Merge commit 'v2.6.31-rc8' into x86/txt

Conflicts:
	arch/x86/kernel/reboot.c
	security/Kconfig

Merge reason: resolve the conflicts, bump up from rc3 to rc8.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

118 files changed, 2980 insertions, 1354 deletions

diff --git a/fs/9p/v9fs.c b/fs/9p/v9fs.c
index 332b5ff02fec..f7003cfac63d 100644
--- a/fs/9p/v9fs.c
+++ b/fs/9p/v9fs.c
@@ -76,7 +76,7 @@ static const match_table_t tokens = {
  * Return 0 upon success, -ERRNO upon failure.
  */
 
-static int v9fs_parse_options(struct v9fs_session_info *v9ses)
+static int v9fs_parse_options(struct v9fs_session_info *v9ses, char *opts)
 {
 	char *options;
 	substring_t args[MAX_OPT_ARGS];
@@ -90,10 +90,10 @@ static int v9fs_parse_options(struct v9fs_session_info *v9ses)
 	v9ses->debug = 0;
 	v9ses->cache = 0;
 
-	if (!v9ses->options)
+	if (!opts)
 		return 0;
 
-	options = kstrdup(v9ses->options, GFP_KERNEL);
+	options = kstrdup(opts, GFP_KERNEL);
 	if (!options) {
 		P9_DPRINTK(P9_DEBUG_ERROR,
 			   "failed to allocate copy of option string\n");
@@ -206,24 +206,14 @@ struct p9_fid *v9fs_session_init(struct v9fs_session_info *v9ses,
 	v9ses->uid = ~0;
 	v9ses->dfltuid = V9FS_DEFUID;
 	v9ses->dfltgid = V9FS_DEFGID;
-	if (data) {
-		v9ses->options = kstrdup(data, GFP_KERNEL);
-		if (!v9ses->options) {
-			P9_DPRINTK(P9_DEBUG_ERROR,
-			   "failed to allocate copy of option string\n");
-			retval = -ENOMEM;
-			goto error;
-		}
-	}
 
-	rc = v9fs_parse_options(v9ses);
+	rc = v9fs_parse_options(v9ses, data);
 	if (rc < 0) {
 		retval = rc;
 		goto error;
 	}
 
-	v9ses->clnt = p9_client_create(dev_name, v9ses->options);
-
+	v9ses->clnt = p9_client_create(dev_name, data);
 	if (IS_ERR(v9ses->clnt)) {
 		retval = PTR_ERR(v9ses->clnt);
 		v9ses->clnt = NULL;
@@ -280,7 +270,6 @@ void v9fs_session_close(struct v9fs_session_info *v9ses)
 
 	__putname(v9ses->uname);
 	__putname(v9ses->aname);
-	kfree(v9ses->options);
 }
 
 /**
diff --git a/fs/9p/v9fs.h b/fs/9p/v9fs.h
index a7d567192998..38762bf102a9 100644
--- a/fs/9p/v9fs.h
+++ b/fs/9p/v9fs.h
@@ -85,7 +85,6 @@ struct v9fs_session_info {
 	unsigned int afid;
 	unsigned int cache;
 
-	char *options;		/* copy of mount options */
 	char *uname;		/* user name to mount as */
 	char *aname;		/* name of remote hierarchy being mounted */
 	unsigned int maxdata;	/* max data for client interface */
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index 81f8bbf12f9f..06a223d50a81 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -171,7 +171,6 @@ int v9fs_uflags2omode(int uflags, int extended)
 
 /**
  * v9fs_blank_wstat - helper function to setup a 9P stat structure
- * @v9ses: 9P session info (for determining extended mode)
  * @wstat: structure to initialize
  *
  */
@@ -207,65 +206,72 @@ v9fs_blank_wstat(struct p9_wstat *wstat)
 
 struct inode *v9fs_get_inode(struct super_block *sb, int mode)
 {
+	int err;
 	struct inode *inode;
 	struct v9fs_session_info *v9ses = sb->s_fs_info;
 
 	P9_DPRINTK(P9_DEBUG_VFS, "super block: %p mode: %o\n", sb, mode);
 
 	inode = new_inode(sb);
-	if (inode) {
-		inode->i_mode = mode;
-		inode->i_uid = current_fsuid();
-		inode->i_gid = current_fsgid();
-		inode->i_blocks = 0;
-		inode->i_rdev = 0;
-		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
-		inode->i_mapping->a_ops = &v9fs_addr_operations;
-
-		switch (mode & S_IFMT) {
-		case S_IFIFO:
-		case S_IFBLK:
-		case S_IFCHR:
-		case S_IFSOCK:
-			if (!v9fs_extended(v9ses)) {
-				P9_DPRINTK(P9_DEBUG_ERROR,
-				      "special files without extended mode\n");
-				return ERR_PTR(-EINVAL);
-			}
-			init_special_inode(inode, inode->i_mode,
-					   inode->i_rdev);
-			break;
-		case S_IFREG:
-			inode->i_op = &v9fs_file_inode_operations;
-			inode->i_fop = &v9fs_file_operations;
-			break;
-		case S_IFLNK:
-			if (!v9fs_extended(v9ses)) {
-				P9_DPRINTK(P9_DEBUG_ERROR,
-					"extended modes used w/o 9P2000.u\n");
-				return ERR_PTR(-EINVAL);
-			}
-			inode->i_op = &v9fs_symlink_inode_operations;
-			break;
-		case S_IFDIR:
-			inc_nlink(inode);
-			if (v9fs_extended(v9ses))
-				inode->i_op = &v9fs_dir_inode_operations_ext;
-			else
-				inode->i_op = &v9fs_dir_inode_operations;
-			inode->i_fop = &v9fs_dir_operations;
-			break;
-		default:
-			P9_DPRINTK(P9_DEBUG_ERROR,
-				"BAD mode 0x%x S_IFMT 0x%x\n",
-				mode, mode & S_IFMT);
-			return ERR_PTR(-EINVAL);
-		}
-	} else {
+	if (!inode) {
 		P9_EPRINTK(KERN_WARNING, "Problem allocating inode\n");
 		return ERR_PTR(-ENOMEM);
 	}
+
+	inode->i_mode = mode;
+	inode->i_uid = current_fsuid();
+	inode->i_gid = current_fsgid();
+	inode->i_blocks = 0;
+	inode->i_rdev = 0;
+	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+	inode->i_mapping->a_ops = &v9fs_addr_operations;
+
+	switch (mode & S_IFMT) {
+	case S_IFIFO:
+	case S_IFBLK:
+	case S_IFCHR:
+	case S_IFSOCK:
+		if (!v9fs_extended(v9ses)) {
+			P9_DPRINTK(P9_DEBUG_ERROR,
+				   "special files without extended mode\n");
+			err = -EINVAL;
+			goto error;
+		}
+		init_special_inode(inode, inode->i_mode, inode->i_rdev);
+		break;
+	case S_IFREG:
+		inode->i_op = &v9fs_file_inode_operations;
+		inode->i_fop = &v9fs_file_operations;
+		break;
+	case S_IFLNK:
+		if (!v9fs_extended(v9ses)) {
+			P9_DPRINTK(P9_DEBUG_ERROR,
+				   "extended modes used w/o 9P2000.u\n");
+			err = -EINVAL;
+			goto error;
+		}
+		inode->i_op = &v9fs_symlink_inode_operations;
+		break;
+	case S_IFDIR:
+		inc_nlink(inode);
+		if (v9fs_extended(v9ses))
+			inode->i_op = &v9fs_dir_inode_operations_ext;
+		else
+			inode->i_op = &v9fs_dir_inode_operations;
+		inode->i_fop = &v9fs_dir_operations;
+		break;
+	default:
+		P9_DPRINTK(P9_DEBUG_ERROR, "BAD mode 0x%x S_IFMT 0x%x\n",
+			   mode, mode & S_IFMT);
+		err = -EINVAL;
+		goto error;
+	}
+
 	return inode;
+
+error:
+	iput(inode);
+	return ERR_PTR(err);
 }
 
 /*
@@ -338,30 +344,25 @@ v9fs_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
 
 	ret = NULL;
 	st = p9_client_stat(fid);
-	if (IS_ERR(st)) {
-		err = PTR_ERR(st);
-		st = NULL;
-		goto error;
-	}
+	if (IS_ERR(st))
+		return ERR_CAST(st);
 
 	umode = p9mode2unixmode(v9ses, st->mode);
 	ret = v9fs_get_inode(sb, umode);
 	if (IS_ERR(ret)) {
 		err = PTR_ERR(ret);
-		ret = NULL;
 		goto error;
 	}
 
 	v9fs_stat2inode(st, ret, sb);
 	ret->i_ino = v9fs_qid2ino(&st->qid);
+	p9stat_free(st);
 	kfree(st);
 	return ret;
 
 error:
+	p9stat_free(st);
 	kfree(st);
-	if (ret)
-		iput(ret);
-
 	return ERR_PTR(err);
 }
 
@@ -403,9 +404,9 @@ v9fs_open_created(struct inode *inode, struct file *file)
  * @v9ses: session information
  * @dir: directory that dentry is being created in
  * @dentry:  dentry that is being created
+ * @extension: 9p2000.u extension string to support devices, etc.
  * @perm: create permissions
  * @mode: open mode
- * @extension: 9p2000.u extension string to support devices, etc.
  *
  */
 static struct p9_fid *
@@ -470,7 +471,10 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 		dentry->d_op = &v9fs_dentry_operations;
 
 	d_instantiate(dentry, inode);
-	v9fs_fid_add(dentry, fid);
+	err = v9fs_fid_add(dentry, fid);
+	if (err < 0)
+		goto error;
+
 	return ofid;
 
 error:
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index 38d695d66a0b..8961f1a8f668 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -81,7 +81,7 @@ static int v9fs_set_super(struct super_block *s, void *data)
 
 static void
 v9fs_fill_super(struct super_block *sb, struct v9fs_session_info *v9ses,
-		int flags)
+		int flags, void *data)
 {
 	sb->s_maxbytes = MAX_LFS_FILESIZE;
 	sb->s_blocksize_bits = fls(v9ses->maxdata - 1);
@@ -91,6 +91,8 @@ v9fs_fill_super(struct super_block *sb, struct v9fs_session_info *v9ses,
 
 	sb->s_flags = flags | MS_ACTIVE | MS_SYNCHRONOUS | MS_DIRSYNC |
 	    MS_NOATIME;
+
+	save_mount_options(sb, data);
 }
 
 /**
@@ -113,14 +115,11 @@ static int v9fs_get_sb(struct file_system_type *fs_type, int flags,
 	struct v9fs_session_info *v9ses = NULL;
 	struct p9_wstat *st = NULL;
 	int mode = S_IRWXUGO | S_ISVTX;
-	uid_t uid = current_fsuid();
-	gid_t gid = current_fsgid();
 	struct p9_fid *fid;
 	int retval = 0;
 
 	P9_DPRINTK(P9_DEBUG_VFS, " \n");
 
-	st = NULL;
 	v9ses = kzalloc(sizeof(struct v9fs_session_info), GFP_KERNEL);
 	if (!v9ses)
 		return -ENOMEM;
@@ -142,7 +141,7 @@ static int v9fs_get_sb(struct file_system_type *fs_type, int flags,
 		retval = PTR_ERR(sb);
 		goto free_stat;
 	}
-	v9fs_fill_super(sb, v9ses, flags);
+	v9fs_fill_super(sb, v9ses, flags, data);
 
 	inode = v9fs_get_inode(sb, S_IFDIR | mode);
 	if (IS_ERR(inode)) {
@@ -150,9 +149,6 @@ static int v9fs_get_sb(struct file_system_type *fs_type, int flags,
 		goto release_sb;
 	}
 
-	inode->i_uid = uid;
-	inode->i_gid = gid;
-
 	root = d_alloc_root(inode);
 	if (!root) {
 		iput(inode);
@@ -173,10 +169,8 @@ P9_DPRINTK(P9_DEBUG_VFS, " simple set mount, return 0\n");
 	simple_set_mnt(mnt, sb);
 	return 0;
 
-release_sb:
-	deactivate_locked_super(sb);
-
 free_stat:
+	p9stat_free(st);
 	kfree(st);
 
 clunk_fid:
@@ -185,7 +179,12 @@ clunk_fid:
 close_session:
 	v9fs_session_close(v9ses);
 	kfree(v9ses);
+	return retval;
 
+release_sb:
+	p9stat_free(st);
+	kfree(st);
+	deactivate_locked_super(sb);
 	return retval;
 }
 
@@ -207,24 +206,10 @@ static void v9fs_kill_super(struct super_block *s)
 
 	v9fs_session_close(v9ses);
 	kfree(v9ses);
+	s->s_fs_info = NULL;
 	P9_DPRINTK(P9_DEBUG_VFS, "exiting kill_super\n");
 }
 
-/**
- * v9fs_show_options - Show mount options in /proc/mounts
- * @m: seq_file to write to
- * @mnt: mount descriptor
- *
- */
-
-static int v9fs_show_options(struct seq_file *m, struct vfsmount *mnt)
-{
-	struct v9fs_session_info *v9ses = mnt->mnt_sb->s_fs_info;
-
-	seq_printf(m, "%s", v9ses->options);
-	return 0;
-}
-
 static void
 v9fs_umount_begin(struct super_block *sb)
 {
@@ -237,7 +222,7 @@ v9fs_umount_begin(struct super_block *sb)
 static const struct super_operations v9fs_super_ops = {
 	.statfs = simple_statfs,
 	.clear_inode = v9fs_clear_inode,
-	.show_options = v9fs_show_options,
+	.show_options = generic_show_options,
 	.umount_begin = v9fs_umount_begin,
 };
 
diff --git a/fs/Kconfig b/fs/Kconfig
index a97263be6a91..0e7da7bb5d93 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -186,32 +186,7 @@ source "fs/romfs/Kconfig"
 source "fs/sysv/Kconfig"
 source "fs/ufs/Kconfig"
 source "fs/exofs/Kconfig"
-
-config NILFS2_FS
-	tristate "NILFS2 file system support (EXPERIMENTAL)"
-	depends on BLOCK && EXPERIMENTAL
-	select CRC32
-	help
-	  NILFS2 is a log-structured file system (LFS) supporting continuous
-	  snapshotting.  In addition to versioning capability of the entire
-	  file system, users can even restore files mistakenly overwritten or
-	  destroyed just a few seconds ago.  Since this file system can keep
-	  consistency like conventional LFS, it achieves quick recovery after
-	  system crashes.
-
-	  NILFS2 creates a number of checkpoints every few seconds or per
-	  synchronous write basis (unless there is no change).  Users can
-	  select significant versions among continuously created checkpoints,
-	  and can change them into snapshots which will be preserved for long
-	  periods until they are changed back to checkpoints.  Each
-	  snapshot is mountable as a read-only file system concurrently with
-	  its writable mount, and this feature is convenient for online backup.
-
-	  Some features including atime, extended attributes, and POSIX ACLs,
-	  are not supported yet.
-
-	  To compile this file system support as a module, choose M here: the
-	  module will be called nilfs2.  If unsure, say N.
+source "fs/nilfs2/Kconfig"
 
 endif # MISC_FILESYSTEMS
 
diff --git a/fs/afs/file.c b/fs/afs/file.c
index 0149dab365e7..681c2a7b013f 100644
--- a/fs/afs/file.c
+++ b/fs/afs/file.c
@@ -134,9 +134,16 @@ static int afs_readpage(struct file *file, struct page *page)
 
 	inode = page->mapping->host;
 
-	ASSERT(file != NULL);
-	key = file->private_data;
-	ASSERT(key != NULL);
+	if (file) {
+		key = file->private_data;
+		ASSERT(key != NULL);
+	} else {
+		key = afs_request_key(AFS_FS_S(inode->i_sb)->volume->cell);
+		if (IS_ERR(key)) {
+			ret = PTR_ERR(key);
+			goto error_nokey;
+		}
+	}
 
 	_enter("{%x},{%lu},{%lu}", key_serial(key), inode->i_ino, page->index);
 
@@ -207,12 +214,17 @@ static int afs_readpage(struct file *file, struct page *page)
 		unlock_page(page);
 	}
 
+	if (!file)
+		key_put(key);
 	_leave(" = 0");
 	return 0;
 
 error:
 	SetPageError(page);
 	unlock_page(page);
+	if (!file)
+		key_put(key);
+error_nokey:
 	_leave(" = %d", ret);
 	return ret;
 }
diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c
index 697f6b5f1313..e92f229e3c6e 100644
--- a/fs/binfmt_flat.c
+++ b/fs/binfmt_flat.c
@@ -828,15 +828,22 @@ static int load_flat_shared_library(int id, struct lib_info *libs)
 	if (IS_ERR(bprm.file))
 		return res;
 
+	bprm.cred = prepare_exec_creds();
+	res = -ENOMEM;
+	if (!bprm.cred)
+		goto out;
+
 	res = prepare_binprm(&bprm);
 
 	if (res <= (unsigned long)-4096)
 		res = load_flat_file(&bprm, libs, id, NULL);
-	if (bprm.file) {
-		allow_write_access(bprm.file);
-		fput(bprm.file);
-		bprm.file = NULL;
-	}
+
+	abort_creds(bprm.cred);
+
+out:
+	allow_write_access(bprm.file);
+	fput(bprm.file);
+
 	return(res);
 }
 
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 3a6d4fb2a329..94dfda24c06e 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -564,6 +564,16 @@ struct block_device *bdget(dev_t dev)
 
 EXPORT_SYMBOL(bdget);
 
+/**
+ * bdgrab -- Grab a reference to an already referenced block device
+ * @bdev:	Block device to grab a reference to.
+ */
+struct block_device *bdgrab(struct block_device *bdev)
+{
+	atomic_inc(&bdev->bd_inode->i_count);
+	return bdev;
+}
+
 long nr_blockdev_pages(void)
 {
 	struct block_device *bdev;
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 6e4f6c50a120..019e8af449ab 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -424,11 +424,11 @@ int btrfs_requeue_work(struct btrfs_work *work)
 	 * list
 	 */
 	if (worker->idle) {
-		spin_lock_irqsave(&worker->workers->lock, flags);
+		spin_lock(&worker->workers->lock);
 		worker->idle = 0;
 		list_move_tail(&worker->worker_list,
 			       &worker->workers->worker_list);
-		spin_unlock_irqrestore(&worker->workers->lock, flags);
+		spin_unlock(&worker->workers->lock);
 	}
 	if (!worker->working) {
 		wake = 1;
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 60a45f3a4e91..3fdcc0512d3a 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -557,19 +557,7 @@ static int comp_keys(struct btrfs_disk_key *disk, struct btrfs_key *k2)
 
 	btrfs_disk_key_to_cpu(&k1, disk);
 
-	if (k1.objectid > k2->objectid)
-		return 1;
-	if (k1.objectid < k2->objectid)
-		return -1;
-	if (k1.type > k2->type)
-		return 1;
-	if (k1.type < k2->type)
-		return -1;
-	if (k1.offset > k2->offset)
-		return 1;
-	if (k1.offset < k2->offset)
-		return -1;
-	return 0;
+	return btrfs_comp_cpu_keys(&k1, k2);
 }
 
 /*
@@ -1052,9 +1040,6 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 	    BTRFS_NODEPTRS_PER_BLOCK(root) / 4)
 		return 0;
 
-	if (btrfs_header_nritems(mid) > 2)
-		return 0;
-
 	if (btrfs_header_nritems(mid) < 2)
 		err_on_enospc = 1;
 
@@ -1701,6 +1686,7 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
 	struct extent_buffer *b;
 	int slot;
 	int ret;
+	int err;
 	int level;
 	int lowest_unlock = 1;
 	u8 lowest_level = 0;
@@ -1737,8 +1723,6 @@ again:
 			p->locks[level] = 1;
 
 		if (cow) {
-			int wret;
-
 			/*
 			 * if we don't really need to cow this block
 			 * then we don't want to set the path blocking,
@@ -1749,12 +1733,12 @@ again:
 
 			btrfs_set_path_blocking(p);
 
-			wret = btrfs_cow_block(trans, root, b,
-					       p->nodes[level + 1],
-					       p->slots[level + 1], &b);
-			if (wret) {
+			err = btrfs_cow_block(trans, root, b,
+					      p->nodes[level + 1],
+					      p->slots[level + 1], &b);
+			if (err) {
 				free_extent_buffer(b);
-				ret = wret;
+				ret = err;
 				goto done;
 			}
 		}
@@ -1793,41 +1777,45 @@ cow_done:
 		ret = bin_search(b, key, level, &slot);
 
 		if (level != 0) {
-			if (ret && slot > 0)
+			int dec = 0;
+			if (ret && slot > 0) {
+				dec = 1;
 				slot -= 1;
+			}
 			p->slots[level] = slot;
-			ret = setup_nodes_for_search(trans, root, p, b, level,
+			err = setup_nodes_for_search(trans, root, p, b, level,
 						     ins_len);
-			if (ret == -EAGAIN)
+			if (err == -EAGAIN)
 				goto again;
-			else if (ret)
+			if (err) {
+				ret = err;
 				goto done;
+			}
 			b = p->nodes[level];
 			slot = p->slots[level];
 
 			unlock_up(p, level, lowest_unlock);
 
-			/* this is only true while dropping a snapshot */
 			if (level == lowest_level) {
-				ret = 0;
+				if (dec)
+					p->slots[level]++;
 				goto done;
 			}
 
-			ret = read_block_for_search(trans, root, p,
+			err = read_block_for_search(trans, root, p,
 						    &b, level, slot, key);
-			if (ret == -EAGAIN)
+			if (err == -EAGAIN)
 				goto again;
-
-			if (ret == -EIO)
+			if (err) {
+				ret = err;
 				goto done;
+			}
 
 			if (!p->skip_locking) {
-				int lret;
-
 				btrfs_clear_path_blocking(p, NULL);
-				lret = btrfs_try_spin_lock(b);
+				err = btrfs_try_spin_lock(b);
 
-				if (!lret) {
+				if (!err) {
 					btrfs_set_path_blocking(p);
 					btrfs_tree_lock(b);
 					btrfs_clear_path_blocking(p, b);
@@ -1837,16 +1825,14 @@ cow_done:
 			p->slots[level] = slot;
 			if (ins_len > 0 &&
 			    btrfs_leaf_free_space(root, b) < ins_len) {
-				int sret;
-
 				btrfs_set_path_blocking(p);
-				sret = split_leaf(trans, root, key,
-						      p, ins_len, ret == 0);
+				err = split_leaf(trans, root, key,
+						 p, ins_len, ret == 0);
 				btrfs_clear_path_blocking(p, NULL);
 
-				BUG_ON(sret > 0);
-				if (sret) {
-					ret = sret;
+				BUG_ON(err > 0);
+				if (err) {
+					ret = err;
 					goto done;
 				}
 			}
@@ -3807,7 +3793,7 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		}
 
 		/* delete the leaf if it is mostly empty */
-		if (used < BTRFS_LEAF_DATA_SIZE(root) / 2) {
+		if (used < BTRFS_LEAF_DATA_SIZE(root) / 3) {
 			/* push_leaf_left fixes the path.
 			 * make sure the path still points to our leaf
 			 * for possible call to del_ptr below
@@ -4042,10 +4028,9 @@ out:
  * calling this function.
  */
 int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
-			struct btrfs_key *key, int lowest_level,
+			struct btrfs_key *key, int level,
 			int cache_only, u64 min_trans)
 {
-	int level = lowest_level;
 	int slot;
 	struct extent_buffer *c;
 
@@ -4058,11 +4043,40 @@ int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
 		c = path->nodes[level];
 next:
 		if (slot >= btrfs_header_nritems(c)) {
-			level++;
-			if (level == BTRFS_MAX_LEVEL)
+			int ret;
+			int orig_lowest;
+			struct btrfs_key cur_key;
+			if (level + 1 >= BTRFS_MAX_LEVEL ||
+			    !path->nodes[level + 1])
 				return 1;
-			continue;
+
+			if (path->locks[level + 1]) {
+				level++;
+				continue;
+			}
+
+			slot = btrfs_header_nritems(c) - 1;
+			if (level == 0)
+				btrfs_item_key_to_cpu(c, &cur_key, slot);
+			else
+				btrfs_node_key_to_cpu(c, &cur_key, slot);
+
+			orig_lowest = path->lowest_level;
+			btrfs_release_path(root, path);
+			path->lowest_level = level;
+			ret = btrfs_search_slot(NULL, root, &cur_key, path,
+						0, 0);
+			path->lowest_level = orig_lowest;
+			if (ret < 0)
+				return ret;
+
+			c = path->nodes[level];
+			slot = path->slots[level];
+			if (ret == 0)
+				slot++;
+			goto next;
 		}
+
 		if (level == 0)
 			btrfs_item_key_to_cpu(c, key, slot);
 		else {
@@ -4146,7 +4160,8 @@ again:
 	 * advance the path if there are now more items available.
 	 */
 	if (nritems > 0 && path->slots[0] < nritems - 1) {
-		path->slots[0]++;
+		if (ret == 0)
+			path->slots[0]++;
 		ret = 0;
 		goto done;
 	}
@@ -4278,10 +4293,10 @@ int btrfs_previous_item(struct btrfs_root *root,
 			path->slots[0]--;
 
 		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
-		if (found_key.type == type)
-			return 0;
 		if (found_key.objectid < min_objectid)
 			break;
+		if (found_key.type == type)
+			return 0;
 		if (found_key.objectid == min_objectid &&
 		    found_key.type < type)
 			break;
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 98a873838717..837435ce84ca 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -481,7 +481,7 @@ struct btrfs_shared_data_ref {
 
 struct btrfs_extent_inline_ref {
 	u8 type;
-	u64 offset;
+	__le64 offset;
 } __attribute__ ((__packed__));
 
 /* old style backrefs item */
@@ -689,6 +689,7 @@ struct btrfs_space_info {
 	struct list_head block_groups;
 	spinlock_t lock;
 	struct rw_semaphore groups_sem;
+	atomic_t caching_threads;
 };
 
 /*
@@ -707,6 +708,9 @@ struct btrfs_free_cluster {
 	/* first extent starting offset */
 	u64 window_start;
 
+	/* if this cluster simply points at a bitmap in the block group */
+	bool points_to_bitmap;
+
 	struct btrfs_block_group_cache *block_group;
 	/*
 	 * when a cluster is allocated from a block group, we put the
@@ -716,24 +720,37 @@ struct btrfs_free_cluster {
 	struct list_head block_group_list;
 };
 
+enum btrfs_caching_type {
+	BTRFS_CACHE_NO		= 0,
+	BTRFS_CACHE_STARTED	= 1,
+	BTRFS_CACHE_FINISHED	= 2,
+};
+
 struct btrfs_block_group_cache {
 	struct btrfs_key key;
 	struct btrfs_block_group_item item;
+	struct btrfs_fs_info *fs_info;
 	spinlock_t lock;
-	struct mutex cache_mutex;
 	u64 pinned;
 	u64 reserved;
 	u64 flags;
-	int cached;
+	u64 sectorsize;
+	int extents_thresh;
+	int free_extents;
+	int total_bitmaps;
 	int ro;
 	int dirty;
 
+	/* cache tracking stuff */
+	wait_queue_head_t caching_q;
+	int cached;
+
 	struct btrfs_space_info *space_info;
 
 	/* free space cache stuff */
 	spinlock_t tree_lock;
-	struct rb_root free_space_bytes;
 	struct rb_root free_space_offset;
+	u64 free_space;
 
 	/* block group cache stuff */
 	struct rb_node cache_node;
@@ -808,6 +825,7 @@ struct btrfs_fs_info {
 	struct mutex drop_mutex;
 	struct mutex volume_mutex;
 	struct mutex tree_reloc_mutex;
+	struct rw_semaphore extent_commit_sem;
 
 	/*
 	 * this protects the ordered operations list only while we are
@@ -1988,6 +2006,7 @@ void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
 				 u64 bytes);
 void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
 			      u64 bytes);
+void btrfs_free_pinned_extents(struct btrfs_fs_info *info);
 /* ctree.c */
 int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key,
 		     int level, int *slot);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index d28d29c95f7c..e83be2e4602c 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1639,6 +1639,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	mutex_init(&fs_info->cleaner_mutex);
 	mutex_init(&fs_info->volume_mutex);
 	mutex_init(&fs_info->tree_reloc_mutex);
+	init_rwsem(&fs_info->extent_commit_sem);
 
 	btrfs_init_free_cluster(&fs_info->meta_alloc_cluster);
 	btrfs_init_free_cluster(&fs_info->data_alloc_cluster);
@@ -1799,6 +1800,11 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 					   btrfs_super_chunk_root(disk_super),
 					   blocksize, generation);
 	BUG_ON(!chunk_root->node);
+	if (!test_bit(EXTENT_BUFFER_UPTODATE, &chunk_root->node->bflags)) {
+		printk(KERN_WARNING "btrfs: failed to read chunk root on %s\n",
+		       sb->s_id);
+		goto fail_chunk_root;
+	}
 	btrfs_set_root_node(&chunk_root->root_item, chunk_root->node);
 	chunk_root->commit_root = btrfs_root_node(chunk_root);
 
@@ -1826,6 +1832,11 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 					  blocksize, generation);
 	if (!tree_root->node)
 		goto fail_chunk_root;
+	if (!test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) {
+		printk(KERN_WARNING "btrfs: failed to read tree root on %s\n",
+		       sb->s_id);
+		goto fail_tree_root;
+	}
 	btrfs_set_root_node(&tree_root->root_item, tree_root->node);
 	tree_root->commit_root = btrfs_root_node(tree_root);
 
@@ -2322,6 +2333,9 @@ int close_ctree(struct btrfs_root *root)
 			printk(KERN_ERR "btrfs: commit super ret %d\n", ret);
 	}
 
+	fs_info->closing = 2;
+	smp_mb();
+
 	if (fs_info->delalloc_bytes) {
 		printk(KERN_INFO "btrfs: at unmount delalloc count %llu\n",
 		       (unsigned long long)fs_info->delalloc_bytes);
@@ -2343,6 +2357,7 @@ int close_ctree(struct btrfs_root *root)
 	free_extent_buffer(root->fs_info->csum_root->commit_root);
 
 	btrfs_free_block_groups(root->fs_info);
+	btrfs_free_pinned_extents(root->fs_info);
 
 	del_fs_roots(fs_info);
 
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index a5aca3997d42..72a2b9c28e9f 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -21,6 +21,7 @@
 #include <linux/blkdev.h>
 #include <linux/sort.h>
 #include <linux/rcupdate.h>
+#include <linux/kthread.h>
 #include "compat.h"
 #include "hash.h"
 #include "ctree.h"
@@ -61,6 +62,13 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force);
 
+static noinline int
+block_group_cache_done(struct btrfs_block_group_cache *cache)
+{
+	smp_mb();
+	return cache->cached == BTRFS_CACHE_FINISHED;
+}
+
 static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
 {
 	return (cache->flags & bits) == bits;
@@ -146,20 +154,70 @@ block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr,
 }
 
 /*
+ * We always set EXTENT_LOCKED for the super mirror extents so we don't
+ * overwrite them, so those bits need to be unset.  Also, if we are unmounting
+ * with pinned extents still sitting there because we had a block group caching,
+ * we need to clear those now, since we are done.
+ */
+void btrfs_free_pinned_extents(struct btrfs_fs_info *info)
+{
+	u64 start, end, last = 0;
+	int ret;
+
+	while (1) {
+		ret = find_first_extent_bit(&info->pinned_extents, last,
+					    &start, &end,
+					    EXTENT_LOCKED|EXTENT_DIRTY);
+		if (ret)
+			break;
+
+		clear_extent_bits(&info->pinned_extents, start, end,
+				  EXTENT_LOCKED|EXTENT_DIRTY, GFP_NOFS);
+		last = end+1;
+	}
+}
+
+static int remove_sb_from_cache(struct btrfs_root *root,
+				struct btrfs_block_group_cache *cache)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	u64 bytenr;
+	u64 *logical;
+	int stripe_len;
+	int i, nr, ret;
+
+	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
+		bytenr = btrfs_sb_offset(i);
+		ret = btrfs_rmap_block(&root->fs_info->mapping_tree,
+				       cache->key.objectid, bytenr,
+				       0, &logical, &nr, &stripe_len);
+		BUG_ON(ret);
+		while (nr--) {
+			try_lock_extent(&fs_info->pinned_extents,
+					logical[nr],
+					logical[nr] + stripe_len - 1, GFP_NOFS);
+		}
+		kfree(logical);
+	}
+
+	return 0;
+}
+
+/*
  * this is only called by cache_block_group, since we could have freed extents
  * we need to check the pinned_extents for any extents that can't be used yet
  * since their free space will be released as soon as the transaction commits.
  */
-static int add_new_free_space(struct btrfs_block_group_cache *block_group,
+static u64 add_new_free_space(struct btrfs_block_group_cache *block_group,
 			      struct btrfs_fs_info *info, u64 start, u64 end)
 {
-	u64 extent_start, extent_end, size;
+	u64 extent_start, extent_end, size, total_added = 0;
 	int ret;
 
 	while (start < end) {
 		ret = find_first_extent_bit(&info->pinned_extents, start,
 					    &extent_start, &extent_end,
-					    EXTENT_DIRTY);
+					    EXTENT_DIRTY|EXTENT_LOCKED);
 		if (ret)
 			break;
 
@@ -167,6 +225,7 @@ static int add_new_free_space(struct btrfs_block_group_cache *block_group,
 			start = extent_end + 1;
 		} else if (extent_start > start && extent_start < end) {
 			size = extent_start - start;
+			total_added += size;
 			ret = btrfs_add_free_space(block_group, start,
 						   size);
 			BUG_ON(ret);
@@ -178,84 +237,93 @@ static int add_new_free_space(struct btrfs_block_group_cache *block_group,
 
 	if (start < end) {
 		size = end - start;
+		total_added += size;
 		ret = btrfs_add_free_space(block_group, start, size);
 		BUG_ON(ret);
 	}
 
-	return 0;
-}
-
-static int remove_sb_from_cache(struct btrfs_root *root,
-				struct btrfs_block_group_cache *cache)
-{
-	u64 bytenr;
-	u64 *logical;
-	int stripe_len;
-	int i, nr, ret;
-
-	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
-		bytenr = btrfs_sb_offset(i);
-		ret = btrfs_rmap_block(&root->fs_info->mapping_tree,
-				       cache->key.objectid, bytenr, 0,
-				       &logical, &nr, &stripe_len);
-		BUG_ON(ret);
-		while (nr--) {
-			btrfs_remove_free_space(cache, logical[nr],
-						stripe_len);
-		}
-		kfree(logical);
-	}
-	return 0;
+	return total_added;
 }
 
-static int cache_block_group(struct btrfs_root *root,
-			     struct btrfs_block_group_cache *block_group)
+static int caching_kthread(void *data)
 {
+	struct btrfs_block_group_cache *block_group = data;
+	struct btrfs_fs_info *fs_info = block_group->fs_info;
+	u64 last = 0;
 	struct btrfs_path *path;
 	int ret = 0;
 	struct btrfs_key key;
 	struct extent_buffer *leaf;
 	int slot;
-	u64 last;
-
-	if (!block_group)
-		return 0;
-
-	root = root->fs_info->extent_root;
+	u64 total_found = 0;
 
-	if (block_group->cached)
-		return 0;
+	BUG_ON(!fs_info);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
-	path->reada = 2;
+	atomic_inc(&block_group->space_info->caching_threads);
+	last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
 	/*
-	 * we get into deadlocks with paths held by callers of this function.
-	 * since the alloc_mutex is protecting things right now, just
-	 * skip the locking here
+	 * We don't want to deadlock with somebody trying to allocate a new
+	 * extent for the extent root while also trying to search the extent
+	 * root to add free space.  So we skip locking and search the commit
+	 * root, since its read-only
 	 */
 	path->skip_locking = 1;
-	last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
+	path->search_commit_root = 1;
+	path->reada = 2;
+
 	key.objectid = last;
 	key.offset = 0;
 	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
-	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+again:
+	/* need to make sure the commit_root doesn't disappear */
+	down_read(&fs_info->extent_commit_sem);
+
+	ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
 	if (ret < 0)
 		goto err;
 
 	while (1) {
+		smp_mb();
+		if (block_group->fs_info->closing > 1) {
+			last = (u64)-1;
+			break;
+		}
+
 		leaf = path->nodes[0];
 		slot = path->slots[0];
 		if (slot >= btrfs_header_nritems(leaf)) {
-			ret = btrfs_next_leaf(root, path);
+			ret = btrfs_next_leaf(fs_info->extent_root, path);
 			if (ret < 0)
 				goto err;
-			if (ret == 0)
-				continue;
-			else
+			else if (ret)
 				break;
+
+			if (need_resched() ||
+			    btrfs_transaction_in_commit(fs_info)) {
+				leaf = path->nodes[0];
+
+				/* this shouldn't happen, but if the
+				 * leaf is empty just move on.
+				 */
+				if (btrfs_header_nritems(leaf) == 0)
+					break;
+				/*
+				 * we need to copy the key out so that
+				 * we are sure the next search advances
+				 * us forward in the btree.
+				 */
+				btrfs_item_key_to_cpu(leaf, &key, 0);
+				btrfs_release_path(fs_info->extent_root, path);
+				up_read(&fs_info->extent_commit_sem);
+				schedule_timeout(1);
+				goto again;
+			}
+
+			continue;
 		}
 		btrfs_item_key_to_cpu(leaf, &key, slot);
 		if (key.objectid < block_group->key.objectid)
@@ -266,24 +334,59 @@ static int cache_block_group(struct btrfs_root *root,
 			break;
 
 		if (btrfs_key_type(&key) == BTRFS_EXTENT_ITEM_KEY) {
-			add_new_free_space(block_group, root->fs_info, last,
-					   key.objectid);
-
+			total_found += add_new_free_space(block_group,
+							  fs_info, last,
+							  key.objectid);
 			last = key.objectid + key.offset;
 		}
+
+		if (total_found > (1024 * 1024 * 2)) {
+			total_found = 0;
+			wake_up(&block_group->caching_q);
+		}
 next:
 		path->slots[0]++;
 	}
+	ret = 0;
 
