	0x0100 /* emulate win95 rule for create */
 #define VFAT_SFN_CREATE_WINNT	0x0200 /* emulate winnt rule for create */
 
+#define FAT_ERRORS_CONT		1      /* ignore error and continue */
+#define FAT_ERRORS_PANIC	2      /* panic on error */
+#define FAT_ERRORS_RO		3      /* remount r/o on error */
+
 struct fat_mount_options {
 	uid_t fs_uid;
 	gid_t fs_gid;
@@ -26,6 +30,7 @@ struct fat_mount_options {
 	char *iocharset;          /* Charset used for filename input/display */
 	unsigned short shortname; /* flags for shortname display/create rule */
 	unsigned char name_check; /* r = relaxed, n = normal, s = strict */
+	unsigned char errors;	  /* On error: continue, panic, remount-ro */
 	unsigned short allow_utime;/* permission for setting the [am]time */
 	unsigned quiet:1,         /* set = fake successful chmods and chowns */
 		 showexec:1,      /* set = only set x bit for com/exe/bat */
@@ -316,7 +321,7 @@ extern int fat_fill_super(struct super_block *sb, void *data, int silent,
 extern int fat_flush_inodes(struct super_block *sb, struct inode *i1,
 		            struct inode *i2);
 /* fat/misc.c */
-extern void fat_fs_panic(struct super_block *s, const char *fmt, ...)
+extern void fat_fs_error(struct super_block *s, const char *fmt, ...)
 	__attribute__ ((format (printf, 2, 3))) __cold;
 extern void fat_clusters_flush(struct super_block *sb);
 extern int fat_chain_add(struct inode *inode, int new_dclus, int nr_cluster);
diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c
index 618f5305c2e4..a81037721a6f 100644
--- a/fs/fat/fatent.c
+++ b/fs/fat/fatent.c
@@ -348,7 +348,7 @@ int fat_ent_read(struct inode *inode, struct fat_entry *fatent, int entry)
 
 	if (entry < FAT_START_ENT || sbi->max_cluster <= entry) {
 		fatent_brelse(fatent);
-		fat_fs_panic(sb, "invalid access to FAT (entry 0x%08x)", entry);
+		fat_fs_error(sb, "invalid access to FAT (entry 0x%08x)", entry);
 		return -EIO;
 	}
 
@@ -560,7 +560,7 @@ int fat_free_clusters(struct inode *inode, int cluster)
 			err = cluster;
 			goto error;
 		} else if (cluster == FAT_ENT_FREE) {
-			fat_fs_panic(sb, "%s: deleting FAT entry beyond EOF",
+			fat_fs_error(sb, "%s: deleting FAT entry beyond EOF",
 				     __func__);
 			err = -EIO;
 			goto error;
diff --git a/fs/fat/file.c b/fs/fat/file.c
index e955a56b4e5e..b28ea646ff60 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -18,106 +18,112 @@
 #include <linux/security.h>
 #include "fat.h"
 
-int fat_generic_ioctl(struct inode *inode, struct file *filp,
-		      unsigned int cmd, unsigned long arg)
+static int fat_ioctl_get_attributes(struct inode *inode, u32 __user *user_attr)
 {
+	u32 attr;
+
+	mutex_lock(&inode->i_mutex);
+	attr = fat_make_attrs(inode);
+	mutex_unlock(&inode->i_mutex);
+
+	return put_user(attr, user_attr);
+}
+
+static int fat_ioctl_set_attributes(struct file *file, u32 __user *user_attr)
+{
+	struct inode *inode = file->f_path.dentry->d_inode;
 	struct msdos_sb_info *sbi = MSDOS_SB(inode->i_sb);
-	u32 __user *user_attr = (u32 __user *)arg;
+	int is_dir = S_ISDIR(inode->i_mode);
+	u32 attr, oldattr;
+	struct iattr ia;
+	int err;
 
-	switch (cmd) {
-	case FAT_IOCTL_GET_ATTRIBUTES:
-	{
-		u32 attr;
+	err = get_user(attr, user_attr);
+	if (err)
+		goto out;
 
-		mutex_lock(&inode->i_mutex);
-		attr = fat_make_attrs(inode);
-		mutex_unlock(&inode->i_mutex);
+	mutex_lock(&inode->i_mutex);
+	err = mnt_want_write(file->f_path.mnt);
+	if (err)
+		goto out_unlock_inode;
 
-		return put_user(attr, user_attr);
+	/*
+	 * ATTR_VOLUME and ATTR_DIR cannot be changed; this also
+	 * prevents the user from turning us into a VFAT
+	 * longname entry.  Also, we obviously can't set
+	 * any of the NTFS attributes in the high 24 bits.
+	 */
+	attr &= 0xff & ~(ATTR_VOLUME | ATTR_DIR);
+	/* Merge in ATTR_VOLUME and ATTR_DIR */
+	attr |= (MSDOS_I(inode)->i_attrs & ATTR_VOLUME) |
+		(is_dir ? ATTR_DIR : 0);
+	oldattr = fat_make_attrs(inode);
+
+	/* Equivalent to a chmod() */
+	ia.ia_valid = ATTR_MODE | ATTR_CTIME;
+	ia.ia_ctime = current_fs_time(inode->i_sb);
+	if (is_dir)
+		ia.ia_mode = fat_make_mode(sbi, attr, S_IRWXUGO);
+	else {
+		ia.ia_mode = fat_make_mode(sbi, attr,
+			S_IRUGO | S_IWUGO | (inode->i_mode & S_IXUGO));
 	}
-	case FAT_IOCTL_SET_ATTRIBUTES:
-	{
-		u32 attr, oldattr;
-		int err, is_dir = S_ISDIR(inode->i_mode);
-		struct iattr ia;
 
-		err = get_user(attr, user_attr);
-		if (err)
-			return err;
+	/* The root directory has no attributes */
+	if (inode->i_ino == MSDOS_ROOT_INO && attr != ATTR_DIR) {
+		err = -EINVAL;
+		goto out_drop_write;
+	}
 
-		mutex_lock(&inode->i_mutex);
-
-		err = mnt_want_write(filp->f_path.mnt);
-		if (err)
-			goto up_no_drop_write;
-
-		/*
-		 * ATTR_VOLUME and ATTR_DIR cannot be changed; this also
-		 * prevents the user from turning us into a VFAT
-		 * longname entry.  Also, we obviously can't set
-		 * any of the NTFS attributes in the high 24 bits.
-		 */
-		attr &= 0xff & ~(ATTR_VOLUME | ATTR_DIR);
-		/* Merge in ATTR_VOLUME and ATTR_DIR */
-		attr |= (MSDOS_I(inode)->i_attrs & ATTR_VOLUME) |
-			(is_dir ? ATTR_DIR : 0);
-		oldattr = fat_make_attrs(inode);
-
-		/* Equivalent to a chmod() */
-		ia.ia_valid = ATTR_MODE | ATTR_CTIME;
-		ia.ia_ctime = current_fs_time(inode->i_sb);
-		if (is_dir)
-			ia.ia_mode = fat_make_mode(sbi, attr, S_IRWXUGO);
-		else {
-			ia.ia_mode = fat_make_mode(sbi, attr,
-				S_IRUGO | S_IWUGO | (inode->i_mode & S_IXUGO));
-		}
+	if (sbi->options.sys_immutable &&
+	    ((attr | oldattr) & ATTR_SYS) &&
+	    !capable(CAP_LINUX_IMMUTABLE)) {
+		err = -EPERM;
+		goto out_drop_write;
+	}
 
-		/* The root directory has no attributes */
-		if (inode->i_ino == MSDOS_ROOT_INO && attr != ATTR_DIR) {
-			err = -EINVAL;
-			goto up;
-		}
+	/*
+	 * The security check is questionable...  We single
+	 * out the RO attribute for checking by the security
+	 * module, just because it maps to a file mode.
+	 */
+	err = security_inode_setattr(file->f_path.dentry, &ia);
+	if (err)
+		goto out_drop_write;
 
-		if (sbi->options.sys_immutable) {
-			if ((attr | oldattr) & ATTR_SYS) {
-				if (!capable(CAP_LINUX_IMMUTABLE)) {
-					err = -EPERM;
-					goto up;
-				}
-			}
-		}
+	/* This MUST be done before doing anything irreversible... */
+	err = fat_setattr(file->f_path.dentry, &ia);
+	if (err)
+		goto out_drop_write;
+
+	fsnotify_change(file->f_path.dentry, ia.ia_valid);
+	if (sbi->options.sys_immutable) {
+		if (attr & ATTR_SYS)
+			inode->i_flags |= S_IMMUTABLE;
+		else
+			inode->i_flags &= S_IMMUTABLE;
+	}
 
-		/*
-		 * The security check is questionable...  We single
-		 * out the RO attribute for checking by the security
-		 * module, just because it maps to a file mode.
-		 */
-		err = security_inode_setattr(filp->f_path.dentry, &ia);
-		if (err)
-			goto up;
-
-		/* This MUST be done before doing anything irreversible... */
-		err = fat_setattr(filp->f_path.dentry, &ia);
-		if (err)
-			goto up;
-
-		fsnotify_change(filp->f_path.dentry, ia.ia_valid);
-		if (sbi->options.sys_immutable) {
-			if (attr & ATTR_SYS)
-				inode->i_flags |= S_IMMUTABLE;
-			else
-				inode->i_flags &= S_IMMUTABLE;
-		}
+	fat_save_attrs(inode, attr);
+	mark_inode_dirty(inode);
+out_drop_write:
+	mnt_drop_write(file->f_path.mnt);
+out_unlock_inode:
+	mutex_unlock(&inode->i_mutex);
+out:
+	return err;
+}
 
-		fat_save_attrs(inode, attr);
-		mark_inode_dirty(inode);
-up:
-		mnt_drop_write(filp->f_path.mnt);
-up_no_drop_write:
-		mutex_unlock(&inode->i_mutex);
-		return err;
-	}
+int fat_generic_ioctl(struct inode *inode, struct file *filp,
+		      unsigned int cmd, unsigned long arg)
+{
+	u32 __user *user_attr = (u32 __user *)arg;
+
+	switch (cmd) {
+	case FAT_IOCTL_GET_ATTRIBUTES:
+		return fat_ioctl_get_attributes(inode, user_attr);
+	case FAT_IOCTL_SET_ATTRIBUTES:
+		return fat_ioctl_set_attributes(filp, user_attr);
 	default:
 		return -ENOTTY;	/* Inappropriate ioctl for device */
 	}
@@ -225,7 +231,7 @@ static int fat_free(struct inode *inode, int skip)
 			fatent_brelse(&fatent);
 			return 0;
 		} else if (ret == FAT_ENT_FREE) {
-			fat_fs_panic(sb,
+			fat_fs_error(sb,
 				     "%s: invalid cluster chain (i_pos %lld)",
 				     __func__, MSDOS_I(inode)->i_pos);
 			ret = -EIO;
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index 51a5ecf9000a..304b411cb8bc 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -76,7 +76,7 @@ static inline int __fat_get_block(struct inode *inode, sector_t iblock,
 		return 0;
 
 	if (iblock != MSDOS_I(inode)->mmu_private >> sb->s_blocksize_bits) {
-		fat_fs_panic(sb, "corrupted file size (i_pos %lld, %lld)",
+		fat_fs_error(sb, "corrupted file size (i_pos %lld, %lld)",
 			MSDOS_I(inode)->i_pos, MSDOS_I(inode)->mmu_private);
 		return -EIO;
 	}
@@ -856,6 +856,12 @@ static int fat_show_options(struct seq_file *m, struct vfsmount *mnt)
 		seq_puts(m, ",flush");
 	if (opts->tz_utc)
 		seq_puts(m, ",tz=UTC");
+	if (opts->errors == FAT_ERRORS_CONT)
+		seq_puts(m, ",errors=continue");
+	else if (opts->errors == FAT_ERRORS_PANIC)
+		seq_puts(m, ",errors=panic");
+	else
+		seq_puts(m, ",errors=remount-ro");
 
 	return 0;
 }
@@ -868,7 +874,8 @@ enum {
 	Opt_charset, Opt_shortname_lower, Opt_shortname_win95,
 	Opt_shortname_winnt, Opt_shortname_mixed, Opt_utf8_no, Opt_utf8_yes,
 	Opt_uni_xl_no, Opt_uni_xl_yes, Opt_nonumtail_no, Opt_nonumtail_yes,
-	Opt_obsolate, Opt_flush, Opt_tz_utc, Opt_rodir, Opt_err,
+	Opt_obsolate, Opt_flush, Opt_tz_utc, Opt_rodir, Opt_err_cont,
+	Opt_err_panic, Opt_err_ro, Opt_err,
 };
 
 static const match_table_t fat_tokens = {
@@ -891,6 +898,11 @@ static const match_table_t fat_tokens = {
 	{Opt_showexec, "showexec"},
 	{Opt_debug, "debug"},
 	{Opt_immutable, "sys_immutable"},
+	{Opt_flush, "flush"},
+	{Opt_tz_utc, "tz=UTC"},
+	{Opt_err_cont, "errors=continue"},
+	{Opt_err_panic, "errors=panic"},
+	{Opt_err_ro, "errors=remount-ro"},
 	{Opt_obsolate, "conv=binary"},
 	{Opt_obsolate, "conv=text"},
 	{Opt_obsolate, "conv=auto"},
@@ -902,8 +914,6 @@ static const match_table_t fat_tokens = {
 	{Opt_obsolate, "cvf_format=%20s"},
 	{Opt_obsolate, "cvf_options=%100s"},
 	{Opt_obsolate, "posix"},
-	{Opt_flush, "flush"},
-	{Opt_tz_utc, "tz=UTC"},
 	{Opt_err, NULL},
 };
 static const match_table_t msdos_tokens = {
@@ -973,6 +983,7 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug,
 	opts->numtail = 1;
 	opts->usefree = opts->nocase = 0;
 	opts->tz_utc = 0;
+	opts->errors = FAT_ERRORS_RO;
 	*debug = 0;
 
 	if (!options)
@@ -1065,6 +1076,15 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug,
 		case Opt_tz_utc:
 			opts->tz_utc = 1;
 			break;
+		case Opt_err_cont:
+			opts->errors = FAT_ERRORS_CONT;
+			break;
+		case Opt_err_panic:
+			opts->errors = FAT_ERRORS_PANIC;
+			break;
+		case Opt_err_ro:
+			opts->errors = FAT_ERRORS_RO;
+			break;
 
 		/* msdos specific */
 		case Opt_dots:
diff --git a/fs/fat/misc.c b/fs/fat/misc.c
index ac39ebcc1496..a6c20473dfd7 100644
--- a/fs/fat/misc.c
+++ b/fs/fat/misc.c
@@ -12,14 +12,19 @@
 #include "fat.h"
 
 /*
- * fat_fs_panic reports a severe file system problem and sets the file system
- * read-only. The file system can be made writable again by remounting it.
+ * fat_fs_error reports a file system problem that might indicate fa data
+ * corruption/inconsistency. Depending on 'errors' mount option the
+ * panic() is called, or error message is printed FAT and nothing is done,
+ * or filesystem is remounted read-only (default behavior).
+ * In case the file system is remounted read-only, it can be made writable
+ * again by remounting it.
  */
-void fat_fs_panic(struct super_block *s, const char *fmt, ...)
+void fat_fs_error(struct super_block *s, const char *fmt, ...)
 {
+	struct fat_mount_options *opts = &MSDOS_SB(s)->options;
 	va_list args;
 
-	printk(KERN_ERR "FAT: Filesystem panic (dev %s)\n", s->s_id);
+	printk(KERN_ERR "FAT: Filesystem error (dev %s)\n", s->s_id);
 
 	printk(KERN_ERR "    ");
 	va_start(args, fmt);
@@ -27,13 +32,14 @@ void fat_fs_panic(struct super_block *s, const char *fmt, ...)
 	va_end(args);
 	printk("\n");
 
-	if (!(s->s_flags & MS_RDONLY)) {
+	if (opts->errors == FAT_ERRORS_PANIC)
+		panic("    FAT fs panic from previous error\n");
+	else if (opts->errors == FAT_ERRORS_RO && !(s->s_flags & MS_RDONLY)) {
 		s->s_flags |= MS_RDONLY;
 		printk(KERN_ERR "    File system has been set read-only\n");
 	}
 }
-
-EXPORT_SYMBOL_GPL(fat_fs_panic);
+EXPORT_SYMBOL_GPL(fat_fs_error);
 
 /* Flushes the number of free clusters on FAT32 */
 /* XXX: Need to write one per FSINFO block.  Currently only writes 1 */
@@ -124,7 +130,7 @@ int fat_chain_add(struct inode *inode, int new_dclus, int nr_cluster)
 			mark_inode_dirty(inode);
 	}
 	if (new_fclus != (inode->i_blocks >> (sbi->cluster_bits - 9))) {
-		fat_fs_panic(sb, "clusters badly computed (%d != %llu)",
+		fat_fs_error(sb, "clusters badly computed (%d != %llu)",
 			     new_fclus,
 			     (llu)(inode->i_blocks >> (sbi->cluster_bits - 9)));
 		fat_cache_inval_inode(inode);
diff --git a/fs/fat/namei_msdos.c b/fs/fat/namei_msdos.c
index 20f522861355..82f88733b681 100644
--- a/fs/fat/namei_msdos.c
+++ b/fs/fat/namei_msdos.c
@@ -608,7 +608,7 @@ error_inode:
 		sinfo.bh = NULL;
 	}
 	if (corrupt < 0) {
-		fat_fs_panic(new_dir->i_sb,
+		fat_fs_error(new_dir->i_sb,
 			     "%s: Filesystem corrupted (i_pos %lld)",
 			     __func__, sinfo.i_pos);
 	}
diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c
index b50ecbe97f83..8d6fdcfd41df 100644
--- a/fs/fat/namei_vfat.c
+++ b/fs/fat/namei_vfat.c
@@ -1030,7 +1030,7 @@ error_inode:
 		sinfo.bh = NULL;
 	}
 	if (corrupt < 0) {
-		fat_fs_panic(new_dir->i_sb,
+		fat_fs_error(new_dir->i_sb,
 			     "%s: Filesystem corrupted (i_pos %lld)",
 			     __func__, sinfo.i_pos);
 	}
diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c
index 064279e33bbb..36df60b6d8a4 100644
--- a/fs/nilfs2/bmap.c
+++ b/fs/nilfs2/bmap.c
@@ -31,21 +31,26 @@
 #include "dat.h"
 #include "alloc.h"
 
+struct inode *nilfs_bmap_get_dat(const struct nilfs_bmap *bmap)
+{
+	return nilfs_dat_inode(NILFS_I_NILFS(bmap->b_inode));
+}
+
 int nilfs_bmap_lookup_at_level(struct nilfs_bmap *bmap, __u64 key, int level,
 			       __u64 *ptrp)
 {
-	__u64 ptr;
+	sector_t blocknr;
 	int ret;
 
 	down_read(&bmap->b_sem);
 	ret = bmap->b_ops->bop_lookup(bmap, key, level, ptrp);
 	if (ret < 0)
 		goto out;
-	if (bmap->b_pops->bpop_translate != NULL) {
-		ret = bmap->b_pops->bpop_translate(bmap, *ptrp, &ptr);
-		if (ret < 0)
-			goto out;
-		*ptrp = ptr;
+	if (NILFS_BMAP_USE_VBN(bmap)) {
+		ret = nilfs_dat_translate(nilfs_bmap_get_dat(bmap), *ptrp,
+					  &blocknr);
+		if (!ret)
+			*ptrp = blocknr;
 	}
 
  out:
@@ -53,6 +58,16 @@ int nilfs_bmap_lookup_at_level(struct nilfs_bmap *bmap, __u64 key, int level,
 	return ret;
 }
 
+int nilfs_bmap_lookup_contig(struct nilfs_bmap *bmap, __u64 key, __u64 *ptrp,
+			     unsigned maxblocks)
+{
+	int ret;
+
+	down_read(&bmap->b_sem);
+	ret = bmap->b_ops->bop_lookup_contig(bmap, key, ptrp, maxblocks);
+	up_read(&bmap->b_sem);
+	return ret;
+}
 
 /**
  * nilfs_bmap_lookup - find a record
@@ -101,8 +116,7 @@ static int nilfs_bmap_do_insert(struct nilfs_bmap *bmap, __u64 key, __u64 ptr)
 			if (n < 0)
 				return n;
 			ret = nilfs_btree_convert_and_insert(
-				bmap, key, ptr, keys, ptrs, n,
-				NILFS_BMAP_LARGE_LOW, NILFS_BMAP_LARGE_HIGH);
+				bmap, key, ptr, keys, ptrs, n);
 			if (ret == 0)
 				bmap->b_u.u_flags |= NILFS_BMAP_LARGE;
 
@@ -158,8 +172,7 @@ static int nilfs_bmap_do_delete(struct nilfs_bmap *bmap, __u64 key)
 			if (n < 0)
 				return n;
 			ret = nilfs_direct_delete_and_convert(
-				bmap, key, keys, ptrs, n,
-				NILFS_BMAP_SMALL_LOW, NILFS_BMAP_SMALL_HIGH);
+				bmap, key, keys, ptrs, n);
 			if (ret == 0)
 				bmap->b_u.u_flags &= ~NILFS_BMAP_LARGE;
 
@@ -417,38 +430,6 @@ void nilfs_bmap_sub_blocks(const struct nilfs_bmap *bmap, int n)
 		mark_inode_dirty(bmap->b_inode);
 }
 
-int nilfs_bmap_get_block(const struct nilfs_bmap *bmap, __u64 ptr,
-			 struct buffer_head **bhp)
-{
-	return nilfs_btnode_get(&NILFS_BMAP_I(bmap)->i_btnode_cache,
-				ptr, 0, bhp, 0);
-}
-
-void nilfs_bmap_put_block(const struct nilfs_bmap *bmap,
-			  struct buffer_head *bh)
-{
-	brelse(bh);
-}
-
-int nilfs_bmap_get_new_block(const struct nilfs_bmap *bmap, __u64 ptr,
-			     struct buffer_head **bhp)
-{
-	int ret;
-
-	ret = nilfs_btnode_get(&NILFS_BMAP_I(bmap)->i_btnode_cache,
-			       ptr, 0, bhp, 1);
-	if (ret < 0)
-		return ret;
-	set_buffer_nilfs_volatile(*bhp);
-	return 0;
-}
-
-void nilfs_bmap_delete_block(const struct nilfs_bmap *bmap,
-			     struct buffer_head *bh)
-{
-	nilfs_btnode_delete(bh);
-}
-
 __u64 nilfs_bmap_data_get_key(const struct nilfs_bmap *bmap,
 			      const struct buffer_head *bh)
 {
@@ -476,11 +457,6 @@ __u64 nilfs_bmap_find_target_seq(const struct nilfs_bmap *bmap, __u64 key)
 		return NILFS_BMAP_INVALID_PTR;
 }
 
-static struct inode *nilfs_bmap_get_dat(const struct nilfs_bmap *bmap)
-{
-	return nilfs_dat_inode(NILFS_I_NILFS(bmap->b_inode));
-}
-
 #define NILFS_BMAP_GROUP_DIV	8
 __u64 nilfs_bmap_find_target_in_group(const struct nilfs_bmap *bmap)
 {
@@ -493,64 +469,51 @@ __u64 nilfs_bmap_find_target_in_group(const struct nilfs_bmap *bmap)
 		(entries_per_group / NILFS_BMAP_GROUP_DIV);
 }
 
-static int nilfs_bmap_prepare_alloc_v(struct nilfs_bmap *bmap,
-				      union nilfs_bmap_ptr_req *req)
+int nilfs_bmap_prepare_alloc_v(struct nilfs_bmap *bmap,
+				 union nilfs_bmap_ptr_req *req)
 {
 	return nilfs_dat_prepare_alloc(nilfs_bmap_get_dat(bmap), &req->bpr_req);
 }
 
-static void nilfs_bmap_commit_alloc_v(struct nilfs_bmap *bmap,
-				      union nilfs_bmap_ptr_req *req)
+void nilfs_bmap_commit_alloc_v(struct nilfs_bmap *bmap,
+				 union nilfs_bmap_ptr_req *req)
 {
 	nilfs_dat_commit_alloc(nilfs_bmap_get_dat(bmap), &req->bpr_req);
 }
 
-static void nilfs_bmap_abort_alloc_v(struct nilfs_bmap *bmap,
-				     union nilfs_bmap_ptr_req *req)
+void nilfs_bmap_abort_alloc_v(struct nilfs_bmap *bmap,
+			      union nilfs_bmap_ptr_req *req)
 {
 	nilfs_dat_abort_alloc(nilfs_bmap_get_dat(bmap), &req->bpr_req);
 }
 
-static int nilfs_bmap_prepare_start_v(struct nilfs_bmap *bmap,
-				      union nilfs_bmap_ptr_req *req)
+int nilfs_bmap_start_v(struct nilfs_bmap *bmap, union nilfs_bmap_ptr_req *req,
+		       sector_t blocknr)
 {
-	return nilfs_dat_prepare_start(nilfs_bmap_get_dat(bmap), &req->bpr_req);
-}
-
-static void nilfs_bmap_commit_start_v(struct nilfs_bmap *bmap,
-				      union nilfs_bmap_ptr_req *req,
-				      sector_t blocknr)
-{
-	nilfs_dat_commit_start(nilfs_bmap_get_dat(bmap), &req->bpr_req,
-			       blocknr);
-}
+	struct inode *dat = nilfs_bmap_get_dat(bmap);
+	int ret;
 
