Merge branch 'master' into for-linus

Conflicts:
	fs/pipe.c

Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

275 files changed, 11458 insertions, 6190 deletions

diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h
index ed835836e0dc..32ef4009d030 100644
--- a/fs/9p/v9fs_vfs.h
+++ b/fs/9p/v9fs_vfs.h
@@ -40,7 +40,9 @@
 extern struct file_system_type v9fs_fs_type;
 extern const struct address_space_operations v9fs_addr_operations;
 extern const struct file_operations v9fs_file_operations;
+extern const struct file_operations v9fs_file_operations_dotl;
 extern const struct file_operations v9fs_dir_operations;
+extern const struct file_operations v9fs_dir_operations_dotl;
 extern const struct dentry_operations v9fs_dentry_operations;
 extern const struct dentry_operations v9fs_cached_dentry_operations;
 
diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c
index 0adfd64dfcee..d61e3b28ce37 100644
--- a/fs/9p/vfs_dir.c
+++ b/fs/9p/vfs_dir.c
@@ -203,3 +203,11 @@ const struct file_operations v9fs_dir_operations = {
 	.open = v9fs_file_open,
 	.release = v9fs_dir_release,
 };
+
+const struct file_operations v9fs_dir_operations_dotl = {
+	.read = generic_read_dir,
+	.llseek = generic_file_llseek,
+	.readdir = v9fs_dir_readdir,
+	.open = v9fs_file_open,
+	.release = v9fs_dir_release,
+};
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index df52d488d2a6..2bedc6c94fc2 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -257,15 +257,13 @@ v9fs_file_write(struct file *filp, const char __user * data,
 	return total;
 }
 
-static int v9fs_file_fsync(struct file *filp, struct dentry *dentry,
-					int datasync)
+static int v9fs_file_fsync(struct file *filp, int datasync)
 {
 	struct p9_fid *fid;
 	struct p9_wstat wstat;
 	int retval;
 
-	P9_DPRINTK(P9_DEBUG_VFS, "filp %p dentry %p datasync %x\n", filp,
-						dentry, datasync);
+	P9_DPRINTK(P9_DEBUG_VFS, "filp %p datasync %x\n", filp, datasync);
 
 	fid = filp->private_data;
 	v9fs_blank_wstat(&wstat);
@@ -296,3 +294,14 @@ const struct file_operations v9fs_file_operations = {
 	.mmap = generic_file_readonly_mmap,
 	.fsync = v9fs_file_fsync,
 };
+
+const struct file_operations v9fs_file_operations_dotl = {
+	.llseek = generic_file_llseek,
+	.read = v9fs_file_read,
+	.write = v9fs_file_write,
+	.open = v9fs_file_open,
+	.release = v9fs_dir_release,
+	.lock = v9fs_file_lock,
+	.mmap = generic_file_readonly_mmap,
+	.fsync = v9fs_file_fsync,
+};
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index 6d4d86187c55..4331b3b5ee1c 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -44,9 +44,12 @@
 #include "cache.h"
 
 static const struct inode_operations v9fs_dir_inode_operations;
-static const struct inode_operations v9fs_dir_inode_operations_ext;
+static const struct inode_operations v9fs_dir_inode_operations_dotu;
+static const struct inode_operations v9fs_dir_inode_operations_dotl;
 static const struct inode_operations v9fs_file_inode_operations;
+static const struct inode_operations v9fs_file_inode_operations_dotl;
 static const struct inode_operations v9fs_symlink_inode_operations;
+static const struct inode_operations v9fs_symlink_inode_operations_dotl;
 
 /**
  * unixmode2p9mode - convert unix mode bits to plan 9
@@ -273,25 +276,44 @@ struct inode *v9fs_get_inode(struct super_block *sb, int mode)
 		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;
+		if (v9fs_proto_dotl(v9ses)) {
+			inode->i_op = &v9fs_file_inode_operations_dotl;
+			inode->i_fop = &v9fs_file_operations_dotl;
+		} else {
+			inode->i_op = &v9fs_file_inode_operations;
+			inode->i_fop = &v9fs_file_operations;
+		}
+
 		break;
+
 	case S_IFLNK:
-		if (!v9fs_proto_dotu(v9ses)) {
-			P9_DPRINTK(P9_DEBUG_ERROR,
-				   "extended modes used w/o 9P2000.u\n");
+		if (!v9fs_proto_dotu(v9ses) && !v9fs_proto_dotl(v9ses)) {
+			P9_DPRINTK(P9_DEBUG_ERROR, "extended modes used with "
+						"legacy protocol.\n");
 			err = -EINVAL;
 			goto error;
 		}
-		inode->i_op = &v9fs_symlink_inode_operations;
+
+		if (v9fs_proto_dotl(v9ses))
+			inode->i_op = &v9fs_symlink_inode_operations_dotl;
+		else
+			inode->i_op = &v9fs_symlink_inode_operations;
+
 		break;
 	case S_IFDIR:
 		inc_nlink(inode);
-		if (v9fs_proto_dotu(v9ses))
-			inode->i_op = &v9fs_dir_inode_operations_ext;
+		if (v9fs_proto_dotl(v9ses))
+			inode->i_op = &v9fs_dir_inode_operations_dotl;
+		else if (v9fs_proto_dotu(v9ses))
+			inode->i_op = &v9fs_dir_inode_operations_dotu;
 		else
 			inode->i_op = &v9fs_dir_inode_operations;
-		inode->i_fop = &v9fs_dir_operations;
+
+		if (v9fs_proto_dotl(v9ses))
+			inode->i_fop = &v9fs_dir_operations_dotl;
+		else
+			inode->i_fop = &v9fs_dir_operations;
+
 		break;
 	default:
 		P9_DPRINTK(P9_DEBUG_ERROR, "BAD mode 0x%x S_IFMT 0x%x\n",
@@ -432,14 +454,12 @@ static int v9fs_remove(struct inode *dir, struct dentry *file, int rmdir)
 {
 	int retval;
 	struct inode *file_inode;
-	struct v9fs_session_info *v9ses;
 	struct p9_fid *v9fid;
 
 	P9_DPRINTK(P9_DEBUG_VFS, "inode: %p dentry: %p rmdir: %d\n", dir, file,
 		rmdir);
 
 	file_inode = file->d_inode;
-	v9ses = v9fs_inode2v9ses(file_inode);
 	v9fid = v9fs_fid_clone(file);
 	if (IS_ERR(v9fid))
 		return PTR_ERR(v9fid);
@@ -482,12 +502,11 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 	ofid = NULL;
 	fid = NULL;
 	name = (char *) dentry->d_name.name;
-	dfid = v9fs_fid_clone(dentry->d_parent);
+	dfid = v9fs_fid_lookup(dentry->d_parent);
 	if (IS_ERR(dfid)) {
 		err = PTR_ERR(dfid);
-		P9_DPRINTK(P9_DEBUG_VFS, "fid clone failed %d\n", err);
-		dfid = NULL;
-		goto error;
+		P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
+		return ERR_PTR(err);
 	}
 
 	/* clone a fid to use for creation */
@@ -495,8 +514,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 	if (IS_ERR(ofid)) {
 		err = PTR_ERR(ofid);
 		P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
-		ofid = NULL;
-		goto error;
+		return ERR_PTR(err);
 	}
 
 	err = p9_client_fcreate(ofid, name, perm, mode, extension);
@@ -506,14 +524,13 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 	}
 
 	/* now walk from the parent so we can get unopened fid */
-	fid = p9_client_walk(dfid, 1, &name, 0);
+	fid = p9_client_walk(dfid, 1, &name, 1);
 	if (IS_ERR(fid)) {
 		err = PTR_ERR(fid);
 		P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
 		fid = NULL;
 		goto error;
-	} else
-		dfid = NULL;
+	}
 
 	/* instantiate inode and assign the unopened fid to the dentry */
 	inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
@@ -536,9 +553,6 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 	return ofid;
 
 error:
-	if (dfid)
-		p9_client_clunk(dfid);
-
 	if (ofid)
 		p9_client_clunk(ofid);
 
@@ -673,8 +687,8 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 	if (IS_ERR(fid)) {
 		result = PTR_ERR(fid);
 		if (result == -ENOENT) {
-			d_add(dentry, NULL);
-			return NULL;
+			inode = NULL;
+			goto inst_out;
 		}
 
 		return ERR_PTR(result);
@@ -691,7 +705,8 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 	if (result < 0)
 		goto error;
 
-	if ((fid->qid.version) && (v9ses->cache))
+inst_out:
+	if (v9ses->cache)
 		dentry->d_op = &v9fs_cached_dentry_operations;
 	else
 		dentry->d_op = &v9fs_dentry_operations;
@@ -770,6 +785,13 @@ v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		goto clunk_olddir;
 	}
 
+	if (v9fs_proto_dotl(v9ses)) {
+		retval = p9_client_rename(oldfid, newdirfid,
+					(char *) new_dentry->d_name.name);
+		if (retval != -ENOSYS)
+			goto clunk_newdir;
+	}
+
 	/* 9P can only handle file rename in the same directory */
 	if (memcmp(&olddirfid->qid, &newdirfid->qid, sizeof(newdirfid->qid))) {
 		P9_DPRINTK(P9_DEBUG_ERROR,
@@ -1195,6 +1217,8 @@ v9fs_vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
 		sprintf(name, "c %u %u", MAJOR(rdev), MINOR(rdev));
 	else if (S_ISFIFO(mode))
 		*name = 0;
+	else if (S_ISSOCK(mode))
+		*name = 0;
 	else {
 		__putname(name);
 		return -EINVAL;
@@ -1206,7 +1230,21 @@ v9fs_vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
 	return retval;
 }
 
-static const struct inode_operations v9fs_dir_inode_operations_ext = {
+static const struct inode_operations v9fs_dir_inode_operations_dotu = {
+	.create = v9fs_vfs_create,
+	.lookup = v9fs_vfs_lookup,
+	.symlink = v9fs_vfs_symlink,
+	.link = v9fs_vfs_link,
+	.unlink = v9fs_vfs_unlink,
+	.mkdir = v9fs_vfs_mkdir,
+	.rmdir = v9fs_vfs_rmdir,
+	.mknod = v9fs_vfs_mknod,
+	.rename = v9fs_vfs_rename,
+	.getattr = v9fs_vfs_getattr,
+	.setattr = v9fs_vfs_setattr,
+};
+
+static const struct inode_operations v9fs_dir_inode_operations_dotl = {
 	.create = v9fs_vfs_create,
 	.lookup = v9fs_vfs_lookup,
 	.symlink = v9fs_vfs_symlink,
@@ -1237,6 +1275,11 @@ static const struct inode_operations v9fs_file_inode_operations = {
 	.setattr = v9fs_vfs_setattr,
 };
 
+static const struct inode_operations v9fs_file_inode_operations_dotl = {
+	.getattr = v9fs_vfs_getattr,
+	.setattr = v9fs_vfs_setattr,
+};
+
 static const struct inode_operations v9fs_symlink_inode_operations = {
 	.readlink = generic_readlink,
 	.follow_link = v9fs_vfs_follow_link,
@@ -1244,3 +1287,11 @@ static const struct inode_operations v9fs_symlink_inode_operations = {
 	.getattr = v9fs_vfs_getattr,
 	.setattr = v9fs_vfs_setattr,
 };
+
+static const struct inode_operations v9fs_symlink_inode_operations_dotl = {
+	.readlink = generic_readlink,
+	.follow_link = v9fs_vfs_follow_link,
+	.put_link = v9fs_vfs_put_link,
+	.getattr = v9fs_vfs_getattr,
+	.setattr = v9fs_vfs_setattr,
+};
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index 806da5d3b3a0..be74d020436e 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -38,6 +38,7 @@
 #include <linux/idr.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
+#include <linux/statfs.h>
 #include <net/9p/9p.h>
 #include <net/9p/client.h>
 
@@ -45,7 +46,7 @@
 #include "v9fs_vfs.h"
 #include "fid.h"
 
-static const struct super_operations v9fs_super_ops;
+static const struct super_operations v9fs_super_ops, v9fs_super_ops_dotl;
 
 /**
  * v9fs_set_super - set the superblock
@@ -76,7 +77,10 @@ v9fs_fill_super(struct super_block *sb, struct v9fs_session_info *v9ses,
 	sb->s_blocksize_bits = fls(v9ses->maxdata - 1);
 	sb->s_blocksize = 1 << sb->s_blocksize_bits;
 	sb->s_magic = V9FS_MAGIC;
-	sb->s_op = &v9fs_super_ops;
+	if (v9fs_proto_dotl(v9ses))
+		sb->s_op = &v9fs_super_ops_dotl;
+	else
+		sb->s_op = &v9fs_super_ops;
 	sb->s_bdi = &v9ses->bdi;
 
 	sb->s_flags = flags | MS_ACTIVE | MS_SYNCHRONOUS | MS_DIRSYNC |
@@ -211,6 +215,42 @@ v9fs_umount_begin(struct super_block *sb)
 	v9fs_session_begin_cancel(v9ses);
 }
 
+static int v9fs_statfs(struct dentry *dentry, struct kstatfs *buf)
+{
+	struct v9fs_session_info *v9ses;
+	struct p9_fid *fid;
+	struct p9_rstatfs rs;
+	int res;
+
+	fid = v9fs_fid_lookup(dentry);
+	if (IS_ERR(fid)) {
+		res = PTR_ERR(fid);
+		goto done;
+	}
+
+	v9ses = v9fs_inode2v9ses(dentry->d_inode);
+	if (v9fs_proto_dotl(v9ses)) {
+		res = p9_client_statfs(fid, &rs);
+		if (res == 0) {
+			buf->f_type = rs.type;
+			buf->f_bsize = rs.bsize;
+			buf->f_blocks = rs.blocks;
+			buf->f_bfree = rs.bfree;
+			buf->f_bavail = rs.bavail;
+			buf->f_files = rs.files;
+			buf->f_ffree = rs.ffree;
+			buf->f_fsid.val[0] = rs.fsid & 0xFFFFFFFFUL;
+			buf->f_fsid.val[1] = (rs.fsid >> 32) & 0xFFFFFFFFUL;
+			buf->f_namelen = rs.namelen;
+		}
+		if (res != -ENOSYS)
+			goto done;
+	}
+	res = simple_statfs(dentry, buf);
+done:
+	return res;
+}
+
 static const struct super_operations v9fs_super_ops = {
 #ifdef CONFIG_9P_FSCACHE
 	.alloc_inode = v9fs_alloc_inode,
@@ -222,6 +262,17 @@ static const struct super_operations v9fs_super_ops = {
 	.umount_begin = v9fs_umount_begin,
 };
 
+static const struct super_operations v9fs_super_ops_dotl = {
+#ifdef CONFIG_9P_FSCACHE
+	.alloc_inode = v9fs_alloc_inode,
+	.destroy_inode = v9fs_destroy_inode,
+#endif
+	.statfs = v9fs_statfs,
+	.clear_inode = v9fs_clear_inode,
+	.show_options = generic_show_options,
+	.umount_begin = v9fs_umount_begin,
+};
+
 struct file_system_type v9fs_fs_type = {
 	.name = "9p",
 	.get_sb = v9fs_get_sb,
diff --git a/fs/adfs/dir.c b/fs/adfs/dir.c
index 23aa52f548a0..f4287e4de744 100644
--- a/fs/adfs/dir.c
+++ b/fs/adfs/dir.c
@@ -197,7 +197,7 @@ const struct file_operations adfs_dir_operations = {
 	.read		= generic_read_dir,
 	.llseek		= generic_file_llseek,
 	.readdir	= adfs_readdir,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 };
 
 static int
diff --git a/fs/adfs/file.c b/fs/adfs/file.c
index 005ea34d1758..a36da5382b40 100644
--- a/fs/adfs/file.c
+++ b/fs/adfs/file.c
@@ -26,7 +26,7 @@ const struct file_operations adfs_file_operations = {
 	.read		= do_sync_read,
 	.aio_read	= generic_file_aio_read,
 	.mmap		= generic_file_mmap,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 	.write		= do_sync_write,
 	.aio_write	= generic_file_aio_write,
 	.splice_read	= generic_file_splice_read,
diff --git a/fs/adfs/inode.c b/fs/adfs/inode.c
index 0f5e30978135..6f850b06ab62 100644
--- a/fs/adfs/inode.c
+++ b/fs/adfs/inode.c
@@ -322,8 +322,9 @@ adfs_notify_change(struct dentry *dentry, struct iattr *attr)
 	if (error)
 		goto out;
 
+	/* XXX: this is missing some actual on-disk truncation.. */
 	if (ia_valid & ATTR_SIZE)
-		error = vmtruncate(inode, attr->ia_size);
+		error = simple_setsize(inode, attr->ia_size);
 
 	if (error)
 		goto out;
diff --git a/fs/affs/affs.h b/fs/affs/affs.h
index 861dae68ac12..f05b6155ccc8 100644
--- a/fs/affs/affs.h
+++ b/fs/affs/affs.h
@@ -183,7 +183,7 @@ extern int			 affs_add_entry(struct inode *dir, struct inode *inode, struct dent
 
 void		affs_free_prealloc(struct inode *inode);
 extern void	affs_truncate(struct inode *);
-int		affs_file_fsync(struct file *, struct dentry *, int);
+int		affs_file_fsync(struct file *, int);
 
 /* dir.c */
 
diff --git a/fs/affs/file.c b/fs/affs/file.c
index 184e55c1c9ba..322710c3eedf 100644
--- a/fs/affs/file.c
+++ b/fs/affs/file.c
@@ -916,9 +916,9 @@ affs_truncate(struct inode *inode)
 	affs_free_prealloc(inode);
 }
 
-int affs_file_fsync(struct file *filp, struct dentry *dentry, int datasync)
+int affs_file_fsync(struct file *filp, int datasync)
 {
-	struct inode * inode = dentry->d_inode;
+	struct inode *inode = filp->f_mapping->host;
 	int ret, err;
 
 	ret = write_inode_now(inode, 0);
diff --git a/fs/affs/namei.c b/fs/affs/namei.c
index d70bbbac6b7b..914d1c0bc07a 100644
--- a/fs/affs/namei.c
+++ b/fs/affs/namei.c
@@ -224,7 +224,7 @@ affs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 		affs_brelse(bh);
 		inode = affs_iget(sb, ino);
 		if (IS_ERR(inode))
-			return ERR_PTR(PTR_ERR(inode));
+			return ERR_CAST(inode);
 	}
 	dentry->d_op = AFFS_SB(sb)->s_flags & SF_INTL ? &affs_intl_dentry_operations : &affs_dentry_operations;
 	d_add(dentry, inode);
diff --git a/fs/afs/internal.h b/fs/afs/internal.h
index 807f284cc75e..5f679b77ce24 100644
--- a/fs/afs/internal.h
+++ b/fs/afs/internal.h
@@ -740,7 +740,7 @@ extern void afs_pages_written_back(struct afs_vnode *, struct afs_call *);
 extern ssize_t afs_file_write(struct kiocb *, const struct iovec *,
 			      unsigned long, loff_t);
 extern int afs_writeback_all(struct afs_vnode *);
-extern int afs_fsync(struct file *, struct dentry *, int);
+extern int afs_fsync(struct file *, int);
 
 
 /*****************************************************************************/
diff --git a/fs/afs/write.c b/fs/afs/write.c
index 3bed54a294d4..3dab9e9948d0 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -701,8 +701,9 @@ int afs_writeback_all(struct afs_vnode *vnode)
  * - the return status from this call provides a reliable indication of
  *   whether any write errors occurred for this process.
  */
-int afs_fsync(struct file *file, struct dentry *dentry, int datasync)
+int afs_fsync(struct file *file, int datasync)
 {
+	struct dentry *dentry = file->f_path.dentry;
 	struct afs_writeback *wb, *xwb;
 	struct afs_vnode *vnode = AFS_FS_I(dentry->d_inode);
 	int ret;
diff --git a/fs/aio.c b/fs/aio.c
index 1cf12b3dd83a..1ccf25cef1f0 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -36,6 +36,7 @@
 #include <linux/blkdev.h>
 #include <linux/mempool.h>
 #include <linux/hash.h>
+#include <linux/compat.h>
 
 #include <asm/kmap_types.h>
 #include <asm/uaccess.h>
@@ -526,7 +527,7 @@ static void aio_fput_routine(struct work_struct *data)
 
 		/* Complete the fput(s) */
 		if (req->ki_filp != NULL)
-			__fput(req->ki_filp);
+			fput(req->ki_filp);
 
 		/* Link the iocb into the context's free list */
 		spin_lock_irq(&ctx->ctx_lock);
@@ -559,11 +560,11 @@ static int __aio_put_req(struct kioctx *ctx, struct kiocb *req)
 
 	/*
 	 * Try to optimize the aio and eventfd file* puts, by avoiding to
-	 * schedule work in case it is not __fput() time. In normal cases,
+	 * schedule work in case it is not final fput() time. In normal cases,
 	 * we would not be holding the last reference to the file*, so
 	 * this function will be executed w/out any aio kthread wakeup.
 	 */
-	if (unlikely(atomic_long_dec_and_test(&req->ki_filp->f_count))) {
+	if (unlikely(!fput_atomic(req->ki_filp))) {
 		get_ioctx(ctx);
 		spin_lock(&fput_lock);
 		list_add(&req->ki_list, &fput_head);
@@ -1384,13 +1385,22 @@ static ssize_t aio_fsync(struct kiocb *iocb)
 	return ret;
 }
 
-static ssize_t aio_setup_vectored_rw(int type, struct kiocb *kiocb)
+static ssize_t aio_setup_vectored_rw(int type, struct kiocb *kiocb, bool compat)
 {
 	ssize_t ret;
 
-	ret = rw_copy_check_uvector(type, (struct iovec __user *)kiocb->ki_buf,
-				    kiocb->ki_nbytes, 1,
-				    &kiocb->ki_inline_vec, &kiocb->ki_iovec);
+#ifdef CONFIG_COMPAT
+	if (compat)
+		ret = compat_rw_copy_check_uvector(type,
+				(struct compat_iovec __user *)kiocb->ki_buf,
+				kiocb->ki_nbytes, 1, &kiocb->ki_inline_vec,
+				&kiocb->ki_iovec);
+	else
+#endif
+		ret = rw_copy_check_uvector(type,
+				(struct iovec __user *)kiocb->ki_buf,
+				kiocb->ki_nbytes, 1, &kiocb->ki_inline_vec,
+				&kiocb->ki_iovec);
 	if (ret < 0)
 		goto out;
 
@@ -1420,7 +1430,7 @@ static ssize_t aio_setup_single_vector(struct kiocb *kiocb)
  *	Performs the initial checks and aio retry method
  *	setup for the kiocb at the time of io submission.
  */
-static ssize_t aio_setup_iocb(struct kiocb *kiocb)
+static ssize_t aio_setup_iocb(struct kiocb *kiocb, bool compat)
 {
 	struct file *file = kiocb->ki_filp;
 	ssize_t ret = 0;
@@ -1469,7 +1479,7 @@ static ssize_t aio_setup_iocb(struct kiocb *kiocb)
 		ret = security_file_permission(file, MAY_READ);
 		if (unlikely(ret))
 			break;
-		ret = aio_setup_vectored_rw(READ, kiocb);
+		ret = aio_setup_vectored_rw(READ, kiocb, compat);
 		if (ret)
 			break;
 		ret = -EINVAL;
@@ -1483,7 +1493,7 @@ static ssize_t aio_setup_iocb(struct kiocb *kiocb)
 		ret = security_file_permission(file, MAY_WRITE);
 		if (unlikely(ret))
 			break;
-		ret = aio_setup_vectored_rw(WRITE, kiocb);
+		ret = aio_setup_vectored_rw(WRITE, kiocb, compat);
 		if (ret)
 			break;
 		ret = -EINVAL;
@@ -1548,7 +1558,8 @@ static void aio_batch_free(struct hlist_head *batch_hash)
 }
 
 static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
-			 struct iocb *iocb, struct hlist_head *batch_hash)
+			 struct iocb *iocb, struct hlist_head *batch_hash,
+			 bool compat)
 {
 	struct kiocb *req;
 	struct file *file;
@@ -1609,7 +1620,7 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 	req->ki_left = req->ki_nbytes = iocb->aio_nbytes;
 	req->ki_opcode = iocb->aio_lio_opcode;
 
-	ret = aio_setup_iocb(req);
+	ret = aio_setup_iocb(req, compat);
 
 	if (ret)
 		goto out_put_req;
@@ -1637,20 +1648,8 @@ out_put_req:
 	return ret;
 }
 
-/* sys_io_submit:
- *	Queue the nr iocbs pointed to by iocbpp for processing.  Returns
- *	the number of iocbs queued.  May return -EINVAL if the aio_context
- *	specified by ctx_id is invalid, if nr is < 0, if the iocb at
- *	*iocbpp[0] is not properly initialized, if the operation specified
- *	is invalid for the file descriptor in the iocb.  May fail with
- *	-EFAULT if any of the data structures point to invalid data.  May
- *	fail with -EBADF if the file descriptor specified in the first
- *	iocb is invalid.  May fail with -EAGAIN if insufficient resources
- *	are available to queue any iocbs.  Will return 0 if nr is 0.  Will
- *	fail with -ENOSYS if not implemented.
- */
-SYSCALL_DEFINE3(io_submit, aio_context_t, ctx_id, long, nr,
-		struct iocb __user * __user *, iocbpp)
+long do_io_submit(aio_context_t ctx_id, long nr,
+		  struct iocb __user *__user *iocbpp, bool compat)
 {
 	struct kioctx *ctx;
 	long ret = 0;
@@ -1687,7 +1686,7 @@ SYSCALL_DEFINE3(io_submit, aio_context_t, ctx_id, long, nr,
 			break;
 		}
 
-		ret = io_submit_one(ctx, user_iocb, &tmp, batch_hash);
+		ret = io_submit_one(ctx, user_iocb, &tmp, batch_hash, compat);
 		if (ret)
 			break;
 	}
@@ -1697,6 +1696,24 @@ SYSCALL_DEFINE3(io_submit, aio_context_t, ctx_id, long, nr,
 	return i ? i : ret;
 }
 
+/* sys_io_submit:
+ *	Queue the nr iocbs pointed to by iocbpp for processing.  Returns
+ *	the number of iocbs queued.  May return -EINVAL if the aio_context
+ *	specified by ctx_id is invalid, if nr is < 0, if the iocb at
+ *	*iocbpp[0] is not properly initialized, if the operation specified
+ *	is invalid for the file descriptor in the iocb.  May fail with
+ *	-EFAULT if any of the data structures point to invalid data.  May
+ *	fail with -EBADF if the file descriptor specified in the first
+ *	iocb is invalid.  May fail with -EAGAIN if insufficient resources
+ *	are available to queue any iocbs.  Will return 0 if nr is 0.  Will
+ *	fail with -ENOSYS if not implemented.
+ */
+SYSCALL_DEFINE3(io_submit, aio_context_t, ctx_id, long, nr,
+		struct iocb __user * __user *, iocbpp)
+{
+	return do_io_submit(ctx_id, nr, iocbpp, 0);
+}
+
 /* lookup_kiocb
  *	Finds a given iocb for cancellation.
  */
diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c
index 9bd4b3876c99..e4b75d6eda83 100644
--- a/fs/anon_inodes.c
+++ b/fs/anon_inodes.c
@@ -205,7 +205,7 @@ static struct inode *anon_inode_mkinode(void)
 	 * that it already _is_ on the dirty list.
 	 */
 	inode->i_state = I_DIRTY;
-	inode->i_mode = S_IFREG | S_IRUSR | S_IWUSR;
+	inode->i_mode = S_IRUSR | S_IWUSR;
 	inode->i_uid = current_fsuid();
 	inode->i_gid = current_fsgid();
 	inode->i_flags |= S_PRIVATE;
diff --git a/fs/attr.c b/fs/attr.c
index 0815e93bb487..b4fa3b0aa596 100644
--- a/fs/attr.c
+++ b/fs/attr.c
@@ -67,14 +67,14 @@ EXPORT_SYMBOL(inode_change_ok);
  * @offset:	the new size to assign to the inode
  * @Returns:	0 on success, -ve errno on failure
  *
+ * inode_newsize_ok must be called with i_mutex held.
+ *
  * inode_newsize_ok will check filesystem limits and ulimits to check that the
  * new inode size is within limits. inode_newsize_ok will also send SIGXFSZ
  * when necessary. Caller must not proceed with inode size change if failure is
  * returned. @inode must be a file (not directory), with appropriate
  * permissions to allow truncate (inode_newsize_ok does NOT check these
  * conditions).
- *
- * inode_newsize_ok must be called with i_mutex held.
  */
 int inode_newsize_ok(const struct inode *inode, loff_t offset)
 {
@@ -104,17 +104,25 @@ out_big:
 }
 EXPORT_SYMBOL(inode_newsize_ok);
 
-int inode_setattr(struct inode * inode, struct iattr * attr)
+/**
+ * generic_setattr - copy simple metadata updates into the generic inode
+ * @inode:	the inode to be updated
+ * @attr:	the new attributes
+ *
+ * generic_setattr must be called with i_mutex held.
+ *
+ * generic_setattr updates the inode's metadata with that specified
+ * in attr. Noticably missing is inode size update, which is more complex
+ * as it requires pagecache updates. See simple_setsize.
+ *
+ * The inode is not marked as dirty after this operation. The rationale is
+ * that for "simple" filesystems, the struct inode is the inode storage.
+ * The caller is free to mark the inode dirty afterwards if needed.
+ */
+void generic_setattr(struct inode *inode, const struct iattr *attr)
 {
 	unsigned int ia_valid = attr->ia_valid;
 
-	if (ia_valid & ATTR_SIZE &&
-	    attr->ia_size != i_size_read(inode)) {
-		int error = vmtruncate(inode, attr->ia_size);
-		if (error)
-			return error;
-	}
-
 	if (ia_valid & ATTR_UID)
 		inode->i_uid = attr->ia_uid;
 	if (ia_valid & ATTR_GID)
@@ -135,6 +143,28 @@ int inode_setattr(struct inode * inode, struct iattr * attr)
 			mode &= ~S_ISGID;
 		inode->i_mode = mode;
 	}
+}
+EXPORT_SYMBOL(generic_setattr);
+
+/*
+ * note this function is deprecated, the new truncate sequence should be
+ * used instead -- see eg. simple_setsize, generic_setattr.
+ */
+int inode_setattr(struct inode *inode, const struct iattr *attr)
+{
+	unsigned int ia_valid = attr->ia_valid;
+
+	if (ia_valid & ATTR_SIZE &&
+	    attr->ia_size != i_size_read(inode)) {
+		int error;
+
+		error = vmtruncate(inode, attr->ia_size);
+		if (error)
+			return error;
+	}
+
+	generic_setattr(inode, attr);
+
 	mark_inode_dirty(inode);
 
 	return 0;
diff --git a/fs/autofs/root.c b/fs/autofs/root.c
index 8713c7cfbc79..9a0520b50663 100644
--- a/fs/autofs/root.c
+++ b/fs/autofs/root.c
@@ -28,6 +28,7 @@ static int autofs_root_mkdir(struct inode *,struct dentry *,int);
 static int autofs_root_ioctl(struct inode *, struct file *,unsigned int,unsigned long);
 
 const struct file_operations autofs_root_operations = {
+	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= autofs_root_readdir,
 	.ioctl		= autofs_root_ioctl,
diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c
index d29b7f6df862..ba4a38b9c22f 100644
--- a/fs/autofs4/dev-ioctl.c
+++ b/fs/autofs4/dev-ioctl.c
@@ -95,7 +95,7 @@ static int check_dev_ioctl_version(int cmd, struct autofs_dev_ioctl *param)
  */
 static struct autofs_dev_ioctl *copy_dev_ioctl(struct autofs_dev_ioctl __user *in)
 {
-	struct autofs_dev_ioctl tmp, *ads;
+	struct autofs_dev_ioctl tmp;
 
 	if (copy_from_user(&tmp, in, sizeof(tmp)))
 		return ERR_PTR(-EFAULT);
@@ -103,16 +103,7 @@ static struct autofs_dev_ioctl *copy_dev_ioctl(struct autofs_dev_ioctl __user *i
 	if (tmp.size < sizeof(tmp))
 		return ERR_PTR(-EINVAL);
 
-	ads = kmalloc(tmp.size, GFP_KERNEL);
-	if (!ads)
-		return ERR_PTR(-ENOMEM);
-
-	if (copy_from_user(ads, in, tmp.size)) {
-		kfree(ads);
-		return ERR_PTR(-EFAULT);
-	}
-
-	return ads;
+	return memdup_user(in, tmp.size);
 }
 
 static inline void free_dev_ioctl(struct autofs_dev_ioctl *param)
@@ -736,11 +727,14 @@ static const struct file_operations _dev_ioctl_fops = {
 };
 
 static struct miscdevice _autofs_dev_ioctl_misc = {
-	.minor 		= MISC_DYNAMIC_MINOR,
+	.minor		= AUTOFS_MINOR,
 	.name  		= AUTOFS_DEVICE_NAME,
 	.fops  		= &_dev_ioctl_fops
 };
 
+MODULE_ALIAS_MISCDEV(AUTOFS_MINOR);
+MODULE_ALIAS("devname:autofs");
+
 /* Register/deregister misc character device */
 int autofs_dev_ioctl_init(void)
 {
diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c
index e8e5e63ac950..db4117ed7803 100644
--- a/fs/autofs4/root.c
+++ b/fs/autofs4/root.c
@@ -18,13 +18,14 @@
 #include <linux/slab.h>
 #include <linux/param.h>
 #include <linux/time.h>
+#include <linux/smp_lock.h>
 #include "autofs_i.h"
 
 static int autofs4_dir_symlink(struct inode *,struct dentry *,const char *);
 static int autofs4_dir_unlink(struct inode *,struct dentry *);
 static int autofs4_dir_rmdir(struct inode *,struct dentry *);
 static int autofs4_dir_mkdir(struct inode *,struct dentry *,int);
-static int autofs4_root_ioctl(struct inode *, struct file *,unsigned int,unsigned long);
+static long autofs4_root_ioctl(struct file *,unsigned int,unsigned long);
 static int autofs4_dir_open(struct inode *inode, struct file *file);
 static struct dentry *autofs4_lookup(struct inode *,struct dentry *, struct nameidata *);
 static void *autofs4_follow_link(struct dentry *, struct nameidata *);
@@ -38,7 +39,7 @@ const struct file_operations autofs4_root_operations = {
 	.read		= generic_read_dir,
 	.readdir	= dcache_readdir,
 	.llseek		= dcache_dir_lseek,
-	.ioctl		= autofs4_root_ioctl,
+	.unlocked_ioctl	= autofs4_root_ioctl,
 };
 
 const struct file_operations autofs4_dir_operations = {
@@ -902,8 +903,8 @@ int is_autofs4_dentry(struct dentry *dentry)
  * ioctl()'s on the root directory is the chief method for the daemon to
  * generate kernel reactions
  */
-static int autofs4_root_ioctl(struct inode *inode, struct file *filp,
-			     unsigned int cmd, unsigned long arg)
+static int autofs4_root_ioctl_unlocked(struct inode *inode, struct file *filp,
+				       unsigned int cmd, unsigned long arg)
 {
 	struct autofs_sb_info *sbi = autofs4_sbi(inode->i_sb);
 	void __user *p = (void __user *)arg;
@@ -947,3 +948,16 @@ static int autofs4_root_ioctl(struct inode *inode, struct file *filp,
 		return -ENOSYS;
 	}
 }
+
+static long autofs4_root_ioctl(struct file *filp,
+			       unsigned int cmd, unsigned long arg)
+{
+	long ret;
+	struct inode *inode = filp->f_dentry->d_inode;
+
+	lock_kernel();
+	ret = autofs4_root_ioctl_unlocked(inode, filp, cmd, arg);
+	unlock_kernel();
+
+	return ret;
+}
diff --git a/fs/bad_inode.c b/fs/bad_inode.c
index a05287a23f62..52e59bf4aa5f 100644
--- a/fs/bad_inode.c
+++ b/fs/bad_inode.c
@@ -93,8 +93,7 @@ static int bad_file_release(struct inode *inode, struct file *filp)
 	return -EIO;
 }
 
-static int bad_file_fsync(struct file *file, struct dentry *dentry,
-			int datasync)
+static int bad_file_fsync(struct file *file, int datasync)
 {
 	return -EIO;
 }
diff --git a/fs/bfs/dir.c b/fs/bfs/dir.c
index 8f73841fc974..d967e052b779 100644
--- a/fs/bfs/dir.c
+++ b/fs/bfs/dir.c
@@ -78,7 +78,7 @@ static int bfs_readdir(struct file *f, void *dirent, filldir_t filldir)
 const struct file_operations bfs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= bfs_readdir,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 	.llseek		= generic_file_llseek,
 };
 
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 26e5f5026620..7346c96308a5 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -172,8 +172,9 @@ blkdev_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file->f_mapping->host;
 
-	return blockdev_direct_IO_no_locking(rw, iocb, inode, I_BDEV(inode),
-				iov, offset, nr_segs, blkdev_get_blocks, NULL);
+	return blockdev_direct_IO_no_locking_newtrunc(rw, iocb, inode,
+				I_BDEV(inode), iov, offset, nr_segs,
+				blkdev_get_blocks, NULL);
 }
 
 int __sync_blockdev(struct block_device *bdev, int wait)
@@ -309,8 +310,8 @@ static int blkdev_write_begin(struct file *file, struct address_space *mapping,
 			struct page **pagep, void **fsdata)
 {
 	*pagep = NULL;
-	return block_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
-				blkdev_get_block);
+	return block_write_begin_newtrunc(file, mapping, pos, len, flags,
+				pagep, fsdata, blkdev_get_block);
 }
 
 static int blkdev_write_end(struct file *file, struct address_space *mapping,
@@ -358,12 +359,7 @@ static loff_t block_llseek(struct file *file, loff_t offset, int origin)
 	return retval;
 }
 	
-/*
- *	Filp is never NULL; the only case when ->fsync() is called with
- *	NULL first argument is nfsd_sync_dir() and that's not a directory.
- */
- 
-int blkdev_fsync(struct file *filp, struct dentry *dentry, int datasync)
+int blkdev_fsync(struct file *filp, int datasync)
 {
 	struct inode *bd_inode = filp->f_mapping->host;
 	struct block_device *bdev = I_BDEV(bd_inode);
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 462859a30141..7ec14097fef1 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -377,6 +377,7 @@ again:
 				if (!list_empty(&worker->pending) ||
 				    !list_empty(&worker->prio_pending)) {
 					spin_unlock_irq(&worker->lock);
+					set_current_state(TASK_RUNNING);
 					goto again;
 				}
 
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 7a4dee199832..6ad63f17eca0 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -137,8 +137,8 @@ struct btrfs_inode {
 	 * of extent items we've reserved metadata for.
 	 */
 	spinlock_t accounting_lock;
+	atomic_t outstanding_extents;
 	int reserved_extents;
-	int outstanding_extents;
 
 	/*
 	 * ordered_data_close is set by truncate when a file that used
@@ -151,6 +151,7 @@ struct btrfs_inode {
 	 * of these.
 	 */
 	unsigned ordered_data_close:1;
+	unsigned orphan_meta_reserved:1;
 	unsigned dummy_inode:1;
 
 	/*
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 6795a713b205..0d1d966b0fe4 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -280,7 +280,8 @@ int btrfs_block_can_be_shared(struct btrfs_root *root,
 static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 				       struct btrfs_root *root,
 				       struct extent_buffer *buf,
-				       struct extent_buffer *cow)
+				       struct extent_buffer *cow,
+				       int *last_ref)
 {
 	u64 refs;
 	u64 owner;
@@ -366,6 +367,7 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 			BUG_ON(ret);
 		}
 		clean_tree_block(trans, root, buf);
+		*last_ref = 1;
 	}
 	return 0;
 }
@@ -392,6 +394,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 	struct btrfs_disk_key disk_key;
 	struct extent_buffer *cow;
 	int level;
+	int last_ref = 0;
 	int unlock_orig = 0;
 	u64 parent_start;
 
@@ -442,7 +445,10 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 			    (unsigned long)btrfs_header_fsid(cow),
 			    BTRFS_FSID_SIZE);
 
-	update_ref_for_cow(trans, root, buf, cow);
+	update_ref_for_cow(trans, root, buf, cow, &last_ref);
+
+	if (root->ref_cows)
+		btrfs_reloc_cow_block(trans, root, buf, cow);
 
 	if (buf == root->node) {
 		WARN_ON(parent && parent != buf);
@@ -457,8 +463,8 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 		extent_buffer_get(cow);
 		spin_unlock(&root->node_lock);
 
-		btrfs_free_tree_block(trans, root, buf->start, buf->len,
-				parent_start, root->root_key.objectid, level);
+		btrfs_free_tree_block(trans, root, buf, parent_start,
+				      last_ref);
 		free_extent_buffer(buf);
 		add_root_to_dirty_list(root);
 	} else {
@@ -473,8 +479,8 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 		btrfs_set_node_ptr_generation(parent, parent_slot,
 					      trans->transid);
 		btrfs_mark_buffer_dirty(parent);
-		btrfs_free_tree_block(trans, root, buf->start, buf->len,
-				parent_start, root->root_key.objectid, level);
+		btrfs_free_tree_block(trans, root, buf, parent_start,
+				      last_ref);
 	}
 	if (unlock_orig)
 		btrfs_tree_unlock(buf);
@@ -949,6 +955,22 @@ int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key,
 	return bin_search(eb, key, level, slot);
 }
 
+static void root_add_used(struct btrfs_root *root, u32 size)
+{
+	spin_lock(&root->accounting_lock);
+	btrfs_set_root_used(&root->root_item,
+			    btrfs_root_used(&root->root_item) + size);
+	spin_unlock(&root->accounting_lock);
+}
+
+static void root_sub_used(struct btrfs_root *root, u32 size)
+{
+	spin_lock(&root->accounting_lock);
+	btrfs_set_root_used(&root->root_item,
+			    btrfs_root_used(&root->root_item) - size);
+	spin_unlock(&root->accounting_lock);
+}
+
 /* given a node and slot number, this reads the blocks it points to.  The
  * extent buffer is returned with a reference taken (but unlocked).
  * NULL is returned on error.
@@ -1019,7 +1041,11 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		btrfs_tree_lock(child);
 		btrfs_set_lock_blocking(child);
 		ret = btrfs_cow_block(trans, root, child, mid, 0, &child);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_tree_unlock(child);
+			free_extent_buffer(child);
+			goto enospc;
+		}
 
 		spin_lock(&root->node_lock);
 		root->node = child;
@@ -1034,11 +1060,12 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		btrfs_tree_unlock(mid);
 		/* once for the path */
 		free_extent_buffer(mid);
-		ret = btrfs_free_tree_block(trans, root, mid->start, mid->len,
-					    0, root->root_key.objectid, level);
+
+		root_sub_used(root, mid->len);
+		btrfs_free_tree_block(trans, root, mid, 0, 1);
 		/* once for the root ptr */
 		free_extent_buffer(mid);
-		return ret;
+		return 0;
 	}
 	if (btrfs_header_nritems(mid) >
 	    BTRFS_NODEPTRS_PER_BLOCK(root) / 4)
@@ -1088,23 +1115,16 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		if (wret < 0 && wret != -ENOSPC)
 			ret = wret;
 		if (btrfs_header_nritems(right) == 0) {
-			u64 bytenr = right->start;
-			u32 blocksize = right->len;
-
 			clean_tree_block(trans, root, right);
 			btrfs_tree_unlock(right);
-			free_extent_buffer(right);
-			right = NULL;
 			wret = del_ptr(trans, root, path, level + 1, pslot +
 				       1);
 			if (wret)
 				ret = wret;
-			wret = btrfs_free_tree_block(trans, root,
-						     bytenr, blocksize, 0,
-						     root->root_key.objectid,
-						     level);
-			if (wret)
-				ret = wret;
+			root_sub_used(root, right->len);
+			btrfs_free_tree_block(trans, root, right, 0, 1);
+			free_extent_buffer(right);
+			right = NULL;
 		} else {
 			struct btrfs_disk_key right_key;
 			btrfs_node_key(right, &right_key, 0);
@@ -1136,21 +1156,15 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		BUG_ON(wret == 1);
 	}
 	if (btrfs_header_nritems(mid) == 0) {
-		/* we've managed to empty the middle node, drop it */
-		u64 bytenr = mid->start;
-		u32 blocksize = mid->len;
-
 		clean_tree_block(trans, root, mid);
 		btrfs_tree_unlock(mid);
-		free_extent_buffer(mid);
-		mid = NULL;
 		wret = del_ptr(trans, root, path, level + 1, pslot);
 		if (wret)
 			ret = wret;
-		wret = btrfs_free_tree_block(trans, root, bytenr, blocksize,
-					 0, root->root_key.objectid, level);
-		if (wret)
-			ret = wret;
+		root_sub_used(root, mid->len);
+		btrfs_free_tree_block(trans, root, mid, 0, 1);
+		free_extent_buffer(mid);
+		mid = NULL;
 	} else {
 		/* update the parent key to reflect our changes */
 		struct btrfs_disk_key mid_key;
@@ -1590,7 +1604,7 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 	btrfs_release_path(NULL, p);
 
 	ret = -EAGAIN;
-	tmp = read_tree_block(root, blocknr, blocksize, gen);
+	tmp = read_tree_block(root, blocknr, blocksize, 0);
 	if (tmp) {
 		/*
 		 * If the read above didn't mark this buffer up to date,
@@ -1740,7 +1754,6 @@ again:
 					      p->nodes[level + 1],
 					      p->slots[level + 1], &b);
 			if (err) {
-				free_extent_buffer(b);
 				ret = err;
 				goto done;
 			}
@@ -2076,6 +2089,8 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans,
 	if (IS_ERR(c))
 		return PTR_ERR(c);
 
+	root_add_used(root, root->nodesize);
+
 	memset_extent_buffer(c, 0, 0, sizeof(struct btrfs_header));
 	btrfs_set_header_nritems(c, 1);
 	btrfs_set_header_level(c, level);
@@ -2134,6 +2149,7 @@ static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root
 	int nritems;
 
 	BUG_ON(!path->nodes[level]);
+	btrfs_assert_tree_locked(path->nodes[level]);
 	lower = path->nodes[level];
 	nritems = btrfs_header_nritems(lower);
 	BUG_ON(slot > nritems);
@@ -2202,6 +2218,8 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 	if (IS_ERR(split))
 		return PTR_ERR(split);
 
+	root_add_used(root, root->nodesize);
+
 	memset_extent_buffer(split, 0, 0, sizeof(struct btrfs_header));
 	btrfs_set_header_level(split, btrfs_header_level(c));
 	btrfs_set_header_bytenr(split, split->start);
@@ -2415,6 +2433,9 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans,
 
 	if (left_nritems)
 		btrfs_mark_buffer_dirty(left);
+	else
+		clean_tree_block(trans, root, left);
+
 	btrfs_mark_buffer_dirty(right);
 
 	btrfs_item_key(right, &disk_key, 0);
@@ -2660,6 +2681,8 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(left);
 	if (right_nritems)
 		btrfs_mark_buffer_dirty(right);
+	else
+		clean_tree_block(trans, root, right);
 
 	btrfs_item_key(right, &disk_key, 0);
 	wret = fixup_low_keys(trans, root, path, &disk_key, 1);
@@ -2669,8 +2692,6 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 	/* then fixup the leaf pointer in the path */
 	if (path->slots[0] < push_items) {
 		path->slots[0] += old_left_nritems;
-		if (btrfs_header_nritems(path->nodes[0]) == 0)
-			clean_tree_block(trans, root, path->nodes[0]);
 		btrfs_tree_unlock(path->nodes[0]);
 		free_extent_buffer(path->nodes[0]);
 		path->nodes[0] = left;
@@ -2932,10 +2953,10 @@ again:
 	right = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
 					root->root_key.objectid,
 					&disk_key, 0, l->start, 0);
-	if (IS_ERR(right)) {
-		BUG_ON(1);
+	if (IS_ERR(right))
 		return PTR_ERR(right);
-	}
+
+	root_add_used(root, root->leafsize);
 
 	memset_extent_buffer(right, 0, 0, sizeof(struct btrfs_header));
 	btrfs_set_header_bytenr(right, right->start);
@@ -3054,7 +3075,8 @@ static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans,
 
 	btrfs_set_path_blocking(path);
 	ret = split_leaf(trans, root, &key, path, ins_len, 1);
-	BUG_ON(ret);
+	if (ret)
+		goto err;
 
 	path->keep_locks = 0;
 	btrfs_unlock_up_safe(path, 1);
@@ -3796,9 +3818,10 @@ static noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans,
 	 */
 	btrfs_unlock_up_safe(path, 0);
 
-	ret = btrfs_free_tree_block(trans, root, leaf->start, leaf->len,
-				    0, root->root_key.objectid, 0);
-	return ret;
+	root_sub_used(root, leaf->len);
+
+	btrfs_free_tree_block(trans, root, leaf, 0, 1);
+	return 0;
 }
 /*
  * delete the item at the leaf level in path.  If that empties
@@ -3865,6 +3888,8 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		if (leaf == root->node) {
 			btrfs_set_header_level(leaf, 0);
 		} else {
+			btrfs_set_path_blocking(path);
+			clean_tree_block(trans, root, leaf);
 			ret = btrfs_del_leaf(trans, root, path, leaf);
 			BUG_ON(ret);
 		}
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 746a7248678e..29c20092847e 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -34,6 +34,7 @@
 
 struct btrfs_trans_handle;
 struct btrfs_transaction;
+struct btrfs_pending_snapshot;
 extern struct kmem_cache *btrfs_trans_handle_cachep;
 extern struct kmem_cache *btrfs_transaction_cachep;
 extern struct kmem_cache *btrfs_bit_radix_cachep;
@@ -663,6 +664,7 @@ struct btrfs_csum_item {
 #define BTRFS_BLOCK_GROUP_RAID1    (1 << 4)
 #define BTRFS_BLOCK_GROUP_DUP	   (1 << 5)
 #define BTRFS_BLOCK_GROUP_RAID10   (1 << 6)
+#define BTRFS_NR_RAID_TYPES	   5
 
 struct btrfs_block_group_item {
 	__le64 used;
@@ -674,42 +676,46 @@ struct btrfs_space_info {
 	u64 flags;
 
 	u64 total_bytes;	/* total bytes in the space */
-	u64 bytes_used;		/* total bytes used on disk */
+	u64 bytes_used;		/* total bytes used,
+				   this does't take mirrors into account */
 	u64 bytes_pinned;	/* total bytes pinned, will be freed when the
 				   transaction finishes */
 	u64 bytes_reserved;	/* total bytes the allocator has reserved for
 				   current allocations */
 	u64 bytes_readonly;	/* total bytes that are read only */
-	u64 bytes_super;	/* total bytes reserved for the super blocks */
-	u64 bytes_root;		/* the number of bytes needed to commit a
-				   transaction */
+
 	u64 bytes_may_use;	/* number of bytes that may be used for
 				   delalloc/allocations */
-	u64 bytes_delalloc;	/* number of bytes currently reserved for
-				   delayed allocation */
+	u64 disk_used;		/* total bytes used on disk */
 
 	int full;		/* indicates that we cannot allocate any more
 				   chunks for this space */
 	int force_alloc;	/* set if we need to force a chunk alloc for
 				   this space */
-	int force_delalloc;	/* make people start doing filemap_flush until
-				   we're under a threshold */
 
 	struct list_head list;
 
-	/* for controlling how we free up space for allocations */
-	wait_queue_head_t allocate_wait;
-	wait_queue_head_t flush_wait;
-	int allocating_chunk;
-	int flushing;
-
 	/* for block groups in our same type */
-	struct list_head block_groups;
+	struct list_head block_groups[BTRFS_NR_RAID_TYPES];
 	spinlock_t lock;
 	struct rw_semaphore groups_sem;
 	atomic_t caching_threads;
 };
 
+struct btrfs_block_rsv {
+	u64 size;
+	u64 reserved;
+	u64 freed[2];
+	struct btrfs_space_info *space_info;
+	struct list_head list;
+	spinlock_t lock;
+	atomic_t usage;
+	unsigned int priority:8;
+	unsigned int durable:1;
+	unsigned int refill_used:1;
+	unsigned int full:1;
+};
+
 /*
  * free clusters are used to claim free space in relatively large chunks,
  * allowing us to do less seeky writes.  They are used for all metadata
@@ -760,6 +766,7 @@ struct btrfs_block_group_cache {
 	spinlock_t lock;
 	u64 pinned;
 	u64 reserved;
+	u64 reserved_pinned;
 	u64 bytes_super;
 	u64 flags;
 	u64 sectorsize;
@@ -825,6 +832,22 @@ struct btrfs_fs_info {
 	/* logical->physical extent mapping */
 	struct btrfs_mapping_tree mapping_tree;
 
+	/* block reservation for extent, checksum and root tree */
+	struct btrfs_block_rsv global_block_rsv;
+	/* block reservation for delay allocation */
+	struct btrfs_block_rsv delalloc_block_rsv;
+	/* block reservation for metadata operations */
+	struct btrfs_block_rsv trans_block_rsv;
+	/* block reservation for chunk tree */
+	struct btrfs_block_rsv chunk_block_rsv;
+
+	struct btrfs_block_rsv empty_block_rsv;
+
+	/* list of block reservations that cross multiple transactions */
+	struct list_head durable_block_rsv_list;
+
+	struct mutex durable_block_rsv_mutex;
+
 	u64 generation;
 	u64 last_trans_committed;
 
@@ -927,7 +950,6 @@ struct btrfs_fs_info {
 	struct btrfs_workers endio_meta_write_workers;
 	struct btrfs_workers endio_write_workers;
 	struct btrfs_workers submit_workers;
-	struct btrfs_workers enospc_workers;
 	/*
 	 * fixup workers take dirty pages that didn't properly go through
 	 * the cow mechanism and make them safe to write.  It happens
@@ -943,6 +965,7 @@ struct btrfs_fs_info {
 	int do_barriers;
 	int closing;
 	int log_root_recovering;
+	int enospc_unlink;
 
 	u64 total_pinned;
 
@@ -1012,6 +1035,9 @@ struct btrfs_root {
 	struct completion kobj_unregister;
 	struct mutex objectid_mutex;
 
+	spinlock_t accounting_lock;
+	struct btrfs_block_rsv *block_rsv;
+
 	struct mutex log_mutex;
 	wait_queue_head_t log_writer_wait;
 	wait_queue_head_t log_commit_wait[2];
@@ -1043,7 +1069,6 @@ struct btrfs_root {
 	int ref_cows;
 	int track_dirty;
 	int in_radix;
-	int clean_orphans;
 
 	u64 defrag_trans_start;
 	struct btrfs_key defrag_progress;
@@ -1057,8 +1082,11 @@ struct btrfs_root {
 
 	struct list_head root_list;
 
-	spinlock_t list_lock;
+	spinlock_t orphan_lock;
 	struct list_head orphan_list;
+	struct btrfs_block_rsv *orphan_block_rsv;
+	int orphan_item_inserted;
+	int orphan_cleanup_state;
 
 	spinlock_t inode_lock;
 	/* red-black tree that keeps track of in-memory inodes */
@@ -1965,6 +1993,9 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache);
 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root, unsigned long count);
 int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len);
+int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, u64 bytenr,
+			     u64 num_bytes, u64 *refs, u64 *flags);
 int btrfs_pin_extent(struct btrfs_root *root,
 		     u64 bytenr, u64 num, int reserved);
 int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
@@ -1984,10 +2015,10 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 					u64 parent, u64 root_objectid,
 					struct btrfs_disk_key *key, int level,
 					u64 hint, u64 empty_size);
-int btrfs_free_tree_block(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root,
-			  u64 bytenr, u32 blocksize,
-			  u64 parent, u64 root_objectid, int level);
+void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   struct extent_buffer *buf,
+			   u64 parent, int last_ref);
 struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 					    struct btrfs_root *root,
 					    u64 bytenr, u32 blocksize,
@@ -2041,27 +2072,49 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 			   u64 size);
 int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root, u64 group_start);
-int btrfs_prepare_block_group_relocation(struct btrfs_root *root,
-				struct btrfs_block_group_cache *group);
-
 u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags);
 void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *ionde);
 void btrfs_clear_space_info_full(struct btrfs_fs_info *info);
-
-int btrfs_reserve_metadata_space(struct btrfs_root *root, int num_items);
-int btrfs_unreserve_metadata_space(struct btrfs_root *root, int num_items);
-int btrfs_unreserve_metadata_for_delalloc(struct btrfs_root *root,
-					  struct inode *inode, int num_items);
-int btrfs_reserve_metadata_for_delalloc(struct btrfs_root *root,
-					struct inode *inode, int num_items);
-int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
-				u64 bytes);
-void btrfs_free_reserved_data_space(struct btrfs_root *root,
-				    struct inode *inode, u64 bytes);
-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);
+int btrfs_check_data_free_space(struct inode *inode, u64 bytes);
+void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes);
+int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root,
+				int num_items, int *retries);
+void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root);
+int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
+				  struct inode *inode);
+void btrfs_orphan_release_metadata(struct inode *inode);
+int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans,
+				struct btrfs_pending_snapshot *pending);
+int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes);
+void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes);
+int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes);
+void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes);
+void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv);
+struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root);
+void btrfs_free_block_rsv(struct btrfs_root *root,
+			  struct btrfs_block_rsv *rsv);
+void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info,
+				 struct btrfs_block_rsv *rsv);
+int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct btrfs_block_rsv *block_rsv,
+			u64 num_bytes, int *retries);
+int btrfs_block_rsv_check(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root,
+			  struct btrfs_block_rsv *block_rsv,
+			  u64 min_reserved, int min_factor);
+int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
+			    struct btrfs_block_rsv *dst_rsv,
+			    u64 num_bytes);
+void btrfs_block_rsv_release(struct btrfs_root *root,
+			     struct btrfs_block_rsv *block_rsv,
+			     u64 num_bytes);
+int btrfs_set_block_group_ro(struct btrfs_root *root,
+			     struct btrfs_block_group_cache *cache);
+int btrfs_set_block_group_rw(struct btrfs_root *root,
+			     struct btrfs_block_group_cache *cache);
 /* ctree.c */
 int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key,
 		     int level, int *slot);
@@ -2152,7 +2205,8 @@ static inline int btrfs_insert_empty_item(struct btrfs_trans_handle *trans,
 int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path);
 int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path);
 int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf);
-int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref);
+int btrfs_drop_snapshot(struct btrfs_root *root,
+			struct btrfs_block_rsv *block_rsv, int update_ref);
 int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
 			struct btrfs_root *root,
 			struct extent_buffer *node,
@@ -2245,6 +2299,12 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
 			   const char *name, int name_len,
 			   u64 inode_objectid, u64 ref_objectid, u64 *index);
+struct btrfs_inode_ref *
+btrfs_lookup_inode_ref(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct btrfs_path *path,
+			const char *name, int name_len,
+			u64 inode_objectid, u64 ref_objectid, int mod);
 int btrfs_insert_empty_inode(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
 			     struct btrfs_path *path, u64 objectid);
@@ -2257,6 +2317,8 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 		    struct btrfs_root *root, u64 bytenr, u64 len);
 int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode,
 			  struct bio *bio, u32 *dst);
+int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode,
+			      struct bio *bio, u64 logical_offset, u32 *dst);
 int btrfs_insert_file_extent(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
 			     u64 objectid, u64 pos,
@@ -2311,6 +2373,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 			       u32 min_type);
 
 int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput);
+int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput);
 int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
 			      struct extent_state **cached_state);
 int btrfs_writepages(struct address_space *mapping,
@@ -2349,10 +2412,20 @@ int btrfs_update_inode(struct btrfs_trans_handle *trans,
 int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode);
 int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode);
 void btrfs_orphan_cleanup(struct btrfs_root *root);
+void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans,
+				struct btrfs_pending_snapshot *pending,
+				u64 *bytes_to_reserve);
+void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans,
+				struct btrfs_pending_snapshot *pending);
+void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root);
 int btrfs_cont_expand(struct inode *inode, loff_t size);
 int btrfs_invalidate_inodes(struct btrfs_root *root);
 void btrfs_add_delayed_iput(struct inode *inode);
 void btrfs_run_delayed_iputs(struct btrfs_root *root);
+int btrfs_prealloc_file_range(struct inode *inode, int mode,
+			      u64 start, u64 num_bytes, u64 min_size,
+			      loff_t actual_len, u64 *alloc_hint);
 extern const struct dentry_operations btrfs_dentry_operations;
 
 /* ioctl.c */
@@ -2361,7 +2434,7 @@ void btrfs_update_iflags(struct inode *inode);
 void btrfs_inherit_iflags(struct inode *inode, struct inode *dir);
 
 /* file.c */
-int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync);
+int btrfs_sync_file(struct file *file, int datasync);
 int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 			    int skip_pinned);
 int btrfs_check_file(struct btrfs_root *root, struct inode *inode);
@@ -2409,4 +2482,12 @@ int btrfs_update_reloc_root(struct btrfs_trans_handle *trans,
 			    struct btrfs_root *root);
 int btrfs_recover_relocation(struct btrfs_root *root);
 int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len);
+void btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, struct extent_buffer *buf,
+			   struct extent_buffer *cow);
+void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans,
+			      struct btrfs_pending_snapshot *pending,
+			      u64 *bytes_to_reserve);
+void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
+			      struct btrfs_pending_snapshot *pending);
 #endif
diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index 902ce507c4e3..e807b143b857 100644
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -319,107 +319,6 @@ out:
 }
 
 /*
- * helper function to lookup reference count and flags of extent.
- *
- * the head node for delayed ref is used to store the sum of all the
- * reference count modifications queued up in the rbtree. the head
- * node may also store the extent flags to set. This way you can check
- * to see what the reference count and extent flags would be if all of
- * the delayed refs are not processed.
- */
-int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *root, u64 bytenr,
-			     u64 num_bytes, u64 *refs, u64 *flags)
-{
-	struct btrfs_delayed_ref_node *ref;
-	struct btrfs_delayed_ref_head *head;
-	struct btrfs_delayed_ref_root *delayed_refs;
-	struct btrfs_path *path;
-	struct btrfs_extent_item *ei;
-	struct extent_buffer *leaf;
-	struct btrfs_key key;
-	u32 item_size;
-	u64 num_refs;
-	u64 extent_flags;
-	int ret;
-
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
-
-	key.objectid = bytenr;
-	key.type = BTRFS_EXTENT_ITEM_KEY;
-	key.offset = num_bytes;
-	delayed_refs = &trans->transaction->delayed_refs;
-again:
-	ret = btrfs_search_slot(trans, root->fs_info->extent_root,
-				&key, path, 0, 0);
-	if (ret < 0)
-		goto out;
-
-	if (ret == 0) {
-		leaf = path->nodes[0];
-		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
-		if (item_size >= sizeof(*ei)) {
-			ei = btrfs_item_ptr(leaf, path->slots[0],
-					    struct btrfs_extent_item);
-			num_refs = btrfs_extent_refs(leaf, ei);
-			extent_flags = btrfs_extent_flags(leaf, ei);
-		} else {
-#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
-			struct btrfs_extent_item_v0 *ei0;
-			BUG_ON(item_size != sizeof(*ei0));
-			ei0 = btrfs_item_ptr(leaf, path->slots[0],
-					     struct btrfs_extent_item_v0);
-			num_refs = btrfs_extent_refs_v0(leaf, ei0);
-			/* FIXME: this isn't correct for data */
-			extent_flags = BTRFS_BLOCK_FLAG_FULL_BACKREF;
-#else
-			BUG();
-#endif
-		}
-		BUG_ON(num_refs == 0);
-	} else {
-		num_refs = 0;
-		extent_flags = 0;
-		ret = 0;
-	}
-
-	spin_lock(&delayed_refs->lock);
-	ref = find_ref_head(&delayed_refs->root, bytenr, NULL);
-	if (ref) {
-		head = btrfs_delayed_node_to_head(ref);
-		if (!mutex_trylock(&head->mutex)) {
-			atomic_inc(&ref->refs);
-			spin_unlock(&delayed_refs->lock);
-
-			btrfs_release_path(root->fs_info->extent_root, path);
-
-			mutex_lock(&head->mutex);
-			mutex_unlock(&head->mutex);
-			btrfs_put_delayed_ref(ref);
-			goto again;
-		}
-		if (head->extent_op && head->extent_op->update_flags)
-			extent_flags |= head->extent_op->flags_to_set;
-		else
-			BUG_ON(num_refs == 0);
-
-		num_refs += ref->ref_mod;
-		mutex_unlock(&head->mutex);
-	}
-	WARN_ON(num_refs == 0);
-	if (refs)
-		*refs = num_refs;
-	if (flags)
-		*flags = extent_flags;
-out:
-	spin_unlock(&delayed_refs->lock);
-	btrfs_free_path(path);
-	return ret;
-}
-
-/*
  * helper function to update an extent delayed ref in the
  * rbtree.  existing and update must both have the same
  * bytenr and parent
diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h
index f6fc67ddad36..50e3cf92fbda 100644
--- a/fs/btrfs/delayed-ref.h
+++ b/fs/btrfs/delayed-ref.h
@@ -167,9 +167,6 @@ int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans,
 struct btrfs_delayed_ref_head *
 btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr);
 int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr);
-int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *root, u64 bytenr,
-			     u64 num_bytes, u64 *refs, u64 *flags);
 int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans,
 			  u64 bytenr, u64 num_bytes, u64 orig_parent,
 			  u64 parent, u64 orig_ref_root, u64 ref_root,
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index feca04197d02..f3b287c22caf 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -74,6 +74,11 @@ struct async_submit_bio {
 	int rw;
 	int mirror_num;
 	unsigned long bio_flags;
+	/*
+	 * bio_offset is optional, can be used if the pages in the bio
+	 * can't tell us where in the file the bio should go
+	 */
+	u64 bio_offset;
 	struct btrfs_work work;
 };
 
@@ -534,7 +539,8 @@ static void run_one_async_start(struct btrfs_work *work)
 	async = container_of(work, struct  async_submit_bio, work);
 	fs_info = BTRFS_I(async->inode)->root->fs_info;
 	async->submit_bio_start(async->inode, async->rw, async->bio,
-			       async->mirror_num, async->bio_flags);
+			       async->mirror_num, async->bio_flags,
+			       async->bio_offset);
 }
 
 static void run_one_async_done(struct btrfs_work *work)
@@ -556,7 +562,8 @@ static void run_one_async_done(struct btrfs_work *work)
 		wake_up(&fs_info->async_submit_wait);
 
 	async->submit_bio_done(async->inode, async->rw, async->bio,
-			       async->mirror_num, async->bio_flags);
+			       async->mirror_num, async->bio_flags,
+			       async->bio_offset);
 }
 
 static void run_one_async_free(struct btrfs_work *work)
@@ -570,6 +577,7 @@ static void run_one_async_free(struct btrfs_work *work)
 int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
 			int rw, struct bio *bio, int mirror_num,
 			unsigned long bio_flags,
+			u64 bio_offset,
 			extent_submit_bio_hook_t *submit_bio_start,
 			extent_submit_bio_hook_t *submit_bio_done)
 {
@@ -592,6 +600,7 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
 
 	async->work.flags = 0;
 	async->bio_flags = bio_flags;
+	async->bio_offset = bio_offset;
 
 	atomic_inc(&fs_info->nr_async_submits);
 
@@ -627,7 +636,8 @@ static int btree_csum_one_bio(struct bio *bio)
 
 static int __btree_submit_bio_start(struct inode *inode, int rw,
 				    struct bio *bio, int mirror_num,
-				    unsigned long bio_flags)
+				    unsigned long bio_flags,
+				    u64 bio_offset)
 {
 	/*
 	 * when we're called for a write, we're already in the async
@@ -638,7 +648,8 @@ static int __btree_submit_bio_start(struct inode *inode, int rw,
 }
 
 static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
-				 int mirror_num, unsigned long bio_flags)
+				 int mirror_num, unsigned long bio_flags,
+				 u64 bio_offset)
 {
 	/*
 	 * when we're called for a write, we're already in the async
@@ -648,7 +659,8 @@ static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
 }
 
 static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
-				 int mirror_num, unsigned long bio_flags)
+				 int mirror_num, unsigned long bio_flags,
+				 u64 bio_offset)
 {
 	int ret;
 
@@ -671,6 +683,7 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 	 */
 	return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
 				   inode, rw, bio, mirror_num, 0,
+				   bio_offset,
 				   __btree_submit_bio_start,
 				   __btree_submit_bio_done);
 }
@@ -894,7 +907,8 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	root->ref_cows = 0;
 	root->track_dirty = 0;
 	root->in_radix = 0;
-	root->clean_orphans = 0;
+	root->orphan_item_inserted = 0;
+	root->orphan_cleanup_state = 0;
 
 	root->fs_info = fs_info;
 	root->objectid = objectid;
@@ -903,13 +917,16 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	root->name = NULL;
 	root->in_sysfs = 0;
 	root->inode_tree = RB_ROOT;
+	root->block_rsv = NULL;
+	root->orphan_block_rsv = NULL;
 
 	INIT_LIST_HEAD(&root->dirty_list);
 	INIT_LIST_HEAD(&root->orphan_list);
 	INIT_LIST_HEAD(&root->root_list);
 	spin_lock_init(&root->node_lock);
-	spin_lock_init(&root->list_lock);
+	spin_lock_init(&root->orphan_lock);
 	spin_lock_init(&root->inode_lock);
+	spin_lock_init(&root->accounting_lock);
 	mutex_init(&root->objectid_mutex);
 	mutex_init(&root->log_mutex);
 	init_waitqueue_head(&root->log_writer_wait);
@@ -968,42 +985,6 @@ static int find_and_setup_root(struct btrfs_root *tree_root,
 	return 0;
 }
 
-int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
-			     struct btrfs_fs_info *fs_info)
-{
-	struct extent_buffer *eb;
-	struct btrfs_root *log_root_tree = fs_info->log_root_tree;
-	u64 start = 0;
-	u64 end = 0;
-	int ret;
-
-	if (!log_root_tree)
-		return 0;
-
-	while (1) {
-		ret = find_first_extent_bit(&log_root_tree->dirty_log_pages,
-				0, &start, &end, EXTENT_DIRTY | EXTENT_NEW);
-		if (ret)
-			break;
-
-		clear_extent_bits(&log_root_tree->dirty_log_pages, start, end,
-				  EXTENT_DIRTY | EXTENT_NEW, GFP_NOFS);
-	}
-	eb = fs_info->log_root_tree->node;
-
-	WARN_ON(btrfs_header_level(eb) != 0);
-	WARN_ON(btrfs_header_nritems(eb) != 0);
-
-	ret = btrfs_free_reserved_extent(fs_info->tree_root,
-				eb->start, eb->len);
-	BUG_ON(ret);
-
-	free_extent_buffer(eb);
-	kfree(fs_info->log_root_tree);
-	fs_info->log_root_tree = NULL;
-	return 0;
-}
-
 static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans,
 					 struct btrfs_fs_info *fs_info)
 {
@@ -1191,19 +1172,23 @@ again:
 	if (root)
 		return root;
 
-	ret = btrfs_find_orphan_item(fs_info->tree_root, location->objectid);
-	if (ret == 0)
-		ret = -ENOENT;
-	if (ret < 0)
-		return ERR_PTR(ret);
-
 	root = btrfs_read_fs_root_no_radix(fs_info->tree_root, location);
 	if (IS_ERR(root))
 		return root;
 
-	WARN_ON(btrfs_root_refs(&root->root_item) == 0);
 	set_anon_super(&root->anon_super, NULL);
 
+	if (btrfs_root_refs(&root->root_item) == 0) {
+		ret = -ENOENT;
+		goto fail;
+	}
+
+	ret = btrfs_find_orphan_item(fs_info->tree_root, location->objectid);
+	if (ret < 0)
+		goto fail;
+	if (ret == 0)
+		root->orphan_item_inserted = 1;
+
 	ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
 	if (ret)
 		goto fail;
@@ -1212,10 +1197,9 @@ again:
 	ret = radix_tree_insert(&fs_info->fs_roots_radix,
 				(unsigned long)root->root_key.objectid,
 				root);
-	if (ret == 0) {
+	if (ret == 0)
 		root->in_radix = 1;
-		root->clean_orphans = 1;
-	}
+
 	spin_unlock(&fs_info->fs_roots_radix_lock);
 	radix_tree_preload_end();
 	if (ret) {
@@ -1461,10 +1445,6 @@ static int cleaner_kthread(void *arg)
 	struct btrfs_root *root = arg;
 
 	do {
-		smp_mb();
-		if (root->fs_info->closing)
-			break;
-
 		vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE);
 
 		if (!(root->fs_info->sb->s_flags & MS_RDONLY) &&
@@ -1477,11 +1457,9 @@ static int cleaner_kthread(void *arg)
 		if (freezing(current)) {
 			refrigerator();
 		} else {
-			smp_mb();
-			if (root->fs_info->closing)
-				break;
 			set_current_state(TASK_INTERRUPTIBLE);
-			schedule();
+			if (!kthread_should_stop())
+				schedule();
 			__set_current_state(TASK_RUNNING);
 		}
 	} while (!kthread_should_stop());
@@ -1493,36 +1471,40 @@ static int transaction_kthread(void *arg)
 	struct btrfs_root *root = arg;
 	struct btrfs_trans_handle *trans;
 	struct btrfs_transaction *cur;
+	u64 transid;
 	unsigned long now;
 	unsigned long delay;
 	int ret;
 
 	do {
-		smp_mb();
-		if (root->fs_info->closing)
-			break;
-
 		delay = HZ * 30;
 		vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE);
 		mutex_lock(&root->fs_info->transaction_kthread_mutex);
 
-		mutex_lock(&root->fs_info->trans_mutex);
+		spin_lock(&root->fs_info->new_trans_lock);
 		cur = root->fs_info->running_transaction;
 		if (!cur) {
-			mutex_unlock(&root->fs_info->trans_mutex);
+			spin_unlock(&root->fs_info->new_trans_lock);
 			goto sleep;
 		}
 
 		now = get_seconds();
-		if (now < cur->start_time || now - cur->start_time < 30) {
-			mutex_unlock(&root->fs_info->trans_mutex);
+		if (!cur->blocked &&
+		    (now < cur->start_time || now - cur->start_time < 30)) {
+			spin_unlock(&root->fs_info->new_trans_lock);
 			delay = HZ * 5;
 			goto sleep;
 		}
-		mutex_unlock(&root->fs_info->trans_mutex);
-		trans = btrfs_start_transaction(root, 1);
-		ret = btrfs_commit_transaction(trans, root);
+		transid = cur->transid;
+		spin_unlock(&root->fs_info->new_trans_lock);
 
+		trans = btrfs_join_transaction(root, 1);
+		if (transid == trans->transid) {
+			ret = btrfs_commit_transaction(trans, root);
+			BUG_ON(ret);
+		} else {
+			btrfs_end_transaction(trans, root);
+		}
 sleep:
 		wake_up_process(root->fs_info->cleaner_kthread);
 		mutex_unlock(&root->fs_info->transaction_kthread_mutex);
@@ -1530,10 +1512,10 @@ sleep:
 		if (freezing(current)) {
 			refrigerator();
 		} else {
-			if (root->fs_info->closing)
-				break;
 			set_current_state(TASK_INTERRUPTIBLE);
-			schedule_timeout(delay);
+			if (!kthread_should_stop() &&
+			    !btrfs_transaction_blocked(root->fs_info))
+				schedule_timeout(delay);
 			__set_current_state(TASK_RUNNING);
 		}
 	} while (!kthread_should_stop());
@@ -1620,6 +1602,13 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots);
 	INIT_LIST_HEAD(&fs_info->space_info);
 	btrfs_mapping_init(&fs_info->mapping_tree);
+	btrfs_init_block_rsv(&fs_info->global_block_rsv);
+	btrfs_init_block_rsv(&fs_info->delalloc_block_rsv);
+	btrfs_init_block_rsv(&fs_info->trans_block_rsv);
+	btrfs_init_block_rsv(&fs_info->chunk_block_rsv);
+	btrfs_init_block_rsv(&fs_info->empty_block_rsv);
+	INIT_LIST_HEAD(&fs_info->durable_block_rsv_list);
+	mutex_init(&fs_info->durable_block_rsv_mutex);
 	atomic_set(&fs_info->nr_async_submits, 0);
 	atomic_set(&fs_info->async_delalloc_pages, 0);
 	atomic_set(&fs_info->async_submit_draining, 0);
@@ -1759,9 +1748,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 			   min_t(u64, fs_devices->num_devices,
 			   fs_info->thread_pool_size),
 			   &fs_info->generic_worker);
-	btrfs_init_workers(&fs_info->enospc_workers, "enospc",
-			   fs_info->thread_pool_size,
-			   &fs_info->generic_worker);
 
 	/* a higher idle thresh on the submit workers makes it much more
 	 * likely that bios will be send down in a sane order to the
@@ -1809,7 +1795,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	btrfs_start_workers(&fs_info->endio_meta_workers, 1);
 	btrfs_start_workers(&fs_info->endio_meta_write_workers, 1);
 	btrfs_start_workers(&fs_info->endio_write_workers, 1);
-	btrfs_start_workers(&fs_info->enospc_workers, 1);
 
 	fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super);
 	fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages,
@@ -1912,17 +1897,18 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	csum_root->track_dirty = 1;
 
+	fs_info->generation = generation;
+	fs_info->last_trans_committed = generation;
+	fs_info->data_alloc_profile = (u64)-1;
+	fs_info->metadata_alloc_profile = (u64)-1;
+	fs_info->system_alloc_profile = fs_info->metadata_alloc_profile;
+
 	ret = btrfs_read_block_groups(extent_root);
 	if (ret) {
 		printk(KERN_ERR "Failed to read block groups: %d\n", ret);
 		goto fail_block_groups;
 	}
 
-	fs_info->generation = generation;
-	fs_info->last_trans_committed = generation;
-	fs_info->data_alloc_profile = (u64)-1;
-	fs_info->metadata_alloc_profile = (u64)-1;
-	fs_info->system_alloc_profile = fs_info->metadata_alloc_profile;
 	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root,
 					       "btrfs-cleaner");
 	if (IS_ERR(fs_info->cleaner_kthread))
@@ -1977,6 +1963,9 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	BUG_ON(ret);
 
 	if (!(sb->s_flags & MS_RDONLY)) {
+		ret = btrfs_cleanup_fs_roots(fs_info);
+		BUG_ON(ret);
+
 		ret = btrfs_recover_relocation(tree_root);
 		if (ret < 0) {
 			printk(KERN_WARNING
@@ -2040,7 +2029,6 @@ fail_sb_buffer:
 	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->submit_workers);
-	btrfs_stop_workers(&fs_info->enospc_workers);
 fail_iput:
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 	iput(fs_info->btree_inode);
@@ -2405,11 +2393,11 @@ int btrfs_commit_super(struct btrfs_root *root)
 	down_write(&root->fs_info->cleanup_work_sem);
 	up_write(&root->fs_info->cleanup_work_sem);
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_join_transaction(root, 1);
 	ret = btrfs_commit_transaction(trans, root);
 	BUG_ON(ret);
 	/* run commit again to drop the original snapshot */
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_join_transaction(root, 1);
 	btrfs_commit_transaction(trans, root);
 	ret = btrfs_write_and_wait_transaction(NULL, root);
 	BUG_ON(ret);
@@ -2426,15 +2414,15 @@ int close_ctree(struct btrfs_root *root)
 	fs_info->closing = 1;
 	smp_mb();
 
-	kthread_stop(root->fs_info->transaction_kthread);
-	kthread_stop(root->fs_info->cleaner_kthread);
-
 	if (!(fs_info->sb->s_flags & MS_RDONLY)) {
 		ret =  btrfs_commit_super(root);
 		if (ret)
 			printk(KERN_ERR "btrfs: commit super ret %d\n", ret);
 	}
 
+	kthread_stop(root->fs_info->transaction_kthread);
+	kthread_stop(root->fs_info->cleaner_kthread);
+
 	fs_info->closing = 2;
 	smp_mb();
 
@@ -2473,7 +2461,6 @@ int close_ctree(struct btrfs_root *root)
 	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->submit_workers);
-	btrfs_stop_workers(&fs_info->enospc_workers);
 
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index c958ecbc1916..88e825a0bf21 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -87,7 +87,7 @@ int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,
 			int metadata);
 int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
 			int rw, struct bio *bio, int mirror_num,
-			unsigned long bio_flags,
+			unsigned long bio_flags, u64 bio_offset,
 			extent_submit_bio_hook_t *submit_bio_start,
 			extent_submit_bio_hook_t *submit_bio_done);
 
@@ -95,8 +95,6 @@ int btrfs_congested_async(struct btrfs_fs_info *info, int iodone);
 unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info);
 int btrfs_write_tree_block(struct extent_buffer *buf);
 int btrfs_wait_tree_block_writeback(struct extent_buffer *buf);
-int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
-			     struct btrfs_fs_info *fs_info);
 int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
 			     struct btrfs_fs_info *fs_info);
 int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index c6a4f459ad76..b9080d71991a 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -35,10 +35,9 @@
 
 static int update_block_group(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
-			      u64 bytenr, u64 num_bytes, int alloc,
-			      int mark_free);
-static int update_reserved_extents(struct btrfs_block_group_cache *cache,
-				   u64 num_bytes, int reserve);
+			      u64 bytenr, u64 num_bytes, int alloc);
+static int update_reserved_bytes(struct btrfs_block_group_cache *cache,
+				 u64 num_bytes, int reserve, int sinfo);
 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root,
 				u64 bytenr, u64 num_bytes, u64 parent,
@@ -61,12 +60,6 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force);
-static int pin_down_bytes(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root,
-			  struct btrfs_path *path,
-			  u64 bytenr, u64 num_bytes,
-			  int is_data, int reserved,
-			  struct extent_buffer **must_clean);
 static int find_next_key(struct btrfs_path *path, int level,
 			 struct btrfs_key *key);
 static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
@@ -91,8 +84,12 @@ void btrfs_get_block_group(struct btrfs_block_group_cache *cache)
 
 void btrfs_put_block_group(struct btrfs_block_group_cache *cache)
 {
-	if (atomic_dec_and_test(&cache->count))
+	if (atomic_dec_and_test(&cache->count)) {
+		WARN_ON(cache->pinned > 0);
+		WARN_ON(cache->reserved > 0);
+		WARN_ON(cache->reserved_pinned > 0);
 		kfree(cache);
+	}
 }
 
 /*
@@ -319,7 +316,7 @@ static int caching_kthread(void *data)
 
 	exclude_super_stripes(extent_root, block_group);
 	spin_lock(&block_group->space_info->lock);
-	block_group->space_info->bytes_super += block_group->bytes_super;
+	block_group->space_info->bytes_readonly += block_group->bytes_super;
 	spin_unlock(&block_group->space_info->lock);
 
 	last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
@@ -507,6 +504,9 @@ static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
 	struct list_head *head = &info->space_info;
 	struct btrfs_space_info *found;
 
+	flags &= BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_SYSTEM |
+		 BTRFS_BLOCK_GROUP_METADATA;
+
 	rcu_read_lock();
 	list_for_each_entry_rcu(found, head, list) {
 		if (found->flags == flags) {
@@ -610,6 +610,113 @@ int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len)
 }
 
 /*
+ * helper function to lookup reference count and flags of extent.
+ *
+ * the head node for delayed ref is used to store the sum of all the
+ * reference count modifications queued up in the rbtree. the head
+ * node may also store the extent flags to set. This way you can check
+ * to see what the reference count and extent flags would be if all of
+ * the delayed refs are not processed.
+ */
+int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, u64 bytenr,
+			     u64 num_bytes, u64 *refs, u64 *flags)
+{
+	struct btrfs_delayed_ref_head *head;
+	struct btrfs_delayed_ref_root *delayed_refs;
+	struct btrfs_path *path;
+	struct btrfs_extent_item *ei;
+	struct extent_buffer *leaf;
+	struct btrfs_key key;
+	u32 item_size;
+	u64 num_refs;
+	u64 extent_flags;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = bytenr;
+	key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.offset = num_bytes;
+	if (!trans) {
+		path->skip_locking = 1;
+		path->search_commit_root = 1;
+	}
+again:
+	ret = btrfs_search_slot(trans, root->fs_info->extent_root,
+				&key, path, 0, 0);
+	if (ret < 0)
+		goto out_free;
+
+	if (ret == 0) {
+		leaf = path->nodes[0];
+		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+		if (item_size >= sizeof(*ei)) {
+			ei = btrfs_item_ptr(leaf, path->slots[0],
+					    struct btrfs_extent_item);
+			num_refs = btrfs_extent_refs(leaf, ei);
+			extent_flags = btrfs_extent_flags(leaf, ei);
+		} else {
+#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
+			struct btrfs_extent_item_v0 *ei0;
+			BUG_ON(item_size != sizeof(*ei0));
+			ei0 = btrfs_item_ptr(leaf, path->slots[0],
+					     struct btrfs_extent_item_v0);
+			num_refs = btrfs_extent_refs_v0(leaf, ei0);
+			/* FIXME: this isn't correct for data */
+			extent_flags = BTRFS_BLOCK_FLAG_FULL_BACKREF;
+#else
+			BUG();
+#endif
+		}
+		BUG_ON(num_refs == 0);
+	} else {
+		num_refs = 0;
+		extent_flags = 0;
+		ret = 0;
+	}
+
+	if (!trans)
+		goto out;
+
+	delayed_refs = &trans->transaction->delayed_refs;
+	spin_lock(&delayed_refs->lock);
+	head = btrfs_find_delayed_ref_head(trans, bytenr);
+	if (head) {
+		if (!mutex_trylock(&head->mutex)) {
+			atomic_inc(&head->node.refs);
+			spin_unlock(&delayed_refs->lock);
+
+			btrfs_release_path(root->fs_info->extent_root, path);
+
+			mutex_lock(&head->mutex);
+			mutex_unlock(&head->mutex);
+			btrfs_put_delayed_ref(&head->node);
+			goto again;
+		}
+		if (head->extent_op && head->extent_op->update_flags)
+			extent_flags |= head->extent_op->flags_to_set;
+		else
+			BUG_ON(num_refs == 0);
+
+		num_refs += head->node.ref_mod;
+		mutex_unlock(&head->mutex);
+	}
+	spin_unlock(&delayed_refs->lock);
+out:
+	WARN_ON(num_refs == 0);
+	if (refs)
+		*refs = num_refs;
+	if (flags)
+		*flags = extent_flags;
+out_free:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
  * Back reference rules.  Back refs have three main goals:
  *
  * 1) differentiate between all holders of references to an extent so that
@@ -1871,7 +1978,6 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-
 /* helper function to actually process a single delayed ref entry */
 static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root,
@@ -1891,32 +1997,14 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 		BUG_ON(extent_op);
 		head = btrfs_delayed_node_to_head(node);
 		if (insert_reserved) {
-			int mark_free = 0;
-			struct extent_buffer *must_clean = NULL;
-
-			ret = pin_down_bytes(trans, root, NULL,
-					     node->bytenr, node->num_bytes,
-					     head->is_data, 1, &must_clean);
-			if (ret > 0)
-				mark_free = 1;
-
-			if (must_clean) {
-				clean_tree_block(NULL, root, must_clean);
-				btrfs_tree_unlock(must_clean);
-				free_extent_buffer(must_clean);
-			}
+			btrfs_pin_extent(root, node->bytenr,
+					 node->num_bytes, 1);
 			if (head->is_data) {
 				ret = btrfs_del_csums(trans, root,
 						      node->bytenr,
 						      node->num_bytes);
 				BUG_ON(ret);
 			}
-			if (mark_free) {
-				ret = btrfs_free_reserved_extent(root,
-							node->bytenr,
-							node->num_bytes);
-				BUG_ON(ret);
-			}
 		}
 		mutex_unlock(&head->mutex);
 		return 0;
@@ -2347,6 +2435,8 @@ int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
 		ret = 0;
 out:
 	btrfs_free_path(path);
+	if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID)
+		WARN_ON(ret > 0);
 	return ret;
 }
 
@@ -2660,12 +2750,21 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 			     struct btrfs_space_info **space_info)
 {
 	struct btrfs_space_info *found;
+	int i;
+	int factor;
+
+	if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 |
+		     BTRFS_BLOCK_GROUP_RAID10))
+		factor = 2;
+	else
+		factor = 1;
 
 	found = __find_space_info(info, flags);
 	if (found) {
 		spin_lock(&found->lock);
 		found->total_bytes += total_bytes;
 		found->bytes_used += bytes_used;
+		found->disk_used += bytes_used * factor;
 		found->full = 0;
 		spin_unlock(&found->lock);
 		*space_info = found;
@@ -2675,18 +2774,20 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	if (!found)
 		return -ENOMEM;
 
-	INIT_LIST_HEAD(&found->block_groups);
+	for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
+		INIT_LIST_HEAD(&found->block_groups[i]);
 	init_rwsem(&found->groups_sem);
-	init_waitqueue_head(&found->flush_wait);
-	init_waitqueue_head(&found->allocate_wait);
 	spin_lock_init(&found->lock);
-	found->flags = flags;
+	found->flags = flags & (BTRFS_BLOCK_GROUP_DATA |
+				BTRFS_BLOCK_GROUP_SYSTEM |
+				BTRFS_BLOCK_GROUP_METADATA);
 	found->total_bytes = total_bytes;
 	found->bytes_used = bytes_used;
+	found->disk_used = bytes_used * factor;
 	found->bytes_pinned = 0;
 	found->bytes_reserved = 0;
 	found->bytes_readonly = 0;
-	found->bytes_delalloc = 0;
+	found->bytes_may_use = 0;
 	found->full = 0;
 	found->force_alloc = 0;
 	*space_info = found;
@@ -2711,19 +2812,6 @@ static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
 	}
 }
 
-static void set_block_group_readonly(struct btrfs_block_group_cache *cache)
-{
-	spin_lock(&cache->space_info->lock);
-	spin_lock(&cache->lock);
-	if (!cache->ro) {
-		cache->space_info->bytes_readonly += cache->key.offset -
-					btrfs_block_group_used(&cache->item);
-		cache->ro = 1;
-	}
-	spin_unlock(&cache->lock);
-	spin_unlock(&cache->space_info->lock);
-}
-
 u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 {
 	u64 num_devices = root->fs_info->fs_devices->rw_devices;
@@ -2752,491 +2840,50 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 	return flags;
 }
 
-static u64 btrfs_get_alloc_profile(struct btrfs_root *root, u64 data)
-{
-	struct btrfs_fs_info *info = root->fs_info;
-	u64 alloc_profile;
-
-	if (data) {
-		alloc_profile = info->avail_data_alloc_bits &
-			info->data_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
-	} else if (root == root->fs_info->chunk_root) {
-		alloc_profile = info->avail_system_alloc_bits &
-			info->system_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
-	} else {
-		alloc_profile = info->avail_metadata_alloc_bits &
-			info->metadata_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
-	}
-
-	return btrfs_reduce_alloc_profile(root, data);
-}
-
-void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode)
-{
-	u64 alloc_target;
-
-	alloc_target = btrfs_get_alloc_profile(root, 1);
-	BTRFS_I(inode)->space_info = __find_space_info(root->fs_info,
-						       alloc_target);
-}
-
-static u64 calculate_bytes_needed(struct btrfs_root *root, int num_items)
-{
-	u64 num_bytes;
-	int level;
-
-	level = BTRFS_MAX_LEVEL - 2;
-	/*
-	 * NOTE: these calculations are absolutely the worst possible case.
-	 * This assumes that _every_ item we insert will require a new leaf, and
-	 * that the tree has grown to its maximum level size.
-	 */
-
-	/*
-	 * for every item we insert we could insert both an extent item and a
-	 * extent ref item.  Then for ever item we insert, we will need to cow
-	 * both the original leaf, plus the leaf to the left and right of it.
-	 *
-	 * Unless we are talking about the extent root, then we just want the
-	 * number of items * 2, since we just need the extent item plus its ref.
-	 */
-	if (root == root->fs_info->extent_root)
-		num_bytes = num_items * 2;
-	else
-		num_bytes = (num_items + (2 * num_items)) * 3;
-
-	/*
-	 * num_bytes is total number of leaves we could need times the leaf
-	 * size, and then for every leaf we could end up cow'ing 2 nodes per
-	 * level, down to the leaf level.
-	 */
-	num_bytes = (num_bytes * root->leafsize) +
-		(num_bytes * (level * 2)) * root->nodesize;
-
-	return num_bytes;
-}
-
-/*
- * Unreserve metadata space for delalloc.  If we have less reserved credits than
- * we have extents, this function does nothing.
- */
-int btrfs_unreserve_metadata_for_delalloc(struct btrfs_root *root,
-					  struct inode *inode, int num_items)
-{
-	struct btrfs_fs_info *info = root->fs_info;
-	struct btrfs_space_info *meta_sinfo;
-	u64 num_bytes;
-	u64 alloc_target;
-	bool bug = false;
-
-	/* get the space info for where the metadata will live */
-	alloc_target = btrfs_get_alloc_profile(root, 0);
-	meta_sinfo = __find_space_info(info, alloc_target);
-
-	num_bytes = calculate_bytes_needed(root->fs_info->extent_root,
-					   num_items);
-
-	spin_lock(&meta_sinfo->lock);
-	spin_lock(&BTRFS_I(inode)->accounting_lock);
-	if (BTRFS_I(inode)->reserved_extents <=
-	    BTRFS_I(inode)->outstanding_extents) {
-		spin_unlock(&BTRFS_I(inode)->accounting_lock);
-		spin_unlock(&meta_sinfo->lock);
-		return 0;
-	}
-	spin_unlock(&BTRFS_I(inode)->accounting_lock);
-
-	BTRFS_I(inode)->reserved_extents -= num_items;
-	BUG_ON(BTRFS_I(inode)->reserved_extents < 0);
-
-	if (meta_sinfo->bytes_delalloc < num_bytes) {
-		bug = true;
-		meta_sinfo->bytes_delalloc = 0;
-	} else {
-		meta_sinfo->bytes_delalloc -= num_bytes;
-	}
-	spin_unlock(&meta_sinfo->lock);
-
-	BUG_ON(bug);
-
-	return 0;
-}
-
-static void check_force_delalloc(struct btrfs_space_info *meta_sinfo)
+static u64 get_alloc_profile(struct btrfs_root *root, u64 flags)
 {
-	u64 thresh;
-
-	thresh = meta_sinfo->bytes_used + meta_sinfo->bytes_reserved +
-		meta_sinfo->bytes_pinned + meta_sinfo->bytes_readonly +
-		meta_sinfo->bytes_super + meta_sinfo->bytes_root +
-		meta_sinfo->bytes_may_use;
-
-	thresh = meta_sinfo->total_bytes - thresh;
-	thresh *= 80;
-	do_div(thresh, 100);
-	if (thresh <= meta_sinfo->bytes_delalloc)
-		meta_sinfo->force_delalloc = 1;
-	else
-		meta_sinfo->force_delalloc = 0;
+	if (flags & BTRFS_BLOCK_GROUP_DATA)
+		flags |= root->fs_info->avail_data_alloc_bits &
+			 root->fs_info->data_alloc_profile;
+	else if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
+		flags |= root->fs_info->avail_system_alloc_bits &
+			 root->fs_info->system_alloc_profile;
+	else if (flags & BTRFS_BLOCK_GROUP_METADATA)
+		flags |= root->fs_info->avail_metadata_alloc_bits &
+			 root->fs_info->metadata_alloc_profile;
+	return btrfs_reduce_alloc_profile(root, flags);
 }
 
-struct async_flush {
-	struct btrfs_root *root;
-	struct btrfs_space_info *info;
-	struct btrfs_work work;
-};
-
-static noinline void flush_delalloc_async(struct btrfs_work *work)
+static u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data)
 {
-	struct async_flush *async;
-	struct btrfs_root *root;
-	struct btrfs_space_info *info;
-
-	async = container_of(work, struct async_flush, work);
-	root = async->root;
-	info = async->info;
-
-	btrfs_start_delalloc_inodes(root, 0);
-	wake_up(&info->flush_wait);
-	btrfs_wait_ordered_extents(root, 0, 0);
-
-	spin_lock(&info->lock);
-	info->flushing = 0;
-	spin_unlock(&info->lock);
-	wake_up(&info->flush_wait);
-
-	kfree(async);
-}
-
-static void wait_on_flush(struct btrfs_space_info *info)
-{
-	DEFINE_WAIT(wait);
-	u64 used;
-
-	while (1) {
-		prepare_to_wait(&info->flush_wait, &wait,
-				TASK_UNINTERRUPTIBLE);
-		spin_lock(&info->lock);
-		if (!info->flushing) {
-			spin_unlock(&info->lock);
-			break;
-		}
-
-		used = info->bytes_used + info->bytes_reserved +
-			info->bytes_pinned + info->bytes_readonly +
-			info->bytes_super + info->bytes_root +
-			info->bytes_may_use + info->bytes_delalloc;
-		if (used < info->total_bytes) {
-			spin_unlock(&info->lock);
-			break;
-		}
-		spin_unlock(&info->lock);
-		schedule();
-	}
-	finish_wait(&info->flush_wait, &wait);
-}
-
-static void flush_delalloc(struct btrfs_root *root,
-				 struct btrfs_space_info *info)
-{
-	struct async_flush *async;
-	bool wait = false;
-
-	spin_lock(&info->lock);
+	u64 flags;
 
-	if (!info->flushing)
-		info->flushing = 1;
+	if (data)
+		flags = BTRFS_BLOCK_GROUP_DATA;
+	else if (root == root->fs_info->chunk_root)
+		flags = BTRFS_BLOCK_GROUP_SYSTEM;
 	else
-		wait = true;
-
-	spin_unlock(&info->lock);
-
-	if (wait) {
-		wait_on_flush(info);
-		return;
-	}
-
-	async = kzalloc(sizeof(*async), GFP_NOFS);
-	if (!async)
-		goto flush;
-
-	async->root = root;
-	async->info = info;
-	async->work.func = flush_delalloc_async;
+		flags = BTRFS_BLOCK_GROUP_METADATA;
 
-	btrfs_queue_worker(&root->fs_info->enospc_workers,
-			   &async->work);
-	wait_on_flush(info);
-	return;
-
-flush:
-	btrfs_start_delalloc_inodes(root, 0);
-	btrfs_wait_ordered_extents(root, 0, 0);
-
-	spin_lock(&info->lock);
-	info->flushing = 0;
-	spin_unlock(&info->lock);
-	wake_up(&info->flush_wait);
+	return get_alloc_profile(root, flags);
 }
 
-static int maybe_allocate_chunk(struct btrfs_root *root,
-				 struct btrfs_space_info *info)
-{
-	struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
-	struct btrfs_trans_handle *trans;
-	bool wait = false;
-	int ret = 0;
-	u64 min_metadata;
-	u64 free_space;
-
-	free_space = btrfs_super_total_bytes(disk_super);
-	/*
-	 * we allow the metadata to grow to a max of either 10gb or 5% of the
-	 * space in the volume.
-	 */
-	min_metadata = min((u64)10 * 1024 * 1024 * 1024,
-			     div64_u64(free_space * 5, 100));
-	if (info->total_bytes >= min_metadata) {
-		spin_unlock(&info->lock);
-		return 0;
-	}
-
-	if (info->full) {
-		spin_unlock(&info->lock);
-		return 0;
-	}
-
-	if (!info->allocating_chunk) {
-		info->force_alloc = 1;
-		info->allocating_chunk = 1;
-	} else {
-		wait = true;
-	}
-
-	spin_unlock(&info->lock);
-
-	if (wait) {
-		wait_event(info->allocate_wait,
-			   !info->allocating_chunk);
-		return 1;
-	}
-
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
-	ret = do_chunk_alloc(trans, root->fs_info->extent_root,
-			     4096 + 2 * 1024 * 1024,
-			     info->flags, 0);
-	btrfs_end_transaction(trans, root);
-	if (ret)
-		goto out;
-out:
-	spin_lock(&info->lock);
-	info->allocating_chunk = 0;
-	spin_unlock(&info->lock);
-	wake_up(&info->allocate_wait);
-
-	if (ret)
-		return 0;
-	return 1;
-}
-
-/*
- * Reserve metadata space for delalloc.
- */
-int btrfs_reserve_metadata_for_delalloc(struct btrfs_root *root,
-					struct inode *inode, int num_items)
-{
-	struct btrfs_fs_info *info = root->fs_info;
-	struct btrfs_space_info *meta_sinfo;
-	u64 num_bytes;
-	u64 used;
-	u64 alloc_target;
-	int flushed = 0;
-	int force_delalloc;
-
-	/* get the space info for where the metadata will live */
-	alloc_target = btrfs_get_alloc_profile(root, 0);
-	meta_sinfo = __find_space_info(info, alloc_target);
-
-	num_bytes = calculate_bytes_needed(root->fs_info->extent_root,
-					   num_items);
-again:
-	spin_lock(&meta_sinfo->lock);
-
-	force_delalloc = meta_sinfo->force_delalloc;
-
-	if (unlikely(!meta_sinfo->bytes_root))
-		meta_sinfo->bytes_root = calculate_bytes_needed(root, 6);
-
-	if (!flushed)
-		meta_sinfo->bytes_delalloc += num_bytes;
-
-	used = meta_sinfo->bytes_used + meta_sinfo->bytes_reserved +
-		meta_sinfo->bytes_pinned + meta_sinfo->bytes_readonly +
-		meta_sinfo->bytes_super + meta_sinfo->bytes_root +
-		meta_sinfo->bytes_may_use + meta_sinfo->bytes_delalloc;
-
-	if (used > meta_sinfo->total_bytes) {
-		flushed++;
-
-		if (flushed == 1) {
-			if (maybe_allocate_chunk(root, meta_sinfo))
-				goto again;
-			flushed++;
-		} else {
-			spin_unlock(&meta_sinfo->lock);
-		}
-
-		if (flushed == 2) {
-			filemap_flush(inode->i_mapping);
-			goto again;
-		} else if (flushed == 3) {
-			flush_delalloc(root, meta_sinfo);
-			goto again;
-		}
-		spin_lock(&meta_sinfo->lock);
-		meta_sinfo->bytes_delalloc -= num_bytes;
-		spin_unlock(&meta_sinfo->lock);
-		printk(KERN_ERR "enospc, has %d, reserved %d\n",
-		       BTRFS_I(inode)->outstanding_extents,
-		       BTRFS_I(inode)->reserved_extents);
-		dump_space_info(meta_sinfo, 0, 0);
-		return -ENOSPC;
-	}
-
-	BTRFS_I(inode)->reserved_extents += num_items;
-	check_force_delalloc(meta_sinfo);
-	spin_unlock(&meta_sinfo->lock);
-
-	if (!flushed && force_delalloc)
-		filemap_flush(inode->i_mapping);
-
-	return 0;
-}
-
-/*
- * unreserve num_items number of items worth of metadata space.  This needs to
- * be paired with btrfs_reserve_metadata_space.
- *
- * NOTE: if you have the option, run this _AFTER_ you do a
- * btrfs_end_transaction, since btrfs_end_transaction will run delayed ref
- * oprations which will result in more used metadata, so we want to make sure we
- * can do that without issue.
- */
-int btrfs_unreserve_metadata_space(struct btrfs_root *root, int num_items)
-{
-	struct btrfs_fs_info *info = root->fs_info;
-	struct btrfs_space_info *meta_sinfo;
-	u64 num_bytes;
-	u64 alloc_target;
-	bool bug = false;
-
-	/* get the space info for where the metadata will live */
-	alloc_target = btrfs_get_alloc_profile(root, 0);
-	meta_sinfo = __find_space_info(info, alloc_target);
-
-	num_bytes = calculate_bytes_needed(root, num_items);
-
-	spin_lock(&meta_sinfo->lock);
-	if (meta_sinfo->bytes_may_use < num_bytes) {
-		bug = true;
-		meta_sinfo->bytes_may_use = 0;
-	} else {
-		meta_sinfo->bytes_may_use -= num_bytes;
-	}
-	spin_unlock(&meta_sinfo->lock);
-
-	BUG_ON(bug);
-
-	return 0;
-}
-
-/*
- * Reserve some metadata space for use.  We'll calculate the worste case number
- * of bytes that would be needed to modify num_items number of items.  If we
- * have space, fantastic, if not, you get -ENOSPC.  Please call
- * btrfs_unreserve_metadata_space when you are done for the _SAME_ number of
- * items you reserved, since whatever metadata you needed should have already
- * been allocated.
- *
- * This will commit the transaction to make more space if we don't have enough
- * metadata space.  THe only time we don't do this is if we're reserving space
- * inside of a transaction, then we will just return -ENOSPC and it is the
- * callers responsibility to handle it properly.
- */
-int btrfs_reserve_metadata_space(struct btrfs_root *root, int num_items)
+void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode)
 {
-	struct btrfs_fs_info *info = root->fs_info;
-	struct btrfs_space_info *meta_sinfo;
-	u64 num_bytes;
-	u64 used;
-	u64 alloc_target;
-	int retries = 0;
-
-	/* get the space info for where the metadata will live */
-	alloc_target = btrfs_get_alloc_profile(root, 0);
-	meta_sinfo = __find_space_info(info, alloc_target);
-
-	num_bytes = calculate_bytes_needed(root, num_items);
-again:
-	spin_lock(&meta_sinfo->lock);
-
-	if (unlikely(!meta_sinfo->bytes_root))
-		meta_sinfo->bytes_root = calculate_bytes_needed(root, 6);
-
-	if (!retries)
-		meta_sinfo->bytes_may_use += num_bytes;
-
-	used = meta_sinfo->bytes_used + meta_sinfo->bytes_reserved +
-		meta_sinfo->bytes_pinned + meta_sinfo->bytes_readonly +
-		meta_sinfo->bytes_super + meta_sinfo->bytes_root +
-		meta_sinfo->bytes_may_use + meta_sinfo->bytes_delalloc;
-
-	if (used > meta_sinfo->total_bytes) {
-		retries++;
-		if (retries == 1) {
-			if (maybe_allocate_chunk(root, meta_sinfo))
-				goto again;
-			retries++;
-		} else {
-			spin_unlock(&meta_sinfo->lock);
-		}
-
-		if (retries == 2) {
-			flush_delalloc(root, meta_sinfo);
-			goto again;
-		}
-		spin_lock(&meta_sinfo->lock);
-		meta_sinfo->bytes_may_use -= num_bytes;
-		spin_unlock(&meta_sinfo->lock);
-
-		dump_space_info(meta_sinfo, 0, 0);
-		return -ENOSPC;
-	}
-
-	check_force_delalloc(meta_sinfo);
-	spin_unlock(&meta_sinfo->lock);
-
-	return 0;
+	BTRFS_I(inode)->space_info = __find_space_info(root->fs_info,
+						       BTRFS_BLOCK_GROUP_DATA);
 }
 
 /*
  * This will check the space that the inode allocates from to make sure we have
  * enough space for bytes.
  */
-int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
-				u64 bytes)
+int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
 {
 	struct btrfs_space_info *data_sinfo;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
 	u64 used;
-	int ret = 0, committed = 0, flushed = 0;
+	int ret = 0, committed = 0;
 
 	/* make sure bytes are sectorsize aligned */
 	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
@@ -3248,21 +2895,13 @@ int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
 again:
 	/* make sure we have enough space to handle the data first */
 	spin_lock(&data_sinfo->lock);
-	used = data_sinfo->bytes_used + data_sinfo->bytes_delalloc +
-		data_sinfo->bytes_reserved + data_sinfo->bytes_pinned +
-		data_sinfo->bytes_readonly + data_sinfo->bytes_may_use +
-		data_sinfo->bytes_super;
+	used = data_sinfo->bytes_used + data_sinfo->bytes_reserved +
+		data_sinfo->bytes_pinned + data_sinfo->bytes_readonly +
+		data_sinfo->bytes_may_use;
 
 	if (used + bytes > data_sinfo->total_bytes) {
 		struct btrfs_trans_handle *trans;
 
-		if (!flushed) {
-			spin_unlock(&data_sinfo->lock);
-			flush_delalloc(root, data_sinfo);
-			flushed = 1;
-			goto again;
-		}
-
 		/*
 		 * if we don't have enough free bytes in this space then we need
 		 * to alloc a new chunk.
@@ -3274,15 +2913,15 @@ again:
 			spin_unlock(&data_sinfo->lock);
 alloc:
 			alloc_target = btrfs_get_alloc_profile(root, 1);
-			trans = btrfs_start_transaction(root, 1);
-			if (!trans)
-				return -ENOMEM;
+			trans = btrfs_join_transaction(root, 1);
+			if (IS_ERR(trans))
+				return PTR_ERR(trans);
 
 			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
 					     bytes + 2 * 1024 * 1024,
 					     alloc_target, 0);
 			btrfs_end_transaction(trans, root);
-			if (ret)
+			if (ret < 0)
 				return ret;
 
 			if (!data_sinfo) {
@@ -3297,25 +2936,26 @@ alloc:
 		if (!committed && !root->fs_info->open_ioctl_trans) {
 			committed = 1;
 			trans = btrfs_join_transaction(root, 1);
-			if (!trans)
-				return -ENOMEM;
+			if (IS_ERR(trans))
+				return PTR_ERR(trans);
 			ret = btrfs_commit_transaction(trans, root);
 			if (ret)
 				return ret;
 			goto again;
 		}
 
-		printk(KERN_ERR "no space left, need %llu, %llu delalloc bytes"
-		       ", %llu bytes_used, %llu bytes_reserved, "
-		       "%llu bytes_pinned, %llu bytes_readonly, %llu may use "
-		       "%llu total\n", (unsigned long long)bytes,
-		       (unsigned long long)data_sinfo->bytes_delalloc,
+#if 0 /* I hope we never need this code again, just in case */
+		printk(KERN_ERR "no space left, need %llu, %llu bytes_used, "
+		       "%llu bytes_reserved, " "%llu bytes_pinned, "
+		       "%llu bytes_readonly, %llu may use %llu total\n",
+		       (unsigned long long)bytes,
 		       (unsigned long long)data_sinfo->bytes_used,
 		       (unsigned long long)data_sinfo->bytes_reserved,
 		       (unsigned long long)data_sinfo->bytes_pinned,
 		       (unsigned long long)data_sinfo->bytes_readonly,
 		       (unsigned long long)data_sinfo->bytes_may_use,
 		       (unsigned long long)data_sinfo->total_bytes);
+#endif
 		return -ENOSPC;
 	}
 	data_sinfo->bytes_may_use += bytes;
@@ -3326,12 +2966,13 @@ alloc:
 }
 
 /*
- * if there was an error for whatever reason after calling
- * btrfs_check_data_free_space, call this so we can cleanup the counters.
+ * called when we are clearing an delalloc extent from the
+ * inode's io_tree or there was an error for whatever reason
+ * after calling btrfs_check_data_free_space
  */
-void btrfs_free_reserved_data_space(struct btrfs_root *root,
-				    struct inode *inode, u64 bytes)
+void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
 {
+	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_space_info *data_sinfo;
 
 	/* make sure bytes are sectorsize aligned */
@@ -3344,48 +2985,6 @@ void btrfs_free_reserved_data_space(struct btrfs_root *root,
 	spin_unlock(&data_sinfo->lock);
 }
 
-/* called when we are adding a delalloc extent to the inode's io_tree */
-void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
-				  u64 bytes)
-{
-	struct btrfs_space_info *data_sinfo;
-
-	/* get the space info for where this inode will be storing its data */
-	data_sinfo = BTRFS_I(inode)->space_info;
-
-	/* make sure we have enough space to handle the data first */
-	spin_lock(&data_sinfo->lock);
-	data_sinfo->bytes_delalloc += bytes;
-
-	/*
-	 * we are adding a delalloc extent without calling
-	 * btrfs_check_data_free_space first.  This happens on a weird
-	 * writepage condition, but shouldn't hurt our accounting
-	 */
-	if (unlikely(bytes > BTRFS_I(inode)->reserved_bytes)) {
-		data_sinfo->bytes_may_use -= BTRFS_I(inode)->reserved_bytes;
-		BTRFS_I(inode)->reserved_bytes = 0;
-	} else {
-		data_sinfo->bytes_may_use -= bytes;
-		BTRFS_I(inode)->reserved_bytes -= bytes;
-	}
-
-	spin_unlock(&data_sinfo->lock);
-}
-
-/* called when we are clearing an delalloc extent from the inode's io_tree */
-void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
-			      u64 bytes)
-{
-	struct btrfs_space_info *info;
-
-	info = BTRFS_I(inode)->space_info;
-
-	spin_lock(&info->lock);
-	info->bytes_delalloc -= bytes;
-	spin_unlock(&info->lock);
-}
-
 static void force_metadata_allocation(struct btrfs_fs_info *info)
 {
 	struct list_head *head = &info->space_info;
@@ -3399,13 +2998,28 @@ static void force_metadata_allocation(struct btrfs_fs_info *info)
 	rcu_read_unlock();
 }
 
+static int should_alloc_chunk(struct btrfs_space_info *sinfo,
+			      u64 alloc_bytes)
+{
+	u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly;
+
+	if (sinfo->bytes_used + sinfo->bytes_reserved +
+	    alloc_bytes + 256 * 1024 * 1024 < num_bytes)
+		return 0;
+
+	if (sinfo->bytes_used + sinfo->bytes_reserved +
+	    alloc_bytes < div_factor(num_bytes, 8))
+		return 0;
+
+	return 1;
+}
+
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force)
 {
 	struct btrfs_space_info *space_info;
 	struct btrfs_fs_info *fs_info = extent_root->fs_info;
-	u64 thresh;
 	int ret = 0;
 
 	mutex_lock(&fs_info->chunk_mutex);
@@ -3428,11 +3042,7 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 		goto out;
 	}
 
-	thresh = space_info->total_bytes - space_info->bytes_readonly;
-	thresh = div_factor(thresh, 8);
-	if (!force &&
-	   (space_info->bytes_used + space_info->bytes_pinned +
-	    space_info->bytes_reserved + alloc_bytes) < thresh) {
+	if (!force && !should_alloc_chunk(space_info, alloc_bytes)) {
 		spin_unlock(&space_info->lock);
 		goto out;
 	}
@@ -3454,6 +3064,8 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 	spin_lock(&space_info->lock);
 	if (ret)
 		space_info->full = 1;
+	else
+		ret = 1;
 	space_info->force_alloc = 0;
 	spin_unlock(&space_info->lock);
 out:
@@ -3461,13 +3073,713 @@ out:
 	return ret;
 }
 
+static int maybe_allocate_chunk(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root,
+				struct btrfs_space_info *sinfo, u64 num_bytes)
+{
+	int ret;
+	int end_trans = 0;
+
+	if (sinfo->full)
+		return 0;
+
+	spin_lock(&sinfo->lock);
+	ret = should_alloc_chunk(sinfo, num_bytes + 2 * 1024 * 1024);
+	spin_unlock(&sinfo->lock);
+	if (!ret)
+		return 0;
+
+	if (!trans) {
+		trans = btrfs_join_transaction(root, 1);
+		BUG_ON(IS_ERR(trans));
+		end_trans = 1;
+	}
+
+	ret = do_chunk_alloc(trans, root->fs_info->extent_root,
+			     num_bytes + 2 * 1024 * 1024,
+			     get_alloc_profile(root, sinfo->flags), 0);
+
+	if (end_trans)
+		btrfs_end_transaction(trans, root);
+
+	return ret == 1 ? 1 : 0;
+}
+
+/*
+ * shrink metadata reservation for delalloc
+ */
+static int shrink_delalloc(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, u64 to_reclaim)
+{
+	struct btrfs_block_rsv *block_rsv;
+	u64 reserved;
+	u64 max_reclaim;
+	u64 reclaimed = 0;
+	int pause = 1;
+	int ret;
+
+	block_rsv = &root->fs_info->delalloc_block_rsv;
+	spin_lock(&block_rsv->lock);
+	reserved = block_rsv->reserved;
+	spin_unlock(&block_rsv->lock);
+
+	if (reserved == 0)
+		return 0;
+
+	max_reclaim = min(reserved, to_reclaim);
+
+	while (1) {
+		ret = btrfs_start_one_delalloc_inode(root, trans ? 1 : 0);
+		if (!ret) {
+			__set_current_state(TASK_INTERRUPTIBLE);
+			schedule_timeout(pause);
+			pause <<= 1;
+			if (pause > HZ / 10)
+				pause = HZ / 10;
+		} else {
+			pause = 1;
+		}
+
+		spin_lock(&block_rsv->lock);
+		if (reserved > block_rsv->reserved)
+			reclaimed = reserved - block_rsv->reserved;
+		reserved = block_rsv->reserved;
+		spin_unlock(&block_rsv->lock);
+
+		if (reserved == 0 || reclaimed >= max_reclaim)
+			break;
+
+		if (trans && trans->transaction->blocked)
+			return -EAGAIN;
+	}
+	return reclaimed >= to_reclaim;
+}
+
+static int should_retry_reserve(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root,
+				struct btrfs_block_rsv *block_rsv,
+				u64 num_bytes, int *retries)
+{
+	struct btrfs_space_info *space_info = block_rsv->space_info;
+	int ret;
+
+	if ((*retries) > 2)
+		return -ENOSPC;
+
+	ret = maybe_allocate_chunk(trans, root, space_info, num_bytes);
+	if (ret)
+		return 1;
+
+	if (trans && trans->transaction->in_commit)
+		return -ENOSPC;
+
+	ret = shrink_delalloc(trans, root, num_bytes);
+	if (ret)
+		return ret;
+
+	spin_lock(&space_info->lock);
+	if (space_info->bytes_pinned < num_bytes)
+		ret = 1;
+	spin_unlock(&space_info->lock);
+	if (ret)
+		return -ENOSPC;
+
+	(*retries)++;
+
+	if (trans)
+		return -EAGAIN;
+
+	trans = btrfs_join_transaction(root, 1);
+	BUG_ON(IS_ERR(trans));
+	ret = btrfs_commit_transaction(trans, root);
+	BUG_ON(ret);
+
+	return 1;
+}
+
+static int reserve_metadata_bytes(struct btrfs_block_rsv *block_rsv,
+				  u64 num_bytes)
+{
+	struct btrfs_space_info *space_info = block_rsv->space_info;
+	u64 unused;
+	int ret = -ENOSPC;
+
+	spin_lock(&space_info->lock);
+	unused = space_info->bytes_used + space_info->bytes_reserved +
+		 space_info->bytes_pinned + space_info->bytes_readonly;
+
+	if (unused < space_info->total_bytes)
+		unused = space_info->total_bytes - unused;
+	else
+		unused = 0;
+
+	if (unused >= num_bytes) {
+		if (block_rsv->priority >= 10) {
+			space_info->bytes_reserved += num_bytes;
+			ret = 0;
+		} else {
+			if ((unused + block_rsv->reserved) *
+			    block_rsv->priority >=
+			    (num_bytes + block_rsv->reserved) * 10) {
+				space_info->bytes_reserved += num_bytes;
+				ret = 0;
+			}
+		}
+	}
+	spin_unlock(&space_info->lock);
+
+	return ret;
+}
+
+static struct btrfs_block_rsv *get_block_rsv(struct btrfs_trans_handle *trans,
+					     struct btrfs_root *root)
+{
+	struct btrfs_block_rsv *block_rsv;
+	if (root->ref_cows)
+		block_rsv = trans->block_rsv;
+	else
+		block_rsv = root->block_rsv;
+
+	if (!block_rsv)
+		block_rsv = &root->fs_info->empty_block_rsv;
+
+	return block_rsv;
+}
+
+static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv,
+			       u64 num_bytes)
+{
+	int ret = -ENOSPC;
+	spin_lock(&block_rsv->lock);
+	if (block_rsv->reserved >= num_bytes) {
+		block_rsv->reserved -= num_bytes;
+		if (block_rsv->reserved < block_rsv->size)
+			block_rsv->full = 0;
+		ret = 0;
+	}
+	spin_unlock(&block_rsv->lock);
+	return ret;
+}
+
+static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv,
+				u64 num_bytes, int update_size)
+{
+	spin_lock(&block_rsv->lock);
+	block_rsv->reserved += num_bytes;
+	if (update_size)
+		block_rsv->size += num_bytes;
+	else if (block_rsv->reserved >= block_rsv->size)
+		block_rsv->full = 1;
+	spin_unlock(&block_rsv->lock);
+}
+
+void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv,
+			     struct btrfs_block_rsv *dest, u64 num_bytes)
+{
+	struct btrfs_space_info *space_info = block_rsv->space_info;
+
+	spin_lock(&block_rsv->lock);
+	if (num_bytes == (u64)-1)
+		num_bytes = block_rsv->size;
+	block_rsv->size -= num_bytes;
+	if (block_rsv->reserved >= block_rsv->size) {
+		num_bytes = block_rsv->reserved - block_rsv->size;
+		block_rsv->reserved = block_rsv->size;
+		block_rsv->full = 1;
+	} else {
+		num_bytes = 0;
+	}
+	spin_unlock(&block_rsv->lock);
+
+	if (num_bytes > 0) {
+		if (dest) {
+			block_rsv_add_bytes(dest, num_bytes, 0);
+		} else {
+			spin_lock(&space_info->lock);
+			space_info->bytes_reserved -= num_bytes;
+			spin_unlock(&space_info->lock);
+		}
+	}
+}
+
+static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src,
+				   struct btrfs_block_rsv *dst, u64 num_bytes)
+{
+	int ret;
+
+	ret = block_rsv_use_bytes(src, num_bytes);
+	if (ret)
+		return ret;
+
+	block_rsv_add_bytes(dst, num_bytes, 1);
+	return 0;
+}
+
+void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv)
+{
+	memset(rsv, 0, sizeof(*rsv));
+	spin_lock_init(&rsv->lock);
+	atomic_set(&rsv->usage, 1);
+	rsv->priority = 6;
+	INIT_LIST_HEAD(&rsv->list);
+}
+
+struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root)
+{
+	struct btrfs_block_rsv *block_rsv;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	u64 alloc_target;
+
+	block_rsv = kmalloc(sizeof(*block_rsv), GFP_NOFS);
+	if (!block_rsv)
+		return NULL;
+
+	btrfs_init_block_rsv(block_rsv);
+
+	alloc_target = btrfs_get_alloc_profile(root, 0);
+	block_rsv->space_info = __find_space_info(fs_info,
+						  BTRFS_BLOCK_GROUP_METADATA);
+
+	return block_rsv;
+}
+
+void btrfs_free_block_rsv(struct btrfs_root *root,
+			  struct btrfs_block_rsv *rsv)
+{
+	if (rsv && atomic_dec_and_test(&rsv->usage)) {
+		btrfs_block_rsv_release(root, rsv, (u64)-1);
+		if (!rsv->durable)
+			kfree(rsv);
+	}
+}
+
+/*
+ * make the block_rsv struct be able to capture freed space.
+ * the captured space will re-add to the the block_rsv struct
+ * after transaction commit
+ */
+void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info,
+				 struct btrfs_block_rsv *block_rsv)
+{
+	block_rsv->durable = 1;
+	mutex_lock(&fs_info->durable_block_rsv_mutex);
+	list_add_tail(&block_rsv->list, &fs_info->durable_block_rsv_list);
+	mutex_unlock(&fs_info->durable_block_rsv_mutex);
+}
+
+int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct btrfs_block_rsv *block_rsv,
+			u64 num_bytes, int *retries)
+{
+	int ret;
+
+	if (num_bytes == 0)
+		return 0;
+again:
+	ret = reserve_metadata_bytes(block_rsv, num_bytes);
+	if (!ret) {
+		block_rsv_add_bytes(block_rsv, num_bytes, 1);
+		return 0;
+	}
+
+	ret = should_retry_reserve(trans, root, block_rsv, num_bytes, retries);
+	if (ret > 0)
+		goto again;
+
+	return ret;
+}
+
+int btrfs_block_rsv_check(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root,
+			  struct btrfs_block_rsv *block_rsv,
+			  u64 min_reserved, int min_factor)
+{
+	u64 num_bytes = 0;
+	int commit_trans = 0;
+	int ret = -ENOSPC;
+
+	if (!block_rsv)
+		return 0;
+
+	spin_lock(&block_rsv->lock);
+	if (min_factor > 0)
+		num_bytes = div_factor(block_rsv->size, min_factor);
+	if (min_reserved > num_bytes)
+		num_bytes = min_reserved;
+
+	if (block_rsv->reserved >= num_bytes) {
+		ret = 0;
+	} else {
+		num_bytes -= block_rsv->reserved;
+		if (block_rsv->durable &&
+		    block_rsv->freed[0] + block_rsv->freed[1] >= num_bytes)
+			commit_trans = 1;
+	}
+	spin_unlock(&block_rsv->lock);
+	if (!ret)
+		return 0;
+
+	if (block_rsv->refill_used) {
+		ret = reserve_metadata_bytes(block_rsv, num_bytes);
+		if (!ret) {
+			block_rsv_add_bytes(block_rsv, num_bytes, 0);
+			return 0;
+		}
+	}
+
+	if (commit_trans) {
+		if (trans)
+			return -EAGAIN;
+
+		trans = btrfs_join_transaction(root, 1);
+		BUG_ON(IS_ERR(trans));
+		ret = btrfs_commit_transaction(trans, root);
+		return 0;
+	}
+
+	WARN_ON(1);
+	printk(KERN_INFO"block_rsv size %llu reserved %llu freed %llu %llu\n",
+		block_rsv->size, block_rsv->reserved,
+		block_rsv->freed[0], block_rsv->freed[1]);
+
+	return -ENOSPC;
+}
+
+int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
+			    struct btrfs_block_rsv *dst_rsv,
+			    u64 num_bytes)
+{
+	return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
+}
+
+void btrfs_block_rsv_release(struct btrfs_root *root,
+			     struct btrfs_block_rsv *block_rsv,
+			     u64 num_bytes)
+{
+	struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv;
+	if (global_rsv->full || global_rsv == block_rsv ||
+	    block_rsv->space_info != global_rsv->space_info)
+		global_rsv = NULL;
+	block_rsv_release_bytes(block_rsv, global_rsv, num_bytes);
+}
+
+/*
+ * helper to calculate size of global block reservation.
+ * the desired value is sum of space used by extent tree,
+ * checksum tree and root tree
+ */
+static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_space_info *sinfo;
+	u64 num_bytes;
+	u64 meta_used;
+	u64 data_used;
+	int csum_size = btrfs_super_csum_size(&fs_info->super_copy);
+#if 0
+	/*
+	 * per tree used space accounting can be inaccuracy, so we
+	 * can't rely on it.
+	 */
+	spin_lock(&fs_info->extent_root->accounting_lock);
+	num_bytes = btrfs_root_used(&fs_info->extent_root->root_item);
+	spin_unlock(&fs_info->extent_root->accounting_lock);
+
+	spin_lock(&fs_info->csum_root->accounting_lock);
+	num_bytes += btrfs_root_used(&fs_info->csum_root->root_item);
+	spin_unlock(&fs_info->csum_root->accounting_lock);
+
+	spin_lock(&fs_info->tree_root->accounting_lock);
+	num_bytes += btrfs_root_used(&fs_info->tree_root->root_item);
+	spin_unlock(&fs_info->tree_root->accounting_lock);
+#endif
+	sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA);
+	spin_lock(&sinfo->lock);
+	data_used = sinfo->bytes_used;
+	spin_unlock(&sinfo->lock);
+
+	sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA);
+	spin_lock(&sinfo->lock);
+	meta_used = sinfo->bytes_used;
+	spin_unlock(&sinfo->lock);
+
+	num_bytes = (data_used >> fs_info->sb->s_blocksize_bits) *
+		    csum_size * 2;
+	num_bytes += div64_u64(data_used + meta_used, 50);
+
+	if (num_bytes * 3 > meta_used)
+		num_bytes = div64_u64(meta_used, 3);
+
+	return ALIGN(num_bytes, fs_info->extent_root->leafsize << 10);
+}
+
+static void update_global_block_rsv(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv;
+	struct btrfs_space_info *sinfo = block_rsv->space_info;
+	u64 num_bytes;
+
+	num_bytes = calc_global_metadata_size(fs_info);
+
+	spin_lock(&block_rsv->lock);
+	spin_lock(&sinfo->lock);
+
+	block_rsv->size = num_bytes;
+
+	num_bytes = sinfo->bytes_used + sinfo->bytes_pinned +
+		    sinfo->bytes_reserved + sinfo->bytes_readonly;
+
+	if (sinfo->total_bytes > num_bytes) {
+		num_bytes = sinfo->total_bytes - num_bytes;
+		block_rsv->reserved += num_bytes;
+		sinfo->bytes_reserved += num_bytes;
+	}
+
+	if (block_rsv->reserved >= block_rsv->size) {
+		num_bytes = block_rsv->reserved - block_rsv->size;
+		sinfo->bytes_reserved -= num_bytes;
+		block_rsv->reserved = block_rsv->size;
+		block_rsv->full = 1;
+	}
+#if 0
+	printk(KERN_INFO"global block rsv size %llu reserved %llu\n",
+		block_rsv->size, block_rsv->reserved);
+#endif
+	spin_unlock(&sinfo->lock);
+	spin_unlock(&block_rsv->lock);
+}
+
+static void init_global_block_rsv(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_space_info *space_info;
+
+	space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM);
+	fs_info->chunk_block_rsv.space_info = space_info;
+	fs_info->chunk_block_rsv.priority = 10;
+
+	space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA);
+	fs_info->global_block_rsv.space_info = space_info;
+	fs_info->global_block_rsv.priority = 10;
+	fs_info->global_block_rsv.refill_used = 1;
+	fs_info->delalloc_block_rsv.space_info = space_info;
+	fs_info->trans_block_rsv.space_info = space_info;
+	fs_info->empty_block_rsv.space_info = space_info;
+	fs_info->empty_block_rsv.priority = 10;
+
+	fs_info->extent_root->block_rsv = &fs_info->global_block_rsv;
+	fs_info->csum_root->block_rsv = &fs_info->global_block_rsv;
+	fs_info->dev_root->block_rsv = &fs_info->global_block_rsv;
+	fs_info->tree_root->block_rsv = &fs_info->global_block_rsv;
+	fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv;
+
+	btrfs_add_durable_block_rsv(fs_info, &fs_info->global_block_rsv);
+
+	btrfs_add_durable_block_rsv(fs_info, &fs_info->delalloc_block_rsv);
+
+	update_global_block_rsv(fs_info);
+}
+
+static void release_global_block_rsv(struct btrfs_fs_info *fs_info)
+{
+	block_rsv_release_bytes(&fs_info->global_block_rsv, NULL, (u64)-1);
+	WARN_ON(fs_info->delalloc_block_rsv.size > 0);
+	WARN_ON(fs_info->delalloc_block_rsv.reserved > 0);
+	WARN_ON(fs_info->trans_block_rsv.size > 0);
+	WARN_ON(fs_info->trans_block_rsv.reserved > 0);
+	WARN_ON(fs_info->chunk_block_rsv.size > 0);
+	WARN_ON(fs_info->chunk_block_rsv.reserved > 0);
+}
+
+static u64 calc_trans_metadata_size(struct btrfs_root *root, int num_items)
+{
+	return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) *
+		3 * num_items;
+}
+
+int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 int num_items, int *retries)
+{
+	u64 num_bytes;
+	int ret;
+
+	if (num_items == 0 || root->fs_info->chunk_root == root)
+		return 0;
+
+	num_bytes = calc_trans_metadata_size(root, num_items);
+	ret = btrfs_block_rsv_add(trans, root, &root->fs_info->trans_block_rsv,
+				  num_bytes, retries);
+	if (!ret) {
+		trans->bytes_reserved += num_bytes;
+		trans->block_rsv = &root->fs_info->trans_block_rsv;
+	}
+	return ret;
+}
+
+void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root)
+{
+	if (!trans->bytes_reserved)
+		return;
+
+	BUG_ON(trans->block_rsv != &root->fs_info->trans_block_rsv);
+	btrfs_block_rsv_release(root, trans->block_rsv,
+				trans->bytes_reserved);
+	trans->bytes_reserved = 0;
+}
+
+int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
+				  struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root);
+	struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv;
+
+	/*
+	 * one for deleting orphan item, one for updating inode and
+	 * two for calling btrfs_truncate_inode_items.
+	 *
+	 * btrfs_truncate_inode_items is a delete operation, it frees
+	 * more space than it uses in most cases. So two units of
+	 * metadata space should be enough for calling it many times.
+	 * If all of the metadata space is used, we can commit
+	 * transaction and use space it freed.
+	 */
+	u64 num_bytes = calc_trans_metadata_size(root, 4);
+	return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
+}
+
+void btrfs_orphan_release_metadata(struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u64 num_bytes = calc_trans_metadata_size(root, 4);
+	btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes);
+}
+
+int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans,
+				struct btrfs_pending_snapshot *pending)
+{
+	struct btrfs_root *root = pending->root;
+	struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root);
+	struct btrfs_block_rsv *dst_rsv = &pending->block_rsv;
+	/*
+	 * two for root back/forward refs, two for directory entries
+	 * and one for root of the snapshot.
+	 */
+	u64 num_bytes = calc_trans_metadata_size(root, 5);
+	dst_rsv->space_info = src_rsv->space_info;
+	return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
+}
+
+static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes)
+{
+	return num_bytes >>= 3;
+}
+
+int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv;
+	u64 to_reserve;
+	int nr_extents;
+	int retries = 0;
+	int ret;
+
+	if (btrfs_transaction_in_commit(root->fs_info))
+		schedule_timeout(1);
+
+	num_bytes = ALIGN(num_bytes, root->sectorsize);
+again:
+	spin_lock(&BTRFS_I(inode)->accounting_lock);
+	nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents) + 1;
+	if (nr_extents > BTRFS_I(inode)->reserved_extents) {
+		nr_extents -= BTRFS_I(inode)->reserved_extents;
+		to_reserve = calc_trans_metadata_size(root, nr_extents);
+	} else {
+		nr_extents = 0;
+		to_reserve = 0;
+	}
+
+	to_reserve += calc_csum_metadata_size(inode, num_bytes);
+	ret = reserve_metadata_bytes(block_rsv, to_reserve);
+	if (ret) {
+		spin_unlock(&BTRFS_I(inode)->accounting_lock);
+		ret = should_retry_reserve(NULL, root, block_rsv, to_reserve,
+					   &retries);
+		if (ret > 0)
+			goto again;
+		return ret;
+	}
+
+	BTRFS_I(inode)->reserved_extents += nr_extents;
+	atomic_inc(&BTRFS_I(inode)->outstanding_extents);
+	spin_unlock(&BTRFS_I(inode)->accounting_lock);
+
+	block_rsv_add_bytes(block_rsv, to_reserve, 1);
+
+	if (block_rsv->size > 512 * 1024 * 1024)
+		shrink_delalloc(NULL, root, to_reserve);
+
+	return 0;
+}
+
+void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u64 to_free;
+	int nr_extents;
+
+	num_bytes = ALIGN(num_bytes, root->sectorsize);
+	atomic_dec(&BTRFS_I(inode)->outstanding_extents);
+
+	spin_lock(&BTRFS_I(inode)->accounting_lock);
+	nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents);
+	if (nr_extents < BTRFS_I(inode)->reserved_extents) {
+		nr_extents = BTRFS_I(inode)->reserved_extents - nr_extents;
+		BTRFS_I(inode)->reserved_extents -= nr_extents;
+	} else {
+		nr_extents = 0;
+	}
+	spin_unlock(&BTRFS_I(inode)->accounting_lock);
+
+	to_free = calc_csum_metadata_size(inode, num_bytes);
+	if (nr_extents > 0)
+		to_free += calc_trans_metadata_size(root, nr_extents);
+
+	btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv,
+				to_free);
+}
+
+int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
+{
+	int ret;
+
+	ret = btrfs_check_data_free_space(inode, num_bytes);
+	if (ret)
+		return ret;
+
+	ret = btrfs_delalloc_reserve_metadata(inode, num_bytes);
+	if (ret) {
+		btrfs_free_reserved_data_space(inode, num_bytes);
+		return ret;
+	}
+
+	return 0;
+}
+
+void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes)
+{
+	btrfs_delalloc_release_metadata(inode, num_bytes);
+	btrfs_free_reserved_data_space(inode, num_bytes);
+}
+
 static int update_block_group(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
-			      u64 bytenr, u64 num_bytes, int alloc,
-			      int mark_free)
+			      u64 bytenr, u64 num_bytes, int alloc)
 {
 	struct btrfs_block_group_cache *cache;
 	struct btrfs_fs_info *info = root->fs_info;
+	int factor;
 	u64 total = num_bytes;
 	u64 old_val;
 	u64 byte_in_group;
@@ -3486,6 +3798,12 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 		cache = btrfs_lookup_block_group(info, bytenr);
 		if (!cache)
 			return -1;
+		if (cache->flags & (BTRFS_BLOCK_GROUP_DUP |
+				    BTRFS_BLOCK_GROUP_RAID1 |
+				    BTRFS_BLOCK_GROUP_RAID10))
+			factor = 2;
+		else
+			factor = 1;
 		byte_in_group = bytenr - cache->key.objectid;
 		WARN_ON(byte_in_group > cache->key.offset);
 
@@ -3498,31 +3816,24 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 			old_val += num_bytes;
 			btrfs_set_block_group_used(&cache->item, old_val);
 			cache->reserved -= num_bytes;
-			cache->space_info->bytes_used += num_bytes;
 			cache->space_info->bytes_reserved -= num_bytes;
-			if (cache->ro)
-				cache->space_info->bytes_readonly -= num_bytes;
+			cache->space_info->bytes_used += num_bytes;
+			cache->space_info->disk_used += num_bytes * factor;
 			spin_unlock(&cache->lock);
 			spin_unlock(&cache->space_info->lock);
 		} else {
 			old_val -= num_bytes;
-			cache->space_info->bytes_used -= num_bytes;
-			if (cache->ro)
-				cache->space_info->bytes_readonly += num_bytes;
 			btrfs_set_block_group_used(&cache->item, old_val);
+			cache->pinned += num_bytes;
+			cache->space_info->bytes_pinned += num_bytes;
+			cache->space_info->bytes_used -= num_bytes;
+			cache->space_info->disk_used -= num_bytes * factor;
 			spin_unlock(&cache->lock);
 			spin_unlock(&cache->space_info->lock);
-			if (mark_free) {
-				int ret;
 
-				ret = btrfs_discard_extent(root, bytenr,
-							   num_bytes);
-				WARN_ON(ret);
-
-				ret = btrfs_add_free_space(cache, bytenr,
-							   num_bytes);
-				WARN_ON(ret);
-			}
+			set_extent_dirty(info->pinned_extents,
+					 bytenr, bytenr + num_bytes - 1,
+					 GFP_NOFS | __GFP_NOFAIL);
 		}
 		btrfs_put_block_group(cache);
 		total -= num_bytes;
@@ -3546,18 +3857,10 @@ static u64 first_logical_byte(struct btrfs_root *root, u64 search_start)
 	return bytenr;
 }
 
-/*
- * this function must be called within transaction
- */
-int btrfs_pin_extent(struct btrfs_root *root,
-		     u64 bytenr, u64 num_bytes, int reserved)
+static int pin_down_extent(struct btrfs_root *root,
+			   struct btrfs_block_group_cache *cache,
+			   u64 bytenr, u64 num_bytes, int reserved)
 {
-	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct btrfs_block_group_cache *cache;
-
-	cache = btrfs_lookup_block_group(fs_info, bytenr);
-	BUG_ON(!cache);
-
 	spin_lock(&cache->space_info->lock);
 	spin_lock(&cache->lock);
 	cache->pinned += num_bytes;
@@ -3569,28 +3872,68 @@ int btrfs_pin_extent(struct btrfs_root *root,
 	spin_unlock(&cache->lock);
 	spin_unlock(&cache->space_info->lock);
 
-	btrfs_put_block_group(cache);
+	set_extent_dirty(root->fs_info->pinned_extents, bytenr,
+			 bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL);
+	return 0;
+}
+
+/*
+ * this function must be called within transaction
+ */
+int btrfs_pin_extent(struct btrfs_root *root,
+		     u64 bytenr, u64 num_bytes, int reserved)
+{
+	struct btrfs_block_group_cache *cache;
+
+	cache = btrfs_lookup_block_group(root->fs_info, bytenr);
+	BUG_ON(!cache);
+
+	pin_down_extent(root, cache, bytenr, num_bytes, reserved);
 
-	set_extent_dirty(fs_info->pinned_extents,
-			 bytenr, bytenr + num_bytes - 1, GFP_NOFS);
+	btrfs_put_block_group(cache);
 	return 0;
 }
 
-static int update_reserved_extents(struct btrfs_block_group_cache *cache,
-				   u64 num_bytes, int reserve)
+/*
+ * update size of reserved extents. this function may return -EAGAIN
+ * if 'reserve' is true or 'sinfo' is false.
+ */
+static int update_reserved_bytes(struct btrfs_block_group_cache *cache,
+				 u64 num_bytes, int reserve, int sinfo)
 {
-	spin_lock(&cache->space_info->lock);
-	spin_lock(&cache->lock);
-	if (reserve) {
-		cache->reserved += num_bytes;
-		cache->space_info->bytes_reserved += num_bytes;
+	int ret = 0;
+	if (sinfo) {
+		struct btrfs_space_info *space_info = cache->space_info;
+		spin_lock(&space_info->lock);
+		spin_lock(&cache->lock);
+		if (reserve) {
+			if (cache->ro) {
+				ret = -EAGAIN;
+			} else {
+				cache->reserved += num_bytes;
+				space_info->bytes_reserved += num_bytes;
+			}
+		} else {
+			if (cache->ro)
+				space_info->bytes_readonly += num_bytes;
+			cache->reserved -= num_bytes;
+			space_info->bytes_reserved -= num_bytes;
+		}
+		spin_unlock(&cache->lock);
+		spin_unlock(&space_info->lock);
 	} else {
-		cache->reserved -= num_bytes;
-		cache->space_info->bytes_reserved -= num_bytes;
+		spin_lock(&cache->lock);
+		if (cache->ro) {
+			ret = -EAGAIN;
+		} else {
+			if (reserve)
+				cache->reserved += num_bytes;
+			else
+				cache->reserved -= num_bytes;
+		}
+		spin_unlock(&cache->lock);
 	}
-	spin_unlock(&cache->lock);
-	spin_unlock(&cache->space_info->lock);
-	return 0;
+	return ret;
 }
 
 int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
@@ -3621,6 +3964,8 @@ int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 		fs_info->pinned_extents = &fs_info->freed_extents[0];
 
 	up_write(&fs_info->extent_commit_sem);
+
+	update_global_block_rsv(fs_info);
 	return 0;
 }
 
@@ -3647,14 +3992,21 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end)
 			btrfs_add_free_space(cache, start, len);
 		}
 