-	add_new_free_space(block_group, root->fs_info, last,
-			   block_group->key.objectid +
-			   block_group->key.offset);
+	total_found += add_new_free_space(block_group, fs_info, last,
+					  block_group->key.objectid +
+					  block_group->key.offset);
+
+	spin_lock(&block_group->lock);
+	block_group->cached = BTRFS_CACHE_FINISHED;
+	spin_unlock(&block_group->lock);
 
-	block_group->cached = 1;
-	remove_sb_from_cache(root, block_group);
-	ret = 0;
 err:
 	btrfs_free_path(path);
+	up_read(&fs_info->extent_commit_sem);
+	atomic_dec(&block_group->space_info->caching_threads);
+	wake_up(&block_group->caching_q);
+
+	return 0;
+}
+
+static int cache_block_group(struct btrfs_block_group_cache *cache)
+{
+	struct task_struct *tsk;
+	int ret = 0;
+
+	spin_lock(&cache->lock);
+	if (cache->cached != BTRFS_CACHE_NO) {
+		spin_unlock(&cache->lock);
+		return ret;
+	}
+	cache->cached = BTRFS_CACHE_STARTED;
+	spin_unlock(&cache->lock);
+
+	tsk = kthread_run(caching_kthread, cache, "btrfs-cache-%llu\n",
+			  cache->key.objectid);
+	if (IS_ERR(tsk)) {
+		ret = PTR_ERR(tsk);
+		printk(KERN_ERR "error running thread %d\n", ret);
+		BUG();
+	}
+
 	return ret;
 }
 
@@ -2387,13 +2490,29 @@ fail:
 
 }
 
+static struct btrfs_block_group_cache *
+next_block_group(struct btrfs_root *root,
+		 struct btrfs_block_group_cache *cache)
+{
+	struct rb_node *node;
+	spin_lock(&root->fs_info->block_group_cache_lock);
+	node = rb_next(&cache->cache_node);
+	btrfs_put_block_group(cache);
+	if (node) {
+		cache = rb_entry(node, struct btrfs_block_group_cache,
+				 cache_node);
+		atomic_inc(&cache->count);
+	} else
+		cache = NULL;
+	spin_unlock(&root->fs_info->block_group_cache_lock);
+	return cache;
+}
+
 int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root)
 {
-	struct btrfs_block_group_cache *cache, *entry;
-	struct rb_node *n;
+	struct btrfs_block_group_cache *cache;
 	int err = 0;
-	int werr = 0;
 	struct btrfs_path *path;
 	u64 last = 0;
 
@@ -2402,39 +2521,35 @@ int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 
 	while (1) {
-		cache = NULL;
-		spin_lock(&root->fs_info->block_group_cache_lock);
-		for (n = rb_first(&root->fs_info->block_group_cache_tree);
-		     n; n = rb_next(n)) {
-			entry = rb_entry(n, struct btrfs_block_group_cache,
-					 cache_node);
-			if (entry->dirty) {
-				cache = entry;
-				break;
-			}
+		if (last == 0) {
+			err = btrfs_run_delayed_refs(trans, root,
+						     (unsigned long)-1);
+			BUG_ON(err);
 		}
-		spin_unlock(&root->fs_info->block_group_cache_lock);
 
-		if (!cache)
-			break;
+		cache = btrfs_lookup_first_block_group(root->fs_info, last);
+		while (cache) {
+			if (cache->dirty)
+				break;
+			cache = next_block_group(root, cache);
+		}
+		if (!cache) {
+			if (last == 0)
+				break;
+			last = 0;
+			continue;
+		}
 
 		cache->dirty = 0;
-		last += cache->key.offset;
+		last = cache->key.objectid + cache->key.offset;
 
-		err = write_one_cache_group(trans, root,
-					    path, cache);
-		/*
-		 * if we fail to write the cache group, we want
-		 * to keep it marked dirty in hopes that a later
-		 * write will work
-		 */
-		if (err) {
-			werr = err;
-			continue;
-		}
+		err = write_one_cache_group(trans, root, path, cache);
+		BUG_ON(err);
+		btrfs_put_block_group(cache);
 	}
+
 	btrfs_free_path(path);
-	return werr;
+	return 0;
 }
 
 int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr)
@@ -2484,6 +2599,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	found->force_alloc = 0;
 	*space_info = found;
 	list_add_rcu(&found->list, &info->space_info);
+	atomic_set(&found->caching_threads, 0);
 	return 0;
 }
 
@@ -2947,13 +3063,9 @@ int btrfs_update_pinned_extents(struct btrfs_root *root,
 	struct btrfs_block_group_cache *cache;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	if (pin) {
+	if (pin)
 		set_extent_dirty(&fs_info->pinned_extents,
 				bytenr, bytenr + num - 1, GFP_NOFS);
-	} else {
-		clear_extent_dirty(&fs_info->pinned_extents,
-				bytenr, bytenr + num - 1, GFP_NOFS);
-	}
 
 	while (num > 0) {
 		cache = btrfs_lookup_block_group(fs_info, bytenr);
@@ -2969,14 +3081,34 @@ int btrfs_update_pinned_extents(struct btrfs_root *root,
 			spin_unlock(&cache->space_info->lock);
 			fs_info->total_pinned += len;
 		} else {
+			int unpin = 0;
+
+			/*
+			 * in order to not race with the block group caching, we
+			 * only want to unpin the extent if we are cached.  If
+			 * we aren't cached, we want to start async caching this
+			 * block group so we can free the extent the next time
+			 * around.
+			 */
 			spin_lock(&cache->space_info->lock);
 			spin_lock(&cache->lock);
-			cache->pinned -= len;
-			cache->space_info->bytes_pinned -= len;
+			unpin = (cache->cached == BTRFS_CACHE_FINISHED);
+			if (likely(unpin)) {
+				cache->pinned -= len;
+				cache->space_info->bytes_pinned -= len;
+				fs_info->total_pinned -= len;
+			}
 			spin_unlock(&cache->lock);
 			spin_unlock(&cache->space_info->lock);
-			fs_info->total_pinned -= len;
-			if (cache->cached)
+
+			if (likely(unpin))
+				clear_extent_dirty(&fs_info->pinned_extents,
+						   bytenr, bytenr + len -1,
+						   GFP_NOFS);
+			else
+				cache_block_group(cache);
+
+			if (unpin)
 				btrfs_add_free_space(cache, bytenr, len);
 		}
 		btrfs_put_block_group(cache);
@@ -3030,6 +3162,7 @@ int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy)
 					    &start, &end, EXTENT_DIRTY);
 		if (ret)
 			break;
+
 		set_extent_dirty(copy, start, end, GFP_NOFS);
 		last = end + 1;
 	}
@@ -3058,6 +3191,7 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 
 		cond_resched();
 	}
+
 	return ret;
 }
 
@@ -3436,6 +3570,45 @@ static u64 stripe_align(struct btrfs_root *root, u64 val)
 }
 
 /*
+ * when we wait for progress in the block group caching, its because
+ * our allocation attempt failed at least once.  So, we must sleep
+ * and let some progress happen before we try again.
+ *
+ * This function will sleep at least once waiting for new free space to
+ * show up, and then it will check the block group free space numbers
+ * for our min num_bytes.  Another option is to have it go ahead
+ * and look in the rbtree for a free extent of a given size, but this
+ * is a good start.
+ */
+static noinline int
+wait_block_group_cache_progress(struct btrfs_block_group_cache *cache,
+				u64 num_bytes)
+{
+	DEFINE_WAIT(wait);
+
+	prepare_to_wait(&cache->caching_q, &wait, TASK_UNINTERRUPTIBLE);
+
+	if (block_group_cache_done(cache)) {
+		finish_wait(&cache->caching_q, &wait);
+		return 0;
+	}
+	schedule();
+	finish_wait(&cache->caching_q, &wait);
+
+	wait_event(cache->caching_q, block_group_cache_done(cache) ||
+		   (cache->free_space >= num_bytes));
+	return 0;
+}
+
+enum btrfs_loop_type {
+	LOOP_CACHED_ONLY = 0,
+	LOOP_CACHING_NOWAIT = 1,
+	LOOP_CACHING_WAIT = 2,
+	LOOP_ALLOC_CHUNK = 3,
+	LOOP_NO_EMPTY_SIZE = 4,
+};
+
+/*
  * walks the btree of allocated extents and find a hole of a given size.
  * The key ins is changed to record the hole:
  * ins->objectid == block start
@@ -3460,6 +3633,7 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 	struct btrfs_space_info *space_info;
 	int last_ptr_loop = 0;
 	int loop = 0;
+	bool found_uncached_bg = false;
 
 	WARN_ON(num_bytes < root->sectorsize);
 	btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
@@ -3491,15 +3665,18 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 	search_start = max(search_start, first_logical_byte(root, 0));
 	search_start = max(search_start, hint_byte);
 
-	if (!last_ptr) {
+	if (!last_ptr)
 		empty_cluster = 0;
-		loop = 1;
-	}
 
 	if (search_start == hint_byte) {
 		block_group = btrfs_lookup_block_group(root->fs_info,
 						       search_start);
-		if (block_group && block_group_bits(block_group, data)) {
+		/*
+		 * we don't want to use the block group if it doesn't match our
+		 * allocation bits, or if its not cached.
+		 */
+		if (block_group && block_group_bits(block_group, data) &&
+		    block_group_cache_done(block_group)) {
 			down_read(&space_info->groups_sem);
 			if (list_empty(&block_group->list) ||
 			    block_group->ro) {
@@ -3522,21 +3699,35 @@ search:
 	down_read(&space_info->groups_sem);
 	list_for_each_entry(block_group, &space_info->block_groups, list) {
 		u64 offset;
+		int cached;
 
 		atomic_inc(&block_group->count);
 		search_start = block_group->key.objectid;
 
 have_block_group:
-		if (unlikely(!block_group->cached)) {
-			mutex_lock(&block_group->cache_mutex);
-			ret = cache_block_group(root, block_group);
-			mutex_unlock(&block_group->cache_mutex);
-			if (ret) {
-				btrfs_put_block_group(block_group);
-				break;
+		if (unlikely(block_group->cached == BTRFS_CACHE_NO)) {
+			/*
+			 * we want to start caching kthreads, but not too many
+			 * right off the bat so we don't overwhelm the system,
+			 * so only start them if there are less than 2 and we're
+			 * in the initial allocation phase.
+			 */
+			if (loop > LOOP_CACHING_NOWAIT ||
+			    atomic_read(&space_info->caching_threads) < 2) {
+				ret = cache_block_group(block_group);
+				BUG_ON(ret);
 			}
 		}
 
+		cached = block_group_cache_done(block_group);
+		if (unlikely(!cached)) {
+			found_uncached_bg = true;
+
+			/* if we only want cached bgs, loop */
+			if (loop == LOOP_CACHED_ONLY)
+				goto loop;
+		}
+
 		if (unlikely(block_group->ro))
 			goto loop;
 
@@ -3615,14 +3806,21 @@ refill_cluster:
 					spin_unlock(&last_ptr->refill_lock);
 					goto checks;
 				}
+			} else if (!cached && loop > LOOP_CACHING_NOWAIT) {
+				spin_unlock(&last_ptr->refill_lock);
+
+				wait_block_group_cache_progress(block_group,
+				       num_bytes + empty_cluster + empty_size);
+				goto have_block_group;
 			}
+
 			/*
 			 * at this point we either didn't find a cluster
 			 * or we weren't able to allocate a block from our
 			 * cluster.  Free the cluster we've been trying
 			 * to use, and go to the next block group
 			 */
-			if (loop < 2) {
+			if (loop < LOOP_NO_EMPTY_SIZE) {
 				btrfs_return_cluster_to_free_space(NULL,
 								   last_ptr);
 				spin_unlock(&last_ptr->refill_lock);
@@ -3633,11 +3831,17 @@ refill_cluster:
 
 		offset = btrfs_find_space_for_alloc(block_group, search_start,
 						    num_bytes, empty_size);
-		if (!offset)
+		if (!offset && (cached || (!cached &&
+					   loop == LOOP_CACHING_NOWAIT))) {
 			goto loop;
+		} else if (!offset && (!cached &&
+				       loop > LOOP_CACHING_NOWAIT)) {
+			wait_block_group_cache_progress(block_group,
+					num_bytes + empty_size);
+			goto have_block_group;
+		}
 checks:
 		search_start = stripe_align(root, offset);
-
 		/* move on to the next group */
 		if (search_start + num_bytes >= search_end) {
 			btrfs_add_free_space(block_group, offset, num_bytes);
@@ -3683,13 +3887,26 @@ loop:
 	}
 	up_read(&space_info->groups_sem);
 
-	/* loop == 0, try to find a clustered alloc in every block group
-	 * loop == 1, try again after forcing a chunk allocation
-	 * loop == 2, set empty_size and empty_cluster to 0 and try again
+	/* LOOP_CACHED_ONLY, only search fully cached block groups
+	 * LOOP_CACHING_NOWAIT, search partially cached block groups, but
+	 *			dont wait foR them to finish caching
+	 * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching
+	 * LOOP_ALLOC_CHUNK, force a chunk allocation and try again
+	 * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try
+	 *			again
 	 */
-	if (!ins->objectid && loop < 3 &&
-	    (empty_size || empty_cluster || allowed_chunk_alloc)) {
-		if (loop >= 2) {
+	if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE &&
+	    (found_uncached_bg || empty_size || empty_cluster ||
+	     allowed_chunk_alloc)) {
+		if (found_uncached_bg) {
+			found_uncached_bg = false;
+			if (loop < LOOP_CACHING_WAIT) {
+				loop++;
+				goto search;
+			}
+		}
+
+		if (loop == LOOP_ALLOC_CHUNK) {
 			empty_size = 0;
 			empty_cluster = 0;
 		}
@@ -3702,7 +3919,7 @@ loop:
 			space_info->force_alloc = 1;
 		}
 
-		if (loop < 3) {
+		if (loop < LOOP_NO_EMPTY_SIZE) {
 			loop++;
 			goto search;
 		}
@@ -3798,7 +4015,7 @@ again:
 			       num_bytes, data, 1);
 		goto again;
 	}
-	if (ret) {
+	if (ret == -ENOSPC) {
 		struct btrfs_space_info *sinfo;
 
 		sinfo = __find_space_info(root->fs_info, data);
@@ -3806,7 +4023,6 @@ again:
 		       "wanted %llu\n", (unsigned long long)data,
 		       (unsigned long long)num_bytes);
 		dump_space_info(sinfo, num_bytes);
-		BUG();
 	}
 
 	return ret;
@@ -3844,7 +4060,9 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 	ret = __btrfs_reserve_extent(trans, root, num_bytes, min_alloc_size,
 				     empty_size, hint_byte, search_end, ins,
 				     data);
-	update_reserved_extents(root, ins->objectid, ins->offset, 1);
+	if (!ret)
+		update_reserved_extents(root, ins->objectid, ins->offset, 1);
+
 	return ret;
 }
 
@@ -4006,9 +4224,9 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
 	struct btrfs_block_group_cache *block_group;
 
 	block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
-	mutex_lock(&block_group->cache_mutex);
-	cache_block_group(root, block_group);
-	mutex_unlock(&block_group->cache_mutex);
+	cache_block_group(block_group);
+	wait_event(block_group->caching_q,
+		   block_group_cache_done(block_group));
 
 	ret = btrfs_remove_free_space(block_group, ins->objectid,
 				      ins->offset);
@@ -4039,7 +4257,8 @@ static int alloc_tree_block(struct btrfs_trans_handle *trans,
 	ret = __btrfs_reserve_extent(trans, root, num_bytes, num_bytes,
 				     empty_size, hint_byte, search_end,
 				     ins, 0);
-	BUG_ON(ret);
+	if (ret)
+		return ret;
 
 	if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
 		if (parent == 0)
@@ -6955,11 +7174,16 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 			 &info->block_group_cache_tree);
 		spin_unlock(&info->block_group_cache_lock);
 
-		btrfs_remove_free_space_cache(block_group);
 		down_write(&block_group->space_info->groups_sem);
 		list_del(&block_group->list);
 		up_write(&block_group->space_info->groups_sem);
 
+		if (block_group->cached == BTRFS_CACHE_STARTED)
+			wait_event(block_group->caching_q,
+				   block_group_cache_done(block_group));
+
+		btrfs_remove_free_space_cache(block_group);
+
 		WARN_ON(atomic_read(&block_group->count) != 1);
 		kfree(block_group);
 
@@ -7025,9 +7249,19 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		atomic_set(&cache->count, 1);
 		spin_lock_init(&cache->lock);
 		spin_lock_init(&cache->tree_lock);
-		mutex_init(&cache->cache_mutex);
+		cache->fs_info = info;
+		init_waitqueue_head(&cache->caching_q);
 		INIT_LIST_HEAD(&cache->list);
 		INIT_LIST_HEAD(&cache->cluster_list);
+
+		/*
+		 * we only want to have 32k of ram per block group for keeping
+		 * track of free space, and if we pass 1/2 of that we want to
+		 * start converting things over to using bitmaps
+		 */
+		cache->extents_thresh = ((1024 * 32) / 2) /
+			sizeof(struct btrfs_free_space);
+
 		read_extent_buffer(leaf, &cache->item,
 				   btrfs_item_ptr_offset(leaf, path->slots[0]),
 				   sizeof(cache->item));
@@ -7036,6 +7270,26 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		key.objectid = found_key.objectid + found_key.offset;
 		btrfs_release_path(root, path);
 		cache->flags = btrfs_block_group_flags(&cache->item);
+		cache->sectorsize = root->sectorsize;
+
+		remove_sb_from_cache(root, cache);
+
+		/*
+		 * check for two cases, either we are full, and therefore
+		 * don't need to bother with the caching work since we won't
+		 * find any space, or we are empty, and we can just add all
+		 * the space in and be done with it.  This saves us _alot_ of
+		 * time, particularly in the full case.
+		 */
+		if (found_key.offset == btrfs_block_group_used(&cache->item)) {
+			cache->cached = BTRFS_CACHE_FINISHED;
+		} else if (btrfs_block_group_used(&cache->item) == 0) {
+			cache->cached = BTRFS_CACHE_FINISHED;
+			add_new_free_space(cache, root->fs_info,
+					   found_key.objectid,
+					   found_key.objectid +
+					   found_key.offset);
+		}
 
 		ret = update_space_info(info, cache->flags, found_key.offset,
 					btrfs_block_group_used(&cache->item),
@@ -7079,10 +7333,19 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	cache->key.objectid = chunk_offset;
 	cache->key.offset = size;
 	cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
+	cache->sectorsize = root->sectorsize;
+
+	/*
+	 * we only want to have 32k of ram per block group for keeping track
+	 * of free space, and if we pass 1/2 of that we want to start
+	 * converting things over to using bitmaps
+	 */
+	cache->extents_thresh = ((1024 * 32) / 2) /
+		sizeof(struct btrfs_free_space);
 	atomic_set(&cache->count, 1);
 	spin_lock_init(&cache->lock);
 	spin_lock_init(&cache->tree_lock);
-	mutex_init(&cache->cache_mutex);
+	init_waitqueue_head(&cache->caching_q);
 	INIT_LIST_HEAD(&cache->list);
 	INIT_LIST_HEAD(&cache->cluster_list);
 
@@ -7091,6 +7354,12 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	cache->flags = type;
 	btrfs_set_block_group_flags(&cache->item, type);
 
+	cache->cached = BTRFS_CACHE_FINISHED;
+	remove_sb_from_cache(root, cache);
+
+	add_new_free_space(cache, root->fs_info, chunk_offset,
+			   chunk_offset + size);
+
 	ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
 				&cache->space_info);
 	BUG_ON(ret);
@@ -7149,7 +7418,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	rb_erase(&block_group->cache_node,
 		 &root->fs_info->block_group_cache_tree);
 	spin_unlock(&root->fs_info->block_group_cache_lock);
-	btrfs_remove_free_space_cache(block_group);
+
 	down_write(&block_group->space_info->groups_sem);
 	/*
 	 * we must use list_del_init so people can check to see if they
@@ -7158,11 +7427,18 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	list_del_init(&block_group->list);
 	up_write(&block_group->space_info->groups_sem);
 
+	if (block_group->cached == BTRFS_CACHE_STARTED)
+		wait_event(block_group->caching_q,
+			   block_group_cache_done(block_group));
+
+	btrfs_remove_free_space_cache(block_group);
+
 	spin_lock(&block_group->space_info->lock);
 	block_group->space_info->total_bytes -= block_group->key.offset;
 	block_group->space_info->bytes_readonly -= block_group->key.offset;
 	spin_unlock(&block_group->space_info->lock);
-	block_group->space_info->full = 0;
+
+	btrfs_clear_space_info_full(root->fs_info);
 
 	btrfs_put_block_group(block_group);
 	btrfs_put_block_group(block_group);
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 4538e48581a5..5edcee3a617f 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -16,45 +16,46 @@
  * Boston, MA 021110-1307, USA.
  */
 
+#include <linux/pagemap.h>
 #include <linux/sched.h>
+#include <linux/math64.h>
 #include "ctree.h"
 #include "free-space-cache.h"
 #include "transaction.h"
 
-struct btrfs_free_space {
-	struct rb_node bytes_index;
-	struct rb_node offset_index;
-	u64 offset;
-	u64 bytes;
-};
+#define BITS_PER_BITMAP		(PAGE_CACHE_SIZE * 8)
+#define MAX_CACHE_BYTES_PER_GIG	(32 * 1024)
 
-static int tree_insert_offset(struct rb_root *root, u64 offset,
-			      struct rb_node *node)
+static inline unsigned long offset_to_bit(u64 bitmap_start, u64 sectorsize,
+					  u64 offset)
 {
-	struct rb_node **p = &root->rb_node;
-	struct rb_node *parent = NULL;
-	struct btrfs_free_space *info;
+	BUG_ON(offset < bitmap_start);
+	offset -= bitmap_start;
+	return (unsigned long)(div64_u64(offset, sectorsize));
+}
 
-	while (*p) {
-		parent = *p;
-		info = rb_entry(parent, struct btrfs_free_space, offset_index);
+static inline unsigned long bytes_to_bits(u64 bytes, u64 sectorsize)
+{
+	return (unsigned long)(div64_u64(bytes, sectorsize));
+}
 
-		if (offset < info->offset)
-			p = &(*p)->rb_left;
-		else if (offset > info->offset)
-			p = &(*p)->rb_right;
-		else
-			return -EEXIST;
-	}
+static inline u64 offset_to_bitmap(struct btrfs_block_group_cache *block_group,
+				   u64 offset)
+{
+	u64 bitmap_start;
+	u64 bytes_per_bitmap;
 
-	rb_link_node(node, parent, p);
-	rb_insert_color(node, root);
+	bytes_per_bitmap = BITS_PER_BITMAP * block_group->sectorsize;
+	bitmap_start = offset - block_group->key.objectid;
+	bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap);
+	bitmap_start *= bytes_per_bitmap;
+	bitmap_start += block_group->key.objectid;
 
-	return 0;
+	return bitmap_start;
 }
 
-static int tree_insert_bytes(struct rb_root *root, u64 bytes,
-			     struct rb_node *node)
+static int tree_insert_offset(struct rb_root *root, u64 offset,
+			      struct rb_node *node, int bitmap)
 {
 	struct rb_node **p = &root->rb_node;
 	struct rb_node *parent = NULL;
@@ -62,12 +63,34 @@ static int tree_insert_bytes(struct rb_root *root, u64 bytes,
 
 	while (*p) {
 		parent = *p;
-		info = rb_entry(parent, struct btrfs_free_space, bytes_index);
+		info = rb_entry(parent, struct btrfs_free_space, offset_index);
 
-		if (bytes < info->bytes)
+		if (offset < info->offset) {
 			p = &(*p)->rb_left;
-		else
+		} else if (offset > info->offset) {
 			p = &(*p)->rb_right;
+		} else {
+			/*
+			 * we could have a bitmap entry and an extent entry
+			 * share the same offset.  If this is the case, we want
+			 * the extent entry to always be found first if we do a
+			 * linear search through the tree, since we want to have
+			 * the quickest allocation time, and allocating from an
+			 * extent is faster than allocating from a bitmap.  So
+			 * if we're inserting a bitmap and we find an entry at
+			 * this offset, we want to go right, or after this entry
+			 * logically.  If we are inserting an extent and we've
+			 * found a bitmap, we want to go left, or before
+			 * logically.
+			 */
+			if (bitmap) {
+				WARN_ON(info->bitmap);
+				p = &(*p)->rb_right;
+			} else {
+				WARN_ON(!info->bitmap);
+				p = &(*p)->rb_left;
+			}
+		}
 	}
 
 	rb_link_node(node, parent, p);
@@ -79,110 +102,143 @@ static int tree_insert_bytes(struct rb_root *root, u64 bytes,
 /*
  * searches the tree for the given offset.
  *
- * fuzzy == 1: this is used for allocations where we are given a hint of where
- * to look for free space.  Because the hint may not be completely on an offset
- * mark, or the hint may no longer point to free space we need to fudge our
- * results a bit.  So we look for free space starting at or after offset with at
- * least bytes size.  We prefer to find as close to the given offset as we can.
- * Also if the offset is within a free space range, then we will return the free
- * space that contains the given offset, which means we can return a free space
- * chunk with an offset before the provided offset.
- *
- * fuzzy == 0: this is just a normal tree search.  Give us the free space that
- * starts at the given offset which is at least bytes size, and if its not there
- * return NULL.
+ * fuzzy - If this is set, then we are trying to make an allocation, and we just
+ * want a section that has at least bytes size and comes at or after the given
+ * offset.
  */
-static struct btrfs_free_space *tree_search_offset(struct rb_root *root,
-						   u64 offset, u64 bytes,
-						   int fuzzy)
+static struct btrfs_free_space *
+tree_search_offset(struct btrfs_block_group_cache *block_group,
+		   u64 offset, int bitmap_only, int fuzzy)
 {
-	struct rb_node *n = root->rb_node;
-	struct btrfs_free_space *entry, *ret = NULL;
+	struct rb_node *n = block_group->free_space_offset.rb_node;
+	struct btrfs_free_space *entry, *prev = NULL;
+
+	/* find entry that is closest to the 'offset' */
+	while (1) {
+		if (!n) {
+			entry = NULL;
+			break;
+		}
 
-	while (n) {
 		entry = rb_entry(n, struct btrfs_free_space, offset_index);
+		prev = entry;
 
-		if (offset < entry->offset) {
-			if (fuzzy &&
-			    (!ret || entry->offset < ret->offset) &&
-			    (bytes <= entry->bytes))
-				ret = entry;
+		if (offset < entry->offset)
 			n = n->rb_left;
-		} else if (offset > entry->offset) {
-			if (fuzzy &&
-			    (entry->offset + entry->bytes - 1) >= offset &&
-			    bytes <= entry->bytes) {
-				ret = entry;
-				break;
-			}
+		else if (offset > entry->offset)
 			n = n->rb_right;
-		} else {
-			if (bytes > entry->bytes) {
-				n = n->rb_right;
-				continue;
-			}
-			ret = entry;
+		else
 			break;
-		}
 	}
 
-	return ret;
-}
+	if (bitmap_only) {
+		if (!entry)
+			return NULL;
+		if (entry->bitmap)
+			return entry;
 
-/*
- * return a chunk at least bytes size, as close to offset that we can get.
- */
-static struct btrfs_free_space *tree_search_bytes(struct rb_root *root,
-						  u64 offset, u64 bytes)
-{
-	struct rb_node *n = root->rb_node;
-	struct btrfs_free_space *entry, *ret = NULL;
-
-	while (n) {
-		entry = rb_entry(n, struct btrfs_free_space, bytes_index);
+		/*
+		 * bitmap entry and extent entry may share same offset,
+		 * in that case, bitmap entry comes after extent entry.
+		 */
+		n = rb_next(n);
+		if (!n)
+			return NULL;
+		entry = rb_entry(n, struct btrfs_free_space, offset_index);
+		if (entry->offset != offset)
+			return NULL;
 
-		if (bytes < entry->bytes) {
+		WARN_ON(!entry->bitmap);
+		return entry;
+	} else if (entry) {
+		if (entry->bitmap) {
 			/*
-			 * We prefer to get a hole size as close to the size we
-			 * are asking for so we don't take small slivers out of
-			 * huge holes, but we also want to get as close to the
-			 * offset as possible so we don't have a whole lot of
-			 * fragmentation.
+			 * if previous extent entry covers the offset,
+			 * we should return it instead of the bitmap entry
 			 */
-			if (offset <= entry->offset) {
-				if (!ret)
-					ret = entry;
-				else if (entry->bytes < ret->bytes)
-					ret = entry;
-				else if (entry->offset < ret->offset)
-					ret = entry;
+			n = &entry->offset_index;
+			while (1) {
+				n = rb_prev(n);
+				if (!n)
+					break;
+				prev = rb_entry(n, struct btrfs_free_space,
+						offset_index);
+				if (!prev->bitmap) {
+					if (prev->offset + prev->bytes > offset)
+						entry = prev;
+					break;
+				}
 			}
-			n = n->rb_left;
-		} else if (bytes > entry->bytes) {
-			n = n->rb_right;
+		}
+		return entry;
+	}
+
+	if (!prev)
+		return NULL;
+
+	/* find last entry before the 'offset' */
+	entry = prev;
+	if (entry->offset > offset) {
+		n = rb_prev(&entry->offset_index);
+		if (n) {
+			entry = rb_entry(n, struct btrfs_free_space,
+					offset_index);
+			BUG_ON(entry->offset > offset);
 		} else {
-			/*
-			 * Ok we may have multiple chunks of the wanted size,
-			 * so we don't want to take the first one we find, we
-			 * want to take the one closest to our given offset, so
-			 * keep searching just in case theres a better match.
-			 */
-			n = n->rb_right;
-			if (offset > entry->offset)
-				continue;
-			else if (!ret || entry->offset < ret->offset)
-				ret = entry;
+			if (fuzzy)
+				return entry;
+			else
+				return NULL;
 		}
 	}
 
-	return ret;
+	if (entry->bitmap) {
+		n = &entry->offset_index;
+		while (1) {
+			n = rb_prev(n);
+			if (!n)
+				break;
+			prev = rb_entry(n, struct btrfs_free_space,
+					offset_index);
+			if (!prev->bitmap) {
+				if (prev->offset + prev->bytes > offset)
+					return prev;
+				break;
+			}
+		}
+		if (entry->offset + BITS_PER_BITMAP *
+		    block_group->sectorsize > offset)
+			return entry;
+	} else if (entry->offset + entry->bytes > offset)
+		return entry;
+
+	if (!fuzzy)
+		return NULL;
+
+	while (1) {
+		if (entry->bitmap) {
+			if (entry->offset + BITS_PER_BITMAP *
+			    block_group->sectorsize > offset)
+				break;
+		} else {
+			if (entry->offset + entry->bytes > offset)
+				break;
+		}
+
+		n = rb_next(&entry->offset_index);
+		if (!n)
+			return NULL;
+		entry = rb_entry(n, struct btrfs_free_space, offset_index);
+	}
+	return entry;
 }
 
 static void unlink_free_space(struct btrfs_block_group_cache *block_group,
 			      struct btrfs_free_space *info)
 {
 	rb_erase(&info->offset_index, &block_group->free_space_offset);
-	rb_erase(&info->bytes_index, &block_group->free_space_bytes);
+	block_group->free_extents--;
+	block_group->free_space -= info->bytes;
 }
 
 static int link_free_space(struct btrfs_block_group_cache *block_group,
@@ -190,17 +246,353 @@ static int link_free_space(struct btrfs_block_group_cache *block_group,
 {
 	int ret = 0;
 
-
-	BUG_ON(!info->bytes);
+	BUG_ON(!info->bitmap && !info->bytes);
 	ret = tree_insert_offset(&block_group->free_space_offset, info->offset,
-				 &info->offset_index);
+				 &info->offset_index, (info->bitmap != NULL));
 	if (ret)
 		return ret;
 
-	ret = tree_insert_bytes(&block_group->free_space_bytes, info->bytes,
-				&info->bytes_index);
-	if (ret)
-		return ret;
+	block_group->free_space += info->bytes;
+	block_group->free_extents++;
+	return ret;
+}
+
+static void recalculate_thresholds(struct btrfs_block_group_cache *block_group)
+{
+	u64 max_bytes, possible_bytes;
+
+	/*
+	 * The goal is to keep the total amount of memory used per 1gb of space
+	 * at or below 32k, so we need to adjust how much memory we allow to be
+	 * used by extent based free space tracking
+	 */
+	max_bytes = MAX_CACHE_BYTES_PER_GIG *
+		(div64_u64(block_group->key.offset, 1024 * 1024 * 1024));
+
+	possible_bytes = (block_group->total_bitmaps * PAGE_CACHE_SIZE) +
+		(sizeof(struct btrfs_free_space) *
+		 block_group->extents_thresh);
+
+	if (possible_bytes > max_bytes) {
+		int extent_bytes = max_bytes -
+			(block_group->total_bitmaps * PAGE_CACHE_SIZE);
+
+		if (extent_bytes <= 0) {
+			block_group->extents_thresh = 0;
+			return;
+		}
+
+		block_group->extents_thresh = extent_bytes /
+			(sizeof(struct btrfs_free_space));
+	}
+}
+
+static void bitmap_clear_bits(struct btrfs_block_group_cache *block_group,
+			      struct btrfs_free_space *info, u64 offset,
+			      u64 bytes)
+{
+	unsigned long start, end;
+	unsigned long i;
+
+	start = offset_to_bit(info->offset, block_group->sectorsize, offset);
+	end = start + bytes_to_bits(bytes, block_group->sectorsize);
+	BUG_ON(end > BITS_PER_BITMAP);
+
+	for (i = start; i < end; i++)
+		clear_bit(i, info->bitmap);
+
+	info->bytes -= bytes;
+	block_group->free_space -= bytes;
+}
+
+static void bitmap_set_bits(struct btrfs_block_group_cache *block_group,
+			    struct btrfs_free_space *info, u64 offset,
+			    u64 bytes)
+{
+	unsigned long start, end;
+	unsigned long i;
+
+	start = offset_to_bit(info->offset, block_group->sectorsize, offset);
+	end = start + bytes_to_bits(bytes, block_group->sectorsize);
+	BUG_ON(end > BITS_PER_BITMAP);
+
+	for (i = start; i < end; i++)
+		set_bit(i, info->bitmap);
+
+	info->bytes += bytes;
+	block_group->free_space += bytes;
+}
+
+static int search_bitmap(struct btrfs_block_group_cache *block_group,
+			 struct btrfs_free_space *bitmap_info, u64 *offset,
+			 u64 *bytes)
+{
+	unsigned long found_bits = 0;
+	unsigned long bits, i;
+	unsigned long next_zero;
+
+	i = offset_to_bit(bitmap_info->offset, block_group->sectorsize,
+			  max_t(u64, *offset, bitmap_info->offset));
+	bits = bytes_to_bits(*bytes, block_group->sectorsize);
+
+	for (i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i);
+	     i < BITS_PER_BITMAP;
+	     i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i + 1)) {
+		next_zero = find_next_zero_bit(bitmap_info->bitmap,
+					       BITS_PER_BITMAP, i);
+		if ((next_zero - i) >= bits) {
+			found_bits = next_zero - i;
+			break;
+		}
+		i = next_zero;
+	}
+
+	if (found_bits) {
+		*offset = (u64)(i * block_group->sectorsize) +
+			bitmap_info->offset;
+		*bytes = (u64)(found_bits) * block_group->sectorsize;
+		return 0;
+	}
+
+	return -1;
+}
+
+static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache
+						*block_group, u64 *offset,
+						u64 *bytes, int debug)
+{
+	struct btrfs_free_space *entry;
+	struct rb_node *node;
+	int ret;
+
+	if (!block_group->free_space_offset.rb_node)
+		return NULL;
+
+	entry = tree_search_offset(block_group,
+				   offset_to_bitmap(block_group, *offset),
+				   0, 1);
+	if (!entry)
+		return NULL;
+
+	for (node = &entry->offset_index; node; node = rb_next(node)) {
+		entry = rb_entry(node, struct btrfs_free_space, offset_index);
+		if (entry->bytes < *bytes)
+			continue;
+
+		if (entry->bitmap) {
+			ret = search_bitmap(block_group, entry, offset, bytes);
+			if (!ret)
+				return entry;
+			continue;
+		}
+
+		*offset = entry->offset;
+		*bytes = entry->bytes;
+		return entry;
+	}
+
+	return NULL;
+}
+
+static void add_new_bitmap(struct btrfs_block_group_cache *block_group,
+			   struct btrfs_free_space *info, u64 offset)
+{
+	u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize;
+	int max_bitmaps = (int)div64_u64(block_group->key.offset +
+					 bytes_per_bg - 1, bytes_per_bg);
+	BUG_ON(block_group->total_bitmaps >= max_bitmaps);
+
+	info->offset = offset_to_bitmap(block_group, offset);
+	link_free_space(block_group, info);
+	block_group->total_bitmaps++;
+
+	recalculate_thresholds(block_group);
+}
+
+static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_group,
+			      struct btrfs_free_space *bitmap_info,
+			      u64 *offset, u64 *bytes)
+{
+	u64 end;
+	u64 search_start, search_bytes;
+	int ret;
+
+again:
+	end = bitmap_info->offset +
+		(u64)(BITS_PER_BITMAP * block_group->sectorsize) - 1;
+
+	/*
+	 * XXX - this can go away after a few releases.
+	 *
+	 * since the only user of btrfs_remove_free_space is the tree logging
+	 * stuff, and the only way to test that is under crash conditions, we
+	 * want to have this debug stuff here just in case somethings not
+	 * working.  Search the bitmap for the space we are trying to use to
+	 * make sure its actually there.  If its not there then we need to stop
+	 * because something has gone wrong.
+	 */
+	search_start = *offset;
+	search_bytes = *bytes;
+	ret = search_bitmap(block_group, bitmap_info, &search_start,
+			    &search_bytes);
+	BUG_ON(ret < 0 || search_start != *offset);
+
+	if (*offset > bitmap_info->offset && *offset + *bytes > end) {
+		bitmap_clear_bits(block_group, bitmap_info, *offset,
+				  end - *offset + 1);
+		*bytes -= end - *offset + 1;
+		*offset = end + 1;
+	} else if (*offset >= bitmap_info->offset && *offset + *bytes <= end) {
+		bitmap_clear_bits(block_group, bitmap_info, *offset, *bytes);
+		*bytes = 0;
+	}
+
+	if (*bytes) {
+		struct rb_node *next = rb_next(&bitmap_info->offset_index);
+		if (!bitmap_info->bytes) {
+			unlink_free_space(block_group, bitmap_info);
+			kfree(bitmap_info->bitmap);
+			kfree(bitmap_info);
+			block_group->total_bitmaps--;
+			recalculate_thresholds(block_group);
+		}
+
+		/*
+		 * no entry after this bitmap, but we still have bytes to
+		 * remove, so something has gone wrong.
+		 */
+		if (!next)
+			return -EINVAL;
+
+		bitmap_info = rb_entry(next, struct btrfs_free_space,
+				       offset_index);
+
+		/*
+		 * if the next entry isn't a bitmap we need to return to let the
+		 * extent stuff do its work.
+		 */
+		if (!bitmap_info->bitmap)
+			return -EAGAIN;
+
+		/*
+		 * Ok the next item is a bitmap, but it may not actually hold
+		 * the information for the rest of this free space stuff, so
+		 * look for it, and if we don't find it return so we can try
+		 * everything over again.
+		 */
+		search_start = *offset;
+		search_bytes = *bytes;
+		ret = search_bitmap(block_group, bitmap_info, &search_start,
+				    &search_bytes);
+		if (ret < 0 || search_start != *offset)
+			return -EAGAIN;
+
+		goto again;
+	} else if (!bitmap_info->bytes) {
+		unlink_free_space(block_group, bitmap_info);
+		kfree(bitmap_info->bitmap);
+		kfree(bitmap_info);
+		block_group->total_bitmaps--;
+		recalculate_thresholds(block_group);
+	}
+
+	return 0;
+}
+
+static int insert_into_bitmap(struct btrfs_block_group_cache *block_group,
+			      struct btrfs_free_space *info)
+{
+	struct btrfs_free_space *bitmap_info;
+	int added = 0;
+	u64 bytes, offset, end;
+	int ret;
+
+	/*
+	 * If we are below the extents threshold then we can add this as an
+	 * extent, and don't have to deal with the bitmap
+	 */
+	if (block_group->free_extents < block_group->extents_thresh &&
+	    info->bytes > block_group->sectorsize * 4)
+		return 0;
+
+	/*
+	 * some block groups are so tiny they can't be enveloped by a bitmap, so
+	 * don't even bother to create a bitmap for this
+	 */
+	if (BITS_PER_BITMAP * block_group->sectorsize >
+	    block_group->key.offset)
+		return 0;
+
+	bytes = info->bytes;
+	offset = info->offset;
+
+again:
+	bitmap_info = tree_search_offset(block_group,
+					 offset_to_bitmap(block_group, offset),
+					 1, 0);
+	if (!bitmap_info) {
+		BUG_ON(added);
+		goto new_bitmap;
+	}
+
+	end = bitmap_info->offset +
+		(u64)(BITS_PER_BITMAP * block_group->sectorsize);
+
+	if (offset >= bitmap_info->offset && offset + bytes > end) {
+		bitmap_set_bits(block_group, bitmap_info, offset,
+				end - offset);
+		bytes -= end - offset;
+		offset = end;
+		added = 0;
+	} else if (offset >= bitmap_info->offset && offset + bytes <= end) {
+		bitmap_set_bits(block_group, bitmap_info, offset, bytes);
+		bytes = 0;
+	} else {
+		BUG();
+	}
+
+	if (!bytes) {
+		ret = 1;
+		goto out;
+	} else
+		goto again;
+
+new_bitmap:
+	if (info && info->bitmap) {
+		add_new_bitmap(block_group, info, offset);
+		added = 1;
+		info = NULL;
+		goto again;
+	} else {
+		spin_unlock(&block_group->tree_lock);
+
+		/* no pre-allocated info, allocate a new one */
+		if (!info) {
+			info = kzalloc(sizeof(struct btrfs_free_space),
+				       GFP_NOFS);
+			if (!info) {
+				spin_lock(&block_group->tree_lock);
+				ret = -ENOMEM;
+				goto out;
+			}
+		}
+
+		/* allocate the bitmap */
+		info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
+		spin_lock(&block_group->tree_lock);
+		if (!info->bitmap) {
+			ret = -ENOMEM;
+			goto out;
+		}
+		goto again;
+	}
+
+out:
+	if (info) {
+		if (info->bitmap)
+			kfree(info->bitmap);
+		kfree(info);
+	}
 
 	return ret;
 }
@@ -208,8 +600,8 @@ static int link_free_space(struct btrfs_block_group_cache *block_group,
 int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
 			 u64 offset, u64 bytes)
 {
-	struct btrfs_free_space *right_info;
-	struct btrfs_free_space *left_info;
+	struct btrfs_free_space *right_info = NULL;
+	struct btrfs_free_space *left_info = NULL;
 	struct btrfs_free_space *info = NULL;
 	int ret = 0;
 