-static void nilfs_bmap_abort_start_v(struct nilfs_bmap *bmap,
-				     union nilfs_bmap_ptr_req *req)
-{
-	nilfs_dat_abort_start(nilfs_bmap_get_dat(bmap), &req->bpr_req);
+	ret = nilfs_dat_prepare_start(dat, &req->bpr_req);
+	if (likely(!ret))
+		nilfs_dat_commit_start(dat, &req->bpr_req, blocknr);
+	return ret;
 }
 
-static int nilfs_bmap_prepare_end_v(struct nilfs_bmap *bmap,
-				    union nilfs_bmap_ptr_req *req)
+int nilfs_bmap_prepare_end_v(struct nilfs_bmap *bmap,
+			     union nilfs_bmap_ptr_req *req)
 {
 	return nilfs_dat_prepare_end(nilfs_bmap_get_dat(bmap), &req->bpr_req);
 }
 
-static void nilfs_bmap_commit_end_v(struct nilfs_bmap *bmap,
-				    union nilfs_bmap_ptr_req *req)
-{
-	nilfs_dat_commit_end(nilfs_bmap_get_dat(bmap), &req->bpr_req, 0);
-}
-
-static void nilfs_bmap_commit_end_vmdt(struct nilfs_bmap *bmap,
-				       union nilfs_bmap_ptr_req *req)
+void nilfs_bmap_commit_end_v(struct nilfs_bmap *bmap,
+			     union nilfs_bmap_ptr_req *req)
 {
-	nilfs_dat_commit_end(nilfs_bmap_get_dat(bmap), &req->bpr_req, 1);
+	nilfs_dat_commit_end(nilfs_bmap_get_dat(bmap), &req->bpr_req,
+			     bmap->b_ptr_type == NILFS_BMAP_PTR_VS);
 }
 
-static void nilfs_bmap_abort_end_v(struct nilfs_bmap *bmap,
-				   union nilfs_bmap_ptr_req *req)
+void nilfs_bmap_abort_end_v(struct nilfs_bmap *bmap,
+			    union nilfs_bmap_ptr_req *req)
 {
 	nilfs_dat_abort_end(nilfs_bmap_get_dat(bmap), &req->bpr_req);
 }
@@ -566,128 +529,44 @@ int nilfs_bmap_mark_dirty(const struct nilfs_bmap *bmap, __u64 vblocknr)
 	return nilfs_dat_mark_dirty(nilfs_bmap_get_dat(bmap), vblocknr);
 }
 
-int nilfs_bmap_prepare_update(struct nilfs_bmap *bmap,
-			      union nilfs_bmap_ptr_req *oldreq,
-			      union nilfs_bmap_ptr_req *newreq)
+int nilfs_bmap_prepare_update_v(struct nilfs_bmap *bmap,
+				union nilfs_bmap_ptr_req *oldreq,
+				union nilfs_bmap_ptr_req *newreq)
 {
+	struct inode *dat = nilfs_bmap_get_dat(bmap);
 	int ret;
 
-	ret = bmap->b_pops->bpop_prepare_end_ptr(bmap, oldreq);
+	ret = nilfs_dat_prepare_end(dat, &oldreq->bpr_req);
 	if (ret < 0)
 		return ret;
-	ret = bmap->b_pops->bpop_prepare_alloc_ptr(bmap, newreq);
+	ret = nilfs_dat_prepare_alloc(dat, &newreq->bpr_req);
 	if (ret < 0)
-		bmap->b_pops->bpop_abort_end_ptr(bmap, oldreq);
+		nilfs_dat_abort_end(dat, &oldreq->bpr_req);
 
 	return ret;
 }
 
-void nilfs_bmap_commit_update(struct nilfs_bmap *bmap,
-			      union nilfs_bmap_ptr_req *oldreq,
-			      union nilfs_bmap_ptr_req *newreq)
+void nilfs_bmap_commit_update_v(struct nilfs_bmap *bmap,
+				union nilfs_bmap_ptr_req *oldreq,
+				union nilfs_bmap_ptr_req *newreq)
 {
-	bmap->b_pops->bpop_commit_end_ptr(bmap, oldreq);
-	bmap->b_pops->bpop_commit_alloc_ptr(bmap, newreq);
-}
+	struct inode *dat = nilfs_bmap_get_dat(bmap);
 
-void nilfs_bmap_abort_update(struct nilfs_bmap *bmap,
-			     union nilfs_bmap_ptr_req *oldreq,
-			     union nilfs_bmap_ptr_req *newreq)
-{
-	bmap->b_pops->bpop_abort_end_ptr(bmap, oldreq);
-	bmap->b_pops->bpop_abort_alloc_ptr(bmap, newreq);
+	nilfs_dat_commit_end(dat, &oldreq->bpr_req,
+			     bmap->b_ptr_type == NILFS_BMAP_PTR_VS);
+	nilfs_dat_commit_alloc(dat, &newreq->bpr_req);
 }
 
-static int nilfs_bmap_translate_v(const struct nilfs_bmap *bmap, __u64 ptr,
-				  __u64 *ptrp)
+void nilfs_bmap_abort_update_v(struct nilfs_bmap *bmap,
+			       union nilfs_bmap_ptr_req *oldreq,
+			       union nilfs_bmap_ptr_req *newreq)
 {
-	sector_t blocknr;
-	int ret;
-
-	ret = nilfs_dat_translate(nilfs_bmap_get_dat(bmap), ptr, &blocknr);
-	if (ret < 0)
-		return ret;
-	if (ptrp != NULL)
-		*ptrp = blocknr;
-	return 0;
-}
+	struct inode *dat = nilfs_bmap_get_dat(bmap);
 
-static int nilfs_bmap_prepare_alloc_p(struct nilfs_bmap *bmap,
-				      union nilfs_bmap_ptr_req *req)
-{
-	/* ignore target ptr */
-	req->bpr_ptr = bmap->b_last_allocated_ptr++;
-	return 0;
+	nilfs_dat_abort_end(dat, &oldreq->bpr_req);
+	nilfs_dat_abort_alloc(dat, &newreq->bpr_req);
 }
 
-static void nilfs_bmap_commit_alloc_p(struct nilfs_bmap *bmap,
-				      union nilfs_bmap_ptr_req *req)
-{
-	/* do nothing */
-}
-
-static void nilfs_bmap_abort_alloc_p(struct nilfs_bmap *bmap,
-				     union nilfs_bmap_ptr_req *req)
-{
-	bmap->b_last_allocated_ptr--;
-}
-
-static const struct nilfs_bmap_ptr_operations nilfs_bmap_ptr_ops_v = {
-	.bpop_prepare_alloc_ptr	=	nilfs_bmap_prepare_alloc_v,
-	.bpop_commit_alloc_ptr	=	nilfs_bmap_commit_alloc_v,
-	.bpop_abort_alloc_ptr	=	nilfs_bmap_abort_alloc_v,
-	.bpop_prepare_start_ptr	=	nilfs_bmap_prepare_start_v,
-	.bpop_commit_start_ptr	=	nilfs_bmap_commit_start_v,
-	.bpop_abort_start_ptr	=	nilfs_bmap_abort_start_v,
-	.bpop_prepare_end_ptr	=	nilfs_bmap_prepare_end_v,
-	.bpop_commit_end_ptr	=	nilfs_bmap_commit_end_v,
-	.bpop_abort_end_ptr	=	nilfs_bmap_abort_end_v,
-
-	.bpop_translate		=	nilfs_bmap_translate_v,
-};
-
-static const struct nilfs_bmap_ptr_operations nilfs_bmap_ptr_ops_vmdt = {
-	.bpop_prepare_alloc_ptr	=	nilfs_bmap_prepare_alloc_v,
-	.bpop_commit_alloc_ptr	=	nilfs_bmap_commit_alloc_v,
-	.bpop_abort_alloc_ptr	=	nilfs_bmap_abort_alloc_v,
-	.bpop_prepare_start_ptr	=	nilfs_bmap_prepare_start_v,
-	.bpop_commit_start_ptr	=	nilfs_bmap_commit_start_v,
-	.bpop_abort_start_ptr	=	nilfs_bmap_abort_start_v,
-	.bpop_prepare_end_ptr	=	nilfs_bmap_prepare_end_v,
-	.bpop_commit_end_ptr	=	nilfs_bmap_commit_end_vmdt,
-	.bpop_abort_end_ptr	=	nilfs_bmap_abort_end_v,
-
-	.bpop_translate		=	nilfs_bmap_translate_v,
-};
-
-static const struct nilfs_bmap_ptr_operations nilfs_bmap_ptr_ops_p = {
-	.bpop_prepare_alloc_ptr	=	nilfs_bmap_prepare_alloc_p,
-	.bpop_commit_alloc_ptr	=	nilfs_bmap_commit_alloc_p,
-	.bpop_abort_alloc_ptr	=	nilfs_bmap_abort_alloc_p,
-	.bpop_prepare_start_ptr	=	NULL,
-	.bpop_commit_start_ptr	=	NULL,
-	.bpop_abort_start_ptr	=	NULL,
-	.bpop_prepare_end_ptr	=	NULL,
-	.bpop_commit_end_ptr	=	NULL,
-	.bpop_abort_end_ptr	=	NULL,
-
-	.bpop_translate		=	NULL,
-};
-
-static const struct nilfs_bmap_ptr_operations nilfs_bmap_ptr_ops_gc = {
-	.bpop_prepare_alloc_ptr	=	NULL,
-	.bpop_commit_alloc_ptr	=	NULL,
-	.bpop_abort_alloc_ptr	=	NULL,
-	.bpop_prepare_start_ptr	=	NULL,
-	.bpop_commit_start_ptr	=	NULL,
-	.bpop_abort_start_ptr	=	NULL,
-	.bpop_prepare_end_ptr	=	NULL,
-	.bpop_commit_end_ptr	=	NULL,
-	.bpop_abort_end_ptr	=	NULL,
-
-	.bpop_translate		=	NULL,
-};
-
 static struct lock_class_key nilfs_bmap_dat_lock_key;
 
 /**
@@ -714,31 +593,26 @@ int nilfs_bmap_read(struct nilfs_bmap *bmap, struct nilfs_inode *raw_inode)
 	bmap->b_inode = &NILFS_BMAP_I(bmap)->vfs_inode;
 	switch (bmap->b_inode->i_ino) {
 	case NILFS_DAT_INO:
-		bmap->b_pops = &nilfs_bmap_ptr_ops_p;
-		bmap->b_last_allocated_key = 0;	/* XXX: use macro */
+		bmap->b_ptr_type = NILFS_BMAP_PTR_P;
+		bmap->b_last_allocated_key = 0;
 		bmap->b_last_allocated_ptr = NILFS_BMAP_NEW_PTR_INIT;
 		lockdep_set_class(&bmap->b_sem, &nilfs_bmap_dat_lock_key);
 		break;
 	case NILFS_CPFILE_INO:
 	case NILFS_SUFILE_INO:
-		bmap->b_pops = &nilfs_bmap_ptr_ops_vmdt;
-		bmap->b_last_allocated_key = 0;	/* XXX: use macro */
+		bmap->b_ptr_type = NILFS_BMAP_PTR_VS;
+		bmap->b_last_allocated_key = 0;
 		bmap->b_last_allocated_ptr = NILFS_BMAP_INVALID_PTR;
 		break;
 	default:
-		bmap->b_pops = &nilfs_bmap_ptr_ops_v;
-		bmap->b_last_allocated_key = 0;	/* XXX: use macro */
+		bmap->b_ptr_type = NILFS_BMAP_PTR_VM;
+		bmap->b_last_allocated_key = 0;
 		bmap->b_last_allocated_ptr = NILFS_BMAP_INVALID_PTR;
 		break;
 	}
 
 	return (bmap->b_u.u_flags & NILFS_BMAP_LARGE) ?
-		nilfs_btree_init(bmap,
-				 NILFS_BMAP_LARGE_LOW,
-				 NILFS_BMAP_LARGE_HIGH) :
-		nilfs_direct_init(bmap,
-				  NILFS_BMAP_SMALL_LOW,
-				  NILFS_BMAP_SMALL_HIGH);
+		nilfs_btree_init(bmap) : nilfs_direct_init(bmap);
 }
 
 /**
@@ -764,7 +638,7 @@ void nilfs_bmap_init_gc(struct nilfs_bmap *bmap)
 	memset(&bmap->b_u, 0, NILFS_BMAP_SIZE);
 	init_rwsem(&bmap->b_sem);
 	bmap->b_inode = &NILFS_BMAP_I(bmap)->vfs_inode;
-	bmap->b_pops = &nilfs_bmap_ptr_ops_gc;
+	bmap->b_ptr_type = NILFS_BMAP_PTR_U;
 	bmap->b_last_allocated_key = 0;
 	bmap->b_last_allocated_ptr = NILFS_BMAP_INVALID_PTR;
 	bmap->b_state = 0;
diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h
index 4f2708abb1ba..b2890cdcef12 100644
--- a/fs/nilfs2/bmap.h
+++ b/fs/nilfs2/bmap.h
@@ -64,6 +64,8 @@ struct nilfs_bmap_stats {
  */
 struct nilfs_bmap_operations {
 	int (*bop_lookup)(const struct nilfs_bmap *, __u64, int, __u64 *);
+	int (*bop_lookup_contig)(const struct nilfs_bmap *, __u64, __u64 *,
+				 unsigned);
 	int (*bop_insert)(struct nilfs_bmap *, __u64, __u64);
 	int (*bop_delete)(struct nilfs_bmap *, __u64);
 	void (*bop_clear)(struct nilfs_bmap *);
@@ -86,34 +88,6 @@ struct nilfs_bmap_operations {
 };
 
 
-/**
- * struct nilfs_bmap_ptr_operations - bmap ptr operation table
- */
-struct nilfs_bmap_ptr_operations {
-	int (*bpop_prepare_alloc_ptr)(struct nilfs_bmap *,
-				      union nilfs_bmap_ptr_req *);
-	void (*bpop_commit_alloc_ptr)(struct nilfs_bmap *,
-				      union nilfs_bmap_ptr_req *);
-	void (*bpop_abort_alloc_ptr)(struct nilfs_bmap *,
-				     union nilfs_bmap_ptr_req *);
-	int (*bpop_prepare_start_ptr)(struct nilfs_bmap *,
-				      union nilfs_bmap_ptr_req *);
-	void (*bpop_commit_start_ptr)(struct nilfs_bmap *,
-				      union nilfs_bmap_ptr_req *,
-				      sector_t);
-	void (*bpop_abort_start_ptr)(struct nilfs_bmap *,
-				     union nilfs_bmap_ptr_req *);
-	int (*bpop_prepare_end_ptr)(struct nilfs_bmap *,
-				    union nilfs_bmap_ptr_req *);
-	void (*bpop_commit_end_ptr)(struct nilfs_bmap *,
-				    union nilfs_bmap_ptr_req *);
-	void (*bpop_abort_end_ptr)(struct nilfs_bmap *,
-				   union nilfs_bmap_ptr_req *);
-
-	int (*bpop_translate)(const struct nilfs_bmap *, __u64, __u64 *);
-};
-
-
 #define NILFS_BMAP_SIZE		(NILFS_INODE_BMAP_SIZE * sizeof(__le64))
 #define NILFS_BMAP_KEY_BIT	(sizeof(unsigned long) * 8 /* CHAR_BIT */)
 #define NILFS_BMAP_NEW_PTR_INIT	\
@@ -131,11 +105,9 @@ static inline int nilfs_bmap_is_new_ptr(unsigned long ptr)
  * @b_sem: semaphore
  * @b_inode: owner of bmap
  * @b_ops: bmap operation table
- * @b_pops: bmap ptr operation table
- * @b_low: low watermark of conversion
- * @b_high: high watermark of conversion
  * @b_last_allocated_key: last allocated key for data block
  * @b_last_allocated_ptr: last allocated ptr for data block
+ * @b_ptr_type: pointer type
  * @b_state: state
  */
 struct nilfs_bmap {
@@ -146,14 +118,22 @@ struct nilfs_bmap {
 	struct rw_semaphore b_sem;
 	struct inode *b_inode;
 	const struct nilfs_bmap_operations *b_ops;
-	const struct nilfs_bmap_ptr_operations *b_pops;
-	__u64 b_low;
-	__u64 b_high;
 	__u64 b_last_allocated_key;
 	__u64 b_last_allocated_ptr;
+	int b_ptr_type;
 	int b_state;
 };
 
+/* pointer type */
+#define NILFS_BMAP_PTR_P	0	/* physical block number (i.e. LBN) */
+#define NILFS_BMAP_PTR_VS	1	/* virtual block number (single
+					   version) */
+#define NILFS_BMAP_PTR_VM	2	/* virtual block number (has multiple
+					   versions) */
+#define NILFS_BMAP_PTR_U	(-1)	/* never perform pointer operations */
+
+#define NILFS_BMAP_USE_VBN(bmap)	((bmap)->b_ptr_type > 0)
+
 /* state */
 #define NILFS_BMAP_DIRTY	0x00000001
 
@@ -162,6 +142,7 @@ int nilfs_bmap_test_and_clear_dirty(struct nilfs_bmap *);
 int nilfs_bmap_read(struct nilfs_bmap *, struct nilfs_inode *);
 void nilfs_bmap_write(struct nilfs_bmap *, struct nilfs_inode *);
 int nilfs_bmap_lookup(struct nilfs_bmap *, unsigned long, unsigned long *);
+int nilfs_bmap_lookup_contig(struct nilfs_bmap *, __u64, __u64 *, unsigned);
 int nilfs_bmap_insert(struct nilfs_bmap *, unsigned long, unsigned long);
 int nilfs_bmap_delete(struct nilfs_bmap *, unsigned long);
 int nilfs_bmap_last_key(struct nilfs_bmap *, unsigned long *);
@@ -182,7 +163,67 @@ void nilfs_bmap_commit_gcdat(struct nilfs_bmap *, struct nilfs_bmap *);
 /*
  * Internal use only
  */
+struct inode *nilfs_bmap_get_dat(const struct nilfs_bmap *);
+int nilfs_bmap_prepare_alloc_v(struct nilfs_bmap *,
+			       union nilfs_bmap_ptr_req *);
+void nilfs_bmap_commit_alloc_v(struct nilfs_bmap *,
+			       union nilfs_bmap_ptr_req *);
+void nilfs_bmap_abort_alloc_v(struct nilfs_bmap *,
+			      union nilfs_bmap_ptr_req *);
 
+static inline int nilfs_bmap_prepare_alloc_ptr(struct nilfs_bmap *bmap,
+					       union nilfs_bmap_ptr_req *req)
+{
+	if (NILFS_BMAP_USE_VBN(bmap))
+		return nilfs_bmap_prepare_alloc_v(bmap, req);
+	/* ignore target ptr */
+	req->bpr_ptr = bmap->b_last_allocated_ptr++;
+	return 0;
+}
+
+static inline void nilfs_bmap_commit_alloc_ptr(struct nilfs_bmap *bmap,
+					       union nilfs_bmap_ptr_req *req)
+{
+	if (NILFS_BMAP_USE_VBN(bmap))
+		nilfs_bmap_commit_alloc_v(bmap, req);
+}
+
+static inline void nilfs_bmap_abort_alloc_ptr(struct nilfs_bmap *bmap,
+					      union nilfs_bmap_ptr_req *req)
+{
+	if (NILFS_BMAP_USE_VBN(bmap))
+		nilfs_bmap_abort_alloc_v(bmap, req);
+	else
+		bmap->b_last_allocated_ptr--;
+}
+
+int nilfs_bmap_prepare_end_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *);
+void nilfs_bmap_commit_end_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *);
+void nilfs_bmap_abort_end_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *);
+
+static inline int nilfs_bmap_prepare_end_ptr(struct nilfs_bmap *bmap,
+					     union nilfs_bmap_ptr_req *req)
+{
+	return NILFS_BMAP_USE_VBN(bmap) ?
+		nilfs_bmap_prepare_end_v(bmap, req) : 0;
+}
+
+static inline void nilfs_bmap_commit_end_ptr(struct nilfs_bmap *bmap,
+					     union nilfs_bmap_ptr_req *req)
+{
+	if (NILFS_BMAP_USE_VBN(bmap))
+		nilfs_bmap_commit_end_v(bmap, req);
+}
+
+static inline void nilfs_bmap_abort_end_ptr(struct nilfs_bmap *bmap,
+					    union nilfs_bmap_ptr_req *req)
+{
+	if (NILFS_BMAP_USE_VBN(bmap))
+		nilfs_bmap_abort_end_v(bmap, req);
+}
+
+int nilfs_bmap_start_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *,
+		       sector_t);
 int nilfs_bmap_move_v(const struct nilfs_bmap *, __u64, sector_t);
 int nilfs_bmap_mark_dirty(const struct nilfs_bmap *, __u64);
 
@@ -193,28 +234,20 @@ __u64 nilfs_bmap_data_get_key(const struct nilfs_bmap *,
 __u64 nilfs_bmap_find_target_seq(const struct nilfs_bmap *, __u64);
 __u64 nilfs_bmap_find_target_in_group(const struct nilfs_bmap *);
 
-int nilfs_bmap_prepare_update(struct nilfs_bmap *,
-			      union nilfs_bmap_ptr_req *,
-			      union nilfs_bmap_ptr_req *);
-void nilfs_bmap_commit_update(struct nilfs_bmap *,
-			      union nilfs_bmap_ptr_req *,
-			      union nilfs_bmap_ptr_req *);
-void nilfs_bmap_abort_update(struct nilfs_bmap *,
-			     union nilfs_bmap_ptr_req *,
-			     union nilfs_bmap_ptr_req *);
+int nilfs_bmap_prepare_update_v(struct nilfs_bmap *,
+				union nilfs_bmap_ptr_req *,
+				union nilfs_bmap_ptr_req *);
+void nilfs_bmap_commit_update_v(struct nilfs_bmap *,
+				union nilfs_bmap_ptr_req *,
+				union nilfs_bmap_ptr_req *);
+void nilfs_bmap_abort_update_v(struct nilfs_bmap *,
+			       union nilfs_bmap_ptr_req *,
+			       union nilfs_bmap_ptr_req *);
 
 void nilfs_bmap_add_blocks(const struct nilfs_bmap *, int);
 void nilfs_bmap_sub_blocks(const struct nilfs_bmap *, int);
 
 
-int nilfs_bmap_get_block(const struct nilfs_bmap *, __u64,
-			 struct buffer_head **);
-void nilfs_bmap_put_block(const struct nilfs_bmap *, struct buffer_head *);
-int nilfs_bmap_get_new_block(const struct nilfs_bmap *, __u64,
-			     struct buffer_head **);
-void nilfs_bmap_delete_block(const struct nilfs_bmap *, struct buffer_head *);
-
-
 /* Assume that bmap semaphore is locked. */
 static inline int nilfs_bmap_dirty(const struct nilfs_bmap *bmap)
 {
diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c
index 4cc07b2c30e0..7e0b61be212e 100644
--- a/fs/nilfs2/btnode.c
+++ b/fs/nilfs2/btnode.c
@@ -46,15 +46,18 @@ void nilfs_btnode_cache_init_once(struct address_space *btnc)
 	INIT_LIST_HEAD(&btnc->i_mmap_nonlinear);
 }
 
-static struct address_space_operations def_btnode_aops;
+static struct address_space_operations def_btnode_aops = {
+	.sync_page		= block_sync_page,
+};
 
-void nilfs_btnode_cache_init(struct address_space *btnc)
+void nilfs_btnode_cache_init(struct address_space *btnc,
+			     struct backing_dev_info *bdi)
 {
 	btnc->host = NULL;  /* can safely set to host inode ? */
 	btnc->flags = 0;
 	mapping_set_gfp_mask(btnc, GFP_NOFS);
 	btnc->assoc_mapping = NULL;
-	btnc->backing_dev_info = &default_backing_dev_info;
+	btnc->backing_dev_info = bdi;
 	btnc->a_ops = &def_btnode_aops;
 }
 
diff --git a/fs/nilfs2/btnode.h b/fs/nilfs2/btnode.h
index 35faa86444a7..3e2275172ed6 100644
--- a/fs/nilfs2/btnode.h
+++ b/fs/nilfs2/btnode.h
@@ -38,7 +38,7 @@ struct nilfs_btnode_chkey_ctxt {
 };
 
 void nilfs_btnode_cache_init_once(struct address_space *);
-void nilfs_btnode_cache_init(struct address_space *);
+void nilfs_btnode_cache_init(struct address_space *, struct backing_dev_info *);
 void nilfs_btnode_cache_clear(struct address_space *);
 int nilfs_btnode_submit_block(struct address_space *, __u64, sector_t,
 			      struct buffer_head **, int);
diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c
index 6b37a2767293..aa412724b64e 100644
--- a/fs/nilfs2/btree.c
+++ b/fs/nilfs2/btree.c
@@ -29,6 +29,7 @@
 #include "btnode.h"
 #include "btree.h"
 #include "alloc.h"
+#include "dat.h"
 