+		start += len;
+
 		spin_lock(&cache->space_info->lock);
 		spin_lock(&cache->lock);
 		cache->pinned -= len;
 		cache->space_info->bytes_pinned -= len;
+		if (cache->ro) {
+			cache->space_info->bytes_readonly += len;
+		} else if (cache->reserved_pinned > 0) {
+			len = min(len, cache->reserved_pinned);
+			cache->reserved_pinned -= len;
+			cache->space_info->bytes_reserved += len;
+		}
 		spin_unlock(&cache->lock);
 		spin_unlock(&cache->space_info->lock);
-
-		start += len;
 	}
 
 	if (cache)
@@ -3667,8 +4019,11 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct extent_io_tree *unpin;
+	struct btrfs_block_rsv *block_rsv;
+	struct btrfs_block_rsv *next_rsv;
 	u64 start;
 	u64 end;
+	int idx;
 	int ret;
 
 	if (fs_info->pinned_extents == &fs_info->freed_extents[0])
@@ -3689,59 +4044,30 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 		cond_resched();
 	}
 
-	return ret;
-}
+	mutex_lock(&fs_info->durable_block_rsv_mutex);
+	list_for_each_entry_safe(block_rsv, next_rsv,
+				 &fs_info->durable_block_rsv_list, list) {
 
-static int pin_down_bytes(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root,
-			  struct btrfs_path *path,
-			  u64 bytenr, u64 num_bytes,
-			  int is_data, int reserved,
-			  struct extent_buffer **must_clean)
-{
-	int err = 0;
-	struct extent_buffer *buf;
-
-	if (is_data)
-		goto pinit;
-
-	/*
-	 * discard is sloooow, and so triggering discards on
-	 * individual btree blocks isn't a good plan.  Just
-	 * pin everything in discard mode.
-	 */
-	if (btrfs_test_opt(root, DISCARD))
-		goto pinit;
-
-	buf = btrfs_find_tree_block(root, bytenr, num_bytes);
-	if (!buf)
-		goto pinit;
+		idx = trans->transid & 0x1;
+		if (block_rsv->freed[idx] > 0) {
+			block_rsv_add_bytes(block_rsv,
+					    block_rsv->freed[idx], 0);
+			block_rsv->freed[idx] = 0;
+		}
+		if (atomic_read(&block_rsv->usage) == 0) {
+			btrfs_block_rsv_release(root, block_rsv, (u64)-1);
 
-	/* we can reuse a block if it hasn't been written
-	 * and it is from this transaction.  We can't
-	 * reuse anything from the tree log root because
-	 * it has tiny sub-transactions.
-	 */
-	if (btrfs_buffer_uptodate(buf, 0) &&
-	    btrfs_try_tree_lock(buf)) {
-		u64 header_owner = btrfs_header_owner(buf);
-		u64 header_transid = btrfs_header_generation(buf);
-		if (header_owner != BTRFS_TREE_LOG_OBJECTID &&
-		    header_transid == trans->transid &&
-		    !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
-			*must_clean = buf;
-			return 1;
+			if (block_rsv->freed[0] == 0 &&
+			    block_rsv->freed[1] == 0) {
+				list_del_init(&block_rsv->list);
+				kfree(block_rsv);
+			}
+		} else {
+			btrfs_block_rsv_release(root, block_rsv, 0);
 		}
-		btrfs_tree_unlock(buf);
 	}
-	free_extent_buffer(buf);
-pinit:
-	if (path)
-		btrfs_set_path_blocking(path);
-	/* unlocks the pinned mutex */
-	btrfs_pin_extent(root, bytenr, num_bytes, reserved);
+	mutex_unlock(&fs_info->durable_block_rsv_mutex);
 
-	BUG_ON(err < 0);
 	return 0;
 }
 
@@ -3902,9 +4228,6 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			BUG_ON(ret);
 		}
 	} else {
-		int mark_free = 0;
-		struct extent_buffer *must_clean = NULL;
-
 		if (found_extent) {
 			BUG_ON(is_data && refs_to_drop !=
 			       extent_data_ref_count(root, path, iref));
@@ -3917,31 +4240,11 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			}
 		}
 
-		ret = pin_down_bytes(trans, root, path, bytenr,
-				     num_bytes, is_data, 0, &must_clean);
-		if (ret > 0)
-			mark_free = 1;
-		BUG_ON(ret < 0);
-		/*
-		 * it is going to be very rare for someone to be waiting
-		 * on the block we're freeing.  del_items might need to
-		 * schedule, so rather than get fancy, just force it
-		 * to blocking here
-		 */
-		if (must_clean)
-			btrfs_set_lock_blocking(must_clean);
-
 		ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
 				      num_to_del);
 		BUG_ON(ret);
 		btrfs_release_path(extent_root, path);
 
-		if (must_clean) {
-			clean_tree_block(NULL, root, must_clean);
-			btrfs_tree_unlock(must_clean);
-			free_extent_buffer(must_clean);
-		}
-
 		if (is_data) {
 			ret = btrfs_del_csums(trans, root, bytenr, num_bytes);
 			BUG_ON(ret);
@@ -3951,8 +4254,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			     (bytenr + num_bytes - 1) >> PAGE_CACHE_SHIFT);
 		}
 
-		ret = update_block_group(trans, root, bytenr, num_bytes, 0,
-					 mark_free);
+		ret = update_block_group(trans, root, bytenr, num_bytes, 0);
 		BUG_ON(ret);
 	}
 	btrfs_free_path(path);
@@ -3960,7 +4262,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 }
 
 /*
- * when we free an extent, it is possible (and likely) that we free the last
+ * when we free an block, it is possible (and likely) that we free the last
  * delayed ref for that extent as well.  This searches the delayed ref tree for
  * a given extent, and if there are no other delayed refs to be processed, it
  * removes it from the tree.
@@ -3972,7 +4274,7 @@ static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
 	struct btrfs_delayed_ref_root *delayed_refs;
 	struct btrfs_delayed_ref_node *ref;
 	struct rb_node *node;
-	int ret;
+	int ret = 0;
 
 	delayed_refs = &trans->transaction->delayed_refs;
 	spin_lock(&delayed_refs->lock);
@@ -4024,17 +4326,99 @@ static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
 	list_del_init(&head->cluster);
 	spin_unlock(&delayed_refs->lock);
 
-	ret = run_one_delayed_ref(trans, root->fs_info->tree_root,
-				  &head->node, head->extent_op,
-				  head->must_insert_reserved);
-	BUG_ON(ret);
+	BUG_ON(head->extent_op);
+	if (head->must_insert_reserved)
+		ret = 1;
+
+	mutex_unlock(&head->mutex);
 	btrfs_put_delayed_ref(&head->node);
-	return 0;
+	return ret;
 out:
 	spin_unlock(&delayed_refs->lock);
 	return 0;
 }
 
+void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   struct extent_buffer *buf,
+			   u64 parent, int last_ref)
+{
+	struct btrfs_block_rsv *block_rsv;
+	struct btrfs_block_group_cache *cache = NULL;
+	int ret;
+
+	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
+		ret = btrfs_add_delayed_tree_ref(trans, buf->start, buf->len,
+						parent, root->root_key.objectid,
+						btrfs_header_level(buf),
+						BTRFS_DROP_DELAYED_REF, NULL);
+		BUG_ON(ret);
+	}
+
+	if (!last_ref)
+		return;
+
+	block_rsv = get_block_rsv(trans, root);
+	cache = btrfs_lookup_block_group(root->fs_info, buf->start);
+	BUG_ON(block_rsv->space_info != cache->space_info);
+
+	if (btrfs_header_generation(buf) == trans->transid) {
+		if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
+			ret = check_ref_cleanup(trans, root, buf->start);
+			if (!ret)
+				goto pin;
+		}
+
+		if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
+			pin_down_extent(root, cache, buf->start, buf->len, 1);
+			goto pin;
+		}
+
+		WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags));
+
+		btrfs_add_free_space(cache, buf->start, buf->len);
+		ret = update_reserved_bytes(cache, buf->len, 0, 0);
+		if (ret == -EAGAIN) {
+			/* block group became read-only */
+			update_reserved_bytes(cache, buf->len, 0, 1);
+			goto out;
+		}
+
+		ret = 1;
+		spin_lock(&block_rsv->lock);
+		if (block_rsv->reserved < block_rsv->size) {
+			block_rsv->reserved += buf->len;
+			ret = 0;
+		}
+		spin_unlock(&block_rsv->lock);
+
+		if (ret) {
+			spin_lock(&cache->space_info->lock);
+			cache->space_info->bytes_reserved -= buf->len;
+			spin_unlock(&cache->space_info->lock);
+		}
+		goto out;
+	}
+pin:
+	if (block_rsv->durable && !cache->ro) {
+		ret = 0;
+		spin_lock(&cache->lock);
+		if (!cache->ro) {
+			cache->reserved_pinned += buf->len;
+			ret = 1;
+		}
+		spin_unlock(&cache->lock);
+
+		if (ret) {
+			spin_lock(&block_rsv->lock);
+			block_rsv->freed[trans->transid & 0x1] += buf->len;
+			spin_unlock(&block_rsv->lock);
+		}
+	}
+out:
+	btrfs_put_block_group(cache);
+}
+
 int btrfs_free_extent(struct btrfs_trans_handle *trans,
 		      struct btrfs_root *root,
 		      u64 bytenr, u64 num_bytes, u64 parent,
@@ -4056,8 +4440,6 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans,
 					parent, root_objectid, (int)owner,
 					BTRFS_DROP_DELAYED_REF, NULL);
 		BUG_ON(ret);
-		ret = check_ref_cleanup(trans, root, bytenr);
-		BUG_ON(ret);
 	} else {
 		ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes,
 					parent, root_objectid, owner,
@@ -4067,21 +4449,6 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-int btrfs_free_tree_block(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root,
-			  u64 bytenr, u32 blocksize,
-			  u64 parent, u64 root_objectid, int level)
-{
-	u64 used;
-	spin_lock(&root->node_lock);
-	used = btrfs_root_used(&root->root_item) - blocksize;
-	btrfs_set_root_used(&root->root_item, used);
-	spin_unlock(&root->node_lock);
-
-	return btrfs_free_extent(trans, root, bytenr, blocksize,
-				 parent, root_objectid, level, 0);
-}
-
 static u64 stripe_align(struct btrfs_root *root, u64 val)
 {
 	u64 mask = ((u64)root->stripesize - 1);
@@ -4134,6 +4501,22 @@ wait_block_group_cache_done(struct btrfs_block_group_cache *cache)
 	return 0;
 }
 
+static int get_block_group_index(struct btrfs_block_group_cache *cache)
+{
+	int index;
+	if (cache->flags & BTRFS_BLOCK_GROUP_RAID10)
+		index = 0;
+	else if (cache->flags & BTRFS_BLOCK_GROUP_RAID1)
+		index = 1;
+	else if (cache->flags & BTRFS_BLOCK_GROUP_DUP)
+		index = 2;
+	else if (cache->flags & BTRFS_BLOCK_GROUP_RAID0)
+		index = 3;
+	else
+		index = 4;
+	return index;
+}
+
 enum btrfs_loop_type {
 	LOOP_FIND_IDEAL = 0,
 	LOOP_CACHING_NOWAIT = 1,
@@ -4155,7 +4538,6 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 				     u64 num_bytes, u64 empty_size,
 				     u64 search_start, u64 search_end,
 				     u64 hint_byte, struct btrfs_key *ins,
-				     u64 exclude_start, u64 exclude_nr,
 				     int data)
 {
 	int ret = 0;
@@ -4168,6 +4550,7 @@ 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;
+	int index = 0;
 	bool found_uncached_bg = false;
 	bool failed_cluster_refill = false;
 	bool failed_alloc = false;
@@ -4237,6 +4620,7 @@ ideal_cache:
 				btrfs_put_block_group(block_group);
 				up_read(&space_info->groups_sem);
 			} else {
+				index = get_block_group_index(block_group);
 				goto have_block_group;
 			}
 		} else if (block_group) {
@@ -4245,7 +4629,8 @@ ideal_cache:
 	}
 search:
 	down_read(&space_info->groups_sem);
-	list_for_each_entry(block_group, &space_info->block_groups, list) {
+	list_for_each_entry(block_group, &space_info->block_groups[index],
+			    list) {
 		u64 offset;
 		int cached;
 
@@ -4436,23 +4821,22 @@ checks:
 			goto loop;
 		}
 
-		if (exclude_nr > 0 &&
-		    (search_start + num_bytes > exclude_start &&
-		     search_start < exclude_start + exclude_nr)) {
-			search_start = exclude_start + exclude_nr;
+		ins->objectid = search_start;
+		ins->offset = num_bytes;
+
+		if (offset < search_start)
+			btrfs_add_free_space(block_group, offset,
+					     search_start - offset);
+		BUG_ON(offset > search_start);
 
+		ret = update_reserved_bytes(block_group, num_bytes, 1,
+					    (data & BTRFS_BLOCK_GROUP_DATA));
+		if (ret == -EAGAIN) {
 			btrfs_add_free_space(block_group, offset, num_bytes);
-			/*
-			 * if search_start is still in this block group
-			 * then we just re-search this block group
-			 */
-			if (search_start >= block_group->key.objectid &&
-			    search_start < (block_group->key.objectid +
-					    block_group->key.offset))
-				goto have_block_group;
 			goto loop;
 		}
 
+		/* we are all good, lets return */
 		ins->objectid = search_start;
 		ins->offset = num_bytes;
 
@@ -4460,18 +4844,18 @@ checks:
 			btrfs_add_free_space(block_group, offset,
 					     search_start - offset);
 		BUG_ON(offset > search_start);
-
-		update_reserved_extents(block_group, num_bytes, 1);
-
-		/* we are all good, lets return */
 		break;
 loop:
 		failed_cluster_refill = false;
 		failed_alloc = false;
+		BUG_ON(index != get_block_group_index(block_group));
 		btrfs_put_block_group(block_group);
 	}
 	up_read(&space_info->groups_sem);
 
+	if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES)
+		goto search;
+
 	/* LOOP_FIND_IDEAL, only search caching/cached bg's, and don't wait for
 	 *			for them to make caching progress.  Also
 	 *			determine the best possible bg to cache
@@ -4485,6 +4869,7 @@ loop:
 	if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE &&
 	    (found_uncached_bg || empty_size || empty_cluster ||
 	     allowed_chunk_alloc)) {
+		index = 0;
 		if (loop == LOOP_FIND_IDEAL && found_uncached_bg) {
 			found_uncached_bg = false;
 			loop++;
@@ -4567,31 +4952,30 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
 			    int dump_block_groups)
 {
 	struct btrfs_block_group_cache *cache;
+	int index = 0;
 
 	spin_lock(&info->lock);
 	printk(KERN_INFO "space_info has %llu free, is %sfull\n",
 	       (unsigned long long)(info->total_bytes - info->bytes_used -
 				    info->bytes_pinned - info->bytes_reserved -
-				    info->bytes_super),
+				    info->bytes_readonly),
 	       (info->full) ? "" : "not ");
-	printk(KERN_INFO "space_info total=%llu, pinned=%llu, delalloc=%llu,"
-	       " may_use=%llu, used=%llu, root=%llu, super=%llu, reserved=%llu"
-	       "\n",
+	printk(KERN_INFO "space_info total=%llu, used=%llu, pinned=%llu, "
+	       "reserved=%llu, may_use=%llu, readonly=%llu\n",
 	       (unsigned long long)info->total_bytes,
+	       (unsigned long long)info->bytes_used,
 	       (unsigned long long)info->bytes_pinned,
-	       (unsigned long long)info->bytes_delalloc,
+	       (unsigned long long)info->bytes_reserved,
 	       (unsigned long long)info->bytes_may_use,
-	       (unsigned long long)info->bytes_used,
-	       (unsigned long long)info->bytes_root,
-	       (unsigned long long)info->bytes_super,
-	       (unsigned long long)info->bytes_reserved);
+	       (unsigned long long)info->bytes_readonly);
 	spin_unlock(&info->lock);
 
 	if (!dump_block_groups)
 		return;
 
 	down_read(&info->groups_sem);
-	list_for_each_entry(cache, &info->block_groups, list) {
+again:
+	list_for_each_entry(cache, &info->block_groups[index], list) {
 		spin_lock(&cache->lock);
 		printk(KERN_INFO "block group %llu has %llu bytes, %llu used "
 		       "%llu pinned %llu reserved\n",
@@ -4603,6 +4987,8 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
 		btrfs_dump_free_space(cache, bytes);
 		spin_unlock(&cache->lock);
 	}
+	if (++index < BTRFS_NR_RAID_TYPES)
+		goto again;
 	up_read(&info->groups_sem);
 }
 
@@ -4628,9 +5014,8 @@ again:
 
 	WARN_ON(num_bytes < root->sectorsize);
 	ret = find_free_extent(trans, root, num_bytes, empty_size,
-			       search_start, search_end, hint_byte, ins,
-			       trans->alloc_exclude_start,
-			       trans->alloc_exclude_nr, data);
+			       search_start, search_end, hint_byte,
+			       ins, data);
 
 	if (ret == -ENOSPC && num_bytes > min_alloc_size) {
 		num_bytes = num_bytes >> 1;
@@ -4668,7 +5053,7 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
 	ret = btrfs_discard_extent(root, start, len);
 
 	btrfs_add_free_space(cache, start, len);
-	update_reserved_extents(cache, len, 0);
+	update_reserved_bytes(cache, len, 0, 1);
 	btrfs_put_block_group(cache);
 
 	return ret;
@@ -4731,8 +5116,7 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(path->nodes[0]);
 	btrfs_free_path(path);
 
-	ret = update_block_group(trans, root, ins->objectid, ins->offset,
-				 1, 0);
+	ret = update_block_group(trans, root, ins->objectid, ins->offset, 1);
 	if (ret) {
 		printk(KERN_ERR "btrfs update block group failed for %llu "
 		       "%llu\n", (unsigned long long)ins->objectid,
@@ -4792,8 +5176,7 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(leaf);
 	btrfs_free_path(path);
 
-	ret = update_block_group(trans, root, ins->objectid, ins->offset,
-				 1, 0);
+	ret = update_block_group(trans, root, ins->objectid, ins->offset, 1);
 	if (ret) {
 		printk(KERN_ERR "btrfs update block group failed for %llu "
 		       "%llu\n", (unsigned long long)ins->objectid,
@@ -4869,73 +5252,14 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
 		put_caching_control(caching_ctl);
 	}
 
-	update_reserved_extents(block_group, ins->offset, 1);
+	ret = update_reserved_bytes(block_group, ins->offset, 1, 1);
+	BUG_ON(ret);
 	btrfs_put_block_group(block_group);
 	ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
 					 0, owner, offset, ins, 1);
 	return ret;
 }
 
-/*
- * finds a free extent and does all the dirty work required for allocation
- * returns the key for the extent through ins, and a tree buffer for
- * the first block of the extent through buf.
- *
- * returns 0 if everything worked, non-zero otherwise.
- */
-static int alloc_tree_block(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root,
-			    u64 num_bytes, u64 parent, u64 root_objectid,
-			    struct btrfs_disk_key *key, int level,
-			    u64 empty_size, u64 hint_byte, u64 search_end,
-			    struct btrfs_key *ins)
-{
-	int ret;
-	u64 flags = 0;
-
-	ret = btrfs_reserve_extent(trans, root, num_bytes, num_bytes,
-				   empty_size, hint_byte, search_end,
-				   ins, 0);
-	if (ret)
-		return ret;
-
-	if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
-		if (parent == 0)
-			parent = ins->objectid;
-		flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
-	} else
-		BUG_ON(parent > 0);
-
-	if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
-		struct btrfs_delayed_extent_op *extent_op;
-		extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
-		BUG_ON(!extent_op);
-		if (key)
-			memcpy(&extent_op->key, key, sizeof(extent_op->key));
-		else
-			memset(&extent_op->key, 0, sizeof(extent_op->key));
-		extent_op->flags_to_set = flags;
-		extent_op->update_key = 1;
-		extent_op->update_flags = 1;
-		extent_op->is_data = 0;
-
-		ret = btrfs_add_delayed_tree_ref(trans, ins->objectid,
-					ins->offset, parent, root_objectid,
-					level, BTRFS_ADD_DELAYED_EXTENT,
-					extent_op);
-		BUG_ON(ret);
-	}
-
-	if (root_objectid == root->root_key.objectid) {
-		u64 used;
-		spin_lock(&root->node_lock);
-		used = btrfs_root_used(&root->root_item) + num_bytes;
-		btrfs_set_root_used(&root->root_item, used);
-		spin_unlock(&root->node_lock);
-	}
-	return ret;
-}
-
 struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 					    struct btrfs_root *root,
 					    u64 bytenr, u32 blocksize,
@@ -4974,8 +5298,45 @@ struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 	return buf;
 }
 
+static struct btrfs_block_rsv *
+use_block_rsv(struct btrfs_trans_handle *trans,
+	      struct btrfs_root *root, u32 blocksize)
+{
+	struct btrfs_block_rsv *block_rsv;
+	int ret;
+
+	block_rsv = get_block_rsv(trans, root);
+
+	if (block_rsv->size == 0) {
+		ret = reserve_metadata_bytes(block_rsv, blocksize);
+		if (ret)
+			return ERR_PTR(ret);
+		return block_rsv;
+	}
+
+	ret = block_rsv_use_bytes(block_rsv, blocksize);
+	if (!ret)
+		return block_rsv;
+
+	WARN_ON(1);
+	printk(KERN_INFO"block_rsv size %llu reserved %llu freed %llu %llu\n",
+		block_rsv->size, block_rsv->reserved,
+		block_rsv->freed[0], block_rsv->freed[1]);
+
+	return ERR_PTR(-ENOSPC);
+}
+
+static void unuse_block_rsv(struct btrfs_block_rsv *block_rsv, u32 blocksize)
+{
+	block_rsv_add_bytes(block_rsv, blocksize, 0);
+	block_rsv_release_bytes(block_rsv, NULL, 0);
+}
+
 /*
- * helper function to allocate a block for a given tree
+ * finds a free extent and does all the dirty work required for allocation
+ * returns the key for the extent through ins, and a tree buffer for
+ * the first block of the extent through buf.
+ *
  * returns the tree buffer or NULL.
  */
 struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
@@ -4985,18 +5346,53 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 					u64 hint, u64 empty_size)
 {
 	struct btrfs_key ins;
-	int ret;
+	struct btrfs_block_rsv *block_rsv;
 	struct extent_buffer *buf;
+	u64 flags = 0;
+	int ret;
+
 
-	ret = alloc_tree_block(trans, root, blocksize, parent, root_objectid,
-			       key, level, empty_size, hint, (u64)-1, &ins);
+	block_rsv = use_block_rsv(trans, root, blocksize);
+	if (IS_ERR(block_rsv))
+		return ERR_CAST(block_rsv);
+
+	ret = btrfs_reserve_extent(trans, root, blocksize, blocksize,
+				   empty_size, hint, (u64)-1, &ins, 0);
 	if (ret) {
-		BUG_ON(ret > 0);
+		unuse_block_rsv(block_rsv, blocksize);
 		return ERR_PTR(ret);
 	}
 
 	buf = btrfs_init_new_buffer(trans, root, ins.objectid,
 				    blocksize, level);
+	BUG_ON(IS_ERR(buf));
+
+	if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
+		if (parent == 0)
+			parent = ins.objectid;
+		flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
+	} else
+		BUG_ON(parent > 0);
+
+	if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
+		struct btrfs_delayed_extent_op *extent_op;
+		extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
+		BUG_ON(!extent_op);
+		if (key)
+			memcpy(&extent_op->key, key, sizeof(extent_op->key));
+		else
+			memset(&extent_op->key, 0, sizeof(extent_op->key));
+		extent_op->flags_to_set = flags;
+		extent_op->update_key = 1;
+		extent_op->update_flags = 1;
+		extent_op->is_data = 0;
+
+		ret = btrfs_add_delayed_tree_ref(trans, ins.objectid,
+					ins.offset, parent, root_objectid,
+					level, BTRFS_ADD_DELAYED_EXTENT,
+					extent_op);
+		BUG_ON(ret);
+	}
 	return buf;
 }
 
@@ -5321,7 +5717,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 				 struct btrfs_path *path,
 				 struct walk_control *wc)
 {
-	int ret = 0;
+	int ret;
 	int level = wc->level;
 	struct extent_buffer *eb = path->nodes[level];
 	u64 parent = 0;
@@ -5399,13 +5795,11 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 			       btrfs_header_owner(path->nodes[level + 1]));
 	}
 
-	ret = btrfs_free_extent(trans, root, eb->start, eb->len, parent,
-				root->root_key.objectid, level, 0);
-	BUG_ON(ret);
+	btrfs_free_tree_block(trans, root, eb, parent, wc->refs[level] == 1);
 out:
 	wc->refs[level] = 0;
 	wc->flags[level] = 0;
-	return ret;
+	return 0;
 }
 
 static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
@@ -5483,7 +5877,8 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
  * also make sure backrefs for the shared block and all lower level
  * blocks are properly updated.
  */
-int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref)
+int btrfs_drop_snapshot(struct btrfs_root *root,
+			struct btrfs_block_rsv *block_rsv, int update_ref)
 {
 	struct btrfs_path *path;
 	struct btrfs_trans_handle *trans;
@@ -5501,7 +5896,9 @@ int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref)
 	wc = kzalloc(sizeof(*wc), GFP_NOFS);
 	BUG_ON(!wc);
 
-	trans = btrfs_start_transaction(tree_root, 1);
+	trans = btrfs_start_transaction(tree_root, 0);
+	if (block_rsv)
+		trans->block_rsv = block_rsv;
 
 	if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
 		level = btrfs_header_level(root->node);
@@ -5589,22 +5986,16 @@ int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref)
 		}
 
 		BUG_ON(wc->level == 0);
-		if (trans->transaction->in_commit ||
-		    trans->transaction->delayed_refs.flushing) {
+		if (btrfs_should_end_transaction(trans, tree_root)) {
 			ret = btrfs_update_root(trans, tree_root,
 						&root->root_key,
 						root_item);
 			BUG_ON(ret);
 
-			btrfs_end_transaction(trans, tree_root);
-			trans = btrfs_start_transaction(tree_root, 1);
-		} else {
-			unsigned long update;
-			update = trans->delayed_ref_updates;
-			trans->delayed_ref_updates = 0;
-			if (update)
-				btrfs_run_delayed_refs(trans, tree_root,
-						       update);
+			btrfs_end_transaction_throttle(trans, tree_root);
+			trans = btrfs_start_transaction(tree_root, 0);
+			if (block_rsv)
+				trans->block_rsv = block_rsv;
 		}
 	}
 	btrfs_release_path(root, path);
@@ -5632,7 +6023,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref)
 		kfree(root);
 	}
 out:
-	btrfs_end_transaction(trans, tree_root);
+	btrfs_end_transaction_throttle(trans, tree_root);
 	kfree(wc);
 	btrfs_free_path(path);
 	return err;
@@ -7228,48 +7619,80 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
 	return flags;
 }
 
-static int __alloc_chunk_for_shrink(struct btrfs_root *root,
-		     struct btrfs_block_group_cache *shrink_block_group,
-		     int force)
+static int set_block_group_ro(struct btrfs_block_group_cache *cache)
 {
-	struct btrfs_trans_handle *trans;
-	u64 new_alloc_flags;
-	u64 calc;
+	struct btrfs_space_info *sinfo = cache->space_info;
+	u64 num_bytes;
+	int ret = -ENOSPC;
 
-	spin_lock(&shrink_block_group->lock);
-	if (btrfs_block_group_used(&shrink_block_group->item) +
-	    shrink_block_group->reserved > 0) {
-		spin_unlock(&shrink_block_group->lock);
+	if (cache->ro)
+		return 0;
 
-		trans = btrfs_start_transaction(root, 1);
-		spin_lock(&shrink_block_group->lock);
+	spin_lock(&sinfo->lock);
+	spin_lock(&cache->lock);
+	num_bytes = cache->key.offset - cache->reserved - cache->pinned -
+		    cache->bytes_super - btrfs_block_group_used(&cache->item);
+
+	if (sinfo->bytes_used + sinfo->bytes_reserved + sinfo->bytes_pinned +
+	    sinfo->bytes_may_use + sinfo->bytes_readonly +
+	    cache->reserved_pinned + num_bytes < sinfo->total_bytes) {
+		sinfo->bytes_readonly += num_bytes;
+		sinfo->bytes_reserved += cache->reserved_pinned;
+		cache->reserved_pinned = 0;
+		cache->ro = 1;
+		ret = 0;
+	}
+	spin_unlock(&cache->lock);
+	spin_unlock(&sinfo->lock);
+	return ret;
+}
 
-		new_alloc_flags = update_block_group_flags(root,
-						   shrink_block_group->flags);
-		if (new_alloc_flags != shrink_block_group->flags) {
-			calc =
-			     btrfs_block_group_used(&shrink_block_group->item);
-		} else {
-			calc = shrink_block_group->key.offset;
-		}
-		spin_unlock(&shrink_block_group->lock);
+int btrfs_set_block_group_ro(struct btrfs_root *root,
+			     struct btrfs_block_group_cache *cache)
 
-		do_chunk_alloc(trans, root->fs_info->extent_root,
-			       calc + 2 * 1024 * 1024, new_alloc_flags, force);
+{
+	struct btrfs_trans_handle *trans;
+	u64 alloc_flags;
+	int ret;
 
-		btrfs_end_transaction(trans, root);
-	} else
-		spin_unlock(&shrink_block_group->lock);
-	return 0;
-}
+	BUG_ON(cache->ro);
+
+	trans = btrfs_join_transaction(root, 1);
+	BUG_ON(IS_ERR(trans));
 
+	alloc_flags = update_block_group_flags(root, cache->flags);
+	if (alloc_flags != cache->flags)
+		do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, 1);
 
-int btrfs_prepare_block_group_relocation(struct btrfs_root *root,
-					 struct btrfs_block_group_cache *group)
+	ret = set_block_group_ro(cache);
+	if (!ret)
+		goto out;
+	alloc_flags = get_alloc_profile(root, cache->space_info->flags);
+	ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, 1);
+	if (ret < 0)
+		goto out;
+	ret = set_block_group_ro(cache);
+out:
+	btrfs_end_transaction(trans, root);
+	return ret;
+}
 
+int btrfs_set_block_group_rw(struct btrfs_root *root,
+			      struct btrfs_block_group_cache *cache)
 {
-	__alloc_chunk_for_shrink(root, group, 1);
-	set_block_group_readonly(group);
+	struct btrfs_space_info *sinfo = cache->space_info;
+	u64 num_bytes;
+
+	BUG_ON(!cache->ro);
+
+	spin_lock(&sinfo->lock);
+	spin_lock(&cache->lock);
+	num_bytes = cache->key.offset - cache->reserved - cache->pinned -
+		    cache->bytes_super - btrfs_block_group_used(&cache->item);
+	sinfo->bytes_readonly -= num_bytes;
+	cache->ro = 0;
+	spin_unlock(&cache->lock);
+	spin_unlock(&sinfo->lock);
 	return 0;
 }
 
@@ -7436,17 +7859,33 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 	 */
 	synchronize_rcu();
 
+	release_global_block_rsv(info);
+
 	while(!list_empty(&info->space_info)) {
 		space_info = list_entry(info->space_info.next,
 					struct btrfs_space_info,
 					list);
-
+		if (space_info->bytes_pinned > 0 ||
+		    space_info->bytes_reserved > 0) {
+			WARN_ON(1);
+			dump_space_info(space_info, 0, 0);
+		}
 		list_del(&space_info->list);
 		kfree(space_info);
 	}
 	return 0;
 }
 
+static void __link_block_group(struct btrfs_space_info *space_info,
+			       struct btrfs_block_group_cache *cache)
+{
+	int index = get_block_group_index(cache);
+
+	down_write(&space_info->groups_sem);
+	list_add_tail(&cache->list, &space_info->block_groups[index]);
+	up_write(&space_info->groups_sem);
+}
+
 int btrfs_read_block_groups(struct btrfs_root *root)
 {
 	struct btrfs_path *path;
@@ -7468,10 +7907,8 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 
 	while (1) {
 		ret = find_first_block_group(root, path, &key);
-		if (ret > 0) {
-			ret = 0;
-			goto error;
-		}
+		if (ret > 0)
+			break;
 		if (ret != 0)
 			goto error;
 
@@ -7480,7 +7917,7 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		cache = kzalloc(sizeof(*cache), GFP_NOFS);
 		if (!cache) {
 			ret = -ENOMEM;
-			break;
+			goto error;
 		}
 
 		atomic_set(&cache->count, 1);
@@ -7537,20 +7974,36 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		BUG_ON(ret);
 		cache->space_info = space_info;
 		spin_lock(&cache->space_info->lock);
-		cache->space_info->bytes_super += cache->bytes_super;
+		cache->space_info->bytes_readonly += cache->bytes_super;
 		spin_unlock(&cache->space_info->lock);
 
-		down_write(&space_info->groups_sem);
-		list_add_tail(&cache->list, &space_info->block_groups);
-		up_write(&space_info->groups_sem);
+		__link_block_group(space_info, cache);
 
 		ret = btrfs_add_block_group_cache(root->fs_info, cache);
 		BUG_ON(ret);
 
 		set_avail_alloc_bits(root->fs_info, cache->flags);
 		if (btrfs_chunk_readonly(root, cache->key.objectid))
-			set_block_group_readonly(cache);
+			set_block_group_ro(cache);
 	}
+
+	list_for_each_entry_rcu(space_info, &root->fs_info->space_info, list) {
+		if (!(get_alloc_profile(root, space_info->flags) &
+		      (BTRFS_BLOCK_GROUP_RAID10 |
+		       BTRFS_BLOCK_GROUP_RAID1 |
+		       BTRFS_BLOCK_GROUP_DUP)))
+			continue;
+		/*
+		 * avoid allocating from un-mirrored block group if there are
+		 * mirrored block groups.
+		 */
+		list_for_each_entry(cache, &space_info->block_groups[3], list)
+			set_block_group_ro(cache);
+		list_for_each_entry(cache, &space_info->block_groups[4], list)
+			set_block_group_ro(cache);
+	}
+
+	init_global_block_rsv(info);
 	ret = 0;
 error:
 	btrfs_free_path(path);
@@ -7611,12 +8064,10 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	BUG_ON(ret);
 
 	spin_lock(&cache->space_info->lock);
-	cache->space_info->bytes_super += cache->bytes_super;
+	cache->space_info->bytes_readonly += cache->bytes_super;
 	spin_unlock(&cache->space_info->lock);
 
-	down_write(&cache->space_info->groups_sem);
-	list_add_tail(&cache->list, &cache->space_info->block_groups);
-	up_write(&cache->space_info->groups_sem);
+	__link_block_group(cache->space_info, cache);
 
 	ret = btrfs_add_block_group_cache(root->fs_info, cache);
 	BUG_ON(ret);
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index d2d03684fab2..a4080c21ec55 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -135,7 +135,7 @@ static struct extent_state *alloc_extent_state(gfp_t mask)
 	return state;
 }
 
-static void free_extent_state(struct extent_state *state)
+void free_extent_state(struct extent_state *state)
 {
 	if (!state)
 		return;
@@ -335,21 +335,18 @@ static int merge_state(struct extent_io_tree *tree,
 }
 
 static int set_state_cb(struct extent_io_tree *tree,
-			 struct extent_state *state,
-			 unsigned long bits)
+			 struct extent_state *state, int *bits)
 {
 	if (tree->ops && tree->ops->set_bit_hook) {
 		return tree->ops->set_bit_hook(tree->mapping->host,
-					       state->start, state->end,
-					       state->state, bits);
+					       state, bits);
 	}
 
 	return 0;
 }
 
 static void clear_state_cb(struct extent_io_tree *tree,
-			   struct extent_state *state,
-			   unsigned long bits)
+			   struct extent_state *state, int *bits)
 {
 	if (tree->ops && tree->ops->clear_bit_hook)
 		tree->ops->clear_bit_hook(tree->mapping->host, state, bits);
@@ -367,9 +364,10 @@ static void clear_state_cb(struct extent_io_tree *tree,
  */
 static int insert_state(struct extent_io_tree *tree,
 			struct extent_state *state, u64 start, u64 end,
-			int bits)
+			int *bits)
 {
 	struct rb_node *node;
+	int bits_to_set = *bits & ~EXTENT_CTLBITS;
 	int ret;
 
 	if (end < start) {
@@ -384,9 +382,9 @@ static int insert_state(struct extent_io_tree *tree,
 	if (ret)
 		return ret;
 
-	if (bits & EXTENT_DIRTY)
+	if (bits_to_set & EXTENT_DIRTY)
 		tree->dirty_bytes += end - start + 1;
-	state->state |= bits;
+	state->state |= bits_to_set;
 	node = tree_insert(&tree->state, end, &state->rb_node);
 	if (node) {
 		struct extent_state *found;
@@ -456,13 +454,13 @@ static int split_state(struct extent_io_tree *tree, struct extent_state *orig,
  * struct is freed and removed from the tree
  */
 static int clear_state_bit(struct extent_io_tree *tree,
-			    struct extent_state *state, int bits, int wake,
-			    int delete)
+			    struct extent_state *state,
+			    int *bits, int wake)
 {
-	int bits_to_clear = bits & ~EXTENT_DO_ACCOUNTING;
+	int bits_to_clear = *bits & ~EXTENT_CTLBITS;
 	int ret = state->state & bits_to_clear;
 
-	if ((bits & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
+	if ((bits_to_clear & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
 		u64 range = state->end - state->start + 1;
 		WARN_ON(range > tree->dirty_bytes);
 		tree->dirty_bytes -= range;
@@ -471,9 +469,8 @@ static int clear_state_bit(struct extent_io_tree *tree,
 	state->state &= ~bits_to_clear;
 	if (wake)
 		wake_up(&state->wq);
-	if (delete || state->state == 0) {
+	if (state->state == 0) {
 		if (state->tree) {
-			clear_state_cb(tree, state, state->state);
 			rb_erase(&state->rb_node, &tree->state);
 			state->tree = NULL;
 			free_extent_state(state);
@@ -514,6 +511,10 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	int set = 0;
 	int clear = 0;
 
+	if (delete)
+		bits |= ~EXTENT_CTLBITS;
+	bits |= EXTENT_FIRST_DELALLOC;
+
 	if (bits & (EXTENT_IOBITS | EXTENT_BOUNDARY))
 		clear = 1;
 again:
@@ -580,8 +581,7 @@ hit_next:
 		if (err)
 			goto out;
 		if (state->end <= end) {
-			set |= clear_state_bit(tree, state, bits, wake,
-					       delete);
+			set |= clear_state_bit(tree, state, &bits, wake);
 			if (last_end == (u64)-1)
 				goto out;
 			start = last_end + 1;
@@ -602,7 +602,7 @@ hit_next:
 		if (wake)
 			wake_up(&state->wq);
 
-		set |= clear_state_bit(tree, prealloc, bits, wake, delete);
+		set |= clear_state_bit(tree, prealloc, &bits, wake);
 
 		prealloc = NULL;
 		goto out;
@@ -613,7 +613,7 @@ hit_next:
 	else
 		next_node = NULL;
 
-	set |= clear_state_bit(tree, state, bits, wake, delete);
+	set |= clear_state_bit(tree, state, &bits, wake);
 	if (last_end == (u64)-1)
 		goto out;
 	start = last_end + 1;
@@ -706,19 +706,19 @@ out:
 
 static int set_state_bits(struct extent_io_tree *tree,
 			   struct extent_state *state,
-			   int bits)
+			   int *bits)
 {
 	int ret;
+	int bits_to_set = *bits & ~EXTENT_CTLBITS;
 
 	ret = set_state_cb(tree, state, bits);
 	if (ret)
 		return ret;
-
-	if ((bits & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
+	if ((bits_to_set & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
 		u64 range = state->end - state->start + 1;
 		tree->dirty_bytes += range;
 	}
-	state->state |= bits;
+	state->state |= bits_to_set;
 
 	return 0;
 }
@@ -745,10 +745,9 @@ static void cache_state(struct extent_state *state,
  * [start, end] is inclusive This takes the tree lock.
  */
 
-static int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
-			  int bits, int exclusive_bits, u64 *failed_start,
-			  struct extent_state **cached_state,
-			  gfp_t mask)
+int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		   int bits, int exclusive_bits, u64 *failed_start,
+		   struct extent_state **cached_state, gfp_t mask)
 {
 	struct extent_state *state;
 	struct extent_state *prealloc = NULL;
@@ -757,6 +756,7 @@ static int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	u64 last_start;
 	u64 last_end;
 
+	bits |= EXTENT_FIRST_DELALLOC;
 again:
 	if (!prealloc && (mask & __GFP_WAIT)) {
 		prealloc = alloc_extent_state(mask);
@@ -778,7 +778,7 @@ again:
 	 */
 	node = tree_search(tree, start);
 	if (!node) {
-		err = insert_state(tree, prealloc, start, end, bits);
+		err = insert_state(tree, prealloc, start, end, &bits);
 		prealloc = NULL;
 		BUG_ON(err == -EEXIST);
 		goto out;
@@ -802,7 +802,7 @@ hit_next:
 			goto out;
 		}
 
-		err = set_state_bits(tree, state, bits);
+		err = set_state_bits(tree, state, &bits);
 		if (err)
 			goto out;
 
@@ -852,7 +852,7 @@ hit_next:
 		if (err)
 			goto out;
 		if (state->end <= end) {
-			err = set_state_bits(tree, state, bits);
+			err = set_state_bits(tree, state, &bits);
 			if (err)
 				goto out;
 			cache_state(state, cached_state);
@@ -877,7 +877,7 @@ hit_next:
 		else
 			this_end = last_start - 1;
 		err = insert_state(tree, prealloc, start, this_end,
-				   bits);
+				   &bits);
 		BUG_ON(err == -EEXIST);
 		if (err) {
 			prealloc = NULL;
@@ -903,7 +903,7 @@ hit_next:
 		err = split_state(tree, state, prealloc, end + 1);
 		BUG_ON(err == -EEXIST);
 
-		err = set_state_bits(tree, prealloc, bits);
+		err = set_state_bits(tree, prealloc, &bits);
 		if (err) {
 			prealloc = NULL;
 			goto out;
@@ -966,8 +966,7 @@ int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
 {
 	return clear_extent_bit(tree, start, end,
 				EXTENT_DIRTY | EXTENT_DELALLOC |
-				EXTENT_DO_ACCOUNTING, 0, 0,
-				NULL, mask);
+				EXTENT_DO_ACCOUNTING, 0, 0, NULL, mask);
 }
 
 int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
@@ -1435,9 +1434,6 @@ int extent_clear_unlock_delalloc(struct inode *inode,
 	if (op & EXTENT_CLEAR_DELALLOC)
 		clear_bits |= EXTENT_DELALLOC;
 
-	if (op & EXTENT_CLEAR_ACCOUNTING)
-		clear_bits |= EXTENT_DO_ACCOUNTING;
-
 	clear_extent_bit(tree, start, end, clear_bits, 1, 0, NULL, GFP_NOFS);
 	if (!(op & (EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
 		    EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK |
@@ -1916,7 +1912,7 @@ static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
 
 	if (tree->ops && tree->ops->submit_bio_hook)
 		tree->ops->submit_bio_hook(page->mapping->host, rw, bio,
-					   mirror_num, bio_flags);
+					   mirror_num, bio_flags, start);
 	else
 		submit_bio(rw, bio);
 	if (bio_flagged(bio, BIO_EOPNOTSUPP))
@@ -2020,6 +2016,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 	sector_t sector;
 	struct extent_map *em;
 	struct block_device *bdev;
+	struct btrfs_ordered_extent *ordered;
 	int ret;
 	int nr = 0;
 	size_t page_offset = 0;
@@ -2031,7 +2028,15 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 	set_page_extent_mapped(page);
 
 	end = page_end;
-	lock_extent(tree, start, end, GFP_NOFS);
+	while (1) {
+		lock_extent(tree, start, end, GFP_NOFS);
+		ordered = btrfs_lookup_ordered_extent(inode, start);
+		if (!ordered)
+			break;
+		unlock_extent(tree, start, end, GFP_NOFS);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		btrfs_put_ordered_extent(ordered);
+	}
 
 	if (page->index == last_byte >> PAGE_CACHE_SHIFT) {
 		char *userpage;
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index bbab4813646f..5691c7b590da 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -16,7 +16,9 @@
 #define EXTENT_BOUNDARY (1 << 9)
 #define EXTENT_NODATASUM (1 << 10)
 #define EXTENT_DO_ACCOUNTING (1 << 11)
+#define EXTENT_FIRST_DELALLOC (1 << 12)
 #define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK)
+#define EXTENT_CTLBITS (EXTENT_DO_ACCOUNTING | EXTENT_FIRST_DELALLOC)
 
 /* flags for bio submission */
 #define EXTENT_BIO_COMPRESSED 1
@@ -47,7 +49,7 @@ struct extent_state;
 
 typedef	int (extent_submit_bio_hook_t)(struct inode *inode, int rw,
 				       struct bio *bio, int mirror_num,
-				       unsigned long bio_flags);
+				       unsigned long bio_flags, u64 bio_offset);
 struct extent_io_ops {
 	int (*fill_delalloc)(struct inode *inode, struct page *locked_page,
 			     u64 start, u64 end, int *page_started,
@@ -69,10 +71,10 @@ struct extent_io_ops {
 				    struct extent_state *state);
 	int (*writepage_end_io_hook)(struct page *page, u64 start, u64 end,
 				      struct extent_state *state, int uptodate);
-	int (*set_bit_hook)(struct inode *inode, u64 start, u64 end,
-			    unsigned long old, unsigned long bits);
+	int (*set_bit_hook)(struct inode *inode, struct extent_state *state,
+			    int *bits);
 	int (*clear_bit_hook)(struct inode *inode, struct extent_state *state,
-			      unsigned long bits);
+			      int *bits);
 	int (*merge_extent_hook)(struct inode *inode,
 				 struct extent_state *new,
 				 struct extent_state *other);
@@ -176,6 +178,7 @@ u64 count_range_bits(struct extent_io_tree *tree,
 		     u64 *start, u64 search_end,
 		     u64 max_bytes, unsigned long bits);
 
+void free_extent_state(struct extent_state *state);
 int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		   int bits, int filled, struct extent_state *cached_state);
 int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
@@ -185,6 +188,9 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		     gfp_t mask);
 int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 		    int bits, gfp_t mask);
+int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		   int bits, int exclusive_bits, u64 *failed_start,
+		   struct extent_state **cached_state, gfp_t mask);
 int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
 			gfp_t mask);
 int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index 54a255065aa3..a562a250ae77 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -149,13 +149,14 @@ int btrfs_lookup_file_extent(struct btrfs_trans_handle *trans,
 }
 
 
-int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode,
-			  struct bio *bio, u32 *dst)
+static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
+				   struct inode *inode, struct bio *bio,
+				   u64 logical_offset, u32 *dst, int dio)
 {
 	u32 sum;
 	struct bio_vec *bvec = bio->bi_io_vec;
 	int bio_index = 0;
-	u64 offset;
+	u64 offset = 0;
 	u64 item_start_offset = 0;
 	u64 item_last_offset = 0;
 	u64 disk_bytenr;
@@ -174,8 +175,11 @@ int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode,
 	WARN_ON(bio->bi_vcnt <= 0);
 
 	disk_bytenr = (u64)bio->bi_sector << 9;
+	if (dio)
+		offset = logical_offset;
 	while (bio_index < bio->bi_vcnt) {
-		offset = page_offset(bvec->bv_page) + bvec->bv_offset;
+		if (!dio)
+			offset = page_offset(bvec->bv_page) + bvec->bv_offset;
 		ret = btrfs_find_ordered_sum(inode, offset, disk_bytenr, &sum);
 		if (ret == 0)
 			goto found;
@@ -238,6 +242,7 @@ found:
 		else
 			set_state_private(io_tree, offset, sum);
 		disk_bytenr += bvec->bv_len;
+		offset += bvec->bv_len;
 		bio_index++;
 		bvec++;
 	}
@@ -245,6 +250,18 @@ found:
 	return 0;
 }
 
+int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode,
+			  struct bio *bio, u32 *dst)
+{
+	return __btrfs_lookup_bio_sums(root, inode, bio, 0, dst, 0);
+}
+
+int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode,
+			      struct bio *bio, u64 offset, u32 *dst)
+{
+	return __btrfs_lookup_bio_sums(root, inode, bio, offset, dst, 1);
+}
+
 int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 			     struct list_head *list)
 {
@@ -657,6 +674,9 @@ again:
 		goto found;
 	}
 	ret = PTR_ERR(item);
+	if (ret != -EFBIG && ret != -ENOENT)
+		goto fail_unlock;
+
 	if (ret == -EFBIG) {
 		u32 item_size;
 		/* we found one, but it isn't big enough yet */
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 29ff749ff4ca..787b50a16a14 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -46,32 +46,42 @@
 static noinline int btrfs_copy_from_user(loff_t pos, int num_pages,
 					 int write_bytes,
 					 struct page **prepared_pages,
-					 const char __user *buf)
+					 struct iov_iter *i)
 {
-	long page_fault = 0;
-	int i;
+	size_t copied;
+	int pg = 0;
 	int offset = pos & (PAGE_CACHE_SIZE - 1);
 
-	for (i = 0; i < num_pages && write_bytes > 0; i++, offset = 0) {
+	while (write_bytes > 0) {
 		size_t count = min_t(size_t,
 				     PAGE_CACHE_SIZE - offset, write_bytes);
-		struct page *page = prepared_pages[i];
-		fault_in_pages_readable(buf, count);
+		struct page *page = prepared_pages[pg];
+again:
+		if (unlikely(iov_iter_fault_in_readable(i, count)))
+			return -EFAULT;
 
 		/* Copy data from userspace to the current page */
-		kmap(page);
-		page_fault = __copy_from_user(page_address(page) + offset,
-					      buf, count);
+		copied = iov_iter_copy_from_user(page, i, offset, count);
+
 		/* Flush processor's dcache for this page */
 		flush_dcache_page(page);
-		kunmap(page);
-		buf += count;
-		write_bytes -= count;
+		iov_iter_advance(i, copied);
+		write_bytes -= copied;
 
-		if (page_fault)
-			break;
+		if (unlikely(copied == 0)) {
+			count = min_t(size_t, PAGE_CACHE_SIZE - offset,
+				      iov_iter_single_seg_count(i));
+			goto again;
+		}
+
+		if (unlikely(copied < PAGE_CACHE_SIZE - offset)) {
+			offset += copied;
+		} else {
+			pg++;
+			offset = 0;
+		}
 	}
-	return page_fault ? -EFAULT : 0;
+	return 0;
 }
 
 /*
@@ -126,8 +136,7 @@ static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans,
 	end_of_last_block = start_pos + num_bytes - 1;
 	err = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block,
 					NULL);
-	if (err)
-		return err;
+	BUG_ON(err);
 
 	for (i = 0; i < num_pages; i++) {
 		struct page *p = pages[i];
@@ -142,7 +151,7 @@ static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans,
 		 * at this time.
 		 */
 	}
-	return err;
+	return 0;
 }
 
 /*
@@ -823,45 +832,46 @@ again:
 	return 0;
 }
 
-static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
-				size_t count, loff_t *ppos)
+static ssize_t btrfs_file_aio_write(struct kiocb *iocb,
+				    const struct iovec *iov,
+				    unsigned long nr_segs, loff_t pos)
 {
-	loff_t pos;
+	struct file *file = iocb->ki_filp;
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct page *pinned[2];
+	struct page **pages = NULL;
+	struct iov_iter i;
+	loff_t *ppos = &iocb->ki_pos;
 	loff_t start_pos;
 	ssize_t num_written = 0;
 	ssize_t err = 0;
+	size_t count;
+	size_t ocount;
 	int ret = 0;
-	struct inode *inode = fdentry(file)->d_inode;
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct page **pages = NULL;
 	int nrptrs;
-	struct page *pinned[2];
 	unsigned long first_index;
 	unsigned long last_index;
 	int will_write;
+	int buffered = 0;
 
 	will_write = ((file->f_flags & O_DSYNC) || IS_SYNC(inode) ||
 		      (file->f_flags & O_DIRECT));
 
-	nrptrs = min((count + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE,
-		     PAGE_CACHE_SIZE / (sizeof(struct page *)));
 	pinned[0] = NULL;
 	pinned[1] = NULL;
 
-	pos = *ppos;
 	start_pos = pos;
 
 	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
 
-	/* do the reserve before the mutex lock in case we have to do some
-	 * flushing.  We wouldn't deadlock, but this is more polite.
-	 */
-	err = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
-	if (err)
-		goto out_nolock;
-
 	mutex_lock(&inode->i_mutex);
 
+	err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ);
+	if (err)
+		goto out;
+	count = ocount;
+
 	current->backing_dev_info = inode->i_mapping->backing_dev_info;
 	err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
 	if (err)
@@ -875,15 +885,53 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		goto out;
 
 	file_update_time(file);
+	BTRFS_I(inode)->sequence++;
+
+	if (unlikely(file->f_flags & O_DIRECT)) {
+		num_written = generic_file_direct_write(iocb, iov, &nr_segs,
+							pos, ppos, count,
+							ocount);
+		/*
+		 * the generic O_DIRECT will update in-memory i_size after the
+		 * DIOs are done.  But our endio handlers that update the on
+		 * disk i_size never update past the in memory i_size.  So we
+		 * need one more update here to catch any additions to the
+		 * file
+		 */
+		if (inode->i_size != BTRFS_I(inode)->disk_i_size) {
+			btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
+			mark_inode_dirty(inode);
+		}
 
+		if (num_written < 0) {
+			ret = num_written;
+			num_written = 0;
+			goto out;
+		} else if (num_written == count) {
+			/* pick up pos changes done by the generic code */
+			pos = *ppos;
+			goto out;
+		}
+		/*
+		 * We are going to do buffered for the rest of the range, so we
+		 * need to make sure to invalidate the buffered pages when we're
+		 * done.
+		 */
+		buffered = 1;
+		pos += num_written;
+	}
+
+	iov_iter_init(&i, iov, nr_segs, count, num_written);
+	nrptrs = min((iov_iter_count(&i) + PAGE_CACHE_SIZE - 1) /
+		     PAGE_CACHE_SIZE, PAGE_CACHE_SIZE /
+		     (sizeof(struct page *)));
 	pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL);
 
 	/* generic_write_checks can change our pos */
 	start_pos = pos;
 
-	BTRFS_I(inode)->sequence++;
 	first_index = pos >> PAGE_CACHE_SHIFT;
-	last_index = (pos + count) >> PAGE_CACHE_SHIFT;
+	last_index = (pos + iov_iter_count(&i)) >> PAGE_CACHE_SHIFT;
 
 	/*
 	 * there are lots of better ways to do this, but this code
@@ -900,7 +948,7 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 			unlock_page(pinned[0]);
 		}
 	}
-	if ((pos + count) & (PAGE_CACHE_SIZE - 1)) {
+	if ((pos + iov_iter_count(&i)) & (PAGE_CACHE_SIZE - 1)) {
 		pinned[1] = grab_cache_page(inode->i_mapping, last_index);
 		if (!PageUptodate(pinned[1])) {
 			ret = btrfs_readpage(NULL, pinned[1]);
@@ -911,10 +959,10 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		}
 	}
 
-	while (count > 0) {
+	while (iov_iter_count(&i) > 0) {
 		size_t offset = pos & (PAGE_CACHE_SIZE - 1);
-		size_t write_bytes = min(count, nrptrs *
-					(size_t)PAGE_CACHE_SIZE -
+		size_t write_bytes = min(iov_iter_count(&i),
+					 nrptrs * (size_t)PAGE_CACHE_SIZE -
 					 offset);
 		size_t num_pages = (write_bytes + PAGE_CACHE_SIZE - 1) >>
 					PAGE_CACHE_SHIFT;
@@ -922,7 +970,7 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		WARN_ON(num_pages > nrptrs);
 		memset(pages, 0, sizeof(struct page *) * nrptrs);
 
-		ret = btrfs_check_data_free_space(root, inode, write_bytes);
+		ret = btrfs_delalloc_reserve_space(inode, write_bytes);
 		if (ret)
 			goto out;
 
@@ -930,26 +978,20 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 				    pos, first_index, last_index,
 				    write_bytes);
 		if (ret) {
-			btrfs_free_reserved_data_space(root, inode,
-						       write_bytes);
+			btrfs_delalloc_release_space(inode, write_bytes);
 			goto out;
 		}
 
 		ret = btrfs_copy_from_user(pos, num_pages,
-					   write_bytes, pages, buf);
-		if (ret) {
-			btrfs_free_reserved_data_space(root, inode,
-						       write_bytes);
-			btrfs_drop_pages(pages, num_pages);
-			goto out;
+					   write_bytes, pages, &i);
+		if (ret == 0) {
+			dirty_and_release_pages(NULL, root, file, pages,
+						num_pages, pos, write_bytes);
 		}
 
-		ret = dirty_and_release_pages(NULL, root, file, pages,
-					      num_pages, pos, write_bytes);
 		btrfs_drop_pages(pages, num_pages);
 		if (ret) {
-			btrfs_free_reserved_data_space(root, inode,
-						       write_bytes);
+			btrfs_delalloc_release_space(inode, write_bytes);
 			goto out;
 		}
 
@@ -965,8 +1007,6 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 			btrfs_throttle(root);
 		}
 
-		buf += write_bytes;
-		count -= write_bytes;
 		pos += write_bytes;
 		num_written += write_bytes;
 
@@ -976,9 +1016,7 @@ out:
 	mutex_unlock(&inode->i_mutex);
 	if (ret)
 		err = ret;
-	btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
 
-out_nolock:
 	kfree(pages);
 	if (pinned[0])
 		page_cache_release(pinned[0]);
@@ -1008,7 +1046,7 @@ out_nolock:
 			num_written = err;
 
 		if ((file->f_flags & O_DSYNC) || IS_SYNC(inode)) {
-			trans = btrfs_start_transaction(root, 1);
+			trans = btrfs_start_transaction(root, 0);
 			ret = btrfs_log_dentry_safe(trans, root,
 						    file->f_dentry);
 			if (ret == 0) {
@@ -1023,7 +1061,7 @@ out_nolock:
 				btrfs_end_transaction(trans, root);
 			}
 		}
-		if (file->f_flags & O_DIRECT) {
+		if (file->f_flags & O_DIRECT && buffered) {
 			invalidate_mapping_pages(inode->i_mapping,
 			      start_pos >> PAGE_CACHE_SHIFT,
 			     (start_pos + num_written - 1) >> PAGE_CACHE_SHIFT);
@@ -1063,8 +1101,9 @@ int btrfs_release_file(struct inode *inode, struct file *filp)
  * important optimization for directories because holding the mutex prevents
  * new operations on the dir while we write to disk.
  */
-int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
+int btrfs_sync_file(struct file *file, int datasync)
 {
+	struct dentry *dentry = file->f_path.dentry;
 	struct inode *inode = dentry->d_inode;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret = 0;
@@ -1104,9 +1143,9 @@ int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
 	if (file && file->private_data)
 		btrfs_ioctl_trans_end(file);
 
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans) {
-		ret = -ENOMEM;
+	trans = btrfs_start_transaction(root, 0);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
 		goto out;
 	}
 
@@ -1161,7 +1200,7 @@ const struct file_operations btrfs_file_operations = {
 	.read		= do_sync_read,
 	.aio_read       = generic_file_aio_read,
 	.splice_read	= generic_file_splice_read,
-	.write		= btrfs_file_write,
+	.aio_write	= btrfs_file_aio_write,
 	.mmap		= btrfs_file_mmap,
 	.open		= generic_file_open,
 	.release	= btrfs_release_file,
diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c
index 72ce3c173d6a..64f1150bb48d 100644
--- a/fs/btrfs/inode-item.c
+++ b/fs/btrfs/inode-item.c
@@ -49,6 +49,33 @@ static int find_name_in_backref(struct btrfs_path *path, const char *name,
 	return 0;
 }
 
+struct btrfs_inode_ref *
+btrfs_lookup_inode_ref(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct btrfs_path *path,
+			const char *name, int name_len,
+			u64 inode_objectid, u64 ref_objectid, int mod)
+{
+	struct btrfs_key key;
+	struct btrfs_inode_ref *ref;
+	int ins_len = mod < 0 ? -1 : 0;
+	int cow = mod != 0;
+	int ret;
+
+	key.objectid = inode_objectid;
+	key.type = BTRFS_INODE_REF_KEY;
+	key.offset = ref_objectid;
+
+	ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
+	if (ret < 0)
+		return ERR_PTR(ret);
+	if (ret > 0)
+		return NULL;
+	if (!find_name_in_backref(path, name, name_len, &ref))
+		return NULL;
+	return ref;
+}
+
 int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
 			   const char *name, int name_len,
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index d601629b85d1..fa6ccc1bfe2a 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -252,6 +252,7 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
 				   inline_len, compressed_size,
 				   compressed_pages);
 	BUG_ON(ret);
+	btrfs_delalloc_release_metadata(inode, end + 1 - start);
 	btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0);
 	return 0;
 }
@@ -414,6 +415,7 @@ again:
 		trans = btrfs_join_transaction(root, 1);
 		BUG_ON(!trans);
 		btrfs_set_trans_block_group(trans, inode);
+		trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 
 		/* lets try to make an inline extent */
 		if (ret || total_in < (actual_end - start)) {
@@ -439,7 +441,6 @@ again:
 			     start, end, NULL,
 			     EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
 			     EXTENT_CLEAR_DELALLOC |
-			     EXTENT_CLEAR_ACCOUNTING |
 			     EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK);
 
 			btrfs_end_transaction(trans, root);
@@ -697,6 +698,38 @@ retry:
 	return 0;
 }
 
+static u64 get_extent_allocation_hint(struct inode *inode, u64 start,
+				      u64 num_bytes)
+{
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct extent_map *em;
+	u64 alloc_hint = 0;
+
+	read_lock(&em_tree->lock);
+	em = search_extent_mapping(em_tree, start, num_bytes);
+	if (em) {
+		/*
+		 * if block start isn't an actual block number then find the
+		 * first block in this inode and use that as a hint.  If that
+		 * block is also bogus then just don't worry about it.
+		 */
+		if (em->block_start >= EXTENT_MAP_LAST_BYTE) {
+			free_extent_map(em);
+			em = search_extent_mapping(em_tree, 0, 0);
+			if (em && em->block_start < EXTENT_MAP_LAST_BYTE)
+				alloc_hint = em->block_start;
+			if (em)
+				free_extent_map(em);
+		} else {
+			alloc_hint = em->block_start;
+			free_extent_map(em);
+		}
+	}
+	read_unlock(&em_tree->lock);
+
+	return alloc_hint;
+}
+
 /*
  * when extent_io.c finds a delayed allocation range in the file,
  * the call backs end up in this code.  The basic idea is to
@@ -734,6 +767,7 @@ static noinline int cow_file_range(struct inode *inode,
 	trans = btrfs_join_transaction(root, 1);
 	BUG_ON(!trans);
 	btrfs_set_trans_block_group(trans, inode);
+	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 
 	actual_end = min_t(u64, isize, end + 1);
 
@@ -753,7 +787,6 @@ static noinline int cow_file_range(struct inode *inode,
 				     EXTENT_CLEAR_UNLOCK_PAGE |
 				     EXTENT_CLEAR_UNLOCK |
 				     EXTENT_CLEAR_DELALLOC |
-				     EXTENT_CLEAR_ACCOUNTING |
 				     EXTENT_CLEAR_DIRTY |
 				     EXTENT_SET_WRITEBACK |
 				     EXTENT_END_WRITEBACK);
@@ -769,29 +802,7 @@ static noinline int cow_file_range(struct inode *inode,
 	BUG_ON(disk_num_bytes >
 	       btrfs_super_total_bytes(&root->fs_info->super_copy));
 
-
-	read_lock(&BTRFS_I(inode)->extent_tree.lock);
-	em = search_extent_mapping(&BTRFS_I(inode)->extent_tree,
-				   start, num_bytes);
-	if (em) {
-		/*
-		 * if block start isn't an actual block number then find the
-		 * first block in this inode and use that as a hint.  If that
-		 * block is also bogus then just don't worry about it.
-		 */
-		if (em->block_start >= EXTENT_MAP_LAST_BYTE) {
-			free_extent_map(em);
-			em = search_extent_mapping(em_tree, 0, 0);
-			if (em && em->block_start < EXTENT_MAP_LAST_BYTE)
-				alloc_hint = em->block_start;
-			if (em)
-				free_extent_map(em);
-		} else {
-			alloc_hint = em->block_start;
-			free_extent_map(em);
-		}
-	}
-	read_unlock(&BTRFS_I(inode)->extent_tree.lock);
+	alloc_hint = get_extent_allocation_hint(inode, start, num_bytes);
 	btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0);
 
 	while (disk_num_bytes > 0) {
@@ -1174,6 +1185,13 @@ out_check:
 					       num_bytes, num_bytes, type);
 		BUG_ON(ret);
 
+		if (root->root_key.objectid ==
+		    BTRFS_DATA_RELOC_TREE_OBJECTID) {
+			ret = btrfs_reloc_clone_csums(inode, cur_offset,
+						      num_bytes);
+			BUG_ON(ret);
+		}
+
 		extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
 				cur_offset, cur_offset + num_bytes - 1,
 				locked_page, EXTENT_CLEAR_UNLOCK_PAGE |
@@ -1226,15 +1244,13 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page,
 }
 
 static int btrfs_split_extent_hook(struct inode *inode,
-				    struct extent_state *orig, u64 split)
+				   struct extent_state *orig, u64 split)
 {
+	/* not delalloc, ignore it */
 	if (!(orig->state & EXTENT_DELALLOC))
 		return 0;
 
-	spin_lock(&BTRFS_I(inode)->accounting_lock);
-	BTRFS_I(inode)->outstanding_extents++;
-	spin_unlock(&BTRFS_I(inode)->accounting_lock);
-
+	atomic_inc(&BTRFS_I(inode)->outstanding_extents);
 	return 0;
 }
 
@@ -1252,10 +1268,7 @@ static int btrfs_merge_extent_hook(struct inode *inode,
 	if (!(other->state & EXTENT_DELALLOC))
 		return 0;
 
-	spin_lock(&BTRFS_I(inode)->accounting_lock);
-	BTRFS_I(inode)->outstanding_extents--;
-	spin_unlock(&BTRFS_I(inode)->accounting_lock);
-
+	atomic_dec(&BTRFS_I(inode)->outstanding_extents);
 	return 0;
 }
 
@@ -1264,8 +1277,8 @@ static int btrfs_merge_extent_hook(struct inode *inode,
  * bytes in this file, and to maintain the list of inodes that
  * have pending delalloc work to be done.
  */
-static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
-		       unsigned long old, unsigned long bits)
+static int btrfs_set_bit_hook(struct inode *inode,
+			      struct extent_state *state, int *bits)
 {
 
 	/*
@@ -1273,17 +1286,18 @@ static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
 	 * but in this case, we are only testeing for the DELALLOC
 	 * bit, which is only set or cleared with irqs on
 	 */
-	if (!(old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
+	if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
 		struct btrfs_root *root = BTRFS_I(inode)->root;
+		u64 len = state->end + 1 - state->start;
 
-		spin_lock(&BTRFS_I(inode)->accounting_lock);
-		BTRFS_I(inode)->outstanding_extents++;
-		spin_unlock(&BTRFS_I(inode)->accounting_lock);
-		btrfs_delalloc_reserve_space(root, inode, end - start + 1);
+		if (*bits & EXTENT_FIRST_DELALLOC)
+			*bits &= ~EXTENT_FIRST_DELALLOC;
+		else
+			atomic_inc(&BTRFS_I(inode)->outstanding_extents);
 
 		spin_lock(&root->fs_info->delalloc_lock);
-		BTRFS_I(inode)->delalloc_bytes += end - start + 1;
-		root->fs_info->delalloc_bytes += end - start + 1;
+		BTRFS_I(inode)->delalloc_bytes += len;
+		root->fs_info->delalloc_bytes += len;
 		if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
 			list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
 				      &root->fs_info->delalloc_inodes);
@@ -1297,45 +1311,32 @@ static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
  * extent_io.c clear_bit_hook, see set_bit_hook for why
  */
 static int btrfs_clear_bit_hook(struct inode *inode,
-				struct extent_state *state, unsigned long bits)
+				struct extent_state *state, int *bits)
 {
 	/*
 	 * set_bit and clear bit hooks normally require _irqsave/restore
 	 * but in this case, we are only testeing for the DELALLOC
 	 * bit, which is only set or cleared with irqs on
 	 */
-	if ((state->state & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
+	if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
 		struct btrfs_root *root = BTRFS_I(inode)->root;
+		u64 len = state->end + 1 - state->start;
 
-		if (bits & EXTENT_DO_ACCOUNTING) {
-			spin_lock(&BTRFS_I(inode)->accounting_lock);
-			WARN_ON(!BTRFS_I(inode)->outstanding_extents);
-			BTRFS_I(inode)->outstanding_extents--;
-			spin_unlock(&BTRFS_I(inode)->accounting_lock);
-			btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
-		}
+		if (*bits & EXTENT_FIRST_DELALLOC)
+			*bits &= ~EXTENT_FIRST_DELALLOC;
+		else if (!(*bits & EXTENT_DO_ACCOUNTING))
+			atomic_dec(&BTRFS_I(inode)->outstanding_extents);
+
+		if (*bits & EXTENT_DO_ACCOUNTING)
+			btrfs_delalloc_release_metadata(inode, len);
+
+		if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID)
+			btrfs_free_reserved_data_space(inode, len);
 
 		spin_lock(&root->fs_info->delalloc_lock);
-		if (state->end - state->start + 1 >
-		    root->fs_info->delalloc_bytes) {
-			printk(KERN_INFO "btrfs warning: delalloc account "
-			       "%llu %llu\n",
-			       (unsigned long long)
-			       state->end - state->start + 1,
-			       (unsigned long long)
-			       root->fs_info->delalloc_bytes);
-			btrfs_delalloc_free_space(root, inode, (u64)-1);
-			root->fs_info->delalloc_bytes = 0;
-			BTRFS_I(inode)->delalloc_bytes = 0;
-		} else {
-			btrfs_delalloc_free_space(root, inode,
-						  state->end -
-						  state->start + 1);
-			root->fs_info->delalloc_bytes -= state->end -
-				state->start + 1;
-			BTRFS_I(inode)->delalloc_bytes -= state->end -
-				state->start + 1;
-		}
+		root->fs_info->delalloc_bytes -= len;
+		BTRFS_I(inode)->delalloc_bytes -= len;
+
 		if (BTRFS_I(inode)->delalloc_bytes == 0 &&
 		    !list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
 			list_del_init(&BTRFS_I(inode)->delalloc_inodes);
@@ -1384,7 +1385,8 @@ int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
  */
 static int __btrfs_submit_bio_start(struct inode *inode, int rw,
 				    struct bio *bio, int mirror_num,
-				    unsigned long bio_flags)
+				    unsigned long bio_flags,
+				    u64 bio_offset)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret = 0;
@@ -1403,7 +1405,8 @@ static int __btrfs_submit_bio_start(struct inode *inode, int rw,
  * are inserted into the btree
  */
 static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
-			  int mirror_num, unsigned long bio_flags)
+			  int mirror_num, unsigned long bio_flags,
+			  u64 bio_offset)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	return btrfs_map_bio(root, rw, bio, mirror_num, 1);
@@ -1414,7 +1417,8 @@ static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
  * on write, or reading the csums from the tree before a read
  */
 static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
-			  int mirror_num, unsigned long bio_flags)
+			  int mirror_num, unsigned long bio_flags,
+			  u64 bio_offset)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret = 0;
@@ -1439,7 +1443,8 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 		/* we're doing a write, do the async checksumming */
 		return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
 				   inode, rw, bio, mirror_num,
-				   bio_flags, __btrfs_submit_bio_start,
+				   bio_flags, bio_offset,
+				   __btrfs_submit_bio_start,
 				   __btrfs_submit_bio_done);
 	}
 
@@ -1520,6 +1525,7 @@ again:
 		goto again;
 	}
 
+	BUG();
 	btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state);
 	ClearPageChecked(page);
 out:
@@ -1650,7 +1656,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct btrfs_trans_handle *trans;
+	struct btrfs_trans_handle *trans = NULL;
 	struct btrfs_ordered_extent *ordered_extent = NULL;
 	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
 	struct extent_state *cached_state = NULL;
@@ -1668,9 +1674,10 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 		ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent);
 		if (!ret) {
 			trans = btrfs_join_transaction(root, 1);
+			btrfs_set_trans_block_group(trans, inode);
+			trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 			ret = btrfs_update_inode(trans, root, inode);
 			BUG_ON(ret);
-			btrfs_end_transaction(trans, root);
 		}
 		goto out;
 	}
@@ -1680,6 +1687,8 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 			 0, &cached_state, GFP_NOFS);
 
 	trans = btrfs_join_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, inode);
+	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 
 	if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
 		compressed = 1;
@@ -1711,12 +1720,13 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 	add_pending_csums(trans, inode, ordered_extent->file_offset,
 			  &ordered_extent->list);
 
-	/* this also removes the ordered extent from the tree */
 	btrfs_ordered_update_i_size(inode, 0, ordered_extent);
 	ret = btrfs_update_inode(trans, root, inode);
 	BUG_ON(ret);
-	btrfs_end_transaction(trans, root);
 out:
+	btrfs_delalloc_release_metadata(inode, ordered_extent->len);
+	if (trans)
+		btrfs_end_transaction(trans, root);
 	/* once for us */
 	btrfs_put_ordered_extent(ordered_extent);
 	/* once for the tree */
@@ -1838,7 +1848,7 @@ static int btrfs_io_failed_hook(struct bio *failed_bio,
 
 	BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio,
 						      failrec->last_mirror,
-						      failrec->bio_flags);
+						      failrec->bio_flags, 0);
 	return 0;
 }
 
@@ -1993,32 +2003,196 @@ void btrfs_run_delayed_iputs(struct btrfs_root *root)
 }
 
 /*
+ * calculate extra metadata reservation when snapshotting a subvolume
+ * contains orphan files.
+ */
+void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans,
+				struct btrfs_pending_snapshot *pending,
+				u64 *bytes_to_reserve)
+{
+	struct btrfs_root *root;
+	struct btrfs_block_rsv *block_rsv;
+	u64 num_bytes;
+	int index;
+
+	root = pending->root;
+	if (!root->orphan_block_rsv || list_empty(&root->orphan_list))
+		return;
+
+	block_rsv = root->orphan_block_rsv;
+
+	/* orphan block reservation for the snapshot */
+	num_bytes = block_rsv->size;
+
+	/*
+	 * after the snapshot is created, COWing tree blocks may use more
+	 * space than it frees. So we should make sure there is enough
+	 * reserved space.
+	 */
+	index = trans->transid & 0x1;
+	if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) {
+		num_bytes += block_rsv->size -
+			     (block_rsv->reserved + block_rsv->freed[index]);
+	}
+
+	*bytes_to_reserve += num_bytes;
+}
+
+void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans,
+				struct btrfs_pending_snapshot *pending)
+{
+	struct btrfs_root *root = pending->root;
+	struct btrfs_root *snap = pending->snap;
+	struct btrfs_block_rsv *block_rsv;
+	u64 num_bytes;
+	int index;
+	int ret;
+
+	if (!root->orphan_block_rsv || list_empty(&root->orphan_list))
+		return;
+
+	/* refill source subvolume's orphan block reservation */
+	block_rsv = root->orphan_block_rsv;
+	index = trans->transid & 0x1;
+	if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) {
+		num_bytes = block_rsv->size -
+			    (block_rsv->reserved + block_rsv->freed[index]);
+		ret = btrfs_block_rsv_migrate(&pending->block_rsv,
+					      root->orphan_block_rsv,
+					      num_bytes);
+		BUG_ON(ret);
+	}
+
+	/* setup orphan block reservation for the snapshot */
+	block_rsv = btrfs_alloc_block_rsv(snap);
+	BUG_ON(!block_rsv);
+
+	btrfs_add_durable_block_rsv(root->fs_info, block_rsv);
+	snap->orphan_block_rsv = block_rsv;
+
+	num_bytes = root->orphan_block_rsv->size;
+	ret = btrfs_block_rsv_migrate(&pending->block_rsv,
+				      block_rsv, num_bytes);
+	BUG_ON(ret);
+
+#if 0
+	/* insert orphan item for the snapshot */
+	WARN_ON(!root->orphan_item_inserted);
+	ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root,
+				       snap->root_key.objectid);
+	BUG_ON(ret);
+	snap->orphan_item_inserted = 1;
+#endif
+}
+
+enum btrfs_orphan_cleanup_state {
+	ORPHAN_CLEANUP_STARTED	= 1,
+	ORPHAN_CLEANUP_DONE	= 2,
+};
+
+/*
+ * This is called in transaction commmit time. If there are no orphan
+ * files in the subvolume, it removes orphan item and frees block_rsv
+ * structure.
+ */
+void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root)
+{
+	int ret;
+
+	if (!list_empty(&root->orphan_list) ||
+	    root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE)
+		return;
+
+	if (root->orphan_item_inserted &&
+	    btrfs_root_refs(&root->root_item) > 0) {
+		ret = btrfs_del_orphan_item(trans, root->fs_info->tree_root,
+					    root->root_key.objectid);
+		BUG_ON(ret);
+		root->orphan_item_inserted = 0;
+	}
+
+	if (root->orphan_block_rsv) {
+		WARN_ON(root->orphan_block_rsv->size > 0);
+		btrfs_free_block_rsv(root, root->orphan_block_rsv);
+		root->orphan_block_rsv = NULL;
+	}
+}
+
+/*
  * This creates an orphan entry for the given inode in case something goes
  * wrong in the middle of an unlink/truncate.
+ *
+ * NOTE: caller of this function should reserve 5 units of metadata for
+ *	 this function.
  */
 int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-	int ret = 0;
+	struct btrfs_block_rsv *block_rsv = NULL;
+	int reserve = 0;
+	int insert = 0;
+	int ret;
 
-	spin_lock(&root->list_lock);
+	if (!root->orphan_block_rsv) {
+		block_rsv = btrfs_alloc_block_rsv(root);
+		BUG_ON(!block_rsv);
+	}
 
-	/* already on the orphan list, we're good */
-	if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
-		spin_unlock(&root->list_lock);
-		return 0;
+	spin_lock(&root->orphan_lock);
+	if (!root->orphan_block_rsv) {
+		root->orphan_block_rsv = block_rsv;
+	} else if (block_rsv) {
+		btrfs_free_block_rsv(root, block_rsv);
+		block_rsv = NULL;
+	}
+
+	if (list_empty(&BTRFS_I(inode)->i_orphan)) {
+		list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+#if 0
+		/*
+		 * For proper ENOSPC handling, we should do orphan
+		 * cleanup when mounting. But this introduces backward
+		 * compatibility issue.
+		 */
+		if (!xchg(&root->orphan_item_inserted, 1))
+			insert = 2;
+		else
+			insert = 1;
+#endif
+		insert = 1;
+	} else {
+		WARN_ON(!BTRFS_I(inode)->orphan_meta_reserved);
 	}
 
-	list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+	if (!BTRFS_I(inode)->orphan_meta_reserved) {
+		BTRFS_I(inode)->orphan_meta_reserved = 1;
+		reserve = 1;
+	}
+	spin_unlock(&root->orphan_lock);
 
-	spin_unlock(&root->list_lock);
+	if (block_rsv)
+		btrfs_add_durable_block_rsv(root->fs_info, block_rsv);
 
-	/*
-	 * insert an orphan item to track this unlinked/truncated file
-	 */
-	ret = btrfs_insert_orphan_item(trans, root, inode->i_ino);
+	/* grab metadata reservation from transaction handle */
+	if (reserve) {
+		ret = btrfs_orphan_reserve_metadata(trans, inode);
+		BUG_ON(ret);
+	}
 
-	return ret;
+	/* insert an orphan item to track this unlinked/truncated file */
+	if (insert >= 1) {
+		ret = btrfs_insert_orphan_item(trans, root, inode->i_ino);
+		BUG_ON(ret);
+	}
+
+	/* insert an orphan item to track subvolume contains orphan files */
+	if (insert >= 2) {
+		ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root,
+					       root->root_key.objectid);
+		BUG_ON(ret);
+	}
+	return 0;
 }
 
 /*
@@ -2028,26 +2202,31 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
 int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int delete_item = 0;
+	int release_rsv = 0;
 	int ret = 0;
 
-	spin_lock(&root->list_lock);
-
-	if (list_empty(&BTRFS_I(inode)->i_orphan)) {
-		spin_unlock(&root->list_lock);
-		return 0;
+	spin_lock(&root->orphan_lock);
+	if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
+		list_del_init(&BTRFS_I(inode)->i_orphan);
+		delete_item = 1;
 	}
 
-	list_del_init(&BTRFS_I(inode)->i_orphan);
-	if (!trans) {
-		spin_unlock(&root->list_lock);
-		return 0;
+	if (BTRFS_I(inode)->orphan_meta_reserved) {
+		BTRFS_I(inode)->orphan_meta_reserved = 0;
+		release_rsv = 1;
 	}
+	spin_unlock(&root->orphan_lock);
 
-	spin_unlock(&root->list_lock);
+	if (trans && delete_item) {
+		ret = btrfs_del_orphan_item(trans, root, inode->i_ino);
+		BUG_ON(ret);
+	}
 
-	ret = btrfs_del_orphan_item(trans, root, inode->i_ino);
+	if (release_rsv)
+		btrfs_orphan_release_metadata(inode);
 
-	return ret;
+	return 0;
 }
 
 /*
@@ -2064,7 +2243,7 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
 	struct inode *inode;
 	int ret = 0, nr_unlink = 0, nr_truncate = 0;
 
-	if (!xchg(&root->clean_orphans, 0))
+	if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED))
 		return;
 
 	path = btrfs_alloc_path();
@@ -2117,16 +2296,15 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
 		found_key.type = BTRFS_INODE_ITEM_KEY;
 		found_key.offset = 0;
 		inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL);
-		if (IS_ERR(inode))
-			break;
+		BUG_ON(IS_ERR(inode));
 
 		/*
 		 * add this inode to the orphan list so btrfs_orphan_del does
 		 * the proper thing when we hit it
 		 */
-		spin_lock(&root->list_lock);
+		spin_lock(&root->orphan_lock);
 		list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
-		spin_unlock(&root->list_lock);
+		spin_unlock(&root->orphan_lock);
 
 		/*
 		 * if this is a bad inode, means we actually succeeded in
@@ -2135,7 +2313,7 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
 		 * do a destroy_inode
 		 */
 		if (is_bad_inode(inode)) {
-			trans = btrfs_start_transaction(root, 1);
+			trans = btrfs_start_transaction(root, 0);
 			btrfs_orphan_del(trans, inode);
 			btrfs_end_transaction(trans, root);
 			iput(inode);
@@ -2153,13 +2331,23 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
 		/* this will do delete_inode and everything for us */
 		iput(inode);
 	}
+	btrfs_free_path(path);
+
+	root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE;
+
+	if (root->orphan_block_rsv)
+		btrfs_block_rsv_release(root, root->orphan_block_rsv,
+					(u64)-1);
+
+	if (root->orphan_block_rsv || root->orphan_item_inserted) {
+		trans = btrfs_join_transaction(root, 1);
+		btrfs_end_transaction(trans, root);
+	}
 
 	if (nr_unlink)
 		printk(KERN_INFO "btrfs: unlinked %d orphans\n", nr_unlink);
 	if (nr_truncate)
 		printk(KERN_INFO "btrfs: truncated %d orphans\n", nr_truncate);
-
-	btrfs_free_path(path);
 }
 
 /*
@@ -2478,29 +2666,201 @@ out:
 	return ret;
 }
 
-static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
+/* helper to check if there is any shared block in the path */
+static int check_path_shared(struct btrfs_root *root,
+			     struct btrfs_path *path)
+{
+	struct extent_buffer *eb;
+	int level;
+	int ret;
+	u64 refs;
+
+	for (level = 0; level < BTRFS_MAX_LEVEL; level++) {
+		if (!path->nodes[level])
+			break;
+		eb = path->nodes[level];
+		if (!btrfs_block_can_be_shared(root, eb))
+			continue;
+		ret = btrfs_lookup_extent_info(NULL, root, eb->start, eb->len,
+					       &refs, NULL);
+		if (refs > 1)
+			return 1;
+	}
+	return 0;
+}
+
+/*
+ * helper to start transaction for unlink and rmdir.
+ *
+ * unlink and rmdir are special in btrfs, they do not always free space.
+ * so in enospc case, we should make sure they will free space before
+ * allowing them to use the global metadata reservation.
+ */
+static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
+						       struct dentry *dentry)
 {
-	struct btrfs_root *root;
 	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_path *path;
+	struct btrfs_inode_ref *ref;
+	struct btrfs_dir_item *di;
 	struct inode *inode = dentry->d_inode;
+	u64 index;
+	int check_link = 1;
+	int err = -ENOSPC;
 	int ret;
-	unsigned long nr = 0;
 
-	root = BTRFS_I(dir)->root;
+	trans = btrfs_start_transaction(root, 10);
+	if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC)
+		return trans;
 
-	/*
-	 * 5 items for unlink inode
-	 * 1 for orphan
-	 */
-	ret = btrfs_reserve_metadata_space(root, 6);
-	if (ret)
-		return ret;
+	if (inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
+		return ERR_PTR(-ENOSPC);
+
+	/* check if there is someone else holds reference */
+	if (S_ISDIR(inode->i_mode) && atomic_read(&inode->i_count) > 1)
+		return ERR_PTR(-ENOSPC);
+
+	if (atomic_read(&inode->i_count) > 2)
+		return ERR_PTR(-ENOSPC);
+
+	if (xchg(&root->fs_info->enospc_unlink, 1))
+		return ERR_PTR(-ENOSPC);
 
-	trans = btrfs_start_transaction(root, 1);
+	path = btrfs_alloc_path();
+	if (!path) {
+		root->fs_info->enospc_unlink = 0;
+		return ERR_PTR(-ENOMEM);
+	}
+
+	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
-		btrfs_unreserve_metadata_space(root, 6);
-		return PTR_ERR(trans);
+		btrfs_free_path(path);
+		root->fs_info->enospc_unlink = 0;
+		return trans;
+	}
+
+	path->skip_locking = 1;
+	path->search_commit_root = 1;
+
+	ret = btrfs_lookup_inode(trans, root, path,
+				&BTRFS_I(dir)->location, 0);
+	if (ret < 0) {
+		err = ret;
+		goto out;
+	}
+	if (ret == 0) {
+		if (check_path_shared(root, path))
+			goto out;
+	} else {
+		check_link = 0;
+	}
+	btrfs_release_path(root, path);
+
+	ret = btrfs_lookup_inode(trans, root, path,
+				&BTRFS_I(inode)->location, 0);
+	if (ret < 0) {
+		err = ret;
+		goto out;
+	}
+	if (ret == 0) {
+		if (check_path_shared(root, path))
+			goto out;
+	} else {
+		check_link = 0;
+	}
+	btrfs_release_path(root, path);
+
+	if (ret == 0 && S_ISREG(inode->i_mode)) {
+		ret = btrfs_lookup_file_extent(trans, root, path,
+					       inode->i_ino, (u64)-1, 0);
+		if (ret < 0) {
+			err = ret;
+			goto out;
+		}
+		BUG_ON(ret == 0);
+		if (check_path_shared(root, path))
+			goto out;
+		btrfs_release_path(root, path);
+	}
+
+	if (!check_link) {
+		err = 0;
+		goto out;
+	}
+
+	di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+				dentry->d_name.name, dentry->d_name.len, 0);
+	if (IS_ERR(di)) {
+		err = PTR_ERR(di);
+		goto out;
+	}
+	if (di) {
+		if (check_path_shared(root, path))
+			goto out;
+	} else {
+		err = 0;
+		goto out;
 	}
+	btrfs_release_path(root, path);
+
+	ref = btrfs_lookup_inode_ref(trans, root, path,
+				dentry->d_name.name, dentry->d_name.len,
+				inode->i_ino, dir->i_ino, 0);
+	if (IS_ERR(ref)) {
+		err = PTR_ERR(ref);
+		goto out;
+	}
+	BUG_ON(!ref);
+	if (check_path_shared(root, path))
+		goto out;
+	index = btrfs_inode_ref_index(path->nodes[0], ref);
+	btrfs_release_path(root, path);
+
+	di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, index,
+				dentry->d_name.name, dentry->d_name.len, 0);
+	if (IS_ERR(di)) {
+		err = PTR_ERR(di);
+		goto out;
+	}
+	BUG_ON(ret == -ENOENT);
+	if (check_path_shared(root, path))
+		goto out;
+
+	err = 0;
+out:
+	btrfs_free_path(path);
+	if (err) {
+		btrfs_end_transaction(trans, root);
+		root->fs_info->enospc_unlink = 0;
+		return ERR_PTR(err);
+	}
+
+	trans->block_rsv = &root->fs_info->global_block_rsv;
+	return trans;
+}
+
+static void __unlink_end_trans(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root)
+{
+	if (trans->block_rsv == &root->fs_info->global_block_rsv) {
+		BUG_ON(!root->fs_info->enospc_unlink);
+		root->fs_info->enospc_unlink = 0;
+	}
+	btrfs_end_transaction_throttle(trans, root);
+}
+
+static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
+{
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_trans_handle *trans;
+	struct inode *inode = dentry->d_inode;
+	int ret;
+	unsigned long nr = 0;
+
+	trans = __unlink_start_trans(dir, dentry);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
 	btrfs_set_trans_block_group(trans, dir);
 
@@ -2508,14 +2868,15 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 
 	ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
 				 dentry->d_name.name, dentry->d_name.len);
+	BUG_ON(ret);
 
-	if (inode->i_nlink == 0)
+	if (inode->i_nlink == 0) {
 		ret = btrfs_orphan_add(trans, inode);
+		BUG_ON(ret);
+	}
 
 	nr = trans->blocks_used;
-
-	btrfs_end_transaction_throttle(trans, root);
-	btrfs_unreserve_metadata_space(root, 6);
+	__unlink_end_trans(trans, root);
 	btrfs_btree_balance_dirty(root, nr);
 	return ret;
 }
@@ -2587,7 +2948,6 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 {
 	struct inode *inode = dentry->d_inode;
 	int err = 0;
-	int ret;
 	struct btrfs_root *root = BTRFS_I(dir)->root;
 	struct btrfs_trans_handle *trans;
 	unsigned long nr = 0;
@@ -2596,15 +2956,9 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 	    inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
 		return -ENOTEMPTY;
 
-	ret = btrfs_reserve_metadata_space(root, 5);
-	if (ret)
-		return ret;
-
-	trans = btrfs_start_transaction(root, 1);
-	if (IS_ERR(trans)) {
-		btrfs_unreserve_metadata_space(root, 5);
+	trans = __unlink_start_trans(dir, dentry);
+	if (IS_ERR(trans))
 		return PTR_ERR(trans);
-	}
 
 	btrfs_set_trans_block_group(trans, dir);
 
@@ -2627,12 +2981,9 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 		btrfs_i_size_write(inode, 0);
 out:
 	nr = trans->blocks_used;
-	ret = btrfs_end_transaction_throttle(trans, root);
-	btrfs_unreserve_metadata_space(root, 5);
+	__unlink_end_trans(trans, root);
 	btrfs_btree_balance_dirty(root, nr);
 
-	if (ret && !err)
-		err = ret;
 	return err;
 }
 
@@ -3029,6 +3380,7 @@ out:
 	if (pending_del_nr) {
 		ret = btrfs_del_items(trans, root, path, pending_del_slot,
 				      pending_del_nr);
+		BUG_ON(ret);
 	}
 	btrfs_free_path(path);
 	return err;
@@ -3056,11 +3408,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from)
 
 	if ((offset & (blocksize - 1)) == 0)
 		goto out;
-	ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
-	if (ret)
-		goto out;
-
-	ret = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
+	ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
 	if (ret)
 		goto out;
 
@@ -3068,8 +3416,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from)
 again:
 	page = grab_cache_page(mapping, index);
 	if (!page) {
-		btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
-		btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
+		btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
 		goto out;
 	}
 
@@ -3132,8 +3479,7 @@ again:
 
 out_unlock:
 	if (ret)
-		btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
-	btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
+		btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
 	unlock_page(page);
 	page_cache_release(page);
 out:
@@ -3145,7 +3491,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
-	struct extent_map *em;
+	struct extent_map *em = NULL;
 	struct extent_state *cached_state = NULL;
 	u64 mask = root->sectorsize - 1;
 	u64 hole_start = (inode->i_size + mask) & ~mask;
@@ -3183,11 +3529,11 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
 			u64 hint_byte = 0;
 			hole_size = last_byte - cur_offset;
 
-			err = btrfs_reserve_metadata_space(root, 2);
-			if (err)
+			trans = btrfs_start_transaction(root, 2);
+			if (IS_ERR(trans)) {
+				err = PTR_ERR(trans);
 				break;
-
-			trans = btrfs_start_transaction(root, 1);
+			}
 			btrfs_set_trans_block_group(trans, inode);
 
 			err = btrfs_drop_extents(trans, inode, cur_offset,
@@ -3205,14 +3551,15 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
 					last_byte - 1, 0);
 
 			btrfs_end_transaction(trans, root);
-			btrfs_unreserve_metadata_space(root, 2);
 		}
 		free_extent_map(em);
+		em = NULL;
 		cur_offset = last_byte;
 		if (cur_offset >= block_end)
 			break;
 	}
 
+	free_extent_map(em);
 	unlock_extent_cached(io_tree, hole_start, block_end - 1, &cached_state,
 			     GFP_NOFS);
 	return err;
@@ -3239,11 +3586,10 @@ static int btrfs_setattr_size(struct inode *inode, struct iattr *attr)
 		}
 	}
 
-	ret = btrfs_reserve_metadata_space(root, 1);
-	if (ret)
-		return ret;
+	trans = btrfs_start_transaction(root, 5);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
-	trans = btrfs_start_transaction(root, 1);
 	btrfs_set_trans_block_group(trans, inode);
 
 	ret = btrfs_orphan_add(trans, inode);
@@ -3251,7 +3597,6 @@ static int btrfs_setattr_size(struct inode *inode, struct iattr *attr)
 
 	nr = trans->blocks_used;
 	btrfs_end_transaction(trans, root);
-	btrfs_unreserve_metadata_space(root, 1);
 	btrfs_btree_balance_dirty(root, nr);
 
 	if (attr->ia_size > inode->i_size) {
@@ -3264,8 +3609,11 @@ static int btrfs_setattr_size(struct inode *inode, struct iattr *attr)
 		i_size_write(inode, attr->ia_size);
 		btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
 
-		trans = btrfs_start_transaction(root, 1);
+		trans = btrfs_start_transaction(root, 0);
+		BUG_ON(IS_ERR(trans));
 		btrfs_set_trans_block_group(trans, inode);
+		trans->block_rsv = root->orphan_block_rsv;
+		BUG_ON(!trans->block_rsv);
 
 		ret = btrfs_update_inode(trans, root, inode);
 		BUG_ON(ret);
@@ -3345,10 +3693,21 @@ void btrfs_delete_inode(struct inode *inode)
 	btrfs_i_size_write(inode, 0);
 
 	while (1) {
-		trans = btrfs_start_transaction(root, 1);
+		trans = btrfs_start_transaction(root, 0);
+		BUG_ON(IS_ERR(trans));
 		btrfs_set_trans_block_group(trans, inode);
-		ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
+		trans->block_rsv = root->orphan_block_rsv;
+
+		ret = btrfs_block_rsv_check(trans, root,
+					    root->orphan_block_rsv, 0, 5);
+		if (ret) {
+			BUG_ON(ret != -EAGAIN);
+			ret = btrfs_commit_transaction(trans, root);
+			BUG_ON(ret);
+			continue;
+		}
 
+		ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
 		if (ret != -EAGAIN)
 			break;
 
@@ -3356,6 +3715,7 @@ void btrfs_delete_inode(struct inode *inode)
 		btrfs_end_transaction(trans, root);
 		trans = NULL;
 		btrfs_btree_balance_dirty(root, nr);
+
 	}
 
 	if (ret == 0) {
@@ -3596,40 +3956,10 @@ again:
 	return 0;
 }
 
-static noinline void init_btrfs_i(struct inode *inode)
-{
-	struct btrfs_inode *bi = BTRFS_I(inode);
-
-	bi->generation = 0;
-	bi->sequence = 0;
-	bi->last_trans = 0;
-	bi->last_sub_trans = 0;
-	bi->logged_trans = 0;
-	bi->delalloc_bytes = 0;
-	bi->reserved_bytes = 0;
-	bi->disk_i_size = 0;
-	bi->flags = 0;
-	bi->index_cnt = (u64)-1;
-	bi->last_unlink_trans = 0;
-	bi->ordered_data_close = 0;
-	bi->force_compress = 0;
-	extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
-	extent_io_tree_init(&BTRFS_I(inode)->io_tree,
-			     inode->i_mapping, GFP_NOFS);
-	extent_io_tree_init(&BTRFS_I(inode)->io_failure_tree,
-			     inode->i_mapping, GFP_NOFS);
-	INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
-	INIT_LIST_HEAD(&BTRFS_I(inode)->ordered_operations);
-	RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
-	btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
-	mutex_init(&BTRFS_I(inode)->log_mutex);
-}
-
 static int btrfs_init_locked_inode(struct inode *inode, void *p)
 {
 	struct btrfs_iget_args *args = p;
 	inode->i_ino = args->ino;
-	init_btrfs_i(inode);
 	BTRFS_I(inode)->root = args->root;
 	btrfs_set_inode_space_info(args->root, inode);
 	return 0;
@@ -3692,8 +4022,6 @@ static struct inode *new_simple_dir(struct super_block *s,
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
 
-	init_btrfs_i(inode);
-
 	BTRFS_I(inode)->root = root;
 	memcpy(&BTRFS_I(inode)->location, key, sizeof(*key));
 	BTRFS_I(inode)->dummy_inode = 1;
@@ -3950,7 +4278,7 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc)
 	struct btrfs_trans_handle *trans;
 	int ret = 0;
 
-	if (root->fs_info->btree_inode == inode)
+	if (BTRFS_I(inode)->dummy_inode)
 		return 0;
 
 	if (wbc->sync_mode == WB_SYNC_ALL) {
@@ -3971,10 +4299,38 @@ void btrfs_dirty_inode(struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_trans_handle *trans;
+	int ret;
+
+	if (BTRFS_I(inode)->dummy_inode)
+		return;
 
 	trans = btrfs_join_transaction(root, 1);
 	btrfs_set_trans_block_group(trans, inode);
-	btrfs_update_inode(trans, root, inode);
+
+	ret = btrfs_update_inode(trans, root, inode);
+	if (ret && ret == -ENOSPC) {
+		/* whoops, lets try again with the full transaction */
+		btrfs_end_transaction(trans, root);
+		trans = btrfs_start_transaction(root, 1);
+		if (IS_ERR(trans)) {
+			if (printk_ratelimit()) {
+				printk(KERN_ERR "btrfs: fail to "
+				       "dirty  inode %lu error %ld\n",
+				       inode->i_ino, PTR_ERR(trans));
+			}
+			return;
+		}
+		btrfs_set_trans_block_group(trans, inode);
+
+		ret = btrfs_update_inode(trans, root, inode);
+		if (ret) {
+			if (printk_ratelimit()) {
+				printk(KERN_ERR "btrfs: fail to "
+				       "dirty  inode %lu error %d\n",
+				       inode->i_ino, ret);
+			}
+		}
+	}
 	btrfs_end_transaction(trans, root);
 }
 
@@ -4092,7 +4448,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 	 * btrfs_get_inode_index_count has an explanation for the magic
 	 * number
 	 */
-	init_btrfs_i(inode);
 	BTRFS_I(inode)->index_cnt = 2;
 	BTRFS_I(inode)->root = root;
 	BTRFS_I(inode)->generation = trans->transid;
@@ -4247,26 +4602,21 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 	if (!new_valid_dev(rdev))
 		return -EINVAL;
 
+	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+	if (err)
+		return err;
+
 	/*
 	 * 2 for inode item and ref
 	 * 2 for dir items
 	 * 1 for xattr if selinux is on
 	 */
-	err = btrfs_reserve_metadata_space(root, 5);
-	if (err)
-		return err;
+	trans = btrfs_start_transaction(root, 5);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans)
-		goto fail;
 	btrfs_set_trans_block_group(trans, dir);
 
-	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
-	if (err) {
-		err = -ENOSPC;
-		goto out_unlock;
-	}
-
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
 				dentry->d_name.len,
 				dentry->d_parent->d_inode->i_ino, objectid,
@@ -4295,13 +4645,11 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 out_unlock:
 	nr = trans->blocks_used;
 	btrfs_end_transaction_throttle(trans, root);
-fail:
-	btrfs_unreserve_metadata_space(root, 5);
+	btrfs_btree_balance_dirty(root, nr);
 	if (drop_inode) {
 		inode_dec_link_count(inode);
 		iput(inode);
 	}
-	btrfs_btree_balance_dirty(root, nr);
 	return err;
 }
 
@@ -4311,32 +4659,26 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(dir)->root;
 	struct inode *inode = NULL;
-	int err;
 	int drop_inode = 0;
+	int err;
 	unsigned long nr = 0;
 	u64 objectid;
 	u64 index = 0;
 
+	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+	if (err)
+		return err;
 	/*
 	 * 2 for inode item and ref
 	 * 2 for dir items
 	 * 1 for xattr if selinux is on
 	 */
-	err = btrfs_reserve_metadata_space(root, 5);
-	if (err)
-		return err;
+	trans = btrfs_start_transaction(root, 5);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans)
-		goto fail;
 	btrfs_set_trans_block_group(trans, dir);
 
-	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
-	if (err) {
-		err = -ENOSPC;
-		goto out_unlock;
-	}
-
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
 				dentry->d_name.len,
 				dentry->d_parent->d_inode->i_ino,
@@ -4368,8 +4710,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
 out_unlock:
 	nr = trans->blocks_used;
 	btrfs_end_transaction_throttle(trans, root);
-fail:
-	btrfs_unreserve_metadata_space(root, 5);
 	if (drop_inode) {
 		inode_dec_link_count(inode);
 		iput(inode);
@@ -4396,21 +4736,21 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 	if (root->objectid != BTRFS_I(inode)->root->objectid)
 		return -EPERM;
 
-	/*
-	 * 1 item for inode ref
-	 * 2 items for dir items
-	 */
-	err = btrfs_reserve_metadata_space(root, 3);
-	if (err)
-		return err;
-
 	btrfs_inc_nlink(inode);
 
 	err = btrfs_set_inode_index(dir, &index);
 	if (err)
 		goto fail;
 
-	trans = btrfs_start_transaction(root, 1);
+	/*
+	 * 1 item for inode ref
+	 * 2 items for dir items
+	 */
+	trans = btrfs_start_transaction(root, 3);
+	if (IS_ERR(trans)) {
+		err = PTR_ERR(trans);
+		goto fail;
+	}
 
 	btrfs_set_trans_block_group(trans, dir);
 	atomic_inc(&inode->i_count);
@@ -4429,7 +4769,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 	nr = trans->blocks_used;
 	btrfs_end_transaction_throttle(trans, root);
 fail:
-	btrfs_unreserve_metadata_space(root, 3);
 	if (drop_inode) {
 		inode_dec_link_count(inode);
 		iput(inode);
@@ -4449,28 +4788,20 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	u64 index = 0;
 	unsigned long nr = 1;
 
+	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+	if (err)
+		return err;
+
 	/*
 	 * 2 items for inode and ref
 	 * 2 items for dir items
 	 * 1 for xattr if selinux is on
 	 */
-	err = btrfs_reserve_metadata_space(root, 5);
-	if (err)
-		return err;
-
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans) {
-		err = -ENOMEM;
-		goto out_unlock;
-	}
+	trans = btrfs_start_transaction(root, 5);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 	btrfs_set_trans_block_group(trans, dir);
 
-	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
-	if (err) {
-		err = -ENOSPC;
-		goto out_fail;
-	}
-
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
 				dentry->d_name.len,
 				dentry->d_parent->d_inode->i_ino, objectid,
@@ -4510,9 +4841,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 out_fail:
 	nr = trans->blocks_used;
 	btrfs_end_transaction_throttle(trans, root);
-
-out_unlock:
-	btrfs_unreserve_metadata_space(root, 5);
 	if (drop_on_err)
 		iput(inode);
 	btrfs_btree_balance_dirty(root, nr);
@@ -4770,6 +5098,7 @@ again:
 			}
 			flush_dcache_page(page);
 		} else if (create && PageUptodate(page)) {
+			WARN_ON(1);
 			if (!trans) {
 				kunmap(page);
 				free_extent_map(em);
@@ -4866,11 +5195,651 @@ out:
 	return em;
 }
 
+static struct extent_map *btrfs_new_extent_direct(struct inode *inode,
+						  u64 start, u64 len)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	struct extent_map *em;
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct btrfs_key ins;
+	u64 alloc_hint;
+	int ret;
+
+	btrfs_drop_extent_cache(inode, start, start + len - 1, 0);
+
+	trans = btrfs_join_transaction(root, 0);
+	if (!trans)
+		return ERR_PTR(-ENOMEM);
+
+	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
+
+	alloc_hint = get_extent_allocation_hint(inode, start, len);
+	ret = btrfs_reserve_extent(trans, root, len, root->sectorsize, 0,
+				   alloc_hint, (u64)-1, &ins, 1);
+	if (ret) {
+		em = ERR_PTR(ret);
+		goto out;
+	}
+
+	em = alloc_extent_map(GFP_NOFS);
+	if (!em) {
+		em = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	em->start = start;
+	em->orig_start = em->start;
+	em->len = ins.offset;
+
+	em->block_start = ins.objectid;
+	em->block_len = ins.offset;
+	em->bdev = root->fs_info->fs_devices->latest_bdev;
+	set_bit(EXTENT_FLAG_PINNED, &em->flags);
+
+	while (1) {
+		write_lock(&em_tree->lock);
+		ret = add_extent_mapping(em_tree, em);
+		write_unlock(&em_tree->lock);
+		if (ret != -EEXIST)
+			break;
+		btrfs_drop_extent_cache(inode, start, start + em->len - 1, 0);
+	}
+
+	ret = btrfs_add_ordered_extent_dio(inode, start, ins.objectid,
+					   ins.offset, ins.offset, 0);
+	if (ret) {
+		btrfs_free_reserved_extent(root, ins.objectid, ins.offset);
+		em = ERR_PTR(ret);
+	}
+out:
+	btrfs_end_transaction(trans, root);
+	return em;
+}
+
+/*
+ * returns 1 when the nocow is safe, < 1 on error, 0 if the
+ * block must be cow'd
+ */
+static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans,
+				      struct inode *inode, u64 offset, u64 len)
+{
+	struct btrfs_path *path;
+	int ret;
+	struct extent_buffer *leaf;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_key key;
+	u64 disk_bytenr;
+	u64 backref_offset;
+	u64 extent_end;
+	u64 num_bytes;
+	int slot;
+	int found_type;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+				       offset, 0);
+	if (ret < 0)
+		goto out;
+
+	slot = path->slots[0];
+	if (ret == 1) {
+		if (slot == 0) {
+			/* can't find the item, must cow */
+			ret = 0;
+			goto out;
+		}
+		slot--;
+	}
+	ret = 0;
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &key, slot);
+	if (key.objectid != inode->i_ino ||
+	    key.type != BTRFS_EXTENT_DATA_KEY) {
+		/* not our file or wrong item type, must cow */
+		goto out;
+	}
+
+	if (key.offset > offset) {
+		/* Wrong offset, must cow */
+		goto out;
+	}
+
+	fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+	found_type = btrfs_file_extent_type(leaf, fi);
+	if (found_type != BTRFS_FILE_EXTENT_REG &&
+	    found_type != BTRFS_FILE_EXTENT_PREALLOC) {
+		/* not a regular extent, must cow */
+		goto out;
+	}
+	disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+	backref_offset = btrfs_file_extent_offset(leaf, fi);
+
+	extent_end = key.offset + btrfs_file_extent_num_bytes(leaf, fi);
+	if (extent_end < offset + len) {
+		/* extent doesn't include our full range, must cow */
+		goto out;
+	}
+
+	if (btrfs_extent_readonly(root, disk_bytenr))
+		goto out;
+
+	/*
+	 * look for other files referencing this extent, if we
+	 * find any we must cow
+	 */
+	if (btrfs_cross_ref_exist(trans, root, inode->i_ino,
+				  key.offset - backref_offset, disk_bytenr))
+		goto out;
+
+	/*
+	 * adjust disk_bytenr and num_bytes to cover just the bytes
+	 * in this extent we are about to write.  If there
+	 * are any csums in that range we have to cow in order
+	 * to keep the csums correct
+	 */
+	disk_bytenr += backref_offset;
+	disk_bytenr += offset - key.offset;
+	num_bytes = min(offset + len, extent_end) - offset;
+	if (csum_exist_in_range(root, disk_bytenr, num_bytes))
+				goto out;
+	/*
+	 * all of the above have passed, it is safe to overwrite this extent
+	 * without cow
+	 */
+	ret = 1;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
+				   struct buffer_head *bh_result, int create)
+{
+	struct extent_map *em;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u64 start = iblock << inode->i_blkbits;
+	u64 len = bh_result->b_size;
+	struct btrfs_trans_handle *trans;
+
+	em = btrfs_get_extent(inode, NULL, 0, start, len, 0);
+	if (IS_ERR(em))
+		return PTR_ERR(em);
+
+	/*
+	 * Ok for INLINE and COMPRESSED extents we need to fallback on buffered
+	 * io.  INLINE is special, and we could probably kludge it in here, but
+	 * it's still buffered so for safety lets just fall back to the generic
+	 * buffered path.
+	 *
+	 * For COMPRESSED we _have_ to read the entire extent in so we can
+	 * decompress it, so there will be buffering required no matter what we
+	 * do, so go ahead and fallback to buffered.
+	 *
+	 * We return -ENOTBLK because thats what makes DIO go ahead and go back
+	 * to buffered IO.  Don't blame me, this is the price we pay for using
+	 * the generic code.
+	 */
+	if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) ||
+	    em->block_start == EXTENT_MAP_INLINE) {
+		free_extent_map(em);
+		return -ENOTBLK;
+	}
+
+	/* Just a good old fashioned hole, return */
+	if (!create && (em->block_start == EXTENT_MAP_HOLE ||
+			test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) {
+		free_extent_map(em);
+		/* DIO will do one hole at a time, so just unlock a sector */
+		unlock_extent(&BTRFS_I(inode)->io_tree, start,
+			      start + root->sectorsize - 1, GFP_NOFS);
+		return 0;
+	}
+
+	/*
+	 * We don't allocate a new extent in the following cases
+	 *
+	 * 1) The inode is marked as NODATACOW.  In this case we'll just use the
+	 * existing extent.
+	 * 2) The extent is marked as PREALLOC.  We're good to go here and can
+	 * just use the extent.
+	 *
+	 */
+	if (!create) {
+		len = em->len - (start - em->start);
+		goto map;
+	}
+
+	if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) ||
+	    ((BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) &&
+	     em->block_start != EXTENT_MAP_HOLE)) {
+		int type;
+		int ret;
+		u64 block_start;
+
+		if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
+			type = BTRFS_ORDERED_PREALLOC;
+		else
+			type = BTRFS_ORDERED_NOCOW;
+		len = min(len, em->len - (start - em->start));
+		block_start = em->block_start + (start - em->start);
+
+		/*
+		 * we're not going to log anything, but we do need
+		 * to make sure the current transaction stays open
+		 * while we look for nocow cross refs
+		 */
+		trans = btrfs_join_transaction(root, 0);
+		if (!trans)
+			goto must_cow;
+
+		if (can_nocow_odirect(trans, inode, start, len) == 1) {
+			ret = btrfs_add_ordered_extent_dio(inode, start,
+					   block_start, len, len, type);
+			btrfs_end_transaction(trans, root);
+			if (ret) {
+				free_extent_map(em);
+				return ret;
+			}
+			goto unlock;
+		}
+		btrfs_end_transaction(trans, root);
+	}
+must_cow:
+	/*
+	 * this will cow the extent, reset the len in case we changed
+	 * it above
+	 */
+	len = bh_result->b_size;
+	free_extent_map(em);
+	em = btrfs_new_extent_direct(inode, start, len);
+	if (IS_ERR(em))
+		return PTR_ERR(em);
+	len = min(len, em->len - (start - em->start));
+unlock:
+	clear_extent_bit(&BTRFS_I(inode)->io_tree, start, start + len - 1,
+			  EXTENT_LOCKED | EXTENT_DELALLOC | EXTENT_DIRTY, 1,
+			  0, NULL, GFP_NOFS);
+map:
+	bh_result->b_blocknr = (em->block_start + (start - em->start)) >>
+		inode->i_blkbits;
+	bh_result->b_size = len;
+	bh_result->b_bdev = em->bdev;
+	set_buffer_mapped(bh_result);
+	if (create && !test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
+		set_buffer_new(bh_result);
+
+	free_extent_map(em);
+
+	return 0;
+}
+
+struct btrfs_dio_private {
+	struct inode *inode;
+	u64 logical_offset;
+	u64 disk_bytenr;
+	u64 bytes;
+	u32 *csums;
+	void *private;
+};
+
+static void btrfs_endio_direct_read(struct bio *bio, int err)
+{
+	struct bio_vec *bvec_end = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct bio_vec *bvec = bio->bi_io_vec;
+	struct btrfs_dio_private *dip = bio->bi_private;
+	struct inode *inode = dip->inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u64 start;
+	u32 *private = dip->csums;
+
+	start = dip->logical_offset;
+	do {
+		if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) {
+			struct page *page = bvec->bv_page;
+			char *kaddr;
+			u32 csum = ~(u32)0;
+			unsigned long flags;
+
+			local_irq_save(flags);
+			kaddr = kmap_atomic(page, KM_IRQ0);
+			csum = btrfs_csum_data(root, kaddr + bvec->bv_offset,
+					       csum, bvec->bv_len);
+			btrfs_csum_final(csum, (char *)&csum);
+			kunmap_atomic(kaddr, KM_IRQ0);
+			local_irq_restore(flags);
+
+			flush_dcache_page(bvec->bv_page);
+			if (csum != *private) {
+				printk(KERN_ERR "btrfs csum failed ino %lu off"
+				      " %llu csum %u private %u\n",
+				      inode->i_ino, (unsigned long long)start,
+				      csum, *private);
+				err = -EIO;
+			}
+		}
+
+		start += bvec->bv_len;
+		private++;
+		bvec++;
+	} while (bvec <= bvec_end);
+
+	unlock_extent(&BTRFS_I(inode)->io_tree, dip->logical_offset,
+		      dip->logical_offset + dip->bytes - 1, GFP_NOFS);
+	bio->bi_private = dip->private;
+
+	kfree(dip->csums);
+	kfree(dip);
+	dio_end_io(bio, err);
+}
+
+static void btrfs_endio_direct_write(struct bio *bio, int err)
+{
+	struct btrfs_dio_private *dip = bio->bi_private;
+	struct inode *inode = dip->inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_ordered_extent *ordered = NULL;
+	struct extent_state *cached_state = NULL;
+	int ret;
+
+	if (err)
+		goto out_done;
+
+	ret = btrfs_dec_test_ordered_pending(inode, &ordered,
+					     dip->logical_offset, dip->bytes);
+	if (!ret)
+		goto out_done;
+
+	BUG_ON(!ordered);
+
+	trans = btrfs_join_transaction(root, 1);
+	if (!trans) {
+		err = -ENOMEM;
+		goto out;
+	}
+	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
+
+	if (test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags)) {
+		ret = btrfs_ordered_update_i_size(inode, 0, ordered);
+		if (!ret)
+			ret = btrfs_update_inode(trans, root, inode);
+		err = ret;
+		goto out;
+	}
+
+	lock_extent_bits(&BTRFS_I(inode)->io_tree, ordered->file_offset,
+			 ordered->file_offset + ordered->len - 1, 0,
+			 &cached_state, GFP_NOFS);
+
+	if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags)) {
+		ret = btrfs_mark_extent_written(trans, inode,
+						ordered->file_offset,
+						ordered->file_offset +
+						ordered->len);
+		if (ret) {
+			err = ret;
+			goto out_unlock;
+		}
+	} else {
+		ret = insert_reserved_file_extent(trans, inode,
+						  ordered->file_offset,
+						  ordered->start,
+						  ordered->disk_len,
+						  ordered->len,
+						  ordered->len,
+						  0, 0, 0,
+						  BTRFS_FILE_EXTENT_REG);
+		unpin_extent_cache(&BTRFS_I(inode)->extent_tree,
+				   ordered->file_offset, ordered->len);
+		if (ret) {
+			err = ret;
+			WARN_ON(1);
+			goto out_unlock;
+		}
+	}
+
+	add_pending_csums(trans, inode, ordered->file_offset, &ordered->list);
+	btrfs_ordered_update_i_size(inode, 0, ordered);
+	btrfs_update_inode(trans, root, inode);
+out_unlock:
+	unlock_extent_cached(&BTRFS_I(inode)->io_tree, ordered->file_offset,
+			     ordered->file_offset + ordered->len - 1,
+			     &cached_state, GFP_NOFS);
+out:
+	btrfs_delalloc_release_metadata(inode, ordered->len);
+	btrfs_end_transaction(trans, root);
+	btrfs_put_ordered_extent(ordered);
+	btrfs_put_ordered_extent(ordered);
+out_done:
+	bio->bi_private = dip->private;
+
+	kfree(dip->csums);
+	kfree(dip);
+	dio_end_io(bio, err);
+}
+
+static int __btrfs_submit_bio_start_direct_io(struct inode *inode, int rw,
+				    struct bio *bio, int mirror_num,
+				    unsigned long bio_flags, u64 offset)
+{
+	int ret;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	ret = btrfs_csum_one_bio(root, inode, bio, offset, 1);
+	BUG_ON(ret);
+	return 0;
+}
+
+static void btrfs_submit_direct(int rw, struct bio *bio, struct inode *inode,
+				loff_t file_offset)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_dio_private *dip;
+	struct bio_vec *bvec = bio->bi_io_vec;
+	u64 start;
+	int skip_sum;
+	int write = rw & (1 << BIO_RW);
+	int ret = 0;
+
+	skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
+
+	dip = kmalloc(sizeof(*dip), GFP_NOFS);
+	if (!dip) {
+		ret = -ENOMEM;
+		goto free_ordered;
+	}
+	dip->csums = NULL;
+
+	if (!skip_sum) {
+		dip->csums = kmalloc(sizeof(u32) * bio->bi_vcnt, GFP_NOFS);
+		if (!dip->csums) {
+			ret = -ENOMEM;
+			goto free_ordered;
+		}
+	}
+
+	dip->private = bio->bi_private;
+	dip->inode = inode;
+	dip->logical_offset = file_offset;
+
+	start = dip->logical_offset;
+	dip->bytes = 0;
+	do {
+		dip->bytes += bvec->bv_len;
+		bvec++;
+	} while (bvec <= (bio->bi_io_vec + bio->bi_vcnt - 1));
+
+	dip->disk_bytenr = (u64)bio->bi_sector << 9;
+	bio->bi_private = dip;
+
+	if (write)
+		bio->bi_end_io = btrfs_endio_direct_write;
+	else
+		bio->bi_end_io = btrfs_endio_direct_read;
+
+	ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
+	if (ret)
+		goto out_err;
+
+	if (write && !skip_sum) {
+		ret = btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
+				   inode, rw, bio, 0, 0,
+				   dip->logical_offset,
+				   __btrfs_submit_bio_start_direct_io,
+				   __btrfs_submit_bio_done);
+		if (ret)
+			goto out_err;
+		return;
+	} else if (!skip_sum)
+		btrfs_lookup_bio_sums_dio(root, inode, bio,
+					  dip->logical_offset, dip->csums);
+
+	ret = btrfs_map_bio(root, rw, bio, 0, 1);
+	if (ret)
+		goto out_err;
+	return;
+out_err:
+	kfree(dip->csums);
+	kfree(dip);
+free_ordered:
+	/*
+	 * If this is a write, we need to clean up the reserved space and kill
+	 * the ordered extent.
+	 */
+	if (write) {
+		struct btrfs_ordered_extent *ordered;
+		ordered = btrfs_lookup_ordered_extent(inode,
+						      dip->logical_offset);
+		if (!test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags) &&
+		    !test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags))
+			btrfs_free_reserved_extent(root, ordered->start,
+						   ordered->disk_len);
+		btrfs_put_ordered_extent(ordered);
+		btrfs_put_ordered_extent(ordered);
+	}
+	bio_endio(bio, ret);
+}
+
+static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *iocb,
+			const struct iovec *iov, loff_t offset,
+			unsigned long nr_segs)
+{
+	int seg;
+	size_t size;
+	unsigned long addr;
+	unsigned blocksize_mask = root->sectorsize - 1;
+	ssize_t retval = -EINVAL;
+	loff_t end = offset;
+
+	if (offset & blocksize_mask)
+		goto out;
+
+	/* Check the memory alignment.  Blocks cannot straddle pages */
+	for (seg = 0; seg < nr_segs; seg++) {
+		addr = (unsigned long)iov[seg].iov_base;
+		size = iov[seg].iov_len;
+		end += size;
+		if ((addr & blocksize_mask) || (size & blocksize_mask)) 
+			goto out;
+	}
+	retval = 0;
+out:
+	return retval;
+}
 static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
 			const struct iovec *iov, loff_t offset,
 			unsigned long nr_segs)
 {
-	return -EINVAL;
+	struct file *file = iocb->ki_filp;
+	struct inode *inode = file->f_mapping->host;
+	struct btrfs_ordered_extent *ordered;
+	struct extent_state *cached_state = NULL;
+	u64 lockstart, lockend;
+	ssize_t ret;
+	int writing = rw & WRITE;
+	int write_bits = 0;
+	size_t count = iov_length(iov, nr_segs);
+
+	if (check_direct_IO(BTRFS_I(inode)->root, rw, iocb, iov,
+			    offset, nr_segs)) {
+		return 0;
+	}
+
+	lockstart = offset;
+	lockend = offset + count - 1;
+
+	if (writing) {
+		ret = btrfs_delalloc_reserve_space(inode, count);
+		if (ret)
+			goto out;
+	}
+
+	while (1) {
+		lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend,
+				 0, &cached_state, GFP_NOFS);
+		/*
+		 * We're concerned with the entire range that we're going to be
+		 * doing DIO to, so we need to make sure theres no ordered
+		 * extents in this range.
+		 */
+		ordered = btrfs_lookup_ordered_range(inode, lockstart,
+						     lockend - lockstart + 1);
+		if (!ordered)
+			break;
+		unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart, lockend,
+				     &cached_state, GFP_NOFS);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		btrfs_put_ordered_extent(ordered);
+		cond_resched();
+	}
+
+	/*
+	 * we don't use btrfs_set_extent_delalloc because we don't want
+	 * the dirty or uptodate bits
+	 */
+	if (writing) {
+		write_bits = EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING;
+		ret = set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend,
+				     EXTENT_DELALLOC, 0, NULL, &cached_state,
+				     GFP_NOFS);
+		if (ret) {
+			clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart,
+					 lockend, EXTENT_LOCKED | write_bits,
+					 1, 0, &cached_state, GFP_NOFS);
+			goto out;
+		}
+	}
+
+	free_extent_state(cached_state);
+	cached_state = NULL;
+
+	ret = __blockdev_direct_IO(rw, iocb, inode,
+		   BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev,
+		   iov, offset, nr_segs, btrfs_get_blocks_direct, NULL,
+		   btrfs_submit_direct, 0);
+
+	if (ret < 0 && ret != -EIOCBQUEUED) {
+		clear_extent_bit(&BTRFS_I(inode)->io_tree, offset,
+			      offset + iov_length(iov, nr_segs) - 1,
+			      EXTENT_LOCKED | write_bits, 1, 0,
+			      &cached_state, GFP_NOFS);
+	} else if (ret >= 0 && ret < iov_length(iov, nr_segs)) {
+		/*
+		 * We're falling back to buffered, unlock the section we didn't
+		 * do IO on.
+		 */
+		clear_extent_bit(&BTRFS_I(inode)->io_tree, offset + ret,
+			      offset + iov_length(iov, nr_segs) - 1,
+			      EXTENT_LOCKED | write_bits, 1, 0,
+			      &cached_state, GFP_NOFS);
+	}
+out:
+	free_extent_state(cached_state);
+	return ret;
 }
 
 static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
@@ -5034,7 +6003,7 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 	u64 page_start;
 	u64 page_end;
 
-	ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
+	ret  = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
 	if (ret) {
 		if (ret == -ENOMEM)
 			ret = VM_FAULT_OOM;
@@ -5043,13 +6012,6 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 		goto out;
 	}
 
-	ret = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
-	if (ret) {
-		btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
-		ret = VM_FAULT_SIGBUS;
-		goto out;
-	}
-
 	ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */
 again:
 	lock_page(page);
@@ -5059,7 +6021,6 @@ again:
 
 	if ((page->mapping != inode->i_mapping) ||
 	    (page_start >= size)) {
-		btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
 		/* page got truncated out from underneath us */
 		goto out_unlock;
 	}
@@ -5100,7 +6061,6 @@ again:
 		unlock_extent_cached(io_tree, page_start, page_end,
 				     &cached_state, GFP_NOFS);
 		ret = VM_FAULT_SIGBUS;
-		btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
 		goto out_unlock;
 	}
 	ret = 0;
@@ -5127,10 +6087,10 @@ again:
 	unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS);
 
 out_unlock:
-	btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
 	if (!ret)
 		return VM_FAULT_LOCKED;
 	unlock_page(page);
+	btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
 out:
 	return ret;
 }
@@ -5155,8 +6115,10 @@ static void btrfs_truncate(struct inode *inode)
 	btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
 	btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_start_transaction(root, 0);
+	BUG_ON(IS_ERR(trans));
 	btrfs_set_trans_block_group(trans, inode);
+	trans->block_rsv = root->orphan_block_rsv;
 
 	/*
 	 * setattr is responsible for setting the ordered_data_close flag,
@@ -5179,6 +6141,23 @@ static void btrfs_truncate(struct inode *inode)
 		btrfs_add_ordered_operation(trans, root, inode);
 
 	while (1) {
+		if (!trans) {
+			trans = btrfs_start_transaction(root, 0);
+			BUG_ON(IS_ERR(trans));
+			btrfs_set_trans_block_group(trans, inode);
+			trans->block_rsv = root->orphan_block_rsv;
+		}
+
+		ret = btrfs_block_rsv_check(trans, root,
+					    root->orphan_block_rsv, 0, 5);
+		if (ret) {
+			BUG_ON(ret != -EAGAIN);
+			ret = btrfs_commit_transaction(trans, root);
+			BUG_ON(ret);
+			trans = NULL;
+			continue;
+		}
+
 		ret = btrfs_truncate_inode_items(trans, root, inode,
 						 inode->i_size,
 						 BTRFS_EXTENT_DATA_KEY);
@@ -5190,10 +6169,8 @@ static void btrfs_truncate(struct inode *inode)
 
 		nr = trans->blocks_used;
 		btrfs_end_transaction(trans, root);
+		trans = NULL;
 		btrfs_btree_balance_dirty(root, nr);
-
-		trans = btrfs_start_transaction(root, 1);
-		btrfs_set_trans_block_group(trans, inode);
 	}
 
 	if (ret == 0 && inode->i_nlink > 0) {
@@ -5254,21 +6231,47 @@ unsigned long btrfs_force_ra(struct address_space *mapping,
 struct inode *btrfs_alloc_inode(struct super_block *sb)
 {
 	struct btrfs_inode *ei;
+	struct inode *inode;
 
 	ei = kmem_cache_alloc(btrfs_inode_cachep, GFP_NOFS);
 	if (!ei)
 		return NULL;
+
+	ei->root = NULL;
+	ei->space_info = NULL;
+	ei->generation = 0;
+	ei->sequence = 0;
 	ei->last_trans = 0;
 	ei->last_sub_trans = 0;
 	ei->logged_trans = 0;
-	ei->outstanding_extents = 0;
-	ei->reserved_extents = 0;
-	ei->root = NULL;
+	ei->delalloc_bytes = 0;
+	ei->reserved_bytes = 0;
+	ei->disk_i_size = 0;
+	ei->flags = 0;
+	ei->index_cnt = (u64)-1;
+	ei->last_unlink_trans = 0;
+
 	spin_lock_init(&ei->accounting_lock);
+	atomic_set(&ei->outstanding_extents, 0);
+	ei->reserved_extents = 0;
+
+	ei->ordered_data_close = 0;
+	ei->orphan_meta_reserved = 0;
+	ei->dummy_inode = 0;
+	ei->force_compress = 0;
+
+	inode = &ei->vfs_inode;
+	extent_map_tree_init(&ei->extent_tree, GFP_NOFS);
+	extent_io_tree_init(&ei->io_tree, &inode->i_data, GFP_NOFS);
+	extent_io_tree_init(&ei->io_failure_tree, &inode->i_data, GFP_NOFS);
+	mutex_init(&ei->log_mutex);
 	btrfs_ordered_inode_tree_init(&ei->ordered_tree);
 	INIT_LIST_HEAD(&ei->i_orphan);
+	INIT_LIST_HEAD(&ei->delalloc_inodes);
 	INIT_LIST_HEAD(&ei->ordered_operations);
-	return &ei->vfs_inode;
+	RB_CLEAR_NODE(&ei->rb_node);
+
+	return inode;
 }
 
 void btrfs_destroy_inode(struct inode *inode)
@@ -5278,6 +6281,8 @@ void btrfs_destroy_inode(struct inode *inode)
 
 	WARN_ON(!list_empty(&inode->i_dentry));
 	WARN_ON(inode->i_data.nrpages);
+	WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents));
+	WARN_ON(BTRFS_I(inode)->reserved_extents);
 
 	/*
 	 * This can happen where we create an inode, but somebody else also
@@ -5298,13 +6303,13 @@ void btrfs_destroy_inode(struct inode *inode)
 		spin_unlock(&root->fs_info->ordered_extent_lock);
 	}
 
-	spin_lock(&root->list_lock);
+	spin_lock(&root->orphan_lock);
 	if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
 		printk(KERN_INFO "BTRFS: inode %lu still on the orphan list\n",
 		       inode->i_ino);
 		list_del_init(&BTRFS_I(inode)->i_orphan);
 	}
-	spin_unlock(&root->list_lock);
+	spin_unlock(&root->orphan_lock);
 
 	while (1) {
 		ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1);
@@ -5425,19 +6430,6 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (S_ISDIR(old_inode->i_mode) && new_inode &&
 	    new_inode->i_size > BTRFS_EMPTY_DIR_SIZE)
 		return -ENOTEMPTY;
-
-	/*
-	 * We want to reserve the absolute worst case amount of items.  So if
-	 * both inodes are subvols and we need to unlink them then that would
-	 * require 4 item modifications, but if they are both normal inodes it
-	 * would require 5 item modifications, so we'll assume their normal
-	 * inodes.  So 5 * 2 is 10, plus 1 for the new link, so 11 total items
-	 * should cover the worst case number of items we'll modify.
-	 */
-	ret = btrfs_reserve_metadata_space(root, 11);
-	if (ret)
-		return ret;
-
 	/*
 	 * we're using rename to replace one file with another.
 	 * and the replacement file is large.  Start IO on it now so
@@ -5450,8 +6442,18 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	/* close the racy window with snapshot create/destroy ioctl */
 	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
 		down_read(&root->fs_info->subvol_sem);
+	/*
+	 * We want to reserve the absolute worst case amount of items.  So if
+	 * both inodes are subvols and we need to unlink them then that would
+	 * require 4 item modifications, but if they are both normal inodes it
+	 * would require 5 item modifications, so we'll assume their normal
+	 * inodes.  So 5 * 2 is 10, plus 1 for the new link, so 11 total items
+	 * should cover the worst case number of items we'll modify.
+	 */
+	trans = btrfs_start_transaction(root, 20);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
-	trans = btrfs_start_transaction(root, 1);
 	btrfs_set_trans_block_group(trans, new_dir);
 
 	if (dest != root)
@@ -5550,7 +6552,6 @@ out_fail:
 	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
 		up_read(&root->fs_info->subvol_sem);
 
-	btrfs_unreserve_metadata_space(root, 11);
 	return ret;
 }
 
@@ -5602,6 +6603,38 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
 	return 0;
 }
 
+int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput)
+{
+	struct btrfs_inode *binode;
+	struct inode *inode = NULL;
+
+	spin_lock(&root->fs_info->delalloc_lock);
+	while (!list_empty(&root->fs_info->delalloc_inodes)) {
+		binode = list_entry(root->fs_info->delalloc_inodes.next,
+				    struct btrfs_inode, delalloc_inodes);
+		inode = igrab(&binode->vfs_inode);
+		if (inode) {
+			list_move_tail(&binode->delalloc_inodes,
+				       &root->fs_info->delalloc_inodes);
+			break;
+		}
+
+		list_del_init(&binode->delalloc_inodes);
+		cond_resched_lock(&root->fs_info->delalloc_lock);
+	}
+	spin_unlock(&root->fs_info->delalloc_lock);
+
+	if (inode) {
+		write_inode_now(inode, 0);
+		if (delay_iput)
+			btrfs_add_delayed_iput(inode);
+		else
+			iput(inode);
+		return 1;
+	}
+	return 0;
+}
+
 static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 			 const char *symname)
 {
@@ -5625,26 +6658,20 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
 		return -ENAMETOOLONG;
 
+	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+	if (err)
+		return err;
 	/*
 	 * 2 items for inode item and ref
 	 * 2 items for dir items
 	 * 1 item for xattr if selinux is on
 	 */
-	err = btrfs_reserve_metadata_space(root, 5);
-	if (err)
-		return err;
+	trans = btrfs_start_transaction(root, 5);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans)
-		goto out_fail;
 	btrfs_set_trans_block_group(trans, dir);
 
-	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
-	if (err) {
-		err = -ENOSPC;
-		goto out_unlock;
-	}
-
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
 				dentry->d_name.len,
 				dentry->d_parent->d_inode->i_ino, objectid,
@@ -5716,8 +6743,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 out_unlock:
 	nr = trans->blocks_used;
 	btrfs_end_transaction_throttle(trans, root);
-out_fail:
-	btrfs_unreserve_metadata_space(root, 5);
 	if (drop_inode) {
 		inode_dec_link_count(inode);
 		iput(inode);
@@ -5726,33 +6751,28 @@ out_fail:
 	return err;
 }
 
-static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
-			u64 alloc_hint, int mode, loff_t actual_len)
+int btrfs_prealloc_file_range(struct inode *inode, int mode,
+			      u64 start, u64 num_bytes, u64 min_size,
+			      loff_t actual_len, u64 *alloc_hint)
 {
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_key ins;
 	u64 cur_offset = start;
-	u64 num_bytes = end - start;
 	int ret = 0;
-	u64 i_size;
 
 	while (num_bytes > 0) {
-		trans = btrfs_start_transaction(root, 1);
-
-		ret = btrfs_reserve_extent(trans, root, num_bytes,
-					   root->sectorsize, 0, alloc_hint,
-					   (u64)-1, &ins, 1);
-		if (ret) {
-			WARN_ON(1);
-			goto stop_trans;
+		trans = btrfs_start_transaction(root, 3);
+		if (IS_ERR(trans)) {
+			ret = PTR_ERR(trans);
+			break;
 		}
 
-		ret = btrfs_reserve_metadata_space(root, 3);
+		ret = btrfs_reserve_extent(trans, root, num_bytes, min_size,
+					   0, *alloc_hint, (u64)-1, &ins, 1);
 		if (ret) {
-			btrfs_free_reserved_extent(root, ins.objectid,
-						   ins.offset);
-			goto stop_trans;
+			btrfs_end_transaction(trans, root);
+			break;
 		}
 
 		ret = insert_reserved_file_extent(trans, inode,
@@ -5766,34 +6786,27 @@ static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
 
 		num_bytes -= ins.offset;
 		cur_offset += ins.offset;
-		alloc_hint = ins.objectid + ins.offset;
+		*alloc_hint = ins.objectid + ins.offset;
 
 		inode->i_ctime = CURRENT_TIME;
 		BTRFS_I(inode)->flags |= BTRFS_INODE_PREALLOC;
 		if (!(mode & FALLOC_FL_KEEP_SIZE) &&
-			(actual_len > inode->i_size) &&
-			(cur_offset > inode->i_size)) {
-
+		    (actual_len > inode->i_size) &&
+		    (cur_offset > inode->i_size)) {
 			if (cur_offset > actual_len)
-				i_size  = actual_len;
+				i_size_write(inode, actual_len);
 			else
-				i_size = cur_offset;
-			i_size_write(inode, i_size);
-			btrfs_ordered_update_i_size(inode, i_size, NULL);
+				i_size_write(inode, cur_offset);
+			i_size_write(inode, cur_offset);
+			btrfs_ordered_update_i_size(inode, cur_offset, NULL);
 		}
 
 		ret = btrfs_update_inode(trans, root, inode);
 		BUG_ON(ret);
 
 		btrfs_end_transaction(trans, root);
-		btrfs_unreserve_metadata_space(root, 3);
 	}
 	return ret;
-
-stop_trans:
-	btrfs_end_transaction(trans, root);
-	return ret;
-
 }
 
 static long btrfs_fallocate(struct inode *inode, int mode,
@@ -5826,8 +6839,7 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 			goto out;
 	}
 
-	ret = btrfs_check_data_free_space(BTRFS_I(inode)->root, inode,
-					  alloc_end - alloc_start);
+	ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start);
 	if (ret)
 		goto out;
 
@@ -5872,16 +6884,16 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 		if (em->block_start == EXTENT_MAP_HOLE ||
 		    (cur_offset >= inode->i_size &&
 		     !test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) {
-			ret = prealloc_file_range(inode,
-						  cur_offset, last_byte,
-						alloc_hint, mode, offset+len);
+			ret = btrfs_prealloc_file_range(inode, 0, cur_offset,
+							last_byte - cur_offset,
+							1 << inode->i_blkbits,
+							offset + len,
+							&alloc_hint);
 			if (ret < 0) {
 				free_extent_map(em);
 				break;
 			}
 		}
-		if (em->block_start <= EXTENT_MAP_LAST_BYTE)
-			alloc_hint = em->block_start;
 		free_extent_map(em);
 
 		cur_offset = last_byte;
@@ -5893,8 +6905,7 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
 			     &cached_state, GFP_NOFS);
 
-	btrfs_free_reserved_data_space(BTRFS_I(inode)->root, inode,
-				       alloc_end - alloc_start);
+	btrfs_free_reserved_data_space(inode, alloc_end - alloc_start);
 out:
 	mutex_unlock(&inode->i_mutex);
 	return ret;
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 97a97839a867..4cdb98cf26de 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -239,23 +239,19 @@ static noinline int create_subvol(struct btrfs_root *root,
 	u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID;
 	u64 index = 0;
 
+	ret = btrfs_find_free_objectid(NULL, root->fs_info->tree_root,
+				       0, &objectid);
+	if (ret)
+		return ret;
 	/*
 	 * 1 - inode item
 	 * 2 - refs
 	 * 1 - root item
 	 * 2 - dir items
 	 */
-	ret = btrfs_reserve_metadata_space(root, 6);
-	if (ret)
-		return ret;
-
-	trans = btrfs_start_transaction(root, 1);
-	BUG_ON(!trans);
-
-	ret = btrfs_find_free_objectid(trans, root->fs_info->tree_root,
-				       0, &objectid);
-	if (ret)
-		goto fail;
+	trans = btrfs_start_transaction(root, 6);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
 	leaf = btrfs_alloc_free_block(trans, root, root->leafsize,
 				      0, objectid, NULL, 0, 0, 0);
@@ -345,13 +341,10 @@ fail:
 	err = btrfs_commit_transaction(trans, root);
 	if (err && !ret)
 		ret = err;
-
-	btrfs_unreserve_metadata_space(root, 6);
 	return ret;
 }
 
-static int create_snapshot(struct btrfs_root *root, struct dentry *dentry,
-			   char *name, int namelen)
+static int create_snapshot(struct btrfs_root *root, struct dentry *dentry)
 {
 	struct inode *inode;
 	struct btrfs_pending_snapshot *pending_snapshot;
@@ -361,40 +354,33 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry,
 	if (!root->ref_cows)
 		return -EINVAL;
 
-	/*
-	 * 1 - inode item
-	 * 2 - refs
-	 * 1 - root item
-	 * 2 - dir items
-	 */
-	ret = btrfs_reserve_metadata_space(root, 6);
-	if (ret)
-		goto fail;
-
 	pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_NOFS);
-	if (!pending_snapshot) {
-		ret = -ENOMEM;
-		btrfs_unreserve_metadata_space(root, 6);
-		goto fail;
-	}
-	pending_snapshot->name = kmalloc(namelen + 1, GFP_NOFS);
-	if (!pending_snapshot->name) {
-		ret = -ENOMEM;
-		kfree(pending_snapshot);
-		btrfs_unreserve_metadata_space(root, 6);
-		goto fail;
-	}
-	memcpy(pending_snapshot->name, name, namelen);
-	pending_snapshot->name[namelen] = '\0';
+	if (!pending_snapshot)
+		return -ENOMEM;
+
+	btrfs_init_block_rsv(&pending_snapshot->block_rsv);
 	pending_snapshot->dentry = dentry;
-	trans = btrfs_start_transaction(root, 1);
-	BUG_ON(!trans);
 	pending_snapshot->root = root;
+
+	trans = btrfs_start_transaction(root->fs_info->extent_root, 5);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		goto fail;
+	}
+
+	ret = btrfs_snap_reserve_metadata(trans, pending_snapshot);
+	BUG_ON(ret);
+
 	list_add(&pending_snapshot->list,
 		 &trans->transaction->pending_snapshots);
-	ret = btrfs_commit_transaction(trans, root);
+	ret = btrfs_commit_transaction(trans, root->fs_info->extent_root);
 	BUG_ON(ret);
-	btrfs_unreserve_metadata_space(root, 6);
+
+	ret = pending_snapshot->error;
+	if (ret)
+		goto fail;
+
+	btrfs_orphan_cleanup(pending_snapshot->snap);
 
 	inode = btrfs_lookup_dentry(dentry->d_parent->d_inode, dentry);
 	if (IS_ERR(inode)) {
@@ -405,6 +391,7 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry,
 	d_instantiate(dentry, inode);
 	ret = 0;
 fail:
+	kfree(pending_snapshot);
 	return ret;
 }
 
@@ -456,8 +443,7 @@ static noinline int btrfs_mksubvol(struct path *parent,
 		goto out_up_read;
 
 	if (snap_src) {
-		error = create_snapshot(snap_src, dentry,
-					name, namelen);
+		error = create_snapshot(snap_src, dentry);
 	} else {
 		error = create_subvol(BTRFS_I(dir)->root, dentry,
 				      name, namelen);
@@ -601,19 +587,9 @@ static int btrfs_defrag_file(struct file *file,
 		if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)
 			BTRFS_I(inode)->force_compress = 1;
 