@@ -227,18 +619,38 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
 	 * are adding, if there is remove that struct and add a new one to
 	 * cover the entire range
 	 */
-	right_info = tree_search_offset(&block_group->free_space_offset,
-					offset+bytes, 0, 0);
-	left_info = tree_search_offset(&block_group->free_space_offset,
-				       offset-1, 0, 1);
+	right_info = tree_search_offset(block_group, offset + bytes, 0, 0);
+	if (right_info && rb_prev(&right_info->offset_index))
+		left_info = rb_entry(rb_prev(&right_info->offset_index),
+				     struct btrfs_free_space, offset_index);
+	else
+		left_info = tree_search_offset(block_group, offset - 1, 0, 0);
+
+	/*
+	 * If there was no extent directly to the left or right of this new
+	 * extent then we know we're going to have to allocate a new extent, so
+	 * before we do that see if we need to drop this into a bitmap
+	 */
+	if ((!left_info || left_info->bitmap) &&
+	    (!right_info || right_info->bitmap)) {
+		ret = insert_into_bitmap(block_group, info);
+
+		if (ret < 0) {
+			goto out;
+		} else if (ret) {
+			ret = 0;
+			goto out;
+		}
+	}
 
-	if (right_info) {
+	if (right_info && !right_info->bitmap) {
 		unlink_free_space(block_group, right_info);
 		info->bytes += right_info->bytes;
 		kfree(right_info);
 	}
 
-	if (left_info && left_info->offset + left_info->bytes == offset) {
+	if (left_info && !left_info->bitmap &&
+	    left_info->offset + left_info->bytes == offset) {
 		unlink_free_space(block_group, left_info);
 		info->offset = left_info->offset;
 		info->bytes += left_info->bytes;
@@ -248,11 +660,11 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
 	ret = link_free_space(block_group, info);
 	if (ret)
 		kfree(info);
-
+out:
 	spin_unlock(&block_group->tree_lock);
 
 	if (ret) {
-		printk(KERN_ERR "btrfs: unable to add free space :%d\n", ret);
+		printk(KERN_CRIT "btrfs: unable to add free space :%d\n", ret);
 		BUG_ON(ret == -EEXIST);
 	}
 
@@ -263,40 +675,74 @@ int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			    u64 offset, u64 bytes)
 {
 	struct btrfs_free_space *info;
+	struct btrfs_free_space *next_info = NULL;
 	int ret = 0;
 
 	spin_lock(&block_group->tree_lock);
 
-	info = tree_search_offset(&block_group->free_space_offset, offset, 0,
-				  1);
-	if (info && info->offset == offset) {
-		if (info->bytes < bytes) {
-			printk(KERN_ERR "Found free space at %llu, size %llu,"
-			       "trying to use %llu\n",
-			       (unsigned long long)info->offset,
-			       (unsigned long long)info->bytes,
-			       (unsigned long long)bytes);
+again:
+	info = tree_search_offset(block_group, offset, 0, 0);
+	if (!info) {
+		/*
+		 * oops didn't find an extent that matched the space we wanted
+		 * to remove, look for a bitmap instead
+		 */
+		info = tree_search_offset(block_group,
+					  offset_to_bitmap(block_group, offset),
+					  1, 0);
+		if (!info) {
+			WARN_ON(1);
+			goto out_lock;
+		}
+	}
+
+	if (info->bytes < bytes && rb_next(&info->offset_index)) {
+		u64 end;
+		next_info = rb_entry(rb_next(&info->offset_index),
+					     struct btrfs_free_space,
+					     offset_index);
+
+		if (next_info->bitmap)
+			end = next_info->offset + BITS_PER_BITMAP *
+				block_group->sectorsize - 1;
+		else
+			end = next_info->offset + next_info->bytes;
+
+		if (next_info->bytes < bytes ||
+		    next_info->offset > offset || offset > end) {
+			printk(KERN_CRIT "Found free space at %llu, size %llu,"
+			      " trying to use %llu\n",
+			      (unsigned long long)info->offset,
+			      (unsigned long long)info->bytes,
+			      (unsigned long long)bytes);
 			WARN_ON(1);
 			ret = -EINVAL;
-			spin_unlock(&block_group->tree_lock);
-			goto out;
+			goto out_lock;
 		}
-		unlink_free_space(block_group, info);
 
-		if (info->bytes == bytes) {
-			kfree(info);
-			spin_unlock(&block_group->tree_lock);
-			goto out;
+		info = next_info;
+	}
+
+	if (info->bytes == bytes) {
+		unlink_free_space(block_group, info);
+		if (info->bitmap) {
+			kfree(info->bitmap);
+			block_group->total_bitmaps--;
 		}
+		kfree(info);
+		goto out_lock;
+	}
 
+	if (!info->bitmap && info->offset == offset) {
+		unlink_free_space(block_group, info);
 		info->offset += bytes;
 		info->bytes -= bytes;
+		link_free_space(block_group, info);
+		goto out_lock;
+	}
 
-		ret = link_free_space(block_group, info);
-		spin_unlock(&block_group->tree_lock);
-		BUG_ON(ret);
-	} else if (info && info->offset < offset &&
-		   info->offset + info->bytes >= offset + bytes) {
+	if (!info->bitmap && info->offset <= offset &&
+	    info->offset + info->bytes >= offset + bytes) {
 		u64 old_start = info->offset;
 		/*
 		 * we're freeing space in the middle of the info,
@@ -312,7 +758,9 @@ int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			info->offset = offset + bytes;
 			info->bytes = old_end - info->offset;
 			ret = link_free_space(block_group, info);
-			BUG_ON(ret);
+			WARN_ON(ret);
+			if (ret)
+				goto out_lock;
 		} else {
 			/* the hole we're creating ends at the end
 			 * of the info struct, just free the info
@@ -320,32 +768,22 @@ int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			kfree(info);
 		}
 		spin_unlock(&block_group->tree_lock);
-		/* step two, insert a new info struct to cover anything
-		 * before the hole
+
+		/* step two, insert a new info struct to cover
+		 * anything before the hole
 		 */
 		ret = btrfs_add_free_space(block_group, old_start,
 					   offset - old_start);
-		BUG_ON(ret);
-	} else {
-		spin_unlock(&block_group->tree_lock);
-		if (!info) {
-			printk(KERN_ERR "couldn't find space %llu to free\n",
-			       (unsigned long long)offset);
-			printk(KERN_ERR "cached is %d, offset %llu bytes %llu\n",
-			       block_group->cached,
-			       (unsigned long long)block_group->key.objectid,
-			       (unsigned long long)block_group->key.offset);
-			btrfs_dump_free_space(block_group, bytes);
-		} else if (info) {
-			printk(KERN_ERR "hmm, found offset=%llu bytes=%llu, "
-			       "but wanted offset=%llu bytes=%llu\n",
-			       (unsigned long long)info->offset,
-			       (unsigned long long)info->bytes,
-			       (unsigned long long)offset,
-			       (unsigned long long)bytes);
-		}
-		WARN_ON(1);
+		WARN_ON(ret);
+		goto out;
 	}
+
+	ret = remove_from_bitmap(block_group, info, &offset, &bytes);
+	if (ret == -EAGAIN)
+		goto again;
+	BUG_ON(ret);
+out_lock:
+	spin_unlock(&block_group->tree_lock);
 out:
 	return ret;
 }
@@ -361,10 +799,13 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 		info = rb_entry(n, struct btrfs_free_space, offset_index);
 		if (info->bytes >= bytes)
 			count++;
-		printk(KERN_ERR "entry offset %llu, bytes %llu\n",
+		printk(KERN_CRIT "entry offset %llu, bytes %llu, bitmap %s\n",
 		       (unsigned long long)info->offset,
-		       (unsigned long long)info->bytes);
+		       (unsigned long long)info->bytes,
+		       (info->bitmap) ? "yes" : "no");
 	}
+	printk(KERN_INFO "block group has cluster?: %s\n",
+	       list_empty(&block_group->cluster_list) ? "no" : "yes");
 	printk(KERN_INFO "%d blocks of free space at or bigger than bytes is"
 	       "\n", count);
 }
@@ -397,26 +838,35 @@ __btrfs_return_cluster_to_free_space(
 {
 	struct btrfs_free_space *entry;
 	struct rb_node *node;
+	bool bitmap;
 
 	spin_lock(&cluster->lock);
 	if (cluster->block_group != block_group)
 		goto out;
 
+	bitmap = cluster->points_to_bitmap;
+	cluster->block_group = NULL;
 	cluster->window_start = 0;
+	list_del_init(&cluster->block_group_list);
+	cluster->points_to_bitmap = false;
+
+	if (bitmap)
+		goto out;
+
 	node = rb_first(&cluster->root);
-	while(node) {
+	while (node) {
 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
 		node = rb_next(&entry->offset_index);
 		rb_erase(&entry->offset_index, &cluster->root);
-		link_free_space(block_group, entry);
+		BUG_ON(entry->bitmap);
+		tree_insert_offset(&block_group->free_space_offset,
+				   entry->offset, &entry->offset_index, 0);
 	}
-	list_del_init(&cluster->block_group_list);
-
-	btrfs_put_block_group(cluster->block_group);
-	cluster->block_group = NULL;
 	cluster->root.rb_node = NULL;
+
 out:
 	spin_unlock(&cluster->lock);
+	btrfs_put_block_group(block_group);
 	return 0;
 }
 
@@ -425,20 +875,28 @@ void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
 	struct btrfs_free_space *info;
 	struct rb_node *node;
 	struct btrfs_free_cluster *cluster;
-	struct btrfs_free_cluster *safe;
+	struct list_head *head;
 
 	spin_lock(&block_group->tree_lock);
-
-	list_for_each_entry_safe(cluster, safe, &block_group->cluster_list,
-				 block_group_list) {
+	while ((head = block_group->cluster_list.next) !=
+	       &block_group->cluster_list) {
+		cluster = list_entry(head, struct btrfs_free_cluster,
+				     block_group_list);
 
 		WARN_ON(cluster->block_group != block_group);
 		__btrfs_return_cluster_to_free_space(block_group, cluster);
+		if (need_resched()) {
+			spin_unlock(&block_group->tree_lock);
+			cond_resched();
+			spin_lock(&block_group->tree_lock);
+		}
 	}
 
-	while ((node = rb_last(&block_group->free_space_bytes)) != NULL) {
-		info = rb_entry(node, struct btrfs_free_space, bytes_index);
+	while ((node = rb_last(&block_group->free_space_offset)) != NULL) {
+		info = rb_entry(node, struct btrfs_free_space, offset_index);
 		unlink_free_space(block_group, info);
+		if (info->bitmap)
+			kfree(info->bitmap);
 		kfree(info);
 		if (need_resched()) {
 			spin_unlock(&block_group->tree_lock);
@@ -446,6 +904,7 @@ void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
 			spin_lock(&block_group->tree_lock);
 		}
 	}
+
 	spin_unlock(&block_group->tree_lock);
 }
 
@@ -453,25 +912,35 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 			       u64 offset, u64 bytes, u64 empty_size)
 {
 	struct btrfs_free_space *entry = NULL;
+	u64 bytes_search = bytes + empty_size;
 	u64 ret = 0;
 
 	spin_lock(&block_group->tree_lock);
-	entry = tree_search_offset(&block_group->free_space_offset, offset,
-				   bytes + empty_size, 1);
+	entry = find_free_space(block_group, &offset, &bytes_search, 0);
 	if (!entry)
-		entry = tree_search_bytes(&block_group->free_space_bytes,
-					  offset, bytes + empty_size);
-	if (entry) {
+		goto out;
+
+	ret = offset;
+	if (entry->bitmap) {
+		bitmap_clear_bits(block_group, entry, offset, bytes);
+		if (!entry->bytes) {
+			unlink_free_space(block_group, entry);
+			kfree(entry->bitmap);
+			kfree(entry);
+			block_group->total_bitmaps--;
+			recalculate_thresholds(block_group);
+		}
+	} else {
 		unlink_free_space(block_group, entry);
-		ret = entry->offset;
 		entry->offset += bytes;
 		entry->bytes -= bytes;
-
 		if (!entry->bytes)
 			kfree(entry);
 		else
 			link_free_space(block_group, entry);
 	}
+
+out:
 	spin_unlock(&block_group->tree_lock);
 
 	return ret;
@@ -517,6 +986,54 @@ int btrfs_return_cluster_to_free_space(
 	return ret;
 }
 
+static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
+				   struct btrfs_free_cluster *cluster,
+				   u64 bytes, u64 min_start)
+{
+	struct btrfs_free_space *entry;
+	int err;
+	u64 search_start = cluster->window_start;
+	u64 search_bytes = bytes;
+	u64 ret = 0;
+
+	spin_lock(&block_group->tree_lock);
+	spin_lock(&cluster->lock);
+
+	if (!cluster->points_to_bitmap)
+		goto out;
+
+	if (cluster->block_group != block_group)
+		goto out;
+
+	/*
+	 * search_start is the beginning of the bitmap, but at some point it may
+	 * be a good idea to point to the actual start of the free area in the
+	 * bitmap, so do the offset_to_bitmap trick anyway, and set bitmap_only
+	 * to 1 to make sure we get the bitmap entry
+	 */
+	entry = tree_search_offset(block_group,
+				   offset_to_bitmap(block_group, search_start),
+				   1, 0);
+	if (!entry || !entry->bitmap)
+		goto out;
+
+	search_start = min_start;
+	search_bytes = bytes;
+
+	err = search_bitmap(block_group, entry, &search_start,
+			    &search_bytes);
+	if (err)
+		goto out;
+
+	ret = search_start;
+	bitmap_clear_bits(block_group, entry, ret, bytes);
+out:
+	spin_unlock(&cluster->lock);
+	spin_unlock(&block_group->tree_lock);
+
+	return ret;
+}
+
 /*
  * given a cluster, try to allocate 'bytes' from it, returns 0
  * if it couldn't find anything suitably large, or a logical disk offset
@@ -530,6 +1047,10 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
 	struct rb_node *node;
 	u64 ret = 0;
 
+	if (cluster->points_to_bitmap)
+		return btrfs_alloc_from_bitmap(block_group, cluster, bytes,
+					       min_start);
+
 	spin_lock(&cluster->lock);
 	if (bytes > cluster->max_size)
 		goto out;
@@ -567,9 +1088,73 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
 	}
 out:
 	spin_unlock(&cluster->lock);
+
 	return ret;
 }
 
+static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
+				struct btrfs_free_space *entry,
+				struct btrfs_free_cluster *cluster,
+				u64 offset, u64 bytes, u64 min_bytes)
+{
+	unsigned long next_zero;
+	unsigned long i;
+	unsigned long search_bits;
+	unsigned long total_bits;
+	unsigned long found_bits;
+	unsigned long start = 0;
+	unsigned long total_found = 0;
+	bool found = false;
+
+	i = offset_to_bit(entry->offset, block_group->sectorsize,
+			  max_t(u64, offset, entry->offset));
+	search_bits = bytes_to_bits(min_bytes, block_group->sectorsize);
+	total_bits = bytes_to_bits(bytes, block_group->sectorsize);
+
+again:
+	found_bits = 0;
+	for (i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, i);
+	     i < BITS_PER_BITMAP;
+	     i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, i + 1)) {
+		next_zero = find_next_zero_bit(entry->bitmap,
+					       BITS_PER_BITMAP, i);
+		if (next_zero - i >= search_bits) {
+			found_bits = next_zero - i;
+			break;
+		}
+		i = next_zero;
+	}
+
+	if (!found_bits)
+		return -1;
+
+	if (!found) {
+		start = i;
+		found = true;
+	}
+
+	total_found += found_bits;
+
+	if (cluster->max_size < found_bits * block_group->sectorsize)
+		cluster->max_size = found_bits * block_group->sectorsize;
+
+	if (total_found < total_bits) {
+		i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, next_zero);
+		if (i - start > total_bits * 2) {
+			total_found = 0;
+			cluster->max_size = 0;
+			found = false;
+		}
+		goto again;
+	}
+
+	cluster->window_start = start * block_group->sectorsize +
+		entry->offset;
+	cluster->points_to_bitmap = true;
+
+	return 0;
+}
+
 /*
  * here we try to find a cluster of blocks in a block group.  The goal
  * is to find at least bytes free and up to empty_size + bytes free.
@@ -587,12 +1172,12 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
 	struct btrfs_free_space *entry = NULL;
 	struct rb_node *node;
 	struct btrfs_free_space *next;
-	struct btrfs_free_space *last;
+	struct btrfs_free_space *last = NULL;
 	u64 min_bytes;
 	u64 window_start;
 	u64 window_free;
 	u64 max_extent = 0;
-	int total_retries = 0;
+	bool found_bitmap = false;
 	int ret;
 
 	/* for metadata, allow allocates with more holes */
@@ -620,31 +1205,80 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
 		goto out;
 	}
 again:
-	min_bytes = min(min_bytes, bytes + empty_size);
-	entry = tree_search_bytes(&block_group->free_space_bytes,
-				  offset, min_bytes);
+	entry = tree_search_offset(block_group, offset, found_bitmap, 1);
 	if (!entry) {
 		ret = -ENOSPC;
 		goto out;
 	}
+
+	/*
+	 * If found_bitmap is true, we exhausted our search for extent entries,
+	 * and we just want to search all of the bitmaps that we can find, and
+	 * ignore any extent entries we find.
+	 */
+	while (entry->bitmap || found_bitmap ||
+	       (!entry->bitmap && entry->bytes < min_bytes)) {
+		struct rb_node *node = rb_next(&entry->offset_index);
+
+		if (entry->bitmap && entry->bytes > bytes + empty_size) {
+			ret = btrfs_bitmap_cluster(block_group, entry, cluster,
+						   offset, bytes + empty_size,
+						   min_bytes);
+			if (!ret)
+				goto got_it;
+		}
+
+		if (!node) {
+			ret = -ENOSPC;
+			goto out;
+		}
+		entry = rb_entry(node, struct btrfs_free_space, offset_index);
+	}
+
+	/*
+	 * We already searched all the extent entries from the passed in offset
+	 * to the end and didn't find enough space for the cluster, and we also
+	 * didn't find any bitmaps that met our criteria, just go ahead and exit
+	 */
+	if (found_bitmap) {
+		ret = -ENOSPC;
+		goto out;
+	}
+
+	cluster->points_to_bitmap = false;
 	window_start = entry->offset;
 	window_free = entry->bytes;
 	last = entry;
 	max_extent = entry->bytes;
 
-	while(1) {
+	while (1) {
 		/* out window is just right, lets fill it */
 		if (window_free >= bytes + empty_size)
 			break;
 
 		node = rb_next(&last->offset_index);
 		if (!node) {
+			if (found_bitmap)
+				goto again;
 			ret = -ENOSPC;
 			goto out;
 		}
 		next = rb_entry(node, struct btrfs_free_space, offset_index);
 
 		/*
+		 * we found a bitmap, so if this search doesn't result in a
+		 * cluster, we know to go and search again for the bitmaps and
+		 * start looking for space there
+		 */
+		if (next->bitmap) {
+			if (!found_bitmap)
+				offset = next->offset;
+			found_bitmap = true;
+			last = next;
+			continue;
+		}
+
+		/*
 		 * we haven't filled the empty size and the window is
 		 * very large.  reset and try again
 		 */
@@ -655,19 +1289,6 @@ again:
 			window_free = entry->bytes;
 			last = entry;
 			max_extent = 0;
-			total_retries++;
-			if (total_retries % 64 == 0) {
-				if (min_bytes >= (bytes + empty_size)) {
-					ret = -ENOSPC;
-					goto out;
-				}
-				/*
-				 * grow our allocation a bit, we're not having
-				 * much luck
-				 */
-				min_bytes *= 2;
-				goto again;
-			}
 		} else {
 			last = next;
 			window_free += next->bytes;
@@ -685,11 +1306,19 @@ again:
 	 * The cluster includes an rbtree, but only uses the offset index
 	 * of each free space cache entry.
 	 */
-	while(1) {
+	while (1) {
 		node = rb_next(&entry->offset_index);
-		unlink_free_space(block_group, entry);
+		if (entry->bitmap && node) {
+			entry = rb_entry(node, struct btrfs_free_space,
+					 offset_index);
+			continue;
+		} else if (entry->bitmap && !node) {
+			break;
+		}
+
+		rb_erase(&entry->offset_index, &block_group->free_space_offset);
 		ret = tree_insert_offset(&cluster->root, entry->offset,
-					 &entry->offset_index);
+					 &entry->offset_index, 0);
 		BUG_ON(ret);
 
 		if (!node || entry == last)
@@ -697,8 +1326,10 @@ again:
 
 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
 	}
-	ret = 0;
+
 	cluster->max_size = max_extent;
+got_it:
+	ret = 0;
 	atomic_inc(&block_group->count);
 	list_add_tail(&cluster->block_group_list, &block_group->cluster_list);
 	cluster->block_group = block_group;
@@ -718,6 +1349,7 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster)
 	spin_lock_init(&cluster->refill_lock);
 	cluster->root.rb_node = NULL;
 	cluster->max_size = 0;
+	cluster->points_to_bitmap = false;
 	INIT_LIST_HEAD(&cluster->block_group_list);
 	cluster->block_group = NULL;
 }
diff --git a/fs/btrfs/free-space-cache.h b/fs/btrfs/free-space-cache.h
index 266fb8764054..890a8e79011b 100644
--- a/fs/btrfs/free-space-cache.h
+++ b/fs/btrfs/free-space-cache.h
@@ -19,6 +19,14 @@
 #ifndef __BTRFS_FREE_SPACE_CACHE
 #define __BTRFS_FREE_SPACE_CACHE
 
+struct btrfs_free_space {
+	struct rb_node offset_index;
+	u64 offset;
+	u64 bytes;
+	unsigned long *bitmap;
+	struct list_head list;
+};
+
 int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
 			 u64 bytenr, u64 size);
 int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 791eab19e330..59cba180fe83 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -2603,8 +2603,8 @@ noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	if (root->ref_cows)
 		btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0);
 	path = btrfs_alloc_path();
-	path->reada = -1;
 	BUG_ON(!path);
+	path->reada = -1;
 
 	/* FIXME, add redo link to tree so we don't leak on crash */
 	key.objectid = inode->i_ino;
@@ -3099,8 +3099,12 @@ static void inode_tree_add(struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_inode *entry;
-	struct rb_node **p = &root->inode_tree.rb_node;
-	struct rb_node *parent = NULL;
+	struct rb_node **p;
+	struct rb_node *parent;
+
+again:
+	p = &root->inode_tree.rb_node;
+	parent = NULL;
 
 	spin_lock(&root->inode_lock);
 	while (*p) {
@@ -3108,13 +3112,16 @@ static void inode_tree_add(struct inode *inode)
 		entry = rb_entry(parent, struct btrfs_inode, rb_node);
 
 		if (inode->i_ino < entry->vfs_inode.i_ino)
-			p = &(*p)->rb_left;
+			p = &parent->rb_left;
 		else if (inode->i_ino > entry->vfs_inode.i_ino)
-			p = &(*p)->rb_right;
+			p = &parent->rb_right;
 		else {
 			WARN_ON(!(entry->vfs_inode.i_state &
 				  (I_WILL_FREE | I_FREEING | I_CLEAR)));
-			break;
+			rb_erase(parent, &root->inode_tree);
+			RB_CLEAR_NODE(parent);
+			spin_unlock(&root->inode_lock);
+			goto again;
 		}
 	}
 	rb_link_node(&BTRFS_I(inode)->rb_node, parent, p);
@@ -3126,12 +3133,12 @@ static void inode_tree_del(struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 
+	spin_lock(&root->inode_lock);
 	if (!RB_EMPTY_NODE(&BTRFS_I(inode)->rb_node)) {
-		spin_lock(&root->inode_lock);
 		rb_erase(&BTRFS_I(inode)->rb_node, &root->inode_tree);
-		spin_unlock(&root->inode_lock);
 		RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
 	}
+	spin_unlock(&root->inode_lock);
 }
 
 static noinline void init_btrfs_i(struct inode *inode)
@@ -4785,8 +4792,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	 * and the replacement file is large.  Start IO on it now so
 	 * we don't add too much work to the end of the transaction
 	 */
-	if (new_inode && old_inode && S_ISREG(old_inode->i_mode) &&
-	    new_inode->i_size &&
+	if (new_inode && S_ISREG(old_inode->i_mode) && new_inode->i_size &&
 	    old_inode->i_size > BTRFS_ORDERED_OPERATIONS_FLUSH_LIMIT)
 		filemap_flush(old_inode->i_mapping);
 
diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c
index 6d6523da0a30..0d126be22b63 100644
--- a/fs/btrfs/print-tree.c
+++ b/fs/btrfs/print-tree.c
@@ -309,7 +309,7 @@ void btrfs_print_tree(struct btrfs_root *root, struct extent_buffer *c)
 	}
 	printk(KERN_INFO "node %llu level %d total ptrs %d free spc %u\n",
 	       (unsigned long long)btrfs_header_bytenr(c),
-	       btrfs_header_level(c), nr,
+	      level, nr,
 	       (u32)BTRFS_NODEPTRS_PER_BLOCK(root) - nr);
 	for (i = 0; i < nr; i++) {
 		btrfs_node_key_to_cpu(c, &key, i);
@@ -326,10 +326,10 @@ void btrfs_print_tree(struct btrfs_root *root, struct extent_buffer *c)
 					btrfs_level_size(root, level - 1),
 					btrfs_node_ptr_generation(c, i));
 		if (btrfs_is_leaf(next) &&
-		    btrfs_header_level(c) != 1)
+		   level != 1)
 			BUG();
 		if (btrfs_header_level(next) !=
-			btrfs_header_level(c) - 1)
+		       level - 1)
 			BUG();
 		btrfs_print_tree(root, next);
 		free_extent_buffer(next);
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 008397934778..c04f7f212602 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -670,6 +670,8 @@ again:
 			err = ret;
 			goto out;
 		}
+		if (ret > 0 && path2->slots[level] > 0)
+			path2->slots[level]--;
 
 		eb = path2->nodes[level];
 		WARN_ON(btrfs_node_blockptr(eb, path2->slots[level]) !=
@@ -1609,6 +1611,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 		BUG_ON(level == 0);
 		path->lowest_level = level;
 		ret = btrfs_search_slot(NULL, reloc_root, &key, path, 0, 0);
+		path->lowest_level = 0;
 		if (ret < 0) {
 			btrfs_free_path(path);
 			return ret;
@@ -2550,8 +2553,13 @@ int relocate_inode_pages(struct inode *inode, u64 start, u64 len)
 	last_index = (start + len - 1) >> PAGE_CACHE_SHIFT;
 
 	/* make sure the dirty trick played by the caller work */
-	ret = invalidate_inode_pages2_range(inode->i_mapping,
-					    first_index, last_index);
+	while (1) {
+		ret = invalidate_inode_pages2_range(inode->i_mapping,
+						    first_index, last_index);
+		if (ret != -EBUSY)
+			break;
+		schedule_timeout(HZ/10);
+	}
 	if (ret)
 		goto out_unlock;
 
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 2dbf1c1f56ee..cdbb5022da52 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -40,6 +40,12 @@ static noinline void put_transaction(struct btrfs_transaction *transaction)
 	}
 }
 
+static noinline void switch_commit_root(struct btrfs_root *root)
+{
+	free_extent_buffer(root->commit_root);
+	root->commit_root = btrfs_root_node(root);
+}
+
 /*
  * either allocate a new transaction or hop into the existing one
  */
@@ -444,9 +450,6 @@ static int update_cowonly_root(struct btrfs_trans_handle *trans,
 
 	btrfs_write_dirty_block_groups(trans, root);
 
-	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
-	BUG_ON(ret);
-
 	while (1) {
 		old_root_bytenr = btrfs_root_bytenr(&root->root_item);
 		if (old_root_bytenr == root->node->start)
@@ -457,13 +460,14 @@ static int update_cowonly_root(struct btrfs_trans_handle *trans,
 					&root->root_key,
 					&root->root_item);
 		BUG_ON(ret);
-		btrfs_write_dirty_block_groups(trans, root);
 
-		ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
+		ret = btrfs_write_dirty_block_groups(trans, root);
 		BUG_ON(ret);
 	}
-	free_extent_buffer(root->commit_root);
-	root->commit_root = btrfs_root_node(root);
+
+	if (root != root->fs_info->extent_root)
+		switch_commit_root(root);
+
 	return 0;
 }
 
@@ -495,10 +499,12 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans,
 		root = list_entry(next, struct btrfs_root, dirty_list);
 
 		update_cowonly_root(trans, root);
-
-		ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
-		BUG_ON(ret);
 	}
+
+	down_write(&fs_info->extent_commit_sem);
+	switch_commit_root(fs_info->extent_root);
+	up_write(&fs_info->extent_commit_sem);
+
 	return 0;
 }
 
@@ -544,8 +550,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 			btrfs_update_reloc_root(trans, root);
 
 			if (root->commit_root != root->node) {
-				free_extent_buffer(root->commit_root);
-				root->commit_root = btrfs_root_node(root);
+				switch_commit_root(root);
 				btrfs_set_root_node(&root->root_item,
 						    root->node);
 			}
@@ -852,6 +857,16 @@ static void update_super_roots(struct btrfs_root *root)
 	super->root_level = root_item->level;
 }
 
+int btrfs_transaction_in_commit(struct btrfs_fs_info *info)
+{
+	int ret = 0;
+	spin_lock(&info->new_trans_lock);
+	if (info->running_transaction)
+		ret = info->running_transaction->in_commit;
+	spin_unlock(&info->new_trans_lock);
+	return ret;
+}
+
 int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root)
 {
@@ -943,9 +958,11 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 		mutex_unlock(&root->fs_info->trans_mutex);
 
-		if (flush_on_commit || snap_pending) {
-			if (flush_on_commit)
-				btrfs_start_delalloc_inodes(root);
+		if (flush_on_commit) {
+			btrfs_start_delalloc_inodes(root);
+			ret = btrfs_wait_ordered_extents(root, 0);
+			BUG_ON(ret);
+		} else if (snap_pending) {
 			ret = btrfs_wait_ordered_extents(root, 1);
 			BUG_ON(ret);
 		}
@@ -1009,15 +1026,11 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	btrfs_set_root_node(&root->fs_info->tree_root->root_item,
 			    root->fs_info->tree_root->node);
-	free_extent_buffer(root->fs_info->tree_root->commit_root);
-	root->fs_info->tree_root->commit_root =
-				btrfs_root_node(root->fs_info->tree_root);
+	switch_commit_root(root->fs_info->tree_root);
 
 	btrfs_set_root_node(&root->fs_info->chunk_root->root_item,
 			    root->fs_info->chunk_root->node);
-	free_extent_buffer(root->fs_info->chunk_root->commit_root);
-	root->fs_info->chunk_root->commit_root =
-				btrfs_root_node(root->fs_info->chunk_root);
+	switch_commit_root(root->fs_info->chunk_root);
 
 	update_super_roots(root);
 