 /**
  * struct nilfs_btree_path - A path on which B-tree operations are executed
@@ -109,8 +110,7 @@ static void nilfs_btree_clear_path(const struct nilfs_btree *btree,
 	     level < NILFS_BTREE_LEVEL_MAX;
 	     level++) {
 		if (path[level].bp_bh != NULL) {
-			nilfs_bmap_put_block(&btree->bt_bmap,
-					     path[level].bp_bh);
+			brelse(path[level].bp_bh);
 			path[level].bp_bh = NULL;
 		}
 		/* sib_bh is released or deleted by prepare or commit
@@ -123,10 +123,29 @@ static void nilfs_btree_clear_path(const struct nilfs_btree *btree,
 	}
 }
 
-
 /*
  * B-tree node operations
  */
+static int nilfs_btree_get_block(const struct nilfs_btree *btree, __u64 ptr,
+				 struct buffer_head **bhp)
+{
+	struct address_space *btnc =
+		&NILFS_BMAP_I((struct nilfs_bmap *)btree)->i_btnode_cache;
+	return nilfs_btnode_get(btnc, ptr, 0, bhp, 0);
+}
+
+static int nilfs_btree_get_new_block(const struct nilfs_btree *btree,
+				     __u64 ptr, struct buffer_head **bhp)
+{
+	struct address_space *btnc =
+		&NILFS_BMAP_I((struct nilfs_bmap *)btree)->i_btnode_cache;
+	int ret;
+
+	ret = nilfs_btnode_get(btnc, ptr, 0, bhp, 1);
+	if (!ret)
+		set_buffer_nilfs_volatile(*bhp);
+	return ret;
+}
 
 static inline int
 nilfs_btree_node_get_flags(const struct nilfs_btree *btree,
@@ -488,8 +507,7 @@ static int nilfs_btree_do_lookup(const struct nilfs_btree *btree,
 	path[level].bp_index = index;
 
 	for (level--; level >= minlevel; level--) {
-		ret = nilfs_bmap_get_block(&btree->bt_bmap, ptr,
-					   &path[level].bp_bh);
+		ret = nilfs_btree_get_block(btree, ptr, &path[level].bp_bh);
 		if (ret < 0)
 			return ret;
 		node = nilfs_btree_get_nonroot_node(btree, path, level);
@@ -535,8 +553,7 @@ static int nilfs_btree_do_lookup_last(const struct nilfs_btree *btree,
 	path[level].bp_index = index;
 
 	for (level--; level > 0; level--) {
-		ret = nilfs_bmap_get_block(&btree->bt_bmap, ptr,
-					   &path[level].bp_bh);
+		ret = nilfs_btree_get_block(btree, ptr, &path[level].bp_bh);
 		if (ret < 0)
 			return ret;
 		node = nilfs_btree_get_nonroot_node(btree, path, level);
@@ -579,6 +596,87 @@ static int nilfs_btree_lookup(const struct nilfs_bmap *bmap,
 	return ret;
 }
 
+static int nilfs_btree_lookup_contig(const struct nilfs_bmap *bmap,
+				     __u64 key, __u64 *ptrp, unsigned maxblocks)
+{
+	struct nilfs_btree *btree = (struct nilfs_btree *)bmap;
+	struct nilfs_btree_path *path;
+	struct nilfs_btree_node *node;
+	struct inode *dat = NULL;
+	__u64 ptr, ptr2;
+	sector_t blocknr;
+	int level = NILFS_BTREE_LEVEL_NODE_MIN;
+	int ret, cnt, index, maxlevel;
+
+	path = nilfs_btree_alloc_path(btree);
+	if (path == NULL)
+		return -ENOMEM;
+	nilfs_btree_init_path(btree, path);
+	ret = nilfs_btree_do_lookup(btree, path, key, &ptr, level);
+	if (ret < 0)
+		goto out;
+
+	if (NILFS_BMAP_USE_VBN(bmap)) {
+		dat = nilfs_bmap_get_dat(bmap);
+		ret = nilfs_dat_translate(dat, ptr, &blocknr);
+		if (ret < 0)
+			goto out;
+		ptr = blocknr;
+	}
+	cnt = 1;
+	if (cnt == maxblocks)
+		goto end;
+
+	maxlevel = nilfs_btree_height(btree) - 1;
+	node = nilfs_btree_get_node(btree, path, level);
+	index = path[level].bp_index + 1;
+	for (;;) {
+		while (index < nilfs_btree_node_get_nchildren(btree, node)) {
+			if (nilfs_btree_node_get_key(btree, node, index) !=
+			    key + cnt)
+				goto end;
+			ptr2 = nilfs_btree_node_get_ptr(btree, node, index);
+			if (dat) {
+				ret = nilfs_dat_translate(dat, ptr2, &blocknr);
+				if (ret < 0)
+					goto out;
+				ptr2 = blocknr;
+			}
+			if (ptr2 != ptr + cnt || ++cnt == maxblocks)
+				goto end;
+			index++;
+			continue;
+		}
+		if (level == maxlevel)
+			break;
+
+		/* look-up right sibling node */
+		node = nilfs_btree_get_node(btree, path, level + 1);
+		index = path[level + 1].bp_index + 1;
+		if (index >= nilfs_btree_node_get_nchildren(btree, node) ||
+		    nilfs_btree_node_get_key(btree, node, index) != key + cnt)
+			break;
+		ptr2 = nilfs_btree_node_get_ptr(btree, node, index);
+		path[level + 1].bp_index = index;
+
+		brelse(path[level].bp_bh);
+		path[level].bp_bh = NULL;
+		ret = nilfs_btree_get_block(btree, ptr2, &path[level].bp_bh);
+		if (ret < 0)
+			goto out;
+		node = nilfs_btree_get_nonroot_node(btree, path, level);
+		index = 0;
+		path[level].bp_index = index;
+	}
+ end:
+	*ptrp = ptr;
+	ret = cnt;
+ out:
+	nilfs_btree_clear_path(btree, path);
+	nilfs_btree_free_path(btree, path);
+	return ret;
+}
+
 static void nilfs_btree_promote_key(struct nilfs_btree *btree,
 				    struct nilfs_btree_path *path,
 				    int level, __u64 key)
@@ -669,13 +767,13 @@ static void nilfs_btree_carry_left(struct nilfs_btree *btree,
 				nilfs_btree_node_get_key(btree, node, 0));
 
 	if (move) {
-		nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_bh);
+		brelse(path[level].bp_bh);
 		path[level].bp_bh = path[level].bp_sib_bh;
 		path[level].bp_sib_bh = NULL;
 		path[level].bp_index += lnchildren;
 		path[level + 1].bp_index--;
 	} else {
-		nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_sib_bh);
+		brelse(path[level].bp_sib_bh);
 		path[level].bp_sib_bh = NULL;
 		path[level].bp_index -= n;
 	}
@@ -722,14 +820,14 @@ static void nilfs_btree_carry_right(struct nilfs_btree *btree,
 	path[level + 1].bp_index--;
 
 	if (move) {
-		nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_bh);
+		brelse(path[level].bp_bh);
 		path[level].bp_bh = path[level].bp_sib_bh;
 		path[level].bp_sib_bh = NULL;
 		path[level].bp_index -=
 			nilfs_btree_node_get_nchildren(btree, node);
 		path[level + 1].bp_index++;
 	} else {
-		nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_sib_bh);
+		brelse(path[level].bp_sib_bh);
 		path[level].bp_sib_bh = NULL;
 	}
 
@@ -781,7 +879,7 @@ static void nilfs_btree_split(struct nilfs_btree *btree,
 		*keyp = nilfs_btree_node_get_key(btree, right, 0);
 		*ptrp = path[level].bp_newreq.bpr_ptr;
 
-		nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_bh);
+		brelse(path[level].bp_bh);
 		path[level].bp_bh = path[level].bp_sib_bh;
 		path[level].bp_sib_bh = NULL;
 	} else {
@@ -790,7 +888,7 @@ static void nilfs_btree_split(struct nilfs_btree *btree,
 		*keyp = nilfs_btree_node_get_key(btree, right, 0);
 		*ptrp = path[level].bp_newreq.bpr_ptr;
 
-		nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_sib_bh);
+		brelse(path[level].bp_sib_bh);
 		path[level].bp_sib_bh = NULL;
 	}
 
@@ -897,12 +995,12 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 	level = NILFS_BTREE_LEVEL_DATA;
 
 	/* allocate a new ptr for data block */
-	if (btree->bt_ops->btop_find_target != NULL)
+	if (NILFS_BMAP_USE_VBN(&btree->bt_bmap))
 		path[level].bp_newreq.bpr_ptr =
-			btree->bt_ops->btop_find_target(btree, path, key);
+			nilfs_btree_find_target_v(btree, path, key);
 
-	ret = btree->bt_bmap.b_pops->bpop_prepare_alloc_ptr(
-		&btree->bt_bmap, &path[level].bp_newreq);
+	ret = nilfs_bmap_prepare_alloc_ptr(&btree->bt_bmap,
+					   &path[level].bp_newreq);
 	if (ret < 0)
 		goto err_out_data;
 
@@ -924,8 +1022,7 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 		if (pindex > 0) {
 			sibptr = nilfs_btree_node_get_ptr(btree, parent,
 							  pindex - 1);
-			ret = nilfs_bmap_get_block(&btree->bt_bmap, sibptr,
-						   &bh);
+			ret = nilfs_btree_get_block(btree, sibptr, &bh);
 			if (ret < 0)
 				goto err_out_child_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
@@ -936,7 +1033,7 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 				stats->bs_nblocks++;
 				goto out;
 			} else
-				nilfs_bmap_put_block(&btree->bt_bmap, bh);
+				brelse(bh);
 		}
 
 		/* right sibling */
@@ -944,8 +1041,7 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 		    nilfs_btree_node_get_nchildren(btree, parent) - 1) {
 			sibptr = nilfs_btree_node_get_ptr(btree, parent,
 							  pindex + 1);
-			ret = nilfs_bmap_get_block(&btree->bt_bmap, sibptr,
-						   &bh);
+			ret = nilfs_btree_get_block(btree, sibptr, &bh);
 			if (ret < 0)
 				goto err_out_child_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
@@ -956,19 +1052,19 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 				stats->bs_nblocks++;
 				goto out;
 			} else
-				nilfs_bmap_put_block(&btree->bt_bmap, bh);
+				brelse(bh);
 		}
 
 		/* split */
 		path[level].bp_newreq.bpr_ptr =
 			path[level - 1].bp_newreq.bpr_ptr + 1;
-		ret = btree->bt_bmap.b_pops->bpop_prepare_alloc_ptr(
-			&btree->bt_bmap, &path[level].bp_newreq);
+		ret = nilfs_bmap_prepare_alloc_ptr(&btree->bt_bmap,
+						   &path[level].bp_newreq);
 		if (ret < 0)
 			goto err_out_child_node;
-		ret = nilfs_bmap_get_new_block(&btree->bt_bmap,
-					       path[level].bp_newreq.bpr_ptr,
-					       &bh);
+		ret = nilfs_btree_get_new_block(btree,
+						path[level].bp_newreq.bpr_ptr,
+						&bh);
 		if (ret < 0)
 			goto err_out_curr_node;
 
@@ -994,12 +1090,12 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 
 	/* grow */
 	path[level].bp_newreq.bpr_ptr = path[level - 1].bp_newreq.bpr_ptr + 1;
-	ret = btree->bt_bmap.b_pops->bpop_prepare_alloc_ptr(
-		&btree->bt_bmap, &path[level].bp_newreq);
+	ret = nilfs_bmap_prepare_alloc_ptr(&btree->bt_bmap,
+					   &path[level].bp_newreq);
 	if (ret < 0)
 		goto err_out_child_node;
-	ret = nilfs_bmap_get_new_block(&btree->bt_bmap,
-				       path[level].bp_newreq.bpr_ptr, &bh);
+	ret = nilfs_btree_get_new_block(btree, path[level].bp_newreq.bpr_ptr,
+					&bh);
 	if (ret < 0)
 		goto err_out_curr_node;
 
@@ -1023,18 +1119,16 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 
 	/* error */
  err_out_curr_node:
-	btree->bt_bmap.b_pops->bpop_abort_alloc_ptr(&btree->bt_bmap,
-						    &path[level].bp_newreq);
+	nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap, &path[level].bp_newreq);
  err_out_child_node:
 	for (level--; level > NILFS_BTREE_LEVEL_DATA; level--) {
-		nilfs_bmap_delete_block(&btree->bt_bmap, path[level].bp_sib_bh);
-		btree->bt_bmap.b_pops->bpop_abort_alloc_ptr(
-			&btree->bt_bmap, &path[level].bp_newreq);
+		nilfs_btnode_delete(path[level].bp_sib_bh);
+		nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap,
+					   &path[level].bp_newreq);
 
 	}
 
-	btree->bt_bmap.b_pops->bpop_abort_alloc_ptr(&btree->bt_bmap,
-						       &path[level].bp_newreq);
+	nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap, &path[level].bp_newreq);
  err_out_data:
 	*levelp = level;
 	stats->bs_nblocks = 0;
@@ -1049,14 +1143,12 @@ static void nilfs_btree_commit_insert(struct nilfs_btree *btree,
 
 	set_buffer_nilfs_volatile((struct buffer_head *)((unsigned long)ptr));
 	ptr = path[NILFS_BTREE_LEVEL_DATA].bp_newreq.bpr_ptr;
-	if (btree->bt_ops->btop_set_target != NULL)
-		btree->bt_ops->btop_set_target(btree, key, ptr);
+	if (NILFS_BMAP_USE_VBN(&btree->bt_bmap))
+		nilfs_btree_set_target_v(btree, key, ptr);
 
 	for (level = NILFS_BTREE_LEVEL_NODE_MIN; level <= maxlevel; level++) {
-		if (btree->bt_bmap.b_pops->bpop_commit_alloc_ptr != NULL) {
-			btree->bt_bmap.b_pops->bpop_commit_alloc_ptr(
-				&btree->bt_bmap, &path[level - 1].bp_newreq);
-		}
+		nilfs_bmap_commit_alloc_ptr(&btree->bt_bmap,
+					    &path[level - 1].bp_newreq);
 		path[level].bp_op(btree, path, level, &key, &ptr);
 	}
 
@@ -1153,7 +1245,7 @@ static void nilfs_btree_borrow_left(struct nilfs_btree *btree,
 	nilfs_btree_promote_key(btree, path, level + 1,
 				nilfs_btree_node_get_key(btree, node, 0));
 
-	nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_sib_bh);
+	brelse(path[level].bp_sib_bh);
 	path[level].bp_sib_bh = NULL;
 	path[level].bp_index += n;
 }
@@ -1192,7 +1284,7 @@ static void nilfs_btree_borrow_right(struct nilfs_btree *btree,
 				nilfs_btree_node_get_key(btree, right, 0));
 	path[level + 1].bp_index--;
 
-	nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_sib_bh);
+	brelse(path[level].bp_sib_bh);
 	path[level].bp_sib_bh = NULL;
 }
 
@@ -1221,7 +1313,7 @@ static void nilfs_btree_concat_left(struct nilfs_btree *btree,
 	unlock_buffer(path[level].bp_bh);
 	unlock_buffer(path[level].bp_sib_bh);
 
-	nilfs_bmap_delete_block(&btree->bt_bmap, path[level].bp_bh);
+	nilfs_btnode_delete(path[level].bp_bh);
 	path[level].bp_bh = path[level].bp_sib_bh;
 	path[level].bp_sib_bh = NULL;
 	path[level].bp_index += nilfs_btree_node_get_nchildren(btree, left);
@@ -1252,7 +1344,7 @@ static void nilfs_btree_concat_right(struct nilfs_btree *btree,
 	unlock_buffer(path[level].bp_bh);
 	unlock_buffer(path[level].bp_sib_bh);
 
-	nilfs_bmap_delete_block(&btree->bt_bmap, path[level].bp_sib_bh);
+	nilfs_btnode_delete(path[level].bp_sib_bh);
 	path[level].bp_sib_bh = NULL;
 	path[level + 1].bp_index++;
 }
@@ -1276,7 +1368,7 @@ static void nilfs_btree_shrink(struct nilfs_btree *btree,
 	nilfs_btree_node_move_left(btree, root, child, n);
 	unlock_buffer(path[level].bp_bh);
 
-	nilfs_bmap_delete_block(&btree->bt_bmap, path[level].bp_bh);
+	nilfs_btnode_delete(path[level].bp_bh);
 	path[level].bp_bh = NULL;
 }
 
@@ -1300,12 +1392,10 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 		path[level].bp_oldreq.bpr_ptr =
 			nilfs_btree_node_get_ptr(btree, node,
 						 path[level].bp_index);
-		if (btree->bt_bmap.b_pops->bpop_prepare_end_ptr != NULL) {
-			ret = btree->bt_bmap.b_pops->bpop_prepare_end_ptr(
-				&btree->bt_bmap, &path[level].bp_oldreq);
-			if (ret < 0)
-				goto err_out_child_node;
-		}
+		ret = nilfs_bmap_prepare_end_ptr(&btree->bt_bmap,
+						 &path[level].bp_oldreq);
+		if (ret < 0)
+			goto err_out_child_node;
 
 		if (nilfs_btree_node_get_nchildren(btree, node) >
 		    nilfs_btree_node_nchildren_min(btree, node)) {
@@ -1321,8 +1411,7 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 			/* left sibling */
 			sibptr = nilfs_btree_node_get_ptr(btree, parent,
 							  pindex - 1);
-			ret = nilfs_bmap_get_block(&btree->bt_bmap, sibptr,
-						   &bh);
+			ret = nilfs_btree_get_block(btree, sibptr, &bh);
 			if (ret < 0)
 				goto err_out_curr_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
@@ -1343,8 +1432,7 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 			/* right sibling */
 			sibptr = nilfs_btree_node_get_ptr(btree, parent,
 							  pindex + 1);
-			ret = nilfs_bmap_get_block(&btree->bt_bmap, sibptr,
-						   &bh);
+			ret = nilfs_btree_get_block(btree, sibptr, &bh);
 			if (ret < 0)
 				goto err_out_curr_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
@@ -1381,12 +1469,12 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 	node = nilfs_btree_get_root(btree);
 	path[level].bp_oldreq.bpr_ptr =
 		nilfs_btree_node_get_ptr(btree, node, path[level].bp_index);
-	if (btree->bt_bmap.b_pops->bpop_prepare_end_ptr != NULL) {
-		ret = btree->bt_bmap.b_pops->bpop_prepare_end_ptr(
-			&btree->bt_bmap, &path[level].bp_oldreq);
-		if (ret < 0)
-			goto err_out_child_node;
-	}
+
+	ret = nilfs_bmap_prepare_end_ptr(&btree->bt_bmap,
+					 &path[level].bp_oldreq);
+	if (ret < 0)
+		goto err_out_child_node;
+
 	/* child of the root node is deleted */
 	path[level].bp_op = nilfs_btree_do_delete;
 	stats->bs_nblocks++;
@@ -1398,15 +1486,12 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 
 	/* error */
  err_out_curr_node:
-	if (btree->bt_bmap.b_pops->bpop_abort_end_ptr != NULL)
-		btree->bt_bmap.b_pops->bpop_abort_end_ptr(
-			&btree->bt_bmap, &path[level].bp_oldreq);
+	nilfs_bmap_abort_end_ptr(&btree->bt_bmap, &path[level].bp_oldreq);
  err_out_child_node:
 	for (level--; level >= NILFS_BTREE_LEVEL_NODE_MIN; level--) {
-		nilfs_bmap_put_block(&btree->bt_bmap, path[level].bp_sib_bh);
-		if (btree->bt_bmap.b_pops->bpop_abort_end_ptr != NULL)
-			btree->bt_bmap.b_pops->bpop_abort_end_ptr(
-				&btree->bt_bmap, &path[level].bp_oldreq);
+		brelse(path[level].bp_sib_bh);
+		nilfs_bmap_abort_end_ptr(&btree->bt_bmap,
+					 &path[level].bp_oldreq);
 	}
 	*levelp = level;
 	stats->bs_nblocks = 0;
@@ -1420,9 +1505,8 @@ static void nilfs_btree_commit_delete(struct nilfs_btree *btree,
 	int level;
 
 	for (level = NILFS_BTREE_LEVEL_NODE_MIN; level <= maxlevel; level++) {
-		if (btree->bt_bmap.b_pops->bpop_commit_end_ptr != NULL)
-			btree->bt_bmap.b_pops->bpop_commit_end_ptr(
-				&btree->bt_bmap, &path[level].bp_oldreq);
+		nilfs_bmap_commit_end_ptr(&btree->bt_bmap,
+					  &path[level].bp_oldreq);
 		path[level].bp_op(btree, path, level, NULL, NULL);
 	}
 
@@ -1501,7 +1585,7 @@ static int nilfs_btree_check_delete(struct nilfs_bmap *bmap, __u64 key)
 		if (nchildren > 1)
 			return 0;
 		ptr = nilfs_btree_node_get_ptr(btree, root, nchildren - 1);
-		ret = nilfs_bmap_get_block(bmap, ptr, &bh);
+		ret = nilfs_btree_get_block(btree, ptr, &bh);
 		if (ret < 0)
 			return ret;
 		node = (struct nilfs_btree_node *)bh->b_data;
@@ -1515,9 +1599,9 @@ static int nilfs_btree_check_delete(struct nilfs_bmap *bmap, __u64 key)
 	nextmaxkey = (nchildren > 1) ?
 		nilfs_btree_node_get_key(btree, node, nchildren - 2) : 0;
 	if (bh != NULL)
-		nilfs_bmap_put_block(bmap, bh);
+		brelse(bh);
 
-	return (maxkey == key) && (nextmaxkey < bmap->b_low);
+	return (maxkey == key) && (nextmaxkey < NILFS_BMAP_LARGE_LOW);
 }
 
 static int nilfs_btree_gather_data(struct nilfs_bmap *bmap,
@@ -1542,7 +1626,7 @@ static int nilfs_btree_gather_data(struct nilfs_bmap *bmap,
 		nchildren = nilfs_btree_node_get_nchildren(btree, root);
 		WARN_ON(nchildren > 1);
 		ptr = nilfs_btree_node_get_ptr(btree, root, nchildren - 1);
-		ret = nilfs_bmap_get_block(bmap, ptr, &bh);
+		ret = nilfs_btree_get_block(btree, ptr, &bh);
 		if (ret < 0)
 			return ret;
 		node = (struct nilfs_btree_node *)bh->b_data;
@@ -1563,7 +1647,7 @@ static int nilfs_btree_gather_data(struct nilfs_bmap *bmap,
 	}
 
 	if (bh != NULL)
-		nilfs_bmap_put_block(bmap, bh);
+		brelse(bh);
 
 	return nitems;
 }
@@ -1584,10 +1668,10 @@ nilfs_btree_prepare_convert_and_insert(struct nilfs_bmap *bmap, __u64 key,
 
 	/* for data */
 	/* cannot find near ptr */
-	if (btree->bt_ops->btop_find_target != NULL)
-		dreq->bpr_ptr
-			= btree->bt_ops->btop_find_target(btree, NULL, key);
-	ret = bmap->b_pops->bpop_prepare_alloc_ptr(bmap, dreq);
+	if (NILFS_BMAP_USE_VBN(bmap))
+		dreq->bpr_ptr = nilfs_btree_find_target_v(btree, NULL, key);
+
+	ret = nilfs_bmap_prepare_alloc_ptr(bmap, dreq);
 	if (ret < 0)
 		return ret;
 
@@ -1595,11 +1679,11 @@ nilfs_btree_prepare_convert_and_insert(struct nilfs_bmap *bmap, __u64 key,
 	stats->bs_nblocks++;
 	if (nreq != NULL) {
 		nreq->bpr_ptr = dreq->bpr_ptr + 1;
-		ret = bmap->b_pops->bpop_prepare_alloc_ptr(bmap, nreq);
+		ret = nilfs_bmap_prepare_alloc_ptr(bmap, nreq);
 		if (ret < 0)
 			goto err_out_dreq;
 
-		ret = nilfs_bmap_get_new_block(bmap, nreq->bpr_ptr, &bh);
+		ret = nilfs_btree_get_new_block(btree, nreq->bpr_ptr, &bh);
 		if (ret < 0)
 			goto err_out_nreq;
 
@@ -1612,9 +1696,9 @@ nilfs_btree_prepare_convert_and_insert(struct nilfs_bmap *bmap, __u64 key,
 
 	/* error */
  err_out_nreq:
-	bmap->b_pops->bpop_abort_alloc_ptr(bmap, nreq);
+	nilfs_bmap_abort_alloc_ptr(bmap, nreq);
  err_out_dreq:
-	bmap->b_pops->bpop_abort_alloc_ptr(bmap, dreq);
+	nilfs_bmap_abort_alloc_ptr(bmap, dreq);
 	stats->bs_nblocks = 0;
 	return ret;
 