-		ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
-		if (ret) {
-			ret = -ENOSPC;
-			break;
-		}
-
-		ret = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
-		if (ret) {
-			btrfs_free_reserved_data_space(root, inode,
-						       PAGE_CACHE_SIZE);
-			ret = -ENOSPC;
-			break;
-		}
+		ret  = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
+		if (ret)
+			goto err_unlock;
 again:
 		if (inode->i_size == 0 ||
 		    i > ((inode->i_size - 1) >> PAGE_CACHE_SHIFT)) {
@@ -622,8 +598,10 @@ again:
 		}
 
 		page = grab_cache_page(inode->i_mapping, i);
-		if (!page)
+		if (!page) {
+			ret = -ENOMEM;
 			goto err_reservations;
+		}
 
 		if (!PageUptodate(page)) {
 			btrfs_readpage(NULL, page);
@@ -631,6 +609,7 @@ again:
 			if (!PageUptodate(page)) {
 				unlock_page(page);
 				page_cache_release(page);
+				ret = -EIO;
 				goto err_reservations;
 			}
 		}
@@ -644,8 +623,7 @@ again:
 		wait_on_page_writeback(page);
 
 		if (PageDirty(page)) {
-			btrfs_free_reserved_data_space(root, inode,
-						       PAGE_CACHE_SIZE);
+			btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
 			goto loop_unlock;
 		}
 
@@ -683,7 +661,6 @@ loop_unlock:
 		page_cache_release(page);
 		mutex_unlock(&inode->i_mutex);
 
-		btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
 		balance_dirty_pages_ratelimited_nr(inode->i_mapping, 1);
 		i++;
 	}
@@ -713,9 +690,9 @@ loop_unlock:
 	return 0;
 
 err_reservations:
+	btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
+err_unlock:
 	mutex_unlock(&inode->i_mutex);
-	btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
-	btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
 	return ret;
 }
 
@@ -811,7 +788,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root,
 		device->name, (unsigned long long)new_size);
 
 	if (new_size > old_size) {
-		trans = btrfs_start_transaction(root, 1);
+		trans = btrfs_start_transaction(root, 0);
 		ret = btrfs_grow_device(trans, device, new_size);
 		btrfs_commit_transaction(trans, root);
 	} else {
@@ -1300,7 +1277,13 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 	if (err)
 		goto out_up_write;
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_start_transaction(root, 0);
+	if (IS_ERR(trans)) {
+		err = PTR_ERR(trans);
+		goto out;
+	}
+	trans->block_rsv = &root->fs_info->global_block_rsv;
+
 	ret = btrfs_unlink_subvol(trans, root, dir,
 				dest->root_key.objectid,
 				dentry->d_name.name,
@@ -1314,10 +1297,12 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 	dest->root_item.drop_level = 0;
 	btrfs_set_root_refs(&dest->root_item, 0);
 
-	ret = btrfs_insert_orphan_item(trans,
-				root->fs_info->tree_root,
-				dest->root_key.objectid);
-	BUG_ON(ret);
+	if (!xchg(&dest->orphan_item_inserted, 1)) {
+		ret = btrfs_insert_orphan_item(trans,
+					root->fs_info->tree_root,
+					dest->root_key.objectid);
+		BUG_ON(ret);
+	}
 
 	ret = btrfs_commit_transaction(trans, root);
 	BUG_ON(ret);
@@ -1358,8 +1343,10 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
 			ret = -EPERM;
 			goto out;
 		}
-		btrfs_defrag_root(root, 0);
-		btrfs_defrag_root(root->fs_info->extent_root, 0);
+		ret = btrfs_defrag_root(root, 0);
+		if (ret)
+			goto out;
+		ret = btrfs_defrag_root(root->fs_info->extent_root, 0);
 		break;
 	case S_IFREG:
 		if (!(file->f_mode & FMODE_WRITE)) {
@@ -1389,9 +1376,11 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
 			/* the rest are all set to zero by kzalloc */
 			range->len = (u64)-1;
 		}
-		btrfs_defrag_file(file, range);
+		ret = btrfs_defrag_file(file, range);
 		kfree(range);
 		break;
+	default:
+		ret = -EINVAL;
 	}
 out:
 	mnt_drop_write(file->f_path.mnt);
@@ -1550,12 +1539,6 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 		btrfs_wait_ordered_range(src, off, off+len);
 	}
 
-	trans = btrfs_start_transaction(root, 1);
-	BUG_ON(!trans);
-
-	/* punch hole in destination first */
-	btrfs_drop_extents(trans, inode, off, off + len, &hint_byte, 1);
-
 	/* clone data */
 	key.objectid = src->i_ino;
 	key.type = BTRFS_EXTENT_DATA_KEY;
@@ -1566,7 +1549,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 		 * note the key will change type as we walk through the
 		 * tree.
 		 */
-		ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 		if (ret < 0)
 			goto out;
 
@@ -1629,12 +1612,31 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 			new_key.objectid = inode->i_ino;
 			new_key.offset = key.offset + destoff - off;
 
+			trans = btrfs_start_transaction(root, 1);
+			if (IS_ERR(trans)) {
+				ret = PTR_ERR(trans);
+				goto out;
+			}
+
 			if (type == BTRFS_FILE_EXTENT_REG ||
 			    type == BTRFS_FILE_EXTENT_PREALLOC) {
+				if (off > key.offset) {
+					datao += off - key.offset;
+					datal -= off - key.offset;
+				}
+
+				if (key.offset + datal > off + len)
+					datal = off + len - key.offset;
+
+				ret = btrfs_drop_extents(trans, inode,
+							 new_key.offset,
+							 new_key.offset + datal,
+							 &hint_byte, 1);
+				BUG_ON(ret);
+
 				ret = btrfs_insert_empty_item(trans, root, path,
 							      &new_key, size);
-				if (ret)
-					goto out;
+				BUG_ON(ret);
 
 				leaf = path->nodes[0];
 				slot = path->slots[0];
@@ -1645,14 +1647,6 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 				extent = btrfs_item_ptr(leaf, slot,
 						struct btrfs_file_extent_item);
 
-				if (off > key.offset) {
-					datao += off - key.offset;
-					datal -= off - key.offset;
-				}
-
-				if (key.offset + datal > off + len)
-					datal = off + len - key.offset;
-
 				/* disko == 0 means it's a hole */
 				if (!disko)
 					datao = 0;
@@ -1683,14 +1677,21 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 
 				if (comp && (skip || trim)) {
 					ret = -EINVAL;
+					btrfs_end_transaction(trans, root);
 					goto out;
 				}
 				size -= skip + trim;
 				datal -= skip + trim;
+
+				ret = btrfs_drop_extents(trans, inode,
+							 new_key.offset,
+							 new_key.offset + datal,
+							 &hint_byte, 1);
+				BUG_ON(ret);
+
 				ret = btrfs_insert_empty_item(trans, root, path,
 							      &new_key, size);
-				if (ret)
-					goto out;
+				BUG_ON(ret);
 
 				if (skip) {
 					u32 start =
@@ -1708,8 +1709,17 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 			}
 
 			btrfs_mark_buffer_dirty(leaf);
-		}
+			btrfs_release_path(root, path);
 
+			inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+			if (new_key.offset + datal > inode->i_size)
+				btrfs_i_size_write(inode,
+						   new_key.offset + datal);
+			BTRFS_I(inode)->flags = BTRFS_I(src)->flags;
+			ret = btrfs_update_inode(trans, root, inode);
+			BUG_ON(ret);
+			btrfs_end_transaction(trans, root);
+		}
 next:
 		btrfs_release_path(root, path);
 		key.offset++;
@@ -1717,17 +1727,7 @@ next:
 	ret = 0;
 out:
 	btrfs_release_path(root, path);
-	if (ret == 0) {
-		inode->i_mtime = inode->i_ctime = CURRENT_TIME;
-		if (destoff + olen > inode->i_size)
-			btrfs_i_size_write(inode, destoff + olen);
-		BTRFS_I(inode)->flags = BTRFS_I(src)->flags;
-		ret = btrfs_update_inode(trans, root, inode);
-	}
-	btrfs_end_transaction(trans, root);
 	unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
-	if (ret)
-		vmtruncate(inode, 0);
 out_unlock:
 	mutex_unlock(&src->i_mutex);
 	mutex_unlock(&inode->i_mutex);
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index a127c0ebb2dc..e56c72bc5add 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -124,6 +124,15 @@ static int offset_in_entry(struct btrfs_ordered_extent *entry, u64 file_offset)
 	return 1;
 }
 
+static int range_overlaps(struct btrfs_ordered_extent *entry, u64 file_offset,
+			  u64 len)
+{
+	if (file_offset + len <= entry->file_offset ||
+	    entry->file_offset + entry->len <= file_offset)
+		return 0;
+	return 1;
+}
+
 /*
  * look find the first ordered struct that has this offset, otherwise
  * the first one less than this offset
@@ -161,8 +170,9 @@ static inline struct rb_node *tree_search(struct btrfs_ordered_inode_tree *tree,
  * The tree is given a single reference on the ordered extent that was
  * inserted.
  */
-int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
-			     u64 start, u64 len, u64 disk_len, int type)
+static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
+				      u64 start, u64 len, u64 disk_len,
+				      int type, int dio)
 {
 	struct btrfs_ordered_inode_tree *tree;
 	struct rb_node *node;
@@ -182,6 +192,9 @@ int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 	if (type != BTRFS_ORDERED_IO_DONE && type != BTRFS_ORDERED_COMPLETE)
 		set_bit(type, &entry->flags);
 
+	if (dio)
+		set_bit(BTRFS_ORDERED_DIRECT, &entry->flags);
+
 	/* one ref for the tree */
 	atomic_set(&entry->refs, 1);
 	init_waitqueue_head(&entry->wait);
@@ -203,6 +216,20 @@ int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 	return 0;
 }
 
+int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
+			     u64 start, u64 len, u64 disk_len, int type)
+{
+	return __btrfs_add_ordered_extent(inode, file_offset, start, len,
+					  disk_len, type, 0);
+}
+
+int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset,
+				 u64 start, u64 len, u64 disk_len, int type)
+{
+	return __btrfs_add_ordered_extent(inode, file_offset, start, len,
+					  disk_len, type, 1);
+}
+
 /*
  * Add a struct btrfs_ordered_sum into the list of checksums to be inserted
  * when an ordered extent is finished.  If the list covers more than one
@@ -311,13 +338,6 @@ static int __btrfs_remove_ordered_extent(struct inode *inode,
 	tree->last = NULL;
 	set_bit(BTRFS_ORDERED_COMPLETE, &entry->flags);
 
-	spin_lock(&BTRFS_I(inode)->accounting_lock);
-	WARN_ON(!BTRFS_I(inode)->outstanding_extents);
-	BTRFS_I(inode)->outstanding_extents--;
-	spin_unlock(&BTRFS_I(inode)->accounting_lock);
-	btrfs_unreserve_metadata_for_delalloc(BTRFS_I(inode)->root,
-					      inode, 1);
-
 	spin_lock(&root->fs_info->ordered_extent_lock);
 	list_del_init(&entry->root_extent_list);
 
@@ -491,7 +511,8 @@ void btrfs_start_ordered_extent(struct inode *inode,
 	 * start IO on any dirty ones so the wait doesn't stall waiting
 	 * for pdflush to find them
 	 */
-	filemap_fdatawrite_range(inode->i_mapping, start, end);
+	if (!test_bit(BTRFS_ORDERED_DIRECT, &entry->flags))
+		filemap_fdatawrite_range(inode->i_mapping, start, end);
 	if (wait) {
 		wait_event(entry->wait, test_bit(BTRFS_ORDERED_COMPLETE,
 						 &entry->flags));
@@ -588,6 +609,47 @@ out:
 	return entry;
 }
 
+/* Since the DIO code tries to lock a wide area we need to look for any ordered
+ * extents that exist in the range, rather than just the start of the range.
+ */
+struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
+							u64 file_offset,
+							u64 len)
+{
+	struct btrfs_ordered_inode_tree *tree;
+	struct rb_node *node;
+	struct btrfs_ordered_extent *entry = NULL;
+
+	tree = &BTRFS_I(inode)->ordered_tree;
+	spin_lock(&tree->lock);
+	node = tree_search(tree, file_offset);
+	if (!node) {
+		node = tree_search(tree, file_offset + len);
+		if (!node)
+			goto out;
+	}
+
+	while (1) {
+		entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
+		if (range_overlaps(entry, file_offset, len))
+			break;
+
+		if (entry->file_offset >= file_offset + len) {
+			entry = NULL;
+			break;
+		}
+		entry = NULL;
+		node = rb_next(node);
+		if (!node)
+			break;
+	}
+out:
+	if (entry)
+		atomic_inc(&entry->refs);
+	spin_unlock(&tree->lock);
+	return entry;
+}
+
 /*
  * lookup and return any extent before 'file_offset'.  NULL is returned
  * if none is found
diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h
index c82f76a9f040..8ac365492a3f 100644
--- a/fs/btrfs/ordered-data.h
+++ b/fs/btrfs/ordered-data.h
@@ -72,6 +72,8 @@ struct btrfs_ordered_sum {
 
 #define BTRFS_ORDERED_PREALLOC 4 /* set when writing to prealloced extent */
 
+#define BTRFS_ORDERED_DIRECT 5 /* set when we're doing DIO with this extent */
+
 struct btrfs_ordered_extent {
 	/* logical offset in the file */
 	u64 file_offset;
@@ -140,7 +142,9 @@ int btrfs_dec_test_ordered_pending(struct inode *inode,
 				   struct btrfs_ordered_extent **cached,
 				   u64 file_offset, u64 io_size);
 int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
-			     u64 start, u64 len, u64 disk_len, int tyep);
+			     u64 start, u64 len, u64 disk_len, int type);
+int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset,
+				 u64 start, u64 len, u64 disk_len, int type);
 int btrfs_add_ordered_sum(struct inode *inode,
 			  struct btrfs_ordered_extent *entry,
 			  struct btrfs_ordered_sum *sum);
@@ -151,6 +155,9 @@ void btrfs_start_ordered_extent(struct inode *inode,
 int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len);
 struct btrfs_ordered_extent *
 btrfs_lookup_first_ordered_extent(struct inode * inode, u64 file_offset);
+struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
+							u64 file_offset,
+							u64 len);
 int btrfs_ordered_update_i_size(struct inode *inode, u64 offset,
 				struct btrfs_ordered_extent *ordered);
 int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, u32 *sum);
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index e558dd941ded..05d41e569236 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -44,8 +44,12 @@ struct tree_entry {
 struct backref_node {
 	struct rb_node rb_node;
 	u64 bytenr;
-	/* objectid tree block owner */
+
+	u64 new_bytenr;
+	/* objectid of tree block owner, can be not uptodate */
 	u64 owner;
+	/* link to pending, changed or detached list */
+	struct list_head list;
 	/* list of upper level blocks reference this block */
 	struct list_head upper;
 	/* list of child blocks in the cache */
@@ -56,9 +60,9 @@ struct backref_node {
 	struct extent_buffer *eb;
 	/* level of tree block */
 	unsigned int level:8;
-	/* 1 if the block is root of old snapshot */
-	unsigned int old_root:1;
-	/* 1 if no child blocks in the cache */
+	/* is the block in non-reference counted tree */
+	unsigned int cowonly:1;
+	/* 1 if no child node in the cache */
 	unsigned int lowest:1;
 	/* is the extent buffer locked */
 	unsigned int locked:1;
@@ -66,6 +70,16 @@ struct backref_node {
 	unsigned int processed:1;
 	/* have backrefs of this block been checked */
 	unsigned int checked:1;
+	/*
+	 * 1 if corresponding block has been cowed but some upper
+	 * level block pointers may not point to the new location
+	 */
+	unsigned int pending:1;
+	/*
+	 * 1 if the backref node isn't connected to any other
+	 * backref node.
+	 */
+	unsigned int detached:1;
 };
 
 /*
@@ -74,7 +88,6 @@ struct backref_node {
 struct backref_edge {
 	struct list_head list[2];
 	struct backref_node *node[2];
-	u64 blockptr;
 };
 
 #define LOWER	0
@@ -83,9 +96,25 @@ struct backref_edge {
 struct backref_cache {
 	/* red black tree of all backref nodes in the cache */
 	struct rb_root rb_root;
-	/* list of backref nodes with no child block in the cache */
+	/* for passing backref nodes to btrfs_reloc_cow_block */
+	struct backref_node *path[BTRFS_MAX_LEVEL];
+	/*
+	 * list of blocks that have been cowed but some block
+	 * pointers in upper level blocks may not reflect the
+	 * new location
+	 */
 	struct list_head pending[BTRFS_MAX_LEVEL];
-	spinlock_t lock;
+	/* list of backref nodes with no child node */
+	struct list_head leaves;
+	/* list of blocks that have been cowed in current transaction */
+	struct list_head changed;
+	/* list of detached backref node. */
+	struct list_head detached;
+
+	u64 last_trans;
+
+	int nr_nodes;
+	int nr_edges;
 };
 
 /*
@@ -113,15 +142,6 @@ struct tree_block {
 	unsigned int key_ready:1;
 };
 
-/* inode vector */
-#define INODEVEC_SIZE 16
-
-struct inodevec {
-	struct list_head list;
-	struct inode *inode[INODEVEC_SIZE];
-	int nr;
-};
-
 #define MAX_EXTENTS 128
 
 struct file_extent_cluster {
@@ -138,36 +158,43 @@ struct reloc_control {
 	struct btrfs_root *extent_root;
 	/* inode for moving data */
 	struct inode *data_inode;
-	struct btrfs_workers workers;
+
+	struct btrfs_block_rsv *block_rsv;
+
+	struct backref_cache backref_cache;
+
+	struct file_extent_cluster cluster;
 	/* tree blocks have been processed */
 	struct extent_io_tree processed_blocks;
 	/* map start of tree root to corresponding reloc tree */
 	struct mapping_tree reloc_root_tree;
 	/* list of reloc trees */
 	struct list_head reloc_roots;
+	/* size of metadata reservation for merging reloc trees */
+	u64 merging_rsv_size;
+	/* size of relocated tree nodes */
+	u64 nodes_relocated;
+
 	u64 search_start;
 	u64 extents_found;
-	u64 extents_skipped;
-	int stage;
-	int create_reloc_root;
+
+	int block_rsv_retries;
+
+	unsigned int stage:8;
+	unsigned int create_reloc_tree:1;
+	unsigned int merge_reloc_tree:1;
 	unsigned int found_file_extent:1;
-	unsigned int found_old_snapshot:1;
+	unsigned int commit_transaction:1;
 };
 
 /* stages of data relocation */
 #define MOVE_DATA_EXTENTS	0
 #define UPDATE_DATA_PTRS	1
 
-/*
- * merge reloc tree to corresponding fs tree in worker threads
- */
-struct async_merge {
-	struct btrfs_work work;
-	struct reloc_control *rc;
-	struct btrfs_root *root;
-	struct completion *done;
-	atomic_t *num_pending;
-};
+static void remove_backref_node(struct backref_cache *cache,
+				struct backref_node *node);
+static void __mark_block_processed(struct reloc_control *rc,
+				   struct backref_node *node);
 
 static void mapping_tree_init(struct mapping_tree *tree)
 {
@@ -181,15 +208,80 @@ static void backref_cache_init(struct backref_cache *cache)
 	cache->rb_root = RB_ROOT;
 	for (i = 0; i < BTRFS_MAX_LEVEL; i++)
 		INIT_LIST_HEAD(&cache->pending[i]);
-	spin_lock_init(&cache->lock);
+	INIT_LIST_HEAD(&cache->changed);
+	INIT_LIST_HEAD(&cache->detached);
+	INIT_LIST_HEAD(&cache->leaves);
+}
+
+static void backref_cache_cleanup(struct backref_cache *cache)
+{
+	struct backref_node *node;
+	int i;
+
+	while (!list_empty(&cache->detached)) {
+		node = list_entry(cache->detached.next,
+				  struct backref_node, list);
+		remove_backref_node(cache, node);
+	}
+
+	while (!list_empty(&cache->leaves)) {
+		node = list_entry(cache->leaves.next,
+				  struct backref_node, lower);
+		remove_backref_node(cache, node);
+	}
+
+	cache->last_trans = 0;
+
+	for (i = 0; i < BTRFS_MAX_LEVEL; i++)
+		BUG_ON(!list_empty(&cache->pending[i]));
+	BUG_ON(!list_empty(&cache->changed));
+	BUG_ON(!list_empty(&cache->detached));
+	BUG_ON(!RB_EMPTY_ROOT(&cache->rb_root));
+	BUG_ON(cache->nr_nodes);
+	BUG_ON(cache->nr_edges);
+}
+
+static struct backref_node *alloc_backref_node(struct backref_cache *cache)
+{
+	struct backref_node *node;
+
+	node = kzalloc(sizeof(*node), GFP_NOFS);
+	if (node) {
+		INIT_LIST_HEAD(&node->list);
+		INIT_LIST_HEAD(&node->upper);
+		INIT_LIST_HEAD(&node->lower);
+		RB_CLEAR_NODE(&node->rb_node);
+		cache->nr_nodes++;
+	}
+	return node;
+}
+
+static void free_backref_node(struct backref_cache *cache,
+			      struct backref_node *node)
+{
+	if (node) {
+		cache->nr_nodes--;
+		kfree(node);
+	}
+}
+
+static struct backref_edge *alloc_backref_edge(struct backref_cache *cache)
+{
+	struct backref_edge *edge;
+
+	edge = kzalloc(sizeof(*edge), GFP_NOFS);
+	if (edge)
+		cache->nr_edges++;
+	return edge;
 }
 
-static void backref_node_init(struct backref_node *node)
+static void free_backref_edge(struct backref_cache *cache,
+			      struct backref_edge *edge)
 {
-	memset(node, 0, sizeof(*node));
-	INIT_LIST_HEAD(&node->upper);
-	INIT_LIST_HEAD(&node->lower);
-	RB_CLEAR_NODE(&node->rb_node);
+	if (edge) {
+		cache->nr_edges--;
+		kfree(edge);
+	}
 }
 
 static struct rb_node *tree_insert(struct rb_root *root, u64 bytenr,
@@ -250,6 +342,7 @@ static struct backref_node *walk_up_backref(struct backref_node *node,
 		edges[idx++] = edge;
 		node = edge->node[UPPER];
 	}
+	BUG_ON(node->detached);
 	*index = idx;
 	return node;
 }
@@ -281,13 +374,18 @@ static struct backref_node *walk_down_backref(struct backref_edge *edges[],
 	return NULL;
 }
 
+static void unlock_node_buffer(struct backref_node *node)
+{
+	if (node->locked) {
+		btrfs_tree_unlock(node->eb);
+		node->locked = 0;
+	}
+}
+
 static void drop_node_buffer(struct backref_node *node)
 {
 	if (node->eb) {
-		if (node->locked) {
-			btrfs_tree_unlock(node->eb);
-			node->locked = 0;
-		}
+		unlock_node_buffer(node);
 		free_extent_buffer(node->eb);
 		node->eb = NULL;
 	}
@@ -296,14 +394,14 @@ static void drop_node_buffer(struct backref_node *node)
 static void drop_backref_node(struct backref_cache *tree,
 			      struct backref_node *node)
 {
-	BUG_ON(!node->lowest);
 	BUG_ON(!list_empty(&node->upper));
 
 	drop_node_buffer(node);
+	list_del(&node->list);
 	list_del(&node->lower);
-
-	rb_erase(&node->rb_node, &tree->rb_root);
-	kfree(node);
+	if (!RB_EMPTY_NODE(&node->rb_node))
+		rb_erase(&node->rb_node, &tree->rb_root);
+	free_backref_node(tree, node);
 }
 
 /*
@@ -318,27 +416,121 @@ static void remove_backref_node(struct backref_cache *cache,
 	if (!node)
 		return;
 
-	BUG_ON(!node->lowest);
+	BUG_ON(!node->lowest && !node->detached);
 	while (!list_empty(&node->upper)) {
 		edge = list_entry(node->upper.next, struct backref_edge,
 				  list[LOWER]);
 		upper = edge->node[UPPER];
 		list_del(&edge->list[LOWER]);
 		list_del(&edge->list[UPPER]);
-		kfree(edge);
+		free_backref_edge(cache, edge);
+
+		if (RB_EMPTY_NODE(&upper->rb_node)) {
+			BUG_ON(!list_empty(&node->upper));
+			drop_backref_node(cache, node);
+			node = upper;
+			node->lowest = 1;
+			continue;
+		}
 		/*
-		 * add the node to pending list if no other
+		 * add the node to leaf node list if no other
 		 * child block cached.
 		 */
 		if (list_empty(&upper->lower)) {
-			list_add_tail(&upper->lower,
-				      &cache->pending[upper->level]);
+			list_add_tail(&upper->lower, &cache->leaves);
 			upper->lowest = 1;
 		}
 	}
+
 	drop_backref_node(cache, node);
 }
 
+static void update_backref_node(struct backref_cache *cache,
+				struct backref_node *node, u64 bytenr)
+{
+	struct rb_node *rb_node;
+	rb_erase(&node->rb_node, &cache->rb_root);
+	node->bytenr = bytenr;
+	rb_node = tree_insert(&cache->rb_root, node->bytenr, &node->rb_node);
+	BUG_ON(rb_node);
+}
+
+/*
+ * update backref cache after a transaction commit
+ */
+static int update_backref_cache(struct btrfs_trans_handle *trans,
+				struct backref_cache *cache)
+{
+	struct backref_node *node;
+	int level = 0;
+
+	if (cache->last_trans == 0) {
+		cache->last_trans = trans->transid;
+		return 0;
+	}
+
+	if (cache->last_trans == trans->transid)
+		return 0;
+
+	/*
+	 * detached nodes are used to avoid unnecessary backref
+	 * lookup. transaction commit changes the extent tree.
+	 * so the detached nodes are no longer useful.
+	 */
+	while (!list_empty(&cache->detached)) {
+		node = list_entry(cache->detached.next,
+				  struct backref_node, list);
+		remove_backref_node(cache, node);
+	}
+
+	while (!list_empty(&cache->changed)) {
+		node = list_entry(cache->changed.next,
+				  struct backref_node, list);
+		list_del_init(&node->list);
+		BUG_ON(node->pending);
+		update_backref_node(cache, node, node->new_bytenr);
+	}
+
+	/*
+	 * some nodes can be left in the pending list if there were
+	 * errors during processing the pending nodes.
+	 */
+	for (level = 0; level < BTRFS_MAX_LEVEL; level++) {
+		list_for_each_entry(node, &cache->pending[level], list) {
+			BUG_ON(!node->pending);
+			if (node->bytenr == node->new_bytenr)
+				continue;
+			update_backref_node(cache, node, node->new_bytenr);
+		}
+	}
+
+	cache->last_trans = 0;
+	return 1;
+}
+
+static int should_ignore_root(struct btrfs_root *root)
+{
+	struct btrfs_root *reloc_root;
+
+	if (!root->ref_cows)
+		return 0;
+
+	reloc_root = root->reloc_root;
+	if (!reloc_root)
+		return 0;
+
+	if (btrfs_root_last_snapshot(&reloc_root->root_item) ==
+	    root->fs_info->running_transaction->transid - 1)
+		return 0;
+	/*
+	 * if there is reloc tree and it was created in previous
+	 * transaction backref lookup can find the reloc tree,
+	 * so backref node for the fs tree root is useless for
+	 * relocation.
+	 */
+	return 1;
+}
+
 /*
  * find reloc tree by address of tree root
  */
@@ -453,11 +645,12 @@ int find_inline_backref(struct extent_buffer *leaf, int slot,
  * for all upper level blocks that directly/indirectly reference the
  * block are also cached.
  */
-static struct backref_node *build_backref_tree(struct reloc_control *rc,
-					       struct backref_cache *cache,
-					       struct btrfs_key *node_key,
-					       int level, u64 bytenr)
+static noinline_for_stack
+struct backref_node *build_backref_tree(struct reloc_control *rc,
+					struct btrfs_key *node_key,
+					int level, u64 bytenr)
 {
+	struct backref_cache *cache = &rc->backref_cache;
 	struct btrfs_path *path1;
 	struct btrfs_path *path2;
 	struct extent_buffer *eb;
@@ -473,6 +666,8 @@ static struct backref_node *build_backref_tree(struct reloc_control *rc,
 	unsigned long end;
 	unsigned long ptr;
 	LIST_HEAD(list);
+	LIST_HEAD(useless);
+	int cowonly;
 	int ret;
 	int err = 0;
 
@@ -483,15 +678,13 @@ static struct backref_node *build_backref_tree(struct reloc_control *rc,
 		goto out;
 	}
 
-	node = kmalloc(sizeof(*node), GFP_NOFS);
+	node = alloc_backref_node(cache);
 	if (!node) {
 		err = -ENOMEM;
 		goto out;
 	}
 
-	backref_node_init(node);
 	node->bytenr = bytenr;
-	node->owner = 0;
 	node->level = level;
 	node->lowest = 1;
 	cur = node;
@@ -587,17 +780,20 @@ again:
 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
 		if (key.type == BTRFS_SHARED_BLOCK_REF_KEY ||
 		    key.type == BTRFS_EXTENT_REF_V0_KEY) {
-			if (key.objectid == key.offset &&
-			    key.type == BTRFS_EXTENT_REF_V0_KEY) {
+			if (key.type == BTRFS_EXTENT_REF_V0_KEY) {
 				struct btrfs_extent_ref_v0 *ref0;
 				ref0 = btrfs_item_ptr(eb, path1->slots[0],
 						struct btrfs_extent_ref_v0);
 				root = find_tree_root(rc, eb, ref0);
-				if (root)
-					cur->root = root;
-				else
-					cur->old_root = 1;
-				break;
+				if (!root->ref_cows)
+					cur->cowonly = 1;
+				if (key.objectid == key.offset) {
+					if (root && !should_ignore_root(root))
+						cur->root = root;
+					else
+						list_add(&cur->list, &useless);
+					break;
+				}
 			}
 #else
 		BUG_ON(key.type == BTRFS_EXTENT_REF_V0_KEY);
@@ -614,22 +810,20 @@ again:
 				break;
 			}
 
-			edge = kzalloc(sizeof(*edge), GFP_NOFS);
+			edge = alloc_backref_edge(cache);
 			if (!edge) {
 				err = -ENOMEM;
 				goto out;
 			}
 			rb_node = tree_search(&cache->rb_root, key.offset);
 			if (!rb_node) {
-				upper = kmalloc(sizeof(*upper), GFP_NOFS);
+				upper = alloc_backref_node(cache);
 				if (!upper) {
-					kfree(edge);
+					free_backref_edge(cache, edge);
 					err = -ENOMEM;
 					goto out;
 				}
-				backref_node_init(upper);
 				upper->bytenr = key.offset;
-				upper->owner = 0;
 				upper->level = cur->level + 1;
 				/*
 				 *  backrefs for the upper level block isn't
@@ -639,11 +833,12 @@ again:
 			} else {
 				upper = rb_entry(rb_node, struct backref_node,
 						 rb_node);
+				BUG_ON(!upper->checked);
 				INIT_LIST_HEAD(&edge->list[UPPER]);
 			}
-			list_add(&edge->list[LOWER], &cur->upper);
-			edge->node[UPPER] = upper;
+			list_add_tail(&edge->list[LOWER], &cur->upper);
 			edge->node[LOWER] = cur;
+			edge->node[UPPER] = upper;
 
 			goto next;
 		} else if (key.type != BTRFS_TREE_BLOCK_REF_KEY) {
@@ -657,11 +852,17 @@ again:
 			goto out;
 		}
 
+		if (!root->ref_cows)
+			cur->cowonly = 1;
+
 		if (btrfs_root_level(&root->root_item) == cur->level) {
 			/* tree root */
 			BUG_ON(btrfs_root_bytenr(&root->root_item) !=
 			       cur->bytenr);
-			cur->root = root;
+			if (should_ignore_root(root))
+				list_add(&cur->list, &useless);
+			else
+				cur->root = root;
 			break;
 		}
 
@@ -692,11 +893,14 @@ again:
 			if (!path2->nodes[level]) {
 				BUG_ON(btrfs_root_bytenr(&root->root_item) !=
 				       lower->bytenr);
-				lower->root = root;
+				if (should_ignore_root(root))
+					list_add(&lower->list, &useless);
+				else
+					lower->root = root;
 				break;
 			}
 
-			edge = kzalloc(sizeof(*edge), GFP_NOFS);
+			edge = alloc_backref_edge(cache);
 			if (!edge) {
 				err = -ENOMEM;
 				goto out;
@@ -705,16 +909,17 @@ again:
 			eb = path2->nodes[level];
 			rb_node = tree_search(&cache->rb_root, eb->start);
 			if (!rb_node) {
-				upper = kmalloc(sizeof(*upper), GFP_NOFS);
+				upper = alloc_backref_node(cache);
 				if (!upper) {
-					kfree(edge);
+					free_backref_edge(cache, edge);
 					err = -ENOMEM;
 					goto out;
 				}
-				backref_node_init(upper);
 				upper->bytenr = eb->start;
 				upper->owner = btrfs_header_owner(eb);
 				upper->level = lower->level + 1;
+				if (!root->ref_cows)
+					upper->cowonly = 1;
 
 				/*
 				 * if we know the block isn't shared
@@ -744,10 +949,12 @@ again:
 						 rb_node);
 				BUG_ON(!upper->checked);
 				INIT_LIST_HEAD(&edge->list[UPPER]);
+				if (!upper->owner)
+					upper->owner = btrfs_header_owner(eb);
 			}
 			list_add_tail(&edge->list[LOWER], &lower->upper);
-			edge->node[UPPER] = upper;
 			edge->node[LOWER] = lower;
+			edge->node[UPPER] = upper;
 
 			if (rb_node)
 				break;
@@ -785,8 +992,13 @@ next:
 	 * into the cache.
 	 */
 	BUG_ON(!node->checked);
-	rb_node = tree_insert(&cache->rb_root, node->bytenr, &node->rb_node);
-	BUG_ON(rb_node);
+	cowonly = node->cowonly;
+	if (!cowonly) {
+		rb_node = tree_insert(&cache->rb_root, node->bytenr,
+				      &node->rb_node);
+		BUG_ON(rb_node);
+		list_add_tail(&node->lower, &cache->leaves);
+	}
 
 	list_for_each_entry(edge, &node->upper, list[LOWER])
 		list_add_tail(&edge->list[UPPER], &list);
@@ -795,6 +1007,14 @@ next:
 		edge = list_entry(list.next, struct backref_edge, list[UPPER]);
 		list_del_init(&edge->list[UPPER]);
 		upper = edge->node[UPPER];
+		if (upper->detached) {
+			list_del(&edge->list[LOWER]);
+			lower = edge->node[LOWER];
+			free_backref_edge(cache, edge);
+			if (list_empty(&lower->upper))
+				list_add(&lower->list, &useless);
+			continue;
+		}
 
 		if (!RB_EMPTY_NODE(&upper->rb_node)) {
 			if (upper->lowest) {
@@ -807,25 +1027,69 @@ next:
 		}
 
 		BUG_ON(!upper->checked);
-		rb_node = tree_insert(&cache->rb_root, upper->bytenr,
-				      &upper->rb_node);
-		BUG_ON(rb_node);
+		BUG_ON(cowonly != upper->cowonly);
+		if (!cowonly) {
+			rb_node = tree_insert(&cache->rb_root, upper->bytenr,
+					      &upper->rb_node);
+			BUG_ON(rb_node);
+		}
 
 		list_add_tail(&edge->list[UPPER], &upper->lower);
 
 		list_for_each_entry(edge, &upper->upper, list[LOWER])
 			list_add_tail(&edge->list[UPPER], &list);
 	}
+	/*
+	 * process useless backref nodes. backref nodes for tree leaves
+	 * are deleted from the cache. backref nodes for upper level
+	 * tree blocks are left in the cache to avoid unnecessary backref
+	 * lookup.
+	 */
+	while (!list_empty(&useless)) {
+		upper = list_entry(useless.next, struct backref_node, list);
+		list_del_init(&upper->list);
+		BUG_ON(!list_empty(&upper->upper));
+		if (upper == node)
+			node = NULL;
+		if (upper->lowest) {
+			list_del_init(&upper->lower);
+			upper->lowest = 0;
+		}
+		while (!list_empty(&upper->lower)) {
+			edge = list_entry(upper->lower.next,
+					  struct backref_edge, list[UPPER]);
+			list_del(&edge->list[UPPER]);
+			list_del(&edge->list[LOWER]);
+			lower = edge->node[LOWER];
+			free_backref_edge(cache, edge);
+
+			if (list_empty(&lower->upper))
+				list_add(&lower->list, &useless);
+		}
+		__mark_block_processed(rc, upper);
+		if (upper->level > 0) {
+			list_add(&upper->list, &cache->detached);
+			upper->detached = 1;
+		} else {
+			rb_erase(&upper->rb_node, &cache->rb_root);
+			free_backref_node(cache, upper);
+		}
+	}
 out:
 	btrfs_free_path(path1);
 	btrfs_free_path(path2);
 	if (err) {
-		INIT_LIST_HEAD(&list);
+		while (!list_empty(&useless)) {
+			lower = list_entry(useless.next,
+					   struct backref_node, upper);
+			list_del_init(&lower->upper);
+		}
 		upper = node;
+		INIT_LIST_HEAD(&list);
 		while (upper) {
 			if (RB_EMPTY_NODE(&upper->rb_node)) {
 				list_splice_tail(&upper->upper, &list);
-				kfree(upper);
+				free_backref_node(cache, upper);
 			}
 
 			if (list_empty(&list))
@@ -833,15 +1097,104 @@ out:
 
 			edge = list_entry(list.next, struct backref_edge,
 					  list[LOWER]);
+			list_del(&edge->list[LOWER]);
 			upper = edge->node[UPPER];
-			kfree(edge);
+			free_backref_edge(cache, edge);
 		}
 		return ERR_PTR(err);
 	}
+	BUG_ON(node && node->detached);
 	return node;
 }
 
 /*
+ * helper to add backref node for the newly created snapshot.
+ * the backref node is created by cloning backref node that
+ * corresponds to root of source tree
+ */
+static int clone_backref_node(struct btrfs_trans_handle *trans,
+			      struct reloc_control *rc,
+			      struct btrfs_root *src,
+			      struct btrfs_root *dest)
+{
+	struct btrfs_root *reloc_root = src->reloc_root;
+	struct backref_cache *cache = &rc->backref_cache;
+	struct backref_node *node = NULL;
+	struct backref_node *new_node;
+	struct backref_edge *edge;
+	struct backref_edge *new_edge;
+	struct rb_node *rb_node;
+
+	if (cache->last_trans > 0)
+		update_backref_cache(trans, cache);
+
+	rb_node = tree_search(&cache->rb_root, src->commit_root->start);
+	if (rb_node) {
+		node = rb_entry(rb_node, struct backref_node, rb_node);
+		if (node->detached)
+			node = NULL;
+		else
+			BUG_ON(node->new_bytenr != reloc_root->node->start);
+	}
+
+	if (!node) {
+		rb_node = tree_search(&cache->rb_root,
+				      reloc_root->commit_root->start);
+		if (rb_node) {
+			node = rb_entry(rb_node, struct backref_node,
+					rb_node);
+			BUG_ON(node->detached);
+		}
+	}
+
+	if (!node)
+		return 0;
+
+	new_node = alloc_backref_node(cache);
+	if (!new_node)
+		return -ENOMEM;
+
+	new_node->bytenr = dest->node->start;
+	new_node->level = node->level;
+	new_node->lowest = node->lowest;
+	new_node->root = dest;
+
+	if (!node->lowest) {
+		list_for_each_entry(edge, &node->lower, list[UPPER]) {
+			new_edge = alloc_backref_edge(cache);
+			if (!new_edge)
+				goto fail;
+
+			new_edge->node[UPPER] = new_node;
+			new_edge->node[LOWER] = edge->node[LOWER];
+			list_add_tail(&new_edge->list[UPPER],
+				      &new_node->lower);
+		}
+	}
+
+	rb_node = tree_insert(&cache->rb_root, new_node->bytenr,
+			      &new_node->rb_node);
+	BUG_ON(rb_node);
+
+	if (!new_node->lowest) {
+		list_for_each_entry(new_edge, &new_node->lower, list[UPPER]) {
+			list_add_tail(&new_edge->list[LOWER],
+				      &new_edge->node[LOWER]->upper);
+		}
+	}
+	return 0;
+fail:
+	while (!list_empty(&new_node->lower)) {
+		new_edge = list_entry(new_node->lower.next,
+				      struct backref_edge, list[UPPER]);
+		list_del(&new_edge->list[UPPER]);
+		free_backref_edge(cache, new_edge);
+	}
+	free_backref_node(cache, new_node);
+	return -ENOMEM;
+}
+
+/*
  * helper to add 'address of tree root -> reloc tree' mapping
  */
 static int __add_reloc_root(struct btrfs_root *root)
@@ -901,12 +1254,8 @@ static int __update_reloc_root(struct btrfs_root *root, int del)
 	return 0;
 }
 
-/*
- * create reloc tree for a given fs tree. reloc tree is just a
- * snapshot of the fs tree with special root objectid.
- */
-int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root)
+static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root, u64 objectid)
 {
 	struct btrfs_root *reloc_root;
 	struct extent_buffer *eb;
@@ -914,36 +1263,45 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
 	struct btrfs_key root_key;
 	int ret;
 
-	if (root->reloc_root) {
-		reloc_root = root->reloc_root;
-		reloc_root->last_trans = trans->transid;
-		return 0;
-	}
-
-	if (!root->fs_info->reloc_ctl ||
-	    !root->fs_info->reloc_ctl->create_reloc_root ||
-	    root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
-		return 0;
-
 	root_item = kmalloc(sizeof(*root_item), GFP_NOFS);
 	BUG_ON(!root_item);
 
 	root_key.objectid = BTRFS_TREE_RELOC_OBJECTID;
 	root_key.type = BTRFS_ROOT_ITEM_KEY;
-	root_key.offset = root->root_key.objectid;
+	root_key.offset = objectid;
 
-	ret = btrfs_copy_root(trans, root, root->commit_root, &eb,
-			      BTRFS_TREE_RELOC_OBJECTID);
-	BUG_ON(ret);
+	if (root->root_key.objectid == objectid) {
+		/* called by btrfs_init_reloc_root */
+		ret = btrfs_copy_root(trans, root, root->commit_root, &eb,
+				      BTRFS_TREE_RELOC_OBJECTID);
+		BUG_ON(ret);
+
+		btrfs_set_root_last_snapshot(&root->root_item,
+					     trans->transid - 1);
+	} else {
+		/*
+		 * called by btrfs_reloc_post_snapshot_hook.
+		 * the source tree is a reloc tree, all tree blocks
+		 * modified after it was created have RELOC flag
+		 * set in their headers. so it's OK to not update
+		 * the 'last_snapshot'.
+		 */
+		ret = btrfs_copy_root(trans, root, root->node, &eb,
+				      BTRFS_TREE_RELOC_OBJECTID);
+		BUG_ON(ret);
+	}
 
-	btrfs_set_root_last_snapshot(&root->root_item, trans->transid - 1);
 	memcpy(root_item, &root->root_item, sizeof(*root_item));
-	btrfs_set_root_refs(root_item, 1);
 	btrfs_set_root_bytenr(root_item, eb->start);
 	btrfs_set_root_level(root_item, btrfs_header_level(eb));
 	btrfs_set_root_generation(root_item, trans->transid);
-	memset(&root_item->drop_progress, 0, sizeof(struct btrfs_disk_key));
-	root_item->drop_level = 0;
+
+	if (root->root_key.objectid == objectid) {
+		btrfs_set_root_refs(root_item, 0);
+		memset(&root_item->drop_progress, 0,
+		       sizeof(struct btrfs_disk_key));
+		root_item->drop_level = 0;
+	}
 
 	btrfs_tree_unlock(eb);
 	free_extent_buffer(eb);
@@ -957,6 +1315,37 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
 						 &root_key);
 	BUG_ON(IS_ERR(reloc_root));
 	reloc_root->last_trans = trans->transid;
+	return reloc_root;
+}
+
+/*
+ * create reloc tree for a given fs tree. reloc tree is just a
+ * snapshot of the fs tree with special root objectid.
+ */
+int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root)
+{
+	struct btrfs_root *reloc_root;
+	struct reloc_control *rc = root->fs_info->reloc_ctl;
+	int clear_rsv = 0;
+
+	if (root->reloc_root) {
+		reloc_root = root->reloc_root;
+		reloc_root->last_trans = trans->transid;
+		return 0;
+	}
+
+	if (!rc || !rc->create_reloc_tree ||
+	    root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
+		return 0;
+
+	if (!trans->block_rsv) {
+		trans->block_rsv = rc->block_rsv;
+		clear_rsv = 1;
+	}
+	reloc_root = create_reloc_root(trans, root, root->root_key.objectid);
+	if (clear_rsv)
+		trans->block_rsv = NULL;
 
 	__add_reloc_root(reloc_root);
 	root->reloc_root = reloc_root;
@@ -980,7 +1369,8 @@ int btrfs_update_reloc_root(struct btrfs_trans_handle *trans,
 	reloc_root = root->reloc_root;
 	root_item = &reloc_root->root_item;
 
-	if (btrfs_root_refs(root_item) == 0) {
+	if (root->fs_info->reloc_ctl->merge_reloc_tree &&
+	    btrfs_root_refs(root_item) == 0) {
 		root->reloc_root = NULL;
 		del = 1;
 	}
@@ -1102,8 +1492,7 @@ static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr,
 		goto out;
 	}
 
-	if (new_bytenr)
-		*new_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+	*new_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
 	ret = 0;
 out:
 	btrfs_free_path(path);
@@ -1114,19 +1503,18 @@ out:
  * update file extent items in the tree leaf to point to
  * the new locations.
  */
-static int replace_file_extents(struct btrfs_trans_handle *trans,
-				struct reloc_control *rc,
-				struct btrfs_root *root,
-				struct extent_buffer *leaf,
-				struct list_head *inode_list)
+static noinline_for_stack
+int replace_file_extents(struct btrfs_trans_handle *trans,
+			 struct reloc_control *rc,
+			 struct btrfs_root *root,
+			 struct extent_buffer *leaf)
 {
 	struct btrfs_key key;
 	struct btrfs_file_extent_item *fi;
 	struct inode *inode = NULL;
-	struct inodevec *ivec = NULL;
 	u64 parent;
 	u64 bytenr;
-	u64 new_bytenr;
+	u64 new_bytenr = 0;
 	u64 num_bytes;
 	u64 end;
 	u32 nritems;
@@ -1166,21 +1554,12 @@ static int replace_file_extents(struct btrfs_trans_handle *trans,
 		 * to complete and drop the extent cache
 		 */
 		if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
-			if (!ivec || ivec->nr == INODEVEC_SIZE) {
-				ivec = kmalloc(sizeof(*ivec), GFP_NOFS);
-				BUG_ON(!ivec);
-				ivec->nr = 0;
-				list_add_tail(&ivec->list, inode_list);
-			}
 			if (first) {
 				inode = find_next_inode(root, key.objectid);
-				if (inode)
-					ivec->inode[ivec->nr++] = inode;
 				first = 0;
 			} else if (inode && inode->i_ino < key.objectid) {
+				btrfs_add_delayed_iput(inode);
 				inode = find_next_inode(root, key.objectid);
-				if (inode)
-					ivec->inode[ivec->nr++] = inode;
 			}
 			if (inode && inode->i_ino == key.objectid) {
 				end = key.offset +
@@ -1204,8 +1583,10 @@ static int replace_file_extents(struct btrfs_trans_handle *trans,
 
 		ret = get_new_location(rc->data_inode, &new_bytenr,
 				       bytenr, num_bytes);
-		if (ret > 0)
+		if (ret > 0) {
+			WARN_ON(1);
 			continue;
+		}
 		BUG_ON(ret < 0);
 
 		btrfs_set_file_extent_disk_bytenr(leaf, fi, new_bytenr);
@@ -1225,6 +1606,8 @@ static int replace_file_extents(struct btrfs_trans_handle *trans,
 	}
 	if (dirty)
 		btrfs_mark_buffer_dirty(leaf);
+	if (inode)
+		btrfs_add_delayed_iput(inode);
 	return 0;
 }
 
@@ -1248,11 +1631,11 @@ int memcmp_node_keys(struct extent_buffer *eb, int slot,
  * if no block got replaced, 0 is returned. if there are other
  * errors, a negative error number is returned.
  */
-static int replace_path(struct btrfs_trans_handle *trans,
-			struct btrfs_root *dest, struct btrfs_root *src,
-			struct btrfs_path *path, struct btrfs_key *next_key,
-			struct extent_buffer **leaf,
-			int lowest_level, int max_level)
+static noinline_for_stack
+int replace_path(struct btrfs_trans_handle *trans,
+		 struct btrfs_root *dest, struct btrfs_root *src,
+		 struct btrfs_path *path, struct btrfs_key *next_key,
+		 int lowest_level, int max_level)
 {
 	struct extent_buffer *eb;
 	struct extent_buffer *parent;
@@ -1263,16 +1646,16 @@ static int replace_path(struct btrfs_trans_handle *trans,
 	u64 new_ptr_gen;
 	u64 last_snapshot;
 	u32 blocksize;
+	int cow = 0;
 	int level;
 	int ret;
 	int slot;
 
 	BUG_ON(src->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
 	BUG_ON(dest->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID);
-	BUG_ON(lowest_level > 1 && leaf);
 
 	last_snapshot = btrfs_root_last_snapshot(&src->root_item);
-
+again:
 	slot = path->slots[lowest_level];
 	btrfs_node_key_to_cpu(path->nodes[lowest_level], &key, slot);
 
@@ -1286,8 +1669,10 @@ static int replace_path(struct btrfs_trans_handle *trans,
 		return 0;
 	}
 
-	ret = btrfs_cow_block(trans, dest, eb, NULL, 0, &eb);
-	BUG_ON(ret);
+	if (cow) {
+		ret = btrfs_cow_block(trans, dest, eb, NULL, 0, &eb);
+		BUG_ON(ret);
+	}
 	btrfs_set_lock_blocking(eb);
 
 	if (next_key) {
@@ -1331,7 +1716,7 @@ static int replace_path(struct btrfs_trans_handle *trans,
 
 		if (new_bytenr == 0 || old_ptr_gen > last_snapshot ||
 		    memcmp_node_keys(parent, slot, path, level)) {
-			if (level <= lowest_level && !leaf) {
+			if (level <= lowest_level) {
 				ret = 0;
 				break;
 			}
@@ -1339,16 +1724,12 @@ static int replace_path(struct btrfs_trans_handle *trans,
 			eb = read_tree_block(dest, old_bytenr, blocksize,
 					     old_ptr_gen);
 			btrfs_tree_lock(eb);
-			ret = btrfs_cow_block(trans, dest, eb, parent,
-					      slot, &eb);
-			BUG_ON(ret);
-			btrfs_set_lock_blocking(eb);
-
-			if (level <= lowest_level) {
-				*leaf = eb;
-				ret = 0;
-				break;
+			if (cow) {
+				ret = btrfs_cow_block(trans, dest, eb, parent,
+						      slot, &eb);
+				BUG_ON(ret);
 			}
+			btrfs_set_lock_blocking(eb);
 
 			btrfs_tree_unlock(parent);
 			free_extent_buffer(parent);
@@ -1357,6 +1738,13 @@ static int replace_path(struct btrfs_trans_handle *trans,
 			continue;
 		}
 
+		if (!cow) {
+			btrfs_tree_unlock(parent);
+			free_extent_buffer(parent);
+			cow = 1;
+			goto again;
+		}
+
 		btrfs_node_key_to_cpu(path->nodes[level], &key,
 				      path->slots[level]);
 		btrfs_release_path(src, path);
@@ -1562,20 +1950,6 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 	return 0;
 }
 
-static void put_inodes(struct list_head *list)
-{
-	struct inodevec *ivec;
-	while (!list_empty(list)) {
-		ivec = list_entry(list->next, struct inodevec, list);
-		list_del(&ivec->list);
-		while (ivec->nr > 0) {
-			ivec->nr--;
-			iput(ivec->inode[ivec->nr]);
-		}
-		kfree(ivec);
-	}
-}
-
 static int find_next_key(struct btrfs_path *path, int level,
 			 struct btrfs_key *key)
 
@@ -1608,13 +1982,14 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 	struct btrfs_root *reloc_root;
 	struct btrfs_root_item *root_item;
 	struct btrfs_path *path;
-	struct extent_buffer *leaf = NULL;
+	struct extent_buffer *leaf;
 	unsigned long nr;
 	int level;
 	int max_level;
 	int replaced = 0;
 	int ret;
 	int err = 0;
+	u32 min_reserved;
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -1648,34 +2023,23 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 		btrfs_unlock_up_safe(path, 0);
 	}
 
-	if (level == 0 && rc->stage == UPDATE_DATA_PTRS) {
-		trans = btrfs_start_transaction(root, 1);
+	min_reserved = root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2;
+	memset(&next_key, 0, sizeof(next_key));
 
-		leaf = path->nodes[0];
-		btrfs_item_key_to_cpu(leaf, &key, 0);
-		btrfs_release_path(reloc_root, path);
+	while (1) {
+		trans = btrfs_start_transaction(root, 0);
+		trans->block_rsv = rc->block_rsv;
 
-		ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
-		if (ret < 0) {
-			err = ret;
-			goto out;
+		ret = btrfs_block_rsv_check(trans, root, rc->block_rsv,
+					    min_reserved, 0);
+		if (ret) {
+			BUG_ON(ret != -EAGAIN);
+			ret = btrfs_commit_transaction(trans, root);
+			BUG_ON(ret);
+			continue;
 		}
 
-		leaf = path->nodes[0];
-		btrfs_unlock_up_safe(path, 1);
-		ret = replace_file_extents(trans, rc, root, leaf,
-					   &inode_list);
-		if (ret < 0)
-			err = ret;
-		goto out;
-	}
-
-	memset(&next_key, 0, sizeof(next_key));
-
-	while (1) {
-		leaf = NULL;
 		replaced = 0;
-		trans = btrfs_start_transaction(root, 1);
 		max_level = level;
 
 		ret = walk_down_reloc_tree(reloc_root, path, &level);
@@ -1689,14 +2053,9 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 		if (!find_next_key(path, level, &key) &&
 		    btrfs_comp_cpu_keys(&next_key, &key) >= 0) {
 			ret = 0;
-		} else if (level == 1 && rc->stage == UPDATE_DATA_PTRS) {
-			ret = replace_path(trans, root, reloc_root,
-					   path, &next_key, &leaf,
-					   level, max_level);
 		} else {
-			ret = replace_path(trans, root, reloc_root,
-					   path, &next_key, NULL,
-					   level, max_level);
+			ret = replace_path(trans, root, reloc_root, path,
+					   &next_key, level, max_level);
 		}
 		if (ret < 0) {
 			err = ret;
@@ -1708,16 +2067,6 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 			btrfs_node_key_to_cpu(path->nodes[level], &key,
 					      path->slots[level]);
 			replaced = 1;
-		} else if (leaf) {
-			/*
-			 * no block got replaced, try replacing file extents
-			 */
-			btrfs_item_key_to_cpu(leaf, &key, 0);
-			ret = replace_file_extents(trans, rc, root, leaf,
-						   &inode_list);
-			btrfs_tree_unlock(leaf);
-			free_extent_buffer(leaf);
-			BUG_ON(ret < 0);
 		}
 
 		ret = walk_up_reloc_tree(reloc_root, path, &level);
@@ -1734,15 +2083,10 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 		root_item->drop_level = level;
 
 		nr = trans->blocks_used;
-		btrfs_end_transaction(trans, root);
+		btrfs_end_transaction_throttle(trans, root);
 
 		btrfs_btree_balance_dirty(root, nr);
 
-		/*
-		 * put inodes outside transaction, otherwise we may deadlock.
-		 */
-		put_inodes(&inode_list);
-
 		if (replaced && rc->stage == UPDATE_DATA_PTRS)
 			invalidate_extent_cache(root, &key, &next_key);
 	}
@@ -1765,87 +2109,125 @@ out:
 		       sizeof(root_item->drop_progress));
 		root_item->drop_level = 0;
 		btrfs_set_root_refs(root_item, 0);
+		btrfs_update_reloc_root(trans, root);
 	}
 
 	nr = trans->blocks_used;
-	btrfs_end_transaction(trans, root);
+	btrfs_end_transaction_throttle(trans, root);
 
 	btrfs_btree_balance_dirty(root, nr);
 
-	put_inodes(&inode_list);
-
 	if (replaced && rc->stage == UPDATE_DATA_PTRS)
 		invalidate_extent_cache(root, &key, &next_key);
 
 	return err;
 }
 
-/*
- * callback for the work threads.
- * this function merges reloc tree with corresponding fs tree,
- * and then drops the reloc tree.
- */
-static void merge_func(struct btrfs_work *work)
+static noinline_for_stack
+int prepare_to_merge(struct reloc_control *rc, int err)
 {
-	struct btrfs_trans_handle *trans;
-	struct btrfs_root *root;
+	struct btrfs_root *root = rc->extent_root;
 	struct btrfs_root *reloc_root;
-	struct async_merge *async;
+	struct btrfs_trans_handle *trans;
+	LIST_HEAD(reloc_roots);
+	u64 num_bytes = 0;
+	int ret;
+	int retries = 0;
+
+	mutex_lock(&root->fs_info->trans_mutex);
+	rc->merging_rsv_size += root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2;
+	rc->merging_rsv_size += rc->nodes_relocated * 2;
+	mutex_unlock(&root->fs_info->trans_mutex);
+again:
+	if (!err) {
+		num_bytes = rc->merging_rsv_size;
+		ret = btrfs_block_rsv_add(NULL, root, rc->block_rsv,
+					  num_bytes, &retries);
+		if (ret)
+			err = ret;
+	}
+
+	trans = btrfs_join_transaction(rc->extent_root, 1);
+
+	if (!err) {
+		if (num_bytes != rc->merging_rsv_size) {
+			btrfs_end_transaction(trans, rc->extent_root);
+			btrfs_block_rsv_release(rc->extent_root,
+						rc->block_rsv, num_bytes);
+			retries = 0;
+			goto again;
+		}
+	}
 
-	async = container_of(work, struct async_merge, work);
-	reloc_root = async->root;
+	rc->merge_reloc_tree = 1;
+
+	while (!list_empty(&rc->reloc_roots)) {
+		reloc_root = list_entry(rc->reloc_roots.next,
+					struct btrfs_root, root_list);
+		list_del_init(&reloc_root->root_list);
 
-	if (btrfs_root_refs(&reloc_root->root_item) > 0) {
 		root = read_fs_root(reloc_root->fs_info,
 				    reloc_root->root_key.offset);
 		BUG_ON(IS_ERR(root));
 		BUG_ON(root->reloc_root != reloc_root);
 
-		merge_reloc_root(async->rc, root);
-
-		trans = btrfs_start_transaction(root, 1);
+		/*
+		 * set reference count to 1, so btrfs_recover_relocation
+		 * knows it should resumes merging
+		 */
+		if (!err)
+			btrfs_set_root_refs(&reloc_root->root_item, 1);
 		btrfs_update_reloc_root(trans, root);
-		btrfs_end_transaction(trans, root);
-	}
 
-	btrfs_drop_snapshot(reloc_root, 0);
+		list_add(&reloc_root->root_list, &reloc_roots);
+	}
 
-	if (atomic_dec_and_test(async->num_pending))
-		complete(async->done);
+	list_splice(&reloc_roots, &rc->reloc_roots);
 
-	kfree(async);
+	if (!err)
+		btrfs_commit_transaction(trans, rc->extent_root);
+	else
+		btrfs_end_transaction(trans, rc->extent_root);
+	return err;
 }
 
-static int merge_reloc_roots(struct reloc_control *rc)
+static noinline_for_stack
+int merge_reloc_roots(struct reloc_control *rc)
 {
-	struct async_merge *async;
 	struct btrfs_root *root;
-	struct completion done;
-	atomic_t num_pending;
+	struct btrfs_root *reloc_root;
+	LIST_HEAD(reloc_roots);
+	int found = 0;
+	int ret;
+again:
+	root = rc->extent_root;
+	mutex_lock(&root->fs_info->trans_mutex);
+	list_splice_init(&rc->reloc_roots, &reloc_roots);
+	mutex_unlock(&root->fs_info->trans_mutex);
 
-	init_completion(&done);
-	atomic_set(&num_pending, 1);
+	while (!list_empty(&reloc_roots)) {
+		found = 1;
+		reloc_root = list_entry(reloc_roots.next,
+					struct btrfs_root, root_list);
 
-	while (!list_empty(&rc->reloc_roots)) {
-		root = list_entry(rc->reloc_roots.next,
-				  struct btrfs_root, root_list);
-		list_del_init(&root->root_list);
+		if (btrfs_root_refs(&reloc_root->root_item) > 0) {
+			root = read_fs_root(reloc_root->fs_info,
+					    reloc_root->root_key.offset);
+			BUG_ON(IS_ERR(root));
+			BUG_ON(root->reloc_root != reloc_root);
 
-		async = kmalloc(sizeof(*async), GFP_NOFS);
-		BUG_ON(!async);
-		async->work.func = merge_func;
-		async->work.flags = 0;
-		async->rc = rc;
-		async->root = root;
-		async->done = &done;
-		async->num_pending = &num_pending;
-		atomic_inc(&num_pending);
-		btrfs_queue_worker(&rc->workers, &async->work);
+			ret = merge_reloc_root(rc, root);
+			BUG_ON(ret);
+		} else {
+			list_del_init(&reloc_root->root_list);
+		}
+		btrfs_drop_snapshot(reloc_root, rc->block_rsv, 0);
 	}
 
-	if (!atomic_dec_and_test(&num_pending))
-		wait_for_completion(&done);
-
+	if (found) {
+		found = 0;
+		goto again;
+	}
 	BUG_ON(!RB_EMPTY_ROOT(&rc->reloc_root_tree.rb_root));
 	return 0;
 }
@@ -1876,119 +2258,169 @@ static int record_reloc_root_in_trans(struct btrfs_trans_handle *trans,
 	return btrfs_record_root_in_trans(trans, root);
 }
 
-/*
- * select one tree from trees that references the block.
- * for blocks in refernce counted trees, we preper reloc tree.
- * if no reloc tree found and reloc_only is true, NULL is returned.
- */
-static struct btrfs_root *__select_one_root(struct btrfs_trans_handle *trans,
-					    struct backref_node *node,
-					    struct backref_edge *edges[],
-					    int *nr, int reloc_only)
+static noinline_for_stack
+struct btrfs_root *select_reloc_root(struct btrfs_trans_handle *trans,
+				     struct reloc_control *rc,
+				     struct backref_node *node,
+				     struct backref_edge *edges[], int *nr)
 {
 	struct backref_node *next;
 	struct btrfs_root *root;
-	int index;
-	int loop = 0;
-again:
-	index = 0;
+	int index = 0;
+
 	next = node;
 	while (1) {
 		cond_resched();
 		next = walk_up_backref(next, edges, &index);
 		root = next->root;
-		if (!root) {
-			BUG_ON(!node->old_root);
-			goto skip;
-		}
-
-		/* no other choice for non-refernce counted tree */
-		if (!root->ref_cows) {
-			BUG_ON(reloc_only);
-			break;
-		}
+		BUG_ON(!root);
+		BUG_ON(!root->ref_cows);
 
 		if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) {
 			record_reloc_root_in_trans(trans, root);
 			break;
 		}
 
-		if (loop) {
-			btrfs_record_root_in_trans(trans, root);
+		btrfs_record_root_in_trans(trans, root);
+		root = root->reloc_root;
+
+		if (next->new_bytenr != root->node->start) {
+			BUG_ON(next->new_bytenr);
+			BUG_ON(!list_empty(&next->list));
+			next->new_bytenr = root->node->start;
+			next->root = root;
+			list_add_tail(&next->list,
+				      &rc->backref_cache.changed);
+			__mark_block_processed(rc, next);
 			break;
 		}
 
-		if (reloc_only || next != node) {
-			if (!root->reloc_root)
-				btrfs_record_root_in_trans(trans, root);
-			root = root->reloc_root;
-			/*
-			 * if the reloc tree was created in current
-			 * transation, there is no node in backref tree
-			 * corresponds to the root of the reloc tree.
-			 */
-			if (btrfs_root_last_snapshot(&root->root_item) ==
-			    trans->transid - 1)
-				break;
-		}
-skip:
+		WARN_ON(1);
 		root = NULL;
 		next = walk_down_backref(edges, &index);
 		if (!next || next->level <= node->level)
 			break;
 	}
+	if (!root)
+		return NULL;
 
-	if (!root && !loop && !reloc_only) {
-		loop = 1;
-		goto again;
+	*nr = index;
+	next = node;
+	/* setup backref node path for btrfs_reloc_cow_block */
+	while (1) {
+		rc->backref_cache.path[next->level] = next;
+		if (--index < 0)
+			break;
+		next = edges[index]->node[UPPER];
 	}
-
-	if (root)
-		*nr = index;
-	else
-		*nr = 0;
-
 	return root;
 }
 
+/*
+ * select a tree root for relocation. return NULL if the block
+ * is reference counted. we should use do_relocation() in this
+ * case. return a tree root pointer if the block isn't reference
+ * counted. return -ENOENT if the block is root of reloc tree.
+ */
 static noinline_for_stack
 struct btrfs_root *select_one_root(struct btrfs_trans_handle *trans,
 				   struct backref_node *node)
 {
+	struct backref_node *next;
+	struct btrfs_root *root;
+	struct btrfs_root *fs_root = NULL;
 	struct backref_edge *edges[BTRFS_MAX_LEVEL - 1];
-	int nr;
-	return __select_one_root(trans, node, edges, &nr, 0);
+	int index = 0;
+
+	next = node;
+	while (1) {
+		cond_resched();
+		next = walk_up_backref(next, edges, &index);
+		root = next->root;
+		BUG_ON(!root);
+
+		/* no other choice for non-refernce counted tree */
+		if (!root->ref_cows)
+			return root;
+
+		if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID)
+			fs_root = root;
+
+		if (next != node)
+			return NULL;
+
+		next = walk_down_backref(edges, &index);
+		if (!next || next->level <= node->level)
+			break;
+	}
+
+	if (!fs_root)
+		return ERR_PTR(-ENOENT);
+	return fs_root;
 }
 
 static noinline_for_stack
-struct btrfs_root *select_reloc_root(struct btrfs_trans_handle *trans,
-				     struct backref_node *node,
-				     struct backref_edge *edges[], int *nr)
+u64 calcu_metadata_size(struct reloc_control *rc,
+			struct backref_node *node, int reserve)
 {
-	return __select_one_root(trans, node, edges, nr, 1);
+	struct backref_node *next = node;
+	struct backref_edge *edge;
+	struct backref_edge *edges[BTRFS_MAX_LEVEL - 1];
+	u64 num_bytes = 0;
+	int index = 0;
+
+	BUG_ON(reserve && node->processed);
+
+	while (next) {
+		cond_resched();
+		while (1) {
+			if (next->processed && (reserve || next != node))
+				break;
+
+			num_bytes += btrfs_level_size(rc->extent_root,
+						      next->level);
+
+			if (list_empty(&next->upper))
+				break;
+
+			edge = list_entry(next->upper.next,
+					  struct backref_edge, list[LOWER]);
+			edges[index++] = edge;
+			next = edge->node[UPPER];
+		}
+		next = walk_down_backref(edges, &index);
+	}
+	return num_bytes;
 }
 
-static void grab_path_buffers(struct btrfs_path *path,
-			      struct backref_node *node,
-			      struct backref_edge *edges[], int nr)
+static int reserve_metadata_space(struct btrfs_trans_handle *trans,
+				  struct reloc_control *rc,
+				  struct backref_node *node)
 {
-	int i = 0;
-	while (1) {
-		drop_node_buffer(node);
-		node->eb = path->nodes[node->level];
-		BUG_ON(!node->eb);
-		if (path->locks[node->level])
-			node->locked = 1;
-		path->nodes[node->level] = NULL;
-		path->locks[node->level] = 0;
-
-		if (i >= nr)
-			break;
+	struct btrfs_root *root = rc->extent_root;
+	u64 num_bytes;
+	int ret;
+
+	num_bytes = calcu_metadata_size(rc, node, 1) * 2;
 
-		edges[i]->blockptr = node->eb->start;
-		node = edges[i]->node[UPPER];
-		i++;
+	trans->block_rsv = rc->block_rsv;
+	ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes,
+				  &rc->block_rsv_retries);
+	if (ret) {
+		if (ret == -EAGAIN)
+			rc->commit_transaction = 1;
+		return ret;
 	}
+
+	rc->block_rsv_retries = 0;
+	return 0;
+}
+
+static void release_metadata_space(struct reloc_control *rc,
+				   struct backref_node *node)
+{
+	u64 num_bytes = calcu_metadata_size(rc, node, 0) * 2;
+	btrfs_block_rsv_release(rc->extent_root, rc->block_rsv, num_bytes);
 }
 
 /*
@@ -1999,6 +2431,7 @@ static void grab_path_buffers(struct btrfs_path *path,
  * in that case this function just updates pointers.
  */
 static int do_relocation(struct btrfs_trans_handle *trans,
+			 struct reloc_control *rc,
 			 struct backref_node *node,
 			 struct btrfs_key *key,
 			 struct btrfs_path *path, int lowest)
@@ -2019,18 +2452,25 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 	BUG_ON(lowest && node->eb);
 
 	path->lowest_level = node->level + 1;
+	rc->backref_cache.path[node->level] = node;
 	list_for_each_entry(edge, &node->upper, list[LOWER]) {
 		cond_resched();
-		if (node->eb && node->eb->start == edge->blockptr)
-			continue;
 
 		upper = edge->node[UPPER];
-		root = select_reloc_root(trans, upper, edges, &nr);
-		if (!root)
-			continue;
-
-		if (upper->eb && !upper->locked)
+		root = select_reloc_root(trans, rc, upper, edges, &nr);
+		BUG_ON(!root);
+
+		if (upper->eb && !upper->locked) {
+			if (!lowest) {
+				ret = btrfs_bin_search(upper->eb, key,
+						       upper->level, &slot);
+				BUG_ON(ret);
+				bytenr = btrfs_node_blockptr(upper->eb, slot);
+				if (node->eb->start == bytenr)
+					goto next;
+			}
 			drop_node_buffer(upper);
+		}
 
 		if (!upper->eb) {
 			ret = btrfs_search_slot(trans, root, key, path, 0, 1);
@@ -2040,11 +2480,17 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 			}
 			BUG_ON(ret > 0);
 
-			slot = path->slots[upper->level];
+			if (!upper->eb) {
+				upper->eb = path->nodes[upper->level];
+				path->nodes[upper->level] = NULL;
+			} else {
+				BUG_ON(upper->eb != path->nodes[upper->level]);
+			}
 
-			btrfs_unlock_up_safe(path, upper->level + 1);
-			grab_path_buffers(path, upper, edges, nr);
+			upper->locked = 1;
+			path->locks[upper->level] = 0;
 
+			slot = path->slots[upper->level];
 			btrfs_release_path(NULL, path);
 		} else {
 			ret = btrfs_bin_search(upper->eb, key, upper->level,
@@ -2053,14 +2499,11 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 		}
 
 		bytenr = btrfs_node_blockptr(upper->eb, slot);
-		if (!lowest) {
-			if (node->eb->start == bytenr) {
-				btrfs_tree_unlock(upper->eb);
-				upper->locked = 0;
-				continue;
-			}
+		if (lowest) {
+			BUG_ON(bytenr != node->bytenr);
 		} else {
-			BUG_ON(node->bytenr != bytenr);
+			if (node->eb->start == bytenr)
+				goto next;
 		}
 
 		blocksize = btrfs_level_size(root, node->level);
@@ -2072,13 +2515,13 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 		if (!node->eb) {
 			ret = btrfs_cow_block(trans, root, eb, upper->eb,
 					      slot, &eb);
+			btrfs_tree_unlock(eb);
+			free_extent_buffer(eb);
 			if (ret < 0) {
 				err = ret;
-				break;
+				goto next;
 			}
-			btrfs_set_lock_blocking(eb);
-			node->eb = eb;
-			node->locked = 1;
+			BUG_ON(node->eb != eb);
 		} else {
 			btrfs_set_node_blockptr(upper->eb, slot,
 						node->eb->start);
@@ -2096,67 +2539,80 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 			ret = btrfs_drop_subtree(trans, root, eb, upper->eb);
 			BUG_ON(ret);
 		}
-		if (!lowest) {
-			btrfs_tree_unlock(upper->eb);
-			upper->locked = 0;
-		}
+next:
+		if (!upper->pending)
+			drop_node_buffer(upper);
+		else
+			unlock_node_buffer(upper);
+		if (err)
+			break;
 	}
+
+	if (!err && node->pending) {
+		drop_node_buffer(node);
+		list_move_tail(&node->list, &rc->backref_cache.changed);
+		node->pending = 0;
+	}
+
 	path->lowest_level = 0;
+	BUG_ON(err == -ENOSPC);
 	return err;
 }
 
 static int link_to_upper(struct btrfs_trans_handle *trans,
+			 struct reloc_control *rc,
 			 struct backref_node *node,
 			 struct btrfs_path *path)
 {
 	struct btrfs_key key;
-	if (!node->eb || list_empty(&node->upper))
-		return 0;
 
 	btrfs_node_key_to_cpu(node->eb, &key, 0);
-	return do_relocation(trans, node, &key, path, 0);
+	return do_relocation(trans, rc, node, &key, path, 0);
 }
 
 static int finish_pending_nodes(struct btrfs_trans_handle *trans,
-				struct backref_cache *cache,
-				struct btrfs_path *path)
+				struct reloc_control *rc,
+				struct btrfs_path *path, int err)
 {
+	LIST_HEAD(list);
+	struct backref_cache *cache = &rc->backref_cache;
 	struct backref_node *node;
 	int level;
 	int ret;
-	int err = 0;
 
 	for (level = 0; level < BTRFS_MAX_LEVEL; level++) {
 		while (!list_empty(&cache->pending[level])) {
 			node = list_entry(cache->pending[level].next,
-					  struct backref_node, lower);
-			BUG_ON(node->level != level);
+					  struct backref_node, list);
+			list_move_tail(&node->list, &list);
+			BUG_ON(!node->pending);
 
-			ret = link_to_upper(trans, node, path);
-			if (ret < 0)
-				err = ret;
-			/*
-			 * this remove the node from the pending list and
-			 * may add some other nodes to the level + 1
-			 * pending list
-			 */
-			remove_backref_node(cache, node);
+			if (!err) {
+				ret = link_to_upper(trans, rc, node, path);
+				if (ret < 0)
+					err = ret;
+			}
 		}
+		list_splice_init(&list, &cache->pending[level]);
 	}
-	BUG_ON(!RB_EMPTY_ROOT(&cache->rb_root));
 	return err;
 }
 
 static void mark_block_processed(struct reloc_control *rc,
-				 struct backref_node *node)
+				 u64 bytenr, u32 blocksize)
+{
+	set_extent_bits(&rc->processed_blocks, bytenr, bytenr + blocksize - 1,
+			EXTENT_DIRTY, GFP_NOFS);
+}
+
+static void __mark_block_processed(struct reloc_control *rc,
+				   struct backref_node *node)
 {
 	u32 blocksize;
 	if (node->level == 0 ||
 	    in_block_group(node->bytenr, rc->block_group)) {
 		blocksize = btrfs_level_size(rc->extent_root, node->level);
-		set_extent_bits(&rc->processed_blocks, node->bytenr,
-				node->bytenr + blocksize - 1, EXTENT_DIRTY,
-				GFP_NOFS);
+		mark_block_processed(rc, node->bytenr, blocksize);
 	}
 	node->processed = 1;
 }
@@ -2179,7 +2635,7 @@ static void update_processed_blocks(struct reloc_control *rc,
 			if (next->processed)
 				break;
 
-			mark_block_processed(rc, next);
+			__mark_block_processed(rc, next);
 
 			if (list_empty(&next->upper))
 				break;
@@ -2202,138 +2658,6 @@ static int tree_block_processed(u64 bytenr, u32 blocksize,
 	return 0;
 }
 
-/*
- * check if there are any file extent pointers in the leaf point to
- * data require processing
- */
-static int check_file_extents(struct reloc_control *rc,
-			      u64 bytenr, u32 blocksize, u64 ptr_gen)
-{
-	struct btrfs_key found_key;
-	struct btrfs_file_extent_item *fi;
-	struct extent_buffer *leaf;
-	u32 nritems;
-	int i;
-	int ret = 0;
-
-	leaf = read_tree_block(rc->extent_root, bytenr, blocksize, ptr_gen);
-
-	nritems = btrfs_header_nritems(leaf);
-	for (i = 0; i < nritems; i++) {
-		cond_resched();
-		btrfs_item_key_to_cpu(leaf, &found_key, i);
-		if (found_key.type != BTRFS_EXTENT_DATA_KEY)
-			continue;
-		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
-		if (btrfs_file_extent_type(leaf, fi) ==
-		    BTRFS_FILE_EXTENT_INLINE)
-			continue;
-		bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
-		if (bytenr == 0)
-			continue;
-		if (in_block_group(bytenr, rc->block_group)) {
-			ret = 1;
-			break;
-		}
-	}
-	free_extent_buffer(leaf);
-	return ret;
-}
-
-/*
- * scan child blocks of a given block to find blocks require processing
- */
-static int add_child_blocks(struct btrfs_trans_handle *trans,
-			    struct reloc_control *rc,
-			    struct backref_node *node,
-			    struct rb_root *blocks)
-{
-	struct tree_block *block;
-	struct rb_node *rb_node;
-	u64 bytenr;
-	u64 ptr_gen;
-	u32 blocksize;
-	u32 nritems;
-	int i;
-	int err = 0;
-
-	nritems = btrfs_header_nritems(node->eb);
-	blocksize = btrfs_level_size(rc->extent_root, node->level - 1);
-	for (i = 0; i < nritems; i++) {
-		cond_resched();
-		bytenr = btrfs_node_blockptr(node->eb, i);
-		ptr_gen = btrfs_node_ptr_generation(node->eb, i);
-		if (ptr_gen == trans->transid)
-			continue;
-		if (!in_block_group(bytenr, rc->block_group) &&
-		    (node->level > 1 || rc->stage == MOVE_DATA_EXTENTS))
-			continue;
-		if (tree_block_processed(bytenr, blocksize, rc))
-			continue;
-
-		readahead_tree_block(rc->extent_root,
-				     bytenr, blocksize, ptr_gen);
-	}
-
-	for (i = 0; i < nritems; i++) {
-		cond_resched();
-		bytenr = btrfs_node_blockptr(node->eb, i);
-		ptr_gen = btrfs_node_ptr_generation(node->eb, i);
-		if (ptr_gen == trans->transid)
-			continue;
-		if (!in_block_group(bytenr, rc->block_group) &&
-		    (node->level > 1 || rc->stage == MOVE_DATA_EXTENTS))
-			continue;
-		if (tree_block_processed(bytenr, blocksize, rc))
-			continue;
-		if (!in_block_group(bytenr, rc->block_group) &&
-		    !check_file_extents(rc, bytenr, blocksize, ptr_gen))
-			continue;
-
-		block = kmalloc(sizeof(*block), GFP_NOFS);
-		if (!block) {
-			err = -ENOMEM;
-			break;
-		}
-		block->bytenr = bytenr;
-		btrfs_node_key_to_cpu(node->eb, &block->key, i);
-		block->level = node->level - 1;
-		block->key_ready = 1;
-		rb_node = tree_insert(blocks, block->bytenr, &block->rb_node);
-		BUG_ON(rb_node);
-	}
-	if (err)
-		free_block_list(blocks);
-	return err;
-}
-
-/*
- * find adjacent blocks require processing
- */
-static noinline_for_stack
-int add_adjacent_blocks(struct btrfs_trans_handle *trans,
-			struct reloc_control *rc,
-			struct backref_cache *cache,
-			struct rb_root *blocks, int level,
-			struct backref_node **upper)
-{
-	struct backref_node *node;
-	int ret = 0;
-
-	WARN_ON(!list_empty(&cache->pending[level]));
-
-	if (list_empty(&cache->pending[level + 1]))
-		return 1;
-
-	node = list_entry(cache->pending[level + 1].next,
-			  struct backref_node, lower);
-	if (node->eb)
-		ret = add_child_blocks(trans, rc, node, blocks);
-
-	*upper = node;
-	return ret;
-}
-
 static int get_tree_block_key(struct reloc_control *rc,
 			      struct tree_block *block)
 {
@@ -2371,40 +2695,53 @@ static int relocate_tree_block(struct btrfs_trans_handle *trans,
 				struct btrfs_path *path)
 {
 	struct btrfs_root *root;
-	int ret;
+	int release = 0;
+	int ret = 0;
 
+	if (!node)
+		return 0;
+
+	BUG_ON(node->processed);
 	root = select_one_root(trans, node);
-	if (unlikely(!root)) {
-		rc->found_old_snapshot = 1;
+	if (root == ERR_PTR(-ENOENT)) {
 		update_processed_blocks(rc, node);
-		return 0;
+		goto out;
 	}
 
-	if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) {
-		ret = do_relocation(trans, node, key, path, 1);
-		if (ret < 0)
-			goto out;
-		if (node->level == 0 && rc->stage == UPDATE_DATA_PTRS) {
-			ret = replace_file_extents(trans, rc, root,
-						   node->eb, NULL);
-			if (ret < 0)
-				goto out;
-		}
-		drop_node_buffer(node);
-	} else if (!root->ref_cows) {
-		path->lowest_level = node->level;
-		ret = btrfs_search_slot(trans, root, key, path, 0, 1);
-		btrfs_release_path(root, path);
-		if (ret < 0)
+	if (!root || root->ref_cows) {
+		ret = reserve_metadata_space(trans, rc, node);
+		if (ret)
 			goto out;
-	} else if (root != node->root) {
-		WARN_ON(node->level > 0 || rc->stage != UPDATE_DATA_PTRS);
+		release = 1;
 	}
 
-	update_processed_blocks(rc, node);
-	ret = 0;
+	if (root) {
+		if (root->ref_cows) {
+			BUG_ON(node->new_bytenr);
+			BUG_ON(!list_empty(&node->list));
+			btrfs_record_root_in_trans(trans, root);
+			root = root->reloc_root;
+			node->new_bytenr = root->node->start;
+			node->root = root;
+			list_add_tail(&node->list, &rc->backref_cache.changed);
+		} else {
+			path->lowest_level = node->level;
+			ret = btrfs_search_slot(trans, root, key, path, 0, 1);
+			btrfs_release_path(root, path);
+			if (ret > 0)
+				ret = 0;
+		}
+		if (!ret)
+			update_processed_blocks(rc, node);
+	} else {
+		ret = do_relocation(trans, rc, node, key, path, 1);
+	}
 out:
-	drop_node_buffer(node);
+	if (ret || node->level == 0 || node->cowonly) {
+		if (release)
+			release_metadata_space(rc, node);
+		remove_backref_node(&rc->backref_cache, node);
+	}
 	return ret;
 }
 
@@ -2415,12 +2752,10 @@ static noinline_for_stack
 int relocate_tree_blocks(struct btrfs_trans_handle *trans,
 			 struct reloc_control *rc, struct rb_root *blocks)
 {
-	struct backref_cache *cache;
 	struct backref_node *node;
 	struct btrfs_path *path;
 	struct tree_block *block;
 	struct rb_node *rb_node;
-	int level = -1;
 	int ret;
 	int err = 0;
 
@@ -2428,21 +2763,9 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
 	if (!path)
 		return -ENOMEM;
 
-	cache = kmalloc(sizeof(*cache), GFP_NOFS);
-	if (!cache) {
-		btrfs_free_path(path);
-		return -ENOMEM;
-	}
-
-	backref_cache_init(cache);
-
 	rb_node = rb_first(blocks);
 	while (rb_node) {
 		block = rb_entry(rb_node, struct tree_block, rb_node);
-		if (level == -1)
-			level = block->level;
-		else
-			BUG_ON(level != block->level);
 		if (!block->key_ready)
 			reada_tree_block(rc, block);
 		rb_node = rb_next(rb_node);
@@ -2460,7 +2783,7 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
 	while (rb_node) {
 		block = rb_entry(rb_node, struct tree_block, rb_node);
 
-		node = build_backref_tree(rc, cache, &block->key,
+		node = build_backref_tree(rc, &block->key,
 					  block->level, block->bytenr);
 		if (IS_ERR(node)) {
 			err = PTR_ERR(node);
@@ -2470,79 +2793,62 @@ int relocate_tree_blocks(struct btrfs_trans_handle *trans,
 		ret = relocate_tree_block(trans, rc, node, &block->key,
 					  path);
 		if (ret < 0) {
-			err = ret;
+			if (ret != -EAGAIN || rb_node == rb_first(blocks))
+				err = ret;
 			goto out;
 		}
-		remove_backref_node(cache, node);
 		rb_node = rb_next(rb_node);
 	}
-
-	if (level > 0)
-		goto out;
-
+out:
 	free_block_list(blocks);
+	err = finish_pending_nodes(trans, rc, path, err);
 
-	/*
-	 * now backrefs of some upper level tree blocks have been cached,
-	 * try relocating blocks referenced by these upper level blocks.
-	 */
-	while (1) {
-		struct backref_node *upper = NULL;
-		if (trans->transaction->in_commit ||
-		    trans->transaction->delayed_refs.flushing)
-			break;
+	btrfs_free_path(path);
+	return err;
+}
 
-		ret = add_adjacent_blocks(trans, rc, cache, blocks, level,
-					  &upper);
-		if (ret < 0)
-			err = ret;
-		if (ret != 0)
-			break;
+static noinline_for_stack
+int prealloc_file_extent_cluster(struct inode *inode,
+				 struct file_extent_cluster *cluster)
+{
+	u64 alloc_hint = 0;
+	u64 start;
+	u64 end;
+	u64 offset = BTRFS_I(inode)->index_cnt;
+	u64 num_bytes;
+	int nr = 0;
+	int ret = 0;
 
-		rb_node = rb_first(blocks);
-		while (rb_node) {
-			block = rb_entry(rb_node, struct tree_block, rb_node);
-			if (trans->transaction->in_commit ||
-			    trans->transaction->delayed_refs.flushing)
-				goto out;
-			BUG_ON(!block->key_ready);
-			node = build_backref_tree(rc, cache, &block->key,
-						  level, block->bytenr);
-			if (IS_ERR(node)) {
-				err = PTR_ERR(node);
-				goto out;
-			}
+	BUG_ON(cluster->start != cluster->boundary[0]);
+	mutex_lock(&inode->i_mutex);
 
-			ret = relocate_tree_block(trans, rc, node,
-						  &block->key, path);
-			if (ret < 0) {
-				err = ret;
-				goto out;
-			}
-			remove_backref_node(cache, node);
-			rb_node = rb_next(rb_node);
-		}
-		free_block_list(blocks);
+	ret = btrfs_check_data_free_space(inode, cluster->end +
+					  1 - cluster->start);
+	if (ret)
+		goto out;
 
-		if (upper) {
-			ret = link_to_upper(trans, upper, path);
-			if (ret < 0) {
-				err = ret;
-				break;
-			}
-			remove_backref_node(cache, upper);
-		}
+	while (nr < cluster->nr) {
+		start = cluster->boundary[nr] - offset;
+		if (nr + 1 < cluster->nr)
+			end = cluster->boundary[nr + 1] - 1 - offset;
+		else
+			end = cluster->end - offset;
+
+		lock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+		num_bytes = end + 1 - start;
+		ret = btrfs_prealloc_file_range(inode, 0, start,
+						num_bytes, num_bytes,
+						end + 1, &alloc_hint);
+		unlock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+		if (ret)
+			break;
+		nr++;
 	}
+	btrfs_free_reserved_data_space(inode, cluster->end +
+				       1 - cluster->start);
 out:
-	free_block_list(blocks);
-
-	ret = finish_pending_nodes(trans, cache, path);
-	if (ret < 0)
-		err = ret;
-
-	kfree(cache);
-	btrfs_free_path(path);
-	return err;
+	mutex_unlock(&inode->i_mutex);
+	return ret;
 }
 
 static noinline_for_stack
@@ -2588,7 +2894,6 @@ static int relocate_file_extent_cluster(struct inode *inode,
 	u64 offset = BTRFS_I(inode)->index_cnt;
 	unsigned long index;
 	unsigned long last_index;
-	unsigned int dirty_page = 0;
 	struct page *page;
 	struct file_ra_state *ra;
 	int nr = 0;
@@ -2601,21 +2906,24 @@ static int relocate_file_extent_cluster(struct inode *inode,
 	if (!ra)
 		return -ENOMEM;
 
-	index = (cluster->start - offset) >> PAGE_CACHE_SHIFT;
-	last_index = (cluster->end - offset) >> PAGE_CACHE_SHIFT;
+	ret = prealloc_file_extent_cluster(inode, cluster);
+	if (ret)
+		goto out;
 
-	mutex_lock(&inode->i_mutex);
+	file_ra_state_init(ra, inode->i_mapping);
 
-	i_size_write(inode, cluster->end + 1 - offset);
 	ret = setup_extent_mapping(inode, cluster->start - offset,
 				   cluster->end - offset, cluster->start);
 	if (ret)
-		goto out_unlock;
-
-	file_ra_state_init(ra, inode->i_mapping);
+		goto out;
 
-	WARN_ON(cluster->start != cluster->boundary[0]);
+	index = (cluster->start - offset) >> PAGE_CACHE_SHIFT;
+	last_index = (cluster->end - offset) >> PAGE_CACHE_SHIFT;
 	while (index <= last_index) {
+		ret = btrfs_delalloc_reserve_metadata(inode, PAGE_CACHE_SIZE);
+		if (ret)
+			goto out;
+
 		page = find_lock_page(inode->i_mapping, index);
 		if (!page) {
 			page_cache_sync_readahead(inode->i_mapping,
@@ -2623,8 +2931,10 @@ static int relocate_file_extent_cluster(struct inode *inode,
 						  last_index + 1 - index);
 			page = grab_cache_page(inode->i_mapping, index);
 			if (!page) {
+				btrfs_delalloc_release_metadata(inode,
+							PAGE_CACHE_SIZE);
 				ret = -ENOMEM;
-				goto out_unlock;
+				goto out;
 			}
 		}
 
@@ -2640,8 +2950,10 @@ static int relocate_file_extent_cluster(struct inode *inode,
 			if (!PageUptodate(page)) {
 				unlock_page(page);
 				page_cache_release(page);
+				btrfs_delalloc_release_metadata(inode,
+							PAGE_CACHE_SIZE);
 				ret = -EIO;
-				goto out_unlock;
+				goto out;
 			}
 		}
 
@@ -2660,10 +2972,9 @@ static int relocate_file_extent_cluster(struct inode *inode,
 					EXTENT_BOUNDARY, GFP_NOFS);
 			nr++;
 		}
-		btrfs_set_extent_delalloc(inode, page_start, page_end, NULL);
 
+		btrfs_set_extent_delalloc(inode, page_start, page_end, NULL);
 		set_page_dirty(page);
-		dirty_page++;
 
 		unlock_extent(&BTRFS_I(inode)->io_tree,
 			      page_start, page_end, GFP_NOFS);
@@ -2671,20 +2982,11 @@ static int relocate_file_extent_cluster(struct inode *inode,
 		page_cache_release(page);
 
 		index++;
-		if (nr < cluster->nr &&
-		    page_end + 1 + offset == cluster->boundary[nr]) {
-			balance_dirty_pages_ratelimited_nr(inode->i_mapping,
-							   dirty_page);
-			dirty_page = 0;
-		}
-	}
-	if (dirty_page) {
-		balance_dirty_pages_ratelimited_nr(inode->i_mapping,
-						   dirty_page);
+		balance_dirty_pages_ratelimited(inode->i_mapping);
+		btrfs_throttle(BTRFS_I(inode)->root);
 	}
 	WARN_ON(nr != cluster->nr);
-out_unlock:
-	mutex_unlock(&inode->i_mutex);
+out:
 	kfree(ra);
 	return ret;
 }
@@ -2870,9 +3172,6 @@ out:
 static int block_use_full_backref(struct reloc_control *rc,
 				  struct extent_buffer *eb)
 {
-	struct btrfs_path *path;
-	struct btrfs_extent_item *ei;
-	struct btrfs_key key;
 	u64 flags;
 	int ret;
 
@@ -2880,28 +3179,14 @@ static int block_use_full_backref(struct reloc_control *rc,
 	    btrfs_header_backref_rev(eb) < BTRFS_MIXED_BACKREF_REV)
 		return 1;
 
-	path = btrfs_alloc_path();
-	BUG_ON(!path);
-
-	key.objectid = eb->start;
-	key.type = BTRFS_EXTENT_ITEM_KEY;
-	key.offset = eb->len;
-
-	path->search_commit_root = 1;
-	path->skip_locking = 1;
-	ret = btrfs_search_slot(NULL, rc->extent_root,
-				&key, path, 0, 0);
+	ret = btrfs_lookup_extent_info(NULL, rc->extent_root,
+				       eb->start, eb->len, NULL, &flags);
 	BUG_ON(ret);
 
-	ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
-			    struct btrfs_extent_item);
-	flags = btrfs_extent_flags(path->nodes[0], ei);
-	BUG_ON(!(flags & BTRFS_EXTENT_FLAG_TREE_BLOCK));
 	if (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)
 		ret = 1;
 	else
 		ret = 0;
-	btrfs_free_path(path);
 	return ret;
 }
 
@@ -3074,22 +3359,10 @@ int add_data_references(struct reloc_control *rc,
 	struct btrfs_extent_inline_ref *iref;
 	unsigned long ptr;
 	unsigned long end;
-	u32 blocksize;
+	u32 blocksize = btrfs_level_size(rc->extent_root, 0);
 	int ret;
 	int err = 0;
 
-	ret = get_new_location(rc->data_inode, NULL, extent_key->objectid,
-			       extent_key->offset);
-	BUG_ON(ret < 0);
-	if (ret > 0) {
-		/* the relocated data is fragmented */
-		rc->extents_skipped++;
-		btrfs_release_path(rc->extent_root, path);
-		return 0;
-	}
-
-	blocksize = btrfs_level_size(rc->extent_root, 0);
-
 	eb = path->nodes[0];
 	ptr = btrfs_item_ptr_offset(eb, path->slots[0]);
 	end = ptr + btrfs_item_size_nr(eb, path->slots[0]);
@@ -3170,7 +3443,8 @@ int add_data_references(struct reloc_control *rc,
  */
 static noinline_for_stack
 int find_next_extent(struct btrfs_trans_handle *trans,
-		     struct reloc_control *rc, struct btrfs_path *path)
+		     struct reloc_control *rc, struct btrfs_path *path,
+		     struct btrfs_key *extent_key)
 {
 	struct btrfs_key key;
 	struct extent_buffer *leaf;
@@ -3225,6 +3499,7 @@ next:
 			rc->search_start = end + 1;
 		} else {
 			rc->search_start = key.objectid + key.offset;
+			memcpy(extent_key, &key, sizeof(key));
 			return 0;
 		}
 	}
@@ -3262,12 +3537,49 @@ static int check_extent_flags(u64 flags)
 	return 0;
 }
 
+static noinline_for_stack
+int prepare_to_relocate(struct reloc_control *rc)
+{
+	struct btrfs_trans_handle *trans;
+	int ret;
+
+	rc->block_rsv = btrfs_alloc_block_rsv(rc->extent_root);
+	if (!rc->block_rsv)
+		return -ENOMEM;
+
+	/*
+	 * reserve some space for creating reloc trees.
+	 * btrfs_init_reloc_root will use them when there
+	 * is no reservation in transaction handle.
+	 */
+	ret = btrfs_block_rsv_add(NULL, rc->extent_root, rc->block_rsv,
+				  rc->extent_root->nodesize * 256,
+				  &rc->block_rsv_retries);
+	if (ret)
+		return ret;
+
+	rc->block_rsv->refill_used = 1;
+	btrfs_add_durable_block_rsv(rc->extent_root->fs_info, rc->block_rsv);
+
+	memset(&rc->cluster, 0, sizeof(rc->cluster));
+	rc->search_start = rc->block_group->key.objectid;
+	rc->extents_found = 0;
+	rc->nodes_relocated = 0;
+	rc->merging_rsv_size = 0;
+	rc->block_rsv_retries = 0;
+
+	rc->create_reloc_tree = 1;
+	set_reloc_control(rc);
+
+	trans = btrfs_join_transaction(rc->extent_root, 1);
+	btrfs_commit_transaction(trans, rc->extent_root);
+	return 0;
+}
 
 static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 {
 	struct rb_root blocks = RB_ROOT;
 	struct btrfs_key key;
-	struct file_extent_cluster *cluster;
 	struct btrfs_trans_handle *trans = NULL;
 	struct btrfs_path *path;
 	struct btrfs_extent_item *ei;
@@ -3277,33 +3589,25 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 	int ret;
 	int err = 0;
 
-	cluster = kzalloc(sizeof(*cluster), GFP_NOFS);
-	if (!cluster)
-		return -ENOMEM;
-
 	path = btrfs_alloc_path();
-	if (!path) {
-		kfree(cluster);
+	if (!path)
 		return -ENOMEM;
-	}
-
-	rc->extents_found = 0;
-	rc->extents_skipped = 0;
-
-	rc->search_start = rc->block_group->key.objectid;
-	clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY,
-			  GFP_NOFS);
-
-	rc->create_reloc_root = 1;
-	set_reloc_control(rc);
 
-	trans = btrfs_start_transaction(rc->extent_root, 1);
-	btrfs_commit_transaction(trans, rc->extent_root);
+	ret = prepare_to_relocate(rc);
+	if (ret) {
+		err = ret;
+		goto out_free;
+	}
 
 	while (1) {
-		trans = btrfs_start_transaction(rc->extent_root, 1);
+		trans = btrfs_start_transaction(rc->extent_root, 0);
+
+		if (update_backref_cache(trans, &rc->backref_cache)) {
+			btrfs_end_transaction(trans, rc->extent_root);
+			continue;
+		}
 
-		ret = find_next_extent(trans, rc, path);
+		ret = find_next_extent(trans, rc, path, &key);
 		if (ret < 0)
 			err = ret;
 		if (ret != 0)
@@ -3313,9 +3617,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 
 		ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
 				    struct btrfs_extent_item);
-		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
-		item_size = btrfs_item_size_nr(path->nodes[0],
-					       path->slots[0]);
+		item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
 		if (item_size >= sizeof(*ei)) {
 			flags = btrfs_extent_flags(path->nodes[0], ei);
 			ret = check_extent_flags(flags);
@@ -3356,73 +3658,100 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 		if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
 			ret = add_tree_block(rc, &key, path, &blocks);
 		} else if (rc->stage == UPDATE_DATA_PTRS &&
-			 (flags & BTRFS_EXTENT_FLAG_DATA)) {
+			   (flags & BTRFS_EXTENT_FLAG_DATA)) {
 			ret = add_data_references(rc, &key, path, &blocks);
 		} else {
 			btrfs_release_path(rc->extent_root, path);
 			ret = 0;
 		}
 		if (ret < 0) {
-			err = 0;
+			err = ret;
 			break;
 		}
 
 		if (!RB_EMPTY_ROOT(&blocks)) {
 			ret = relocate_tree_blocks(trans, rc, &blocks);
 			if (ret < 0) {
+				if (ret != -EAGAIN) {
+					err = ret;
+					break;
+				}
+				rc->extents_found--;
+				rc->search_start = key.objectid;
+			}
+		}
+
+		ret = btrfs_block_rsv_check(trans, rc->extent_root,
+					    rc->block_rsv, 0, 5);
+		if (ret < 0) {
+			if (ret != -EAGAIN) {
 				err = ret;
+				WARN_ON(1);
 				break;
 			}
+			rc->commit_transaction = 1;
 		}
 
-		nr = trans->blocks_used;
-		btrfs_end_transaction(trans, rc->extent_root);
+		if (rc->commit_transaction) {
+			rc->commit_transaction = 0;
+			ret = btrfs_commit_transaction(trans, rc->extent_root);
+			BUG_ON(ret);
+		} else {
+			nr = trans->blocks_used;
+			btrfs_end_transaction_throttle(trans, rc->extent_root);
+			btrfs_btree_balance_dirty(rc->extent_root, nr);
+		}
 		trans = NULL;
-		btrfs_btree_balance_dirty(rc->extent_root, nr);
 
 		if (rc->stage == MOVE_DATA_EXTENTS &&
 		    (flags & BTRFS_EXTENT_FLAG_DATA)) {
 			rc->found_file_extent = 1;
 			ret = relocate_data_extent(rc->data_inode,
-						   &key, cluster);
+						   &key, &rc->cluster);
 			if (ret < 0) {
 				err = ret;
 				break;
 			}
 		}
 	}
-	btrfs_free_path(path);
+
+	btrfs_release_path(rc->extent_root, path);
+	clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY,
+			  GFP_NOFS);
 
 	if (trans) {
 		nr = trans->blocks_used;
-		btrfs_end_transaction(trans, rc->extent_root);
+		btrfs_end_transaction_throttle(trans, rc->extent_root);
 		btrfs_btree_balance_dirty(rc->extent_root, nr);
 	}
 
 	if (!err) {
-		ret = relocate_file_extent_cluster(rc->data_inode, cluster);
+		ret = relocate_file_extent_cluster(rc->data_inode,
+						   &rc->cluster);
 		if (ret < 0)
 			err = ret;
 	}
 
-	kfree(cluster);
+	rc->create_reloc_tree = 0;
+	set_reloc_control(rc);
 
-	rc->create_reloc_root = 0;
-	smp_mb();
+	backref_cache_cleanup(&rc->backref_cache);
+	btrfs_block_rsv_release(rc->extent_root, rc->block_rsv, (u64)-1);
 
-	if (rc->extents_found > 0) {
-		trans = btrfs_start_transaction(rc->extent_root, 1);
-		btrfs_commit_transaction(trans, rc->extent_root);
-	}
+	err = prepare_to_merge(rc, err);
 
 	merge_reloc_roots(rc);
 
+	rc->merge_reloc_tree = 0;
 	unset_reloc_control(rc);
+	btrfs_block_rsv_release(rc->extent_root, rc->block_rsv, (u64)-1);
 
 	/* get rid of pinned extents */
-	trans = btrfs_start_transaction(rc->extent_root, 1);
+	trans = btrfs_join_transaction(rc->extent_root, 1);
 	btrfs_commit_transaction(trans, rc->extent_root);
-
+out_free:
+	btrfs_free_block_rsv(rc->extent_root, rc->block_rsv);
+	btrfs_free_path(path);
 	return err;
 }
 
@@ -3448,7 +3777,8 @@ static int __insert_orphan_inode(struct btrfs_trans_handle *trans,
 	btrfs_set_inode_generation(leaf, item, 1);
 	btrfs_set_inode_size(leaf, item, 0);
 	btrfs_set_inode_mode(leaf, item, S_IFREG | 0600);
-	btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS);
+	btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS |
+					  BTRFS_INODE_PREALLOC);
 	btrfs_mark_buffer_dirty(leaf);
 	btrfs_release_path(root, path);
 out:
@@ -3460,8 +3790,9 @@ out:
  * helper to create inode for data relocation.
  * the inode is in data relocation tree and its link count is 0
  */
-static struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
-					struct btrfs_block_group_cache *group)
+static noinline_for_stack
+struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
+				 struct btrfs_block_group_cache *group)
 {
 	struct inode *inode = NULL;
 	struct btrfs_trans_handle *trans;
@@ -3475,8 +3806,9 @@ static struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
 	if (IS_ERR(root))
 		return ERR_CAST(root);
 
-	trans = btrfs_start_transaction(root, 1);
-	BUG_ON(!trans);
+	trans = btrfs_start_transaction(root, 6);
+	if (IS_ERR(trans))
+		return ERR_CAST(trans);
 
 	err = btrfs_find_free_objectid(trans, root, objectid, &objectid);
 	if (err)
@@ -3496,7 +3828,6 @@ static struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
 out:
 	nr = trans->blocks_used;
 	btrfs_end_transaction(trans, root);
-
 	btrfs_btree_balance_dirty(root, nr);
 	if (err) {
 		if (inode)
@@ -3506,6 +3837,21 @@ out:
 	return inode;
 }
 
+static struct reloc_control *alloc_reloc_control(void)
+{
+	struct reloc_control *rc;
+
+	rc = kzalloc(sizeof(*rc), GFP_NOFS);
+	if (!rc)
+		return NULL;
+
+	INIT_LIST_HEAD(&rc->reloc_roots);
+	backref_cache_init(&rc->backref_cache);
+	mapping_tree_init(&rc->reloc_root_tree);
+	extent_io_tree_init(&rc->processed_blocks, NULL, GFP_NOFS);
+	return rc;
+}
+
 /*
  * function to relocate all extents in a block group.
  */
@@ -3514,24 +3860,26 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 	struct btrfs_fs_info *fs_info = extent_root->fs_info;
 	struct reloc_control *rc;
 	int ret;
+	int rw = 0;
 	int err = 0;
 
-	rc = kzalloc(sizeof(*rc), GFP_NOFS);
+	rc = alloc_reloc_control();
 	if (!rc)
 		return -ENOMEM;
 
-	mapping_tree_init(&rc->reloc_root_tree);
-	extent_io_tree_init(&rc->processed_blocks, NULL, GFP_NOFS);
-	INIT_LIST_HEAD(&rc->reloc_roots);
+	rc->extent_root = extent_root;
 
 	rc->block_group = btrfs_lookup_block_group(fs_info, group_start);
 	BUG_ON(!rc->block_group);
 
-	btrfs_init_workers(&rc->workers, "relocate",
-			   fs_info->thread_pool_size, NULL);
-
-	rc->extent_root = extent_root;
-	btrfs_prepare_block_group_relocation(extent_root, rc->block_group);
+	if (!rc->block_group->ro) {
+		ret = btrfs_set_block_group_ro(extent_root, rc->block_group);
+		if (ret) {
+			err = ret;
+			goto out;
+		}
+		rw = 1;
+	}
 
 	rc->data_inode = create_reloc_inode(fs_info, rc->block_group);
 	if (IS_ERR(rc->data_inode)) {
@@ -3548,9 +3896,6 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 	btrfs_wait_ordered_extents(fs_info->tree_root, 0, 0);
 
 	while (1) {
-		rc->extents_found = 0;
-		rc->extents_skipped = 0;
-
 		mutex_lock(&fs_info->cleaner_mutex);
 
 		btrfs_clean_old_snapshots(fs_info->tree_root);
@@ -3559,7 +3904,7 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 		mutex_unlock(&fs_info->cleaner_mutex);
 		if (ret < 0) {
 			err = ret;
-			break;
+			goto out;
 		}
 
 		if (rc->extents_found == 0)
@@ -3573,18 +3918,6 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 			invalidate_mapping_pages(rc->data_inode->i_mapping,
 						 0, -1);
 			rc->stage = UPDATE_DATA_PTRS;
-		} else if (rc->stage == UPDATE_DATA_PTRS &&
-			   rc->extents_skipped >= rc->extents_found) {
-			iput(rc->data_inode);
-			rc->data_inode = create_reloc_inode(fs_info,
-							    rc->block_group);
-			if (IS_ERR(rc->data_inode)) {
-				err = PTR_ERR(rc->data_inode);
-				rc->data_inode = NULL;
-				break;
-			}
-			rc->stage = MOVE_DATA_EXTENTS;
-			rc->found_file_extent = 0;
 		}
 	}
 
@@ -3597,8 +3930,9 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 	WARN_ON(rc->block_group->reserved > 0);
 	WARN_ON(btrfs_block_group_used(&rc->block_group->item) > 0);
 out:
+	if (err && rw)
+		btrfs_set_block_group_rw(extent_root, rc->block_group);
 	iput(rc->data_inode);
-	btrfs_stop_workers(&rc->workers);
 	btrfs_put_block_group(rc->block_group);
 	kfree(rc);
 	return err;
@@ -3609,7 +3943,7 @@ static noinline_for_stack int mark_garbage_root(struct btrfs_root *root)
 	struct btrfs_trans_handle *trans;
 	int ret;
 
-	trans = btrfs_start_transaction(root->fs_info->tree_root, 1);
+	trans = btrfs_start_transaction(root->fs_info->tree_root, 0);
 
 	memset(&root->root_item.drop_progress, 0,
 		sizeof(root->root_item.drop_progress));
@@ -3702,20 +4036,20 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 	if (list_empty(&reloc_roots))
 		goto out;
 
-	rc = kzalloc(sizeof(*rc), GFP_NOFS);
+	rc = alloc_reloc_control();
 	if (!rc) {
 		err = -ENOMEM;
 		goto out;
 	}
 
-	mapping_tree_init(&rc->reloc_root_tree);
-	INIT_LIST_HEAD(&rc->reloc_roots);
-	btrfs_init_workers(&rc->workers, "relocate",
-			   root->fs_info->thread_pool_size, NULL);
 	rc->extent_root = root->fs_info->extent_root;
 
 	set_reloc_control(rc);
 
+	trans = btrfs_join_transaction(rc->extent_root, 1);
+
+	rc->merge_reloc_tree = 1;
+
 	while (!list_empty(&reloc_roots)) {
 		reloc_root = list_entry(reloc_roots.next,
 					struct btrfs_root, root_list);
@@ -3735,20 +4069,16 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 		fs_root->reloc_root = reloc_root;
 	}
 
-	trans = btrfs_start_transaction(rc->extent_root, 1);
 	btrfs_commit_transaction(trans, rc->extent_root);
 
 	merge_reloc_roots(rc);
 
 	unset_reloc_control(rc);
 
-	trans = btrfs_start_transaction(rc->extent_root, 1);
+	trans = btrfs_join_transaction(rc->extent_root, 1);
 	btrfs_commit_transaction(trans, rc->extent_root);
 out:
-	if (rc) {
-		btrfs_stop_workers(&rc->workers);
-		kfree(rc);
-	}
+	kfree(rc);
 	while (!list_empty(&reloc_roots)) {
 		reloc_root = list_entry(reloc_roots.next,
 					struct btrfs_root, root_list);
@@ -3814,3 +4144,130 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 	btrfs_put_ordered_extent(ordered);
 	return 0;
 }
+
+void btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, struct extent_buffer *buf,
+			   struct extent_buffer *cow)
+{
+	struct reloc_control *rc;
+	struct backref_node *node;
+	int first_cow = 0;
+	int level;
+	int ret;
+
+	rc = root->fs_info->reloc_ctl;
+	if (!rc)
+		return;
+
+	BUG_ON(rc->stage == UPDATE_DATA_PTRS &&
+	       root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID);
+
+	level = btrfs_header_level(buf);
+	if (btrfs_header_generation(buf) <=
+	    btrfs_root_last_snapshot(&root->root_item))
+		first_cow = 1;
+
+	if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID &&
+	    rc->create_reloc_tree) {
+		WARN_ON(!first_cow && level == 0);
+
+		node = rc->backref_cache.path[level];
+		BUG_ON(node->bytenr != buf->start &&
+		       node->new_bytenr != buf->start);
+
+		drop_node_buffer(node);
+		extent_buffer_get(cow);
+		node->eb = cow;
+		node->new_bytenr = cow->start;
+
+		if (!node->pending) {
+			list_move_tail(&node->list,
+				       &rc->backref_cache.pending[level]);
+			node->pending = 1;
+		}
+
+		if (first_cow)
+			__mark_block_processed(rc, node);
+
+		if (first_cow && level > 0)
+			rc->nodes_relocated += buf->len;
+	}
+
+	if (level == 0 && first_cow && rc->stage == UPDATE_DATA_PTRS) {
+		ret = replace_file_extents(trans, rc, root, cow);
+		BUG_ON(ret);
+	}
+}
+
+/*
+ * called before creating snapshot. it calculates metadata reservation
+ * requried for relocating tree blocks in the snapshot
+ */
+void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans,
+			      struct btrfs_pending_snapshot *pending,
+			      u64 *bytes_to_reserve)
+{
+	struct btrfs_root *root;
+	struct reloc_control *rc;
+
+	root = pending->root;
+	if (!root->reloc_root)
+		return;
+
+	rc = root->fs_info->reloc_ctl;
+	if (!rc->merge_reloc_tree)
+		return;
+
+	root = root->reloc_root;
+	BUG_ON(btrfs_root_refs(&root->root_item) == 0);
+	/*
+	 * relocation is in the stage of merging trees. the space
+	 * used by merging a reloc tree is twice the size of
+	 * relocated tree nodes in the worst case. half for cowing
+	 * the reloc tree, half for cowing the fs tree. the space
+	 * used by cowing the reloc tree will be freed after the
+	 * tree is dropped. if we create snapshot, cowing the fs
+	 * tree may use more space than it frees. so we need
+	 * reserve extra space.
+	 */
+	*bytes_to_reserve += rc->nodes_relocated;
+}
+
+/*
+ * called after snapshot is created. migrate block reservation
+ * and create reloc root for the newly created snapshot
+ */
+void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
+			       struct btrfs_pending_snapshot *pending)
+{
+	struct btrfs_root *root = pending->root;
+	struct btrfs_root *reloc_root;
+	struct btrfs_root *new_root;
+	struct reloc_control *rc;
+	int ret;
+
+	if (!root->reloc_root)
+		return;
+
+	rc = root->fs_info->reloc_ctl;
+	rc->merging_rsv_size += rc->nodes_relocated;
+
+	if (rc->merge_reloc_tree) {
+		ret = btrfs_block_rsv_migrate(&pending->block_rsv,
+					      rc->block_rsv,
+					      rc->nodes_relocated);
+		BUG_ON(ret);
+	}
+
+	new_root = pending->snap;
+	reloc_root = create_reloc_root(trans, root->reloc_root,
+				       new_root->root_key.objectid);
+
+	__add_reloc_root(reloc_root);
+	new_root->reloc_root = reloc_root;
+
+	if (rc->create_reloc_tree) {
+		ret = clone_backref_node(trans, rc, root, reloc_root);
+		BUG_ON(ret);
+	}
+}
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index 67fa2d29d663..b91ccd972644 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -259,6 +259,8 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 	struct extent_buffer *leaf;
 	struct btrfs_path *path;
 	struct btrfs_key key;
+	struct btrfs_key root_key;
+	struct btrfs_root *root;
 	int err = 0;
 	int ret;
 
@@ -270,6 +272,9 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 	key.type = BTRFS_ORPHAN_ITEM_KEY;
 	key.offset = 0;
 
+	root_key.type = BTRFS_ROOT_ITEM_KEY;
+	root_key.offset = (u64)-1;
+
 	while (1) {
 		ret = btrfs_search_slot(NULL, tree_root, &key, path, 0, 0);
 		if (ret < 0) {
@@ -294,13 +299,25 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 		    key.type != BTRFS_ORPHAN_ITEM_KEY)
 			break;
 
-		ret = btrfs_find_dead_roots(tree_root, key.offset);
-		if (ret) {
+		root_key.objectid = key.offset;
+		key.offset++;
+
+		root = btrfs_read_fs_root_no_name(tree_root->fs_info,
+						  &root_key);
+		if (!IS_ERR(root))
+			continue;
+
+		ret = PTR_ERR(root);
+		if (ret != -ENOENT) {
 			err = ret;
 			break;
 		}
 
-		key.offset++;
+		ret = btrfs_find_dead_roots(tree_root, root_key.objectid);
+		if (ret) {
+			err = ret;
+			break;
+		}
 	}
 
 	btrfs_free_path(path);
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 1866dff0538e..d34b2dfc9628 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -498,7 +498,7 @@ int btrfs_sync_fs(struct super_block *sb, int wait)
 	btrfs_start_delalloc_inodes(root, 0);
 	btrfs_wait_ordered_extents(root, 0, 0);
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_start_transaction(root, 0);
 	ret = btrfs_commit_transaction(trans, root);
 	return ret;
 }
@@ -694,11 +694,11 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 		if (btrfs_super_log_root(&root->fs_info->super_copy) != 0)
 			return -EINVAL;
 
-		/* recover relocation */
-		ret = btrfs_recover_relocation(root);
+		ret = btrfs_cleanup_fs_roots(root->fs_info);
 		WARN_ON(ret);
 
-		ret = btrfs_cleanup_fs_roots(root->fs_info);
+		/* recover relocation */
+		ret = btrfs_recover_relocation(root);
 		WARN_ON(ret);
 
 		sb->s_flags &= ~MS_RDONLY;
@@ -714,34 +714,18 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	struct list_head *head = &root->fs_info->space_info;
 	struct btrfs_space_info *found;
 	u64 total_used = 0;
-	u64 data_used = 0;
 	int bits = dentry->d_sb->s_blocksize_bits;
 	__be32 *fsid = (__be32 *)root->fs_info->fsid;
 
 	rcu_read_lock();
-	list_for_each_entry_rcu(found, head, list) {
-		if (found->flags & (BTRFS_BLOCK_GROUP_DUP|
-				    BTRFS_BLOCK_GROUP_RAID10|
-				    BTRFS_BLOCK_GROUP_RAID1)) {
-			total_used += found->bytes_used;
-			if (found->flags & BTRFS_BLOCK_GROUP_DATA)
-				data_used += found->bytes_used;
-			else
-				data_used += found->total_bytes;
-		}
-
-		total_used += found->bytes_used;
-		if (found->flags & BTRFS_BLOCK_GROUP_DATA)
-			data_used += found->bytes_used;
-		else
-			data_used += found->total_bytes;
-	}
+	list_for_each_entry_rcu(found, head, list)
+		total_used += found->disk_used;
 	rcu_read_unlock();
 
 	buf->f_namelen = BTRFS_NAME_LEN;
 	buf->f_blocks = btrfs_super_total_bytes(disk_super) >> bits;
 	buf->f_bfree = buf->f_blocks - (total_used >> bits);
-	buf->f_bavail = buf->f_blocks - (data_used >> bits);
+	buf->f_bavail = buf->f_bfree;
 	buf->f_bsize = dentry->d_sb->s_blocksize;
 	buf->f_type = BTRFS_SUPER_MAGIC;
 
@@ -832,11 +816,14 @@ static const struct file_operations btrfs_ctl_fops = {
 };
 
 static struct miscdevice btrfs_misc = {
-	.minor		= MISC_DYNAMIC_MINOR,
+	.minor		= BTRFS_MINOR,
 	.name		= "btrfs-control",
 	.fops		= &btrfs_ctl_fops
 };
 
+MODULE_ALIAS_MISCDEV(BTRFS_MINOR);
+MODULE_ALIAS("devname:btrfs-control");
+
 static int btrfs_interface_init(void)
 {
 	return misc_register(&btrfs_misc);
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 2cb116099b90..66e4c66cc63b 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -165,54 +165,89 @@ enum btrfs_trans_type {
 	TRANS_USERSPACE,
 };
 
+static int may_wait_transaction(struct btrfs_root *root, int type)
+{
+	if (!root->fs_info->log_root_recovering &&
+	    ((type == TRANS_START && !root->fs_info->open_ioctl_trans) ||
+	     type == TRANS_USERSPACE))
+		return 1;
+	return 0;
+}
+
 static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
-					     int num_blocks, int type)
+						    u64 num_items, int type)
 {
-	struct btrfs_trans_handle *h =
-		kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
+	struct btrfs_trans_handle *h;
+	struct btrfs_transaction *cur_trans;
+	int retries = 0;
 	int ret;
+again:
+	h = kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
+	if (!h)
+		return ERR_PTR(-ENOMEM);
 
 	mutex_lock(&root->fs_info->trans_mutex);
-	if (!root->fs_info->log_root_recovering &&
-	    ((type == TRANS_START && !root->fs_info->open_ioctl_trans) ||
-	     type == TRANS_USERSPACE))
+	if (may_wait_transaction(root, type))
 		wait_current_trans(root);
+
 	ret = join_transaction(root);
 	BUG_ON(ret);
 
-	h->transid = root->fs_info->running_transaction->transid;
-	h->transaction = root->fs_info->running_transaction;
-	h->blocks_reserved = num_blocks;
+	cur_trans = root->fs_info->running_transaction;
+	cur_trans->use_count++;
+	mutex_unlock(&root->fs_info->trans_mutex);
+
+	h->transid = cur_trans->transid;
+	h->transaction = cur_trans;
 	h->blocks_used = 0;
 	h->block_group = 0;
-	h->alloc_exclude_nr = 0;
-	h->alloc_exclude_start = 0;
+	h->bytes_reserved = 0;
 	h->delayed_ref_updates = 0;
+	h->block_rsv = NULL;
 
-	if (!current->journal_info && type != TRANS_USERSPACE)
-		current->journal_info = h;
+	smp_mb();
+	if (cur_trans->blocked && may_wait_transaction(root, type)) {
+		btrfs_commit_transaction(h, root);
+		goto again;
+	}
+
+	if (num_items > 0) {
+		ret = btrfs_trans_reserve_metadata(h, root, num_items,
+						   &retries);
+		if (ret == -EAGAIN) {
+			btrfs_commit_transaction(h, root);
+			goto again;
+		}
+		if (ret < 0) {
+			btrfs_end_transaction(h, root);
+			return ERR_PTR(ret);
+		}
+	}
 
-	root->fs_info->running_transaction->use_count++;
+	mutex_lock(&root->fs_info->trans_mutex);
 	record_root_in_trans(h, root);
 	mutex_unlock(&root->fs_info->trans_mutex);
+
+	if (!current->journal_info && type != TRANS_USERSPACE)
+		current->journal_info = h;
 	return h;
 }
 
 struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
-						   int num_blocks)
+						   int num_items)
 {
-	return start_transaction(root, num_blocks, TRANS_START);
+	return start_transaction(root, num_items, TRANS_START);
 }
 struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
 						   int num_blocks)
 {
-	return start_transaction(root, num_blocks, TRANS_JOIN);
+	return start_transaction(root, 0, TRANS_JOIN);
 }
 
 struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
 							 int num_blocks)
 {
-	return start_transaction(r, num_blocks, TRANS_USERSPACE);
+	return start_transaction(r, 0, TRANS_USERSPACE);
 }
 
 /* wait for a transaction commit to be fully complete */
@@ -286,10 +321,36 @@ void btrfs_throttle(struct btrfs_root *root)
 	mutex_unlock(&root->fs_info->trans_mutex);
 }
 
+static int should_end_transaction(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root)
+{
+	int ret;
+	ret = btrfs_block_rsv_check(trans, root,
+				    &root->fs_info->global_block_rsv, 0, 5);
+	return ret ? 1 : 0;
+}
+
+int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root)
+{
+	struct btrfs_transaction *cur_trans = trans->transaction;
+	int updates;
+
+	if (cur_trans->blocked || cur_trans->delayed_refs.flushing)
+		return 1;
+
+	updates = trans->delayed_ref_updates;
+	trans->delayed_ref_updates = 0;
+	if (updates)
+		btrfs_run_delayed_refs(trans, root, updates);
+
+	return should_end_transaction(trans, root);
+}
+
 static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root, int throttle)
 {
-	struct btrfs_transaction *cur_trans;
+	struct btrfs_transaction *cur_trans = trans->transaction;
 	struct btrfs_fs_info *info = root->fs_info;
 	int count = 0;
 
@@ -313,9 +374,21 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 		count++;
 	}
 
+	btrfs_trans_release_metadata(trans, root);
+
+	if (!root->fs_info->open_ioctl_trans &&
+	    should_end_transaction(trans, root))
+		trans->transaction->blocked = 1;
+
+	if (cur_trans->blocked && !cur_trans->in_commit) {
+		if (throttle)
+			return btrfs_commit_transaction(trans, root);
+		else
+			wake_up_process(info->transaction_kthread);
+	}
+
 	mutex_lock(&info->trans_mutex);
-	cur_trans = info->running_transaction;
-	WARN_ON(cur_trans != trans->transaction);
+	WARN_ON(cur_trans != info->running_transaction);
 	WARN_ON(cur_trans->num_writers < 1);
 	cur_trans->num_writers--;
 
@@ -603,6 +676,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 
 			btrfs_free_log(trans, root);
 			btrfs_update_reloc_root(trans, root);
+			btrfs_orphan_commit_root(trans, root);
 
 			if (root->commit_root != root->node) {
 				switch_commit_root(root);
@@ -627,30 +701,30 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 int btrfs_defrag_root(struct btrfs_root *root, int cacheonly)
 {
 	struct btrfs_fs_info *info = root->fs_info;
-	int ret;
 	struct btrfs_trans_handle *trans;
+	int ret;
 	unsigned long nr;
 
-	smp_mb();
-	if (root->defrag_running)
+	if (xchg(&root->defrag_running, 1))
 		return 0;
-	trans = btrfs_start_transaction(root, 1);
+
 	while (1) {
-		root->defrag_running = 1;
+		trans = btrfs_start_transaction(root, 0);
+		if (IS_ERR(trans))
+			return PTR_ERR(trans);
+
 		ret = btrfs_defrag_leaves(trans, root, cacheonly);
+
 		nr = trans->blocks_used;
 		btrfs_end_transaction(trans, root);
 		btrfs_btree_balance_dirty(info->tree_root, nr);
 		cond_resched();
 
-		trans = btrfs_start_transaction(root, 1);
 		if (root->fs_info->closing || ret != -EAGAIN)
 			break;
 	}
 	root->defrag_running = 0;
-	smp_mb();
-	btrfs_end_transaction(trans, root);
-	return 0;
+	return ret;
 }
 
 #if 0
@@ -758,47 +832,63 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	struct btrfs_root *root = pending->root;
 	struct btrfs_root *parent_root;
 	struct inode *parent_inode;
+	struct dentry *dentry;
 	struct extent_buffer *tmp;
 	struct extent_buffer *old;
 	int ret;
-	u64 objectid;
-	int namelen;
+	int retries = 0;
+	u64 to_reserve = 0;
 	u64 index = 0;
-
-	parent_inode = pending->dentry->d_parent->d_inode;
-	parent_root = BTRFS_I(parent_inode)->root;
+	u64 objectid;
 
 	new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS);
 	if (!new_root_item) {
-		ret = -ENOMEM;
+		pending->error = -ENOMEM;
 		goto fail;
 	}
+
 	ret = btrfs_find_free_objectid(trans, tree_root, 0, &objectid);
-	if (ret)
+	if (ret) {
+		pending->error = ret;
 		goto fail;
+	}
+
+	btrfs_reloc_pre_snapshot(trans, pending, &to_reserve);
+	btrfs_orphan_pre_snapshot(trans, pending, &to_reserve);
+
+	if (to_reserve > 0) {
+		ret = btrfs_block_rsv_add(trans, root, &pending->block_rsv,
+					  to_reserve, &retries);
+		if (ret) {
+			pending->error = ret;
+			goto fail;
+		}
+	}
 
 	key.objectid = objectid;
-	/* record when the snapshot was created in key.offset */
-	key.offset = trans->transid;
-	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+	key.offset = (u64)-1;
+	key.type = BTRFS_ROOT_ITEM_KEY;
 
-	memcpy(&pending->root_key, &key, sizeof(key));
-	pending->root_key.offset = (u64)-1;
+	trans->block_rsv = &pending->block_rsv;
 
+	dentry = pending->dentry;
+	parent_inode = dentry->d_parent->d_inode;
+	parent_root = BTRFS_I(parent_inode)->root;
 	record_root_in_trans(trans, parent_root);
+
 	/*
 	 * insert the directory item
 	 */
-	namelen = strlen(pending->name);
 	ret = btrfs_set_inode_index(parent_inode, &index);
 	BUG_ON(ret);
 	ret = btrfs_insert_dir_item(trans, parent_root,
-			    pending->name, namelen,
-			    parent_inode->i_ino,
-			    &pending->root_key, BTRFS_FT_DIR, index);
+				dentry->d_name.name, dentry->d_name.len,
+				parent_inode->i_ino, &key,
+				BTRFS_FT_DIR, index);
 	BUG_ON(ret);
 
-	btrfs_i_size_write(parent_inode, parent_inode->i_size + namelen * 2);
+	btrfs_i_size_write(parent_inode, parent_inode->i_size +
+					 dentry->d_name.len * 2);
 	ret = btrfs_update_inode(trans, parent_root, parent_inode);
 	BUG_ON(ret);
 
@@ -815,22 +905,32 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	free_extent_buffer(old);
 
 	btrfs_set_root_node(new_root_item, tmp);
-	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
-				new_root_item);
-	BUG_ON(ret);
+	/* record when the snapshot was created in key.offset */
+	key.offset = trans->transid;
+	ret = btrfs_insert_root(trans, tree_root, &key, new_root_item);
 	btrfs_tree_unlock(tmp);
 	free_extent_buffer(tmp);
+	BUG_ON(ret);
 
-	ret = btrfs_add_root_ref(trans, parent_root->fs_info->tree_root,
-				 pending->root_key.objectid,
+	/*
+	 * insert root back/forward references
+	 */
+	ret = btrfs_add_root_ref(trans, tree_root, objectid,
 				 parent_root->root_key.objectid,
-				 parent_inode->i_ino, index, pending->name,
-				 namelen);
+				 parent_inode->i_ino, index,
+				 dentry->d_name.name, dentry->d_name.len);
 	BUG_ON(ret);
 
+	key.offset = (u64)-1;
+	pending->snap = btrfs_read_fs_root_no_name(root->fs_info, &key);
+	BUG_ON(IS_ERR(pending->snap));
+
+	btrfs_reloc_post_snapshot(trans, pending);
+	btrfs_orphan_post_snapshot(trans, pending);
 fail:
 	kfree(new_root_item);
-	return ret;
+	btrfs_block_rsv_release(root, &pending->block_rsv, (u64)-1);
+	return 0;
 }
 
 /*
@@ -878,6 +978,16 @@ int btrfs_transaction_in_commit(struct btrfs_fs_info *info)
 	return ret;
 }
 
+int btrfs_transaction_blocked(struct btrfs_fs_info *info)
+{
+	int ret = 0;
+	spin_lock(&info->new_trans_lock);
+	if (info->running_transaction)
+		ret = info->running_transaction->blocked;
+	spin_unlock(&info->new_trans_lock);
+	return ret;
+}
+
 int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root)
 {
@@ -899,6 +1009,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	ret = btrfs_run_delayed_refs(trans, root, 0);
 	BUG_ON(ret);
 
+	btrfs_trans_release_metadata(trans, root);
+
 	cur_trans = trans->transaction;
 	/*
 	 * set the flushing flag so procs in this transaction have to
@@ -951,9 +1063,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 			snap_pending = 1;
 
 		WARN_ON(cur_trans != trans->transaction);
-		prepare_to_wait(&cur_trans->writer_wait, &wait,
-				TASK_UNINTERRUPTIBLE);
-
 		if (cur_trans->num_writers > 1)
 			timeout = MAX_SCHEDULE_TIMEOUT;
 		else if (should_grow)
@@ -976,6 +1085,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		 */
 		btrfs_run_ordered_operations(root, 1);
 
+		prepare_to_wait(&cur_trans->writer_wait, &wait,
+				TASK_UNINTERRUPTIBLE);
+
 		smp_mb();
 		if (cur_trans->num_writers > 1 || should_grow)
 			schedule_timeout(timeout);
@@ -1103,9 +1215,9 @@ int btrfs_clean_old_snapshots(struct btrfs_root *root)
 
 		if (btrfs_header_backref_rev(root->node) <
 		    BTRFS_MIXED_BACKREF_REV)
-			btrfs_drop_snapshot(root, 0);
+			btrfs_drop_snapshot(root, NULL, 0);
 		else
-			btrfs_drop_snapshot(root, 1);
+			btrfs_drop_snapshot(root, NULL, 1);
 	}
 	return 0;
 }
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index 93c7ccb33118..e104986d0bfd 100644
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -45,20 +45,23 @@ struct btrfs_transaction {
 
 struct btrfs_trans_handle {
 	u64 transid;
+	u64 block_group;
+	u64 bytes_reserved;
 	unsigned long blocks_reserved;
 	unsigned long blocks_used;
-	struct btrfs_transaction *transaction;
-	u64 block_group;
-	u64 alloc_exclude_start;
-	u64 alloc_exclude_nr;
 	unsigned long delayed_ref_updates;
+	struct btrfs_transaction *transaction;
+	struct btrfs_block_rsv *block_rsv;
 };
 
 struct btrfs_pending_snapshot {
 	struct dentry *dentry;
 	struct btrfs_root *root;
-	char *name;
-	struct btrfs_key root_key;
+	struct btrfs_root *snap;
+	/* block reservation for the operation */
+	struct btrfs_block_rsv block_rsv;
+	/* extra metadata reseration for relocation */
+	int error;
 	struct list_head list;
 };
 
@@ -85,11 +88,11 @@ static inline void btrfs_set_inode_last_trans(struct btrfs_trans_handle *trans,
 int btrfs_end_transaction(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root);
 struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
-						   int num_blocks);
+						   int num_items);
 struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
-						   int num_blocks);
+						  int num_blocks);
 struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
-						   int num_blocks);
+							 int num_blocks);
 int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root);
 int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
@@ -103,6 +106,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root);
 int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root);
+int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root);
 void btrfs_throttle(struct btrfs_root *root);
 int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root);
@@ -112,5 +117,6 @@ int btrfs_write_marked_extents(struct btrfs_root *root,
 				struct extent_io_tree *dirty_pages, int mark);
 int btrfs_wait_marked_extents(struct btrfs_root *root,
 				struct extent_io_tree *dirty_pages, int mark);
+int btrfs_transaction_blocked(struct btrfs_fs_info *info);
 int btrfs_transaction_in_commit(struct btrfs_fs_info *info);
 #endif
diff --git a/fs/btrfs/tree-defrag.c b/fs/btrfs/tree-defrag.c
index b10eacdb1620..f7ac8e013ed7 100644
--- a/fs/btrfs/tree-defrag.c
+++ b/fs/btrfs/tree-defrag.c
@@ -117,13 +117,14 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 				 path->nodes[1], 0,
 				 cache_only, &last_ret,
 				 &root->defrag_progress);
-	WARN_ON(ret && ret != -EAGAIN);
+	if (ret) {
+		WARN_ON(ret == -EAGAIN);
+		goto out;
+	}
 	if (next_key_ret == 0) {
 		memcpy(&root->defrag_progress, &key, sizeof(key));
 		ret = -EAGAIN;
 	}
-
-	btrfs_release_path(root, path);
 out:
 	if (path)
 		btrfs_free_path(path);
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index af57dd2b43d4..fb102a9aee9c 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -135,6 +135,7 @@ static int start_log_trans(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root)
 {
 	int ret;
+	int err = 0;
 
 	mutex_lock(&root->log_mutex);
 	if (root->log_root) {
@@ -155,17 +156,19 @@ static int start_log_trans(struct btrfs_trans_handle *trans,
 	mutex_lock(&root->fs_info->tree_log_mutex);
 	if (!root->fs_info->log_root_tree) {
 		ret = btrfs_init_log_root_tree(trans, root->fs_info);
-		BUG_ON(ret);
+		if (ret)
+			err = ret;
 	}
-	if (!root->log_root) {
+	if (err == 0 && !root->log_root) {
 		ret = btrfs_add_log_tree(trans, root);
-		BUG_ON(ret);
+		if (ret)
+			err = ret;
 	}
 	mutex_unlock(&root->fs_info->tree_log_mutex);
 	root->log_batch++;
 	atomic_inc(&root->log_writers);
 	mutex_unlock(&root->log_mutex);
-	return 0;
+	return err;
 }
 
 /*
@@ -376,7 +379,7 @@ insert:
 			BUG_ON(ret);
 		}
 	} else if (ret) {
-		BUG();
+		return ret;
 	}
 	dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
 					path->slots[0]);
@@ -1699,9 +1702,9 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 		next = btrfs_find_create_tree_block(root, bytenr, blocksize);
 
-		wc->process_func(root, next, wc, ptr_gen);
-
 		if (*level == 1) {
+			wc->process_func(root, next, wc, ptr_gen);
+
 			path->slots[*level]++;
 			if (wc->free) {
 				btrfs_read_buffer(next, ptr_gen);
@@ -1734,35 +1737,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 	WARN_ON(*level < 0);
 	WARN_ON(*level >= BTRFS_MAX_LEVEL);
 
-	if (path->nodes[*level] == root->node)
-		parent = path->nodes[*level];
-	else
-		parent = path->nodes[*level + 1];
-
-	bytenr = path->nodes[*level]->start;
-
-	blocksize = btrfs_level_size(root, *level);
-	root_owner = btrfs_header_owner(parent);
-	root_gen = btrfs_header_generation(parent);
-
-	wc->process_func(root, path->nodes[*level], wc,
-			 btrfs_header_generation(path->nodes[*level]));
-
-	if (wc->free) {
-		next = path->nodes[*level];
-		btrfs_tree_lock(next);
-		clean_tree_block(trans, root, next);
-		btrfs_set_lock_blocking(next);
-		btrfs_wait_tree_block_writeback(next);
-		btrfs_tree_unlock(next);
-
-		WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
-		ret = btrfs_free_reserved_extent(root, bytenr, blocksize);
-		BUG_ON(ret);
-	}
-	free_extent_buffer(path->nodes[*level]);
-	path->nodes[*level] = NULL;
-	*level += 1;
+	path->slots[*level] = btrfs_header_nritems(path->nodes[*level]);
 
 	cond_resched();
 	return 0;
@@ -1781,7 +1756,7 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
 
 	for (i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
 		slot = path->slots[i];
-		if (slot < btrfs_header_nritems(path->nodes[i]) - 1) {
+		if (slot + 1 < btrfs_header_nritems(path->nodes[i])) {
 			struct extent_buffer *node;
 			node = path->nodes[i];
 			path->slots[i]++;
@@ -2047,7 +2022,6 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	mutex_unlock(&log_root_tree->log_mutex);
 
 	ret = update_log_root(trans, log);
-	BUG_ON(ret);
 
 	mutex_lock(&log_root_tree->log_mutex);
 	if (atomic_dec_and_test(&log_root_tree->log_writers)) {
@@ -2056,6 +2030,15 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 			wake_up(&log_root_tree->log_writer_wait);
 	}
 
+	if (ret) {
+		BUG_ON(ret != -ENOSPC);
+		root->fs_info->last_trans_log_full_commit = trans->transid;
+		btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
+		mutex_unlock(&log_root_tree->log_mutex);
+		ret = -EAGAIN;
+		goto out;
+	}
+
 	index2 = log_root_tree->log_transid % 2;
 	if (atomic_read(&log_root_tree->log_commit[index2])) {
 		btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
@@ -2129,15 +2112,10 @@ out:
 	return 0;
 }
 
-/*
- * free all the extents used by the tree log.  This should be called
- * at commit time of the full transaction
- */
-int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
+static void free_log_tree(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *log)
 {
 	int ret;
-	struct btrfs_root *log;
-	struct key;
 	u64 start;
 	u64 end;
 	struct walk_control wc = {
@@ -2145,10 +2123,6 @@ int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
 		.process_func = process_one_buffer
 	};
 
-	if (!root->log_root || root->fs_info->log_root_recovering)
-		return 0;
-
-	log = root->log_root;
 	ret = walk_log_tree(trans, log, &wc);
 	BUG_ON(ret);
 
@@ -2162,14 +2136,30 @@ int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
 				  EXTENT_DIRTY | EXTENT_NEW, GFP_NOFS);
 	}
 
-	if (log->log_transid > 0) {
-		ret = btrfs_del_root(trans, root->fs_info->log_root_tree,
-				     &log->root_key);
-		BUG_ON(ret);
-	}
-	root->log_root = NULL;
 	free_extent_buffer(log->node);
 	kfree(log);
+}
+
+/*
+ * free all the extents used by the tree log.  This should be called
+ * at commit time of the full transaction
+ */
+int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
+{
+	if (root->log_root) {
+		free_log_tree(trans, root->log_root);
+		root->log_root = NULL;
+	}
+	return 0;
+}
+
+int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info)
+{
+	if (fs_info->log_root_tree) {
+		free_log_tree(trans, fs_info->log_root_tree);
+		fs_info->log_root_tree = NULL;
+	}
 	return 0;
 }
 
@@ -2203,6 +2193,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	struct btrfs_dir_item *di;
 	struct btrfs_path *path;
 	int ret;
+	int err = 0;
 	int bytes_del = 0;
 
 	if (BTRFS_I(dir)->logged_trans < trans->transid)
@@ -2218,7 +2209,11 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	path = btrfs_alloc_path();
 	di = btrfs_lookup_dir_item(trans, log, path, dir->i_ino,
 				   name, name_len, -1);
-	if (di && !IS_ERR(di)) {
+	if (IS_ERR(di)) {
+		err = PTR_ERR(di);
+		goto fail;
+	}
+	if (di) {
 		ret = btrfs_delete_one_dir_name(trans, log, path, di);
 		bytes_del += name_len;
 		BUG_ON(ret);
@@ -2226,7 +2221,11 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	btrfs_release_path(log, path);
 	di = btrfs_lookup_dir_index_item(trans, log, path, dir->i_ino,
 					 index, name, name_len, -1);
-	if (di && !IS_ERR(di)) {
+	if (IS_ERR(di)) {
+		err = PTR_ERR(di);
+		goto fail;
+	}
+	if (di) {
 		ret = btrfs_delete_one_dir_name(trans, log, path, di);
 		bytes_del += name_len;
 		BUG_ON(ret);
@@ -2244,6 +2243,10 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 		btrfs_release_path(log, path);
 
 		ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
+		if (ret < 0) {
+			err = ret;
+			goto fail;
+		}
 		if (ret == 0) {
 			struct btrfs_inode_item *item;
 			u64 i_size;
@@ -2261,9 +2264,13 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 			ret = 0;
 		btrfs_release_path(log, path);
 	}
-
+fail:
 	btrfs_free_path(path);
 	mutex_unlock(&BTRFS_I(dir)->log_mutex);
+	if (ret == -ENOSPC) {
+		root->fs_info->last_trans_log_full_commit = trans->transid;
+		ret = 0;
+	}
 	btrfs_end_log_trans(root);
 
 	return 0;
@@ -2291,6 +2298,10 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
 	ret = btrfs_del_inode_ref(trans, log, name, name_len, inode->i_ino,
 				  dirid, &index);
 	mutex_unlock(&BTRFS_I(inode)->log_mutex);
+	if (ret == -ENOSPC) {
+		root->fs_info->last_trans_log_full_commit = trans->transid;
+		ret = 0;
+	}
 	btrfs_end_log_trans(root);
 
 	return ret;
@@ -2318,7 +2329,8 @@ static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
 	else
 		key.type = BTRFS_DIR_LOG_INDEX_KEY;
 	ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
-	BUG_ON(ret);
+	if (ret)
+		return ret;
 