@@ -1057,6 +1070,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	cur_trans->commit_done = 1;
 
 	root->fs_info->last_trans_committed = cur_trans->transid;
+
 	wake_up(&cur_trans->commit_wait);
 
 	put_transaction(cur_trans);
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index 961c3ee5a2e1..663c67404918 100644
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -107,4 +107,5 @@ int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root);
 int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
 					struct extent_io_tree *dirty_pages);
+int btrfs_transaction_in_commit(struct btrfs_fs_info *info);
 #endif
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index c13922206d1b..d91b0de7c502 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -797,7 +797,7 @@ static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
 		return -ENOENT;
 
 	inode = read_one_inode(root, key->objectid);
-	BUG_ON(!dir);
+	BUG_ON(!inode);
 
 	ref_ptr = btrfs_item_ptr_offset(eb, slot);
 	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 3ab80e9cd767..5dbefd11b4af 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -721,7 +721,8 @@ error:
  */
 static noinline int find_free_dev_extent(struct btrfs_trans_handle *trans,
 					 struct btrfs_device *device,
-					 u64 num_bytes, u64 *start)
+					 u64 num_bytes, u64 *start,
+					 u64 *max_avail)
 {
 	struct btrfs_key key;
 	struct btrfs_root *root = device->dev_root;
@@ -758,9 +759,13 @@ static noinline int find_free_dev_extent(struct btrfs_trans_handle *trans,
 	ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
 	if (ret < 0)
 		goto error;
-	ret = btrfs_previous_item(root, path, 0, key.type);
-	if (ret < 0)
-		goto error;
+	if (ret > 0) {
+		ret = btrfs_previous_item(root, path, key.objectid, key.type);
+		if (ret < 0)
+			goto error;
+		if (ret > 0)
+			start_found = 1;
+	}
 	l = path->nodes[0];
 	btrfs_item_key_to_cpu(l, &key, path->slots[0]);
 	while (1) {
@@ -803,6 +808,10 @@ no_more_items:
 			if (last_byte < search_start)
 				last_byte = search_start;
 			hole_size = key.offset - last_byte;
+
+			if (hole_size > *max_avail)
+				*max_avail = hole_size;
+
 			if (key.offset > last_byte &&
 			    hole_size >= num_bytes) {
 				*start = last_byte;
@@ -1621,6 +1630,7 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
 	device->fs_devices->total_rw_bytes += diff;
 
 	device->total_bytes = new_size;
+	device->disk_total_bytes = new_size;
 	btrfs_clear_space_info_full(device->dev_root->fs_info);
 
 	return btrfs_update_device(trans, device);
@@ -2007,7 +2017,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 			goto done;
 		if (ret) {
 			ret = 0;
-			goto done;
+			break;
 		}
 
 		l = path->nodes[0];
@@ -2015,7 +2025,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 		btrfs_item_key_to_cpu(l, &key, path->slots[0]);
 
 		if (key.objectid != device->devid)
-			goto done;
+			break;
 
 		dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
 		length = btrfs_dev_extent_length(l, dev_extent);
@@ -2171,6 +2181,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 			     max_chunk_size);
 
 again:
+	max_avail = 0;
 	if (!map || map->num_stripes != num_stripes) {
 		kfree(map);
 		map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
@@ -2219,7 +2230,8 @@ again:
 
 		if (device->in_fs_metadata && avail >= min_free) {
 			ret = find_free_dev_extent(trans, device,
-						   min_free, &dev_offset);
+						   min_free, &dev_offset,
+						   &max_avail);
 			if (ret == 0) {
 				list_move_tail(&device->dev_alloc_list,
 					       &private_devs);
@@ -2795,26 +2807,6 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
 		}
 	}
 
-	for (i = 0; i > nr; i++) {
-		struct btrfs_multi_bio *multi;
-		struct btrfs_bio_stripe *stripe;
-		int ret;
-
-		length = 1;
-		ret = btrfs_map_block(map_tree, WRITE, buf[i],
-				      &length, &multi, 0);
-		BUG_ON(ret);
-
-		stripe = multi->stripes;
-		for (j = 0; j < multi->num_stripes; j++) {
-			if (stripe->physical >= physical &&
-			    physical < stripe->physical + length)
-				break;
-		}
-		BUG_ON(j >= multi->num_stripes);
-		kfree(multi);
-	}
-
 	*logical = buf;
 	*naddrs = nr;
 	*stripe_len = map->stripe_len;
diff --git a/fs/btrfs/zlib.c b/fs/btrfs/zlib.c
index ecfbce836d32..3e2b90eaa239 100644
--- a/fs/btrfs/zlib.c
+++ b/fs/btrfs/zlib.c
@@ -208,7 +208,7 @@ int btrfs_zlib_compress_pages(struct address_space *mapping,
 	*total_in = 0;
 
 	workspace = find_zlib_workspace();
-	if (!workspace)
+	if (IS_ERR(workspace))
 		return -1;
 
 	if (Z_OK != zlib_deflateInit(&workspace->def_strm, 3)) {
@@ -366,7 +366,7 @@ int btrfs_zlib_decompress_biovec(struct page **pages_in,
 	char *kaddr;
 
 	workspace = find_zlib_workspace();
-	if (!workspace)
+	if (IS_ERR(workspace))
 		return -ENOMEM;
 
 	data_in = kmap(pages_in[page_in_index]);
@@ -547,7 +547,7 @@ int btrfs_zlib_decompress(unsigned char *data_in,
 		return -ENOMEM;
 
 	workspace = find_zlib_workspace();
-	if (!workspace)
+	if (IS_ERR(workspace))
 		return -ENOMEM;
 
 	workspace->inf_strm.next_in = data_in;
diff --git a/fs/buffer.c b/fs/buffer.c
index a3ef091a45bd..28f320fac4d4 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -1165,8 +1165,11 @@ void mark_buffer_dirty(struct buffer_head *bh)
 
 	if (!test_set_buffer_dirty(bh)) {
 		struct page *page = bh->b_page;
-		if (!TestSetPageDirty(page))
-			__set_page_dirty(page, page_mapping(page), 0);
+		if (!TestSetPageDirty(page)) {
+			struct address_space *mapping = page_mapping(page);
+			if (mapping)
+				__set_page_dirty(page, mapping, 0);
+		}
 	}
 }
 
diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES
index 92888aa90749..e85b1e4389e0 100644
--- a/fs/cifs/CHANGES
+++ b/fs/cifs/CHANGES
@@ -1,3 +1,10 @@
+Version 1.60
+-------------
+Fix memory leak in reconnect.  Fix oops in DFS mount error path.
+Set s_maxbytes to smaller (the max that vfs can handle) so that
+sendfile will now work over cifs mounts again.  Add noforcegid
+and noforceuid mount parameters.
+
 Version 1.59
 ------------
 Client uses server inode numbers (which are persistent) rather than
diff --git a/fs/cifs/README b/fs/cifs/README
index ad92921dbde4..79c1a93400be 100644
--- a/fs/cifs/README
+++ b/fs/cifs/README
@@ -262,11 +262,11 @@ A partial list of the supported mount options follows:
 		mount.	
   domain	Set the SMB/CIFS workgroup name prepended to the
 		username during CIFS session establishment
-  forceuid	Set the default uid for inodes based on the uid
-		passed in. For mounts to servers
+  forceuid	Set the default uid for inodes to the uid
+		passed in on mount. For mounts to servers
 		which do support the CIFS Unix extensions, such as a
 		properly configured Samba server, the server provides
-		the uid, gid and mode so this parameter should  not be
+		the uid, gid and mode so this parameter should not be
 		specified unless the server and clients uid and gid
 		numbering differ.  If the server and client are in the
 		same domain (e.g. running winbind or nss_ldap) and
@@ -278,11 +278,7 @@ A partial list of the supported mount options follows:
 		of existing files will be the uid (gid) of the person
 		who executed the mount (root, except when mount.cifs
 		is configured setuid for user mounts) unless the "uid=" 
-		(gid) mount option is specified.  For the uid (gid) of newly
-		created files and directories, ie files created since 
-		the last mount of the server share, the expected uid 
-		(gid) is cached as long as the inode remains in 
-		memory on the client.   Also note that permission
+		(gid) mount option is specified. Also note that permission
 		checks (authorization checks) on accesses to a file occur
 		at the server, but there are cases in which an administrator
 		may want to restrict at the client as well.  For those
@@ -290,12 +286,15 @@ A partial list of the supported mount options follows:
 		(such as Windows), permissions can also be checked at the
 		client, and a crude form of client side permission checking 
 		can be enabled by specifying file_mode and dir_mode on 
-		the client.  Note that the mount.cifs helper must be
-		at version 1.10 or higher to support specifying the uid
-		(or gid) in non-numeric form.
-  forcegid	(similar to above but for the groupid instead of uid)
+		the client.  (default)
+  forcegid	(similar to above but for the groupid instead of uid) (default)
+  noforceuid	Fill in file owner information (uid) by requesting it from
+		the server if possible. With this option, the value given in
+		the uid= option (on mount) will only be used if the server
+		can not support returning uids on inodes.
+  noforcegid	(similar to above but for the group owner, gid, instead of uid)
   uid		Set the default uid for inodes, and indicate to the
-		cifs kernel driver which local user mounted . If the server
+		cifs kernel driver which local user mounted. If the server
 		supports the unix extensions the default uid is
 		not used to fill in the owner fields of inodes (files)
 		unless the "forceuid" parameter is specified.
diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c
index 3bb11be8b6a8..606912d8f2a8 100644
--- a/fs/cifs/cifs_dfs_ref.c
+++ b/fs/cifs/cifs_dfs_ref.c
@@ -55,7 +55,7 @@ void cifs_dfs_release_automount_timer(void)
  * i.e. strips from UNC trailing path that is not part of share
  * name and fixup missing '\' in the begining of DFS node refferal
  * if neccessary.
- * Returns pointer to share name on success or NULL on error.
+ * Returns pointer to share name on success or ERR_PTR on error.
  * Caller is responsible for freeing returned string.
  */
 static char *cifs_get_share_name(const char *node_name)
@@ -68,7 +68,7 @@ static char *cifs_get_share_name(const char *node_name)
 	UNC = kmalloc(len+2 /*for term null and additional \ if it's missed */,
 			 GFP_KERNEL);
 	if (!UNC)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
 	/* get share name and server name */
 	if (node_name[1] != '\\') {
@@ -87,7 +87,7 @@ static char *cifs_get_share_name(const char *node_name)
 		cERROR(1, ("%s: no server name end in node name: %s",
 			__func__, node_name));
 		kfree(UNC);
-		return NULL;
+		return ERR_PTR(-EINVAL);
 	}
 
 	/* find sharename end */
@@ -133,6 +133,12 @@ char *cifs_compose_mount_options(const char *sb_mountdata,
 		return ERR_PTR(-EINVAL);
 
 	*devname = cifs_get_share_name(ref->node_name);
+	if (IS_ERR(*devname)) {
+		rc = PTR_ERR(*devname);
+		*devname = NULL;
+		goto compose_mount_options_err;
+	}
+
 	rc = dns_resolve_server_name_to_ip(*devname, &srvIP);
 	if (rc != 0) {
 		cERROR(1, ("%s: Failed to resolve server part of %s to IP: %d",
diff --git a/fs/cifs/cifs_unicode.c b/fs/cifs/cifs_unicode.c
index 60e3c4253de0..714a542cbafc 100644
--- a/fs/cifs/cifs_unicode.c
+++ b/fs/cifs/cifs_unicode.c
@@ -44,7 +44,7 @@ cifs_ucs2_bytes(const __le16 *from, int maxbytes,
 	int maxwords = maxbytes / 2;
 	char tmp[NLS_MAX_CHARSET_SIZE];
 
-	for (i = 0; from[i] && i < maxwords; i++) {
+	for (i = 0; i < maxwords && from[i]; i++) {
 		charlen = codepage->uni2char(le16_to_cpu(from[i]), tmp,
 					     NLS_MAX_CHARSET_SIZE);
 		if (charlen > 0)
diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index 44f30504b82d..84b75253b05a 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -376,10 +376,14 @@ cifs_show_options(struct seq_file *s, struct vfsmount *m)
 	seq_printf(s, ",uid=%d", cifs_sb->mnt_uid);
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_UID)
 		seq_printf(s, ",forceuid");
+	else
+		seq_printf(s, ",noforceuid");
 
 	seq_printf(s, ",gid=%d", cifs_sb->mnt_gid);
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_GID)
 		seq_printf(s, ",forcegid");
+	else
+		seq_printf(s, ",noforcegid");
 
 	cifs_show_address(s, tcon->ses->server);
 
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 9bb5c8750736..1f3345d7fa79 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -803,6 +803,10 @@ cifs_parse_mount_options(char *options, const char *devname,
 	char *data;
 	unsigned int  temp_len, i, j;
 	char separator[2];
+	short int override_uid = -1;
+	short int override_gid = -1;
+	bool uid_specified = false;
+	bool gid_specified = false;
 
 	separator[0] = ',';
 	separator[1] = 0;
@@ -1093,18 +1097,20 @@ cifs_parse_mount_options(char *options, const char *devname,
 						    "too long.\n");
 				return 1;
 			}
-		} else if (strnicmp(data, "uid", 3) == 0) {
-			if (value && *value)
-				vol->linux_uid =
-					simple_strtoul(value, &value, 0);
-		} else if (strnicmp(data, "forceuid", 8) == 0) {
-				vol->override_uid = 1;
-		} else if (strnicmp(data, "gid", 3) == 0) {
-			if (value && *value)
-				vol->linux_gid =
-					simple_strtoul(value, &value, 0);
-		} else if (strnicmp(data, "forcegid", 8) == 0) {
-				vol->override_gid = 1;
+		} else if (!strnicmp(data, "uid", 3) && value && *value) {
+			vol->linux_uid = simple_strtoul(value, &value, 0);
+			uid_specified = true;
+		} else if (!strnicmp(data, "forceuid", 8)) {
+			override_uid = 1;
+		} else if (!strnicmp(data, "noforceuid", 10)) {
+			override_uid = 0;
+		} else if (!strnicmp(data, "gid", 3) && value && *value) {
+			vol->linux_gid = simple_strtoul(value, &value, 0);
+			gid_specified = true;
+		} else if (!strnicmp(data, "forcegid", 8)) {
+			override_gid = 1;
+		} else if (!strnicmp(data, "noforcegid", 10)) {
+			override_gid = 0;
 		} else if (strnicmp(data, "file_mode", 4) == 0) {
 			if (value && *value) {
 				vol->file_mode =
@@ -1355,6 +1361,18 @@ cifs_parse_mount_options(char *options, const char *devname,
 	if (vol->UNCip == NULL)
 		vol->UNCip = &vol->UNC[2];
 
+	if (uid_specified)
+		vol->override_uid = override_uid;
+	else if (override_uid == 1)
+		printk(KERN_NOTICE "CIFS: ignoring forceuid mount option "
+				   "specified with no uid= option.\n");
+
+	if (gid_specified)
+		vol->override_gid = override_gid;
+	else if (override_gid == 1)
+		printk(KERN_NOTICE "CIFS: ignoring forcegid mount option "
+				   "specified with no gid= option.\n");
+
 	return 0;
 }
 
@@ -2452,10 +2470,10 @@ try_mount_again:
 		tcon->local_lease = volume_info->local_lease;
 	}
 	if (pSesInfo) {
-		if (pSesInfo->capabilities & CAP_LARGE_FILES) {
-			sb->s_maxbytes = (u64) 1 << 63;
-		} else
-			sb->s_maxbytes = (u64) 1 << 31;	/* 2 GB */
+		if (pSesInfo->capabilities & CAP_LARGE_FILES)
+			sb->s_maxbytes = MAX_LFS_FILESIZE;
+		else
+			sb->s_maxbytes = MAX_NON_LFS;
 	}
 
 	/* BB FIXME fix time_gran to be larger for LANMAN sessions */
@@ -2544,11 +2562,20 @@ remote_path_check:
 
 			if (mount_data != mount_data_global)
 				kfree(mount_data);
+
 			mount_data = cifs_compose_mount_options(
 					cifs_sb->mountdata, full_path + 1,
 					referrals, &fake_devname);
-			kfree(fake_devname);
+
 			free_dfs_info_array(referrals, num_referrals);
+			kfree(fake_devname);
+			kfree(full_path);
+
+			if (IS_ERR(mount_data)) {
+				rc = PTR_ERR(mount_data);
+				mount_data = NULL;
+				goto mount_fail_check;
+			}
 
 			if (tcon)
 				cifs_put_tcon(tcon);
@@ -2556,8 +2583,6 @@ remote_path_check:
 				cifs_put_smb_ses(pSesInfo);
 
 			cleanup_volume_info(&volume_info);
-			FreeXid(xid);
-			kfree(full_path);
 			referral_walks_count++;
 			goto try_mount_again;
 		}
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 18afe57b2461..82d83839655e 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -212,7 +212,7 @@ cifs_unix_basic_to_fattr(struct cifs_fattr *fattr, FILE_UNIX_BASIC_INFO *info,
  * junction to the new submount (ie to setup the fake directory
  * which represents a DFS referral).
  */
-void
+static void
 cifs_create_dfs_fattr(struct cifs_fattr *fattr, struct super_block *sb)
 {
 	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
@@ -388,7 +388,7 @@ static int cifs_sfu_mode(struct cifs_fattr *fattr, const unsigned char *path,
 }
 
 /* Fill a cifs_fattr struct with info from FILE_ALL_INFO */
-void
+static void
 cifs_all_info_to_fattr(struct cifs_fattr *fattr, FILE_ALL_INFO *info,
 		       struct cifs_sb_info *cifs_sb, bool adjust_tz)
 {
@@ -513,9 +513,12 @@ int cifs_get_inode_info(struct inode **pinode,
 					cifs_sb->mnt_cifs_flags &
 						CIFS_MOUNT_MAP_SPECIAL_CHR);
 			if (rc1) {
-				/* BB EOPNOSUPP disable SERVER_INUM? */
 				cFYI(1, ("GetSrvInodeNum rc %d", rc1));
 				fattr.cf_uniqueid = iunique(sb, ROOT_I);
+				/* disable serverino if call not supported */
+				if (rc1 == -EINVAL)
+					cifs_sb->mnt_cifs_flags &=
+							~CIFS_MOUNT_SERVER_INUM;
 			}
 		} else {
 			fattr.cf_uniqueid = iunique(sb, ROOT_I);
diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c
index f28f070a60fc..f91fd51b32e3 100644
--- a/fs/compat_ioctl.c
+++ b/fs/compat_ioctl.c
@@ -1905,6 +1905,7 @@ COMPATIBLE_IOCTL(FIONCLEX)
 COMPATIBLE_IOCTL(FIOASYNC)
 COMPATIBLE_IOCTL(FIONBIO)
 COMPATIBLE_IOCTL(FIONREAD)  /* This is also TIOCINQ */
+COMPATIBLE_IOCTL(FS_IOC_FIEMAP)
 /* 0x00 */
 COMPATIBLE_IOCTL(FIBMAP)
 COMPATIBLE_IOCTL(FIGETBSZ)
diff --git a/fs/ecryptfs/keystore.c b/fs/ecryptfs/keystore.c
index af737bb56cb7..259525c9abb8 100644
--- a/fs/ecryptfs/keystore.c
+++ b/fs/ecryptfs/keystore.c
@@ -1303,6 +1303,13 @@ parse_tag_3_packet(struct ecryptfs_crypt_stat *crypt_stat,
 	}
 	(*new_auth_tok)->session_key.encrypted_key_size =
 		(body_size - (ECRYPTFS_SALT_SIZE + 5));
+	if ((*new_auth_tok)->session_key.encrypted_key_size
+	    > ECRYPTFS_MAX_ENCRYPTED_KEY_BYTES) {
+		printk(KERN_WARNING "Tag 3 packet contains key larger "
+		       "than ECRYPTFS_MAX_ENCRYPTED_KEY_BYTES\n");
+		rc = -EINVAL;
+		goto out_free;
+	}
 	if (unlikely(data[(*packet_size)++] != 0x04)) {
 		printk(KERN_WARNING "Unknown version number [%d]\n",
 		       data[(*packet_size) - 1]);
@@ -1449,6 +1456,12 @@ parse_tag_11_packet(unsigned char *data, unsigned char *contents,
 		rc = -EINVAL;
 		goto out;
 	}
+	if (unlikely((*tag_11_contents_size) > max_contents_bytes)) {
+		printk(KERN_ERR "Literal data section in tag 11 packet exceeds "
+		       "expected size\n");
+		rc = -EINVAL;
+		goto out;
+	}
 	if (data[(*packet_size)++] != 0x62) {
 		printk(KERN_WARNING "Unrecognizable packet\n");
 		rc = -EINVAL;
diff --git a/fs/exec.c b/fs/exec.c
index 4a8849e45b21..fb4f3cdda78c 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -678,8 +678,8 @@ exit:
 }
 EXPORT_SYMBOL(open_exec);
 
-int kernel_read(struct file *file, unsigned long offset,
-	char *addr, unsigned long count)
+int kernel_read(struct file *file, loff_t offset,
+		char *addr, unsigned long count)
 {
 	mm_segment_t old_fs;
 	loff_t pos = offset;
diff --git a/fs/ext3/Kconfig b/fs/ext3/Kconfig
index fb3c1a21b135..522b15498f45 100644
--- a/fs/ext3/Kconfig
+++ b/fs/ext3/Kconfig
@@ -29,23 +29,25 @@ config EXT3_FS
 	  module will be called ext3.
 
 config EXT3_DEFAULTS_TO_ORDERED
-	bool "Default to 'data=ordered' in ext3 (legacy option)"
+	bool "Default to 'data=ordered' in ext3"
 	depends on EXT3_FS
 	help
-	  If a filesystem does not explicitly specify a data ordering
-	  mode, and the journal capability allowed it, ext3 used to
-	  historically default to 'data=ordered'.
-
-	  That was a rather unfortunate choice, because it leads to all
-	  kinds of latency problems, and the 'data=writeback' mode is more
-	  appropriate these days.
-
-	  You should probably always answer 'n' here, and if you really
-	  want to use 'data=ordered' mode, set it in the filesystem itself
-	  with 'tune2fs -o journal_data_ordered'.
-
-	  But if you really want to enable the legacy default, you can do
-	  so by answering 'y' to this question.
+	  The journal mode options for ext3 have different tradeoffs
+	  between when data is guaranteed to be on disk and
+	  performance.	The use of "data=writeback" can cause
+	  unwritten data to appear in files after an system crash or
+	  power failure, which can be a security issue.	 However,
+	  "data=ordered" mode can also result in major performance
+	  problems, including seconds-long delays before an fsync()
+	  call returns.	 For details, see:
+
+	  http://ext4.wiki.kernel.org/index.php/Ext3_data_mode_tradeoffs
+
+	  If you have been historically happy with ext3's performance,
+	  data=ordered mode will be a safe choice and you should
+	  answer 'y' here.  If you understand the reliability and data
+	  privacy issues of data=writeback and are willing to make
+	  that trade off, answer 'n'.
 
 config EXT3_FS_XATTR
 	bool "Ext3 extended attributes"
diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c
index 3d724a95882f..373fa90c796a 100644
--- a/fs/ext3/dir.c
+++ b/fs/ext3/dir.c
@@ -130,8 +130,7 @@ static int ext3_readdir(struct file * filp,
 		struct buffer_head *bh = NULL;
 
 		map_bh.b_state = 0;
-		err = ext3_get_blocks_handle(NULL, inode, blk, 1,
-						&map_bh, 0, 0);
+		err = ext3_get_blocks_handle(NULL, inode, blk, 1, &map_bh, 0);
 		if (err > 0) {
 			pgoff_t index = map_bh.b_blocknr >>
 					(PAGE_CACHE_SHIFT - inode->i_blkbits);
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index 5f51fed5c750..b49908a167ae 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -788,7 +788,7 @@ err_out:
 int ext3_get_blocks_handle(handle_t *handle, struct inode *inode,
 		sector_t iblock, unsigned long maxblocks,
 		struct buffer_head *bh_result,
-		int create, int extend_disksize)
+		int create)
 {
 	int err = -EIO;
 	int offsets[4];
@@ -911,13 +911,6 @@ int ext3_get_blocks_handle(handle_t *handle, struct inode *inode,
 	if (!err)
 		err = ext3_splice_branch(handle, inode, iblock,
 					partial, indirect_blks, count);
-	/*
-	 * i_disksize growing is protected by truncate_mutex.  Don't forget to
-	 * protect it if you're about to implement concurrent
-	 * ext3_get_block() -bzzz
-	*/
-	if (!err && extend_disksize && inode->i_size > ei->i_disksize)
-		ei->i_disksize = inode->i_size;
 	mutex_unlock(&ei->truncate_mutex);
 	if (err)
 		goto cleanup;
@@ -972,7 +965,7 @@ static int ext3_get_block(struct inode *inode, sector_t iblock,
 	}
 
 	ret = ext3_get_blocks_handle(handle, inode, iblock,
-					max_blocks, bh_result, create, 0);
+					max_blocks, bh_result, create);
 	if (ret > 0) {
 		bh_result->b_size = (ret << inode->i_blkbits);
 		ret = 0;
@@ -1005,7 +998,7 @@ struct buffer_head *ext3_getblk(handle_t *handle, struct inode *inode,
 	dummy.b_blocknr = -1000;
 	buffer_trace_init(&dummy.b_history);
 	err = ext3_get_blocks_handle(handle, inode, block, 1,
-					&dummy, create, 1);
+					&dummy, create);
 	/*
 	 * ext3_get_blocks_handle() returns number of blocks
 	 * mapped. 0 in case of a HOLE.
@@ -1193,15 +1186,16 @@ write_begin_failed:
 		 * i_size_read because we hold i_mutex.
 		 *
 		 * Add inode to orphan list in case we crash before truncate
-		 * finishes.
+		 * finishes. Do this only if ext3_can_truncate() agrees so
+		 * that orphan processing code is happy.
 		 */
-		if (pos + len > inode->i_size)
+		if (pos + len > inode->i_size && ext3_can_truncate(inode))
 			ext3_orphan_add(handle, inode);
 		ext3_journal_stop(handle);
 		unlock_page(page);
 		page_cache_release(page);
 		if (pos + len > inode->i_size)
-			vmtruncate(inode, inode->i_size);
+			ext3_truncate(inode);
 	}
 	if (ret == -ENOSPC && ext3_should_retry_alloc(inode->i_sb, &retries))
 		goto retry;
@@ -1287,7 +1281,7 @@ static int ext3_ordered_write_end(struct file *file,
 	 * There may be allocated blocks outside of i_size because
 	 * we failed to copy some data. Prepare for truncate.
 	 */
-	if (pos + len > inode->i_size)
+	if (pos + len > inode->i_size && ext3_can_truncate(inode))
 		ext3_orphan_add(handle, inode);
 	ret2 = ext3_journal_stop(handle);
 	if (!ret)
@@ -1296,7 +1290,7 @@ static int ext3_ordered_write_end(struct file *file,
 	page_cache_release(page);
 
 	if (pos + len > inode->i_size)
-		vmtruncate(inode, inode->i_size);
+		ext3_truncate(inode);
 	return ret ? ret : copied;
 }
 
@@ -1315,14 +1309,14 @@ static int ext3_writeback_write_end(struct file *file,
 	 * There may be allocated blocks outside of i_size because
 	 * we failed to copy some data. Prepare for truncate.
 	 */
-	if (pos + len > inode->i_size)
+	if (pos + len > inode->i_size && ext3_can_truncate(inode))
 		ext3_orphan_add(handle, inode);
 	ret = ext3_journal_stop(handle);
 	unlock_page(page);
 	page_cache_release(page);
 
 	if (pos + len > inode->i_size)
-		vmtruncate(inode, inode->i_size);
+		ext3_truncate(inode);
 	return ret ? ret : copied;
 }
 
@@ -1358,7 +1352,7 @@ static int ext3_journalled_write_end(struct file *file,
 	 * There may be allocated blocks outside of i_size because
 	 * we failed to copy some data. Prepare for truncate.
 	 */
-	if (pos + len > inode->i_size)
+	if (pos + len > inode->i_size && ext3_can_truncate(inode))
 		ext3_orphan_add(handle, inode);
 	EXT3_I(inode)->i_state |= EXT3_STATE_JDATA;
 	if (inode->i_size > EXT3_I(inode)->i_disksize) {
@@ -1375,7 +1369,7 @@ static int ext3_journalled_write_end(struct file *file,
 	page_cache_release(page);
 
 	if (pos + len > inode->i_size)
-		vmtruncate(inode, inode->i_size);
+		ext3_truncate(inode);
 	return ret ? ret : copied;
 }
 
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 524b349c6299..a8d80a7f1105 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -543,6 +543,19 @@ static inline void ext3_show_quota_options(struct seq_file *seq, struct super_bl
 #endif
 }
 
+static char *data_mode_string(unsigned long mode)
+{
+	switch (mode) {
+	case EXT3_MOUNT_JOURNAL_DATA:
+		return "journal";
+	case EXT3_MOUNT_ORDERED_DATA:
+		return "ordered";
+	case EXT3_MOUNT_WRITEBACK_DATA:
+		return "writeback";
+	}
+	return "unknown";
+}
+
 /*
  * Show an option if
  *  - it's set to a non-default value OR
@@ -616,13 +629,8 @@ static int ext3_show_options(struct seq_file *seq, struct vfsmount *vfs)
 	if (test_opt(sb, NOBH))
 		seq_puts(seq, ",nobh");
 
-	if (test_opt(sb, DATA_FLAGS) == EXT3_MOUNT_JOURNAL_DATA)
-		seq_puts(seq, ",data=journal");
-	else if (test_opt(sb, DATA_FLAGS) == EXT3_MOUNT_ORDERED_DATA)
-		seq_puts(seq, ",data=ordered");
-	else if (test_opt(sb, DATA_FLAGS) == EXT3_MOUNT_WRITEBACK_DATA)
-		seq_puts(seq, ",data=writeback");
-
+	seq_printf(seq, ",data=%s", data_mode_string(sbi->s_mount_opt &
+						     EXT3_MOUNT_DATA_FLAGS));
 	if (test_opt(sb, DATA_ERR_ABORT))
 		seq_puts(seq, ",data_err=abort");
 
@@ -1024,12 +1032,18 @@ static int parse_options (char *options, struct super_block *sb,
 		datacheck:
 			if (is_remount) {
 				if ((sbi->s_mount_opt & EXT3_MOUNT_DATA_FLAGS)
-						!= data_opt) {
-					printk(KERN_ERR
-						"EXT3-fs: cannot change data "
-						"mode on remount\n");
-					return 0;
-				}
+						== data_opt)
+					break;
+				printk(KERN_ERR
+					"EXT3-fs (device %s): Cannot change "
+					"data mode on remount. The filesystem "
+					"is mounted in data=%s mode and you "
+					"try to remount it in data=%s mode.\n",
+					sb->s_id,
+					data_mode_string(sbi->s_mount_opt &
+							EXT3_MOUNT_DATA_FLAGS),
+					data_mode_string(data_opt));
+				return 0;
 			} else {
 				sbi->s_mount_opt &= ~EXT3_MOUNT_DATA_FLAGS;
 				sbi->s_mount_opt |= data_opt;
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index 03ebb439ace0..7ebae9a4ecc0 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -624,6 +624,7 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 {
 	struct gfs2_inode *ip = GFS2_I(mapping->host);
 	struct gfs2_sbd *sdp = GFS2_SB(mapping->host);
+	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
 	unsigned int data_blocks = 0, ind_blocks = 0, rblocks;
 	int alloc_required;
 	int error = 0;
@@ -637,6 +638,14 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 	error = gfs2_glock_nq(&ip->i_gh);
 	if (unlikely(error))
 		goto out_uninit;
+	if (&ip->i_inode == sdp->sd_rindex) {
+		error = gfs2_glock_nq_init(m_ip->i_gl, LM_ST_EXCLUSIVE,
+					   GL_NOCACHE, &m_ip->i_gh);
+		if (unlikely(error)) {
+			gfs2_glock_dq(&ip->i_gh);
+			goto out_uninit;
+		}
+	}
 
 	error = gfs2_write_alloc_required(ip, pos, len, &alloc_required);
 	if (error)
@@ -667,6 +676,8 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 		rblocks += data_blocks ? data_blocks : 1;
 	if (ind_blocks || data_blocks)
 		rblocks += RES_STATFS + RES_QUOTA;
+	if (&ip->i_inode == sdp->sd_rindex)
+		rblocks += 2 * RES_STATFS;
 
 	error = gfs2_trans_begin(sdp, rblocks,
 				 PAGE_CACHE_SIZE/sdp->sd_sb.sb_bsize);
@@ -712,6 +723,10 @@ out_alloc_put:
 		gfs2_alloc_put(ip);
 	}
 out_unlock:
+	if (&ip->i_inode == sdp->sd_rindex) {
+		gfs2_glock_dq(&m_ip->i_gh);
+		gfs2_holder_uninit(&m_ip->i_gh);
+	}
 	gfs2_glock_dq(&ip->i_gh);
 out_uninit:
 	gfs2_holder_uninit(&ip->i_gh);
@@ -725,14 +740,21 @@ out_uninit:
 static void adjust_fs_space(struct inode *inode)
 {
 	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
+	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
+	struct gfs2_inode *l_ip = GFS2_I(sdp->sd_sc_inode);
 	struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
 	struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
+	struct buffer_head *m_bh, *l_bh;
 	u64 fs_total, new_free;
 
 	/* Total up the file system space, according to the latest rindex. */
 	fs_total = gfs2_ri_total(sdp);
+	if (gfs2_meta_inode_buffer(m_ip, &m_bh) != 0)
+		return;
 
 	spin_lock(&sdp->sd_statfs_spin);
+	gfs2_statfs_change_in(m_sc, m_bh->b_data +
+			      sizeof(struct gfs2_dinode));
 	if (fs_total > (m_sc->sc_total + l_sc->sc_total))
 		new_free = fs_total - (m_sc->sc_total + l_sc->sc_total);
 	else
@@ -741,6 +763,13 @@ static void adjust_fs_space(struct inode *inode)
 	fs_warn(sdp, "File system extended by %llu blocks.\n",
 		(unsigned long long)new_free);
 	gfs2_statfs_change(sdp, new_free, new_free, 0);
+
+	if (gfs2_meta_inode_buffer(l_ip, &l_bh) != 0)
+		goto out;
+	update_statfs(sdp, m_bh, l_bh);
+	brelse(l_bh);
+out:
+	brelse(m_bh);
 }
 
 /**
@@ -763,6 +792,7 @@ static int gfs2_stuffed_write_end(struct inode *inode, struct buffer_head *dibh,
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
+	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
 	u64 to = pos + copied;
 	void *kaddr;
 	unsigned char *buf = dibh->b_data + sizeof(struct gfs2_dinode);
@@ -794,6 +824,10 @@ static int gfs2_stuffed_write_end(struct inode *inode, struct buffer_head *dibh,
 
 	brelse(dibh);
 	gfs2_trans_end(sdp);
+	if (inode == sdp->sd_rindex) {
+		gfs2_glock_dq(&m_ip->i_gh);
+		gfs2_holder_uninit(&m_ip->i_gh);
+	}
 	gfs2_glock_dq(&ip->i_gh);
 	gfs2_holder_uninit(&ip->i_gh);
 	return copied;
@@ -823,6 +857,7 @@ static int gfs2_write_end(struct file *file, struct address_space *mapping,
 	struct inode *inode = page->mapping->host;
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
+	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
 	struct buffer_head *dibh;
 	struct gfs2_alloc *al = ip->i_alloc;
 	unsigned int from = pos & (PAGE_CACHE_SIZE - 1);
@@ -865,6 +900,10 @@ failed:
 		gfs2_quota_unlock(ip);
 		gfs2_alloc_put(ip);
 	}
+	if (inode == sdp->sd_rindex) {
+		gfs2_glock_dq(&m_ip->i_gh);
+		gfs2_holder_uninit(&m_ip->i_gh);
+	}
 	gfs2_glock_dq(&ip->i_gh);
 	gfs2_holder_uninit(&ip->i_gh);
 	return ret;
diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 297421c0427a..8b674b1f3a55 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -63,6 +63,7 @@ static void do_xmote(struct gfs2_glock *gl, struct gfs2_holder *gh, unsigned int
 static DECLARE_RWSEM(gfs2_umount_flush_sem);
 static struct dentry *gfs2_root;
 static struct workqueue_struct *glock_workqueue;
+struct workqueue_struct *gfs2_delete_workqueue;
 static LIST_HEAD(lru_list);
 static atomic_t lru_count = ATOMIC_INIT(0);
 static DEFINE_SPINLOCK(lru_lock);
@@ -167,13 +168,33 @@ static void glock_free(struct gfs2_glock *gl)
  *
  */
 
-static void gfs2_glock_hold(struct gfs2_glock *gl)
+void gfs2_glock_hold(struct gfs2_glock *gl)
 {
 	GLOCK_BUG_ON(gl, atomic_read(&gl->gl_ref) == 0);
 	atomic_inc(&gl->gl_ref);
 }
 
 /**
+ * demote_ok - Check to see if it's ok to unlock a glock
+ * @gl: the glock
+ *
+ * Returns: 1 if it's ok
+ */
+
+static int demote_ok(const struct gfs2_glock *gl)
+{
+	const struct gfs2_glock_operations *glops = gl->gl_ops;
+
+	if (gl->gl_state == LM_ST_UNLOCKED)
+		return 0;
+	if (!list_empty(&gl->gl_holders))
+		return 0;
+	if (glops->go_demote_ok)
+		return glops->go_demote_ok(gl);
+	return 1;
+}
+
+/**
  * gfs2_glock_schedule_for_reclaim - Add a glock to the reclaim list
  * @gl: the glock
  *
@@ -181,8 +202,13 @@ static void gfs2_glock_hold(struct gfs2_glock *gl)
 
 static void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
 {
+	int may_reclaim;
+	may_reclaim = (demote_ok(gl) &&
+		       (atomic_read(&gl->gl_ref) == 1 ||
+			(gl->gl_name.ln_type == LM_TYPE_INODE &&
+			 atomic_read(&gl->gl_ref) <= 2)));
 	spin_lock(&lru_lock);
-	if (list_empty(&gl->gl_lru) && gl->gl_state != LM_ST_UNLOCKED) {
+	if (list_empty(&gl->gl_lru) && may_reclaim) {
 		list_add_tail(&gl->gl_lru, &lru_list);
 		atomic_inc(&lru_count);
 	}
@@ -190,6 +216,21 @@ static void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
 }
 