@@ -1624,7 +1708,7 @@ static void
 nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 				      __u64 key, __u64 ptr,
 				      const __u64 *keys, const __u64 *ptrs,
-				      int n, __u64 low, __u64 high,
+				      int n,
 				      union nilfs_bmap_ptr_req *dreq,
 				      union nilfs_bmap_ptr_req *nreq,
 				      struct buffer_head *bh)
@@ -1642,12 +1726,10 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 
 	/* convert and insert */
 	btree = (struct nilfs_btree *)bmap;
-	nilfs_btree_init(bmap, low, high);
+	nilfs_btree_init(bmap);
 	if (nreq != NULL) {
-		if (bmap->b_pops->bpop_commit_alloc_ptr != NULL) {
-			bmap->b_pops->bpop_commit_alloc_ptr(bmap, dreq);
-			bmap->b_pops->bpop_commit_alloc_ptr(bmap, nreq);
-		}
+		nilfs_bmap_commit_alloc_ptr(bmap, dreq);
+		nilfs_bmap_commit_alloc_ptr(bmap, nreq);
 
 		/* create child node at level 1 */
 		lock_buffer(bh);
@@ -1661,7 +1743,7 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 			nilfs_bmap_set_dirty(bmap);
 
 		unlock_buffer(bh);
-		nilfs_bmap_put_block(bmap, bh);
+		brelse(bh);
 
 		/* create root node at level 2 */
 		node = nilfs_btree_get_root(btree);
@@ -1669,8 +1751,7 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 		nilfs_btree_node_init(btree, node, NILFS_BTREE_NODE_ROOT,
 				      2, 1, &keys[0], &tmpptr);
 	} else {
-		if (bmap->b_pops->bpop_commit_alloc_ptr != NULL)
-			bmap->b_pops->bpop_commit_alloc_ptr(bmap, dreq);
+		nilfs_bmap_commit_alloc_ptr(bmap, dreq);
 
 		/* create root node at level 1 */
 		node = nilfs_btree_get_root(btree);
@@ -1682,8 +1763,8 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 			nilfs_bmap_set_dirty(bmap);
 	}
 
-	if (btree->bt_ops->btop_set_target != NULL)
-		btree->bt_ops->btop_set_target(btree, key, dreq->bpr_ptr);
+	if (NILFS_BMAP_USE_VBN(bmap))
+		nilfs_btree_set_target_v(btree, key, dreq->bpr_ptr);
 }
 
 /**
@@ -1694,13 +1775,10 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
  * @keys:
  * @ptrs:
  * @n:
- * @low:
- * @high:
  */
 int nilfs_btree_convert_and_insert(struct nilfs_bmap *bmap,
 				   __u64 key, __u64 ptr,
-				   const __u64 *keys, const __u64 *ptrs,
-				   int n, __u64 low, __u64 high)
+				   const __u64 *keys, const __u64 *ptrs, int n)
 {
 	struct buffer_head *bh;
 	union nilfs_bmap_ptr_req dreq, nreq, *di, *ni;
@@ -1725,7 +1803,7 @@ int nilfs_btree_convert_and_insert(struct nilfs_bmap *bmap,
 	if (ret < 0)
 		return ret;
 	nilfs_btree_commit_convert_and_insert(bmap, key, ptr, keys, ptrs, n,
-					      low, high, di, ni, bh);
+					      di, ni, bh);
 	nilfs_bmap_add_blocks(bmap, stats.bs_nblocks);
 	return 0;
 }
@@ -1754,9 +1832,9 @@ static int nilfs_btree_prepare_update_v(struct nilfs_btree *btree,
 		nilfs_btree_node_get_ptr(btree, parent,
 					 path[level + 1].bp_index);
 	path[level].bp_newreq.bpr_ptr = path[level].bp_oldreq.bpr_ptr + 1;
-	ret = nilfs_bmap_prepare_update(&btree->bt_bmap,
-					&path[level].bp_oldreq,
-					&path[level].bp_newreq);
+	ret = nilfs_bmap_prepare_update_v(&btree->bt_bmap,
+					  &path[level].bp_oldreq,
+					  &path[level].bp_newreq);
 	if (ret < 0)
 		return ret;
 
@@ -1768,9 +1846,9 @@ static int nilfs_btree_prepare_update_v(struct nilfs_btree *btree,
 			&NILFS_BMAP_I(&btree->bt_bmap)->i_btnode_cache,
 			&path[level].bp_ctxt);
 		if (ret < 0) {
-			nilfs_bmap_abort_update(&btree->bt_bmap,
-						&path[level].bp_oldreq,
-						&path[level].bp_newreq);
+			nilfs_bmap_abort_update_v(&btree->bt_bmap,
+						  &path[level].bp_oldreq,
+						  &path[level].bp_newreq);
 			return ret;
 		}
 	}
@@ -1784,9 +1862,9 @@ static void nilfs_btree_commit_update_v(struct nilfs_btree *btree,
 {
 	struct nilfs_btree_node *parent;
 
-	nilfs_bmap_commit_update(&btree->bt_bmap,
-				 &path[level].bp_oldreq,
-				 &path[level].bp_newreq);
+	nilfs_bmap_commit_update_v(&btree->bt_bmap,
+				   &path[level].bp_oldreq,
+				   &path[level].bp_newreq);
 
 	if (buffer_nilfs_node(path[level].bp_bh)) {
 		nilfs_btnode_commit_change_key(
@@ -1805,9 +1883,9 @@ static void nilfs_btree_abort_update_v(struct nilfs_btree *btree,
 				       struct nilfs_btree_path *path,
 				       int level)
 {
-	nilfs_bmap_abort_update(&btree->bt_bmap,
-				&path[level].bp_oldreq,
-				&path[level].bp_newreq);
+	nilfs_bmap_abort_update_v(&btree->bt_bmap,
+				  &path[level].bp_oldreq,
+				  &path[level].bp_newreq);
 	if (buffer_nilfs_node(path[level].bp_bh))
 		nilfs_btnode_abort_change_key(
 			&NILFS_BMAP_I(&btree->bt_bmap)->i_btnode_cache,
@@ -1930,7 +2008,9 @@ static int nilfs_btree_propagate(const struct nilfs_bmap *bmap,
 		goto out;
 	}
 
-	ret = btree->bt_ops->btop_propagate(btree, path, level, bh);
+	ret = NILFS_BMAP_USE_VBN(bmap) ?
+		nilfs_btree_propagate_v(btree, path, level, bh) :
+		nilfs_btree_propagate_p(btree, path, level, bh);
 
  out:
 	nilfs_btree_clear_path(btree, path);
@@ -2066,12 +2146,9 @@ static int nilfs_btree_assign_v(struct nilfs_btree *btree,
 	ptr = nilfs_btree_node_get_ptr(btree, parent,
 				       path[level + 1].bp_index);
 	req.bpr_ptr = ptr;
-	ret = btree->bt_bmap.b_pops->bpop_prepare_start_ptr(&btree->bt_bmap,
-							       &req);
-	if (ret < 0)
+	ret = nilfs_bmap_start_v(&btree->bt_bmap, &req, blocknr);
+	if (unlikely(ret < 0))
 		return ret;
-	btree->bt_bmap.b_pops->bpop_commit_start_ptr(&btree->bt_bmap,
-							&req, blocknr);
 
 	key = nilfs_btree_node_get_key(btree, parent,
 				       path[level + 1].bp_index);
@@ -2114,8 +2191,9 @@ static int nilfs_btree_assign(struct nilfs_bmap *bmap,
 		goto out;
 	}
 
-	ret = btree->bt_ops->btop_assign(btree, path, level, bh,
-					    blocknr, binfo);
+	ret = NILFS_BMAP_USE_VBN(bmap) ?
+		nilfs_btree_assign_v(btree, path, level, bh, blocknr, binfo) :
+		nilfs_btree_assign_p(btree, path, level, bh, blocknr, binfo);
 
  out:
 	nilfs_btree_clear_path(btree, path);
@@ -2171,7 +2249,7 @@ static int nilfs_btree_mark(struct nilfs_bmap *bmap, __u64 key, int level)
 		WARN_ON(ret == -ENOENT);
 		goto out;
 	}
-	ret = nilfs_bmap_get_block(&btree->bt_bmap, ptr, &bh);
+	ret = nilfs_btree_get_block(btree, ptr, &bh);
 	if (ret < 0) {
 		WARN_ON(ret == -ENOENT);
 		goto out;
@@ -2179,7 +2257,7 @@ static int nilfs_btree_mark(struct nilfs_bmap *bmap, __u64 key, int level)
 
 	if (!buffer_dirty(bh))
 		nilfs_btnode_mark_dirty(bh);
-	nilfs_bmap_put_block(&btree->bt_bmap, bh);
+	brelse(bh);
 	if (!nilfs_bmap_dirty(&btree->bt_bmap))
 		nilfs_bmap_set_dirty(&btree->bt_bmap);
 
@@ -2191,6 +2269,7 @@ static int nilfs_btree_mark(struct nilfs_bmap *bmap, __u64 key, int level)
 
 static const struct nilfs_bmap_operations nilfs_btree_ops = {
 	.bop_lookup		=	nilfs_btree_lookup,
+	.bop_lookup_contig	=	nilfs_btree_lookup_contig,
 	.bop_insert		=	nilfs_btree_insert,
 	.bop_delete		=	nilfs_btree_delete,
 	.bop_clear		=	NULL,
@@ -2210,6 +2289,7 @@ static const struct nilfs_bmap_operations nilfs_btree_ops = {
 
 static const struct nilfs_bmap_operations nilfs_btree_ops_gc = {
 	.bop_lookup		=	NULL,
+	.bop_lookup_contig	=	NULL,
 	.bop_insert		=	NULL,
 	.bop_delete		=	NULL,
 	.bop_clear		=	NULL,
@@ -2227,43 +2307,13 @@ static const struct nilfs_bmap_operations nilfs_btree_ops_gc = {
 	.bop_gather_data	=	NULL,
 };
 
-static const struct nilfs_btree_operations nilfs_btree_ops_v = {
-	.btop_find_target	=	nilfs_btree_find_target_v,
-	.btop_set_target	=	nilfs_btree_set_target_v,
-	.btop_propagate		=	nilfs_btree_propagate_v,
-	.btop_assign		=	nilfs_btree_assign_v,
-};
-
-static const struct nilfs_btree_operations nilfs_btree_ops_p = {
-	.btop_find_target	=	NULL,
-	.btop_set_target	=	NULL,
-	.btop_propagate		=	nilfs_btree_propagate_p,
-	.btop_assign		=	nilfs_btree_assign_p,
-};
-
-int nilfs_btree_init(struct nilfs_bmap *bmap, __u64 low, __u64 high)
+int nilfs_btree_init(struct nilfs_bmap *bmap)
 {
-	struct nilfs_btree *btree;
-
-	btree = (struct nilfs_btree *)bmap;
 	bmap->b_ops = &nilfs_btree_ops;
-	bmap->b_low = low;
-	bmap->b_high = high;
-	switch (bmap->b_inode->i_ino) {
-	case NILFS_DAT_INO:
-		btree->bt_ops = &nilfs_btree_ops_p;
-		break;
-	default:
-		btree->bt_ops = &nilfs_btree_ops_v;
-		break;
-	}
-
 	return 0;
 }
 
 void nilfs_btree_init_gc(struct nilfs_bmap *bmap)
 {
-	bmap->b_low = NILFS_BMAP_LARGE_LOW;
-	bmap->b_high = NILFS_BMAP_LARGE_HIGH;
 	bmap->b_ops = &nilfs_btree_ops_gc;
 }
diff --git a/fs/nilfs2/btree.h b/fs/nilfs2/btree.h
index 4766deb52fb1..0e72bbbc6b64 100644
--- a/fs/nilfs2/btree.h
+++ b/fs/nilfs2/btree.h
@@ -34,28 +34,6 @@ struct nilfs_btree;
 struct nilfs_btree_path;
 
 /**
- * struct nilfs_btree_operations - B-tree operation table
- */
-struct nilfs_btree_operations {
-	__u64 (*btop_find_target)(const struct nilfs_btree *,
-				  const struct nilfs_btree_path *, __u64);
-	void (*btop_set_target)(struct nilfs_btree *, __u64, __u64);
-
-	struct the_nilfs *(*btop_get_nilfs)(struct nilfs_btree *);
-
-	int (*btop_propagate)(struct nilfs_btree *,
-			      struct nilfs_btree_path *,
-			      int,
-			      struct buffer_head *);
-	int (*btop_assign)(struct nilfs_btree *,
-			   struct nilfs_btree_path *,
-			   int,
-			   struct buffer_head **,
-			   sector_t,
-			   union nilfs_binfo *);
-};
-
-/**
  * struct nilfs_btree_node - B-tree node
  * @bn_flags: flags
  * @bn_level: level
@@ -80,13 +58,9 @@ struct nilfs_btree_node {
 /**
  * struct nilfs_btree - B-tree structure
  * @bt_bmap: bmap base structure
- * @bt_ops: B-tree operation table
  */
 struct nilfs_btree {
 	struct nilfs_bmap bt_bmap;
-
-	/* B-tree-specific members */
-	const struct nilfs_btree_operations *bt_ops;
 };
 
 
@@ -108,10 +82,9 @@ struct nilfs_btree {
 
 int nilfs_btree_path_cache_init(void);
 void nilfs_btree_path_cache_destroy(void);
-int nilfs_btree_init(struct nilfs_bmap *, __u64, __u64);
+int nilfs_btree_init(struct nilfs_bmap *);
 int nilfs_btree_convert_and_insert(struct nilfs_bmap *, __u64, __u64,
-				   const __u64 *, const __u64 *,
-				   int, __u64, __u64);
+				   const __u64 *, const __u64 *, int);
 void nilfs_btree_init_gc(struct nilfs_bmap *);
 
 #endif	/* _NILFS_BTREE_H */
diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c
index cadd36b14d07..7d49813f66d6 100644
--- a/fs/nilfs2/cpfile.c
+++ b/fs/nilfs2/cpfile.c
@@ -295,10 +295,6 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile,
 		return -EINVAL;
 	}
 
-	/* cannot delete the latest checkpoint */
-	if (start == nilfs_mdt_cno(cpfile) - 1)
-		return -EPERM;
-
 	down_write(&NILFS_MDT(cpfile)->mi_sem);
 
 	ret = nilfs_cpfile_get_header_block(cpfile, &header_bh);
@@ -384,9 +380,10 @@ static void nilfs_cpfile_checkpoint_to_cpinfo(struct inode *cpfile,
 }
 
 static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop,
-					  struct nilfs_cpinfo *ci, size_t nci)
+					  void *buf, unsigned cisz, size_t nci)
 {
 	struct nilfs_checkpoint *cp;
+	struct nilfs_cpinfo *ci = buf;
 	struct buffer_head *bh;
 	size_t cpsz = NILFS_MDT(cpfile)->mi_entry_size;
 	__u64 cur_cno = nilfs_mdt_cno(cpfile), cno = *cnop;
@@ -410,17 +407,22 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop,
 		kaddr = kmap_atomic(bh->b_page, KM_USER0);
 		cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, bh, kaddr);
 		for (i = 0; i < ncps && n < nci; i++, cp = (void *)cp + cpsz) {
-			if (!nilfs_checkpoint_invalid(cp))
-				nilfs_cpfile_checkpoint_to_cpinfo(
-					cpfile, cp, &ci[n++]);
+			if (!nilfs_checkpoint_invalid(cp)) {
+				nilfs_cpfile_checkpoint_to_cpinfo(cpfile, cp,
+								  ci);
+				ci = (void *)ci + cisz;
+				n++;
+			}
 		}
 		kunmap_atomic(kaddr, KM_USER0);
 		brelse(bh);
 	}
 
 	ret = n;
-	if (n > 0)
-		*cnop = ci[n - 1].ci_cno + 1;
+	if (n > 0) {
+		ci = (void *)ci - cisz;
+		*cnop = ci->ci_cno + 1;
+	}
 
  out:
 	up_read(&NILFS_MDT(cpfile)->mi_sem);
@@ -428,11 +430,12 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop,
 }
 
 static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
-					  struct nilfs_cpinfo *ci, size_t nci)
+					  void *buf, unsigned cisz, size_t nci)
 {
 	struct buffer_head *bh;
 	struct nilfs_cpfile_header *header;
 	struct nilfs_checkpoint *cp;
+	struct nilfs_cpinfo *ci = buf;
 	__u64 curr = *cnop, next;
 	unsigned long curr_blkoff, next_blkoff;
 	void *kaddr;
@@ -472,7 +475,9 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
 		if (unlikely(nilfs_checkpoint_invalid(cp) ||
 			     !nilfs_checkpoint_snapshot(cp)))
 			break;
-		nilfs_cpfile_checkpoint_to_cpinfo(cpfile, cp, &ci[n++]);
+		nilfs_cpfile_checkpoint_to_cpinfo(cpfile, cp, ci);
+		ci = (void *)ci + cisz;
+		n++;
 		next = le64_to_cpu(cp->cp_snapshot_list.ssl_next);
 		if (next == 0)
 			break; /* reach end of the snapshot list */
@@ -511,13 +516,13 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
  */
 
 ssize_t nilfs_cpfile_get_cpinfo(struct inode *cpfile, __u64 *cnop, int mode,
-				struct nilfs_cpinfo *ci, size_t nci)
+				void *buf, unsigned cisz, size_t nci)
 {
 	switch (mode) {
 	case NILFS_CHECKPOINT:
-		return nilfs_cpfile_do_get_cpinfo(cpfile, cnop, ci, nci);
+		return nilfs_cpfile_do_get_cpinfo(cpfile, cnop, buf, cisz, nci);
 	case NILFS_SNAPSHOT:
-		return nilfs_cpfile_do_get_ssinfo(cpfile, cnop, ci, nci);
+		return nilfs_cpfile_do_get_ssinfo(cpfile, cnop, buf, cisz, nci);
 	default:
 		return -EINVAL;
 	}
@@ -533,20 +538,14 @@ int nilfs_cpfile_delete_checkpoint(struct inode *cpfile, __u64 cno)
 	struct nilfs_cpinfo ci;
 	__u64 tcno = cno;
 	ssize_t nci;
-	int ret;
 
-	nci = nilfs_cpfile_do_get_cpinfo(cpfile, &tcno, &ci, 1);
+	nci = nilfs_cpfile_do_get_cpinfo(cpfile, &tcno, &ci, sizeof(ci), 1);
 	if (nci < 0)
 		return nci;
 	else if (nci == 0 || ci.ci_cno != cno)
 		return -ENOENT;
-
-	/* cannot delete the latest checkpoint nor snapshots */
-	ret = nilfs_cpinfo_snapshot(&ci);
-	if (ret < 0)
-		return ret;
-	else if (ret > 0 || cno == nilfs_mdt_cno(cpfile) - 1)
-		return -EPERM;
+	else if (nilfs_cpinfo_snapshot(&ci))
+		return -EBUSY;
 
 	return nilfs_cpfile_delete_checkpoints(cpfile, cno, cno + 1);
 }
diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h
index 1a8a1008c342..788a45950197 100644
--- a/fs/nilfs2/cpfile.h
+++ b/fs/nilfs2/cpfile.h
@@ -39,7 +39,7 @@ int nilfs_cpfile_delete_checkpoint(struct inode *, __u64);
 int nilfs_cpfile_change_cpmode(struct inode *, __u64, int);
 int nilfs_cpfile_is_snapshot(struct inode *, __u64);
 int nilfs_cpfile_get_stat(struct inode *, struct nilfs_cpstat *);
-ssize_t nilfs_cpfile_get_cpinfo(struct inode *, __u64 *, int,
-				struct nilfs_cpinfo *, size_t);
+ssize_t nilfs_cpfile_get_cpinfo(struct inode *, __u64 *, int, void *, unsigned,
+				size_t);
 
 #endif	/* _NILFS_CPFILE_H */
diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c
index bb8a5818e7f1..0b2710e2d565 100644
--- a/fs/nilfs2/dat.c
+++ b/fs/nilfs2/dat.c
@@ -92,21 +92,6 @@ void nilfs_dat_abort_alloc(struct inode *dat, struct nilfs_palloc_req *req)
 	nilfs_palloc_abort_alloc_entry(dat, req);
 }
 
-int nilfs_dat_prepare_free(struct inode *dat, struct nilfs_palloc_req *req)
-{
-	int ret;
-
-	ret = nilfs_palloc_prepare_free_entry(dat, req);
-	if (ret < 0)
-		return ret;
-	ret = nilfs_dat_prepare_entry(dat, req, 0);
-	if (ret < 0) {
-		nilfs_palloc_abort_free_entry(dat, req);
-		return ret;
-	}
-	return 0;
-}
-
 void nilfs_dat_commit_free(struct inode *dat, struct nilfs_palloc_req *req)
 {
 	struct nilfs_dat_entry *entry;
@@ -391,36 +376,37 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp)
 	return ret;
 }
 
-ssize_t nilfs_dat_get_vinfo(struct inode *dat, struct nilfs_vinfo *vinfo,
+ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned visz,
 			    size_t nvi)
 {
 	struct buffer_head *entry_bh;
 	struct nilfs_dat_entry *entry;
+	struct nilfs_vinfo *vinfo = buf;
 	__u64 first, last;
 	void *kaddr;
 	unsigned long entries_per_block = NILFS_MDT(dat)->mi_entries_per_block;
 	int i, j, n, ret;
 
 	for (i = 0; i < nvi; i += n) {
-		ret = nilfs_palloc_get_entry_block(dat, vinfo[i].vi_vblocknr,
+		ret = nilfs_palloc_get_entry_block(dat, vinfo->vi_vblocknr,
 						   0, &entry_bh);
 		if (ret < 0)
 			return ret;
 		kaddr = kmap_atomic(entry_bh->b_page, KM_USER0);
 		/* last virtual block number in this block */
-		first = vinfo[i].vi_vblocknr;
+		first = vinfo->vi_vblocknr;
 		do_div(first, entries_per_block);
 		first *= entries_per_block;
 		last = first + entries_per_block - 1;
 		for (j = i, n = 0;
-		     j < nvi && vinfo[j].vi_vblocknr >= first &&
-			     vinfo[j].vi_vblocknr <= last;
-		     j++, n++) {
+		     j < nvi && vinfo->vi_vblocknr >= first &&
+			     vinfo->vi_vblocknr <= last;
+		     j++, n++, vinfo = (void *)vinfo + visz) {
 			entry = nilfs_palloc_block_get_entry(
-				dat, vinfo[j].vi_vblocknr, entry_bh, kaddr);
-			vinfo[j].vi_start = le64_to_cpu(entry->de_start);
-			vinfo[j].vi_end = le64_to_cpu(entry->de_end);
-			vinfo[j].vi_blocknr = le64_to_cpu(entry->de_blocknr);
+				dat, vinfo->vi_vblocknr, entry_bh, kaddr);
+			vinfo->vi_start = le64_to_cpu(entry->de_start);
+			vinfo->vi_end = le64_to_cpu(entry->de_end);
+			vinfo->vi_blocknr = le64_to_cpu(entry->de_blocknr);
 		}
 		kunmap_atomic(kaddr, KM_USER0);
 		brelse(entry_bh);
diff --git a/fs/nilfs2/dat.h b/fs/nilfs2/dat.h
index d9560654a4b7..d328b81eead4 100644
--- a/fs/nilfs2/dat.h
+++ b/fs/nilfs2/dat.h
@@ -47,6 +47,6 @@ void nilfs_dat_abort_end(struct inode *, struct nilfs_palloc_req *);
 int nilfs_dat_mark_dirty(struct inode *, __u64);
 int nilfs_dat_freev(struct inode *, __u64 *, size_t);
 int nilfs_dat_move(struct inode *, __u64, sector_t);
-ssize_t nilfs_dat_get_vinfo(struct inode *, struct nilfs_vinfo *, size_t);
+ssize_t nilfs_dat_get_vinfo(struct inode *, void *, unsigned, size_t);
 
 #endif	/* _NILFS_DAT_H */
diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c
index c6379e482781..342d9765df8d 100644
--- a/fs/nilfs2/direct.c
+++ b/fs/nilfs2/direct.c
@@ -25,6 +25,7 @@
 #include "page.h"
 #include "direct.h"
 #include "alloc.h"
+#include "dat.h"
 
 static inline __le64 *nilfs_direct_dptrs(const struct nilfs_direct *direct)
 {
@@ -62,6 +63,47 @@ static int nilfs_direct_lookup(const struct nilfs_bmap *bmap,
 	return 0;
 }
 