 	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
 			      struct btrfs_dir_log_item);
@@ -2343,6 +2355,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 	struct btrfs_key max_key;
 	struct btrfs_root *log = root->log_root;
 	struct extent_buffer *src;
+	int err = 0;
 	int ret;
 	int i;
 	int nritems;
@@ -2405,6 +2418,10 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 			ret = overwrite_item(trans, log, dst_path,
 					     path->nodes[0], path->slots[0],
 					     &tmp);
+			if (ret) {
+				err = ret;
+				goto done;
+			}
 		}
 	}
 	btrfs_release_path(root, path);
@@ -2432,7 +2449,10 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 				goto done;
 			ret = overwrite_item(trans, log, dst_path, src, i,
 					     &min_key);
-			BUG_ON(ret);
+			if (ret) {
+				err = ret;
+				goto done;
+			}
 		}
 		path->slots[0] = nritems;
 
@@ -2454,22 +2474,30 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 			ret = overwrite_item(trans, log, dst_path,
 					     path->nodes[0], path->slots[0],
 					     &tmp);
-
-			BUG_ON(ret);
-			last_offset = tmp.offset;
+			if (ret)
+				err = ret;
+			else
+				last_offset = tmp.offset;
 			goto done;
 		}
 	}
 done:
-	*last_offset_ret = last_offset;
 	btrfs_release_path(root, path);
 	btrfs_release_path(log, dst_path);
 
-	/* insert the log range keys to indicate where the log is valid */
-	ret = insert_dir_log_key(trans, log, path, key_type, inode->i_ino,
-				 first_offset, last_offset);
-	BUG_ON(ret);
-	return 0;
+	if (err == 0) {
+		*last_offset_ret = last_offset;
+		/*
+		 * insert the log range keys to indicate where the log
+		 * is valid
+		 */
+		ret = insert_dir_log_key(trans, log, path, key_type,
+					 inode->i_ino, first_offset,
+					 last_offset);
+		if (ret)
+			err = ret;
+	}
+	return err;
 }
 
 /*
@@ -2501,7 +2529,8 @@ again:
 		ret = log_dir_items(trans, root, inode, path,
 				    dst_path, key_type, min_key,
 				    &max_key);
-		BUG_ON(ret);
+		if (ret)
+			return ret;
 		if (max_key == (u64)-1)
 			break;
 		min_key = max_key + 1;
@@ -2535,8 +2564,8 @@ static int drop_objectid_items(struct btrfs_trans_handle *trans,
 
 	while (1) {
 		ret = btrfs_search_slot(trans, log, &key, path, -1, 1);
-
-		if (ret != 1)
+		BUG_ON(ret == 0);
+		if (ret < 0)
 			break;
 
 		if (path->slots[0] == 0)
@@ -2554,7 +2583,7 @@ static int drop_objectid_items(struct btrfs_trans_handle *trans,
 		btrfs_release_path(log, path);
 	}
 	btrfs_release_path(log, path);
-	return 0;
+	return ret;
 }
 
 static noinline int copy_items(struct btrfs_trans_handle *trans,
@@ -2587,7 +2616,10 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 	}
 	ret = btrfs_insert_empty_items(trans, log, dst_path,
 				       ins_keys, ins_sizes, nr);
-	BUG_ON(ret);
+	if (ret) {
+		kfree(ins_data);
+		return ret;
+	}
 
 	for (i = 0; i < nr; i++, dst_path->slots[0]++) {
 		dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
@@ -2660,16 +2692,17 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 	 * we have to do this after the loop above to avoid changing the
 	 * log tree while trying to change the log tree.
 	 */
+	ret = 0;
 	while (!list_empty(&ordered_sums)) {
 		struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
 						   struct btrfs_ordered_sum,
 						   list);
-		ret = btrfs_csum_file_blocks(trans, log, sums);
-		BUG_ON(ret);
+		if (!ret)
+			ret = btrfs_csum_file_blocks(trans, log, sums);
 		list_del(&sums->list);
 		kfree(sums);
 	}
-	return 0;
+	return ret;
 }
 
 /* log a single inode in the tree log.
@@ -2697,6 +2730,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 	struct btrfs_root *log = root->log_root;
 	struct extent_buffer *src = NULL;
 	u32 size;
+	int err = 0;
 	int ret;
 	int nritems;
 	int ins_start_slot = 0;
@@ -2739,7 +2773,10 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 	} else {
 		ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0);
 	}
-	BUG_ON(ret);
+	if (ret) {
+		err = ret;
+		goto out_unlock;
+	}
 	path->keep_locks = 1;
 
 	while (1) {
@@ -2768,7 +2805,10 @@ again:
 
 		ret = copy_items(trans, log, dst_path, src, ins_start_slot,
 				 ins_nr, inode_only);
-		BUG_ON(ret);
+		if (ret) {
+			err = ret;
+			goto out_unlock;
+		}
 		ins_nr = 1;
 		ins_start_slot = path->slots[0];
 next_slot:
@@ -2784,7 +2824,10 @@ next_slot:
 			ret = copy_items(trans, log, dst_path, src,
 					 ins_start_slot,
 					 ins_nr, inode_only);
-			BUG_ON(ret);
+			if (ret) {
+				err = ret;
+				goto out_unlock;
+			}
 			ins_nr = 0;
 		}
 		btrfs_release_path(root, path);
@@ -2802,7 +2845,10 @@ next_slot:
 		ret = copy_items(trans, log, dst_path, src,
 				 ins_start_slot,
 				 ins_nr, inode_only);
-		BUG_ON(ret);
+		if (ret) {
+			err = ret;
+			goto out_unlock;
+		}
 		ins_nr = 0;
 	}
 	WARN_ON(ins_nr);
@@ -2810,14 +2856,18 @@ next_slot:
 		btrfs_release_path(root, path);
 		btrfs_release_path(log, dst_path);
 		ret = log_directory_changes(trans, root, inode, path, dst_path);
-		BUG_ON(ret);
+		if (ret) {
+			err = ret;
+			goto out_unlock;
+		}
 	}
 	BTRFS_I(inode)->logged_trans = trans->transid;
+out_unlock:
 	mutex_unlock(&BTRFS_I(inode)->log_mutex);
 
 	btrfs_free_path(path);
 	btrfs_free_path(dst_path);
-	return 0;
+	return err;
 }
 
 /*
@@ -2942,10 +2992,13 @@ int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
 		goto end_no_trans;
 	}
 
-	start_log_trans(trans, root);
+	ret = start_log_trans(trans, root);
+	if (ret)
+		goto end_trans;
 
 	ret = btrfs_log_inode(trans, root, inode, inode_only);
-	BUG_ON(ret);
+	if (ret)
+		goto end_trans;
 
 	/*
 	 * for regular files, if its inode is already on disk, we don't
@@ -2955,8 +3008,10 @@ int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
 	 */
 	if (S_ISREG(inode->i_mode) &&
 	    BTRFS_I(inode)->generation <= last_committed &&
-	    BTRFS_I(inode)->last_unlink_trans <= last_committed)
-			goto no_parent;
+	    BTRFS_I(inode)->last_unlink_trans <= last_committed) {
+		ret = 0;
+		goto end_trans;
+	}
 
 	inode_only = LOG_INODE_EXISTS;
 	while (1) {
@@ -2970,15 +3025,21 @@ int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
 		if (BTRFS_I(inode)->generation >
 		    root->fs_info->last_trans_committed) {
 			ret = btrfs_log_inode(trans, root, inode, inode_only);
-			BUG_ON(ret);
+			if (ret)
+				goto end_trans;
 		}
 		if (IS_ROOT(parent))
 			break;
 
 		parent = parent->d_parent;
 	}
-no_parent:
 	ret = 0;
+end_trans:
+	if (ret < 0) {
+		BUG_ON(ret != -ENOSPC);
+		root->fs_info->last_trans_log_full_commit = trans->transid;
+		ret = 1;
+	}
 	btrfs_end_log_trans(root);
 end_no_trans:
 	return ret;
@@ -3020,7 +3081,7 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
 
-	trans = btrfs_start_transaction(fs_info->tree_root, 1);
+	trans = btrfs_start_transaction(fs_info->tree_root, 0);
 
 	wc.trans = trans;
 	wc.pin = 1;
diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h
index 0776eacb5083..3dfae84c8cc8 100644
--- a/fs/btrfs/tree-log.h
+++ b/fs/btrfs/tree-log.h
@@ -25,6 +25,8 @@
 int btrfs_sync_log(struct btrfs_trans_handle *trans,
 		   struct btrfs_root *root);
 int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root);
+int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info);
 int btrfs_recover_log_trees(struct btrfs_root *tree_root);
 int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root, struct dentry *dentry);
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 8db7b14bbae8..d6e3af8be95b 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1097,7 +1097,7 @@ static int btrfs_rm_dev_item(struct btrfs_root *root,
 	if (!path)
 		return -ENOMEM;
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_start_transaction(root, 0);
 	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
 	key.type = BTRFS_DEV_ITEM_KEY;
 	key.offset = device->devid;
@@ -1486,7 +1486,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 		goto error;
 	}
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_start_transaction(root, 0);
 	lock_chunks(root);
 
 	device->barriers = 1;
@@ -1751,9 +1751,10 @@ static int btrfs_relocate_chunk(struct btrfs_root *root,
 
 	/* step one, relocate all the extents inside this chunk */
 	ret = btrfs_relocate_block_group(extent_root, chunk_offset);
-	BUG_ON(ret);
+	if (ret)
+		return ret;
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_start_transaction(root, 0);
 	BUG_ON(!trans);
 
 	lock_chunks(root);
@@ -1925,7 +1926,7 @@ int btrfs_balance(struct btrfs_root *dev_root)
 			break;
 		BUG_ON(ret);
 
-		trans = btrfs_start_transaction(dev_root, 1);
+		trans = btrfs_start_transaction(dev_root, 0);
 		BUG_ON(!trans);
 
 		ret = btrfs_grow_device(trans, device, old_size);
@@ -2094,11 +2095,7 @@ again:
 	}
 
 	/* Shrinking succeeded, else we would be at "done". */
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans) {
-		ret = -ENOMEM;
-		goto done;
-	}
+	trans = btrfs_start_transaction(root, 0);
 	lock_chunks(root);
 
 	device->disk_total_bytes = new_size;
diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c
index 59acd3eb288a..88ecbb215878 100644
--- a/fs/btrfs/xattr.c
+++ b/fs/btrfs/xattr.c
@@ -154,15 +154,10 @@ int __btrfs_setxattr(struct btrfs_trans_handle *trans,
 	if (trans)
 		return do_setxattr(trans, inode, name, value, size, flags);
 
-	ret = btrfs_reserve_metadata_space(root, 2);
-	if (ret)
-		return ret;
+	trans = btrfs_start_transaction(root, 2);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
-	trans = btrfs_start_transaction(root, 1);
-	if (!trans) {
-		ret = -ENOMEM;
-		goto out;
-	}
 	btrfs_set_trans_block_group(trans, inode);
 
 	ret = do_setxattr(trans, inode, name, value, size, flags);
@@ -174,7 +169,6 @@ int __btrfs_setxattr(struct btrfs_trans_handle *trans,
 	BUG_ON(ret);
 out:
 	btrfs_end_transaction_throttle(trans, root);
-	btrfs_unreserve_metadata_space(root, 2);
 	return ret;
 }
 
diff --git a/fs/buffer.c b/fs/buffer.c
index e8aa7081d25c..d54812b198e9 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -1949,14 +1949,11 @@ static int __block_commit_write(struct inode *inode, struct page *page,
 }
 
 /*
- * block_write_begin takes care of the basic task of block allocation and
- * bringing partial write blocks uptodate first.
- *
- * If *pagep is not NULL, then block_write_begin uses the locked page
- * at *pagep rather than allocating its own. In this case, the page will
- * not be unlocked or deallocated on failure.
+ * Filesystems implementing the new truncate sequence should use the
+ * _newtrunc postfix variant which won't incorrectly call vmtruncate.
+ * The filesystem needs to handle block truncation upon failure.
  */
-int block_write_begin(struct file *file, struct address_space *mapping,
+int block_write_begin_newtrunc(struct file *file, struct address_space *mapping,
 			loff_t pos, unsigned len, unsigned flags,
 			struct page **pagep, void **fsdata,
 			get_block_t *get_block)
@@ -1992,20 +1989,50 @@ int block_write_begin(struct file *file, struct address_space *mapping,
 			unlock_page(page);
 			page_cache_release(page);
 			*pagep = NULL;
-
-			/*
-			 * prepare_write() may have instantiated a few blocks
-			 * outside i_size.  Trim these off again. Don't need
-			 * i_size_read because we hold i_mutex.
-			 */
-			if (pos + len > inode->i_size)
-				vmtruncate(inode, inode->i_size);
 		}
 	}
 
 out:
 	return status;
 }
+EXPORT_SYMBOL(block_write_begin_newtrunc);
+
+/*
+ * block_write_begin takes care of the basic task of block allocation and
+ * bringing partial write blocks uptodate first.
+ *
+ * If *pagep is not NULL, then block_write_begin uses the locked page
+ * at *pagep rather than allocating its own. In this case, the page will
+ * not be unlocked or deallocated on failure.
+ */
+int block_write_begin(struct file *file, struct address_space *mapping,
+			loff_t pos, unsigned len, unsigned flags,
+			struct page **pagep, void **fsdata,
+			get_block_t *get_block)
+{
+	int ret;
+
+	ret = block_write_begin_newtrunc(file, mapping, pos, len, flags,
+					pagep, fsdata, get_block);
+
+	/*
+	 * prepare_write() may have instantiated a few blocks
+	 * outside i_size.  Trim these off again. Don't need
+	 * i_size_read because we hold i_mutex.
+	 *
+	 * Filesystems which pass down their own page also cannot
+	 * call into vmtruncate here because it would lead to lock
+	 * inversion problems (*pagep is locked). This is a further
+	 * example of where the old truncate sequence is inadequate.
+	 */
+	if (unlikely(ret) && *pagep == NULL) {
+		loff_t isize = mapping->host->i_size;
+		if (pos + len > isize)
+			vmtruncate(mapping->host, isize);
+	}
+
+	return ret;
+}
 EXPORT_SYMBOL(block_write_begin);
 
 int block_write_end(struct file *file, struct address_space *mapping,
@@ -2324,7 +2351,7 @@ out:
  * For moronic filesystems that do not allow holes in file.
  * We may have to extend the file.
  */
-int cont_write_begin(struct file *file, struct address_space *mapping,
+int cont_write_begin_newtrunc(struct file *file, struct address_space *mapping,
 			loff_t pos, unsigned len, unsigned flags,
 			struct page **pagep, void **fsdata,
 			get_block_t *get_block, loff_t *bytes)
@@ -2345,11 +2372,30 @@ int cont_write_begin(struct file *file, struct address_space *mapping,
 	}
 
 	*pagep = NULL;
-	err = block_write_begin(file, mapping, pos, len,
+	err = block_write_begin_newtrunc(file, mapping, pos, len,
 				flags, pagep, fsdata, get_block);
 out:
 	return err;
 }
+EXPORT_SYMBOL(cont_write_begin_newtrunc);
+
+int cont_write_begin(struct file *file, struct address_space *mapping,
+			loff_t pos, unsigned len, unsigned flags,
+			struct page **pagep, void **fsdata,
+			get_block_t *get_block, loff_t *bytes)
+{
+	int ret;
+
+	ret = cont_write_begin_newtrunc(file, mapping, pos, len, flags,
+					pagep, fsdata, get_block, bytes);
+	if (unlikely(ret)) {
+		loff_t isize = mapping->host->i_size;
+		if (pos + len > isize)
+			vmtruncate(mapping->host, isize);
+	}
+
+	return ret;
+}
 EXPORT_SYMBOL(cont_write_begin);
 
 int block_prepare_write(struct page *page, unsigned from, unsigned to,
@@ -2381,7 +2427,7 @@ EXPORT_SYMBOL(block_commit_write);
  *
  * We are not allowed to take the i_mutex here so we have to play games to
  * protect against truncate races as the page could now be beyond EOF.  Because
- * vmtruncate() writes the inode size before removing pages, once we have the
+ * truncate writes the inode size before removing pages, once we have the
  * page lock we can determine safely if the page is beyond EOF. If it is not
  * beyond EOF, then the page is guaranteed safe against truncation until we
  * unlock the page.
@@ -2464,10 +2510,11 @@ static void attach_nobh_buffers(struct page *page, struct buffer_head *head)
 }
 
 /*
- * On entry, the page is fully not uptodate.
- * On exit the page is fully uptodate in the areas outside (from,to)
+ * Filesystems implementing the new truncate sequence should use the
+ * _newtrunc postfix variant which won't incorrectly call vmtruncate.
+ * The filesystem needs to handle block truncation upon failure.
  */
-int nobh_write_begin(struct file *file, struct address_space *mapping,
+int nobh_write_begin_newtrunc(struct file *file, struct address_space *mapping,
 			loff_t pos, unsigned len, unsigned flags,
 			struct page **pagep, void **fsdata,
 			get_block_t *get_block)
@@ -2500,8 +2547,8 @@ int nobh_write_begin(struct file *file, struct address_space *mapping,
 		unlock_page(page);
 		page_cache_release(page);
 		*pagep = NULL;
-		return block_write_begin(file, mapping, pos, len, flags, pagep,
-					fsdata, get_block);
+		return block_write_begin_newtrunc(file, mapping, pos, len,
+					flags, pagep, fsdata, get_block);
 	}
 
 	if (PageMappedToDisk(page))
@@ -2605,8 +2652,34 @@ out_release:
 	page_cache_release(page);
 	*pagep = NULL;
 
-	if (pos + len > inode->i_size)
-		vmtruncate(inode, inode->i_size);
+	return ret;
+}
+EXPORT_SYMBOL(nobh_write_begin_newtrunc);
+
+/*
+ * On entry, the page is fully not uptodate.
+ * On exit the page is fully uptodate in the areas outside (from,to)
+ */
+int nobh_write_begin(struct file *file, struct address_space *mapping,
+			loff_t pos, unsigned len, unsigned flags,
+			struct page **pagep, void **fsdata,
+			get_block_t *get_block)
+{
+	int ret;
+
+	ret = nobh_write_begin_newtrunc(file, mapping, pos, len, flags,
+					pagep, fsdata, get_block);
+
+	/*
+	 * prepare_write() may have instantiated a few blocks
+	 * outside i_size.  Trim these off again. Don't need
+	 * i_size_read because we hold i_mutex.
+	 */
+	if (unlikely(ret)) {
+		loff_t isize = mapping->host->i_size;
+		if (pos + len > isize)
+			vmtruncate(mapping->host, isize);
+	}
 
 	return ret;
 }
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index a9005d862ed4..d9c60b84949a 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -274,7 +274,6 @@ static int ceph_readpages(struct file *file, struct address_space *mapping,
 	struct ceph_osd_client *osdc = &ceph_inode_to_client(inode)->osdc;
 	int rc = 0;
 	struct page **pages;
-	struct pagevec pvec;
 	loff_t offset;
 	u64 len;
 
@@ -297,8 +296,6 @@ static int ceph_readpages(struct file *file, struct address_space *mapping,
 	if (rc < 0)
 		goto out;
 
-	/* set uptodate and add to lru in pagevec-sized chunks */
-	pagevec_init(&pvec, 0);
 	for (; !list_empty(page_list) && len > 0;
 	     rc -= PAGE_CACHE_SIZE, len -= PAGE_CACHE_SIZE) {
 		struct page *page =
@@ -312,7 +309,7 @@ static int ceph_readpages(struct file *file, struct address_space *mapping,
 			zero_user_segment(page, s, PAGE_CACHE_SIZE);
 		}
 
-		if (add_to_page_cache(page, mapping, page->index, GFP_NOFS)) {
+		if (add_to_page_cache_lru(page, mapping, page->index, GFP_NOFS)) {
 			page_cache_release(page);
 			dout("readpages %p add_to_page_cache failed %p\n",
 			     inode, page);
@@ -323,10 +320,8 @@ static int ceph_readpages(struct file *file, struct address_space *mapping,
 		flush_dcache_page(page);
 		SetPageUptodate(page);
 		unlock_page(page);
-		if (pagevec_add(&pvec, page) == 0)
-			pagevec_lru_add_file(&pvec);   /* add to lru */
+		page_cache_release(page);
 	}
-	pagevec_lru_add_file(&pvec);
 	rc = 0;
 
 out:
@@ -568,7 +563,7 @@ static void writepages_finish(struct ceph_osd_request *req,
 	ceph_release_pages(req->r_pages, req->r_num_pages);
 	if (req->r_pages_from_pool)
 		mempool_free(req->r_pages,
-			     ceph_client(inode->i_sb)->wb_pagevec_pool);
+			     ceph_sb_to_client(inode->i_sb)->wb_pagevec_pool);
 	else
 		kfree(req->r_pages);
 	ceph_osdc_put_request(req);
diff --git a/fs/ceph/auth.c b/fs/ceph/auth.c
index 818afe72e6c7..89490beaf537 100644
--- a/fs/ceph/auth.c
+++ b/fs/ceph/auth.c
@@ -1,7 +1,6 @@
 #include "ceph_debug.h"
 
 #include <linux/module.h>
-#include <linux/slab.h>
 #include <linux/err.h>
 #include <linux/slab.h>
 
@@ -150,7 +149,8 @@ int ceph_build_auth_request(struct ceph_auth_client *ac,
 
 	ret = ac->ops->build_request(ac, p + sizeof(u32), end);
 	if (ret < 0) {
-		pr_err("error %d building request\n", ret);
+		pr_err("error %d building auth method %s request\n", ret,
+		       ac->ops->name);
 		return ret;
 	}
 	dout(" built request %d bytes\n", ret);
@@ -229,7 +229,7 @@ int ceph_handle_auth_reply(struct ceph_auth_client *ac,
 	if (ret == -EAGAIN) {
 		return ceph_build_auth_request(ac, reply_buf, reply_len);
 	} else if (ret) {
-		pr_err("authentication error %d\n", ret);
+		pr_err("auth method '%s' error %d\n", ac->ops->name, ret);
 		return ret;
 	}
 	return 0;
@@ -246,7 +246,7 @@ int ceph_build_auth(struct ceph_auth_client *ac,
 	if (!ac->protocol)
 		return ceph_auth_build_hello(ac, msg_buf, msg_len);
 	BUG_ON(!ac->ops);
-	if (!ac->ops->is_authenticated(ac))
+	if (ac->ops->should_authenticate(ac))
 		return ceph_build_auth_request(ac, msg_buf, msg_len);
 	return 0;
 }
diff --git a/fs/ceph/auth.h b/fs/ceph/auth.h
index ca4f57cfb267..d38a2fb4a137 100644
--- a/fs/ceph/auth.h
+++ b/fs/ceph/auth.h
@@ -15,6 +15,8 @@ struct ceph_auth_client;
 struct ceph_authorizer;
 
 struct ceph_auth_client_ops {
+	const char *name;
+
 	/*
 	 * true if we are authenticated and can connect to
 	 * services.
@@ -22,6 +24,12 @@ struct ceph_auth_client_ops {
 	int (*is_authenticated)(struct ceph_auth_client *ac);
 
 	/*
+	 * true if we should (re)authenticate, e.g., when our tickets
+	 * are getting old and crusty.
+	 */
+	int (*should_authenticate)(struct ceph_auth_client *ac);
+
+	/*
 	 * build requests and process replies during monitor
 	 * handshake.  if handle_reply returns -EAGAIN, we build
 	 * another request.
diff --git a/fs/ceph/auth_none.c b/fs/ceph/auth_none.c
index 8cd9e3af07f7..ad1dc21286c7 100644
--- a/fs/ceph/auth_none.c
+++ b/fs/ceph/auth_none.c
@@ -31,6 +31,13 @@ static int is_authenticated(struct ceph_auth_client *ac)
 	return !xi->starting;
 }
 
+static int should_authenticate(struct ceph_auth_client *ac)
+{
+	struct ceph_auth_none_info *xi = ac->private;
+
+	return xi->starting;
+}
+
 /*
  * the generic auth code decode the global_id, and we carry no actual
  * authenticate state, so nothing happens here.
@@ -94,9 +101,11 @@ static void ceph_auth_none_destroy_authorizer(struct ceph_auth_client *ac,
 }
 
 static const struct ceph_auth_client_ops ceph_auth_none_ops = {
+	.name = "none",
 	.reset = reset,
 	.destroy = destroy,
 	.is_authenticated = is_authenticated,
+	.should_authenticate = should_authenticate,
 	.handle_reply = handle_reply,
 	.create_authorizer = ceph_auth_none_create_authorizer,
 	.destroy_authorizer = ceph_auth_none_destroy_authorizer,
diff --git a/fs/ceph/auth_x.c b/fs/ceph/auth_x.c
index fee5a08da881..83d4d2785ffe 100644
--- a/fs/ceph/auth_x.c
+++ b/fs/ceph/auth_x.c
@@ -27,6 +27,17 @@ static int ceph_x_is_authenticated(struct ceph_auth_client *ac)
 	return (ac->want_keys & xi->have_keys) == ac->want_keys;
 }
 
+static int ceph_x_should_authenticate(struct ceph_auth_client *ac)
+{
+	struct ceph_x_info *xi = ac->private;
+	int need;
+
+	ceph_x_validate_tickets(ac, &need);
+	dout("ceph_x_should_authenticate want=%d need=%d have=%d\n",
+	     ac->want_keys, need, xi->have_keys);
+	return need != 0;
+}
+
 static int ceph_x_encrypt_buflen(int ilen)
 {
 	return sizeof(struct ceph_x_encrypt_header) + ilen + 16 +
@@ -127,7 +138,7 @@ static int ceph_x_proc_ticket_reply(struct ceph_auth_client *ac,
 	int ret;
 	char *dbuf;
 	char *ticket_buf;
-	u8 struct_v;
+	u8 reply_struct_v;
 
 	dbuf = kmalloc(TEMP_TICKET_BUF_LEN, GFP_NOFS);
 	if (!dbuf)
@@ -139,14 +150,14 @@ static int ceph_x_proc_ticket_reply(struct ceph_auth_client *ac,
 		goto out_dbuf;
 
 	ceph_decode_need(&p, end, 1 + sizeof(u32), bad);
-	struct_v = ceph_decode_8(&p);
-	if (struct_v != 1)
+	reply_struct_v = ceph_decode_8(&p);
+	if (reply_struct_v != 1)
 		goto bad;
 	num = ceph_decode_32(&p);
 	dout("%d tickets\n", num);
 	while (num--) {
 		int type;
-		u8 struct_v;
+		u8 tkt_struct_v, blob_struct_v;
 		struct ceph_x_ticket_handler *th;
 		void *dp, *dend;
 		int dlen;
@@ -165,8 +176,8 @@ static int ceph_x_proc_ticket_reply(struct ceph_auth_client *ac,
 		type = ceph_decode_32(&p);
 		dout(" ticket type %d %s\n", type, ceph_entity_type_name(type));
 
-		struct_v = ceph_decode_8(&p);
-		if (struct_v != 1)
+		tkt_struct_v = ceph_decode_8(&p);
+		if (tkt_struct_v != 1)
 			goto bad;
 
 		th = get_ticket_handler(ac, type);
@@ -186,8 +197,8 @@ static int ceph_x_proc_ticket_reply(struct ceph_auth_client *ac,
 		dend = dbuf + dlen;
 		dp = dbuf;
 
-		struct_v = ceph_decode_8(&dp);
-		if (struct_v != 1)
+		tkt_struct_v = ceph_decode_8(&dp);
+		if (tkt_struct_v != 1)
 			goto bad;
 
 		memcpy(&old_key, &th->session_key, sizeof(old_key));
@@ -224,7 +235,7 @@ static int ceph_x_proc_ticket_reply(struct ceph_auth_client *ac,
 		tpend = tp + dlen;
 		dout(" ticket blob is %d bytes\n", dlen);
 		ceph_decode_need(&tp, tpend, 1 + sizeof(u64), bad);
-		struct_v = ceph_decode_8(&tp);
+		blob_struct_v = ceph_decode_8(&tp);
 		new_secret_id = ceph_decode_64(&tp);
 		ret = ceph_decode_buffer(&new_ticket_blob, &tp, tpend);
 		if (ret)
@@ -618,7 +629,9 @@ static void ceph_x_invalidate_authorizer(struct ceph_auth_client *ac,
 
 
 static const struct ceph_auth_client_ops ceph_x_ops = {
+	.name = "x",
 	.is_authenticated = ceph_x_is_authenticated,
+	.should_authenticate = ceph_x_should_authenticate,
 	.build_request = ceph_x_build_request,
 	.handle_reply = ceph_x_handle_reply,
 	.create_authorizer = ceph_x_create_authorizer,
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index d9400534b279..ae3e3a306445 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -867,7 +867,8 @@ void __ceph_remove_cap(struct ceph_cap *cap)
 {
 	struct ceph_mds_session *session = cap->session;
 	struct ceph_inode_info *ci = cap->ci;
-	struct ceph_mds_client *mdsc = &ceph_client(ci->vfs_inode.i_sb)->mdsc;
+	struct ceph_mds_client *mdsc =
+		&ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
 	int removed = 0;
 
 	dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode);
@@ -937,9 +938,9 @@ static int send_cap_msg(struct ceph_mds_session *session,
 	     seq, issue_seq, mseq, follows, size, max_size,
 	     xattr_version, xattrs_buf ? (int)xattrs_buf->vec.iov_len : 0);
 
-	msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPS, sizeof(*fc), 0, 0, NULL);
-	if (IS_ERR(msg))
-		return PTR_ERR(msg);
+	msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPS, sizeof(*fc), GFP_NOFS);
+	if (!msg)
+		return -ENOMEM;
 
 	msg->hdr.tid = cpu_to_le64(flush_tid);
 
@@ -1298,7 +1299,8 @@ static void ceph_flush_snaps(struct ceph_inode_info *ci)
  */
 void __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask)
 {
-	struct ceph_mds_client *mdsc = &ceph_client(ci->vfs_inode.i_sb)->mdsc;
+	struct ceph_mds_client *mdsc =
+		&ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
 	struct inode *inode = &ci->vfs_inode;
 	int was = ci->i_dirty_caps;
 	int dirty = 0;
@@ -1336,7 +1338,7 @@ void __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask)
 static int __mark_caps_flushing(struct inode *inode,
 				 struct ceph_mds_session *session)
 {
-	struct ceph_mds_client *mdsc = &ceph_client(inode->i_sb)->mdsc;
+	struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	int flushing;
 
@@ -1663,7 +1665,7 @@ ack:
 static int try_flush_caps(struct inode *inode, struct ceph_mds_session *session,
 			  unsigned *flush_tid)
 {
-	struct ceph_mds_client *mdsc = &ceph_client(inode->i_sb)->mdsc;
+	struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	int unlock_session = session ? 0 : 1;
 	int flushing = 0;
@@ -1716,10 +1718,9 @@ out_unlocked:
 static int caps_are_flushed(struct inode *inode, unsigned tid)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	int dirty, i, ret = 1;
+	int i, ret = 1;
 
 	spin_lock(&inode->i_lock);
-	dirty = __ceph_caps_dirty(ci);
 	for (i = 0; i < CEPH_CAP_BITS; i++)
 		if ((ci->i_flushing_caps & (1 << i)) &&
 		    ci->i_cap_flush_tid[i] <= tid) {
@@ -1775,9 +1776,9 @@ out:
 	spin_unlock(&ci->i_unsafe_lock);
 }
 
-int ceph_fsync(struct file *file, struct dentry *dentry, int datasync)
+int ceph_fsync(struct file *file, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	unsigned flush_tid;
 	int ret;
@@ -1829,7 +1830,8 @@ int ceph_write_inode(struct inode *inode, struct writeback_control *wbc)
 			err = wait_event_interruptible(ci->i_cap_wq,
 				       caps_are_flushed(inode, flush_tid));
 	} else {
-		struct ceph_mds_client *mdsc = &ceph_client(inode->i_sb)->mdsc;
+		struct ceph_mds_client *mdsc =
+			&ceph_sb_to_client(inode->i_sb)->mdsc;
 
 		spin_lock(&inode->i_lock);
 		if (__ceph_caps_dirty(ci))
@@ -2411,7 +2413,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
 	__releases(inode->i_lock)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	struct ceph_mds_client *mdsc = &ceph_client(inode->i_sb)->mdsc;
+	struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc;
 	unsigned seq = le32_to_cpu(m->seq);
 	int dirty = le32_to_cpu(m->dirty);
 	int cleaned = 0;
diff --git a/fs/ceph/ceph_fs.h b/fs/ceph/ceph_fs.h
index 0c2241ef3653..2fa992eaf7da 100644
--- a/fs/ceph/ceph_fs.h
+++ b/fs/ceph/ceph_fs.h
@@ -19,7 +19,7 @@
  * Ceph release version
  */
 #define CEPH_VERSION_MAJOR 0
-#define CEPH_VERSION_MINOR 19
+#define CEPH_VERSION_MINOR 20
 #define CEPH_VERSION_PATCH 0
 
 #define _CEPH_STRINGIFY(x) #x
@@ -36,7 +36,7 @@
  * client-facing protocol.
  */
 #define CEPH_OSD_PROTOCOL     8 /* cluster internal */
-#define CEPH_MDS_PROTOCOL     9 /* cluster internal */
+#define CEPH_MDS_PROTOCOL    12 /* cluster internal */
 #define CEPH_MON_PROTOCOL     5 /* cluster internal */
 #define CEPH_OSDC_PROTOCOL   24 /* server/client */
 #define CEPH_MDSC_PROTOCOL   32 /* server/client */
@@ -53,8 +53,18 @@
 /*
  * feature bits
  */
-#define CEPH_FEATURE_SUPPORTED  0
-#define CEPH_FEATURE_REQUIRED   0
+#define CEPH_FEATURE_UID        1
+#define CEPH_FEATURE_NOSRCADDR  2
+#define CEPH_FEATURE_FLOCK      4
+
+#define CEPH_FEATURE_SUPPORTED_MON  CEPH_FEATURE_UID|CEPH_FEATURE_NOSRCADDR
+#define CEPH_FEATURE_REQUIRED_MON   CEPH_FEATURE_UID
+#define CEPH_FEATURE_SUPPORTED_MDS  CEPH_FEATURE_UID|CEPH_FEATURE_NOSRCADDR|CEPH_FEATURE_FLOCK
+#define CEPH_FEATURE_REQUIRED_MDS   CEPH_FEATURE_UID
+#define CEPH_FEATURE_SUPPORTED_OSD  CEPH_FEATURE_UID|CEPH_FEATURE_NOSRCADDR
+#define CEPH_FEATURE_REQUIRED_OSD   CEPH_FEATURE_UID
+#define CEPH_FEATURE_SUPPORTED_CLIENT CEPH_FEATURE_NOSRCADDR
+#define CEPH_FEATURE_REQUIRED_CLIENT CEPH_FEATURE_NOSRCADDR
 
 
 /*
@@ -91,6 +101,8 @@ int ceph_file_layout_is_valid(const struct ceph_file_layout *layout);
 #define CEPH_AUTH_NONE	 	0x1
 #define CEPH_AUTH_CEPHX	 	0x2
 
+#define CEPH_AUTH_UID_DEFAULT ((__u64) -1)
+
 
 /*********************************************
  * message layer
@@ -128,11 +140,27 @@ int ceph_file_layout_is_valid(const struct ceph_file_layout *layout);
 #define CEPH_MSG_CLIENT_SNAP            0x312
 #define CEPH_MSG_CLIENT_CAPRELEASE      0x313
 
+/* pool ops */
+#define CEPH_MSG_POOLOP_REPLY           48
+#define CEPH_MSG_POOLOP                 49
+
+
 /* osd */
 #define CEPH_MSG_OSD_MAP          41
 #define CEPH_MSG_OSD_OP           42
 #define CEPH_MSG_OSD_OPREPLY      43
 
+/* pool operations */
+enum {
+  POOL_OP_CREATE			= 0x01,
+  POOL_OP_DELETE			= 0x02,
+  POOL_OP_AUID_CHANGE			= 0x03,
+  POOL_OP_CREATE_SNAP			= 0x11,
+  POOL_OP_DELETE_SNAP			= 0x12,
+  POOL_OP_CREATE_UNMANAGED_SNAP		= 0x21,
+  POOL_OP_DELETE_UNMANAGED_SNAP		= 0x22,
+};
+
 struct ceph_mon_request_header {
 	__le64 have_version;
 	__le16 session_mon;
@@ -155,6 +183,31 @@ struct ceph_mon_statfs_reply {
 	struct ceph_statfs st;
 } __attribute__ ((packed));
 
+const char *ceph_pool_op_name(int op);
+
+struct ceph_mon_poolop {
+	struct ceph_mon_request_header monhdr;
+	struct ceph_fsid fsid;
+	__le32 pool;
+	__le32 op;
+	__le64 auid;
+	__le64 snapid;
+	__le32 name_len;
+} __attribute__ ((packed));
+
+struct ceph_mon_poolop_reply {
+	struct ceph_mon_request_header monhdr;
+	struct ceph_fsid fsid;
+	__le32 reply_code;
+	__le32 epoch;
+	char has_data;
+	char data[0];
+} __attribute__ ((packed));
+
+struct ceph_mon_unmanaged_snap {
+	__le64 snapid;
+} __attribute__ ((packed));
+
 struct ceph_osd_getmap {
 	struct ceph_mon_request_header monhdr;
 	struct ceph_fsid fsid;
@@ -212,16 +265,17 @@ extern const char *ceph_mds_state_name(int s);
  *  - they also define the lock ordering by the MDS
  *  - a few of these are internal to the mds
  */
-#define CEPH_LOCK_DN          1
-#define CEPH_LOCK_ISNAP       2
-#define CEPH_LOCK_IVERSION    4     /* mds internal */
-#define CEPH_LOCK_IFILE       8     /* mds internal */
-#define CEPH_LOCK_IAUTH       32
-#define CEPH_LOCK_ILINK       64
-#define CEPH_LOCK_IDFT        128   /* dir frag tree */
-#define CEPH_LOCK_INEST       256   /* mds internal */
-#define CEPH_LOCK_IXATTR      512
-#define CEPH_LOCK_INO         2048  /* immutable inode bits; not a lock */
+#define CEPH_LOCK_DVERSION    1
+#define CEPH_LOCK_DN          2
+#define CEPH_LOCK_ISNAP       16
+#define CEPH_LOCK_IVERSION    32    /* mds internal */
+#define CEPH_LOCK_IFILE       64
+#define CEPH_LOCK_IAUTH       128
+#define CEPH_LOCK_ILINK       256
+#define CEPH_LOCK_IDFT        512   /* dir frag tree */
+#define CEPH_LOCK_INEST       1024  /* mds internal */
+#define CEPH_LOCK_IXATTR      2048
+#define CEPH_LOCK_INO         8192  /* immutable inode bits; not a lock */
 
 /* client_session ops */
 enum {
@@ -308,6 +362,7 @@ union ceph_mds_request_args {
 	struct {
 		__le32 frag;                 /* which dir fragment */
 		__le32 max_entries;          /* how many dentries to grab */
+		__le32 max_bytes;
 	} __attribute__ ((packed)) readdir;
 	struct {
 		__le32 mode;
diff --git a/fs/ceph/ceph_strings.c b/fs/ceph/ceph_strings.c
index 8e4be6a80c62..7503aee828ce 100644
--- a/fs/ceph/ceph_strings.c
+++ b/fs/ceph/ceph_strings.c
@@ -10,7 +10,6 @@ const char *ceph_entity_type_name(int type)
 	case CEPH_ENTITY_TYPE_OSD: return "osd";
 	case CEPH_ENTITY_TYPE_MON: return "mon";
 	case CEPH_ENTITY_TYPE_CLIENT: return "client";
-	case CEPH_ENTITY_TYPE_ADMIN: return "admin";
 	case CEPH_ENTITY_TYPE_AUTH: return "auth";
 	default: return "unknown";
 	}
@@ -45,6 +44,7 @@ const char *ceph_osd_op_name(int op)
 	case CEPH_OSD_OP_SETXATTRS: return "setxattrs";
 	case CEPH_OSD_OP_RESETXATTRS: return "resetxattrs";
 	case CEPH_OSD_OP_RMXATTR: return "rmxattr";
+	case CEPH_OSD_OP_CMPXATTR: return "cmpxattr";
 
 	case CEPH_OSD_OP_PULL: return "pull";
 	case CEPH_OSD_OP_PUSH: return "push";
@@ -174,3 +174,17 @@ const char *ceph_snap_op_name(int o)
 	}
 	return "???";
 }
+
+const char *ceph_pool_op_name(int op)
+{
+	switch (op) {
+	case POOL_OP_CREATE: return "create";
+	case POOL_OP_DELETE: return "delete";
+	case POOL_OP_AUID_CHANGE: return "auid change";
+	case POOL_OP_CREATE_SNAP: return "create snap";
+	case POOL_OP_DELETE_SNAP: return "delete snap";
+	case POOL_OP_CREATE_UNMANAGED_SNAP: return "create unmanaged snap";
+	case POOL_OP_DELETE_UNMANAGED_SNAP: return "delete unmanaged snap";
+	}
+	return "???";
+}
diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c
index f7048da92acc..3be33fb066cc 100644
--- a/fs/ceph/debugfs.c
+++ b/fs/ceph/debugfs.c
@@ -113,7 +113,7 @@ static int osdmap_show(struct seq_file *s, void *p)
 static int monc_show(struct seq_file *s, void *p)
 {
 	struct ceph_client *client = s->private;
-	struct ceph_mon_statfs_request *req;
+	struct ceph_mon_generic_request *req;
 	struct ceph_mon_client *monc = &client->monc;
 	struct rb_node *rp;
 
@@ -126,9 +126,14 @@ static int monc_show(struct seq_file *s, void *p)
 	if (monc->want_next_osdmap)
 		seq_printf(s, "want next osdmap\n");
 
-	for (rp = rb_first(&monc->statfs_request_tree); rp; rp = rb_next(rp)) {
-		req = rb_entry(rp, struct ceph_mon_statfs_request, node);
-		seq_printf(s, "%lld statfs\n", req->tid);
+	for (rp = rb_first(&monc->generic_request_tree); rp; rp = rb_next(rp)) {
+		__u16 op;
+		req = rb_entry(rp, struct ceph_mon_generic_request, node);
+		op = le16_to_cpu(req->request->hdr.type);
+		if (op == CEPH_MSG_STATFS)
+			seq_printf(s, "%lld statfs\n", req->tid);
+		else
+			seq_printf(s, "%lld unknown\n", req->tid);
 	}
 
 	mutex_unlock(&monc->mutex);
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 650d2db5ed26..f85719310db2 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -51,8 +51,11 @@ int ceph_init_dentry(struct dentry *dentry)
 		return -ENOMEM;          /* oh well */
 
 	spin_lock(&dentry->d_lock);
-	if (dentry->d_fsdata) /* lost a race */
+	if (dentry->d_fsdata) {
+		/* lost a race */
+		kmem_cache_free(ceph_dentry_cachep, di);
 		goto out_unlock;
+	}
 	di->dentry = dentry;
 	di->lease_session = NULL;
 	dentry->d_fsdata = di;
@@ -125,7 +128,8 @@ more:
 	dentry = list_entry(p, struct dentry, d_u.d_child);
 	di = ceph_dentry(dentry);
 	while (1) {
-		dout(" p %p/%p d_subdirs %p/%p\n", p->prev, p->next,
+		dout(" p %p/%p %s d_subdirs %p/%p\n", p->prev, p->next,
+		     d_unhashed(dentry) ? "!hashed" : "hashed",
 		     parent->d_subdirs.prev, parent->d_subdirs.next);
 		if (p == &parent->d_subdirs) {
 			fi->at_end = 1;
@@ -229,6 +233,7 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	u32 ftype;
 	struct ceph_mds_reply_info_parsed *rinfo;
 	const int max_entries = client->mount_args->max_readdir;
+	const int max_bytes = client->mount_args->max_readdir_bytes;
 
 	dout("readdir %p filp %p frag %u off %u\n", inode, filp, frag, off);
 	if (fi->at_end)
@@ -312,6 +317,7 @@ more:
 		req->r_readdir_offset = fi->next_offset;
 		req->r_args.readdir.frag = cpu_to_le32(frag);
 		req->r_args.readdir.max_entries = cpu_to_le32(max_entries);
+		req->r_args.readdir.max_bytes = cpu_to_le32(max_bytes);
 		req->r_num_caps = max_entries + 1;
 		err = ceph_mdsc_do_request(mdsc, NULL, req);
 		if (err < 0) {
@@ -335,7 +341,7 @@ more:
 		if (req->r_reply_info.dir_end) {
 			kfree(fi->last_name);
 			fi->last_name = NULL;
-			fi->next_offset = 0;
+			fi->next_offset = 2;
 		} else {
 			rinfo = &req->r_reply_info;
 			err = note_last_dentry(fi,
@@ -478,7 +484,7 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int origin)
 struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
 				  struct dentry *dentry, int err)
 {
-	struct ceph_client *client = ceph_client(dentry->d_sb);
+	struct ceph_client *client = ceph_sb_to_client(dentry->d_sb);
 	struct inode *parent = dentry->d_parent->d_inode;
 
 	/* .snap dir? */
@@ -568,7 +574,6 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
 		    !is_root_ceph_dentry(dir, dentry) &&
 		    (ci->i_ceph_flags & CEPH_I_COMPLETE) &&
 		    (__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1))) {
-			di->offset = ci->i_max_offset++;
 			spin_unlock(&dir->i_lock);
 			dout(" dir %p complete, -ENOENT\n", dir);
 			d_add(dentry, NULL);
@@ -582,7 +587,7 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
 		CEPH_MDS_OP_LOOKUPSNAP : CEPH_MDS_OP_LOOKUP;
 	req = ceph_mdsc_create_request(mdsc, op, USE_ANY_MDS);
 	if (IS_ERR(req))
-		return ERR_PTR(PTR_ERR(req));
+		return ERR_CAST(req);
 	req->r_dentry = dget(dentry);
 	req->r_num_caps = 2;
 	/* we only need inode linkage */
@@ -888,13 +893,22 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry,
 
 		/* ensure target dentry is invalidated, despite
 		   rehashing bug in vfs_rename_dir */
-		new_dentry->d_time = jiffies;
-		ceph_dentry(new_dentry)->lease_shared_gen = 0;
+		ceph_invalidate_dentry_lease(new_dentry);
 	}
 	ceph_mdsc_put_request(req);
 	return err;
 }
 
+/*
+ * Ensure a dentry lease will no longer revalidate.
+ */
+void ceph_invalidate_dentry_lease(struct dentry *dentry)
+{
+	spin_lock(&dentry->d_lock);
+	dentry->d_time = jiffies;
+	ceph_dentry(dentry)->lease_shared_gen = 0;
+	spin_unlock(&dentry->d_lock);
+}
 
 /*
  * Check if dentry lease is valid.  If not, delete the lease.  Try to
@@ -972,8 +986,9 @@ static int ceph_d_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
 	struct inode *dir = dentry->d_parent->d_inode;
 
-	dout("d_revalidate %p '%.*s' inode %p\n", dentry,
-	     dentry->d_name.len, dentry->d_name.name, dentry->d_inode);
+	dout("d_revalidate %p '%.*s' inode %p offset %lld\n", dentry,
+	     dentry->d_name.len, dentry->d_name.name, dentry->d_inode,
+	     ceph_dentry(dentry)->offset);
 
 	/* always trust cached snapped dentries, snapdir dentry */
 	if (ceph_snap(dir) != CEPH_NOSNAP) {
@@ -1050,7 +1065,7 @@ static ssize_t ceph_read_dir(struct file *file, char __user *buf, size_t size,
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	int left;
 
-	if (!ceph_test_opt(ceph_client(inode->i_sb), DIRSTAT))
+	if (!ceph_test_opt(ceph_sb_to_client(inode->i_sb), DIRSTAT))
 		return -EISDIR;
 
 	if (!cf->dir_info) {
@@ -1092,10 +1107,9 @@ static ssize_t ceph_read_dir(struct file *file, char __user *buf, size_t size,
  * an fsync() on a dir will wait for any uncommitted directory
  * operations to commit.
  */
-static int ceph_dir_fsync(struct file *file, struct dentry *dentry,
-			  int datasync)
+static int ceph_dir_fsync(struct file *file, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_path.dentry->d_inode;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	struct list_head *head = &ci->i_unsafe_dirops;
 	struct ceph_mds_request *req;
@@ -1152,7 +1166,7 @@ void ceph_dentry_lru_add(struct dentry *dn)
 	dout("dentry_lru_add %p %p '%.*s'\n", di, dn,
 	     dn->d_name.len, dn->d_name.name);
 	if (di) {
-		mdsc = &ceph_client(dn->d_sb)->mdsc;
+		mdsc = &ceph_sb_to_client(dn->d_sb)->mdsc;
 		spin_lock(&mdsc->dentry_lru_lock);
 		list_add_tail(&di->lru, &mdsc->dentry_lru);
 		mdsc->num_dentry++;
@@ -1165,10 +1179,10 @@ void ceph_dentry_lru_touch(struct dentry *dn)
 	struct ceph_dentry_info *di = ceph_dentry(dn);
 	struct ceph_mds_client *mdsc;
 
-	dout("dentry_lru_touch %p %p '%.*s'\n", di, dn,
-	     dn->d_name.len, dn->d_name.name);
+	dout("dentry_lru_touch %p %p '%.*s' (offset %lld)\n", di, dn,
+	     dn->d_name.len, dn->d_name.name, di->offset);
 	if (di) {
-		mdsc = &ceph_client(dn->d_sb)->mdsc;
+		mdsc = &ceph_sb_to_client(dn->d_sb)->mdsc;
 		spin_lock(&mdsc->dentry_lru_lock);
 		list_move_tail(&di->lru, &mdsc->dentry_lru);
 		spin_unlock(&mdsc->dentry_lru_lock);
@@ -1183,7 +1197,7 @@ void ceph_dentry_lru_del(struct dentry *dn)
 	dout("dentry_lru_del %p %p '%.*s'\n", di, dn,
 	     dn->d_name.len, dn->d_name.name);
 	if (di) {
-		mdsc = &ceph_client(dn->d_sb)->mdsc;
+		mdsc = &ceph_sb_to_client(dn->d_sb)->mdsc;
 		spin_lock(&mdsc->dentry_lru_lock);
 		list_del_init(&di->lru);
 		mdsc->num_dentry--;
diff --git a/fs/ceph/export.c b/fs/ceph/export.c
index 9d67572fb328..4480cb1c63e7 100644
--- a/fs/ceph/export.c
+++ b/fs/ceph/export.c
@@ -93,11 +93,11 @@ static struct dentry *__fh_to_dentry(struct super_block *sb,
 		return ERR_PTR(-ESTALE);
 
 	dentry = d_obtain_alias(inode);
-	if (!dentry) {
+	if (IS_ERR(dentry)) {
 		pr_err("fh_to_dentry %llx -- inode %p but ENOMEM\n",
 		       fh->ino, inode);
 		iput(inode);
-		return ERR_PTR(-ENOMEM);
+		return dentry;
 	}
 	err = ceph_init_dentry(dentry);
 
@@ -115,7 +115,7 @@ static struct dentry *__fh_to_dentry(struct super_block *sb,
 static struct dentry *__cfh_to_dentry(struct super_block *sb,
 				      struct ceph_nfs_confh *cfh)
 {
-	struct ceph_mds_client *mdsc = &ceph_client(sb)->mdsc;
+	struct ceph_mds_client *mdsc = &ceph_sb_to_client(sb)->mdsc;
 	struct inode *inode;
 	struct dentry *dentry;
 	struct ceph_vino vino;
@@ -133,7 +133,7 @@ static struct dentry *__cfh_to_dentry(struct super_block *sb,
 		req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_LOOKUPHASH,
 					       USE_ANY_MDS);
 		if (IS_ERR(req))
-			return ERR_PTR(PTR_ERR(req));
+			return ERR_CAST(req);
 
 		req->r_ino1 = vino;
 		req->r_ino2.ino = cfh->parent_ino;
@@ -149,11 +149,11 @@ static struct dentry *__cfh_to_dentry(struct super_block *sb,
 	}
 
 	dentry = d_obtain_alias(inode);
-	if (!dentry) {
+	if (IS_ERR(dentry)) {
 		pr_err("cfh_to_dentry %llx -- inode %p but ENOMEM\n",
 		       cfh->ino, inode);
 		iput(inode);
-		return ERR_PTR(-ENOMEM);
+		return dentry;
 	}
 	err = ceph_init_dentry(dentry);
 	if (err < 0) {
@@ -202,11 +202,11 @@ static struct dentry *ceph_fh_to_parent(struct super_block *sb,
 		return ERR_PTR(-ESTALE);
 
 	dentry = d_obtain_alias(inode);
-	if (!dentry) {
+	if (IS_ERR(dentry)) {
 		pr_err("fh_to_parent %llx -- inode %p but ENOMEM\n",
 		       cfh->ino, inode);
 		iput(inode);
-		return ERR_PTR(-ENOMEM);
+		return dentry;
 	}
 	err = ceph_init_dentry(dentry);
 	if (err < 0) {
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 7d634938edc9..6251a1574b94 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -230,7 +230,7 @@ struct dentry *ceph_lookup_open(struct inode *dir, struct dentry *dentry,
 	/* do the open */
 	req = prepare_open_request(dir->i_sb, flags, mode);
 	if (IS_ERR(req))
-		return ERR_PTR(PTR_ERR(req));
+		return ERR_CAST(req);
 	req->r_dentry = dget(dentry);
 	req->r_num_caps = 2;
 	if (flags & O_CREAT) {
@@ -317,16 +317,16 @@ void ceph_release_page_vector(struct page **pages, int num_pages)
 /*
  * allocate a vector new pages
  */
-static struct page **alloc_page_vector(int num_pages)
+struct page **ceph_alloc_page_vector(int num_pages, gfp_t flags)
 {
 	struct page **pages;
 	int i;
 
-	pages = kmalloc(sizeof(*pages) * num_pages, GFP_NOFS);
+	pages = kmalloc(sizeof(*pages) * num_pages, flags);
 	if (!pages)
 		return ERR_PTR(-ENOMEM);
 	for (i = 0; i < num_pages; i++) {
-		pages[i] = alloc_page(GFP_NOFS);
+		pages[i] = __page_cache_alloc(flags);
 		if (pages[i] == NULL) {
 			ceph_release_page_vector(pages, i);
 			return ERR_PTR(-ENOMEM);
@@ -540,7 +540,7 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data,
 		 * in sequence.
 		 */
 	} else {
-		pages = alloc_page_vector(num_pages);
+		pages = ceph_alloc_page_vector(num_pages, GFP_NOFS);
 	}
 	if (IS_ERR(pages))
 		return PTR_ERR(pages);
@@ -649,8 +649,8 @@ more:
 				    do_sync,
 				    ci->i_truncate_seq, ci->i_truncate_size,
 				    &mtime, false, 2);
-	if (IS_ERR(req))
-		return PTR_ERR(req);
+	if (!req)
+		return -ENOMEM;
 
 	num_pages = calc_pages_for(pos, len);
 
@@ -668,7 +668,7 @@ more:
 		truncate_inode_pages_range(inode->i_mapping, pos, 
 					   (pos+len) | (PAGE_CACHE_SIZE-1));
 	} else {
-		pages = alloc_page_vector(num_pages);
+		pages = ceph_alloc_page_vector(num_pages, GFP_NOFS);
 		if (IS_ERR(pages)) {
 			ret = PTR_ERR(pages);
 			goto out;
@@ -809,7 +809,7 @@ static ssize_t ceph_aio_write(struct kiocb *iocb, const struct iovec *iov,
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file->f_dentry->d_inode;
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	struct ceph_osd_client *osdc = &ceph_client(inode->i_sb)->osdc;
+	struct ceph_osd_client *osdc = &ceph_sb_to_client(inode->i_sb)->osdc;
 	loff_t endoff = pos + iov->iov_len;
 	int got = 0;
 	int ret, err;
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 85b4d2ffdeba..226f5a50d362 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -69,7 +69,7 @@ struct inode *ceph_get_snapdir(struct inode *parent)
 
 	BUG_ON(!S_ISDIR(parent->i_mode));
 	if (IS_ERR(inode))
-		return ERR_PTR(PTR_ERR(inode));
+		return inode;
 	inode->i_mode = parent->i_mode;
 	inode->i_uid = parent->i_uid;
 	inode->i_gid = parent->i_gid;
@@ -384,7 +384,7 @@ void ceph_destroy_inode(struct inode *inode)
 	 */
 	if (ci->i_snap_realm) {
 		struct ceph_mds_client *mdsc =
-			&ceph_client(ci->vfs_inode.i_sb)->mdsc;
+			&ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
 		struct ceph_snap_realm *realm = ci->i_snap_realm;
 
 		dout(" dropping residual ref to snap realm %p\n", realm);
@@ -619,11 +619,12 @@ static int fill_inode(struct inode *inode,
 			memcpy(ci->i_xattrs.blob->vec.iov_base,
 			       iinfo->xattr_data, iinfo->xattr_len);
 		ci->i_xattrs.version = le64_to_cpu(info->xattr_version);
+		xattr_blob = NULL;
 	}
 
 	inode->i_mapping->a_ops = &ceph_aops;
 	inode->i_mapping->backing_dev_info =
-		&ceph_client(inode->i_sb)->backing_dev_info;
+		&ceph_sb_to_client(inode->i_sb)->backing_dev_info;
 
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFIFO:
@@ -674,14 +675,15 @@ static int fill_inode(struct inode *inode,
 		/* set dir completion flag? */
 		if (ci->i_files == 0 && ci->i_subdirs == 0 &&
 		    ceph_snap(inode) == CEPH_NOSNAP &&
-		    (le32_to_cpu(info->cap.caps) & CEPH_CAP_FILE_SHARED)) {
+		    (le32_to_cpu(info->cap.caps) & CEPH_CAP_FILE_SHARED) &&
+		    (ci->i_ceph_flags & CEPH_I_COMPLETE) == 0) {
 			dout(" marking %p complete (empty)\n", inode);
 			ci->i_ceph_flags |= CEPH_I_COMPLETE;
 			ci->i_max_offset = 2;
 		}
 
 		/* it may be better to set st_size in getattr instead? */
-		if (ceph_test_opt(ceph_client(inode->i_sb), RBYTES))
+		if (ceph_test_opt(ceph_sb_to_client(inode->i_sb), RBYTES))
 			inode->i_size = ci->i_rbytes;
 		break;
 	default:
@@ -802,6 +804,37 @@ out_unlock:
 }
 
 /*
+ * Set dentry's directory position based on the current dir's max, and
+ * order it in d_subdirs, so that dcache_readdir behaves.
+ */
+static void ceph_set_dentry_offset(struct dentry *dn)
+{
+	struct dentry *dir = dn->d_parent;
+	struct inode *inode = dn->d_parent->d_inode;
+	struct ceph_dentry_info *di;
+
+	BUG_ON(!inode);
+
+	di = ceph_dentry(dn);
+
+	spin_lock(&inode->i_lock);
+	if ((ceph_inode(inode)->i_ceph_flags & CEPH_I_COMPLETE) == 0) {
+		spin_unlock(&inode->i_lock);
+		return;
+	}
+	di->offset = ceph_inode(inode)->i_max_offset++;
+	spin_unlock(&inode->i_lock);
+
+	spin_lock(&dcache_lock);
+	spin_lock(&dn->d_lock);
+	list_move_tail(&dir->d_subdirs, &dn->d_u.d_child);
+	dout("set_dentry_offset %p %lld (%p %p)\n", dn, di->offset,
+	     dn->d_u.d_child.prev, dn->d_u.d_child.next);
+	spin_unlock(&dn->d_lock);
+	spin_unlock(&dcache_lock);
+}
+
+/*
  * splice a dentry to an inode.
  * caller must hold directory i_mutex for this to be safe.
  *
@@ -814,6 +847,8 @@ static struct dentry *splice_dentry(struct dentry *dn, struct inode *in,
 {
 	struct dentry *realdn;
 
+	BUG_ON(dn->d_inode);
+
 	/* dn must be unhashed */
 	if (!d_unhashed(dn))
 		d_drop(dn);
@@ -835,44 +870,17 @@ static struct dentry *splice_dentry(struct dentry *dn, struct inode *in,
 		dn = realdn;
 	} else {
 		BUG_ON(!ceph_dentry(dn));
-
 		dout("dn %p attached to %p ino %llx.%llx\n",
 		     dn, dn->d_inode, ceph_vinop(dn->d_inode));
 	}
 	if ((!prehash || *prehash) && d_unhashed(dn))
 		d_rehash(dn);
+	ceph_set_dentry_offset(dn);
 out:
 	return dn;
 }
 
 /*
- * Set dentry's directory position based on the current dir's max, and
- * order it in d_subdirs, so that dcache_readdir behaves.
- */
-static void ceph_set_dentry_offset(struct dentry *dn)
-{
-	struct dentry *dir = dn->d_parent;
-	struct inode *inode = dn->d_parent->d_inode;
-	struct ceph_dentry_info *di;
-
-	BUG_ON(!inode);
-
-	di = ceph_dentry(dn);
-
-	spin_lock(&inode->i_lock);
-	di->offset = ceph_inode(inode)->i_max_offset++;
-	spin_unlock(&inode->i_lock);
-
-	spin_lock(&dcache_lock);
-	spin_lock(&dn->d_lock);
-	list_move_tail(&dir->d_subdirs, &dn->d_u.d_child);
-	dout("set_dentry_offset %p %lld (%p %p)\n", dn, di->offset,
-	     dn->d_u.d_child.prev, dn->d_u.d_child.next);
-	spin_unlock(&dn->d_lock);
-	spin_unlock(&dcache_lock);
-}
-
-/*
  * Incorporate results into the local cache.  This is either just
  * one inode, or a directory, dentry, and possibly linked-to inode (e.g.,
  * after a lookup).
@@ -933,14 +941,8 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 
 	if (!rinfo->head->is_target && !rinfo->head->is_dentry) {
 		dout("fill_trace reply is empty!\n");
-		if (rinfo->head->result == 0 && req->r_locked_dir) {
-			struct ceph_inode_info *ci =
-				ceph_inode(req->r_locked_dir);
-			dout(" clearing %p complete (empty trace)\n",
-			     req->r_locked_dir);
-			ci->i_ceph_flags &= ~CEPH_I_COMPLETE;
-			ci->i_release_count++;
-		}
+		if (rinfo->head->result == 0 && req->r_locked_dir)
+			ceph_invalidate_dir_request(req);
 		return 0;
 	}
 
@@ -1011,13 +1013,18 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 			     req->r_old_dentry->d_name.len,
 			     req->r_old_dentry->d_name.name,
 			     dn, dn->d_name.len, dn->d_name.name);
+
 			/* ensure target dentry is invalidated, despite
 			   rehashing bug in vfs_rename_dir */
-			dn->d_time = jiffies;
-			ceph_dentry(dn)->lease_shared_gen = 0;
+			ceph_invalidate_dentry_lease(dn);
+
 			/* take overwritten dentry's readdir offset */
+			dout("dn %p gets %p offset %lld (old offset %lld)\n",
+			     req->r_old_dentry, dn, ceph_dentry(dn)->offset,
+			     ceph_dentry(req->r_old_dentry)->offset);
 			ceph_dentry(req->r_old_dentry)->offset =
 				ceph_dentry(dn)->offset;
+
 			dn = req->r_old_dentry;  /* use old_dentry */
 			in = dn->d_inode;
 		}
@@ -1059,7 +1066,6 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 				goto done;
 			}
 			req->r_dentry = dn;  /* may have spliced */
-			ceph_set_dentry_offset(dn);
 			igrab(in);
 		} else if (ceph_ino(in) == vino.ino &&
 			   ceph_snap(in) == vino.snap) {
@@ -1102,7 +1108,6 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 			err = PTR_ERR(dn);
 			goto done;
 		}
-		ceph_set_dentry_offset(dn);
 		req->r_dentry = dn;  /* may have spliced */
 		igrab(in);
 		rinfo->head->is_dentry = 1;  /* fool notrace handlers */
@@ -1429,7 +1434,7 @@ void ceph_queue_vmtruncate(struct inode *inode)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
 
-	if (queue_work(ceph_client(inode->i_sb)->trunc_wq,
+	if (queue_work(ceph_sb_to_client(inode->i_sb)->trunc_wq,
 		       &ci->i_vmtruncate_work)) {
 		dout("ceph_queue_vmtruncate %p\n", inode);
 		igrab(inode);
@@ -1518,7 +1523,7 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr)
 	struct inode *parent_inode = dentry->d_parent->d_inode;
 	const unsigned int ia_valid = attr->ia_valid;
 	struct ceph_mds_request *req;
-	struct ceph_mds_client *mdsc = &ceph_client(dentry->d_sb)->mdsc;
+	struct ceph_mds_client *mdsc = &ceph_sb_to_client(dentry->d_sb)->mdsc;
 	int issued;
 	int release = 0, dirtied = 0;
 	int mask = 0;
diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index 8a5bcae62846..d085f07756b4 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -98,7 +98,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
 	struct ceph_ioctl_dataloc dl;
 	struct inode *inode = file->f_dentry->d_inode;
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	struct ceph_osd_client *osdc = &ceph_client(inode->i_sb)->osdc;
+	struct ceph_osd_client *osdc = &ceph_sb_to_client(inode->i_sb)->osdc;
 	u64 len = 1, olen;
 	u64 tmp;
 	struct ceph_object_layout ol;
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 24561a557e01..b49f12822cbc 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -40,7 +40,7 @@
 static void __wake_requests(struct ceph_mds_client *mdsc,
 			    struct list_head *head);
 
-const static struct ceph_connection_operations mds_con_ops;
+static const struct ceph_connection_operations mds_con_ops;
 
 
 /*
@@ -665,10 +665,10 @@ static struct ceph_msg *create_session_msg(u32 op, u64 seq)
 	struct ceph_msg *msg;
 	struct ceph_mds_session_head *h;
 
-	msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), 0, 0, NULL);
-	if (IS_ERR(msg)) {
+	msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), GFP_NOFS);
+	if (!msg) {
 		pr_err("create_session_msg ENOMEM creating msg\n");
-		return ERR_PTR(PTR_ERR(msg));
+		return NULL;
 	}
 	h = msg->front.iov_base;
 	h->op = cpu_to_le32(op);
@@ -687,7 +687,6 @@ static int __open_session(struct ceph_mds_client *mdsc,
 	struct ceph_msg *msg;
 	int mstate;
 	int mds = session->s_mds;
-	int err = 0;
 
 	/* wait for mds to go active? */
 	mstate = ceph_mdsmap_get_state(mdsc->mdsmap, mds);
@@ -698,13 +697,9 @@ static int __open_session(struct ceph_mds_client *mdsc,
 
 	/* send connect message */
 	msg = create_session_msg(CEPH_SESSION_REQUEST_OPEN, session->s_seq);
-	if (IS_ERR(msg)) {
-		err = PTR_ERR(msg);
-		goto out;
-	}
+	if (!msg)
+		return -ENOMEM;
 	ceph_con_send(&session->s_con, msg);
-
-out:
 	return 0;
 }
 
@@ -804,12 +799,49 @@ out:
 }
 
 static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
-				   void *arg)
+				  void *arg)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
+	int drop = 0;
+
 	dout("removing cap %p, ci is %p, inode is %p\n",
 	     cap, ci, &ci->vfs_inode);
-	ceph_remove_cap(cap);
+	spin_lock(&inode->i_lock);
+	__ceph_remove_cap(cap);
+	if (!__ceph_is_any_real_caps(ci)) {
+		struct ceph_mds_client *mdsc =
+			&ceph_sb_to_client(inode->i_sb)->mdsc;
+
+		spin_lock(&mdsc->cap_dirty_lock);
+		if (!list_empty(&ci->i_dirty_item)) {
+			pr_info(" dropping dirty %s state for %p %lld\n",
+				ceph_cap_string(ci->i_dirty_caps),
+				inode, ceph_ino(inode));
+			ci->i_dirty_caps = 0;
+			list_del_init(&ci->i_dirty_item);
+			drop = 1;
+		}
+		if (!list_empty(&ci->i_flushing_item)) {
+			pr_info(" dropping dirty+flushing %s state for %p %lld\n",
+				ceph_cap_string(ci->i_flushing_caps),
+				inode, ceph_ino(inode));
+			ci->i_flushing_caps = 0;
+			list_del_init(&ci->i_flushing_item);
+			mdsc->num_cap_flushing--;
+			drop = 1;
+		}
+		if (drop && ci->i_wrbuffer_ref) {
+			pr_info(" dropping dirty data for %p %lld\n",
+				inode, ceph_ino(inode));
+			ci->i_wrbuffer_ref = 0;
+			ci->i_wrbuffer_ref_head = 0;
+			drop++;
+		}
+		spin_unlock(&mdsc->cap_dirty_lock);
+	}
+	spin_unlock(&inode->i_lock);
+	while (drop--)
+		iput(inode);
 	return 0;
 }
 
@@ -821,6 +853,7 @@ static void remove_session_caps(struct ceph_mds_session *session)
 	dout("remove_session_caps on %p\n", session);
 	iterate_session_caps(session, remove_session_caps_cb, NULL);
 	BUG_ON(session->s_nr_caps > 0);
+	BUG_ON(!list_empty(&session->s_cap_flushing));
 	cleanup_cap_releases(session);
 }
 
@@ -883,8 +916,8 @@ static int send_renew_caps(struct ceph_mds_client *mdsc,
 		ceph_mds_state_name(state));
 	msg = create_session_msg(CEPH_SESSION_REQUEST_RENEWCAPS,
 				 ++session->s_renew_seq);
-	if (IS_ERR(msg))
-		return PTR_ERR(msg);
+	if (!msg)
+		return -ENOMEM;
 	ceph_con_send(&session->s_con, msg);
 	return 0;
 }
@@ -931,17 +964,15 @@ static int request_close_session(struct ceph_mds_client *mdsc,
 				 struct ceph_mds_session *session)
 {
 	struct ceph_msg *msg;
-	int err = 0;
 
 	dout("request_close_session mds%d state %s seq %lld\n",
 	     session->s_mds, session_state_name(session->s_state),
 	     session->s_seq);
 	msg = create_session_msg(CEPH_SESSION_REQUEST_CLOSE, session->s_seq);
-	if (IS_ERR(msg))
-		err = PTR_ERR(msg);
-	else
-		ceph_con_send(&session->s_con, msg);
-	return err;
+	if (!msg)
+		return -ENOMEM;
+	ceph_con_send(&session->s_con, msg);
+	return 0;
 }
 
 /*
@@ -1059,7 +1090,7 @@ static int add_cap_releases(struct ceph_mds_client *mdsc,
 	while (session->s_num_cap_releases < session->s_nr_caps + extra) {
 		spin_unlock(&session->s_cap_lock);
 		msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPRELEASE, PAGE_CACHE_SIZE,
-				   0, 0, NULL);
+				   GFP_NOFS);
 		if (!msg)
 			goto out_unlocked;
 		dout("add_cap_releases %p msg %p now %d\n", session, msg,
@@ -1151,10 +1182,8 @@ static void send_cap_releases(struct ceph_mds_client *mdsc,
 	struct ceph_msg *msg;
 
 	dout("send_cap_releases mds%d\n", session->s_mds);
-	while (1) {
-		spin_lock(&session->s_cap_lock);
-		if (list_empty(&session->s_cap_releases_done))
-			break;
+	spin_lock(&session->s_cap_lock);
+	while (!list_empty(&session->s_cap_releases_done)) {
 		msg = list_first_entry(&session->s_cap_releases_done,
 				 struct ceph_msg, list_head);
 		list_del_init(&msg->list_head);
@@ -1162,10 +1191,49 @@ static void send_cap_releases(struct ceph_mds_client *mdsc,
 		msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
 		dout("send_cap_releases mds%d %p\n", session->s_mds, msg);
 		ceph_con_send(&session->s_con, msg);
+		spin_lock(&session->s_cap_lock);
 	}
 	spin_unlock(&session->s_cap_lock);
 }
 
+static void discard_cap_releases(struct ceph_mds_client *mdsc,
+				 struct ceph_mds_session *session)
+{
+	struct ceph_msg *msg;
+	struct ceph_mds_cap_release *head;
+	unsigned num;
+
+	dout("discard_cap_releases mds%d\n", session->s_mds);
+	spin_lock(&session->s_cap_lock);
+
+	/* zero out the in-progress message */
+	msg = list_first_entry(&session->s_cap_releases,
+			       struct ceph_msg, list_head);
+	head = msg->front.iov_base;
+	num = le32_to_cpu(head->num);
+	dout("discard_cap_releases mds%d %p %u\n", session->s_mds, msg, num);
+	head->num = cpu_to_le32(0);
+	session->s_num_cap_releases += num;
+
+	/* requeue completed messages */
+	while (!list_empty(&session->s_cap_releases_done)) {
+		msg = list_first_entry(&session->s_cap_releases_done,
+				 struct ceph_msg, list_head);
+		list_del_init(&msg->list_head);
+
+		head = msg->front.iov_base;
+		num = le32_to_cpu(head->num);
+		dout("discard_cap_releases mds%d %p %u\n", session->s_mds, msg,
+		     num);
+		session->s_num_cap_releases += num;
+		head->num = cpu_to_le32(0);
+		msg->front.iov_len = sizeof(*head);
+		list_add(&msg->list_head, &session->s_cap_releases);
+	}
+
+	spin_unlock(&session->s_cap_lock);
+}
+
 /*
  * requests
  */
@@ -1181,6 +1249,7 @@ ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode)
 	if (!req)
 		return ERR_PTR(-ENOMEM);
 
+	mutex_init(&req->r_fill_mutex);
 	req->r_started = jiffies;
 	req->r_resend_mds = -1;
 	INIT_LIST_HEAD(&req->r_unsafe_dir_item);
@@ -1251,7 +1320,7 @@ retry:
 			len += 1 + temp->d_name.len;
 		temp = temp->d_parent;
 		if (temp == NULL) {
-			pr_err("build_path_dentry corrupt dentry %p\n", dentry);
+			pr_err("build_path corrupt dentry %p\n", dentry);
 			return ERR_PTR(-EINVAL);
 		}
 	}
@@ -1267,7 +1336,7 @@ retry:
 		struct inode *inode = temp->d_inode;
 
 		if (inode && ceph_snap(inode) == CEPH_SNAPDIR) {
-			dout("build_path_dentry path+%d: %p SNAPDIR\n",
+			dout("build_path path+%d: %p SNAPDIR\n",
 			     pos, temp);
 		} else if (stop_on_nosnap && inode &&
 			   ceph_snap(inode) == CEPH_NOSNAP) {
@@ -1278,20 +1347,18 @@ retry:
 				break;
 			strncpy(path + pos, temp->d_name.name,
 				temp->d_name.len);
-			dout("build_path_dentry path+%d: %p '%.*s'\n",
-			     pos, temp, temp->d_name.len, path + pos);
 		}
 		if (pos)
 			path[--pos] = '/';
 		temp = temp->d_parent;
 		if (temp == NULL) {
-			pr_err("build_path_dentry corrupt dentry\n");
+			pr_err("build_path corrupt dentry\n");
 			kfree(path);
 			return ERR_PTR(-EINVAL);
 		}
 	}
 	if (pos != 0) {
-		pr_err("build_path_dentry did not end path lookup where "
+		pr_err("build_path did not end path lookup where "
 		       "expected, namelen is %d, pos is %d\n", len, pos);
 		/* presumably this is only possible if racing with a
 		   rename of one of the parent directories (we can not
@@ -1303,7 +1370,7 @@ retry:
 
 	*base = ceph_ino(temp->d_inode);
 	*plen = len;
-	dout("build_path_dentry on %p %d built %llx '%.*s'\n",
+	dout("build_path on %p %d built %llx '%.*s'\n",
 	     dentry, atomic_read(&dentry->d_count), *base, len, path);
 	return path;
 }
@@ -1426,9 +1493,11 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc,
 	if (req->r_old_dentry_drop)
 		len += req->r_old_dentry->d_name.len;
 
-	msg = ceph_msg_new(CEPH_MSG_CLIENT_REQUEST, len, 0, 0, NULL);
-	if (IS_ERR(msg))
+	msg = ceph_msg_new(CEPH_MSG_CLIENT_REQUEST, len, GFP_NOFS);
+	if (!msg) {
+		msg = ERR_PTR(-ENOMEM);
 		goto out_free2;
+	}
 
 	msg->hdr.tid = cpu_to_le64(req->r_tid);
 
@@ -1517,9 +1586,9 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc,
 	}
 	msg = create_request_message(mdsc, req, mds);
 	if (IS_ERR(msg)) {
-		req->r_reply = ERR_PTR(PTR_ERR(msg));
+		req->r_err = PTR_ERR(msg);
 		complete_request(mdsc, req);
-		return -PTR_ERR(msg);
+		return PTR_ERR(msg);
 	}
 	req->r_request = msg;
 
@@ -1552,7 +1621,7 @@ static int __do_request(struct ceph_mds_client *mdsc,
 	int mds = -1;
 	int err = -EAGAIN;
 
-	if (req->r_reply)
+	if (req->r_err || req->r_got_result)
 		goto out;
 
 	if (req->r_timeout &&
@@ -1609,7 +1678,7 @@ out:
 	return err;
 
 finish:
-	req->r_reply = ERR_PTR(err);
+	req->r_err = err;
 	complete_request(mdsc, req);
 	goto out;
 }
@@ -1630,10 +1699,9 @@ static void __wake_requests(struct ceph_mds_client *mdsc,
 
 /*
  * Wake up threads with requests pending for @mds, so that they can
- * resubmit their requests to a possibly different mds.  If @all is set,
- * wake up if their requests has been forwarded to @mds, too.
+ * resubmit their requests to a possibly different mds.
  */
-static void kick_requests(struct ceph_mds_client *mdsc, int mds, int all)
+static void kick_requests(struct ceph_mds_client *mdsc, int mds)
 {
 	struct ceph_mds_request *req;
 	struct rb_node *p;
@@ -1689,64 +1757,78 @@ int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
 	__register_request(mdsc, req, dir);
 	__do_request(mdsc, req);
 
-	/* wait */
-	if (!req->r_reply) {
-		mutex_unlock(&mdsc->mutex);
-		if (req->r_timeout) {
-			err = (long)wait_for_completion_interruptible_timeout(
-				&req->r_completion, req->r_timeout);
-			if (err == 0)
-				req->r_reply = ERR_PTR(-EIO);
-			else if (err < 0)
-				req->r_reply = ERR_PTR(err);
-		} else {
-                        err = wait_for_completion_interruptible(
-                                &req->r_completion);
-                        if (err)
-                                req->r_reply = ERR_PTR(err);
-		}
-		mutex_lock(&mdsc->mutex);
+	if (req->r_err) {
+		err = req->r_err;
+		__unregister_request(mdsc, req);
+		dout("do_request early error %d\n", err);
+		goto out;
 	}
 
-	if (IS_ERR(req->r_reply)) {
-		err = PTR_ERR(req->r_reply);
-		req->r_reply = NULL;
+	/* wait */
+	mutex_unlock(&mdsc->mutex);
+	dout("do_request waiting\n");
+	if (req->r_timeout) {
+		err = (long)wait_for_completion_killable_timeout(
+			&req->r_completion, req->r_timeout);
+		if (err == 0)
+			err = -EIO;
+	} else {
+		err = wait_for_completion_killable(&req->r_completion);
+	}
+	dout("do_request waited, got %d\n", err);
+	mutex_lock(&mdsc->mutex);
 
-		if (err == -ERESTARTSYS) {
-			/* aborted */
-			req->r_aborted = true;
+	/* only abort if we didn't race with a real reply */
+	if (req->r_got_result) {
+		err = le32_to_cpu(req->r_reply_info.head->result);
+	} else if (err < 0) {
+		dout("aborted request %lld with %d\n", req->r_tid, err);
 
-			if (req->r_locked_dir &&
-			    (req->r_op & CEPH_MDS_OP_WRITE)) {
-				struct ceph_inode_info *ci =
-					ceph_inode(req->r_locked_dir);
+		/*
+		 * ensure we aren't running concurrently with
+		 * ceph_fill_trace or ceph_readdir_prepopulate, which
+		 * rely on locks (dir mutex) held by our caller.
+		 */
+		mutex_lock(&req->r_fill_mutex);
+		req->r_err = err;
+		req->r_aborted = true;
+		mutex_unlock(&req->r_fill_mutex);
 
-				dout("aborted, clearing I_COMPLETE on %p\n", 
-				     req->r_locked_dir);
-				spin_lock(&req->r_locked_dir->i_lock);
-				ci->i_ceph_flags &= ~CEPH_I_COMPLETE;
-				ci->i_release_count++;
-				spin_unlock(&req->r_locked_dir->i_lock);
-			}
-		} else {
-			/* clean up this request */
-			__unregister_request(mdsc, req);
-			if (!list_empty(&req->r_unsafe_item))
-				list_del_init(&req->r_unsafe_item);
-			complete(&req->r_safe_completion);
-		}
-	} else if (req->r_err) {
-		err = req->r_err;
+		if (req->r_locked_dir &&
+		    (req->r_op & CEPH_MDS_OP_WRITE))
+			ceph_invalidate_dir_request(req);
 	} else {
-		err = le32_to_cpu(req->r_reply_info.head->result);
+		err = req->r_err;
 	}
-	mutex_unlock(&mdsc->mutex);
 
+out:
+	mutex_unlock(&mdsc->mutex);
 	dout("do_request %p done, result %d\n", req, err);
 	return err;
 }
 
 /*
+ * Invalidate dir I_COMPLETE, dentry lease state on an aborted MDS
+ * namespace request.
+ */
+void ceph_invalidate_dir_request(struct ceph_mds_request *req)
+{
+	struct inode *inode = req->r_locked_dir;
+	struct ceph_inode_info *ci = ceph_inode(inode);
+
+	dout("invalidate_dir_request %p (I_COMPLETE, lease(s))\n", inode);
+	spin_lock(&inode->i_lock);
+	ci->i_ceph_flags &= ~CEPH_I_COMPLETE;
+	ci->i_release_count++;
+	spin_unlock(&inode->i_lock);
+
+	if (req->r_dentry)
+		ceph_invalidate_dentry_lease(req->r_dentry);
+	if (req->r_old_dentry)
+		ceph_invalidate_dentry_lease(req->r_old_dentry);
+}
+
+/*
  * Handle mds reply.
  *
  * We take the session mutex and parse and process the reply immediately.
@@ -1797,6 +1879,12 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
 		mutex_unlock(&mdsc->mutex);
 		goto out;
 	}
+	if (req->r_got_safe && !head->safe) {
+		pr_warning("got unsafe after safe on %llu from mds%d\n",
+			   tid, mds);
+		mutex_unlock(&mdsc->mutex);
+		goto out;
+	}
 
 	result = le32_to_cpu(head->result);
 
@@ -1838,11 +1926,7 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
 			mutex_unlock(&mdsc->mutex);
 			goto out;
 		}
-	}
-
-	BUG_ON(req->r_reply);
-
-	if (!head->safe) {
+	} else {
 		req->r_got_unsafe = true;
 		list_add_tail(&req->r_unsafe_item, &req->r_session->s_unsafe);
 	}
@@ -1871,21 +1955,30 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
 	}
 
 	/* insert trace into our cache */
+	mutex_lock(&req->r_fill_mutex);
 	err = ceph_fill_trace(mdsc->client->sb, req, req->r_session);
 	if (err == 0) {
 		if (result == 0 && rinfo->dir_nr)
 			ceph_readdir_prepopulate(req, req->r_session);
 		ceph_unreserve_caps(&req->r_caps_reservation);
 	}
+	mutex_unlock(&req->r_fill_mutex);
 
 	up_read(&mdsc->snap_rwsem);
 out_err:
-	if (err) {
-		req->r_err = err;
+	mutex_lock(&mdsc->mutex);
+	if (!req->r_aborted) {
+		if (err) {
+			req->r_err = err;
+		} else {
+			req->r_reply = msg;
+			ceph_msg_get(msg);
+			req->r_got_result = true;
+		}
 	} else {
-		req->r_reply = msg;
-		ceph_msg_get(msg);
+		dout("reply arrived after request %lld was aborted\n", tid);
 	}
+	mutex_unlock(&mdsc->mutex);
 
 	add_cap_releases(mdsc, req->r_session, -1);
 	mutex_unlock(&session->s_mutex);
@@ -1921,16 +2014,21 @@ static void handle_forward(struct ceph_mds_client *mdsc,
 	mutex_lock(&mdsc->mutex);
 	req = __lookup_request(mdsc, tid);
 	if (!req) {
-		dout("forward %llu to mds%d - req dne\n", tid, next_mds);
+		dout("forward tid %llu to mds%d - req dne\n", tid, next_mds);
 		goto out;  /* dup reply? */
 	}
 
-	if (fwd_seq <= req->r_num_fwd) {
-		dout("forward %llu to mds%d - old seq %d <= %d\n",
+	if (req->r_aborted) {
+		dout("forward tid %llu aborted, unregistering\n", tid);
+		__unregister_request(mdsc, req);
+	} else if (fwd_seq <= req->r_num_fwd) {
+		dout("forward tid %llu to mds%d - old seq %d <= %d\n",
 		     tid, next_mds, req->r_num_fwd, fwd_seq);
 	} else {
 		/* resend. forward race not possible; mds would drop */
-		dout("forward %llu to mds%d (we resend)\n", tid, next_mds);
+		dout("forward tid %llu to mds%d (we resend)\n", tid, next_mds);
+		BUG_ON(req->r_err);
+		BUG_ON(req->r_got_result);
 		req->r_num_fwd = fwd_seq;
 		req->r_resend_mds = next_mds;
 		put_request_session(req);
@@ -1984,6 +2082,8 @@ static void handle_session(struct ceph_mds_session *session,
 
 	switch (op) {
 	case CEPH_SESSION_OPEN:
+		if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
+			pr_info("mds%d reconnect success\n", session->s_mds);
 		session->s_state = CEPH_MDS_SESSION_OPEN;
 		renewed_caps(mdsc, session, 0);
 		wake = 1;
@@ -1997,10 +2097,12 @@ static void handle_session(struct ceph_mds_session *session,
 		break;
 
 	case CEPH_SESSION_CLOSE:
+		if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
+			pr_info("mds%d reconnect denied\n", session->s_mds);
 		remove_session_caps(session);
 		wake = 1; /* for good measure */
 		complete(&mdsc->session_close_waiters);
-		kick_requests(mdsc, mds, 0);      /* cur only */
+		kick_requests(mdsc, mds);
 		break;
 
 	case CEPH_SESSION_STALE:
@@ -2132,54 +2234,44 @@ out:
  *
  * called with mdsc->mutex held.
  */
-static void send_mds_reconnect(struct ceph_mds_client *mdsc, int mds)
+static void send_mds_reconnect(struct ceph_mds_client *mdsc,
+			       struct ceph_mds_session *session)
 {
-	struct ceph_mds_session *session = NULL;
 	struct ceph_msg *reply;
 	struct rb_node *p;
+	int mds = session->s_mds;
 	int err = -ENOMEM;
 	struct ceph_pagelist *pagelist;
 
-	pr_info("reconnect to recovering mds%d\n", mds);
+	pr_info("mds%d reconnect start\n", mds);
 
 	pagelist = kmalloc(sizeof(*pagelist), GFP_NOFS);
 	if (!pagelist)
 		goto fail_nopagelist;
 	ceph_pagelist_init(pagelist);
 
-	reply = ceph_msg_new(CEPH_MSG_CLIENT_RECONNECT, 0, 0, 0, NULL);
-	if (IS_ERR(reply)) {
-		err = PTR_ERR(reply);
+	reply = ceph_msg_new(CEPH_MSG_CLIENT_RECONNECT, 0, GFP_NOFS);
+	if (!reply)
 		goto fail_nomsg;
-	}
-
-	/* find session */
-	session = __ceph_lookup_mds_session(mdsc, mds);
-	mutex_unlock(&mdsc->mutex);    /* drop lock for duration */
 
-	if (session) {
-		mutex_lock(&session->s_mutex);
+	mutex_lock(&session->s_mutex);
+	session->s_state = CEPH_MDS_SESSION_RECONNECTING;
+	session->s_seq = 0;
 
-		session->s_state = CEPH_MDS_SESSION_RECONNECTING;
-		session->s_seq = 0;
+	ceph_con_open(&session->s_con,
+		      ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
 
-		ceph_con_open(&session->s_con,
-			      ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
-
-		/* replay unsafe requests */
-		replay_unsafe_requests(mdsc, session);
-	} else {
-		dout("no session for mds%d, will send short reconnect\n",
-		     mds);
-	}
+	/* replay unsafe requests */
+	replay_unsafe_requests(mdsc, session);
 
 	down_read(&mdsc->snap_rwsem);
 
-	if (!session)
-		goto send;
 	dout("session %p state %s\n", session,
 	     session_state_name(session->s_state));
 
+	/* drop old cap expires; we're about to reestablish that state */
+	discard_cap_releases(mdsc, session);
+
 	/* traverse this session's caps */
 	err = ceph_pagelist_encode_32(pagelist, session->s_nr_caps);
 	if (err)
@@ -2208,36 +2300,29 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, int mds)
 			goto fail;
 	}
 
-send:
 	reply->pagelist = pagelist;
 	reply->hdr.data_len = cpu_to_le32(pagelist->length);
 	reply->nr_pages = calc_pages_for(0, pagelist->length);
 	ceph_con_send(&session->s_con, reply);
 
-	session->s_state = CEPH_MDS_SESSION_OPEN;
 	mutex_unlock(&session->s_mutex);
 
 	mutex_lock(&mdsc->mutex);
 	__wake_requests(mdsc, &session->s_waiting);
 	mutex_unlock(&mdsc->mutex);
 
-	ceph_put_mds_session(session);
-
 	up_read(&mdsc->snap_rwsem);
-	mutex_lock(&mdsc->mutex);
 	return;
 
 fail:
 	ceph_msg_put(reply);
 	up_read(&mdsc->snap_rwsem);
 	mutex_unlock(&session->s_mutex);
-	ceph_put_mds_session(session);
 fail_nomsg:
 	ceph_pagelist_release(pagelist);
 	kfree(pagelist);
 fail_nopagelist:
 	pr_err("error %d preparing reconnect for mds%d\n", err, mds);
-	mutex_lock(&mdsc->mutex);
 	return;
 }
 
@@ -2290,7 +2375,7 @@ static void check_new_map(struct ceph_mds_client *mdsc,
 			}
 
 			/* kick any requests waiting on the recovering mds */
-			kick_requests(mdsc, i, 1);
+			kick_requests(mdsc, i);
 		} else if (oldstate == newstate) {
 			continue;  /* nothing new with this mds */
 		}
@@ -2299,22 +2384,21 @@ static void check_new_map(struct ceph_mds_client *mdsc,
 		 * send reconnect?
 		 */
 		if (s->s_state == CEPH_MDS_SESSION_RESTARTING &&
-		    newstate >= CEPH_MDS_STATE_RECONNECT)
-			send_mds_reconnect(mdsc, i);
+		    newstate >= CEPH_MDS_STATE_RECONNECT) {
+			mutex_unlock(&mdsc->mutex);
+			send_mds_reconnect(mdsc, s);
+			mutex_lock(&mdsc->mutex);
+		}
 
 		/*
-		 * kick requests on any mds that has gone active.
-		 *
-		 * kick requests on cur or forwarder: we may have sent
-		 * the request to mds1, mds1 told us it forwarded it
-		 * to mds2, but then we learn mds1 failed and can't be
-		 * sure it successfully forwarded our request before
-		 * it died.
+		 * kick request on any mds that has gone active.
 		 */
 		if (oldstate < CEPH_MDS_STATE_ACTIVE &&
 		    newstate >= CEPH_MDS_STATE_ACTIVE) {
-			pr_info("mds%d reconnect completed\n", s->s_mds);
-			kick_requests(mdsc, i, 1);
+			if (oldstate != CEPH_MDS_STATE_CREATING &&
+			    oldstate != CEPH_MDS_STATE_STARTING)
+				pr_info("mds%d recovery completed\n", s->s_mds);
+			kick_requests(mdsc, i);
 			ceph_kick_flushing_caps(mdsc, s);
 			wake_up_session_caps(s, 1);
 		}
@@ -2457,12 +2541,12 @@ void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
 	dnamelen = dentry->d_name.len;
 	len += dnamelen;
 
-	msg = ceph_msg_new(CEPH_MSG_CLIENT_LEASE, len, 0, 0, NULL);
-	if (IS_ERR(msg))
+	msg = ceph_msg_new(CEPH_MSG_CLIENT_LEASE, len, GFP_NOFS);
+	if (!msg)
 		return;
 	lease = msg->front.iov_base;
 	lease->action = action;
-	lease->mask = cpu_to_le16(CEPH_LOCK_DN);
+	lease->mask = cpu_to_le16(1);
 	lease->ino = cpu_to_le64(ceph_vino(inode).ino);
 	lease->first = lease->last = cpu_to_le64(ceph_vino(inode).snap);
 	lease->seq = cpu_to_le32(seq);
@@ -2492,7 +2576,7 @@ void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode,
 
 	BUG_ON(inode == NULL);
 	BUG_ON(dentry == NULL);
-	BUG_ON(mask != CEPH_LOCK_DN);
+	BUG_ON(mask == 0);
 
 	/* is dentry lease valid? */
 	spin_lock(&dentry->d_lock);
@@ -2603,7 +2687,9 @@ static void delayed_work(struct work_struct *work)
 		else
 			ceph_con_keepalive(&s->s_con);
 		add_cap_releases(mdsc, s, -1);
-		send_cap_releases(mdsc, s);
+		if (s->s_state == CEPH_MDS_SESSION_OPEN ||
+		    s->s_state == CEPH_MDS_SESSION_HUNG)
+			send_cap_releases(mdsc, s);
 		mutex_unlock(&s->s_mutex);
 		ceph_put_mds_session(s);
 
@@ -2620,6 +2706,9 @@ int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client)
 	mdsc->client = client;
 	mutex_init(&mdsc->mutex);
 	mdsc->mdsmap = kzalloc(sizeof(*mdsc->mdsmap), GFP_NOFS);
+	if (mdsc->mdsmap == NULL)
+		return -ENOMEM;
+
 	init_completion(&mdsc->safe_umount_waiters);
 	init_completion(&mdsc->session_close_waiters);
 	INIT_LIST_HEAD(&mdsc->waiting_for_map);
@@ -2645,6 +2734,7 @@ int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client)
 	init_waitqueue_head(&mdsc->cap_flushing_wq);
 	spin_lock_init(&mdsc->dentry_lru_lock);
 	INIT_LIST_HEAD(&mdsc->dentry_lru);
+
 	return 0;
 }
 
@@ -2740,6 +2830,9 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc)
 {
 	u64 want_tid, want_flush;
 
+	if (mdsc->client->mount_state == CEPH_MOUNT_SHUTDOWN)
+		return;
+
 	dout("sync\n");
 	mutex_lock(&mdsc->mutex);
 	want_tid = mdsc->last_tid;
@@ -2922,9 +3015,10 @@ static void con_put(struct ceph_connection *con)
 static void peer_reset(struct ceph_connection *con)
 {
 	struct ceph_mds_session *s = con->private;
+	struct ceph_mds_client *mdsc = s->s_mdsc;
 
-	pr_err("mds%d gave us the boot.  IMPLEMENT RECONNECT.\n",
-	       s->s_mds);
+	pr_warning("mds%d closed our session\n", s->s_mds);
+	send_mds_reconnect(mdsc, s);
 }
 
 static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
@@ -3031,7 +3125,7 @@ static int invalidate_authorizer(struct ceph_connection *con)
 	return ceph_monc_validate_auth(&mdsc->client->monc);
 }
 
-const static struct ceph_connection_operations mds_con_ops = {
+static const struct ceph_connection_operations mds_con_ops = {
 	.get = con_get,
 	.put = con_put,
 	.dispatch = dispatch,
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
index 961cc6f65878..d9936c4f1212 100644
--- a/fs/ceph/mds_client.h
+++ b/fs/ceph/mds_client.h
@@ -165,6 +165,8 @@ struct ceph_mds_request {
 	struct inode *r_locked_dir; /* dir (if any) i_mutex locked by vfs */
 	struct inode *r_target_inode;       /* resulting inode */
 
+	struct mutex r_fill_mutex;
+
 	union ceph_mds_request_args r_args;
 	int r_fmode;        /* file mode, if expecting cap */
 
@@ -213,7 +215,7 @@ struct ceph_mds_request {
 	struct completion r_safe_completion;
 	ceph_mds_request_callback_t r_callback;
 	struct list_head  r_unsafe_item;  /* per-session unsafe list item */
-	bool		  r_got_unsafe, r_got_safe;
+	bool		  r_got_unsafe, r_got_safe, r_got_result;
 
 	bool              r_did_prepopulate;
 	u32               r_readdir_offset;
@@ -301,6 +303,8 @@ extern void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc,
 				    struct inode *inode,
 				    struct dentry *dn, int mask);
 
+extern void ceph_invalidate_dir_request(struct ceph_mds_request *req);
+
 extern struct ceph_mds_request *
 ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode);
 extern void ceph_mdsc_submit_request(struct ceph_mds_client *mdsc,
diff --git a/fs/ceph/messenger.c b/fs/ceph/messenger.c
index cd4fadb6491a..64b8b1f7863d 100644
--- a/fs/ceph/messenger.c
+++ b/fs/ceph/messenger.c
@@ -39,18 +39,6 @@ static void queue_con(struct ceph_connection *con);
 static void con_work(struct work_struct *);
 static void ceph_fault(struct ceph_connection *con);
 
-const char *ceph_name_type_str(int t)
-{
-	switch (t) {
-	case CEPH_ENTITY_TYPE_MON: return "mon";
-	case CEPH_ENTITY_TYPE_MDS: return "mds";
-	case CEPH_ENTITY_TYPE_OSD: return "osd";
-	case CEPH_ENTITY_TYPE_CLIENT: return "client";
-	case CEPH_ENTITY_TYPE_ADMIN: return "admin";
-	default: return "???";
-	}
-}
-
 /*
  * nicely render a sockaddr as a string.
  */
@@ -132,6 +120,12 @@ void ceph_msgr_exit(void)
 	destroy_workqueue(ceph_msgr_wq);
 }
 
+void ceph_msgr_flush()
+{
+	flush_workqueue(ceph_msgr_wq);
+}
+
+
 /*
  * socket callback functions
  */
@@ -340,6 +334,7 @@ static void reset_connection(struct ceph_connection *con)
 		ceph_msg_put(con->out_msg);
 		con->out_msg = NULL;
 	}
+	con->out_keepalive_pending = false;
 	con->in_seq = 0;
 	con->in_seq_acked = 0;
 }
@@ -357,6 +352,7 @@ void ceph_con_close(struct ceph_connection *con)
 	clear_bit(WRITE_PENDING, &con->state);
 	mutex_lock(&con->mutex);
 	reset_connection(con);
+	con->peer_global_seq = 0;
 	cancel_delayed_work(&con->work);
 	mutex_unlock(&con->mutex);
 	queue_con(con);
@@ -661,7 +657,7 @@ static void prepare_write_connect(struct ceph_messenger *msgr,
 	dout("prepare_write_connect %p cseq=%d gseq=%d proto=%d\n", con,
 	     con->connect_seq, global_seq, proto);
 
-	con->out_connect.features = CEPH_FEATURE_SUPPORTED;
+	con->out_connect.features = CEPH_FEATURE_SUPPORTED_CLIENT;
 	con->out_connect.host_type = cpu_to_le32(CEPH_ENTITY_TYPE_CLIENT);
 	con->out_connect.connect_seq = cpu_to_le32(con->connect_seq);
 	con->out_connect.global_seq = cpu_to_le32(global_seq);
@@ -1124,8 +1120,8 @@ static void fail_protocol(struct ceph_connection *con)
 
 static int process_connect(struct ceph_connection *con)
 {
-	u64 sup_feat = CEPH_FEATURE_SUPPORTED;
-	u64 req_feat = CEPH_FEATURE_REQUIRED;
+	u64 sup_feat = CEPH_FEATURE_SUPPORTED_CLIENT;
+	u64 req_feat = CEPH_FEATURE_REQUIRED_CLIENT;
 	u64 server_feat = le64_to_cpu(con->in_reply.features);
 
 	dout("process_connect on %p tag %d\n", con, (int)con->in_tag);
@@ -1233,6 +1229,7 @@ static int process_connect(struct ceph_connection *con)
 		clear_bit(CONNECTING, &con->state);
 		con->peer_global_seq = le32_to_cpu(con->in_reply.global_seq);
 		con->connect_seq++;
+		con->peer_features = server_feat;
 		dout("process_connect got READY gseq %d cseq %d (%d)\n",
 		     con->peer_global_seq,
 		     le32_to_cpu(con->in_reply.connect_seq),
@@ -1402,19 +1399,17 @@ static int read_partial_message(struct ceph_connection *con)
 		con->in_msg = ceph_alloc_msg(con, &con->in_hdr, &skip);
 		if (skip) {
 			/* skip this message */
-			dout("alloc_msg returned NULL, skipping message\n");
+			dout("alloc_msg said skip message\n");
 			con->in_base_pos = -front_len - middle_len - data_len -
 				sizeof(m->footer);
 			con->in_tag = CEPH_MSGR_TAG_READY;
 			con->in_seq++;
 			return 0;
 		}
-		if (IS_ERR(con->in_msg)) {
-			ret = PTR_ERR(con->in_msg);
-			con->in_msg = NULL;
+		if (!con->in_msg) {
 			con->error_msg =
 				"error allocating memory for incoming message";
-			return ret;
+			return -ENOMEM;
 		}
 		m = con->in_msg;
 		m->front.iov_len = 0;    /* haven't read it yet */
@@ -1514,14 +1509,14 @@ static void process_message(struct ceph_connection *con)
 
 	/* if first message, set peer_name */
 	if (con->peer_name.type == 0)
-		con->peer_name = msg->hdr.src.name;
+		con->peer_name = msg->hdr.src;
 
 	con->in_seq++;
 	mutex_unlock(&con->mutex);
 
 	dout("===== %p %llu from %s%lld %d=%s len %d+%d (%u %u %u) =====\n",
 	     msg, le64_to_cpu(msg->hdr.seq),
-	     ENTITY_NAME(msg->hdr.src.name),
+	     ENTITY_NAME(msg->hdr.src),
 	     le16_to_cpu(msg->hdr.type),
 	     ceph_msg_type_name(le16_to_cpu(msg->hdr.type)),
 	     le32_to_cpu(msg->hdr.front_len),
@@ -1546,7 +1541,6 @@ static int try_write(struct ceph_connection *con)
 	dout("try_write start %p state %lu nref %d\n", con, con->state,
 	     atomic_read(&con->nref));
 
-	mutex_lock(&con->mutex);
 more:
 	dout("try_write out_kvec_bytes %d\n", con->out_kvec_bytes);
 
@@ -1639,7 +1633,6 @@ do_next:
 done:
 	ret = 0;
 out:
-	mutex_unlock(&con->mutex);
 	dout("try_write done on %p\n", con);
 	return ret;
 }
@@ -1651,7 +1644,6 @@ out:
  */
 static int try_read(struct ceph_connection *con)
 {
-	struct ceph_messenger *msgr;
 	int ret = -1;
 
 	if (!con->sock)
@@ -1661,9 +1653,6 @@ static int try_read(struct ceph_connection *con)
 		return 0;
 
 	dout("try_read start on %p\n", con);
-	msgr = con->msgr;
-
-	mutex_lock(&con->mutex);
 
 more:
 	dout("try_read tag %d in_base_pos %d\n", (int)con->in_tag,
@@ -1758,7 +1747,6 @@ more:
 done:
 	ret = 0;
 out:
-	mutex_unlock(&con->mutex);
 	dout("try_read done on %p\n", con);
 	return ret;
 
@@ -1830,6 +1818,8 @@ more:
 	dout("con_work %p start, clearing QUEUED\n", con);
 	clear_bit(QUEUED, &con->state);
 
+	mutex_lock(&con->mutex);
+
 	if (test_bit(CLOSED, &con->state)) { /* e.g. if we are replaced */
 		dout("con_work CLOSED\n");
 		con_close_socket(con);
@@ -1844,11 +1834,16 @@ more:
 	if (test_and_clear_bit(SOCK_CLOSED, &con->state) ||
 	    try_read(con) < 0 ||
 	    try_write(con) < 0) {
+		mutex_unlock(&con->mutex);
 		backoff = 1;
 		ceph_fault(con);     /* error/fault path */
+		goto done_unlocked;
 	}
 
 done:
+	mutex_unlock(&con->mutex);
+
+done_unlocked:
 	clear_bit(BUSY, &con->state);
 	dout("con->state=%lu\n", con->state);
 	if (test_bit(QUEUED, &con->state)) {
@@ -1947,7 +1942,7 @@ struct ceph_messenger *ceph_messenger_create(struct ceph_entity_addr *myaddr)
 
 	/* the zero page is needed if a request is "canceled" while the message
 	 * is being written over the socket */
-	msgr->zero_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
+	msgr->zero_page = __page_cache_alloc(GFP_KERNEL | __GFP_ZERO);
 	if (!msgr->zero_page) {
 		kfree(msgr);
 		return ERR_PTR(-ENOMEM);
@@ -1987,9 +1982,7 @@ void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg)
 	}
 
 	/* set src+dst */
-	msg->hdr.src.name = con->msgr->inst.name;
-	msg->hdr.src.addr = con->msgr->my_enc_addr;
-	msg->hdr.orig_src = msg->hdr.src;
+	msg->hdr.src = con->msgr->inst.name;
 
 	BUG_ON(msg->front.iov_len != le32_to_cpu(msg->hdr.front_len));
 
@@ -2083,12 +2076,11 @@ void ceph_con_keepalive(struct ceph_connection *con)
  * construct a new message with given type, size
  * the new msg has a ref count of 1.
  */
-struct ceph_msg *ceph_msg_new(int type, int front_len,
-			      int page_len, int page_off, struct page **pages)
+struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags)
 {
 	struct ceph_msg *m;
 
-	m = kmalloc(sizeof(*m), GFP_NOFS);
+	m = kmalloc(sizeof(*m), flags);
 	if (m == NULL)
 		goto out;
 	kref_init(&m->kref);
@@ -2100,8 +2092,8 @@ struct ceph_msg *ceph_msg_new(int type, int front_len,
 	m->hdr.version = 0;
 	m->hdr.front_len = cpu_to_le32(front_len);
 	m->hdr.middle_len = 0;
-	m->hdr.data_len = cpu_to_le32(page_len);
-	m->hdr.data_off = cpu_to_le16(page_off);
+	m->hdr.data_len = 0;
+	m->hdr.data_off = 0;
 	m->hdr.reserved = 0;
 	m->footer.front_crc = 0;
 	m->footer.middle_crc = 0;
@@ -2115,11 +2107,11 @@ struct ceph_msg *ceph_msg_new(int type, int front_len,
 	/* front */
 	if (front_len) {
 		if (front_len > PAGE_CACHE_SIZE) {
-			m->front.iov_base = __vmalloc(front_len, GFP_NOFS,
+			m->front.iov_base = __vmalloc(front_len, flags,
 						      PAGE_KERNEL);
 			m->front_is_vmalloc = true;
 		} else {
-			m->front.iov_base = kmalloc(front_len, GFP_NOFS);
+			m->front.iov_base = kmalloc(front_len, flags);
 		}
 		if (m->front.iov_base == NULL) {
 			pr_err("msg_new can't allocate %d bytes\n",
@@ -2135,19 +2127,18 @@ struct ceph_msg *ceph_msg_new(int type, int front_len,
 	m->middle = NULL;
 
 	/* data */
-	m->nr_pages = calc_pages_for(page_off, page_len);
-	m->pages = pages;
+	m->nr_pages = 0;
+	m->pages = NULL;
 	m->pagelist = NULL;
 
-	dout("ceph_msg_new %p page %d~%d -> %d\n", m, page_off, page_len,
-	     m->nr_pages);
+	dout("ceph_msg_new %p front %d\n", m, front_len);
 	return m;
 
 out2:
 	ceph_msg_put(m);
 out:
-	pr_err("msg_new can't create type %d len %d\n", type, front_len);
-	return ERR_PTR(-ENOMEM);
+	pr_err("msg_new can't create type %d front %d\n", type, front_len);
+	return NULL;
 }
 
 /*
@@ -2190,29 +2181,25 @@ static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con,
 		mutex_unlock(&con->mutex);
 		msg = con->ops->alloc_msg(con, hdr, skip);
 		mutex_lock(&con->mutex);
-		if (IS_ERR(msg))
-			return msg;
-
-		if (*skip)
+		if (!msg || *skip)
 			return NULL;
 	}
 	if (!msg) {
 		*skip = 0;
-		msg = ceph_msg_new(type, front_len, 0, 0, NULL);
+		msg = ceph_msg_new(type, front_len, GFP_NOFS);
 		if (!msg) {
 			pr_err("unable to allocate msg type %d len %d\n",
 			       type, front_len);
-			return ERR_PTR(-ENOMEM);
+			return NULL;
 		}
 	}
 	memcpy(&msg->hdr, &con->in_hdr, sizeof(con->in_hdr));
 
-	if (middle_len) {
+	if (middle_len && !msg->middle) {
 		ret = ceph_alloc_middle(con, msg);
-
 		if (ret < 0) {
 			ceph_msg_put(msg);
-			return msg;
+			return NULL;
 		}
 	}
 
diff --git a/fs/ceph/messenger.h b/fs/ceph/messenger.h
index a5caf91cc971..76fbc957bc13 100644
--- a/fs/ceph/messenger.h
+++ b/fs/ceph/messenger.h
@@ -49,10 +49,8 @@ struct ceph_connection_operations {
 					int *skip);
 };
 
-extern const char *ceph_name_type_str(int t);
-
 /* use format string %s%d */
-#define ENTITY_NAME(n) ceph_name_type_str((n).type), le64_to_cpu((n).num)
+#define ENTITY_NAME(n) ceph_entity_type_name((n).type), le64_to_cpu((n).num)
 
 struct ceph_messenger {
 	struct ceph_entity_inst inst;    /* my name+address */
@@ -144,6 +142,7 @@ struct ceph_connection {
 	struct ceph_entity_addr peer_addr; /* peer address */
 	struct ceph_entity_name peer_name; /* peer name */
 	struct ceph_entity_addr peer_addr_for_me;
+	unsigned peer_features;
 	u32 connect_seq;      /* identify the most recent connection
 				 attempt for this connection, client */
 	u32 peer_global_seq;  /* peer's global seq for this connection */
@@ -158,7 +157,6 @@ struct ceph_connection {
 	struct list_head out_queue;
 	struct list_head out_sent;   /* sending or sent but unacked */
 	u64 out_seq;		     /* last message queued for send */
-	u64 out_seq_sent;            /* last message sent */
 	bool out_keepalive_pending;
 
 	u64 in_seq, in_seq_acked;  /* last message received, acked */
@@ -215,6 +213,7 @@ extern int ceph_parse_ips(const char *c, const char *end,
 
 extern int ceph_msgr_init(void);
 extern void ceph_msgr_exit(void);
+extern void ceph_msgr_flush(void);
 
 extern struct ceph_messenger *ceph_messenger_create(
 	struct ceph_entity_addr *myaddr);
@@ -234,9 +233,7 @@ extern void ceph_con_keepalive(struct ceph_connection *con);
 extern struct ceph_connection *ceph_con_get(struct ceph_connection *con);
 extern void ceph_con_put(struct ceph_connection *con);
 
-extern struct ceph_msg *ceph_msg_new(int type, int front_len,
-				     int page_len, int page_off,
-				     struct page **pages);
+extern struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags);
 extern void ceph_msg_kfree(struct ceph_msg *m);
 
 
diff --git a/fs/ceph/mon_client.c b/fs/ceph/mon_client.c
index 8fdc011ca956..21c62e9b7d1d 100644
--- a/fs/ceph/mon_client.c
+++ b/fs/ceph/mon_client.c
@@ -28,7 +28,7 @@
  * resend any outstanding requests.
  */
 
-const static struct ceph_connection_operations mon_con_ops;
+static const struct ceph_connection_operations mon_con_ops;
 
 static int __validate_auth(struct ceph_mon_client *monc);
 
@@ -104,6 +104,7 @@ static void __send_prepared_auth_request(struct ceph_mon_client *monc, int len)
 	monc->pending_auth = 1;
 	monc->m_auth->front.iov_len = len;
 	monc->m_auth->hdr.front_len = cpu_to_le32(len);
+	ceph_con_revoke(monc->con, monc->m_auth);
 	ceph_msg_get(monc->m_auth);  /* keep our ref */
 	ceph_con_send(monc->con, monc->m_auth);
 }
@@ -187,16 +188,12 @@ static void __send_subscribe(struct ceph_mon_client *monc)
 	     monc->want_next_osdmap);
 	if ((__sub_expired(monc) && !monc->sub_sent) ||
 	    monc->want_next_osdmap == 1) {
-		struct ceph_msg *msg;
+		struct ceph_msg *msg = monc->m_subscribe;
 		struct ceph_mon_subscribe_item *i;
 		void *p, *end;
 
-		msg = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE, 96, 0, 0, NULL);
-		if (!msg)
-			return;
-
 		p = msg->front.iov_base;
-		end = p + msg->front.iov_len;
+		end = p + msg->front_max;
 
 		dout("__send_subscribe to 'mdsmap' %u+\n",
 		     (unsigned)monc->have_mdsmap);
@@ -226,7 +223,8 @@ static void __send_subscribe(struct ceph_mon_client *monc)
 
 		msg->front.iov_len = p - msg->front.iov_base;
 		msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
-		ceph_con_send(monc->con, msg);
+		ceph_con_revoke(monc->con, msg);
+		ceph_con_send(monc->con, ceph_msg_get(msg));
 
 		monc->sub_sent = jiffies | 1;  /* never 0 */
 	}
@@ -353,14 +351,14 @@ out:
 /*
  * statfs
  */
-static struct ceph_mon_statfs_request *__lookup_statfs(
+static struct ceph_mon_generic_request *__lookup_generic_req(
 	struct ceph_mon_client *monc, u64 tid)
 {
-	struct ceph_mon_statfs_request *req;
-	struct rb_node *n = monc->statfs_request_tree.rb_node;
+	struct ceph_mon_generic_request *req;
+	struct rb_node *n = monc->generic_request_tree.rb_node;
 
 	while (n) {
-		req = rb_entry(n, struct ceph_mon_statfs_request, node);
+		req = rb_entry(n, struct ceph_mon_generic_request, node);
 		if (tid < req->tid)
 			n = n->rb_left;
 		else if (tid > req->tid)
@@ -371,16 +369,16 @@ static struct ceph_mon_statfs_request *__lookup_statfs(
 	return NULL;
 }
 
-static void __insert_statfs(struct ceph_mon_client *monc,
-			    struct ceph_mon_statfs_request *new)
+static void __insert_generic_request(struct ceph_mon_client *monc,
+			    struct ceph_mon_generic_request *new)
 {
-	struct rb_node **p = &monc->statfs_request_tree.rb_node;
+	struct rb_node **p = &monc->generic_request_tree.rb_node;
 	struct rb_node *parent = NULL;
-	struct ceph_mon_statfs_request *req = NULL;
+	struct ceph_mon_generic_request *req = NULL;
 
 	while (*p) {
 		parent = *p;
-		req = rb_entry(parent, struct ceph_mon_statfs_request, node);
+		req = rb_entry(parent, struct ceph_mon_generic_request, node);
 		if (new->tid < req->tid)
 			p = &(*p)->rb_left;
 		else if (new->tid > req->tid)
@@ -390,113 +388,157 @@ static void __insert_statfs(struct ceph_mon_client *monc,
 	}
 
 	rb_link_node(&new->node, parent, p);
-	rb_insert_color(&new->node, &monc->statfs_request_tree);
+	rb_insert_color(&new->node, &monc->generic_request_tree);
+}
+
+static void release_generic_request(struct kref *kref)
+{
+	struct ceph_mon_generic_request *req =
+		container_of(kref, struct ceph_mon_generic_request, kref);
+
+	if (req->reply)
+		ceph_msg_put(req->reply);
+	if (req->request)
+		ceph_msg_put(req->request);
+}
+
+static void put_generic_request(struct ceph_mon_generic_request *req)
+{
+	kref_put(&req->kref, release_generic_request);
+}
+
+static void get_generic_request(struct ceph_mon_generic_request *req)
+{
+	kref_get(&req->kref);
+}
+
+static struct ceph_msg *get_generic_reply(struct ceph_connection *con,
+					 struct ceph_msg_header *hdr,
+					 int *skip)
+{
+	struct ceph_mon_client *monc = con->private;
+	struct ceph_mon_generic_request *req;
+	u64 tid = le64_to_cpu(hdr->tid);
+	struct ceph_msg *m;
+
+	mutex_lock(&monc->mutex);
+	req = __lookup_generic_req(monc, tid);
+	if (!req) {
+		dout("get_generic_reply %lld dne\n", tid);
+		*skip = 1;
+		m = NULL;
+	} else {
+		dout("get_generic_reply %lld got %p\n", tid, req->reply);
+		m = ceph_msg_get(req->reply);
+		/*
+		 * we don't need to track the connection reading into
+		 * this reply because we only have one open connection
+		 * at a time, ever.
+		 */
+	}
+	mutex_unlock(&monc->mutex);
+	return m;
 }
 
 static void handle_statfs_reply(struct ceph_mon_client *monc,
 				struct ceph_msg *msg)
 {
-	struct ceph_mon_statfs_request *req;
+	struct ceph_mon_generic_request *req;
 	struct ceph_mon_statfs_reply *reply = msg->front.iov_base;
-	u64 tid;
+	u64 tid = le64_to_cpu(msg->hdr.tid);
 
 	if (msg->front.iov_len != sizeof(*reply))
 		goto bad;
-	tid = le64_to_cpu(msg->hdr.tid);
 	dout("handle_statfs_reply %p tid %llu\n", msg, tid);
 
 	mutex_lock(&monc->mutex);
-	req = __lookup_statfs(monc, tid);
+	req = __lookup_generic_req(monc, tid);
 	if (req) {
-		*req->buf = reply->st;
+		*(struct ceph_statfs *)req->buf = reply->st;
 		req->result = 0;
+		get_generic_request(req);
 	}
 	mutex_unlock(&monc->mutex);
-	if (req)
+	if (req) {
 		complete(&req->completion);
+		put_generic_request(req);
+	}
 	return;
 
 bad:
-	pr_err("corrupt statfs reply, no tid\n");
+	pr_err("corrupt generic reply, no tid\n");
 	ceph_msg_dump(msg);
 }
 
 /*
- * (re)send a statfs request
+ * Do a synchronous statfs().
  */
-static int send_statfs(struct ceph_mon_client *monc,
-		       struct ceph_mon_statfs_request *req)
+int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf)
 {
-	struct ceph_msg *msg;
+	struct ceph_mon_generic_request *req;
 	struct ceph_mon_statfs *h;
+	int err;
 
-	dout("send_statfs tid %llu\n", req->tid);
-	msg = ceph_msg_new(CEPH_MSG_STATFS, sizeof(*h), 0, 0, NULL);
-	if (IS_ERR(msg))
-		return PTR_ERR(msg);
-	req->request = msg;
-	msg->hdr.tid = cpu_to_le64(req->tid);
-	h = msg->front.iov_base;
+	req = kzalloc(sizeof(*req), GFP_NOFS);
+	if (!req)
+		return -ENOMEM;
+
+	kref_init(&req->kref);
+	req->buf = buf;
+	init_completion(&req->completion);
+
+	err = -ENOMEM;
+	req->request = ceph_msg_new(CEPH_MSG_STATFS, sizeof(*h), GFP_NOFS);
+	if (!req->request)
+		goto out;
+	req->reply = ceph_msg_new(CEPH_MSG_STATFS_REPLY, 1024, GFP_NOFS);
+	if (!req->reply)
+		goto out;
+
+	/* fill out request */
+	h = req->request->front.iov_base;
 	h->monhdr.have_version = 0;
 	h->monhdr.session_mon = cpu_to_le16(-1);
 	h->monhdr.session_mon_tid = 0;
 	h->fsid = monc->monmap->fsid;
-	ceph_con_send(monc->con, msg);
-	return 0;
-}
-
-/*
- * Do a synchronous statfs().
- */
-int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf)
-{
-	struct ceph_mon_statfs_request req;
-	int err;
-
-	req.buf = buf;
-	init_completion(&req.completion);
-
-	/* allocate memory for reply */
-	err = ceph_msgpool_resv(&monc->msgpool_statfs_reply, 1);
-	if (err)
-		return err;
 
 	/* register request */
 	mutex_lock(&monc->mutex);
-	req.tid = ++monc->last_tid;
-	req.last_attempt = jiffies;
-	req.delay = BASE_DELAY_INTERVAL;
-	__insert_statfs(monc, &req);
-	monc->num_statfs_requests++;
+	req->tid = ++monc->last_tid;
+	req->request->hdr.tid = cpu_to_le64(req->tid);
+	__insert_generic_request(monc, req);
+	monc->num_generic_requests++;
 	mutex_unlock(&monc->mutex);
 
 	/* send request and wait */
-	err = send_statfs(monc, &req);
-	if (!err)
-		err = wait_for_completion_interruptible(&req.completion);
+	ceph_con_send(monc->con, ceph_msg_get(req->request));
+	err = wait_for_completion_interruptible(&req->completion);
 
 	mutex_lock(&monc->mutex);
-	rb_erase(&req.node, &monc->statfs_request_tree);
-	monc->num_statfs_requests--;
-	ceph_msgpool_resv(&monc->msgpool_statfs_reply, -1);
+	rb_erase(&req->node, &monc->generic_request_tree);
+	monc->num_generic_requests--;
 	mutex_unlock(&monc->mutex);
 
 	if (!err)
-		err = req.result;
+		err = req->result;
+
+out:
+	kref_put(&req->kref, release_generic_request);
 	return err;
 }
 
 /*
  * Resend pending statfs requests.
  */
-static void __resend_statfs(struct ceph_mon_client *monc)
+static void __resend_generic_request(struct ceph_mon_client *monc)
 {
-	struct ceph_mon_statfs_request *req;
+	struct ceph_mon_generic_request *req;
 	struct rb_node *p;
 
-	for (p = rb_first(&monc->statfs_request_tree); p; p = rb_next(p)) {
-		req = rb_entry(p, struct ceph_mon_statfs_request, node);
-		send_statfs(monc, req);
+	for (p = rb_first(&monc->generic_request_tree); p; p = rb_next(p)) {
+		req = rb_entry(p, struct ceph_mon_generic_request, node);
+		ceph_con_revoke(monc->con, req->request);
+		ceph_con_send(monc->con, ceph_msg_get(req->request));
 	}
 }
 
@@ -586,26 +628,26 @@ int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl)
 		CEPH_ENTITY_TYPE_AUTH | CEPH_ENTITY_TYPE_MON |
 		CEPH_ENTITY_TYPE_OSD | CEPH_ENTITY_TYPE_MDS;
 
-	/* msg pools */
-	err = ceph_msgpool_init(&monc->msgpool_subscribe_ack,
-			       sizeof(struct ceph_mon_subscribe_ack), 1, false);
-	if (err < 0)
+	/* msgs */
+	err = -ENOMEM;
+	monc->m_subscribe_ack = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE_ACK,
+				     sizeof(struct ceph_mon_subscribe_ack),
+				     GFP_NOFS);
+	if (!monc->m_subscribe_ack)
 		goto out_monmap;
-	err = ceph_msgpool_init(&monc->msgpool_statfs_reply,
-				sizeof(struct ceph_mon_statfs_reply), 0, false);
-	if (err < 0)
-		goto out_pool1;
-	err = ceph_msgpool_init(&monc->msgpool_auth_reply, 4096, 1, false);
-	if (err < 0)
-		goto out_pool2;
-
-	monc->m_auth = ceph_msg_new(CEPH_MSG_AUTH, 4096, 0, 0, NULL);
+
+	monc->m_subscribe = ceph_msg_new(CEPH_MSG_MON_SUBSCRIBE, 96, GFP_NOFS);
+	if (!monc->m_subscribe)
+		goto out_subscribe_ack;
+
+	monc->m_auth_reply = ceph_msg_new(CEPH_MSG_AUTH_REPLY, 4096, GFP_NOFS);
+	if (!monc->m_auth_reply)
+		goto out_subscribe;
+
+	monc->m_auth = ceph_msg_new(CEPH_MSG_AUTH, 4096, GFP_NOFS);
 	monc->pending_auth = 0;
-	if (IS_ERR(monc->m_auth)) {
-		err = PTR_ERR(monc->m_auth);
-		monc->m_auth = NULL;
-		goto out_pool3;
-	}
+	if (!monc->m_auth)
+		goto out_auth_reply;
 
 	monc->cur_mon = -1;
 	monc->hunting = true;
@@ -613,8 +655,8 @@ int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl)
 	monc->sub_sent = 0;
 
 	INIT_DELAYED_WORK(&monc->delayed_work, delayed_work);
-	monc->statfs_request_tree = RB_ROOT;
-	monc->num_statfs_requests = 0;
+	monc->generic_request_tree = RB_ROOT;
+	monc->num_generic_requests = 0;
 	monc->last_tid = 0;
 
 	monc->have_mdsmap = 0;
@@ -622,12 +664,12 @@ int ceph_monc_init(struct ceph_mon_client *monc, struct ceph_client *cl)
 	monc->want_next_osdmap = 1;
 	return 0;
 
-out_pool3:
-	ceph_msgpool_destroy(&monc->msgpool_auth_reply);
-out_pool2:
-	ceph_msgpool_destroy(&monc->msgpool_subscribe_ack);
-out_pool1:
-	ceph_msgpool_destroy(&monc->msgpool_statfs_reply);
+out_auth_reply:
+	ceph_msg_put(monc->m_auth_reply);
+out_subscribe:
+	ceph_msg_put(monc->m_subscribe);
+out_subscribe_ack:
+	ceph_msg_put(monc->m_subscribe_ack);
 out_monmap:
 	kfree(monc->monmap);
 out:
@@ -651,9 +693,9 @@ void ceph_monc_stop(struct ceph_mon_client *monc)
 	ceph_auth_destroy(monc->auth);
 
 	ceph_msg_put(monc->m_auth);
-	ceph_msgpool_destroy(&monc->msgpool_subscribe_ack);
-	ceph_msgpool_destroy(&monc->msgpool_statfs_reply);
-	ceph_msgpool_destroy(&monc->msgpool_auth_reply);
+	ceph_msg_put(monc->m_auth_reply);
+	ceph_msg_put(monc->m_subscribe);
+	ceph_msg_put(monc->m_subscribe_ack);
 
 	kfree(monc->monmap);
 }
@@ -662,8 +704,11 @@ static void handle_auth_reply(struct ceph_mon_client *monc,
 			      struct ceph_msg *msg)
 {
 	int ret;
+	int was_auth = 0;
 
 	mutex_lock(&monc->mutex);
+	if (monc->auth->ops)
+		was_auth = monc->auth->ops->is_authenticated(monc->auth);
 	monc->pending_auth = 0;
 	ret = ceph_handle_auth_reply(monc->auth, msg->front.iov_base,
 				     msg->front.iov_len,
@@ -674,14 +719,14 @@ static void handle_auth_reply(struct ceph_mon_client *monc,
 		wake_up(&monc->client->auth_wq);
 	} else if (ret > 0) {
 		__send_prepared_auth_request(monc, ret);
-	} else if (monc->auth->ops->is_authenticated(monc->auth)) {
+	} else if (!was_auth && monc->auth->ops->is_authenticated(monc->auth)) {
 		dout("authenticated, starting session\n");
 
 		monc->client->msgr->inst.name.type = CEPH_ENTITY_TYPE_CLIENT;
 		monc->client->msgr->inst.name.num = monc->auth->global_id;
 
 		__send_subscribe(monc);
-		__resend_statfs(monc);
+		__resend_generic_request(monc);
 	}
 	mutex_unlock(&monc->mutex);
 }
@@ -770,18 +815,17 @@ static struct ceph_msg *mon_alloc_msg(struct ceph_connection *con,
 
 	switch (type) {
 	case CEPH_MSG_MON_SUBSCRIBE_ACK:
-		m = ceph_msgpool_get(&monc->msgpool_subscribe_ack, front_len);
+		m = ceph_msg_get(monc->m_subscribe_ack);
 		break;
 	case CEPH_MSG_STATFS_REPLY:
-		m = ceph_msgpool_get(&monc->msgpool_statfs_reply, front_len);
-		break;
+		return get_generic_reply(con, hdr, skip);
 	case CEPH_MSG_AUTH_REPLY:
-		m = ceph_msgpool_get(&monc->msgpool_auth_reply, front_len);
+		m = ceph_msg_get(monc->m_auth_reply);
 		break;
 	case CEPH_MSG_MON_MAP:
 	case CEPH_MSG_MDS_MAP:
 	case CEPH_MSG_OSD_MAP:
-		m = ceph_msg_new(type, front_len, 0, 0, NULL);
+		m = ceph_msg_new(type, front_len, GFP_NOFS);
 		break;
 	}
 
@@ -826,7 +870,7 @@ out:
 	mutex_unlock(&monc->mutex);
 }
 
-const static struct ceph_connection_operations mon_con_ops = {
+static const struct ceph_connection_operations mon_con_ops = {
 	.get = ceph_con_get,
 	.put = ceph_con_put,
 	.dispatch = dispatch,
diff --git a/fs/ceph/mon_client.h b/fs/ceph/mon_client.h
index b958ad5afa06..174d794321d0 100644
--- a/fs/ceph/mon_client.h
+++ b/fs/ceph/mon_client.h
@@ -2,10 +2,10 @@
 #define _FS_CEPH_MON_CLIENT_H
 
 #include <linux/completion.h>
+#include <linux/kref.h>
 #include <linux/rbtree.h>
 
 #include "messenger.h"
-#include "msgpool.h"
 
 struct ceph_client;
 struct ceph_mount_args;
@@ -22,7 +22,7 @@ struct ceph_monmap {
 };
 
 struct ceph_mon_client;
-struct ceph_mon_statfs_request;
+struct ceph_mon_generic_request;
 
 
 /*
@@ -40,17 +40,19 @@ struct ceph_mon_request {
 };
 
 /*
- * statfs() is done a bit differently because we need to get data back
+ * ceph_mon_generic_request is being used for the statfs and poolop requests
+ * which are bening done a bit differently because we need to get data back
  * to the caller
  */
-struct ceph_mon_statfs_request {
+struct ceph_mon_generic_request {
+	struct kref kref;
 	u64 tid;
 	struct rb_node node;
 	int result;
-	struct ceph_statfs *buf;
+	void *buf;
 	struct completion completion;
-	unsigned long last_attempt, delay; /* jiffies */
 	struct ceph_msg *request;  /* original request */
+	struct ceph_msg *reply;    /* and reply */
 };
 
 struct ceph_mon_client {
@@ -61,7 +63,7 @@ struct ceph_mon_client {
 	struct delayed_work delayed_work;
 
 	struct ceph_auth_client *auth;
-	struct ceph_msg *m_auth;
+	struct ceph_msg *m_auth, *m_auth_reply, *m_subscribe, *m_subscribe_ack;
 	int pending_auth;
 
 	bool hunting;
@@ -70,14 +72,9 @@ struct ceph_mon_client {
 	struct ceph_connection *con;
 	bool have_fsid;
 
-	/* msg pools */
-	struct ceph_msgpool msgpool_subscribe_ack;
-	struct ceph_msgpool msgpool_statfs_reply;
-	struct ceph_msgpool msgpool_auth_reply;
-
-	/* pending statfs requests */
-	struct rb_root statfs_request_tree;
-	int num_statfs_requests;
+	/* pending generic requests */
+	struct rb_root generic_request_tree;
+	int num_generic_requests;
 	u64 last_tid;
 
 	/* mds/osd map */
diff --git a/fs/ceph/msgpool.c b/fs/ceph/msgpool.c
index ca3b44a89f2d..dd65a6438131 100644
--- a/fs/ceph/msgpool.c
+++ b/fs/ceph/msgpool.c
@@ -7,180 +7,58 @@
 
 #include "msgpool.h"
 
-/*
- * We use msg pools to preallocate memory for messages we expect to
- * receive over the wire, to avoid getting ourselves into OOM
- * conditions at unexpected times.  We take use a few different
- * strategies:
- *
- *  - for request/response type interactions, we preallocate the
- * memory needed for the response when we generate the request.
- *
- *  - for messages we can receive at any time from the MDS, we preallocate
- * a pool of messages we can re-use.
- *
- *  - for writeback, we preallocate some number of messages to use for
- * requests and their replies, so that we always make forward
- * progress.
- *
- * The msgpool behaves like a mempool_t, but keeps preallocated
- * ceph_msgs strung together on a list_head instead of using a pointer
- * vector.  This avoids vector reallocation when we adjust the number
- * of preallocated items (which happens frequently).
- */
+static void *alloc_fn(gfp_t gfp_mask, void *arg)
+{
+	struct ceph_msgpool *pool = arg;
+	void *p;
 
+	p = ceph_msg_new(0, pool->front_len, gfp_mask);
+	if (!p)
+		pr_err("msgpool %s alloc failed\n", pool->name);
+	return p;
+}
 
-/*
- * Allocate or release as necessary to meet our target pool size.
- */
-static int __fill_msgpool(struct ceph_msgpool *pool)
+static void free_fn(void *element, void *arg)
 {
-	struct ceph_msg *msg;
-
-	while (pool->num < pool->min) {
-		dout("fill_msgpool %p %d/%d allocating\n", pool, pool->num,
-		     pool->min);
-		spin_unlock(&pool->lock);
-		msg = ceph_msg_new(0, pool->front_len, 0, 0, NULL);
-		spin_lock(&pool->lock);
-		if (IS_ERR(msg))
-			return PTR_ERR(msg);
-		msg->pool = pool;
-		list_add(&msg->list_head, &pool->msgs);
-		pool->num++;
-	}
-	while (pool->num > pool->min) {
-		msg = list_first_entry(&pool->msgs, struct ceph_msg, list_head);
-		dout("fill_msgpool %p %d/%d releasing %p\n", pool, pool->num,
-		     pool->min, msg);
-		list_del_init(&msg->list_head);
-		pool->num--;
-		ceph_msg_kfree(msg);
-	}
-	return 0;
+	ceph_msg_put(element);
 }
 
 int ceph_msgpool_init(struct ceph_msgpool *pool,
-		      int front_len, int min, bool blocking)
+		      int front_len, int size, bool blocking, const char *name)
 {
-	int ret;
-
-	dout("msgpool_init %p front_len %d min %d\n", pool, front_len, min);
-	spin_lock_init(&pool->lock);
 	pool->front_len = front_len;
-	INIT_LIST_HEAD(&pool->msgs);
-	pool->num = 0;
-	pool->min = min;
-	pool->blocking = blocking;
-	init_waitqueue_head(&pool->wait);
-
-	spin_lock(&pool->lock);
-	ret = __fill_msgpool(pool);
-	spin_unlock(&pool->lock);
-	return ret;
+	pool->pool = mempool_create(size, alloc_fn, free_fn, pool);
+	if (!pool->pool)
+		return -ENOMEM;
+	pool->name = name;
+	return 0;
 }
 
 void ceph_msgpool_destroy(struct ceph_msgpool *pool)
 {
-	dout("msgpool_destroy %p\n", pool);
-	spin_lock(&pool->lock);
-	pool->min = 0;
-	__fill_msgpool(pool);
-	spin_unlock(&pool->lock);
+	mempool_destroy(pool->pool);
 }
 
-int ceph_msgpool_resv(struct ceph_msgpool *pool, int delta)
+struct ceph_msg *ceph_msgpool_get(struct ceph_msgpool *pool,
+				  int front_len)
 {
-	int ret;
-
-	spin_lock(&pool->lock);
-	dout("msgpool_resv %p delta %d\n", pool, delta);
-	pool->min += delta;
-	ret = __fill_msgpool(pool);
-	spin_unlock(&pool->lock);
-	return ret;
-}
-
-struct ceph_msg *ceph_msgpool_get(struct ceph_msgpool *pool, int front_len)
-{
-	wait_queue_t wait;
-	struct ceph_msg *msg;
-
-	if (front_len && front_len > pool->front_len) {
-		pr_err("msgpool_get pool %p need front %d, pool size is %d\n",
-		       pool, front_len, pool->front_len);
+	if (front_len > pool->front_len) {
+		pr_err("msgpool_get pool %s need front %d, pool size is %d\n",
+		       pool->name, front_len, pool->front_len);
 		WARN_ON(1);
 
 		/* try to alloc a fresh message */
-		msg = ceph_msg_new(0, front_len, 0, 0, NULL);
-		if (!IS_ERR(msg))
-			return msg;
-	}
-
-	if (!front_len)
-		front_len = pool->front_len;
-
-	if (pool->blocking) {
-		/* mempool_t behavior; first try to alloc */
-		msg = ceph_msg_new(0, front_len, 0, 0, NULL);
-		if (!IS_ERR(msg))
-			return msg;
+		return ceph_msg_new(0, front_len, GFP_NOFS);
 	}
 
-	while (1) {
-		spin_lock(&pool->lock);
-		if (likely(pool->num)) {
-			msg = list_entry(pool->msgs.next, struct ceph_msg,
-					 list_head);
-			list_del_init(&msg->list_head);
-			pool->num--;
-			dout("msgpool_get %p got %p, now %d/%d\n", pool, msg,
-			     pool->num, pool->min);
-			spin_unlock(&pool->lock);
-			return msg;
-		}
-		pr_err("msgpool_get %p now %d/%d, %s\n", pool, pool->num,
-		       pool->min, pool->blocking ? "waiting" : "may fail");
-		spin_unlock(&pool->lock);
-
-		if (!pool->blocking) {
-			WARN_ON(1);
-
-			/* maybe we can allocate it now? */
-			msg = ceph_msg_new(0, front_len, 0, 0, NULL);
-			if (!IS_ERR(msg))
-				return msg;
-
-			pr_err("msgpool_get %p empty + alloc failed\n", pool);
-			return ERR_PTR(-ENOMEM);
-		}
-
-		init_wait(&wait);
-		prepare_to_wait(&pool->wait, &wait, TASK_UNINTERRUPTIBLE);
-		schedule();
-		finish_wait(&pool->wait, &wait);
-	}
+	return mempool_alloc(pool->pool, GFP_NOFS);
 }
 
 void ceph_msgpool_put(struct ceph_msgpool *pool, struct ceph_msg *msg)
 {
-	spin_lock(&pool->lock);
-	if (pool->num < pool->min) {
-		/* reset msg front_len; user may have changed it */
-		msg->front.iov_len = pool->front_len;
-		msg->hdr.front_len = cpu_to_le32(pool->front_len);
+	/* reset msg front_len; user may have changed it */
+	msg->front.iov_len = pool->front_len;
+	msg->hdr.front_len = cpu_to_le32(pool->front_len);
 
-		kref_set(&msg->kref, 1);  /* retake a single ref */
-		list_add(&msg->list_head, &pool->msgs);
-		pool->num++;
-		dout("msgpool_put %p reclaim %p, now %d/%d\n", pool, msg,
-		     pool->num, pool->min);
-		spin_unlock(&pool->lock);
-		wake_up(&pool->wait);
-	} else {
-		dout("msgpool_put %p drop %p, at %d/%d\n", pool, msg,
-		     pool->num, pool->min);
-		spin_unlock(&pool->lock);
-		ceph_msg_kfree(msg);
-	}
+	kref_init(&msg->kref);  /* retake single ref */
 }
diff --git a/fs/ceph/msgpool.h b/fs/ceph/msgpool.h
index bc834bfcd720..a362605f9368 100644
--- a/fs/ceph/msgpool.h
+++ b/fs/ceph/msgpool.h
@@ -1,6 +1,7 @@
 #ifndef _FS_CEPH_MSGPOOL
 #define _FS_CEPH_MSGPOOL
 
+#include <linux/mempool.h>
 #include "messenger.h"
 
 /*
@@ -8,18 +9,15 @@
  * avoid unexpected OOM conditions.
  */
 struct ceph_msgpool {
-	spinlock_t lock;
+	const char *name;
+	mempool_t *pool;
 	int front_len;          /* preallocated payload size */
-	struct list_head msgs;  /* msgs in the pool; each has 1 ref */
-	int num, min;           /* cur, min # msgs in the pool */
-	bool blocking;
-	wait_queue_head_t wait;
 };
 
 extern int ceph_msgpool_init(struct ceph_msgpool *pool,
-			     int front_len, int size, bool blocking);
+			     int front_len, int size, bool blocking,
+			     const char *name);
 extern void ceph_msgpool_destroy(struct ceph_msgpool *pool);
-extern int ceph_msgpool_resv(struct ceph_msgpool *, int delta);
 extern struct ceph_msg *ceph_msgpool_get(struct ceph_msgpool *,
 					 int front_len);
 extern void ceph_msgpool_put(struct ceph_msgpool *, struct ceph_msg *);
diff --git a/fs/ceph/msgr.h b/fs/ceph/msgr.h
index 8aaab414f3f8..892a0298dfdf 100644
--- a/fs/ceph/msgr.h
+++ b/fs/ceph/msgr.h
@@ -50,7 +50,6 @@ struct ceph_entity_name {
 #define CEPH_ENTITY_TYPE_MDS    0x02
 #define CEPH_ENTITY_TYPE_OSD    0x04
 #define CEPH_ENTITY_TYPE_CLIENT 0x08
-#define CEPH_ENTITY_TYPE_ADMIN  0x10
 #define CEPH_ENTITY_TYPE_AUTH   0x20
 
 #define CEPH_ENTITY_TYPE_ANY    0xFF
@@ -120,7 +119,7 @@ struct ceph_msg_connect_reply {
 /*
  * message header
  */
-struct ceph_msg_header {
+struct ceph_msg_header_old {
 	__le64 seq;       /* message seq# for this session */
 	__le64 tid;       /* transaction id */
 	__le16 type;      /* message type */
@@ -138,6 +137,24 @@ struct ceph_msg_header {
 	__le32 crc;       /* header crc32c */
 } __attribute__ ((packed));
 
+struct ceph_msg_header {
+	__le64 seq;       /* message seq# for this session */
+	__le64 tid;       /* transaction id */
+	__le16 type;      /* message type */
+	__le16 priority;  /* priority.  higher value == higher priority */
+	__le16 version;   /* version of message encoding */
+
+	__le32 front_len; /* bytes in main payload */
+	__le32 middle_len;/* bytes in middle payload */
+	__le32 data_len;  /* bytes of data payload */
+	__le16 data_off;  /* sender: include full offset;
+			     receiver: mask against ~PAGE_MASK */
+
+	struct ceph_entity_name src;
+	__le32 reserved;
+	__le32 crc;       /* header crc32c */
+} __attribute__ ((packed));
+
 #define CEPH_MSG_PRIO_LOW     64
 #define CEPH_MSG_PRIO_DEFAULT 127
 #define CEPH_MSG_PRIO_HIGH    196
diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c
index 3514f71ff85f..d25b4add85b4 100644
--- a/fs/ceph/osd_client.c
+++ b/fs/ceph/osd_client.c
@@ -16,7 +16,7 @@
 #define OSD_OP_FRONT_LEN	4096
 #define OSD_OPREPLY_FRONT_LEN	512
 