 /**
+ * gfs2_glock_put_nolock() - Decrement reference count on glock
+ * @gl: The glock to put
+ *
+ * This function should only be used if the caller has its own reference
+ * to the glock, in addition to the one it is dropping.
+ */
+
+void gfs2_glock_put_nolock(struct gfs2_glock *gl)
+{
+	if (atomic_dec_and_test(&gl->gl_ref))
+		GLOCK_BUG_ON(gl, 1);
+	gfs2_glock_schedule_for_reclaim(gl);
+}
+
+/**
  * gfs2_glock_put() - Decrement reference count on glock
  * @gl: The glock to put
  *
@@ -214,9 +255,9 @@ int gfs2_glock_put(struct gfs2_glock *gl)
 		rv = 1;
 		goto out;
 	}
-	/* 1 for being hashed, 1 for having state != LM_ST_UNLOCKED */
-	if (atomic_read(&gl->gl_ref) == 2)
-		gfs2_glock_schedule_for_reclaim(gl);
+	spin_lock(&gl->gl_spin);
+	gfs2_glock_schedule_for_reclaim(gl);
+	spin_unlock(&gl->gl_spin);
 	write_unlock(gl_lock_addr(gl->gl_hash));
 out:
 	return rv;
@@ -398,7 +439,7 @@ static void state_change(struct gfs2_glock *gl, unsigned int new_state)
 		if (held2)
 			gfs2_glock_hold(gl);
 		else
-			gfs2_glock_put(gl);
+			gfs2_glock_put_nolock(gl);
 	}
 
 	gl->gl_state = new_state;
@@ -633,12 +674,35 @@ out:
 out_sched:
 	gfs2_glock_hold(gl);
 	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
-		gfs2_glock_put(gl);
+		gfs2_glock_put_nolock(gl);
 out_unlock:
 	clear_bit(GLF_LOCK, &gl->gl_flags);
 	goto out;
 }
 
+static void delete_work_func(struct work_struct *work)
+{
+	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_delete);
+	struct gfs2_sbd *sdp = gl->gl_sbd;
+	struct gfs2_inode *ip = NULL;
+	struct inode *inode;
+	u64 no_addr = 0;
+
+	spin_lock(&gl->gl_spin);
+	ip = (struct gfs2_inode *)gl->gl_object;
+	if (ip)
+		no_addr = ip->i_no_addr;
+	spin_unlock(&gl->gl_spin);
+	if (ip) {
+		inode = gfs2_ilookup(sdp->sd_vfs, no_addr);
+		if (inode) {
+			d_prune_aliases(inode);
+			iput(inode);
+		}
+	}
+	gfs2_glock_put(gl);
+}
+
 static void glock_work_func(struct work_struct *work)
 {
 	unsigned long delay = 0;
@@ -717,6 +781,7 @@ int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
 	gl->gl_sbd = sdp;
 	gl->gl_aspace = NULL;
 	INIT_DELAYED_WORK(&gl->gl_work, glock_work_func);
+	INIT_WORK(&gl->gl_delete, delete_work_func);
 
 	/* If this glock protects actual on-disk data or metadata blocks,
 	   create a VFS inode to manage the pages/buffers holding them. */
@@ -858,6 +923,8 @@ static void handle_callback(struct gfs2_glock *gl, unsigned int state,
 			gl->gl_demote_state != state) {
 		gl->gl_demote_state = LM_ST_UNLOCKED;
 	}
+	if (gl->gl_ops->go_callback)
+		gl->gl_ops->go_callback(gl);
 	trace_gfs2_demote_rq(gl);
 }
 
@@ -1274,33 +1341,12 @@ void gfs2_glock_complete(struct gfs2_glock *gl, int ret)
 		gfs2_glock_put(gl);
 }
 
-/**
- * demote_ok - Check to see if it's ok to unlock a glock
- * @gl: the glock
- *
- * Returns: 1 if it's ok
- */
-
-static int demote_ok(const struct gfs2_glock *gl)
-{
-	const struct gfs2_glock_operations *glops = gl->gl_ops;
-
-	if (gl->gl_state == LM_ST_UNLOCKED)
-		return 0;
-	if (!list_empty(&gl->gl_holders))
-		return 0;
-	if (glops->go_demote_ok)
-		return glops->go_demote_ok(gl);
-	return 1;
-}
-
 
 static int gfs2_shrink_glock_memory(int nr, gfp_t gfp_mask)
 {
 	struct gfs2_glock *gl;
 	int may_demote;
 	int nr_skipped = 0;
-	int got_ref = 0;
 	LIST_HEAD(skipped);
 
 	if (nr == 0)
@@ -1315,37 +1361,29 @@ static int gfs2_shrink_glock_memory(int nr, gfp_t gfp_mask)
 		list_del_init(&gl->gl_lru);
 		atomic_dec(&lru_count);
 
+		/* Check if glock is about to be freed */
+		if (atomic_read(&gl->gl_ref) == 0)
+			continue;
+
 		/* Test for being demotable */
 		if (!test_and_set_bit(GLF_LOCK, &gl->gl_flags)) {
 			gfs2_glock_hold(gl);
-			got_ref = 1;
 			spin_unlock(&lru_lock);
 			spin_lock(&gl->gl_spin);
 			may_demote = demote_ok(gl);
-			spin_unlock(&gl->gl_spin);
-			clear_bit(GLF_LOCK, &gl->gl_flags);
 			if (may_demote) {
 				handle_callback(gl, LM_ST_UNLOCKED, 0);
 				nr--;
-				if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
-					gfs2_glock_put(gl);
-				got_ref = 0;
 			}
+			if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
+				gfs2_glock_put_nolock(gl);
+			spin_unlock(&gl->gl_spin);
+			clear_bit(GLF_LOCK, &gl->gl_flags);
 			spin_lock(&lru_lock);
-			if (may_demote)
-				continue;
-		}
-		if (list_empty(&gl->gl_lru) &&
-		    (atomic_read(&gl->gl_ref) <= (2 + got_ref))) {
-			nr_skipped++;
-			list_add(&gl->gl_lru, &skipped);
-		}
-		if (got_ref) {
-			spin_unlock(&lru_lock);
-			gfs2_glock_put(gl);
-			spin_lock(&lru_lock);
-			got_ref = 0;
+			continue;
 		}
+		nr_skipped++;
+		list_add(&gl->gl_lru, &skipped);
 	}
 	list_splice(&skipped, &lru_list);
 	atomic_add(nr_skipped, &lru_count);
@@ -1727,6 +1765,11 @@ int __init gfs2_glock_init(void)
 	glock_workqueue = create_workqueue("glock_workqueue");
 	if (IS_ERR(glock_workqueue))
 		return PTR_ERR(glock_workqueue);
+	gfs2_delete_workqueue = create_workqueue("delete_workqueue");
+	if (IS_ERR(gfs2_delete_workqueue)) {
+		destroy_workqueue(glock_workqueue);
+		return PTR_ERR(gfs2_delete_workqueue);
+	}
 
 	register_shrinker(&glock_shrinker);
 
@@ -1737,6 +1780,7 @@ void gfs2_glock_exit(void)
 {
 	unregister_shrinker(&glock_shrinker);
 	destroy_workqueue(glock_workqueue);
+	destroy_workqueue(gfs2_delete_workqueue);
 }
 
 static int gfs2_glock_iter_next(struct gfs2_glock_iter *gi)
diff --git a/fs/gfs2/glock.h b/fs/gfs2/glock.h
index a602a28f6f08..c609894ec0d0 100644
--- a/fs/gfs2/glock.h
+++ b/fs/gfs2/glock.h
@@ -143,6 +143,7 @@ struct lm_lockops {
 
 #define GLR_TRYFAILED		13
 
+extern struct workqueue_struct *gfs2_delete_workqueue;
 static inline struct gfs2_holder *gfs2_glock_is_locked_by_me(struct gfs2_glock *gl)
 {
 	struct gfs2_holder *gh;
@@ -191,6 +192,8 @@ static inline int gfs2_glock_is_blocking(struct gfs2_glock *gl)
 int gfs2_glock_get(struct gfs2_sbd *sdp,
 		   u64 number, const struct gfs2_glock_operations *glops,
 		   int create, struct gfs2_glock **glp);
+void gfs2_glock_hold(struct gfs2_glock *gl);
+void gfs2_glock_put_nolock(struct gfs2_glock *gl);
 int gfs2_glock_put(struct gfs2_glock *gl);
 void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, unsigned flags,
 		      struct gfs2_holder *gh);
diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index d5e4ab155ca0..6985eef06c39 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -323,6 +323,7 @@ static void trans_go_sync(struct gfs2_glock *gl)
 
 	if (gl->gl_state != LM_ST_UNLOCKED &&
 	    test_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags)) {
+		flush_workqueue(gfs2_delete_workqueue);
 		gfs2_meta_syncfs(sdp);
 		gfs2_log_shutdown(sdp);
 	}
@@ -372,6 +373,25 @@ static int trans_go_demote_ok(const struct gfs2_glock *gl)
 	return 0;
 }
 
+/**
+ * iopen_go_callback - schedule the dcache entry for the inode to be deleted
+ * @gl: the glock
+ *
+ * gl_spin lock is held while calling this
+ */
+static void iopen_go_callback(struct gfs2_glock *gl)
+{
+	struct gfs2_inode *ip = (struct gfs2_inode *)gl->gl_object;
+
+	if (gl->gl_demote_state == LM_ST_UNLOCKED &&
+	    gl->gl_state == LM_ST_SHARED &&
+	    ip && test_bit(GIF_USER, &ip->i_flags)) {
+		gfs2_glock_hold(gl);
+		if (queue_work(gfs2_delete_workqueue, &gl->gl_delete) == 0)
+			gfs2_glock_put_nolock(gl);
+	}
+}
+
 const struct gfs2_glock_operations gfs2_meta_glops = {
 	.go_type = LM_TYPE_META,
 };
@@ -406,6 +426,7 @@ const struct gfs2_glock_operations gfs2_trans_glops = {
 
 const struct gfs2_glock_operations gfs2_iopen_glops = {
 	.go_type = LM_TYPE_IOPEN,
+	.go_callback = iopen_go_callback,
 };
 
 const struct gfs2_glock_operations gfs2_flock_glops = {
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 225347fbff3c..61801ada36f0 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -159,6 +159,7 @@ struct gfs2_glock_operations {
 	int (*go_lock) (struct gfs2_holder *gh);
 	void (*go_unlock) (struct gfs2_holder *gh);
 	int (*go_dump)(struct seq_file *seq, const struct gfs2_glock *gl);
+	void (*go_callback) (struct gfs2_glock *gl);
 	const int go_type;
 	const unsigned long go_min_hold_time;
 };
@@ -228,6 +229,7 @@ struct gfs2_glock {
 	struct list_head gl_ail_list;
 	atomic_t gl_ail_count;
 	struct delayed_work gl_work;
+	struct work_struct gl_delete;
 };
 
 #define GFS2_MIN_LVB_SIZE 32	/* Min size of LVB that gfs2 supports */
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index daa4ae341a29..fba795798d3a 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -285,27 +285,19 @@ void gfs2_rgrp_verify(struct gfs2_rgrpd *rgd)
 	}
 
 	tmp = rgd->rd_data - rgd->rd_free - rgd->rd_dinodes;
-	if (count[1] + count[2] != tmp) {
+	if (count[1] != tmp) {
 		if (gfs2_consist_rgrpd(rgd))
 			fs_err(sdp, "used data mismatch:  %u != %u\n",
 			       count[1], tmp);
 		return;
 	}
 
-	if (count[3] != rgd->rd_dinodes) {
+	if (count[2] + count[3] != rgd->rd_dinodes) {
 		if (gfs2_consist_rgrpd(rgd))
 			fs_err(sdp, "used metadata mismatch:  %u != %u\n",
-			       count[3], rgd->rd_dinodes);
+			       count[2] + count[3], rgd->rd_dinodes);
 		return;
 	}
-
-	if (count[2] > count[3]) {
-		if (gfs2_consist_rgrpd(rgd))
-			fs_err(sdp, "unlinked inodes > inodes:  %u\n",
-			       count[2]);
-		return;
-	}
-
 }
 
 static inline int rgrp_contains_block(struct gfs2_rgrpd *rgd, u64 block)
@@ -961,7 +953,8 @@ static int try_rgrp_fit(struct gfs2_rgrpd *rgd, struct gfs2_alloc *al)
  * Returns: The inode, if one has been found
  */
 
-static struct inode *try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked)
+static struct inode *try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked,
+				     u64 skip)
 {
 	struct inode *inode;
 	u32 goal = 0, block;
@@ -985,6 +978,8 @@ static struct inode *try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked)
 		goal++;
 		if (*last_unlinked != NO_BLOCK && no_addr <= *last_unlinked)
 			continue;
+		if (no_addr == skip)
+			continue;
 		*last_unlinked = no_addr;
 		inode = gfs2_inode_lookup(rgd->rd_sbd->sd_vfs, DT_UNKNOWN,
 					  no_addr, -1, 1);
@@ -1104,7 +1099,7 @@ static struct inode *get_local_rgrp(struct gfs2_inode *ip, u64 *last_unlinked)
 			if (try_rgrp_fit(rgd, al))
 				goto out;
 			if (rgd->rd_flags & GFS2_RDF_CHECK)
-				inode = try_rgrp_unlink(rgd, last_unlinked);
+				inode = try_rgrp_unlink(rgd, last_unlinked, ip->i_no_addr);
 			if (!rg_locked)
 				gfs2_glock_dq_uninit(&al->al_rgd_gh);
 			if (inode)
@@ -1138,7 +1133,7 @@ static struct inode *get_local_rgrp(struct gfs2_inode *ip, u64 *last_unlinked)
 			if (try_rgrp_fit(rgd, al))
 				goto out;
 			if (rgd->rd_flags & GFS2_RDF_CHECK)
-				inode = try_rgrp_unlink(rgd, last_unlinked);
+				inode = try_rgrp_unlink(rgd, last_unlinked, ip->i_no_addr);
 			if (!rg_locked)
 				gfs2_glock_dq_uninit(&al->al_rgd_gh);
 			if (inode)
diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index 0a6801336470..f522bb017973 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -353,7 +353,7 @@ fail:
 	return error;
 }
 
-static void gfs2_statfs_change_in(struct gfs2_statfs_change_host *sc, const void *buf)
+void gfs2_statfs_change_in(struct gfs2_statfs_change_host *sc, const void *buf)
 {
 	const struct gfs2_statfs_change *str = buf;
 
@@ -441,6 +441,29 @@ void gfs2_statfs_change(struct gfs2_sbd *sdp, s64 total, s64 free,
 	brelse(l_bh);
 }
 
+void update_statfs(struct gfs2_sbd *sdp, struct buffer_head *m_bh,
+		   struct buffer_head *l_bh)
+{
+	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
+	struct gfs2_inode *l_ip = GFS2_I(sdp->sd_sc_inode);
+	struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
+	struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
+
+	gfs2_trans_add_bh(l_ip->i_gl, l_bh, 1);
+
+	spin_lock(&sdp->sd_statfs_spin);
+	m_sc->sc_total += l_sc->sc_total;
+	m_sc->sc_free += l_sc->sc_free;
+	m_sc->sc_dinodes += l_sc->sc_dinodes;
+	memset(l_sc, 0, sizeof(struct gfs2_statfs_change));
+	memset(l_bh->b_data + sizeof(struct gfs2_dinode),
+	       0, sizeof(struct gfs2_statfs_change));
+	spin_unlock(&sdp->sd_statfs_spin);
+
+	gfs2_trans_add_bh(m_ip->i_gl, m_bh, 1);
+	gfs2_statfs_change_out(m_sc, m_bh->b_data + sizeof(struct gfs2_dinode));
+}
+
 int gfs2_statfs_sync(struct gfs2_sbd *sdp)
 {
 	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
@@ -477,19 +500,7 @@ int gfs2_statfs_sync(struct gfs2_sbd *sdp)
 	if (error)
 		goto out_bh2;
 
-	gfs2_trans_add_bh(l_ip->i_gl, l_bh, 1);
-
-	spin_lock(&sdp->sd_statfs_spin);
-	m_sc->sc_total += l_sc->sc_total;
-	m_sc->sc_free += l_sc->sc_free;
-	m_sc->sc_dinodes += l_sc->sc_dinodes;
-	memset(l_sc, 0, sizeof(struct gfs2_statfs_change));
-	memset(l_bh->b_data + sizeof(struct gfs2_dinode),
-	       0, sizeof(struct gfs2_statfs_change));
-	spin_unlock(&sdp->sd_statfs_spin);
-
-	gfs2_trans_add_bh(m_ip->i_gl, m_bh, 1);
-	gfs2_statfs_change_out(m_sc, m_bh->b_data + sizeof(struct gfs2_dinode));
+	update_statfs(sdp, m_bh, l_bh);
 
 	gfs2_trans_end(sdp);
 
@@ -680,6 +691,7 @@ static int gfs2_make_fs_ro(struct gfs2_sbd *sdp)
 	struct gfs2_holder t_gh;
 	int error;
 
+	flush_workqueue(gfs2_delete_workqueue);
 	gfs2_quota_sync(sdp);
 	gfs2_statfs_sync(sdp);
 
diff --git a/fs/gfs2/super.h b/fs/gfs2/super.h
index b56413e3e40d..22e0417ed996 100644
--- a/fs/gfs2/super.h
+++ b/fs/gfs2/super.h
@@ -40,6 +40,10 @@ extern int gfs2_make_fs_rw(struct gfs2_sbd *sdp);
 extern int gfs2_statfs_init(struct gfs2_sbd *sdp);
 extern void gfs2_statfs_change(struct gfs2_sbd *sdp, s64 total, s64 free,
 			       s64 dinodes);
+extern void gfs2_statfs_change_in(struct gfs2_statfs_change_host *sc,
+				  const void *buf);
+extern void update_statfs(struct gfs2_sbd *sdp, struct buffer_head *m_bh,
+			  struct buffer_head *l_bh);
 extern int gfs2_statfs_sync(struct gfs2_sbd *sdp);
 
 extern int gfs2_freeze_fs(struct gfs2_sbd *sdp);
diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c
index 23419dc3027b..a7cbfbd340c7 100644
--- a/fs/gfs2/sys.c
+++ b/fs/gfs2/sys.c
@@ -386,16 +386,16 @@ static ssize_t jid_show(struct gfs2_sbd *sdp, char *buf)
 #define GDLM_ATTR(_name,_mode,_show,_store) \
 static struct gfs2_attr gdlm_attr_##_name = __ATTR(_name,_mode,_show,_store)
 
-GDLM_ATTR(proto_name,     0444, proto_name_show,	NULL);
-GDLM_ATTR(block,          0644, block_show,		block_store);
-GDLM_ATTR(withdraw,       0644, withdraw_show,		withdraw_store);
-GDLM_ATTR(id,             0444, lkid_show,		NULL);
-GDLM_ATTR(jid,		  0444, jid_show,		NULL);
-GDLM_ATTR(first,          0444, lkfirst_show,		NULL);
-GDLM_ATTR(first_done,     0444, first_done_show,	NULL);
-GDLM_ATTR(recover,        0200, NULL,			recover_store);
-GDLM_ATTR(recover_done,   0444, recover_done_show,	NULL);
-GDLM_ATTR(recover_status, 0444, recover_status_show,	NULL);
+GDLM_ATTR(proto_name,		0444, proto_name_show,		NULL);
+GDLM_ATTR(block,		0644, block_show,		block_store);
+GDLM_ATTR(withdraw,		0644, withdraw_show,		withdraw_store);
+GDLM_ATTR(id,			0444, lkid_show,		NULL);
+GDLM_ATTR(jid,			0444, jid_show,			NULL);
+GDLM_ATTR(first,		0444, lkfirst_show,		NULL);
+GDLM_ATTR(first_done,		0444, first_done_show,		NULL);
+GDLM_ATTR(recover,		0600, NULL,			recover_store);
+GDLM_ATTR(recover_done,		0444, recover_done_show,	NULL);
+GDLM_ATTR(recover_status,	0444, recover_status_show,	NULL);
 
 static struct attribute *lock_module_attrs[] = {
 	&gdlm_attr_proto_name.attr,
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index 941c8425c10b..cb88dac8ccaa 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -935,26 +935,28 @@ static int can_do_hugetlb_shm(void)
 	return capable(CAP_IPC_LOCK) || in_group_p(sysctl_hugetlb_shm_group);
 }
 
-struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag)
+struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag,
+						struct user_struct **user)
 {
 	int error = -ENOMEM;
-	int unlock_shm = 0;
 	struct file *file;
 	struct inode *inode;
 	struct dentry *dentry, *root;
 	struct qstr quick_string;
-	struct user_struct *user = current_user();
 
+	*user = NULL;
 	if (!hugetlbfs_vfsmount)
 		return ERR_PTR(-ENOENT);
 
 	if (!can_do_hugetlb_shm()) {
-		if (user_shm_lock(size, user)) {
-			unlock_shm = 1;
+		*user = current_user();
+		if (user_shm_lock(size, *user)) {
 			WARN_ONCE(1,
 			  "Using mlock ulimits for SHM_HUGETLB deprecated\n");
-		} else
+		} else {
+			*user = NULL;
 			return ERR_PTR(-EPERM);
+		}
 	}
 
 	root = hugetlbfs_vfsmount->mnt_root;
@@ -996,8 +998,10 @@ out_inode:
 out_dentry:
 	dput(dentry);
 out_shm_unlock:
-	if (unlock_shm)
-		user_shm_unlock(size, user);
+	if (*user) {
+		user_shm_unlock(size, *user);
+		*user = NULL;
+	}
 	return ERR_PTR(error);
 }
 
diff --git a/fs/inode.c b/fs/inode.c
index 901bad1e5f12..ae7b67e48661 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -120,12 +120,11 @@ static void wake_up_inode(struct inode *inode)
  * These are initializations that need to be done on every inode
  * allocation as the fields are not initialised by slab allocation.
  */
-struct inode *inode_init_always(struct super_block *sb, struct inode *inode)
+int inode_init_always(struct super_block *sb, struct inode *inode)
 {
 	static const struct address_space_operations empty_aops;
 	static struct inode_operations empty_iops;
 	static const struct file_operations empty_fops;
-
 	struct address_space *const mapping = &inode->i_data;
 
 	inode->i_sb = sb;
@@ -152,7 +151,7 @@ struct inode *inode_init_always(struct super_block *sb, struct inode *inode)
 	inode->dirtied_when = 0;
 
 	if (security_inode_alloc(inode))
-		goto out_free_inode;
+		goto out;
 
 	/* allocate and initialize an i_integrity */
 	if (ima_inode_alloc(inode))
@@ -198,16 +197,12 @@ struct inode *inode_init_always(struct super_block *sb, struct inode *inode)
 	inode->i_fsnotify_mask = 0;
 #endif
 
-	return inode;
+	return 0;
 
 out_free_security:
 	security_inode_free(inode);
-out_free_inode:
-	if (inode->i_sb->s_op->destroy_inode)
-		inode->i_sb->s_op->destroy_inode(inode);
-	else
-		kmem_cache_free(inode_cachep, (inode));
-	return NULL;
+out:
+	return -ENOMEM;
 }
 EXPORT_SYMBOL(inode_init_always);
 
@@ -220,12 +215,21 @@ static struct inode *alloc_inode(struct super_block *sb)
 	else
 		inode = kmem_cache_alloc(inode_cachep, GFP_KERNEL);
 
-	if (inode)
-		return inode_init_always(sb, inode);
-	return NULL;
+	if (!inode)
+		return NULL;
+
+	if (unlikely(inode_init_always(sb, inode))) {
+		if (inode->i_sb->s_op->destroy_inode)
+			inode->i_sb->s_op->destroy_inode(inode);
+		else
+			kmem_cache_free(inode_cachep, inode);
+		return NULL;
+	}
+
+	return inode;
 }
 
-void destroy_inode(struct inode *inode)
+void __destroy_inode(struct inode *inode)
 {
 	BUG_ON(inode_has_buffers(inode));
 	ima_inode_free(inode);
@@ -237,13 +241,17 @@ void destroy_inode(struct inode *inode)
 	if (inode->i_default_acl && inode->i_default_acl != ACL_NOT_CACHED)
 		posix_acl_release(inode->i_default_acl);
 #endif
+}
+EXPORT_SYMBOL(__destroy_inode);
+
+void destroy_inode(struct inode *inode)
+{
+	__destroy_inode(inode);
 	if (inode->i_sb->s_op->destroy_inode)
 		inode->i_sb->s_op->destroy_inode(inode);
 	else
 		kmem_cache_free(inode_cachep, (inode));
 }
-EXPORT_SYMBOL(destroy_inode);
-
 
 /*
  * These are initializations that only need to be done
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index 737f7246a4b5..f96f85092d1c 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -287,6 +287,7 @@ int journal_write_metadata_buffer(transaction_t *transaction,
 	struct page *new_page;
 	unsigned int new_offset;
 	struct buffer_head *bh_in = jh2bh(jh_in);
+	journal_t *journal = transaction->t_journal;
 
 	/*
 	 * The buffer really shouldn't be locked: only the current committing
@@ -300,6 +301,11 @@ int journal_write_metadata_buffer(transaction_t *transaction,
 	J_ASSERT_BH(bh_in, buffer_jbddirty(bh_in));
 
 	new_bh = alloc_buffer_head(GFP_NOFS|__GFP_NOFAIL);
+	/* keep subsequent assertions sane */
+	new_bh->b_state = 0;
+	init_buffer(new_bh, NULL, NULL);
+	atomic_set(&new_bh->b_count, 1);
+	new_jh = journal_add_journal_head(new_bh);	/* This sleeps */
 
 	/*
 	 * If a new transaction has already done a buffer copy-out, then
@@ -361,14 +367,6 @@ repeat:
 		kunmap_atomic(mapped_data, KM_USER0);
 	}
 
-	/* keep subsequent assertions sane */
-	new_bh->b_state = 0;
-	init_buffer(new_bh, NULL, NULL);
-	atomic_set(&new_bh->b_count, 1);
-	jbd_unlock_bh_state(bh_in);
-
-	new_jh = journal_add_journal_head(new_bh);	/* This sleeps */
-
 	set_bh_page(new_bh, new_page, new_offset);
 	new_jh->b_transaction = NULL;
 	new_bh->b_size = jh2bh(jh_in)->b_size;
@@ -385,7 +383,11 @@ repeat:
 	 * copying is moved to the transaction's shadow queue.
 	 */
 	JBUFFER_TRACE(jh_in, "file as BJ_Shadow");
-	journal_file_buffer(jh_in, transaction, BJ_Shadow);
+	spin_lock(&journal->j_list_lock);
+	__journal_file_buffer(jh_in, transaction, BJ_Shadow);
+	spin_unlock(&journal->j_list_lock);
+	jbd_unlock_bh_state(bh_in);
+
 	JBUFFER_TRACE(new_jh, "file as BJ_IO");
 	journal_file_buffer(new_jh, transaction, BJ_IO);
 
@@ -848,6 +850,12 @@ static int journal_reset(journal_t *journal)
 
 	first = be32_to_cpu(sb->s_first);
 	last = be32_to_cpu(sb->s_maxlen);
+	if (first + JFS_MIN_JOURNAL_BLOCKS > last + 1) {
+		printk(KERN_ERR "JBD: Journal too short (blocks %lu-%lu).\n",
+		       first, last);
+		journal_fail_superblock(journal);
+		return -EINVAL;
+	}
 
 	journal->j_first = first;
 	journal->j_last = last;
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index 73242ba7c7b1..c03ac11f74be 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -489,34 +489,15 @@ void journal_unlock_updates (journal_t *journal)
 	wake_up(&journal->j_wait_transaction_locked);
 }
 
-/*
- * Report any unexpected dirty buffers which turn up.  Normally those
- * indicate an error, but they can occur if the user is running (say)
- * tune2fs to modify the live filesystem, so we need the option of
- * continuing as gracefully as possible.  #
- *
- * The caller should already hold the journal lock and
- * j_list_lock spinlock: most callers will need those anyway
- * in order to probe the buffer's journaling state safely.
- */
-static void jbd_unexpected_dirty_buffer(struct journal_head *jh)
+static void warn_dirty_buffer(struct buffer_head *bh)
 {
-	int jlist;
-
-	/* If this buffer is one which might reasonably be dirty
-	 * --- ie. data, or not part of this journal --- then
-	 * we're OK to leave it alone, but otherwise we need to
-	 * move the dirty bit to the journal's own internal
-	 * JBDDirty bit. */
-	jlist = jh->b_jlist;
+	char b[BDEVNAME_SIZE];
 
-	if (jlist == BJ_Metadata || jlist == BJ_Reserved ||
-	    jlist == BJ_Shadow || jlist == BJ_Forget) {
-		struct buffer_head *bh = jh2bh(jh);
-
-		if (test_clear_buffer_dirty(bh))
-			set_buffer_jbddirty(bh);
-	}
+	printk(KERN_WARNING
+	       "JBD: Spotted dirty metadata buffer (dev = %s, blocknr = %llu). "
+	       "There's a risk of filesystem corruption in case of system "
+	       "crash.\n",
+	       bdevname(bh->b_bdev, b), (unsigned long long)bh->b_blocknr);
 }
 
 /*
@@ -583,14 +564,16 @@ repeat:
 			if (jh->b_next_transaction)
 				J_ASSERT_JH(jh, jh->b_next_transaction ==
 							transaction);
+			warn_dirty_buffer(bh);
 		}
 		/*
 		 * In any case we need to clean the dirty flag and we must
 		 * do it under the buffer lock to be sure we don't race
 		 * with running write-out.
 		 */
-		JBUFFER_TRACE(jh, "Unexpected dirty buffer");
-		jbd_unexpected_dirty_buffer(jh);
+		JBUFFER_TRACE(jh, "Journalling dirty buffer");
+		clear_buffer_dirty(bh);
+		set_buffer_jbddirty(bh);
 	}
 
 	unlock_buffer(bh);
@@ -826,6 +809,15 @@ int journal_get_create_access(handle_t *handle, struct buffer_head *bh)
 	J_ASSERT_JH(jh, buffer_locked(jh2bh(jh)));
 
 	if (jh->b_transaction == NULL) {
+		/*
+		 * Previous journal_forget() could have left the buffer
+		 * with jbddirty bit set because it was being committed. When
+		 * the commit finished, we've filed the buffer for
+		 * checkpointing and marked it dirty. Now we are reallocating
+		 * the buffer so the transaction freeing it must have
+		 * committed and so it's safe to clear the dirty bit.
+		 */
+		clear_buffer_dirty(jh2bh(jh));
 		jh->b_transaction = transaction;
 
 		/* first access by this transaction */
@@ -1782,8 +1774,13 @@ static int __dispose_buffer(struct journal_head *jh, transaction_t *transaction)
 
 	if (jh->b_cp_transaction) {
 		JBUFFER_TRACE(jh, "on running+cp transaction");
+		/*
+		 * We don't want to write the buffer anymore, clear the
+		 * bit so that we don't confuse checks in
+		 * __journal_file_buffer
+		 */
+		clear_buffer_dirty(bh);
 		__journal_file_buffer(jh, transaction, BJ_Forget);
-		clear_buffer_jbddirty(bh);
 		may_free = 0;
 	} else {
 		JBUFFER_TRACE(jh, "on running transaction");
@@ -2041,12 +2038,17 @@ void __journal_file_buffer(struct journal_head *jh,
 	if (jh->b_transaction && jh->b_jlist == jlist)
 		return;
 
-	/* The following list of buffer states needs to be consistent
-	 * with __jbd_unexpected_dirty_buffer()'s handling of dirty
-	 * state. */
-
 	if (jlist == BJ_Metadata || jlist == BJ_Reserved ||
 	    jlist == BJ_Shadow || jlist == BJ_Forget) {
+		/*
+		 * For metadata buffers, we track dirty bit in buffer_jbddirty
+		 * instead of buffer_dirty. We should not see a dirty bit set
+		 * here because we clear it in do_get_write_access but e.g.
+		 * tune2fs can modify the sb and set the dirty bit at any time
+		 * so we try to gracefully handle that.
+		 */
+		if (buffer_dirty(bh))
+			warn_dirty_buffer(bh);
 		if (test_clear_buffer_dirty(bh) ||
 		    test_clear_buffer_jbddirty(bh))
 			was_dirty = 1;
diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index 5edc2bf20581..23c947539864 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -99,7 +99,7 @@ static int jffs2_do_readpage_nolock (struct inode *inode, struct page *pg)
 	kunmap(pg);
 
 	D2(printk(KERN_DEBUG "readpage finished\n"));
-	return 0;
+	return ret;
 }
 
 int jffs2_do_readpage_unlock(struct inode *inode, struct page *pg)
diff --git a/fs/jfs/acl.c b/fs/jfs/acl.c
index 91fa3ad6e8c2..a29c7c3e3fb8 100644
--- a/fs/jfs/acl.c
+++ b/fs/jfs/acl.c
@@ -67,10 +67,8 @@ static struct posix_acl *jfs_get_acl(struct inode *inode, int type)
 		acl = posix_acl_from_xattr(value, size);
 	}
 	kfree(value);
-	if (!IS_ERR(acl)) {
+	if (!IS_ERR(acl))
 		set_cached_acl(inode, type, acl);
-		posix_acl_release(acl);
-	}
 	return acl;
 }
 
diff --git a/fs/libfs.c b/fs/libfs.c
index ddfa89948c3f..dcec3d3ea64f 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -217,7 +217,7 @@ int get_sb_pseudo(struct file_system_type *fs_type, char *name,
 		return PTR_ERR(s);
 
 	s->s_flags = MS_NOUSER;
-	s->s_maxbytes = ~0ULL;
+	s->s_maxbytes = MAX_LFS_FILESIZE;
 	s->s_blocksize = PAGE_SIZE;
 	s->s_blocksize_bits = PAGE_SHIFT;
 	s->s_magic = magic;
diff --git a/fs/namespace.c b/fs/namespace.c
index 277c28a63ead..7230787d18b0 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -316,7 +316,8 @@ EXPORT_SYMBOL_GPL(mnt_clone_write);
  */
 int mnt_want_write_file(struct file *file)
 {
-	if (!(file->f_mode & FMODE_WRITE))
+	struct inode *inode = file->f_dentry->d_inode;
+	if (!(file->f_mode & FMODE_WRITE) || special_file(inode->i_mode))
 		return mnt_want_write(file->f_path.mnt);
 	else
 		return mnt_clone_write(file->f_path.mnt);
diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index c2d061675d80..8d25ccb2d51d 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -1242,20 +1242,6 @@ error:
 	return error;
 }
 
-/*
- * Initialize a session.
- * Note: save the mount rsize and wsize for create_server negotiation.
- */
-static void nfs4_init_session(struct nfs_client *clp,
-			      unsigned int wsize, unsigned int rsize)
-{
-#if defined(CONFIG_NFS_V4_1)
-	if (nfs4_has_session(clp)) {
-		clp->cl_session->fc_attrs.max_rqst_sz = wsize;
-		clp->cl_session->fc_attrs.max_resp_sz = rsize;
-	}
-#endif /* CONFIG_NFS_V4_1 */
-}
 
 /*
  * Session has been established, and the client marked ready.
@@ -1350,7 +1336,9 @@ struct nfs_server *nfs4_create_server(const struct nfs_parsed_mount_data *data,
 	BUG_ON(!server->nfs_client->rpc_ops);
 	BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops);
 
-	nfs4_init_session(server->nfs_client, server->wsize, server->rsize);
+	error = nfs4_init_session(server);
+	if (error < 0)
+		goto error;
 
 	/* Probe the root fh to retrieve its FSID */
 	error = nfs4_path_walk(server, mntfh, data->nfs_server.export_path);
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 38d42c29fb92..32062c33c859 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1025,12 +1025,12 @@ static struct dentry *nfs_atomic_lookup(struct inode *dir, struct dentry *dentry
 				res = NULL;
 				goto out;
 			/* This turned out not to be a regular file */
-			case -EISDIR:
 			case -ENOTDIR:
 				goto no_open;
 			case -ELOOP:
 				if (!(nd->intent.open.flags & O_NOFOLLOW))
 					goto no_open;
+			/* case -EISDIR: */
 			/* case -EINVAL: */
 			default:
 				goto out;
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index 489fc01a3204..e4e089a8f294 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -255,7 +255,7 @@ static void nfs_direct_read_release(void *calldata)
 
 	if (put_dreq(dreq))
 		nfs_direct_complete(dreq);
-	nfs_readdata_release(calldata);
+	nfs_readdata_free(data);
 }
 
 static const struct rpc_call_ops nfs_read_direct_ops = {
@@ -314,14 +314,14 @@ static ssize_t nfs_direct_read_schedule_segment(struct nfs_direct_req *dreq,
 					data->npages, 1, 0, data->pagevec, NULL);
 		up_read(&current->mm->mmap_sem);
 		if (result < 0) {
-			nfs_readdata_release(data);
+			nfs_readdata_free(data);
 			break;
 		}
 		if ((unsigned)result < data->npages) {
 			bytes = result * PAGE_SIZE;
 			if (bytes <= pgbase) {
 				nfs_direct_release_pages(data->pagevec, result);
-				nfs_readdata_release(data);
+				nfs_readdata_free(data);
 				break;
 			}
 			bytes -= pgbase;
@@ -334,7 +334,7 @@ static ssize_t nfs_direct_read_schedule_segment(struct nfs_direct_req *dreq,
 		data->inode = inode;
 		data->cred = msg.rpc_cred;
 		data->args.fh = NFS_FH(inode);
-		data->args.context = get_nfs_open_context(ctx);
+		data->args.context = ctx;
 		data->args.offset = pos;
 		data->args.pgbase = pgbase;
 		data->args.pages = data->pagevec;
@@ -441,7 +441,7 @@ static void nfs_direct_free_writedata(struct nfs_direct_req *dreq)
 		struct nfs_write_data *data = list_entry(dreq->rewrite_list.next, struct nfs_write_data, pages);
 		list_del(&data->pages);
 		nfs_direct_release_pages(data->pagevec, data->npages);
-		nfs_writedata_release(data);
+		nfs_writedata_free(data);
 	}
 }
 