+static int nilfs_direct_lookup_contig(const struct nilfs_bmap *bmap,
+				      __u64 key, __u64 *ptrp,
+				      unsigned maxblocks)
+{
+	struct nilfs_direct *direct = (struct nilfs_direct *)bmap;
+	struct inode *dat = NULL;
+	__u64 ptr, ptr2;
+	sector_t blocknr;
+	int ret, cnt;
+
+	if (key > NILFS_DIRECT_KEY_MAX ||
+	    (ptr = nilfs_direct_get_ptr(direct, key)) ==
+	    NILFS_BMAP_INVALID_PTR)
+		return -ENOENT;
+
+	if (NILFS_BMAP_USE_VBN(bmap)) {
+		dat = nilfs_bmap_get_dat(bmap);
+		ret = nilfs_dat_translate(dat, ptr, &blocknr);
+		if (ret < 0)
+			return ret;
+		ptr = blocknr;
+	}
+
+	maxblocks = min_t(unsigned, maxblocks, NILFS_DIRECT_KEY_MAX - key + 1);
+	for (cnt = 1; cnt < maxblocks &&
+		     (ptr2 = nilfs_direct_get_ptr(direct, key + cnt)) !=
+		     NILFS_BMAP_INVALID_PTR;
+	     cnt++) {
+		if (dat) {
+			ret = nilfs_dat_translate(dat, ptr2, &blocknr);
+			if (ret < 0)
+				return ret;
+			ptr2 = blocknr;
+		}
+		if (ptr2 != ptr + cnt)
+			break;
+	}
+	*ptrp = ptr;
+	return cnt;
+}
+
 static __u64
 nilfs_direct_find_target_v(const struct nilfs_direct *direct, __u64 key)
 {
@@ -90,10 +132,9 @@ static int nilfs_direct_prepare_insert(struct nilfs_direct *direct,
 {
 	int ret;
 
-	if (direct->d_ops->dop_find_target != NULL)
-		req->bpr_ptr = direct->d_ops->dop_find_target(direct, key);
-	ret = direct->d_bmap.b_pops->bpop_prepare_alloc_ptr(&direct->d_bmap,
-							       req);
+	if (NILFS_BMAP_USE_VBN(&direct->d_bmap))
+		req->bpr_ptr = nilfs_direct_find_target_v(direct, key);
+	ret = nilfs_bmap_prepare_alloc_ptr(&direct->d_bmap, req);
 	if (ret < 0)
 		return ret;
 
@@ -111,16 +152,14 @@ static void nilfs_direct_commit_insert(struct nilfs_direct *direct,
 	bh = (struct buffer_head *)((unsigned long)ptr);
 	set_buffer_nilfs_volatile(bh);
 
-	if (direct->d_bmap.b_pops->bpop_commit_alloc_ptr != NULL)
-		direct->d_bmap.b_pops->bpop_commit_alloc_ptr(
-			&direct->d_bmap, req);
+	nilfs_bmap_commit_alloc_ptr(&direct->d_bmap, req);
 	nilfs_direct_set_ptr(direct, key, req->bpr_ptr);
 
 	if (!nilfs_bmap_dirty(&direct->d_bmap))
 		nilfs_bmap_set_dirty(&direct->d_bmap);
 
-	if (direct->d_ops->dop_set_target != NULL)
-		direct->d_ops->dop_set_target(direct, key, req->bpr_ptr);
+	if (NILFS_BMAP_USE_VBN(&direct->d_bmap))
+		nilfs_direct_set_target_v(direct, key, req->bpr_ptr);
 }
 
 static int nilfs_direct_insert(struct nilfs_bmap *bmap, __u64 key, __u64 ptr)
@@ -152,25 +191,18 @@ static int nilfs_direct_prepare_delete(struct nilfs_direct *direct,
 {
 	int ret;
 
-	if (direct->d_bmap.b_pops->bpop_prepare_end_ptr != NULL) {
-		req->bpr_ptr = nilfs_direct_get_ptr(direct, key);
-		ret = direct->d_bmap.b_pops->bpop_prepare_end_ptr(
-			&direct->d_bmap, req);
-		if (ret < 0)
-			return ret;
-	}
-
-	stats->bs_nblocks = 1;
-	return 0;
+	req->bpr_ptr = nilfs_direct_get_ptr(direct, key);
+	ret = nilfs_bmap_prepare_end_ptr(&direct->d_bmap, req);
+	if (!ret)
+		stats->bs_nblocks = 1;
+	return ret;
 }
 
 static void nilfs_direct_commit_delete(struct nilfs_direct *direct,
 				       union nilfs_bmap_ptr_req *req,
 				       __u64 key)
 {
-	if (direct->d_bmap.b_pops->bpop_commit_end_ptr != NULL)
-		direct->d_bmap.b_pops->bpop_commit_end_ptr(
-			&direct->d_bmap, req);
+	nilfs_bmap_commit_end_ptr(&direct->d_bmap, req);
 	nilfs_direct_set_ptr(direct, key, NILFS_BMAP_INVALID_PTR);
 }
 
@@ -244,8 +276,7 @@ static int nilfs_direct_gather_data(struct nilfs_bmap *bmap,
 }
 
 int nilfs_direct_delete_and_convert(struct nilfs_bmap *bmap,
-				    __u64 key, __u64 *keys, __u64 *ptrs,
-				    int n, __u64 low, __u64 high)
+				    __u64 key, __u64 *keys, __u64 *ptrs, int n)
 {
 	struct nilfs_direct *direct;
 	__le64 *dptrs;
@@ -275,8 +306,7 @@ int nilfs_direct_delete_and_convert(struct nilfs_bmap *bmap,
 			dptrs[i] = NILFS_BMAP_INVALID_PTR;
 	}
 
-	nilfs_direct_init(bmap, low, high);
-
+	nilfs_direct_init(bmap);
 	return 0;
 }
 
@@ -293,11 +323,11 @@ static int nilfs_direct_propagate_v(struct nilfs_direct *direct,
 	if (!buffer_nilfs_volatile(bh)) {
 		oldreq.bpr_ptr = ptr;
 		newreq.bpr_ptr = ptr;
-		ret = nilfs_bmap_prepare_update(&direct->d_bmap, &oldreq,
-						&newreq);
+		ret = nilfs_bmap_prepare_update_v(&direct->d_bmap, &oldreq,
+						  &newreq);
 		if (ret < 0)
 			return ret;
-		nilfs_bmap_commit_update(&direct->d_bmap, &oldreq, &newreq);
+		nilfs_bmap_commit_update_v(&direct->d_bmap, &oldreq, &newreq);
 		set_buffer_nilfs_volatile(bh);
 		nilfs_direct_set_ptr(direct, key, newreq.bpr_ptr);
 	} else
@@ -309,12 +339,10 @@ static int nilfs_direct_propagate_v(struct nilfs_direct *direct,
 static int nilfs_direct_propagate(const struct nilfs_bmap *bmap,
 				  struct buffer_head *bh)
 {
-	struct nilfs_direct *direct;
+	struct nilfs_direct *direct = (struct nilfs_direct *)bmap;
 
-	direct = (struct nilfs_direct *)bmap;
-	return (direct->d_ops->dop_propagate != NULL) ?
-		direct->d_ops->dop_propagate(direct, bh) :
-		0;
+	return NILFS_BMAP_USE_VBN(bmap) ?
+		nilfs_direct_propagate_v(direct, bh) : 0;
 }
 
 static int nilfs_direct_assign_v(struct nilfs_direct *direct,
@@ -327,12 +355,9 @@ static int nilfs_direct_assign_v(struct nilfs_direct *direct,
 	int ret;
 
 	req.bpr_ptr = ptr;
-	ret = direct->d_bmap.b_pops->bpop_prepare_start_ptr(
-		&direct->d_bmap, &req);
-	if (ret < 0)
+	ret = nilfs_bmap_start_v(&direct->d_bmap, &req, blocknr);
+	if (unlikely(ret < 0))
 		return ret;
-	direct->d_bmap.b_pops->bpop_commit_start_ptr(&direct->d_bmap,
-						     &req, blocknr);
 
 	binfo->bi_v.bi_vblocknr = nilfs_bmap_ptr_to_dptr(ptr);
 	binfo->bi_v.bi_blkoff = nilfs_bmap_key_to_dkey(key);
@@ -377,12 +402,14 @@ static int nilfs_direct_assign(struct nilfs_bmap *bmap,
 		return -EINVAL;
 	}
 
-	return direct->d_ops->dop_assign(direct, key, ptr, bh,
-					 blocknr, binfo);
+	return NILFS_BMAP_USE_VBN(bmap) ?
+		nilfs_direct_assign_v(direct, key, ptr, bh, blocknr, binfo) :
+		nilfs_direct_assign_p(direct, key, ptr, bh, blocknr, binfo);
 }
 
 static const struct nilfs_bmap_operations nilfs_direct_ops = {
 	.bop_lookup		=	nilfs_direct_lookup,
+	.bop_lookup_contig	=	nilfs_direct_lookup_contig,
 	.bop_insert		=	nilfs_direct_insert,
 	.bop_delete		=	nilfs_direct_delete,
 	.bop_clear		=	NULL,
@@ -401,36 +428,8 @@ static const struct nilfs_bmap_operations nilfs_direct_ops = {
 };
 
 
-static const struct nilfs_direct_operations nilfs_direct_ops_v = {
-	.dop_find_target	=	nilfs_direct_find_target_v,
-	.dop_set_target		=	nilfs_direct_set_target_v,
-	.dop_propagate		=	nilfs_direct_propagate_v,
-	.dop_assign		=	nilfs_direct_assign_v,
-};
-
-static const struct nilfs_direct_operations nilfs_direct_ops_p = {
-	.dop_find_target	=	NULL,
-	.dop_set_target		=	NULL,
-	.dop_propagate		=	NULL,
-	.dop_assign		=	nilfs_direct_assign_p,
-};
-
-int nilfs_direct_init(struct nilfs_bmap *bmap, __u64 low, __u64 high)
+int nilfs_direct_init(struct nilfs_bmap *bmap)
 {
-	struct nilfs_direct *direct;
-
-	direct = (struct nilfs_direct *)bmap;
 	bmap->b_ops = &nilfs_direct_ops;
-	bmap->b_low = low;
-	bmap->b_high = high;
-	switch (bmap->b_inode->i_ino) {
-	case NILFS_DAT_INO:
-		direct->d_ops = &nilfs_direct_ops_p;
-		break;
-	default:
-		direct->d_ops = &nilfs_direct_ops_v;
-		break;
-	}
-
 	return 0;
 }
diff --git a/fs/nilfs2/direct.h b/fs/nilfs2/direct.h
index 45d2c5cda812..a5ffd66e25d0 100644
--- a/fs/nilfs2/direct.h
+++ b/fs/nilfs2/direct.h
@@ -31,18 +31,6 @@
 struct nilfs_direct;
 
 /**
- * struct nilfs_direct_operations - direct mapping operation table
- */
-struct nilfs_direct_operations {
-	__u64 (*dop_find_target)(const struct nilfs_direct *, __u64);
-	void (*dop_set_target)(struct nilfs_direct *, __u64, __u64);
-	int (*dop_propagate)(struct nilfs_direct *, struct buffer_head *);
-	int (*dop_assign)(struct nilfs_direct *, __u64, __u64,
-			  struct buffer_head **, sector_t,
-			  union nilfs_binfo *);
-};
-
-/**
  * struct nilfs_direct_node - direct node
  * @dn_flags: flags
  * @dn_pad: padding
@@ -55,13 +43,9 @@ struct nilfs_direct_node {
 /**
  * struct nilfs_direct - direct mapping
  * @d_bmap: bmap structure
- * @d_ops: direct mapping operation table
  */
 struct nilfs_direct {
 	struct nilfs_bmap d_bmap;
-
-	/* direct-mapping-specific members */
-	const struct nilfs_direct_operations *d_ops;
 };
 
 
@@ -70,9 +54,9 @@ struct nilfs_direct {
 #define NILFS_DIRECT_KEY_MAX	(NILFS_DIRECT_NBLOCKS - 1)
 
 
-int nilfs_direct_init(struct nilfs_bmap *, __u64, __u64);
+int nilfs_direct_init(struct nilfs_bmap *);
 int nilfs_direct_delete_and_convert(struct nilfs_bmap *, __u64, __u64 *,
-				    __u64 *, int, __u64, __u64);
+				    __u64 *, int);
 
 
 #endif	/* _NILFS_DIRECT_H */
diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c
index 19d2102b6a69..1b3c2bb20da9 100644
--- a/fs/nilfs2/gcinode.c
+++ b/fs/nilfs2/gcinode.c
@@ -52,8 +52,9 @@
 #include "dat.h"
 #include "ifile.h"
 
-static struct address_space_operations def_gcinode_aops = {};
-/* XXX need def_gcinode_iops/fops? */
+static struct address_space_operations def_gcinode_aops = {
+	.sync_page		= block_sync_page,
+};
 
 /*
  * nilfs_gccache_submit_read_data() - add data buffer and submit read request
diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c
index 49ab4a49bb4f..2696d6b513b7 100644
--- a/fs/nilfs2/inode.c
+++ b/fs/nilfs2/inode.c
@@ -43,22 +43,23 @@
  *
  * This function does not issue actual read request of the specified data
  * block. It is done by VFS.
- * Bulk read for direct-io is not supported yet. (should be supported)
  */
 int nilfs_get_block(struct inode *inode, sector_t blkoff,
 		    struct buffer_head *bh_result, int create)
 {
 	struct nilfs_inode_info *ii = NILFS_I(inode);
-	unsigned long blknum = 0;
+	__u64 blknum = 0;
 	int err = 0, ret;
 	struct inode *dat = nilfs_dat_inode(NILFS_I_NILFS(inode));
+	unsigned maxblocks = bh_result->b_size >> inode->i_blkbits;
 
-	/* This exclusion control is a workaround; should be revised */
-	down_read(&NILFS_MDT(dat)->mi_sem);	/* XXX */
-	ret = nilfs_bmap_lookup(ii->i_bmap, (unsigned long)blkoff, &blknum);
-	up_read(&NILFS_MDT(dat)->mi_sem);	/* XXX */
-	if (ret == 0) {	/* found */
+	down_read(&NILFS_MDT(dat)->mi_sem);
+	ret = nilfs_bmap_lookup_contig(ii->i_bmap, blkoff, &blknum, maxblocks);
+	up_read(&NILFS_MDT(dat)->mi_sem);
+	if (ret >= 0) {	/* found */
 		map_bh(bh_result, inode->i_sb, blknum);
+		if (ret > 0)
+			bh_result->b_size = (ret << inode->i_blkbits);
 		goto out;
 	}
 	/* data block was not found */
@@ -240,7 +241,7 @@ nilfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 struct address_space_operations nilfs_aops = {
 	.writepage		= nilfs_writepage,
 	.readpage		= nilfs_readpage,
-	/* .sync_page		= nilfs_sync_page, */
+	.sync_page		= block_sync_page,
 	.writepages		= nilfs_writepages,
 	.set_page_dirty		= nilfs_set_page_dirty,
 	.readpages		= nilfs_readpages,
@@ -249,6 +250,7 @@ struct address_space_operations nilfs_aops = {
 	/* .releasepage		= nilfs_releasepage, */
 	.invalidatepage		= block_invalidatepage,
 	.direct_IO		= nilfs_direct_IO,
+	.is_partially_uptodate  = block_is_partially_uptodate,
 };
 
 struct inode *nilfs_new_inode(struct inode *dir, int mode)
diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c
index d6759b92006f..6ea5f872e2de 100644
--- a/fs/nilfs2/ioctl.c
+++ b/fs/nilfs2/ioctl.c
@@ -152,7 +152,7 @@ nilfs_ioctl_do_get_cpinfo(struct the_nilfs *nilfs, __u64 *posp, int flags,
 
 	down_read(&nilfs->ns_segctor_sem);
 	ret = nilfs_cpfile_get_cpinfo(nilfs->ns_cpfile, posp, flags, buf,
-				      nmembs);
+				      size, nmembs);
 	up_read(&nilfs->ns_segctor_sem);
 	return ret;
 }
@@ -182,7 +182,8 @@ nilfs_ioctl_do_get_suinfo(struct the_nilfs *nilfs, __u64 *posp, int flags,
 	int ret;
 
 	down_read(&nilfs->ns_segctor_sem);
-	ret = nilfs_sufile_get_suinfo(nilfs->ns_sufile, *posp, buf, nmembs);
+	ret = nilfs_sufile_get_suinfo(nilfs->ns_sufile, *posp, buf, size,
+				      nmembs);
 	up_read(&nilfs->ns_segctor_sem);
 	return ret;
 }
@@ -212,7 +213,7 @@ nilfs_ioctl_do_get_vinfo(struct the_nilfs *nilfs, __u64 *posp, int flags,
 	int ret;
 
 	down_read(&nilfs->ns_segctor_sem);
-	ret = nilfs_dat_get_vinfo(nilfs_dat_inode(nilfs), buf, nmembs);
+	ret = nilfs_dat_get_vinfo(nilfs_dat_inode(nilfs), buf, size, nmembs);
 	up_read(&nilfs->ns_segctor_sem);
 	return ret;
 }
@@ -435,24 +436,6 @@ static int nilfs_ioctl_mark_blocks_dirty(struct the_nilfs *nilfs,
 	return nmembs;
 }
 
-static int nilfs_ioctl_free_segments(struct the_nilfs *nilfs,
-				     struct nilfs_argv *argv, void *buf)
-{
-	size_t nmembs = argv->v_nmembs;
-	struct nilfs_sb_info *sbi = nilfs->ns_writer;
-	int ret;
-
-	if (unlikely(!sbi)) {
-		/* never happens because called for a writable mount */
-		WARN_ON(1);
-		return -EROFS;
-	}
-	ret = nilfs_segctor_add_segments_to_be_freed(
-		NILFS_SC(sbi), buf, nmembs);
-
-	return (ret < 0) ? ret : nmembs;
-}
-
 int nilfs_ioctl_prepare_clean_segments(struct the_nilfs *nilfs,
 				       struct nilfs_argv *argv, void **kbufs)
 {
@@ -491,14 +474,6 @@ int nilfs_ioctl_prepare_clean_segments(struct the_nilfs *nilfs,
 		msg = "cannot mark copying blocks dirty";
 		goto failed;
 	}
-	ret = nilfs_ioctl_free_segments(nilfs, &argv[4], kbufs[4]);
-	if (ret < 0) {
-		/*
-		 * can safely abort because this operation is atomic.
-		 */
-		msg = "cannot set segments to be freed";
-		goto failed;
-	}
 	return 0;
 
  failed:
@@ -615,7 +590,7 @@ static int nilfs_ioctl_get_info(struct inode *inode, struct file *filp,
 	if (copy_from_user(&argv, argp, sizeof(argv)))
 		return -EFAULT;
 
-	if (argv.v_size != membsz)
+	if (argv.v_size < membsz)
 		return -EINVAL;
 
 	ret = nilfs_ioctl_wrap_copy(nilfs, &argv, _IOC_DIR(cmd), dofunc);
diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index bb78745a0e30..3d3ddb3f5177 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -430,6 +430,7 @@ nilfs_mdt_write_page(struct page *page, struct writeback_control *wbc)
 
 static struct address_space_operations def_mdt_aops = {
 	.writepage		= nilfs_mdt_write_page,
+	.sync_page		= block_sync_page,
 };
 
 static struct inode_operations def_mdt_iops;
@@ -449,7 +450,7 @@ struct inode *
 nilfs_mdt_new_common(struct the_nilfs *nilfs, struct super_block *sb,
 		     ino_t ino, gfp_t gfp_mask)
 {
-	struct inode *inode = nilfs_alloc_inode(sb);
+	struct inode *inode = nilfs_alloc_inode_common(nilfs);
 
 	if (!inode)
 		return NULL;
diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h
index da6fc0bba2e5..edf6a59d9f2a 100644
--- a/fs/nilfs2/nilfs.h
+++ b/fs/nilfs2/nilfs.h
@@ -263,6 +263,7 @@ extern void nilfs_dirty_inode(struct inode *);
 extern struct dentry *nilfs_get_parent(struct dentry *);
 
 /* super.c */
+extern struct inode *nilfs_alloc_inode_common(struct the_nilfs *);
 extern struct inode *nilfs_alloc_inode(struct super_block *);
 extern void nilfs_destroy_inode(struct inode *);
 extern void nilfs_error(struct super_block *, const char *, const char *, ...)
diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c
index 57afa9d24061..d80cc71be749 100644
--- a/fs/nilfs2/recovery.c
+++ b/fs/nilfs2/recovery.c
@@ -28,7 +28,6 @@
 #include "segment.h"
 #include "sufile.h"
 #include "page.h"
-#include "seglist.h"
 #include "segbuf.h"
 
 /*
@@ -395,6 +394,24 @@ static void dispose_recovery_list(struct list_head *head)
 	}
 }
 
+struct nilfs_segment_entry {
+	struct list_head	list;
+	__u64			segnum;
+};
+
+static int nilfs_segment_list_add(struct list_head *head, __u64 segnum)
+{
+	struct nilfs_segment_entry *ent = kmalloc(sizeof(*ent), GFP_NOFS);
+
+	if (unlikely(!ent))
+		return -ENOMEM;
+
+	ent->segnum = segnum;
+	INIT_LIST_HEAD(&ent->list);
+	list_add_tail(&ent->list, head);
+	return 0;
+}
+
 void nilfs_dispose_segment_list(struct list_head *head)
 {
 	while (!list_empty(head)) {
@@ -402,7 +419,7 @@ void nilfs_dispose_segment_list(struct list_head *head)
 			= list_entry(head->next,
 				     struct nilfs_segment_entry, list);
 		list_del(&ent->list);
-		nilfs_free_segment_entry(ent);
+		kfree(ent);
 	}
 }
 
@@ -431,12 +448,10 @@ static int nilfs_prepare_segment_for_recovery(struct the_nilfs *nilfs,
 	if (unlikely(err))
 		goto failed;
 
-	err = -ENOMEM;
 	for (i = 1; i < 4; i++) {
-		ent = nilfs_alloc_segment_entry(segnum[i]);
-		if (unlikely(!ent))
+		err = nilfs_segment_list_add(head, segnum[i]);
+		if (unlikely(err))
 			goto failed;
-		list_add_tail(&ent->list, head);
 	}
 
 	/*
@@ -450,7 +465,7 @@ static int nilfs_prepare_segment_for_recovery(struct the_nilfs *nilfs,
 				goto failed;
 		}
 		list_del(&ent->list);
-		nilfs_free_segment_entry(ent);
+		kfree(ent);
 	}
 
 	/* Allocate new segments for recovery */
@@ -791,7 +806,6 @@ int nilfs_search_super_root(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi,
 	u64 seg_seq;
 	__u64 segnum, nextnum = 0;
 	__u64 cno;
-	struct nilfs_segment_entry *ent;
 	LIST_HEAD(segments);
 	int empty_seg = 0, scan_newer = 0;
 	int ret;
@@ -892,12 +906,9 @@ int nilfs_search_super_root(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi,
 		if (empty_seg++)
 			goto super_root_found; /* found a valid super root */
 
-		ent = nilfs_alloc_segment_entry(segnum);
-		if (unlikely(!ent)) {
-			ret = -ENOMEM;
+		ret = nilfs_segment_list_add(&segments, segnum);
+		if (unlikely(ret))
 			goto failed;
-		}
-		list_add_tail(&ent->list, &segments);
 
 		seg_seq++;
 		segnum = nextnum;
diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c
index 1e68821b4a9b..9e3fe17bb96b 100644
--- a/fs/nilfs2/segbuf.c
+++ b/fs/nilfs2/segbuf.c
@@ -26,7 +26,6 @@
 #include <linux/crc32.h>
 #include "page.h"
 #include "segbuf.h"
-#include "seglist.h"
 
 
 static struct kmem_cache *nilfs_segbuf_cachep;
@@ -394,7 +393,7 @@ int nilfs_segbuf_write(struct nilfs_segment_buffer *segbuf,
 		 * Last BIO is always sent through the following
 		 * submission.
 		 */
-		rw |= (1 << BIO_RW_SYNCIO);
+		rw |= (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG);
 		res = nilfs_submit_seg_bio(wi, rw);
 		if (unlikely(res))
 			goto failed_bio;
diff --git a/fs/nilfs2/seglist.h b/fs/nilfs2/seglist.h
deleted file mode 100644
index d39df9144e99..000000000000
--- a/fs/nilfs2/seglist.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * seglist.h - expediential structure and routines to handle list of segments
- *             (would be removed in a future release)
- *
- * Copyright (C) 2005-2008 Nippon Telegraph and Telephone Corporation.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- *
- * Written by Ryusuke Konishi <ryusuke@osrg.net>
- *
- */
-#ifndef _NILFS_SEGLIST_H
-#define _NILFS_SEGLIST_H
-
-#include <linux/fs.h>
-#include <linux/buffer_head.h>
-#include <linux/nilfs2_fs.h>
-#include "sufile.h"
-
-struct nilfs_segment_entry {
-	__u64			segnum;
-
-#define NILFS_SLH_FREED		0x0001	/* The segment was freed provisonally.
-					   It must be cancelled if
-					   construction aborted */
-
-	unsigned		flags;
-	struct list_head	list;
-	struct buffer_head     *bh_su;
-	struct nilfs_segment_usage *raw_su;
-};
-
-
-void nilfs_dispose_segment_list(struct list_head *);
-
-static inline struct nilfs_segment_entry *
-nilfs_alloc_segment_entry(__u64 segnum)
-{
-	struct nilfs_segment_entry *ent = kmalloc(sizeof(*ent), GFP_NOFS);
-
-	if (likely(ent)) {
-		ent->segnum = segnum;
-		ent->flags = 0;
-		ent->bh_su = NULL;
-		ent->raw_su = NULL;
-		INIT_LIST_HEAD(&ent->list);
-	}
-	return ent;
-}
-
-static inline int nilfs_open_segment_entry(struct nilfs_segment_entry *ent,
-					   struct inode *sufile)
-{
-	return nilfs_sufile_get_segment_usage(sufile, ent->segnum,
-					      &ent->raw_su, &ent->bh_su);
-}
-
-static inline void nilfs_close_segment_entry(struct nilfs_segment_entry *ent,
-					     struct inode *sufile)
-{
-	if (!ent->bh_su)
-		return;
-	nilfs_sufile_put_segment_usage(sufile, ent->segnum, ent->bh_su);
-	ent->bh_su = NULL;
-	ent->raw_su = NULL;
-}
-
-static inline void nilfs_free_segment_entry(struct nilfs_segment_entry *ent)
-{
-	kfree(ent);
-}
-
-#endif /* _NILFS_SEGLIST_H */
diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 22c7f65c2403..aa977549919e 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -39,7 +39,6 @@
 #include "sufile.h"
 #include "cpfile.h"
 #include "ifile.h"
-#include "seglist.h"
 #include "segbuf.h"
 