-const static struct ceph_connection_operations osd_con_ops;
+static const struct ceph_connection_operations osd_con_ops;
 static int __kick_requests(struct ceph_osd_client *osdc,
 			  struct ceph_osd *kickosd);
 
@@ -147,7 +147,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 		req = kzalloc(sizeof(*req), GFP_NOFS);
 	}
 	if (req == NULL)
-		return ERR_PTR(-ENOMEM);
+		return NULL;
 
 	req->r_osdc = osdc;
 	req->r_mempool = use_mempool;
@@ -164,10 +164,10 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 		msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0);
 	else
 		msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY,
-				   OSD_OPREPLY_FRONT_LEN, 0, 0, NULL);
-	if (IS_ERR(msg)) {
+				   OSD_OPREPLY_FRONT_LEN, GFP_NOFS);
+	if (!msg) {
 		ceph_osdc_put_request(req);
-		return ERR_PTR(PTR_ERR(msg));
+		return NULL;
 	}
 	req->r_reply = msg;
 
@@ -178,10 +178,10 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
 	if (use_mempool)
 		msg = ceph_msgpool_get(&osdc->msgpool_op, 0);
 	else
-		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, 0, 0, NULL);
-	if (IS_ERR(msg)) {
+		msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, GFP_NOFS);
+	if (!msg) {
 		ceph_osdc_put_request(req);
-		return ERR_PTR(PTR_ERR(msg));
+		return NULL;
 	}
 	msg->hdr.type = cpu_to_le16(CEPH_MSG_OSD_OP);
 	memset(msg->front.iov_base, 0, msg->front.iov_len);
@@ -361,8 +361,13 @@ static void put_osd(struct ceph_osd *osd)
 {
 	dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref),
 	     atomic_read(&osd->o_ref) - 1);
-	if (atomic_dec_and_test(&osd->o_ref))
+	if (atomic_dec_and_test(&osd->o_ref)) {
+		struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth;
+
+		if (osd->o_authorizer)
+			ac->ops->destroy_authorizer(ac, osd->o_authorizer);
 		kfree(osd);
+	}
 }
 
 /*
@@ -715,7 +720,7 @@ static void handle_timeout(struct work_struct *work)
 	 * should mark the osd as failed and we should find out about
 	 * it from an updated osd map.
 	 */
-	while (!list_empty(&osdc->req_lru)) {
+	while (timeout && !list_empty(&osdc->req_lru)) {
 		req = list_entry(osdc->req_lru.next, struct ceph_osd_request,
 				 r_req_lru_item);
 
@@ -1078,6 +1083,7 @@ done:
 	if (newmap)
 		kick_requests(osdc, NULL);
 	up_read(&osdc->map_sem);
+	wake_up(&osdc->client->auth_wq);
 	return;
 
 bad:
@@ -1087,45 +1093,6 @@ bad:
 	return;
 }
 
-
-/*
- * A read request prepares specific pages that data is to be read into.
- * When a message is being read off the wire, we call prepare_pages to
- * find those pages.
- *  0 = success, -1 failure.
- */
-static int __prepare_pages(struct ceph_connection *con,
-			 struct ceph_msg_header *hdr,
-			 struct ceph_osd_request *req,
-			 u64 tid,
-			 struct ceph_msg *m)
-{
-	struct ceph_osd *osd = con->private;
-	struct ceph_osd_client *osdc;
-	int ret = -1;
-	int data_len = le32_to_cpu(hdr->data_len);
-	unsigned data_off = le16_to_cpu(hdr->data_off);
-
-	int want = calc_pages_for(data_off & ~PAGE_MASK, data_len);
-
-	if (!osd)
-		return -1;
-
-	osdc = osd->o_osdc;
-
-	dout("__prepare_pages on msg %p tid %llu, has %d pages, want %d\n", m,
-	     tid, req->r_num_pages, want);
-	if (unlikely(req->r_num_pages < want))
-		goto out;
-	m->pages = req->r_pages;
-	m->nr_pages = req->r_num_pages;
-	ret = 0; /* success */
-out:
-	BUG_ON(ret < 0 || m->nr_pages < want);
-
-	return ret;
-}
-
 /*
  * Register request, send initial attempt.
  */
@@ -1252,11 +1219,13 @@ int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client)
 	if (!osdc->req_mempool)
 		goto out;
 
-	err = ceph_msgpool_init(&osdc->msgpool_op, OSD_OP_FRONT_LEN, 10, true);
+	err = ceph_msgpool_init(&osdc->msgpool_op, OSD_OP_FRONT_LEN, 10, true,
+				"osd_op");
 	if (err < 0)
 		goto out_mempool;
 	err = ceph_msgpool_init(&osdc->msgpool_op_reply,
-				OSD_OPREPLY_FRONT_LEN, 10, true);
+				OSD_OPREPLY_FRONT_LEN, 10, true,
+				"osd_op_reply");
 	if (err < 0)
 		goto out_msgpool;
 	return 0;
@@ -1302,8 +1271,8 @@ int ceph_osdc_readpages(struct ceph_osd_client *osdc,
 				    CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
 				    NULL, 0, truncate_seq, truncate_size, NULL,
 				    false, 1);
-	if (IS_ERR(req))
-		return PTR_ERR(req);
+	if (!req)
+		return -ENOMEM;
 
 	/* it may be a short read due to an object boundary */
 	req->r_pages = pages;
@@ -1345,8 +1314,8 @@ int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino,
 				    snapc, do_sync,
 				    truncate_seq, truncate_size, mtime,
 				    nofail, 1);
-	if (IS_ERR(req))
-		return PTR_ERR(req);
+	if (!req)
+		return -ENOMEM;
 
 	/* it may be a short write due to an object boundary */
 	req->r_pages = pages;
@@ -1394,7 +1363,8 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
 }
 
 /*
- * lookup and return message for incoming reply
+ * lookup and return message for incoming reply.  set up reply message
+ * pages.
  */
 static struct ceph_msg *get_reply(struct ceph_connection *con,
 				  struct ceph_msg_header *hdr,
@@ -1407,7 +1377,6 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
 	int front = le32_to_cpu(hdr->front_len);
 	int data_len = le32_to_cpu(hdr->data_len);
 	u64 tid;
-	int err;
 
 	tid = le64_to_cpu(hdr->tid);
 	mutex_lock(&osdc->request_mutex);
@@ -1425,13 +1394,14 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
 		     req->r_reply, req->r_con_filling_msg);
 		ceph_con_revoke_message(req->r_con_filling_msg, req->r_reply);
 		ceph_con_put(req->r_con_filling_msg);
+		req->r_con_filling_msg = NULL;
 	}
 
 	if (front > req->r_reply->front.iov_len) {
 		pr_warning("get_reply front %d > preallocated %d\n",
 			   front, (int)req->r_reply->front.iov_len);
-		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, 0, 0, NULL);
-		if (IS_ERR(m))
+		m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front, GFP_NOFS);
+		if (!m)
 			goto out;
 		ceph_msg_put(req->r_reply);
 		req->r_reply = m;
@@ -1439,12 +1409,19 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
 	m = ceph_msg_get(req->r_reply);
 
 	if (data_len > 0) {
-		err = __prepare_pages(con, hdr, req, tid, m);
-		if (err < 0) {
+		unsigned data_off = le16_to_cpu(hdr->data_off);
+		int want = calc_pages_for(data_off & ~PAGE_MASK, data_len);
+
+		if (unlikely(req->r_num_pages < want)) {
+			pr_warning("tid %lld reply %d > expected %d pages\n",
+				   tid, want, m->nr_pages);
 			*skip = 1;
 			ceph_msg_put(m);
-			m = ERR_PTR(err);
+			m = NULL;
+			goto out;
 		}
+		m->pages = req->r_pages;
+		m->nr_pages = req->r_num_pages;
 	}
 	*skip = 0;
 	req->r_con_filling_msg = ceph_con_get(con);
@@ -1466,7 +1443,7 @@ static struct ceph_msg *alloc_msg(struct ceph_connection *con,
 
 	switch (type) {
 	case CEPH_MSG_OSD_MAP:
-		return ceph_msg_new(type, front, 0, 0, NULL);
+		return ceph_msg_new(type, front, GFP_NOFS);
 	case CEPH_MSG_OSD_OPREPLY:
 		return get_reply(con, hdr, skip);
 	default:
@@ -1552,7 +1529,7 @@ static int invalidate_authorizer(struct ceph_connection *con)
 	return ceph_monc_validate_auth(&osdc->client->monc);
 }
 
-const static struct ceph_connection_operations osd_con_ops = {
+static const struct ceph_connection_operations osd_con_ops = {
 	.get = get_osd_con,
 	.put = put_osd_con,
 	.dispatch = dispatch,
diff --git a/fs/ceph/osdmap.c b/fs/ceph/osdmap.c
index cfdd8f4388b7..ddc656fb5c05 100644
--- a/fs/ceph/osdmap.c
+++ b/fs/ceph/osdmap.c
@@ -706,7 +706,7 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 		     len, *p, end);
 		newcrush = crush_decode(*p, min(*p+len, end));
 		if (IS_ERR(newcrush))
-			return ERR_PTR(PTR_ERR(newcrush));
+			return ERR_CAST(newcrush);
 	}
 
 	/* new flags? */
diff --git a/fs/ceph/pagelist.c b/fs/ceph/pagelist.c
index 5f8dbf7c745a..b6859f47d364 100644
--- a/fs/ceph/pagelist.c
+++ b/fs/ceph/pagelist.c
@@ -20,7 +20,7 @@ int ceph_pagelist_release(struct ceph_pagelist *pl)
 
 static int ceph_pagelist_addpage(struct ceph_pagelist *pl)
 {
-	struct page *page = alloc_page(GFP_NOFS);
+	struct page *page = __page_cache_alloc(GFP_NOFS);
 	if (!page)
 		return -ENOMEM;
 	pl->room += PAGE_SIZE;
diff --git a/fs/ceph/rados.h b/fs/ceph/rados.h
index fd56451a871f..8fcc023056c7 100644
--- a/fs/ceph/rados.h
+++ b/fs/ceph/rados.h
@@ -101,8 +101,8 @@ struct ceph_pg_pool {
 	__le64 snap_seq;          /* seq for per-pool snapshot */
 	__le32 snap_epoch;        /* epoch of last snap */
 	__le32 num_snaps;
-	__le32 num_removed_snap_intervals;
-	__le64 uid;
+	__le32 num_removed_snap_intervals; /* if non-empty, NO per-pool snaps */
+	__le64 auid;               /* who owns the pg */
 } __attribute__ ((packed));
 
 /*
@@ -208,6 +208,7 @@ enum {
 	/* read */
 	CEPH_OSD_OP_GETXATTR  = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 1,
 	CEPH_OSD_OP_GETXATTRS = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 2,
+	CEPH_OSD_OP_CMPXATTR  = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 3,
 
 	/* write */
 	CEPH_OSD_OP_SETXATTR  = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 1,
@@ -305,6 +306,22 @@ enum {
 #define EOLDSNAPC    ERESTART  /* ORDERSNAP flag set; writer has old snapc*/
 #define EBLACKLISTED ESHUTDOWN /* blacklisted */
 
+/* xattr comparison */
+enum {
+	CEPH_OSD_CMPXATTR_OP_NOP = 0,
+	CEPH_OSD_CMPXATTR_OP_EQ  = 1,
+	CEPH_OSD_CMPXATTR_OP_NE  = 2,
+	CEPH_OSD_CMPXATTR_OP_GT  = 3,
+	CEPH_OSD_CMPXATTR_OP_GTE = 4,
+	CEPH_OSD_CMPXATTR_OP_LT  = 5,
+	CEPH_OSD_CMPXATTR_OP_LTE = 6
+};
+
+enum {
+	CEPH_OSD_CMPXATTR_MODE_STRING = 1,
+	CEPH_OSD_CMPXATTR_MODE_U64    = 2
+};
+
 /*
  * an individual object operation.  each may be accompanied by some data
  * payload
@@ -321,6 +338,8 @@ struct ceph_osd_op {
 		struct {
 			__le32 name_len;
 			__le32 value_len;
+			__u8 cmp_op;       /* CEPH_OSD_CMPXATTR_OP_* */
+			__u8 cmp_mode;     /* CEPH_OSD_CMPXATTR_MODE_* */
 		} __attribute__ ((packed)) xattr;
 		struct {
 			__u8 class_len;
diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c
index d5114db70453..c0b26b6badba 100644
--- a/fs/ceph/snap.c
+++ b/fs/ceph/snap.c
@@ -512,7 +512,7 @@ int __ceph_finish_cap_snap(struct ceph_inode_info *ci,
 			    struct ceph_cap_snap *capsnap)
 {
 	struct inode *inode = &ci->vfs_inode;
-	struct ceph_mds_client *mdsc = &ceph_client(inode->i_sb)->mdsc;
+	struct ceph_mds_client *mdsc = &ceph_sb_to_client(inode->i_sb)->mdsc;
 
 	BUG_ON(capsnap->writing);
 	capsnap->size = inode->i_size;
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 9307bbee6fbe..4e0bee240b9d 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -8,14 +8,11 @@
 #include <linux/module.h>
 #include <linux/mount.h>
 #include <linux/parser.h>
-#include <linux/rwsem.h>
 #include <linux/sched.h>
 #include <linux/seq_file.h>
 #include <linux/slab.h>
 #include <linux/statfs.h>
 #include <linux/string.h>
-#include <linux/version.h>
-#include <linux/vmalloc.h>
 
 #include "decode.h"
 #include "super.h"
@@ -107,12 +104,40 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf)
 static int ceph_syncfs(struct super_block *sb, int wait)
 {
 	dout("sync_fs %d\n", wait);
-	ceph_osdc_sync(&ceph_client(sb)->osdc);
-	ceph_mdsc_sync(&ceph_client(sb)->mdsc);
+	ceph_osdc_sync(&ceph_sb_to_client(sb)->osdc);
+	ceph_mdsc_sync(&ceph_sb_to_client(sb)->mdsc);
 	dout("sync_fs %d done\n", wait);
 	return 0;
 }
 
+static int default_congestion_kb(void)
+{
+	int congestion_kb;
+
+	/*
+	 * Copied from NFS
+	 *
+	 * congestion size, scale with available memory.
+	 *
+	 *  64MB:    8192k
+	 * 128MB:   11585k
+	 * 256MB:   16384k
+	 * 512MB:   23170k
+	 *   1GB:   32768k
+	 *   2GB:   46340k
+	 *   4GB:   65536k
+	 *   8GB:   92681k
+	 *  16GB:  131072k
+	 *
+	 * This allows larger machines to have larger/more transfers.
+	 * Limit the default to 256M
+	 */
+	congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10);
+	if (congestion_kb > 256*1024)
+		congestion_kb = 256*1024;
+
+	return congestion_kb;
+}
 
 /**
  * ceph_show_options - Show mount options in /proc/mounts
@@ -138,6 +163,35 @@ static int ceph_show_options(struct seq_file *m, struct vfsmount *mnt)
 		seq_puts(m, ",nocrc");
 	if (args->flags & CEPH_OPT_NOASYNCREADDIR)
 		seq_puts(m, ",noasyncreaddir");
+
+	if (args->mount_timeout != CEPH_MOUNT_TIMEOUT_DEFAULT)
+		seq_printf(m, ",mount_timeout=%d", args->mount_timeout);
+	if (args->osd_idle_ttl != CEPH_OSD_IDLE_TTL_DEFAULT)
+		seq_printf(m, ",osd_idle_ttl=%d", args->osd_idle_ttl);
+	if (args->osd_timeout != CEPH_OSD_TIMEOUT_DEFAULT)
+		seq_printf(m, ",osdtimeout=%d", args->osd_timeout);
+	if (args->osd_keepalive_timeout != CEPH_OSD_KEEPALIVE_DEFAULT)
+		seq_printf(m, ",osdkeepalivetimeout=%d",
+			 args->osd_keepalive_timeout);
+	if (args->wsize)
+		seq_printf(m, ",wsize=%d", args->wsize);
+	if (args->rsize != CEPH_MOUNT_RSIZE_DEFAULT)
+		seq_printf(m, ",rsize=%d", args->rsize);
+	if (args->congestion_kb != default_congestion_kb())
+		seq_printf(m, ",write_congestion_kb=%d", args->congestion_kb);
+	if (args->caps_wanted_delay_min != CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT)
+		seq_printf(m, ",caps_wanted_delay_min=%d",
+			 args->caps_wanted_delay_min);
+	if (args->caps_wanted_delay_max != CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT)
+		seq_printf(m, ",caps_wanted_delay_max=%d",
+			   args->caps_wanted_delay_max);
+	if (args->cap_release_safety != CEPH_CAP_RELEASE_SAFETY_DEFAULT)
+		seq_printf(m, ",cap_release_safety=%d",
+			   args->cap_release_safety);
+	if (args->max_readdir != CEPH_MAX_READDIR_DEFAULT)
+		seq_printf(m, ",readdir_max_entries=%d", args->max_readdir);
+	if (args->max_readdir_bytes != CEPH_MAX_READDIR_BYTES_DEFAULT)
+		seq_printf(m, ",readdir_max_bytes=%d", args->max_readdir_bytes);
 	if (strcmp(args->snapdir_name, CEPH_SNAPDIRNAME_DEFAULT))
 		seq_printf(m, ",snapdirname=%s", args->snapdir_name);
 	if (args->name)
@@ -161,35 +215,6 @@ static void ceph_inode_init_once(void *foo)
 	inode_init_once(&ci->vfs_inode);
 }
 
-static int default_congestion_kb(void)
-{
-	int congestion_kb;
-
-	/*
-	 * Copied from NFS
-	 *
-	 * congestion size, scale with available memory.
-	 *
-	 *  64MB:    8192k
-	 * 128MB:   11585k
-	 * 256MB:   16384k
-	 * 512MB:   23170k
-	 *   1GB:   32768k
-	 *   2GB:   46340k
-	 *   4GB:   65536k
-	 *   8GB:   92681k
-	 *  16GB:  131072k
-	 *
-	 * This allows larger machines to have larger/more transfers.
-	 * Limit the default to 256M
-	 */
-	congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10);
-	if (congestion_kb > 256*1024)
-		congestion_kb = 256*1024;
-
-	return congestion_kb;
-}
-
 static int __init init_caches(void)
 {
 	ceph_inode_cachep = kmem_cache_create("ceph_inode_info",
@@ -308,7 +333,9 @@ enum {
 	Opt_osd_idle_ttl,
 	Opt_caps_wanted_delay_min,
 	Opt_caps_wanted_delay_max,
+	Opt_cap_release_safety,
 	Opt_readdir_max_entries,
+	Opt_readdir_max_bytes,
 	Opt_congestion_kb,
 	Opt_last_int,
 	/* int args above */
@@ -339,7 +366,9 @@ static match_table_t arg_tokens = {
 	{Opt_osd_idle_ttl, "osd_idle_ttl=%d"},
 	{Opt_caps_wanted_delay_min, "caps_wanted_delay_min=%d"},
 	{Opt_caps_wanted_delay_max, "caps_wanted_delay_max=%d"},
+	{Opt_cap_release_safety, "cap_release_safety=%d"},
 	{Opt_readdir_max_entries, "readdir_max_entries=%d"},
+	{Opt_readdir_max_bytes, "readdir_max_bytes=%d"},
 	{Opt_congestion_kb, "write_congestion_kb=%d"},
 	/* int args above */
 	{Opt_snapdirname, "snapdirname=%s"},
@@ -388,8 +417,9 @@ static struct ceph_mount_args *parse_mount_args(int flags, char *options,
 	args->caps_wanted_delay_max = CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT;
 	args->rsize = CEPH_MOUNT_RSIZE_DEFAULT;
 	args->snapdir_name = kstrdup(CEPH_SNAPDIRNAME_DEFAULT, GFP_KERNEL);
-	args->cap_release_safety = CEPH_CAPS_PER_RELEASE * 4;
-	args->max_readdir = 1024;
+	args->cap_release_safety = CEPH_CAP_RELEASE_SAFETY_DEFAULT;
+	args->max_readdir = CEPH_MAX_READDIR_DEFAULT;
+	args->max_readdir_bytes = CEPH_MAX_READDIR_BYTES_DEFAULT;
 	args->congestion_kb = default_congestion_kb();
 
 	/* ip1[:port1][,ip2[:port2]...]:/subdir/in/fs */
@@ -497,6 +527,9 @@ static struct ceph_mount_args *parse_mount_args(int flags, char *options,
 		case Opt_readdir_max_entries:
 			args->max_readdir = intval;
 			break;
+		case Opt_readdir_max_bytes:
+			args->max_readdir_bytes = intval;
+			break;
 		case Opt_congestion_kb:
 			args->congestion_kb = intval;
 			break;
@@ -636,9 +669,17 @@ static void ceph_destroy_client(struct ceph_client *client)
 
 	/* unmount */
 	ceph_mdsc_stop(&client->mdsc);
-	ceph_monc_stop(&client->monc);
 	ceph_osdc_stop(&client->osdc);
 
+	/*
+	 * make sure mds and osd connections close out before destroying
+	 * the auth module, which is needed to free those connections'
+	 * ceph_authorizers.
+	 */
+	ceph_msgr_flush();
+
+	ceph_monc_stop(&client->monc);
+
 	ceph_adjust_min_caps(-client->min_caps);
 
 	ceph_debugfs_client_cleanup(client);
@@ -682,9 +723,10 @@ int ceph_check_fsid(struct ceph_client *client, struct ceph_fsid *fsid)
 /*
  * true if we have the mon map (and have thus joined the cluster)
  */
-static int have_mon_map(struct ceph_client *client)
+static int have_mon_and_osd_map(struct ceph_client *client)
 {
-	return client->monc.monmap && client->monc.monmap->epoch;
+	return client->monc.monmap && client->monc.monmap->epoch &&
+	       client->osdc.osdmap && client->osdc.osdmap->epoch;
 }
 
 /*
@@ -704,7 +746,7 @@ static struct dentry *open_root_dentry(struct ceph_client *client,
 	dout("open_root_inode opening '%s'\n", path);
 	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETATTR, USE_ANY_MDS);
 	if (IS_ERR(req))
-		return ERR_PTR(PTR_ERR(req));
+		return ERR_CAST(req);
 	req->r_path1 = kstrdup(path, GFP_NOFS);
 	req->r_ino1.ino = CEPH_INO_ROOT;
 	req->r_ino1.snap = CEPH_NOSNAP;
@@ -762,7 +804,7 @@ static int ceph_mount(struct ceph_client *client, struct vfsmount *mnt,
 	if (err < 0)
 		goto out;
 
-	while (!have_mon_map(client)) {
+	while (!have_mon_and_osd_map(client)) {
 		err = -EIO;
 		if (timeout && time_after_eq(jiffies, started + timeout))
 			goto out;
@@ -770,8 +812,8 @@ static int ceph_mount(struct ceph_client *client, struct vfsmount *mnt,
 		/* wait */
 		dout("mount waiting for mon_map\n");
 		err = wait_event_interruptible_timeout(client->auth_wq,
-			       have_mon_map(client) || (client->auth_err < 0),
-			       timeout);
+		       have_mon_and_osd_map(client) || (client->auth_err < 0),
+		       timeout);
 		if (err == -EINTR || err == -ERESTARTSYS)
 			goto out;
 		if (client->auth_err < 0) {
@@ -884,6 +926,8 @@ static int ceph_compare_super(struct super_block *sb, void *data)
 /*
  * construct our own bdi so we can control readahead, etc.
  */
+static atomic_long_t bdi_seq = ATOMIC_INIT(0);
+
 static int ceph_register_bdi(struct super_block *sb, struct ceph_client *client)
 {
 	int err;
@@ -893,7 +937,8 @@ static int ceph_register_bdi(struct super_block *sb, struct ceph_client *client)
 		client->backing_dev_info.ra_pages =
 			(client->mount_args->rsize + PAGE_CACHE_SIZE - 1)
 			>> PAGE_SHIFT;
-	err = bdi_register_dev(&client->backing_dev_info, sb->s_dev);
+	err = bdi_register(&client->backing_dev_info, NULL, "ceph-%d",
+			   atomic_long_inc_return(&bdi_seq));
 	if (!err)
 		sb->s_bdi = &client->backing_dev_info;
 	return err;
@@ -932,9 +977,9 @@ static int ceph_get_sb(struct file_system_type *fs_type,
 		goto out;
 	}
 
-	if (ceph_client(sb) != client) {
+	if (ceph_sb_to_client(sb) != client) {
 		ceph_destroy_client(client);
-		client = ceph_client(sb);
+		client = ceph_sb_to_client(sb);
 		dout("get_sb got existing client %p\n", client);
 	} else {
 		dout("get_sb using new client %p\n", client);
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 13513b80d87f..10a4a406e887 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -10,7 +10,6 @@
 #include <linux/fs.h>
 #include <linux/mempool.h>
 #include <linux/pagemap.h>
-#include <linux/slab.h>
 #include <linux/wait.h>
 #include <linux/writeback.h>
 #include <linux/slab.h>
@@ -52,24 +51,25 @@
 
 struct ceph_mount_args {
 	int sb_flags;
+	int flags;
+	struct ceph_fsid fsid;
+	struct ceph_entity_addr my_addr;
 	int num_mon;
 	struct ceph_entity_addr *mon_addr;
-	int flags;
 	int mount_timeout;
 	int osd_idle_ttl;
-	int caps_wanted_delay_min, caps_wanted_delay_max;
-	struct ceph_fsid fsid;
-	struct ceph_entity_addr my_addr;
-	int wsize;
-	int rsize;            /* max readahead */
-	int max_readdir;      /* max readdir size */
-	int congestion_kb;      /* max readdir size */
 	int osd_timeout;
 	int osd_keepalive_timeout;
+	int wsize;
+	int rsize;            /* max readahead */
+	int congestion_kb;    /* max writeback in flight */
+	int caps_wanted_delay_min, caps_wanted_delay_max;
+	int cap_release_safety;
+	int max_readdir;       /* max readdir result (entires) */
+	int max_readdir_bytes; /* max readdir result (bytes) */
 	char *snapdir_name;   /* default ".snap" */
 	char *name;
 	char *secret;
-	int cap_release_safety;
 };
 
 /*
@@ -80,13 +80,14 @@ struct ceph_mount_args {
 #define CEPH_OSD_KEEPALIVE_DEFAULT  5
 #define CEPH_OSD_IDLE_TTL_DEFAULT    60
 #define CEPH_MOUNT_RSIZE_DEFAULT    (512*1024) /* readahead */
+#define CEPH_MAX_READDIR_DEFAULT    1024
+#define CEPH_MAX_READDIR_BYTES_DEFAULT    (512*1024)
 
 #define CEPH_MSG_MAX_FRONT_LEN	(16*1024*1024)
 #define CEPH_MSG_MAX_DATA_LEN	(16*1024*1024)
 
 #define CEPH_SNAPDIRNAME_DEFAULT ".snap"
 #define CEPH_AUTH_NAME_DEFAULT   "guest"
-
 /*
  * Delay telling the MDS we no longer want caps, in case we reopen
  * the file.  Delay a minimum amount of time, even if we send a cap
@@ -96,6 +97,7 @@ struct ceph_mount_args {
 #define CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT      5  /* cap release delay */
 #define CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT     60  /* cap release delay */
 
+#define CEPH_CAP_RELEASE_SAFETY_DEFAULT        (CEPH_CAPS_PER_RELEASE * 4)
 
 /* mount state */
 enum {
@@ -160,12 +162,6 @@ struct ceph_client {
 #endif
 };
 
-static inline struct ceph_client *ceph_client(struct super_block *sb)
-{
-	return sb->s_fs_info;
-}
-
-
 /*
  * File i/o capability.  This tracks shared state with the metadata
  * server that allows us to cache or writeback attributes or to read
@@ -814,7 +810,7 @@ extern void ceph_put_cap(struct ceph_cap *cap);
 
 extern void ceph_queue_caps_release(struct inode *inode);
 extern int ceph_write_inode(struct inode *inode, struct writeback_control *wbc);
-extern int ceph_fsync(struct file *file, struct dentry *dentry, int datasync);
+extern int ceph_fsync(struct file *file, int datasync);
 extern void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc,
 				    struct ceph_mds_session *session);
 extern int ceph_get_cap_mds(struct inode *inode);
@@ -871,6 +867,7 @@ extern struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
 extern void ceph_dentry_lru_add(struct dentry *dn);
 extern void ceph_dentry_lru_touch(struct dentry *dn);
 extern void ceph_dentry_lru_del(struct dentry *dn);
+extern void ceph_invalidate_dentry_lease(struct dentry *dentry);
 
 /*
  * our d_ops vary depending on whether the inode is live,
diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c
index 2845422907fc..68aeebc69681 100644
--- a/fs/ceph/xattr.c
+++ b/fs/ceph/xattr.c
@@ -7,7 +7,8 @@
 
 static bool ceph_is_valid_xattr(const char *name)
 {
-	return !strncmp(name, XATTR_SECURITY_PREFIX,
+	return !strncmp(name, "ceph.", 5) ||
+	       !strncmp(name, XATTR_SECURITY_PREFIX,
 			XATTR_SECURITY_PREFIX_LEN) ||
 	       !strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) ||
 	       !strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN);
@@ -76,14 +77,14 @@ static size_t ceph_vxattrcb_rctime(struct ceph_inode_info *ci, char *val,
 }
 
 static struct ceph_vxattr_cb ceph_dir_vxattrs[] = {
-	{ true, "user.ceph.dir.entries", ceph_vxattrcb_entries},
-	{ true, "user.ceph.dir.files", ceph_vxattrcb_files},
-	{ true, "user.ceph.dir.subdirs", ceph_vxattrcb_subdirs},
-	{ true, "user.ceph.dir.rentries", ceph_vxattrcb_rentries},
-	{ true, "user.ceph.dir.rfiles", ceph_vxattrcb_rfiles},
-	{ true, "user.ceph.dir.rsubdirs", ceph_vxattrcb_rsubdirs},
-	{ true, "user.ceph.dir.rbytes", ceph_vxattrcb_rbytes},
-	{ true, "user.ceph.dir.rctime", ceph_vxattrcb_rctime},
+	{ true, "ceph.dir.entries", ceph_vxattrcb_entries},
+	{ true, "ceph.dir.files", ceph_vxattrcb_files},
+	{ true, "ceph.dir.subdirs", ceph_vxattrcb_subdirs},
+	{ true, "ceph.dir.rentries", ceph_vxattrcb_rentries},
+	{ true, "ceph.dir.rfiles", ceph_vxattrcb_rfiles},
+	{ true, "ceph.dir.rsubdirs", ceph_vxattrcb_rsubdirs},
+	{ true, "ceph.dir.rbytes", ceph_vxattrcb_rbytes},
+	{ true, "ceph.dir.rctime", ceph_vxattrcb_rctime},
 	{ true, NULL, NULL }
 };
 
@@ -107,7 +108,7 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val,
 }
 
 static struct ceph_vxattr_cb ceph_file_vxattrs[] = {
-	{ true, "user.ceph.layout", ceph_vxattrcb_layout},
+	{ true, "ceph.layout", ceph_vxattrcb_layout},
 	{ NULL, NULL }
 };
 
@@ -186,12 +187,6 @@ static int __set_xattr(struct ceph_inode_info *ci,
 		ci->i_xattrs.names_size -= xattr->name_len;
 		ci->i_xattrs.vals_size -= xattr->val_len;
 	}
-	if (!xattr) {
-		pr_err("__set_xattr ENOMEM on %p %llx.%llx xattr %s=%s\n",
-		       &ci->vfs_inode, ceph_vinop(&ci->vfs_inode), name,
-		       xattr->val);
-		return -ENOMEM;
-	}
 	ci->i_xattrs.names_size += name_len;
 	ci->i_xattrs.vals_size += val_len;
 	if (val)
@@ -574,7 +569,7 @@ ssize_t ceph_listxattr(struct dentry *dentry, char *names, size_t size)
 	     ci->i_xattrs.version, ci->i_xattrs.index_version);
 
 	if (__ceph_caps_issued_mask(ci, CEPH_CAP_XATTR_SHARED, 1) &&
-	    (ci->i_xattrs.index_version > ci->i_xattrs.version)) {
+	    (ci->i_xattrs.index_version >= ci->i_xattrs.version)) {
 		goto list_xattr;
 	} else {
 		spin_unlock(&inode->i_lock);
@@ -622,7 +617,7 @@ out:
 static int ceph_sync_setxattr(struct dentry *dentry, const char *name,
 			      const char *value, size_t size, int flags)
 {
-	struct ceph_client *client = ceph_client(dentry->d_sb);
+	struct ceph_client *client = ceph_sb_to_client(dentry->d_sb);
 	struct inode *inode = dentry->d_inode;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	struct inode *parent_inode = dentry->d_parent->d_inode;
@@ -641,7 +636,7 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name,
 			return -ENOMEM;
 		err = -ENOMEM;
 		for (i = 0; i < nr_pages; i++) {
-			pages[i] = alloc_page(GFP_NOFS);
+			pages[i] = __page_cache_alloc(GFP_NOFS);
 			if (!pages[i]) {
 				nr_pages = i;
 				goto out;
@@ -779,7 +774,7 @@ out:
 
 static int ceph_send_removexattr(struct dentry *dentry, const char *name)
 {
-	struct ceph_client *client = ceph_client(dentry->d_sb);
+	struct ceph_client *client = ceph_sb_to_client(dentry->d_sb);
 	struct ceph_mds_client *mdsc = &client->mdsc;
 	struct inode *inode = dentry->d_inode;
 	struct inode *parent_inode = dentry->d_parent->d_inode;
diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h
index 0242ff9cbf41..a7eb65c84b1c 100644
--- a/fs/cifs/cifsfs.h
+++ b/fs/cifs/cifsfs.h
@@ -84,7 +84,7 @@ extern ssize_t cifs_user_read(struct file *file, char __user *read_data,
 extern ssize_t cifs_user_write(struct file *file, const char __user *write_data,
 			 size_t write_size, loff_t *poffset);
 extern int cifs_lock(struct file *, int, struct file_lock *);
-extern int cifs_fsync(struct file *, struct dentry *, int);
+extern int cifs_fsync(struct file *, int);
 extern int cifs_flush(struct file *, fl_owner_t id);
 extern int cifs_file_mmap(struct file * , struct vm_area_struct *);
 extern const struct file_operations cifs_dir_ops;
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index a83541ec9713..f1ff785b2292 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -1676,7 +1676,7 @@ static int cifs_write_end(struct file *file, struct address_space *mapping,
 	return rc;
 }
 
-int cifs_fsync(struct file *file, struct dentry *dentry, int datasync)
+int cifs_fsync(struct file *file, int datasync)
 {
 	int xid;
 	int rc = 0;
@@ -1688,7 +1688,7 @@ int cifs_fsync(struct file *file, struct dentry *dentry, int datasync)
 	xid = GetXid();
 
 	cFYI(1, "Sync file - name: %s datasync: 0x%x",
-		dentry->d_name.name, datasync);
+		file->f_path.dentry->d_name.name, datasync);
 
 	rc = filemap_write_and_wait(inode->i_mapping);
 	if (rc == 0) {
diff --git a/fs/coda/coda_int.h b/fs/coda/coda_int.h
index d99860a33890..6b443ff43a19 100644
--- a/fs/coda/coda_int.h
+++ b/fs/coda/coda_int.h
@@ -11,8 +11,7 @@ extern int coda_fake_statfs;
 
 void coda_destroy_inodecache(void);
 int coda_init_inodecache(void);
-int coda_fsync(struct file *coda_file, struct dentry *coda_dentry,
-	       int datasync);
+int coda_fsync(struct file *coda_file, int datasync);
 void coda_sysctl_init(void);
 void coda_sysctl_clean(void);
 
diff --git a/fs/coda/file.c b/fs/coda/file.c
index 7196077b1688..ad3cd2abeeb4 100644
--- a/fs/coda/file.c
+++ b/fs/coda/file.c
@@ -202,10 +202,10 @@ int coda_release(struct inode *coda_inode, struct file *coda_file)
 	return 0;
 }
 
-int coda_fsync(struct file *coda_file, struct dentry *coda_dentry, int datasync)
+int coda_fsync(struct file *coda_file, int datasync)
 {
 	struct file *host_file;
-	struct inode *coda_inode = coda_dentry->d_inode;
+	struct inode *coda_inode = coda_file->f_path.dentry->d_inode;
 	struct coda_file_info *cfi;
 	int err = 0;
 
diff --git a/fs/coda/pioctl.c b/fs/coda/pioctl.c
index 773f2ce9aa06..ca25d96d45c9 100644
--- a/fs/coda/pioctl.c
+++ b/fs/coda/pioctl.c
@@ -1,6 +1,6 @@
 /*
  * Pioctl operations for Coda.
- * Original version: (C) 1996 Peter Braam 
+ * Original version: (C) 1996 Peter Braam
  * Rewritten for Linux 2.1: (C) 1997 Carnegie Mellon University
  *
  * Carnegie Mellon encourages users of this code to contribute improvements
@@ -23,21 +23,22 @@
 #include <linux/coda_fs_i.h>
 #include <linux/coda_psdev.h>
 
+#include <linux/smp_lock.h>
+
 /* pioctl ops */
 static int coda_ioctl_permission(struct inode *inode, int mask);
-static int coda_pioctl(struct inode * inode, struct file * filp, 
-                       unsigned int cmd, unsigned long user_data);
+static long coda_pioctl(struct file *filp, unsigned int cmd,
+			unsigned long user_data);
 
 /* exported from this file */
-const struct inode_operations coda_ioctl_inode_operations =
-{
+const struct inode_operations coda_ioctl_inode_operations = {
 	.permission	= coda_ioctl_permission,
 	.setattr	= coda_setattr,
 };
 
 const struct file_operations coda_ioctl_operations = {
 	.owner		= THIS_MODULE,
-	.ioctl		= coda_pioctl,
+	.unlocked_ioctl	= coda_pioctl,
 };
 
 /* the coda pioctl inode ops */
@@ -46,48 +47,53 @@ static int coda_ioctl_permission(struct inode *inode, int mask)
 	return (mask & MAY_EXEC) ? -EACCES : 0;
 }
 
-static int coda_pioctl(struct inode * inode, struct file * filp, 
-                       unsigned int cmd, unsigned long user_data)
+static long coda_pioctl(struct file *filp, unsigned int cmd,
+			unsigned long user_data)
 {
 	struct path path;
-        int error;
+	int error;
 	struct PioctlData data;
-        struct inode *target_inode = NULL;
-        struct coda_inode_info *cnp;
+	struct inode *inode = filp->f_dentry->d_inode;
+	struct inode *target_inode = NULL;
+	struct coda_inode_info *cnp;
 
-        /* get the Pioctl data arguments from user space */
-        if (copy_from_user(&data, (void __user *)user_data, sizeof(data))) {
-	    return -EINVAL;
-	}
-       
-        /* 
-         * Look up the pathname. Note that the pathname is in 
-         * user memory, and namei takes care of this
-         */
-        if (data.follow) {
-                error = user_path(data.path, &path);
-	} else {
-	        error = user_lpath(data.path, &path);
+	lock_kernel();
+
+	/* get the Pioctl data arguments from user space */
+	if (copy_from_user(&data, (void __user *)user_data, sizeof(data))) {
+		error = -EINVAL;
+		goto out;
 	}
-		
-	if ( error ) {
-		return error;
-        } else {
+
+	/*
+	 * Look up the pathname. Note that the pathname is in
+	 * user memory, and namei takes care of this
+	 */
+	if (data.follow)
+		error = user_path(data.path, &path);
+	else
+		error = user_lpath(data.path, &path);
+
+	if (error)
+		goto out;
+	else
 		target_inode = path.dentry->d_inode;
-	}
-	
+
 	/* return if it is not a Coda inode */
-	if ( target_inode->i_sb != inode->i_sb ) {
+	if (target_inode->i_sb != inode->i_sb) {
 		path_put(&path);
-	        return  -EINVAL;
+		error = -EINVAL;
+		goto out;
 	}
 
 	/* now proceed to make the upcall */
-        cnp = ITOC(target_inode);
+	cnp = ITOC(target_inode);
 
 	error = venus_pioctl(inode->i_sb, &(cnp->c_fid), cmd, &data);
 
 	path_put(&path);
-        return error;
-}
 
+out:
+	unlock_kernel();
+	return error;
+}
diff --git a/fs/coda/psdev.c b/fs/coda/psdev.c
index be4392ca2098..66b9cf79c5ba 100644
--- a/fs/coda/psdev.c
+++ b/fs/coda/psdev.c
@@ -73,8 +73,7 @@ static unsigned int coda_psdev_poll(struct file *file, poll_table * wait)
 	return mask;
 }
 
-static int coda_psdev_ioctl(struct inode * inode, struct file * filp, 
-			    unsigned int cmd, unsigned long arg)
+static long coda_psdev_ioctl(struct file * filp, unsigned int cmd, unsigned long arg)
 {
 	unsigned int data;
 
@@ -344,7 +343,7 @@ static const struct file_operations coda_psdev_fops = {
 	.read		= coda_psdev_read,
 	.write		= coda_psdev_write,
 	.poll		= coda_psdev_poll,
-	.ioctl		= coda_psdev_ioctl,
+	.unlocked_ioctl	= coda_psdev_ioctl,
 	.open		= coda_psdev_open,
 	.release	= coda_psdev_release,
 };
diff --git a/fs/compat.c b/fs/compat.c
index 05448730f840..f0b391c50552 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -568,6 +568,79 @@ out:
 	return ret;
 }
 
+/* A write operation does a read from user space and vice versa */
+#define vrfy_dir(type) ((type) == READ ? VERIFY_WRITE : VERIFY_READ)
+
+ssize_t compat_rw_copy_check_uvector(int type,
+		const struct compat_iovec __user *uvector, unsigned long nr_segs,
+		unsigned long fast_segs, struct iovec *fast_pointer,
+		struct iovec **ret_pointer)
+{
+	compat_ssize_t tot_len;
+	struct iovec *iov = *ret_pointer = fast_pointer;
+	ssize_t ret = 0;
+	int seg;
+
+	/*
+	 * SuS says "The readv() function *may* fail if the iovcnt argument
+	 * was less than or equal to 0, or greater than {IOV_MAX}.  Linux has
+	 * traditionally returned zero for zero segments, so...
+	 */
+	if (nr_segs == 0)
+		goto out;
+
+	ret = -EINVAL;
+	if (nr_segs > UIO_MAXIOV || nr_segs < 0)
+		goto out;
+	if (nr_segs > fast_segs) {
+		ret = -ENOMEM;
+		iov = kmalloc(nr_segs*sizeof(struct iovec), GFP_KERNEL);
+		if (iov == NULL) {
+			*ret_pointer = fast_pointer;
+			goto out;
+		}
+	}
+	*ret_pointer = iov;
+
+	/*
+	 * Single unix specification:
+	 * We should -EINVAL if an element length is not >= 0 and fitting an
+	 * ssize_t.  The total length is fitting an ssize_t
+	 *
+	 * Be careful here because iov_len is a size_t not an ssize_t
+	 */
+	tot_len = 0;
+	ret = -EINVAL;
+	for (seg = 0; seg < nr_segs; seg++) {
+		compat_ssize_t tmp = tot_len;
+		compat_uptr_t buf;
+		compat_ssize_t len;
+
+		if (__get_user(len, &uvector->iov_len) ||
+		   __get_user(buf, &uvector->iov_base)) {
+			ret = -EFAULT;
+			goto out;
+		}
+		if (len < 0)	/* size_t not fitting in compat_ssize_t .. */
+			goto out;
+		tot_len += len;
+		if (tot_len < tmp) /* maths overflow on the compat_ssize_t */
+			goto out;
+		if (!access_ok(vrfy_dir(type), buf, len)) {
+			ret = -EFAULT;
+			goto out;
+		}
+		iov->iov_base = compat_ptr(buf);
+		iov->iov_len = (compat_size_t) len;
+		uvector++;
+		iov++;
+	}
+	ret = tot_len;
+
+out:
+	return ret;
+}
+
 static inline long
 copy_iocb(long nr, u32 __user *ptr32, struct iocb __user * __user *ptr64)
 {
@@ -600,7 +673,7 @@ compat_sys_io_submit(aio_context_t ctx_id, int nr, u32 __user *iocb)
 	iocb64 = compat_alloc_user_space(nr * sizeof(*iocb64));
 	ret = copy_iocb(nr, iocb, iocb64);
 	if (!ret)
-		ret = sys_io_submit(ctx_id, nr, iocb64);
+		ret = do_io_submit(ctx_id, nr, iocb64, 1);
 	return ret;
 }
 
@@ -1077,70 +1150,21 @@ static ssize_t compat_do_readv_writev(int type, struct file *file,
 {
 	compat_ssize_t tot_len;
 	struct iovec iovstack[UIO_FASTIOV];
-	struct iovec *iov=iovstack, *vector;
+	struct iovec *iov;
 	ssize_t ret;
-	int seg;
 	io_fn_t fn;
 	iov_fn_t fnv;
 
-	/*
-	 * SuS says "The readv() function *may* fail if the iovcnt argument
-	 * was less than or equal to 0, or greater than {IOV_MAX}.  Linux has
-	 * traditionally returned zero for zero segments, so...
-	 */
-	ret = 0;
-	if (nr_segs == 0)
-		goto out;
-
-	/*
-	 * First get the "struct iovec" from user memory and
-	 * verify all the pointers
-	 */
 	ret = -EINVAL;
-	if ((nr_segs > UIO_MAXIOV) || (nr_segs <= 0))
-		goto out;
 	if (!file->f_op)
 		goto out;
-	if (nr_segs > UIO_FASTIOV) {
-		ret = -ENOMEM;
-		iov = kmalloc(nr_segs*sizeof(struct iovec), GFP_KERNEL);
-		if (!iov)
-			goto out;
-	}
+
 	ret = -EFAULT;
 	if (!access_ok(VERIFY_READ, uvector, nr_segs*sizeof(*uvector)))
 		goto out;
 
-	/*
-	 * Single unix specification:
-	 * We should -EINVAL if an element length is not >= 0 and fitting an
-	 * ssize_t.  The total length is fitting an ssize_t
-	 *
-	 * Be careful here because iov_len is a size_t not an ssize_t
-	 */
-	tot_len = 0;
-	vector = iov;
-	ret = -EINVAL;
-	for (seg = 0 ; seg < nr_segs; seg++) {
-		compat_ssize_t tmp = tot_len;
-		compat_ssize_t len;
-		compat_uptr_t buf;
-
-		if (__get_user(len, &uvector->iov_len) ||
-		    __get_user(buf, &uvector->iov_base)) {
-			ret = -EFAULT;
-			goto out;
-		}
-		if (len < 0)	/* size_t not fitting an compat_ssize_t .. */
-			goto out;
-		tot_len += len;
-		if (tot_len < tmp) /* maths overflow on the compat_ssize_t */
-			goto out;
-		vector->iov_base = compat_ptr(buf);
-		vector->iov_len = (compat_size_t) len;
-		uvector++;
-		vector++;
-	}
+	tot_len = compat_rw_copy_check_uvector(type, uvector, nr_segs,
+					       UIO_FASTIOV, iovstack, &iov);
 	if (tot_len == 0) {
 		ret = 0;
 		goto out;
diff --git a/fs/configfs/inode.c b/fs/configfs/inode.c
index c8af2d91174b..41645142b88b 100644
--- a/fs/configfs/inode.c
+++ b/fs/configfs/inode.c
@@ -72,16 +72,11 @@ int configfs_setattr(struct dentry * dentry, struct iattr * iattr)
 	if (!sd)
 		return -EINVAL;
 
-	sd_iattr = sd->s_iattr;
-
-	error = inode_change_ok(inode, iattr);
-	if (error)
-		return error;
-
-	error = inode_setattr(inode, iattr);
+	error = simple_setattr(dentry, iattr);
 	if (error)
 		return error;
 
+	sd_iattr = sd->s_iattr;
 	if (!sd_iattr) {
 		/* setting attributes for the first time, allocate now */
 		sd_iattr = kzalloc(sizeof(struct iattr), GFP_KERNEL);
diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c
index 4d74fc72c195..0210898458b2 100644
--- a/fs/debugfs/file.c
+++ b/fs/debugfs/file.c
@@ -277,8 +277,10 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_x32, debugfs_u32_get, debugfs_u32_set, "0x%08llx\n"
 DEFINE_SIMPLE_ATTRIBUTE(fops_x32_ro, debugfs_u32_get, NULL, "0x%08llx\n");
 DEFINE_SIMPLE_ATTRIBUTE(fops_x32_wo, NULL, debugfs_u32_set, "0x%08llx\n");
 
+DEFINE_SIMPLE_ATTRIBUTE(fops_x64, debugfs_u64_get, debugfs_u64_set, "0x%016llx\n");
+
 /*
- * debugfs_create_x{8,16,32} - create a debugfs file that is used to read and write an unsigned {8,16,32}-bit value
+ * debugfs_create_x{8,16,32,64} - create a debugfs file that is used to read and write an unsigned {8,16,32,64}-bit value
  *
  * These functions are exactly the same as the above functions (but use a hex
  * output for the decimal challenged). For details look at the above unsigned
@@ -357,6 +359,23 @@ struct dentry *debugfs_create_x32(const char *name, mode_t mode,
 }
 EXPORT_SYMBOL_GPL(debugfs_create_x32);
 
+/**
+ * debugfs_create_x64 - create a debugfs file that is used to read and write an unsigned 64-bit value
+ * @name: a pointer to a string containing the name of the file to create.
+ * @mode: the permission that the file should have
+ * @parent: a pointer to the parent dentry for this file.  This should be a
+ *          directory dentry if set.  If this parameter is %NULL, then the
+ *          file will be created in the root of the debugfs filesystem.
+ * @value: a pointer to the variable that the file should read to and write
+ *         from.
+ */
+struct dentry *debugfs_create_x64(const char *name, mode_t mode,
+				 struct dentry *parent, u64 *value)
+{
+	return debugfs_create_file(name, mode, parent, value, &fops_x64);
+}
+EXPORT_SYMBOL_GPL(debugfs_create_x64);
+
 
 static int debugfs_size_t_set(void *data, u64 val)
 {
diff --git a/fs/direct-io.c b/fs/direct-io.c
index e82adc2debb7..7600aacf531d 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -82,6 +82,8 @@ struct dio {
 	int reap_counter;		/* rate limit reaping */
 	get_block_t *get_block;		/* block mapping function */
 	dio_iodone_t *end_io;		/* IO completion function */
+	dio_submit_t *submit_io;	/* IO submition function */
+	loff_t logical_offset_in_bio;	/* current first logical block in bio */
 	sector_t final_block_in_bio;	/* current final block in bio + 1 */
 	sector_t next_block_for_io;	/* next block to be put under IO,
 					   in dio_blocks units */
@@ -96,6 +98,7 @@ struct dio {
 	unsigned cur_page_offset;	/* Offset into it, in bytes */
 	unsigned cur_page_len;		/* Nr of bytes at cur_page_offset */
 	sector_t cur_page_block;	/* Where it starts */
+	loff_t cur_page_fs_offset;	/* Offset in file */
 
 	/* BIO completion state */
 	spinlock_t bio_lock;		/* protects BIO fields below */
@@ -300,6 +303,26 @@ static void dio_bio_end_io(struct bio *bio, int error)
 	spin_unlock_irqrestore(&dio->bio_lock, flags);
 }
 
+/**
+ * dio_end_io - handle the end io action for the given bio
+ * @bio: The direct io bio thats being completed
+ * @error: Error if there was one
+ *
+ * This is meant to be called by any filesystem that uses their own dio_submit_t
+ * so that the DIO specific endio actions are dealt with after the filesystem
+ * has done it's completion work.
+ */
+void dio_end_io(struct bio *bio, int error)
+{
+	struct dio *dio = bio->bi_private;
+
+	if (dio->is_async)
+		dio_bio_end_aio(bio, error);
+	else
+		dio_bio_end_io(bio, error);
+}
+EXPORT_SYMBOL_GPL(dio_end_io);
+
 static int
 dio_bio_alloc(struct dio *dio, struct block_device *bdev,
 		sector_t first_sector, int nr_vecs)
@@ -316,6 +339,7 @@ dio_bio_alloc(struct dio *dio, struct block_device *bdev,
 		bio->bi_end_io = dio_bio_end_io;
 
 	dio->bio = bio;
+	dio->logical_offset_in_bio = dio->cur_page_fs_offset;
 	return 0;
 }
 
@@ -340,10 +364,15 @@ static void dio_bio_submit(struct dio *dio)
 	if (dio->is_async && dio->rw == READ)
 		bio_set_pages_dirty(bio);
 
-	submit_bio(dio->rw, bio);
+	if (dio->submit_io)
+		dio->submit_io(dio->rw, bio, dio->inode,
+			       dio->logical_offset_in_bio);
+	else
+		submit_bio(dio->rw, bio);
 
 	dio->bio = NULL;
 	dio->boundary = 0;
+	dio->logical_offset_in_bio = 0;
 }
 
 /*
@@ -603,10 +632,26 @@ static int dio_send_cur_page(struct dio *dio)
 	int ret = 0;
 
 	if (dio->bio) {
+		loff_t cur_offset = dio->block_in_file << dio->blkbits;
+		loff_t bio_next_offset = dio->logical_offset_in_bio +
+			dio->bio->bi_size;
+
 		/*
-		 * See whether this new request is contiguous with the old
+		 * See whether this new request is contiguous with the old.
+		 *
+		 * Btrfs cannot handl having logically non-contiguous requests
+		 * submitted.  For exmple if you have
+		 *
+		 * Logical:  [0-4095][HOLE][8192-12287]
+		 * Phyiscal: [0-4095]      [4096-8181]
+		 *
+		 * We cannot submit those pages together as one BIO.  So if our
+		 * current logical offset in the file does not equal what would
+		 * be the next logical offset in the bio, submit the bio we
+		 * have.
 		 */
-		if (dio->final_block_in_bio != dio->cur_page_block)
+		if (dio->final_block_in_bio != dio->cur_page_block ||
+		    cur_offset != bio_next_offset)
 			dio_bio_submit(dio);
 		/*
 		 * Submit now if the underlying fs is about to perform a
@@ -701,6 +746,7 @@ submit_page_section(struct dio *dio, struct page *page,
 	dio->cur_page_offset = offset;
 	dio->cur_page_len = len;
 	dio->cur_page_block = blocknr;
+	dio->cur_page_fs_offset = dio->block_in_file << dio->blkbits;
 out:
 	return ret;
 }
@@ -935,7 +981,7 @@ static ssize_t
 direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode, 
 	const struct iovec *iov, loff_t offset, unsigned long nr_segs, 
 	unsigned blkbits, get_block_t get_block, dio_iodone_t end_io,
-	struct dio *dio)
+	dio_submit_t submit_io, struct dio *dio)
 {
 	unsigned long user_addr; 
 	unsigned long flags;
@@ -952,6 +998,7 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
 
 	dio->get_block = get_block;
 	dio->end_io = end_io;
+	dio->submit_io = submit_io;
 	dio->final_block_in_bio = -1;
 	dio->next_block_for_io = -1;
 
@@ -1008,7 +1055,7 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
 		}
 	} /* end iovec loop */
 
-	if (ret == -ENOTBLK && (rw & WRITE)) {
+	if (ret == -ENOTBLK) {
 		/*
 		 * The remaining part of the request will be
 		 * be handled by buffered I/O when we return
@@ -1087,30 +1134,11 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
 	return ret;
 }
 
-/*
- * This is a library function for use by filesystem drivers.
- *
- * The locking rules are governed by the flags parameter:
- *  - if the flags value contains DIO_LOCKING we use a fancy locking
- *    scheme for dumb filesystems.
- *    For writes this function is called under i_mutex and returns with
- *    i_mutex held, for reads, i_mutex is not held on entry, but it is
- *    taken and dropped again before returning.
- *    For reads and writes i_alloc_sem is taken in shared mode and released
- *    on I/O completion (which may happen asynchronously after returning to
- *    the caller).
- *
- *  - if the flags value does NOT contain DIO_LOCKING we don't use any
- *    internal locking but rather rely on the filesystem to synchronize
- *    direct I/O reads/writes versus each other and truncate.
- *    For reads and writes both i_mutex and i_alloc_sem are not held on
- *    entry and are never taken.
- */
 ssize_t
-__blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
+__blockdev_direct_IO_newtrunc(int rw, struct kiocb *iocb, struct inode *inode,
 	struct block_device *bdev, const struct iovec *iov, loff_t offset, 
 	unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
-	int flags)
+	dio_submit_t submit_io,	int flags)
 {
 	int seg;
 	size_t size;
@@ -1197,11 +1225,49 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 		(end > i_size_read(inode)));
 
 	retval = direct_io_worker(rw, iocb, inode, iov, offset,
-				nr_segs, blkbits, get_block, end_io, dio);
+				nr_segs, blkbits, get_block, end_io,
+				submit_io, dio);
+
+out:
+	return retval;
+}
+EXPORT_SYMBOL(__blockdev_direct_IO_newtrunc);
+
+/*
+ * This is a library function for use by filesystem drivers.
+ *
+ * The locking rules are governed by the flags parameter:
+ *  - if the flags value contains DIO_LOCKING we use a fancy locking
+ *    scheme for dumb filesystems.
+ *    For writes this function is called under i_mutex and returns with
+ *    i_mutex held, for reads, i_mutex is not held on entry, but it is
+ *    taken and dropped again before returning.
+ *    For reads and writes i_alloc_sem is taken in shared mode and released
+ *    on I/O completion (which may happen asynchronously after returning to
+ *    the caller).
+ *
+ *  - if the flags value does NOT contain DIO_LOCKING we don't use any
+ *    internal locking but rather rely on the filesystem to synchronize
+ *    direct I/O reads/writes versus each other and truncate.
+ *    For reads and writes both i_mutex and i_alloc_sem are not held on
+ *    entry and are never taken.
+ */
+ssize_t
+__blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
+	struct block_device *bdev, const struct iovec *iov, loff_t offset,
+	unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
+	dio_submit_t submit_io,	int flags)
+{
+	ssize_t retval;
 
+	retval = __blockdev_direct_IO_newtrunc(rw, iocb, inode, bdev, iov,
+			offset, nr_segs, get_block, end_io, submit_io, flags);
 	/*
 	 * In case of error extending write may have instantiated a few
 	 * blocks outside i_size. Trim these off again for DIO_LOCKING.
+	 * NOTE: DIO_NO_LOCK/DIO_OWN_LOCK callers have to handle this in
+	 * their own manner. This is a further example of where the old
+	 * truncate sequence is inadequate.
 	 *
 	 * NOTE: filesystems with their own locking have to handle this
 	 * on their own.
@@ -1209,12 +1275,13 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 	if (flags & DIO_LOCKING) {
 		if (unlikely((rw & WRITE) && retval < 0)) {
 			loff_t isize = i_size_read(inode);
+			loff_t end = offset + iov_length(iov, nr_segs);
+
 			if (end > isize)
 				vmtruncate(inode, isize);
 		}
 	}
 
-out:
 	return retval;
 }
 EXPORT_SYMBOL(__blockdev_direct_IO);
diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c
index 3bdddbcc785f..e8fcf4e2ed7d 100644
--- a/fs/ecryptfs/file.c
+++ b/fs/ecryptfs/file.c
@@ -274,7 +274,7 @@ static int ecryptfs_release(struct inode *inode, struct file *file)
 }
 
 static int
-ecryptfs_fsync(struct file *file, struct dentry *dentry, int datasync)
+ecryptfs_fsync(struct file *file, int datasync)
 {
 	return vfs_fsync(ecryptfs_file_to_lower(file), datasync);
 }
diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c
index 65dee2f336ae..31ef5252f0fe 100644
--- a/fs/ecryptfs/inode.c
+++ b/fs/ecryptfs/inode.c
@@ -805,7 +805,7 @@ static int truncate_upper(struct dentry *dentry, struct iattr *ia,
 				    - (ia->ia_size & ~PAGE_CACHE_MASK));
 
 		if (!(crypt_stat->flags & ECRYPTFS_ENCRYPTED)) {
-			rc = vmtruncate(inode, ia->ia_size);
+			rc = simple_setsize(inode, ia->ia_size);
 			if (rc)
 				goto out;
 			lower_ia->ia_size = ia->ia_size;
@@ -830,7 +830,7 @@ static int truncate_upper(struct dentry *dentry, struct iattr *ia,
 				goto out;
 			}
 		}
-		vmtruncate(inode, ia->ia_size);
+		simple_setsize(inode, ia->ia_size);
 		rc = ecryptfs_write_inode_size_to_metadata(inode);
 		if (rc) {
 			printk(KERN_ERR	"Problem with "
diff --git a/fs/exec.c b/fs/exec.c
index e6e94c626c2c..e19de6a80339 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -242,9 +242,10 @@ static int __bprm_mm_init(struct linux_binprm *bprm)
 	 * use STACK_TOP because that can depend on attributes which aren't
 	 * configured yet.
 	 */
+	BUG_ON(VM_STACK_FLAGS & VM_STACK_INCOMPLETE_SETUP);
 	vma->vm_end = STACK_TOP_MAX;
 	vma->vm_start = vma->vm_end - PAGE_SIZE;
-	vma->vm_flags = VM_STACK_FLAGS;
+	vma->vm_flags = VM_STACK_FLAGS | VM_STACK_INCOMPLETE_SETUP;
 	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
 	INIT_LIST_HEAD(&vma->anon_vma_chain);
 	err = insert_vm_struct(mm, vma);
@@ -616,6 +617,7 @@ int setup_arg_pages(struct linux_binprm *bprm,
 	else if (executable_stack == EXSTACK_DISABLE_X)
 		vm_flags &= ~VM_EXEC;
 	vm_flags |= mm->def_flags;
+	vm_flags |= VM_STACK_INCOMPLETE_SETUP;
 
 	ret = mprotect_fixup(vma, &prev, vma->vm_start, vma->vm_end,
 			vm_flags);
@@ -630,6 +632,9 @@ int setup_arg_pages(struct linux_binprm *bprm,
 			goto out_unlock;
 	}
 
+	/* mprotect_fixup is overkill to remove the temporary stack flags */
+	vma->vm_flags &= ~VM_STACK_INCOMPLETE_SETUP;
+
 	stack_expand = 131072UL; /* randomly 32*4k (or 2*64k) pages */
 	stack_size = vma->vm_end - vma->vm_start;
 	/*
@@ -763,7 +768,6 @@ static int de_thread(struct task_struct *tsk)
 	struct signal_struct *sig = tsk->signal;
 	struct sighand_struct *oldsighand = tsk->sighand;
 	spinlock_t *lock = &oldsighand->siglock;
-	int count;
 
 	if (thread_group_empty(tsk))
 		goto no_thread_group;
@@ -780,13 +784,13 @@ static int de_thread(struct task_struct *tsk)
 		spin_unlock_irq(lock);
 		return -EAGAIN;
 	}
+
 	sig->group_exit_task = tsk;
-	zap_other_threads(tsk);
+	sig->notify_count = zap_other_threads(tsk);
+	if (!thread_group_leader(tsk))
+		sig->notify_count--;
 
-	/* Account for the thread group leader hanging around: */
-	count = thread_group_leader(tsk) ? 1 : 2;
-	sig->notify_count = count;
-	while (atomic_read(&sig->count) > count) {
+	while (sig->notify_count) {
 		__set_current_state(TASK_UNINTERRUPTIBLE);
 		spin_unlock_irq(lock);
 		schedule();
@@ -1657,12 +1661,15 @@ static int coredump_wait(int exit_code, struct core_state *core_state)
 	struct task_struct *tsk = current;
 	struct mm_struct *mm = tsk->mm;
 	struct completion *vfork_done;
-	int core_waiters;
+	int core_waiters = -EBUSY;
 
 	init_completion(&core_state->startup);
 	core_state->dumper.task = tsk;
 	core_state->dumper.next = NULL;
-	core_waiters = zap_threads(tsk, mm, core_state, exit_code);
+
+	down_write(&mm->mmap_sem);
+	if (!mm->core_state)
+		core_waiters = zap_threads(tsk, mm, core_state, exit_code);
 	up_write(&mm->mmap_sem);
 
 	if (unlikely(core_waiters < 0))
@@ -1782,21 +1789,61 @@ static void wait_for_dump_helpers(struct file *file)
 }
 
 
+/*
+ * uhm_pipe_setup
+ * helper function to customize the process used
+ * to collect the core in userspace.  Specifically
+ * it sets up a pipe and installs it as fd 0 (stdin)
+ * for the process.  Returns 0 on success, or
+ * PTR_ERR on failure.
+ * Note that it also sets the core limit to 1.  This
+ * is a special value that we use to trap recursive
+ * core dumps
+ */
+static int umh_pipe_setup(struct subprocess_info *info)
+{
+	struct file *rp, *wp;
+	struct fdtable *fdt;
+	struct coredump_params *cp = (struct coredump_params *)info->data;
+	struct files_struct *cf = current->files;
+
+	wp = create_write_pipe(0);
+	if (IS_ERR(wp))
+		return PTR_ERR(wp);
+
+	rp = create_read_pipe(wp, 0);
+	if (IS_ERR(rp)) {
+		free_write_pipe(wp);
+		return PTR_ERR(rp);
+	}
+
+	cp->file = wp;
+
+	sys_close(0);
+	fd_install(0, rp);
+	spin_lock(&cf->file_lock);
+	fdt = files_fdtable(cf);
+	FD_SET(0, fdt->open_fds);
+	FD_CLR(0, fdt->close_on_exec);
+	spin_unlock(&cf->file_lock);
+
+	/* and disallow core files too */
+	current->signal->rlim[RLIMIT_CORE] = (struct rlimit){1, 1};
+
+	return 0;
+}
+
 void do_coredump(long signr, int exit_code, struct pt_regs *regs)
 {
 	struct core_state core_state;
 	char corename[CORENAME_MAX_SIZE + 1];
 	struct mm_struct *mm = current->mm;
 	struct linux_binfmt * binfmt;
-	struct inode * inode;
 	const struct cred *old_cred;
 	struct cred *cred;
 	int retval = 0;
 	int flag = 0;
-	int ispipe = 0;
-	char **helper_argv = NULL;
-	int helper_argc = 0;
-	int dump_count = 0;
+	int ispipe;
 	static atomic_t core_dump_count = ATOMIC_INIT(0);
 	struct coredump_params cprm = {
 		.signr = signr,
@@ -1815,23 +1862,12 @@ void do_coredump(long signr, int exit_code, struct pt_regs *regs)
 	binfmt = mm->binfmt;
 	if (!binfmt || !binfmt->core_dump)
 		goto fail;
-
-	cred = prepare_creds();
-	if (!cred) {
-		retval = -ENOMEM;
+	if (!__get_dumpable(cprm.mm_flags))
 		goto fail;
-	}
 
-	down_write(&mm->mmap_sem);
-	/*
-	 * If another thread got here first, or we are not dumpable, bail out.
-	 */
-	if (mm->core_state || !__get_dumpable(cprm.mm_flags)) {
-		up_write(&mm->mmap_sem);
-		put_cred(cred);
+	cred = prepare_creds();
+	if (!cred)
 		goto fail;
-	}
-
 	/*
 	 *	We cannot trust fsuid as being the "true" uid of the
 	 *	process nor do we know its entire history. We only know it
@@ -1844,10 +1880,8 @@ void do_coredump(long signr, int exit_code, struct pt_regs *regs)
 	}
 
 	retval = coredump_wait(exit_code, &core_state);
-	if (retval < 0) {
-		put_cred(cred);
-		goto fail;
-	}
+	if (retval < 0)
+		goto fail_creds;
 
 	old_cred = override_creds(cred);
 
@@ -1865,19 +1899,19 @@ void do_coredump(long signr, int exit_code, struct pt_regs *regs)
 	ispipe = format_corename(corename, signr);
 	unlock_kernel();
 
-	if ((!ispipe) && (cprm.limit < binfmt->min_coredump))
-		goto fail_unlock;
-
  	if (ispipe) {
-		if (cprm.limit == 0) {
+		int dump_count;
+		char **helper_argv;
+
+		if (cprm.limit == 1) {
 			/*
 			 * Normally core limits are irrelevant to pipes, since
 			 * we're not writing to the file system, but we use
-			 * cprm.limit of 0 here as a speacial value. Any
-			 * non-zero limit gets set to RLIM_INFINITY below, but
+			 * cprm.limit of 1 here as a speacial value. Any
+			 * non-1 limit gets set to RLIM_INFINITY below, but
 			 * a limit of 0 skips the dump.  This is a consistent
 			 * way to catch recursive crashes.  We can still crash
-			 * if the core_pattern binary sets RLIM_CORE =  !0
+			 * if the core_pattern binary sets RLIM_CORE =  !1
 			 * but it runs as root, and can do lots of stupid things
 			 * Note that we use task_tgid_vnr here to grab the pid
 			 * of the process group leader.  That way we get the
@@ -1885,11 +1919,12 @@ void do_coredump(long signr, int exit_code, struct pt_regs *regs)
 			 * core_pattern process dies.
 			 */
 			printk(KERN_WARNING
-				"Process %d(%s) has RLIMIT_CORE set to 0\n",
+				"Process %d(%s) has RLIMIT_CORE set to 1\n",
 				task_tgid_vnr(current), current->comm);
 			printk(KERN_WARNING "Aborting core\n");
 			goto fail_unlock;
 		}
+		cprm.limit = RLIM_INFINITY;
 
 		dump_count = atomic_inc_return(&core_dump_count);
 		if (core_pipe_limit && (core_pipe_limit < dump_count)) {
@@ -1899,71 +1934,74 @@ void do_coredump(long signr, int exit_code, struct pt_regs *regs)
 			goto fail_dropcount;
 		}
 
-		helper_argv = argv_split(GFP_KERNEL, corename+1, &helper_argc);
+		helper_argv = argv_split(GFP_KERNEL, corename+1, NULL);
 		if (!helper_argv) {
 			printk(KERN_WARNING "%s failed to allocate memory\n",
 			       __func__);
 			goto fail_dropcount;
 		}
 
-		cprm.limit = RLIM_INFINITY;
-
-		/* SIGPIPE can happen, but it's just never processed */
-		if (call_usermodehelper_pipe(helper_argv[0], helper_argv, NULL,
-				&cprm.file)) {
+		retval = call_usermodehelper_fns(helper_argv[0], helper_argv,
+					NULL, UMH_WAIT_EXEC, umh_pipe_setup,
+					NULL, &cprm);
+		argv_free(helper_argv);
+		if (retval) {
  			printk(KERN_INFO "Core dump to %s pipe failed\n",
 			       corename);
-			goto fail_dropcount;
+			goto close_fail;
  		}
- 	} else
+	} else {
+		struct inode *inode;
+
+		if (cprm.limit < binfmt->min_coredump)
+			goto fail_unlock;
+
 		cprm.file = filp_open(corename,
 				 O_CREAT | 2 | O_NOFOLLOW | O_LARGEFILE | flag,
 				 0600);
-	if (IS_ERR(cprm.file))
-		goto fail_dropcount;
-	inode = cprm.file->f_path.dentry->d_inode;
-	if (inode->i_nlink > 1)
-		goto close_fail;	/* multiple links - don't dump */
-	if (!ispipe && d_unhashed(cprm.file->f_path.dentry))
-		goto close_fail;
-
-	/* AK: actually i see no reason to not allow this for named pipes etc.,
-	   but keep the previous behaviour for now. */
-	if (!ispipe && !S_ISREG(inode->i_mode))
-		goto close_fail;
-	/*
-	 * Dont allow local users get cute and trick others to coredump
-	 * into their pre-created files:
-	 * Note, this is not relevant for pipes
-	 */
-	if (!ispipe && (inode->i_uid != current_fsuid()))
-		goto close_fail;
-	if (!cprm.file->f_op)
-		goto close_fail;
-	if (!cprm.file->f_op->write)
-		goto close_fail;
-	if (!ispipe &&
-	    do_truncate(cprm.file->f_path.dentry, 0, 0, cprm.file) != 0)
-		goto close_fail;
+		if (IS_ERR(cprm.file))
+			goto fail_unlock;
 
-	retval = binfmt->core_dump(&cprm);
+		inode = cprm.file->f_path.dentry->d_inode;
+		if (inode->i_nlink > 1)
+			goto close_fail;
+		if (d_unhashed(cprm.file->f_path.dentry))
+			goto close_fail;
+		/*
+		 * AK: actually i see no reason to not allow this for named
+		 * pipes etc, but keep the previous behaviour for now.
+		 */
+		if (!S_ISREG(inode->i_mode))
+			goto close_fail;
+		/*
+		 * Dont allow local users get cute and trick others to coredump
+		 * into their pre-created files.
+		 */
+		if (inode->i_uid != current_fsuid())
+			goto close_fail;
+		if (!cprm.file->f_op || !cprm.file->f_op->write)
+			goto close_fail;
+		if (do_truncate(cprm.file->f_path.dentry, 0, 0, cprm.file))
+			goto close_fail;
+	}
 
+	retval = binfmt->core_dump(&cprm);
 	if (retval)
 		current->signal->group_exit_code |= 0x80;
-close_fail:
+
 	if (ispipe && core_pipe_limit)
 		wait_for_dump_helpers(cprm.file);
-	filp_close(cprm.file, NULL);
+close_fail:
+	if (cprm.file)
+		filp_close(cprm.file, NULL);
 fail_dropcount:
-	if (dump_count)
+	if (ispipe)
 		atomic_dec(&core_dump_count);
 fail_unlock:
-	if (helper_argv)
-		argv_free(helper_argv);
-
+	coredump_finish(mm);
 	revert_creds(old_cred);
+fail_creds:
 	put_cred(cred);
-	coredump_finish(mm);
 fail:
 	return;
 }
diff --git a/fs/exofs/dir.c b/fs/exofs/dir.c
index 4cfab1cc75c0..d91e9d829bc1 100644
--- a/fs/exofs/dir.c
+++ b/fs/exofs/dir.c
@@ -608,7 +608,7 @@ int exofs_make_empty(struct inode *inode, struct inode *parent)
 	de->inode_no = cpu_to_le64(parent->i_ino);
 	memcpy(de->name, PARENT_DIR, sizeof(PARENT_DIR));
 	exofs_set_de_type(de, inode);
-	kunmap_atomic(page, KM_USER0);
+	kunmap_atomic(kaddr, KM_USER0);
 	err = exofs_commit_chunk(page, 0, chunk_size);
 fail:
 	page_cache_release(page);
diff --git a/fs/exofs/file.c b/fs/exofs/file.c
index 839b9dc1e70f..fef6899be397 100644
--- a/fs/exofs/file.c
+++ b/fs/exofs/file.c
@@ -40,12 +40,11 @@ static int exofs_release_file(struct inode *inode, struct file *filp)
 	return 0;
 }
 
-static int exofs_file_fsync(struct file *filp, struct dentry *dentry,
-			    int datasync)
+static int exofs_file_fsync(struct file *filp, int datasync)
 {
 	int ret;
 	struct address_space *mapping = filp->f_mapping;
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = mapping->host;
 	struct super_block *sb;
 
 	ret = filemap_write_and_wait(mapping);
@@ -66,7 +65,7 @@ static int exofs_file_fsync(struct file *filp, struct dentry *dentry,
 
 static int exofs_flush(struct file *file, fl_owner_t id)
 {
-	exofs_file_fsync(file, file->f_path.dentry, 1);
+	exofs_file_fsync(file, 1);
 	/* TODO: Flush the OSD target */
 	return 0;
 }
diff --git a/fs/exofs/inode.c b/fs/exofs/inode.c
index d7c6afa79754..4bb6ef822e46 100644
--- a/fs/exofs/inode.c
+++ b/fs/exofs/inode.c
@@ -755,6 +755,21 @@ static int exofs_write_end(struct file *file, struct address_space *mapping,
 	return ret;
 }
 
+static int exofs_releasepage(struct page *page, gfp_t gfp)
+{
+	EXOFS_DBGMSG("page 0x%lx\n", page->index);
+	WARN_ON(1);
+	return try_to_free_buffers(page);
+}
+
+static void exofs_invalidatepage(struct page *page, unsigned long offset)
+{
+	EXOFS_DBGMSG("page_has_buffers=>%d\n", page_has_buffers(page));
+	WARN_ON(1);
+
+	block_invalidatepage(page, offset);
+}
+
 const struct address_space_operations exofs_aops = {
 	.readpage	= exofs_readpage,
 	.readpages	= exofs_readpages,
@@ -762,6 +777,21 @@ const struct address_space_operations exofs_aops = {
 	.writepages	= exofs_writepages,
 	.write_begin	= exofs_write_begin_export,
 	.write_end	= exofs_write_end,
+	.releasepage	= exofs_releasepage,
+	.set_page_dirty	= __set_page_dirty_nobuffers,
+	.invalidatepage = exofs_invalidatepage,
+
+	/* Not implemented Yet */
+	.bmap		= NULL, /* TODO: use osd's OSD_ACT_READ_MAP */
+	.direct_IO	= NULL, /* TODO: Should be trivial to do */
+
+	/* With these NULL has special meaning or default is not exported */
+	.sync_page	= NULL,
+	.get_xip_mem	= NULL,
+	.migratepage	= NULL,
+	.launder_page	= NULL,
+	.is_partially_uptodate = NULL,
+	.error_remove_page = NULL,
 };
 
 /******************************************************************************
diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h
index 0b038e47ad2f..52b34f1d2738 100644
--- a/fs/ext2/ext2.h
+++ b/fs/ext2/ext2.h
@@ -122,7 +122,6 @@ extern int ext2_write_inode (struct inode *, struct writeback_control *);
 extern void ext2_delete_inode (struct inode *);
 extern int ext2_sync_inode (struct inode *);
 extern int ext2_get_block(struct inode *, sector_t, struct buffer_head *, int);
-extern void ext2_truncate (struct inode *);
 extern int ext2_setattr (struct dentry *, struct iattr *);
 extern void ext2_set_inode_flags(struct inode *inode);
 extern void ext2_get_inode_flags(struct ext2_inode_info *);
@@ -155,7 +154,7 @@ extern void ext2_write_super (struct super_block *);
 extern const struct file_operations ext2_dir_operations;
 
 /* file.c */
-extern int ext2_fsync(struct file *file, struct dentry *dentry, int datasync);
+extern int ext2_fsync(struct file *file, int datasync);
 extern const struct inode_operations ext2_file_inode_operations;
 extern const struct file_operations ext2_file_operations;
 extern const struct file_operations ext2_xip_file_operations;
diff --git a/fs/ext2/file.c b/fs/ext2/file.c
index 5d198d0697fb..49eec9456c5b 100644
--- a/fs/ext2/file.c
+++ b/fs/ext2/file.c
@@ -40,13 +40,13 @@ static int ext2_release_file (struct inode * inode, struct file * filp)
 	return 0;
 }
 
-int ext2_fsync(struct file *file, struct dentry *dentry, int datasync)
+int ext2_fsync(struct file *file, int datasync)
 {
 	int ret;
-	struct super_block *sb = dentry->d_inode->i_sb;
+	struct super_block *sb = file->f_mapping->host->i_sb;
 	struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
 
-	ret = simple_fsync(file, dentry, datasync);
+	ret = generic_file_fsync(file, datasync);
 	if (ret == -EIO || test_and_clear_bit(AS_EIO, &mapping->flags)) {
 		/* We don't really know where the IO error happened... */
 		ext2_error(sb, __func__,
@@ -95,7 +95,6 @@ const struct file_operations ext2_xip_file_operations = {
 #endif
 
 const struct inode_operations ext2_file_inode_operations = {
-	.truncate	= ext2_truncate,
 #ifdef CONFIG_EXT2_FS_XATTR
 	.setxattr	= generic_setxattr,
 	.getxattr	= generic_getxattr,
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index 527c46d9bc1f..19214435b752 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -54,6 +54,18 @@ static inline int ext2_inode_is_fast_symlink(struct inode *inode)
 		inode->i_blocks - ea_blocks == 0);
 }
 
+static void ext2_truncate_blocks(struct inode *inode, loff_t offset);
+
+static void ext2_write_failed(struct address_space *mapping, loff_t to)
+{
+	struct inode *inode = mapping->host;
+
+	if (to > inode->i_size) {
+		truncate_pagecache(inode, to, inode->i_size);
+		ext2_truncate_blocks(inode, inode->i_size);
+	}
+}
+
 /*
  * Called at the last iput() if i_nlink is zero.
  */
@@ -71,7 +83,7 @@ void ext2_delete_inode (struct inode * inode)
 
 	inode->i_size = 0;
 	if (inode->i_blocks)
-		ext2_truncate (inode);
+		ext2_truncate_blocks(inode, 0);
 	ext2_free_inode (inode);
 
 	return;
@@ -757,8 +769,8 @@ int __ext2_write_begin(struct file *file, struct address_space *mapping,
 		loff_t pos, unsigned len, unsigned flags,
 		struct page **pagep, void **fsdata)
 {
-	return block_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
-							ext2_get_block);
+	return block_write_begin_newtrunc(file, mapping, pos, len, flags,
+					pagep, fsdata, ext2_get_block);
 }
 
 static int
@@ -766,8 +778,25 @@ ext2_write_begin(struct file *file, struct address_space *mapping,
 		loff_t pos, unsigned len, unsigned flags,
 		struct page **pagep, void **fsdata)
 {
+	int ret;
+
 	*pagep = NULL;
-	return __ext2_write_begin(file, mapping, pos, len, flags, pagep,fsdata);
+	ret = __ext2_write_begin(file, mapping, pos, len, flags, pagep, fsdata);
+	if (ret < 0)
+		ext2_write_failed(mapping, pos + len);
+	return ret;
+}
+
+static int ext2_write_end(struct file *file, struct address_space *mapping,
+			loff_t pos, unsigned len, unsigned copied,
+			struct page *page, void *fsdata)
+{
+	int ret;
+
+	ret = generic_write_end(file, mapping, pos, len, copied, page, fsdata);
+	if (ret < len)
+		ext2_write_failed(mapping, pos + len);
+	return ret;
 }
 
 static int
@@ -775,13 +804,18 @@ ext2_nobh_write_begin(struct file *file, struct address_space *mapping,
 		loff_t pos, unsigned len, unsigned flags,
 		struct page **pagep, void **fsdata)
 {
+	int ret;
+
 	/*
 	 * Dir-in-pagecache still uses ext2_write_begin. Would have to rework
 	 * directory handling code to pass around offsets rather than struct
 	 * pages in order to make this work easily.
 	 */
-	return nobh_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
-							ext2_get_block);
+	ret = nobh_write_begin_newtrunc(file, mapping, pos, len, flags, pagep,
+						fsdata, ext2_get_block);
+	if (ret < 0)
+		ext2_write_failed(mapping, pos + len);
+	return ret;
 }
 
 static int ext2_nobh_writepage(struct page *page,
@@ -800,10 +834,15 @@ ext2_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 			loff_t offset, unsigned long nr_segs)
 {
 	struct file *file = iocb->ki_filp;
-	struct inode *inode = file->f_mapping->host;
-
-	return blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
-				offset, nr_segs, ext2_get_block, NULL);
+	struct address_space *mapping = file->f_mapping;
+	struct inode *inode = mapping->host;
+	ssize_t ret;
+
+	ret = blockdev_direct_IO_newtrunc(rw, iocb, inode, inode->i_sb->s_bdev,
+				iov, offset, nr_segs, ext2_get_block, NULL);
+	if (ret < 0 && (rw & WRITE))
+		ext2_write_failed(mapping, offset + iov_length(iov, nr_segs));
+	return ret;
 }
 
 static int
@@ -818,7 +857,7 @@ const struct address_space_operations ext2_aops = {
 	.writepage		= ext2_writepage,
 	.sync_page		= block_sync_page,
 	.write_begin		= ext2_write_begin,
-	.write_end		= generic_write_end,
+	.write_end		= ext2_write_end,
 	.bmap			= ext2_bmap,
 	.direct_IO		= ext2_direct_IO,
 	.writepages		= ext2_writepages,
@@ -1027,7 +1066,7 @@ static void ext2_free_branches(struct inode *inode, __le32 *p, __le32 *q, int de
 		ext2_free_data(inode, p, q);
 }
 
-void ext2_truncate(struct inode *inode)
+static void __ext2_truncate_blocks(struct inode *inode, loff_t offset)
 {
 	__le32 *i_data = EXT2_I(inode)->i_data;
 	struct ext2_inode_info *ei = EXT2_I(inode);
@@ -1039,27 +1078,8 @@ void ext2_truncate(struct inode *inode)
 	int n;
 	long iblock;
 	unsigned blocksize;
-
-	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
-	    S_ISLNK(inode->i_mode)))
-		return;
-	if (ext2_inode_is_fast_symlink(inode))
-		return;
-	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
-		return;
-
 	blocksize = inode->i_sb->s_blocksize;
-	iblock = (inode->i_size + blocksize-1)
-					>> EXT2_BLOCK_SIZE_BITS(inode->i_sb);
-
-	if (mapping_is_xip(inode->i_mapping))
-		xip_truncate_page(inode->i_mapping, inode->i_size);
-	else if (test_opt(inode->i_sb, NOBH))
-		nobh_truncate_page(inode->i_mapping,
-				inode->i_size, ext2_get_block);
-	else
-		block_truncate_page(inode->i_mapping,
-				inode->i_size, ext2_get_block);
+	iblock = (offset + blocksize-1) >> EXT2_BLOCK_SIZE_BITS(inode->i_sb);
 
 	n = ext2_block_to_path(inode, iblock, offsets, NULL);
 	if (n == 0)
@@ -1127,6 +1147,62 @@ do_indirects:
 	ext2_discard_reservation(inode);
 
 	mutex_unlock(&ei->truncate_mutex);
+}
+
+static void ext2_truncate_blocks(struct inode *inode, loff_t offset)
+{
+	/*
+	 * XXX: it seems like a bug here that we don't allow
+	 * IS_APPEND inode to have blocks-past-i_size trimmed off.
+	 * review and fix this.
+	 *
+	 * Also would be nice to be able to handle IO errors and such,
+	 * but that's probably too much to ask.
+	 */
+	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
+	    S_ISLNK(inode->i_mode)))
+		return;
+	if (ext2_inode_is_fast_symlink(inode))
+		return;
+	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
+		return;
+	__ext2_truncate_blocks(inode, offset);
+}
+
+int ext2_setsize(struct inode *inode, loff_t newsize)
+{
+	loff_t oldsize;
+	int error;
+
+	error = inode_newsize_ok(inode, newsize);
+	if (error)
+		return error;
+
+	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
+	    S_ISLNK(inode->i_mode)))
+		return -EINVAL;
+	if (ext2_inode_is_fast_symlink(inode))
+		return -EINVAL;
+	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
+		return -EPERM;
+
+	if (mapping_is_xip(inode->i_mapping))
+		error = xip_truncate_page(inode->i_mapping, newsize);
+	else if (test_opt(inode->i_sb, NOBH))
+		error = nobh_truncate_page(inode->i_mapping,
+				newsize, ext2_get_block);
+	else
+		error = block_truncate_page(inode->i_mapping,
+				newsize, ext2_get_block);
+	if (error)
+		return error;
+
+	oldsize = inode->i_size;
+	i_size_write(inode, newsize);
+	truncate_pagecache(inode, oldsize, newsize);
+
+	__ext2_truncate_blocks(inode, newsize);
+
 	inode->i_mtime = inode->i_ctime = CURRENT_TIME_SEC;
 	if (inode_needs_sync(inode)) {
 		sync_mapping_buffers(inode->i_mapping);
@@ -1134,6 +1210,8 @@ do_indirects:
 	} else {
 		mark_inode_dirty(inode);
 	}
+
+	return 0;
 }
 
 static struct ext2_inode *ext2_get_inode(struct super_block *sb, ino_t ino,
@@ -1474,8 +1552,15 @@ int ext2_setattr(struct dentry *dentry, struct iattr *iattr)
 		if (error)
 			return error;
 	}
-	error = inode_setattr(inode, iattr);
-	if (!error && (iattr->ia_valid & ATTR_MODE))
+	if (iattr->ia_valid & ATTR_SIZE) {
+		error = ext2_setsize(inode, iattr->ia_size);
+		if (error)
+			return error;
+	}
+	generic_setattr(inode, iattr);
+	if (iattr->ia_valid & ATTR_MODE)
 		error = ext2_acl_chmod(inode);
+	mark_inode_dirty(inode);
+
 	return error;
 }
diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index 71e9eb1fa696..7ff43f4a59cd 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -119,6 +119,8 @@ static void ext2_put_super (struct super_block * sb)
 	int i;
 	struct ext2_sb_info *sbi = EXT2_SB(sb);
 
+	dquot_disable(sb, -1, DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
+
 	if (sb->s_dirt)
 		ext2_write_super(sb);
 
@@ -1063,6 +1065,12 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent)
 	sb->s_op = &ext2_sops;
 	sb->s_export_op = &ext2_export_ops;
 	sb->s_xattr = ext2_xattr_handlers;
+
+#ifdef CONFIG_QUOTA
+	sb->dq_op = &dquot_operations;
+	sb->s_qcop = &dquot_quotactl_ops;
+#endif
+
 	root = ext2_iget(sb, EXT2_ROOT_INO);
 	if (IS_ERR(root)) {
 		ret = PTR_ERR(root);
@@ -1241,6 +1249,7 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data)
 			spin_unlock(&sbi->s_lock);
 			return 0;
 		}
+
 		/*
 		 * OK, we are remounting a valid rw partition rdonly, so set
 		 * the rdonly flag and then mark the partition as valid again.
@@ -1248,6 +1257,13 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data)
 		es->s_state = cpu_to_le16(sbi->s_mount_state);
 		es->s_mtime = cpu_to_le32(get_seconds());
 		spin_unlock(&sbi->s_lock);
+
+		err = dquot_suspend(sb, -1);
+		if (err < 0) {
+			spin_lock(&sbi->s_lock);
+			goto restore_opts;
+		}
+
 		ext2_sync_super(sb, es, 1);
 	} else {
 		__le32 ret = EXT2_HAS_RO_COMPAT_FEATURE(sb,
@@ -1269,8 +1285,12 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data)
 		if (!ext2_setup_super (sb, es, 0))
 			sb->s_flags &= ~MS_RDONLY;
 		spin_unlock(&sbi->s_lock);
+
 		ext2_write_super(sb);
+
+		dquot_resume(sb, -1);
 	}
+
 	return 0;
 restore_opts:
 	sbi->s_mount_opt = old_opts.s_mount_opt;
diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c
index 373fa90c796a..e2e72c367cf6 100644
--- a/fs/ext3/dir.c
+++ b/fs/ext3/dir.c
@@ -297,7 +297,7 @@ static void free_rb_tree_fname(struct rb_root *root)
 			kfree (old);
 		}
 		if (!parent)
-			root->rb_node = NULL;
+			*root = RB_ROOT;
 		else if (parent->rb_left == n)
 			parent->rb_left = NULL;
 		else if (parent->rb_right == n)
diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c
index fcf7487734b6..d7e9f74dc3a6 100644
--- a/fs/ext3/fsync.c
+++ b/fs/ext3/fsync.c
@@ -43,9 +43,9 @@
  * inode to disk.
  */
 
-int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync)
+int ext3_sync_file(struct file *file, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	struct ext3_inode_info *ei = EXT3_I(inode);
 	journal_t *journal = EXT3_SB(inode->i_sb)->s_journal;
 	int ret, needs_barrier = 0;
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 0fc1293d0e96..6c953bb255e7 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -410,6 +410,8 @@ static void ext3_put_super (struct super_block * sb)
 	struct ext3_super_block *es = sbi->s_es;
 	int i, err;
 
+	dquot_disable(sb, -1, DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
+
 	lock_kernel();
 
 	ext3_xattr_put_super(sb);
@@ -748,7 +750,7 @@ static int ext3_release_dquot(struct dquot *dquot);
 static int ext3_mark_dquot_dirty(struct dquot *dquot);
 static int ext3_write_info(struct super_block *sb, int type);
 static int ext3_quota_on(struct super_block *sb, int type, int format_id,
-				char *path, int remount);
+				char *path);
 static int ext3_quota_on_mount(struct super_block *sb, int type);
 static ssize_t ext3_quota_read(struct super_block *sb, int type, char *data,
 			       size_t len, loff_t off);
@@ -767,12 +769,12 @@ static const struct dquot_operations ext3_quota_operations = {
 
 static const struct quotactl_ops ext3_qctl_operations = {
 	.quota_on	= ext3_quota_on,
-	.quota_off	= vfs_quota_off,
-	.quota_sync	= vfs_quota_sync,
-	.get_info	= vfs_get_dqinfo,
-	.set_info	= vfs_set_dqinfo,
-	.get_dqblk	= vfs_get_dqblk,
-	.set_dqblk	= vfs_set_dqblk
+	.quota_off	= dquot_quota_off,
+	.quota_sync	= dquot_quota_sync,
+	.get_info	= dquot_get_dqinfo,
+	.set_info	= dquot_set_dqinfo,
+	.get_dqblk	= dquot_get_dqblk,
+	.set_dqblk	= dquot_set_dqblk
 };
 #endif
 
@@ -1527,7 +1529,7 @@ static void ext3_orphan_cleanup (struct super_block * sb,
 	/* Turn quotas off */
 	for (i = 0; i < MAXQUOTAS; i++) {
 		if (sb_dqopt(sb)->files[i])
-			vfs_quota_off(sb, i, 0);
+			dquot_quota_off(sb, i);
 	}
 #endif
 	sb->s_flags = s_flags; /* Restore MS_RDONLY status */
@@ -2551,6 +2553,7 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data)
 	ext3_fsblk_t n_blocks_count = 0;
 	unsigned long old_sb_flags;
 	struct ext3_mount_options old_opts;
+	int enable_quota = 0;
 	int err;
 #ifdef CONFIG_QUOTA
 	int i;
@@ -2597,6 +2600,10 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data)
 		}
 
 		if (*flags & MS_RDONLY) {
+			err = dquot_suspend(sb, -1);
+			if (err < 0)
+				goto restore_opts;
+
 			/*
 			 * First of all, the unconditional stuff we have to do
 			 * to disable replay of the journal when we next remount
@@ -2651,6 +2658,7 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data)
 				goto restore_opts;
 			if (!ext3_setup_super (sb, es, 0))
 				sb->s_flags &= ~MS_RDONLY;
+			enable_quota = 1;
 		}
 	}
 #ifdef CONFIG_QUOTA
@@ -2662,6 +2670,9 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data)
 #endif
 	unlock_super(sb);
 	unlock_kernel();
+
+	if (enable_quota)
+		dquot_resume(sb, -1);
 	return 0;
 restore_opts:
 	sb->s_flags = old_sb_flags;
@@ -2851,24 +2862,21 @@ static int ext3_write_info(struct super_block *sb, int type)
  */
 static int ext3_quota_on_mount(struct super_block *sb, int type)
 {
-	return vfs_quota_on_mount(sb, EXT3_SB(sb)->s_qf_names[type],
-			EXT3_SB(sb)->s_jquota_fmt, type);
+	return dquot_quota_on_mount(sb, EXT3_SB(sb)->s_qf_names[type],
+					EXT3_SB(sb)->s_jquota_fmt, type);
 }
 
 /*
  * Standard function to be called on quota_on
  */
 static int ext3_quota_on(struct super_block *sb, int type, int format_id,
-			 char *name, int remount)
+			 char *name)
 {
 	int err;
 	struct path path;
 
 	if (!test_opt(sb, QUOTA))
 		return -EINVAL;
-	/* When remounting, no checks are needed and in fact, name is NULL */
-	if (remount)
-		return vfs_quota_on(sb, type, format_id, name, remount);
 
 	err = kern_path(name, LOOKUP_FOLLOW, &path);
 	if (err)
@@ -2906,7 +2914,7 @@ static int ext3_quota_on(struct super_block *sb, int type, int format_id,
 		}
 	}
 
-	err = vfs_quota_on_path(sb, type, format_id, &path);
+	err = dquot_quota_on_path(sb, type, format_id, &path);
 	path_put(&path);
 	return err;
 }
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index d2f37a5516c7..95b7594c76f9 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -591,14 +591,15 @@ ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
 	ret = ext4_mb_new_blocks(handle, &ar, errp);
 	if (count)
 		*count = ar.len;
-
 	/*
-	 * Account for the allocated meta blocks
+	 * Account for the allocated meta blocks.  We will never
+	 * fail EDQUOT for metdata, but we do account for it.
 	 */
 	if (!(*errp) && EXT4_I(inode)->i_delalloc_reserved_flag) {
 		spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
 		EXT4_I(inode)->i_allocated_meta_blocks += ar.len;
 		spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
+		dquot_alloc_block_nofail(inode, ar.len);
 	}
 	return ret;
 }
diff --git a/fs/ext4/block_validity.c b/fs/ext4/block_validity.c
index 538c48655084..5b6973fbf1bd 100644
--- a/fs/ext4/block_validity.c
+++ b/fs/ext4/block_validity.c
@@ -72,9 +72,9 @@ static int add_system_zone(struct ext4_sb_info *sbi,
 		else if (start_blk >= (entry->start_blk + entry->count))
 			n = &(*n)->rb_right;
 		else {
-			if (start_blk + count > (entry->start_blk + 
+			if (start_blk + count > (entry->start_blk +
 						 entry->count))
-				entry->count = (start_blk + count - 
+				entry->count = (start_blk + count -
 						entry->start_blk);
 			new_node = *n;
 			new_entry = rb_entry(new_node, struct ext4_system_zone,
diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c
index 86cb6d86a048..ea5e6cb7e2a5 100644
--- a/fs/ext4/dir.c
+++ b/fs/ext4/dir.c
@@ -83,11 +83,10 @@ int ext4_check_dir_entry(const char *function, struct inode *dir,
 		error_msg = "inode out of bounds";
 
 	if (error_msg != NULL)
-		__ext4_error(dir->i_sb, function,
-			"bad entry in directory #%lu: %s - block=%llu"
+		ext4_error_inode(function, dir,
+			"bad entry in directory: %s - block=%llu"
 			"offset=%u(%u), inode=%u, rec_len=%d, name_len=%d",
-			dir->i_ino, error_msg, 
-			(unsigned long long) bh->b_blocknr,     
+			error_msg, (unsigned long long) bh->b_blocknr,
 			(unsigned) (offset%bh->b_size), offset,
 			le32_to_cpu(de->inode),
 			rlen, de->name_len);
@@ -111,7 +110,7 @@ static int ext4_readdir(struct file *filp,
 
 	if (EXT4_HAS_COMPAT_FEATURE(inode->i_sb,
 				    EXT4_FEATURE_COMPAT_DIR_INDEX) &&
-	    ((EXT4_I(inode)->i_flags & EXT4_INDEX_FL) ||
+	    ((ext4_test_inode_flag(inode, EXT4_INODE_INDEX)) ||
 	     ((inode->i_size >> sb->s_blocksize_bits) == 1))) {
 		err = ext4_dx_readdir(filp, dirent, filldir);
 		if (err != ERR_BAD_DX_DIR) {
@@ -122,20 +121,20 @@ static int ext4_readdir(struct file *filp,
 		 * We don't set the inode dirty flag since it's not
 		 * critical that it get flushed back to the disk.
 		 */
-		EXT4_I(filp->f_path.dentry->d_inode)->i_flags &= ~EXT4_INDEX_FL;
+		ext4_clear_inode_flag(filp->f_path.dentry->d_inode, EXT4_INODE_INDEX);
 	}
 	stored = 0;
 	offset = filp->f_pos & (sb->s_blocksize - 1);
 
 	while (!error && !stored && filp->f_pos < inode->i_size) {
-		ext4_lblk_t blk = filp->f_pos >> EXT4_BLOCK_SIZE_BITS(sb);
-		struct buffer_head map_bh;
+		struct ext4_map_blocks map;
 		struct buffer_head *bh = NULL;
 
-		map_bh.b_state = 0;
-		err = ext4_get_blocks(NULL, inode, blk, 1, &map_bh, 0);
+		map.m_lblk = filp->f_pos >> EXT4_BLOCK_SIZE_BITS(sb);
+		map.m_len = 1;
+		err = ext4_map_blocks(NULL, inode, &map, 0);
 		if (err > 0) {
-			pgoff_t index = map_bh.b_blocknr >>
+			pgoff_t index = map.m_pblk >>
 					(PAGE_CACHE_SHIFT - inode->i_blkbits);
 			if (!ra_has_index(&filp->f_ra, index))
 				page_cache_sync_readahead(
@@ -143,7 +142,7 @@ static int ext4_readdir(struct file *filp,
 					&filp->f_ra, filp,
 					index, 1);
 			filp->f_ra.prev_pos = (loff_t)index << PAGE_CACHE_SHIFT;
-			bh = ext4_bread(NULL, inode, blk, 0, &err);
+			bh = ext4_bread(NULL, inode, map.m_lblk, 0, &err);
 		}
 
 		/*
@@ -152,9 +151,8 @@ static int ext4_readdir(struct file *filp,
 		 */
 		if (!bh) {
 			if (!dir_has_error) {
-				ext4_error(sb, "directory #%lu "
+				EXT4_ERROR_INODE(inode, "directory "
 					   "contains a hole at offset %Lu",
-					   inode->i_ino,
 					   (unsigned long long) filp->f_pos);
 				dir_has_error = 1;
 			}
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index bf938cf7c5f0..19a4de57128a 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -29,6 +29,9 @@
 #include <linux/wait.h>
 #include <linux/blockgroup_lock.h>
 #include <linux/percpu_counter.h>
+#ifdef __KERNEL__
+#include <linux/compat.h>
+#endif
 
 /*
  * The fourth extended filesystem constants/structures
@@ -54,10 +57,10 @@
 #endif
 
 #define EXT4_ERROR_INODE(inode, fmt, a...) \
-	ext4_error_inode(__func__, (inode), (fmt), ## a);
+	ext4_error_inode(__func__, (inode), (fmt), ## a)
 
 #define EXT4_ERROR_FILE(file, fmt, a...)	\
-	ext4_error_file(__func__, (file), (fmt), ## a);
+	ext4_error_file(__func__, (file), (fmt), ## a)
 
 /* data type for block offset of block group */
 typedef int ext4_grpblk_t;
@@ -72,7 +75,7 @@ typedef __u32 ext4_lblk_t;
 typedef unsigned int ext4_group_t;
 
 /*
- * Flags used in mballoc's allocation_context flags field.  
+ * Flags used in mballoc's allocation_context flags field.
  *
  * Also used to show what's going on for debugging purposes when the
  * flag field is exported via the traceport interface
@@ -126,6 +129,29 @@ struct ext4_allocation_request {
 };
 
 /*
+ * Logical to physical block mapping, used by ext4_map_blocks()
+ *
+ * This structure is used to pass requests into ext4_map_blocks() as
+ * well as to store the information returned by ext4_map_blocks().  It
+ * takes less room on the stack than a struct buffer_head.
+ */
+#define EXT4_MAP_NEW		(1 << BH_New)
+#define EXT4_MAP_MAPPED		(1 << BH_Mapped)
+#define EXT4_MAP_UNWRITTEN	(1 << BH_Unwritten)
+#define EXT4_MAP_BOUNDARY	(1 << BH_Boundary)
+#define EXT4_MAP_UNINIT		(1 << BH_Uninit)
+#define EXT4_MAP_FLAGS		(EXT4_MAP_NEW | EXT4_MAP_MAPPED |\
+				 EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY |\
+				 EXT4_MAP_UNINIT)
+
+struct ext4_map_blocks {
+	ext4_fsblk_t m_pblk;
+	ext4_lblk_t m_lblk;
+	unsigned int m_len;
+	unsigned int m_flags;
+};
+
+/*
  * For delayed allocation tracking
  */
 struct mpage_da_data {
@@ -321,6 +347,83 @@ static inline __u32 ext4_mask_flags(umode_t mode, __u32 flags)
 		return flags & EXT4_OTHER_FLMASK;
 }
 
+/*
+ * Inode flags used for atomic set/get
+ */
+enum {
+	EXT4_INODE_SECRM	= 0,	/* Secure deletion */
+	EXT4_INODE_UNRM		= 1,	/* Undelete */
+	EXT4_INODE_COMPR	= 2,	/* Compress file */
+	EXT4_INODE_SYNC		= 3,	/* Synchronous updates */
+	EXT4_INODE_IMMUTABLE	= 4,	/* Immutable file */
+	EXT4_INODE_APPEND	= 5,	/* writes to file may only append */
+	EXT4_INODE_NODUMP	= 6,	/* do not dump file */
+	EXT4_INODE_NOATIME	= 7,	/* do not update atime */
+/* Reserved for compression usage... */
+	EXT4_INODE_DIRTY	= 8,
+	EXT4_INODE_COMPRBLK	= 9,	/* One or more compressed clusters */
+	EXT4_INODE_NOCOMPR	= 10,	/* Don't compress */
+	EXT4_INODE_ECOMPR	= 11,	/* Compression error */
+/* End compression flags --- maybe not all used */
+	EXT4_INODE_INDEX	= 12,	/* hash-indexed directory */
+	EXT4_INODE_IMAGIC	= 13,	/* AFS directory */
+	EXT4_INODE_JOURNAL_DATA	= 14,	/* file data should be journaled */
+	EXT4_INODE_NOTAIL	= 15,	/* file tail should not be merged */
+	EXT4_INODE_DIRSYNC	= 16,	/* dirsync behaviour (directories only) */
+	EXT4_INODE_TOPDIR	= 17,	/* Top of directory hierarchies*/
+	EXT4_INODE_HUGE_FILE	= 18,	/* Set to each huge file */
+	EXT4_INODE_EXTENTS	= 19,	/* Inode uses extents */
+	EXT4_INODE_EA_INODE	= 21,	/* Inode used for large EA */
+	EXT4_INODE_EOFBLOCKS	= 22,	/* Blocks allocated beyond EOF */
+	EXT4_INODE_RESERVED	= 31,	/* reserved for ext4 lib */
+};
+
+#define TEST_FLAG_VALUE(FLAG) (EXT4_##FLAG##_FL == (1 << EXT4_INODE_##FLAG))
+#define CHECK_FLAG_VALUE(FLAG) if (!TEST_FLAG_VALUE(FLAG)) { \
+	printk(KERN_EMERG "EXT4 flag fail: " #FLAG ": %d %d\n", \
+		EXT4_##FLAG##_FL, EXT4_INODE_##FLAG); BUG_ON(1); }
+
+/*
+ * Since it's pretty easy to mix up bit numbers and hex values, and we
+ * can't do a compile-time test for ENUM values, we use a run-time
+ * test to make sure that EXT4_XXX_FL is consistent with respect to
+ * EXT4_INODE_XXX.  If all is well the printk and BUG_ON will all drop
+ * out so it won't cost any extra space in the compiled kernel image.
+ * But it's important that these values are the same, since we are
+ * using EXT4_INODE_XXX to test for the flag values, but EXT4_XX_FL
+ * must be consistent with the values of FS_XXX_FL defined in
+ * include/linux/fs.h and the on-disk values found in ext2, ext3, and
+ * ext4 filesystems, and of course the values defined in e2fsprogs.
+ *
+ * It's not paranoia if the Murphy's Law really *is* out to get you.  :-)
+ */
+static inline void ext4_check_flag_values(void)
+{
+	CHECK_FLAG_VALUE(SECRM);
+	CHECK_FLAG_VALUE(UNRM);
+	CHECK_FLAG_VALUE(COMPR);
+	CHECK_FLAG_VALUE(SYNC);
+	CHECK_FLAG_VALUE(IMMUTABLE);
+	CHECK_FLAG_VALUE(APPEND);
+	CHECK_FLAG_VALUE(NODUMP);
+	CHECK_FLAG_VALUE(NOATIME);
+	CHECK_FLAG_VALUE(DIRTY);
+	CHECK_FLAG_VALUE(COMPRBLK);
+	CHECK_FLAG_VALUE(NOCOMPR);
+	CHECK_FLAG_VALUE(ECOMPR);
+	CHECK_FLAG_VALUE(INDEX);
+	CHECK_FLAG_VALUE(IMAGIC);
+	CHECK_FLAG_VALUE(JOURNAL_DATA);
+	CHECK_FLAG_VALUE(NOTAIL);
+	CHECK_FLAG_VALUE(DIRSYNC);
+	CHECK_FLAG_VALUE(TOPDIR);
+	CHECK_FLAG_VALUE(HUGE_FILE);
+	CHECK_FLAG_VALUE(EXTENTS);
+	CHECK_FLAG_VALUE(EA_INODE);
+	CHECK_FLAG_VALUE(EOFBLOCKS);
+	CHECK_FLAG_VALUE(RESERVED);
+}
+
 /* Used to pass group descriptor data when online resize is done */
 struct ext4_new_group_input {
 	__u32 group;		/* Group number for this data */
@@ -332,6 +435,18 @@ struct ext4_new_group_input {
 	__u16 unused;
 };
 
+#if defined(__KERNEL__) && defined(CONFIG_COMPAT)
+struct compat_ext4_new_group_input {
+	u32 group;
+	compat_u64 block_bitmap;
+	compat_u64 inode_bitmap;
+	compat_u64 inode_table;
+	u32 blocks_count;
+	u16 reserved_blocks;
+	u16 unused;
+};
+#endif
+
 /* The struct ext4_new_group_input in kernel space, with free_blocks_count */
 struct ext4_new_group_data {
 	__u32 group;
@@ -355,7 +470,7 @@ struct ext4_new_group_data {
 #define EXT4_GET_BLOCKS_CREATE_UNINIT_EXT	(EXT4_GET_BLOCKS_UNINIT_EXT|\
 						 EXT4_GET_BLOCKS_CREATE)
 	/* Caller is from the delayed allocation writeout path,
-	   so set the magic i_delalloc_reserve_flag after taking the 
+	   so set the magic i_delalloc_reserve_flag after taking the
 	   inode allocation semaphore for */
 #define EXT4_GET_BLOCKS_DELALLOC_RESERVE	0x0004
 	/* caller is from the direct IO path, request to creation of an
@@ -398,6 +513,7 @@ struct ext4_new_group_data {
 #define EXT4_IOC_ALLOC_DA_BLKS		_IO('f', 12)
 #define EXT4_IOC_MOVE_EXT		_IOWR('f', 15, struct move_extent)
 
+#if defined(__KERNEL__) && defined(CONFIG_COMPAT)
 /*
  * ioctl commands in 32 bit emulation
  */
@@ -408,11 +524,13 @@ struct ext4_new_group_data {
 #define EXT4_IOC32_GETRSVSZ		_IOR('f', 5, int)
 #define EXT4_IOC32_SETRSVSZ		_IOW('f', 6, int)
 #define EXT4_IOC32_GROUP_EXTEND		_IOW('f', 7, unsigned int)
+#define EXT4_IOC32_GROUP_ADD		_IOW('f', 8, struct compat_ext4_new_group_input)
 #ifdef CONFIG_JBD2_DEBUG
 #define EXT4_IOC32_WAIT_FOR_READONLY	_IOR('f', 99, int)
 #endif
 #define EXT4_IOC32_GETVERSION_OLD	FS_IOC32_GETVERSION
 #define EXT4_IOC32_SETVERSION_OLD	FS_IOC32_SETVERSION
+#endif
 
 
 /*
@@ -616,9 +734,8 @@ struct ext4_ext_cache {
  */
 struct ext4_inode_info {
 	__le32	i_data[15];	/* unconverted */
-	__u32	i_flags;
-	ext4_fsblk_t	i_file_acl;
 	__u32	i_dtime;
+	ext4_fsblk_t	i_file_acl;
 
 	/*
 	 * i_block_group is the number of the block group which contains
@@ -629,6 +746,7 @@ struct ext4_inode_info {
 	 */
 	ext4_group_t	i_block_group;
 	unsigned long	i_state_flags;		/* Dynamic state flags */
+	unsigned long	i_flags;
 
 	ext4_lblk_t		i_dir_start_lookup;
 #ifdef CONFIG_EXT4_FS_XATTR
@@ -1062,22 +1180,25 @@ enum {
 	EXT4_STATE_DA_ALLOC_CLOSE,	/* Alloc DA blks on close */
 	EXT4_STATE_EXT_MIGRATE,		/* Inode is migrating */
 	EXT4_STATE_DIO_UNWRITTEN,	/* need convert on dio done*/
+	EXT4_STATE_NEWENTRY,		/* File just added to dir */
 };
 
-static inline int ext4_test_inode_state(struct inode *inode, int bit)
-{
-	return test_bit(bit, &EXT4_I(inode)->i_state_flags);
-}
-
-static inline void ext4_set_inode_state(struct inode *inode, int bit)
-{
-	set_bit(bit, &EXT4_I(inode)->i_state_flags);
+#define EXT4_INODE_BIT_FNS(name, field)					\
+static inline int ext4_test_inode_##name(struct inode *inode, int bit)	\
+{									\
+	return test_bit(bit, &EXT4_I(inode)->i_##field);		\
+}									\
+static inline void ext4_set_inode_##name(struct inode *inode, int bit)	\
+{									\
+	set_bit(bit, &EXT4_I(inode)->i_##field);			\
+}									\
+static inline void ext4_clear_inode_##name(struct inode *inode, int bit) \
+{									\
+	clear_bit(bit, &EXT4_I(inode)->i_##field);			\
 }
 
-static inline void ext4_clear_inode_state(struct inode *inode, int bit)
-{
-	clear_bit(bit, &EXT4_I(inode)->i_state_flags);
-}
+EXT4_INODE_BIT_FNS(flag, flags)
+EXT4_INODE_BIT_FNS(state, state_flags)
 #else
 /* Assume that user mode programs are passing in an ext4fs superblock, not
  * a kernel struct super_block.  This will allow us to call the feature-test
@@ -1264,7 +1385,7 @@ struct ext4_dir_entry_2 {
 
 #define is_dx(dir) (EXT4_HAS_COMPAT_FEATURE(dir->i_sb, \
 				      EXT4_FEATURE_COMPAT_DIR_INDEX) && \
-		      (EXT4_I(dir)->i_flags & EXT4_INDEX_FL))
+		    ext4_test_inode_flag((dir), EXT4_INODE_INDEX))
 #define EXT4_DIR_LINK_MAX(dir) (!is_dx(dir) && (dir)->i_nlink >= EXT4_LINK_MAX)
 #define EXT4_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1)
 
@@ -1398,7 +1519,7 @@ extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash,
 extern void ext4_htree_free_dir_info(struct dir_private_info *p);
 
 /* fsync.c */
-extern int ext4_sync_file(struct file *, struct dentry *, int);
+extern int ext4_sync_file(struct file *, int);
 
 /* hash.c */
 extern int ext4fs_dirhash(const char *name, int len, struct
@@ -1678,6 +1799,7 @@ struct ext4_group_info {
 	ext4_grpblk_t	bb_first_free;	/* first free block */
 	ext4_grpblk_t	bb_free;	/* total free blocks */
 	ext4_grpblk_t	bb_fragments;	/* nr of freespace fragments */
+	ext4_grpblk_t	bb_largest_free_order;/* order of largest frag in BG */
 	struct          list_head bb_prealloc_list;
 #ifdef DOUBLE_CHECK
 	void            *bb_bitmap;
@@ -1772,9 +1894,8 @@ extern int ext4_ext_tree_init(handle_t *handle, struct inode *);
 extern int ext4_ext_writepage_trans_blocks(struct inode *, int);
 extern int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks,
 				       int chunk);
-extern int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
-			       ext4_lblk_t iblock, unsigned int max_blocks,
-			       struct buffer_head *bh_result, int flags);
+extern int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
+			       struct ext4_map_blocks *map, int flags);
 extern void ext4_ext_truncate(struct inode *);
 extern void ext4_ext_init(struct super_block *);
 extern void ext4_ext_release(struct super_block *);
@@ -1782,6 +1903,8 @@ extern long ext4_fallocate(struct inode *inode, int mode, loff_t offset,
 			  loff_t len);
 extern int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset,
 			  ssize_t len);
+extern int ext4_map_blocks(handle_t *handle, struct inode *inode,
+			   struct ext4_map_blocks *map, int flags);
 extern int ext4_get_blocks(handle_t *handle, struct inode *inode,
 			   sector_t block, unsigned int max_blocks,
 			   struct buffer_head *bh, int flags);
diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h
index b79ad5126468..dade0c024797 100644
--- a/fs/ext4/ext4_jbd2.h
+++ b/fs/ext4/ext4_jbd2.h
@@ -273,7 +273,7 @@ static inline int ext4_should_journal_data(struct inode *inode)
 		return 1;
 	if (test_opt(inode->i_sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA)
 		return 1;
-	if (EXT4_I(inode)->i_flags & EXT4_JOURNAL_DATA_FL)
+	if (ext4_test_inode_flag(inode, EXT4_INODE_JOURNAL_DATA))
 		return 1;
 	return 0;
 }
@@ -284,7 +284,7 @@ static inline int ext4_should_order_data(struct inode *inode)
 		return 0;
 	if (!S_ISREG(inode->i_mode))
 		return 0;
-	if (EXT4_I(inode)->i_flags & EXT4_JOURNAL_DATA_FL)
+	if (ext4_test_inode_flag(inode, EXT4_INODE_JOURNAL_DATA))
 		return 0;
 	if (test_opt(inode->i_sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA)
 		return 1;
@@ -297,7 +297,7 @@ static inline int ext4_should_writeback_data(struct inode *inode)
 		return 0;
 	if (EXT4_JOURNAL(inode) == NULL)
 		return 1;
-	if (EXT4_I(inode)->i_flags & EXT4_JOURNAL_DATA_FL)
+	if (ext4_test_inode_flag(inode, EXT4_INODE_JOURNAL_DATA))
 		return 0;
 	if (test_opt(inode->i_sb, DATA_FLAGS) == EXT4_MOUNT_WRITEBACK_DATA)
 		return 1;
@@ -321,7 +321,7 @@ static inline int ext4_should_dioread_nolock(struct inode *inode)
 		return 0;
 	if (!S_ISREG(inode->i_mode))
 		return 0;
-	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 		return 0;
 	if (ext4_should_journal_data(inode))
 		return 0;
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 236b834b4ca8..377309c1af65 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -107,11 +107,8 @@ static int ext4_ext_truncate_extend_restart(handle_t *handle,
 	if (err <= 0)
 		return err;
 	err = ext4_truncate_restart_trans(handle, inode, needed);
-	/*
-	 * We have dropped i_data_sem so someone might have cached again
-	 * an extent we are going to truncate.
-	 */
-	ext4_ext_invalidate_cache(inode);
+	if (err == 0)
+		err = -EAGAIN;
 
 	return err;
 }
@@ -185,10 +182,10 @@ static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
 	if (flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) {
 		/*
 		 * If there are at least EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME
-		 * block groups per flexgroup, reserve the first block 
-		 * group for directories and special files.  Regular 
+		 * block groups per flexgroup, reserve the first block
+		 * group for directories and special files.  Regular
 		 * files will start at the second block group.  This
-		 * tends to speed up directory access and improves 
+		 * tends to speed up directory access and improves
 		 * fsck times.
 		 */
 		block_group &= ~(flex_size-1);
@@ -439,10 +436,10 @@ static int __ext4_ext_check(const char *function, struct inode *inode,
 	return 0;
 
 corrupted:
-	__ext4_error(inode->i_sb, function,
-			"bad header/extent in inode #%lu: %s - magic %x, "
+	ext4_error_inode(function, inode,
+			"bad header/extent: %s - magic %x, "
 			"entries %u, max %u(%u), depth %u(%u)",
-			inode->i_ino, error_msg, le16_to_cpu(eh->eh_magic),
+			error_msg, le16_to_cpu(eh->eh_magic),
 			le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max),
 			max, le16_to_cpu(eh->eh_depth), depth);
 
@@ -1622,9 +1619,7 @@ int ext4_ext_try_to_merge(struct inode *inode,
 		merge_done = 1;
 		WARN_ON(eh->eh_entries == 0);
 		if (!eh->eh_entries)
-			ext4_error(inode->i_sb,
-				   "inode#%lu, eh->eh_entries = 0!",
-				   inode->i_ino);
+			EXT4_ERROR_INODE(inode, "eh->eh_entries = 0!");
 	}
 
 	return merge_done;
@@ -2039,7 +2034,7 @@ ext4_ext_in_cache(struct inode *inode, ext4_lblk_t block,
 	struct ext4_ext_cache *cex;
 	int ret = EXT4_EXT_CACHE_NO;
 
-	/* 
+	/*
 	 * We borrow i_block_reservation_lock to protect i_cached_extent
 	 */
 	spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
@@ -2361,7 +2356,7 @@ static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
 	int depth = ext_depth(inode);
 	struct ext4_ext_path *path;
 	handle_t *handle;
-	int i = 0, err = 0;
+	int i, err;
 
 	ext_debug("truncate since %u\n", start);
 
@@ -2370,23 +2365,26 @@ static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
 	if (IS_ERR(handle))
 		return PTR_ERR(handle);
 
+again:
 	ext4_ext_invalidate_cache(inode);
 
 	/*
 	 * We start scanning from right side, freeing all the blocks
 	 * after i_size and walking into the tree depth-wise.
 	 */
+	depth = ext_depth(inode);
 	path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1), GFP_NOFS);
 	if (path == NULL) {
 		ext4_journal_stop(handle);
 		return -ENOMEM;
 	}
+	path[0].p_depth = depth;
 	path[0].p_hdr = ext_inode_hdr(inode);
 	if (ext4_ext_check(inode, path[0].p_hdr, depth)) {
 		err = -EIO;
 		goto out;
 	}
-	path[0].p_depth = depth;
+	i = err = 0;
 
 	while (i >= 0 && err == 0) {
 		if (i == depth) {
@@ -2480,6 +2478,8 @@ static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
 out:
 	ext4_ext_drop_refs(path);
 	kfree(path);
+	if (err == -EAGAIN)
+		goto again;
 	ext4_journal_stop(handle);
 
 	return err;
@@ -2544,7 +2544,7 @@ static void bi_complete(struct bio *bio, int error)
 /* FIXME!! we need to try to merge to left or right after zero-out  */
 static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
 {
-	int ret = -EIO;
+	int ret;
 	struct bio *bio;
 	int blkbits, blocksize;
 	sector_t ee_pblock;
@@ -2568,6 +2568,9 @@ static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
 			len = ee_len;
 
 		bio = bio_alloc(GFP_NOIO, len);
+		if (!bio)
+			return -ENOMEM;
+
 		bio->bi_sector = ee_pblock;
 		bio->bi_bdev   = inode->i_sb->s_bdev;
 
@@ -2595,22 +2598,20 @@ static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
 		submit_bio(WRITE, bio);
 		wait_for_completion(&event);
 
-		if (test_bit(BIO_UPTODATE, &bio->bi_flags))
-			ret = 0;
-		else {
-			ret = -EIO;
-			break;
+		if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) {
+			bio_put(bio);
+			return -EIO;
 		}
 		bio_put(bio);
 		ee_len    -= done;
 		ee_pblock += done  << (blkbits - 9);
 	}
-	return ret;
+	return 0;
 }
 
 #define EXT4_EXT_ZERO_LEN 7
 /*
- * This function is called by ext4_ext_get_blocks() if someone tries to write
+ * This function is called by ext4_ext_map_blocks() if someone tries to write
  * to an uninitialized extent. It may result in splitting the uninitialized
  * extent into multiple extents (upto three - one initialized and two
  * uninitialized).
@@ -2620,39 +2621,55 @@ static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
  *   c> Splits in three extents: Somone is writing in middle of the extent
  */
 static int ext4_ext_convert_to_initialized(handle_t *handle,
-						struct inode *inode,
-						struct ext4_ext_path *path,
-						ext4_lblk_t iblock,
-						unsigned int max_blocks)
+					   struct inode *inode,
+					   struct ext4_map_blocks *map,
+					   struct ext4_ext_path *path)
 {
 	struct ext4_extent *ex, newex, orig_ex;
 	struct ext4_extent *ex1 = NULL;
 	struct ext4_extent *ex2 = NULL;
 	struct ext4_extent *ex3 = NULL;
 	struct ext4_extent_header *eh;
-	ext4_lblk_t ee_block;
+	ext4_lblk_t ee_block, eof_block;
 	unsigned int allocated, ee_len, depth;
 	ext4_fsblk_t newblock;
 	int err = 0;
 	int ret = 0;
+	int may_zeroout;
+
+	ext_debug("ext4_ext_convert_to_initialized: inode %lu, logical"
+		"block %llu, max_blocks %u\n", inode->i_ino,
+		(unsigned long long)map->m_lblk, map->m_len);
+
+	eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
+		inode->i_sb->s_blocksize_bits;
+	if (eof_block < map->m_lblk + map->m_len)
+		eof_block = map->m_lblk + map->m_len;
 
 	depth = ext_depth(inode);
 	eh = path[depth].p_hdr;
 	ex = path[depth].p_ext;
 	ee_block = le32_to_cpu(ex->ee_block);
 	ee_len = ext4_ext_get_actual_len(ex);
-	allocated = ee_len - (iblock - ee_block);
-	newblock = iblock - ee_block + ext_pblock(ex);
+	allocated = ee_len - (map->m_lblk - ee_block);
+	newblock = map->m_lblk - ee_block + ext_pblock(ex);
+
 	ex2 = ex;
 	orig_ex.ee_block = ex->ee_block;
 	orig_ex.ee_len   = cpu_to_le16(ee_len);
 	ext4_ext_store_pblock(&orig_ex, ext_pblock(ex));
 
+	/*
+	 * It is safe to convert extent to initialized via explicit
+	 * zeroout only if extent is fully insde i_size or new_size.
+	 */
+	may_zeroout = ee_block + ee_len <= eof_block;
+
 	err = ext4_ext_get_access(handle, inode, path + depth);
 	if (err)
 		goto out;
 	/* If extent has less than 2*EXT4_EXT_ZERO_LEN zerout directly */
-	if (ee_len <= 2*EXT4_EXT_ZERO_LEN) {
+	if (ee_len <= 2*EXT4_EXT_ZERO_LEN && may_zeroout) {
 		err =  ext4_ext_zeroout(inode, &orig_ex);
 		if (err)
 			goto fix_extent_len;
@@ -2665,10 +2682,10 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 		return allocated;
 	}
 
-	/* ex1: ee_block to iblock - 1 : uninitialized */
-	if (iblock > ee_block) {
+	/* ex1: ee_block to map->m_lblk - 1 : uninitialized */
+	if (map->m_lblk > ee_block) {
 		ex1 = ex;
-		ex1->ee_len = cpu_to_le16(iblock - ee_block);
+		ex1->ee_len = cpu_to_le16(map->m_lblk - ee_block);
 		ext4_ext_mark_uninitialized(ex1);
 		ex2 = &newex;
 	}
@@ -2677,15 +2694,15 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 	 * we insert ex3, if ex1 is NULL. This is to avoid temporary
 	 * overlap of blocks.
 	 */
-	if (!ex1 && allocated > max_blocks)
-		ex2->ee_len = cpu_to_le16(max_blocks);
+	if (!ex1 && allocated > map->m_len)
+		ex2->ee_len = cpu_to_le16(map->m_len);
 	/* ex3: to ee_block + ee_len : uninitialised */
-	if (allocated > max_blocks) {
+	if (allocated > map->m_len) {
 		unsigned int newdepth;
 		/* If extent has less than EXT4_EXT_ZERO_LEN zerout directly */
-		if (allocated <= EXT4_EXT_ZERO_LEN) {
+		if (allocated <= EXT4_EXT_ZERO_LEN && may_zeroout) {
 			/*
-			 * iblock == ee_block is handled by the zerouout
+			 * map->m_lblk == ee_block is handled by the zerouout
 			 * at the beginning.
 			 * Mark first half uninitialized.
 			 * Mark second half initialized and zero out the
@@ -2698,7 +2715,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 			ext4_ext_dirty(handle, inode, path + depth);
 
 			ex3 = &newex;
-			ex3->ee_block = cpu_to_le32(iblock);
+			ex3->ee_block = cpu_to_le32(map->m_lblk);
 			ext4_ext_store_pblock(ex3, newblock);
 			ex3->ee_len = cpu_to_le16(allocated);
 			err = ext4_ext_insert_extent(handle, inode, path,
@@ -2711,7 +2728,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 				ex->ee_len   = orig_ex.ee_len;
 				ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
 				ext4_ext_dirty(handle, inode, path + depth);
-				/* blocks available from iblock */
+				/* blocks available from map->m_lblk */
 				return allocated;
 
 			} else if (err)
@@ -2733,8 +2750,8 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 				 */
 				depth = ext_depth(inode);
 				ext4_ext_drop_refs(path);
-				path = ext4_ext_find_extent(inode,
-								iblock, path);
+				path = ext4_ext_find_extent(inode, map->m_lblk,
+							    path);
 				if (IS_ERR(path)) {
 					err = PTR_ERR(path);
 					return err;
@@ -2754,12 +2771,12 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 			return allocated;
 		}
 		ex3 = &newex;
-		ex3->ee_block = cpu_to_le32(iblock + max_blocks);
-		ext4_ext_store_pblock(ex3, newblock + max_blocks);
-		ex3->ee_len = cpu_to_le16(allocated - max_blocks);
+		ex3->ee_block = cpu_to_le32(map->m_lblk + map->m_len);
+		ext4_ext_store_pblock(ex3, newblock + map->m_len);
+		ex3->ee_len = cpu_to_le16(allocated - map->m_len);
 		ext4_ext_mark_uninitialized(ex3);
 		err = ext4_ext_insert_extent(handle, inode, path, ex3, 0);
-		if (err == -ENOSPC) {
+		if (err == -ENOSPC && may_zeroout) {
 			err =  ext4_ext_zeroout(inode, &orig_ex);
 			if (err)
 				goto fix_extent_len;
@@ -2769,7 +2786,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 			ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
 			ext4_ext_dirty(handle, inode, path + depth);
 			/* zeroed the full extent */
-			/* blocks available from iblock */
+			/* blocks available from map->m_lblk */
 			return allocated;
 
 		} else if (err)
@@ -2783,11 +2800,13 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 		 * update the extent length after successful insert of the
 		 * split extent
 		 */
-		orig_ex.ee_len = cpu_to_le16(ee_len -
-						ext4_ext_get_actual_len(ex3));
+		ee_len -= ext4_ext_get_actual_len(ex3);
+		orig_ex.ee_len = cpu_to_le16(ee_len);
+		may_zeroout = ee_block + ee_len <= eof_block;
+
 		depth = newdepth;
 		ext4_ext_drop_refs(path);
-		path = ext4_ext_find_extent(inode, iblock, path);
+		path = ext4_ext_find_extent(inode, map->m_lblk, path);
 		if (IS_ERR(path)) {
 			err = PTR_ERR(path);
 			goto out;
@@ -2801,14 +2820,14 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 		if (err)
 			goto out;
 
-		allocated = max_blocks;
+		allocated = map->m_len;
 
 		/* If extent has less than EXT4_EXT_ZERO_LEN and we are trying
 		 * to insert a extent in the middle zerout directly
 		 * otherwise give the extent a chance to merge to left
 		 */
 		if (le16_to_cpu(orig_ex.ee_len) <= EXT4_EXT_ZERO_LEN &&
-							iblock != ee_block) {
+			map->m_lblk != ee_block && may_zeroout) {
 			err =  ext4_ext_zeroout(inode, &orig_ex);
 			if (err)
 				goto fix_extent_len;
@@ -2818,7 +2837,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 			ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
 			ext4_ext_dirty(handle, inode, path + depth);
 			/* zero out the first half */
-			/* blocks available from iblock */
+			/* blocks available from map->m_lblk */
 			return allocated;
 		}
 	}
@@ -2829,12 +2848,12 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 	 */
 	if (ex1 && ex1 != ex) {
 		ex1 = ex;
-		ex1->ee_len = cpu_to_le16(iblock - ee_block);
+		ex1->ee_len = cpu_to_le16(map->m_lblk - ee_block);
 		ext4_ext_mark_uninitialized(ex1);
 		ex2 = &newex;
 	}
-	/* ex2: iblock to iblock + maxblocks-1 : initialised */
-	ex2->ee_block = cpu_to_le32(iblock);
+	/* ex2: map->m_lblk to map->m_lblk + maxblocks-1 : initialised */
+	ex2->ee_block = cpu_to_le32(map->m_lblk);
 	ext4_ext_store_pblock(ex2, newblock);
 	ex2->ee_len = cpu_to_le16(allocated);
 	if (ex2 != ex)
@@ -2877,7 +2896,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 	goto out;
 insert:
 	err = ext4_ext_insert_extent(handle, inode, path, &newex, 0);
-	if (err == -ENOSPC) {
+	if (err == -ENOSPC && may_zeroout) {
 		err =  ext4_ext_zeroout(inode, &orig_ex);
 		if (err)
 			goto fix_extent_len;
@@ -2904,7 +2923,7 @@ fix_extent_len:
 }
 
 /*
- * This function is called by ext4_ext_get_blocks() from
+ * This function is called by ext4_ext_map_blocks() from
  * ext4_get_blocks_dio_write() when DIO to write
  * to an uninitialized extent.
  *
@@ -2927,9 +2946,8 @@ fix_extent_len:
  */
 static int ext4_split_unwritten_extents(handle_t *handle,
 					struct inode *inode,
+					struct ext4_map_blocks *map,
 					struct ext4_ext_path *path,
-					ext4_lblk_t iblock,
-					unsigned int max_blocks,
 					int flags)
 {
 	struct ext4_extent *ex, newex, orig_ex;
@@ -2937,41 +2955,55 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 	struct ext4_extent *ex2 = NULL;
 	struct ext4_extent *ex3 = NULL;
 	struct ext4_extent_header *eh;
-	ext4_lblk_t ee_block;
+	ext4_lblk_t ee_block, eof_block;
 	unsigned int allocated, ee_len, depth;
 	ext4_fsblk_t newblock;
 	int err = 0;
+	int may_zeroout;
+
+	ext_debug("ext4_split_unwritten_extents: inode %lu, logical"
+		"block %llu, max_blocks %u\n", inode->i_ino,
+		(unsigned long long)map->m_lblk, map->m_len);
+
+	eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
+		inode->i_sb->s_blocksize_bits;
+	if (eof_block < map->m_lblk + map->m_len)
+		eof_block = map->m_lblk + map->m_len;
 
-	ext_debug("ext4_split_unwritten_extents: inode %lu,"
-		  "iblock %llu, max_blocks %u\n", inode->i_ino,
-		  (unsigned long long)iblock, max_blocks);
 	depth = ext_depth(inode);
 	eh = path[depth].p_hdr;
 	ex = path[depth].p_ext;
 	ee_block = le32_to_cpu(ex->ee_block);
 	ee_len = ext4_ext_get_actual_len(ex);
-	allocated = ee_len - (iblock - ee_block);
-	newblock = iblock - ee_block + ext_pblock(ex);
+	allocated = ee_len - (map->m_lblk - ee_block);
+	newblock = map->m_lblk - ee_block + ext_pblock(ex);
+
 	ex2 = ex;
 	orig_ex.ee_block = ex->ee_block;
 	orig_ex.ee_len   = cpu_to_le16(ee_len);
 	ext4_ext_store_pblock(&orig_ex, ext_pblock(ex));
 
 	/*
+	 * It is safe to convert extent to initialized via explicit
+	 * zeroout only if extent is fully insde i_size or new_size.
+	 */
+	may_zeroout = ee_block + ee_len <= eof_block;
+
+	/*
  	 * If the uninitialized extent begins at the same logical
  	 * block where the write begins, and the write completely
  	 * covers the extent, then we don't need to split it.
  	 */
-	if ((iblock == ee_block) && (allocated <= max_blocks))
+	if ((map->m_lblk == ee_block) && (allocated <= map->m_len))
 		return allocated;
 
 	err = ext4_ext_get_access(handle, inode, path + depth);
 	if (err)
 		goto out;
-	/* ex1: ee_block to iblock - 1 : uninitialized */
-	if (iblock > ee_block) {
+	/* ex1: ee_block to map->m_lblk - 1 : uninitialized */
+	if (map->m_lblk > ee_block) {
 		ex1 = ex;
-		ex1->ee_len = cpu_to_le16(iblock - ee_block);
+		ex1->ee_len = cpu_to_le16(map->m_lblk - ee_block);
 		ext4_ext_mark_uninitialized(ex1);
 		ex2 = &newex;
 	}
@@ -2980,18 +3012,18 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 	 * we insert ex3, if ex1 is NULL. This is to avoid temporary
 	 * overlap of blocks.
 	 */
-	if (!ex1 && allocated > max_blocks)
-		ex2->ee_len = cpu_to_le16(max_blocks);
+	if (!ex1 && allocated > map->m_len)
+		ex2->ee_len = cpu_to_le16(map->m_len);
 	/* ex3: to ee_block + ee_len : uninitialised */
-	if (allocated > max_blocks) {
+	if (allocated > map->m_len) {
 		unsigned int newdepth;
 		ex3 = &newex;
-		ex3->ee_block = cpu_to_le32(iblock + max_blocks);
-		ext4_ext_store_pblock(ex3, newblock + max_blocks);
-		ex3->ee_len = cpu_to_le16(allocated - max_blocks);
+		ex3->ee_block = cpu_to_le32(map->m_lblk + map->m_len);
+		ext4_ext_store_pblock(ex3, newblock + map->m_len);
+		ex3->ee_len = cpu_to_le16(allocated - map->m_len);
 		ext4_ext_mark_uninitialized(ex3);
 		err = ext4_ext_insert_extent(handle, inode, path, ex3, flags);
-		if (err == -ENOSPC) {
+		if (err == -ENOSPC && may_zeroout) {
 			err =  ext4_ext_zeroout(inode, &orig_ex);
 			if (err)
 				goto fix_extent_len;
@@ -3001,7 +3033,7 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 			ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
 			ext4_ext_dirty(handle, inode, path + depth);
 			/* zeroed the full extent */
-			/* blocks available from iblock */
+			/* blocks available from map->m_lblk */
 			return allocated;
 
 		} else if (err)
@@ -3015,11 +3047,13 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 		 * update the extent length after successful insert of the
 		 * split extent
 		 */
-		orig_ex.ee_len = cpu_to_le16(ee_len -
-						ext4_ext_get_actual_len(ex3));
+		ee_len -= ext4_ext_get_actual_len(ex3);
+		orig_ex.ee_len = cpu_to_le16(ee_len);
+		may_zeroout = ee_block + ee_len <= eof_block;
+
 		depth = newdepth;
 		ext4_ext_drop_refs(path);
-		path = ext4_ext_find_extent(inode, iblock, path);
+		path = ext4_ext_find_extent(inode, map->m_lblk, path);
 		if (IS_ERR(path)) {
 			err = PTR_ERR(path);
 			goto out;
@@ -3033,7 +3067,7 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 		if (err)
 			goto out;
 
-		allocated = max_blocks;
+		allocated = map->m_len;
 	}
 	/*
 	 * If there was a change of depth as part of the
@@ -3042,15 +3076,15 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 	 */
 	if (ex1 && ex1 != ex) {
 		ex1 = ex;
-		ex1->ee_len = cpu_to_le16(iblock - ee_block);
+		ex1->ee_len = cpu_to_le16(map->m_lblk - ee_block);
 		ext4_ext_mark_uninitialized(ex1);
 		ex2 = &newex;
 	}
 	/*
-	 * ex2: iblock to iblock + maxblocks-1 : to be direct IO written,
-	 * uninitialised still.
+	 * ex2: map->m_lblk to map->m_lblk + map->m_len-1 : to be written
+	 * using direct I/O, uninitialised still.
 	 */
-	ex2->ee_block = cpu_to_le32(iblock);
+	ex2->ee_block = cpu_to_le32(map->m_lblk);
 	ext4_ext_store_pblock(ex2, newblock);
 	ex2->ee_len = cpu_to_le16(allocated);
 	ext4_ext_mark_uninitialized(ex2);
@@ -3062,7 +3096,7 @@ static int ext4_split_unwritten_extents(handle_t *handle,
 	goto out;
 insert:
 	err = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
-	if (err == -ENOSPC) {
+	if (err == -ENOSPC && may_zeroout) {
 		err =  ext4_ext_zeroout(inode, &orig_ex);
 		if (err)
 			goto fix_extent_len;
@@ -3152,10 +3186,9 @@ static void unmap_underlying_metadata_blocks(struct block_device *bdev,
 
 static int
 ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
-			ext4_lblk_t iblock, unsigned int max_blocks,
+			struct ext4_map_blocks *map,
 			struct ext4_ext_path *path, int flags,
-			unsigned int allocated, struct buffer_head *bh_result,
-			ext4_fsblk_t newblock)
+			unsigned int allocated, ext4_fsblk_t newblock)
 {
 	int ret = 0;
 	int err = 0;
@@ -3163,15 +3196,14 @@ ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
 
 	ext_debug("ext4_ext_handle_uninitialized_extents: inode %lu, logical"
 		  "block %llu, max_blocks %u, flags %d, allocated %u",
-		  inode->i_ino, (unsigned long long)iblock, max_blocks,
+		  inode->i_ino, (unsigned long long)map->m_lblk, map->m_len,
 		  flags, allocated);
 	ext4_ext_show_leaf(inode, path);
 
 	/* get_block() before submit the IO, split the extent */
 	if ((flags & EXT4_GET_BLOCKS_PRE_IO)) {
-		ret = ext4_split_unwritten_extents(handle,
-						inode, path, iblock,
-						max_blocks, flags);
+		ret = ext4_split_unwritten_extents(handle, inode, map,
+						   path, flags);
 		/*
 		 * Flag the inode(non aio case) or end_io struct (aio case)
 		 * that this IO needs to convertion to written when IO is
@@ -3182,7 +3214,7 @@ ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
 		else
 			ext4_set_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
 		if (ext4_should_dioread_nolock(inode))
-			set_buffer_uninit(bh_result);
+			map->m_flags |= EXT4_MAP_UNINIT;
 		goto out;
 	}
 	/* IO end_io complete, convert the filled extent to written */
@@ -3210,14 +3242,12 @@ ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
 		 * the buffer head will be unmapped so that
 		 * a read from the block returns 0s.
 		 */
-		set_buffer_unwritten(bh_result);
+		map->m_flags |= EXT4_MAP_UNWRITTEN;
 		goto out1;
 	}
 