@@ -534,7 +534,7 @@ static void nfs_direct_commit_release(void *calldata)
 
 	dprintk("NFS: %5u commit returned %d\n", data->task.tk_pid, status);
 	nfs_direct_write_complete(dreq, data->inode);
-	nfs_commitdata_release(calldata);
+	nfs_commit_free(data);
 }
 
 static const struct rpc_call_ops nfs_commit_direct_ops = {
@@ -570,7 +570,7 @@ static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq)
 	data->args.fh = NFS_FH(data->inode);
 	data->args.offset = 0;
 	data->args.count = 0;
-	data->args.context = get_nfs_open_context(dreq->ctx);
+	data->args.context = dreq->ctx;
 	data->res.count = 0;
 	data->res.fattr = &data->fattr;
 	data->res.verf = &data->verf;
@@ -734,14 +734,14 @@ static ssize_t nfs_direct_write_schedule_segment(struct nfs_direct_req *dreq,
 					data->npages, 0, 0, data->pagevec, NULL);
 		up_read(&current->mm->mmap_sem);
 		if (result < 0) {
-			nfs_writedata_release(data);
+			nfs_writedata_free(data);
 			break;
 		}
 		if ((unsigned)result < data->npages) {
 			bytes = result * PAGE_SIZE;
 			if (bytes <= pgbase) {
 				nfs_direct_release_pages(data->pagevec, result);
-				nfs_writedata_release(data);
+				nfs_writedata_free(data);
 				break;
 			}
 			bytes -= pgbase;
@@ -756,7 +756,7 @@ static ssize_t nfs_direct_write_schedule_segment(struct nfs_direct_req *dreq,
 		data->inode = inode;
 		data->cred = msg.rpc_cred;
 		data->args.fh = NFS_FH(inode);
-		data->args.context = get_nfs_open_context(ctx);
+		data->args.context = ctx;
 		data->args.offset = pos;
 		data->args.pgbase = pgbase;
 		data->args.pages = data->pagevec;
diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index 61bc3a32e1e2..6ea07a3c75d4 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -220,6 +220,7 @@ extern void nfs4_destroy_session(struct nfs4_session *session);
 extern struct nfs4_session *nfs4_alloc_session(struct nfs_client *clp);
 extern int nfs4_proc_create_session(struct nfs_client *, int reset);
 extern int nfs4_proc_destroy_session(struct nfs4_session *);
+extern int nfs4_init_session(struct nfs_server *server);
 #else /* CONFIG_NFS_v4_1 */
 static inline int nfs4_setup_sequence(struct nfs_client *clp,
 		struct nfs4_sequence_args *args, struct nfs4_sequence_res *res,
@@ -227,6 +228,11 @@ static inline int nfs4_setup_sequence(struct nfs_client *clp,
 {
 	return 0;
 }
+
+static inline int nfs4_init_session(struct nfs_server *server)
+{
+	return 0;
+}
 #endif /* CONFIG_NFS_V4_1 */
 
 extern struct nfs4_state_maintenance_ops *nfs4_state_renewal_ops[];
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index ff0c080db59b..6917311f201c 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -2040,15 +2040,9 @@ static int _nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
 		.rpc_argp = &args,
 		.rpc_resp = &res,
 	};
-	int status;
 
 	nfs_fattr_init(info->fattr);
-	status = nfs4_recover_expired_lease(server);
-	if (!status)
-		status = nfs4_check_client_ready(server->nfs_client);
-	if (!status)
-		status = nfs4_call_sync(server, &msg, &args, &res, 0);
-	return status;
+	return nfs4_call_sync(server, &msg, &args, &res, 0);
 }
 
 static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
@@ -4099,15 +4093,23 @@ nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
 	if (request->fl_start < 0 || request->fl_end < 0)
 		return -EINVAL;
 
-	if (IS_GETLK(cmd))
-		return nfs4_proc_getlk(state, F_GETLK, request);
+	if (IS_GETLK(cmd)) {
+		if (state != NULL)
+			return nfs4_proc_getlk(state, F_GETLK, request);
+		return 0;
+	}
 
 	if (!(IS_SETLK(cmd) || IS_SETLKW(cmd)))
 		return -EINVAL;
 
-	if (request->fl_type == F_UNLCK)
-		return nfs4_proc_unlck(state, cmd, request);
+	if (request->fl_type == F_UNLCK) {
+		if (state != NULL)
+			return nfs4_proc_unlck(state, cmd, request);
+		return 0;
+	}
 
+	if (state == NULL)
+		return -ENOLCK;
 	do {
 		status = nfs4_proc_setlk(state, cmd, request);
 		if ((status != -EAGAIN) || IS_SETLK(cmd))
@@ -4793,6 +4795,22 @@ int nfs4_proc_destroy_session(struct nfs4_session *session)
 	return status;
 }
 
+int nfs4_init_session(struct nfs_server *server)
+{
+	struct nfs_client *clp = server->nfs_client;
+	int ret;
+
+	if (!nfs4_has_session(clp))
+		return 0;
+
+	clp->cl_session->fc_attrs.max_rqst_sz = server->wsize;
+	clp->cl_session->fc_attrs.max_resp_sz = server->rsize;
+	ret = nfs4_recover_expired_lease(server);
+	if (!ret)
+		ret = nfs4_check_client_ready(clp);
+	return ret;
+}
+
 /*
  * Renew the cl_session lease.
  */
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index b73c5a728655..1434080aefeb 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -553,6 +553,7 @@ static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, f
 	INIT_LIST_HEAD(&lsp->ls_sequence.list);
 	lsp->ls_seqid.sequence = &lsp->ls_sequence;
 	atomic_set(&lsp->ls_count, 1);
+	lsp->ls_state = state;
 	lsp->ls_owner = fl_owner;
 	spin_lock(&clp->cl_lock);
 	nfs_alloc_unique_id(&clp->cl_lockowner_id, &lsp->ls_id, 1, 64);
@@ -587,7 +588,6 @@ static struct nfs4_lock_state *nfs4_get_lock_state(struct nfs4_state *state, fl_
 		if (lsp != NULL)
 			break;
 		if (new != NULL) {
-			new->ls_state = state;
 			list_add(&new->ls_locks, &state->lock_states);
 			set_bit(LK_STATE_IN_USE, &state->flags);
 			lsp = new;
@@ -1250,8 +1250,8 @@ static void nfs4_state_manager(struct nfs_client *clp)
 				continue;
 		}
 		/* Initialize or reset the session */
-		if (nfs4_has_session(clp) &&
-		   test_and_clear_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state)) {
+		if (test_and_clear_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state)
+		   && nfs4_has_session(clp)) {
 			if (clp->cl_cons_state == NFS_CS_SESSION_INITING)
 				status = nfs4_initialize_session(clp);
 			else
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 73ea5e8d66ce..12c9e66d3f1d 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -60,17 +60,15 @@ struct nfs_read_data *nfs_readdata_alloc(unsigned int pagecount)
 	return p;
 }
 
-static void nfs_readdata_free(struct nfs_read_data *p)
+void nfs_readdata_free(struct nfs_read_data *p)
 {
 	if (p && (p->pagevec != &p->page_array[0]))
 		kfree(p->pagevec);
 	mempool_free(p, nfs_rdata_mempool);
 }
 
-void nfs_readdata_release(void *data)
+static void nfs_readdata_release(struct nfs_read_data *rdata)
 {
-	struct nfs_read_data *rdata = data;
-
 	put_nfs_open_context(rdata->args.context);
 	nfs_readdata_free(rdata);
 }
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 0a0a2ff767c3..a34fae21fe10 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -87,17 +87,15 @@ struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount)
 	return p;
 }
 
-static void nfs_writedata_free(struct nfs_write_data *p)
+void nfs_writedata_free(struct nfs_write_data *p)
 {
 	if (p && (p->pagevec != &p->page_array[0]))
 		kfree(p->pagevec);
 	mempool_free(p, nfs_wdata_mempool);
 }
 
-void nfs_writedata_release(void *data)
+static void nfs_writedata_release(struct nfs_write_data *wdata)
 {
-	struct nfs_write_data *wdata = data;
-
 	put_nfs_open_context(wdata->args.context);
 	nfs_writedata_free(wdata);
 }
diff --git a/fs/nilfs2/Kconfig b/fs/nilfs2/Kconfig
new file mode 100644
index 000000000000..72da095d4009
--- /dev/null
+++ b/fs/nilfs2/Kconfig
@@ -0,0 +1,25 @@
+config NILFS2_FS
+	tristate "NILFS2 file system support (EXPERIMENTAL)"
+	depends on BLOCK && EXPERIMENTAL
+	select CRC32
+	help
+	  NILFS2 is a log-structured file system (LFS) supporting continuous
+	  snapshotting.  In addition to versioning capability of the entire
+	  file system, users can even restore files mistakenly overwritten or
+	  destroyed just a few seconds ago.  Since this file system can keep
+	  consistency like conventional LFS, it achieves quick recovery after
+	  system crashes.
+
+	  NILFS2 creates a number of checkpoints every few seconds or per
+	  synchronous write basis (unless there is no change).  Users can
+	  select significant versions among continuously created checkpoints,
+	  and can change them into snapshots which will be preserved for long
+	  periods until they are changed back to checkpoints.  Each
+	  snapshot is mountable as a read-only file system concurrently with
+	  its writable mount, and this feature is convenient for online backup.
+
+	  Some features including atime, extended attributes, and POSIX ACLs,
+	  are not supported yet.
+
+	  To compile this file system support as a module, choose M here: the
+	  module will be called nilfs2.  If unsure, say N.
diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index 3d3ddb3f5177..2dfd47714ae5 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -412,8 +412,10 @@ nilfs_mdt_write_page(struct page *page, struct writeback_control *wbc)
 		return 0; /* Do not request flush for shadow page cache */
 	if (!sb) {
 		writer = nilfs_get_writer(NILFS_MDT(inode)->mi_nilfs);
-		if (!writer)
+		if (!writer) {
+			nilfs_put_writer(NILFS_MDT(inode)->mi_nilfs);
 			return -EROFS;
+		}
 		sb = writer->s_super;
 	}
 
diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 8b5e4778cf28..51ff3d0a4ee2 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -1859,12 +1859,26 @@ static void nilfs_end_page_io(struct page *page, int err)
 	if (!page)
 		return;
 
-	if (buffer_nilfs_node(page_buffers(page)) && !PageWriteback(page))
+	if (buffer_nilfs_node(page_buffers(page)) && !PageWriteback(page)) {
 		/*
 		 * For b-tree node pages, this function may be called twice
 		 * or more because they might be split in a segment.
 		 */
+		if (PageDirty(page)) {
+			/*
+			 * For pages holding split b-tree node buffers, dirty
+			 * flag on the buffers may be cleared discretely.
+			 * In that case, the page is once redirtied for
+			 * remaining buffers, and it must be cancelled if
+			 * all the buffers get cleaned later.
+			 */
+			lock_page(page);
+			if (nilfs_page_buffers_clean(page))
+				__nilfs_clear_page_dirty(page);
+			unlock_page(page);
+		}
 		return;
+	}
 
 	__nilfs_end_page_io(page, err);
 }
diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c
index 8e2ec43b18f4..151964f0de4c 100644
--- a/fs/nilfs2/super.c
+++ b/fs/nilfs2/super.c
@@ -416,8 +416,10 @@ int nilfs_attach_checkpoint(struct nilfs_sb_info *sbi, __u64 cno)
 	if (unlikely(err))
 		goto failed;
 
+	down_read(&nilfs->ns_segctor_sem);
 	err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, cno, 0, &raw_cp,
 					  &bh_cp);
+	up_read(&nilfs->ns_segctor_sem);
 	if (unlikely(err)) {
 		if (err == -ENOENT || err == -EINVAL) {
 			printk(KERN_ERR
diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h
index e8adbffc626f..1b9caafb8662 100644
--- a/fs/nilfs2/the_nilfs.h
+++ b/fs/nilfs2/the_nilfs.h
@@ -253,7 +253,7 @@ nilfs_detach_writer(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi)
 
 static inline void nilfs_put_sbinfo(struct nilfs_sb_info *sbi)
 {
-	if (!atomic_dec_and_test(&sbi->s_count))
+	if (atomic_dec_and_test(&sbi->s_count))
 		kfree(sbi);
 }
 
diff --git a/fs/notify/Kconfig b/fs/notify/Kconfig
index 31dac7e3b0f1..dffbb0911d02 100644
--- a/fs/notify/Kconfig
+++ b/fs/notify/Kconfig
@@ -1,15 +1,5 @@
 config FSNOTIFY
-	bool "Filesystem notification backend"
-	default y
-	---help---
-	   fsnotify is a backend for filesystem notification.  fsnotify does
-	   not provide any userspace interface but does provide the basis
-	   needed for other notification schemes such as dnotify, inotify,
-	   and fanotify.
-
-	   Say Y here to enable fsnotify suport.
-
-	   If unsure, say Y.
+	def_bool n
 
 source "fs/notify/dnotify/Kconfig"
 source "fs/notify/inotify/Kconfig"
diff --git a/fs/notify/dnotify/Kconfig b/fs/notify/dnotify/Kconfig
index 904ff8d5405a..f9c1ca139d8f 100644
--- a/fs/notify/dnotify/Kconfig
+++ b/fs/notify/dnotify/Kconfig
@@ -1,6 +1,6 @@
 config DNOTIFY
 	bool "Dnotify support"
-	depends on FSNOTIFY
+	select FSNOTIFY
 	default y
 	help
 	  Dnotify is a directory-based per-fd file change notification system
diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c
index ec2f7bd76818..037e878e03fc 100644
--- a/fs/notify/fsnotify.c
+++ b/fs/notify/fsnotify.c
@@ -159,7 +159,9 @@ void fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is, const
 			if (!group->ops->should_send_event(group, to_tell, mask))
 				continue;
 			if (!event) {
-				event = fsnotify_create_event(to_tell, mask, data, data_is, file_name, cookie);
+				event = fsnotify_create_event(to_tell, mask, data,
+							      data_is, file_name, cookie,
+							      GFP_KERNEL);
 				/* shit, we OOM'd and now we can't tell, maybe
 				 * someday someone else will want to do something
 				 * here */
diff --git a/fs/notify/inotify/Kconfig b/fs/notify/inotify/Kconfig
index 5356884289a1..3e56dbffe729 100644
--- a/fs/notify/inotify/Kconfig
+++ b/fs/notify/inotify/Kconfig
@@ -15,7 +15,7 @@ config INOTIFY
 
 config INOTIFY_USER
 	bool "Inotify support for userspace"
-	depends on FSNOTIFY
+	select FSNOTIFY
 	default y
 	---help---
 	  Say Y here to enable inotify support for userspace, including the
diff --git a/fs/notify/inotify/inotify_fsnotify.c b/fs/notify/inotify/inotify_fsnotify.c
index 47cd258fd24d..c9ee67b442e1 100644
--- a/fs/notify/inotify/inotify_fsnotify.c
+++ b/fs/notify/inotify/inotify_fsnotify.c
@@ -62,13 +62,14 @@ static int inotify_handle_event(struct fsnotify_group *group, struct fsnotify_ev
 	event_priv->wd = wd;
 
 	ret = fsnotify_add_notify_event(group, event, fsn_event_priv);
-	/* EEXIST is not an error */
-	if (ret == -EEXIST)
-		ret = 0;
-
-	/* did event_priv get attached? */
-	if (list_empty(&fsn_event_priv->event_list))
+	if (ret) {
 		inotify_free_event_priv(fsn_event_priv);
+		/* EEXIST says we tail matched, EOVERFLOW isn't something
+		 * to report up the stack. */
+		if ((ret == -EEXIST) ||
+		    (ret == -EOVERFLOW))
+			ret = 0;
+	}
 
 	/*
 	 * If we hold the entry until after the event is on the queue
@@ -104,16 +105,45 @@ static bool inotify_should_send_event(struct fsnotify_group *group, struct inode
 	return send;
 }
 
+/*
+ * This is NEVER supposed to be called.  Inotify marks should either have been
+ * removed from the idr when the watch was removed or in the
+ * fsnotify_destroy_mark_by_group() call when the inotify instance was being
+ * torn down.  This is only called if the idr is about to be freed but there
+ * are still marks in it.
+ */
 static int idr_callback(int id, void *p, void *data)
 {
-	BUG();
+	struct fsnotify_mark_entry *entry;
+	struct inotify_inode_mark_entry *ientry;
+	static bool warned = false;
+
+	if (warned)
+		return 0;
+
+	warned = false;
+	entry = p;
+	ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry);
+
+	WARN(1, "inotify closing but id=%d for entry=%p in group=%p still in "
+		"idr.  Probably leaking memory\n", id, p, data);
+
+	/*
+	 * I'm taking the liberty of assuming that the mark in question is a
+	 * valid address and I'm dereferencing it.  This might help to figure
+	 * out why we got here and the panic is no worse than the original
+	 * BUG() that was here.
+	 */
+	if (entry)
+		printk(KERN_WARNING "entry->group=%p inode=%p wd=%d\n",
+			entry->group, entry->inode, ientry->wd);
 	return 0;
 }
 
 static void inotify_free_group_priv(struct fsnotify_group *group)
 {
 	/* ideally the idr is empty and we won't hit the BUG in teh callback */
-	idr_for_each(&group->inotify_data.idr, idr_callback, NULL);
+	idr_for_each(&group->inotify_data.idr, idr_callback, group);
 	idr_remove_all(&group->inotify_data.idr);
 	idr_destroy(&group->inotify_data.idr);
 }
diff --git a/fs/notify/inotify/inotify_user.c b/fs/notify/inotify/inotify_user.c
index ff27a2965844..0e781bc88d1e 100644
--- a/fs/notify/inotify/inotify_user.c
+++ b/fs/notify/inotify/inotify_user.c
@@ -47,9 +47,6 @@
 
 static struct vfsmount *inotify_mnt __read_mostly;
 
-/* this just sits here and wastes global memory.  used to just pad userspace messages with zeros */
-static struct inotify_event nul_inotify_event;
-
 /* these are configurable via /proc/sys/fs/inotify/ */
 static int inotify_max_user_instances __read_mostly;
 static int inotify_max_queued_events __read_mostly;
@@ -57,7 +54,6 @@ int inotify_max_user_watches __read_mostly;
 
 static struct kmem_cache *inotify_inode_mark_cachep __read_mostly;
 struct kmem_cache *event_priv_cachep __read_mostly;
-static struct fsnotify_event *inotify_ignored_event;
 
 /*
  * When inotify registers a new group it increments this and uses that
@@ -200,8 +196,10 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
 		inotify_free_event_priv(fsn_priv);
 	}
 
-	/* round up event->name_len so it is a multiple of event_size */
-	name_len = roundup(event->name_len, event_size);
+	/* round up event->name_len so it is a multiple of event_size
+	 * plus an extra byte for the terminating '\0'.
+	 */
+	name_len = roundup(event->name_len + 1, event_size);
 	inotify_event.len = name_len;
 
 	inotify_event.mask = inotify_mask_to_arg(event->mask);
@@ -225,8 +223,8 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
 			return -EFAULT;
 		buf += event->name_len;
 
-		/* fill userspace with 0's from nul_inotify_event */
-		if (copy_to_user(buf, &nul_inotify_event, len_to_zero))
+		/* fill userspace with 0's */
+		if (clear_user(buf, len_to_zero))
 			return -EFAULT;
 		buf += len_to_zero;
 		event_size += name_len;
@@ -366,20 +364,71 @@ static int inotify_find_inode(const char __user *dirname, struct path *path, uns
 }
 
 /*
- * Send IN_IGNORED for this wd, remove this wd from the idr, and drop the
- * internal reference help on the mark because it is in the idr.
+ * Remove the mark from the idr (if present) and drop the reference
+ * on the mark because it was in the idr.
+ */
+static void inotify_remove_from_idr(struct fsnotify_group *group,
+				    struct inotify_inode_mark_entry *ientry)
+{
+	struct idr *idr;
+	struct fsnotify_mark_entry *entry;
+	struct inotify_inode_mark_entry *found_ientry;
+	int wd;
+
+	spin_lock(&group->inotify_data.idr_lock);
+	idr = &group->inotify_data.idr;
+	wd = ientry->wd;
+
+	if (wd == -1)
+		goto out;
+
+	entry = idr_find(&group->inotify_data.idr, wd);
+	if (unlikely(!entry))
+		goto out;
+
+	found_ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry);
+	if (unlikely(found_ientry != ientry)) {
+		/* We found an entry in the idr with the right wd, but it's
+		 * not the entry we were told to remove.  eparis seriously
+		 * fucked up somewhere. */
+		WARN_ON(1);
+		ientry->wd = -1;
+		goto out;
+	}
+
+	/* One ref for being in the idr, one ref held by the caller */
+	BUG_ON(atomic_read(&entry->refcnt) < 2);
+
+	idr_remove(idr, wd);
+	ientry->wd = -1;
+
+	/* removed from the idr, drop that ref */
+	fsnotify_put_mark(entry);
+out:
+	spin_unlock(&group->inotify_data.idr_lock);
+}
+
+/*
+ * Send IN_IGNORED for this wd, remove this wd from the idr.
  */
 void inotify_ignored_and_remove_idr(struct fsnotify_mark_entry *entry,
 				    struct fsnotify_group *group)
 {
 	struct inotify_inode_mark_entry *ientry;
+	struct fsnotify_event *ignored_event;
 	struct inotify_event_private_data *event_priv;
 	struct fsnotify_event_private_data *fsn_event_priv;
-	struct idr *idr;
+	int ret;
+
+	ignored_event = fsnotify_create_event(NULL, FS_IN_IGNORED, NULL,
+					      FSNOTIFY_EVENT_NONE, NULL, 0,
+					      GFP_NOFS);
+	if (!ignored_event)
+		return;
 
 	ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry);
 
-	event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL);
+	event_priv = kmem_cache_alloc(event_priv_cachep, GFP_NOFS);
 	if (unlikely(!event_priv))
 		goto skip_send_ignore;
 
@@ -388,22 +437,19 @@ void inotify_ignored_and_remove_idr(struct fsnotify_mark_entry *entry,
 	fsn_event_priv->group = group;
 	event_priv->wd = ientry->wd;
 
-	fsnotify_add_notify_event(group, inotify_ignored_event, fsn_event_priv);
-
-	/* did the private data get added? */
-	if (list_empty(&fsn_event_priv->event_list))
+	ret = fsnotify_add_notify_event(group, ignored_event, fsn_event_priv);
+	if (ret)
 		inotify_free_event_priv(fsn_event_priv);
 
 skip_send_ignore:
 
+	/* matches the reference taken when the event was created */
+	fsnotify_put_event(ignored_event);
+
 	/* remove this entry from the idr */
-	spin_lock(&group->inotify_data.idr_lock);
-	idr = &group->inotify_data.idr;
-	idr_remove(idr, ientry->wd);
-	spin_unlock(&group->inotify_data.idr_lock);
+	inotify_remove_from_idr(group, ientry);
 
-	/* removed from idr, drop that reference */
-	fsnotify_put_mark(entry);
+	atomic_dec(&group->inotify_data.user->inotify_watches);
 }
 
 /* ding dong the mark is dead */
@@ -414,67 +460,29 @@ static void inotify_free_mark(struct fsnotify_mark_entry *entry)
 	kmem_cache_free(inotify_inode_mark_cachep, ientry);
 }
 
-static int inotify_update_watch(struct fsnotify_group *group, struct inode *inode, u32 arg)
+static int inotify_update_existing_watch(struct fsnotify_group *group,
+					 struct inode *inode,
+					 u32 arg)
 {
-	struct fsnotify_mark_entry *entry = NULL;
+	struct fsnotify_mark_entry *entry;
 	struct inotify_inode_mark_entry *ientry;
-	int ret = 0;
-	int add = (arg & IN_MASK_ADD);
-	__u32 mask;
 	__u32 old_mask, new_mask;
+	__u32 mask;
+	int add = (arg & IN_MASK_ADD);
+	int ret;
 
 	/* don't allow invalid bits: we don't want flags set */
 	mask = inotify_arg_to_mask(arg);
 	if (unlikely(!mask))
 		return -EINVAL;
 
-	ientry = kmem_cache_alloc(inotify_inode_mark_cachep, GFP_KERNEL);
-	if (unlikely(!ientry))
-		return -ENOMEM;
-	/* we set the mask at the end after attaching it */
-	fsnotify_init_mark(&ientry->fsn_entry, inotify_free_mark);
-	ientry->wd = 0;
-
-find_entry:
 	spin_lock(&inode->i_lock);
 	entry = fsnotify_find_mark_entry(group, inode);
 	spin_unlock(&inode->i_lock);
-	if (entry) {
-		kmem_cache_free(inotify_inode_mark_cachep, ientry);
-		ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry);
-	} else {
-		if (atomic_read(&group->inotify_data.user->inotify_watches) >= inotify_max_user_watches) {
-			ret = -ENOSPC;
-			goto out_err;
-		}
-
-		ret = fsnotify_add_mark(&ientry->fsn_entry, group, inode);
-		if (ret == -EEXIST)
-			goto find_entry;
-		else if (ret)
-			goto out_err;
+	if (!entry)
+		return -ENOENT;
 
-		entry = &ientry->fsn_entry;
-retry:
-		ret = -ENOMEM;
-		if (unlikely(!idr_pre_get(&group->inotify_data.idr, GFP_KERNEL)))
-			goto out_err;
-
-		spin_lock(&group->inotify_data.idr_lock);
-		/* if entry is added to the idr we keep the reference obtained
-		 * through fsnotify_mark_add.  remember to drop this reference
-		 * when entry is removed from idr */
-		ret = idr_get_new_above(&group->inotify_data.idr, entry,
-					++group->inotify_data.last_wd,
-					&ientry->wd);
-		spin_unlock(&group->inotify_data.idr_lock);
-		if (ret) {
-			if (ret == -EAGAIN)
-				goto retry;
-			goto out_err;
-		}
-		atomic_inc(&group->inotify_data.user->inotify_watches);
-	}
+	ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry);
 
 	spin_lock(&entry->lock);
 
@@ -506,14 +514,104 @@ retry:
 			fsnotify_recalc_group_mask(group);
 	}
 
-	return ientry->wd;
+	/* return the wd */
+	ret = ientry->wd;
 
-out_err:
-	/* see this isn't supposed to happen, just kill the watch */
-	if (entry) {
-		fsnotify_destroy_mark_by_entry(entry);
-		fsnotify_put_mark(entry);
+	/* match the get from fsnotify_find_mark_entry() */
+	fsnotify_put_mark(entry);
+
+	return ret;
+}
+
+static int inotify_new_watch(struct fsnotify_group *group,
+			     struct inode *inode,
+			     u32 arg)
+{
+	struct inotify_inode_mark_entry *tmp_ientry;
+	__u32 mask;
+	int ret;
+
+	/* don't allow invalid bits: we don't want flags set */
+	mask = inotify_arg_to_mask(arg);
+	if (unlikely(!mask))
+		return -EINVAL;
+
+	tmp_ientry = kmem_cache_alloc(inotify_inode_mark_cachep, GFP_KERNEL);
+	if (unlikely(!tmp_ientry))
+		return -ENOMEM;
+
+	fsnotify_init_mark(&tmp_ientry->fsn_entry, inotify_free_mark);
+	tmp_ientry->fsn_entry.mask = mask;
+	tmp_ientry->wd = -1;
+
+	ret = -ENOSPC;
+	if (atomic_read(&group->inotify_data.user->inotify_watches) >= inotify_max_user_watches)
+		goto out_err;
+retry:
+	ret = -ENOMEM;
+	if (unlikely(!idr_pre_get(&group->inotify_data.idr, GFP_KERNEL)))
+		goto out_err;
+
+	spin_lock(&group->inotify_data.idr_lock);
+	ret = idr_get_new_above(&group->inotify_data.idr, &tmp_ientry->fsn_entry,
+				group->inotify_data.last_wd,
+				&tmp_ientry->wd);
+	spin_unlock(&group->inotify_data.idr_lock);
+	if (ret) {
+		/* idr was out of memory allocate and try again */
+		if (ret == -EAGAIN)
+			goto retry;
+		goto out_err;
+	}
+
+	/* we put the mark on the idr, take a reference */
+	fsnotify_get_mark(&tmp_ientry->fsn_entry);
+
+	/* we are on the idr, now get on the inode */
+	ret = fsnotify_add_mark(&tmp_ientry->fsn_entry, group, inode);
+	if (ret) {
+		/* we failed to get on the inode, get off the idr */
+		inotify_remove_from_idr(group, tmp_ientry);
+		goto out_err;
 	}
+
+	/* update the idr hint, who cares about races, it's just a hint */
+	group->inotify_data.last_wd = tmp_ientry->wd;
+
+	/* increment the number of watches the user has */
+	atomic_inc(&group->inotify_data.user->inotify_watches);
+
+	/* return the watch descriptor for this new entry */
+	ret = tmp_ientry->wd;
+
+	/* match the ref from fsnotify_init_markentry() */
+	fsnotify_put_mark(&tmp_ientry->fsn_entry);
+
+out_err:
+	if (ret < 0)
+		kmem_cache_free(inotify_inode_mark_cachep, tmp_ientry);
+
+	return ret;
+}
+
+static int inotify_update_watch(struct fsnotify_group *group, struct inode *inode, u32 arg)
+{
+	int ret = 0;
+
+retry:
+	/* try to update and existing watch with the new arg */
+	ret = inotify_update_existing_watch(group, inode, arg);
+	/* no mark present, try to add a new one */
+	if (ret == -ENOENT)
+		ret = inotify_new_watch(group, inode, arg);
+	/*
+	 * inotify_new_watch could race with another thread which did an
+	 * inotify_new_watch between the update_existing and the add watch
+	 * here, go back and try to update an existing mark again.
+	 */
+	if (ret == -EEXIST)
+		goto retry;
+
 	return ret;
 }
 
@@ -532,7 +630,7 @@ static struct fsnotify_group *inotify_new_group(struct user_struct *user, unsign
 
 	spin_lock_init(&group->inotify_data.idr_lock);
 	idr_init(&group->inotify_data.idr);
-	group->inotify_data.last_wd = 0;
+	group->inotify_data.last_wd = 1;
 	group->inotify_data.user = user;
 	group->inotify_data.fa = NULL;
 
@@ -721,9 +819,6 @@ static int __init inotify_user_setup(void)
 
 	inotify_inode_mark_cachep = KMEM_CACHE(inotify_inode_mark_entry, SLAB_PANIC);
 	event_priv_cachep = KMEM_CACHE(inotify_event_private_data, SLAB_PANIC);
-	inotify_ignored_event = fsnotify_create_event(NULL, FS_IN_IGNORED, NULL, FSNOTIFY_EVENT_NONE, NULL, 0);
-	if (!inotify_ignored_event)
-		panic("unable to allocate the inotify ignored event\n");
 
 	inotify_max_queued_events = 16384;
 	inotify_max_user_instances = 128;
diff --git a/fs/notify/notification.c b/fs/notify/notification.c
index 959b73e756fd..3816d5750dd5 100644
--- a/fs/notify/notification.c
+++ b/fs/notify/notification.c
@@ -136,18 +136,28 @@ static bool event_compare(struct fsnotify_event *old, struct fsnotify_event *new
 {
 	if ((old->mask == new->mask) &&
 	    (old->to_tell == new->to_tell) &&
-	    (old->data_type == new->data_type)) {
+	    (old->data_type == new->data_type) &&
+	    (old->name_len == new->name_len)) {
 		switch (old->data_type) {
 		case (FSNOTIFY_EVENT_INODE):
-			if (old->inode == new->inode)
+			/* remember, after old was put on the wait_q we aren't
+			 * allowed to look at the inode any more, only thing
+			 * left to check was if the file_name is the same */
+			if (old->name_len &&
+			    !strcmp(old->file_name, new->file_name))
 				return true;
 			break;
 		case (FSNOTIFY_EVENT_PATH):
 			if ((old->path.mnt == new->path.mnt) &&
 			    (old->path.dentry == new->path.dentry))
 				return true;
+			break;
 		case (FSNOTIFY_EVENT_NONE):
-			return true;
+			if (old->mask & FS_Q_OVERFLOW)
+				return true;
+			else if (old->mask & FS_IN_IGNORED)
+				return false;
+			return false;
 		};
 	}
 	return false;
@@ -165,9 +175,7 @@ int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_even
 	struct list_head *list = &group->notification_list;
 	struct fsnotify_event_holder *last_holder;
 	struct fsnotify_event *last_event;
-
-	/* easy to tell if priv was attached to the event */
-	INIT_LIST_HEAD(&priv->event_list);
+	int ret = 0;
 
 	/*
 	 * There is one fsnotify_event_holder embedded inside each fsnotify_event.
@@ -188,6 +196,7 @@ alloc_holder:
 
 	if (group->q_len >= group->max_events) {
 		event = &q_overflow_event;
+		ret = -EOVERFLOW;
 		/* sorry, no private data on the overflow event */
 		priv = NULL;
 	}
@@ -229,7 +238,7 @@ alloc_holder:
 	mutex_unlock(&group->notification_mutex);
 
 	wake_up(&group->notification_waitq);
-	return 0;
+	return ret;
 }
 
 /*
@@ -339,18 +348,19 @@ static void initialize_event(struct fsnotify_event *event)
  * @name the filename, if available
  */
 struct fsnotify_event *fsnotify_create_event(struct inode *to_tell, __u32 mask, void *data,
-					     int data_type, const char *name, u32 cookie)
+					     int data_type, const char *name, u32 cookie,
+					     gfp_t gfp)
 {
 	struct fsnotify_event *event;
 
-	event = kmem_cache_alloc(fsnotify_event_cachep, GFP_KERNEL);
+	event = kmem_cache_alloc(fsnotify_event_cachep, gfp);
 	if (!event)
 		return NULL;
 
 	initialize_event(event);
 
 	if (name) {
-		event->file_name = kstrdup(name, GFP_KERNEL);
+		event->file_name = kstrdup(name, gfp);
 		if (!event->file_name) {
 			kmem_cache_free(fsnotify_event_cachep, event);
 			return NULL;
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 9edcde4974aa..ab513ddaeff2 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -1914,7 +1914,8 @@ static void ocfs2_adjust_adjacent_records(struct ocfs2_extent_rec *left_rec,
 	 * immediately to their right.
 	 */
 	left_clusters = le32_to_cpu(right_child_el->l_recs[0].e_cpos);
-	if (ocfs2_is_empty_extent(&right_child_el->l_recs[0])) {
+	if (!ocfs2_rec_clusters(right_child_el, &right_child_el->l_recs[0])) {
+		BUG_ON(right_child_el->l_tree_depth);
 		BUG_ON(le16_to_cpu(right_child_el->l_next_free_rec) <= 1);
 		left_clusters = le32_to_cpu(right_child_el->l_recs[1].e_cpos);
 	}
@@ -2476,15 +2477,37 @@ out_ret_path:
 	return ret;
 }
 
-static void ocfs2_update_edge_lengths(struct inode *inode, handle_t *handle,
-				      struct ocfs2_path *path)
+static int ocfs2_update_edge_lengths(struct inode *inode, handle_t *handle,
+				     int subtree_index, struct ocfs2_path *path)
 {
-	int i, idx;
+	int i, idx, ret;
 	struct ocfs2_extent_rec *rec;
 	struct ocfs2_extent_list *el;
 	struct ocfs2_extent_block *eb;
 	u32 range;
 
+	/*
+	 * In normal tree rotation process, we will never touch the
+	 * tree branch above subtree_index and ocfs2_extend_rotate_transaction
+	 * doesn't reserve the credits for them either.
+	 *
+	 * But we do have a special case here which will update the rightmost
+	 * records for all the bh in the path.
+	 * So we have to allocate extra credits and access them.
+	 */
+	ret = ocfs2_extend_trans(handle,
+				 handle->h_buffer_credits + subtree_index);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_path(inode, handle, path);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
 	/* Path should always be rightmost. */
 	eb = (struct ocfs2_extent_block *)path_leaf_bh(path)->b_data;
 	BUG_ON(eb->h_next_leaf_blk != 0ULL);
@@ -2505,6 +2528,8 @@ static void ocfs2_update_edge_lengths(struct inode *inode, handle_t *handle,
 
 		ocfs2_journal_dirty(handle, path->p_node[i].bh);
 	}
+out:
+	return ret;
 }
 
 static void ocfs2_unlink_path(struct inode *inode, handle_t *handle,
@@ -2717,7 +2742,12 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 	if (del_right_subtree) {
 		ocfs2_unlink_subtree(inode, handle, left_path, right_path,
 				     subtree_index, dealloc);
-		ocfs2_update_edge_lengths(inode, handle, left_path);
+		ret = ocfs2_update_edge_lengths(inode, handle, subtree_index,
+						left_path);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
 
 		eb = (struct ocfs2_extent_block *)path_leaf_bh(left_path)->b_data;
 		ocfs2_et_set_last_eb_blk(et, le64_to_cpu(eb->h_blkno));
@@ -3034,7 +3064,12 @@ static int ocfs2_remove_rightmost_path(struct inode *inode, handle_t *handle,
 
 		ocfs2_unlink_subtree(inode, handle, left_path, path,
 				     subtree_index, dealloc);
-		ocfs2_update_edge_lengths(inode, handle, left_path);
+		ret = ocfs2_update_edge_lengths(inode, handle, subtree_index,
+						left_path);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
 
 		eb = (struct ocfs2_extent_block *)path_leaf_bh(left_path)->b_data;
 		ocfs2_et_set_last_eb_blk(et, le64_to_cpu(eb->h_blkno));
@@ -6816,7 +6851,7 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 	}
 	status = 0;
 bail:
-
+	brelse(last_eb_bh);
 	mlog_exit(status);
 	return status;
 }
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index b2c52b3a1484..b401654011a2 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -193,6 +193,7 @@ static int ocfs2_get_block(struct inode *inode, sector_t iblock,
 			     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 			mlog(ML_ERROR, "Size %llu, clusters %u\n", (unsigned long long)i_size_read(inode), OCFS2_I(inode)->ip_clusters);
 			dump_stack();
+			goto bail;
 		}
 
 		past_eof = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
@@ -894,18 +895,17 @@ struct ocfs2_write_cluster_desc {
 	 */
 	unsigned	c_new;
 	unsigned	c_unwritten;
+	unsigned	c_needs_zero;
 };
 