 
@@ -79,7 +78,8 @@ enum {
 /* State flags of collection */
 #define NILFS_CF_NODE		0x0001	/* Collecting node blocks */
 #define NILFS_CF_IFILE_STARTED	0x0002	/* IFILE stage has started */
-#define NILFS_CF_HISTORY_MASK	(NILFS_CF_IFILE_STARTED)
+#define NILFS_CF_SUFREED	0x0004	/* segment usages has been freed */
+#define NILFS_CF_HISTORY_MASK	(NILFS_CF_IFILE_STARTED | NILFS_CF_SUFREED)
 
 /* Operations depending on the construction mode and file type */
 struct nilfs_sc_operations {
@@ -810,7 +810,7 @@ static int nilfs_segctor_clean(struct nilfs_sc_info *sci)
 {
 	return list_empty(&sci->sc_dirty_files) &&
 		!test_bit(NILFS_SC_DIRTY, &sci->sc_flags) &&
-		list_empty(&sci->sc_cleaning_segments) &&
+		sci->sc_nfreesegs == 0 &&
 		(!nilfs_doing_gc() || list_empty(&sci->sc_gc_inodes));
 }
 
@@ -1005,44 +1005,6 @@ static void nilfs_drop_collected_inodes(struct list_head *head)
 	}
 }
 
-static void nilfs_segctor_cancel_free_segments(struct nilfs_sc_info *sci,
-					       struct inode *sufile)
-
-{
-	struct list_head *head = &sci->sc_cleaning_segments;
-	struct nilfs_segment_entry *ent;
-	int err;
-
-	list_for_each_entry(ent, head, list) {
-		if (!(ent->flags & NILFS_SLH_FREED))
-			break;
-		err = nilfs_sufile_cancel_free(sufile, ent->segnum);
-		WARN_ON(err); /* do not happen */
-		ent->flags &= ~NILFS_SLH_FREED;
-	}
-}
-
-static int nilfs_segctor_prepare_free_segments(struct nilfs_sc_info *sci,
-					       struct inode *sufile)
-{
-	struct list_head *head = &sci->sc_cleaning_segments;
-	struct nilfs_segment_entry *ent;
-	int err;
-
-	list_for_each_entry(ent, head, list) {
-		err = nilfs_sufile_free(sufile, ent->segnum);
-		if (unlikely(err))
-			return err;
-		ent->flags |= NILFS_SLH_FREED;
-	}
-	return 0;
-}
-
-static void nilfs_segctor_commit_free_segments(struct nilfs_sc_info *sci)
-{
-	nilfs_dispose_segment_list(&sci->sc_cleaning_segments);
-}
-
 static int nilfs_segctor_apply_buffers(struct nilfs_sc_info *sci,
 				       struct inode *inode,
 				       struct list_head *listp,
@@ -1161,6 +1123,7 @@ static int nilfs_segctor_collect_blocks(struct nilfs_sc_info *sci, int mode)
 	struct the_nilfs *nilfs = sbi->s_nilfs;
 	struct list_head *head;
 	struct nilfs_inode_info *ii;
+	size_t ndone;
 	int err = 0;
 
 	switch (sci->sc_stage.scnt) {
@@ -1250,10 +1213,16 @@ static int nilfs_segctor_collect_blocks(struct nilfs_sc_info *sci, int mode)
 			break;
 		sci->sc_stage.scnt++;  /* Fall through */
 	case NILFS_ST_SUFILE:
-		err = nilfs_segctor_prepare_free_segments(sci,
-							  nilfs->ns_sufile);
-		if (unlikely(err))
+		err = nilfs_sufile_freev(nilfs->ns_sufile, sci->sc_freesegs,
+					 sci->sc_nfreesegs, &ndone);
+		if (unlikely(err)) {
+			nilfs_sufile_cancel_freev(nilfs->ns_sufile,
+						  sci->sc_freesegs, ndone,
+						  NULL);
 			break;
+		}
+		sci->sc_stage.flags |= NILFS_CF_SUFREED;
+
 		err = nilfs_segctor_scan_file(sci, nilfs->ns_sufile,
 					      &nilfs_sc_file_ops);
 		if (unlikely(err))
@@ -1486,7 +1455,15 @@ static void nilfs_segctor_end_construction(struct nilfs_sc_info *sci,
 {
 	if (unlikely(err)) {
 		nilfs_segctor_free_incomplete_segments(sci, nilfs);
-		nilfs_segctor_cancel_free_segments(sci, nilfs->ns_sufile);
+		if (sci->sc_stage.flags & NILFS_CF_SUFREED) {
+			int ret;
+
+			ret = nilfs_sufile_cancel_freev(nilfs->ns_sufile,
+							sci->sc_freesegs,
+							sci->sc_nfreesegs,
+							NULL);
+			WARN_ON(ret); /* do not happen */
+		}
 	}
 	nilfs_segctor_clear_segment_buffers(sci);
 }
@@ -1585,7 +1562,13 @@ static int nilfs_segctor_collect(struct nilfs_sc_info *sci,
 		if (mode != SC_LSEG_SR || sci->sc_stage.scnt < NILFS_ST_CPFILE)
 			break;
 
-		nilfs_segctor_cancel_free_segments(sci, nilfs->ns_sufile);
+		if (sci->sc_stage.flags & NILFS_CF_SUFREED) {
+			err = nilfs_sufile_cancel_freev(nilfs->ns_sufile,
+							sci->sc_freesegs,
+							sci->sc_nfreesegs,
+							NULL);
+			WARN_ON(err); /* do not happen */
+		}
 		nilfs_segctor_clear_segment_buffers(sci);
 
 		err = nilfs_segctor_extend_segments(sci, nilfs, nadd);
@@ -2224,10 +2207,8 @@ static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode)
 		nilfs_segctor_complete_write(sci);
 
 		/* Commit segments */
-		if (has_sr) {
-			nilfs_segctor_commit_free_segments(sci);
+		if (has_sr)
 			nilfs_segctor_clear_metadata_dirty(sci);
-		}
 
 		nilfs_segctor_end_construction(sci, nilfs, 0);
 
@@ -2301,48 +2282,6 @@ void nilfs_flush_segment(struct super_block *sb, ino_t ino)
 					/* assign bit 0 to data files */
 }
 
-int nilfs_segctor_add_segments_to_be_freed(struct nilfs_sc_info *sci,
-					   __u64 *segnum, size_t nsegs)
-{
-	struct nilfs_segment_entry *ent;
-	struct the_nilfs *nilfs = sci->sc_sbi->s_nilfs;
-	struct inode *sufile = nilfs->ns_sufile;
-	LIST_HEAD(list);
-	__u64 *pnum;
-	size_t i;
-	int err;
-
-	for (pnum = segnum, i = 0; i < nsegs; pnum++, i++) {
-		ent = nilfs_alloc_segment_entry(*pnum);
-		if (unlikely(!ent)) {
-			err = -ENOMEM;
-			goto failed;
-		}
-		list_add_tail(&ent->list, &list);
-
-		err = nilfs_open_segment_entry(ent, sufile);
-		if (unlikely(err))
-			goto failed;
-
-		if (unlikely(!nilfs_segment_usage_dirty(ent->raw_su)))
-			printk(KERN_WARNING "NILFS: unused segment is "
-			       "requested to be cleaned (segnum=%llu)\n",
-			       (unsigned long long)ent->segnum);
-		nilfs_close_segment_entry(ent, sufile);
-	}
-	list_splice(&list, sci->sc_cleaning_segments.prev);
-	return 0;
-
- failed:
-	nilfs_dispose_segment_list(&list);
-	return err;
-}
-
-void nilfs_segctor_clear_segments_to_be_freed(struct nilfs_sc_info *sci)
-{
-	nilfs_dispose_segment_list(&sci->sc_cleaning_segments);
-}
-
 struct nilfs_segctor_wait_request {
 	wait_queue_t	wq;
 	__u32		seq;
@@ -2607,10 +2546,13 @@ int nilfs_clean_segments(struct super_block *sb, struct nilfs_argv *argv,
 	err = nilfs_init_gcdat_inode(nilfs);
 	if (unlikely(err))
 		goto out_unlock;
+
 	err = nilfs_ioctl_prepare_clean_segments(nilfs, argv, kbufs);
 	if (unlikely(err))
 		goto out_unlock;
 
+	sci->sc_freesegs = kbufs[4];
+	sci->sc_nfreesegs = argv[4].v_nmembs;
 	list_splice_init(&nilfs->ns_gc_inodes, sci->sc_gc_inodes.prev);
 
 	for (;;) {
@@ -2629,6 +2571,8 @@ int nilfs_clean_segments(struct super_block *sb, struct nilfs_argv *argv,
 	}
 
  out_unlock:
+	sci->sc_freesegs = NULL;
+	sci->sc_nfreesegs = 0;
 	nilfs_clear_gcdat_inode(nilfs);
 	nilfs_transaction_unlock(sbi);
 	return err;
@@ -2835,7 +2779,6 @@ static struct nilfs_sc_info *nilfs_segctor_new(struct nilfs_sb_info *sbi)
 	INIT_LIST_HEAD(&sci->sc_dirty_files);
 	INIT_LIST_HEAD(&sci->sc_segbufs);
 	INIT_LIST_HEAD(&sci->sc_gc_inodes);
-	INIT_LIST_HEAD(&sci->sc_cleaning_segments);
 	INIT_LIST_HEAD(&sci->sc_copied_buffers);
 
 	sci->sc_interval = HZ * NILFS_SC_DEFAULT_TIMEOUT;
@@ -2901,9 +2844,6 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci)
 		nilfs_dispose_list(sbi, &sci->sc_dirty_files, 1);
 	}
 
-	if (!list_empty(&sci->sc_cleaning_segments))
-		nilfs_dispose_segment_list(&sci->sc_cleaning_segments);
-
 	WARN_ON(!list_empty(&sci->sc_segbufs));
 
 	down_write(&sbi->s_nilfs->ns_segctor_sem);
diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h
index 476bdd5df5be..0d2a475a741b 100644
--- a/fs/nilfs2/segment.h
+++ b/fs/nilfs2/segment.h
@@ -90,8 +90,9 @@ struct nilfs_segsum_pointer {
  * @sc_nblk_inc: Block count of current generation
  * @sc_dirty_files: List of files to be written
  * @sc_gc_inodes: List of GC inodes having blocks to be written
- * @sc_cleaning_segments: List of segments to be freed through construction
  * @sc_copied_buffers: List of copied buffers (buffer heads) to freeze data
+ * @sc_freesegs: array of segment numbers to be freed
+ * @sc_nfreesegs: number of segments on @sc_freesegs
  * @sc_dsync_inode: inode whose data pages are written for a sync operation
  * @sc_dsync_start: start byte offset of data pages
  * @sc_dsync_end: end byte offset of data pages (inclusive)
@@ -131,9 +132,11 @@ struct nilfs_sc_info {
 
 	struct list_head	sc_dirty_files;
 	struct list_head	sc_gc_inodes;
-	struct list_head	sc_cleaning_segments;
 	struct list_head	sc_copied_buffers;
 
+	__u64		       *sc_freesegs;
+	size_t			sc_nfreesegs;
+
 	struct nilfs_inode_info *sc_dsync_inode;
 	loff_t			sc_dsync_start;
 	loff_t			sc_dsync_end;
@@ -225,10 +228,6 @@ extern void nilfs_flush_segment(struct super_block *, ino_t);
 extern int nilfs_clean_segments(struct super_block *, struct nilfs_argv *,
 				void **);
 
-extern int nilfs_segctor_add_segments_to_be_freed(struct nilfs_sc_info *,
-						  __u64 *, size_t);
-extern void nilfs_segctor_clear_segments_to_be_freed(struct nilfs_sc_info *);
-
 extern int nilfs_attach_segment_constructor(struct nilfs_sb_info *);
 extern void nilfs_detach_segment_constructor(struct nilfs_sb_info *);
 
@@ -240,5 +239,6 @@ extern int nilfs_search_super_root(struct the_nilfs *, struct nilfs_sb_info *,
 extern int nilfs_recover_logical_segments(struct the_nilfs *,
 					  struct nilfs_sb_info *,
 					  struct nilfs_recovery_info *);
+extern void nilfs_dispose_segment_list(struct list_head *);
 
 #endif /* _NILFS_SEGMENT_H */
diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c
index 98e68677f045..37994d4a59cc 100644
--- a/fs/nilfs2/sufile.c
+++ b/fs/nilfs2/sufile.c
@@ -18,6 +18,7 @@
  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  *
  * Written by Koji Sato <koji@osrg.net>.
+ * Rivised by Ryusuke Konishi <ryusuke@osrg.net>.
  */
 
 #include <linux/kernel.h>
@@ -108,6 +109,102 @@ static void nilfs_sufile_mod_counter(struct buffer_head *header_bh,
 	nilfs_mdt_mark_buffer_dirty(header_bh);
 }
 
+/**
+ * nilfs_sufile_updatev - modify multiple segment usages at a time
+ * @sufile: inode of segment usage file
+ * @segnumv: array of segment numbers
+ * @nsegs: size of @segnumv array
+ * @create: creation flag
+ * @ndone: place to store number of modified segments on @segnumv
+ * @dofunc: primitive operation for the update
+ *
+ * Description: nilfs_sufile_updatev() repeatedly calls @dofunc
+ * against the given array of segments.  The @dofunc is called with
+ * buffers of a header block and the sufile block in which the target
+ * segment usage entry is contained.  If @ndone is given, the number
+ * of successfully modified segments from the head is stored in the
+ * place @ndone points to.
+ *
+ * Return Value: On success, zero is returned.  On error, one of the
+ * following negative error codes is returned.
+ *
+ * %-EIO - I/O error.
+ *
+ * %-ENOMEM - Insufficient amount of memory available.
+ *
+ * %-ENOENT - Given segment usage is in hole block (may be returned if
+ *            @create is zero)
+ *
+ * %-EINVAL - Invalid segment usage number
+ */
+int nilfs_sufile_updatev(struct inode *sufile, __u64 *segnumv, size_t nsegs,
+			 int create, size_t *ndone,
+			 void (*dofunc)(struct inode *, __u64,
+					struct buffer_head *,
+					struct buffer_head *))
+{
+	struct buffer_head *header_bh, *bh;
+	unsigned long blkoff, prev_blkoff;
+	__u64 *seg;
+	size_t nerr = 0, n = 0;
+	int ret = 0;
+
+	if (unlikely(nsegs == 0))
+		goto out;
+
+	down_write(&NILFS_MDT(sufile)->mi_sem);
+	for (seg = segnumv; seg < segnumv + nsegs; seg++) {
+		if (unlikely(*seg >= nilfs_sufile_get_nsegments(sufile))) {
+			printk(KERN_WARNING
+			       "%s: invalid segment number: %llu\n", __func__,
+			       (unsigned long long)*seg);
+			nerr++;
+		}
+	}
+	if (nerr > 0) {
+		ret = -EINVAL;
+		goto out_sem;
+	}
+
+	ret = nilfs_sufile_get_header_block(sufile, &header_bh);
+	if (ret < 0)
+		goto out_sem;
+
+	seg = segnumv;
+	blkoff = nilfs_sufile_get_blkoff(sufile, *seg);
+	ret = nilfs_mdt_get_block(sufile, blkoff, create, NULL, &bh);
+	if (ret < 0)
+		goto out_header;
+
+	for (;;) {
+		dofunc(sufile, *seg, header_bh, bh);
+
+		if (++seg >= segnumv + nsegs)
+			break;
+		prev_blkoff = blkoff;
+		blkoff = nilfs_sufile_get_blkoff(sufile, *seg);
+		if (blkoff == prev_blkoff)
+			continue;
+
+		/* get different block */
+		brelse(bh);
+		ret = nilfs_mdt_get_block(sufile, blkoff, create, NULL, &bh);
+		if (unlikely(ret < 0))
+			goto out_header;
+	}
+	brelse(bh);
+
+ out_header:
+	n = seg - segnumv;
+	brelse(header_bh);
+ out_sem:
+	up_write(&NILFS_MDT(sufile)->mi_sem);
+ out:
+	if (ndone)
+		*ndone = n;
+	return ret;
+}
+
 int nilfs_sufile_update(struct inode *sufile, __u64 segnum, int create,
 			void (*dofunc)(struct inode *, __u64,
 				       struct buffer_head *,
@@ -490,7 +587,8 @@ void nilfs_sufile_do_set_error(struct inode *sufile, __u64 segnum,
  * nilfs_sufile_get_suinfo -
  * @sufile: inode of segment usage file
  * @segnum: segment number to start looking
- * @si: array of suinfo
+ * @buf: array of suinfo
+ * @sisz: byte size of suinfo
  * @nsi: size of suinfo array
  *
  * Description:
@@ -502,11 +600,12 @@ void nilfs_sufile_do_set_error(struct inode *sufile, __u64 segnum,
  *
  * %-ENOMEM - Insufficient amount of memory available.
  */
-ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum,
-				struct nilfs_suinfo *si, size_t nsi)
+ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf,
+				unsigned sisz, size_t nsi)
 {
 	struct buffer_head *su_bh;
 	struct nilfs_segment_usage *su;
+	struct nilfs_suinfo *si = buf;
 	size_t susz = NILFS_MDT(sufile)->mi_entry_size;
 	struct the_nilfs *nilfs = NILFS_MDT(sufile)->mi_nilfs;
 	void *kaddr;
@@ -531,20 +630,22 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum,
 			if (ret != -ENOENT)
 				goto out;
 			/* hole */
-			memset(&si[i], 0, sizeof(struct nilfs_suinfo) * n);
+			memset(si, 0, sisz * n);
+			si = (void *)si + sisz * n;
 			continue;
 		}
 
 		kaddr = kmap_atomic(su_bh->b_page, KM_USER0);
 		su = nilfs_sufile_block_get_segment_usage(
 			sufile, segnum, su_bh, kaddr);
-		for (j = 0; j < n; j++, su = (void *)su + susz) {
-			si[i + j].sui_lastmod = le64_to_cpu(su->su_lastmod);
-			si[i + j].sui_nblocks = le32_to_cpu(su->su_nblocks);
-			si[i + j].sui_flags = le32_to_cpu(su->su_flags) &
+		for (j = 0; j < n;
+		     j++, su = (void *)su + susz, si = (void *)si + sisz) {
+			si->sui_lastmod = le64_to_cpu(su->su_lastmod);
+			si->sui_nblocks = le32_to_cpu(su->su_nblocks);
+			si->sui_flags = le32_to_cpu(su->su_flags) &
 				~(1UL << NILFS_SEGMENT_USAGE_ACTIVE);
 			if (nilfs_segment_is_active(nilfs, segnum + j))
-				si[i + j].sui_flags |=
+				si->sui_flags |=
 					(1UL << NILFS_SEGMENT_USAGE_ACTIVE);
 		}
 		kunmap_atomic(kaddr, KM_USER0);
diff --git a/fs/nilfs2/sufile.h b/fs/nilfs2/sufile.h
index a2e2efd4ade1..a2c4d76c3366 100644
--- a/fs/nilfs2/sufile.h
+++ b/fs/nilfs2/sufile.h
@@ -43,43 +43,27 @@ void nilfs_sufile_put_segment_usage(struct inode *, __u64,
 				    struct buffer_head *);
 int nilfs_sufile_get_stat(struct inode *, struct nilfs_sustat *);
 int nilfs_sufile_get_ncleansegs(struct inode *, unsigned long *);
-ssize_t nilfs_sufile_get_suinfo(struct inode *, __u64, struct nilfs_suinfo *,
+ssize_t nilfs_sufile_get_suinfo(struct inode *, __u64, void *, unsigned,
 				size_t);
 
+int nilfs_sufile_updatev(struct inode *, __u64 *, size_t, int, size_t *,
+			 void (*dofunc)(struct inode *, __u64,
+					struct buffer_head *,
+					struct buffer_head *));
 int nilfs_sufile_update(struct inode *, __u64, int,
 			void (*dofunc)(struct inode *, __u64,
 				       struct buffer_head *,
 				       struct buffer_head *));
-void nilfs_sufile_do_cancel_free(struct inode *, __u64, struct buffer_head *,
-				 struct buffer_head *);
 void nilfs_sufile_do_scrap(struct inode *, __u64, struct buffer_head *,
 			   struct buffer_head *);
 void nilfs_sufile_do_free(struct inode *, __u64, struct buffer_head *,
 			  struct buffer_head *);
+void nilfs_sufile_do_cancel_free(struct inode *, __u64, struct buffer_head *,
+				 struct buffer_head *);
 void nilfs_sufile_do_set_error(struct inode *, __u64, struct buffer_head *,
 			       struct buffer_head *);
 
 /**
- * nilfs_sufile_cancel_free -
- * @sufile: inode of segment usage file
- * @segnum: segment number
- *
- * Description:
- *
- * Return Value: On success, 0 is returned. On error, one of the following
- * negative error codes is returned.
- *
- * %-EIO - I/O error.
- *
- * %-ENOMEM - Insufficient amount of memory available.
- */
-static inline int nilfs_sufile_cancel_free(struct inode *sufile, __u64 segnum)
-{
-	return nilfs_sufile_update(sufile, segnum, 0,
-				   nilfs_sufile_do_cancel_free);
-}
-
-/**
  * nilfs_sufile_scrap - make a segment garbage
  * @sufile: inode of segment usage file
  * @segnum: segment number to be freed
@@ -100,6 +84,38 @@ static inline int nilfs_sufile_free(struct inode *sufile, __u64 segnum)
 }
 
 /**
+ * nilfs_sufile_freev - free segments
+ * @sufile: inode of segment usage file
+ * @segnumv: array of segment numbers
+ * @nsegs: size of @segnumv array
+ * @ndone: place to store the number of freed segments
+ */
+static inline int nilfs_sufile_freev(struct inode *sufile, __u64 *segnumv,
+				     size_t nsegs, size_t *ndone)
+{
+	return nilfs_sufile_updatev(sufile, segnumv, nsegs, 0, ndone,
+				    nilfs_sufile_do_free);
+}
+
+/**
+ * nilfs_sufile_cancel_freev - reallocate freeing segments
+ * @sufile: inode of segment usage file
+ * @segnumv: array of segment numbers
+ * @nsegs: size of @segnumv array
+ * @ndone: place to store the number of cancelled segments
+ *
+ * Return Value: On success, 0 is returned. On error, a negative error codes
+ * is returned.
+ */
+static inline int nilfs_sufile_cancel_freev(struct inode *sufile,
+					    __u64 *segnumv, size_t nsegs,
+					    size_t *ndone)
+{
+	return nilfs_sufile_updatev(sufile, segnumv, nsegs, 0, ndone,
+				    nilfs_sufile_do_cancel_free);
+}
+
+/**
  * nilfs_sufile_set_error - mark a segment as erroneous
  * @sufile: inode of segment usage file
  * @segnum: segment number
diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c
index 1777a3467bd2..ab785f85aa50 100644
--- a/fs/nilfs2/super.c
+++ b/fs/nilfs2/super.c
@@ -133,7 +133,7 @@ void nilfs_warning(struct super_block *sb, const char *function,
 
 static struct kmem_cache *nilfs_inode_cachep;
 
-struct inode *nilfs_alloc_inode(struct super_block *sb)
+struct inode *nilfs_alloc_inode_common(struct the_nilfs *nilfs)
 {
 	struct nilfs_inode_info *ii;
 
@@ -143,10 +143,15 @@ struct inode *nilfs_alloc_inode(struct super_block *sb)
 	ii->i_bh = NULL;
 	ii->i_state = 0;
 	ii->vfs_inode.i_version = 1;
-	nilfs_btnode_cache_init(&ii->i_btnode_cache);
+	nilfs_btnode_cache_init(&ii->i_btnode_cache, nilfs->ns_bdi);
 	return &ii->vfs_inode;
 }
 
+struct inode *nilfs_alloc_inode(struct super_block *sb)
+{
+	return nilfs_alloc_inode_common(NILFS_SB(sb)->s_nilfs);
+}
+
 void nilfs_destroy_inode(struct inode *inode)
 {
 	kmem_cache_free(nilfs_inode_cachep, NILFS_I(inode));
diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c
index e4e5c78bcc93..8b8889825716 100644
--- a/fs/nilfs2/the_nilfs.c
+++ b/fs/nilfs2/the_nilfs.c
@@ -32,7 +32,6 @@
 #include "cpfile.h"
 #include "sufile.h"
 #include "dat.h"
-#include "seglist.h"
 #include "segbuf.h"
 