 	/* buffered write, writepage time, convert*/
-	ret = ext4_ext_convert_to_initialized(handle, inode,
-						path, iblock,
-						max_blocks);
+	ret = ext4_ext_convert_to_initialized(handle, inode, map, path);
 	if (ret >= 0)
 		ext4_update_inode_fsync_trans(handle, inode, 1);
 out:
@@ -3226,7 +3256,7 @@ out:
 		goto out2;
 	} else
 		allocated = ret;
-	set_buffer_new(bh_result);
+	map->m_flags |= EXT4_MAP_NEW;
 	/*
 	 * if we allocated more blocks than requested
 	 * we need to make sure we unmap the extra block
@@ -3234,11 +3264,11 @@ out:
 	 * unmapped later when we find the buffer_head marked
 	 * new.
 	 */
-	if (allocated > max_blocks) {
+	if (allocated > map->m_len) {
 		unmap_underlying_metadata_blocks(inode->i_sb->s_bdev,
-					newblock + max_blocks,
-					allocated - max_blocks);
-		allocated = max_blocks;
+					newblock + map->m_len,
+					allocated - map->m_len);
+		allocated = map->m_len;
 	}
 
 	/*
@@ -3252,13 +3282,13 @@ out:
 		ext4_da_update_reserve_space(inode, allocated, 0);
 
 map_out:
-	set_buffer_mapped(bh_result);
+	map->m_flags |= EXT4_MAP_MAPPED;
 out1:
-	if (allocated > max_blocks)
-		allocated = max_blocks;
+	if (allocated > map->m_len)
+		allocated = map->m_len;
 	ext4_ext_show_leaf(inode, path);
-	bh_result->b_bdev = inode->i_sb->s_bdev;
-	bh_result->b_blocknr = newblock;
+	map->m_pblk = newblock;
+	map->m_len = allocated;
 out2:
 	if (path) {
 		ext4_ext_drop_refs(path);
@@ -3284,26 +3314,23 @@ out2:
  *
  * return < 0, error case.
  */
-int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
-			ext4_lblk_t iblock,
-			unsigned int max_blocks, struct buffer_head *bh_result,
-			int flags)
+int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
+			struct ext4_map_blocks *map, int flags)
 {
 	struct ext4_ext_path *path = NULL;
 	struct ext4_extent_header *eh;
 	struct ext4_extent newex, *ex, *last_ex;
 	ext4_fsblk_t newblock;
-	int err = 0, depth, ret, cache_type;
+	int i, err = 0, depth, ret, cache_type;
 	unsigned int allocated = 0;
 	struct ext4_allocation_request ar;
 	ext4_io_end_t *io = EXT4_I(inode)->cur_aio_dio;
 
-	__clear_bit(BH_New, &bh_result->b_state);
 	ext_debug("blocks %u/%u requested for inode %lu\n",
-			iblock, max_blocks, inode->i_ino);
+		  map->m_lblk, map->m_len, inode->i_ino);
 
 	/* check in cache */
-	cache_type = ext4_ext_in_cache(inode, iblock, &newex);
+	cache_type = ext4_ext_in_cache(inode, map->m_lblk, &newex);
 	if (cache_type) {
 		if (cache_type == EXT4_EXT_CACHE_GAP) {
 			if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
@@ -3316,12 +3343,12 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 			/* we should allocate requested block */
 		} else if (cache_type == EXT4_EXT_CACHE_EXTENT) {
 			/* block is already allocated */
-			newblock = iblock
+			newblock = map->m_lblk
 				   - le32_to_cpu(newex.ee_block)
 				   + ext_pblock(&newex);
 			/* number of remaining blocks in the extent */
 			allocated = ext4_ext_get_actual_len(&newex) -
-					(iblock - le32_to_cpu(newex.ee_block));
+				(map->m_lblk - le32_to_cpu(newex.ee_block));
 			goto out;
 		} else {
 			BUG();
@@ -3329,7 +3356,7 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	}
 
 	/* find extent for this block */
-	path = ext4_ext_find_extent(inode, iblock, NULL);
+	path = ext4_ext_find_extent(inode, map->m_lblk, NULL);
 	if (IS_ERR(path)) {
 		err = PTR_ERR(path);
 		path = NULL;
@@ -3345,8 +3372,9 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	 */
 	if (unlikely(path[depth].p_ext == NULL && depth != 0)) {
 		EXT4_ERROR_INODE(inode, "bad extent address "
-				 "iblock: %d, depth: %d pblock %lld",
-				 iblock, depth, path[depth].p_block);
+				 "lblock: %lu, depth: %d pblock %lld",
+				 (unsigned long) map->m_lblk, depth,
+				 path[depth].p_block);
 		err = -EIO;
 		goto out2;
 	}
@@ -3364,12 +3392,12 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 		 */
 		ee_len = ext4_ext_get_actual_len(ex);
 		/* if found extent covers block, simply return it */
-		if (in_range(iblock, ee_block, ee_len)) {
-			newblock = iblock - ee_block + ee_start;
+		if (in_range(map->m_lblk, ee_block, ee_len)) {
+			newblock = map->m_lblk - ee_block + ee_start;
 			/* number of remaining blocks in the extent */
-			allocated = ee_len - (iblock - ee_block);
-			ext_debug("%u fit into %u:%d -> %llu\n", iblock,
-					ee_block, ee_len, newblock);
+			allocated = ee_len - (map->m_lblk - ee_block);
+			ext_debug("%u fit into %u:%d -> %llu\n", map->m_lblk,
+				  ee_block, ee_len, newblock);
 
 			/* Do not put uninitialized extent in the cache */
 			if (!ext4_ext_is_uninitialized(ex)) {
@@ -3379,8 +3407,8 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 				goto out;
 			}
 			ret = ext4_ext_handle_uninitialized_extents(handle,
-					inode, iblock, max_blocks, path,
-					flags, allocated, bh_result, newblock);
+					inode, map, path, flags, allocated,
+					newblock);
 			return ret;
 		}
 	}
@@ -3394,7 +3422,7 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 		 * put just found gap into cache to speed up
 		 * subsequent requests
 		 */
-		ext4_ext_put_gap_in_cache(inode, path, iblock);
+		ext4_ext_put_gap_in_cache(inode, path, map->m_lblk);
 		goto out2;
 	}
 	/*
@@ -3402,11 +3430,11 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	 */
 
 	/* find neighbour allocated blocks */
-	ar.lleft = iblock;
+	ar.lleft = map->m_lblk;
 	err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft);
 	if (err)
 		goto out2;
-	ar.lright = iblock;
+	ar.lright = map->m_lblk;
 	err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright);
 	if (err)
 		goto out2;
@@ -3417,26 +3445,26 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	 * EXT_INIT_MAX_LEN and for an uninitialized extent this limit is
 	 * EXT_UNINIT_MAX_LEN.
 	 */
-	if (max_blocks > EXT_INIT_MAX_LEN &&
+	if (map->m_len > EXT_INIT_MAX_LEN &&
 	    !(flags & EXT4_GET_BLOCKS_UNINIT_EXT))
-		max_blocks = EXT_INIT_MAX_LEN;
-	else if (max_blocks > EXT_UNINIT_MAX_LEN &&
+		map->m_len = EXT_INIT_MAX_LEN;
+	else if (map->m_len > EXT_UNINIT_MAX_LEN &&
 		 (flags & EXT4_GET_BLOCKS_UNINIT_EXT))
-		max_blocks = EXT_UNINIT_MAX_LEN;
+		map->m_len = EXT_UNINIT_MAX_LEN;
 
-	/* Check if we can really insert (iblock)::(iblock+max_blocks) extent */
-	newex.ee_block = cpu_to_le32(iblock);
-	newex.ee_len = cpu_to_le16(max_blocks);
+	/* Check if we can really insert (m_lblk)::(m_lblk + m_len) extent */
+	newex.ee_block = cpu_to_le32(map->m_lblk);
+	newex.ee_len = cpu_to_le16(map->m_len);
 	err = ext4_ext_check_overlap(inode, &newex, path);
 	if (err)
 		allocated = ext4_ext_get_actual_len(&newex);
 	else
-		allocated = max_blocks;
+		allocated = map->m_len;
 
 	/* allocate new block */
 	ar.inode = inode;
-	ar.goal = ext4_ext_find_goal(inode, path, iblock);
-	ar.logical = iblock;
+	ar.goal = ext4_ext_find_goal(inode, path, map->m_lblk);
+	ar.logical = map->m_lblk;
 	ar.len = allocated;
 	if (S_ISREG(inode->i_mode))
 		ar.flags = EXT4_MB_HINT_DATA;
@@ -3470,21 +3498,33 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 						     EXT4_STATE_DIO_UNWRITTEN);
 		}
 		if (ext4_should_dioread_nolock(inode))
-			set_buffer_uninit(bh_result);
+			map->m_flags |= EXT4_MAP_UNINIT;
 	}
 
-	if (unlikely(EXT4_I(inode)->i_flags & EXT4_EOFBLOCKS_FL)) {
+	if (unlikely(ext4_test_inode_flag(inode, EXT4_INODE_EOFBLOCKS))) {
 		if (unlikely(!eh->eh_entries)) {
 			EXT4_ERROR_INODE(inode,
-					 "eh->eh_entries == 0 ee_block %d",
-					 ex->ee_block);
+					 "eh->eh_entries == 0 and "
+					 "EOFBLOCKS_FL set");
 			err = -EIO;
 			goto out2;
 		}
 		last_ex = EXT_LAST_EXTENT(eh);
-		if (iblock + ar.len > le32_to_cpu(last_ex->ee_block)
-		    + ext4_ext_get_actual_len(last_ex))
-			EXT4_I(inode)->i_flags &= ~EXT4_EOFBLOCKS_FL;
+		/*
+		 * If the current leaf block was reached by looking at
+		 * the last index block all the way down the tree, and
+		 * we are extending the inode beyond the last extent
+		 * in the current leaf block, then clear the
+		 * EOFBLOCKS_FL flag.
+		 */
+		for (i = depth-1; i >= 0; i--) {
+			if (path[i].p_idx != EXT_LAST_INDEX(path[i].p_hdr))
+				break;
+		}
+		if ((i < 0) &&
+		    (map->m_lblk + ar.len > le32_to_cpu(last_ex->ee_block) +
+		     ext4_ext_get_actual_len(last_ex)))
+			ext4_clear_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
 	}
 	err = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
 	if (err) {
@@ -3500,9 +3540,9 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	/* previous routine could use block we allocated */
 	newblock = ext_pblock(&newex);
 	allocated = ext4_ext_get_actual_len(&newex);
-	if (allocated > max_blocks)
-		allocated = max_blocks;
-	set_buffer_new(bh_result);
+	if (allocated > map->m_len)
+		allocated = map->m_len;
+	map->m_flags |= EXT4_MAP_NEW;
 
 	/*
 	 * Update reserved blocks/metadata blocks after successful
@@ -3516,18 +3556,18 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	 * when it is _not_ an uninitialized extent.
 	 */
 	if ((flags & EXT4_GET_BLOCKS_UNINIT_EXT) == 0) {
-		ext4_ext_put_in_cache(inode, iblock, allocated, newblock,
+		ext4_ext_put_in_cache(inode, map->m_lblk, allocated, newblock,
 						EXT4_EXT_CACHE_EXTENT);
 		ext4_update_inode_fsync_trans(handle, inode, 1);
 	} else
 		ext4_update_inode_fsync_trans(handle, inode, 0);
 out:
-	if (allocated > max_blocks)
-		allocated = max_blocks;
+	if (allocated > map->m_len)
+		allocated = map->m_len;
 	ext4_ext_show_leaf(inode, path);
-	set_buffer_mapped(bh_result);
-	bh_result->b_bdev = inode->i_sb->s_bdev;
-	bh_result->b_blocknr = newblock;
+	map->m_flags |= EXT4_MAP_MAPPED;
+	map->m_pblk = newblock;
+	map->m_len = allocated;
 out2:
 	if (path) {
 		ext4_ext_drop_refs(path);
@@ -3625,7 +3665,7 @@ static void ext4_falloc_update_inode(struct inode *inode,
 		 * can proceed even if the new size is the same as i_size.
 		 */
 		if (new_size > i_size_read(inode))
-			EXT4_I(inode)->i_flags |= EXT4_EOFBLOCKS_FL;
+			ext4_set_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
 	}
 
 }
@@ -3640,55 +3680,57 @@ static void ext4_falloc_update_inode(struct inode *inode,
 long ext4_fallocate(struct inode *inode, int mode, loff_t offset, loff_t len)
 {
 	handle_t *handle;
-	ext4_lblk_t block;
 	loff_t new_size;
 	unsigned int max_blocks;
 	int ret = 0;
 	int ret2 = 0;
 	int retries = 0;
-	struct buffer_head map_bh;
+	struct ext4_map_blocks map;
 	unsigned int credits, blkbits = inode->i_blkbits;
 
 	/*
 	 * currently supporting (pre)allocate mode for extent-based
 	 * files _only_
 	 */
-	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 		return -EOPNOTSUPP;
 
 	/* preallocation to directories is currently not supported */
 	if (S_ISDIR(inode->i_mode))
 		return -ENODEV;
 
-	block = offset >> blkbits;
+	map.m_lblk = offset >> blkbits;
 	/*
 	 * We can't just convert len to max_blocks because
 	 * If blocksize = 4096 offset = 3072 and len = 2048
 	 */
 	max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
-							- block;
+		- map.m_lblk;
 	/*
 	 * credits to insert 1 extent into extent tree
 	 */
 	credits = ext4_chunk_trans_blocks(inode, max_blocks);
 	mutex_lock(&inode->i_mutex);
+	ret = inode_newsize_ok(inode, (len + offset));
+	if (ret) {
+		mutex_unlock(&inode->i_mutex);
+		return ret;
+	}
 retry:
 	while (ret >= 0 && ret < max_blocks) {
-		block = block + ret;
-		max_blocks = max_blocks - ret;
+		map.m_lblk = map.m_lblk + ret;
+		map.m_len = max_blocks = max_blocks - ret;
 		handle = ext4_journal_start(inode, credits);
 		if (IS_ERR(handle)) {
 			ret = PTR_ERR(handle);
 			break;
 		}
-		map_bh.b_state = 0;
-		ret = ext4_get_blocks(handle, inode, block,
-				      max_blocks, &map_bh,
+		ret = ext4_map_blocks(handle, inode, &map,
 				      EXT4_GET_BLOCKS_CREATE_UNINIT_EXT);
 		if (ret <= 0) {
 #ifdef EXT4FS_DEBUG
 			WARN_ON(ret <= 0);
-			printk(KERN_ERR "%s: ext4_ext_get_blocks "
+			printk(KERN_ERR "%s: ext4_ext_map_blocks "
 				    "returned error inode#%lu, block=%u, "
 				    "max_blocks=%u", __func__,
 				    inode->i_ino, block, max_blocks);
@@ -3697,14 +3739,14 @@ retry:
 			ret2 = ext4_journal_stop(handle);
 			break;
 		}
-		if ((block + ret) >= (EXT4_BLOCK_ALIGN(offset + len,
+		if ((map.m_lblk + ret) >= (EXT4_BLOCK_ALIGN(offset + len,
 						blkbits) >> blkbits))
 			new_size = offset + len;
 		else
-			new_size = (block + ret) << blkbits;
+			new_size = (map.m_lblk + ret) << blkbits;
 
 		ext4_falloc_update_inode(inode, mode, new_size,
-						buffer_new(&map_bh));
+					 (map.m_flags & EXT4_MAP_NEW));
 		ext4_mark_inode_dirty(handle, inode);
 		ret2 = ext4_journal_stop(handle);
 		if (ret2)
@@ -3733,42 +3775,39 @@ int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset,
 				    ssize_t len)
 {
 	handle_t *handle;
-	ext4_lblk_t block;
 	unsigned int max_blocks;
 	int ret = 0;
 	int ret2 = 0;
-	struct buffer_head map_bh;
+	struct ext4_map_blocks map;
 	unsigned int credits, blkbits = inode->i_blkbits;
 
-	block = offset >> blkbits;
+	map.m_lblk = offset >> blkbits;
 	/*
 	 * We can't just convert len to max_blocks because
 	 * If blocksize = 4096 offset = 3072 and len = 2048
 	 */
-	max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
-							- block;
+	max_blocks = ((EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits) -
+		      map.m_lblk);
 	/*
 	 * credits to insert 1 extent into extent tree
 	 */
 	credits = ext4_chunk_trans_blocks(inode, max_blocks);
 	while (ret >= 0 && ret < max_blocks) {
-		block = block + ret;
-		max_blocks = max_blocks - ret;
+		map.m_lblk += ret;
+		map.m_len = (max_blocks -= ret);
 		handle = ext4_journal_start(inode, credits);
 		if (IS_ERR(handle)) {
 			ret = PTR_ERR(handle);
 			break;
 		}
-		map_bh.b_state = 0;
-		ret = ext4_get_blocks(handle, inode, block,
-				      max_blocks, &map_bh,
+		ret = ext4_map_blocks(handle, inode, &map,
 				      EXT4_GET_BLOCKS_IO_CONVERT_EXT);
 		if (ret <= 0) {
 			WARN_ON(ret <= 0);
-			printk(KERN_ERR "%s: ext4_ext_get_blocks "
+			printk(KERN_ERR "%s: ext4_ext_map_blocks "
 				    "returned error inode#%lu, block=%u, "
 				    "max_blocks=%u", __func__,
-				    inode->i_ino, block, max_blocks);
+				    inode->i_ino, map.m_lblk, map.m_len);
 		}
 		ext4_mark_inode_dirty(handle, inode);
 		ret2 = ext4_journal_stop(handle);
@@ -3898,7 +3937,7 @@ int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	int error = 0;
 
 	/* fallback to generic here if not in extents fmt */
-	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 		return generic_block_fiemap(inode, fieinfo, start, len,
 			ext4_get_block);
 
diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index d0776e410f34..5313ae4cda2d 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -66,7 +66,7 @@ ext4_file_write(struct kiocb *iocb, const struct iovec *iov,
 	 * is smaller than s_maxbytes, which is for extent-mapped files.
 	 */
 
-	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL)) {
+	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
 		struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 		size_t length = iov_length(iov, nr_segs);
 
diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c
index ef3d980e67cb..592adf2e546e 100644
--- a/fs/ext4/fsync.c
+++ b/fs/ext4/fsync.c
@@ -35,6 +35,29 @@
 #include <trace/events/ext4.h>
 
 /*
+ * If we're not journaling and this is a just-created file, we have to
+ * sync our parent directory (if it was freshly created) since
+ * otherwise it will only be written by writeback, leaving a huge
+ * window during which a crash may lose the file.  This may apply for
+ * the parent directory's parent as well, and so on recursively, if
+ * they are also freshly created.
+ */
+static void ext4_sync_parent(struct inode *inode)
+{
+	struct dentry *dentry = NULL;
+
+	while (inode && ext4_test_inode_state(inode, EXT4_STATE_NEWENTRY)) {
+		ext4_clear_inode_state(inode, EXT4_STATE_NEWENTRY);
+		dentry = list_entry(inode->i_dentry.next,
+				    struct dentry, d_alias);
+		if (!dentry || !dentry->d_parent || !dentry->d_parent->d_inode)
+			break;
+		inode = dentry->d_parent->d_inode;
+		sync_mapping_buffers(inode->i_mapping);
+	}
+}
+
+/*
  * akpm: A new design for ext4_sync_file().
  *
  * This is only called from sys_fsync(), sys_fdatasync() and sys_msync().
@@ -48,9 +71,9 @@
  * i_mutex lock is held when entering and exiting this function
  */
 
-int ext4_sync_file(struct file *file, struct dentry *dentry, int datasync)
+int ext4_sync_file(struct file *file, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	struct ext4_inode_info *ei = EXT4_I(inode);
 	journal_t *journal = EXT4_SB(inode->i_sb)->s_journal;
 	int ret;
@@ -58,7 +81,7 @@ int ext4_sync_file(struct file *file, struct dentry *dentry, int datasync)
 
 	J_ASSERT(ext4_journal_current_handle() == NULL);
 
-	trace_ext4_sync_file(file, dentry, datasync);
+	trace_ext4_sync_file(file, datasync);
 
 	if (inode->i_sb->s_flags & MS_RDONLY)
 		return 0;
@@ -66,9 +89,13 @@ int ext4_sync_file(struct file *file, struct dentry *dentry, int datasync)
 	ret = flush_completed_IO(inode);
 	if (ret < 0)
 		return ret;
-	
-	if (!journal)
-		return simple_fsync(file, dentry, datasync);
+
+	if (!journal) {
+		ret = generic_file_fsync(file, datasync);
+		if (!ret && !list_empty(&inode->i_dentry))
+			ext4_sync_parent(inode);
+		return ret;
+	}
 
 	/*
 	 * data=writeback,ordered:
@@ -102,7 +129,7 @@ int ext4_sync_file(struct file *file, struct dentry *dentry, int datasync)
 		    (journal->j_flags & JBD2_BARRIER))
 			blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL,
 					NULL, BLKDEV_IFL_WAIT);
-		jbd2_log_wait_commit(journal, commit_tid);
+		ret = jbd2_log_wait_commit(journal, commit_tid);
 	} else if (journal->j_flags & JBD2_BARRIER)
 		blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL,
 			BLKDEV_IFL_WAIT);
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index 1a0e183a2f04..25c4b3173fd9 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -240,56 +240,49 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
 	if (fatal)
 		goto error_return;
 
-	/* Ok, now we can actually update the inode bitmaps.. */
-	cleared = ext4_clear_bit_atomic(ext4_group_lock_ptr(sb, block_group),
-					bit, bitmap_bh->b_data);
-	if (!cleared)
-		ext4_error(sb, "bit already cleared for inode %lu", ino);
-	else {
-		gdp = ext4_get_group_desc(sb, block_group, &bh2);
-
+	fatal = -ESRCH;
+	gdp = ext4_get_group_desc(sb, block_group, &bh2);
+	if (gdp) {
 		BUFFER_TRACE(bh2, "get_write_access");
 		fatal = ext4_journal_get_write_access(handle, bh2);
-		if (fatal) goto error_return;
-
-		if (gdp) {
-			ext4_lock_group(sb, block_group);
-			count = ext4_free_inodes_count(sb, gdp) + 1;
-			ext4_free_inodes_set(sb, gdp, count);
-			if (is_directory) {
-				count = ext4_used_dirs_count(sb, gdp) - 1;
-				ext4_used_dirs_set(sb, gdp, count);
-				if (sbi->s_log_groups_per_flex) {
-					ext4_group_t f;
-
-					f = ext4_flex_group(sbi, block_group);
-					atomic_dec(&sbi->s_flex_groups[f].used_dirs);
-				}
+	}
+	ext4_lock_group(sb, block_group);
+	cleared = ext4_clear_bit(bit, bitmap_bh->b_data);
+	if (fatal || !cleared) {
+		ext4_unlock_group(sb, block_group);
+		goto out;
+	}
 
-			}
-			gdp->bg_checksum = ext4_group_desc_csum(sbi,
-							block_group, gdp);
-			ext4_unlock_group(sb, block_group);
-			percpu_counter_inc(&sbi->s_freeinodes_counter);
-			if (is_directory)
-				percpu_counter_dec(&sbi->s_dirs_counter);
-
-			if (sbi->s_log_groups_per_flex) {
-				ext4_group_t f;
-
-				f = ext4_flex_group(sbi, block_group);
-				atomic_inc(&sbi->s_flex_groups[f].free_inodes);
-			}
-		}
-		BUFFER_TRACE(bh2, "call ext4_handle_dirty_metadata");
-		err = ext4_handle_dirty_metadata(handle, NULL, bh2);
-		if (!fatal) fatal = err;
+	count = ext4_free_inodes_count(sb, gdp) + 1;
+	ext4_free_inodes_set(sb, gdp, count);
+	if (is_directory) {
+		count = ext4_used_dirs_count(sb, gdp) - 1;
+		ext4_used_dirs_set(sb, gdp, count);
+		percpu_counter_dec(&sbi->s_dirs_counter);
 	}
-	BUFFER_TRACE(bitmap_bh, "call ext4_handle_dirty_metadata");
-	err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
-	if (!fatal)
-		fatal = err;
-	sb->s_dirt = 1;
+	gdp->bg_checksum = ext4_group_desc_csum(sbi, block_group, gdp);
+	ext4_unlock_group(sb, block_group);
+
+	percpu_counter_inc(&sbi->s_freeinodes_counter);
+	if (sbi->s_log_groups_per_flex) {
+		ext4_group_t f = ext4_flex_group(sbi, block_group);
+
+		atomic_inc(&sbi->s_flex_groups[f].free_inodes);
+		if (is_directory)
+			atomic_dec(&sbi->s_flex_groups[f].used_dirs);
+	}
+	BUFFER_TRACE(bh2, "call ext4_handle_dirty_metadata");
+	fatal = ext4_handle_dirty_metadata(handle, NULL, bh2);
+out:
+	if (cleared) {
+		BUFFER_TRACE(bitmap_bh, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
+		if (!fatal)
+			fatal = err;
+		sb->s_dirt = 1;
+	} else
+		ext4_error(sb, "bit already cleared for inode %lu", ino);
+
 error_return:
 	brelse(bitmap_bh);
 	ext4_std_error(sb, fatal);
@@ -499,7 +492,7 @@ static int find_group_orlov(struct super_block *sb, struct inode *parent,
 
 	if (S_ISDIR(mode) &&
 	    ((parent == sb->s_root->d_inode) ||
-	     (EXT4_I(parent)->i_flags & EXT4_TOPDIR_FL))) {
+	     (ext4_test_inode_flag(parent, EXT4_INODE_TOPDIR)))) {
 		int best_ndir = inodes_per_group;
 		int ret = -1;
 
@@ -1041,7 +1034,7 @@ got:
 	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
 		/* set extent flag only for directory, file and normal symlink*/
 		if (S_ISDIR(mode) || S_ISREG(mode) || S_ISLNK(mode)) {
-			EXT4_I(inode)->i_flags |= EXT4_EXTENTS_FL;
+			ext4_set_inode_flag(inode, EXT4_INODE_EXTENTS);
 			ext4_ext_tree_init(handle, inode);
 		}
 	}
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 3e0f6af9d08d..19df61c321fd 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -149,7 +149,7 @@ int ext4_truncate_restart_trans(handle_t *handle, struct inode *inode,
 	int ret;
 
 	/*
-	 * Drop i_data_sem to avoid deadlock with ext4_get_blocks At this
+	 * Drop i_data_sem to avoid deadlock with ext4_map_blocks.  At this
 	 * moment, get_block can be called only for blocks inside i_size since
 	 * page cache has been already dropped and writes are blocked by
 	 * i_mutex. So we can safely drop the i_data_sem here.
@@ -348,9 +348,8 @@ static int __ext4_check_blockref(const char *function, struct inode *inode,
 		if (blk &&
 		    unlikely(!ext4_data_block_valid(EXT4_SB(inode->i_sb),
 						    blk, 1))) {
-			__ext4_error(inode->i_sb, function,
-				   "invalid block reference %u "
-				   "in inode #%lu", blk, inode->i_ino);
+			ext4_error_inode(function, inode,
+					 "invalid block reference %u", blk);
 			return -EIO;
 		}
 	}
@@ -785,7 +784,7 @@ failed:
 	/* Allocation failed, free what we already allocated */
 	ext4_free_blocks(handle, inode, 0, new_blocks[0], 1, 0);
 	for (i = 1; i <= n ; i++) {
-		/* 
+		/*
 		 * branch[i].bh is newly allocated, so there is no
 		 * need to revoke the block, which is why we don't
 		 * need to set EXT4_FREE_BLOCKS_METADATA.
@@ -875,7 +874,7 @@ static int ext4_splice_branch(handle_t *handle, struct inode *inode,
 
 err_out:
 	for (i = 1; i <= num; i++) {
-		/* 
+		/*
 		 * branch[i].bh is newly allocated, so there is no
 		 * need to revoke the block, which is why we don't
 		 * need to set EXT4_FREE_BLOCKS_METADATA.
@@ -890,9 +889,9 @@ err_out:
 }
 
 /*
- * The ext4_ind_get_blocks() function handles non-extents inodes
+ * The ext4_ind_map_blocks() function handles non-extents inodes
  * (i.e., using the traditional indirect/double-indirect i_blocks
- * scheme) for ext4_get_blocks().
+ * scheme) for ext4_map_blocks().
  *
  * Allocation strategy is simple: if we have to allocate something, we will
  * have to go the whole way to leaf. So let's do it before attaching anything
@@ -917,9 +916,8 @@ err_out:
  * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system
  * blocks.
  */
-static int ext4_ind_get_blocks(handle_t *handle, struct inode *inode,
-			       ext4_lblk_t iblock, unsigned int maxblocks,
-			       struct buffer_head *bh_result,
+static int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
+			       struct ext4_map_blocks *map,
 			       int flags)
 {
 	int err = -EIO;
@@ -933,9 +931,9 @@ static int ext4_ind_get_blocks(handle_t *handle, struct inode *inode,
 	int count = 0;
 	ext4_fsblk_t first_block = 0;
 
-	J_ASSERT(!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL));
+	J_ASSERT(!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)));
 	J_ASSERT(handle != NULL || (flags & EXT4_GET_BLOCKS_CREATE) == 0);
-	depth = ext4_block_to_path(inode, iblock, offsets,
+	depth = ext4_block_to_path(inode, map->m_lblk, offsets,
 				   &blocks_to_boundary);
 
 	if (depth == 0)
@@ -946,10 +944,9 @@ static int ext4_ind_get_blocks(handle_t *handle, struct inode *inode,
 	/* Simplest case - block found, no allocation needed */
 	if (!partial) {
 		first_block = le32_to_cpu(chain[depth - 1].key);
-		clear_buffer_new(bh_result);
 		count++;
 		/*map more blocks*/
-		while (count < maxblocks && count <= blocks_to_boundary) {
+		while (count < map->m_len && count <= blocks_to_boundary) {
 			ext4_fsblk_t blk;
 
 			blk = le32_to_cpu(*(chain[depth-1].p + count));
@@ -969,7 +966,7 @@ static int ext4_ind_get_blocks(handle_t *handle, struct inode *inode,
 	/*
 	 * Okay, we need to do block allocation.
 	*/
-	goal = ext4_find_goal(inode, iblock, partial);
+	goal = ext4_find_goal(inode, map->m_lblk, partial);
 
 	/* the number of blocks need to allocate for [d,t]indirect blocks */
 	indirect_blks = (chain + depth) - partial - 1;
@@ -979,11 +976,11 @@ static int ext4_ind_get_blocks(handle_t *handle, struct inode *inode,
 	 * direct blocks to allocate for this branch.
 	 */
 	count = ext4_blks_to_allocate(partial, indirect_blks,
-					maxblocks, blocks_to_boundary);
+				      map->m_len, blocks_to_boundary);
 	/*
 	 * Block out ext4_truncate while we alter the tree
 	 */
-	err = ext4_alloc_branch(handle, inode, iblock, indirect_blks,
+	err = ext4_alloc_branch(handle, inode, map->m_lblk, indirect_blks,
 				&count, goal,
 				offsets + (partial - chain), partial);
 
@@ -995,18 +992,20 @@ static int ext4_ind_get_blocks(handle_t *handle, struct inode *inode,
 	 * may need to return -EAGAIN upwards in the worst case.  --sct
 	 */
 	if (!err)
-		err = ext4_splice_branch(handle, inode, iblock,
+		err = ext4_splice_branch(handle, inode, map->m_lblk,
 					 partial, indirect_blks, count);
 	if (err)
 		goto cleanup;
 
-	set_buffer_new(bh_result);
+	map->m_flags |= EXT4_MAP_NEW;
 
 	ext4_update_inode_fsync_trans(handle, inode, 1);
 got_it:
-	map_bh(bh_result, inode->i_sb, le32_to_cpu(chain[depth-1].key));
+	map->m_flags |= EXT4_MAP_MAPPED;
+	map->m_pblk = le32_to_cpu(chain[depth-1].key);
+	map->m_len = count;
 	if (count > blocks_to_boundary)
-		set_buffer_boundary(bh_result);
+		map->m_flags |= EXT4_MAP_BOUNDARY;
 	err = count;
 	/* Clean up and exit */
 	partial = chain + depth - 1;	/* the whole chain */
@@ -1016,7 +1015,6 @@ cleanup:
 		brelse(partial->bh);
 		partial--;
 	}
-	BUFFER_TRACE(bh_result, "returned");
 out:
 	return err;
 }
@@ -1061,7 +1059,7 @@ static int ext4_indirect_calc_metadata_amount(struct inode *inode,
  */
 static int ext4_calc_metadata_amount(struct inode *inode, sector_t lblock)
 {
-	if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL)
+	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
 		return ext4_ext_calc_metadata_amount(inode, lblock);
 
 	return ext4_indirect_calc_metadata_amount(inode, lblock);
@@ -1076,7 +1074,6 @@ void ext4_da_update_reserve_space(struct inode *inode,
 {
 	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 	struct ext4_inode_info *ei = EXT4_I(inode);
-	int mdb_free = 0, allocated_meta_blocks = 0;
 
 	spin_lock(&ei->i_block_reservation_lock);
 	trace_ext4_da_update_reserve_space(inode, used);
@@ -1091,11 +1088,10 @@ void ext4_da_update_reserve_space(struct inode *inode,
 
 	/* Update per-inode reservations */
 	ei->i_reserved_data_blocks -= used;
-	used += ei->i_allocated_meta_blocks;
 	ei->i_reserved_meta_blocks -= ei->i_allocated_meta_blocks;
-	allocated_meta_blocks = ei->i_allocated_meta_blocks;
+	percpu_counter_sub(&sbi->s_dirtyblocks_counter,
+			   used + ei->i_allocated_meta_blocks);
 	ei->i_allocated_meta_blocks = 0;
-	percpu_counter_sub(&sbi->s_dirtyblocks_counter, used);
 
 	if (ei->i_reserved_data_blocks == 0) {
 		/*
@@ -1103,30 +1099,23 @@ void ext4_da_update_reserve_space(struct inode *inode,
 		 * only when we have written all of the delayed
 		 * allocation blocks.
 		 */
-		mdb_free = ei->i_reserved_meta_blocks;
+		percpu_counter_sub(&sbi->s_dirtyblocks_counter,
+				   ei->i_reserved_meta_blocks);
 		ei->i_reserved_meta_blocks = 0;
 		ei->i_da_metadata_calc_len = 0;
-		percpu_counter_sub(&sbi->s_dirtyblocks_counter, mdb_free);
 	}
 	spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
 
-	/* Update quota subsystem */
-	if (quota_claim) {
+	/* Update quota subsystem for data blocks */
+	if (quota_claim)
 		dquot_claim_block(inode, used);
-		if (mdb_free)
-			dquot_release_reservation_block(inode, mdb_free);
-	} else {
+	else {
 		/*
 		 * We did fallocate with an offset that is already delayed
 		 * allocated. So on delayed allocated writeback we should
-		 * not update the quota for allocated blocks. But then
-		 * converting an fallocate region to initialized region would
-		 * have caused a metadata allocation. So claim quota for
-		 * that
+		 * not re-claim the quota for fallocated blocks.
 		 */
-		if (allocated_meta_blocks)
-			dquot_claim_block(inode, allocated_meta_blocks);
-		dquot_release_reservation_block(inode, mdb_free + used);
+		dquot_release_reservation_block(inode, used);
 	}
 
 	/*
@@ -1139,15 +1128,15 @@ void ext4_da_update_reserve_space(struct inode *inode,
 		ext4_discard_preallocations(inode);
 }
 
-static int check_block_validity(struct inode *inode, const char *msg,
-				sector_t logical, sector_t phys, int len)
+static int check_block_validity(struct inode *inode, const char *func,
+				struct ext4_map_blocks *map)
 {
-	if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), phys, len)) {
-		__ext4_error(inode->i_sb, msg,
-			   "inode #%lu logical block %llu mapped to %llu "
-			   "(size %d)", inode->i_ino,
-			   (unsigned long long) logical,
-			   (unsigned long long) phys, len);
+	if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), map->m_pblk,
+				   map->m_len)) {
+		ext4_error_inode(func, inode,
+			   "lblock %lu mapped to illegal pblock %llu "
+			   "(length %d)", (unsigned long) map->m_lblk,
+				 map->m_pblk, map->m_len);
 		return -EIO;
 	}
 	return 0;
@@ -1212,15 +1201,15 @@ static pgoff_t ext4_num_dirty_pages(struct inode *inode, pgoff_t idx,
 }
 
 /*
- * The ext4_get_blocks() function tries to look up the requested blocks,
+ * The ext4_map_blocks() function tries to look up the requested blocks,
  * and returns if the blocks are already mapped.
  *
  * Otherwise it takes the write lock of the i_data_sem and allocate blocks
  * and store the allocated blocks in the result buffer head and mark it
  * mapped.
  *
- * If file type is extents based, it will call ext4_ext_get_blocks(),
- * Otherwise, call with ext4_ind_get_blocks() to handle indirect mapping
+ * If file type is extents based, it will call ext4_ext_map_blocks(),
+ * Otherwise, call with ext4_ind_map_blocks() to handle indirect mapping
  * based files
  *
  * On success, it returns the number of blocks being mapped or allocate.
@@ -1233,35 +1222,29 @@ static pgoff_t ext4_num_dirty_pages(struct inode *inode, pgoff_t idx,
  *
  * It returns the error in case of allocation failure.
  */
-int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
-		    unsigned int max_blocks, struct buffer_head *bh,
-		    int flags)
+int ext4_map_blocks(handle_t *handle, struct inode *inode,
+		    struct ext4_map_blocks *map, int flags)
 {
 	int retval;
 
-	clear_buffer_mapped(bh);
-	clear_buffer_unwritten(bh);
-
-	ext_debug("ext4_get_blocks(): inode %lu, flag %d, max_blocks %u,"
-		  "logical block %lu\n", inode->i_ino, flags, max_blocks,
-		  (unsigned long)block);
+	map->m_flags = 0;
+	ext_debug("ext4_map_blocks(): inode %lu, flag %d, max_blocks %u,"
+		  "logical block %lu\n", inode->i_ino, flags, map->m_len,
+		  (unsigned long) map->m_lblk);
 	/*
 	 * Try to see if we can get the block without requesting a new
 	 * file system block.
 	 */
 	down_read((&EXT4_I(inode)->i_data_sem));
-	if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL) {
-		retval =  ext4_ext_get_blocks(handle, inode, block, max_blocks,
-				bh, 0);
+	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
+		retval = ext4_ext_map_blocks(handle, inode, map, 0);
 	} else {
-		retval = ext4_ind_get_blocks(handle, inode, block, max_blocks,
-					     bh, 0);
+		retval = ext4_ind_map_blocks(handle, inode, map, 0);
 	}
 	up_read((&EXT4_I(inode)->i_data_sem));
 
-	if (retval > 0 && buffer_mapped(bh)) {
-		int ret = check_block_validity(inode, "file system corruption",
-					       block, bh->b_blocknr, retval);
+	if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) {
+		int ret = check_block_validity(inode, __func__, map);
 		if (ret != 0)
 			return ret;
 	}
@@ -1277,7 +1260,7 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 	 * ext4_ext_get_block() returns th create = 0
 	 * with buffer head unmapped.
 	 */
-	if (retval > 0 && buffer_mapped(bh))
+	if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED)
 		return retval;
 
 	/*
@@ -1290,7 +1273,7 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 	 * of BH_Unwritten and BH_Mapped flags being simultaneously
 	 * set on the buffer_head.
 	 */
-	clear_buffer_unwritten(bh);
+	map->m_flags &= ~EXT4_MAP_UNWRITTEN;
 
 	/*
 	 * New blocks allocate and/or writing to uninitialized extent
@@ -1312,14 +1295,12 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 	 * We need to check for EXT4 here because migrate
 	 * could have changed the inode type in between
 	 */
-	if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL) {
-		retval =  ext4_ext_get_blocks(handle, inode, block, max_blocks,
-					      bh, flags);
+	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
+		retval = ext4_ext_map_blocks(handle, inode, map, flags);
 	} else {
-		retval = ext4_ind_get_blocks(handle, inode, block,
-					     max_blocks, bh, flags);
+		retval = ext4_ind_map_blocks(handle, inode, map, flags);
 
-		if (retval > 0 && buffer_new(bh)) {
+		if (retval > 0 && map->m_flags & EXT4_MAP_NEW) {
 			/*
 			 * We allocated new blocks which will result in
 			 * i_data's format changing.  Force the migrate
@@ -1342,10 +1323,10 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 		EXT4_I(inode)->i_delalloc_reserved_flag = 0;
 
 	up_write((&EXT4_I(inode)->i_data_sem));
-	if (retval > 0 && buffer_mapped(bh)) {
-		int ret = check_block_validity(inode, "file system "
-					       "corruption after allocation",
-					       block, bh->b_blocknr, retval);
+	if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) {
+		int ret = check_block_validity(inode,
+					       "ext4_map_blocks_after_alloc",
+					       map);
 		if (ret != 0)
 			return ret;
 	}
@@ -1355,109 +1336,109 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 /* Maximum number of blocks we map for direct IO at once. */
 #define DIO_MAX_BLOCKS 4096
 
-int ext4_get_block(struct inode *inode, sector_t iblock,
-		   struct buffer_head *bh_result, int create)
+static int _ext4_get_block(struct inode *inode, sector_t iblock,
+			   struct buffer_head *bh, int flags)
 {
 	handle_t *handle = ext4_journal_current_handle();
+	struct ext4_map_blocks map;
 	int ret = 0, started = 0;
-	unsigned max_blocks = bh_result->b_size >> inode->i_blkbits;
 	int dio_credits;
 
-	if (create && !handle) {
+	map.m_lblk = iblock;
+	map.m_len = bh->b_size >> inode->i_blkbits;
+
+	if (flags && !handle) {
 		/* Direct IO write... */
-		if (max_blocks > DIO_MAX_BLOCKS)
-			max_blocks = DIO_MAX_BLOCKS;
-		dio_credits = ext4_chunk_trans_blocks(inode, max_blocks);
+		if (map.m_len > DIO_MAX_BLOCKS)
+			map.m_len = DIO_MAX_BLOCKS;
+		dio_credits = ext4_chunk_trans_blocks(inode, map.m_len);
 		handle = ext4_journal_start(inode, dio_credits);
 		if (IS_ERR(handle)) {
 			ret = PTR_ERR(handle);
-			goto out;
+			return ret;
 		}
 		started = 1;
 	}
 
-	ret = ext4_get_blocks(handle, inode, iblock, max_blocks, bh_result,
-			      create ? EXT4_GET_BLOCKS_CREATE : 0);
+	ret = ext4_map_blocks(handle, inode, &map, flags);
 	if (ret > 0) {
-		bh_result->b_size = (ret << inode->i_blkbits);
+		map_bh(bh, inode->i_sb, map.m_pblk);
+		bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | map.m_flags;
+		bh->b_size = inode->i_sb->s_blocksize * map.m_len;
 		ret = 0;
 	}
 	if (started)
 		ext4_journal_stop(handle);
-out:
 	return ret;
 }
 
+int ext4_get_block(struct inode *inode, sector_t iblock,
+		   struct buffer_head *bh, int create)
+{
+	return _ext4_get_block(inode, iblock, bh,
+			       create ? EXT4_GET_BLOCKS_CREATE : 0);
+}
+
 /*
  * `handle' can be NULL if create is zero
  */
 struct buffer_head *ext4_getblk(handle_t *handle, struct inode *inode,
 				ext4_lblk_t block, int create, int *errp)
 {
-	struct buffer_head dummy;
+	struct ext4_map_blocks map;
+	struct buffer_head *bh;
 	int fatal = 0, err;
-	int flags = 0;
 
 	J_ASSERT(handle != NULL || create == 0);
 
-	dummy.b_state = 0;
-	dummy.b_blocknr = -1000;
-	buffer_trace_init(&dummy.b_history);
-	if (create)
-		flags |= EXT4_GET_BLOCKS_CREATE;
-	err = ext4_get_blocks(handle, inode, block, 1, &dummy, flags);
-	/*
-	 * ext4_get_blocks() returns number of blocks mapped. 0 in
-	 * case of a HOLE.
-	 */
-	if (err > 0) {
-		if (err > 1)
-			WARN_ON(1);
-		err = 0;
+	map.m_lblk = block;
+	map.m_len = 1;
+	err = ext4_map_blocks(handle, inode, &map,
+			      create ? EXT4_GET_BLOCKS_CREATE : 0);
+
+	if (err < 0)
+		*errp = err;
+	if (err <= 0)
+		return NULL;
+	*errp = 0;
+
+	bh = sb_getblk(inode->i_sb, map.m_pblk);
+	if (!bh) {
+		*errp = -EIO;
+		return NULL;
 	}
-	*errp = err;
-	if (!err && buffer_mapped(&dummy)) {
-		struct buffer_head *bh;
-		bh = sb_getblk(inode->i_sb, dummy.b_blocknr);
-		if (!bh) {
-			*errp = -EIO;
-			goto err;
-		}
-		if (buffer_new(&dummy)) {
-			J_ASSERT(create != 0);
-			J_ASSERT(handle != NULL);
+	if (map.m_flags & EXT4_MAP_NEW) {
+		J_ASSERT(create != 0);
+		J_ASSERT(handle != NULL);
 
-			/*
-			 * Now that we do not always journal data, we should
-			 * keep in mind whether this should always journal the
-			 * new buffer as metadata.  For now, regular file
-			 * writes use ext4_get_block instead, so it's not a
-			 * problem.
-			 */
-			lock_buffer(bh);
-			BUFFER_TRACE(bh, "call get_create_access");
-			fatal = ext4_journal_get_create_access(handle, bh);
-			if (!fatal && !buffer_uptodate(bh)) {
-				memset(bh->b_data, 0, inode->i_sb->s_blocksize);
-				set_buffer_uptodate(bh);
-			}
-			unlock_buffer(bh);
-			BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
-			err = ext4_handle_dirty_metadata(handle, inode, bh);
-			if (!fatal)
-				fatal = err;
-		} else {
-			BUFFER_TRACE(bh, "not a new buffer");
-		}
-		if (fatal) {
-			*errp = fatal;
-			brelse(bh);
-			bh = NULL;
+		/*
+		 * Now that we do not always journal data, we should
+		 * keep in mind whether this should always journal the
+		 * new buffer as metadata.  For now, regular file
+		 * writes use ext4_get_block instead, so it's not a
+		 * problem.
+		 */
+		lock_buffer(bh);
+		BUFFER_TRACE(bh, "call get_create_access");
+		fatal = ext4_journal_get_create_access(handle, bh);
+		if (!fatal && !buffer_uptodate(bh)) {
+			memset(bh->b_data, 0, inode->i_sb->s_blocksize);
+			set_buffer_uptodate(bh);
 		}
-		return bh;
+		unlock_buffer(bh);
+		BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, inode, bh);
+		if (!fatal)
+			fatal = err;
+	} else {
+		BUFFER_TRACE(bh, "not a new buffer");
 	}
-err:
-	return NULL;
+	if (fatal) {
+		*errp = fatal;
+		brelse(bh);
+		bh = NULL;
+	}
+	return bh;
 }
 
 struct buffer_head *ext4_bread(handle_t *handle, struct inode *inode,
@@ -1860,7 +1841,7 @@ static int ext4_da_reserve_space(struct inode *inode, sector_t lblock)
 	int retries = 0;
 	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 	struct ext4_inode_info *ei = EXT4_I(inode);
-	unsigned long md_needed, md_reserved;
+	unsigned long md_needed;
 	int ret;
 
 	/*
@@ -1870,22 +1851,24 @@ static int ext4_da_reserve_space(struct inode *inode, sector_t lblock)
 	 */
 repeat:
 	spin_lock(&ei->i_block_reservation_lock);
-	md_reserved = ei->i_reserved_meta_blocks;
 	md_needed = ext4_calc_metadata_amount(inode, lblock);
 	trace_ext4_da_reserve_space(inode, md_needed);
 	spin_unlock(&ei->i_block_reservation_lock);
 
 	/*
-	 * Make quota reservation here to prevent quota overflow
-	 * later. Real quota accounting is done at pages writeout
-	 * time.
+	 * We will charge metadata quota at writeout time; this saves
+	 * us from metadata over-estimation, though we may go over by
+	 * a small amount in the end.  Here we just reserve for data.
 	 */
-	ret = dquot_reserve_block(inode, md_needed + 1);
+	ret = dquot_reserve_block(inode, 1);
 	if (ret)
 		return ret;
-
+	/*
+	 * We do still charge estimated metadata to the sb though;
+	 * we cannot afford to run out of free blocks.
+	 */
 	if (ext4_claim_free_blocks(sbi, md_needed + 1)) {
-		dquot_release_reservation_block(inode, md_needed + 1);
+		dquot_release_reservation_block(inode, 1);
 		if (ext4_should_retry_alloc(inode->i_sb, &retries)) {
 			yield();
 			goto repeat;
@@ -1910,6 +1893,7 @@ static void ext4_da_release_space(struct inode *inode, int to_free)
 
 	spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
 
+	trace_ext4_da_release_space(inode, to_free);
 	if (unlikely(to_free > ei->i_reserved_data_blocks)) {
 		/*
 		 * if there aren't enough reserved blocks, then the
@@ -1932,12 +1916,13 @@ static void ext4_da_release_space(struct inode *inode, int to_free)
 		 * only when we have written all of the delayed
 		 * allocation blocks.
 		 */
-		to_free += ei->i_reserved_meta_blocks;
+		percpu_counter_sub(&sbi->s_dirtyblocks_counter,
+				   ei->i_reserved_meta_blocks);
 		ei->i_reserved_meta_blocks = 0;
 		ei->i_da_metadata_calc_len = 0;
 	}
 
-	/* update fs dirty blocks counter */
+	/* update fs dirty data blocks counter */
 	percpu_counter_sub(&sbi->s_dirtyblocks_counter, to_free);
 
 	spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
@@ -2042,28 +2027,23 @@ static int mpage_da_submit_io(struct mpage_da_data *mpd)
 /*
  * mpage_put_bnr_to_bhs - walk blocks and assign them actual numbers
  *
- * @mpd->inode - inode to walk through
- * @exbh->b_blocknr - first block on a disk
- * @exbh->b_size - amount of space in bytes
- * @logical - first logical block to start assignment with
- *
  * the function goes through all passed space and put actual disk
  * block numbers into buffer heads, dropping BH_Delay and BH_Unwritten
  */
-static void mpage_put_bnr_to_bhs(struct mpage_da_data *mpd, sector_t logical,
-				 struct buffer_head *exbh)
+static void mpage_put_bnr_to_bhs(struct mpage_da_data *mpd,
+				 struct ext4_map_blocks *map)
 {
 	struct inode *inode = mpd->inode;
 	struct address_space *mapping = inode->i_mapping;
-	int blocks = exbh->b_size >> inode->i_blkbits;
-	sector_t pblock = exbh->b_blocknr, cur_logical;
+	int blocks = map->m_len;
+	sector_t pblock = map->m_pblk, cur_logical;
 	struct buffer_head *head, *bh;
 	pgoff_t index, end;
 	struct pagevec pvec;
 	int nr_pages, i;
 
-	index = logical >> (PAGE_CACHE_SHIFT - inode->i_blkbits);
-	end = (logical + blocks - 1) >> (PAGE_CACHE_SHIFT - inode->i_blkbits);
+	index = map->m_lblk >> (PAGE_CACHE_SHIFT - inode->i_blkbits);
+	end = (map->m_lblk + blocks - 1) >> (PAGE_CACHE_SHIFT - inode->i_blkbits);
 	cur_logical = index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
 
 	pagevec_init(&pvec, 0);
@@ -2090,17 +2070,16 @@ static void mpage_put_bnr_to_bhs(struct mpage_da_data *mpd, sector_t logical,
 
 			/* skip blocks out of the range */
 			do {
-				if (cur_logical >= logical)
+				if (cur_logical >= map->m_lblk)
 					break;
 				cur_logical++;
 			} while ((bh = bh->b_this_page) != head);
 
 			do {
-				if (cur_logical >= logical + blocks)
+				if (cur_logical >= map->m_lblk + blocks)
 					break;
 
-				if (buffer_delay(bh) ||
-						buffer_unwritten(bh)) {
+				if (buffer_delay(bh) || buffer_unwritten(bh)) {
 
 					BUG_ON(bh->b_bdev != inode->i_sb->s_bdev);
 
@@ -2119,7 +2098,7 @@ static void mpage_put_bnr_to_bhs(struct mpage_da_data *mpd, sector_t logical,
 				} else if (buffer_mapped(bh))
 					BUG_ON(bh->b_blocknr != pblock);
 
-				if (buffer_uninit(exbh))
+				if (map->m_flags & EXT4_MAP_UNINIT)
 					set_buffer_uninit(bh);
 				cur_logical++;
 				pblock++;
@@ -2130,21 +2109,6 @@ static void mpage_put_bnr_to_bhs(struct mpage_da_data *mpd, sector_t logical,
 }
 
 
-/*
- * __unmap_underlying_blocks - just a helper function to unmap
- * set of blocks described by @bh
- */
-static inline void __unmap_underlying_blocks(struct inode *inode,
-					     struct buffer_head *bh)
-{
-	struct block_device *bdev = inode->i_sb->s_bdev;
-	int blocks, i;
-
-	blocks = bh->b_size >> inode->i_blkbits;
-	for (i = 0; i < blocks; i++)
-		unmap_underlying_metadata(bdev, bh->b_blocknr + i);
-}
-
 static void ext4_da_block_invalidatepages(struct mpage_da_data *mpd,
 					sector_t logical, long blk_cnt)
 {
@@ -2206,7 +2170,7 @@ static void ext4_print_free_blocks(struct inode *inode)
 static int mpage_da_map_blocks(struct mpage_da_data *mpd)
 {
 	int err, blks, get_blocks_flags;
-	struct buffer_head new;
+	struct ext4_map_blocks map;
 	sector_t next = mpd->b_blocknr;
 	unsigned max_blocks = mpd->b_size >> mpd->inode->i_blkbits;
 	loff_t disksize = EXT4_I(mpd->inode)->i_disksize;
@@ -2247,15 +2211,15 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd)
 	 * EXT4_GET_BLOCKS_DELALLOC_RESERVE so the delalloc accounting
 	 * variables are updated after the blocks have been allocated.
 	 */
-	new.b_state = 0;
+	map.m_lblk = next;
+	map.m_len = max_blocks;
 	get_blocks_flags = EXT4_GET_BLOCKS_CREATE;
 	if (ext4_should_dioread_nolock(mpd->inode))
 		get_blocks_flags |= EXT4_GET_BLOCKS_IO_CREATE_EXT;
 	if (mpd->b_state & (1 << BH_Delay))
 		get_blocks_flags |= EXT4_GET_BLOCKS_DELALLOC_RESERVE;
 
-	blks = ext4_get_blocks(handle, mpd->inode, next, max_blocks,
-			       &new, get_blocks_flags);
+	blks = ext4_map_blocks(handle, mpd->inode, &map, get_blocks_flags);
 	if (blks < 0) {
 		err = blks;
 		/*
@@ -2282,7 +2246,7 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd)
 		ext4_msg(mpd->inode->i_sb, KERN_CRIT,
 			 "delayed block allocation failed for inode %lu at "
 			 "logical offset %llu with max blocks %zd with "
-			 "error %d\n", mpd->inode->i_ino,
+			 "error %d", mpd->inode->i_ino,
 			 (unsigned long long) next,
 			 mpd->b_size >> mpd->inode->i_blkbits, err);
 		printk(KERN_CRIT "This should not happen!!  "
@@ -2297,10 +2261,13 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd)
 	}
 	BUG_ON(blks == 0);
 
-	new.b_size = (blks << mpd->inode->i_blkbits);
+	if (map.m_flags & EXT4_MAP_NEW) {
+		struct block_device *bdev = mpd->inode->i_sb->s_bdev;
+		int i;
 
-	if (buffer_new(&new))
-		__unmap_underlying_blocks(mpd->inode, &new);
+		for (i = 0; i < map.m_len; i++)
+			unmap_underlying_metadata(bdev, map.m_pblk + i);
+	}
 
 	/*
 	 * If blocks are delayed marked, we need to
@@ -2308,7 +2275,7 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd)
 	 */
 	if ((mpd->b_state & (1 << BH_Delay)) ||
 	    (mpd->b_state & (1 << BH_Unwritten)))
-		mpage_put_bnr_to_bhs(mpd, next, &new);
+		mpage_put_bnr_to_bhs(mpd, &map);
 
 	if (ext4_should_order_data(mpd->inode)) {
 		err = ext4_jbd2_file_inode(handle, mpd->inode);
@@ -2349,8 +2316,17 @@ static void mpage_add_bh_to_extent(struct mpage_da_data *mpd,
 	sector_t next;
 	int nrblocks = mpd->b_size >> mpd->inode->i_blkbits;
 
+	/*
+	 * XXX Don't go larger than mballoc is willing to allocate
+	 * This is a stopgap solution.  We eventually need to fold
+	 * mpage_da_submit_io() into this function and then call
+	 * ext4_get_blocks() multiple times in a loop
+	 */
+	if (nrblocks >= 8*1024*1024/mpd->inode->i_sb->s_blocksize)
+		goto flush_it;
+
 	/* check if thereserved journal credits might overflow */
-	if (!(EXT4_I(mpd->inode)->i_flags & EXT4_EXTENTS_FL)) {
+	if (!(ext4_test_inode_flag(mpd->inode, EXT4_INODE_EXTENTS))) {
 		if (nrblocks >= EXT4_MAX_TRANS_DATA) {
 			/*
 			 * With non-extent format we are limited by the journal
@@ -2423,17 +2399,6 @@ static int __mpage_da_writepage(struct page *page,
 	struct buffer_head *bh, *head;
 	sector_t logical;
 
-	if (mpd->io_done) {
-		/*
-		 * Rest of the page in the page_vec
-		 * redirty then and skip then. We will
-		 * try to write them again after
-		 * starting a new transaction
-		 */
-		redirty_page_for_writepage(wbc, page);
-		unlock_page(page);
-		return MPAGE_DA_EXTENT_TAIL;
-	}
 	/*
 	 * Can we merge this page to current extent?
 	 */
@@ -2528,8 +2493,9 @@ static int __mpage_da_writepage(struct page *page,
  * initialized properly.
  */
 static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
-				  struct buffer_head *bh_result, int create)
+				  struct buffer_head *bh, int create)
 {
+	struct ext4_map_blocks map;
 	int ret = 0;
 	sector_t invalid_block = ~((sector_t) 0xffff);
 
@@ -2537,16 +2503,22 @@ static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
 		invalid_block = ~0;
 
 	BUG_ON(create == 0);
-	BUG_ON(bh_result->b_size != inode->i_sb->s_blocksize);
+	BUG_ON(bh->b_size != inode->i_sb->s_blocksize);
+
+	map.m_lblk = iblock;
+	map.m_len = 1;
 
 	/*
 	 * first, we need to know whether the block is allocated already
 	 * preallocated blocks are unmapped but should treated
 	 * the same as allocated blocks.
 	 */
-	ret = ext4_get_blocks(NULL, inode, iblock, 1,  bh_result, 0);
-	if ((ret == 0) && !buffer_delay(bh_result)) {
-		/* the block isn't (pre)allocated yet, let's reserve space */
+	ret = ext4_map_blocks(NULL, inode, &map, 0);
+	if (ret < 0)
+		return ret;
+	if (ret == 0) {
+		if (buffer_delay(bh))
+			return 0; /* Not sure this could or should happen */
 		/*
 		 * XXX: __block_prepare_write() unmaps passed block,
 		 * is it OK?
@@ -2556,26 +2528,26 @@ static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
 			/* not enough space to reserve */
 			return ret;
 
-		map_bh(bh_result, inode->i_sb, invalid_block);
-		set_buffer_new(bh_result);
-		set_buffer_delay(bh_result);
-	} else if (ret > 0) {
-		bh_result->b_size = (ret << inode->i_blkbits);
-		if (buffer_unwritten(bh_result)) {
-			/* A delayed write to unwritten bh should
-			 * be marked new and mapped.  Mapped ensures
-			 * that we don't do get_block multiple times
-			 * when we write to the same offset and new
-			 * ensures that we do proper zero out for
-			 * partial write.
-			 */
-			set_buffer_new(bh_result);
-			set_buffer_mapped(bh_result);
-		}
-		ret = 0;
+		map_bh(bh, inode->i_sb, invalid_block);
+		set_buffer_new(bh);
+		set_buffer_delay(bh);
+		return 0;
 	}
 
-	return ret;
+	map_bh(bh, inode->i_sb, map.m_pblk);
+	bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | map.m_flags;
+
+	if (buffer_unwritten(bh)) {
+		/* A delayed write to unwritten bh should be marked
+		 * new and mapped.  Mapped ensures that we don't do
+		 * get_block multiple times when we write to the same
+		 * offset and new ensures that we do proper zero out
+		 * for partial write.
+		 */
+		set_buffer_new(bh);
+		set_buffer_mapped(bh);
+	}
+	return 0;
 }
 
 /*
@@ -2597,21 +2569,8 @@ static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
 static int noalloc_get_block_write(struct inode *inode, sector_t iblock,
 				   struct buffer_head *bh_result, int create)
 {
-	int ret = 0;
-	unsigned max_blocks = bh_result->b_size >> inode->i_blkbits;
-
 	BUG_ON(bh_result->b_size != inode->i_sb->s_blocksize);
-
-	/*
-	 * we don't want to do block allocation in writepage
-	 * so call get_block_wrap with create = 0
-	 */
-	ret = ext4_get_blocks(NULL, inode, iblock, max_blocks, bh_result, 0);
-	if (ret > 0) {
-		bh_result->b_size = (ret << inode->i_blkbits);
-		ret = 0;
-	}
-	return ret;
+	return _ext4_get_block(inode, iblock, bh_result, 0);
 }
 
 static int bget_one(handle_t *handle, struct buffer_head *bh)
@@ -2821,13 +2780,131 @@ static int ext4_da_writepages_trans_blocks(struct inode *inode)
 	 * number of contiguous block. So we will limit
 	 * number of contiguous block to a sane value
 	 */
-	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL) &&
+	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) &&
 	    (max_blocks > EXT4_MAX_TRANS_DATA))
 		max_blocks = EXT4_MAX_TRANS_DATA;
 
 	return ext4_chunk_trans_blocks(inode, max_blocks);
 }
 
+/*
+ * write_cache_pages_da - walk the list of dirty pages of the given
+ * address space and call the callback function (which usually writes
+ * the pages).
+ *
+ * This is a forked version of write_cache_pages().  Differences:
+ *	Range cyclic is ignored.
+ *	no_nrwrite_index_update is always presumed true
+ */
+static int write_cache_pages_da(struct address_space *mapping,
+				struct writeback_control *wbc,
+				struct mpage_da_data *mpd)
+{
+	int ret = 0;
+	int done = 0;
+	struct pagevec pvec;
+	int nr_pages;
+	pgoff_t index;
+	pgoff_t end;		/* Inclusive */
+	long nr_to_write = wbc->nr_to_write;
+
+	pagevec_init(&pvec, 0);
+	index = wbc->range_start >> PAGE_CACHE_SHIFT;
+	end = wbc->range_end >> PAGE_CACHE_SHIFT;
+
+	while (!done && (index <= end)) {
+		int i;
+
+		nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
+			      PAGECACHE_TAG_DIRTY,
+			      min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
+		if (nr_pages == 0)
+			break;
+
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+
+			/*
+			 * At this point, the page may be truncated or
+			 * invalidated (changing page->mapping to NULL), or
+			 * even swizzled back from swapper_space to tmpfs file
+			 * mapping. However, page->index will not change
+			 * because we have a reference on the page.
+			 */
+			if (page->index > end) {
+				done = 1;
+				break;
+			}
+
+			lock_page(page);
+
+			/*
+			 * Page truncated or invalidated. We can freely skip it
+			 * then, even for data integrity operations: the page
+			 * has disappeared concurrently, so there could be no
+			 * real expectation of this data interity operation
+			 * even if there is now a new, dirty page at the same
+			 * pagecache address.
+			 */
+			if (unlikely(page->mapping != mapping)) {
+continue_unlock:
+				unlock_page(page);
+				continue;
+			}
+
+			if (!PageDirty(page)) {
+				/* someone wrote it for us */
+				goto continue_unlock;
+			}
+
+			if (PageWriteback(page)) {
+				if (wbc->sync_mode != WB_SYNC_NONE)
+					wait_on_page_writeback(page);
+				else
+					goto continue_unlock;
+			}
+
+			BUG_ON(PageWriteback(page));
+			if (!clear_page_dirty_for_io(page))
+				goto continue_unlock;
+
+			ret = __mpage_da_writepage(page, wbc, mpd);
+			if (unlikely(ret)) {
+				if (ret == AOP_WRITEPAGE_ACTIVATE) {
+					unlock_page(page);
+					ret = 0;
+				} else {
+					done = 1;
+					break;
+				}
+			}
+
+			if (nr_to_write > 0) {
+				nr_to_write--;
+				if (nr_to_write == 0 &&
+				    wbc->sync_mode == WB_SYNC_NONE) {
+					/*
+					 * We stop writing back only if we are
+					 * not doing integrity sync. In case of
+					 * integrity sync we have to keep going
+					 * because someone may be concurrently
+					 * dirtying pages, and we might have
+					 * synced a lot of newly appeared dirty
+					 * pages, but have not synced all of the
+					 * old dirty pages.
+					 */
+					done = 1;
+					break;
+				}
+			}
+		}
+		pagevec_release(&pvec);
+		cond_resched();
+	}
+	return ret;
+}
+
+
 static int ext4_da_writepages(struct address_space *mapping,
 			      struct writeback_control *wbc)
 {
@@ -2836,7 +2913,6 @@ static int ext4_da_writepages(struct address_space *mapping,
 	handle_t *handle = NULL;
 	struct mpage_da_data mpd;
 	struct inode *inode = mapping->host;
-	int no_nrwrite_index_update;
 	int pages_written = 0;
 	long pages_skipped;
 	unsigned int max_pages;
@@ -2916,12 +2992,6 @@ static int ext4_da_writepages(struct address_space *mapping,
 	mpd.wbc = wbc;
 	mpd.inode = mapping->host;
 
-	/*
-	 * we don't want write_cache_pages to update
-	 * nr_to_write and writeback_index
-	 */
-	no_nrwrite_index_update = wbc->no_nrwrite_index_update;
-	wbc->no_nrwrite_index_update = 1;
 	pages_skipped = wbc->pages_skipped;
 
 retry:
@@ -2941,7 +3011,7 @@ retry:
 		if (IS_ERR(handle)) {
 			ret = PTR_ERR(handle);
 			ext4_msg(inode->i_sb, KERN_CRIT, "%s: jbd2_start: "
-			       "%ld pages, ino %lu; err %d\n", __func__,
+			       "%ld pages, ino %lu; err %d", __func__,
 				wbc->nr_to_write, inode->i_ino, ret);
 			goto out_writepages;
 		}
@@ -2963,8 +3033,7 @@ retry:
 		mpd.io_done = 0;
 		mpd.pages_written = 0;
 		mpd.retval = 0;
-		ret = write_cache_pages(mapping, wbc, __mpage_da_writepage,
-					&mpd);
+		ret = write_cache_pages_da(mapping, wbc, &mpd);
 		/*
 		 * If we have a contiguous extent of pages and we
 		 * haven't done the I/O yet, map the blocks and submit
@@ -3016,7 +3085,7 @@ retry:
 	if (pages_skipped != wbc->pages_skipped)
 		ext4_msg(inode->i_sb, KERN_CRIT,
 			 "This should not happen leaving %s "
-			 "with nr_to_write = %ld ret = %d\n",
+			 "with nr_to_write = %ld ret = %d",
 			 __func__, wbc->nr_to_write, ret);
 
 	/* Update index */
@@ -3030,8 +3099,6 @@ retry:
 		mapping->writeback_index = index;
 
 out_writepages:
-	if (!no_nrwrite_index_update)
-		wbc->no_nrwrite_index_update = 0;
 	wbc->nr_to_write -= nr_to_writebump;
 	wbc->range_start = range_start;
 	trace_ext4_da_writepages_result(inode, wbc, ret, pages_written);
@@ -3076,7 +3143,7 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping,
 			       loff_t pos, unsigned len, unsigned flags,
 			       struct page **pagep, void **fsdata)
 {
-	int ret, retries = 0, quota_retries = 0;
+	int ret, retries = 0;
 	struct page *page;
 	pgoff_t index;
 	unsigned from, to;
@@ -3135,22 +3202,6 @@ retry:
 
 	if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries))
 		goto retry;
-
-	if ((ret == -EDQUOT) &&
-	    EXT4_I(inode)->i_reserved_meta_blocks &&
-	    (quota_retries++ < 3)) {
-		/*
-		 * Since we often over-estimate the number of meta
-		 * data blocks required, we may sometimes get a
-		 * spurios out of quota error even though there would
-		 * be enough space once we write the data blocks and
-		 * find out how many meta data blocks were _really_
-		 * required.  So try forcing the inode write to see if
-		 * that helps.
-		 */
-		write_inode_now(inode, (quota_retries == 3));
-		goto retry;
-	}
 out:
 	return ret;
 }
@@ -3546,46 +3597,18 @@ out:
 	return ret;
 }
 
+/*
+ * ext4_get_block used when preparing for a DIO write or buffer write.
+ * We allocate an uinitialized extent if blocks haven't been allocated.
+ * The extent will be converted to initialized after the IO is complete.
+ */
 static int ext4_get_block_write(struct inode *inode, sector_t iblock,
 		   struct buffer_head *bh_result, int create)
 {
-	handle_t *handle = ext4_journal_current_handle();
-	int ret = 0;
-	unsigned max_blocks = bh_result->b_size >> inode->i_blkbits;
-	int dio_credits;
-	int started = 0;
-
 	ext4_debug("ext4_get_block_write: inode %lu, create flag %d\n",
 		   inode->i_ino, create);
-	/*
-	 * ext4_get_block in prepare for a DIO write or buffer write.
-	 * We allocate an uinitialized extent if blocks haven't been allocated.
-	 * The extent will be converted to initialized after IO complete.
-	 */
-	create = EXT4_GET_BLOCKS_IO_CREATE_EXT;
-
-	if (!handle) {
-		if (max_blocks > DIO_MAX_BLOCKS)
-			max_blocks = DIO_MAX_BLOCKS;
-		dio_credits = ext4_chunk_trans_blocks(inode, max_blocks);
-		handle = ext4_journal_start(inode, dio_credits);
-		if (IS_ERR(handle)) {
-			ret = PTR_ERR(handle);
-			goto out;
-		}
-		started = 1;
-	}
-
-	ret = ext4_get_blocks(handle, inode, iblock, max_blocks, bh_result,
-			      create);
-	if (ret > 0) {
-		bh_result->b_size = (ret << inode->i_blkbits);
-		ret = 0;
-	}
-	if (started)
-		ext4_journal_stop(handle);
-out:
-	return ret;
+	return _ext4_get_block(inode, iblock, bh_result,
+			       EXT4_GET_BLOCKS_IO_CREATE_EXT);
 }
 
 static void dump_completed_IO(struct inode * inode)
@@ -3973,7 +3996,7 @@ static ssize_t ext4_direct_IO(int rw, struct kiocb *iocb,
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file->f_mapping->host;
 
-	if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL)
+	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
 		return ext4_ext_direct_IO(rw, iocb, iov, offset, nr_segs);
 
 	return ext4_ind_direct_IO(rw, iocb, iov, offset, nr_segs);
@@ -4302,10 +4325,9 @@ static int ext4_clear_blocks(handle_t *handle, struct inode *inode,
 
 	if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), block_to_free,
 				   count)) {
-		ext4_error(inode->i_sb, "inode #%lu: "
-			   "attempt to clear blocks %llu len %lu, invalid",
-			   inode->i_ino, (unsigned long long) block_to_free,
-			   count);
+		EXT4_ERROR_INODE(inode, "attempt to clear invalid "
+				 "blocks %llu len %lu",
+				 (unsigned long long) block_to_free, count);
 		return 1;
 	}
 
@@ -4410,11 +4432,10 @@ static void ext4_free_data(handle_t *handle, struct inode *inode,
 		if ((EXT4_JOURNAL(inode) == NULL) || bh2jh(this_bh))
 			ext4_handle_dirty_metadata(handle, inode, this_bh);
 		else
-			ext4_error(inode->i_sb,
-				   "circular indirect block detected, "
-				   "inode=%lu, block=%llu",
-				   inode->i_ino,
-				   (unsigned long long) this_bh->b_blocknr);
+			EXT4_ERROR_INODE(inode,
+					 "circular indirect block detected at "
+					 "block %llu",
+				(unsigned long long) this_bh->b_blocknr);
 	}
 }
 
@@ -4452,11 +4473,10 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 
 			if (!ext4_data_block_valid(EXT4_SB(inode->i_sb),
 						   nr, 1)) {
-				ext4_error(inode->i_sb,
-					   "indirect mapped block in inode "
-					   "#%lu invalid (level %d, blk #%lu)",
-					   inode->i_ino, depth,
-					   (unsigned long) nr);
+				EXT4_ERROR_INODE(inode,
+						 "invalid indirect mapped "
+						 "block %lu (level %d)",
+						 (unsigned long) nr, depth);
 				break;
 			}
 
@@ -4468,9 +4488,9 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 			 * (should be rare).
 			 */
 			if (!bh) {
-				ext4_error(inode->i_sb,
-					   "Read failure, inode=%lu, block=%llu",
-					   inode->i_ino, nr);
+				EXT4_ERROR_INODE(inode,
+						 "Read failure block=%llu",
+						 (unsigned long long) nr);
 				continue;
 			}
 
@@ -4612,12 +4632,12 @@ void ext4_truncate(struct inode *inode)
 	if (!ext4_can_truncate(inode))
 		return;
 
-	EXT4_I(inode)->i_flags &= ~EXT4_EOFBLOCKS_FL;
+	ext4_clear_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
 
 	if (inode->i_size == 0 && !test_opt(inode->i_sb, NO_AUTO_DA_ALLOC))
 		ext4_set_inode_state(inode, EXT4_STATE_DA_ALLOC_CLOSE);
 
-	if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL) {
+	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
 		ext4_ext_truncate(inode);
 		return;
 	}
@@ -4785,8 +4805,8 @@ static int __ext4_get_inode_loc(struct inode *inode,
 
 	bh = sb_getblk(sb, block);
 	if (!bh) {
-		ext4_error(sb, "unable to read inode block - "
-			   "inode=%lu, block=%llu", inode->i_ino, block);
+		EXT4_ERROR_INODE(inode, "unable to read inode block - "
+				 "block %llu", block);
 		return -EIO;
 	}
 	if (!buffer_uptodate(bh)) {
@@ -4884,8 +4904,8 @@ make_io:
 		submit_bh(READ_META, bh);
 		wait_on_buffer(bh);
 		if (!buffer_uptodate(bh)) {
-			ext4_error(sb, "unable to read inode block - inode=%lu,"
-				   " block=%llu", inode->i_ino, block);
+			EXT4_ERROR_INODE(inode, "unable to read inode "
+					 "block %llu", block);
 			brelse(bh);
 			return -EIO;
 		}
@@ -5096,8 +5116,8 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
 	ret = 0;
 	if (ei->i_file_acl &&
 	    !ext4_data_block_valid(EXT4_SB(sb), ei->i_file_acl, 1)) {
-		ext4_error(sb, "bad extended attribute block %llu inode #%lu",
-			   ei->i_file_acl, inode->i_ino);
+		EXT4_ERROR_INODE(inode, "bad extended attribute block %llu",
+				 ei->i_file_acl);
 		ret = -EIO;
 		goto bad_inode;
 	} else if (ei->i_flags & EXT4_EXTENTS_FL) {
@@ -5142,8 +5162,7 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
 			   new_decode_dev(le32_to_cpu(raw_inode->i_block[1])));
 	} else {
 		ret = -EIO;
-		ext4_error(inode->i_sb, "bogus i_mode (%o) for inode=%lu",
-			   inode->i_mode, inode->i_ino);
+		EXT4_ERROR_INODE(inode, "bogus i_mode (%o)", inode->i_mode);
 		goto bad_inode;
 	}
 	brelse(iloc.bh);
@@ -5381,9 +5400,9 @@ int ext4_write_inode(struct inode *inode, struct writeback_control *wbc)
 		if (wbc->sync_mode == WB_SYNC_ALL)
 			sync_dirty_buffer(iloc.bh);
 		if (buffer_req(iloc.bh) && !buffer_uptodate(iloc.bh)) {
-			ext4_error(inode->i_sb, "IO error syncing inode, "
-				   "inode=%lu, block=%llu", inode->i_ino,
-				   (unsigned long long)iloc.bh->b_blocknr);
+			EXT4_ERROR_INODE(inode,
+				"IO error syncing inode (block=%llu)",
+				(unsigned long long) iloc.bh->b_blocknr);
 			err = -EIO;
 		}
 		brelse(iloc.bh);
@@ -5455,7 +5474,7 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
 	}
 
 	if (attr->ia_valid & ATTR_SIZE) {
-		if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL)) {
+		if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
 			struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 
 			if (attr->ia_size > sbi->s_bitmap_maxbytes) {
@@ -5468,7 +5487,7 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
 	if (S_ISREG(inode->i_mode) &&
 	    attr->ia_valid & ATTR_SIZE &&
 	    (attr->ia_size < inode->i_size ||
-	     (EXT4_I(inode)->i_flags & EXT4_EOFBLOCKS_FL))) {
+	     (ext4_test_inode_flag(inode, EXT4_INODE_EOFBLOCKS)))) {
 		handle_t *handle;
 
 		handle = ext4_journal_start(inode, 3);
@@ -5500,7 +5519,7 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
 			}
 		}
 		/* ext4_truncate will clear the flag */
-		if ((EXT4_I(inode)->i_flags & EXT4_EOFBLOCKS_FL))
+		if ((ext4_test_inode_flag(inode, EXT4_INODE_EOFBLOCKS)))
 			ext4_truncate(inode);
 	}
 
@@ -5576,7 +5595,7 @@ static int ext4_indirect_trans_blocks(struct inode *inode, int nrblocks,
 
 static int ext4_index_trans_blocks(struct inode *inode, int nrblocks, int chunk)
 {
-	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 		return ext4_indirect_trans_blocks(inode, nrblocks, chunk);
 	return ext4_ext_index_trans_blocks(inode, nrblocks, chunk);
 }
@@ -5911,9 +5930,9 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val)
 	 */
 
 	if (val)
-		EXT4_I(inode)->i_flags |= EXT4_JOURNAL_DATA_FL;
+		ext4_set_inode_flag(inode, EXT4_INODE_JOURNAL_DATA);
 	else
-		EXT4_I(inode)->i_flags &= ~EXT4_JOURNAL_DATA_FL;
+		ext4_clear_inode_flag(inode, EXT4_INODE_JOURNAL_DATA);
 	ext4_set_aops(inode);
 
 	jbd2_journal_unlock_updates(journal);
diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
index 016d0249294f..bf5ae883b1bd 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -258,7 +258,7 @@ setversion_out:
 		if (me.moved_len > 0)
 			file_remove_suid(donor_filp);
 
-		if (copy_to_user((struct move_extent __user *)arg, 
+		if (copy_to_user((struct move_extent __user *)arg,
 				 &me, sizeof(me)))
 			err = -EFAULT;
 mext_out:
@@ -373,7 +373,30 @@ long ext4_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 	case EXT4_IOC32_SETRSVSZ:
 		cmd = EXT4_IOC_SETRSVSZ;
 		break;
-	case EXT4_IOC_GROUP_ADD:
+	case EXT4_IOC32_GROUP_ADD: {
+		struct compat_ext4_new_group_input __user *uinput;
+		struct ext4_new_group_input input;
+		mm_segment_t old_fs;
+		int err;
+
+		uinput = compat_ptr(arg);
+		err = get_user(input.group, &uinput->group);
+		err |= get_user(input.block_bitmap, &uinput->block_bitmap);
+		err |= get_user(input.inode_bitmap, &uinput->inode_bitmap);
+		err |= get_user(input.inode_table, &uinput->inode_table);
+		err |= get_user(input.blocks_count, &uinput->blocks_count);
+		err |= get_user(input.reserved_blocks,
+				&uinput->reserved_blocks);
+		if (err)
+			return -EFAULT;
+		old_fs = get_fs();
+		set_fs(KERNEL_DS);
+		err = ext4_ioctl(file, EXT4_IOC_GROUP_ADD,
+				 (unsigned long) &input);
+		set_fs(old_fs);
+		return err;
+	}
+	case EXT4_IOC_MOVE_EXT:
 		break;
 	default:
 		return -ENOIOCTLCMD;
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index b423a364dca3..12b3bc026a68 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -658,6 +658,27 @@ static void ext4_mb_mark_free_simple(struct super_block *sb,
 	}
 }
 
+/*
+ * Cache the order of the largest free extent we have available in this block
+ * group.
+ */
+static void
+mb_set_largest_free_order(struct super_block *sb, struct ext4_group_info *grp)
+{
+	int i;
+	int bits;
+
+	grp->bb_largest_free_order = -1; /* uninit */
+
+	bits = sb->s_blocksize_bits + 1;
+	for (i = bits; i >= 0; i--) {
+		if (grp->bb_counters[i] > 0) {
+			grp->bb_largest_free_order = i;
+			break;
+		}
+	}
+}
+
 static noinline_for_stack
 void ext4_mb_generate_buddy(struct super_block *sb,
 				void *buddy, void *bitmap, ext4_group_t group)
@@ -700,6 +721,7 @@ void ext4_mb_generate_buddy(struct super_block *sb,
 		 */
 		grp->bb_free = free;
 	}
+	mb_set_largest_free_order(sb, grp);
 
 	clear_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(grp->bb_state));
 
@@ -725,6 +747,9 @@ void ext4_mb_generate_buddy(struct super_block *sb,
  * contain blocks_per_page (PAGE_CACHE_SIZE / blocksize)  blocks.
  * So it can have information regarding groups_per_page which
  * is blocks_per_page/2
+ *
+ * Locking note:  This routine takes the block group lock of all groups
+ * for this page; do not hold this lock when calling this routine!
  */
 
 static int ext4_mb_init_cache(struct page *page, char *incore)
@@ -865,6 +890,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 			BUG_ON(incore == NULL);
 			mb_debug(1, "put buddy for group %u in page %lu/%x\n",
 				group, page->index, i * blocksize);
+			trace_ext4_mb_buddy_bitmap_load(sb, group);
 			grinfo = ext4_get_group_info(sb, group);
 			grinfo->bb_fragments = 0;
 			memset(grinfo->bb_counters, 0,
@@ -882,6 +908,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 			BUG_ON(incore != NULL);
 			mb_debug(1, "put bitmap for group %u in page %lu/%x\n",
 				group, page->index, i * blocksize);
+			trace_ext4_mb_bitmap_load(sb, group);
 
 			/* see comments in ext4_mb_put_pa() */
 			ext4_lock_group(sb, group);
@@ -910,6 +937,11 @@ out:
 	return err;
 }
 
+/*
+ * Locking note:  This routine calls ext4_mb_init_cache(), which takes the
+ * block group lock of all groups for this page; do not hold the BG lock when
+ * calling this routine!
+ */
 static noinline_for_stack
 int ext4_mb_init_group(struct super_block *sb, ext4_group_t group)
 {
@@ -1004,6 +1036,11 @@ err:
 	return ret;
 }
 
+/*
+ * Locking note:  This routine calls ext4_mb_init_cache(), which takes the
+ * block group lock of all groups for this page; do not hold the BG lock when
+ * calling this routine!
+ */
 static noinline_for_stack int
 ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
 					struct ext4_buddy *e4b)
@@ -1150,7 +1187,7 @@ err:
 	return ret;
 }
 
-static void ext4_mb_release_desc(struct ext4_buddy *e4b)
+static void ext4_mb_unload_buddy(struct ext4_buddy *e4b)
 {
 	if (e4b->bd_bitmap_page)
 		page_cache_release(e4b->bd_bitmap_page);
@@ -1299,6 +1336,7 @@ static void mb_free_blocks(struct inode *inode, struct ext4_buddy *e4b,
 			buddy = buddy2;
 		} while (1);
 	}
+	mb_set_largest_free_order(sb, e4b->bd_info);
 	mb_check_buddy(e4b);
 }
 
@@ -1427,6 +1465,7 @@ static int mb_mark_used(struct ext4_buddy *e4b, struct ext4_free_extent *ex)
 		e4b->bd_info->bb_counters[ord]++;
 		e4b->bd_info->bb_counters[ord]++;
 	}
+	mb_set_largest_free_order(e4b->bd_sb, e4b->bd_info);
 
 	mb_set_bits(EXT4_MB_BITMAP(e4b), ex->fe_start, len0);
 	mb_check_buddy(e4b);
@@ -1617,7 +1656,7 @@ int ext4_mb_try_best_found(struct ext4_allocation_context *ac,
 	}
 
 	ext4_unlock_group(ac->ac_sb, group);
-	ext4_mb_release_desc(e4b);
+	ext4_mb_unload_buddy(e4b);
 
 	return 0;
 }
@@ -1672,7 +1711,7 @@ int ext4_mb_find_by_goal(struct ext4_allocation_context *ac,
 		ext4_mb_use_best_found(ac, e4b);
 	}
 	ext4_unlock_group(ac->ac_sb, group);
-	ext4_mb_release_desc(e4b);
+	ext4_mb_unload_buddy(e4b);
 
 	return 0;
 }
@@ -1821,16 +1860,22 @@ void ext4_mb_scan_aligned(struct ext4_allocation_context *ac,
 	}
 }
 
+/* This is now called BEFORE we load the buddy bitmap. */
 static int ext4_mb_good_group(struct ext4_allocation_context *ac,
 				ext4_group_t group, int cr)
 {
 	unsigned free, fragments;
-	unsigned i, bits;
 	int flex_size = ext4_flex_bg_size(EXT4_SB(ac->ac_sb));
 	struct ext4_group_info *grp = ext4_get_group_info(ac->ac_sb, group);
 
 	BUG_ON(cr < 0 || cr >= 4);
-	BUG_ON(EXT4_MB_GRP_NEED_INIT(grp));
+
+	/* We only do this if the grp has never been initialized */
+	if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) {
+		int ret = ext4_mb_init_group(ac->ac_sb, group);
+		if (ret)
+			return 0;
+	}
 
 	free = grp->bb_free;
 	fragments = grp->bb_fragments;
@@ -1843,17 +1888,16 @@ static int ext4_mb_good_group(struct ext4_allocation_context *ac,
 	case 0:
 		BUG_ON(ac->ac_2order == 0);
 
+		if (grp->bb_largest_free_order < ac->ac_2order)
+			return 0;
+
 		/* Avoid using the first bg of a flexgroup for data files */
 		if ((ac->ac_flags & EXT4_MB_HINT_DATA) &&
 		    (flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) &&
 		    ((group % flex_size) == 0))
 			return 0;
 
-		bits = ac->ac_sb->s_blocksize_bits + 1;
-		for (i = ac->ac_2order; i <= bits; i++)
-			if (grp->bb_counters[i] > 0)
-				return 1;
-		break;
+		return 1;
 	case 1:
 		if ((free / fragments) >= ac->ac_g_ex.fe_len)
 			return 1;
@@ -1964,7 +2008,7 @@ ext4_mb_regular_allocator(struct ext4_allocation_context *ac)
 	sbi = EXT4_SB(sb);
 	ngroups = ext4_get_groups_count(sb);
 	/* non-extent files are limited to low blocks/groups */
-	if (!(EXT4_I(ac->ac_inode)->i_flags & EXT4_EXTENTS_FL))
+	if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS)))
 		ngroups = sbi->s_blockfile_groups;
 
 	BUG_ON(ac->ac_status == AC_STATUS_FOUND);
@@ -2024,15 +2068,11 @@ repeat:
 		group = ac->ac_g_ex.fe_group;
 
 		for (i = 0; i < ngroups; group++, i++) {
-			struct ext4_group_info *grp;
-			struct ext4_group_desc *desc;
-
 			if (group == ngroups)
 				group = 0;
 
-			/* quick check to skip empty groups */
-			grp = ext4_get_group_info(sb, group);
-			if (grp->bb_free == 0)
+			/* This now checks without needing the buddy page */
+			if (!ext4_mb_good_group(ac, group, cr))
 				continue;
 
 			err = ext4_mb_load_buddy(sb, group, &e4b);
@@ -2040,15 +2080,18 @@ repeat:
 				goto out;
 
 			ext4_lock_group(sb, group);
+
+			/*
+			 * We need to check again after locking the
+			 * block group
+			 */
 			if (!ext4_mb_good_group(ac, group, cr)) {
-				/* someone did allocation from this group */
 				ext4_unlock_group(sb, group);
-				ext4_mb_release_desc(&e4b);
+				ext4_mb_unload_buddy(&e4b);
 				continue;
 			}
 
 			ac->ac_groups_scanned++;
-			desc = ext4_get_group_desc(sb, group, NULL);
 			if (cr == 0)
 				ext4_mb_simple_scan_group(ac, &e4b);
 			else if (cr == 1 &&
@@ -2058,7 +2101,7 @@ repeat:
 				ext4_mb_complex_scan_group(ac, &e4b);
 
 			ext4_unlock_group(sb, group);
-			ext4_mb_release_desc(&e4b);
+			ext4_mb_unload_buddy(&e4b);
 
 			if (ac->ac_status != AC_STATUS_CONTINUE)
 				break;
@@ -2148,7 +2191,7 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
 	ext4_lock_group(sb, group);
 	memcpy(&sg, ext4_get_group_info(sb, group), i);
 	ext4_unlock_group(sb, group);
-	ext4_mb_release_desc(&e4b);
+	ext4_mb_unload_buddy(&e4b);
 
 	seq_printf(seq, "#%-5u: %-5u %-5u %-5u [", group, sg.info.bb_free,
 			sg.info.bb_fragments, sg.info.bb_first_free);
@@ -2255,6 +2298,7 @@ int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group,
 	INIT_LIST_HEAD(&meta_group_info[i]->bb_prealloc_list);
 	init_rwsem(&meta_group_info[i]->alloc_sem);
 	meta_group_info[i]->bb_free_root = RB_ROOT;
+	meta_group_info[i]->bb_largest_free_order = -1;  /* uninit */
 
 #ifdef DOUBLE_CHECK
 	{
@@ -2536,6 +2580,7 @@ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn)
 			 entry->count, entry->group, entry);
 
 		if (test_opt(sb, DISCARD)) {
+			int ret;
 			ext4_fsblk_t discard_block;
 
 			discard_block = entry->start_blk +
@@ -2543,7 +2588,12 @@ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn)
 			trace_ext4_discard_blocks(sb,
 					(unsigned long long)discard_block,
 					entry->count);
-			sb_issue_discard(sb, discard_block, entry->count);
+			ret = sb_issue_discard(sb, discard_block, entry->count);
+			if (ret == EOPNOTSUPP) {
+				ext4_warning(sb,
+					"discard not supported, disabling");
+				clear_opt(EXT4_SB(sb)->s_mount_opt, DISCARD);
+			}
 		}
 
 		err = ext4_mb_load_buddy(sb, entry->group, &e4b);
@@ -2568,7 +2618,7 @@ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn)
 		}
 		ext4_unlock_group(sb, entry->group);
 		kmem_cache_free(ext4_free_ext_cachep, entry);
-		ext4_mb_release_desc(&e4b);
+		ext4_mb_unload_buddy(&e4b);
 	}
 
 	mb_debug(1, "freed %u blocks in %u structures\n", count, count2);
@@ -2641,7 +2691,7 @@ int __init init_ext4_mballoc(void)
 
 void exit_ext4_mballoc(void)
 {
-	/* 
+	/*
 	 * Wait for completion of call_rcu()'s on ext4_pspace_cachep
 	 * before destroying the slab cache.
 	 */
@@ -2981,7 +3031,7 @@ static void ext4_mb_collect_stats(struct ext4_allocation_context *ac)
 	if (sbi->s_mb_stats && ac->ac_g_ex.fe_len > 1) {
 		atomic_inc(&sbi->s_bal_reqs);
 		atomic_add(ac->ac_b_ex.fe_len, &sbi->s_bal_allocated);
-		if (ac->ac_o_ex.fe_len >= ac->ac_g_ex.fe_len)
+		if (ac->ac_b_ex.fe_len >= ac->ac_o_ex.fe_len)
 			atomic_inc(&sbi->s_bal_success);
 		atomic_add(ac->ac_found, &sbi->s_bal_ex_scanned);
 		if (ac->ac_g_ex.fe_start == ac->ac_b_ex.fe_start &&
@@ -3123,7 +3173,7 @@ ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
 			continue;
 
 		/* non-extent files can't have physical blocks past 2^32 */
-		if (!(EXT4_I(ac->ac_inode)->i_flags & EXT4_EXTENTS_FL) &&
+		if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS)) &&
 			pa->pa_pstart + pa->pa_len > EXT4_MAX_BLOCK_FILE_PHYS)
 			continue;
 
@@ -3280,7 +3330,7 @@ static void ext4_mb_put_pa(struct ext4_allocation_context *ac,
 	spin_unlock(&pa->pa_lock);
 
 	grp_blk = pa->pa_pstart;
-	/* 
+	/*
 	 * If doing group-based preallocation, pa_pstart may be in the
 	 * next group when pa is used up
 	 */
@@ -3697,7 +3747,7 @@ out:
 	ext4_unlock_group(sb, group);
 	if (ac)
 		kmem_cache_free(ext4_ac_cachep, ac);
-	ext4_mb_release_desc(&e4b);
+	ext4_mb_unload_buddy(&e4b);
 	put_bh(bitmap_bh);
 	return free;
 }
@@ -3801,7 +3851,7 @@ repeat:
 		if (bitmap_bh == NULL) {
 			ext4_error(sb, "Error reading block bitmap for %u",
 					group);
-			ext4_mb_release_desc(&e4b);
+			ext4_mb_unload_buddy(&e4b);
 			continue;
 		}
 
@@ -3810,7 +3860,7 @@ repeat:
 		ext4_mb_release_inode_pa(&e4b, bitmap_bh, pa, ac);
 		ext4_unlock_group(sb, group);
 
-		ext4_mb_release_desc(&e4b);
+		ext4_mb_unload_buddy(&e4b);
 		put_bh(bitmap_bh);
 
 		list_del(&pa->u.pa_tmp_list);
@@ -4074,7 +4124,7 @@ ext4_mb_discard_lg_preallocations(struct super_block *sb,
 		ext4_mb_release_group_pa(&e4b, pa, ac);
 		ext4_unlock_group(sb, group);
 
-		ext4_mb_release_desc(&e4b);
+		ext4_mb_unload_buddy(&e4b);
 		list_del(&pa->u.pa_tmp_list);
 		call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback);
 	}
@@ -4484,12 +4534,12 @@ void ext4_free_blocks(handle_t *handle, struct inode *inode,
 			if (!bh)
 				tbh = sb_find_get_block(inode->i_sb,
 							block + i);
-			ext4_forget(handle, flags & EXT4_FREE_BLOCKS_METADATA, 
+			ext4_forget(handle, flags & EXT4_FREE_BLOCKS_METADATA,
 				    inode, tbh, block + i);
 		}
 	}
 
-	/* 
+	/*
 	 * We need to make sure we don't reuse the freed block until
 	 * after the transaction is committed, which we can do by
 	 * treating the block as metadata, below.  We make an
@@ -4610,7 +4660,7 @@ do_more:
 		atomic_add(count, &sbi->s_flex_groups[flex_group].free_blocks);
 	}
 
-	ext4_mb_release_desc(&e4b);
+	ext4_mb_unload_buddy(&e4b);
 
 	freed += count;
 
diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c
index 34dcfc52ef44..6f3a27ec30bf 100644
--- a/fs/ext4/migrate.c
+++ b/fs/ext4/migrate.c
@@ -475,7 +475,7 @@ int ext4_ext_migrate(struct inode *inode)
 	 */
 	if (!EXT4_HAS_INCOMPAT_FEATURE(inode->i_sb,
 				       EXT4_FEATURE_INCOMPAT_EXTENTS) ||
-	    (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+	    (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 		return -EINVAL;
 
 	if (S_ISLNK(inode->i_mode) && inode->i_blocks == 0)
diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index d1fc662cc311..3a6c92ac131c 100644
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -482,6 +482,7 @@ mext_leaf_block(handle_t *handle, struct inode *orig_inode,
 	int depth = ext_depth(orig_inode);
 	int ret;
 
+	start_ext.ee_block = end_ext.ee_block = 0;
 	o_start = o_end = oext = orig_path[depth].p_ext;
 	oext_alen = ext4_ext_get_actual_len(oext);
 	start_ext.ee_len = end_ext.ee_len = 0;
@@ -529,7 +530,7 @@ mext_leaf_block(handle_t *handle, struct inode *orig_inode,
 	 * new_ext       |-------|
 	 */
 	if (le32_to_cpu(oext->ee_block) + oext_alen - 1 < new_ext_end) {
-		ext4_error(orig_inode->i_sb,
+		EXT4_ERROR_INODE(orig_inode,
 			"new_ext_end(%u) should be less than or equal to "
 			"oext->ee_block(%u) + oext_alen(%d) - 1",
 			new_ext_end, le32_to_cpu(oext->ee_block),
@@ -692,12 +693,12 @@ mext_replace_branches(handle_t *handle, struct inode *orig_inode,
 	while (1) {
 		/* The extent for donor must be found. */
 		if (!dext) {
-			ext4_error(donor_inode->i_sb,
+			EXT4_ERROR_INODE(donor_inode,
 				   "The extent for donor must be found");
 			*err = -EIO;
 			goto out;
 		} else if (donor_off != le32_to_cpu(tmp_dext.ee_block)) {
-			ext4_error(donor_inode->i_sb,
+			EXT4_ERROR_INODE(donor_inode,
 				"Donor offset(%u) and the first block of donor "
 				"extent(%u) should be equal",
 				donor_off,
@@ -976,11 +977,11 @@ mext_check_arguments(struct inode *orig_inode,
 	}
 
 	/* Ext4 move extent supports only extent based file */
-	if (!(EXT4_I(orig_inode)->i_flags & EXT4_EXTENTS_FL)) {
+	if (!(ext4_test_inode_flag(orig_inode, EXT4_INODE_EXTENTS))) {
 		ext4_debug("ext4 move extent: orig file is not extents "
 			"based file [ino:orig %lu]\n", orig_inode->i_ino);
 		return -EOPNOTSUPP;
-	} else if (!(EXT4_I(donor_inode)->i_flags & EXT4_EXTENTS_FL)) {
+	} else if (!(ext4_test_inode_flag(donor_inode, EXT4_INODE_EXTENTS))) {
 		ext4_debug("ext4 move extent: donor file is not extents "
 			"based file [ino:donor %lu]\n", donor_inode->i_ino);
 		return -EOPNOTSUPP;
@@ -1354,7 +1355,7 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp,
 			if (ret1 < 0)
 				break;
 			if (*moved_len > len) {
-				ext4_error(orig_inode->i_sb,
+				EXT4_ERROR_INODE(orig_inode,
 					"We replaced blocks too much! "
 					"sum of replaced: %llu requested: %llu",
 					*moved_len, len);
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 0c070fabd108..a43e6617b351 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -187,7 +187,7 @@ unsigned int ext4_rec_len_from_disk(__le16 dlen, unsigned blocksize)
 		return blocksize;
 	return (len & 65532) | ((len & 3) << 16);
 }
-  
+
 __le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize)
 {
 	if ((len > blocksize) || (blocksize > (1 << 18)) || (len & 3))
@@ -197,7 +197,7 @@ __le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize)
 	if (len == blocksize) {
 		if (blocksize == 65536)
 			return cpu_to_le16(EXT4_MAX_REC_LEN);
-		else 
+		else
 			return cpu_to_le16(0);
 	}
 	return cpu_to_le16((len & 65532) | ((len >> 16) & 3));
@@ -349,7 +349,7 @@ struct stats dx_show_entries(struct dx_hash_info *hinfo, struct inode *dir,
 		brelse(bh);
 	}
 	if (bcount)
-		printk(KERN_DEBUG "%snames %u, fullness %u (%u%%)\n", 
+		printk(KERN_DEBUG "%snames %u, fullness %u (%u%%)\n",
 		       levels ? "" : "   ", names, space/bcount,
 		       (space/bcount)*100/blocksize);
 	return (struct stats) { names, space, bcount};
@@ -653,10 +653,10 @@ int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash,
 	int ret, err;
 	__u32 hashval;
 
-	dxtrace(printk(KERN_DEBUG "In htree_fill_tree, start hash: %x:%x\n", 
+	dxtrace(printk(KERN_DEBUG "In htree_fill_tree, start hash: %x:%x\n",
 		       start_hash, start_minor_hash));
 	dir = dir_file->f_path.dentry->d_inode;
-	if (!(EXT4_I(dir)->i_flags & EXT4_INDEX_FL)) {
+	if (!(ext4_test_inode_flag(dir, EXT4_INODE_INDEX))) {
 		hinfo.hash_version = EXT4_SB(dir->i_sb)->s_def_hash_version;
 		if (hinfo.hash_version <= DX_HASH_TEA)
 			hinfo.hash_version +=
@@ -801,7 +801,7 @@ static void ext4_update_dx_flag(struct inode *inode)
 {
 	if (!EXT4_HAS_COMPAT_FEATURE(inode->i_sb,
 				     EXT4_FEATURE_COMPAT_DIR_INDEX))
-		EXT4_I(inode)->i_flags &= ~EXT4_INDEX_FL;
+		ext4_clear_inode_flag(inode, EXT4_INODE_INDEX);
 }
 
 /*
@@ -943,8 +943,8 @@ restart:
 		wait_on_buffer(bh);
 		if (!buffer_uptodate(bh)) {
 			/* read error, skip block & hope for the best */
-			ext4_error(sb, "reading directory #%lu offset %lu",
-				   dir->i_ino, (unsigned long)block);
+			EXT4_ERROR_INODE(dir, "reading directory lblock %lu",
+					 (unsigned long) block);
 			brelse(bh);
 			goto next;
 		}
@@ -1066,15 +1066,15 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, stru
 		__u32 ino = le32_to_cpu(de->inode);
 		brelse(bh);
 		if (!ext4_valid_inum(dir->i_sb, ino)) {
-			ext4_error(dir->i_sb, "bad inode number: %u", ino);
+			EXT4_ERROR_INODE(dir, "bad inode number: %u", ino);
 			return ERR_PTR(-EIO);
 		}
 		inode = ext4_iget(dir->i_sb, ino);
 		if (unlikely(IS_ERR(inode))) {
 			if (PTR_ERR(inode) == -ESTALE) {
-				ext4_error(dir->i_sb,
-						"deleted inode referenced: %u",
-						ino);
+				EXT4_ERROR_INODE(dir,
+						 "deleted inode referenced: %u",
+						 ino);
 				return ERR_PTR(-EIO);
 			} else {
 				return ERR_CAST(inode);
@@ -1104,8 +1104,8 @@ struct dentry *ext4_get_parent(struct dentry *child)
 	brelse(bh);
 
 	if (!ext4_valid_inum(child->d_inode->i_sb, ino)) {
-		ext4_error(child->d_inode->i_sb,
-			   "bad inode number: %u", ino);
+		EXT4_ERROR_INODE(child->d_inode,
+				 "bad parent inode number: %u", ino);
 		return ERR_PTR(-EIO);
 	}
 
@@ -1141,7 +1141,7 @@ dx_move_dirents(char *from, char *to, struct dx_map_entry *map, int count,
 	unsigned rec_len = 0;
 
 	while (count--) {
-		struct ext4_dir_entry_2 *de = (struct ext4_dir_entry_2 *) 
+		struct ext4_dir_entry_2 *de = (struct ext4_dir_entry_2 *)
 						(from + (map->offs<<2));
 		rec_len = EXT4_DIR_REC_LEN(de->name_len);
 		memcpy (to, de, rec_len);
@@ -1404,9 +1404,7 @@ static int make_indexed_dir(handle_t *handle, struct dentry *dentry,
 	de = (struct ext4_dir_entry_2 *)((char *)fde +
 		ext4_rec_len_from_disk(fde->rec_len, blocksize));
 	if ((char *) de >= (((char *) root) + blocksize)) {
-		ext4_error(dir->i_sb,
-			   "invalid rec_len for '..' in inode %lu",
-			   dir->i_ino);
+		EXT4_ERROR_INODE(dir, "invalid rec_len for '..'");
 		brelse(bh);
 		return -EIO;
 	}
@@ -1418,7 +1416,7 @@ static int make_indexed_dir(handle_t *handle, struct dentry *dentry,
 		brelse(bh);
 		return retval;
 	}
-	EXT4_I(dir)->i_flags |= EXT4_INDEX_FL;
+	ext4_set_inode_flag(dir, EXT4_INODE_INDEX);
 	data1 = bh2->b_data;
 
 	memcpy (data1, de, len);
@@ -1491,7 +1489,7 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 		retval = ext4_dx_add_entry(handle, dentry, inode);
 		if (!retval || (retval != ERR_BAD_DX_DIR))
 			return retval;
-		EXT4_I(dir)->i_flags &= ~EXT4_INDEX_FL;
+		ext4_clear_inode_flag(dir, EXT4_INODE_INDEX);
 		dx_fallback++;
 		ext4_mark_inode_dirty(handle, dir);
 	}
@@ -1519,6 +1517,8 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 	de->rec_len = ext4_rec_len_to_disk(blocksize, blocksize);
 	retval = add_dirent_to_buf(handle, dentry, inode, de, bh);
 	brelse(bh);
+	if (retval == 0)
+		ext4_set_inode_state(inode, EXT4_STATE_NEWENTRY);
 	return retval;
 }
 
@@ -1915,9 +1915,8 @@ static int empty_dir(struct inode *inode)
 	if (inode->i_size < EXT4_DIR_REC_LEN(1) + EXT4_DIR_REC_LEN(2) ||
 	    !(bh = ext4_bread(NULL, inode, 0, 0, &err))) {
 		if (err)
-			ext4_error(inode->i_sb,
-				   "error %d reading directory #%lu offset 0",
-				   err, inode->i_ino);
+			EXT4_ERROR_INODE(inode,
+				"error %d reading directory lblock 0", err);
 		else
 			ext4_warning(inode->i_sb,
 				     "bad directory (dir #%lu) - no data block",
@@ -1941,17 +1940,17 @@ static int empty_dir(struct inode *inode)
 	de = ext4_next_entry(de1, sb->s_blocksize);
 	while (offset < inode->i_size) {
 		if (!bh ||
-			(void *) de >= (void *) (bh->b_data+sb->s_blocksize)) {
+		    (void *) de >= (void *) (bh->b_data+sb->s_blocksize)) {
+			unsigned int lblock;
 			err = 0;
 			brelse(bh);
-			bh = ext4_bread(NULL, inode,
-				offset >> EXT4_BLOCK_SIZE_BITS(sb), 0, &err);
+			lblock = offset >> EXT4_BLOCK_SIZE_BITS(sb);
+			bh = ext4_bread(NULL, inode, lblock, 0, &err);
 			if (!bh) {
 				if (err)
-					ext4_error(sb,
-						   "error %d reading directory"
-						   " #%lu offset %u",
-						   err, inode->i_ino, offset);
+					EXT4_ERROR_INODE(inode,
+						"error %d reading directory "
+						"lblock %u", err, lblock);
 				offset += sb->s_blocksize;
 				continue;
 			}
@@ -2297,7 +2296,7 @@ retry:
 		}
 	} else {
 		/* clear the extent format for fast symlink */
-		EXT4_I(inode)->i_flags &= ~EXT4_EXTENTS_FL;
+		ext4_clear_inode_flag(inode, EXT4_INODE_EXTENTS);
 		inode->i_op = &ext4_fast_symlink_inode_operations;
 		memcpy((char *)&EXT4_I(inode)->i_data, symname, l);
 		inode->i_size = l-1;
diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
index 5692c48754a0..6df797eb9aeb 100644
--- a/fs/ext4/resize.c
+++ b/fs/ext4/resize.c
@@ -911,7 +911,8 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 	percpu_counter_add(&sbi->s_freeinodes_counter,
 			   EXT4_INODES_PER_GROUP(sb));
 
-	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG)) {
+	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG) &&
+	    sbi->s_log_groups_per_flex) {
 		ext4_group_t flex_group;
 		flex_group = ext4_flex_group(sbi, input->group);
 		atomic_add(input->free_blocks_count,
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index e14d22c170d5..4e8983a9811b 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -241,6 +241,7 @@ handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks)
 	if (sb->s_flags & MS_RDONLY)
 		return ERR_PTR(-EROFS);
 
+	vfs_check_frozen(sb, SB_FREEZE_WRITE);
 	/* Special case here: if the journal has aborted behind our
 	 * backs (eg. EIO in the commit thread), then we still need to
 	 * take the FS itself readonly cleanly. */
@@ -645,6 +646,8 @@ static void ext4_put_super(struct super_block *sb)
 	struct ext4_super_block *es = sbi->s_es;
 	int i, err;
 
+	dquot_disable(sb, -1, DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
+
 	flush_workqueue(sbi->dio_unwritten_wq);
 	destroy_workqueue(sbi->dio_unwritten_wq);
 
@@ -941,6 +944,8 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs)
 	seq_puts(seq, test_opt(sb, BARRIER) ? "1" : "0");
 	if (test_opt(sb, JOURNAL_ASYNC_COMMIT))
 		seq_puts(seq, ",journal_async_commit");
+	else if (test_opt(sb, JOURNAL_CHECKSUM))
+		seq_puts(seq, ",journal_checksum");
 	if (test_opt(sb, NOBH))
 		seq_puts(seq, ",nobh");
 	if (test_opt(sb, I_VERSION))
@@ -1059,7 +1064,7 @@ static int ext4_release_dquot(struct dquot *dquot);
 static int ext4_mark_dquot_dirty(struct dquot *dquot);
 static int ext4_write_info(struct super_block *sb, int type);
 static int ext4_quota_on(struct super_block *sb, int type, int format_id,
-				char *path, int remount);
+				char *path);
 static int ext4_quota_on_mount(struct super_block *sb, int type);
 static ssize_t ext4_quota_read(struct super_block *sb, int type, char *data,
 			       size_t len, loff_t off);
@@ -1081,12 +1086,12 @@ static const struct dquot_operations ext4_quota_operations = {
 
 static const struct quotactl_ops ext4_qctl_operations = {
 	.quota_on	= ext4_quota_on,
-	.quota_off	= vfs_quota_off,
-	.quota_sync	= vfs_quota_sync,
-	.get_info	= vfs_get_dqinfo,
-	.set_info	= vfs_set_dqinfo,
-	.get_dqblk	= vfs_get_dqblk,
-	.set_dqblk	= vfs_set_dqblk
+	.quota_off	= dquot_quota_off,
+	.quota_sync	= dquot_quota_sync,
+	.get_info	= dquot_get_dqinfo,
+	.set_info	= dquot_set_dqinfo,
+	.get_dqblk	= dquot_get_dqblk,
+	.set_dqblk	= dquot_set_dqblk
 };
 #endif
 
@@ -2051,7 +2056,7 @@ static void ext4_orphan_cleanup(struct super_block *sb,
 	/* Turn quotas off */
 	for (i = 0; i < MAXQUOTAS; i++) {
 		if (sb_dqopt(sb)->files[i])
-			vfs_quota_off(sb, i, 0);
+			dquot_quota_off(sb, i);
 	}
 #endif
 	sb->s_flags = s_flags; /* Restore MS_RDONLY status */
@@ -2213,7 +2218,7 @@ static unsigned long ext4_get_stripe_size(struct ext4_sb_info *sbi)
 struct ext4_attr {
 	struct attribute attr;
 	ssize_t (*show)(struct ext4_attr *, struct ext4_sb_info *, char *);
-	ssize_t (*store)(struct ext4_attr *, struct ext4_sb_info *, 
+	ssize_t (*store)(struct ext4_attr *, struct ext4_sb_info *,
 			 const char *, size_t);
 	int offset;
 };
@@ -2430,6 +2435,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 				__releases(kernel_lock)
 				__acquires(kernel_lock)
 {
+	char *orig_data = kstrdup(data, GFP_KERNEL);
 	struct buffer_head *bh;
 	struct ext4_super_block *es = NULL;
 	struct ext4_sb_info *sbi;
@@ -2793,24 +2799,6 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	get_random_bytes(&sbi->s_next_generation, sizeof(u32));
 	spin_lock_init(&sbi->s_next_gen_lock);
 
-	err = percpu_counter_init(&sbi->s_freeblocks_counter,
-			ext4_count_free_blocks(sb));
-	if (!err) {
-		err = percpu_counter_init(&sbi->s_freeinodes_counter,
-				ext4_count_free_inodes(sb));
-	}
-	if (!err) {
-		err = percpu_counter_init(&sbi->s_dirs_counter,
-				ext4_count_dirs(sb));
-	}
-	if (!err) {
-		err = percpu_counter_init(&sbi->s_dirtyblocks_counter, 0);
-	}
-	if (err) {
-		ext4_msg(sb, KERN_ERR, "insufficient memory");
-		goto failed_mount3;
-	}
-
 	sbi->s_stripe = ext4_get_stripe_size(sbi);
 	sbi->s_max_writeback_mb_bump = 128;
 
@@ -2910,6 +2898,20 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	set_task_ioprio(sbi->s_journal->j_task, journal_ioprio);
 
 no_journal:
+	err = percpu_counter_init(&sbi->s_freeblocks_counter,
+				  ext4_count_free_blocks(sb));
+	if (!err)
+		err = percpu_counter_init(&sbi->s_freeinodes_counter,
+					  ext4_count_free_inodes(sb));
+	if (!err)
+		err = percpu_counter_init(&sbi->s_dirs_counter,
+					  ext4_count_dirs(sb));
+	if (!err)
+		err = percpu_counter_init(&sbi->s_dirtyblocks_counter, 0);
+	if (err) {
+		ext4_msg(sb, KERN_ERR, "insufficient memory");
+		goto failed_mount_wq;
+	}
 	if (test_opt(sb, NOBH)) {
 		if (!(test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_WRITEBACK_DATA)) {
 			ext4_msg(sb, KERN_WARNING, "Ignoring nobh option - "
@@ -3001,7 +3003,7 @@ no_journal:
 	err = ext4_setup_system_zone(sb);
 	if (err) {
 		ext4_msg(sb, KERN_ERR, "failed to initialize system "
-			 "zone (%d)\n", err);
+			 "zone (%d)", err);
 		goto failed_mount4;
 	}
 
@@ -3040,9 +3042,11 @@ no_journal:
 	} else
 		descr = "out journal";
 
-	ext4_msg(sb, KERN_INFO, "mounted filesystem with%s", descr);
+	ext4_msg(sb, KERN_INFO, "mounted filesystem with%s. "
+		"Opts: %s", descr, orig_data);
 
 	lock_kernel();
+	kfree(orig_data);
 	return 0;
 
 cantfind_ext4:
@@ -3059,6 +3063,10 @@ failed_mount_wq:
 		jbd2_journal_destroy(sbi->s_journal);
 		sbi->s_journal = NULL;
 	}
+	percpu_counter_destroy(&sbi->s_freeblocks_counter);
+	percpu_counter_destroy(&sbi->s_freeinodes_counter);
+	percpu_counter_destroy(&sbi->s_dirs_counter);
+	percpu_counter_destroy(&sbi->s_dirtyblocks_counter);
 failed_mount3:
 	if (sbi->s_flex_groups) {
 		if (is_vmalloc_addr(sbi->s_flex_groups))
@@ -3066,10 +3074,6 @@ failed_mount3:
 		else
 			kfree(sbi->s_flex_groups);
 	}
-	percpu_counter_destroy(&sbi->s_freeblocks_counter);
-	percpu_counter_destroy(&sbi->s_freeinodes_counter);
-	percpu_counter_destroy(&sbi->s_dirs_counter);
-	percpu_counter_destroy(&sbi->s_dirtyblocks_counter);
 failed_mount2:
 	for (i = 0; i < db_count; i++)
 		brelse(sbi->s_group_desc[i]);
@@ -3089,6 +3093,7 @@ out_fail:
 	kfree(sbi->s_blockgroup_lock);
 	kfree(sbi);
 	lock_kernel();
+	kfree(orig_data);
 	return ret;
 }
 
@@ -3380,7 +3385,7 @@ static int ext4_commit_super(struct super_block *sb, int sync)
 	if (!(sb->s_flags & MS_RDONLY))
 		es->s_wtime = cpu_to_le32(get_seconds());
 	es->s_kbytes_written =
-		cpu_to_le64(EXT4_SB(sb)->s_kbytes_written + 
+		cpu_to_le64(EXT4_SB(sb)->s_kbytes_written +
 			    ((part_stat_read(sb->s_bdev->bd_part, sectors[1]) -
 			      EXT4_SB(sb)->s_sectors_written_start) >> 1));
 	ext4_free_blocks_count_set(es, percpu_counter_sum_positive(
@@ -3485,8 +3490,10 @@ int ext4_force_commit(struct super_block *sb)
 		return 0;
 
 	journal = EXT4_SB(sb)->s_journal;
-	if (journal)
+	if (journal) {
+		vfs_check_frozen(sb, SB_FREEZE_WRITE);
 		ret = ext4_journal_force_commit(journal);
+	}
 
 	return ret;
 }
@@ -3535,18 +3542,16 @@ static int ext4_freeze(struct super_block *sb)
 	 * the journal.
 	 */
 	error = jbd2_journal_flush(journal);
-	if (error < 0) {
-	out:
-		jbd2_journal_unlock_updates(journal);
-		return error;
-	}
+	if (error < 0)
+		goto out;
 
 	/* Journal blocked and flushed, clear needs_recovery flag. */
 	EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
 	error = ext4_commit_super(sb, 1);
-	if (error)
-		goto out;
-	return 0;
+out:
+	/* we rely on s_frozen to stop further updates */
+	jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
+	return error;
 }
 
 /*
@@ -3563,7 +3568,6 @@ static int ext4_unfreeze(struct super_block *sb)
 	EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
 	ext4_commit_super(sb, 1);
 	unlock_super(sb);
-	jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
 	return 0;
 }
 
@@ -3574,12 +3578,14 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 	ext4_fsblk_t n_blocks_count = 0;
 	unsigned long old_sb_flags;
 	struct ext4_mount_options old_opts;
+	int enable_quota = 0;
 	ext4_group_t g;
 	unsigned int journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
 	int err;
 #ifdef CONFIG_QUOTA
 	int i;
 #endif
+	char *orig_data = kstrdup(data, GFP_KERNEL);
 
 	lock_kernel();
 
@@ -3630,6 +3636,10 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 		}
 
 		if (*flags & MS_RDONLY) {
+			err = dquot_suspend(sb, -1);
+			if (err < 0)
+				goto restore_opts;
+
 			/*
 			 * First of all, the unconditional stuff we have to do
 			 * to disable replay of the journal when we next remount
@@ -3698,6 +3708,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 				goto restore_opts;
 			if (!ext4_setup_super(sb, es, 0))
 				sb->s_flags &= ~MS_RDONLY;
+			enable_quota = 1;
 		}
 	}
 	ext4_setup_system_zone(sb);
@@ -3713,6 +3724,11 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 #endif
 	unlock_super(sb);
 	unlock_kernel();
+	if (enable_quota)
+		dquot_resume(sb, -1);
+
+	ext4_msg(sb, KERN_INFO, "re-mounted. Opts: %s", orig_data);
+	kfree(orig_data);
 	return 0;
 
 restore_opts:
@@ -3734,6 +3750,7 @@ restore_opts:
 #endif
 	unlock_super(sb);
 	unlock_kernel();
+	kfree(orig_data);
 	return err;
 }
 
@@ -3906,24 +3923,21 @@ static int ext4_write_info(struct super_block *sb, int type)
  */
 static int ext4_quota_on_mount(struct super_block *sb, int type)
 {
-	return vfs_quota_on_mount(sb, EXT4_SB(sb)->s_qf_names[type],
-				  EXT4_SB(sb)->s_jquota_fmt, type);
+	return dquot_quota_on_mount(sb, EXT4_SB(sb)->s_qf_names[type],
+					EXT4_SB(sb)->s_jquota_fmt, type);
 }
 
 /*
  * Standard function to be called on quota_on
  */
 static int ext4_quota_on(struct super_block *sb, int type, int format_id,
-			 char *name, int remount)
+			 char *name)
 {
 	int err;
 	struct path path;
 
 	if (!test_opt(sb, QUOTA))
 		return -EINVAL;
-	/* When remounting, no checks are needed and in fact, name is NULL */
-	if (remount)
-		return vfs_quota_on(sb, type, format_id, name, remount);
 
 	err = kern_path(name, LOOKUP_FOLLOW, &path);
 	if (err)
@@ -3962,7 +3976,7 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id,
 		}
 	}
 
-	err = vfs_quota_on_path(sb, type, format_id, &path);
+	err = dquot_quota_on_path(sb, type, format_id, &path);
 	path_put(&path);
 	return err;
 }
@@ -4141,6 +4155,7 @@ static int __init init_ext4_fs(void)
 {
 	int err;
 
+	ext4_check_flag_values();
 	err = init_ext4_system_zone();
 	if (err)
 		return err;
diff --git a/fs/ext4/symlink.c b/fs/ext4/symlink.c
index 00740cb32be3..ed9354aff279 100644
--- a/fs/ext4/symlink.c
+++ b/fs/ext4/symlink.c
@@ -34,6 +34,7 @@ const struct inode_operations ext4_symlink_inode_operations = {
 	.readlink	= generic_readlink,
 	.follow_link	= page_follow_link_light,
 	.put_link	= page_put_link,
+	.setattr	= ext4_setattr,
 #ifdef CONFIG_EXT4_FS_XATTR
 	.setxattr	= generic_setxattr,
 	.getxattr	= generic_getxattr,
@@ -45,6 +46,7 @@ const struct inode_operations ext4_symlink_inode_operations = {
 const struct inode_operations ext4_fast_symlink_inode_operations = {
 	.readlink	= generic_readlink,
 	.follow_link	= ext4_follow_link,
+	.setattr	= ext4_setattr,
 #ifdef CONFIG_EXT4_FS_XATTR
 	.setxattr	= generic_setxattr,
 	.getxattr	= generic_getxattr,
diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
index 2de0e9515089..04338009793a 100644
--- a/fs/ext4/xattr.c
+++ b/fs/ext4/xattr.c
@@ -228,9 +228,8 @@ ext4_xattr_block_get(struct inode *inode, int name_index, const char *name,
 		atomic_read(&(bh->b_count)), le32_to_cpu(BHDR(bh)->h_refcount));
 	if (ext4_xattr_check_block(bh)) {
 bad_block:
-		ext4_error(inode->i_sb,
-			   "inode %lu: bad block %llu", inode->i_ino,
-			   EXT4_I(inode)->i_file_acl);
+		EXT4_ERROR_INODE(inode, "bad block %llu",
+				 EXT4_I(inode)->i_file_acl);
 		error = -EIO;
 		goto cleanup;
 	}
@@ -372,9 +371,8 @@ ext4_xattr_block_list(struct dentry *dentry, char *buffer, size_t buffer_size)
 	ea_bdebug(bh, "b_count=%d, refcount=%d",
 		atomic_read(&(bh->b_count)), le32_to_cpu(BHDR(bh)->h_refcount));
 	if (ext4_xattr_check_block(bh)) {
-		ext4_error(inode->i_sb,
-			   "inode %lu: bad block %llu", inode->i_ino,
-			   EXT4_I(inode)->i_file_acl);
+		EXT4_ERROR_INODE(inode, "bad block %llu",
+				 EXT4_I(inode)->i_file_acl);
 		error = -EIO;
 		goto cleanup;
 	}
@@ -666,8 +664,8 @@ ext4_xattr_block_find(struct inode *inode, struct ext4_xattr_info *i,
 			atomic_read(&(bs->bh->b_count)),
 			le32_to_cpu(BHDR(bs->bh)->h_refcount));
 		if (ext4_xattr_check_block(bs->bh)) {
-			ext4_error(sb, "inode %lu: bad block %llu",
-				   inode->i_ino, EXT4_I(inode)->i_file_acl);
+			EXT4_ERROR_INODE(inode, "bad block %llu",
+					 EXT4_I(inode)->i_file_acl);
 			error = -EIO;
 			goto cleanup;
 		}
@@ -820,7 +818,7 @@ inserted:
 						EXT4_I(inode)->i_block_group);
 
 			/* non-extent files can't have physical blocks past 2^32 */
-			if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+			if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 				goal = goal & EXT4_MAX_BLOCK_FILE_PHYS;
 
 			block = ext4_new_meta_blocks(handle, inode,
@@ -828,7 +826,7 @@ inserted:
 			if (error)
 				goto cleanup;
 
-			if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+			if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 				BUG_ON(block > EXT4_MAX_BLOCK_FILE_PHYS);
 
 			ea_idebug(inode, "creating block %d", block);
@@ -880,8 +878,8 @@ cleanup_dquot:
 	goto cleanup;
 
 bad_block:
-	ext4_error(inode->i_sb, "inode %lu: bad block %llu",
-		   inode->i_ino, EXT4_I(inode)->i_file_acl);
+	EXT4_ERROR_INODE(inode, "bad block %llu",
+			 EXT4_I(inode)->i_file_acl);
 	goto cleanup;
 
 #undef header
@@ -1194,8 +1192,8 @@ retry:
 		if (!bh)
 			goto cleanup;
 		if (ext4_xattr_check_block(bh)) {
-			ext4_error(inode->i_sb, "inode %lu: bad block %llu",
-				   inode->i_ino, EXT4_I(inode)->i_file_acl);
+			EXT4_ERROR_INODE(inode, "bad block %llu",
+					 EXT4_I(inode)->i_file_acl);
 			error = -EIO;
 			goto cleanup;
 		}
@@ -1372,14 +1370,14 @@ ext4_xattr_delete_inode(handle_t *handle, struct inode *inode)
 		goto cleanup;
 	bh = sb_bread(inode->i_sb, EXT4_I(inode)->i_file_acl);
 	if (!bh) {
-		ext4_error(inode->i_sb, "inode %lu: block %llu read error",
-			   inode->i_ino, EXT4_I(inode)->i_file_acl);
+		EXT4_ERROR_INODE(inode, "block %llu read error",
+				 EXT4_I(inode)->i_file_acl);
 		goto cleanup;
 	}
 	if (BHDR(bh)->h_magic != cpu_to_le32(EXT4_XATTR_MAGIC) ||
 	    BHDR(bh)->h_blocks != cpu_to_le32(1)) {
-		ext4_error(inode->i_sb, "inode %lu: bad block %llu",
-			   inode->i_ino, EXT4_I(inode)->i_file_acl);
+		EXT4_ERROR_INODE(inode, "bad block %llu",
+				 EXT4_I(inode)->i_file_acl);
 		goto cleanup;
 	}
 	ext4_xattr_release_block(handle, inode, bh);
@@ -1504,9 +1502,8 @@ again:
 		}
 		bh = sb_bread(inode->i_sb, ce->e_block);
 		if (!bh) {
-			ext4_error(inode->i_sb,
-				"inode %lu: block %lu read error",
-				inode->i_ino, (unsigned long) ce->e_block);
+			EXT4_ERROR_INODE(inode, "block %lu read error",
+					 (unsigned long) ce->e_block);
 		} else if (le32_to_cpu(BHDR(bh)->h_refcount) >=
 				EXT4_XATTR_REFCOUNT_MAX) {
 			ea_idebug(inode, "block %lu refcount %d>=%d",
diff --git a/fs/fat/cache.c b/fs/fat/cache.c
index 113f0a1e565d..ae8200f84e39 100644
--- a/fs/fat/cache.c
+++ b/fs/fat/cache.c
@@ -242,9 +242,10 @@ int fat_get_cluster(struct inode *inode, int cluster, int *fclus, int *dclus)
 	while (*fclus < cluster) {
 		/* prevent the infinite loop of cluster chain */
 		if (*fclus > limit) {
-			fat_fs_error(sb, "%s: detected the cluster chain loop"
-				     " (i_pos %lld)", __func__,
-				     MSDOS_I(inode)->i_pos);
+			fat_fs_error_ratelimit(sb,
+					"%s: detected the cluster chain loop"
+					" (i_pos %lld)", __func__,
+					MSDOS_I(inode)->i_pos);
 			nr = -EIO;
 			goto out;
 		}
@@ -253,9 +254,9 @@ int fat_get_cluster(struct inode *inode, int cluster, int *fclus, int *dclus)
 		if (nr < 0)
 			goto out;
 		else if (nr == FAT_ENT_FREE) {
-			fat_fs_error(sb, "%s: invalid cluster chain"
-				     " (i_pos %lld)", __func__,
-				     MSDOS_I(inode)->i_pos);
+			fat_fs_error_ratelimit(sb, "%s: invalid cluster chain"
+					       " (i_pos %lld)", __func__,
+					       MSDOS_I(inode)->i_pos);
 			nr = -EIO;
 			goto out;
 		} else if (nr == FAT_ENT_EOF) {
diff --git a/fs/fat/dir.c b/fs/fat/dir.c
index 530b4ca01510..ee42b9e0b16a 100644
--- a/fs/fat/dir.c
+++ b/fs/fat/dir.c
@@ -19,6 +19,7 @@
 #include <linux/buffer_head.h>
 #include <linux/compat.h>
 #include <asm/uaccess.h>
+#include <linux/kernel.h>
 #include "fat.h"
 
 /*
@@ -140,28 +141,22 @@ static int uni16_to_x8(unsigned char *ascii, const wchar_t *uni, int len,
 {
 	const wchar_t *ip;
 	wchar_t ec;
-	unsigned char *op, nc;
+	unsigned char *op;
 	int charlen;
-	int k;
 
 	ip = uni;
 	op = ascii;
 
 	while (*ip && ((len - NLS_MAX_CHARSET_SIZE) > 0)) {
 		ec = *ip++;
-		if ( (charlen = nls->uni2char(ec, op, NLS_MAX_CHARSET_SIZE)) > 0) {
+		if ((charlen = nls->uni2char(ec, op, NLS_MAX_CHARSET_SIZE)) > 0) {
 			op += charlen;
 			len -= charlen;
 		} else {
 			if (uni_xlate == 1) {
-				*op = ':';
-				for (k = 4; k > 0; k--) {
-					nc = ec & 0xF;
-					op[k] = nc > 9	? nc + ('a' - 10)
-							: nc + '0';
-					ec >>= 4;
-				}
-				op += 5;
+				*op++ = ':';
+				op = pack_hex_byte(op, ec >> 8);
+				op = pack_hex_byte(op, ec);
 				len -= 5;
 			} else {
 				*op++ = '?';
@@ -758,9 +753,10 @@ static int fat_ioctl_readdir(struct inode *inode, struct file *filp,
 	return ret;
 }
 
-static int fat_dir_ioctl(struct inode *inode, struct file *filp,
-			 unsigned int cmd, unsigned long arg)
+static long fat_dir_ioctl(struct file *filp, unsigned int cmd,
+			  unsigned long arg)
 {
+	struct inode *inode = filp->f_path.dentry->d_inode;
 	struct __fat_dirent __user *d1 = (struct __fat_dirent __user *)arg;
 	int short_only, both;
 
@@ -774,7 +770,7 @@ static int fat_dir_ioctl(struct inode *inode, struct file *filp,
 		both = 1;
 		break;
 	default:
-		return fat_generic_ioctl(inode, filp, cmd, arg);
+		return fat_generic_ioctl(filp, cmd, arg);
 	}
 
 	if (!access_ok(VERIFY_WRITE, d1, sizeof(struct __fat_dirent[2])))
@@ -814,7 +810,7 @@ static long fat_compat_dir_ioctl(struct file *filp, unsigned cmd,
 		both = 1;
 		break;
 	default:
-		return -ENOIOCTLCMD;
+		return fat_generic_ioctl(filp, cmd, (unsigned long)arg);
 	}
 
 	if (!access_ok(VERIFY_WRITE, d1, sizeof(struct compat_dirent[2])))
@@ -836,7 +832,7 @@ const struct file_operations fat_dir_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= fat_readdir,
-	.ioctl		= fat_dir_ioctl,
+	.unlocked_ioctl	= fat_dir_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= fat_compat_dir_ioctl,
 #endif
diff --git a/fs/fat/fat.h b/fs/fat/fat.h
index e6efdfa0f6db..27ac25725954 100644
--- a/fs/fat/fat.h
+++ b/fs/fat/fat.h
@@ -6,6 +6,7 @@
 #include <linux/nls.h>
 #include <linux/fs.h>
 #include <linux/mutex.h>
+#include <linux/ratelimit.h>
 #include <linux/msdos_fs.h>
 
 /*
@@ -82,6 +83,8 @@ struct msdos_sb_info {
 	struct fatent_operations *fatent_ops;
 	struct inode *fat_inode;
 
+	struct ratelimit_state ratelimit;
+
 	spinlock_t inode_hash_lock;
 	struct hlist_head inode_hashtable[FAT_HASH_SIZE];
 };
@@ -298,16 +301,16 @@ extern int fat_free_clusters(struct inode *inode, int cluster);
 extern int fat_count_free_clusters(struct super_block *sb);
 
 /* fat/file.c */
-extern int fat_generic_ioctl(struct inode *inode, struct file *filp,
-			     unsigned int cmd, unsigned long arg);
+extern long fat_generic_ioctl(struct file *filp, unsigned int cmd,
+			      unsigned long arg);
 extern const struct file_operations fat_file_operations;
 extern const struct inode_operations fat_file_inode_operations;
 extern int fat_setattr(struct dentry * dentry, struct iattr * attr);
-extern void fat_truncate(struct inode *inode);
+extern int fat_setsize(struct inode *inode, loff_t offset);
+extern void fat_truncate_blocks(struct inode *inode, loff_t offset);
 extern int fat_getattr(struct vfsmount *mnt, struct dentry *dentry,
 		       struct kstat *stat);
-extern int fat_file_fsync(struct file *file, struct dentry *dentry,
-			  int datasync);
+extern int fat_file_fsync(struct file *file, int datasync);
 
 /* fat/inode.c */
 extern void fat_attach(struct inode *inode, loff_t i_pos);
@@ -322,8 +325,13 @@ extern int fat_fill_super(struct super_block *sb, void *data, int silent,
 extern int fat_flush_inodes(struct super_block *sb, struct inode *i1,
 		            struct inode *i2);
 /* fat/misc.c */
-extern void fat_fs_error(struct super_block *s, const char *fmt, ...)
-	__attribute__ ((format (printf, 2, 3))) __cold;
+extern void
+__fat_fs_error(struct super_block *s, int report, const char *fmt, ...)
+	__attribute__ ((format (printf, 3, 4))) __cold;
+#define fat_fs_error(s, fmt, args...)		\
+	__fat_fs_error(s, 1, fmt , ## args)
+#define fat_fs_error_ratelimit(s, fmt, args...) \
+	__fat_fs_error(s, __ratelimit(&MSDOS_SB(s)->ratelimit), fmt , ## args)
 extern int fat_clusters_flush(struct super_block *sb);
 extern int fat_chain_add(struct inode *inode, int new_dclus, int nr_cluster);
 extern void fat_time_fat2unix(struct msdos_sb_info *sbi, struct timespec *ts,
diff --git a/fs/fat/file.c b/fs/fat/file.c
index e8c159de236b..990dfae022e5 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -8,6 +8,7 @@
 
 #include <linux/capability.h>
 #include <linux/module.h>
+#include <linux/compat.h>
 #include <linux/mount.h>
 #include <linux/time.h>
 #include <linux/buffer_head.h>
@@ -114,9 +115,9 @@ out:
 	return err;
 }
 
-int fat_generic_ioctl(struct inode *inode, struct file *filp,
-		      unsigned int cmd, unsigned long arg)
+long fat_generic_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
+	struct inode *inode = filp->f_path.dentry->d_inode;
 	u32 __user *user_attr = (u32 __user *)arg;
 
 	switch (cmd) {
@@ -129,6 +130,15 @@ int fat_generic_ioctl(struct inode *inode, struct file *filp,
 	}
 }
 
+#ifdef CONFIG_COMPAT
+static long fat_generic_compat_ioctl(struct file *filp, unsigned int cmd,
+				      unsigned long arg)
+
+{
+	return fat_generic_ioctl(filp, cmd, (unsigned long)compat_ptr(arg));
+}
+#endif
+
 static int fat_file_release(struct inode *inode, struct file *filp)
 {
 	if ((filp->f_mode & FMODE_WRITE) &&
@@ -139,12 +149,12 @@ static int fat_file_release(struct inode *inode, struct file *filp)
 	return 0;
 }
 
-int fat_file_fsync(struct file *filp, struct dentry *dentry, int datasync)
+int fat_file_fsync(struct file *filp, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = filp->f_mapping->host;
 	int res, err;
 
-	res = simple_fsync(filp, dentry, datasync);
+	res = generic_file_fsync(filp, datasync);
 	err = sync_mapping_buffers(MSDOS_SB(inode->i_sb)->fat_inode->i_mapping);
 
 	return res ? res : err;
@@ -159,7 +169,10 @@ const struct file_operations fat_file_operations = {
 	.aio_write	= generic_file_aio_write,
 	.mmap		= generic_file_mmap,
 	.release	= fat_file_release,
-	.ioctl		= fat_generic_ioctl,
+	.unlocked_ioctl	= fat_generic_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl	= fat_generic_compat_ioctl,
+#endif
 	.fsync		= fat_file_fsync,
 	.splice_read	= generic_file_splice_read,
 };
@@ -270,7 +283,7 @@ static int fat_free(struct inode *inode, int skip)
 	return fat_free_clusters(inode, free_start);
 }
 
-void fat_truncate(struct inode *inode)
+void fat_truncate_blocks(struct inode *inode, loff_t offset)
 {
 	struct msdos_sb_info *sbi = MSDOS_SB(inode->i_sb);
 	const unsigned int cluster_size = sbi->cluster_size;
@@ -280,10 +293,10 @@ void fat_truncate(struct inode *inode)
 	 * This protects against truncating a file bigger than it was then
 	 * trying to write into the hole.
 	 */
-	if (MSDOS_I(inode)->mmu_private > inode->i_size)
-		MSDOS_I(inode)->mmu_private = inode->i_size;
+	if (MSDOS_I(inode)->mmu_private > offset)
+		MSDOS_I(inode)->mmu_private = offset;
 
-	nr_clusters = (inode->i_size + (cluster_size - 1)) >> sbi->cluster_bits;
+	nr_clusters = (offset + (cluster_size - 1)) >> sbi->cluster_bits;
 
 	fat_free(inode, nr_clusters);
 	fat_flush_inodes(inode->i_sb, inode, NULL);
@@ -351,6 +364,18 @@ static int fat_allow_set_time(struct msdos_sb_info *sbi, struct inode *inode)
 	return 0;
 }
 
+int fat_setsize(struct inode *inode, loff_t offset)
+{
+	int error;
+
+	error = simple_setsize(inode, offset);
+	if (error)
+		return error;
+	fat_truncate_blocks(inode, offset);
+
+	return error;
+}
+
 #define TIMES_SET_FLAGS	(ATTR_MTIME_SET | ATTR_ATIME_SET | ATTR_TIMES_SET)
 /* valid file mode bits */
 #define FAT_VALID_MODE	(S_IFREG | S_IFDIR | S_IRWXUGO)
@@ -365,7 +390,8 @@ int fat_setattr(struct dentry *dentry, struct iattr *attr)
 	/*
 	 * Expand the file. Since inode_setattr() updates ->i_size
 	 * before calling the ->truncate(), but FAT needs to fill the
-	 * hole before it.
+	 * hole before it. XXX: this is no longer true with new truncate
+	 * sequence.
 	 */
 	if (attr->ia_valid & ATTR_SIZE) {
 		if (attr->ia_size > inode->i_size) {
@@ -414,15 +440,20 @@ int fat_setattr(struct dentry *dentry, struct iattr *attr)
 			attr->ia_valid &= ~ATTR_MODE;
 	}
 
-	if (attr->ia_valid)
-		error = inode_setattr(inode, attr);
+	if (attr->ia_valid & ATTR_SIZE) {
+		error = fat_setsize(inode, attr->ia_size);
+		if (error)
+			goto out;
+	}
+
+	generic_setattr(inode, attr);
+	mark_inode_dirty(inode);
 out:
 	return error;
 }
 EXPORT_SYMBOL_GPL(fat_setattr);
 
 const struct inode_operations fat_file_inode_operations = {
-	.truncate	= fat_truncate,
 	.setattr	= fat_setattr,
 	.getattr	= fat_getattr,
 };
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index 0ce143bd7d56..7bf45aee56d7 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -142,14 +142,29 @@ static int fat_readpages(struct file *file, struct address_space *mapping,
 	return mpage_readpages(mapping, pages, nr_pages, fat_get_block);
 }
 
+static void fat_write_failed(struct address_space *mapping, loff_t to)
+{
+	struct inode *inode = mapping->host;
+
+	if (to > inode->i_size) {
+		truncate_pagecache(inode, to, inode->i_size);
+		fat_truncate_blocks(inode, inode->i_size);
+	}
+}
+
 static int fat_write_begin(struct file *file, struct address_space *mapping,
 			loff_t pos, unsigned len, unsigned flags,
 			struct page **pagep, void **fsdata)
 {
+	int err;
+
 	*pagep = NULL;
-	return cont_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
-				fat_get_block,
+	err = cont_write_begin_newtrunc(file, mapping, pos, len, flags,
+				pagep, fsdata, fat_get_block,
 				&MSDOS_I(mapping->host)->mmu_private);
+	if (err < 0)
+		fat_write_failed(mapping, pos + len);
+	return err;
 }
 
 static int fat_write_end(struct file *file, struct address_space *mapping,
@@ -159,6 +174,8 @@ static int fat_write_end(struct file *file, struct address_space *mapping,
 	struct inode *inode = mapping->host;
 	int err;
 	err = generic_write_end(file, mapping, pos, len, copied, pagep, fsdata);
+	if (err < len)
+		fat_write_failed(mapping, pos + len);
 	if (!(err < 0) && !(MSDOS_I(inode)->i_attrs & ATTR_ARCH)) {
 		inode->i_mtime = inode->i_ctime = CURRENT_TIME_SEC;
 		MSDOS_I(inode)->i_attrs |= ATTR_ARCH;
@@ -172,7 +189,9 @@ static ssize_t fat_direct_IO(int rw, struct kiocb *iocb,
 			     loff_t offset, unsigned long nr_segs)
 {
 	struct file *file = iocb->ki_filp;
-	struct inode *inode = file->f_mapping->host;
+	struct address_space *mapping = file->f_mapping;
+	struct inode *inode = mapping->host;
+	ssize_t ret;
 
 	if (rw == WRITE) {
 		/*
@@ -193,8 +212,12 @@ static ssize_t fat_direct_IO(int rw, struct kiocb *iocb,
 	 * FAT need to use the DIO_LOCKING for avoiding the race
 	 * condition of fat_get_block() and ->truncate().
 	 */
-	return blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
-				  offset, nr_segs, fat_get_block, NULL);
+	ret = blockdev_direct_IO_newtrunc(rw, iocb, inode, inode->i_sb->s_bdev,
+				iov, offset, nr_segs, fat_get_block, NULL);
+	if (ret < 0 && (rw & WRITE))
+		fat_write_failed(mapping, offset + iov_length(iov, nr_segs));
+
+	return ret;
 }
 
 static sector_t _fat_bmap(struct address_space *mapping, sector_t block)
@@ -429,7 +452,7 @@ static void fat_delete_inode(struct inode *inode)
 {
 	truncate_inode_pages(&inode->i_data, 0);
 	inode->i_size = 0;
-	fat_truncate(inode);
+	fat_truncate_blocks(inode, 0);
 	clear_inode(inode);
 }
 