-static inline int ocfs2_should_zero_cluster(struct ocfs2_write_cluster_desc *d)
-{
-	return d->c_new || d->c_unwritten;
-}
-
 struct ocfs2_write_ctxt {
 	/* Logical cluster position / len of write */
 	u32				w_cpos;
 	u32				w_clen;
 
+	/* First cluster allocated in a nonsparse extend */
+	u32				w_first_new_cpos;
+
 	struct ocfs2_write_cluster_desc	w_desc[OCFS2_MAX_CLUSTERS_PER_PAGE];
 
 	/*
@@ -983,6 +983,7 @@ static int ocfs2_alloc_write_ctxt(struct ocfs2_write_ctxt **wcp,
 		return -ENOMEM;
 
 	wc->w_cpos = pos >> osb->s_clustersize_bits;
+	wc->w_first_new_cpos = UINT_MAX;
 	cend = (pos + len - 1) >> osb->s_clustersize_bits;
 	wc->w_clen = cend - wc->w_cpos + 1;
 	get_bh(di_bh);
@@ -1217,20 +1218,18 @@ out:
  */
 static int ocfs2_write_cluster(struct address_space *mapping,
 			       u32 phys, unsigned int unwritten,
+			       unsigned int should_zero,
 			       struct ocfs2_alloc_context *data_ac,
 			       struct ocfs2_alloc_context *meta_ac,
 			       struct ocfs2_write_ctxt *wc, u32 cpos,
 			       loff_t user_pos, unsigned user_len)
 {
-	int ret, i, new, should_zero = 0;
+	int ret, i, new;
 	u64 v_blkno, p_blkno;
 	struct inode *inode = mapping->host;
 	struct ocfs2_extent_tree et;
 
 	new = phys == 0 ? 1 : 0;
-	if (new || unwritten)
-		should_zero = 1;
-
 	if (new) {
 		u32 tmp_pos;
 
@@ -1301,7 +1300,7 @@ static int ocfs2_write_cluster(struct address_space *mapping,
 		if (tmpret) {
 			mlog_errno(tmpret);
 			if (ret == 0)
-				tmpret = ret;
+				ret = tmpret;
 		}
 	}
 
@@ -1341,7 +1340,9 @@ static int ocfs2_write_cluster_by_desc(struct address_space *mapping,
 			local_len = osb->s_clustersize - cluster_off;
 
 		ret = ocfs2_write_cluster(mapping, desc->c_phys,
-					  desc->c_unwritten, data_ac, meta_ac,
+					  desc->c_unwritten,
+					  desc->c_needs_zero,
+					  data_ac, meta_ac,
 					  wc, desc->c_cpos, pos, local_len);
 		if (ret) {
 			mlog_errno(ret);
@@ -1391,14 +1392,14 @@ static void ocfs2_set_target_boundaries(struct ocfs2_super *osb,
 		 * newly allocated cluster.
 		 */
 		desc = &wc->w_desc[0];
-		if (ocfs2_should_zero_cluster(desc))
+		if (desc->c_needs_zero)
 			ocfs2_figure_cluster_boundaries(osb,
 							desc->c_cpos,
 							&wc->w_target_from,
 							NULL);
 
 		desc = &wc->w_desc[wc->w_clen - 1];
-		if (ocfs2_should_zero_cluster(desc))
+		if (desc->c_needs_zero)
 			ocfs2_figure_cluster_boundaries(osb,
 							desc->c_cpos,
 							NULL,
@@ -1466,13 +1467,28 @@ static int ocfs2_populate_write_desc(struct inode *inode,
 			phys++;
 		}
 
+		/*
+		 * If w_first_new_cpos is < UINT_MAX, we have a non-sparse
+		 * file that got extended.  w_first_new_cpos tells us
+		 * where the newly allocated clusters are so we can
+		 * zero them.
+		 */
+		if (desc->c_cpos >= wc->w_first_new_cpos) {
+			BUG_ON(phys == 0);
+			desc->c_needs_zero = 1;
+		}
+
 		desc->c_phys = phys;
 		if (phys == 0) {
 			desc->c_new = 1;
+			desc->c_needs_zero = 1;
 			*clusters_to_alloc = *clusters_to_alloc + 1;
 		}
-		if (ext_flags & OCFS2_EXT_UNWRITTEN)
+
+		if (ext_flags & OCFS2_EXT_UNWRITTEN) {
 			desc->c_unwritten = 1;
+			desc->c_needs_zero = 1;
+		}
 
 		num_clusters--;
 	}
@@ -1632,10 +1648,13 @@ static int ocfs2_expand_nonsparse_inode(struct inode *inode, loff_t pos,
 	if (newsize <= i_size_read(inode))
 		return 0;
 
-	ret = ocfs2_extend_no_holes(inode, newsize, newsize - len);
+	ret = ocfs2_extend_no_holes(inode, newsize, pos);
 	if (ret)
 		mlog_errno(ret);
 
+	wc->w_first_new_cpos =
+		ocfs2_clusters_for_bytes(inode->i_sb, i_size_read(inode));
+
 	return ret;
 }
 
@@ -1644,7 +1663,7 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 			     struct page **pagep, void **fsdata,
 			     struct buffer_head *di_bh, struct page *mmap_page)
 {
-	int ret, credits = OCFS2_INODE_UPDATE_CREDITS;
+	int ret, cluster_of_pages, credits = OCFS2_INODE_UPDATE_CREDITS;
 	unsigned int clusters_to_alloc, extents_to_split;
 	struct ocfs2_write_ctxt *wc;
 	struct inode *inode = mapping->host;
@@ -1722,8 +1741,19 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 
 	}
 
-	ocfs2_set_target_boundaries(osb, wc, pos, len,
-				    clusters_to_alloc + extents_to_split);
+	/*
+	 * We have to zero sparse allocated clusters, unwritten extent clusters,
+	 * and non-sparse clusters we just extended.  For non-sparse writes,
+	 * we know zeros will only be needed in the first and/or last cluster.
+	 */
+	if (clusters_to_alloc || extents_to_split ||
+	    wc->w_desc[0].c_needs_zero ||
+	    wc->w_desc[wc->w_clen - 1].c_needs_zero)
+		cluster_of_pages = 1;
+	else
+		cluster_of_pages = 0;
+
+	ocfs2_set_target_boundaries(osb, wc, pos, len, cluster_of_pages);
 
 	handle = ocfs2_start_trans(osb, credits);
 	if (IS_ERR(handle)) {
@@ -1756,8 +1786,7 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 	 * extent.
 	 */
 	ret = ocfs2_grab_pages_for_write(mapping, wc, wc->w_cpos, pos,
-					 clusters_to_alloc + extents_to_split,
-					 mmap_page);
+					 cluster_of_pages, mmap_page);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_quota;
diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c
index b574431a031d..2f28b7de2c8d 100644
--- a/fs/ocfs2/dcache.c
+++ b/fs/ocfs2/dcache.c
@@ -310,22 +310,19 @@ out_attach:
 	return ret;
 }
 
-static DEFINE_SPINLOCK(dentry_list_lock);
+DEFINE_SPINLOCK(dentry_list_lock);
 
 /* We limit the number of dentry locks to drop in one go. We have
  * this limit so that we don't starve other users of ocfs2_wq. */
 #define DL_INODE_DROP_COUNT 64
 
 /* Drop inode references from dentry locks */
-void ocfs2_drop_dl_inodes(struct work_struct *work)
+static void __ocfs2_drop_dl_inodes(struct ocfs2_super *osb, int drop_count)
 {
-	struct ocfs2_super *osb = container_of(work, struct ocfs2_super,
-					       dentry_lock_work);
 	struct ocfs2_dentry_lock *dl;
-	int drop_count = DL_INODE_DROP_COUNT;
 
 	spin_lock(&dentry_list_lock);
-	while (osb->dentry_lock_list && drop_count--) {
+	while (osb->dentry_lock_list && (drop_count < 0 || drop_count--)) {
 		dl = osb->dentry_lock_list;
 		osb->dentry_lock_list = dl->dl_next;
 		spin_unlock(&dentry_list_lock);
@@ -333,11 +330,32 @@ void ocfs2_drop_dl_inodes(struct work_struct *work)
 		kfree(dl);
 		spin_lock(&dentry_list_lock);
 	}
-	if (osb->dentry_lock_list)
+	spin_unlock(&dentry_list_lock);
+}
+
+void ocfs2_drop_dl_inodes(struct work_struct *work)
+{
+	struct ocfs2_super *osb = container_of(work, struct ocfs2_super,
+					       dentry_lock_work);
+
+	__ocfs2_drop_dl_inodes(osb, DL_INODE_DROP_COUNT);
+	/*
+	 * Don't queue dropping if umount is in progress. We flush the
+	 * list in ocfs2_dismount_volume
+	 */
+	spin_lock(&dentry_list_lock);
+	if (osb->dentry_lock_list &&
+	    !ocfs2_test_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED))
 		queue_work(ocfs2_wq, &osb->dentry_lock_work);
 	spin_unlock(&dentry_list_lock);
 }
 
+/* Flush the whole work queue */
+void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb)
+{
+	__ocfs2_drop_dl_inodes(osb, -1);
+}
+
 /*
  * ocfs2_dentry_iput() and friends.
  *
@@ -368,7 +386,8 @@ static void ocfs2_drop_dentry_lock(struct ocfs2_super *osb,
 	/* We leave dropping of inode reference to ocfs2_wq as that can
 	 * possibly lead to inode deletion which gets tricky */
 	spin_lock(&dentry_list_lock);
-	if (!osb->dentry_lock_list)
+	if (!osb->dentry_lock_list &&
+	    !ocfs2_test_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED))
 		queue_work(ocfs2_wq, &osb->dentry_lock_work);
 	dl->dl_next = osb->dentry_lock_list;
 	osb->dentry_lock_list = dl;
diff --git a/fs/ocfs2/dcache.h b/fs/ocfs2/dcache.h
index faa12e75f98d..f5dd1789acf1 100644
--- a/fs/ocfs2/dcache.h
+++ b/fs/ocfs2/dcache.h
@@ -49,10 +49,13 @@ struct ocfs2_dentry_lock {
 int ocfs2_dentry_attach_lock(struct dentry *dentry, struct inode *inode,
 			     u64 parent_blkno);
 
+extern spinlock_t dentry_list_lock;
+
 void ocfs2_dentry_lock_put(struct ocfs2_super *osb,
 			   struct ocfs2_dentry_lock *dl);
 
 void ocfs2_drop_dl_inodes(struct work_struct *work);
+void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb);
 
 struct dentry *ocfs2_find_local_alias(struct inode *inode, u64 parent_blkno,
 				      int skip_unhashed);
diff --git a/fs/ocfs2/dlm/dlmast.c b/fs/ocfs2/dlm/dlmast.c
index d07ddbe4b283..81eff8e58322 100644
--- a/fs/ocfs2/dlm/dlmast.c
+++ b/fs/ocfs2/dlm/dlmast.c
@@ -103,7 +103,6 @@ static void __dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
 		     lock->ast_pending, lock->ml.type);
 		BUG();
 	}
-	BUG_ON(!list_empty(&lock->ast_list));
 	if (lock->ast_pending)
 		mlog(0, "lock has an ast getting flushed right now\n");
 
diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c
index bcb9260c3735..43e6e3280569 100644
--- a/fs/ocfs2/dlm/dlmrecovery.c
+++ b/fs/ocfs2/dlm/dlmrecovery.c
@@ -1118,7 +1118,7 @@ static int dlm_send_mig_lockres_msg(struct dlm_ctxt *dlm,
 
 	mlog(0, "%s:%.*s: sending mig lockres (%s) to %u\n",
 	     dlm->name, res->lockname.len, res->lockname.name,
-	     orig_flags & DLM_MRES_MIGRATION ? "migrate" : "recovery",
+	     orig_flags & DLM_MRES_MIGRATION ? "migration" : "recovery",
 	     send_to);
 
 	/* send it */
diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c
index fcf879ed6930..756f5b0998e0 100644
--- a/fs/ocfs2/dlm/dlmunlock.c
+++ b/fs/ocfs2/dlm/dlmunlock.c
@@ -122,7 +122,7 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm,
 	 * that still has AST's pending... */
 	in_use = !list_empty(&lock->ast_list);
 	spin_unlock(&dlm->ast_lock);
-	if (in_use) {
+	if (in_use && !(flags & LKM_CANCEL)) {
 	       mlog(ML_ERROR, "lockres %.*s: Someone is calling dlmunlock "
 		    "while waiting for an ast!", res->lockname.len,
 		    res->lockname.name);
@@ -131,7 +131,7 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm,
 
 	spin_lock(&res->spinlock);
 	if (res->state & DLM_LOCK_RES_IN_PROGRESS) {
-		if (master_node) {
+		if (master_node && !(flags & LKM_CANCEL)) {
 			mlog(ML_ERROR, "lockres in progress!\n");
 			spin_unlock(&res->spinlock);
 			return DLM_FORWARD;
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 62442e413a00..aa501d3f93f1 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -1851,6 +1851,7 @@ relock:
 		if (ret)
 			goto out_dio;
 
+		count = ocount;
 		ret = generic_write_checks(file, ppos, &count,
 					   S_ISBLK(inode->i_mode));
 		if (ret)
@@ -1918,8 +1919,10 @@ out_sems:
 
 	mutex_unlock(&inode->i_mutex);
 
+	if (written)
+		ret = written;
 	mlog_exit(ret);
-	return written ? written : ret;
+	return ret;
 }
 
 static int ocfs2_splice_to_file(struct pipe_inode_info *pipe,
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index f033760ecbea..c48b93ac6b65 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -1954,10 +1954,16 @@ void ocfs2_orphan_scan_init(struct ocfs2_super *osb)
 	os->os_osb = osb;
 	os->os_count = 0;
 	os->os_seqno = 0;
-	os->os_scantime = CURRENT_TIME;
 	mutex_init(&os->os_lock);
 	INIT_DELAYED_WORK(&os->os_orphan_scan_work, ocfs2_orphan_scan_work);
+}
 
+void ocfs2_orphan_scan_start(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+
+	os = &osb->osb_orphan_scan;
+	os->os_scantime = CURRENT_TIME;
 	if (ocfs2_is_hard_readonly(osb) || ocfs2_mount_local(osb))
 		atomic_set(&os->os_state, ORPHAN_SCAN_INACTIVE);
 	else {
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index 5432c7f79cc6..2c3222aec622 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -145,6 +145,7 @@ 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. */
 void ocfs2_orphan_scan_init(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_start(struct ocfs2_super *osb);
 void ocfs2_orphan_scan_stop(struct ocfs2_super *osb);
 void ocfs2_orphan_scan_exit(struct ocfs2_super *osb);
 
@@ -329,20 +330,27 @@ int                  ocfs2_journal_dirty(handle_t *handle,
 /* extended attribute block update */
 #define OCFS2_XATTR_BLOCK_UPDATE_CREDITS 1
 
+/* Update of a single quota block */
+#define OCFS2_QUOTA_BLOCK_UPDATE_CREDITS 1
+
 /* global quotafile inode update, data block */
-#define OCFS2_QINFO_WRITE_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 1)
+#define OCFS2_QINFO_WRITE_CREDITS (OCFS2_INODE_UPDATE_CREDITS + \
+				   OCFS2_QUOTA_BLOCK_UPDATE_CREDITS)
 
+#define OCFS2_LOCAL_QINFO_WRITE_CREDITS OCFS2_QUOTA_BLOCK_UPDATE_CREDITS
 /*
  * The two writes below can accidentally see global info dirty due
  * to set_info() quotactl so make them prepared for the writes.
  */
 /* quota data block, global info */
 /* Write to local quota file */
-#define OCFS2_QWRITE_CREDITS (OCFS2_QINFO_WRITE_CREDITS + 1)
+#define OCFS2_QWRITE_CREDITS (OCFS2_QINFO_WRITE_CREDITS + \
+			      OCFS2_QUOTA_BLOCK_UPDATE_CREDITS)
 
 /* global quota data block, local quota data block, global quota inode,
  * global quota info */
-#define OCFS2_QSYNC_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 3)
+#define OCFS2_QSYNC_CREDITS (OCFS2_QINFO_WRITE_CREDITS + \
+			     2 * OCFS2_QUOTA_BLOCK_UPDATE_CREDITS)
 
 static inline int ocfs2_quota_trans_credits(struct super_block *sb)
 {
@@ -355,11 +363,6 @@ static inline int ocfs2_quota_trans_credits(struct super_block *sb)
 	return credits;
 }
 
-/* Number of credits needed for removing quota structure from file */
-int ocfs2_calc_qdel_credits(struct super_block *sb, int type);
-/* Number of credits needed for initialization of new quota structure */
-int ocfs2_calc_qinit_credits(struct super_block *sb, int type);
-
 /* group extend. inode update and last group update. */
 #define OCFS2_GROUP_EXTEND_CREDITS	(OCFS2_INODE_UPDATE_CREDITS + 1)
 
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index c9345ebb8493..39e1d5a39505 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -224,10 +224,12 @@ enum ocfs2_mount_options
 	OCFS2_MOUNT_GRPQUOTA = 1 << 10, /* We support group quotas */
 };
 
-#define OCFS2_OSB_SOFT_RO	0x0001
-#define OCFS2_OSB_HARD_RO	0x0002
-#define OCFS2_OSB_ERROR_FS	0x0004
-#define OCFS2_DEFAULT_ATIME_QUANTUM	60
+#define OCFS2_OSB_SOFT_RO			0x0001
+#define OCFS2_OSB_HARD_RO			0x0002
+#define OCFS2_OSB_ERROR_FS			0x0004
+#define OCFS2_OSB_DROP_DENTRY_LOCK_IMMED	0x0008
+
+#define OCFS2_DEFAULT_ATIME_QUANTUM		60
 
 struct ocfs2_journal;
 struct ocfs2_slot_info;
@@ -490,6 +492,18 @@ static inline void ocfs2_set_osb_flag(struct ocfs2_super *osb,
 	spin_unlock(&osb->osb_lock);
 }
 
+
+static inline unsigned long  ocfs2_test_osb_flag(struct ocfs2_super *osb,
+						 unsigned long flag)
+{
+	unsigned long ret;
+
+	spin_lock(&osb->osb_lock);
+	ret = osb->osb_flags & flag;
+	spin_unlock(&osb->osb_lock);
+	return ret;
+}
+
 static inline void ocfs2_set_ro_flag(struct ocfs2_super *osb,
 				     int hard)
 {
diff --git a/fs/ocfs2/ocfs2_lockid.h b/fs/ocfs2/ocfs2_lockid.h
index fcdba091af3d..c212cf5a2bdf 100644
--- a/fs/ocfs2/ocfs2_lockid.h
+++ b/fs/ocfs2/ocfs2_lockid.h
@@ -108,6 +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_NFS_SYNC] = "NFSSync",
 	[OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
 };
 
diff --git a/fs/ocfs2/quota.h b/fs/ocfs2/quota.h
index 7365e2e08706..3fb96fcd4c81 100644
--- a/fs/ocfs2/quota.h
+++ b/fs/ocfs2/quota.h
@@ -50,7 +50,6 @@ struct ocfs2_mem_dqinfo {
 	unsigned int dqi_chunks;	/* Number of chunks in local quota file */
 	unsigned int dqi_blocks;	/* Number of blocks allocated for local quota file */
 	unsigned int dqi_syncms;	/* How often should we sync with other nodes */
-	unsigned int dqi_syncjiff;	/* Precomputed dqi_syncms in jiffies */
 	struct list_head dqi_chunk;	/* List of chunks */
 	struct inode *dqi_gqinode;	/* Global quota file inode */
 	struct ocfs2_lock_res dqi_gqlock;	/* Lock protecting quota information structure */
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index edfa60cd155c..44f2a5e1d042 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -23,6 +23,7 @@
 #include "sysfile.h"
 #include "dlmglue.h"
 #include "uptodate.h"
+#include "super.h"
 #include "quota.h"
 
 static struct workqueue_struct *ocfs2_quota_wq = NULL;
@@ -69,6 +70,7 @@ static void ocfs2_global_mem2diskdqb(void *dp, struct dquot *dquot)
 	d->dqb_curspace = cpu_to_le64(m->dqb_curspace);
 	d->dqb_btime = cpu_to_le64(m->dqb_btime);
 	d->dqb_itime = cpu_to_le64(m->dqb_itime);
+	d->dqb_pad1 = d->dqb_pad2 = 0;
 }
 
 static int ocfs2_global_is_id(void *dp, struct dquot *dquot)
@@ -113,6 +115,15 @@ int ocfs2_read_quota_block(struct inode *inode, u64 v_block,
 	int rc = 0;
 	struct buffer_head *tmp = *bh;
 
+	if (i_size_read(inode) >> inode->i_sb->s_blocksize_bits <= v_block) {
+		ocfs2_error(inode->i_sb,
+			    "Quota file %llu is probably corrupted! Requested "
+			    "to read block %Lu but file has size only %Lu\n",
+			    (unsigned long long)OCFS2_I(inode)->ip_blkno,
+			    (unsigned long long)v_block,
+			    (unsigned long long)i_size_read(inode));
+		return -EIO;
+	}
 	rc = ocfs2_read_virt_blocks(inode, v_block, 1, &tmp, 0,
 				    ocfs2_validate_quota_block);
 	if (rc)
@@ -211,14 +222,13 @@ ssize_t ocfs2_quota_write(struct super_block *sb, int type,
 
 	mutex_lock_nested(&gqinode->i_mutex, I_MUTEX_QUOTA);
 	if (gqinode->i_size < off + len) {
-		down_write(&OCFS2_I(gqinode)->ip_alloc_sem);
-		err = ocfs2_extend_no_holes(gqinode, off + len, off);
-		up_write(&OCFS2_I(gqinode)->ip_alloc_sem);
-		if (err < 0)
-			goto out;
+		loff_t rounded_end =
+				ocfs2_align_bytes_to_blocks(sb, off + len);
+
+		/* Space is already allocated in ocfs2_global_read_dquot() */
 		err = ocfs2_simple_size_update(gqinode,
 					       oinfo->dqi_gqi_bh,
-					       off + len);
+					       rounded_end);
 		if (err < 0)
 			goto out;
 		new = 1;
@@ -234,7 +244,7 @@ ssize_t ocfs2_quota_write(struct super_block *sb, int type,
 	}
 	if (err) {
 		mlog_errno(err);
-		return err;
+		goto out;
 	}
 	lock_buffer(bh);
 	if (new)
@@ -342,7 +352,6 @@ int ocfs2_global_read_info(struct super_block *sb, int type)
 	info->dqi_bgrace = le32_to_cpu(dinfo.dqi_bgrace);
 	info->dqi_igrace = le32_to_cpu(dinfo.dqi_igrace);
 	oinfo->dqi_syncms = le32_to_cpu(dinfo.dqi_syncms);
-	oinfo->dqi_syncjiff = msecs_to_jiffies(oinfo->dqi_syncms);
 	oinfo->dqi_gi.dqi_blocks = le32_to_cpu(dinfo.dqi_blocks);
 	oinfo->dqi_gi.dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk);
 	oinfo->dqi_gi.dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry);
@@ -352,7 +361,7 @@ int ocfs2_global_read_info(struct super_block *sb, int type)
 	oinfo->dqi_gi.dqi_qtree_depth = qtree_depth(&oinfo->dqi_gi);
 	INIT_DELAYED_WORK(&oinfo->dqi_sync_work, qsync_work_fn);
 	queue_delayed_work(ocfs2_quota_wq, &oinfo->dqi_sync_work,
-			   oinfo->dqi_syncjiff);
+			   msecs_to_jiffies(oinfo->dqi_syncms));
 
 out_err:
 	mlog_exit(status);
@@ -402,13 +411,36 @@ int ocfs2_global_write_info(struct super_block *sb, int type)
 	return err;
 }
 
+static int ocfs2_global_qinit_alloc(struct super_block *sb, int type)
+{
+	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
+
+	/*
+	 * We may need to allocate tree blocks and a leaf block but not the
+	 * root block
+	 */
+	return oinfo->dqi_gi.dqi_qtree_depth;
+}
+
+static int ocfs2_calc_global_qinit_credits(struct super_block *sb, int type)
+{
+	/* We modify all the allocated blocks, tree root, and info block */
+	return (ocfs2_global_qinit_alloc(sb, type) + 2) *
+			OCFS2_QUOTA_BLOCK_UPDATE_CREDITS;
+}
+
 /* Read in information from global quota file and acquire a reference to it.
  * dquot_acquire() has already started the transaction and locked quota file */
 int ocfs2_global_read_dquot(struct dquot *dquot)
 {
 	int err, err2, ex = 0;
-	struct ocfs2_mem_dqinfo *info =
-			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
+	struct super_block *sb = dquot->dq_sb;
+	int type = dquot->dq_type;
+	struct ocfs2_mem_dqinfo *info = sb_dqinfo(sb, type)->dqi_priv;
+	struct ocfs2_super *osb = OCFS2_SB(sb);
+	struct inode *gqinode = info->dqi_gqinode;
+	int need_alloc = ocfs2_global_qinit_alloc(sb, type);
+	handle_t *handle = NULL;
 
 	err = ocfs2_qinfo_lock(info, 0);
 	if (err < 0)
@@ -419,14 +451,33 @@ int ocfs2_global_read_dquot(struct dquot *dquot)
 	OCFS2_DQUOT(dquot)->dq_use_count++;
 	OCFS2_DQUOT(dquot)->dq_origspace = dquot->dq_dqb.dqb_curspace;
 	OCFS2_DQUOT(dquot)->dq_originodes = dquot->dq_dqb.dqb_curinodes;
+	ocfs2_qinfo_unlock(info, 0);
+
 	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;
 		ex = 1;
+		/*
+		 * Add blocks to quota file before we start a transaction since
+		 * locking allocators ranks above a transaction start
+		 */
+		WARN_ON(journal_current_handle());
+		down_write(&OCFS2_I(gqinode)->ip_alloc_sem);
+		err = ocfs2_extend_no_holes(gqinode,
+			gqinode->i_size + (need_alloc << sb->s_blocksize_bits),
+			gqinode->i_size);
+		up_write(&OCFS2_I(gqinode)->ip_alloc_sem);
+		if (err < 0)
+			goto out;
 	}
+
+	handle = ocfs2_start_trans(osb,
+				   ocfs2_calc_global_qinit_credits(sb, type));
+	if (IS_ERR(handle)) {
+		err = PTR_ERR(handle);
+		goto out;
+	}
+	err = ocfs2_qinfo_lock(info, ex);
+	if (err < 0)
+		goto out_trans;
 	err = qtree_write_dquot(&info->dqi_gi, dquot);
 	if (ex && info_dirty(sb_dqinfo(dquot->dq_sb, dquot->dq_type))) {
 		err2 = __ocfs2_global_write_info(dquot->dq_sb, dquot->dq_type);
@@ -438,6 +489,9 @@ out_qlock:
 		ocfs2_qinfo_unlock(info, 1);
 	else
 		ocfs2_qinfo_unlock(info, 0);
+out_trans:
+	if (handle)
+		ocfs2_commit_trans(osb, handle);
 out:
 	if (err < 0)
 		mlog_errno(err);
@@ -607,7 +661,7 @@ static void qsync_work_fn(struct work_struct *work)
 
 	dquot_scan_active(sb, ocfs2_sync_dquot_helper, oinfo->dqi_type);
 	queue_delayed_work(ocfs2_quota_wq, &oinfo->dqi_sync_work,
-			   oinfo->dqi_syncjiff);
+			   msecs_to_jiffies(oinfo->dqi_syncms));
 }
 
 /*
@@ -635,20 +689,18 @@ out:
 	return status;
 }
 
-int ocfs2_calc_qdel_credits(struct super_block *sb, int type)
+static int ocfs2_calc_qdel_credits(struct super_block *sb, int type)
 {
-	struct ocfs2_mem_dqinfo *oinfo;
-	int features[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
-				    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA };
-
-	if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, features[type]))
-		return 0;
-
-	oinfo = sb_dqinfo(sb, type)->dqi_priv;
-	/* We modify tree, leaf block, global info, local chunk header,
-	 * global and local inode */
-	return oinfo->dqi_gi.dqi_qtree_depth + 2 + 1 +
-	       2 * OCFS2_INODE_UPDATE_CREDITS;
+	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
+	/*
+	 * We modify tree, leaf block, global info, local chunk header,
+	 * global and local inode; OCFS2_QINFO_WRITE_CREDITS already
+	 * accounts for inode update
+	 */
+	return (oinfo->dqi_gi.dqi_qtree_depth + 2) *
+	       OCFS2_QUOTA_BLOCK_UPDATE_CREDITS +
+	       OCFS2_QINFO_WRITE_CREDITS +
+	       OCFS2_INODE_UPDATE_CREDITS;
 }
 
 static int ocfs2_release_dquot(struct dquot *dquot)
@@ -680,33 +732,10 @@ out:
 	return status;
 }
 
-int ocfs2_calc_qinit_credits(struct super_block *sb, int type)
-{
-	struct ocfs2_mem_dqinfo *oinfo;
-	int features[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
-				    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA };
-	struct ocfs2_dinode *lfe, *gfe;
-
-	if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, features[type]))
-		return 0;
-
-	oinfo = sb_dqinfo(sb, type)->dqi_priv;
-	gfe = (struct ocfs2_dinode *)oinfo->dqi_gqi_bh->b_data;
-	lfe = (struct ocfs2_dinode *)oinfo->dqi_lqi_bh->b_data;
-	/* We can extend local file + global file. In local file we
-	 * can modify info, chunk header block and dquot block. In
-	 * global file we can modify info, tree and leaf block */
-	return ocfs2_calc_extend_credits(sb, &lfe->id2.i_list, 0) +
-	       ocfs2_calc_extend_credits(sb, &gfe->id2.i_list, 0) +
-	       3 + oinfo->dqi_gi.dqi_qtree_depth + 2;
-}
-
 static int ocfs2_acquire_dquot(struct dquot *dquot)
 {
-	handle_t *handle;
 	struct ocfs2_mem_dqinfo *oinfo =
 			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
-	struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
 	int status = 0;
 
 	mlog_entry("id=%u, type=%d", dquot->dq_id, dquot->dq_type);
@@ -715,16 +744,7 @@ static int ocfs2_acquire_dquot(struct dquot *dquot)
 	status = ocfs2_lock_global_qf(oinfo, 1);
 	if (status < 0)
 		goto out;
-	handle = ocfs2_start_trans(osb,
-		ocfs2_calc_qinit_credits(dquot->dq_sb, dquot->dq_type));
-	if (IS_ERR(handle)) {
-		status = PTR_ERR(handle);
-		mlog_errno(status);
-		goto out_ilock;
-	}
 	status = dquot_acquire(dquot);
-	ocfs2_commit_trans(osb, handle);
-out_ilock:
 	ocfs2_unlock_global_qf(oinfo, 1);
 out:
 	mlog_exit(status);
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index 5a460fa82553..bdb09cb6e1fe 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -20,6 +20,7 @@
 #include "sysfile.h"
 #include "dlmglue.h"
 #include "quota.h"
+#include "uptodate.h"
 
 /* Number of local quota structures per block */
 static inline unsigned int ol_quota_entries_per_block(struct super_block *sb)
@@ -100,7 +101,8 @@ static int ocfs2_modify_bh(struct inode *inode, struct buffer_head *bh,
 	handle_t *handle;
 	int status;
 
-	handle = ocfs2_start_trans(OCFS2_SB(sb), 1);
+	handle = ocfs2_start_trans(OCFS2_SB(sb),
+				   OCFS2_QUOTA_BLOCK_UPDATE_CREDITS);
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
 		mlog_errno(status);
@@ -610,7 +612,8 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb,
 			goto out_bh;
 		/* Mark quota file as clean if we are recovering quota file of
 		 * some other node. */
-		handle = ocfs2_start_trans(osb, 1);
+		handle = ocfs2_start_trans(osb,
+					   OCFS2_LOCAL_QINFO_WRITE_CREDITS);
 		if (IS_ERR(handle)) {
 			status = PTR_ERR(handle);
 			mlog_errno(status);
@@ -940,7 +943,7 @@ static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk(
 	struct ocfs2_local_disk_chunk *dchunk;
 	int status;
 	handle_t *handle;
-	struct buffer_head *bh = NULL;
+	struct buffer_head *bh = NULL, *dbh = NULL;
 	u64 p_blkno;
 
 	/* We are protected by dqio_sem so no locking needed */
@@ -964,32 +967,35 @@ static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk(
 		mlog_errno(status);
 		goto out;
 	}
+	/* Local quota info and two new blocks we initialize */
+	handle = ocfs2_start_trans(OCFS2_SB(sb),
+			OCFS2_LOCAL_QINFO_WRITE_CREDITS +
+			2 * OCFS2_QUOTA_BLOCK_UPDATE_CREDITS);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out;
+	}
 
+	/* Initialize chunk header */
 	down_read(&OCFS2_I(lqinode)->ip_alloc_sem);
 	status = ocfs2_extent_map_get_blocks(lqinode, oinfo->dqi_blocks,
 					     &p_blkno, NULL, NULL);
 	up_read(&OCFS2_I(lqinode)->ip_alloc_sem);
 	if (status < 0) {
 		mlog_errno(status);
-		goto out;
+		goto out_trans;
 	}
 	bh = sb_getblk(sb, p_blkno);
 	if (!bh) {
 		status = -ENOMEM;
 		mlog_errno(status);
-		goto out;
+		goto out_trans;
 	}
 	dchunk = (struct ocfs2_local_disk_chunk *)bh->b_data;
-
-	handle = ocfs2_start_trans(OCFS2_SB(sb), 2);
-	if (IS_ERR(handle)) {
-		status = PTR_ERR(handle);
-		mlog_errno(status);
-		goto out;
-	}
-
+	ocfs2_set_new_buffer_uptodate(lqinode, bh);
 	status = ocfs2_journal_access_dq(handle, lqinode, bh,
-					 OCFS2_JOURNAL_ACCESS_WRITE);
+					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_trans;
@@ -999,7 +1005,6 @@ static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk(
 	memset(dchunk->dqc_bitmap, 0,
 	       sb->s_blocksize - sizeof(struct ocfs2_local_disk_chunk) -
 	       OCFS2_QBLK_RESERVED_SPACE);
-	set_buffer_uptodate(bh);
 	unlock_buffer(bh);
 	status = ocfs2_journal_dirty(handle, bh);
 	if (status < 0) {
@@ -1007,6 +1012,38 @@ static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk(
 		goto out_trans;
 	}
 
+	/* Initialize new block with structures */
+	down_read(&OCFS2_I(lqinode)->ip_alloc_sem);
+	status = ocfs2_extent_map_get_blocks(lqinode, oinfo->dqi_blocks + 1,
+					     &p_blkno, NULL, NULL);
+	up_read(&OCFS2_I(lqinode)->ip_alloc_sem);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	dbh = sb_getblk(sb, p_blkno);
+	if (!dbh) {
+		status = -ENOMEM;
+		mlog_errno(status);
+		goto out_trans;
+	}
+	ocfs2_set_new_buffer_uptodate(lqinode, dbh);
+	status = ocfs2_journal_access_dq(handle, lqinode, dbh,
+					 OCFS2_JOURNAL_ACCESS_CREATE);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	lock_buffer(dbh);
+	memset(dbh->b_data, 0, sb->s_blocksize - OCFS2_QBLK_RESERVED_SPACE);
+	unlock_buffer(dbh);
+	status = ocfs2_journal_dirty(handle, dbh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+
+	/* Update local quotafile info */
 	oinfo->dqi_blocks += 2;
 	oinfo->dqi_chunks++;
 	status = ocfs2_local_write_info(sb, type);
@@ -1031,6 +1068,7 @@ out_trans:
 	ocfs2_commit_trans(OCFS2_SB(sb), handle);
 out:
 	brelse(bh);
+	brelse(dbh);
 	kmem_cache_free(ocfs2_qf_chunk_cachep, chunk);
 	return ERR_PTR(status);
 }
@@ -1048,6 +1086,8 @@ static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file(
 	struct ocfs2_local_disk_chunk *dchunk;
 	int epb = ol_quota_entries_per_block(sb);
 	unsigned int chunk_blocks;
+	struct buffer_head *bh;
+	u64 p_blkno;
 	int status;
 	handle_t *handle;
 
@@ -1075,12 +1115,49 @@ static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file(
 		mlog_errno(status);
 		goto out;
 	}
-	handle = ocfs2_start_trans(OCFS2_SB(sb), 2);
+
+	/* Get buffer from the just added block */
+	down_read(&OCFS2_I(lqinode)->ip_alloc_sem);
+	status = ocfs2_extent_map_get_blocks(lqinode, oinfo->dqi_blocks,
+					     &p_blkno, NULL, NULL);
+	up_read(&OCFS2_I(lqinode)->ip_alloc_sem);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	bh = sb_getblk(sb, p_blkno);
+	if (!bh) {
+		status = -ENOMEM;
+		mlog_errno(status);
+		goto out;
+	}
+	ocfs2_set_new_buffer_uptodate(lqinode, bh);
+
+	/* Local quota info, chunk header and the new block we initialize */
+	handle = ocfs2_start_trans(OCFS2_SB(sb),
+			OCFS2_LOCAL_QINFO_WRITE_CREDITS +
+			2 * OCFS2_QUOTA_BLOCK_UPDATE_CREDITS);
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
 		mlog_errno(status);
 		goto out;
 	}
+	/* Zero created block */
+	status = ocfs2_journal_access_dq(handle, lqinode, bh,
+				 OCFS2_JOURNAL_ACCESS_CREATE);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	lock_buffer(bh);
+	memset(bh->b_data, 0, sb->s_blocksize);
+	unlock_buffer(bh);
+	status = ocfs2_journal_dirty(handle, bh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	/* Update chunk header */
 	status = ocfs2_journal_access_dq(handle, lqinode, chunk->qc_headerbh,
 				 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
@@ -1097,6 +1174,7 @@ static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file(
 		mlog_errno(status);
 		goto out_trans;
 	}
+	/* Update file header */
 	oinfo->dqi_blocks++;
 	status = ocfs2_local_write_info(sb, type);
 	if (status < 0) {
diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c
index 3f661376a2de..e49c41050264 100644
--- a/fs/ocfs2/stack_o2cb.c
+++ b/fs/ocfs2/stack_o2cb.c
@@ -17,6 +17,7 @@
  * General Public License for more details.
  */
 