 
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 678a067d9251..9edcde4974aa 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -475,6 +475,12 @@ struct ocfs2_path {
 #define path_leaf_el(_path) ((_path)->p_node[(_path)->p_tree_depth].el)
 #define path_num_items(_path) ((_path)->p_tree_depth + 1)
 
+static int ocfs2_find_path(struct inode *inode, struct ocfs2_path *path,
+			   u32 cpos);
+static void ocfs2_adjust_rightmost_records(struct inode *inode,
+					   handle_t *handle,
+					   struct ocfs2_path *path,
+					   struct ocfs2_extent_rec *insert_rec);
 /*
  * Reset the actual path elements so that we can re-use the structure
  * to build another path. Generally, this involves freeing the buffer
@@ -1013,6 +1019,54 @@ static inline u32 ocfs2_sum_rightmost_rec(struct ocfs2_extent_list  *el)
 }
 
 /*
+ * Change range of the branches in the right most path according to the leaf
+ * extent block's rightmost record.
+ */
+static int ocfs2_adjust_rightmost_branch(handle_t *handle,
+					 struct inode *inode,
+					 struct ocfs2_extent_tree *et)
+{
+	int status;
+	struct ocfs2_path *path = NULL;
+	struct ocfs2_extent_list *el;
+	struct ocfs2_extent_rec *rec;
+
+	path = ocfs2_new_path_from_et(et);
+	if (!path) {
+		status = -ENOMEM;
+		return status;
+	}
+
+	status = ocfs2_find_path(inode, path, UINT_MAX);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	status = ocfs2_extend_trans(handle, path_num_items(path) +
+				    handle->h_buffer_credits);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	status = ocfs2_journal_access_path(inode, handle, path);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	el = path_leaf_el(path);
+	rec = &el->l_recs[le32_to_cpu(el->l_next_free_rec) - 1];
+
+	ocfs2_adjust_rightmost_records(inode, handle, path, rec);
+
+out:
+	ocfs2_free_path(path);
+	return status;
+}
+
+/*
  * Add an entire tree branch to our inode. eb_bh is the extent block
  * to start at, if we don't want to start the branch at the dinode
  * structure.
@@ -1038,7 +1092,7 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 	struct ocfs2_extent_block *eb;
 	struct ocfs2_extent_list  *eb_el;
 	struct ocfs2_extent_list  *el;
-	u32 new_cpos;
+	u32 new_cpos, root_end;
 
 	mlog_entry_void();
 
@@ -1055,6 +1109,27 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 
 	new_blocks = le16_to_cpu(el->l_tree_depth);
 
+	eb = (struct ocfs2_extent_block *)(*last_eb_bh)->b_data;
+	new_cpos = ocfs2_sum_rightmost_rec(&eb->h_list);
+	root_end = ocfs2_sum_rightmost_rec(et->et_root_el);
+
+	/*
+	 * If there is a gap before the root end and the real end
+	 * of the righmost leaf block, we need to remove the gap
+	 * between new_cpos and root_end first so that the tree
+	 * is consistent after we add a new branch(it will start
+	 * from new_cpos).
+	 */
+	if (root_end > new_cpos) {
+		mlog(0, "adjust the cluster end from %u to %u\n",
+		     root_end, new_cpos);
+		status = ocfs2_adjust_rightmost_branch(handle, inode, et);
+		if (status) {
+			mlog_errno(status);
+			goto bail;
+		}
+	}
+
 	/* allocate the number of new eb blocks we need */
 	new_eb_bhs = kcalloc(new_blocks, sizeof(struct buffer_head *),
 			     GFP_KERNEL);
@@ -1071,9 +1146,6 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	eb = (struct ocfs2_extent_block *)(*last_eb_bh)->b_data;
-	new_cpos = ocfs2_sum_rightmost_rec(&eb->h_list);
-
 	/* Note: new_eb_bhs[new_blocks - 1] is the guy which will be
 	 * linked with the rest of the tree.
 	 * conversly, new_eb_bhs[0] is the new bottommost leaf.
diff --git a/fs/ocfs2/blockcheck.c b/fs/ocfs2/blockcheck.c
index 2a947c44e594..a1163b8b417c 100644
--- a/fs/ocfs2/blockcheck.c
+++ b/fs/ocfs2/blockcheck.c
@@ -22,6 +22,9 @@
 #include <linux/crc32.h>
 #include <linux/buffer_head.h>
 #include <linux/bitops.h>
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/fs.h>
 #include <asm/byteorder.h>
 
 #include <cluster/masklog.h>
@@ -222,6 +225,155 @@ void ocfs2_hamming_fix_block(void *data, unsigned int blocksize,
 	ocfs2_hamming_fix(data, blocksize * 8, 0, fix);
 }
 
+
+/*
+ * Debugfs handling.
+ */
+
+#ifdef CONFIG_DEBUG_FS
+
+static int blockcheck_u64_get(void *data, u64 *val)
+{
+	*val = *(u64 *)data;
+	return 0;
+}
+DEFINE_SIMPLE_ATTRIBUTE(blockcheck_fops, blockcheck_u64_get, NULL, "%llu\n");
+
+static struct dentry *blockcheck_debugfs_create(const char *name,
+						struct dentry *parent,
+						u64 *value)
+{
+	return debugfs_create_file(name, S_IFREG | S_IRUSR, parent, value,
+				   &blockcheck_fops);
+}
+
+static void ocfs2_blockcheck_debug_remove(struct ocfs2_blockcheck_stats *stats)
+{
+	if (stats) {
+		debugfs_remove(stats->b_debug_check);
+		stats->b_debug_check = NULL;
+		debugfs_remove(stats->b_debug_failure);
+		stats->b_debug_failure = NULL;
+		debugfs_remove(stats->b_debug_recover);
+		stats->b_debug_recover = NULL;
+		debugfs_remove(stats->b_debug_dir);
+		stats->b_debug_dir = NULL;
+	}
+}
+
+static int ocfs2_blockcheck_debug_install(struct ocfs2_blockcheck_stats *stats,
+					  struct dentry *parent)
+{
+	int rc = -EINVAL;
+
+	if (!stats)
+		goto out;
+
+	stats->b_debug_dir = debugfs_create_dir("blockcheck", parent);
+	if (!stats->b_debug_dir)
+		goto out;
+
+	stats->b_debug_check =
+		blockcheck_debugfs_create("blocks_checked",
+					  stats->b_debug_dir,
+					  &stats->b_check_count);
+
+	stats->b_debug_failure =
+		blockcheck_debugfs_create("checksums_failed",
+					  stats->b_debug_dir,
+					  &stats->b_failure_count);
+
+	stats->b_debug_recover =
+		blockcheck_debugfs_create("ecc_recoveries",
+					  stats->b_debug_dir,
+					  &stats->b_recover_count);
+	if (stats->b_debug_check && stats->b_debug_failure &&
+	    stats->b_debug_recover)
+		rc = 0;
+
+out:
+	if (rc)
+		ocfs2_blockcheck_debug_remove(stats);
+	return rc;
+}
+#else
+static inline int ocfs2_blockcheck_debug_install(struct ocfs2_blockcheck_stats *stats,
+						 struct dentry *parent)
+{
+	return 0;
+}
+
+static inline void ocfs2_blockcheck_debug_remove(struct ocfs2_blockcheck_stats *stats)
+{
+}
+#endif  /* CONFIG_DEBUG_FS */
+
+/* Always-called wrappers for starting and stopping the debugfs files */
+int ocfs2_blockcheck_stats_debugfs_install(struct ocfs2_blockcheck_stats *stats,
+					   struct dentry *parent)
+{
+	return ocfs2_blockcheck_debug_install(stats, parent);
+}
+
+void ocfs2_blockcheck_stats_debugfs_remove(struct ocfs2_blockcheck_stats *stats)
+{
+	ocfs2_blockcheck_debug_remove(stats);
+}
+
+static void ocfs2_blockcheck_inc_check(struct ocfs2_blockcheck_stats *stats)
+{
+	u64 new_count;
+
+	if (!stats)
+		return;
+
+	spin_lock(&stats->b_lock);
+	stats->b_check_count++;
+	new_count = stats->b_check_count;
+	spin_unlock(&stats->b_lock);
+
+	if (!new_count)
+		mlog(ML_NOTICE, "Block check count has wrapped\n");
+}
+
+static void ocfs2_blockcheck_inc_failure(struct ocfs2_blockcheck_stats *stats)
+{
+	u64 new_count;
+
+	if (!stats)
+		return;
+
+	spin_lock(&stats->b_lock);
+	stats->b_failure_count++;
+	new_count = stats->b_failure_count;
+	spin_unlock(&stats->b_lock);
+
+	if (!new_count)
+		mlog(ML_NOTICE, "Checksum failure count has wrapped\n");
+}
+
+static void ocfs2_blockcheck_inc_recover(struct ocfs2_blockcheck_stats *stats)
+{
+	u64 new_count;
+
+	if (!stats)
+		return;
+
+	spin_lock(&stats->b_lock);
+	stats->b_recover_count++;
+	new_count = stats->b_recover_count;
+	spin_unlock(&stats->b_lock);
+
+	if (!new_count)
+		mlog(ML_NOTICE, "ECC recovery count has wrapped\n");
+}
+
+
+
+/*
+ * These are the low-level APIs for using the ocfs2_block_check structure.
+ */
+
 /*
  * This function generates check information for a block.
  * data is the block to be checked.  bc is a pointer to the
@@ -266,12 +418,15 @@ void ocfs2_block_check_compute(void *data, size_t blocksize,
  * Again, the data passed in should be the on-disk endian.
  */
 int ocfs2_block_check_validate(void *data, size_t blocksize,
-			       struct ocfs2_block_check *bc)
+			       struct ocfs2_block_check *bc,
+			       struct ocfs2_blockcheck_stats *stats)
 {
 	int rc = 0;
 	struct ocfs2_block_check check;
 	u32 crc, ecc;
 
+	ocfs2_blockcheck_inc_check(stats);
+
 	check.bc_crc32e = le32_to_cpu(bc->bc_crc32e);
 	check.bc_ecc = le16_to_cpu(bc->bc_ecc);
 
@@ -282,6 +437,7 @@ int ocfs2_block_check_validate(void *data, size_t blocksize,
 	if (crc == check.bc_crc32e)
 		goto out;
 
+	ocfs2_blockcheck_inc_failure(stats);
 	mlog(ML_ERROR,
 	     "CRC32 failed: stored: %u, computed %u.  Applying ECC.\n",
 	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
@@ -292,8 +448,10 @@ int ocfs2_block_check_validate(void *data, size_t blocksize,
 
 	/* And check the crc32 again */
 	crc = crc32_le(~0, data, blocksize);
-	if (crc == check.bc_crc32e)
+	if (crc == check.bc_crc32e) {
+		ocfs2_blockcheck_inc_recover(stats);
 		goto out;
+	}
 
 	mlog(ML_ERROR, "Fixed CRC32 failed: stored: %u, computed %u\n",
 	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
@@ -366,7 +524,8 @@ void ocfs2_block_check_compute_bhs(struct buffer_head **bhs, int nr,
  * Again, the data passed in should be the on-disk endian.
  */
 int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr,
-				   struct ocfs2_block_check *bc)
+				   struct ocfs2_block_check *bc,
+				   struct ocfs2_blockcheck_stats *stats)
 {
 	int i, rc = 0;
 	struct ocfs2_block_check check;
@@ -377,6 +536,8 @@ int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr,
 	if (!nr)
 		return 0;
 
+	ocfs2_blockcheck_inc_check(stats);
+
 	check.bc_crc32e = le32_to_cpu(bc->bc_crc32e);
 	check.bc_ecc = le16_to_cpu(bc->bc_ecc);
 
@@ -388,6 +549,7 @@ int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr,
 	if (crc == check.bc_crc32e)
 		goto out;
 
+	ocfs2_blockcheck_inc_failure(stats);
 	mlog(ML_ERROR,
 	     "CRC32 failed: stored: %u, computed %u.  Applying ECC.\n",
 	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
@@ -416,8 +578,10 @@ int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr,
 	/* And check the crc32 again */
 	for (i = 0, crc = ~0; i < nr; i++)
 		crc = crc32_le(crc, bhs[i]->b_data, bhs[i]->b_size);
-	if (crc == check.bc_crc32e)
+	if (crc == check.bc_crc32e) {
+		ocfs2_blockcheck_inc_recover(stats);
 		goto out;
+	}
 
 	mlog(ML_ERROR, "Fixed CRC32 failed: stored: %u, computed %u\n",
 	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
@@ -448,9 +612,11 @@ int ocfs2_validate_meta_ecc(struct super_block *sb, void *data,
 			    struct ocfs2_block_check *bc)
 {
 	int rc = 0;
+	struct ocfs2_super *osb = OCFS2_SB(sb);
 
-	if (ocfs2_meta_ecc(OCFS2_SB(sb)))
-		rc = ocfs2_block_check_validate(data, sb->s_blocksize, bc);
+	if (ocfs2_meta_ecc(osb))
+		rc = ocfs2_block_check_validate(data, sb->s_blocksize, bc,
+						&osb->osb_ecc_stats);
 
 	return rc;
 }
@@ -468,9 +634,11 @@ int ocfs2_validate_meta_ecc_bhs(struct super_block *sb,
 				struct ocfs2_block_check *bc)
 {
 	int rc = 0;
+	struct ocfs2_super *osb = OCFS2_SB(sb);
 
-	if (ocfs2_meta_ecc(OCFS2_SB(sb)))
-		rc = ocfs2_block_check_validate_bhs(bhs, nr, bc);
+	if (ocfs2_meta_ecc(osb))
+		rc = ocfs2_block_check_validate_bhs(bhs, nr, bc,
+						    &osb->osb_ecc_stats);
 
 	return rc;
 }
diff --git a/fs/ocfs2/blockcheck.h b/fs/ocfs2/blockcheck.h
index 70ec3feda32f..d4b69febf70a 100644
--- a/fs/ocfs2/blockcheck.h
+++ b/fs/ocfs2/blockcheck.h
@@ -21,6 +21,24 @@
 #define OCFS2_BLOCKCHECK_H
 
 
+/* Count errors and error correction from blockcheck.c */
+struct ocfs2_blockcheck_stats {
+	spinlock_t b_lock;
+	u64 b_check_count;	/* Number of blocks we've checked */
+	u64 b_failure_count;	/* Number of failed checksums */
+	u64 b_recover_count;	/* Number of blocks fixed by ecc */
+
+	/*
+	 * debugfs entries, used if this is passed to
+	 * ocfs2_blockcheck_stats_debugfs_install()
+	 */
+	struct dentry *b_debug_dir;	/* Parent of the debugfs  files */
+	struct dentry *b_debug_check;	/* Exposes b_check_count */
+	struct dentry *b_debug_failure;	/* Exposes b_failure_count */
+	struct dentry *b_debug_recover;	/* Exposes b_recover_count */
+};
+
+
 /* High level block API */
 void ocfs2_compute_meta_ecc(struct super_block *sb, void *data,
 			    struct ocfs2_block_check *bc);
@@ -37,11 +55,18 @@ int ocfs2_validate_meta_ecc_bhs(struct super_block *sb,
 void ocfs2_block_check_compute(void *data, size_t blocksize,
 			       struct ocfs2_block_check *bc);
 int ocfs2_block_check_validate(void *data, size_t blocksize,
-			       struct ocfs2_block_check *bc);
+			       struct ocfs2_block_check *bc,
+			       struct ocfs2_blockcheck_stats *stats);
 void ocfs2_block_check_compute_bhs(struct buffer_head **bhs, int nr,
 				   struct ocfs2_block_check *bc);
 int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr,
-				   struct ocfs2_block_check *bc);
+				   struct ocfs2_block_check *bc,
+				   struct ocfs2_blockcheck_stats *stats);
+
+/* Debug Initialization */
+int ocfs2_blockcheck_stats_debugfs_install(struct ocfs2_blockcheck_stats *stats,
+					   struct dentry *parent);
+void ocfs2_blockcheck_stats_debugfs_remove(struct ocfs2_blockcheck_stats *stats);
 
 /*
  * Hamming code functions
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 7e72a81bc2d4..696c32e50716 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -48,34 +48,33 @@
  * only emit the appropriage printk() when the caller passes in a constant
  * mask, as is almost always the case.
  *
- * All this bitmask nonsense is hidden from the /proc interface so that Joel
- * doesn't have an aneurism.  Reading the file gives a straight forward
- * indication of which bits are on or off:
- * 	ENTRY off
- * 	EXIT off
+ * All this bitmask nonsense is managed from the files under
+ * /sys/fs/o2cb/logmask/.  Reading the files gives a straightforward
+ * indication of which bits are allowed (allow) or denied (off/deny).
+ * 	ENTRY deny
+ * 	EXIT deny
  * 	TCP off
  * 	MSG off
  * 	SOCKET off
- * 	ERROR off
- * 	NOTICE on
+ * 	ERROR allow
+ * 	NOTICE allow
  *
  * Writing changes the state of a given bit and requires a strictly formatted
  * single write() call:
  *
- * 	write(fd, "ENTRY on", 8);
+ * 	write(fd, "allow", 5);
  *
- * would turn the entry bit on.  "1" is also accepted in the place of "on", and
- * "off" and "0" behave as expected.
+ * Echoing allow/deny/off string into the logmask files can flip the bits
+ * on or off as expected; here is the bash script for example:
  *
- * Some trivial shell can flip all the bits on or off:
+ * log_mask="/sys/fs/o2cb/log_mask"
+ * for node in ENTRY EXIT TCP MSG SOCKET ERROR NOTICE; do
+ *	echo allow >"$log_mask"/"$node"
+ * done
  *
- * log_mask="/proc/fs/ocfs2_nodemanager/log_mask"
- * cat $log_mask | (
- * 	while read bit status; do
- * 		# $1 is "on" or "off", say
- * 		echo "$bit $1" > $log_mask
- * 	done
- * )
+ * The debugfs.ocfs2 tool can also flip the bits with the -l option:
+ *
+ * debugfs.ocfs2 -l TCP allow
  */
 
 /* for task_struct */
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index 9fbe849f6344..334f231a422c 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -974,7 +974,7 @@ static int o2net_tx_can_proceed(struct o2net_node *nn,
 int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec,
 			   size_t caller_veclen, u8 target_node, int *status)
 {
-	int ret, error = 0;
+	int ret;
 	struct o2net_msg *msg = NULL;
 	size_t veclen, caller_bytes = 0;
 	struct kvec *vec = NULL;
@@ -1015,10 +1015,7 @@ int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec,
 
 	o2net_set_nst_sock_time(&nst);
 
-	ret = wait_event_interruptible(nn->nn_sc_wq,
-				       o2net_tx_can_proceed(nn, &sc, &error));
-	if (!ret && error)
-		ret = error;
+	wait_event(nn->nn_sc_wq, o2net_tx_can_proceed(nn, &sc, &ret));
 	if (ret)
 		goto out;
 
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index c5752305627c..b358f3bf896d 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -2900,6 +2900,8 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	alloc = ocfs2_clusters_for_bytes(sb, bytes);
 	dx_alloc = 0;
 
+	down_write(&oi->ip_alloc_sem);
+
 	if (ocfs2_supports_indexed_dirs(osb)) {
 		credits += ocfs2_add_dir_index_credits(sb);
 
@@ -2940,8 +2942,6 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 		goto out;
 	}
 
-	down_write(&oi->ip_alloc_sem);
-
 	/*
 	 * Prepare for worst case allocation scenario of two separate
 	 * extents in the unindexed tree.
@@ -2953,7 +2953,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	if (IS_ERR(handle)) {
 		ret = PTR_ERR(handle);
 		mlog_errno(ret);
-		goto out_sem;
+		goto out;
 	}
 
 	if (vfs_dq_alloc_space_nodirty(dir,
@@ -3172,10 +3172,8 @@ out_commit:
 
 	ocfs2_commit_trans(osb, handle);
 
-out_sem:
-	up_write(&oi->ip_alloc_sem);
-
 out:
+	up_write(&oi->ip_alloc_sem);
 	if (data_ac)
 		ocfs2_free_alloc_context(data_ac);
 	if (meta_ac)
@@ -3322,11 +3320,15 @@ static int ocfs2_extend_dir(struct ocfs2_super *osb,
 		brelse(new_bh);
 		new_bh = NULL;
 
+		down_write(&OCFS2_I(dir)->ip_alloc_sem);
+		drop_alloc_sem = 1;
 		dir_i_size = i_size_read(dir);
 		credits = OCFS2_SIMPLE_DIR_EXTEND_CREDITS;
 		goto do_extend;
 	}
 
+	down_write(&OCFS2_I(dir)->ip_alloc_sem);
+	drop_alloc_sem = 1;
 	dir_i_size = i_size_read(dir);
 	mlog(0, "extending dir %llu (i_size = %lld)\n",
 	     (unsigned long long)OCFS2_I(dir)->ip_blkno, dir_i_size);
@@ -3370,9 +3372,6 @@ do_extend:
 		credits++; /* For attaching the new dirent block to the
 			    * dx_root */
 
-	down_write(&OCFS2_I(dir)->ip_alloc_sem);
-	drop_alloc_sem = 1;
-
 	handle = ocfs2_start_trans(osb, credits);
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
@@ -3435,10 +3434,10 @@ bail_bh:
 	*new_de_bh = new_bh;
 	get_bh(*new_de_bh);
 bail:
-	if (drop_alloc_sem)
-		up_write(&OCFS2_I(dir)->ip_alloc_sem);
 	if (handle)
 		ocfs2_commit_trans(osb, handle);
+	if (drop_alloc_sem)
+		up_write(&OCFS2_I(dir)->ip_alloc_sem);
 
 	if (data_ac)
 		ocfs2_free_alloc_context(data_ac);
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index e15fc7d50827..6cdeaa76f27f 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -248,6 +248,10 @@ static struct ocfs2_lock_res_ops ocfs2_nfs_sync_lops = {
 	.flags		= 0,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
+	.flags		= LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
+};
+
 static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
 	.get_osb	= ocfs2_get_dentry_osb,
 	.post_unlock	= ocfs2_dentry_post_unlock,
@@ -637,6 +641,19 @@ static void ocfs2_nfs_sync_lock_res_init(struct ocfs2_lock_res *res,
 				   &ocfs2_nfs_sync_lops, osb);
 }
 
+static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
+					    struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan_lvb *lvb;
+
+	ocfs2_lock_res_init_once(res);
+	ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, res->l_name);
+	ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
+				   &ocfs2_orphan_scan_lops, osb);
+	lvb = ocfs2_dlm_lvb(&res->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+}
+
 void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
 			      struct ocfs2_file_private *fp)
 {
@@ -2352,6 +2369,37 @@ void ocfs2_inode_unlock(struct inode *inode,
 	mlog_exit_void();
 }
 
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+	int status = 0;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
+	if (status < 0)
+		return status;
+
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	if (lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
+		*seqno = be32_to_cpu(lvb->lvb_os_seqno);
+	return status;
+}
+
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+	lvb->lvb_os_seqno = cpu_to_be32(seqno);
+	ocfs2_cluster_unlock(osb, lockres, level);
+}
+
 int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex)
 {
@@ -2842,6 +2890,7 @@ local:
 	ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
 	ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
 	ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
+	ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
 
 	osb->cconn = conn;
 
@@ -2878,6 +2927,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
 	ocfs2_lock_res_free(&osb->osb_super_lockres);
 	ocfs2_lock_res_free(&osb->osb_rename_lockres);
 	ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
+	ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
 
 	ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
 	osb->cconn = NULL;
@@ -3061,6 +3111,7 @@ static void ocfs2_drop_osb_locks(struct ocfs2_super *osb)
 	ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
+	ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
 }
 
 int ocfs2_drop_inode_locks(struct inode *inode)
diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h
index e1fd5721cd7f..31b90d7b8f51 100644
--- a/fs/ocfs2/dlmglue.h
+++ b/fs/ocfs2/dlmglue.h
@@ -62,6 +62,14 @@ struct ocfs2_qinfo_lvb {
 	__be32	lvb_free_entry;
 };
 
+#define OCFS2_ORPHAN_LVB_VERSION 1
+
+struct ocfs2_orphan_scan_lvb {
+	__u8	lvb_version;
+	__u8	lvb_reserved[3];
+	__be32	lvb_os_seqno;
+};
+
 /* ocfs2_inode_lock_full() 'arg_flags' flags */
 /* don't wait on recovery. */
 #define OCFS2_META_LOCK_RECOVERY	(0x01)
@@ -113,6 +121,9 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex);
 void ocfs2_super_unlock(struct ocfs2_super *osb,
 			int ex);
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex);
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex);
+
 int ocfs2_rename_lock(struct ocfs2_super *osb);
 void ocfs2_rename_unlock(struct ocfs2_super *osb);
 int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex);
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index c2a87c885b73..07267e0da909 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -187,6 +187,9 @@ static int ocfs2_sync_file(struct file *file,
 	if (err)
 		goto bail;
 