@@ -1250,6 +1273,8 @@ int fat_fill_super(struct super_block *sb, void *data, int silent,
 	sb->s_op = &fat_sops;
 	sb->s_export_op = &fat_export_ops;
 	sbi->dir_ops = fs_dir_inode_ops;
+	ratelimit_state_init(&sbi->ratelimit, DEFAULT_RATELIMIT_INTERVAL,
+			     DEFAULT_RATELIMIT_BURST);
 
 	error = parse_options(data, isvfat, silent, &debug, &sbi->options);
 	if (error)
@@ -1497,10 +1522,8 @@ out_fail:
 		iput(fat_inode);
 	if (root_inode)
 		iput(root_inode);
-	if (sbi->nls_io)
-		unload_nls(sbi->nls_io);
-	if (sbi->nls_disk)
-		unload_nls(sbi->nls_disk);
+	unload_nls(sbi->nls_io);
+	unload_nls(sbi->nls_disk);
 	if (sbi->options.iocharset != fat_default_iocharset)
 		kfree(sbi->options.iocharset);
 	sb->s_fs_info = NULL;
diff --git a/fs/fat/misc.c b/fs/fat/misc.c
index d3da05f26465..1fa23f6ffba5 100644
--- a/fs/fat/misc.c
+++ b/fs/fat/misc.c
@@ -20,27 +20,29 @@
  * In case the file system is remounted read-only, it can be made writable
  * again by remounting it.
  */
-void fat_fs_error(struct super_block *s, const char *fmt, ...)
+void __fat_fs_error(struct super_block *s, int report, const char *fmt, ...)
 {
 	struct fat_mount_options *opts = &MSDOS_SB(s)->options;
 	va_list args;
 
-	printk(KERN_ERR "FAT: Filesystem error (dev %s)\n", s->s_id);
+	if (report) {
+		printk(KERN_ERR "FAT: Filesystem error (dev %s)\n", s->s_id);
 
-	printk(KERN_ERR "    ");
-	va_start(args, fmt);
-	vprintk(fmt, args);
-	va_end(args);
-	printk("\n");
+		printk(KERN_ERR "    ");
+		va_start(args, fmt);
+		vprintk(fmt, args);
+		va_end(args);
+		printk("\n");
+	}
 
 	if (opts->errors == FAT_ERRORS_PANIC)
-		panic("    FAT fs panic from previous error\n");
+		panic("FAT: fs panic from previous error\n");
 	else if (opts->errors == FAT_ERRORS_RO && !(s->s_flags & MS_RDONLY)) {
 		s->s_flags |= MS_RDONLY;
-		printk(KERN_ERR "    File system has been set read-only\n");
+		printk(KERN_ERR "FAT: Filesystem has been set read-only\n");
 	}
 }
-EXPORT_SYMBOL_GPL(fat_fs_error);
+EXPORT_SYMBOL_GPL(__fat_fs_error);
 
 /* Flushes the number of free clusters on FAT32 */
 /* XXX: Need to write one per FSINFO block.  Currently only writes 1 */
diff --git a/fs/file_table.c b/fs/file_table.c
index 32d12b78bac8..5c7d10ead4ad 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -194,14 +194,6 @@ struct file *alloc_file(struct path *path, fmode_t mode,
 }
 EXPORT_SYMBOL(alloc_file);
 
-void fput(struct file *file)
-{
-	if (atomic_long_dec_and_test(&file->f_count))
-		__fput(file);
-}
-
-EXPORT_SYMBOL(fput);
-
 /**
  * drop_file_write_access - give up ability to write to a file
  * @file: the file to which we will stop writing
@@ -227,10 +219,9 @@ void drop_file_write_access(struct file *file)
 }
 EXPORT_SYMBOL_GPL(drop_file_write_access);
 
-/* __fput is called from task context when aio completion releases the last
- * last use of a struct file *.  Do not use otherwise.
+/* the real guts of fput() - releasing the last reference to file
  */
-void __fput(struct file *file)
+static void __fput(struct file *file)
 {
 	struct dentry *dentry = file->f_path.dentry;
 	struct vfsmount *mnt = file->f_path.mnt;
@@ -268,6 +259,14 @@ void __fput(struct file *file)
 	mntput(mnt);
 }
 
+void fput(struct file *file)
+{
+	if (atomic_long_dec_and_test(&file->f_count))
+		__fput(file);
+}
+
+EXPORT_SYMBOL(fput);
+
 struct file *fget(unsigned int fd)
 {
 	struct file *file;
diff --git a/fs/freevxfs/vxfs_lookup.c b/fs/freevxfs/vxfs_lookup.c
index aee049cb9f84..0ec7bb2c95c6 100644
--- a/fs/freevxfs/vxfs_lookup.c
+++ b/fs/freevxfs/vxfs_lookup.c
@@ -57,6 +57,8 @@ const struct inode_operations vxfs_dir_inode_ops = {
 };
 
 const struct file_operations vxfs_dir_operations = {
+	.llseek =		generic_file_llseek,
+	.read =			generic_read_dir,
 	.readdir =		vxfs_readdir,
 };
 
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 408a7877b79d..1d1088f48bc2 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -398,11 +398,11 @@ static void inode_wait_for_writeback(struct inode *inode)
 	wait_queue_head_t *wqh;
 
 	wqh = bit_waitqueue(&inode->i_state, __I_SYNC);
-	do {
+	 while (inode->i_state & I_SYNC) {
 		spin_unlock(&inode_lock);
 		__wait_on_bit(wqh, &wq, inode_wait, TASK_UNINTERRUPTIBLE);
 		spin_lock(&inode_lock);
-	} while (inode->i_state & I_SYNC);
+	}
 }
 
 /*
diff --git a/fs/fscache/object-list.c b/fs/fscache/object-list.c
index 1e1f286dd70e..4a8eb31c5338 100644
--- a/fs/fscache/object-list.c
+++ b/fs/fscache/object-list.c
@@ -103,7 +103,7 @@ static struct fscache_object *fscache_objlist_lookup(loff_t *_pos)
 	/* banners (can't represent line 0 by pos 0 as that would involve
 	 * returning a NULL pointer) */
 	if (pos == 0)
-		return (struct fscache_object *) ++(*_pos);
+		return (struct fscache_object *)(long)++(*_pos);
 	if (pos < 3)
 		return (struct fscache_object *)pos;
 
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index eb7e9423691f..9424796d6634 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -16,8 +16,12 @@
 #include <linux/pagemap.h>
 #include <linux/file.h>
 #include <linux/slab.h>
+#include <linux/pipe_fs_i.h>
+#include <linux/swap.h>
+#include <linux/splice.h>
 
 MODULE_ALIAS_MISCDEV(FUSE_MINOR);
+MODULE_ALIAS("devname:fuse");
 
 static struct kmem_cache *fuse_req_cachep;
 
@@ -498,6 +502,9 @@ struct fuse_copy_state {
 	int write;
 	struct fuse_req *req;
 	const struct iovec *iov;
+	struct pipe_buffer *pipebufs;
+	struct pipe_buffer *currbuf;
+	struct pipe_inode_info *pipe;
 	unsigned long nr_segs;
 	unsigned long seglen;
 	unsigned long addr;
@@ -505,16 +512,16 @@ struct fuse_copy_state {
 	void *mapaddr;
 	void *buf;
 	unsigned len;
+	unsigned move_pages:1;
 };
 
 static void fuse_copy_init(struct fuse_copy_state *cs, struct fuse_conn *fc,
-			   int write, struct fuse_req *req,
+			   int write,
 			   const struct iovec *iov, unsigned long nr_segs)
 {
 	memset(cs, 0, sizeof(*cs));
 	cs->fc = fc;
 	cs->write = write;
-	cs->req = req;
 	cs->iov = iov;
 	cs->nr_segs = nr_segs;
 }
@@ -522,7 +529,18 @@ static void fuse_copy_init(struct fuse_copy_state *cs, struct fuse_conn *fc,
 /* Unmap and put previous page of userspace buffer */
 static void fuse_copy_finish(struct fuse_copy_state *cs)
 {
-	if (cs->mapaddr) {
+	if (cs->currbuf) {
+		struct pipe_buffer *buf = cs->currbuf;
+
+		if (!cs->write) {
+			buf->ops->unmap(cs->pipe, buf, cs->mapaddr);
+		} else {
+			kunmap_atomic(cs->mapaddr, KM_USER0);
+			buf->len = PAGE_SIZE - cs->len;
+		}
+		cs->currbuf = NULL;
+		cs->mapaddr = NULL;
+	} else if (cs->mapaddr) {
 		kunmap_atomic(cs->mapaddr, KM_USER0);
 		if (cs->write) {
 			flush_dcache_page(cs->pg);
@@ -544,26 +562,61 @@ static int fuse_copy_fill(struct fuse_copy_state *cs)
 
 	unlock_request(cs->fc, cs->req);
 	fuse_copy_finish(cs);
-	if (!cs->seglen) {
-		BUG_ON(!cs->nr_segs);
-		cs->seglen = cs->iov[0].iov_len;
-		cs->addr = (unsigned long) cs->iov[0].iov_base;
-		cs->iov++;
-		cs->nr_segs--;
+	if (cs->pipebufs) {
+		struct pipe_buffer *buf = cs->pipebufs;
+
+		if (!cs->write) {
+			err = buf->ops->confirm(cs->pipe, buf);
+			if (err)
+				return err;
+
+			BUG_ON(!cs->nr_segs);
+			cs->currbuf = buf;
+			cs->mapaddr = buf->ops->map(cs->pipe, buf, 1);
+			cs->len = buf->len;
+			cs->buf = cs->mapaddr + buf->offset;
+			cs->pipebufs++;
+			cs->nr_segs--;
+		} else {
+			struct page *page;
+
+			if (cs->nr_segs == cs->pipe->buffers)
+				return -EIO;
+
+			page = alloc_page(GFP_HIGHUSER);
+			if (!page)
+				return -ENOMEM;
+
+			buf->page = page;
+			buf->offset = 0;
+			buf->len = 0;
+
+			cs->currbuf = buf;
+			cs->mapaddr = kmap_atomic(page, KM_USER0);
+			cs->buf = cs->mapaddr;
+			cs->len = PAGE_SIZE;
+			cs->pipebufs++;
+			cs->nr_segs++;
+		}
+	} else {
+		if (!cs->seglen) {
+			BUG_ON(!cs->nr_segs);
+			cs->seglen = cs->iov[0].iov_len;
+			cs->addr = (unsigned long) cs->iov[0].iov_base;
+			cs->iov++;
+			cs->nr_segs--;
+		}
+		err = get_user_pages_fast(cs->addr, 1, cs->write, &cs->pg);
+		if (err < 0)
+			return err;
+		BUG_ON(err != 1);
+		offset = cs->addr % PAGE_SIZE;
+		cs->mapaddr = kmap_atomic(cs->pg, KM_USER0);
+		cs->buf = cs->mapaddr + offset;
+		cs->len = min(PAGE_SIZE - offset, cs->seglen);
+		cs->seglen -= cs->len;
+		cs->addr += cs->len;
 	}
-	down_read(&current->mm->mmap_sem);
-	err = get_user_pages(current, current->mm, cs->addr, 1, cs->write, 0,
-			     &cs->pg, NULL);
-	up_read(&current->mm->mmap_sem);
-	if (err < 0)
-		return err;
-	BUG_ON(err != 1);
-	offset = cs->addr % PAGE_SIZE;
-	cs->mapaddr = kmap_atomic(cs->pg, KM_USER0);
-	cs->buf = cs->mapaddr + offset;
-	cs->len = min(PAGE_SIZE - offset, cs->seglen);
-	cs->seglen -= cs->len;
-	cs->addr += cs->len;
 
 	return lock_request(cs->fc, cs->req);
 }
@@ -585,23 +638,178 @@ static int fuse_copy_do(struct fuse_copy_state *cs, void **val, unsigned *size)
 	return ncpy;
 }
 
+static int fuse_check_page(struct page *page)
+{
+	if (page_mapcount(page) ||
+	    page->mapping != NULL ||
+	    page_count(page) != 1 ||
+	    (page->flags & PAGE_FLAGS_CHECK_AT_PREP &
+	     ~(1 << PG_locked |
+	       1 << PG_referenced |
+	       1 << PG_uptodate |
+	       1 << PG_lru |
+	       1 << PG_active |
+	       1 << PG_reclaim))) {
+		printk(KERN_WARNING "fuse: trying to steal weird page\n");
+		printk(KERN_WARNING "  page=%p index=%li flags=%08lx, count=%i, mapcount=%i, mapping=%p\n", page, page->index, page->flags, page_count(page), page_mapcount(page), page->mapping);
+		return 1;
+	}
+	return 0;
+}
+
+static int fuse_try_move_page(struct fuse_copy_state *cs, struct page **pagep)
+{
+	int err;
+	struct page *oldpage = *pagep;
+	struct page *newpage;
+	struct pipe_buffer *buf = cs->pipebufs;
+	struct address_space *mapping;
+	pgoff_t index;
+
+	unlock_request(cs->fc, cs->req);
+	fuse_copy_finish(cs);
+
+	err = buf->ops->confirm(cs->pipe, buf);
+	if (err)
+		return err;
+
+	BUG_ON(!cs->nr_segs);
+	cs->currbuf = buf;
+	cs->len = buf->len;
+	cs->pipebufs++;
+	cs->nr_segs--;
+
+	if (cs->len != PAGE_SIZE)
+		goto out_fallback;
+
+	if (buf->ops->steal(cs->pipe, buf) != 0)
+		goto out_fallback;
+
+	newpage = buf->page;
+
+	if (WARN_ON(!PageUptodate(newpage)))
+		return -EIO;
+
+	ClearPageMappedToDisk(newpage);
+
+	if (fuse_check_page(newpage) != 0)
+		goto out_fallback_unlock;
+
+	mapping = oldpage->mapping;
+	index = oldpage->index;
+
+	/*
+	 * This is a new and locked page, it shouldn't be mapped or
+	 * have any special flags on it
+	 */
+	if (WARN_ON(page_mapped(oldpage)))
+		goto out_fallback_unlock;
+	if (WARN_ON(page_has_private(oldpage)))
+		goto out_fallback_unlock;
+	if (WARN_ON(PageDirty(oldpage) || PageWriteback(oldpage)))
+		goto out_fallback_unlock;
+	if (WARN_ON(PageMlocked(oldpage)))
+		goto out_fallback_unlock;
+
+	remove_from_page_cache(oldpage);
+	page_cache_release(oldpage);
+
+	err = add_to_page_cache_locked(newpage, mapping, index, GFP_KERNEL);
+	if (err) {
+		printk(KERN_WARNING "fuse_try_move_page: failed to add page");
+		goto out_fallback_unlock;
+	}
+	page_cache_get(newpage);
+
+	if (!(buf->flags & PIPE_BUF_FLAG_LRU))
+		lru_cache_add_file(newpage);
+
+	err = 0;
+	spin_lock(&cs->fc->lock);
+	if (cs->req->aborted)
+		err = -ENOENT;
+	else
+		*pagep = newpage;
+	spin_unlock(&cs->fc->lock);
+
+	if (err) {
+		unlock_page(newpage);
+		page_cache_release(newpage);
+		return err;
+	}
+
+	unlock_page(oldpage);
+	page_cache_release(oldpage);
+	cs->len = 0;
+
+	return 0;
+
+out_fallback_unlock:
+	unlock_page(newpage);
+out_fallback:
+	cs->mapaddr = buf->ops->map(cs->pipe, buf, 1);
+	cs->buf = cs->mapaddr + buf->offset;
+
+	err = lock_request(cs->fc, cs->req);
+	if (err)
+		return err;
+
+	return 1;
+}
+
+static int fuse_ref_page(struct fuse_copy_state *cs, struct page *page,
+			 unsigned offset, unsigned count)
+{
+	struct pipe_buffer *buf;
+
+	if (cs->nr_segs == cs->pipe->buffers)
+		return -EIO;
+
+	unlock_request(cs->fc, cs->req);
+	fuse_copy_finish(cs);
+
+	buf = cs->pipebufs;
+	page_cache_get(page);
+	buf->page = page;
+	buf->offset = offset;
+	buf->len = count;
+
+	cs->pipebufs++;
+	cs->nr_segs++;
+	cs->len = 0;
+
+	return 0;
+}
+
 /*
  * Copy a page in the request to/from the userspace buffer.  Must be
  * done atomically
  */
-static int fuse_copy_page(struct fuse_copy_state *cs, struct page *page,
+static int fuse_copy_page(struct fuse_copy_state *cs, struct page **pagep,
 			  unsigned offset, unsigned count, int zeroing)
 {
+	int err;
+	struct page *page = *pagep;
+
 	if (page && zeroing && count < PAGE_SIZE) {
 		void *mapaddr = kmap_atomic(page, KM_USER1);
 		memset(mapaddr, 0, PAGE_SIZE);
 		kunmap_atomic(mapaddr, KM_USER1);
 	}
 	while (count) {
-		if (!cs->len) {
-			int err = fuse_copy_fill(cs);
-			if (err)
-				return err;
+		if (cs->write && cs->pipebufs && page) {
+			return fuse_ref_page(cs, page, offset, count);
+		} else if (!cs->len) {
+			if (cs->move_pages && page &&
+			    offset == 0 && count == PAGE_SIZE) {
+				err = fuse_try_move_page(cs, pagep);
+				if (err <= 0)
+					return err;
+			} else {
+				err = fuse_copy_fill(cs);
+				if (err)
+					return err;
+			}
 		}
 		if (page) {
 			void *mapaddr = kmap_atomic(page, KM_USER1);
@@ -626,8 +834,10 @@ static int fuse_copy_pages(struct fuse_copy_state *cs, unsigned nbytes,
 	unsigned count = min(nbytes, (unsigned) PAGE_SIZE - offset);
 
 	for (i = 0; i < req->num_pages && (nbytes || zeroing); i++) {
-		struct page *page = req->pages[i];
-		int err = fuse_copy_page(cs, page, offset, count, zeroing);
+		int err;
+
+		err = fuse_copy_page(cs, &req->pages[i], offset, count,
+				     zeroing);
 		if (err)
 			return err;
 
@@ -704,11 +914,10 @@ __acquires(&fc->lock)
  *
  * Called with fc->lock held, releases it
  */
-static int fuse_read_interrupt(struct fuse_conn *fc, struct fuse_req *req,
-			       const struct iovec *iov, unsigned long nr_segs)
+static int fuse_read_interrupt(struct fuse_conn *fc, struct fuse_copy_state *cs,
+			       size_t nbytes, struct fuse_req *req)
 __releases(&fc->lock)
 {
-	struct fuse_copy_state cs;
 	struct fuse_in_header ih;
 	struct fuse_interrupt_in arg;
 	unsigned reqsize = sizeof(ih) + sizeof(arg);
@@ -724,14 +933,13 @@ __releases(&fc->lock)
 	arg.unique = req->in.h.unique;
 
 	spin_unlock(&fc->lock);
-	if (iov_length(iov, nr_segs) < reqsize)
+	if (nbytes < reqsize)
 		return -EINVAL;
 
-	fuse_copy_init(&cs, fc, 1, NULL, iov, nr_segs);
-	err = fuse_copy_one(&cs, &ih, sizeof(ih));
+	err = fuse_copy_one(cs, &ih, sizeof(ih));
 	if (!err)
-		err = fuse_copy_one(&cs, &arg, sizeof(arg));
-	fuse_copy_finish(&cs);
+		err = fuse_copy_one(cs, &arg, sizeof(arg));
+	fuse_copy_finish(cs);
 
 	return err ? err : reqsize;
 }
@@ -745,18 +953,13 @@ __releases(&fc->lock)
  * request_end().  Otherwise add it to the processing list, and set
  * the 'sent' flag.
  */
-static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
-			      unsigned long nr_segs, loff_t pos)
+static ssize_t fuse_dev_do_read(struct fuse_conn *fc, struct file *file,
+				struct fuse_copy_state *cs, size_t nbytes)
 {
 	int err;
 	struct fuse_req *req;
 	struct fuse_in *in;
-	struct fuse_copy_state cs;
 	unsigned reqsize;
-	struct file *file = iocb->ki_filp;
-	struct fuse_conn *fc = fuse_get_conn(file);
-	if (!fc)
-		return -EPERM;
 
  restart:
 	spin_lock(&fc->lock);
@@ -776,7 +979,7 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
 	if (!list_empty(&fc->interrupts)) {
 		req = list_entry(fc->interrupts.next, struct fuse_req,
 				 intr_entry);
-		return fuse_read_interrupt(fc, req, iov, nr_segs);
+		return fuse_read_interrupt(fc, cs, nbytes, req);
 	}
 
 	req = list_entry(fc->pending.next, struct fuse_req, list);
@@ -786,7 +989,7 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
 	in = &req->in;
 	reqsize = in->h.len;
 	/* If request is too large, reply with an error and restart the read */
-	if (iov_length(iov, nr_segs) < reqsize) {
+	if (nbytes < reqsize) {
 		req->out.h.error = -EIO;
 		/* SETXATTR is special, since it may contain too large data */
 		if (in->h.opcode == FUSE_SETXATTR)
@@ -795,12 +998,12 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
 		goto restart;
 	}
 	spin_unlock(&fc->lock);
-	fuse_copy_init(&cs, fc, 1, req, iov, nr_segs);
-	err = fuse_copy_one(&cs, &in->h, sizeof(in->h));
+	cs->req = req;
+	err = fuse_copy_one(cs, &in->h, sizeof(in->h));
 	if (!err)
-		err = fuse_copy_args(&cs, in->numargs, in->argpages,
+		err = fuse_copy_args(cs, in->numargs, in->argpages,
 				     (struct fuse_arg *) in->args, 0);
-	fuse_copy_finish(&cs);
+	fuse_copy_finish(cs);
 	spin_lock(&fc->lock);
 	req->locked = 0;
 	if (req->aborted) {
@@ -828,6 +1031,110 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
 	return err;
 }
 
+static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
+			      unsigned long nr_segs, loff_t pos)
+{
+	struct fuse_copy_state cs;
+	struct file *file = iocb->ki_filp;
+	struct fuse_conn *fc = fuse_get_conn(file);
+	if (!fc)
+		return -EPERM;
+
+	fuse_copy_init(&cs, fc, 1, iov, nr_segs);
+
+	return fuse_dev_do_read(fc, file, &cs, iov_length(iov, nr_segs));
+}
+
+static int fuse_dev_pipe_buf_steal(struct pipe_inode_info *pipe,
+				   struct pipe_buffer *buf)
+{
+	return 1;
+}
+
+static const struct pipe_buf_operations fuse_dev_pipe_buf_ops = {
+	.can_merge = 0,
+	.map = generic_pipe_buf_map,
+	.unmap = generic_pipe_buf_unmap,
+	.confirm = generic_pipe_buf_confirm,
+	.release = generic_pipe_buf_release,
+	.steal = fuse_dev_pipe_buf_steal,
+	.get = generic_pipe_buf_get,
+};
+
+static ssize_t fuse_dev_splice_read(struct file *in, loff_t *ppos,
+				    struct pipe_inode_info *pipe,
+				    size_t len, unsigned int flags)
+{
+	int ret;
+	int page_nr = 0;
+	int do_wakeup = 0;
+	struct pipe_buffer *bufs;
+	struct fuse_copy_state cs;
+	struct fuse_conn *fc = fuse_get_conn(in);
+	if (!fc)
+		return -EPERM;
+
+	bufs = kmalloc(pipe->buffers * sizeof (struct pipe_buffer), GFP_KERNEL);
+	if (!bufs)
+		return -ENOMEM;
+
+	fuse_copy_init(&cs, fc, 1, NULL, 0);
+	cs.pipebufs = bufs;
+	cs.pipe = pipe;
+	ret = fuse_dev_do_read(fc, in, &cs, len);
+	if (ret < 0)
+		goto out;
+
+	ret = 0;
+	pipe_lock(pipe);
+
+	if (!pipe->readers) {
+		send_sig(SIGPIPE, current, 0);
+		if (!ret)
+			ret = -EPIPE;
+		goto out_unlock;
+	}
+
+	if (pipe->nrbufs + cs.nr_segs > pipe->buffers) {
+		ret = -EIO;
+		goto out_unlock;
+	}
+
+	while (page_nr < cs.nr_segs) {
+		int newbuf = (pipe->curbuf + pipe->nrbufs) & (pipe->buffers - 1);
+		struct pipe_buffer *buf = pipe->bufs + newbuf;
+
+		buf->page = bufs[page_nr].page;
+		buf->offset = bufs[page_nr].offset;
+		buf->len = bufs[page_nr].len;
+		buf->ops = &fuse_dev_pipe_buf_ops;
+
+		pipe->nrbufs++;
+		page_nr++;
+		ret += buf->len;
+
+		if (pipe->inode)
+			do_wakeup = 1;
+	}
+
+out_unlock:
+	pipe_unlock(pipe);
+
+	if (do_wakeup) {
+		smp_mb();
+		if (waitqueue_active(&pipe->wait))
+			wake_up_interruptible(&pipe->wait);
+		kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
+	}
+
+out:
+	for (; page_nr < cs.nr_segs; page_nr++)
+		page_cache_release(bufs[page_nr].page);
+
+	kfree(bufs);
+	return ret;
+}
+
 static int fuse_notify_poll(struct fuse_conn *fc, unsigned int size,
 			    struct fuse_copy_state *cs)
 {
@@ -987,23 +1294,17 @@ static int copy_out_args(struct fuse_copy_state *cs, struct fuse_out *out,
  * it from the list and copy the rest of the buffer to the request.
  * The request is finished by calling request_end()
  */
-static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
-			       unsigned long nr_segs, loff_t pos)
+static ssize_t fuse_dev_do_write(struct fuse_conn *fc,
+				 struct fuse_copy_state *cs, size_t nbytes)
 {
 	int err;
-	size_t nbytes = iov_length(iov, nr_segs);
 	struct fuse_req *req;
 	struct fuse_out_header oh;
-	struct fuse_copy_state cs;
-	struct fuse_conn *fc = fuse_get_conn(iocb->ki_filp);
-	if (!fc)
-		return -EPERM;
 
-	fuse_copy_init(&cs, fc, 0, NULL, iov, nr_segs);
 	if (nbytes < sizeof(struct fuse_out_header))
 		return -EINVAL;
 
-	err = fuse_copy_one(&cs, &oh, sizeof(oh));
+	err = fuse_copy_one(cs, &oh, sizeof(oh));
 	if (err)
 		goto err_finish;
 
@@ -1016,7 +1317,7 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
 	 * and error contains notification code.
 	 */
 	if (!oh.unique) {
-		err = fuse_notify(fc, oh.error, nbytes - sizeof(oh), &cs);
+		err = fuse_notify(fc, oh.error, nbytes - sizeof(oh), cs);
 		return err ? err : nbytes;
 	}
 
@@ -1035,7 +1336,7 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
 
 	if (req->aborted) {
 		spin_unlock(&fc->lock);
-		fuse_copy_finish(&cs);
+		fuse_copy_finish(cs);
 		spin_lock(&fc->lock);
 		request_end(fc, req);
 		return -ENOENT;
@@ -1052,7 +1353,7 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
 			queue_interrupt(fc, req);
 
 		spin_unlock(&fc->lock);
-		fuse_copy_finish(&cs);
+		fuse_copy_finish(cs);
 		return nbytes;
 	}
 
@@ -1060,11 +1361,13 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
 	list_move(&req->list, &fc->io);
 	req->out.h = oh;
 	req->locked = 1;
-	cs.req = req;
+	cs->req = req;
+	if (!req->out.page_replace)
+		cs->move_pages = 0;
 	spin_unlock(&fc->lock);
 
-	err = copy_out_args(&cs, &req->out, nbytes);
-	fuse_copy_finish(&cs);
+	err = copy_out_args(cs, &req->out, nbytes);
+	fuse_copy_finish(cs);
 
 	spin_lock(&fc->lock);
 	req->locked = 0;
@@ -1080,10 +1383,101 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
  err_unlock:
 	spin_unlock(&fc->lock);
  err_finish:
-	fuse_copy_finish(&cs);
+	fuse_copy_finish(cs);
 	return err;
 }
 
+static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
+			      unsigned long nr_segs, loff_t pos)
+{
+	struct fuse_copy_state cs;
+	struct fuse_conn *fc = fuse_get_conn(iocb->ki_filp);
+	if (!fc)
+		return -EPERM;
+
+	fuse_copy_init(&cs, fc, 0, iov, nr_segs);
+
+	return fuse_dev_do_write(fc, &cs, iov_length(iov, nr_segs));
+}
+
+static ssize_t fuse_dev_splice_write(struct pipe_inode_info *pipe,
+				     struct file *out, loff_t *ppos,
+				     size_t len, unsigned int flags)
+{
+	unsigned nbuf;
+	unsigned idx;
+	struct pipe_buffer *bufs;
+	struct fuse_copy_state cs;
+	struct fuse_conn *fc;
+	size_t rem;
+	ssize_t ret;
+
+	fc = fuse_get_conn(out);
+	if (!fc)
+		return -EPERM;
+
+	bufs = kmalloc(pipe->buffers * sizeof (struct pipe_buffer), GFP_KERNEL);
+	if (!bufs)
+		return -ENOMEM;
+
+	pipe_lock(pipe);
+	nbuf = 0;
+	rem = 0;
+	for (idx = 0; idx < pipe->nrbufs && rem < len; idx++)
+		rem += pipe->bufs[(pipe->curbuf + idx) & (pipe->buffers - 1)].len;
+
+	ret = -EINVAL;
+	if (rem < len) {
+		pipe_unlock(pipe);
+		goto out;
+	}
+
+	rem = len;
+	while (rem) {
+		struct pipe_buffer *ibuf;
+		struct pipe_buffer *obuf;
+
+		BUG_ON(nbuf >= pipe->buffers);
+		BUG_ON(!pipe->nrbufs);
+		ibuf = &pipe->bufs[pipe->curbuf];
+		obuf = &bufs[nbuf];
+
+		if (rem >= ibuf->len) {
+			*obuf = *ibuf;
+			ibuf->ops = NULL;
+			pipe->curbuf = (pipe->curbuf + 1) & (pipe->buffers - 1);
+			pipe->nrbufs--;
+		} else {
+			ibuf->ops->get(pipe, ibuf);
+			*obuf = *ibuf;
+			obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
+			obuf->len = rem;
+			ibuf->offset += obuf->len;
+			ibuf->len -= obuf->len;
+		}
+		nbuf++;
+		rem -= obuf->len;
+	}
+	pipe_unlock(pipe);
+
+	fuse_copy_init(&cs, fc, 0, NULL, nbuf);
+	cs.pipebufs = bufs;
+	cs.pipe = pipe;
+
+	if (flags & SPLICE_F_MOVE)
+		cs.move_pages = 1;
+
+	ret = fuse_dev_do_write(fc, &cs, len);
+
+	for (idx = 0; idx < nbuf; idx++) {
+		struct pipe_buffer *buf = &bufs[idx];
+		buf->ops->release(pipe, buf);
+	}
+out:
+	kfree(bufs);
+	return ret;
+}
+
 static unsigned fuse_dev_poll(struct file *file, poll_table *wait)
 {
 	unsigned mask = POLLOUT | POLLWRNORM;
@@ -1225,8 +1619,10 @@ const struct file_operations fuse_dev_operations = {
 	.llseek		= no_llseek,
 	.read		= do_sync_read,
 	.aio_read	= fuse_dev_read,
+	.splice_read	= fuse_dev_splice_read,
 	.write		= do_sync_write,
 	.aio_write	= fuse_dev_write,
+	.splice_write	= fuse_dev_splice_write,
 	.poll		= fuse_dev_poll,
 	.release	= fuse_dev_release,
 	.fasync		= fuse_dev_fasync,
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index 4787ae6c5c1c..3cdc5f78a406 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -1156,10 +1156,9 @@ static int fuse_dir_release(struct inode *inode, struct file *file)
 	return 0;
 }
 
-static int fuse_dir_fsync(struct file *file, struct dentry *de, int datasync)
+static int fuse_dir_fsync(struct file *file, int datasync)
 {
-	/* nfsd can call this with no file */
-	return file ? fuse_fsync_common(file, de, datasync, 1) : 0;
+	return fuse_fsync_common(file, datasync, 1);
 }
 
 static bool update_mtime(unsigned ivalid)
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index a9f5e137f1d3..ada0adeb3bb5 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -351,10 +351,9 @@ static void fuse_sync_writes(struct inode *inode)
 	fuse_release_nowrite(inode);
 }
 
-int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
-		      int isdir)
+int fuse_fsync_common(struct file *file, int datasync, int isdir)
 {
-	struct inode *inode = de->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	struct fuse_conn *fc = get_fuse_conn(inode);
 	struct fuse_file *ff = file->private_data;
 	struct fuse_req *req;
@@ -403,9 +402,9 @@ int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
 	return err;
 }
 
-static int fuse_fsync(struct file *file, struct dentry *de, int datasync)
+static int fuse_fsync(struct file *file, int datasync)
 {
-	return fuse_fsync_common(file, de, datasync, 0);
+	return fuse_fsync_common(file, datasync, 0);
 }
 
 void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos,
@@ -517,17 +516,26 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
 	int i;
 	size_t count = req->misc.read.in.size;
 	size_t num_read = req->out.args[0].size;
-	struct inode *inode = req->pages[0]->mapping->host;
+	struct address_space *mapping = NULL;
 
-	/*
-	 * Short read means EOF.  If file size is larger, truncate it
-	 */
-	if (!req->out.h.error && num_read < count) {
-		loff_t pos = page_offset(req->pages[0]) + num_read;
-		fuse_read_update_size(inode, pos, req->misc.read.attr_ver);
-	}
+	for (i = 0; mapping == NULL && i < req->num_pages; i++)
+		mapping = req->pages[i]->mapping;
 
-	fuse_invalidate_attr(inode); /* atime changed */
+	if (mapping) {
+		struct inode *inode = mapping->host;
+
+		/*
+		 * Short read means EOF. If file size is larger, truncate it
+		 */
+		if (!req->out.h.error && num_read < count) {
+			loff_t pos;
+
+			pos = page_offset(req->pages[0]) + num_read;
+			fuse_read_update_size(inode, pos,
+					      req->misc.read.attr_ver);
+		}
+		fuse_invalidate_attr(inode); /* atime changed */
+	}
 
 	for (i = 0; i < req->num_pages; i++) {
 		struct page *page = req->pages[i];
@@ -536,6 +544,7 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
 		else
 			SetPageError(page);
 		unlock_page(page);
+		page_cache_release(page);
 	}
 	if (req->ff)
 		fuse_file_put(req->ff);
@@ -550,6 +559,7 @@ static void fuse_send_readpages(struct fuse_req *req, struct file *file)
 
 	req->out.argpages = 1;
 	req->out.page_zeroing = 1;
+	req->out.page_replace = 1;
 	fuse_read_fill(req, file, pos, count, FUSE_READ);
 	req->misc.read.attr_ver = fuse_get_attr_version(fc);
 	if (fc->async_read) {
@@ -589,6 +599,7 @@ static int fuse_readpages_fill(void *_data, struct page *page)
 			return PTR_ERR(req);
 		}
 	}
+	page_cache_get(page);
 	req->pages[req->num_pages] = page;
 	req->num_pages++;
 	return 0;
@@ -994,10 +1005,7 @@ static int fuse_get_user_pages(struct fuse_req *req, const char __user *buf,
 	nbytes = min_t(size_t, nbytes, FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT);
 	npages = (nbytes + offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
 	npages = clamp(npages, 1, FUSE_MAX_PAGES_PER_REQ);
-	down_read(&current->mm->mmap_sem);
-	npages = get_user_pages(current, current->mm, user_addr, npages, !write,
-				0, req->pages, NULL);
-	up_read(&current->mm->mmap_sem);
+	npages = get_user_pages_fast(user_addr, npages, !write, req->pages);
 	if (npages < 0)
 		return npages;
 
@@ -1580,9 +1588,9 @@ static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov,
 	while (iov_iter_count(&ii)) {
 		struct page *page = pages[page_idx++];
 		size_t todo = min_t(size_t, PAGE_SIZE, iov_iter_count(&ii));
-		void *kaddr, *map;
+		void *kaddr;
 
-		kaddr = map = kmap(page);
+		kaddr = kmap(page);
 
 		while (todo) {
 			char __user *uaddr = ii.iov->iov_base + ii.iov_offset;
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 01cc462ff45d..8f309f04064e 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -177,6 +177,9 @@ struct fuse_out {
 	/** Zero partially or not copied pages */
 	unsigned page_zeroing:1;
 
+	/** Pages may be replaced with new ones */
+	unsigned page_replace:1;
+
 	/** Number or arguments */
 	unsigned numargs;
 
@@ -568,8 +571,7 @@ void fuse_release_common(struct file *file, int opcode);
 /**
  * Send FSYNC or FSYNCDIR request
  */
-int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
-		      int isdir);
+int fuse_fsync_common(struct file *file, int datasync, int isdir);
 
 /**
  * Notify poll wakeup
diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c
index 9fb76b0a0485..48171f4c943d 100644
--- a/fs/gfs2/acl.c
+++ b/fs/gfs2/acl.c
@@ -236,10 +236,14 @@ static int gfs2_xattr_system_get(struct dentry *dentry, const char *name,
 				 void *buffer, size_t size, int xtype)
 {
 	struct inode *inode = dentry->d_inode;
+	struct gfs2_sbd *sdp = GFS2_SB(inode);
 	struct posix_acl *acl;
 	int type;
 	int error;
 
+	if (!sdp->sd_args.ar_posix_acl)
+		return -EOPNOTSUPP;
+
 	type = gfs2_acl_type(name);
 	if (type < 0)
 		return type;
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index a739a0a48067..9f8b52500d63 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -700,8 +700,14 @@ out:
 		return 0;
 
 	page_cache_release(page);
+
+	/*
+	 * XXX(hch): the call below should probably be replaced with
+	 * a call to the gfs2-specific truncate blocks helper to actually
+	 * release disk blocks..
+	 */
 	if (pos + len > ip->i_inode.i_size)
-		vmtruncate(&ip->i_inode, ip->i_inode.i_size);
+		simple_setsize(&ip->i_inode, ip->i_inode.i_size);
 out_endtrans:
 	gfs2_trans_end(sdp);
 out_trans_fail:
diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index e6dd2aec6f82..ed9a94f0ef15 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -218,6 +218,11 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 	if (error)
 		goto out_drop_write;
 
+	error = -EACCES;
+	if (!is_owner_or_cap(inode))
+		goto out;
+
+	error = 0;
 	flags = ip->i_diskflags;
 	new_flags = (flags & ~mask) | (reqflags & mask);
 	if ((new_flags ^ flags) == 0)
@@ -275,8 +280,10 @@ static int gfs2_set_flags(struct file *filp, u32 __user *ptr)
 {
 	struct inode *inode = filp->f_path.dentry->d_inode;
 	u32 fsflags, gfsflags;
+
 	if (get_user(fsflags, ptr))
 		return -EFAULT;
+
 	gfsflags = fsflags_cvt(fsflags_to_gfs2, fsflags);
 	if (!S_ISDIR(inode->i_mode)) {
 		if (gfsflags & GFS2_DIF_INHERIT_JDATA)
@@ -547,9 +554,9 @@ static int gfs2_close(struct inode *inode, struct file *file)
  * Returns: errno
  */
 
-static int gfs2_fsync(struct file *file, struct dentry *dentry, int datasync)
+static int gfs2_fsync(struct file *file, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	int sync_state = inode->i_state & (I_DIRTY_SYNC|I_DIRTY_DATASYNC);
 	int ret = 0;
 
diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 51d8061fa07a..b5612cbb62a5 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -242,34 +242,38 @@ fail:
 }
 
 /**
- * gfs2_unlinked_inode_lookup - Lookup an unlinked inode for reclamation
+ * gfs2_process_unlinked_inode - Lookup an unlinked inode for reclamation
+ *                               and try to reclaim it by doing iput.
+ *
+ * This function assumes no rgrp locks are currently held.
+ *
  * @sb: The super block
  * no_addr: The inode number
- * @@inode: A pointer to the inode found, if any
  *
- * Returns: 0 and *inode if no errors occurred.  If an error occurs,
- *          the resulting *inode may or may not be NULL.
  */
 
-int gfs2_unlinked_inode_lookup(struct super_block *sb, u64 no_addr,
-			       struct inode **inode)
+void gfs2_process_unlinked_inode(struct super_block *sb, u64 no_addr)
 {
 	struct gfs2_sbd *sdp;
 	struct gfs2_inode *ip;
 	struct gfs2_glock *io_gl;
 	int error;
 	struct gfs2_holder gh;
+	struct inode *inode;
 
-	*inode = gfs2_iget_skip(sb, no_addr);
+	inode = gfs2_iget_skip(sb, no_addr);
 
-	if (!(*inode))
-		return -ENOBUFS;
+	if (!inode)
+		return;
 
-	if (!((*inode)->i_state & I_NEW))
-		return -ENOBUFS;
+	/* If it's not a new inode, someone's using it, so leave it alone. */
+	if (!(inode->i_state & I_NEW)) {
+		iput(inode);
+		return;
+	}
 
-	ip = GFS2_I(*inode);
-	sdp = GFS2_SB(*inode);
+	ip = GFS2_I(inode);
+	sdp = GFS2_SB(inode);
 	ip->i_no_formal_ino = -1;
 
 	error = gfs2_glock_get(sdp, no_addr, &gfs2_inode_glops, CREATE, &ip->i_gl);
@@ -284,15 +288,13 @@ int gfs2_unlinked_inode_lookup(struct super_block *sb, u64 no_addr,
 	set_bit(GIF_INVALID, &ip->i_flags);
 	error = gfs2_glock_nq_init(io_gl, LM_ST_SHARED, LM_FLAG_TRY | GL_EXACT,
 				   &ip->i_iopen_gh);
-	if (unlikely(error)) {
-		if (error == GLR_TRYFAILED)
-			error = 0;
+	if (unlikely(error))
 		goto fail_iopen;
-	}
+
 	ip->i_iopen_gh.gh_gl->gl_object = ip;
 	gfs2_glock_put(io_gl);
 
-	(*inode)->i_mode = DT2IF(DT_UNKNOWN);
+	inode->i_mode = DT2IF(DT_UNKNOWN);
 
 	/*
 	 * We must read the inode in order to work out its type in
@@ -303,16 +305,17 @@ int gfs2_unlinked_inode_lookup(struct super_block *sb, u64 no_addr,
 	 */
 	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, LM_FLAG_TRY,
 				   &gh);
-	if (unlikely(error)) {
-		if (error == GLR_TRYFAILED)
-			error = 0;
+	if (unlikely(error))
 		goto fail_glock;
-	}
+
 	/* Inode is now uptodate */
 	gfs2_glock_dq_uninit(&gh);
-	gfs2_set_iop(*inode);
+	gfs2_set_iop(inode);
+
+	/* The iput will cause it to be deleted. */
+	iput(inode);
+	return;
 
-	return 0;
 fail_glock:
 	gfs2_glock_dq(&ip->i_iopen_gh);
 fail_iopen:
@@ -321,7 +324,8 @@ fail_put:
 	ip->i_gl->gl_object = NULL;
 	gfs2_glock_put(ip->i_gl);
 fail:
-	return error;
+	iget_failed(inode);
+	return;
 }
 
 static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf)
diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h
index e161461d4c57..300ada3f21de 100644
--- a/fs/gfs2/inode.h
+++ b/fs/gfs2/inode.h
@@ -84,8 +84,7 @@ static inline void gfs2_inum_out(const struct gfs2_inode *ip,
 extern void gfs2_set_iop(struct inode *inode);
 extern struct inode *gfs2_inode_lookup(struct super_block *sb, unsigned type, 
 				       u64 no_addr, u64 no_formal_ino);
-extern int gfs2_unlinked_inode_lookup(struct super_block *sb, u64 no_addr,
-				      struct inode **inode);
+extern void gfs2_process_unlinked_inode(struct super_block *sb, u64 no_addr);
 extern struct inode *gfs2_ilookup(struct super_block *sb, u64 no_addr);
 
 extern int gfs2_inode_refresh(struct gfs2_inode *ip);
diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c
index b593f0e28f25..6a857e24f947 100644
--- a/fs/gfs2/log.c
+++ b/fs/gfs2/log.c
@@ -696,7 +696,7 @@ static void gfs2_ordered_wait(struct gfs2_sbd *sdp)
  *
  */
 
-void __gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl)
+void gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl)
 {
 	struct gfs2_ail *ai;
 
diff --git a/fs/gfs2/log.h b/fs/gfs2/log.h
index eb570b4ad443..0d007f920234 100644
--- a/fs/gfs2/log.h
+++ b/fs/gfs2/log.h
@@ -47,28 +47,21 @@ static inline void gfs2_log_pointers_init(struct gfs2_sbd *sdp,
 	sdp->sd_log_head = sdp->sd_log_tail = value;
 }
 
-unsigned int gfs2_struct2blk(struct gfs2_sbd *sdp, unsigned int nstruct,
+extern unsigned int gfs2_struct2blk(struct gfs2_sbd *sdp, unsigned int nstruct,
 			    unsigned int ssize);
 
-int gfs2_log_reserve(struct gfs2_sbd *sdp, unsigned int blks);
-void gfs2_log_incr_head(struct gfs2_sbd *sdp);
+extern int gfs2_log_reserve(struct gfs2_sbd *sdp, unsigned int blks);
+extern void gfs2_log_incr_head(struct gfs2_sbd *sdp);
 
-struct buffer_head *gfs2_log_get_buf(struct gfs2_sbd *sdp);
-struct buffer_head *gfs2_log_fake_buf(struct gfs2_sbd *sdp,
+extern struct buffer_head *gfs2_log_get_buf(struct gfs2_sbd *sdp);
+extern struct buffer_head *gfs2_log_fake_buf(struct gfs2_sbd *sdp,
 				      struct buffer_head *real);
-void __gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl);
+extern void gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl);
+extern void gfs2_log_commit(struct gfs2_sbd *sdp, struct gfs2_trans *trans);
+extern void gfs2_remove_from_ail(struct gfs2_bufdata *bd);
 
-static inline void gfs2_log_flush(struct gfs2_sbd *sbd, struct gfs2_glock *gl)
-{
-	if (!gl || test_bit(GLF_LFLUSH, &gl->gl_flags))
-		__gfs2_log_flush(sbd, gl);
-}
-
-void gfs2_log_commit(struct gfs2_sbd *sdp, struct gfs2_trans *trans);
-void gfs2_remove_from_ail(struct gfs2_bufdata *bd);
-
-void gfs2_log_shutdown(struct gfs2_sbd *sdp);
-void gfs2_meta_syncfs(struct gfs2_sbd *sdp);
-int gfs2_logd(void *data);
+extern void gfs2_log_shutdown(struct gfs2_sbd *sdp);
+extern void gfs2_meta_syncfs(struct gfs2_sbd *sdp);
+extern int gfs2_logd(void *data);
 
 #endif /* __LOG_DOT_H__ */
diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c
index 4e64352d49de..98cdd05f3316 100644
--- a/fs/gfs2/ops_inode.c
+++ b/fs/gfs2/ops_inode.c
@@ -1071,6 +1071,9 @@ int gfs2_permission(struct inode *inode, int mask)
 	return error;
 }
 
+/*
+ * XXX: should be changed to have proper ordering by opencoding simple_setsize
+ */
 static int setattr_size(struct inode *inode, struct iattr *attr)
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
@@ -1081,7 +1084,7 @@ static int setattr_size(struct inode *inode, struct iattr *attr)
 		error = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks);
 		if (error)
 			return error;
-		error = vmtruncate(inode, attr->ia_size);
+		error = simple_setsize(inode, attr->ia_size);
 		gfs2_trans_end(sdp);
 		if (error) 
 			return error;
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index 117fa4171f62..171a744f8e45 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -1192,7 +1192,6 @@ int gfs2_inplace_reserve_i(struct gfs2_inode *ip, char *file, unsigned int line)
 {
 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
 	struct gfs2_alloc *al = ip->i_alloc;
-	struct inode *inode;
 	int error = 0;
 	u64 last_unlinked = NO_BLOCK, unlinked;
 
@@ -1210,22 +1209,27 @@ try_again:
 	if (error)
 		return error;
 
+	/* Find an rgrp suitable for allocation.  If it encounters any unlinked
+	   dinodes along the way, error will equal -EAGAIN and unlinked will
+	   contains it block address. We then need to look up that inode and
+	   try to free it, and try the allocation again. */
 	error = get_local_rgrp(ip, &unlinked, &last_unlinked);
 	if (error) {
 		if (ip != GFS2_I(sdp->sd_rindex))
 			gfs2_glock_dq_uninit(&al->al_ri_gh);
 		if (error != -EAGAIN)
 			return error;
-		error = gfs2_unlinked_inode_lookup(ip->i_inode.i_sb,
-						   unlinked, &inode);
-		if (inode)
-			iput(inode);
+
+		gfs2_process_unlinked_inode(ip->i_inode.i_sb, unlinked);
+		/* regardless of whether or not gfs2_process_unlinked_inode
+		   was successful, we don't want to repeat it again. */
+		last_unlinked = unlinked;
 		gfs2_log_flush(sdp, NULL);
-		if (error == GLR_TRYFAILED)
-			error = 0;
+		error = 0;
+
 		goto try_again;
 	}
-
+	/* no error, so we have the rgrp set in the inode's allocation. */
 	al->al_file = file;
 	al->al_line = line;
 
diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c
index 5f4023678251..764fd1bdca88 100644
--- a/fs/hfsplus/dir.c
+++ b/fs/hfsplus/dir.c
@@ -494,7 +494,7 @@ const struct inode_operations hfsplus_dir_inode_operations = {
 const struct file_operations hfsplus_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= hfsplus_readdir,
-	.ioctl          = hfsplus_ioctl,
+	.unlocked_ioctl = hfsplus_ioctl,
 	.llseek		= generic_file_llseek,
 	.release	= hfsplus_dir_release,
 };
diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h
index 5c10d803d9df..6505c30ad965 100644
--- a/fs/hfsplus/hfsplus_fs.h
+++ b/fs/hfsplus/hfsplus_fs.h
@@ -337,8 +337,7 @@ struct inode *hfsplus_new_inode(struct super_block *, int);
 void hfsplus_delete_inode(struct inode *);
 
 /* ioctl.c */
-int hfsplus_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
-		  unsigned long arg);
+long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
 int hfsplus_setxattr(struct dentry *dentry, const char *name,
 		     const void *value, size_t size, int flags);
 ssize_t hfsplus_getxattr(struct dentry *dentry, const char *name,
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index 1bcf597c0562..9bbb82924a22 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -285,7 +285,7 @@ static const struct file_operations hfsplus_file_operations = {
 	.fsync		= file_fsync,
 	.open		= hfsplus_file_open,
 	.release	= hfsplus_file_release,
-	.ioctl          = hfsplus_ioctl,
+	.unlocked_ioctl = hfsplus_ioctl,
 };
 
 struct inode *hfsplus_new_inode(struct super_block *sb, int mode)
diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c
index f457d2ca51ab..ac405f099026 100644
--- a/fs/hfsplus/ioctl.c
+++ b/fs/hfsplus/ioctl.c
@@ -17,14 +17,16 @@
 #include <linux/mount.h>
 #include <linux/sched.h>
 #include <linux/xattr.h>
+#include <linux/smp_lock.h>
 #include <asm/uaccess.h>
 #include "hfsplus_fs.h"
 
-int hfsplus_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
-		  unsigned long arg)
+long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
+	struct inode *inode = filp->f_path.dentry->d_inode;
 	unsigned int flags;
 
+	lock_kernel();
 	switch (cmd) {
 	case HFSPLUS_IOC_EXT2_GETFLAGS:
 		flags = 0;
@@ -38,8 +40,10 @@ int hfsplus_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
 	case HFSPLUS_IOC_EXT2_SETFLAGS: {
 		int err = 0;
 		err = mnt_want_write(filp->f_path.mnt);
-		if (err)
+		if (err) {
+			unlock_kernel();
 			return err;
+		}
 
 		if (!is_owner_or_cap(inode)) {
 			err = -EACCES;
@@ -85,9 +89,11 @@ int hfsplus_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
 		mark_inode_dirty(inode);
 setflags_out:
 		mnt_drop_write(filp->f_path.mnt);
+		unlock_kernel();
 		return err;
 	}
 	default:
+		unlock_kernel();
 		return -ENOTTY;
 	}
 }
diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c
index 3a029d8f4cf1..87ac1891a185 100644
--- a/fs/hostfs/hostfs_kern.c
+++ b/fs/hostfs/hostfs_kern.c
@@ -411,9 +411,9 @@ int hostfs_file_open(struct inode *ino, struct file *file)
 	return 0;
 }
 
-int hostfs_fsync(struct file *file, struct dentry *dentry, int datasync)
+int hostfs_fsync(struct file *file, int datasync)
 {
-	return fsync_file(HOSTFS_I(dentry->d_inode)->fd, datasync);
+	return fsync_file(HOSTFS_I(file->f_mapping->host)->fd, datasync);
 }
 
 static const struct file_operations hostfs_file_fops = {
diff --git a/fs/hpfs/file.c b/fs/hpfs/file.c
index 3efabff00367..a9ae9bfa752f 100644
--- a/fs/hpfs/file.c
+++ b/fs/hpfs/file.c
@@ -19,9 +19,9 @@ static int hpfs_file_release(struct inode *inode, struct file *file)
 	return 0;
 }
 
-int hpfs_file_fsync(struct file *file, struct dentry *dentry, int datasync)
+int hpfs_file_fsync(struct file *file, int datasync)
 {
-	/*return file_fsync(file, dentry);*/
+	/*return file_fsync(file, datasync);*/
 	return 0; /* Don't fsync :-) */
 }
 
diff --git a/fs/hpfs/hpfs_fn.h b/fs/hpfs/hpfs_fn.h
index 97bf738cd5d6..75f9d4324851 100644
--- a/fs/hpfs/hpfs_fn.h
+++ b/fs/hpfs/hpfs_fn.h
@@ -268,7 +268,7 @@ void hpfs_set_ea(struct inode *, struct fnode *, const char *,
 
 /* file.c */
 
-int hpfs_file_fsync(struct file *, struct dentry *, int);
+int hpfs_file_fsync(struct file *, int);
 extern const struct file_operations hpfs_file_ops;
 extern const struct inode_operations hpfs_file_iops;
 extern const struct address_space_operations hpfs_aops;
diff --git a/fs/hppfs/hppfs.c b/fs/hppfs/hppfs.c
index 2e4dfa8593da..826c3f9d29ac 100644
--- a/fs/hppfs/hppfs.c
+++ b/fs/hppfs/hppfs.c
@@ -587,7 +587,7 @@ static int hppfs_readdir(struct file *file, void *ent, filldir_t filldir)
 	return err;
 }
 
-static int hppfs_fsync(struct file *file, struct dentry *dentry, int datasync)
+static int hppfs_fsync(struct file *file, int datasync)
 {
 	return 0;
 }
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index a0bbd3d1b41a..a4e9a7ec3691 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -688,7 +688,7 @@ static void init_once(void *foo)
 const struct file_operations hugetlbfs_file_operations = {
 	.read			= hugetlbfs_read,
 	.mmap			= hugetlbfs_file_mmap,
-	.fsync			= simple_sync_file,
+	.fsync			= noop_fsync,
 	.get_unmapped_area	= hugetlb_get_unmapped_area,
 };
 
diff --git a/fs/isofs/dir.c b/fs/isofs/dir.c
index b9ab69b3a482..e0aca9a0ac68 100644
--- a/fs/isofs/dir.c
+++ b/fs/isofs/dir.c
@@ -272,6 +272,7 @@ static int isofs_readdir(struct file *filp,
 
 const struct file_operations isofs_dir_operations =
 {
+	.llseek = generic_file_llseek,
 	.read = generic_read_dir,
 	.readdir = isofs_readdir,
 };
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index bfc70f57900f..e214d68620ac 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -1311,7 +1311,6 @@ int jbd2_journal_stop(handle_t *handle)
 	if (handle->h_sync)
 		transaction->t_synchronous_commit = 1;
 	current->journal_info = NULL;
-	spin_lock(&journal->j_state_lock);
 	spin_lock(&transaction->t_handle_lock);
 	transaction->t_outstanding_credits -= handle->h_buffer_credits;
 	transaction->t_updates--;
@@ -1340,8 +1339,7 @@ int jbd2_journal_stop(handle_t *handle)
 		jbd_debug(2, "transaction too old, requesting commit for "
 					"handle %p\n", handle);
 		/* This is non-blocking */
-		__jbd2_log_start_commit(journal, transaction->t_tid);
-		spin_unlock(&journal->j_state_lock);
+		jbd2_log_start_commit(journal, transaction->t_tid);
 
 		/*
 		 * Special case: JBD2_SYNC synchronous updates require us
@@ -1351,7 +1349,6 @@ int jbd2_journal_stop(handle_t *handle)
 			err = jbd2_log_wait_commit(journal, tid);
 	} else {
 		spin_unlock(&transaction->t_handle_lock);
-		spin_unlock(&journal->j_state_lock);
 	}
 
 	lock_map_release(&handle->h_lockdep_map);
diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index e7291c161a19..813497024437 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -26,9 +26,9 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 			struct page **pagep, void **fsdata);
 static int jffs2_readpage (struct file *filp, struct page *pg);
 
-int jffs2_fsync(struct file *filp, struct dentry *dentry, int datasync)
+int jffs2_fsync(struct file *filp, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = filp->f_mapping->host;
 	struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
 
 	/* Trigger GC to flush any pending writes for this inode */
diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c
index 86e0821fc989..8bc2c80ab159 100644
--- a/fs/jffs2/fs.c
+++ b/fs/jffs2/fs.c
@@ -169,13 +169,13 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr)
 	mutex_unlock(&f->sem);
 	jffs2_complete_reservation(c);
 
-	/* We have to do the vmtruncate() without f->sem held, since
+	/* We have to do the simple_setsize() without f->sem held, since
 	   some pages may be locked and waiting for it in readpage().
 	   We are protected from a simultaneous write() extending i_size
 	   back past iattr->ia_size, because do_truncate() holds the
 	   generic inode semaphore. */
 	if (ivalid & ATTR_SIZE && inode->i_size > iattr->ia_size) {
-		vmtruncate(inode, iattr->ia_size);	
+		simple_setsize(inode, iattr->ia_size);
 		inode->i_blocks = (inode->i_size + 511) >> 9;
 	}	
 
diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h
index 035a767f958b..4791aacf3084 100644
--- a/fs/jffs2/os-linux.h
+++ b/fs/jffs2/os-linux.h
@@ -158,7 +158,7 @@ extern const struct inode_operations jffs2_dir_inode_operations;
 extern const struct file_operations jffs2_file_operations;
 extern const struct inode_operations jffs2_file_inode_operations;
 extern const struct address_space_operations jffs2_file_address_operations;
-int jffs2_fsync(struct file *, struct dentry *, int);
+int jffs2_fsync(struct file *, int);
 int jffs2_do_readpage_unlock (struct inode *inode, struct page *pg);
 
 /* ioctl.c */
diff --git a/fs/jfs/file.c b/fs/jfs/file.c
index 85d9ec659225..127263cc8657 100644
--- a/fs/jfs/file.c
+++ b/fs/jfs/file.c
@@ -27,9 +27,9 @@
 #include "jfs_acl.h"
 #include "jfs_debug.h"
 
-int jfs_fsync(struct file *file, struct dentry *dentry, int datasync)
+int jfs_fsync(struct file *file, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	int rc = 0;
 
 	if (!(inode->i_state & I_DIRTY) ||
diff --git a/fs/jfs/jfs_inode.h b/fs/jfs/jfs_inode.h
index 9e6bda30a6e8..11042b1f44b5 100644
--- a/fs/jfs/jfs_inode.h
+++ b/fs/jfs/jfs_inode.h
@@ -21,7 +21,7 @@
 struct fid;
 
 extern struct inode *ialloc(struct inode *, umode_t);
-extern int jfs_fsync(struct file *, struct dentry *, int);
+extern int jfs_fsync(struct file *, int);
 extern long jfs_ioctl(struct file *, unsigned int, unsigned long);
 extern long jfs_compat_ioctl(struct file *, unsigned int, unsigned long);
 extern struct inode *jfs_iget(struct super_block *, unsigned long);
diff --git a/fs/jfs/super.c b/fs/jfs/super.c
index b66832ac33ac..b38f96bef829 100644
--- a/fs/jfs/super.c
+++ b/fs/jfs/super.c
@@ -179,6 +179,8 @@ static void jfs_put_super(struct super_block *sb)
 
 	jfs_info("In jfs_put_super");
 
+	dquot_disable(sb, -1, DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
+
 	lock_kernel();
 
 	rc = jfs_umount(sb);
@@ -396,10 +398,20 @@ static int jfs_remount(struct super_block *sb, int *flags, char *data)
 
 		JFS_SBI(sb)->flag = flag;
 		ret = jfs_mount_rw(sb, 1);
+
+		/* mark the fs r/w for quota activity */
+		sb->s_flags &= ~MS_RDONLY;
+
 		unlock_kernel();
+		dquot_resume(sb, -1);
 		return ret;
 	}
 	if ((!(sb->s_flags & MS_RDONLY)) && (*flags & MS_RDONLY)) {
+		rc = dquot_suspend(sb, -1);
+		if (rc < 0) {
+			unlock_kernel();
+			return rc;
+		}
 		rc = jfs_umount_rw(sb);
 		JFS_SBI(sb)->flag = flag;
 		unlock_kernel();
@@ -469,6 +481,10 @@ static int jfs_fill_super(struct super_block *sb, void *data, int silent)
 	 */
 	sb->s_op = &jfs_super_operations;
 	sb->s_export_op = &jfs_export_operations;
+#ifdef CONFIG_QUOTA
+	sb->dq_op = &dquot_operations;
+	sb->s_qcop = &dquot_quotactl_ops;
+#endif
 
 	/*
 	 * Initialize direct-mapping inode/address-space
diff --git a/fs/libfs.c b/fs/libfs.c
index 232bea425b09..09e1016eb774 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -8,6 +8,7 @@
 #include <linux/slab.h>
 #include <linux/mount.h>
 #include <linux/vfs.h>
+#include <linux/quotaops.h>
 #include <linux/mutex.h>
 #include <linux/exportfs.h>
 #include <linux/writeback.h>
@@ -58,11 +59,6 @@ struct dentry *simple_lookup(struct inode *dir, struct dentry *dentry, struct na
 	return NULL;
 }
 
-int simple_sync_file(struct file * file, struct dentry *dentry, int datasync)
-{
-	return 0;
-}
- 
 int dcache_dir_open(struct inode *inode, struct file *file)
 {
 	static struct qstr cursor_name = {.len = 1, .name = "."};
@@ -190,7 +186,7 @@ const struct file_operations simple_dir_operations = {
 	.llseek		= dcache_dir_lseek,
 	.read		= generic_read_dir,
 	.readdir	= dcache_readdir,
-	.fsync		= simple_sync_file,
+	.fsync		= noop_fsync,
 };
 
 const struct inode_operations simple_dir_inode_operations = {
@@ -330,6 +326,81 @@ int simple_rename(struct inode *old_dir, struct dentry *old_dentry,
 	return 0;
 }
 
+/**
+ * simple_setsize - handle core mm and vfs requirements for file size change
+ * @inode: inode
+ * @newsize: new file size
+ *
+ * Returns 0 on success, -error on failure.
+ *
+ * simple_setsize must be called with inode_mutex held.
+ *
+ * simple_setsize will check that the requested new size is OK (see
+ * inode_newsize_ok), and then will perform the necessary i_size update
+ * and pagecache truncation (if necessary). It will be typically be called
+ * from the filesystem's setattr function when ATTR_SIZE is passed in.
+ *
+ * The inode itself must have correct permissions and attributes to allow
+ * i_size to be changed, this function then just checks that the new size
+ * requested is valid.
+ *
+ * In the case of simple in-memory filesystems with inodes stored solely
+ * in the inode cache, and file data in the pagecache, nothing more needs
+ * to be done to satisfy a truncate request. Filesystems with on-disk
+ * blocks for example will need to free them in the case of truncate, in
+ * that case it may be easier not to use simple_setsize (but each of its
+ * components will likely be required at some point to update pagecache
+ * and inode etc).
+ */
+int simple_setsize(struct inode *inode, loff_t newsize)
+{
+	loff_t oldsize;
+	int error;
+
+	error = inode_newsize_ok(inode, newsize);
+	if (error)
+		return error;
+
+	oldsize = inode->i_size;
+	i_size_write(inode, newsize);
+	truncate_pagecache(inode, oldsize, newsize);
+
+	return error;
+}
+EXPORT_SYMBOL(simple_setsize);
+
+/**
+ * simple_setattr - setattr for simple in-memory filesystem
+ * @dentry: dentry
+ * @iattr: iattr structure
+ *
+ * Returns 0 on success, -error on failure.
+ *
+ * simple_setattr implements setattr for an in-memory filesystem which
+ * does not store its own file data or metadata (eg. uses the page cache
+ * and inode cache as its data store).
+ */
+int simple_setattr(struct dentry *dentry, struct iattr *iattr)
+{
+	struct inode *inode = dentry->d_inode;
+	int error;
+
+	error = inode_change_ok(inode, iattr);
+	if (error)
+		return error;
+
+	if (iattr->ia_valid & ATTR_SIZE) {
+		error = simple_setsize(inode, iattr->ia_size);
+		if (error)
+			return error;
+	}
+
+	generic_setattr(inode, iattr);
+
+	return error;
+}
+EXPORT_SYMBOL(simple_setattr);
+
 int simple_readpage(struct file *file, struct page *page)
 {
 	clear_highpage(page);
@@ -851,13 +922,22 @@ struct dentry *generic_fh_to_parent(struct super_block *sb, struct fid *fid,
 }
 EXPORT_SYMBOL_GPL(generic_fh_to_parent);
 
-int simple_fsync(struct file *file, struct dentry *dentry, int datasync)
+/**
+ * generic_file_fsync - generic fsync implementation for simple filesystems
+ * @file:	file to synchronize
+ * @datasync:	only synchronize essential metadata if true
+ *
+ * This is a generic implementation of the fsync method for simple
+ * filesystems which track all non-inode metadata in the buffers list
+ * hanging off the address_space structure.
+ */
+int generic_file_fsync(struct file *file, int datasync)
 {
 	struct writeback_control wbc = {
 		.sync_mode = WB_SYNC_ALL,
 		.nr_to_write = 0, /* metadata-only; caller takes care of data */
 	};
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	int err;
 	int ret;
 
@@ -872,7 +952,15 @@ int simple_fsync(struct file *file, struct dentry *dentry, int datasync)
 		ret = err;
 	return ret;
 }
-EXPORT_SYMBOL(simple_fsync);
+EXPORT_SYMBOL(generic_file_fsync);
+
+/*
+ * No-op implementation of ->fsync for in-memory filesystems.
+ */
+int noop_fsync(struct file *file, int datasync)
+{
+	return 0;
+}
 
 EXPORT_SYMBOL(dcache_dir_close);
 EXPORT_SYMBOL(dcache_dir_lseek);
@@ -895,7 +983,7 @@ EXPORT_SYMBOL(simple_release_fs);
 EXPORT_SYMBOL(simple_rename);
 EXPORT_SYMBOL(simple_rmdir);
 EXPORT_SYMBOL(simple_statfs);
-EXPORT_SYMBOL(simple_sync_file);
+EXPORT_SYMBOL(noop_fsync);
 EXPORT_SYMBOL(simple_unlink);
 EXPORT_SYMBOL(simple_read_from_buffer);
 EXPORT_SYMBOL(simple_write_to_buffer);
diff --git a/fs/logfs/file.c b/fs/logfs/file.c
index 0de524071870..abe1cafbd4c2 100644
--- a/fs/logfs/file.c
+++ b/fs/logfs/file.c
@@ -219,9 +219,9 @@ int logfs_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
 	}
 }
 
-int logfs_fsync(struct file *file, struct dentry *dentry, int datasync)
+int logfs_fsync(struct file *file, int datasync)
 {
-	struct super_block *sb = dentry->d_inode->i_sb;
+	struct super_block *sb = file->f_mapping->host->i_sb;
 
 	logfs_write_anchor(sb);
 	return 0;
diff --git a/fs/logfs/logfs.h b/fs/logfs/logfs.h
index 1a9db84f8d8f..c838c4d72111 100644
--- a/fs/logfs/logfs.h
+++ b/fs/logfs/logfs.h
@@ -506,7 +506,7 @@ extern const struct address_space_operations logfs_reg_aops;
 int logfs_readpage(struct file *file, struct page *page);
 int logfs_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
 		unsigned long arg);
-int logfs_fsync(struct file *file, struct dentry *dentry, int datasync);
+int logfs_fsync(struct file *file, int datasync);
 
 /* gc.c */
 u32 get_best_cand(struct super_block *sb, struct candidate_list *list, u32 *ec);
diff --git a/fs/minix/dir.c b/fs/minix/dir.c
index 6198731d7fcd..91969589131c 100644
--- a/fs/minix/dir.c
+++ b/fs/minix/dir.c
@@ -22,7 +22,7 @@ const struct file_operations minix_dir_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= minix_readdir,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 };
 
 static inline void dir_put_page(struct page *page)
@@ -72,11 +72,8 @@ static struct page * dir_get_page(struct inode *dir, unsigned long n)
 {
 	struct address_space *mapping = dir->i_mapping;
 	struct page *page = read_mapping_page(mapping, n, NULL);
-	if (!IS_ERR(page)) {
+	if (!IS_ERR(page))
 		kmap(page);
-		if (!PageUptodate(page))
-			goto fail;
-	}
 	return page;
 
 fail:
diff --git a/fs/minix/file.c b/fs/minix/file.c
index 3eec3e607a87..d5320ff23faf 100644
--- a/fs/minix/file.c
+++ b/fs/minix/file.c
@@ -19,7 +19,7 @@ const struct file_operations minix_file_operations = {
 	.write		= do_sync_write,
 	.aio_write	= generic_file_aio_write,
 	.mmap		= generic_file_mmap,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 	.splice_read	= generic_file_splice_read,
 };
 
diff --git a/fs/minix/itree_v2.c b/fs/minix/itree_v2.c
index f23010969369..13487ad16894 100644
--- a/fs/minix/itree_v2.c
+++ b/fs/minix/itree_v2.c
@@ -20,6 +20,9 @@ static inline block_t *i_data(struct inode *inode)
 	return (block_t *)minix_i(inode)->u.i2_data;
 }
 
+#define DIRCOUNT 7
+#define INDIRCOUNT(sb) (1 << ((sb)->s_blocksize_bits - 2))
+
 static int block_to_path(struct inode * inode, long block, int offsets[DEPTH])
 {
 	int n = 0;
@@ -34,21 +37,21 @@ static int block_to_path(struct inode * inode, long block, int offsets[DEPTH])
 			printk("MINIX-fs: block_to_path: "
 			       "block %ld too big on dev %s\n",
 				block, bdevname(sb->s_bdev, b));
-	} else if (block < 7) {
+	} else if (block < DIRCOUNT) {
 		offsets[n++] = block;
-	} else if ((block -= 7) < 256) {
-		offsets[n++] = 7;
+	} else if ((block -= DIRCOUNT) < INDIRCOUNT(sb)) {
+		offsets[n++] = DIRCOUNT;
 		offsets[n++] = block;
-	} else if ((block -= 256) < 256*256) {
-		offsets[n++] = 8;
-		offsets[n++] = block>>8;
-		offsets[n++] = block & 255;
+	} else if ((block -= INDIRCOUNT(sb)) < INDIRCOUNT(sb) * INDIRCOUNT(sb)) {
+		offsets[n++] = DIRCOUNT + 1;
+		offsets[n++] = block / INDIRCOUNT(sb);
+		offsets[n++] = block % INDIRCOUNT(sb);
 	} else {
-		block -= 256*256;
-		offsets[n++] = 9;
-		offsets[n++] = block>>16;
-		offsets[n++] = (block>>8) & 255;
-		offsets[n++] = block & 255;
+		block -= INDIRCOUNT(sb) * INDIRCOUNT(sb);
+		offsets[n++] = DIRCOUNT + 2;
+		offsets[n++] = (block / INDIRCOUNT(sb)) / INDIRCOUNT(sb);
+		offsets[n++] = (block / INDIRCOUNT(sb)) % INDIRCOUNT(sb);
+		offsets[n++] = block % INDIRCOUNT(sb);
 	}
 	return n;
 }
diff --git a/fs/namei.c b/fs/namei.c
index 48e1f60520ea..868d0cb9d473 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -1621,6 +1621,7 @@ static struct file *do_last(struct nameidata *nd, struct path *path,
 	case LAST_DOTDOT:
 		follow_dotdot(nd);
 		dir = nd->path.dentry;
+	case LAST_DOT:
 		if (nd->path.mnt->mnt_sb->s_type->fs_flags & FS_REVAL_DOT) {
 			if (!dir->d_op->d_revalidate(dir, nd)) {
 				error = -ESTALE;
@@ -1628,7 +1629,6 @@ static struct file *do_last(struct nameidata *nd, struct path *path,
 			}
 		}
 		/* fallthrough */
-	case LAST_DOT:
 	case LAST_ROOT:
 		if (open_flag & O_CREAT)
 			goto exit;
diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c
index 7edfcd4d5e52..9578cbe0cd58 100644
--- a/fs/ncpfs/dir.c
+++ b/fs/ncpfs/dir.c
@@ -49,9 +49,10 @@ extern int ncp_symlink(struct inode *, struct dentry *, const char *);
 		      
 const struct file_operations ncp_dir_operations =
 {
+	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= ncp_readdir,
-	.ioctl		= ncp_ioctl,
+	.unlocked_ioctl	= ncp_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= ncp_compat_ioctl,
 #endif
diff --git a/fs/ncpfs/file.c b/fs/ncpfs/file.c
index 1daabb90e0a5..3639cc5cbdae 100644
--- a/fs/ncpfs/file.c
+++ b/fs/ncpfs/file.c
@@ -22,7 +22,7 @@
 #include <linux/ncp_fs.h>
 #include "ncplib_kernel.h"
 
-static int ncp_fsync(struct file *file, struct dentry *dentry, int datasync)
+static int ncp_fsync(struct file *file, int datasync)
 {
 	return 0;
 }
@@ -295,7 +295,7 @@ const struct file_operations ncp_file_operations =
 	.llseek 	= ncp_remote_llseek,
 	.read		= ncp_file_read,
 	.write		= ncp_file_write,
-	.ioctl		= ncp_ioctl,
+	.unlocked_ioctl	= ncp_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= ncp_compat_ioctl,
 #endif
diff --git a/fs/ncpfs/ioctl.c b/fs/ncpfs/ioctl.c
index 60a5e2864ea8..023c03d02070 100644
--- a/fs/ncpfs/ioctl.c
+++ b/fs/ncpfs/ioctl.c
@@ -20,6 +20,7 @@
 #include <linux/smp_lock.h>
 #include <linux/vmalloc.h>
 #include <linux/sched.h>
+#include <linux/smp_lock.h>
 
 #include <linux/ncp_fs.h>
 
@@ -261,9 +262,9 @@ ncp_get_charsets(struct ncp_server* server, struct ncp_nls_ioctl __user *arg)
 }
 #endif /* CONFIG_NCPFS_NLS */
 
-static int __ncp_ioctl(struct inode *inode, struct file *filp,
-	      unsigned int cmd, unsigned long arg)
+static long __ncp_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
+	struct inode *inode = filp->f_dentry->d_inode;
 	struct ncp_server *server = NCP_SERVER(inode);
 	int result;
 	struct ncp_ioctl_request request;
@@ -841,11 +842,11 @@ static int ncp_ioctl_need_write(unsigned int cmd)
 	}
 }
 
-int ncp_ioctl(struct inode *inode, struct file *filp,
-	      unsigned int cmd, unsigned long arg)
+long ncp_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
-	int ret;
+	long ret;
 
+	lock_kernel();
 	if (ncp_ioctl_need_write(cmd)) {
 		/*
 		 * inside the ioctl(), any failures which
@@ -853,24 +854,28 @@ int ncp_ioctl(struct inode *inode, struct file *filp,
 		 * -EACCESS, so it seems consistent to keep
 		 *  that here.
 		 */
-		if (mnt_want_write(filp->f_path.mnt))
-			return -EACCES;
+		if (mnt_want_write(filp->f_path.mnt)) {
+			ret = -EACCES;
+			goto out;
+		}
 	}
-	ret = __ncp_ioctl(inode, filp, cmd, arg);
+	ret = __ncp_ioctl(filp, cmd, arg);
 	if (ncp_ioctl_need_write(cmd))
 		mnt_drop_write(filp->f_path.mnt);
+
+out:
+	unlock_kernel();
 	return ret;
 }
 
 #ifdef CONFIG_COMPAT
 long ncp_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 {
-	struct inode *inode = file->f_path.dentry->d_inode;
-	int ret;
+	long ret;
 
 	lock_kernel();
 	arg = (unsigned long) compat_ptr(arg);
-	ret = ncp_ioctl(inode, file, cmd, arg);
+	ret = ncp_ioctl(file, cmd, arg);
 	unlock_kernel();
 	return ret;
 }
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index ee9a179ebdf3..782b431ef91c 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -53,7 +53,7 @@ static int nfs_link(struct dentry *, struct inode *, struct dentry *);
 static int nfs_mknod(struct inode *, struct dentry *, int, dev_t);
 static int nfs_rename(struct inode *, struct dentry *,
 		      struct inode *, struct dentry *);
-static int nfs_fsync_dir(struct file *, struct dentry *, int);
+static int nfs_fsync_dir(struct file *, int);
 static loff_t nfs_llseek_dir(struct file *, loff_t, int);
 
 const struct file_operations nfs_dir_operations = {
@@ -641,8 +641,10 @@ out:
  * All directory operations under NFS are synchronous, so fsync()
  * is a dummy operation.
  */
-static int nfs_fsync_dir(struct file *filp, struct dentry *dentry, int datasync)
+static int nfs_fsync_dir(struct file *filp, int datasync)
 {
+	struct dentry *dentry = filp->f_path.dentry;
+
 	dfprintk(FILE, "NFS: fsync dir(%s/%s) datasync %d\n",
 			dentry->d_parent->d_name.name, dentry->d_name.name,
 			datasync);
@@ -1741,6 +1743,7 @@ remove_lru_entry:
 			clear_bit(NFS_INO_ACL_LRU_SET, &nfsi->flags);
 			smp_mb__after_clear_bit();
 		}
+		spin_unlock(&inode->i_lock);
 	}
 	spin_unlock(&nfs_access_lru_lock);
 	nfs_access_free_list(&head);
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index cac96bcc91e4..36a5e74f51b4 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -53,7 +53,7 @@ static ssize_t nfs_file_splice_write(struct pipe_inode_info *pipe,
 static ssize_t nfs_file_write(struct kiocb *, const struct iovec *iov,
 				unsigned long nr_segs, loff_t pos);
 static int  nfs_file_flush(struct file *, fl_owner_t id);
-static int  nfs_file_fsync(struct file *, struct dentry *dentry, int datasync);
+static int  nfs_file_fsync(struct file *, int datasync);
 static int nfs_check_flags(int flags);
 static int nfs_lock(struct file *filp, int cmd, struct file_lock *fl);
 static int nfs_flock(struct file *filp, int cmd, struct file_lock *fl);
@@ -322,8 +322,9 @@ nfs_file_mmap(struct file * file, struct vm_area_struct * vma)
  * whether any write errors occurred for this process.
  */
 static int
-nfs_file_fsync(struct file *file, struct dentry *dentry, int datasync)
+nfs_file_fsync(struct file *file, int datasync)
 {
+	struct dentry *dentry = file->f_path.dentry;
 	struct nfs_open_context *ctx = nfs_file_open_context(file);
 	struct inode *inode = dentry->d_inode;
 
diff --git a/fs/nfs/super.c b/fs/nfs/super.c
index 2f8b1157daa2..04214fc5c304 100644
--- a/fs/nfs/super.c
+++ b/fs/nfs/super.c
@@ -1060,7 +1060,7 @@ static int nfs_parse_mount_options(char *raw,
 				goto out_nomem;
 			rc = strict_strtoul(string, 10, &option);
 			kfree(string);
-			if (rc != 0 || option > USHORT_MAX)
+			if (rc != 0 || option > USHRT_MAX)
 				goto out_invalid_value;
 			mnt->nfs_server.port = option;
 			break;
@@ -1181,7 +1181,7 @@ static int nfs_parse_mount_options(char *raw,
 				goto out_nomem;
 			rc = strict_strtoul(string, 10, &option);
 			kfree(string);
-			if (rc != 0 || option > USHORT_MAX)
+			if (rc != 0 || option > USHRT_MAX)
 				goto out_invalid_value;
 			mnt->mount_server.port = option;
 			break;
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 3aea3ca98ab7..91679e2631ee 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1386,7 +1386,7 @@ static int nfs_commit_inode(struct inode *inode, int how)
 	int res = 0;
 
 	if (!nfs_commit_set_lock(NFS_I(inode), may_wait))
-		goto out;
+		goto out_mark_dirty;
 	spin_lock(&inode->i_lock);
 	res = nfs_scan_commit(inode, &head, 0, 0);
 	spin_unlock(&inode->i_lock);
@@ -1398,9 +1398,18 @@ static int nfs_commit_inode(struct inode *inode, int how)
 			wait_on_bit(&NFS_I(inode)->flags, NFS_INO_COMMIT,
 					nfs_wait_bit_killable,
 					TASK_KILLABLE);
+		else
+			goto out_mark_dirty;
 	} else
 		nfs_commit_clear_lock(NFS_I(inode));
-out:
+	return res;
+	/* Note: If we exit without ensuring that the commit is complete,
+	 * we must mark the inode as dirty. Otherwise, future calls to
+	 * sync_inode() with the WB_SYNC_ALL flag set will fail to ensure
+	 * that the data is on the disk.
+	 */
+out_mark_dirty:
+	__mark_inode_dirty(inode, I_DIRTY_DATASYNC);
 	return res;
 }
 
@@ -1509,14 +1518,17 @@ int nfs_wb_page(struct inode *inode, struct page *page)
 	};
 	int ret;
 
-	while(PagePrivate(page)) {
+	for (;;) {
 		wait_on_page_writeback(page);
 		if (clear_page_dirty_for_io(page)) {
 			ret = nfs_writepage_locked(page, &wbc);
 			if (ret < 0)
 				goto out_error;
+			continue;
 		}
-		ret = sync_inode(inode, &wbc);
+		if (!PagePrivate(page))
+			break;
+		ret = nfs_commit_inode(inode, FLUSH_SYNC);
 		if (ret < 0)
 			goto out_error;
 	}
diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
index bc3194ea01f5..508941c23af7 100644
--- a/fs/nfsd/nfsctl.c
+++ b/fs/nfsd/nfsctl.c
@@ -998,7 +998,7 @@ static ssize_t __write_ports_addxprt(char *buf)
 	if (sscanf(buf, "%15s %4u", transport, &port) != 2)
 		return -EINVAL;
 
-	if (port < 1 || port > USHORT_MAX)
+	if (port < 1 || port > USHRT_MAX)
 		return -EINVAL;
 
 	err = nfsd_create_serv();
@@ -1040,7 +1040,7 @@ static ssize_t __write_ports_delxprt(char *buf)
 	if (sscanf(&buf[1], "%15s %4u", transport, &port) != 2)
 		return -EINVAL;
 
-	if (port < 1 || port > USHORT_MAX || nfsd_serv == NULL)
+	if (port < 1 || port > USHRT_MAX || nfsd_serv == NULL)
 		return -EINVAL;
 
 	xprt = svc_find_xprt(nfsd_serv, transport, AF_UNSPEC, port);
diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c
index 30292df443ce..c9a30d7ff6fc 100644
--- a/fs/nilfs2/file.c
+++ b/fs/nilfs2/file.c
@@ -27,7 +27,7 @@
 #include "nilfs.h"
 #include "segment.h"
 
-int nilfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
+int nilfs_sync_file(struct file *file, int datasync)
 {
 	/*
 	 * Called from fsync() system call
@@ -37,7 +37,7 @@ int nilfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
 	 * This function should be implemented when the writeback function
 	 * will be implemented.
 	 */
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	int err;
 
 	if (!nilfs_inode_dirty(inode))
diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h
index 8723e5bfd071..47d6d7928122 100644
--- a/fs/nilfs2/nilfs.h
+++ b/fs/nilfs2/nilfs.h
@@ -228,7 +228,7 @@ extern void nilfs_set_link(struct inode *, struct nilfs_dir_entry *,
 			   struct page *, struct inode *);
 
 /* file.c */
-extern int nilfs_sync_file(struct file *, struct dentry *, int);
+extern int nilfs_sync_file(struct file *, int);
 
 /* ioctl.c */
 long nilfs_ioctl(struct file *, unsigned int, unsigned long);
diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c
index fe44d3feee4a..0f48e7c5d9e1 100644
--- a/fs/ntfs/dir.c
+++ b/fs/ntfs/dir.c
@@ -1527,10 +1527,9 @@ static int ntfs_dir_open(struct inode *vi, struct file *filp)
  * this problem for now.  We do write the $BITMAP attribute if it is present
  * which is the important one for a directory so things are not too bad.
  */
-static int ntfs_dir_fsync(struct file *filp, struct dentry *dentry,
-		int datasync)
+static int ntfs_dir_fsync(struct file *filp, int datasync)
 {
-	struct inode *bmp_vi, *vi = dentry->d_inode;
+	struct inode *bmp_vi, *vi = filp->f_mapping->host;
 	int err, ret;
 	ntfs_attr na;
 
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index 8804f093ba75..113ebd9f25a4 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -98,9 +98,6 @@ static int ntfs_file_open(struct inode *vi, struct file *filp)
  * the page at all.  For a more detailed explanation see ntfs_truncate() in
  * fs/ntfs/inode.c.
  *
- * @cached_page and @lru_pvec are just optimizations for dealing with multiple
- * pages.
- *
  * Return 0 on success and -errno on error.  In the case that an error is
  * encountered it is possible that the initialized size will already have been
  * incremented some way towards @new_init_size but it is guaranteed that if
@@ -110,8 +107,7 @@ static int ntfs_file_open(struct inode *vi, struct file *filp)
  * Locking: i_mutex on the vfs inode corrseponsind to the ntfs inode @ni must be
  *	    held by the caller.
  */
-static int ntfs_attr_extend_initialized(ntfs_inode *ni, const s64 new_init_size,
-		struct page **cached_page, struct pagevec *lru_pvec)
+static int ntfs_attr_extend_initialized(ntfs_inode *ni, const s64 new_init_size)
 {
 	s64 old_init_size;
 	loff_t old_i_size;
@@ -403,18 +399,13 @@ static inline void ntfs_fault_in_pages_readable_iovec(const struct iovec *iov,
  * Obtain @nr_pages locked page cache pages from the mapping @mapping and
  * starting at index @index.
  *
- * If a page is newly created, increment its refcount and add it to the
- * caller's lru-buffering pagevec @lru_pvec.
- *
- * This is the same as mm/filemap.c::__grab_cache_page(), except that @nr_pages
- * are obtained at once instead of just one page and that 0 is returned on
- * success and -errno on error.
+ * If a page is newly created, add it to lru list
  *
  * Note, the page locks are obtained in ascending page index order.
  */
 static inline int __ntfs_grab_cache_pages(struct address_space *mapping,
 		pgoff_t index, const unsigned nr_pages, struct page **pages,
-		struct page **cached_page, struct pagevec *lru_pvec)
+		struct page **cached_page)
 {
 	int err, nr;
 
@@ -430,7 +421,7 @@ static inline int __ntfs_grab_cache_pages(struct address_space *mapping,
 					goto err_out;
 				}
 			}
-			err = add_to_page_cache(*cached_page, mapping, index,
+			err = add_to_page_cache_lru(*cached_page, mapping, index,
 					GFP_KERNEL);
 			if (unlikely(err)) {
 				if (err == -EEXIST)
@@ -438,9 +429,6 @@ static inline int __ntfs_grab_cache_pages(struct address_space *mapping,
 				goto err_out;
 			}
 			pages[nr] = *cached_page;
-			page_cache_get(*cached_page);
-			if (unlikely(!pagevec_add(lru_pvec, *cached_page)))
-				__pagevec_lru_add_file(lru_pvec);
 			*cached_page = NULL;
 		}
 		index++;
@@ -1800,7 +1788,6 @@ static ssize_t ntfs_file_buffered_write(struct kiocb *iocb,
 	ssize_t status, written;
 	unsigned nr_pages;
 	int err;
-	struct pagevec lru_pvec;
 
 	ntfs_debug("Entering for i_ino 0x%lx, attribute type 0x%x, "
 			"pos 0x%llx, count 0x%lx.",
@@ -1912,7 +1899,6 @@ static ssize_t ntfs_file_buffered_write(struct kiocb *iocb,
 			}
 		}
 	}
-	pagevec_init(&lru_pvec, 0);
 	written = 0;
 	/*
 	 * If the write starts beyond the initialized size, extend it up to the
@@ -1925,8 +1911,7 @@ static ssize_t ntfs_file_buffered_write(struct kiocb *iocb,
 	ll = ni->initialized_size;
 	read_unlock_irqrestore(&ni->size_lock, flags);
 	if (pos > ll) {
-		err = ntfs_attr_extend_initialized(ni, pos, &cached_page,
-				&lru_pvec);
+		err = ntfs_attr_extend_initialized(ni, pos);
 		if (err < 0) {
 			ntfs_error(vol->sb, "Cannot perform write to inode "
 					"0x%lx, attribute type 0x%x, because "
@@ -2012,7 +1997,7 @@ static ssize_t ntfs_file_buffered_write(struct kiocb *iocb,
 			ntfs_fault_in_pages_readable_iovec(iov, iov_ofs, bytes);
 		/* Get and lock @do_pages starting at index @start_idx. */
 		status = __ntfs_grab_cache_pages(mapping, start_idx, do_pages,
-				pages, &cached_page, &lru_pvec);
+				pages, &cached_page);
 		if (unlikely(status))
 			break;
 		/*
@@ -2077,7 +2062,6 @@ err_out:
 	*ppos = pos;
 	if (cached_page)
 		page_cache_release(cached_page);
-	pagevec_lru_add_file(&lru_pvec);
 	ntfs_debug("Done.  Returning %s (written 0x%lx, status %li).",
 			written ? "written" : "status", (unsigned long)written,
 			(long)status);
@@ -2149,7 +2133,6 @@ static ssize_t ntfs_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
 /**
  * ntfs_file_fsync - sync a file to disk
  * @filp:	file to be synced
- * @dentry:	dentry describing the file to sync
  * @datasync:	if non-zero only flush user data and not metadata
  *
  * Data integrity sync of a file to disk.  Used for fsync, fdatasync, and msync
@@ -2165,19 +2148,15 @@ static ssize_t ntfs_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
  * Also, if @datasync is true, we do not wait on the inode to be written out
  * but we always wait on the page cache pages to be written out.
  *
- * Note: In the past @filp could be NULL so we ignore it as we don't need it
- * anyway.
- *
  * Locking: Caller must hold i_mutex on the inode.
  *
  * TODO: We should probably also write all attribute/index inodes associated
  * with this inode but since we have no simple way of getting to them we ignore
  * this problem for now.
  */
-static int ntfs_file_fsync(struct file *filp, struct dentry *dentry,
-		int datasync)
+static int ntfs_file_fsync(struct file *filp, int datasync)
 {
-	struct inode *vi = dentry->d_inode;
+	struct inode *vi = filp->f_mapping->host;
 	int err, ret = 0;
 
 	ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
diff --git a/fs/ocfs2/blockcheck.c b/fs/ocfs2/blockcheck.c
index b7428c5d0d3b..ec6d12339593 100644
--- a/fs/ocfs2/blockcheck.c
+++ b/fs/ocfs2/blockcheck.c
@@ -403,7 +403,7 @@ void ocfs2_block_check_compute(void *data, size_t blocksize,
 	 * No ecc'd ocfs2 structure is larger than 4K, so ecc will be no
 	 * larger than 16 bits.
 	 */
-	BUG_ON(ecc > USHORT_MAX);
+	BUG_ON(ecc > USHRT_MAX);
 
 	bc->bc_crc32e = cpu_to_le32(crc);
 	bc->bc_ecc = cpu_to_le16((u16)ecc);
@@ -508,7 +508,7 @@ void ocfs2_block_check_compute_bhs(struct buffer_head **bhs, int nr,
 	 * No ecc'd ocfs2 structure is larger than 4K, so ecc will be no
 	 * larger than 16 bits.
 	 */
-	BUG_ON(ecc > USHORT_MAX);
+	BUG_ON(ecc > USHRT_MAX);
 
 	bc->bc_crc32e = cpu_to_le32(crc);
 	bc->bc_ecc = cpu_to_le16((u16)ecc);
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 97e54b9e654b..6a13ea64c447 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -175,13 +175,12 @@ static int ocfs2_dir_release(struct inode *inode, struct file *file)
 	return 0;
 }
 
-static int ocfs2_sync_file(struct file *file,
-			   struct dentry *dentry,
-			   int datasync)
+static int ocfs2_sync_file(struct file *file, int datasync)
 {
 	int err = 0;
 	journal_t *journal;
-	struct inode *inode = dentry->d_inode;
+	struct dentry *dentry = file->f_path.dentry;
+	struct inode *inode = file->f_mapping->host;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", file, dentry, datasync,
@@ -1053,7 +1052,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 	}
 
 	/*
-	 * This will intentionally not wind up calling vmtruncate(),
+	 * This will intentionally not wind up calling simple_setsize(),
 	 * since all the work for a size change has been done above.
 	 * Otherwise, we could get into problems with truncate as
 	 * ip_alloc_sem is used there to protect against i_size
@@ -2119,9 +2118,13 @@ relock:
 			 * direct write may have instantiated a few
 			 * blocks outside i_size. Trim these off again.
 			 * Don't need i_size_read because we hold i_mutex.
+			 *
+			 * XXX(hch): this looks buggy because ocfs2 did not
+			 * actually implement ->truncate.  Take a look at
+			 * the new truncate sequence and update this accordingly
 			 */
 			if (*ppos + count > inode->i_size)
-				vmtruncate(inode, inode->i_size);
+				simple_setsize(inode, inode->i_size);
 			ret = written;
 			goto out_dio;
 		}
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 2c26ce251cb3..0eaa929a4dbf 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -879,13 +879,15 @@ static int ocfs2_susp_quotas(struct ocfs2_super *osb, int unsuspend)
 		if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, feature[type]))
 			continue;
 		if (unsuspend)
-			status = vfs_quota_enable(
-					sb_dqopt(sb)->files[type],
-					type, QFMT_OCFS2,
-					DQUOT_SUSPENDED);
-		else
-			status = vfs_quota_disable(sb, type,
-						   DQUOT_SUSPENDED);
+			status = dquot_resume(sb, type);
+		else {
+			struct ocfs2_mem_dqinfo *oinfo;
+
+			/* Cancel periodic syncing before suspending */
+			oinfo = sb_dqinfo(sb, type)->dqi_priv;
+			cancel_delayed_work_sync(&oinfo->dqi_sync_work);
+			status = dquot_suspend(sb, type);
+		}
 		if (status < 0)
 			break;
 	}
@@ -916,8 +918,8 @@ static int ocfs2_enable_quotas(struct ocfs2_super *osb)
 			status = -ENOENT;
 			goto out_quota_off;
 		}
-		status = vfs_quota_enable(inode[type], type, QFMT_OCFS2,
-						DQUOT_USAGE_ENABLED);
+		status = dquot_enable(inode[type], type, QFMT_OCFS2,
+				      DQUOT_USAGE_ENABLED);
 		if (status < 0)
 			goto out_quota_off;
 	}
@@ -952,8 +954,8 @@ static void ocfs2_disable_quotas(struct ocfs2_super *osb)
 		/* Turn off quotas. This will remove all dquot structures from
 		 * memory and so they will be automatically synced to global
 		 * quota files */
-		vfs_quota_disable(sb, type, DQUOT_USAGE_ENABLED |
-					    DQUOT_LIMITS_ENABLED);
+		dquot_disable(sb, type, DQUOT_USAGE_ENABLED |
+					DQUOT_LIMITS_ENABLED);
 		if (!inode)
 			continue;
 		iput(inode);
@@ -962,7 +964,7 @@ static void ocfs2_disable_quotas(struct ocfs2_super *osb)
 
 /* Handle quota on quotactl */
 static int ocfs2_quota_on(struct super_block *sb, int type, int format_id,
-			  char *path, int remount)
+			  char *path)
 {
 	unsigned int feature[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
 					     OCFS2_FEATURE_RO_COMPAT_GRPQUOTA};
@@ -970,30 +972,24 @@ static int ocfs2_quota_on(struct super_block *sb, int type, int format_id,
 	if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, feature[type]))
 		return -EINVAL;
 
-	if (remount)
-		return 0;	/* Just ignore it has been handled in
-				 * ocfs2_remount() */
-	return vfs_quota_enable(sb_dqopt(sb)->files[type], type,
-				    format_id, DQUOT_LIMITS_ENABLED);
+	return dquot_enable(sb_dqopt(sb)->files[type], type,
+			    format_id, DQUOT_LIMITS_ENABLED);
 }
 
 /* Handle quota off quotactl */
-static int ocfs2_quota_off(struct super_block *sb, int type, int remount)
+static int ocfs2_quota_off(struct super_block *sb, int type)
 {
-	if (remount)
-		return 0;	/* Ignore now and handle later in
-				 * ocfs2_remount() */
-	return vfs_quota_disable(sb, type, DQUOT_LIMITS_ENABLED);
+	return dquot_disable(sb, type, DQUOT_LIMITS_ENABLED);
 }
 
 static const struct quotactl_ops ocfs2_quotactl_ops = {
 	.quota_on	= ocfs2_quota_on,
 	.quota_off	= ocfs2_quota_off,
-	.quota_sync	= vfs_quota_sync,
-	.get_info	= vfs_get_dqinfo,
-	.set_info	= vfs_set_dqinfo,
-	.get_dqblk	= vfs_get_dqblk,
-	.set_dqblk	= vfs_set_dqblk,
+	.quota_sync	= dquot_quota_sync,
+	.get_info	= dquot_get_dqinfo,
+	.set_info	= dquot_set_dqinfo,
+	.get_dqblk	= dquot_get_dqblk,
+	.set_dqblk	= dquot_set_dqblk,
 };
 
 static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
diff --git a/fs/omfs/file.c b/fs/omfs/file.c
index 399487c09364..6e7a3291bbe8 100644
--- a/fs/omfs/file.c
+++ b/fs/omfs/file.c
@@ -329,7 +329,7 @@ const struct file_operations omfs_file_operations = {
 	.aio_read = generic_file_aio_read,
 	.aio_write = generic_file_aio_write,
 	.mmap = generic_file_mmap,
-	.fsync = simple_fsync,
+	.fsync = generic_file_fsync,
 	.splice_read = generic_file_splice_read,
 };
 
diff --git a/fs/partitions/ldm.c b/fs/partitions/ldm.c
index 3ceca05b668c..648c9d8f3357 100644
--- a/fs/partitions/ldm.c
+++ b/fs/partitions/ldm.c
@@ -26,6 +26,7 @@
 #include <linux/slab.h>
 #include <linux/pagemap.h>
 #include <linux/stringify.h>
+#include <linux/kernel.h>
 #include "ldm.h"
 #include "check.h"
 #include "msdos.h"
@@ -77,17 +78,16 @@ static int ldm_parse_hexbyte (const u8 *src)
 	int h;
 
 	/* high part */
-	if      ((x = src[0] - '0') <= '9'-'0') h = x;
-	else if ((x = src[0] - 'a') <= 'f'-'a') h = x+10;
-	else if ((x = src[0] - 'A') <= 'F'-'A') h = x+10;
-	else return -1;
-	h <<= 4;
+	x = h = hex_to_bin(src[0]);
+	if (h < 0)
+		return -1;
 
 	/* low part */
-	if ((x = src[1] - '0') <= '9'-'0') return h | x;
-	if ((x = src[1] - 'a') <= 'f'-'a') return h | (x+10);
-	if ((x = src[1] - 'A') <= 'F'-'A') return h | (x+10);
-	return -1;
+	h = hex_to_bin(src[1]);
+	if (h < 0)
+		return -1;
+
+	return (x << 4) + h;
 }
 
 /**
diff --git a/fs/pipe.c b/fs/pipe.c
index bdd3f96054b9..541d6626f9d9 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -230,6 +230,7 @@ void *generic_pipe_buf_map(struct pipe_inode_info *pipe,
 
 	return kmap(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_map);
 
 /**
  * generic_pipe_buf_unmap - unmap a previously mapped pipe buffer
@@ -249,6 +250,7 @@ void generic_pipe_buf_unmap(struct pipe_inode_info *pipe,
 	} else
 		kunmap(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_unmap);
 
 /**
  * generic_pipe_buf_steal - attempt to take ownership of a &pipe_buffer
@@ -279,6 +281,7 @@ int generic_pipe_buf_steal(struct pipe_inode_info *pipe,
 
 	return 1;
 }
+EXPORT_SYMBOL(generic_pipe_buf_steal);
 
 /**
  * generic_pipe_buf_get - get a reference to a &struct pipe_buffer
@@ -294,6 +297,7 @@ void generic_pipe_buf_get(struct pipe_inode_info *pipe, struct pipe_buffer *buf)
 {
 	page_cache_get(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_get);
 
 /**
  * generic_pipe_buf_confirm - verify contents of the pipe buffer
@@ -309,6 +313,7 @@ int generic_pipe_buf_confirm(struct pipe_inode_info *info,
 {
 	return 0;
 }
+EXPORT_SYMBOL(generic_pipe_buf_confirm);
 
 /**
  * generic_pipe_buf_release - put a reference to a &struct pipe_buffer
@@ -323,6 +328,7 @@ void generic_pipe_buf_release(struct pipe_inode_info *pipe,
 {
 	page_cache_release(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_release);
 
 static const struct pipe_buf_operations anon_pipe_buf_ops = {
 	.can_merge = 1,
@@ -1172,16 +1178,20 @@ long pipe_fcntl(struct file *file, unsigned int cmd, unsigned long arg)
 		nr_pages = (arg + PAGE_SIZE - 1) >> PAGE_SHIFT;
 		nr_pages = roundup_pow_of_two(nr_pages);
 
-		if (!capable(CAP_SYS_ADMIN) && nr_pages > pipe_max_pages)
-			return -EPERM;
+		if (!capable(CAP_SYS_ADMIN) && nr_pages > pipe_max_pages) {
+			ret = -EPERM;
+			goto out;
+		}
 
 		/*
 		 * The pipe needs to be at least 2 pages large to
 		 * guarantee POSIX behaviour.
 		 */
-		if (nr_pages < 2)
-			return -EINVAL;
-		ret = pipe_set_size(pipe, nr_pages);
+		if (arg < 2) {
+			ret = -EINVAL;
+			goto out;
+		}
+		ret = pipe_set_size(pipe, arg);
 		break;
 		}
 	case F_GETPIPE_SZ:
@@ -1192,6 +1202,7 @@ long pipe_fcntl(struct file *file, unsigned int cmd, unsigned long arg)
 		break;
 	}
 
+out:
 	mutex_unlock(&pipe->inode->i_mutex);
 	return ret;
 }
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 885ab5513ac5..9b58d38bc911 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -267,7 +267,7 @@ static inline void task_sig(struct seq_file *m, struct task_struct *p)
 		shpending = p->signal->shared_pending.signal;
 		blocked = p->blocked;
 		collect_sigign_sigcatch(p, &ignored, &caught);
-		num_threads = atomic_read(&p->signal->count);
+		num_threads = get_nr_threads(p);
 		rcu_read_lock();  /* FIXME: is this correct? */
 		qsize = atomic_read(&__task_cred(p)->user->sigpending);
 		rcu_read_unlock();
@@ -410,7 +410,7 @@ static int do_task_stat(struct seq_file *m, struct pid_namespace *ns,
 			tty_nr = new_encode_dev(tty_devnum(sig->tty));
 		}
 
-		num_threads = atomic_read(&sig->count);
+		num_threads = get_nr_threads(task);
 		collect_sigign_sigcatch(task, &sigign, &sigcatch);
 
 		cmin_flt = sig->cmin_flt;
diff --git a/fs/proc/base.c b/fs/proc/base.c
index c7f9f23449dc..acb7ef80ea4f 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -166,18 +166,6 @@ static int get_fs_path(struct task_struct *task, struct path *path, bool root)
 	return result;
 }
 
-static int get_nr_threads(struct task_struct *tsk)
-{
-	unsigned long flags;
-	int count = 0;
-
-	if (lock_task_sighand(tsk, &flags)) {
-		count = atomic_read(&tsk->signal->count);
-		unlock_task_sighand(tsk, &flags);
-	}
-	return count;
-}
-
 static int proc_cwd_link(struct inode *inode, struct path *path)
 {
 	struct task_struct *task = get_proc_task(inode);
@@ -2444,7 +2432,7 @@ static struct dentry *proc_base_instantiate(struct inode *dir,
 	const struct pid_entry *p = ptr;
 	struct inode *inode;
 	struct proc_inode *ei;
-	struct dentry *error = ERR_PTR(-EINVAL);
+	struct dentry *error;
 
 	/* Allocate the inode */
 	error = ERR_PTR(-ENOMEM);
@@ -2794,7 +2782,7 @@ out:
 
 struct dentry *proc_pid_lookup(struct inode *dir, struct dentry * dentry, struct nameidata *nd)
 {
-	struct dentry *result = ERR_PTR(-ENOENT);
+	struct dentry *result;
 	struct task_struct *task;
 	unsigned tgid;
 	struct pid_namespace *ns;
diff --git a/fs/proc/generic.c b/fs/proc/generic.c
index 43c127490606..2791907744ed 100644
--- a/fs/proc/generic.c
+++ b/fs/proc/generic.c
@@ -343,21 +343,6 @@ static DEFINE_SPINLOCK(proc_inum_lock); /* protects the above */
 /*
  * Return an inode number between PROC_DYNAMIC_FIRST and
  * 0xffffffff, or zero on failure.
- *
- * Current inode allocations in the proc-fs (hex-numbers):
- *
- * 00000000		reserved
- * 00000001-00000fff	static entries	(goners)
- *      001		root-ino
- *
- * 00001000-00001fff	unused
- * 0001xxxx-7fffxxxx	pid-dir entries for pid 1-7fff
- * 80000000-efffffff	unused
- * f0000000-ffffffff	dynamic entries
- *
- * Goal:
- *	Once we split the thing into several virtual filesystems,
- *	we will get rid of magical ranges (and this comment, BTW).
  */
 static unsigned int get_inode_number(void)
 {
diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c
index c837a77351be..6f37c391468d 100644
--- a/fs/proc/kcore.c
+++ b/fs/proc/kcore.c
@@ -588,7 +588,7 @@ static struct kcore_list kcore_text;
  */
 static void __init proc_kcore_text_init(void)
 {
-	kclist_add(&kcore_text, _stext, _end - _stext, KCORE_TEXT);
+	kclist_add(&kcore_text, _text, _end - _text, KCORE_TEXT);
 }
 #else
 static void __init proc_kcore_text_init(void)
diff --git a/fs/proc/root.c b/fs/proc/root.c
index 757c069f2a65..4258384ed22d 100644
--- a/fs/proc/root.c
+++ b/fs/proc/root.c
@@ -110,7 +110,6 @@ void __init proc_root_init(void)
 	if (err)
 		return;
 	proc_mnt = kern_mount_data(&proc_fs_type, &init_pid_ns);
-	err = PTR_ERR(proc_mnt);
 	if (IS_ERR(proc_mnt)) {
 		unregister_filesystem(&proc_fs_type);
 		return;
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 47f5b145f56e..aea1d3f1ffb5 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -634,6 +634,7 @@ static int pagemap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
 	return err;
 }
 
+#ifdef CONFIG_HUGETLB_PAGE
 static u64 huge_pte_to_pagemap_entry(pte_t pte, int offset)
 {
 	u64 pme = 0;
@@ -664,6 +665,7 @@ static int pagemap_hugetlb_range(pte_t *pte, unsigned long hmask,
 
 	return err;
 }
+#endif /* HUGETLB_PAGE */
 
 /*
  * /proc/pid/pagemap - an array mapping virtual pages to pfns
@@ -733,7 +735,9 @@ static ssize_t pagemap_read(struct file *file, char __user *buf,
 
 	pagemap_walk.pmd_entry = pagemap_pte_range;
 	pagemap_walk.pte_hole = pagemap_pte_hole;
+#ifdef CONFIG_HUGETLB_PAGE
 	pagemap_walk.hugetlb_entry = pagemap_hugetlb_range;
+#endif
 	pagemap_walk.mm = mm;
 	pagemap_walk.private = &pm;
 
diff --git a/fs/qnx4/dir.c b/fs/qnx4/dir.c
index 6f30c3d5bcbf..6e8fc62b40a8 100644
--- a/fs/qnx4/dir.c
+++ b/fs/qnx4/dir.c
@@ -77,9 +77,10 @@ out:
 
 const struct file_operations qnx4_dir_operations =
 {
+	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= qnx4_readdir,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 };
 
 const struct inode_operations qnx4_dir_inode_operations =
diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c
index 655a4c52b8c3..12c233da1b6b 100644
--- a/fs/quota/dquot.c
+++ b/fs/quota/dquot.c
@@ -228,10 +228,6 @@ static struct hlist_head *dquot_hash;
 
 struct dqstats dqstats;
 EXPORT_SYMBOL(dqstats);
-#ifdef CONFIG_SMP
-struct dqstats *dqstats_pcpu;
-EXPORT_SYMBOL(dqstats_pcpu);
-#endif
 
 static qsize_t inode_get_rsv_space(struct inode *inode);
 static void __dquot_initialize(struct inode *inode, int type);
@@ -584,7 +580,7 @@ out:
 }
 EXPORT_SYMBOL(dquot_scan_active);
 
-int vfs_quota_sync(struct super_block *sb, int type, int wait)
+int dquot_quota_sync(struct super_block *sb, int type, int wait)
 {
 	struct list_head *dirty;
 	struct dquot *dquot;
@@ -656,7 +652,7 @@ int vfs_quota_sync(struct super_block *sb, int type, int wait)
 
 	return 0;
 }
-EXPORT_SYMBOL(vfs_quota_sync);
+EXPORT_SYMBOL(dquot_quota_sync);
 
 /* Free unused dquots from cache */
 static void prune_dqcache(int count)
@@ -676,27 +672,10 @@ static void prune_dqcache(int count)
 	}
 }
 
-static int dqstats_read(unsigned int type)
-{
-	int count = 0;
-#ifdef CONFIG_SMP
-	int cpu;
-	for_each_possible_cpu(cpu)
-		count += per_cpu_ptr(dqstats_pcpu, cpu)->stat[type];
-	/* Statistics reading is racy, but absolute accuracy isn't required */
-	if (count < 0)
-		count = 0;
-#else
-	count = dqstats.stat[type];
-#endif
-	return count;
-}
-
 /*
  * This is called from kswapd when we think we need some
  * more memory
  */
-
 static int shrink_dqcache_memory(int nr, gfp_t gfp_mask)
 {
 	if (nr) {
@@ -704,7 +683,9 @@ static int shrink_dqcache_memory(int nr, gfp_t gfp_mask)
 		prune_dqcache(nr);
 		spin_unlock(&dq_list_lock);
 	}
-	return (dqstats_read(DQST_FREE_DQUOTS)/100) * sysctl_vfs_cache_pressure;
+	return ((unsigned)
+		percpu_counter_read_positive(&dqstats.counter[DQST_FREE_DQUOTS])
+		/100) * sysctl_vfs_cache_pressure;
 }
 
 static struct shrinker dqcache_shrinker = {
@@ -1514,11 +1495,13 @@ static void inode_decr_space(struct inode *inode, qsize_t number, int reserve)
 /*
  * This operation can block, but only after everything is updated
  */
-int __dquot_alloc_space(struct inode *inode, qsize_t number,
-		int warn, int reserve)
+int __dquot_alloc_space(struct inode *inode, qsize_t number, int flags)
 {
 	int cnt, ret = 0;
 	char warntype[MAXQUOTAS];
+	int warn = flags & DQUOT_SPACE_WARN;
+	int reserve = flags & DQUOT_SPACE_RESERVE;
+	int nofail = flags & DQUOT_SPACE_NOFAIL;
 
 	/*
 	 * First test before acquiring mutex - solves deadlocks when we
@@ -1539,7 +1522,7 @@ int __dquot_alloc_space(struct inode *inode, qsize_t number,
 			continue;
 		ret = check_bdq(inode->i_dquot[cnt], number, !warn,
 				warntype+cnt);
-		if (ret) {
+		if (ret && !nofail) {
 			spin_unlock(&dq_data_lock);
 			goto out_flush_warn;
 		}
@@ -1638,10 +1621,11 @@ EXPORT_SYMBOL(dquot_claim_space_nodirty);
 /*
  * This operation can block, but only after everything is updated
  */
-void __dquot_free_space(struct inode *inode, qsize_t number, int reserve)
+void __dquot_free_space(struct inode *inode, qsize_t number, int flags)
 {
 	unsigned int cnt;
 	char warntype[MAXQUOTAS];
+	int reserve = flags & DQUOT_SPACE_RESERVE;
 
 	/* First test before acquiring mutex - solves deadlocks when we
          * re-enter the quota code and are already holding the mutex */
@@ -1812,7 +1796,7 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr)
 	if (iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid)
 		transfer_to[USRQUOTA] = dqget(sb, iattr->ia_uid, USRQUOTA);
 	if (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)
-		transfer_to[GRPQUOTA] = dqget(sb, iattr->ia_uid, GRPQUOTA);
+		transfer_to[GRPQUOTA] = dqget(sb, iattr->ia_gid, GRPQUOTA);
 
 	ret = __dquot_transfer(inode, transfer_to);
 	dqput_all(transfer_to);
@@ -1847,6 +1831,7 @@ const struct dquot_operations dquot_operations = {
 	.alloc_dquot	= dquot_alloc,
 	.destroy_dquot	= dquot_destroy,
 };
+EXPORT_SYMBOL(dquot_operations);
 
 /*
  * Generic helper for ->open on filesystems supporting disk quotas.
@@ -1865,7 +1850,7 @@ EXPORT_SYMBOL(dquot_file_open);
 /*
  * Turn quota off on a device. type == -1 ==> quotaoff for all types (umount)
  */
-int vfs_quota_disable(struct super_block *sb, int type, unsigned int flags)
+int dquot_disable(struct super_block *sb, int type, unsigned int flags)
 {
 	int cnt, ret = 0;
 	struct quota_info *dqopt = sb_dqopt(sb);
@@ -1995,14 +1980,15 @@ put_inodes:
 		}
 	return ret;
 }
-EXPORT_SYMBOL(vfs_quota_disable);
+EXPORT_SYMBOL(dquot_disable);
 
-int vfs_quota_off(struct super_block *sb, int type, int remount)
+int dquot_quota_off(struct super_block *sb, int type)
 {
-	return vfs_quota_disable(sb, type, remount ? DQUOT_SUSPENDED :
-				 (DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED));
+	return dquot_disable(sb, type,
+			     DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
 }
-EXPORT_SYMBOL(vfs_quota_off);
+EXPORT_SYMBOL(dquot_quota_off);
+
 /*
  *	Turn quotas on on a device
  */
@@ -2120,36 +2106,43 @@ out_fmt:
 }
 
 /* Reenable quotas on remount RW */
-static int vfs_quota_on_remount(struct super_block *sb, int type)
+int dquot_resume(struct super_block *sb, int type)
 {
 	struct quota_info *dqopt = sb_dqopt(sb);
 	struct inode *inode;
-	int ret;
+	int ret = 0, cnt;
 	unsigned int flags;
 
-	mutex_lock(&dqopt->dqonoff_mutex);
-	if (!sb_has_quota_suspended(sb, type)) {
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
+		if (type != -1 && cnt != type)
+			continue;
+
+		mutex_lock(&dqopt->dqonoff_mutex);
+		if (!sb_has_quota_suspended(sb, cnt)) {
+			mutex_unlock(&dqopt->dqonoff_mutex);
+			continue;
+		}
+		inode = dqopt->files[cnt];
+		dqopt->files[cnt] = NULL;
+		spin_lock(&dq_state_lock);
+		flags = dqopt->flags & dquot_state_flag(DQUOT_USAGE_ENABLED |
+							DQUOT_LIMITS_ENABLED,
+							cnt);
+		dqopt->flags &= ~dquot_state_flag(DQUOT_STATE_FLAGS, cnt);
+		spin_unlock(&dq_state_lock);
 		mutex_unlock(&dqopt->dqonoff_mutex);
-		return 0;
-	}
-	inode = dqopt->files[type];
-	dqopt->files[type] = NULL;
-	spin_lock(&dq_state_lock);
-	flags = dqopt->flags & dquot_state_flag(DQUOT_USAGE_ENABLED |
-						DQUOT_LIMITS_ENABLED, type);
-	dqopt->flags &= ~dquot_state_flag(DQUOT_STATE_FLAGS, type);
-	spin_unlock(&dq_state_lock);
-	mutex_unlock(&dqopt->dqonoff_mutex);
 
-	flags = dquot_generic_flag(flags, type);
-	ret = vfs_load_quota_inode(inode, type, dqopt->info[type].dqi_fmt_id,
-				   flags);
-	iput(inode);
+		flags = dquot_generic_flag(flags, cnt);
+		ret = vfs_load_quota_inode(inode, cnt,
+				dqopt->info[cnt].dqi_fmt_id, flags);
+		iput(inode);
+	}
 
 	return ret;
 }
+EXPORT_SYMBOL(dquot_resume);
 
-int vfs_quota_on_path(struct super_block *sb, int type, int format_id,
+int dquot_quota_on_path(struct super_block *sb, int type, int format_id,
 		      struct path *path)
 {
 	int error = security_quota_on(path->dentry);
@@ -2164,40 +2157,36 @@ int vfs_quota_on_path(struct super_block *sb, int type, int format_id,
 					     DQUOT_LIMITS_ENABLED);
 	return error;
 }
-EXPORT_SYMBOL(vfs_quota_on_path);
+EXPORT_SYMBOL(dquot_quota_on_path);
 
-int vfs_quota_on(struct super_block *sb, int type, int format_id, char *name,
-		 int remount)
+int dquot_quota_on(struct super_block *sb, int type, int format_id, char *name)
 {
 	struct path path;
 	int error;
 
-	if (remount)
-		return vfs_quota_on_remount(sb, type);
-
 	error = kern_path(name, LOOKUP_FOLLOW, &path);
 	if (!error) {
-		error = vfs_quota_on_path(sb, type, format_id, &path);
+		error = dquot_quota_on_path(sb, type, format_id, &path);
 		path_put(&path);
 	}
 	return error;
 }
-EXPORT_SYMBOL(vfs_quota_on);
+EXPORT_SYMBOL(dquot_quota_on);
 
 /*
  * More powerful function for turning on quotas allowing setting
  * of individual quota flags
  */
-int vfs_quota_enable(struct inode *inode, int type, int format_id,
-		unsigned int flags)
+int dquot_enable(struct inode *inode, int type, int format_id,
+		 unsigned int flags)
 {
 	int ret = 0;
 	struct super_block *sb = inode->i_sb;
 	struct quota_info *dqopt = sb_dqopt(sb);
 
 	/* Just unsuspend quotas? */
-	if (flags & DQUOT_SUSPENDED)
-		return vfs_quota_on_remount(sb, type);
+	BUG_ON(flags & DQUOT_SUSPENDED);
+
 	if (!flags)
 		return 0;
 	/* Just updating flags needed? */
@@ -2229,13 +2218,13 @@ out_lock:
 load_quota:
 	return vfs_load_quota_inode(inode, type, format_id, flags);
 }
-EXPORT_SYMBOL(vfs_quota_enable);
+EXPORT_SYMBOL(dquot_enable);
 
 /*
  * This function is used when filesystem needs to initialize quotas
  * during mount time.
  */
-int vfs_quota_on_mount(struct super_block *sb, char *qf_name,
+int dquot_quota_on_mount(struct super_block *sb, char *qf_name,
 		int format_id, int type)
 {
 	struct dentry *dentry;
@@ -2261,24 +2250,7 @@ out:
 	dput(dentry);
 	return error;
 }
-EXPORT_SYMBOL(vfs_quota_on_mount);
-
-/* Wrapper to turn on quotas when remounting rw */
-int vfs_dq_quota_on_remount(struct super_block *sb)
-{
-	int cnt;
-	int ret = 0, err;
-
-	if (!sb->s_qcop || !sb->s_qcop->quota_on)
-		return -ENOSYS;
-	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
-		err = sb->s_qcop->quota_on(sb, cnt, 0, NULL, 1);
-		if (err < 0 && !ret)
-			ret = err;
-	}
-	return ret;
-}
-EXPORT_SYMBOL(vfs_dq_quota_on_remount);
+EXPORT_SYMBOL(dquot_quota_on_mount);
 
 static inline qsize_t qbtos(qsize_t blocks)
 {
@@ -2313,8 +2285,8 @@ static void do_get_dqblk(struct dquot *dquot, struct fs_disk_quota *di)
 	spin_unlock(&dq_data_lock);
 }
 
-int vfs_get_dqblk(struct super_block *sb, int type, qid_t id,
-		  struct fs_disk_quota *di)
+int dquot_get_dqblk(struct super_block *sb, int type, qid_t id,
+		    struct fs_disk_quota *di)
 {
 	struct dquot *dquot;
 
@@ -2326,7 +2298,7 @@ int vfs_get_dqblk(struct super_block *sb, int type, qid_t id,
 
 	return 0;
 }
-EXPORT_SYMBOL(vfs_get_dqblk);
+EXPORT_SYMBOL(dquot_get_dqblk);
 
 #define VFS_FS_DQ_MASK \
 	(FS_DQ_BCOUNT | FS_DQ_BSOFT | FS_DQ_BHARD | \
@@ -2425,7 +2397,7 @@ static int do_set_dqblk(struct dquot *dquot, struct fs_disk_quota *di)
 	return 0;
 }
 
-int vfs_set_dqblk(struct super_block *sb, int type, qid_t id,
+int dquot_set_dqblk(struct super_block *sb, int type, qid_t id,
 		  struct fs_disk_quota *di)
 {
 	struct dquot *dquot;
@@ -2441,10 +2413,10 @@ int vfs_set_dqblk(struct super_block *sb, int type, qid_t id,
 out:
 	return rc;
 }
-EXPORT_SYMBOL(vfs_set_dqblk);
+EXPORT_SYMBOL(dquot_set_dqblk);
 
 /* Generic routine for getting common part of quota file information */
-int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
+int dquot_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 {
 	struct mem_dqinfo *mi;
   
@@ -2463,10 +2435,10 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
-EXPORT_SYMBOL(vfs_get_dqinfo);
+EXPORT_SYMBOL(dquot_get_dqinfo);
 
 /* Generic routine for setting common part of quota file information */
-int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
+int dquot_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 {
 	struct mem_dqinfo *mi;
 	int err = 0;
@@ -2493,27 +2465,27 @@ out:
 	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return err;
 }
-EXPORT_SYMBOL(vfs_set_dqinfo);
+EXPORT_SYMBOL(dquot_set_dqinfo);
 
-const struct quotactl_ops vfs_quotactl_ops = {
-	.quota_on	= vfs_quota_on,
-	.quota_off	= vfs_quota_off,
-	.quota_sync	= vfs_quota_sync,
-	.get_info	= vfs_get_dqinfo,
-	.set_info	= vfs_set_dqinfo,
-	.get_dqblk	= vfs_get_dqblk,
-	.set_dqblk	= vfs_set_dqblk
+const struct quotactl_ops dquot_quotactl_ops = {
+	.quota_on	= dquot_quota_on,
+	.quota_off	= dquot_quota_off,
+	.quota_sync	= dquot_quota_sync,
+	.get_info	= dquot_get_dqinfo,
+	.set_info	= dquot_set_dqinfo,
+	.get_dqblk	= dquot_get_dqblk,
+	.set_dqblk	= dquot_set_dqblk
 };
-
+EXPORT_SYMBOL(dquot_quotactl_ops);
 
 static int do_proc_dqstats(struct ctl_table *table, int write,
 		     void __user *buffer, size_t *lenp, loff_t *ppos)
 {
-#ifdef CONFIG_SMP
-	/* Update global table */
 	unsigned int type = (int *)table->data - dqstats.stat;
-	dqstats.stat[type] = dqstats_read(type);
-#endif
+
+	/* Update global table */
+	dqstats.stat[type] =
+			percpu_counter_sum_positive(&dqstats.counter[type]);
 	return proc_dointvec(table, write, buffer, lenp, ppos);
 }
 
@@ -2606,7 +2578,7 @@ static ctl_table sys_table[] = {
 
 static int __init dquot_init(void)
 {
-	int i;
+	int i, ret;
 	unsigned long nr_hash, order;
 
 	printk(KERN_NOTICE "VFS: Disk quotas %s\n", __DQUOT_VERSION__);
@@ -2624,12 +2596,11 @@ static int __init dquot_init(void)
 	if (!dquot_hash)
 		panic("Cannot create dquot hash table");
 
-#ifdef CONFIG_SMP
-	dqstats_pcpu = alloc_percpu(struct dqstats);
-	if (!dqstats_pcpu)
-		panic("Cannot create dquot stats table");
-#endif
-	memset(&dqstats, 0, sizeof(struct dqstats));
+	for (i = 0; i < _DQST_DQSTAT_LAST; i++) {
+		ret = percpu_counter_init(&dqstats.counter[i], 0);
+		if (ret)
+			panic("Cannot create dquot stat counters");
+	}
 
 	/* Find power-of-two hlist_heads which can fit into allocation */
 	nr_hash = (1UL << order) * PAGE_SIZE / sizeof(struct hlist_head);
diff --git a/fs/quota/quota.c b/fs/quota/quota.c
index ce3dfd066f59..b299961e1edb 100644
--- a/fs/quota/quota.c
+++ b/fs/quota/quota.c
@@ -73,7 +73,7 @@ static int quota_quotaon(struct super_block *sb, int type, int cmd, qid_t id,
 	if (IS_ERR(pathname))
 		return PTR_ERR(pathname);
 	if (sb->s_qcop->quota_on)
-		ret = sb->s_qcop->quota_on(sb, type, id, pathname, 0);
+		ret = sb->s_qcop->quota_on(sb, type, id, pathname);
 	putname(pathname);
 	return ret;
 }
@@ -260,7 +260,7 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id,
 	case Q_QUOTAOFF:
 		if (!sb->s_qcop->quota_off)
 			return -ENOSYS;
-		return sb->s_qcop->quota_off(sb, type, 0);
+		return sb->s_qcop->quota_off(sb, type);
 	case Q_GETFMT:
 		return quota_getfmt(sb, type, addr);
 	case Q_GETINFO:
diff --git a/fs/ramfs/file-mmu.c b/fs/ramfs/file-mmu.c
index 78f613cb9c76..4884ac5ae9be 100644
--- a/fs/ramfs/file-mmu.c
+++ b/fs/ramfs/file-mmu.c
@@ -43,12 +43,13 @@ const struct file_operations ramfs_file_operations = {
 	.write		= do_sync_write,
 	.aio_write	= generic_file_aio_write,
 	.mmap		= generic_file_mmap,
-	.fsync		= simple_sync_file,
+	.fsync		= noop_fsync,
 	.splice_read	= generic_file_splice_read,
 	.splice_write	= generic_file_splice_write,
 	.llseek		= generic_file_llseek,
 };
 
 const struct inode_operations ramfs_file_inode_operations = {
+	.setattr	= simple_setattr,
 	.getattr	= simple_getattr,
 };
diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c
index 5ea4ad81a429..d532c20fc179 100644
--- a/fs/ramfs/file-nommu.c
+++ b/fs/ramfs/file-nommu.c
@@ -42,7 +42,7 @@ const struct file_operations ramfs_file_operations = {
 	.aio_read		= generic_file_aio_read,
 	.write			= do_sync_write,
 	.aio_write		= generic_file_aio_write,
-	.fsync			= simple_sync_file,
+	.fsync			= noop_fsync,
 	.splice_read		= generic_file_splice_read,
 	.splice_write		= generic_file_splice_write,
 	.llseek			= generic_file_llseek,
@@ -146,7 +146,7 @@ static int ramfs_nommu_resize(struct inode *inode, loff_t newsize, loff_t size)
 			return ret;
 	}
 
-	ret = vmtruncate(inode, newsize);
+	ret = simple_setsize(inode, newsize);
 
 	return ret;
 }
@@ -169,7 +169,8 @@ static int ramfs_nommu_setattr(struct dentry *dentry, struct iattr *ia)
 
 	/* pick out size-changing events */
 	if (ia->ia_valid & ATTR_SIZE) {
-		loff_t size = i_size_read(inode);
+		loff_t size = inode->i_size;
+
 		if (ia->ia_size != size) {
 			ret = ramfs_nommu_resize(inode, ia->ia_size, size);
 			if (ret < 0 || ia->ia_valid == ATTR_SIZE)
@@ -182,7 +183,7 @@ static int ramfs_nommu_setattr(struct dentry *dentry, struct iattr *ia)
 		}
 	}
 
-	ret = inode_setattr(inode, ia);
+	generic_setattr(inode, ia);
  out:
 	ia->ia_valid = old_ia_valid;
 	return ret;
diff --git a/fs/read_write.c b/fs/read_write.c
index 113386d6fd2d..9c0485236e68 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
@@ -97,6 +97,23 @@ loff_t generic_file_llseek(struct file *file, loff_t offset, int origin)
 }
 EXPORT_SYMBOL(generic_file_llseek);
 
+/**
+ * noop_llseek - No Operation Performed llseek implementation
+ * @file:	file structure to seek on
+ * @offset:	file offset to seek to
+ * @origin:	type of seek
+ *
+ * This is an implementation of ->llseek useable for the rare special case when
+ * userspace expects the seek to succeed but the (device) file is actually not
+ * able to perform the seek. In this case you use noop_llseek() instead of
+ * falling back to the default implementation of ->llseek.
+ */
+loff_t noop_llseek(struct file *file, loff_t offset, int origin)
+{
+	return file->f_pos;
+}
+EXPORT_SYMBOL(noop_llseek);
+
 loff_t no_llseek(struct file *file, loff_t offset, int origin)
 {
 	return -ESPIPE;
diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c
index 07930449a958..198dabf1b2bb 100644
--- a/fs/reiserfs/dir.c
+++ b/fs/reiserfs/dir.c
@@ -14,10 +14,10 @@
 extern const struct reiserfs_key MIN_KEY;
 
 static int reiserfs_readdir(struct file *, void *, filldir_t);
-static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry,
-			      int datasync);
+static int reiserfs_dir_fsync(struct file *filp, int datasync);
 
 const struct file_operations reiserfs_dir_operations = {
+	.llseek = generic_file_llseek,
 	.read = generic_read_dir,
 	.readdir = reiserfs_readdir,
 	.fsync = reiserfs_dir_fsync,
@@ -27,10 +27,9 @@ const struct file_operations reiserfs_dir_operations = {
 #endif
 };
 
-static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry,
-			      int datasync)
+static int reiserfs_dir_fsync(struct file *filp, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = filp->f_mapping->host;
 	int err;
 	reiserfs_write_lock(inode->i_sb);
 	err = reiserfs_commit_for_inode(inode);
diff --git a/fs/reiserfs/file.c b/fs/reiserfs/file.c
index 9977df9f3a54..b82cdd8a45dd 100644
--- a/fs/reiserfs/file.c
+++ b/fs/reiserfs/file.c
@@ -134,10 +134,9 @@ static void reiserfs_vfs_truncate_file(struct inode *inode)
  * be removed...
  */
 
-static int reiserfs_sync_file(struct file *filp,
-			      struct dentry *dentry, int datasync)
+static int reiserfs_sync_file(struct file *filp, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = filp->f_mapping->host;
 	int err;
 	int barrier_done;
 
diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c
index 59125fb36d42..9822fa15118b 100644
--- a/fs/reiserfs/super.c
+++ b/fs/reiserfs/super.c
@@ -158,6 +158,7 @@ static int finish_unfinished(struct super_block *s)
 #ifdef CONFIG_QUOTA
 	int i;
 	int ms_active_set;
+	int quota_enabled[MAXQUOTAS];
 #endif
 
 	/* compose key to look for "save" links */
@@ -179,8 +180,15 @@ static int finish_unfinished(struct super_block *s)
 	}
 	/* Turn on quotas so that they are updated correctly */
 	for (i = 0; i < MAXQUOTAS; i++) {
+		quota_enabled[i] = 1;
 		if (REISERFS_SB(s)->s_qf_names[i]) {
-			int ret = reiserfs_quota_on_mount(s, i);
+			int ret;
+
+			if (sb_has_quota_active(s, i)) {
+				quota_enabled[i] = 0;
+				continue;
+			}
+			ret = reiserfs_quota_on_mount(s, i);
 			if (ret < 0)
 				reiserfs_warning(s, "reiserfs-2500",
 						 "cannot turn on journaled "
@@ -304,8 +312,8 @@ static int finish_unfinished(struct super_block *s)
 #ifdef CONFIG_QUOTA
 	/* Turn quotas off */
 	for (i = 0; i < MAXQUOTAS; i++) {
-		if (sb_dqopt(s)->files[i])
-			vfs_quota_off(s, i, 0);
+		if (sb_dqopt(s)->files[i] && quota_enabled[i])
+			dquot_quota_off(s, i);
 	}
 	if (ms_active_set)
 		/* Restore the flag back */
@@ -466,6 +474,8 @@ static void reiserfs_put_super(struct super_block *s)
 	struct reiserfs_transaction_handle th;
 	th.t_trans_id = 0;
 
+	dquot_disable(s, -1, DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
+
 	reiserfs_write_lock(s);
 
 	if (s->s_dirt)
@@ -620,7 +630,7 @@ static int reiserfs_acquire_dquot(struct dquot *);
 static int reiserfs_release_dquot(struct dquot *);
 static int reiserfs_mark_dquot_dirty(struct dquot *);
 static int reiserfs_write_info(struct super_block *, int);
-static int reiserfs_quota_on(struct super_block *, int, int, char *, int);
+static int reiserfs_quota_on(struct super_block *, int, int, char *);
 
 static const struct dquot_operations reiserfs_quota_operations = {
 	.write_dquot = reiserfs_write_dquot,
@@ -634,12 +644,12 @@ static const struct dquot_operations reiserfs_quota_operations = {
 
 static const struct quotactl_ops reiserfs_qctl_operations = {
 	.quota_on = reiserfs_quota_on,
-	.quota_off = vfs_quota_off,
-	.quota_sync = vfs_quota_sync,
-	.get_info = vfs_get_dqinfo,
-	.set_info = vfs_set_dqinfo,
-	.get_dqblk = vfs_get_dqblk,
-	.set_dqblk = vfs_set_dqblk,
+	.quota_off = dquot_quota_off,
+	.quota_sync = dquot_quota_sync,
+	.get_info = dquot_get_dqinfo,
+	.set_info = dquot_set_dqinfo,
+	.get_dqblk = dquot_get_dqblk,
+	.set_dqblk = dquot_set_dqblk,
 };
 #endif
 
@@ -1242,6 +1252,11 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg)
 		if (s->s_flags & MS_RDONLY)
 			/* it is read-only already */
 			goto out_ok;
+
+		err = dquot_suspend(s, -1);
+		if (err < 0)
+			goto out_err;
+
 		/* try to remount file system with read-only permissions */
 		if (sb_umount_state(rs) == REISERFS_VALID_FS
 		    || REISERFS_SB(s)->s_mount_state != REISERFS_VALID_FS) {
@@ -1295,6 +1310,7 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg)
 	s->s_dirt = 0;
 
 	if (!(*mount_flags & MS_RDONLY)) {
+		dquot_resume(s, -1);
 		finish_unfinished(s);
 		reiserfs_xattr_init(s, *mount_flags);
 	}
@@ -2022,15 +2038,15 @@ static int reiserfs_write_info(struct super_block *sb, int type)
  */
 static int reiserfs_quota_on_mount(struct super_block *sb, int type)
 {
-	return vfs_quota_on_mount(sb, REISERFS_SB(sb)->s_qf_names[type],
-				  REISERFS_SB(sb)->s_jquota_fmt, type);
+	return dquot_quota_on_mount(sb, REISERFS_SB(sb)->s_qf_names[type],
+					REISERFS_SB(sb)->s_jquota_fmt, type);
 }
 
 /*
  * Standard function to be called on quota_on
  */
 static int reiserfs_quota_on(struct super_block *sb, int type, int format_id,
-			     char *name, int remount)
+			     char *name)
 {
 	int err;
 	struct path path;
@@ -2039,9 +2055,7 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id,
 
 	if (!(REISERFS_SB(sb)->s_mount_opt & (1 << REISERFS_QUOTA)))
 		return -EINVAL;
-	/* No more checks needed? Path and format_id are bogus anyway... */
-	if (remount)
-		return vfs_quota_on(sb, type, format_id, name, 1);
+
 	err = kern_path(name, LOOKUP_FOLLOW, &path);
 	if (err)
 		return err;
@@ -2085,7 +2099,7 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id,
 		if (err)
 			goto out;
 	}
-	err = vfs_quota_on_path(sb, type, format_id, &path);
+	err = dquot_quota_on_path(sb, type, format_id, &path);
 out:
 	path_put(&path);
 	return err;
diff --git a/fs/smbfs/dir.c b/fs/smbfs/dir.c
index 3e4803b4427e..00a70cab1f36 100644
--- a/fs/smbfs/dir.c
+++ b/fs/smbfs/dir.c
@@ -37,9 +37,10 @@ static int smb_link(struct dentry *, struct inode *, struct dentry *);
 
 const struct file_operations smb_dir_operations =
 {
+	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= smb_readdir,
-	.ioctl		= smb_ioctl,
+	.unlocked_ioctl	= smb_ioctl,
 	.open		= smb_dir_open,
 };
 
diff --git a/fs/smbfs/file.c b/fs/smbfs/file.c
index dbf6548bbf06..8e187a0f94bb 100644
--- a/fs/smbfs/file.c
+++ b/fs/smbfs/file.c
@@ -28,8 +28,9 @@
 #include "proto.h"
 
 static int
-smb_fsync(struct file *file, struct dentry * dentry, int datasync)
+smb_fsync(struct file *file, int datasync)
 {
+	struct dentry *dentry = file->f_path.dentry;
 	struct smb_sb_info *server = server_from_dentry(dentry);
 	int result;
 
@@ -437,7 +438,7 @@ const struct file_operations smb_file_operations =
 	.aio_read	= smb_file_aio_read,
 	.write		= do_sync_write,
 	.aio_write	= smb_file_aio_write,
-	.ioctl		= smb_ioctl,
+	.unlocked_ioctl	= smb_ioctl,
 	.mmap		= smb_file_mmap,
 	.open		= smb_file_open,
 	.release	= smb_file_release,
diff --git a/fs/smbfs/inode.c b/fs/smbfs/inode.c
index dfa1d67f8fca..9551cb6f7fe4 100644
--- a/fs/smbfs/inode.c
+++ b/fs/smbfs/inode.c
@@ -714,7 +714,7 @@ smb_notify_change(struct dentry *dentry, struct iattr *attr)
 		error = server->ops->truncate(inode, attr->ia_size);
 		if (error)
 			goto out;
-		error = vmtruncate(inode, attr->ia_size);
+		error = simple_setsize(inode, attr->ia_size);
 		if (error)
 			goto out;
 		refresh = 1;
diff --git a/fs/smbfs/ioctl.c b/fs/smbfs/ioctl.c
index dbae1f8ea26f..07215312ad39 100644
--- a/fs/smbfs/ioctl.c
+++ b/fs/smbfs/ioctl.c
@@ -13,6 +13,7 @@
 #include <linux/time.h>
 #include <linux/mm.h>
 #include <linux/highuid.h>
+#include <linux/smp_lock.h>
 #include <linux/net.h>
 
 #include <linux/smb_fs.h>
@@ -22,14 +23,14 @@
 
 #include "proto.h"
 
-int
-smb_ioctl(struct inode *inode, struct file *filp,
-	  unsigned int cmd, unsigned long arg)
+long
+smb_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
-	struct smb_sb_info *server = server_from_inode(inode);
+	struct smb_sb_info *server = server_from_inode(filp->f_path.dentry->d_inode);
 	struct smb_conn_opt opt;
 	int result = -EINVAL;
 
+	lock_kernel();
 	switch (cmd) {
 		uid16_t uid16;
 		uid_t uid32;
@@ -62,6 +63,7 @@ smb_ioctl(struct inode *inode, struct file *filp,
 	default:
 		break;
 	}
+	unlock_kernel();
 
 	return result;
 }
diff --git a/fs/smbfs/proto.h b/fs/smbfs/proto.h
index 03f456c1b7d4..05939a6f43e6 100644
--- a/fs/smbfs/proto.h
+++ b/fs/smbfs/proto.h
@@ -67,7 +67,7 @@ extern const struct address_space_operations smb_file_aops;
 extern const struct file_operations smb_file_operations;
 extern const struct inode_operations smb_file_inode_operations;
 /* ioctl.c */
-extern int smb_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg);
+extern long smb_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
 /* smbiod.c */
 extern void smbiod_wake_up(void);
 extern int smbiod_register_server(struct smb_sb_info *server);
diff --git a/fs/smbfs/symlink.c b/fs/smbfs/symlink.c
index 54350b59046b..00b2909bd469 100644
--- a/fs/smbfs/symlink.c
+++ b/fs/smbfs/symlink.c
@@ -15,7 +15,6 @@
 #include <linux/pagemap.h>
 #include <linux/net.h>
 #include <linux/namei.h>
-#include <linux/slab.h>
 
 #include <asm/uaccess.h>
 #include <asm/system.h>
diff --git a/fs/squashfs/Kconfig b/fs/squashfs/Kconfig
index 25a00d19d686..cc6ce8a84c21 100644
--- a/fs/squashfs/Kconfig
+++ b/fs/squashfs/Kconfig
@@ -26,6 +26,17 @@ config SQUASHFS
 
 	  If unsure, say N.
 