+#include <linux/kernel.h>
 #include <linux/crc32.h>
 #include <linux/module.h>
 
@@ -153,7 +154,7 @@ static int status_map[] = {
 
 static int dlm_status_to_errno(enum dlm_status status)
 {
-	BUG_ON(status > (sizeof(status_map) / sizeof(status_map[0])));
+	BUG_ON(status < 0 || status >= ARRAY_SIZE(status_map));
 
 	return status_map[status];
 }
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 7efb349fb9bd..a3f8871d21fd 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -777,6 +777,7 @@ static int ocfs2_sb_probe(struct super_block *sb,
 		}
 		di = (struct ocfs2_dinode *) (*bh)->b_data;
 		memset(stats, 0, sizeof(struct ocfs2_blockcheck_stats));
+		spin_lock_init(&stats->b_lock);
 		status = ocfs2_verify_volume(di, *bh, blksize, stats);
 		if (status >= 0)
 			goto bail;
@@ -1182,7 +1183,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	wake_up(&osb->osb_mount_event);
 
 	/* Start this when the mount is almost sure of being successful */
-	ocfs2_orphan_scan_init(osb);
+	ocfs2_orphan_scan_start(osb);
 
 	mlog_exit(status);
 	return status;
@@ -1213,14 +1214,31 @@ static int ocfs2_get_sb(struct file_system_type *fs_type,
 			   mnt);
 }
 
+static void ocfs2_kill_sb(struct super_block *sb)
+{
+	struct ocfs2_super *osb = OCFS2_SB(sb);
+
+	/* Failed mount? */
+	if (!osb || atomic_read(&osb->vol_state) == VOLUME_DISABLED)
+		goto out;
+
+	/* Prevent further queueing of inode drop events */
+	spin_lock(&dentry_list_lock);
+	ocfs2_set_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED);
+	spin_unlock(&dentry_list_lock);
+	/* Wait for work to finish and/or remove it */
+	cancel_work_sync(&osb->dentry_lock_work);
+out:
+	kill_block_super(sb);
+}
+
 static struct file_system_type ocfs2_fs_type = {
 	.owner          = THIS_MODULE,
 	.name           = "ocfs2",
 	.get_sb         = ocfs2_get_sb, /* is this called when we mount
 					* the fs? */
-	.kill_sb        = kill_block_super, /* set to the generic one
-					     * right now, but do we
-					     * need to change that? */
+	.kill_sb        = ocfs2_kill_sb,
+
 	.fs_flags       = FS_REQUIRES_DEV|FS_RENAME_DOES_D_MOVE,
 	.next           = NULL
 };
@@ -1819,6 +1837,12 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	debugfs_remove(osb->osb_ctxt);
 
+	/*
+	 * Flush inode dropping work queue so that deletes are
+	 * performed while the filesystem is still working
+	 */
+	ocfs2_drop_all_dl_inodes(osb);
+
 	/* Orphan scan should be stopped as early as possible */
 	ocfs2_orphan_scan_stop(osb);
 
@@ -1981,6 +2005,8 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	snprintf(osb->dev_str, sizeof(osb->dev_str), "%u,%u",
 		 MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));
 
+	ocfs2_orphan_scan_init(osb);
+
 	status = ocfs2_recovery_init(osb);
 	if (status) {
 		mlog(ML_ERROR, "Unable to initialize recovery state\n");
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index ba320e250747..d1a27cda984f 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -1052,7 +1052,8 @@ static int ocfs2_xattr_block_get(struct inode *inode,
 	struct ocfs2_xattr_block *xb;
 	struct ocfs2_xattr_value_root *xv;
 	size_t size;
-	int ret = -ENODATA, name_offset, name_len, block_off, i;
+	int ret = -ENODATA, name_offset, name_len, i;
+	int uninitialized_var(block_off);
 
 	xs->bucket = ocfs2_xattr_bucket_new(inode);
 	if (!xs->bucket) {
diff --git a/fs/pipe.c b/fs/pipe.c
index f7dd21ad85a6..52c415114838 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -68,8 +68,8 @@ void pipe_double_lock(struct pipe_inode_info *pipe1,
 		pipe_lock_nested(pipe1, I_MUTEX_PARENT);
 		pipe_lock_nested(pipe2, I_MUTEX_CHILD);
 	} else {
-		pipe_lock_nested(pipe2, I_MUTEX_CHILD);
-		pipe_lock_nested(pipe1, I_MUTEX_PARENT);
+		pipe_lock_nested(pipe2, I_MUTEX_PARENT);
+		pipe_lock_nested(pipe1, I_MUTEX_CHILD);
 	}
 }
 
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 3ce5ae9e3d2d..6f742f6658a9 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -234,23 +234,20 @@ static int check_mem_permission(struct task_struct *task)
 
 struct mm_struct *mm_for_maps(struct task_struct *task)
 {
-	struct mm_struct *mm = get_task_mm(task);
-	if (!mm)
+	struct mm_struct *mm;
+
+	if (mutex_lock_killable(&task->cred_guard_mutex))
 		return NULL;
-	down_read(&mm->mmap_sem);
-	task_lock(task);
-	if (task->mm != mm)
-		goto out;
-	if (task->mm != current->mm &&
-	    __ptrace_may_access(task, PTRACE_MODE_READ) < 0)
-		goto out;
-	task_unlock(task);
+
+	mm = get_task_mm(task);
+	if (mm && mm != current->mm &&
+			!ptrace_may_access(task, PTRACE_MODE_READ)) {
+		mmput(mm);
+		mm = NULL;
+	}
+	mutex_unlock(&task->cred_guard_mutex);
+
 	return mm;
-out:
-	task_unlock(task);
-	up_read(&mm->mmap_sem);
-	mmput(mm);
-	return NULL;
 }
 
 static int proc_pid_cmdline(struct task_struct *task, char * buffer)
@@ -1006,12 +1003,7 @@ static ssize_t oom_adjust_read(struct file *file, char __user *buf,
 
 	if (!task)
 		return -ESRCH;
-	task_lock(task);
-	if (task->mm)
-		oom_adjust = task->mm->oom_adj;
-	else
-		oom_adjust = OOM_DISABLE;
-	task_unlock(task);
+	oom_adjust = task->oomkilladj;
 	put_task_struct(task);
 
 	len = snprintf(buffer, sizeof(buffer), "%i\n", oom_adjust);
@@ -1040,19 +1032,11 @@ static ssize_t oom_adjust_write(struct file *file, const char __user *buf,
 	task = get_proc_task(file->f_path.dentry->d_inode);
 	if (!task)
 		return -ESRCH;
-	task_lock(task);
-	if (!task->mm) {
-		task_unlock(task);
-		put_task_struct(task);
-		return -EINVAL;
-	}
-	if (oom_adjust < task->mm->oom_adj && !capable(CAP_SYS_RESOURCE)) {
-		task_unlock(task);
+	if (oom_adjust < task->oomkilladj && !capable(CAP_SYS_RESOURCE)) {
 		put_task_struct(task);
 		return -EACCES;
 	}
-	task->mm->oom_adj = oom_adjust;
-	task_unlock(task);
+	task->oomkilladj = oom_adjust;
 	put_task_struct(task);
 	if (end - buffer == 0)
 		return -EIO;
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 6f61b7cc32e0..9bd8be1d235c 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -119,6 +119,7 @@ static void *m_start(struct seq_file *m, loff_t *pos)
 	mm = mm_for_maps(priv->task);
 	if (!mm)
 		return NULL;
+	down_read(&mm->mmap_sem);
 
 	tail_vma = get_gate_vma(priv->task);
 	priv->tail_vma = tail_vma;
diff --git a/fs/proc/task_nommu.c b/fs/proc/task_nommu.c
index 64a72e2e7650..8f5c05d3dbd3 100644
--- a/fs/proc/task_nommu.c
+++ b/fs/proc/task_nommu.c
@@ -189,6 +189,7 @@ static void *m_start(struct seq_file *m, loff_t *pos)
 		priv->task = NULL;
 		return NULL;
 	}
+	down_read(&mm->mmap_sem);
 
 	/* start from the Nth VMA */
 	for (p = rb_first(&mm->mm_rb); p; p = rb_next(p))
diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c
index 70f36c043d62..38f7bd559f35 100644
--- a/fs/quota/dquot.c
+++ b/fs/quota/dquot.c
@@ -2043,7 +2043,6 @@ static int vfs_load_quota_inode(struct inode *inode, int type, int format_id,
 		invalidate_bdev(sb->s_bdev);
 	}
 	mutex_lock(&dqopt->dqonoff_mutex);
-	mutex_lock_nested(&inode->i_mutex, I_MUTEX_QUOTA);
 	if (sb_has_quota_loaded(sb, type)) {
 		error = -EBUSY;
 		goto out_lock;
@@ -2054,9 +2053,11 @@ static int vfs_load_quota_inode(struct inode *inode, int type, int format_id,
 		 * possible) Also nobody should write to the file - we use
 		 * special IO operations which ignore the immutable bit. */
 		down_write(&dqopt->dqptr_sem);
+		mutex_lock_nested(&inode->i_mutex, I_MUTEX_QUOTA);
 		oldflags = inode->i_flags & (S_NOATIME | S_IMMUTABLE |
 					     S_NOQUOTA);
 		inode->i_flags |= S_NOQUOTA | S_NOATIME | S_IMMUTABLE;
+		mutex_unlock(&inode->i_mutex);
 		up_write(&dqopt->dqptr_sem);
 		sb->dq_op->drop(inode);
 	}
@@ -2080,7 +2081,6 @@ static int vfs_load_quota_inode(struct inode *inode, int type, int format_id,
 		goto out_file_init;
 	}
 	mutex_unlock(&dqopt->dqio_mutex);
-	mutex_unlock(&inode->i_mutex);
 	spin_lock(&dq_state_lock);
 	dqopt->flags |= dquot_state_flag(flags, type);
 	spin_unlock(&dq_state_lock);
@@ -2096,13 +2096,14 @@ out_file_init:
 out_lock:
 	if (oldflags != -1) {
 		down_write(&dqopt->dqptr_sem);
+		mutex_lock_nested(&inode->i_mutex, I_MUTEX_QUOTA);
 		/* Set the flags back (in the case of accidental quotaon()
 		 * on a wrong file we don't want to mess up the flags) */
 		inode->i_flags &= ~(S_NOATIME | S_NOQUOTA | S_IMMUTABLE);
 		inode->i_flags |= oldflags;
+		mutex_unlock(&inode->i_mutex);
 		up_write(&dqopt->dqptr_sem);
 	}
-	mutex_unlock(&inode->i_mutex);
 	mutex_unlock(&dqopt->dqonoff_mutex);
 out_fmt:
 	put_quota_format(fmt);
diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c
index ebb2c417912c..11f0c06316de 100644
--- a/fs/ramfs/file-nommu.c
+++ b/fs/ramfs/file-nommu.c
@@ -20,6 +20,7 @@
 #include <linux/ramfs.h>
 #include <linux/pagevec.h>
 #include <linux/mman.h>
+#include <linux/sched.h>
 
 #include <asm/uaccess.h>
 #include "internal.h"
diff --git a/fs/select.c b/fs/select.c
index d870237e42c7..8084834e123e 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -110,6 +110,7 @@ void poll_initwait(struct poll_wqueues *pwq)
 {
 	init_poll_funcptr(&pwq->pt, __pollwait);
 	pwq->polling_task = current;
+	pwq->triggered = 0;
 	pwq->error = 0;
 	pwq->table = NULL;
 	pwq->inline_index = 0;
diff --git a/fs/sysfs/dir.c b/fs/sysfs/dir.c
index d88d0fac9fa5..14f2d71ea3ce 100644
--- a/fs/sysfs/dir.c
+++ b/fs/sysfs/dir.c
@@ -939,8 +939,10 @@ again:
 	/* Remove from old parent's list and insert into new parent's list. */
 	sysfs_unlink_sibling(sd);
 	sysfs_get(new_parent_sd);
+	drop_nlink(old_parent->d_inode);
 	sysfs_put(sd->s_parent);
 	sd->s_parent = new_parent_sd;
+	inc_nlink(new_parent->d_inode);
 	sysfs_link_sibling(sd);
 
  out_unlock:
diff --git a/fs/udf/super.c b/fs/udf/super.c
index 6832135159b6..9d1b8c2e6c45 100644
--- a/fs/udf/super.c
+++ b/fs/udf/super.c
@@ -1087,11 +1087,23 @@ static int udf_load_vat(struct super_block *sb, int p_index, int type1_index)
 	struct udf_inode_info *vati;
 	uint32_t pos;
 	struct virtualAllocationTable20 *vat20;
+	sector_t blocks = sb->s_bdev->bd_inode->i_size >> sb->s_blocksize_bits;
 
 	/* VAT file entry is in the last recorded block */
 	ino.partitionReferenceNum = type1_index;
 	ino.logicalBlockNum = sbi->s_last_block - map->s_partition_root;
 	sbi->s_vat_inode = udf_iget(sb, &ino);
+	if (!sbi->s_vat_inode &&
+	    sbi->s_last_block != blocks - 1) {
+		printk(KERN_NOTICE "UDF-fs: Failed to read VAT inode from the"
+		       " last recorded block (%lu), retrying with the last "
+		       "block of the device (%lu).\n",
+		       (unsigned long)sbi->s_last_block,
+		       (unsigned long)blocks - 1);
+		ino.partitionReferenceNum = type1_index;
+		ino.logicalBlockNum = blocks - 1 - map->s_partition_root;
+		sbi->s_vat_inode = udf_iget(sb, &ino);
+	}
 	if (!sbi->s_vat_inode)
 		return 1;
 
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index 7ec89fc05b2b..aecf2519db76 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -1268,6 +1268,14 @@ xfs_vm_writepage(
 	if (!page_has_buffers(page))
 		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
 
+
+	/*
+	 *  VM calculation for nr_to_write seems off.  Bump it way
+	 *  up, this gets simple streaming writes zippy again.
+	 *  To be reviewed again after Jens' writeback changes.
+	 */
+	wbc->nr_to_write *= 4;
+
 	/*
 	 * Convert delayed allocate, unwritten or unmapped space
 	 * to real space and flush out to disk.
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index 0c93c7ef3d18..965df1227d64 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -770,7 +770,7 @@ xfs_buf_associate_memory(
 	bp->b_pages = NULL;
 	bp->b_addr = mem;
 
-	rval = _xfs_buf_get_pages(bp, page_count, 0);
+	rval = _xfs_buf_get_pages(bp, page_count, XBF_DONT_BLOCK);
 	if (rval)
 		return rval;
 
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index 58973bb46038..8070b34cc287 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -680,8 +680,8 @@ xfs_vn_fiemap(
 	else
 		bm.bmv_length = BTOBB(length);
 
-	/* our formatter will tell xfs_getbmap when to stop. */
-	bm.bmv_count = MAXEXTNUM;
+	/* We add one because in getbmap world count includes the header */
+	bm.bmv_count = fieinfo->fi_extents_max + 1;
 	bm.bmv_iflags = BMV_IF_PREALLOC;
 	if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR)
 		bm.bmv_iflags |= BMV_IF_ATTRFORK;
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
index b619d6b8ca43..98ef624d9baf 100644
--- a/fs/xfs/linux-2.6/xfs_sync.c
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -708,6 +708,16 @@ xfs_reclaim_inode(
 	return 0;
 }
 
+void
+__xfs_inode_set_reclaim_tag(
+	struct xfs_perag	*pag,
+	struct xfs_inode	*ip)
+{
+	radix_tree_tag_set(&pag->pag_ici_root,
+			   XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino),
+			   XFS_ICI_RECLAIM_TAG);
+}
+
 /*
  * We set the inode flag atomically with the radix tree tag.
  * Once we get tag lookups on the radix tree, this inode flag
@@ -722,8 +732,7 @@ xfs_inode_set_reclaim_tag(
 
 	read_lock(&pag->pag_ici_lock);
 	spin_lock(&ip->i_flags_lock);
-	radix_tree_tag_set(&pag->pag_ici_root,
-			XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
+	__xfs_inode_set_reclaim_tag(pag, ip);
 	__xfs_iflags_set(ip, XFS_IRECLAIMABLE);
 	spin_unlock(&ip->i_flags_lock);
 	read_unlock(&pag->pag_ici_lock);
diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h
index 2a10301c99c7..59120602588a 100644
--- a/fs/xfs/linux-2.6/xfs_sync.h
+++ b/fs/xfs/linux-2.6/xfs_sync.h
@@ -48,6 +48,7 @@ int xfs_reclaim_inode(struct xfs_inode *ip, int locked, int sync_mode);
 int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
 
 void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
+void __xfs_inode_set_reclaim_tag(struct xfs_perag *pag, struct xfs_inode *ip);
 void xfs_inode_clear_reclaim_tag(struct xfs_inode *ip);
 void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, struct xfs_perag *pag,
 				struct xfs_inode *ip);
diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c
index db15feb906ff..4ece1906bd41 100644
--- a/fs/xfs/xfs_attr.c
+++ b/fs/xfs/xfs_attr.c
@@ -2010,7 +2010,9 @@ xfs_attr_rmtval_get(xfs_da_args_t *args)
 			dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
 			blkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
 			error = xfs_read_buf(mp, mp->m_ddev_targp, dblkno,
-					     blkcnt, XFS_BUF_LOCK, &bp);
+					     blkcnt,
+					     XFS_BUF_LOCK | XBF_DONT_BLOCK,
+					     &bp);
 			if (error)
 				return(error);
 
@@ -2141,8 +2143,8 @@ xfs_attr_rmtval_set(xfs_da_args_t *args)
 		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
 		blkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
 
-		bp = xfs_buf_get_flags(mp->m_ddev_targp, dblkno,
-							blkcnt, XFS_BUF_LOCK);
+		bp = xfs_buf_get_flags(mp->m_ddev_targp, dblkno, blkcnt,
+				       XFS_BUF_LOCK | XBF_DONT_BLOCK);
 		ASSERT(bp);
 		ASSERT(!XFS_BUF_GETERROR(bp));
 
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 7928b9983c1d..8ee5b5a76a2a 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -6009,7 +6009,7 @@ xfs_getbmap(
 	 */
 	error = ENOMEM;
 	subnex = 16;
-	map = kmem_alloc(subnex * sizeof(*map), KM_MAYFAIL);
+	map = kmem_alloc(subnex * sizeof(*map), KM_MAYFAIL | KM_NOFS);
 	if (!map)
 		goto out_unlock_ilock;
 
diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c
index e9df99574829..26717388acf5 100644
--- a/fs/xfs/xfs_btree.c
+++ b/fs/xfs/xfs_btree.c
@@ -120,8 +120,8 @@ xfs_btree_check_sblock(
 			XFS_RANDOM_BTREE_CHECK_SBLOCK))) {
 		if (bp)
 			xfs_buftrace("SBTREE ERROR", bp);
-		XFS_ERROR_REPORT("xfs_btree_check_sblock", XFS_ERRLEVEL_LOW,
-				 cur->bc_mp);
+		XFS_CORRUPTION_ERROR("xfs_btree_check_sblock",
+			XFS_ERRLEVEL_LOW, cur->bc_mp, block);
 		return XFS_ERROR(EFSCORRUPTED);
 	}
 	return 0;
diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c
index 9ff6e57a5075..2847bbc1c534 100644
--- a/fs/xfs/xfs_da_btree.c
+++ b/fs/xfs/xfs_da_btree.c
@@ -2201,7 +2201,7 @@ kmem_zone_t *xfs_dabuf_zone;		/* dabuf zone */
 xfs_da_state_t *
 xfs_da_state_alloc(void)
 {
-	return kmem_zone_zalloc(xfs_da_state_zone, KM_SLEEP);
+	return kmem_zone_zalloc(xfs_da_state_zone, KM_NOFS);
 }
 
 /*
@@ -2261,9 +2261,9 @@ xfs_da_buf_make(int nbuf, xfs_buf_t **bps, inst_t *ra)
 	int		off;
 
 	if (nbuf == 1)
-		dabuf = kmem_zone_alloc(xfs_dabuf_zone, KM_SLEEP);
+		dabuf = kmem_zone_alloc(xfs_dabuf_zone, KM_NOFS);
 	else
-		dabuf = kmem_alloc(XFS_DA_BUF_SIZE(nbuf), KM_SLEEP);
+		dabuf = kmem_alloc(XFS_DA_BUF_SIZE(nbuf), KM_NOFS);
 	dabuf->dirty = 0;
 #ifdef XFS_DABUF_DEBUG
 	dabuf->ra = ra;
diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c
index c657bec6d951..bb1d58eb3982 100644
--- a/fs/xfs/xfs_dir2.c
+++ b/fs/xfs/xfs_dir2.c
@@ -256,7 +256,7 @@ xfs_dir_cilookup_result(
 					!(args->op_flags & XFS_DA_OP_CILOOKUP))
 		return EEXIST;
 
-	args->value = kmem_alloc(len, KM_MAYFAIL);
+	args->value = kmem_alloc(len, KM_NOFS | KM_MAYFAIL);
 	if (!args->value)
 		return ENOMEM;
 
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index cbd451bb4848..2d0b3e1da9e6 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -167,17 +167,25 @@ xfs_growfs_data_private(
 	new = nb - mp->m_sb.sb_dblocks;
 	oagcount = mp->m_sb.sb_agcount;
 	if (nagcount > oagcount) {
+		void *new_perag, *old_perag;
+
 		xfs_filestream_flush(mp);
+
+		new_perag = kmem_zalloc(sizeof(xfs_perag_t) * nagcount,
+					KM_MAYFAIL);
+		if (!new_perag)
+			return XFS_ERROR(ENOMEM);
+
 		down_write(&mp->m_peraglock);
-		mp->m_perag = kmem_realloc(mp->m_perag,
-			sizeof(xfs_perag_t) * nagcount,
-			sizeof(xfs_perag_t) * oagcount,
-			KM_SLEEP);
-		memset(&mp->m_perag[oagcount], 0,
-			(nagcount - oagcount) * sizeof(xfs_perag_t));
+		memcpy(new_perag, mp->m_perag, sizeof(xfs_perag_t) * oagcount);
+		old_perag = mp->m_perag;
+		mp->m_perag = new_perag;
+
 		mp->m_flags |= XFS_MOUNT_32BITINODES;
 		nagimax = xfs_initialize_perag(mp, nagcount);
 		up_write(&mp->m_peraglock);
+
+		kmem_free(old_perag);
 	}
 	tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFS);
 	tp->t_flags |= XFS_TRANS_RESERVE;
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index 5fcec6f020a7..ecbf8b4d2e2e 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -64,6 +64,10 @@ xfs_inode_alloc(
 	ip = kmem_zone_alloc(xfs_inode_zone, KM_SLEEP);
 	if (!ip)
 		return NULL;
+	if (inode_init_always(mp->m_super, VFS_I(ip))) {
+		kmem_zone_free(xfs_inode_zone, ip);
+		return NULL;
+	}
 
 	ASSERT(atomic_read(&ip->i_iocount) == 0);
 	ASSERT(atomic_read(&ip->i_pincount) == 0);
@@ -105,17 +109,6 @@ xfs_inode_alloc(
 #ifdef XFS_DIR2_TRACE
 	ip->i_dir_trace = ktrace_alloc(XFS_DIR2_KTRACE_SIZE, KM_NOFS);
 #endif
-	/*
-	* Now initialise the VFS inode. We do this after the xfs_inode
-	* initialisation as internal failures will result in ->destroy_inode
-	* being called and that will pass down through the reclaim path and
-	* free the XFS inode. This path requires the XFS inode to already be
-	* initialised. Hence if this call fails, the xfs_inode has already
-	* been freed and we should not reference it at all in the error
-	* handling.
-	*/
-	if (!inode_init_always(mp->m_super, VFS_I(ip)))
-		return NULL;
 
 	/* prevent anyone from using this yet */
 	VFS_I(ip)->i_state = I_NEW|I_LOCK;
@@ -123,6 +116,71 @@ xfs_inode_alloc(
 	return ip;
 }
 
+STATIC void
+xfs_inode_free(
+	struct xfs_inode	*ip)
+{
+	switch (ip->i_d.di_mode & S_IFMT) {
+	case S_IFREG:
+	case S_IFDIR:
+	case S_IFLNK:
+		xfs_idestroy_fork(ip, XFS_DATA_FORK);
+		break;
+	}
+
+	if (ip->i_afp)
+		xfs_idestroy_fork(ip, XFS_ATTR_FORK);
+
+#ifdef XFS_INODE_TRACE
+	ktrace_free(ip->i_trace);
+#endif
+#ifdef XFS_BMAP_TRACE
+	ktrace_free(ip->i_xtrace);
+#endif
+#ifdef XFS_BTREE_TRACE
+	ktrace_free(ip->i_btrace);
+#endif
+#ifdef XFS_RW_TRACE
+	ktrace_free(ip->i_rwtrace);
+#endif
+#ifdef XFS_ILOCK_TRACE
+	ktrace_free(ip->i_lock_trace);
+#endif
+#ifdef XFS_DIR2_TRACE
+	ktrace_free(ip->i_dir_trace);
+#endif
+
+	if (ip->i_itemp) {
+		/*
+		 * Only if we are shutting down the fs will we see an
+		 * inode still in the AIL. If it is there, we should remove
+		 * it to prevent a use-after-free from occurring.
+		 */
+		xfs_log_item_t	*lip = &ip->i_itemp->ili_item;
+		struct xfs_ail	*ailp = lip->li_ailp;
+
+		ASSERT(((lip->li_flags & XFS_LI_IN_AIL) == 0) ||
+				       XFS_FORCED_SHUTDOWN(ip->i_mount));
+		if (lip->li_flags & XFS_LI_IN_AIL) {
+			spin_lock(&ailp->xa_lock);
+			if (lip->li_flags & XFS_LI_IN_AIL)
+				xfs_trans_ail_delete(ailp, lip);
+			else
+				spin_unlock(&ailp->xa_lock);
+		}
+		xfs_inode_item_destroy(ip);
+		ip->i_itemp = NULL;
+	}
+
+	/* asserts to verify all state is correct here */
+	ASSERT(atomic_read(&ip->i_iocount) == 0);
+	ASSERT(atomic_read(&ip->i_pincount) == 0);
+	ASSERT(!spin_is_locked(&ip->i_flags_lock));
+	ASSERT(completion_done(&ip->i_flush));
+
+	kmem_zone_free(xfs_inode_zone, ip);
+}
+
 /*
  * Check the validity of the inode we just found it the cache
  */
@@ -133,80 +191,82 @@ xfs_iget_cache_hit(
 	int			flags,
 	int			lock_flags) __releases(pag->pag_ici_lock)
 {
+	struct inode		*inode = VFS_I(ip);
 	struct xfs_mount	*mp = ip->i_mount;
-	int			error = EAGAIN;
+	int			error;
+
+	spin_lock(&ip->i_flags_lock);
 
 	/*
-	 * If INEW is set this inode is being set up
-	 * If IRECLAIM is set this inode is being torn down
-	 * Pause and try again.
+	 * If we are racing with another cache hit that is currently
+	 * instantiating this inode or currently recycling it out of
+	 * reclaimabe state, wait for the initialisation to complete
+	 * before continuing.
+	 *
+	 * XXX(hch): eventually we should do something equivalent to
+	 *	     wait_on_inode to wait for these flags to be cleared
+	 *	     instead of polling for it.
 	 */
-	if (xfs_iflags_test(ip, (XFS_INEW|XFS_IRECLAIM))) {
+	if (ip->i_flags & (XFS_INEW|XFS_IRECLAIM)) {
 		XFS_STATS_INC(xs_ig_frecycle);
+		error = EAGAIN;
 		goto out_error;
 	}
 
-	/* If IRECLAIMABLE is set, we've torn down the vfs inode part */
-	if (xfs_iflags_test(ip, XFS_IRECLAIMABLE)) {
-
-		/*
-		 * If lookup is racing with unlink, then we should return an
-		 * error immediately so we don't remove it from the reclaim
-		 * list and potentially leak the inode.
-		 */
-		if ((ip->i_d.di_mode == 0) && !(flags & XFS_IGET_CREATE)) {
-			error = ENOENT;
-			goto out_error;
-		}
+	/*
+	 * If lookup is racing with unlink return an error immediately.
+	 */
+	if (ip->i_d.di_mode == 0 && !(flags & XFS_IGET_CREATE)) {
+		error = ENOENT;
+		goto out_error;
+	}
 
+	/*
+	 * If IRECLAIMABLE is set, we've torn down the VFS inode already.
+	 * Need to carefully get it back into useable state.
+	 */
+	if (ip->i_flags & XFS_IRECLAIMABLE) {
 		xfs_itrace_exit_tag(ip, "xfs_iget.alloc");
 
 		/*
-		 * We need to re-initialise the VFS inode as it has been
-		 * 'freed' by the VFS. Do this here so we can deal with
-		 * errors cleanly, then tag it so it can be set up correctly
-		 * later.
+		 * We need to set XFS_INEW atomically with clearing the
+		 * reclaimable tag so that we do have an indicator of the
+		 * inode still being initialized.
 		 */
-		if (!inode_init_always(mp->m_super, VFS_I(ip))) {
-			error = ENOMEM;
-			goto out_error;
-		}
+		ip->i_flags |= XFS_INEW;
+		ip->i_flags &= ~XFS_IRECLAIMABLE;
+		__xfs_inode_clear_reclaim_tag(mp, pag, ip);
 
-		/*
-		 * We must set the XFS_INEW flag before clearing the
-		 * XFS_IRECLAIMABLE flag so that if a racing lookup does
-		 * not find the XFS_IRECLAIMABLE above but has the igrab()
-		 * below succeed we can safely check XFS_INEW to detect
-		 * that this inode is still being initialised.
-		 */
-		xfs_iflags_set(ip, XFS_INEW);
-		xfs_iflags_clear(ip, XFS_IRECLAIMABLE);
+		spin_unlock(&ip->i_flags_lock);
+		read_unlock(&pag->pag_ici_lock);
 
-		/* clear the radix tree reclaim flag as well. */
-		__xfs_inode_clear_reclaim_tag(mp, pag, ip);
-	} else if (!igrab(VFS_I(ip))) {
+		error = -inode_init_always(mp->m_super, inode);
+		if (error) {
+			/*
+			 * Re-initializing the inode failed, and we are in deep
+			 * trouble.  Try to re-add it to the reclaim list.
+			 */
+			read_lock(&pag->pag_ici_lock);
+			spin_lock(&ip->i_flags_lock);
+
+			ip->i_flags &= ~XFS_INEW;
+			ip->i_flags |= XFS_IRECLAIMABLE;
+			__xfs_inode_set_reclaim_tag(pag, ip);
+			goto out_error;
+		}
+		inode->i_state = I_LOCK|I_NEW;
+	} else {
 		/* If the VFS inode is being torn down, pause and try again. */
-		XFS_STATS_INC(xs_ig_frecycle);
-		goto out_error;
-	} else if (xfs_iflags_test(ip, XFS_INEW)) {
-		/*
-		 * We are racing with another cache hit that is
-		 * currently recycling this inode out of the XFS_IRECLAIMABLE
-		 * state. Wait for the initialisation to complete before
-		 * continuing.
-		 */
-		wait_on_inode(VFS_I(ip));
-	}
+		if (!igrab(inode)) {
+			error = EAGAIN;
+			goto out_error;
+		}
 
-	if (ip->i_d.di_mode == 0 && !(flags & XFS_IGET_CREATE)) {
-		error = ENOENT;
-		iput(VFS_I(ip));
-		goto out_error;
+		/* We've got a live one. */
+		spin_unlock(&ip->i_flags_lock);
+		read_unlock(&pag->pag_ici_lock);
 	}
 
-	/* We've got a live one. */
-	read_unlock(&pag->pag_ici_lock);
-
 	if (lock_flags != 0)
 		xfs_ilock(ip, lock_flags);
 
@@ -216,6 +276,7 @@ xfs_iget_cache_hit(
 	return 0;
 
 out_error:
+	spin_unlock(&ip->i_flags_lock);
 	read_unlock(&pag->pag_ici_lock);
 	return error;
 }
@@ -299,7 +360,8 @@ out_preload_end:
 	if (lock_flags)
 		xfs_iunlock(ip, lock_flags);
 out_destroy:
-	xfs_destroy_inode(ip);
+	__destroy_inode(VFS_I(ip));
+	xfs_inode_free(ip);
 	return error;
 }
 
@@ -504,62 +566,7 @@ xfs_ireclaim(
 	xfs_qm_dqdetach(ip);
 	xfs_iunlock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
 
-	switch (ip->i_d.di_mode & S_IFMT) {
-	case S_IFREG:
-	case S_IFDIR:
-	case S_IFLNK:
-		xfs_idestroy_fork(ip, XFS_DATA_FORK);
-		break;
-	}
-
-	if (ip->i_afp)
-		xfs_idestroy_fork(ip, XFS_ATTR_FORK);
-
-#ifdef XFS_INODE_TRACE
-	ktrace_free(ip->i_trace);
-#endif
-#ifdef XFS_BMAP_TRACE
-	ktrace_free(ip->i_xtrace);
-#endif
-#ifdef XFS_BTREE_TRACE
-	ktrace_free(ip->i_btrace);
-#endif
-#ifdef XFS_RW_TRACE
-	ktrace_free(ip->i_rwtrace);
-#endif
-#ifdef XFS_ILOCK_TRACE
-	ktrace_free(ip->i_lock_trace);
-#endif
-#ifdef XFS_DIR2_TRACE
-	ktrace_free(ip->i_dir_trace);
-#endif
-	if (ip->i_itemp) {
-		/*
-		 * Only if we are shutting down the fs will we see an
-		 * inode still in the AIL. If it is there, we should remove
-		 * it to prevent a use-after-free from occurring.
-		 */
-		xfs_log_item_t	*lip = &ip->i_itemp->ili_item;
-		struct xfs_ail	*ailp = lip->li_ailp;
-
-		ASSERT(((lip->li_flags & XFS_LI_IN_AIL) == 0) ||
-				       XFS_FORCED_SHUTDOWN(ip->i_mount));
-		if (lip->li_flags & XFS_LI_IN_AIL) {
-			spin_lock(&ailp->xa_lock);
-			if (lip->li_flags & XFS_LI_IN_AIL)
-				xfs_trans_ail_delete(ailp, lip);
-			else
-				spin_unlock(&ailp->xa_lock);
-		}
-		xfs_inode_item_destroy(ip);
-		ip->i_itemp = NULL;
-	}
-	/* asserts to verify all state is correct here */
-	ASSERT(atomic_read(&ip->i_iocount) == 0);
-	ASSERT(atomic_read(&ip->i_pincount) == 0);
-	ASSERT(!spin_is_locked(&ip->i_flags_lock));
-	ASSERT(completion_done(&ip->i_flush));
-	kmem_zone_free(xfs_inode_zone, ip);
+	xfs_inode_free(ip);
 }
 
 /*
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 1f22d65fed0a..da428b3fe0f5 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -343,6 +343,16 @@ xfs_iformat(
 		return XFS_ERROR(EFSCORRUPTED);
 	}
 
+	if (unlikely((ip->i_d.di_flags & XFS_DIFLAG_REALTIME) &&
+		     !ip->i_mount->m_rtdev_targp)) {
+		xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
+			"corrupt dinode %Lu, has realtime flag set.",
+			ip->i_ino);
+		XFS_CORRUPTION_ERROR("xfs_iformat(realtime)",
+				     XFS_ERRLEVEL_LOW, ip->i_mount, dip);
+		return XFS_ERROR(EFSCORRUPTED);
+	}
+
 	switch (ip->i_d.di_mode & S_IFMT) {
 	case S_IFIFO:
 	case S_IFCHR:
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1804f866a71d..65f24a3cc992 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -310,23 +310,6 @@ static inline struct inode *VFS_I(struct xfs_inode *ip)
 }
 
 /*
- * Get rid of a partially initialized inode.
- *
- * We have to go through destroy_inode to make sure allocations
- * from init_inode_always like the security data are undone.
- *
- * We mark the inode bad so that it takes the short cut in
- * the reclaim path instead of going through the flush path
- * which doesn't make sense for an inode that has never seen the
- * light of day.
- */
-static inline void xfs_destroy_inode(struct xfs_inode *ip)
-{
-	make_bad_inode(VFS_I(ip));
-	return destroy_inode(VFS_I(ip));
-}
-
-/*
  * i_flags helper functions
  */
 static inline void
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index 3750f04ede0b..9dbdff3ea484 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -3180,7 +3180,7 @@ try_again:
 STATIC void
 xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog)
 {
-	ASSERT(spin_is_locked(&log->l_icloglock));
+	assert_spin_locked(&log->l_icloglock);
 
 	if (iclog->ic_state == XLOG_STATE_ACTIVE) {
 		xlog_state_switch_iclogs(log, iclog, 0);
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index c4eca5ed5dab..492d75bae2bf 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -538,7 +538,9 @@ xfs_readlink_bmap(
 		d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
 		byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
 
-		bp = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0);
+		bp = xfs_buf_read_flags(mp->m_ddev_targp, d, BTOBB(byte_cnt),
+					XBF_LOCK | XBF_MAPPED |
+					XBF_DONT_BLOCK);
 		error = XFS_BUF_GETERROR(bp);
 		if (error) {
 			xfs_ioerror_alert("xfs_readlink",