+	if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
+		goto bail;
+
 	journal = osb->journal->j_journal;
 	err = jbd2_journal_force_commit(journal);
 
@@ -894,9 +897,9 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 	struct ocfs2_super *osb = OCFS2_SB(sb);
 	struct buffer_head *bh = NULL;
 	handle_t *handle = NULL;
-	int locked[MAXQUOTAS] = {0, 0};
-	int credits, qtype;
-	struct ocfs2_mem_dqinfo *oinfo;
+	int qtype;
+	struct dquot *transfer_from[MAXQUOTAS] = { };
+	struct dquot *transfer_to[MAXQUOTAS] = { };
 
 	mlog_entry("(0x%p, '%.*s')\n", dentry,
 	           dentry->d_name.len, dentry->d_name.name);
@@ -969,30 +972,37 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 
 	if ((attr->ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
 	    (attr->ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
-		credits = OCFS2_INODE_UPDATE_CREDITS;
+		/*
+		 * Gather pointers to quota structures so that allocation /
+		 * freeing of quota structures happens here and not inside
+		 * vfs_dq_transfer() where we have problems with lock ordering
+		 */
 		if (attr->ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid
 		    && OCFS2_HAS_RO_COMPAT_FEATURE(sb,
 		    OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
-			oinfo = sb_dqinfo(sb, USRQUOTA)->dqi_priv;
-			status = ocfs2_lock_global_qf(oinfo, 1);
-			if (status < 0)
+			transfer_to[USRQUOTA] = dqget(sb, attr->ia_uid,
+						      USRQUOTA);
+			transfer_from[USRQUOTA] = dqget(sb, inode->i_uid,
+							USRQUOTA);
+			if (!transfer_to[USRQUOTA] || !transfer_from[USRQUOTA]) {
+				status = -ESRCH;
 				goto bail_unlock;
-			credits += ocfs2_calc_qinit_credits(sb, USRQUOTA) +
-				ocfs2_calc_qdel_credits(sb, USRQUOTA);
-			locked[USRQUOTA] = 1;
+			}
 		}
 		if (attr->ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid
 		    && OCFS2_HAS_RO_COMPAT_FEATURE(sb,
 		    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
-			oinfo = sb_dqinfo(sb, GRPQUOTA)->dqi_priv;
-			status = ocfs2_lock_global_qf(oinfo, 1);
-			if (status < 0)
+			transfer_to[GRPQUOTA] = dqget(sb, attr->ia_gid,
+						      GRPQUOTA);
+			transfer_from[GRPQUOTA] = dqget(sb, inode->i_gid,
+							GRPQUOTA);
+			if (!transfer_to[GRPQUOTA] || !transfer_from[GRPQUOTA]) {
+				status = -ESRCH;
 				goto bail_unlock;
-			credits += ocfs2_calc_qinit_credits(sb, GRPQUOTA) +
-				   ocfs2_calc_qdel_credits(sb, GRPQUOTA);
-			locked[GRPQUOTA] = 1;
+			}
 		}
-		handle = ocfs2_start_trans(osb, credits);
+		handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS +
+					   2 * ocfs2_quota_trans_credits(sb));
 		if (IS_ERR(handle)) {
 			status = PTR_ERR(handle);
 			mlog_errno(status);
@@ -1030,12 +1040,6 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 bail_commit:
 	ocfs2_commit_trans(osb, handle);
 bail_unlock:
-	for (qtype = 0; qtype < MAXQUOTAS; qtype++) {
-		if (!locked[qtype])
-			continue;
-		oinfo = sb_dqinfo(sb, qtype)->dqi_priv;
-		ocfs2_unlock_global_qf(oinfo, 1);
-	}
 	ocfs2_inode_unlock(inode, 1);
 bail_unlock_rw:
 	if (size_change)
@@ -1043,6 +1047,12 @@ bail_unlock_rw:
 bail:
 	brelse(bh);
 
+	/* Release quota pointers in case we acquired them */
+	for (qtype = 0; qtype < MAXQUOTAS; qtype++) {
+		dqput(transfer_to[qtype]);
+		dqput(transfer_from[qtype]);
+	}
+
 	if (!status && attr->ia_valid & ATTR_MODE) {
 		status = ocfs2_acl_chmod(inode);
 		if (status < 0)
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index a20a0f1e37fd..4a3b9e6b31ad 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -28,6 +28,8 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/kthread.h>
+#include <linux/time.h>
+#include <linux/random.h>
 
 #define MLOG_MASK_PREFIX ML_JOURNAL
 #include <cluster/masklog.h>
@@ -52,6 +54,8 @@
 
 DEFINE_SPINLOCK(trans_inc_lock);
 
+#define ORPHAN_SCAN_SCHEDULE_TIMEOUT 300000
+
 static int ocfs2_force_read_journal(struct inode *inode);
 static int ocfs2_recover_node(struct ocfs2_super *osb,
 			      int node_num, int slot_num);
@@ -1841,6 +1845,113 @@ bail:
 	return status;
 }
 
+/*
+ * Scan timer should get fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT. Add some
+ * randomness to the timeout to minimize multple nodes firing the timer at the
+ * same time.
+ */
+static inline unsigned long ocfs2_orphan_scan_timeout(void)
+{
+	unsigned long time;
+
+	get_random_bytes(&time, sizeof(time));
+	time = ORPHAN_SCAN_SCHEDULE_TIMEOUT + (time % 5000);
+	return msecs_to_jiffies(time);
+}
+
+/*
+ * ocfs2_queue_orphan_scan calls ocfs2_queue_recovery_completion for
+ * every slot, queuing a recovery of the slot on the ocfs2_wq thread. This
+ * is done to catch any orphans that are left over in orphan directories.
+ *
+ * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT
+ * seconds.  It gets an EX lock on os_lockres and checks sequence number
+ * stored in LVB. If the sequence number has changed, it means some other
+ * node has done the scan.  This node skips the scan and tracks the
+ * sequence number.  If the sequence number didn't change, it means a scan
+ * hasn't happened.  The node queues a scan and increments the
+ * sequence number in the LVB.
+ */
+void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+	int status, i;
+	u32 seqno = 0;
+
+	os = &osb->osb_orphan_scan;
+
+	status = ocfs2_orphan_scan_lock(osb, &seqno, DLM_LOCK_EX);
+	if (status < 0) {
+		if (status != -EAGAIN)
+			mlog_errno(status);
+		goto out;
+	}
+
+	if (os->os_seqno != seqno) {
+		os->os_seqno = seqno;
+		goto unlock;
+	}
+
+	for (i = 0; i < osb->max_slots; i++)
+		ocfs2_queue_recovery_completion(osb->journal, i, NULL, NULL,
+						NULL);
+	/*
+	 * We queued a recovery on orphan slots, increment the sequence
+	 * number and update LVB so other node will skip the scan for a while
+	 */
+	seqno++;
+	os->os_count++;
+	os->os_scantime = CURRENT_TIME;
+unlock:
+	ocfs2_orphan_scan_unlock(osb, seqno, DLM_LOCK_EX);
+out:
+	return;
+}
+
+/* Worker task that gets fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT millsec */
+void ocfs2_orphan_scan_work(struct work_struct *work)
+{
+	struct ocfs2_orphan_scan *os;
+	struct ocfs2_super *osb;
+
+	os = container_of(work, struct ocfs2_orphan_scan,
+			  os_orphan_scan_work.work);
+	osb = os->os_osb;
+
+	mutex_lock(&os->os_lock);
+	ocfs2_queue_orphan_scan(osb);
+	schedule_delayed_work(&os->os_orphan_scan_work,
+			      ocfs2_orphan_scan_timeout());
+	mutex_unlock(&os->os_lock);
+}
+
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+
+	os = &osb->osb_orphan_scan;
+	mutex_lock(&os->os_lock);
+	cancel_delayed_work(&os->os_orphan_scan_work);
+	mutex_unlock(&os->os_lock);
+}
+
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+
+	os = &osb->osb_orphan_scan;
+	os->os_osb = osb;
+	os->os_count = 0;
+	os->os_scantime = CURRENT_TIME;
+	mutex_init(&os->os_lock);
+
+	INIT_DELAYED_WORK(&os->os_orphan_scan_work,
+			  ocfs2_orphan_scan_work);
+	schedule_delayed_work(&os->os_orphan_scan_work,
+			      ocfs2_orphan_scan_timeout());
+	return 0;
+}
+
 struct ocfs2_orphan_filldir_priv {
 	struct inode		*head;
 	struct ocfs2_super	*osb;
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index eb7b76331eb7..61045eeb3f6e 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -144,6 +144,10 @@ static inline void ocfs2_inode_set_new(struct ocfs2_super *osb,
 }
 
 /* Exported only for the journal struct init code in super.c. Do not call. */
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_exit(struct ocfs2_super *osb);
+
 void ocfs2_complete_recovery(struct work_struct *work);
 void ocfs2_wait_for_recovery(struct ocfs2_super *osb);
 
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index 1386281950db..18c1d9ec1c93 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -47,6 +47,9 @@
 #include "ocfs2_fs.h"
 #include "ocfs2_lockid.h"
 
+/* For struct ocfs2_blockcheck_stats */
+#include "blockcheck.h"
+
 /* Most user visible OCFS2 inodes will have very few pieces of
  * metadata, but larger files (including bitmaps, etc) must be taken
  * into account when designing an access scheme. We allow a small
@@ -151,6 +154,16 @@ struct ocfs2_lock_res {
 #endif
 };
 
+struct ocfs2_orphan_scan {
+	struct mutex 		os_lock;
+	struct ocfs2_super 	*os_osb;
+	struct ocfs2_lock_res 	os_lockres;     /* lock to synchronize scans */
+	struct delayed_work 	os_orphan_scan_work;
+	struct timespec		os_scantime;  /* time this node ran the scan */
+	u32			os_count;      /* tracks node specific scans */
+	u32  			os_seqno;       /* tracks cluster wide scans */
+};
+
 struct ocfs2_dlm_debug {
 	struct kref d_refcnt;
 	struct dentry *d_locking_state;
@@ -295,6 +308,7 @@ struct ocfs2_super
 	struct ocfs2_dinode *local_alloc_copy;
 	struct ocfs2_quota_recovery *quota_rec;
 
+	struct ocfs2_blockcheck_stats osb_ecc_stats;
 	struct ocfs2_alloc_stats alloc_stats;
 	char dev_str[20];		/* "major,minor" of the device */
 
@@ -341,6 +355,8 @@ struct ocfs2_super
 	unsigned int			*osb_orphan_wipes;
 	wait_queue_head_t		osb_wipe_event;
 
+	struct ocfs2_orphan_scan	osb_orphan_scan;
+
 	/* used to protect metaecc calculation check of xattr. */
 	spinlock_t osb_xattr_lock;
 
diff --git a/fs/ocfs2/ocfs2_lockid.h b/fs/ocfs2/ocfs2_lockid.h
index a53ce87481bf..fcdba091af3d 100644
--- a/fs/ocfs2/ocfs2_lockid.h
+++ b/fs/ocfs2/ocfs2_lockid.h
@@ -48,6 +48,7 @@ enum ocfs2_lock_type {
 	OCFS2_LOCK_TYPE_FLOCK,
 	OCFS2_LOCK_TYPE_QINFO,
 	OCFS2_LOCK_TYPE_NFS_SYNC,
+	OCFS2_LOCK_TYPE_ORPHAN_SCAN,
 	OCFS2_NUM_LOCK_TYPES
 };
 
@@ -85,6 +86,9 @@ static inline char ocfs2_lock_type_char(enum ocfs2_lock_type type)
 		case OCFS2_LOCK_TYPE_NFS_SYNC:
 			c = 'Y';
 			break;
+		case OCFS2_LOCK_TYPE_ORPHAN_SCAN:
+			c = 'P';
+			break;
 		default:
 			c = '\0';
 	}
@@ -104,6 +108,7 @@ static char *ocfs2_lock_type_strings[] = {
 	[OCFS2_LOCK_TYPE_OPEN] = "Open",
 	[OCFS2_LOCK_TYPE_FLOCK] = "Flock",
 	[OCFS2_LOCK_TYPE_QINFO] = "Quota",
+	[OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
 };
 
 static inline const char *ocfs2_lock_type_string(enum ocfs2_lock_type type)
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index 1ed0f7c86869..edfa60cd155c 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -421,6 +421,7 @@ int ocfs2_global_read_dquot(struct dquot *dquot)
 	OCFS2_DQUOT(dquot)->dq_originodes = dquot->dq_dqb.dqb_curinodes;
 	if (!dquot->dq_off) {	/* No real quota entry? */
 		/* Upgrade to exclusive lock for allocation */
+		ocfs2_qinfo_unlock(info, 0);
 		err = ocfs2_qinfo_lock(info, 1);
 		if (err < 0)
 			goto out_qlock;
@@ -435,7 +436,8 @@ int ocfs2_global_read_dquot(struct dquot *dquot)
 out_qlock:
 	if (ex)
 		ocfs2_qinfo_unlock(info, 1);
-	ocfs2_qinfo_unlock(info, 0);
+	else
+		ocfs2_qinfo_unlock(info, 0);
 out:
 	if (err < 0)
 		mlog_errno(err);
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index 07deec5e9721..5a460fa82553 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -444,10 +444,6 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
 
 	mlog_entry("ino=%lu type=%u", (unsigned long)lqinode->i_ino, type);
 
-	status = ocfs2_lock_global_qf(oinfo, 1);
-	if (status < 0)
-		goto out;
-
 	list_for_each_entry_safe(rchunk, next, &(rec->r_list[type]), rc_list) {
 		chunk = rchunk->rc_chunk;
 		hbh = NULL;
@@ -480,12 +476,18 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
 				     type);
 				goto out_put_bh;
 			}
+			status = ocfs2_lock_global_qf(oinfo, 1);
+			if (status < 0) {
+				mlog_errno(status);
+				goto out_put_dquot;
+			}
+
 			handle = ocfs2_start_trans(OCFS2_SB(sb),
 						   OCFS2_QSYNC_CREDITS);
 			if (IS_ERR(handle)) {
 				status = PTR_ERR(handle);
 				mlog_errno(status);
-				goto out_put_dquot;
+				goto out_drop_lock;
 			}
 			mutex_lock(&sb_dqopt(sb)->dqio_mutex);
 			spin_lock(&dq_data_lock);
@@ -523,6 +525,8 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
 out_commit:
 			mutex_unlock(&sb_dqopt(sb)->dqio_mutex);
 			ocfs2_commit_trans(OCFS2_SB(sb), handle);
+out_drop_lock:
+			ocfs2_unlock_global_qf(oinfo, 1);
 out_put_dquot:
 			dqput(dquot);
 out_put_bh:
@@ -537,8 +541,6 @@ out_put_bh:
 		if (status < 0)
 			break;
 	}
-	ocfs2_unlock_global_qf(oinfo, 1);
-out:
 	if (status < 0)
 		free_recovery_list(&(rec->r_list[type]));
 	mlog_exit(status);
@@ -655,6 +657,9 @@ static int ocfs2_local_read_info(struct super_block *sb, int type)
 	struct ocfs2_quota_recovery *rec;
 	int locked = 0;
 
+	/* We don't need the lock and we have to acquire quota file locks
+	 * which will later depend on this lock */
+	mutex_unlock(&sb_dqopt(sb)->dqio_mutex);
 	info->dqi_maxblimit = 0x7fffffffffffffffLL;
 	info->dqi_maxilimit = 0x7fffffffffffffffLL;
 	oinfo = kmalloc(sizeof(struct ocfs2_mem_dqinfo), GFP_NOFS);
@@ -733,6 +738,7 @@ static int ocfs2_local_read_info(struct super_block *sb, int type)
 		goto out_err;
 	}
 
+	mutex_lock(&sb_dqopt(sb)->dqio_mutex);
 	return 0;
 out_err:
 	if (oinfo) {
@@ -746,6 +752,7 @@ out_err:
 		kfree(oinfo);
 	}
 	brelse(bh);
+	mutex_lock(&sb_dqopt(sb)->dqio_mutex);
 	return -1;
 }
 
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 201b40a441fe..d33767f17ba3 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -119,10 +119,12 @@ static void ocfs2_release_system_inodes(struct ocfs2_super *osb);
 static int ocfs2_check_volume(struct ocfs2_super *osb);
 static int ocfs2_verify_volume(struct ocfs2_dinode *di,
 			       struct buffer_head *bh,
-			       u32 sectsize);
+			       u32 sectsize,
+			       struct ocfs2_blockcheck_stats *stats);
 static int ocfs2_initialize_super(struct super_block *sb,
 				  struct buffer_head *bh,
-				  int sector_size);
+				  int sector_size,
+				  struct ocfs2_blockcheck_stats *stats);
 static int ocfs2_get_sector(struct super_block *sb,
 			    struct buffer_head **bh,
 			    int block,
@@ -207,6 +209,7 @@ static int ocfs2_osb_dump(struct ocfs2_super *osb, char *buf, int len)
 	int i;
 	struct ocfs2_cluster_connection *cconn = osb->cconn;
 	struct ocfs2_recovery_map *rm = osb->recovery_map;
+	struct ocfs2_orphan_scan *os;
 
 	out += snprintf(buf + out, len - out,
 			"%10s => Id: %-s  Uuid: %-s  Gen: 0x%X  Label: %-s\n",
@@ -308,6 +311,13 @@ static int ocfs2_osb_dump(struct ocfs2_super *osb, char *buf, int len)
 				i, osb->slot_recovery_generations[i]);
 	}
 
+	os = &osb->osb_orphan_scan;
+	out += snprintf(buf + out, len - out, "Orphan Scan=> ");
+	out += snprintf(buf + out, len - out, "Local: %u  Global: %u ",
+			os->os_count, os->os_seqno);
+	out += snprintf(buf + out, len - out, " Last Scan: %lu seconds ago\n",
+			(get_seconds() - os->os_scantime.tv_sec));
+
 	return out;
 }
 
@@ -693,7 +703,8 @@ out:
 
 static int ocfs2_sb_probe(struct super_block *sb,
 			  struct buffer_head **bh,
-			  int *sector_size)
+			  int *sector_size,
+			  struct ocfs2_blockcheck_stats *stats)
 {
 	int status, tmpstat;
 	struct ocfs1_vol_disk_hdr *hdr;
@@ -759,7 +770,8 @@ static int ocfs2_sb_probe(struct super_block *sb,
 			goto bail;
 		}
 		di = (struct ocfs2_dinode *) (*bh)->b_data;
-		status = ocfs2_verify_volume(di, *bh, blksize);
+		memset(stats, 0, sizeof(struct ocfs2_blockcheck_stats));
+		status = ocfs2_verify_volume(di, *bh, blksize, stats);
 		if (status >= 0)
 			goto bail;
 		brelse(*bh);
@@ -965,6 +977,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	struct ocfs2_super *osb = NULL;
 	struct buffer_head *bh = NULL;
 	char nodestr[8];
+	struct ocfs2_blockcheck_stats stats;
 
 	mlog_entry("%p, %p, %i", sb, data, silent);
 
@@ -974,13 +987,13 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	}
 
 	/* probe for superblock */
-	status = ocfs2_sb_probe(sb, &bh, &sector_size);
+	status = ocfs2_sb_probe(sb, &bh, &sector_size, &stats);
 	if (status < 0) {
 		mlog(ML_ERROR, "superblock probe failed!\n");
 		goto read_super_error;
 	}
 
-	status = ocfs2_initialize_super(sb, bh, sector_size);
+	status = ocfs2_initialize_super(sb, bh, sector_size, &stats);
 	osb = OCFS2_SB(sb);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1090,6 +1103,18 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 		goto read_super_error;
 	}
 
+	if (ocfs2_meta_ecc(osb)) {
+		status = ocfs2_blockcheck_stats_debugfs_install(
+						&osb->osb_ecc_stats,
+						osb->osb_debug_root);
+		if (status) {
+			mlog(ML_ERROR,
+			     "Unable to create blockcheck statistics "
+			     "files\n");
+			goto read_super_error;
+		}
+	}
+
 	status = ocfs2_mount_volume(sb);
 	if (osb->root_inode)
 		inode = igrab(osb->root_inode);
@@ -1760,13 +1785,8 @@ static int ocfs2_mount_volume(struct super_block *sb)
 	}
 
 	status = ocfs2_truncate_log_init(osb);
-	if (status < 0) {
+	if (status < 0)
 		mlog_errno(status);
-		goto leave;
-	}
-
-	if (ocfs2_mount_local(osb))
-		goto leave;
 
 leave:
 	if (unlock_super)
@@ -1796,6 +1816,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	ocfs2_truncate_log_shutdown(osb);
 
+	ocfs2_orphan_scan_stop(osb);
+
 	/* This will disable recovery and flush any recovery work. */
 	ocfs2_recovery_exit(osb);
 
@@ -1833,6 +1855,7 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 	if (osb->cconn)
 		ocfs2_dlm_shutdown(osb, hangup_needed);
 
+	ocfs2_blockcheck_stats_debugfs_remove(&osb->osb_ecc_stats);
 	debugfs_remove(osb->osb_debug_root);
 
 	if (hangup_needed)
@@ -1880,7 +1903,8 @@ static int ocfs2_setup_osb_uuid(struct ocfs2_super *osb, const unsigned char *uu
 
 static int ocfs2_initialize_super(struct super_block *sb,
 				  struct buffer_head *bh,
-				  int sector_size)
+				  int sector_size,
+				  struct ocfs2_blockcheck_stats *stats)
 {
 	int status;
 	int i, cbits, bbits;
@@ -1939,6 +1963,9 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	atomic_set(&osb->alloc_stats.bg_allocs, 0);
 	atomic_set(&osb->alloc_stats.bg_extends, 0);
 
+	/* Copy the blockcheck stats from the superblock probe */
+	osb->osb_ecc_stats = *stats;
+
 	ocfs2_init_node_maps(osb);
 
 	snprintf(osb->dev_str, sizeof(osb->dev_str), "%u,%u",
@@ -1951,6 +1978,13 @@ static int ocfs2_initialize_super(struct super_block *sb,
 		goto bail;
 	}
 
+	status = ocfs2_orphan_scan_init(osb);
+	if (status) {
+		mlog(ML_ERROR, "Unable to initialize delayed orphan scan\n");
+		mlog_errno(status);
+		goto bail;
+	}
+
 	init_waitqueue_head(&osb->checkpoint_event);
 	atomic_set(&osb->needs_checkpoint, 0);
 
@@ -2169,7 +2203,8 @@ bail:
  */
 static int ocfs2_verify_volume(struct ocfs2_dinode *di,
 			       struct buffer_head *bh,
-			       u32 blksz)
+			       u32 blksz,
+			       struct ocfs2_blockcheck_stats *stats)
 {
 	int status = -EAGAIN;
 
@@ -2182,7 +2217,8 @@ static int ocfs2_verify_volume(struct ocfs2_dinode *di,
 		    OCFS2_FEATURE_INCOMPAT_META_ECC) {
 			status = ocfs2_block_check_validate(bh->b_data,
 							    bh->b_size,
-							    &di->i_check);
+							    &di->i_check,
+							    stats);
 			if (status)
 				goto out;
 		}
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 15631019dc63..ba320e250747 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -3154,7 +3154,7 @@ static int ocfs2_iterate_xattr_buckets(struct inode *inode,
 		     le32_to_cpu(bucket_xh(bucket)->xh_entries[0].xe_name_hash));
 		if (func) {
 			ret = func(inode, bucket, para);
-			if (ret)
+			if (ret && ret != -ERANGE)
 				mlog_errno(ret);
 			/* Fall through to bucket_relse() */
 		}
@@ -3261,7 +3261,8 @@ static int ocfs2_xattr_tree_list_index_block(struct inode *inode,
 						  ocfs2_list_xattr_bucket,
 						  &xl);
 		if (ret) {
-			mlog_errno(ret);
+			if (ret != -ERANGE)
+				mlog_errno(ret);
 			goto out;
 		}
 
diff --git a/fs/ramfs/inode.c b/fs/ramfs/inode.c
index 3a6b193d8444..0ff7566c767c 100644
--- a/fs/ramfs/inode.c
+++ b/fs/ramfs/inode.c
@@ -202,9 +202,12 @@ static int ramfs_parse_options(char *data, struct ramfs_mount_opts *opts)
 				return -EINVAL;
 			opts->mode = option & S_IALLUGO;
 			break;
-		default:
-			printk(KERN_ERR "ramfs: bad mount option: %s\n", p);
-			return -EINVAL;
+		/*
+		 * We might like to report bad mount options here;
+		 * but traditionally ramfs has ignored all mount options,
+		 * and as it is used as a !CONFIG_SHMEM simple substitute
+		 * for tmpfs, better continue to ignore other mount options.
+		 */
 		}
 	}
 
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 3589eab02a2f..3260b73abe29 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -1937,6 +1937,9 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 	err  = bdi_init(&c->bdi);
 	if (err)
 		goto out_close;
+	err = bdi_register(&c->bdi, NULL, "ubifs");
+	if (err)
+		goto out_bdi;
 
 	err = ubifs_parse_options(c, data, 0);
 	if (err)