+config SQUASHFS_XATTRS
+	bool "Squashfs XATTR support"
+	depends on SQUASHFS
+	default n
+	help
+	  Saying Y here includes support for extended attributes (xattrs).
+	  Xattrs are name:value pairs associated with inodes by
+	  the kernel or by users (see the attr(5) manual page).
+
+	  If unsure, say N.
+
 config SQUASHFS_EMBEDDED
 
 	bool "Additional option for memory-constrained systems" 
diff --git a/fs/squashfs/Makefile b/fs/squashfs/Makefile
index df8a19ef870d..2cee3e9fa452 100644
--- a/fs/squashfs/Makefile
+++ b/fs/squashfs/Makefile
@@ -5,3 +5,5 @@
 obj-$(CONFIG_SQUASHFS) += squashfs.o
 squashfs-y += block.o cache.o dir.o export.o file.o fragment.o id.o inode.o
 squashfs-y += namei.o super.o symlink.o zlib_wrapper.o decompressor.o
+squashfs-$(CONFIG_SQUASHFS_XATTRS) += xattr.o xattr_id.o
+
diff --git a/fs/squashfs/inode.c b/fs/squashfs/inode.c
index 49daaf669e41..62e63ad25075 100644
--- a/fs/squashfs/inode.c
+++ b/fs/squashfs/inode.c
@@ -40,11 +40,13 @@
 
 #include <linux/fs.h>
 #include <linux/vfs.h>
+#include <linux/xattr.h>
 
 #include "squashfs_fs.h"
 #include "squashfs_fs_sb.h"
 #include "squashfs_fs_i.h"
 #include "squashfs.h"
+#include "xattr.h"
 
 /*
  * Initialise VFS inode with the base inode information common to all
@@ -111,6 +113,7 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 	int err, type, offset = SQUASHFS_INODE_OFFSET(ino);
 	union squashfs_inode squashfs_ino;
 	struct squashfs_base_inode *sqshb_ino = &squashfs_ino.base;
+	int xattr_id = SQUASHFS_INVALID_XATTR;
 
 	TRACE("Entered squashfs_read_inode\n");
 
@@ -199,8 +202,10 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 			frag_offset = 0;
 		}
 
+		xattr_id = le32_to_cpu(sqsh_ino->xattr);
 		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
 		inode->i_size = le64_to_cpu(sqsh_ino->file_size);
+		inode->i_op = &squashfs_inode_ops;
 		inode->i_fop = &generic_ro_fops;
 		inode->i_mode |= S_IFREG;
 		inode->i_blocks = ((inode->i_size -
@@ -251,6 +256,7 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 		if (err < 0)
 			goto failed_read;
 
+		xattr_id = le32_to_cpu(sqsh_ino->xattr);
 		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
 		inode->i_size = le32_to_cpu(sqsh_ino->file_size);
 		inode->i_op = &squashfs_dir_inode_ops;
@@ -280,21 +286,33 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 
 		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
 		inode->i_size = le32_to_cpu(sqsh_ino->symlink_size);
-		inode->i_op = &page_symlink_inode_operations;
+		inode->i_op = &squashfs_symlink_inode_ops;
 		inode->i_data.a_ops = &squashfs_symlink_aops;
 		inode->i_mode |= S_IFLNK;
 		squashfs_i(inode)->start = block;
 		squashfs_i(inode)->offset = offset;
 
+		if (type == SQUASHFS_LSYMLINK_TYPE) {
+			__le32 xattr;
+
+			err = squashfs_read_metadata(sb, NULL, &block,
+						&offset, inode->i_size);
+			if (err < 0)
+				goto failed_read;
+			err = squashfs_read_metadata(sb, &xattr, &block,
+						&offset, sizeof(xattr));
+			if (err < 0)
+				goto failed_read;
+			xattr_id = le32_to_cpu(xattr);
+		}
+
 		TRACE("Symbolic link inode %x:%x, start_block %llx, offset "
 				"%x\n", SQUASHFS_INODE_BLK(ino), offset,
 				block, offset);
 		break;
 	}
 	case SQUASHFS_BLKDEV_TYPE:
-	case SQUASHFS_CHRDEV_TYPE:
-	case SQUASHFS_LBLKDEV_TYPE:
-	case SQUASHFS_LCHRDEV_TYPE: {
+	case SQUASHFS_CHRDEV_TYPE: {
 		struct squashfs_dev_inode *sqsh_ino = &squashfs_ino.dev;
 		unsigned int rdev;
 
@@ -315,10 +333,32 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 				SQUASHFS_INODE_BLK(ino), offset, rdev);
 		break;
 	}
+	case SQUASHFS_LBLKDEV_TYPE:
+	case SQUASHFS_LCHRDEV_TYPE: {
+		struct squashfs_ldev_inode *sqsh_ino = &squashfs_ino.ldev;
+		unsigned int rdev;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+				sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		if (type == SQUASHFS_LCHRDEV_TYPE)
+			inode->i_mode |= S_IFCHR;
+		else
+			inode->i_mode |= S_IFBLK;
+		xattr_id = le32_to_cpu(sqsh_ino->xattr);
+		inode->i_op = &squashfs_inode_ops;
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		rdev = le32_to_cpu(sqsh_ino->rdev);
+		init_special_inode(inode, inode->i_mode, new_decode_dev(rdev));
+
+		TRACE("Device inode %x:%x, rdev %x\n",
+				SQUASHFS_INODE_BLK(ino), offset, rdev);
+		break;
+	}
 	case SQUASHFS_FIFO_TYPE:
-	case SQUASHFS_SOCKET_TYPE:
-	case SQUASHFS_LFIFO_TYPE:
-	case SQUASHFS_LSOCKET_TYPE: {
+	case SQUASHFS_SOCKET_TYPE: {
 		struct squashfs_ipc_inode *sqsh_ino = &squashfs_ino.ipc;
 
 		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
@@ -334,14 +374,52 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 		init_special_inode(inode, inode->i_mode, 0);
 		break;
 	}
+	case SQUASHFS_LFIFO_TYPE:
+	case SQUASHFS_LSOCKET_TYPE: {
+		struct squashfs_lipc_inode *sqsh_ino = &squashfs_ino.lipc;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+				sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		if (type == SQUASHFS_LFIFO_TYPE)
+			inode->i_mode |= S_IFIFO;
+		else
+			inode->i_mode |= S_IFSOCK;
+		xattr_id = le32_to_cpu(sqsh_ino->xattr);
+		inode->i_op = &squashfs_inode_ops;
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		init_special_inode(inode, inode->i_mode, 0);
+		break;
+	}
 	default:
 		ERROR("Unknown inode type %d in squashfs_iget!\n", type);
 		return -EINVAL;
 	}
 
+	if (xattr_id != SQUASHFS_INVALID_XATTR && msblk->xattr_id_table) {
+		err = squashfs_xattr_lookup(sb, xattr_id,
+					&squashfs_i(inode)->xattr_count,
+					&squashfs_i(inode)->xattr_size,
+					&squashfs_i(inode)->xattr);
+		if (err < 0)
+			goto failed_read;
+		inode->i_blocks += ((squashfs_i(inode)->xattr_size - 1) >> 9)
+				+ 1;
+	} else
+		squashfs_i(inode)->xattr_count = 0;
+
 	return 0;
 
 failed_read:
 	ERROR("Unable to read inode 0x%llx\n", ino);
 	return err;
 }
+
+
+const struct inode_operations squashfs_inode_ops = {
+	.getxattr = generic_getxattr,
+	.listxattr = squashfs_listxattr
+};
+
diff --git a/fs/squashfs/namei.c b/fs/squashfs/namei.c
index 5266bd8ad932..7a9464d08cf6 100644
--- a/fs/squashfs/namei.c
+++ b/fs/squashfs/namei.c
@@ -57,11 +57,13 @@
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/dcache.h>
+#include <linux/xattr.h>
 
 #include "squashfs_fs.h"
 #include "squashfs_fs_sb.h"
 #include "squashfs_fs_i.h"
 #include "squashfs.h"
+#include "xattr.h"
 
 /*
  * Lookup name in the directory index, returning the location of the metadata
@@ -237,5 +239,7 @@ failed:
 
 
 const struct inode_operations squashfs_dir_inode_ops = {
-	.lookup = squashfs_lookup
+	.lookup = squashfs_lookup,
+	.getxattr = generic_getxattr,
+	.listxattr = squashfs_listxattr
 };
diff --git a/fs/squashfs/squashfs.h b/fs/squashfs/squashfs.h
index fe2587af5512..733a17c42945 100644
--- a/fs/squashfs/squashfs.h
+++ b/fs/squashfs/squashfs.h
@@ -73,8 +73,11 @@ extern struct inode *squashfs_iget(struct super_block *, long long,
 				unsigned int);
 extern int squashfs_read_inode(struct inode *, long long);
 
+/* xattr.c */
+extern ssize_t squashfs_listxattr(struct dentry *, char *, size_t);
+
 /*
- * Inodes, files and decompressor operations
+ * Inodes, files,  decompressor and xattr operations
  */
 
 /* dir.c */
@@ -86,11 +89,18 @@ extern const struct export_operations squashfs_export_ops;
 /* file.c */
 extern const struct address_space_operations squashfs_aops;
 
+/* inode.c */
+extern const struct inode_operations squashfs_inode_ops;
+
 /* namei.c */
 extern const struct inode_operations squashfs_dir_inode_ops;
 
 /* symlink.c */
 extern const struct address_space_operations squashfs_symlink_aops;
+extern const struct inode_operations squashfs_symlink_inode_ops;
+
+/* xattr.c */
+extern const struct xattr_handler *squashfs_xattr_handlers[];
 
 /* zlib_wrapper.c */
 extern const struct squashfs_decompressor squashfs_zlib_comp_ops;
diff --git a/fs/squashfs/squashfs_fs.h b/fs/squashfs/squashfs_fs.h
index 79024245ea00..8eabb808b78d 100644
--- a/fs/squashfs/squashfs_fs.h
+++ b/fs/squashfs/squashfs_fs.h
@@ -46,6 +46,7 @@
 #define SQUASHFS_NAME_LEN		256
 
 #define SQUASHFS_INVALID_FRAG		(0xffffffffU)
+#define SQUASHFS_INVALID_XATTR		(0xffffffffU)
 #define SQUASHFS_INVALID_BLK		(-1LL)
 
 /* Filesystem flags */
@@ -96,6 +97,13 @@
 #define SQUASHFS_LFIFO_TYPE		13
 #define SQUASHFS_LSOCKET_TYPE		14
 
+/* Xattr types */
+#define SQUASHFS_XATTR_USER             0
+#define SQUASHFS_XATTR_TRUSTED          1
+#define SQUASHFS_XATTR_SECURITY         2
+#define SQUASHFS_XATTR_VALUE_OOL        256
+#define SQUASHFS_XATTR_PREFIX_MASK      0xff
+
 /* Flag whether block is compressed or uncompressed, bit is set if block is
  * uncompressed */
 #define SQUASHFS_COMPRESSED_BIT		(1 << 15)
@@ -174,6 +182,24 @@
 
 #define SQUASHFS_ID_BLOCK_BYTES(A)	(SQUASHFS_ID_BLOCKS(A) *\
 					sizeof(u64))
+/* xattr id lookup table defines */
+#define SQUASHFS_XATTR_BYTES(A)		((A) * sizeof(struct squashfs_xattr_id))
+
+#define SQUASHFS_XATTR_BLOCK(A)		(SQUASHFS_XATTR_BYTES(A) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_XATTR_BLOCK_OFFSET(A)	(SQUASHFS_XATTR_BYTES(A) % \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_XATTR_BLOCKS(A)	((SQUASHFS_XATTR_BYTES(A) + \
+					SQUASHFS_METADATA_SIZE - 1) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_XATTR_BLOCK_BYTES(A)	(SQUASHFS_XATTR_BLOCKS(A) *\
+					sizeof(u64))
+#define SQUASHFS_XATTR_BLK(A)		((unsigned int) ((A) >> 16))
+
+#define SQUASHFS_XATTR_OFFSET(A)	((unsigned int) ((A) & 0xffff))
 
 /* cached data constants for filesystem */
 #define SQUASHFS_CACHED_BLKS		8
@@ -228,7 +254,7 @@ struct squashfs_super_block {
 	__le64			root_inode;
 	__le64			bytes_used;
 	__le64			id_table_start;
-	__le64			xattr_table_start;
+	__le64			xattr_id_table_start;
 	__le64			inode_table_start;
 	__le64			directory_table_start;
 	__le64			fragment_table_start;
@@ -261,6 +287,17 @@ struct squashfs_ipc_inode {
 	__le32			nlink;
 };
 
+struct squashfs_lipc_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			nlink;
+	__le32			xattr;
+};
+
 struct squashfs_dev_inode {
 	__le16			inode_type;
 	__le16			mode;
@@ -272,6 +309,18 @@ struct squashfs_dev_inode {
 	__le32			rdev;
 };
 
+struct squashfs_ldev_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			nlink;
+	__le32			rdev;
+	__le32			xattr;
+};
+
 struct squashfs_symlink_inode {
 	__le16			inode_type;
 	__le16			mode;
@@ -349,12 +398,14 @@ struct squashfs_ldir_inode {
 union squashfs_inode {
 	struct squashfs_base_inode		base;
 	struct squashfs_dev_inode		dev;
+	struct squashfs_ldev_inode		ldev;
 	struct squashfs_symlink_inode		symlink;
 	struct squashfs_reg_inode		reg;
 	struct squashfs_lreg_inode		lreg;
 	struct squashfs_dir_inode		dir;
 	struct squashfs_ldir_inode		ldir;
 	struct squashfs_ipc_inode		ipc;
+	struct squashfs_lipc_inode		lipc;
 };
 
 struct squashfs_dir_entry {
@@ -377,4 +428,27 @@ struct squashfs_fragment_entry {
 	unsigned int		unused;
 };
 
+struct squashfs_xattr_entry {
+	__le16			type;
+	__le16			size;
+	char			data[0];
+};
+
+struct squashfs_xattr_val {
+	__le32			vsize;
+	char			value[0];
+};
+
+struct squashfs_xattr_id {
+	__le64			xattr;
+	__le32			count;
+	__le32			size;
+};
+
+struct squashfs_xattr_id_table {
+	__le64			xattr_table_start;
+	__le32			xattr_ids;
+	__le32			unused;
+};
+
 #endif
diff --git a/fs/squashfs/squashfs_fs_i.h b/fs/squashfs/squashfs_fs_i.h
index fbfca30c0c68..d3e3a37f28a1 100644
--- a/fs/squashfs/squashfs_fs_i.h
+++ b/fs/squashfs/squashfs_fs_i.h
@@ -26,6 +26,9 @@
 struct squashfs_inode_info {
 	u64		start;
 	int		offset;
+	u64		xattr;
+	unsigned int	xattr_size;
+	int		xattr_count;
 	union {
 		struct {
 			u64		fragment_block;
diff --git a/fs/squashfs/squashfs_fs_sb.h b/fs/squashfs/squashfs_fs_sb.h
index 2e77dc547e25..d9037a5215f0 100644
--- a/fs/squashfs/squashfs_fs_sb.h
+++ b/fs/squashfs/squashfs_fs_sb.h
@@ -61,6 +61,7 @@ struct squashfs_sb_info {
 	int					next_meta_index;
 	__le64					*id_table;
 	__le64					*fragment_index;
+	__le64					*xattr_id_table;
 	struct mutex				read_data_mutex;
 	struct mutex				meta_index_mutex;
 	struct meta_index			*meta_index;
@@ -68,9 +69,11 @@ struct squashfs_sb_info {
 	__le64					*inode_lookup_table;
 	u64					inode_table;
 	u64					directory_table;
+	u64					xattr_table;
 	unsigned int				block_size;
 	unsigned short				block_log;
 	long long				bytes_used;
 	unsigned int				inodes;
+	int					xattr_ids;
 };
 #endif
diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c
index 48b6f4a385a6..88b4f8606652 100644
--- a/fs/squashfs/super.c
+++ b/fs/squashfs/super.c
@@ -36,12 +36,14 @@
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/magic.h>
+#include <linux/xattr.h>
 
 #include "squashfs_fs.h"
 #include "squashfs_fs_sb.h"
 #include "squashfs_fs_i.h"
 #include "squashfs.h"
 #include "decompressor.h"
+#include "xattr.h"
 
 static struct file_system_type squashfs_fs_type;
 static const struct super_operations squashfs_super_ops;
@@ -82,7 +84,7 @@ static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
 	long long root_inode;
 	unsigned short flags;
 	unsigned int fragments;
-	u64 lookup_table_start;
+	u64 lookup_table_start, xattr_id_table_start;
 	int err;
 
 	TRACE("Entered squashfs_fill_superblock\n");
@@ -139,13 +141,6 @@ static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
 	if (msblk->decompressor == NULL)
 		goto failed_mount;
 
-	/*
-	 * Check if there's xattrs in the filesystem.  These are not
-	 * supported in this version, so warn that they will be ignored.
-	 */
-	if (le64_to_cpu(sblk->xattr_table_start) != SQUASHFS_INVALID_BLK)
-		ERROR("Xattrs in filesystem, these will be ignored\n");
-
 	/* Check the filesystem does not extend beyond the end of the
 	   block device */
 	msblk->bytes_used = le64_to_cpu(sblk->bytes_used);
@@ -253,7 +248,7 @@ static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
 allocate_lookup_table:
 	lookup_table_start = le64_to_cpu(sblk->lookup_table_start);
 	if (lookup_table_start == SQUASHFS_INVALID_BLK)
-		goto allocate_root;
+		goto allocate_xattr_table;
 
 	/* Allocate and read inode lookup table */
 	msblk->inode_lookup_table = squashfs_read_inode_lookup_table(sb,
@@ -266,6 +261,21 @@ allocate_lookup_table:
 
 	sb->s_export_op = &squashfs_export_ops;
 
+allocate_xattr_table:
+	sb->s_xattr = squashfs_xattr_handlers;
+	xattr_id_table_start = le64_to_cpu(sblk->xattr_id_table_start);
+	if (xattr_id_table_start == SQUASHFS_INVALID_BLK)
+		goto allocate_root;
+
+	/* Allocate and read xattr id lookup table */
+	msblk->xattr_id_table = squashfs_read_xattr_id_table(sb,
+		xattr_id_table_start, &msblk->xattr_table, &msblk->xattr_ids);
+	if (IS_ERR(msblk->xattr_id_table)) {
+		err = PTR_ERR(msblk->xattr_id_table);
+		msblk->xattr_id_table = NULL;
+		if (err != -ENOTSUPP)
+			goto failed_mount;
+	}
 allocate_root:
 	root = new_inode(sb);
 	if (!root) {
@@ -301,6 +311,7 @@ failed_mount:
 	kfree(msblk->inode_lookup_table);
 	kfree(msblk->fragment_index);
 	kfree(msblk->id_table);
+	kfree(msblk->xattr_id_table);
 	kfree(sb->s_fs_info);
 	sb->s_fs_info = NULL;
 	kfree(sblk);
@@ -355,6 +366,7 @@ static void squashfs_put_super(struct super_block *sb)
 		kfree(sbi->fragment_index);
 		kfree(sbi->meta_index);
 		kfree(sbi->inode_lookup_table);
+		kfree(sbi->xattr_id_table);
 		kfree(sb->s_fs_info);
 		sb->s_fs_info = NULL;
 	}
diff --git a/fs/squashfs/symlink.c b/fs/squashfs/symlink.c
index 32b911f4ee39..ec86434921e1 100644
--- a/fs/squashfs/symlink.c
+++ b/fs/squashfs/symlink.c
@@ -35,11 +35,13 @@
 #include <linux/kernel.h>
 #include <linux/string.h>
 #include <linux/pagemap.h>
+#include <linux/xattr.h>
 
 #include "squashfs_fs.h"
 #include "squashfs_fs_sb.h"
 #include "squashfs_fs_i.h"
 #include "squashfs.h"
+#include "xattr.h"
 
 static int squashfs_symlink_readpage(struct file *file, struct page *page)
 {
@@ -114,3 +116,12 @@ error_out:
 const struct address_space_operations squashfs_symlink_aops = {
 	.readpage = squashfs_symlink_readpage
 };
+
+const struct inode_operations squashfs_symlink_inode_ops = {
+	.readlink = generic_readlink,
+	.follow_link = page_follow_link_light,
+	.put_link = page_put_link,
+	.getxattr = generic_getxattr,
+	.listxattr = squashfs_listxattr
+};
+
diff --git a/fs/squashfs/xattr.c b/fs/squashfs/xattr.c
new file mode 100644
index 000000000000..c7655e8b31cd
--- /dev/null
+++ b/fs/squashfs/xattr.c
@@ -0,0 +1,323 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2010
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * xattr_id.c
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/string.h>
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/xattr.h>
+#include <linux/slab.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+static const struct xattr_handler *squashfs_xattr_handler(int);
+
+ssize_t squashfs_listxattr(struct dentry *d, char *buffer,
+	size_t buffer_size)
+{
+	struct inode *inode = d->d_inode;
+	struct super_block *sb = inode->i_sb;
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	u64 start = SQUASHFS_XATTR_BLK(squashfs_i(inode)->xattr)
+						 + msblk->xattr_table;
+	int offset = SQUASHFS_XATTR_OFFSET(squashfs_i(inode)->xattr);
+	int count = squashfs_i(inode)->xattr_count;
+	size_t rest = buffer_size;
+	int err;
+
+	/* check that the file system has xattrs */
+	if (msblk->xattr_id_table == NULL)
+		return -EOPNOTSUPP;
+
+	/* loop reading each xattr name */
+	while (count--) {
+		struct squashfs_xattr_entry entry;
+		struct squashfs_xattr_val val;
+		const struct xattr_handler *handler;
+		int name_size, prefix_size = 0;
+
+		err = squashfs_read_metadata(sb, &entry, &start, &offset,
+							sizeof(entry));
+		if (err < 0)
+			goto failed;
+
+		name_size = le16_to_cpu(entry.size);
+		handler = squashfs_xattr_handler(le16_to_cpu(entry.type));
+		if (handler)
+			prefix_size = handler->list(d, buffer, rest, NULL,
+				name_size, handler->flags);
+		if (prefix_size) {
+			if (buffer) {
+				if (prefix_size + name_size + 1 > rest) {
+					err = -ERANGE;
+					goto failed;
+				}
+				buffer += prefix_size;
+			}
+			err = squashfs_read_metadata(sb, buffer, &start,
+				&offset, name_size);
+			if (err < 0)
+				goto failed;
+			if (buffer) {
+				buffer[name_size] = '\0';
+				buffer += name_size + 1;
+			}
+			rest -= prefix_size + name_size + 1;
+		} else  {
+			/* no handler or insuffficient privileges, so skip */
+			err = squashfs_read_metadata(sb, NULL, &start,
+				&offset, name_size);
+			if (err < 0)
+				goto failed;
+		}
+
+
+		/* skip remaining xattr entry */
+		err = squashfs_read_metadata(sb, &val, &start, &offset,
+						sizeof(val));
+		if (err < 0)
+			goto failed;
+
+		err = squashfs_read_metadata(sb, NULL, &start, &offset,
+						le32_to_cpu(val.vsize));
+		if (err < 0)
+			goto failed;
+	}
+	err = buffer_size - rest;
+
+failed:
+	return err;
+}
+
+
+static int squashfs_xattr_get(struct inode *inode, int name_index,
+	const char *name, void *buffer, size_t buffer_size)
+{
+	struct super_block *sb = inode->i_sb;
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	u64 start = SQUASHFS_XATTR_BLK(squashfs_i(inode)->xattr)
+						 + msblk->xattr_table;
+	int offset = SQUASHFS_XATTR_OFFSET(squashfs_i(inode)->xattr);
+	int count = squashfs_i(inode)->xattr_count;
+	int name_len = strlen(name);
+	int err, vsize;
+	char *target = kmalloc(name_len, GFP_KERNEL);
+
+	if (target == NULL)
+		return  -ENOMEM;
+
+	/* loop reading each xattr name */
+	for (; count; count--) {
+		struct squashfs_xattr_entry entry;
+		struct squashfs_xattr_val val;
+		int type, prefix, name_size;
+
+		err = squashfs_read_metadata(sb, &entry, &start, &offset,
+							sizeof(entry));
+		if (err < 0)
+			goto failed;
+
+		name_size = le16_to_cpu(entry.size);
+		type = le16_to_cpu(entry.type);
+		prefix = type & SQUASHFS_XATTR_PREFIX_MASK;
+
+		if (prefix == name_index && name_size == name_len)
+			err = squashfs_read_metadata(sb, target, &start,
+						&offset, name_size);
+		else
+			err = squashfs_read_metadata(sb, NULL, &start,
+						&offset, name_size);
+		if (err < 0)
+			goto failed;
+
+		if (prefix == name_index && name_size == name_len &&
+					strncmp(target, name, name_size) == 0) {
+			/* found xattr */
+			if (type & SQUASHFS_XATTR_VALUE_OOL) {
+				__le64 xattr;
+				/* val is a reference to the real location */
+				err = squashfs_read_metadata(sb, &val, &start,
+						&offset, sizeof(val));
+				if (err < 0)
+					goto failed;
+				err = squashfs_read_metadata(sb, &xattr, &start,
+					 &offset, sizeof(xattr));
+				if (err < 0)
+					goto failed;
+				xattr = le64_to_cpu(xattr);
+				start = SQUASHFS_XATTR_BLK(xattr) +
+							msblk->xattr_table;
+				offset = SQUASHFS_XATTR_OFFSET(xattr);
+			}
+			/* read xattr value */
+			err = squashfs_read_metadata(sb, &val, &start, &offset,
+							sizeof(val));
+			if (err < 0)
+				goto failed;
+
+			vsize = le32_to_cpu(val.vsize);
+			if (buffer) {
+				if (vsize > buffer_size) {
+					err = -ERANGE;
+					goto failed;
+				}
+				err = squashfs_read_metadata(sb, buffer, &start,
+					 &offset, vsize);
+				if (err < 0)
+					goto failed;
+			}
+			break;
+		}
+
+		/* no match, skip remaining xattr entry */
+		err = squashfs_read_metadata(sb, &val, &start, &offset,
+							sizeof(val));
+		if (err < 0)
+			goto failed;
+		err = squashfs_read_metadata(sb, NULL, &start, &offset,
+						le32_to_cpu(val.vsize));
+		if (err < 0)
+			goto failed;
+	}
+	err = count ? vsize : -ENODATA;
+
+failed:
+	kfree(target);
+	return err;
+}
+
+
+/*
+ * User namespace support
+ */
+static size_t squashfs_user_list(struct dentry *d, char *list, size_t list_size,
+	const char *name, size_t name_len, int type)
+{
+	if (list && XATTR_USER_PREFIX_LEN <= list_size)
+		memcpy(list, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN);
+	return XATTR_USER_PREFIX_LEN;
+}
+
+static int squashfs_user_get(struct dentry *d, const char *name, void *buffer,
+	size_t size, int type)
+{
+	if (name[0] == '\0')
+		return  -EINVAL;
+
+	return squashfs_xattr_get(d->d_inode, SQUASHFS_XATTR_USER, name,
+		buffer, size);
+}
+
+static const struct xattr_handler squashfs_xattr_user_handler = {
+	.prefix	= XATTR_USER_PREFIX,
+	.list	= squashfs_user_list,
+	.get	= squashfs_user_get
+};
+
+/*
+ * Trusted namespace support
+ */
+static size_t squashfs_trusted_list(struct dentry *d, char *list,
+	size_t list_size, const char *name, size_t name_len, int type)
+{
+	if (!capable(CAP_SYS_ADMIN))
+		return 0;
+
+	if (list && XATTR_TRUSTED_PREFIX_LEN <= list_size)
+		memcpy(list, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN);
+	return XATTR_TRUSTED_PREFIX_LEN;
+}
+
+static int squashfs_trusted_get(struct dentry *d, const char *name,
+	void *buffer, size_t size, int type)
+{
+	if (name[0] == '\0')
+		return  -EINVAL;
+
+	return squashfs_xattr_get(d->d_inode, SQUASHFS_XATTR_TRUSTED, name,
+		buffer, size);
+}
+
+static const struct xattr_handler squashfs_xattr_trusted_handler = {
+	.prefix	= XATTR_TRUSTED_PREFIX,
+	.list	= squashfs_trusted_list,
+	.get	= squashfs_trusted_get
+};
+
+/*
+ * Security namespace support
+ */
+static size_t squashfs_security_list(struct dentry *d, char *list,
+	size_t list_size, const char *name, size_t name_len, int type)
+{
+	if (list && XATTR_SECURITY_PREFIX_LEN <= list_size)
+		memcpy(list, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN);
+	return XATTR_SECURITY_PREFIX_LEN;
+}
+
+static int squashfs_security_get(struct dentry *d, const char *name,
+	void *buffer, size_t size, int type)
+{
+	if (name[0] == '\0')
+		return  -EINVAL;
+
+	return squashfs_xattr_get(d->d_inode, SQUASHFS_XATTR_SECURITY, name,
+		buffer, size);
+}
+
+static const struct xattr_handler squashfs_xattr_security_handler = {
+	.prefix	= XATTR_SECURITY_PREFIX,
+	.list	= squashfs_security_list,
+	.get	= squashfs_security_get
+};
+
+static inline const struct xattr_handler *squashfs_xattr_handler(int type)
+{
+	if (type & ~(SQUASHFS_XATTR_PREFIX_MASK | SQUASHFS_XATTR_VALUE_OOL))
+		/* ignore unrecognised type */
+		return NULL;
+
+	switch (type & SQUASHFS_XATTR_PREFIX_MASK) {
+	case SQUASHFS_XATTR_USER:
+		return &squashfs_xattr_user_handler;
+	case SQUASHFS_XATTR_TRUSTED:
+		return &squashfs_xattr_trusted_handler;
+	case SQUASHFS_XATTR_SECURITY:
+		return &squashfs_xattr_security_handler;
+	default:
+		/* ignore unrecognised type */
+		return NULL;
+	}
+}
+
+const struct xattr_handler *squashfs_xattr_handlers[] = {
+	&squashfs_xattr_user_handler,
+	&squashfs_xattr_trusted_handler,
+	&squashfs_xattr_security_handler,
+	NULL
+};
+
diff --git a/fs/squashfs/xattr.h b/fs/squashfs/xattr.h
new file mode 100644
index 000000000000..9da071ae181c
--- /dev/null
+++ b/fs/squashfs/xattr.h
@@ -0,0 +1,46 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2010
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * xattr.h
+ */
+
+#ifdef CONFIG_SQUASHFS_XATTRS
+extern __le64 *squashfs_read_xattr_id_table(struct super_block *, u64,
+		u64 *, int *);
+extern int squashfs_xattr_lookup(struct super_block *, unsigned int, int *,
+		int *, unsigned long long *);
+#else
+static inline __le64 *squashfs_read_xattr_id_table(struct super_block *sb,
+		u64 start, u64 *xattr_table_start, int *xattr_ids)
+{
+	ERROR("Xattrs in filesystem, these will be ignored\n");
+	return ERR_PTR(-ENOTSUPP);
+}
+
+static inline int squashfs_xattr_lookup(struct super_block *sb,
+		unsigned int index, int *count, int *size,
+		unsigned long long *xattr)
+{
+	return 0;
+}
+#define squashfs_listxattr NULL
+#define generic_getxattr NULL
+#define squashfs_xattr_handlers NULL
+#endif
diff --git a/fs/squashfs/xattr_id.c b/fs/squashfs/xattr_id.c
new file mode 100644
index 000000000000..cfb41106098f
--- /dev/null
+++ b/fs/squashfs/xattr_id.c
@@ -0,0 +1,100 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2010
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * xattr_id.c
+ */
+
+/*
+ * This file implements code to map the 32-bit xattr id stored in the inode
+ * into the on disk location of the xattr data.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Map xattr id using the xattr id look up table
+ */
+int squashfs_xattr_lookup(struct super_block *sb, unsigned int index,
+		int *count, unsigned int *size, unsigned long long *xattr)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int block = SQUASHFS_XATTR_BLOCK(index);
+	int offset = SQUASHFS_XATTR_BLOCK_OFFSET(index);
+	u64 start_block = le64_to_cpu(msblk->xattr_id_table[block]);
+	struct squashfs_xattr_id id;
+	int err;
+
+	err = squashfs_read_metadata(sb, &id, &start_block, &offset,
+							sizeof(id));
+	if (err < 0)
+		return err;
+
+	*xattr = le64_to_cpu(id.xattr);
+	*size = le32_to_cpu(id.size);
+	*count = le32_to_cpu(id.count);
+	return 0;
+}
+
+
+/*
+ * Read uncompressed xattr id lookup table indexes from disk into memory
+ */
+__le64 *squashfs_read_xattr_id_table(struct super_block *sb, u64 start,
+		u64 *xattr_table_start, int *xattr_ids)
+{
+	unsigned int len;
+	__le64 *xid_table;
+	struct squashfs_xattr_id_table id_table;
+	int err;
+
+	err = squashfs_read_table(sb, &id_table, start, sizeof(id_table));
+	if (err < 0) {
+		ERROR("unable to read xattr id table\n");
+		return ERR_PTR(err);
+	}
+	*xattr_table_start = le64_to_cpu(id_table.xattr_table_start);
+	*xattr_ids = le32_to_cpu(id_table.xattr_ids);
+	len = SQUASHFS_XATTR_BLOCK_BYTES(*xattr_ids);
+
+	TRACE("In read_xattr_index_table, length %d\n", len);
+
+	/* Allocate xattr id lookup table indexes */
+	xid_table = kmalloc(len, GFP_KERNEL);
+	if (xid_table == NULL) {
+		ERROR("Failed to allocate xattr id index table\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	err = squashfs_read_table(sb, xid_table, start + sizeof(id_table), len);
+	if (err < 0) {
+		ERROR("unable to read xattr id index table\n");
+		kfree(xid_table);
+		return ERR_PTR(err);
+	}
+
+	return xid_table;
+}
diff --git a/fs/super.c b/fs/super.c
index 69688b15f1fa..5c35bc7a499e 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -24,7 +24,6 @@
 #include <linux/slab.h>
 #include <linux/acct.h>
 #include <linux/blkdev.h>
-#include <linux/quotaops.h>
 #include <linux/mount.h>
 #include <linux/security.h>
 #include <linux/writeback.h>		/* for the emergency remount stuff */
@@ -94,8 +93,6 @@ static struct super_block *alloc_super(struct file_system_type *type)
 		init_rwsem(&s->s_dquot.dqptr_sem);
 		init_waitqueue_head(&s->s_wait_unfrozen);
 		s->s_maxbytes = MAX_NON_LFS;
-		s->dq_op = sb_dquot_ops;
-		s->s_qcop = sb_quotactl_ops;
 		s->s_op = &default_op;
 		s->s_time_gran = 1000000000;
 	}
@@ -160,7 +157,6 @@ void deactivate_locked_super(struct super_block *s)
 {
 	struct file_system_type *fs = s->s_type;
 	if (atomic_dec_and_test(&s->s_active)) {
-		vfs_dq_off(s, 0);
 		fs->kill_sb(s);
 		put_filesystem(fs);
 		put_super(s);
@@ -524,7 +520,7 @@ rescan:
 int do_remount_sb(struct super_block *sb, int flags, void *data, int force)
 {
 	int retval;
-	int remount_rw, remount_ro;
+	int remount_ro;
 
 	if (sb->s_frozen != SB_UNFROZEN)
 		return -EBUSY;
@@ -540,7 +536,6 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force)
 	sync_filesystem(sb);
 
 	remount_ro = (flags & MS_RDONLY) && !(sb->s_flags & MS_RDONLY);
-	remount_rw = !(flags & MS_RDONLY) && (sb->s_flags & MS_RDONLY);
 
 	/* If we are remounting RDONLY and current sb is read/write,
 	   make sure there are no rw files opened */
@@ -549,9 +544,6 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force)
 			mark_files_ro(sb);
 		else if (!fs_may_remount_ro(sb))
 			return -EBUSY;
-		retval = vfs_dq_off(sb, 1);
-		if (retval < 0 && retval != -ENOSYS)
-			return -EBUSY;
 	}
 
 	if (sb->s_op->remount_fs) {
@@ -560,8 +552,7 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force)
 			return retval;
 	}
 	sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (flags & MS_RMT_MASK);
-	if (remount_rw)
-		vfs_dq_quota_on_remount(sb);
+
 	/*
 	 * Some filesystems modify their metadata via some other path than the
 	 * bdev buffer cache (eg. use a private mapping, or directories in
@@ -946,8 +937,8 @@ out:
 EXPORT_SYMBOL_GPL(vfs_kern_mount);
 
 /**
- * freeze_super -- lock the filesystem and force it into a consistent state
- * @super: the super to lock
+ * freeze_super - lock the filesystem and force it into a consistent state
+ * @sb: the super to lock
  *
  * Syncs the super to make sure the filesystem is consistent and calls the fs's
  * freeze_fs.  Subsequent calls to this without first thawing the fs will return
diff --git a/fs/sync.c b/fs/sync.c
index 5a537ccd2e85..15aa6f03b2da 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -130,12 +130,10 @@ void emergency_sync(void)
 
 /*
  * Generic function to fsync a file.
- *
- * filp may be NULL if called via the msync of a vma.
  */
-int file_fsync(struct file *filp, struct dentry *dentry, int datasync)
+int file_fsync(struct file *filp, int datasync)
 {
-	struct inode * inode = dentry->d_inode;
+	struct inode *inode = filp->f_mapping->host;
 	struct super_block * sb;
 	int ret, err;
 
@@ -183,7 +181,7 @@ int vfs_fsync_range(struct file *file, loff_t start, loff_t end, int datasync)
 	 * livelocks in fsync_buffers_list().
 	 */
 	mutex_lock(&mapping->host->i_mutex);
-	err = file->f_op->fsync(file, file->f_path.dentry, datasync);
+	err = file->f_op->fsync(file, datasync);
 	if (!ret)
 		ret = err;
 	mutex_unlock(&mapping->host->i_mutex);
diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c
index bbd77e95cf7f..bde1a4c3679a 100644
--- a/fs/sysfs/inode.c
+++ b/fs/sysfs/inode.c
@@ -117,13 +117,11 @@ int sysfs_setattr(struct dentry *dentry, struct iattr *iattr)
 	if (error)
 		goto out;
 
-	iattr->ia_valid &= ~ATTR_SIZE; /* ignore size changes */
-
-	error = inode_setattr(inode, iattr);
-	if (error)
-		goto out;
+	/* this ignores size changes */
+	generic_setattr(inode, iattr);
 
 	error = sysfs_sd_setattr(sd, iattr);
+
 out:
 	mutex_unlock(&sysfs_mutex);
 	return error;
diff --git a/fs/sysv/dir.c b/fs/sysv/dir.c
index 1dabed286b4c..79941e4964a4 100644
--- a/fs/sysv/dir.c
+++ b/fs/sysv/dir.c
@@ -24,7 +24,7 @@ const struct file_operations sysv_dir_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= sysv_readdir,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 };
 
 static inline void dir_put_page(struct page *page)
diff --git a/fs/sysv/file.c b/fs/sysv/file.c
index 96340c01f4a7..750cc22349bd 100644
--- a/fs/sysv/file.c
+++ b/fs/sysv/file.c
@@ -26,7 +26,7 @@ const struct file_operations sysv_file_operations = {
 	.write		= do_sync_write,
 	.aio_write	= generic_file_aio_write,
 	.mmap		= generic_file_mmap,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 	.splice_read	= generic_file_splice_read,
 };
 
diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c
index 4573734d723d..d4a5380b5669 100644
--- a/fs/sysv/inode.c
+++ b/fs/sysv/inode.c
@@ -43,6 +43,7 @@ static int sysv_sync_fs(struct super_block *sb, int wait)
 	 * then attach current time stamp.
 	 * But if the filesystem was marked clean, keep it clean.
 	 */
+	sb->s_dirt = 0;
 	old_time = fs32_to_cpu(sbi, *sbi->s_sb_time);
 	if (sbi->s_type == FSTYPE_SYSV4) {
 		if (*sbi->s_sb_state == cpu_to_fs32(sbi, 0x7c269d38 - old_time))
diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
index 5692cf72b807..12f445cee9f7 100644
--- a/fs/ubifs/file.c
+++ b/fs/ubifs/file.c
@@ -967,12 +967,15 @@ static int do_writepage(struct page *page, int len)
  * the page locked, and it locks @ui_mutex. However, write-back does take inode
  * @i_mutex, which means other VFS operations may be run on this inode at the
  * same time. And the problematic one is truncation to smaller size, from where
- * we have to call 'vmtruncate()', which first changes @inode->i_size, then
+ * we have to call 'simple_setsize()', which first changes @inode->i_size, then
  * drops the truncated pages. And while dropping the pages, it takes the page
- * lock. This means that 'do_truncation()' cannot call 'vmtruncate()' with
+ * lock. This means that 'do_truncation()' cannot call 'simple_setsize()' with
  * @ui_mutex locked, because it would deadlock with 'ubifs_writepage()'. This
  * means that @inode->i_size is changed while @ui_mutex is unlocked.
  *
+ * XXX: with the new truncate the above is not true anymore, the simple_setsize
+ * calls can be replaced with the individual components.
+ *
  * But in 'ubifs_writepage()' we have to guarantee that we do not write beyond
  * inode size. How do we do this if @inode->i_size may became smaller while we
  * are in the middle of 'ubifs_writepage()'? The UBIFS solution is the
@@ -1125,7 +1128,7 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode,
 		budgeted = 0;
 	}
 
-	err = vmtruncate(inode, new_size);
+	err = simple_setsize(inode, new_size);
 	if (err)
 		goto out_budg;
 
@@ -1214,7 +1217,7 @@ static int do_setattr(struct ubifs_info *c, struct inode *inode,
 
 	if (attr->ia_valid & ATTR_SIZE) {
 		dbg_gen("size %lld -> %lld", inode->i_size, new_size);
-		err = vmtruncate(inode, new_size);
+		err = simple_setsize(inode, new_size);
 		if (err)
 			goto out;
 	}
@@ -1223,7 +1226,7 @@ static int do_setattr(struct ubifs_info *c, struct inode *inode,
 	if (attr->ia_valid & ATTR_SIZE) {
 		/* Truncation changes inode [mc]time */
 		inode->i_mtime = inode->i_ctime = ubifs_current_time(inode);
-		/* 'vmtruncate()' changed @i_size, update @ui_size */
+		/* 'simple_setsize()' changed @i_size, update @ui_size */
 		ui->ui_size = inode->i_size;
 	}
 
@@ -1304,9 +1307,9 @@ static void *ubifs_follow_link(struct dentry *dentry, struct nameidata *nd)
 	return NULL;
 }
 
-int ubifs_fsync(struct file *file, struct dentry *dentry, int datasync)
+int ubifs_fsync(struct file *file, int datasync)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode = file->f_mapping->host;
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
 	int err;
 
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index bd2542dad014..2eef553d50c8 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -379,7 +379,7 @@ struct ubifs_gced_idx_leb {
  * The @ui_size is a "shadow" variable for @inode->i_size and UBIFS uses
  * @ui_size instead of @inode->i_size. The reason for this is that UBIFS cannot
  * make sure @inode->i_size is always changed under @ui_mutex, because it
- * cannot call 'vmtruncate()' with @ui_mutex locked, because it would deadlock
+ * cannot call 'simple_setsize()' with @ui_mutex locked, because it would deadlock
  * with 'ubifs_writepage()' (see file.c). All the other inode fields are
  * changed under @ui_mutex, so they do not need "shadow" fields. Note, one
  * could consider to rework locking and base it on "shadow" fields.
@@ -1678,7 +1678,7 @@ const struct ubifs_lprops *ubifs_fast_find_frdi_idx(struct ubifs_info *c);
 int ubifs_calc_dark(const struct ubifs_info *c, int spc);
 
 /* file.c */
-int ubifs_fsync(struct file *file, struct dentry *dentry, int datasync);
+int ubifs_fsync(struct file *file, int datasync);
 int ubifs_setattr(struct dentry *dentry, struct iattr *attr);
 
 /* dir.c */
diff --git a/fs/udf/balloc.c b/fs/udf/balloc.c
index 9a9378b4eb5a..b608efaa4cee 100644
--- a/fs/udf/balloc.c
+++ b/fs/udf/balloc.c
@@ -21,7 +21,6 @@
 
 #include "udfdecl.h"
 
-#include <linux/quotaops.h>
 #include <linux/buffer_head.h>
 #include <linux/bitops.h>
 
@@ -159,8 +158,6 @@ static void udf_bitmap_free_blocks(struct super_block *sb,
 				udf_debug("byte=%2x\n",
 					((char *)bh->b_data)[(bit + i) >> 3]);
 			} else {
-				if (inode)
-					dquot_free_block(inode, 1);
 				udf_add_free_space(sb, sbi->s_partition, 1);
 			}
 		}
@@ -210,15 +207,8 @@ static int udf_bitmap_prealloc_blocks(struct super_block *sb,
 		bit = block % (sb->s_blocksize << 3);
 
 		while (bit < (sb->s_blocksize << 3) && block_count > 0) {
-			if (!udf_test_bit(bit, bh->b_data))
+			if (!udf_clear_bit(bit, bh->b_data))
 				goto out;
-			else if (dquot_prealloc_block(inode, 1))
-				goto out;
-			else if (!udf_clear_bit(bit, bh->b_data)) {
-				udf_debug("bit already cleared for block %d\n", bit);
-				dquot_free_block(inode, 1);
-				goto out;
-			}
 			block_count--;
 			alloc_count++;
 			bit++;
@@ -338,20 +328,6 @@ search_back:
 	}
 
 got_block:
-
-	/*
-	 * Check quota for allocation of this block.
-	 */
-	if (inode) {
-		int ret = dquot_alloc_block(inode, 1);
-
-		if (ret) {
-			mutex_unlock(&sbi->s_alloc_mutex);
-			*err = ret;
-			return 0;
-		}
-	}
-
 	newblock = bit + (block_group << (sb->s_blocksize_bits + 3)) -
 		(sizeof(struct spaceBitmapDesc) << 3);
 
@@ -401,10 +377,6 @@ static void udf_table_free_blocks(struct super_block *sb,
 	}
 
 	iinfo = UDF_I(table);
-	/* We do this up front - There are some error conditions that
-	   could occure, but.. oh well */
-	if (inode)
-		dquot_free_block(inode, count);
 	udf_add_free_space(sb, sbi->s_partition, count);
 
 	start = bloc->logicalBlockNum + offset;
@@ -649,10 +621,7 @@ static int udf_table_prealloc_blocks(struct super_block *sb,
 		epos.offset -= adsize;
 
 		alloc_count = (elen >> sb->s_blocksize_bits);
-		if (inode && dquot_prealloc_block(inode,
-			alloc_count > block_count ? block_count : alloc_count))
-			alloc_count = 0;
-		else if (alloc_count > block_count) {
+		if (alloc_count > block_count) {
 			alloc_count = block_count;
 			eloc.logicalBlockNum += alloc_count;
 			elen -= (alloc_count << sb->s_blocksize_bits);
@@ -752,14 +721,6 @@ static int udf_table_new_block(struct super_block *sb,
 	newblock = goal_eloc.logicalBlockNum;
 	goal_eloc.logicalBlockNum++;
 	goal_elen -= sb->s_blocksize;
-	if (inode) {
-		*err = dquot_alloc_block(inode, 1);
-		if (*err) {
-			brelse(goal_epos.bh);
-			mutex_unlock(&sbi->s_alloc_mutex);
-			return 0;
-		}
-	}
 
 	if (goal_elen)
 		udf_write_aext(table, &goal_epos, &goal_eloc, goal_elen, 1);
diff --git a/fs/udf/dir.c b/fs/udf/dir.c
index 3a84455c2a77..51552bf50225 100644
--- a/fs/udf/dir.c
+++ b/fs/udf/dir.c
@@ -207,8 +207,9 @@ static int udf_readdir(struct file *filp, void *dirent, filldir_t filldir)
 
 /* readdir and lookup functions */
 const struct file_operations udf_dir_operations = {
+	.llseek			= generic_file_llseek,
 	.read			= generic_read_dir,
 	.readdir		= udf_readdir,
 	.unlocked_ioctl		= udf_ioctl,
-	.fsync			= simple_fsync,
+	.fsync			= generic_file_fsync,
 };
diff --git a/fs/udf/file.c b/fs/udf/file.c
index baae3a723946..94e06d6bddbd 100644
--- a/fs/udf/file.c
+++ b/fs/udf/file.c
@@ -34,7 +34,6 @@
 #include <linux/errno.h>
 #include <linux/smp_lock.h>
 #include <linux/pagemap.h>
-#include <linux/quotaops.h>
 #include <linux/buffer_head.h>
 #include <linux/aio.h>
 #include <linux/smp_lock.h>
@@ -219,39 +218,16 @@ const struct file_operations udf_file_operations = {
 	.read			= do_sync_read,
 	.aio_read		= generic_file_aio_read,
 	.unlocked_ioctl		= udf_ioctl,
-	.open			= dquot_file_open,
+	.open			= generic_file_open,
 	.mmap			= generic_file_mmap,
 	.write			= do_sync_write,
 	.aio_write		= udf_file_aio_write,
 	.release		= udf_release_file,
-	.fsync			= simple_fsync,
+	.fsync			= generic_file_fsync,
 	.splice_read		= generic_file_splice_read,
 	.llseek			= generic_file_llseek,
 };
 
-int udf_setattr(struct dentry *dentry, struct iattr *iattr)
-{
-	struct inode *inode = dentry->d_inode;
-	int error;
-
-	error = inode_change_ok(inode, iattr);
-	if (error)
-		return error;
-
-	if (is_quota_modification(inode, iattr))
-		dquot_initialize(inode);
-
-	if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
-            (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
-		error = dquot_transfer(inode, iattr);
-		if (error)
-			return error;
-	}
-
-	return inode_setattr(inode, iattr);
-}
-
 const struct inode_operations udf_file_inode_operations = {
 	.truncate		= udf_truncate,
-	.setattr		= udf_setattr,
 };
diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c
index 2b5586c7f02a..18cd7111185d 100644
--- a/fs/udf/ialloc.c
+++ b/fs/udf/ialloc.c
@@ -20,7 +20,6 @@
 
 #include "udfdecl.h"
 #include <linux/fs.h>
-#include <linux/quotaops.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
 
@@ -32,13 +31,6 @@ void udf_free_inode(struct inode *inode)
 	struct super_block *sb = inode->i_sb;
 	struct udf_sb_info *sbi = UDF_SB(sb);
 
-	/*
-	 * Note: we must free any quota before locking the superblock,
-	 * as writing the quota to disk may need the lock as well.
-	 */
-	dquot_free_inode(inode);
-	dquot_drop(inode);
-
 	clear_inode(inode);
 
 	mutex_lock(&sbi->s_alloc_mutex);
@@ -61,7 +53,7 @@ struct inode *udf_new_inode(struct inode *dir, int mode, int *err)
 	struct super_block *sb = dir->i_sb;
 	struct udf_sb_info *sbi = UDF_SB(sb);
 	struct inode *inode;
-	int block, ret;
+	int block;
 	uint32_t start = UDF_I(dir)->i_location.logicalBlockNum;
 	struct udf_inode_info *iinfo;
 	struct udf_inode_info *dinfo = UDF_I(dir);
@@ -146,17 +138,6 @@ struct inode *udf_new_inode(struct inode *dir, int mode, int *err)
 	insert_inode_hash(inode);
 	mark_inode_dirty(inode);
 
-	dquot_initialize(inode);
-	ret = dquot_alloc_inode(inode);
-	if (ret) {
-		dquot_drop(inode);
-		inode->i_flags |= S_NOQUOTA;
-		inode->i_nlink = 0;
-		iput(inode);
-		*err = ret;
-		return NULL;
-	}
-
 	*err = 0;
 	return inode;
 }
diff --git a/fs/udf/inode.c b/fs/udf/inode.c
index 8a3fbd177cab..124852bcf6fe 100644
--- a/fs/udf/inode.c
+++ b/fs/udf/inode.c
@@ -36,7 +36,6 @@
 #include <linux/pagemap.h>
 #include <linux/buffer_head.h>
 #include <linux/writeback.h>
-#include <linux/quotaops.h>
 #include <linux/slab.h>
 #include <linux/crc-itu-t.h>
 
@@ -71,9 +70,6 @@ static int udf_get_block(struct inode *, sector_t, struct buffer_head *, int);
 
 void udf_delete_inode(struct inode *inode)
 {
-	if (!is_bad_inode(inode))
-		dquot_initialize(inode);
-
 	truncate_inode_pages(&inode->i_data, 0);
 
 	if (is_bad_inode(inode))
@@ -113,7 +109,6 @@ void udf_clear_inode(struct inode *inode)
 			(unsigned long long)iinfo->i_lenExtents);
 	}
 
-	dquot_drop(inode);
 	kfree(iinfo->i_ext.i_data);
 	iinfo->i_ext.i_data = NULL;
 }
diff --git a/fs/udf/namei.c b/fs/udf/namei.c
index 585f733615dc..bf5fc674193c 100644
--- a/fs/udf/namei.c
+++ b/fs/udf/namei.c
@@ -27,7 +27,6 @@
 #include <linux/errno.h>
 #include <linux/mm.h>
 #include <linux/slab.h>
-#include <linux/quotaops.h>
 #include <linux/smp_lock.h>
 #include <linux/buffer_head.h>
 #include <linux/sched.h>
@@ -563,8 +562,6 @@ static int udf_create(struct inode *dir, struct dentry *dentry, int mode,
 	int err;
 	struct udf_inode_info *iinfo;
 
-	dquot_initialize(dir);
-
 	lock_kernel();
 	inode = udf_new_inode(dir, mode, &err);
 	if (!inode) {
@@ -617,8 +614,6 @@ static int udf_mknod(struct inode *dir, struct dentry *dentry, int mode,
 	if (!old_valid_dev(rdev))
 		return -EINVAL;
 
-	dquot_initialize(dir);
-
 	lock_kernel();
 	err = -EIO;
 	inode = udf_new_inode(dir, mode, &err);
@@ -664,8 +659,6 @@ static int udf_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	struct udf_inode_info *dinfo = UDF_I(dir);
 	struct udf_inode_info *iinfo;
 
-	dquot_initialize(dir);
-
 	lock_kernel();
 	err = -EMLINK;
 	if (dir->i_nlink >= (256 << sizeof(dir->i_nlink)) - 1)
@@ -800,8 +793,6 @@ static int udf_rmdir(struct inode *dir, struct dentry *dentry)
 	struct fileIdentDesc *fi, cfi;
 	struct kernel_lb_addr tloc;
 
-	dquot_initialize(dir);
-
 	retval = -ENOENT;
 	lock_kernel();
 	fi = udf_find_entry(dir, &dentry->d_name, &fibh, &cfi);
@@ -848,8 +839,6 @@ static int udf_unlink(struct inode *dir, struct dentry *dentry)
 	struct fileIdentDesc cfi;
 	struct kernel_lb_addr tloc;
 
-	dquot_initialize(dir);
-
 	retval = -ENOENT;
 	lock_kernel();
 	fi = udf_find_entry(dir, &dentry->d_name, &fibh, &cfi);
@@ -904,8 +893,6 @@ static int udf_symlink(struct inode *dir, struct dentry *dentry,
 	struct buffer_head *bh;
 	struct udf_inode_info *iinfo;
 
-	dquot_initialize(dir);
-
 	lock_kernel();
 	inode = udf_new_inode(dir, S_IFLNK | S_IRWXUGO, &err);
 	if (!inode)
@@ -1075,8 +1062,6 @@ static int udf_link(struct dentry *old_dentry, struct inode *dir,
 	int err;
 	struct buffer_head *bh;
 
-	dquot_initialize(dir);
-
 	lock_kernel();
 	if (inode->i_nlink >= (256 << sizeof(inode->i_nlink)) - 1) {
 		unlock_kernel();
@@ -1139,9 +1124,6 @@ static int udf_rename(struct inode *old_dir, struct dentry *old_dentry,
 	struct kernel_lb_addr tloc;
 	struct udf_inode_info *old_iinfo = UDF_I(old_inode);
 
-	dquot_initialize(old_dir);
-	dquot_initialize(new_dir);
-
 	lock_kernel();
 	ofi = udf_find_entry(old_dir, &old_dentry->d_name, &ofibh, &ocfi);
 	if (ofi) {
@@ -1387,7 +1369,6 @@ const struct export_operations udf_export_ops = {
 const struct inode_operations udf_dir_inode_operations = {
 	.lookup				= udf_lookup,
 	.create				= udf_create,
-	.setattr			= udf_setattr,
 	.link				= udf_link,
 	.unlink				= udf_unlink,
 	.symlink			= udf_symlink,
@@ -1400,5 +1381,4 @@ const struct inode_operations udf_symlink_inode_operations = {
 	.readlink	= generic_readlink,
 	.follow_link	= page_follow_link_light,
 	.put_link	= page_put_link,
-	.setattr	= udf_setattr,
 };
diff --git a/fs/udf/super.c b/fs/udf/super.c
index 1e4543cbcd27..612d1e2e285a 100644
--- a/fs/udf/super.c
+++ b/fs/udf/super.c
@@ -557,6 +557,7 @@ static int udf_remount_fs(struct super_block *sb, int *flags, char *options)
 {
 	struct udf_options uopt;
 	struct udf_sb_info *sbi = UDF_SB(sb);
+	int error = 0;
 
 	uopt.flags = sbi->s_flags;
 	uopt.uid   = sbi->s_uid;
@@ -582,17 +583,17 @@ static int udf_remount_fs(struct super_block *sb, int *flags, char *options)
 			*flags |= MS_RDONLY;
 	}
 
-	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) {
-		unlock_kernel();
-		return 0;
-	}
+	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
+		goto out_unlock;
+
 	if (*flags & MS_RDONLY)
 		udf_close_lvid(sb);
 	else
 		udf_open_lvid(sb);
 
+out_unlock:
 	unlock_kernel();
-	return 0;
+	return error;
 }
 
 /* Check Volume Structure Descriptors (ECMA 167 2/9.1) */
@@ -1939,7 +1940,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent)
 	/* Fill in the rest of the superblock */
 	sb->s_op = &udf_sb_ops;
 	sb->s_export_op = &udf_export_ops;
-	sb->dq_op = NULL;
+
 	sb->s_dirt = 0;
 	sb->s_magic = UDF_SUPER_MAGIC;
 	sb->s_time_gran = 1000;
diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h
index 9079ff7d6255..2bac0354891f 100644
--- a/fs/udf/udfdecl.h
+++ b/fs/udf/udfdecl.h
@@ -131,7 +131,6 @@ extern int udf_write_fi(struct inode *inode, struct fileIdentDesc *,
 
 /* file.c */
 extern long udf_ioctl(struct file *, unsigned int, unsigned long);
-extern int udf_setattr(struct dentry *dentry, struct iattr *iattr);
 /* inode.c */
 extern struct inode *udf_iget(struct super_block *, struct kernel_lb_addr *);
 extern int udf_sync_inode(struct inode *);
diff --git a/fs/ufs/balloc.c b/fs/ufs/balloc.c
index 5cfa4d85ccf2..048484fb10d2 100644
--- a/fs/ufs/balloc.c
+++ b/fs/ufs/balloc.c
@@ -12,7 +12,6 @@
 #include <linux/stat.h>
 #include <linux/time.h>
 #include <linux/string.h>
-#include <linux/quotaops.h>
 #include <linux/buffer_head.h>
 #include <linux/capability.h>
 #include <linux/bitops.h>
@@ -85,9 +84,6 @@ void ufs_free_fragments(struct inode *inode, u64 fragment, unsigned count)
 				   "bit already cleared for fragment %u", i);
 	}
 	
-	dquot_free_block(inode, count);
-
-	
 	fs32_add(sb, &ucg->cg_cs.cs_nffree, count);
 	uspi->cs_total.cs_nffree += count;
 	fs32_add(sb, &UFS_SB(sb)->fs_cs(cgno).cs_nffree, count);
@@ -195,7 +191,6 @@ do_more:
 		ubh_setblock(UCPI_UBH(ucpi), ucpi->c_freeoff, blkno);
 		if ((UFS_SB(sb)->s_flags & UFS_CG_MASK) == UFS_CG_44BSD)
 			ufs_clusteracct (sb, ucpi, blkno, 1);
-		dquot_free_block(inode, uspi->s_fpb);
 
 		fs32_add(sb, &ucg->cg_cs.cs_nbfree, 1);
 		uspi->cs_total.cs_nbfree++;
@@ -511,7 +506,6 @@ static u64 ufs_add_fragments(struct inode *inode, u64 fragment,
 	struct ufs_cg_private_info * ucpi;
 	struct ufs_cylinder_group * ucg;
 	unsigned cgno, fragno, fragoff, count, fragsize, i;
-	int ret;
 	
 	UFSD("ENTER, fragment %llu, oldcount %u, newcount %u\n",
 	     (unsigned long long)fragment, oldcount, newcount);
@@ -557,11 +551,6 @@ static u64 ufs_add_fragments(struct inode *inode, u64 fragment,
 		fs32_add(sb, &ucg->cg_frsum[fragsize - count], 1);
 	for (i = oldcount; i < newcount; i++)
 		ubh_clrbit (UCPI_UBH(ucpi), ucpi->c_freeoff, fragno + i);
-	ret = dquot_alloc_block(inode, count);
-	if (ret) {
-		*err = ret;
-		return 0;
-	}
 
 	fs32_sub(sb, &ucg->cg_cs.cs_nffree, count);
 	fs32_sub(sb, &UFS_SB(sb)->fs_cs(cgno).cs_nffree, count);
@@ -598,7 +587,6 @@ static u64 ufs_alloc_fragments(struct inode *inode, unsigned cgno,
 	struct ufs_cylinder_group * ucg;
 	unsigned oldcg, i, j, k, allocsize;
 	u64 result;
-	int ret;
 	
 	UFSD("ENTER, ino %lu, cgno %u, goal %llu, count %u\n",
 	     inode->i_ino, cgno, (unsigned long long)goal, count);
@@ -667,7 +655,6 @@ cg_found:
 		for (i = count; i < uspi->s_fpb; i++)
 			ubh_setbit (UCPI_UBH(ucpi), ucpi->c_freeoff, goal + i);
 		i = uspi->s_fpb - count;
-		dquot_free_block(inode, i);
 
 		fs32_add(sb, &ucg->cg_cs.cs_nffree, i);
 		uspi->cs_total.cs_nffree += i;
@@ -679,11 +666,6 @@ cg_found:
 	result = ufs_bitmap_search (sb, ucpi, goal, allocsize);
 	if (result == INVBLOCK)
 		return 0;
-	ret = dquot_alloc_block(inode, count);
-	if (ret) {
-		*err = ret;
-		return 0;
-	}
 	for (i = 0; i < count; i++)
 		ubh_clrbit (UCPI_UBH(ucpi), ucpi->c_freeoff, result + i);
 	
@@ -718,7 +700,6 @@ static u64 ufs_alloccg_block(struct inode *inode,
 	struct ufs_super_block_first * usb1;
 	struct ufs_cylinder_group * ucg;
 	u64 result, blkno;
-	int ret;
 
 	UFSD("ENTER, goal %llu\n", (unsigned long long)goal);
 
@@ -752,11 +733,6 @@ gotit:
 	ubh_clrblock (UCPI_UBH(ucpi), ucpi->c_freeoff, blkno);
 	if ((UFS_SB(sb)->s_flags & UFS_CG_MASK) == UFS_CG_44BSD)
 		ufs_clusteracct (sb, ucpi, blkno, -1);
-	ret = dquot_alloc_block(inode, uspi->s_fpb);
-	if (ret) {
-		*err = ret;
-		return INVBLOCK;
-	}
 
 	fs32_sub(sb, &ucg->cg_cs.cs_nbfree, 1);
 	uspi->cs_total.cs_nbfree--;
diff --git a/fs/ufs/dir.c b/fs/ufs/dir.c
index 317a0d444f6b..ec784756dc65 100644
--- a/fs/ufs/dir.c
+++ b/fs/ufs/dir.c
@@ -666,6 +666,6 @@ not_empty:
 const struct file_operations ufs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= ufs_readdir,
-	.fsync		= simple_fsync,
+	.fsync		= generic_file_fsync,
 	.llseek		= generic_file_llseek,
 };
diff --git a/fs/ufs/file.c b/fs/ufs/file.c
index a8962cecde5b..33afa20d4509 100644
--- a/fs/ufs/file.c
+++ b/fs/ufs/file.c
@@ -24,7 +24,6 @@
  */
 
 #include <linux/fs.h>
-#include <linux/quotaops.h>
 
 #include "ufs_fs.h"
 #include "ufs.h"
@@ -41,7 +40,7 @@ const struct file_operations ufs_file_operations = {
 	.write		= do_sync_write,
 	.aio_write	= generic_file_aio_write,
 	.mmap		= generic_file_mmap,
-	.open           = dquot_file_open,
-	.fsync		= simple_fsync,
+	.open           = generic_file_open,
+	.fsync		= generic_file_fsync,
 	.splice_read	= generic_file_splice_read,
 };
diff --git a/fs/ufs/ialloc.c b/fs/ufs/ialloc.c
index 3a959d55084d..594480e537d2 100644
--- a/fs/ufs/ialloc.c
+++ b/fs/ufs/ialloc.c
@@ -27,7 +27,6 @@
 #include <linux/time.h>
 #include <linux/stat.h>
 #include <linux/string.h>
-#include <linux/quotaops.h>
 #include <linux/buffer_head.h>
 #include <linux/sched.h>
 #include <linux/bitops.h>
@@ -95,9 +94,6 @@ void ufs_free_inode (struct inode * inode)
 
 	is_directory = S_ISDIR(inode->i_mode);
 
-	dquot_free_inode(inode);
-	dquot_drop(inode);
-
 	clear_inode (inode);
 
 	if (ubh_isclr (UCPI_UBH(ucpi), ucpi->c_iusedoff, bit))
@@ -347,21 +343,12 @@ cg_found:
 
 	unlock_super (sb);
 
-	dquot_initialize(inode);
-	err = dquot_alloc_inode(inode);
-	if (err) {
-		dquot_drop(inode);
-		goto fail_without_unlock;
-	}
-
 	UFSD("allocating inode %lu\n", inode->i_ino);
 	UFSD("EXIT\n");
 	return inode;
 
 fail_remove_inode:
 	unlock_super(sb);
-fail_without_unlock:
-	inode->i_flags |= S_NOQUOTA;
 	inode->i_nlink = 0;
 	iput(inode);
 	UFSD("EXIT (FAILED): err %d\n", err);
diff --git a/fs/ufs/inode.c b/fs/ufs/inode.c
index cffa756f1047..73fe773aa034 100644
--- a/fs/ufs/inode.c
+++ b/fs/ufs/inode.c
@@ -37,7 +37,6 @@
 #include <linux/smp_lock.h>
 #include <linux/buffer_head.h>
 #include <linux/writeback.h>
-#include <linux/quotaops.h>
 
 #include "ufs_fs.h"
 #include "ufs.h"
@@ -910,9 +909,6 @@ void ufs_delete_inode (struct inode * inode)
 {
 	loff_t old_i_size;
 
-	if (!is_bad_inode(inode))
-		dquot_initialize(inode);
-
 	truncate_inode_pages(&inode->i_data, 0);
 	if (is_bad_inode(inode))
 		goto no_delete;
diff --git a/fs/ufs/namei.c b/fs/ufs/namei.c
index eabc02eb1294..b056f02b1fb3 100644
--- a/fs/ufs/namei.c
+++ b/fs/ufs/namei.c
@@ -30,7 +30,6 @@
 #include <linux/time.h>
 #include <linux/fs.h>
 #include <linux/smp_lock.h>
-#include <linux/quotaops.h>
 
 #include "ufs_fs.h"
 #include "ufs.h"
@@ -86,8 +85,6 @@ static int ufs_create (struct inode * dir, struct dentry * dentry, int mode,
 
 	UFSD("BEGIN\n");
 
-	dquot_initialize(dir);
-
 	inode = ufs_new_inode(dir, mode);
 	err = PTR_ERR(inode);
 
@@ -112,8 +109,6 @@ static int ufs_mknod (struct inode * dir, struct dentry *dentry, int mode, dev_t
 	if (!old_valid_dev(rdev))
 		return -EINVAL;
 
-	dquot_initialize(dir);
-
 	inode = ufs_new_inode(dir, mode);
 	err = PTR_ERR(inode);
 	if (!IS_ERR(inode)) {
@@ -138,8 +133,6 @@ static int ufs_symlink (struct inode * dir, struct dentry * dentry,
 	if (l > sb->s_blocksize)
 		goto out_notlocked;
 
-	dquot_initialize(dir);
-
 	lock_kernel();
 	inode = ufs_new_inode(dir, S_IFLNK | S_IRWXUGO);
 	err = PTR_ERR(inode);
@@ -185,8 +178,6 @@ static int ufs_link (struct dentry * old_dentry, struct inode * dir,
 		return -EMLINK;
 	}
 
-	dquot_initialize(dir);
-
 	inode->i_ctime = CURRENT_TIME_SEC;
 	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
@@ -204,8 +195,6 @@ static int ufs_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 	if (dir->i_nlink >= UFS_LINK_MAX)
 		goto out;
 
-	dquot_initialize(dir);
-
 	lock_kernel();
 	inode_inc_link_count(dir);
 
@@ -250,8 +239,6 @@ static int ufs_unlink(struct inode *dir, struct dentry *dentry)
 	struct page *page;
 	int err = -ENOENT;
 
-	dquot_initialize(dir);
-
 	de = ufs_find_entry(dir, &dentry->d_name, &page);
 	if (!de)
 		goto out;
@@ -296,9 +283,6 @@ static int ufs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	struct ufs_dir_entry *old_de;
 	int err = -ENOENT;
 
-	dquot_initialize(old_dir);
-	dquot_initialize(new_dir);
-
 	old_de = ufs_find_entry(old_dir, &old_dentry->d_name, &old_page);
 	if (!old_de)
 		goto out;
diff --git a/fs/ufs/super.c b/fs/ufs/super.c
index 14743d935a93..3ec5a9eb6efb 100644
--- a/fs/ufs/super.c
+++ b/fs/ufs/super.c
@@ -77,7 +77,6 @@
 
 #include <linux/errno.h>
 #include <linux/fs.h>
-#include <linux/quotaops.h>
 #include <linux/slab.h>
 #include <linux/time.h>
 #include <linux/stat.h>
@@ -918,6 +917,7 @@ again:
 	sbi->s_bytesex = BYTESEX_LE;
 	switch ((uspi->fs_magic = fs32_to_cpu(sb, usb3->fs_magic))) {
 		case UFS_MAGIC:
+		case UFS_MAGIC_BW:
 		case UFS2_MAGIC:
 		case UFS_MAGIC_LFN:
 	        case UFS_MAGIC_FEA:
@@ -927,6 +927,7 @@ again:
 	sbi->s_bytesex = BYTESEX_BE;
 	switch ((uspi->fs_magic = fs32_to_cpu(sb, usb3->fs_magic))) {
 		case UFS_MAGIC:
+		case UFS_MAGIC_BW:
 		case UFS2_MAGIC:
 		case UFS_MAGIC_LFN:
 	        case UFS_MAGIC_FEA:
@@ -1045,7 +1046,7 @@ magic_found:
 	 */
 	sb->s_op = &ufs_super_ops;
 	sb->s_export_op = &ufs_export_ops;
-	sb->dq_op = NULL; /***/
+
 	sb->s_magic = fs32_to_cpu(sb, usb3->fs_magic);
 
 	uspi->s_sblkno = fs32_to_cpu(sb, usb1->fs_sblkno);
@@ -1435,126 +1436,19 @@ static void destroy_inodecache(void)
 	kmem_cache_destroy(ufs_inode_cachep);
 }
 
-static void ufs_clear_inode(struct inode *inode)
-{
-	dquot_drop(inode);
-}
-
-#ifdef CONFIG_QUOTA
-static ssize_t ufs_quota_read(struct super_block *, int, char *,size_t, loff_t);
-static ssize_t ufs_quota_write(struct super_block *, int, const char *, size_t, loff_t);
-#endif
-
 static const struct super_operations ufs_super_ops = {
 	.alloc_inode	= ufs_alloc_inode,
 	.destroy_inode	= ufs_destroy_inode,
 	.write_inode	= ufs_write_inode,
 	.delete_inode	= ufs_delete_inode,
-	.clear_inode	= ufs_clear_inode,
 	.put_super	= ufs_put_super,
 	.write_super	= ufs_write_super,
 	.sync_fs	= ufs_sync_fs,
 	.statfs		= ufs_statfs,
 	.remount_fs	= ufs_remount,
 	.show_options   = ufs_show_options,
-#ifdef CONFIG_QUOTA
-	.quota_read	= ufs_quota_read,
-	.quota_write	= ufs_quota_write,
-#endif
 };
 
-#ifdef CONFIG_QUOTA
-
-/* Read data from quotafile - avoid pagecache and such because we cannot afford
- * acquiring the locks... As quota files are never truncated and quota code
- * itself serializes the operations (and noone else should touch the files)
- * we don't have to be afraid of races */
-static ssize_t ufs_quota_read(struct super_block *sb, int type, char *data,
-			       size_t len, loff_t off)
-{
-	struct inode *inode = sb_dqopt(sb)->files[type];
-	sector_t blk = off >> sb->s_blocksize_bits;
-	int err = 0;
-	int offset = off & (sb->s_blocksize - 1);
-	int tocopy;
-	size_t toread;
-	struct buffer_head *bh;
-	loff_t i_size = i_size_read(inode);
-
-	if (off > i_size)
-		return 0;
-	if (off+len > i_size)
-		len = i_size-off;
-	toread = len;
-	while (toread > 0) {
-		tocopy = sb->s_blocksize - offset < toread ?
-				sb->s_blocksize - offset : toread;
-
-		bh = ufs_bread(inode, blk, 0, &err);
-		if (err)
-			return err;
-		if (!bh)	/* A hole? */
-			memset(data, 0, tocopy);
-		else {
-			memcpy(data, bh->b_data+offset, tocopy);
-			brelse(bh);
-		}
-		offset = 0;
-		toread -= tocopy;
-		data += tocopy;
-		blk++;
-	}
-	return len;
-}
-
-/* Write to quotafile */
-static ssize_t ufs_quota_write(struct super_block *sb, int type,
-				const char *data, size_t len, loff_t off)
-{
-	struct inode *inode = sb_dqopt(sb)->files[type];
-	sector_t blk = off >> sb->s_blocksize_bits;
-	int err = 0;
-	int offset = off & (sb->s_blocksize - 1);
-	int tocopy;
-	size_t towrite = len;
-	struct buffer_head *bh;
-
-	mutex_lock_nested(&inode->i_mutex, I_MUTEX_QUOTA);
-	while (towrite > 0) {
-		tocopy = sb->s_blocksize - offset < towrite ?
-				sb->s_blocksize - offset : towrite;
-
-		bh = ufs_bread(inode, blk, 1, &err);
-		if (!bh)
-			goto out;
-		lock_buffer(bh);
-		memcpy(bh->b_data+offset, data, tocopy);
-		flush_dcache_page(bh->b_page);
-		set_buffer_uptodate(bh);
-		mark_buffer_dirty(bh);
-		unlock_buffer(bh);
-		brelse(bh);
-		offset = 0;
-		towrite -= tocopy;
-		data += tocopy;
-		blk++;
-	}
-out:
-	if (len == towrite) {
-		mutex_unlock(&inode->i_mutex);
-		return err;
-	}
-	if (inode->i_size < off+len-towrite)
-		i_size_write(inode, off+len-towrite);
-	inode->i_version++;
-	inode->i_mtime = inode->i_ctime = CURRENT_TIME_SEC;
-	mark_inode_dirty(inode);
-	mutex_unlock(&inode->i_mutex);
-	return len - towrite;
-}
-
-#endif
-
 static int ufs_get_sb(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *data, struct vfsmount *mnt)
 {
diff --git a/fs/ufs/truncate.c b/fs/ufs/truncate.c
index f294c44577dc..589e01a465ba 100644
--- a/fs/ufs/truncate.c
+++ b/fs/ufs/truncate.c
@@ -44,7 +44,6 @@
 #include <linux/buffer_head.h>
 #include <linux/blkdev.h>
 #include <linux/sched.h>
-#include <linux/quotaops.h>
 
 #include "ufs_fs.h"
 #include "ufs.h"
@@ -501,12 +500,10 @@ out:
 	return err;
 }
 
-
 /*
- * We don't define our `inode->i_op->truncate', and call it here,
- * because of:
- * - there is no way to know old size
- * - there is no way inform user about error, if it happens in `truncate'
+ * TODO:
+ *	- truncate case should use proper ordering instead of using
+ *	  simple_setsize
  */
 int ufs_setattr(struct dentry *dentry, struct iattr *attr)
 {
@@ -518,19 +515,10 @@ int ufs_setattr(struct dentry *dentry, struct iattr *attr)
 	if (error)
 		return error;
 
-	if (is_quota_modification(inode, attr))
-		dquot_initialize(inode);
-
-	if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
-	    (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
-		error = dquot_transfer(inode, attr);
-		if (error)
-			return error;
-	}
 	if (ia_valid & ATTR_SIZE && attr->ia_size != inode->i_size) {
 		loff_t old_i_size = inode->i_size;
 
-		error = vmtruncate(inode, attr->ia_size);
+		error = simple_setsize(inode, attr->ia_size);
 		if (error)
 			return error;
 		error = ufs_truncate(inode, old_i_size);
diff --git a/fs/ufs/ufs_fs.h b/fs/ufs/ufs_fs.h
index 6943ec677c0b..8aba544f9fad 100644
--- a/fs/ufs/ufs_fs.h
+++ b/fs/ufs/ufs_fs.h
@@ -48,6 +48,7 @@ typedef __u16 __bitwise __fs16;
 #define UFS_SECTOR_SIZE 512
 #define UFS_SECTOR_BITS 9
 #define UFS_MAGIC  0x00011954
+#define UFS_MAGIC_BW 0x0f242697
 #define UFS2_MAGIC 0x19540119
 #define UFS_CIGAM  0x54190100 /* byteswapped MAGIC */
 
diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index b4769e40e8bc..c8fb13f83b3f 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -77,6 +77,7 @@ xfs-y				+= xfs_alloc.o \
 				   xfs_itable.o \
 				   xfs_dfrag.o \
 				   xfs_log.o \
+				   xfs_log_cil.o \
 				   xfs_log_recover.o \
 				   xfs_mount.o \
 				   xfs_mru_cache.o \
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index f01de3c55c43..649ade8ef598 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -37,6 +37,7 @@
 
 #include "xfs_sb.h"
 #include "xfs_inum.h"
+#include "xfs_log.h"
 #include "xfs_ag.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
@@ -850,6 +851,12 @@ xfs_buf_lock_value(
  *	Note that this in no way locks the underlying pages, so it is only
  *	useful for synchronizing concurrent use of buffer objects, not for
  *	synchronizing independent access to the underlying pages.
+ *
+ *	If we come across a stale, pinned, locked buffer, we know that we
+ *	are being asked to lock a buffer that has been reallocated. Because
+ *	it is pinned, we know that the log has not been pushed to disk and
+ *	hence it will still be locked. Rather than sleeping until someone
+ *	else pushes the log, push it ourselves before trying to get the lock.
  */
 void
 xfs_buf_lock(
@@ -857,6 +864,8 @@ xfs_buf_lock(
 {
 	trace_xfs_buf_lock(bp, _RET_IP_);
 
+	if (atomic_read(&bp->b_pin_count) && (bp->b_flags & XBF_STALE))
+		xfs_log_force(bp->b_mount, 0);
 	if (atomic_read(&bp->b_io_remaining))
 		blk_run_address_space(bp->b_target->bt_mapping);
 	down(&bp->b_sema);
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index d8fb1b5d6cb5..257a56b127cf 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -100,10 +100,10 @@ xfs_iozero(
 STATIC int
 xfs_file_fsync(
 	struct file		*file,
-	struct dentry		*dentry,
 	int			datasync)
 {
-	struct xfs_inode	*ip = XFS_I(dentry->d_inode);
+	struct inode		*inode = file->f_mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_trans	*tp;
 	int			error = 0;
 	int			log_flushed = 0;
@@ -140,8 +140,8 @@ xfs_file_fsync(
 	 * might gets cleared when the inode gets written out via the AIL
 	 * or xfs_iflush_cluster.
 	 */
-	if (((dentry->d_inode->i_state & I_DIRTY_DATASYNC) ||
-	    ((dentry->d_inode->i_state & I_DIRTY_SYNC) && !datasync)) &&
+	if (((inode->i_state & I_DIRTY_DATASYNC) ||
+	    ((inode->i_state & I_DIRTY_SYNC) && !datasync)) &&
 	    ip->i_update_core) {
 		/*
 		 * Kick off a transaction to log the inode core to get the
@@ -868,7 +868,7 @@ write_retry:
 			mutex_lock(&inode->i_mutex);
 		xfs_ilock(ip, iolock);
 
-		error2 = -xfs_file_fsync(file, file->f_path.dentry,
+		error2 = -xfs_file_fsync(file,
 					 (file->f_flags & __O_SYNC) ? 0 : 1);
 		if (!error)
 			error = error2;
diff --git a/fs/xfs/linux-2.6/xfs_quotaops.c b/fs/xfs/linux-2.6/xfs_quotaops.c
index e31bf21fe5d3..9ac8aea91529 100644
--- a/fs/xfs/linux-2.6/xfs_quotaops.c
+++ b/fs/xfs/linux-2.6/xfs_quotaops.c
@@ -19,6 +19,7 @@
 #include "xfs_dmapi.h"
 #include "xfs_sb.h"
 #include "xfs_inum.h"
+#include "xfs_log.h"
 #include "xfs_ag.h"
 #include "xfs_mount.h"
 #include "xfs_quota.h"
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index f24dbe5efde3..f2d1718c9165 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -119,6 +119,8 @@ mempool_t *xfs_ioend_pool;
 #define MNTOPT_DMAPI	"dmapi"		/* DMI enabled (DMAPI / XDSM) */
 #define MNTOPT_XDSM	"xdsm"		/* DMI enabled (DMAPI / XDSM) */
 #define MNTOPT_DMI	"dmi"		/* DMI enabled (DMAPI / XDSM) */
+#define MNTOPT_DELAYLOG   "delaylog"	/* Delayed loging enabled */
+#define MNTOPT_NODELAYLOG "nodelaylog"	/* Delayed loging disabled */
 
 /*
  * Table driven mount option parser.
@@ -374,6 +376,13 @@ xfs_parseargs(
 			mp->m_flags |= XFS_MOUNT_DMAPI;
 		} else if (!strcmp(this_char, MNTOPT_DMI)) {
 			mp->m_flags |= XFS_MOUNT_DMAPI;
+		} else if (!strcmp(this_char, MNTOPT_DELAYLOG)) {
+			mp->m_flags |= XFS_MOUNT_DELAYLOG;
+			cmn_err(CE_WARN,
+				"Enabling EXPERIMENTAL delayed logging feature "
+				"- use at your own risk.\n");
+		} else if (!strcmp(this_char, MNTOPT_NODELAYLOG)) {
+			mp->m_flags &= ~XFS_MOUNT_DELAYLOG;
 		} else if (!strcmp(this_char, "ihashsize")) {
 			cmn_err(CE_WARN,
 	"XFS: ihashsize no longer used, option is deprecated.");
@@ -535,6 +544,7 @@ xfs_showargs(
 		{ XFS_MOUNT_FILESTREAMS,	"," MNTOPT_FILESTREAM },
 		{ XFS_MOUNT_DMAPI,		"," MNTOPT_DMAPI },
 		{ XFS_MOUNT_GRPID,		"," MNTOPT_GRPID },
+		{ XFS_MOUNT_DELAYLOG,		"," MNTOPT_DELAYLOG },
 		{ 0, NULL }
 	};
 	static struct proc_xfs_info xfs_info_unset[] = {
@@ -1755,7 +1765,7 @@ xfs_init_zones(void)
 	 * but it is much faster.
 	 */
 	xfs_buf_item_zone = kmem_zone_init((sizeof(xfs_buf_log_item_t) +
-				(((XFS_MAX_BLOCKSIZE / XFS_BLI_CHUNK) /
+				(((XFS_MAX_BLOCKSIZE / XFS_BLF_CHUNK) /
 				  NBWORD) * sizeof(int))), "xfs_buf_item");
 	if (!xfs_buf_item_zone)
 		goto out_destroy_trans_zone;
diff --git a/fs/xfs/linux-2.6/xfs_trace.h b/fs/xfs/linux-2.6/xfs_trace.h
index 8a319cfd2901..ff6bc797baf2 100644
--- a/fs/xfs/linux-2.6/xfs_trace.h
+++ b/fs/xfs/linux-2.6/xfs_trace.h
@@ -1059,83 +1059,112 @@ TRACE_EVENT(xfs_bunmap,
 
 );
 
+#define XFS_BUSY_SYNC \
+	{ 0,	"async" }, \
+	{ 1,	"sync" }
+
 TRACE_EVENT(xfs_alloc_busy,
-	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno,
-		 xfs_extlen_t len, int slot),
-	TP_ARGS(mp, agno, agbno, len, slot),
+	TP_PROTO(struct xfs_trans *trans, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_extlen_t len, int sync),
+	TP_ARGS(trans, agno, agbno, len, sync),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(struct xfs_trans *, tp)
+		__field(int, tid)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, agbno)
 		__field(xfs_extlen_t, len)
-		__field(int, slot)
+		__field(int, sync)
 	),
 	TP_fast_assign(
-		__entry->dev = mp->m_super->s_dev;
+		__entry->dev = trans->t_mountp->m_super->s_dev;
+		__entry->tp = trans;
+		__entry->tid = trans->t_ticket->t_tid;
 		__entry->agno = agno;
 		__entry->agbno = agbno;
 		__entry->len = len;
-		__entry->slot = slot;
+		__entry->sync = sync;
 	),
-	TP_printk("dev %d:%d agno %u agbno %u len %u slot %d",
+	TP_printk("dev %d:%d trans 0x%p tid 0x%x agno %u agbno %u len %u %s",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->tp,
+		  __entry->tid,
 		  __entry->agno,
 		  __entry->agbno,
 		  __entry->len,
-		  __entry->slot)
+		  __print_symbolic(__entry->sync, XFS_BUSY_SYNC))
 
 );
 
-#define XFS_BUSY_STATES \
-	{ 0,	"found" }, \
-	{ 1,	"missing" }
-
 TRACE_EVENT(xfs_alloc_unbusy,
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
-		 int slot, int found),
-	TP_ARGS(mp, agno, slot, found),
+		 xfs_agblock_t agbno, xfs_extlen_t len),
+	TP_ARGS(mp, agno, agbno, len),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(xfs_agnumber_t, agno)
-		__field(int, slot)
-		__field(int, found)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
 	),
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
 		__entry->agno = agno;
-		__entry->slot = slot;
-		__entry->found = found;
+		__entry->agbno = agbno;
+		__entry->len = len;
 	),
-	TP_printk("dev %d:%d agno %u slot %d %s",
+	TP_printk("dev %d:%d agno %u agbno %u len %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->agno,
-		  __entry->slot,
-		  __print_symbolic(__entry->found, XFS_BUSY_STATES))
+		  __entry->agbno,
+		  __entry->len)
 );
 
+#define XFS_BUSY_STATES \
+	{ 0,	"missing" }, \
+	{ 1,	"found" }
+
 TRACE_EVENT(xfs_alloc_busysearch,
-	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno,
-		 xfs_extlen_t len, xfs_lsn_t lsn),
-	TP_ARGS(mp, agno, agbno, len, lsn),
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_extlen_t len, int found),
+	TP_ARGS(mp, agno, agbno, len, found),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, agbno)
 		__field(xfs_extlen_t, len)
-		__field(xfs_lsn_t, lsn)
+		__field(int, found)
 	),
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
 		__entry->agno = agno;
 		__entry->agbno = agbno;
 		__entry->len = len;
-		__entry->lsn = lsn;
+		__entry->found = found;
 	),
-	TP_printk("dev %d:%d agno %u agbno %u len %u force lsn 0x%llx",
+	TP_printk("dev %d:%d agno %u agbno %u len %u %s",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->agno,
 		  __entry->agbno,
 		  __entry->len,
+		  __print_symbolic(__entry->found, XFS_BUSY_STATES))
+);
+
+TRACE_EVENT(xfs_trans_commit_lsn,
+	TP_PROTO(struct xfs_trans *trans),
+	TP_ARGS(trans),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(struct xfs_trans *, tp)
+		__field(xfs_lsn_t, lsn)
+	),
+	TP_fast_assign(
+		__entry->dev = trans->t_mountp->m_super->s_dev;
+		__entry->tp = trans;
+		__entry->lsn = trans->t_commit_lsn;
+	),
+	TP_printk("dev %d:%d trans 0x%p commit_lsn 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->tp,
 		  __entry->lsn)
 );
 
diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c
index b89ec5df0129..585e7633dfc7 100644
--- a/fs/xfs/quota/xfs_dquot.c
+++ b/fs/xfs/quota/xfs_dquot.c
@@ -344,9 +344,9 @@ xfs_qm_init_dquot_blk(
 	for (i = 0; i < q->qi_dqperchunk; i++, d++, curid++)
 		xfs_qm_dqinit_core(curid, type, d);
 	xfs_trans_dquot_buf(tp, bp,
-			    (type & XFS_DQ_USER ? XFS_BLI_UDQUOT_BUF :
-			    ((type & XFS_DQ_PROJ) ? XFS_BLI_PDQUOT_BUF :
-			     XFS_BLI_GDQUOT_BUF)));
+			    (type & XFS_DQ_USER ? XFS_BLF_UDQUOT_BUF :
+			    ((type & XFS_DQ_PROJ) ? XFS_BLF_PDQUOT_BUF :
+			     XFS_BLF_GDQUOT_BUF)));
 	xfs_trans_log_buf(tp, bp, 0, BBTOB(q->qi_dqchunklen) - 1);
 }
 
diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h
index abb8222b88c9..401f364ad36c 100644
--- a/fs/xfs/xfs_ag.h
+++ b/fs/xfs/xfs_ag.h
@@ -175,14 +175,20 @@ typedef struct xfs_agfl {
 } xfs_agfl_t;
 
 /*
- * Busy block/extent entry.  Used in perag to mark blocks that have been freed
- * but whose transactions aren't committed to disk yet.
+ * Busy block/extent entry.  Indexed by a rbtree in perag to mark blocks that
+ * have been freed but whose transactions aren't committed to disk yet.
+ *
+ * Note that we use the transaction ID to record the transaction, not the
+ * transaction structure itself. See xfs_alloc_busy_insert() for details.
  */
-typedef struct xfs_perag_busy {
-	xfs_agblock_t	busy_start;
-	xfs_extlen_t	busy_length;
-	struct xfs_trans *busy_tp;	/* transaction that did the free */
-} xfs_perag_busy_t;
+struct xfs_busy_extent {
+	struct rb_node	rb_node;	/* ag by-bno indexed search tree */
+	struct list_head list;		/* transaction busy extent list */
+	xfs_agnumber_t	agno;
+	xfs_agblock_t	bno;
+	xfs_extlen_t	length;
+	xlog_tid_t	tid;		/* transaction that created this */
+};
 
 /*
  * Per-ag incore structure, copies of information in agf and agi,
@@ -216,7 +222,8 @@ typedef struct xfs_perag {
 	xfs_agino_t	pagl_leftrec;
 	xfs_agino_t	pagl_rightrec;
 #ifdef __KERNEL__
-	spinlock_t	pagb_lock;	/* lock for pagb_list */
+	spinlock_t	pagb_lock;	/* lock for pagb_tree */
+	struct rb_root	pagb_tree;	/* ordered tree of busy extents */
 
 	atomic_t        pagf_fstrms;    /* # of filestreams active in this AG */
 
@@ -226,7 +233,6 @@ typedef struct xfs_perag {
 	int		pag_ici_reclaimable;	/* reclaimable inodes */
 #endif
 	int		pagb_count;	/* pagb slots in use */
-	xfs_perag_busy_t pagb_list[XFS_PAGB_NUM_SLOTS];	/* unstable blocks */
 } xfs_perag_t;
 
 /*
diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c
index 94cddbfb2560..a7fbe8a99b12 100644
--- a/fs/xfs/xfs_alloc.c
+++ b/fs/xfs/xfs_alloc.c
@@ -46,11 +46,9 @@
 #define	XFSA_FIXUP_BNO_OK	1
 #define	XFSA_FIXUP_CNT_OK	2
 
-STATIC void
-xfs_alloc_search_busy(xfs_trans_t *tp,
-		    xfs_agnumber_t agno,
-		    xfs_agblock_t bno,
-		    xfs_extlen_t len);
+static int
+xfs_alloc_busy_search(struct xfs_mount *mp, xfs_agnumber_t agno,
+		    xfs_agblock_t bno, xfs_extlen_t len);
 
 /*
  * Prototypes for per-ag allocation routines
@@ -540,9 +538,16 @@ xfs_alloc_ag_vextent(
 				be32_to_cpu(agf->agf_length));
 			xfs_alloc_log_agf(args->tp, args->agbp,
 						XFS_AGF_FREEBLKS);
-			/* search the busylist for these blocks */
-			xfs_alloc_search_busy(args->tp, args->agno,
-					args->agbno, args->len);
+			/*
+			 * Search the busylist for these blocks and mark the
+			 * transaction as synchronous if blocks are found. This
+			 * avoids the need to block due to a synchronous log
+			 * force to ensure correct ordering as the synchronous
+			 * transaction will guarantee that for us.
+			 */
+			if (xfs_alloc_busy_search(args->mp, args->agno,
+						args->agbno, args->len))
+				xfs_trans_set_sync(args->tp);
 		}
 		if (!args->isfl)
 			xfs_trans_mod_sb(args->tp,
@@ -1693,7 +1698,7 @@ xfs_free_ag_extent(
 	 * when the iclog commits to disk.  If a busy block is allocated,
 	 * the iclog is pushed up to the LSN that freed the block.
 	 */
-	xfs_alloc_mark_busy(tp, agno, bno, len);
+	xfs_alloc_busy_insert(tp, agno, bno, len);
 	return 0;
 
  error0:
@@ -1989,14 +1994,20 @@ xfs_alloc_get_freelist(
 	*bnop = bno;
 
 	/*
-	 * As blocks are freed, they are added to the per-ag busy list
-	 * and remain there until the freeing transaction is committed to
-	 * disk.  Now that we have allocated blocks, this list must be
-	 * searched to see if a block is being reused.  If one is, then
-	 * the freeing transaction must be pushed to disk NOW by forcing
-	 * to disk all iclogs up that transaction's LSN.
+	 * As blocks are freed, they are added to the per-ag busy list and
+	 * remain there until the freeing transaction is committed to disk.
+	 * Now that we have allocated blocks, this list must be searched to see
+	 * if a block is being reused.  If one is, then the freeing transaction
+	 * must be pushed to disk before this transaction.
+	 *
+	 * We do this by setting the current transaction to a sync transaction
+	 * which guarantees that the freeing transaction is on disk before this
+	 * transaction. This is done instead of a synchronous log force here so
+	 * that we don't sit and wait with the AGF locked in the transaction
+	 * during the log force.
 	 */
-	xfs_alloc_search_busy(tp, be32_to_cpu(agf->agf_seqno), bno, 1);
+	if (xfs_alloc_busy_search(mp, be32_to_cpu(agf->agf_seqno), bno, 1))
+		xfs_trans_set_sync(tp);
 	return 0;
 }
 
@@ -2201,7 +2212,7 @@ xfs_alloc_read_agf(
 			be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]);
 		spin_lock_init(&pag->pagb_lock);
 		pag->pagb_count = 0;
-		memset(pag->pagb_list, 0, sizeof(pag->pagb_list));
+		pag->pagb_tree = RB_ROOT;
 		pag->pagf_init = 1;
 	}
 #ifdef DEBUG
@@ -2479,127 +2490,263 @@ error0:
  * list is reused, the transaction that freed it must be forced to disk
  * before continuing to use the block.
  *
- * xfs_alloc_mark_busy - add to the per-ag busy list
- * xfs_alloc_clear_busy - remove an item from the per-ag busy list
+ * xfs_alloc_busy_insert - add to the per-ag busy list
+ * xfs_alloc_busy_clear - remove an item from the per-ag busy list
+ * xfs_alloc_busy_search - search for a busy extent
+ */
+
+/*
+ * Insert a new extent into the busy tree.
+ *
+ * The busy extent tree is indexed by the start block of the busy extent.
+ * there can be multiple overlapping ranges in the busy extent tree but only
+ * ever one entry at a given start block. The reason for this is that
+ * multi-block extents can be freed, then smaller chunks of that extent
+ * allocated and freed again before the first transaction commit is on disk.
+ * If the exact same start block is freed a second time, we have to wait for
+ * that busy extent to pass out of the tree before the new extent is inserted.
+ * There are two main cases we have to handle here.
+ *
+ * The first case is a transaction that triggers a "free - allocate - free"
+ * cycle. This can occur during btree manipulations as a btree block is freed
+ * to the freelist, then allocated from the free list, then freed again. In
+ * this case, the second extxpnet free is what triggers the duplicate and as
+ * such the transaction IDs should match. Because the extent was allocated in
+ * this transaction, the transaction must be marked as synchronous. This is
+ * true for all cases where the free/alloc/free occurs in the one transaction,
+ * hence the addition of the ASSERT(tp->t_flags & XFS_TRANS_SYNC) to this case.
+ * This serves to catch violations of the second case quite effectively.
+ *
+ * The second case is where the free/alloc/free occur in different
+ * transactions. In this case, the thread freeing the extent the second time
+ * can't mark the extent busy immediately because it is already tracked in a
+ * transaction that may be committing.  When the log commit for the existing
+ * busy extent completes, the busy extent will be removed from the tree. If we
+ * allow the second busy insert to continue using that busy extent structure,
+ * it can be freed before this transaction is safely in the log.  Hence our
+ * only option in this case is to force the log to remove the existing busy
+ * extent from the list before we insert the new one with the current
+ * transaction ID.
+ *
+ * The problem we are trying to avoid in the free-alloc-free in separate
+ * transactions is most easily described with a timeline:
+ *
+ *      Thread 1	Thread 2	Thread 3	xfslogd
+ *	xact alloc
+ *	free X
+ *	mark busy
+ *	commit xact
+ *	free xact
+ *			xact alloc
+ *			alloc X
+ *			busy search
+ *			mark xact sync
+ *			commit xact
+ *			free xact
+ *			force log
+ *			checkpoint starts
+ *			....
+ *					xact alloc
+ *					free X
+ *					mark busy
+ *					finds match
+ *					*** KABOOM! ***
+ *					....
+ *							log IO completes
+ *							unbusy X
+ *			checkpoint completes
+ *
+ * By issuing a log force in thread 3 @ "KABOOM", the thread will block until
+ * the checkpoint completes, and the busy extent it matched will have been
+ * removed from the tree when it is woken. Hence it can then continue safely.
+ *
+ * However, to ensure this matching process is robust, we need to use the
+ * transaction ID for identifying transaction, as delayed logging results in
+ * the busy extent and transaction lifecycles being different. i.e. the busy
+ * extent is active for a lot longer than the transaction.  Hence the
+ * transaction structure can be freed and reallocated, then mark the same
+ * extent busy again in the new transaction. In this case the new transaction
+ * will have a different tid but can have the same address, and hence we need
+ * to check against the tid.
+ *
+ * Future: for delayed logging, we could avoid the log force if the extent was
+ * first freed in the current checkpoint sequence. This, however, requires the
+ * ability to pin the current checkpoint in memory until this transaction
+ * commits to ensure that both the original free and the current one combine
+ * logically into the one checkpoint. If the checkpoint sequences are
+ * different, however, we still need to wait on a log force.
  */
 void
-xfs_alloc_mark_busy(xfs_trans_t *tp,
-		    xfs_agnumber_t agno,
-		    xfs_agblock_t bno,
-		    xfs_extlen_t len)
+xfs_alloc_busy_insert(
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	xfs_agblock_t		bno,
+	xfs_extlen_t		len)
 {
-	xfs_perag_busy_t	*bsy;
+	struct xfs_busy_extent	*new;
+	struct xfs_busy_extent	*busyp;
 	struct xfs_perag	*pag;
-	int			n;
+	struct rb_node		**rbp;
+	struct rb_node		*parent;
+	int			match;
 
-	pag = xfs_perag_get(tp->t_mountp, agno);
-	spin_lock(&pag->pagb_lock);
 
-	/* search pagb_list for an open slot */
-	for (bsy = pag->pagb_list, n = 0;
-	     n < XFS_PAGB_NUM_SLOTS;
-	     bsy++, n++) {
-		if (bsy->busy_tp == NULL) {
-			break;
-		}
+	new = kmem_zalloc(sizeof(struct xfs_busy_extent), KM_MAYFAIL);
+	if (!new) {
+		/*
+		 * No Memory!  Since it is now not possible to track the free
+		 * block, make this a synchronous transaction to insure that
+		 * the block is not reused before this transaction commits.
+		 */
+		trace_xfs_alloc_busy(tp, agno, bno, len, 1);
+		xfs_trans_set_sync(tp);
+		return;
 	}
 
-	trace_xfs_alloc_busy(tp->t_mountp, agno, bno, len, n);
+	new->agno = agno;
+	new->bno = bno;
+	new->length = len;
+	new->tid = xfs_log_get_trans_ident(tp);
 
-	if (n < XFS_PAGB_NUM_SLOTS) {
-		bsy = &pag->pagb_list[n];
-		pag->pagb_count++;
-		bsy->busy_start = bno;
-		bsy->busy_length = len;
-		bsy->busy_tp = tp;
-		xfs_trans_add_busy(tp, agno, n);
-	} else {
+	INIT_LIST_HEAD(&new->list);
+
+	/* trace before insert to be able to see failed inserts */
+	trace_xfs_alloc_busy(tp, agno, bno, len, 0);
+
+	pag = xfs_perag_get(tp->t_mountp, new->agno);
+restart:
+	spin_lock(&pag->pagb_lock);
+	rbp = &pag->pagb_tree.rb_node;
+	parent = NULL;
+	busyp = NULL;
+	match = 0;
+	while (*rbp && match >= 0) {
+		parent = *rbp;
+		busyp = rb_entry(parent, struct xfs_busy_extent, rb_node);
+
+		if (new->bno < busyp->bno) {
+			/* may overlap, but exact start block is lower */
+			rbp = &(*rbp)->rb_left;
+			if (new->bno + new->length > busyp->bno)
+				match = busyp->tid == new->tid ? 1 : -1;
+		} else if (new->bno > busyp->bno) {
+			/* may overlap, but exact start block is higher */
+			rbp = &(*rbp)->rb_right;
+			if (bno < busyp->bno + busyp->length)
+				match = busyp->tid == new->tid ? 1 : -1;
+		} else {
+			match = busyp->tid == new->tid ? 1 : -1;
+			break;
+		}
+	}
+	if (match < 0) {
+		/* overlap marked busy in different transaction */
+		spin_unlock(&pag->pagb_lock);
+		xfs_log_force(tp->t_mountp, XFS_LOG_SYNC);
+		goto restart;
+	}
+	if (match > 0) {
 		/*
-		 * The busy list is full!  Since it is now not possible to
-		 * track the free block, make this a synchronous transaction
-		 * to insure that the block is not reused before this
-		 * transaction commits.
+		 * overlap marked busy in same transaction. Update if exact
+		 * start block match, otherwise combine the busy extents into
+		 * a single range.
 		 */
-		xfs_trans_set_sync(tp);
-	}
+		if (busyp->bno == new->bno) {
+			busyp->length = max(busyp->length, new->length);
+			spin_unlock(&pag->pagb_lock);
+			ASSERT(tp->t_flags & XFS_TRANS_SYNC);
+			xfs_perag_put(pag);
+			kmem_free(new);
+			return;
+		}
+		rb_erase(&busyp->rb_node, &pag->pagb_tree);
+		new->length = max(busyp->bno + busyp->length,
+					new->bno + new->length) -
+				min(busyp->bno, new->bno);
+		new->bno = min(busyp->bno, new->bno);
+	} else
+		busyp = NULL;
 
+	rb_link_node(&new->rb_node, parent, rbp);
+	rb_insert_color(&new->rb_node, &pag->pagb_tree);
+
+	list_add(&new->list, &tp->t_busy);
 	spin_unlock(&pag->pagb_lock);
 	xfs_perag_put(pag);
+	kmem_free(busyp);
 }
 
-void
-xfs_alloc_clear_busy(xfs_trans_t *tp,
-		     xfs_agnumber_t agno,
-		     int idx)
+/*
+ * Search for a busy extent within the range of the extent we are about to
+ * allocate.  You need to be holding the busy extent tree lock when calling
+ * xfs_alloc_busy_search(). This function returns 0 for no overlapping busy
+ * extent, -1 for an overlapping but not exact busy extent, and 1 for an exact
+ * match. This is done so that a non-zero return indicates an overlap that
+ * will require a synchronous transaction, but it can still be
+ * used to distinguish between a partial or exact match.
+ */
+static int
+xfs_alloc_busy_search(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno,
+	xfs_agblock_t		bno,
+	xfs_extlen_t		len)
 {
 	struct xfs_perag	*pag;
-	xfs_perag_busy_t	*list;
+	struct rb_node		*rbp;
+	struct xfs_busy_extent	*busyp;
+	int			match = 0;
 
-	ASSERT(idx < XFS_PAGB_NUM_SLOTS);
-	pag = xfs_perag_get(tp->t_mountp, agno);
+	pag = xfs_perag_get(mp, agno);
 	spin_lock(&pag->pagb_lock);
-	list = pag->pagb_list;
 
-	trace_xfs_alloc_unbusy(tp->t_mountp, agno, idx, list[idx].busy_tp == tp);
-
-	if (list[idx].busy_tp == tp) {
-		list[idx].busy_tp = NULL;
-		pag->pagb_count--;
+	rbp = pag->pagb_tree.rb_node;
+
+	/* find closest start bno overlap */
+	while (rbp) {
+		busyp = rb_entry(rbp, struct xfs_busy_extent, rb_node);
+		if (bno < busyp->bno) {
+			/* may overlap, but exact start block is lower */
+			if (bno + len > busyp->bno)
+				match = -1;
+			rbp = rbp->rb_left;
+		} else if (bno > busyp->bno) {
+			/* may overlap, but exact start block is higher */
+			if (bno < busyp->bno + busyp->length)
+				match = -1;
+			rbp = rbp->rb_right;
+		} else {
+			/* bno matches busyp, length determines exact match */
+			match = (busyp->length == len) ? 1 : -1;
+			break;
+		}
 	}
-
 	spin_unlock(&pag->pagb_lock);
+	trace_xfs_alloc_busysearch(mp, agno, bno, len, !!match);
 	xfs_perag_put(pag);
+	return match;
 }
 
-
-/*
- * If we find the extent in the busy list, force the log out to get the
- * extent out of the busy list so the caller can use it straight away.
- */
-STATIC void
-xfs_alloc_search_busy(xfs_trans_t *tp,
-		    xfs_agnumber_t agno,
-		    xfs_agblock_t bno,
-		    xfs_extlen_t len)
+void
+xfs_alloc_busy_clear(
+	struct xfs_mount	*mp,
+	struct xfs_busy_extent	*busyp)
 {
 	struct xfs_perag	*pag;
-	xfs_perag_busy_t	*bsy;
-	xfs_agblock_t		uend, bend;
-	xfs_lsn_t		lsn = 0;
-	int			cnt;
 
-	pag = xfs_perag_get(tp->t_mountp, agno);
-	spin_lock(&pag->pagb_lock);
-	cnt = pag->pagb_count;
+	trace_xfs_alloc_unbusy(mp, busyp->agno, busyp->bno,
+						busyp->length);
 
-	/*
-	 * search pagb_list for this slot, skipping open slots. We have to
-	 * search the entire array as there may be multiple overlaps and
-	 * we have to get the most recent LSN for the log force to push out
-	 * all the transactions that span the range.
-	 */
-	uend = bno + len - 1;
-	for (cnt = 0; cnt < pag->pagb_count; cnt++) {
-		bsy = &pag->pagb_list[cnt];
-		if (!bsy->busy_tp)
-			continue;
+	ASSERT(xfs_alloc_busy_search(mp, busyp->agno, busyp->bno,
+						busyp->length) == 1);
 
-		bend = bsy->busy_start + bsy->busy_length - 1;
-		if (bno > bend || uend < bsy->busy_start)
-			continue;
+	list_del_init(&busyp->list);
 
-		/* (start1,length1) within (start2, length2) */
-		if (XFS_LSN_CMP(bsy->busy_tp->t_commit_lsn, lsn) > 0)
-			lsn = bsy->busy_tp->t_commit_lsn;
-	}
+	pag = xfs_perag_get(mp, busyp->agno);
+	spin_lock(&pag->pagb_lock);
+	rb_erase(&busyp->rb_node, &pag->pagb_tree);
 	spin_unlock(&pag->pagb_lock);
 	xfs_perag_put(pag);
-	trace_xfs_alloc_busysearch(tp->t_mountp, agno, bno, len, lsn);
 
-	/*
-	 * If a block was found, force the log through the LSN of the
-	 * transaction that freed the block
-	 */
-	if (lsn)
-		xfs_log_force_lsn(tp->t_mountp, lsn, XFS_LOG_SYNC);
+	kmem_free(busyp);
 }
diff --git a/fs/xfs/xfs_alloc.h b/fs/xfs/xfs_alloc.h
index 599bffa39784..6d05199b667c 100644
--- a/fs/xfs/xfs_alloc.h
+++ b/fs/xfs/xfs_alloc.h
@@ -22,6 +22,7 @@ struct xfs_buf;
 struct xfs_mount;
 struct xfs_perag;
 struct xfs_trans;
+struct xfs_busy_extent;
 
 /*
  * Freespace allocation types.  Argument to xfs_alloc_[v]extent.
@@ -119,15 +120,13 @@ xfs_alloc_longest_free_extent(struct xfs_mount *mp,
 #ifdef __KERNEL__
 
 void
-xfs_alloc_mark_busy(xfs_trans_t *tp,
+xfs_alloc_busy_insert(xfs_trans_t *tp,
 		xfs_agnumber_t agno,
 		xfs_agblock_t bno,
 		xfs_extlen_t len);
 
 void
-xfs_alloc_clear_busy(xfs_trans_t *tp,
-		xfs_agnumber_t ag,
-		int idx);
+xfs_alloc_busy_clear(struct xfs_mount *mp, struct xfs_busy_extent *busyp);
 
 #endif	/* __KERNEL__ */
 
diff --git a/fs/xfs/xfs_alloc_btree.c b/fs/xfs/xfs_alloc_btree.c
index b726e10d2c1c..83f494218759 100644
--- a/fs/xfs/xfs_alloc_btree.c
+++ b/fs/xfs/xfs_alloc_btree.c
@@ -134,7 +134,7 @@ xfs_allocbt_free_block(
 	 * disk. If a busy block is allocated, the iclog is pushed up to the
 	 * LSN that freed the block.
 	 */
-	xfs_alloc_mark_busy(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1);
+	xfs_alloc_busy_insert(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1);
 	xfs_trans_agbtree_delta(cur->bc_tp, -1);
 	return 0;
 }
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 240340a4727b..02a80984aa05 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -64,7 +64,7 @@ xfs_buf_item_log_debug(
 	nbytes = last - first + 1;
 	bfset(bip->bli_logged, first, nbytes);
 	for (x = 0; x < nbytes; x++) {
-		chunk_num = byte >> XFS_BLI_SHIFT;
+		chunk_num = byte >> XFS_BLF_SHIFT;
 		word_num = chunk_num >> BIT_TO_WORD_SHIFT;
 		bit_num = chunk_num & (NBWORD - 1);
 		wordp = &(bip->bli_format.blf_data_map[word_num]);
@@ -166,7 +166,7 @@ xfs_buf_item_size(
 		 * cancel flag in it.
 		 */
 		trace_xfs_buf_item_size_stale(bip);
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
+		ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
 		return 1;
 	}
 
@@ -197,9 +197,9 @@ xfs_buf_item_size(
 		} else if (next_bit != last_bit + 1) {
 			last_bit = next_bit;
 			nvecs++;
-		} else if (xfs_buf_offset(bp, next_bit * XFS_BLI_CHUNK) !=
-			   (xfs_buf_offset(bp, last_bit * XFS_BLI_CHUNK) +
-			    XFS_BLI_CHUNK)) {
+		} else if (xfs_buf_offset(bp, next_bit * XFS_BLF_CHUNK) !=
+			   (xfs_buf_offset(bp, last_bit * XFS_BLF_CHUNK) +
+			    XFS_BLF_CHUNK)) {
 			last_bit = next_bit;
 			nvecs++;
 		} else {
@@ -254,6 +254,20 @@ xfs_buf_item_format(
 	vecp++;
 	nvecs = 1;
 
+	/*
+	 * If it is an inode buffer, transfer the in-memory state to the
+	 * format flags and clear the in-memory state. We do not transfer
+	 * this state if the inode buffer allocation has not yet been committed
+	 * to the log as setting the XFS_BLI_INODE_BUF flag will prevent
+	 * correct replay of the inode allocation.
+	 */
+	if (bip->bli_flags & XFS_BLI_INODE_BUF) {
+		if (!((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) &&
+		      xfs_log_item_in_current_chkpt(&bip->bli_item)))
+			bip->bli_format.blf_flags |= XFS_BLF_INODE_BUF;
+		bip->bli_flags &= ~XFS_BLI_INODE_BUF;
+	}
+
 	if (bip->bli_flags & XFS_BLI_STALE) {
 		/*
 		 * The buffer is stale, so all we need to log
@@ -261,7 +275,7 @@ xfs_buf_item_format(
 		 * cancel flag in it.
 		 */
 		trace_xfs_buf_item_format_stale(bip);
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
+		ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
 		bip->bli_format.blf_size = nvecs;
 		return;
 	}
@@ -294,28 +308,28 @@ xfs_buf_item_format(
 		 * keep counting and scanning.
 		 */
 		if (next_bit == -1) {
-			buffer_offset = first_bit * XFS_BLI_CHUNK;
+			buffer_offset = first_bit * XFS_BLF_CHUNK;
 			vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
-			vecp->i_len = nbits * XFS_BLI_CHUNK;
+			vecp->i_len = nbits * XFS_BLF_CHUNK;
 			vecp->i_type = XLOG_REG_TYPE_BCHUNK;
 			nvecs++;
 			break;
 		} else if (next_bit != last_bit + 1) {
-			buffer_offset = first_bit * XFS_BLI_CHUNK;
+			buffer_offset = first_bit * XFS_BLF_CHUNK;
 			vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
-			vecp->i_len = nbits * XFS_BLI_CHUNK;
+			vecp->i_len = nbits * XFS_BLF_CHUNK;
 			vecp->i_type = XLOG_REG_TYPE_BCHUNK;
 			nvecs++;
 			vecp++;
 			first_bit = next_bit;
 			last_bit = next_bit;
 			nbits = 1;
-		} else if (xfs_buf_offset(bp, next_bit << XFS_BLI_SHIFT) !=
-			   (xfs_buf_offset(bp, last_bit << XFS_BLI_SHIFT) +
-			    XFS_BLI_CHUNK)) {
-			buffer_offset = first_bit * XFS_BLI_CHUNK;
+		} else if (xfs_buf_offset(bp, next_bit << XFS_BLF_SHIFT) !=
+			   (xfs_buf_offset(bp, last_bit << XFS_BLF_SHIFT) +
+			    XFS_BLF_CHUNK)) {
+			buffer_offset = first_bit * XFS_BLF_CHUNK;
 			vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
-			vecp->i_len = nbits * XFS_BLI_CHUNK;
+			vecp->i_len = nbits * XFS_BLF_CHUNK;
 			vecp->i_type = XLOG_REG_TYPE_BCHUNK;
 /* You would think we need to bump the nvecs here too, but we do not
  * this number is used by recovery, and it gets confused by the boundary
@@ -341,10 +355,15 @@ xfs_buf_item_format(
 }
 
 /*
- * This is called to pin the buffer associated with the buf log
- * item in memory so it cannot be written out.  Simply call bpin()
- * on the buffer to do this.
+ * This is called to pin the buffer associated with the buf log item in memory
+ * so it cannot be written out.  Simply call bpin() on the buffer to do this.
+ *
+ * We also always take a reference to the buffer log item here so that the bli
+ * is held while the item is pinned in memory. This means that we can
+ * unconditionally drop the reference count a transaction holds when the
+ * transaction is completed.
  */
+
 STATIC void
 xfs_buf_item_pin(
 	xfs_buf_log_item_t	*bip)
@@ -356,6 +375,7 @@ xfs_buf_item_pin(
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 	ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
 	       (bip->bli_flags & XFS_BLI_STALE));
+	atomic_inc(&bip->bli_refcount);
 	trace_xfs_buf_item_pin(bip);
 	xfs_bpin(bp);
 }
@@ -393,7 +413,7 @@ xfs_buf_item_unpin(
 		ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
 		ASSERT(!(XFS_BUF_ISDELAYWRITE(bp)));
 		ASSERT(XFS_BUF_ISSTALE(bp));
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
+		ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
 		trace_xfs_buf_item_unpin_stale(bip);
 
 		/*
@@ -489,20 +509,23 @@ xfs_buf_item_trylock(
 }
 
 /*
- * Release the buffer associated with the buf log item.
- * If there is no dirty logged data associated with the
- * buffer recorded in the buf log item, then free the
- * buf log item and remove the reference to it in the
- * buffer.
+ * Release the buffer associated with the buf log item.  If there is no dirty
+ * logged data associated with the buffer recorded in the buf log item, then
+ * free the buf log item and remove the reference to it in the buffer.
+ *
+ * This call ignores the recursion count.  It is only called when the buffer
+ * should REALLY be unlocked, regardless of the recursion count.
  *
- * This call ignores the recursion count.  It is only called
- * when the buffer should REALLY be unlocked, regardless
- * of the recursion count.
+ * We unconditionally drop the transaction's reference to the log item. If the
+ * item was logged, then another reference was taken when it was pinned, so we
+ * can safely drop the transaction reference now.  This also allows us to avoid
+ * potential races with the unpin code freeing the bli by not referencing the
+ * bli after we've dropped the reference count.
  *
- * If the XFS_BLI_HOLD flag is set in the buf log item, then
- * free the log item if necessary but do not unlock the buffer.
- * This is for support of xfs_trans_bhold(). Make sure the
- * XFS_BLI_HOLD field is cleared if we don't free the item.
+ * If the XFS_BLI_HOLD flag is set in the buf log item, then free the log item
+ * if necessary but do not unlock the buffer.  This is for support of
+ * xfs_trans_bhold(). Make sure the XFS_BLI_HOLD field is cleared if we don't
+ * free the item.
  */
 STATIC void
 xfs_buf_item_unlock(
@@ -514,73 +537,54 @@ xfs_buf_item_unlock(
 
 	bp = bip->bli_buf;
 
-	/*
-	 * Clear the buffer's association with this transaction.
-	 */
+	/* Clear the buffer's association with this transaction. */
 	XFS_BUF_SET_FSPRIVATE2(bp, NULL);
 
 	/*
-	 * If this is a transaction abort, don't return early.
-	 * Instead, allow the brelse to happen.
-	 * Normally it would be done for stale (cancelled) buffers
-	 * at unpin time, but we'll never go through the pin/unpin
-	 * cycle if we abort inside commit.
+	 * If this is a transaction abort, don't return early.  Instead, allow
+	 * the brelse to happen.  Normally it would be done for stale
+	 * (cancelled) buffers at unpin time, but we'll never go through the
+	 * pin/unpin cycle if we abort inside commit.
 	 */
 	aborted = (bip->bli_item.li_flags & XFS_LI_ABORTED) != 0;
 
 	/*
-	 * If the buf item is marked stale, then don't do anything.
-	 * We'll unlock the buffer and free the buf item when the
-	 * buffer is unpinned for the last time.
+	 * Before possibly freeing the buf item, determine if we should
+	 * release the buffer at the end of this routine.
 	 */
-	if (bip->bli_flags & XFS_BLI_STALE) {
-		bip->bli_flags &= ~XFS_BLI_LOGGED;
-		trace_xfs_buf_item_unlock_stale(bip);
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
-		if (!aborted)
-			return;
-	}
+	hold = bip->bli_flags & XFS_BLI_HOLD;
+
+	/* Clear the per transaction state. */
+	bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_HOLD);
 
 	/*
-	 * Drop the transaction's reference to the log item if
-	 * it was not logged as part of the transaction.  Otherwise
-	 * we'll drop the reference in xfs_buf_item_unpin() when
-	 * the transaction is really through with the buffer.
+	 * If the buf item is marked stale, then don't do anything.  We'll
+	 * unlock the buffer and free the buf item when the buffer is unpinned
+	 * for the last time.
 	 */
-	if (!(bip->bli_flags & XFS_BLI_LOGGED)) {
-		atomic_dec(&bip->bli_refcount);
-	} else {
-		/*
-		 * Clear the logged flag since this is per
-		 * transaction state.
-		 */
-		bip->bli_flags &= ~XFS_BLI_LOGGED;
+	if (bip->bli_flags & XFS_BLI_STALE) {
+		trace_xfs_buf_item_unlock_stale(bip);
+		ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
+		if (!aborted) {
+			atomic_dec(&bip->bli_refcount);
+			return;
+		}
 	}
 
-	/*
-	 * Before possibly freeing the buf item, determine if we should
-	 * release the buffer at the end of this routine.
-	 */
-	hold = bip->bli_flags & XFS_BLI_HOLD;
 	trace_xfs_buf_item_unlock(bip);
 
 	/*
-	 * If the buf item isn't tracking any data, free it.
-	 * Otherwise, if XFS_BLI_HOLD is set clear it.
+	 * If the buf item isn't tracking any data, free it, otherwise drop the
+	 * reference we hold to it.
 	 */
 	if (xfs_bitmap_empty(bip->bli_format.blf_data_map,
-			     bip->bli_format.blf_map_size)) {
+			     bip->bli_format.blf_map_size))
 		xfs_buf_item_relse(bp);
-	} else if (hold) {
-		bip->bli_flags &= ~XFS_BLI_HOLD;
-	}
+	else
+		atomic_dec(&bip->bli_refcount);
 
-	/*
-	 * Release the buffer if XFS_BLI_HOLD was not set.
-	 */
-	if (!hold) {
+	if (!hold)
 		xfs_buf_relse(bp);
-	}
 }
 
 /*
@@ -717,12 +721,12 @@ xfs_buf_item_init(
 	}
 
 	/*
-	 * chunks is the number of XFS_BLI_CHUNK size pieces
+	 * chunks is the number of XFS_BLF_CHUNK size pieces
 	 * the buffer can be divided into. Make sure not to
 	 * truncate any pieces.  map_size is the size of the
 	 * bitmap needed to describe the chunks of the buffer.
 	 */
-	chunks = (int)((XFS_BUF_COUNT(bp) + (XFS_BLI_CHUNK - 1)) >> XFS_BLI_SHIFT);
+	chunks = (int)((XFS_BUF_COUNT(bp) + (XFS_BLF_CHUNK - 1)) >> XFS_BLF_SHIFT);
 	map_size = (int)((chunks + NBWORD) >> BIT_TO_WORD_SHIFT);
 
 	bip = (xfs_buf_log_item_t*)kmem_zone_zalloc(xfs_buf_item_zone,
@@ -790,8 +794,8 @@ xfs_buf_item_log(
 	/*
 	 * Convert byte offsets to bit numbers.
 	 */
-	first_bit = first >> XFS_BLI_SHIFT;
-	last_bit = last >> XFS_BLI_SHIFT;
+	first_bit = first >> XFS_BLF_SHIFT;
+	last_bit = last >> XFS_BLF_SHIFT;
 
 	/*
 	 * Calculate the total number of bits to be set.
diff --git a/fs/xfs/xfs_buf_item.h b/fs/xfs/xfs_buf_item.h
index df4454511f73..f20bb472d582 100644
--- a/fs/xfs/xfs_buf_item.h
+++ b/fs/xfs/xfs_buf_item.h
@@ -41,22 +41,22 @@ typedef struct xfs_buf_log_format {
  * This flag indicates that the buffer contains on disk inodes
  * and requires special recovery handling.
  */
-#define	XFS_BLI_INODE_BUF	0x1
+#define	XFS_BLF_INODE_BUF	0x1
 /*
  * This flag indicates that the buffer should not be replayed
  * during recovery because its blocks are being freed.
  */
-#define	XFS_BLI_CANCEL		0x2
+#define	XFS_BLF_CANCEL		0x2
 /*
  * This flag indicates that the buffer contains on disk
  * user or group dquots and may require special recovery handling.
  */
-#define	XFS_BLI_UDQUOT_BUF	0x4
-#define XFS_BLI_PDQUOT_BUF	0x8
-#define	XFS_BLI_GDQUOT_BUF	0x10
+#define	XFS_BLF_UDQUOT_BUF	0x4
+#define XFS_BLF_PDQUOT_BUF	0x8
+#define	XFS_BLF_GDQUOT_BUF	0x10
 
-#define	XFS_BLI_CHUNK		128
-#define	XFS_BLI_SHIFT		7
+#define	XFS_BLF_CHUNK		128
+#define	XFS_BLF_SHIFT		7
 #define	BIT_TO_WORD_SHIFT	5
 #define	NBWORD			(NBBY * sizeof(unsigned int))
 
@@ -69,6 +69,7 @@ typedef struct xfs_buf_log_format {
 #define	XFS_BLI_LOGGED		0x08
 #define	XFS_BLI_INODE_ALLOC_BUF	0x10
 #define XFS_BLI_STALE_INODE	0x20
+#define	XFS_BLI_INODE_BUF	0x40
 
 #define XFS_BLI_FLAGS \
 	{ XFS_BLI_HOLD,		"HOLD" }, \
@@ -76,7 +77,8 @@ typedef struct xfs_buf_log_format {
 	{ XFS_BLI_STALE,	"STALE" }, \
 	{ XFS_BLI_LOGGED,	"LOGGED" }, \
 	{ XFS_BLI_INODE_ALLOC_BUF, "INODE_ALLOC" }, \
-	{ XFS_BLI_STALE_INODE,	"STALE_INODE" }
+	{ XFS_BLI_STALE_INODE,	"STALE_INODE" }, \
+	{ XFS_BLI_INODE_BUF,	"INODE_BUF" }
 
 
 #ifdef __KERNEL__
diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
index ef96175c0744..047b8a8e5c29 100644
--- a/fs/xfs/xfs_error.c
+++ b/fs/xfs/xfs_error.c
@@ -170,7 +170,7 @@ xfs_cmn_err(int panic_tag, int level, xfs_mount_t *mp, char *fmt, ...)
 	va_list ap;
 
 #ifdef DEBUG
-	xfs_panic_mask |= XFS_PTAG_SHUTDOWN_CORRUPT;
+	xfs_panic_mask |= (XFS_PTAG_SHUTDOWN_CORRUPT | XFS_PTAG_LOGRES);
 #endif
 
 	if (xfs_panic_mask && (xfs_panic_mask & panic_tag)
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index 3038dd52c72a..5215abc8023a 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -54,9 +54,6 @@ STATIC xlog_t *  xlog_alloc_log(xfs_mount_t	*mp,
 STATIC int	 xlog_space_left(xlog_t *log, int cycle, int bytes);
 STATIC int	 xlog_sync(xlog_t *log, xlog_in_core_t *iclog);
 STATIC void	 xlog_dealloc_log(xlog_t *log);
-STATIC int	 xlog_write(struct log *log, struct xfs_log_vec *log_vector,
-			    struct xlog_ticket *tic, xfs_lsn_t *start_lsn,
-			    xlog_in_core_t **commit_iclog, uint flags);
 
 /* local state machine functions */
 STATIC void xlog_state_done_syncing(xlog_in_core_t *iclog, int);
@@ -86,14 +83,6 @@ STATIC int xlog_regrant_write_log_space(xlog_t		*log,
 STATIC void xlog_ungrant_log_space(xlog_t	 *log,
 				   xlog_ticket_t *ticket);
 
-
-/* local ticket functions */
-STATIC xlog_ticket_t	*xlog_ticket_alloc(xlog_t *log,
-					 int	unit_bytes,
-					 int	count,
-					 char	clientid,
-					 uint	flags);
-
 #if defined(DEBUG)
 STATIC void	xlog_verify_dest_ptr(xlog_t *log, char *ptr);
 STATIC void	xlog_verify_grant_head(xlog_t *log, int equals);
@@ -360,6 +349,15 @@ xfs_log_reserve(
 		ASSERT(flags & XFS_LOG_PERM_RESERV);
 		internal_ticket = *ticket;
 
+		/*
+		 * this is a new transaction on the ticket, so we need to
+		 * change the transaction ID so that the next transaction has a
+		 * different TID in the log. Just add one to the existing tid
+		 * so that we can see chains of rolling transactions in the log
+		 * easily.
+		 */
+		internal_ticket->t_tid++;
+
 		trace_xfs_log_reserve(log, internal_ticket);
 
 		xlog_grant_push_ail(mp, internal_ticket->t_unit_res);
@@ -367,7 +365,8 @@ xfs_log_reserve(
 	} else {
 		/* may sleep if need to allocate more tickets */
 		internal_ticket = xlog_ticket_alloc(log, unit_bytes, cnt,
-						  client, flags);
+						  client, flags,
+						  KM_SLEEP|KM_MAYFAIL);
 		if (!internal_ticket)
 			return XFS_ERROR(ENOMEM);
 		internal_ticket->t_trans_type = t_type;
@@ -452,6 +451,13 @@ xfs_log_mount(
 	/* Normal transactions can now occur */
 	mp->m_log->l_flags &= ~XLOG_ACTIVE_RECOVERY;
 
+	/*
+	 * Now the log has been fully initialised and we know were our
+	 * space grant counters are, we can initialise the permanent ticket
+	 * needed for delayed logging to work.
+	 */
+	xlog_cil_init_post_recovery(mp->m_log);
+
 	return 0;
 
 out_destroy_ail:
@@ -658,6 +664,10 @@ xfs_log_item_init(
 	item->li_ailp = mp->m_ail;
 	item->li_type = type;
 	item->li_ops = ops;
+	item->li_lv = NULL;
+
+	INIT_LIST_HEAD(&item->li_ail);
+	INIT_LIST_HEAD(&item->li_cil);
 }
 
 /*
@@ -1168,6 +1178,9 @@ xlog_alloc_log(xfs_mount_t	*mp,
 	*iclogp = log->l_iclog;			/* complete ring */
 	log->l_iclog->ic_prev = prev_iclog;	/* re-write 1st prev ptr */
 
+	error = xlog_cil_init(log);
+	if (error)
+		goto out_free_iclog;
 	return log;
 
 out_free_iclog:
@@ -1494,6 +1507,8 @@ xlog_dealloc_log(xlog_t *log)
 	xlog_in_core_t	*iclog, *next_iclog;
 	int		i;
 
+	xlog_cil_destroy(log);
+
 	iclog = log->l_iclog;
 	for (i=0; i<log->l_iclog_bufs; i++) {
 		sv_destroy(&iclog->ic_force_wait);
@@ -1536,8 +1551,10 @@ xlog_state_finish_copy(xlog_t		*log,
  * print out info relating to regions written which consume
  * the reservation
  */
-STATIC void
-xlog_print_tic_res(xfs_mount_t *mp, xlog_ticket_t *ticket)
+void
+xlog_print_tic_res(
+	struct xfs_mount	*mp,
+	struct xlog_ticket	*ticket)
 {
 	uint i;
 	uint ophdr_spc = ticket->t_res_num_ophdrs * (uint)sizeof(xlog_op_header_t);
@@ -1637,6 +1654,10 @@ xlog_print_tic_res(xfs_mount_t *mp, xlog_ticket_t *ticket)
 			    "bad-rtype" : res_type_str[r_type-1]),
 			    ticket->t_res_arr[i].r_len);
 	}
+
+	xfs_cmn_err(XFS_PTAG_LOGRES, CE_ALERT, mp,
+		"xfs_log_write: reservation ran out. Need to up reservation");
+	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
 }
 
 /*
@@ -1865,7 +1886,7 @@ xlog_write_copy_finish(
  *	we don't update ic_offset until the end when we know exactly how many
  *	bytes have been written out.
  */
-STATIC int
+int
 xlog_write(
 	struct log		*log,
 	struct xfs_log_vec	*log_vector,
@@ -1889,22 +1910,26 @@ xlog_write(
 	*start_lsn = 0;
 
 	len = xlog_write_calc_vec_length(ticket, log_vector);
-	if (ticket->t_curr_res < len) {
-		xlog_print_tic_res(log->l_mp, ticket);
-#ifdef DEBUG
-		xlog_panic(
-	"xfs_log_write: reservation ran out. Need to up reservation");
-#else
-		/* Customer configurable panic */
-		xfs_cmn_err(XFS_PTAG_LOGRES, CE_ALERT, log->l_mp,
-	"xfs_log_write: reservation ran out. Need to up reservation");
+	if (log->l_cilp) {
+		/*
+		 * Region headers and bytes are already accounted for.
+		 * We only need to take into account start records and
+		 * split regions in this function.
+		 */
+		if (ticket->t_flags & XLOG_TIC_INITED)
+			ticket->t_curr_res -= sizeof(xlog_op_header_t);
 
-		/* If we did not panic, shutdown the filesystem */
-		xfs_force_shutdown(log->l_mp, SHUTDOWN_CORRUPT_INCORE);
-#endif
-	}
+		/*
+		 * Commit record headers need to be accounted for. These
+		 * come in as separate writes so are easy to detect.
+		 */
+		if (flags & (XLOG_COMMIT_TRANS | XLOG_UNMOUNT_TRANS))
+			ticket->t_curr_res -= sizeof(xlog_op_header_t);
+	} else
+		ticket->t_curr_res -= len;
 
-	ticket->t_curr_res -= len;
+	if (ticket->t_curr_res < 0)
+		xlog_print_tic_res(log->l_mp, ticket);
 
 	index = 0;
 	lv = log_vector;
@@ -3000,6 +3025,8 @@ _xfs_log_force(
 
 	XFS_STATS_INC(xs_log_force);
 
+	xlog_cil_push(log, 1);
+
 	spin_lock(&log->l_icloglock);
 
 	iclog = log->l_iclog;
@@ -3149,6 +3176,12 @@ _xfs_log_force_lsn(
 
 	XFS_STATS_INC(xs_log_force);
 
+	if (log->l_cilp) {
+		lsn = xlog_cil_push_lsn(log, lsn);
+		if (lsn == NULLCOMMITLSN)
+			return 0;
+	}
+
 try_again:
 	spin_lock(&log->l_icloglock);
 	iclog = log->l_iclog;
@@ -3313,22 +3346,30 @@ xfs_log_ticket_get(
 	return ticket;
 }
 
+xlog_tid_t
+xfs_log_get_trans_ident(
+	struct xfs_trans	*tp)
+{
+	return tp->t_ticket->t_tid;
+}
+
 /*
  * Allocate and initialise a new log ticket.
  */
-STATIC xlog_ticket_t *
+xlog_ticket_t *
 xlog_ticket_alloc(
 	struct log	*log,
 	int		unit_bytes,
 	int		cnt,
 	char		client,
-	uint		xflags)
+	uint		xflags,
+	int		alloc_flags)
 {
 	struct xlog_ticket *tic;
 	uint		num_headers;
 	int		iclog_space;
 
-	tic = kmem_zone_zalloc(xfs_log_ticket_zone, KM_SLEEP|KM_MAYFAIL);
+	tic = kmem_zone_zalloc(xfs_log_ticket_zone, alloc_flags);
 	if (!tic)
 		return NULL;
 
@@ -3647,6 +3688,11 @@ xlog_state_ioerror(
  *	c. nothing new gets queued up after (a) and (b) are done.
  *	d. if !logerror, flush the iclogs to disk, then seal them off
  *	   for business.
+ *
+ * Note: for delayed logging the !logerror case needs to flush the regions
+ * held in memory out to the iclogs before flushing them to disk. This needs
+ * to be done before the log is marked as shutdown, otherwise the flush to the
+ * iclogs will fail.
  */
 int
 xfs_log_force_umount(
@@ -3680,6 +3726,16 @@ xfs_log_force_umount(
 		return 1;
 	}
 	retval = 0;
+
+	/*
+	 * Flush the in memory commit item list before marking the log as
+	 * being shut down. We need to do it in this order to ensure all the
+	 * completed transactions are flushed to disk with the xfs_log_force()
+	 * call below.
+	 */
+	if (!logerror && (mp->m_flags & XFS_MOUNT_DELAYLOG))
+		xlog_cil_push(log, 1);
+
 	/*
 	 * We must hold both the GRANT lock and the LOG lock,
 	 * before we mark the filesystem SHUTDOWN and wake
diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h
index 229d1f36ba9a..04c78e642cc8 100644
--- a/fs/xfs/xfs_log.h
+++ b/fs/xfs/xfs_log.h
@@ -19,7 +19,6 @@
 #define __XFS_LOG_H__
 
 /* get lsn fields */
-
 #define CYCLE_LSN(lsn) ((uint)((lsn)>>32))
 #define BLOCK_LSN(lsn) ((uint)(lsn))
 
@@ -114,6 +113,9 @@ struct xfs_log_vec {
 	struct xfs_log_vec	*lv_next;	/* next lv in build list */
 	int			lv_niovecs;	/* number of iovecs in lv */
 	struct xfs_log_iovec	*lv_iovecp;	/* iovec array */
+	struct xfs_log_item	*lv_item;	/* owner */
+	char			*lv_buf;	/* formatted buffer */
+	int			lv_buf_len;	/* size of formatted buffer */
 };
 
 /*
@@ -134,6 +136,7 @@ struct xlog_in_core;
 struct xlog_ticket;
 struct xfs_log_item;
 struct xfs_item_ops;
+struct xfs_trans;
 
 void	xfs_log_item_init(struct xfs_mount	*mp,
 			struct xfs_log_item	*item,
@@ -187,9 +190,16 @@ int	  xfs_log_need_covered(struct xfs_mount *mp);
 
 void	  xlog_iodone(struct xfs_buf *);
 
-struct xlog_ticket * xfs_log_ticket_get(struct xlog_ticket *ticket);
+struct xlog_ticket *xfs_log_ticket_get(struct xlog_ticket *ticket);
 void	  xfs_log_ticket_put(struct xlog_ticket *ticket);
 
+xlog_tid_t xfs_log_get_trans_ident(struct xfs_trans *tp);
+
+int	xfs_log_commit_cil(struct xfs_mount *mp, struct xfs_trans *tp,
+				struct xfs_log_vec *log_vector,
+				xfs_lsn_t *commit_lsn, int flags);
+bool	xfs_log_item_in_current_chkpt(struct xfs_log_item *lip);
+
 #endif
 
 
diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c
new file mode 100644
index 000000000000..bb17cc044bf3
--- /dev/null
+++ b/fs/xfs/xfs_log_cil.c
@@ -0,0 +1,725 @@
+/*
+ * Copyright (c) 2010 Red Hat, Inc. All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_types.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_trans_priv.h"
+#include "xfs_log_priv.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_dir2.h"
+#include "xfs_dmapi.h"
+#include "xfs_mount.h"
+#include "xfs_error.h"
+#include "xfs_alloc.h"
+
+/*
+ * Perform initial CIL structure initialisation. If the CIL is not
+ * enabled in this filesystem, ensure the log->l_cilp is null so
+ * we can check this conditional to determine if we are doing delayed
+ * logging or not.
+ */
+int
+xlog_cil_init(
+	struct log	*log)
+{
+	struct xfs_cil	*cil;
+	struct xfs_cil_ctx *ctx;
+
+	log->l_cilp = NULL;
+	if (!(log->l_mp->m_flags & XFS_MOUNT_DELAYLOG))
+		return 0;
+
+	cil = kmem_zalloc(sizeof(*cil), KM_SLEEP|KM_MAYFAIL);
+	if (!cil)
+		return ENOMEM;
+
+	ctx = kmem_zalloc(sizeof(*ctx), KM_SLEEP|KM_MAYFAIL);
+	if (!ctx) {
+		kmem_free(cil);
+		return ENOMEM;
+	}
+
+	INIT_LIST_HEAD(&cil->xc_cil);
+	INIT_LIST_HEAD(&cil->xc_committing);
+	spin_lock_init(&cil->xc_cil_lock);
+	init_rwsem(&cil->xc_ctx_lock);
+	sv_init(&cil->xc_commit_wait, SV_DEFAULT, "cilwait");
+
+	INIT_LIST_HEAD(&ctx->committing);
+	INIT_LIST_HEAD(&ctx->busy_extents);
+	ctx->sequence = 1;
+	ctx->cil = cil;
+	cil->xc_ctx = ctx;
+
+	cil->xc_log = log;
+	log->l_cilp = cil;
+	return 0;
+}
+
+void
+xlog_cil_destroy(
+	struct log	*log)
+{
+	if (!log->l_cilp)
+		return;
+
+	if (log->l_cilp->xc_ctx) {
+		if (log->l_cilp->xc_ctx->ticket)
+			xfs_log_ticket_put(log->l_cilp->xc_ctx->ticket);
+		kmem_free(log->l_cilp->xc_ctx);
+	}
+
+	ASSERT(list_empty(&log->l_cilp->xc_cil));
+	kmem_free(log->l_cilp);
+}
+
+/*
+ * Allocate a new ticket. Failing to get a new ticket makes it really hard to
+ * recover, so we don't allow failure here. Also, we allocate in a context that
+ * we don't want to be issuing transactions from, so we need to tell the
+ * allocation code this as well.
+ *
+ * We don't reserve any space for the ticket - we are going to steal whatever
+ * space we require from transactions as they commit. To ensure we reserve all
+ * the space required, we need to set the current reservation of the ticket to
+ * zero so that we know to steal the initial transaction overhead from the
+ * first transaction commit.
+ */
+static struct xlog_ticket *
+xlog_cil_ticket_alloc(
+	struct log	*log)
+{
+	struct xlog_ticket *tic;
+
+	tic = xlog_ticket_alloc(log, 0, 1, XFS_TRANSACTION, 0,
+				KM_SLEEP|KM_NOFS);
+	tic->t_trans_type = XFS_TRANS_CHECKPOINT;
+
+	/*
+	 * set the current reservation to zero so we know to steal the basic
+	 * transaction overhead reservation from the first transaction commit.
+	 */
+	tic->t_curr_res = 0;
+	return tic;
+}
+
+/*
+ * After the first stage of log recovery is done, we know where the head and
+ * tail of the log are. We need this log initialisation done before we can
+ * initialise the first CIL checkpoint context.
+ *
+ * Here we allocate a log ticket to track space usage during a CIL push.  This
+ * ticket is passed to xlog_write() directly so that we don't slowly leak log
+ * space by failing to account for space used by log headers and additional
+ * region headers for split regions.
+ */
+void
+xlog_cil_init_post_recovery(
+	struct log	*log)
+{
+	if (!log->l_cilp)
+		return;
+
+	log->l_cilp->xc_ctx->ticket = xlog_cil_ticket_alloc(log);
+	log->l_cilp->xc_ctx->sequence = 1;
+	log->l_cilp->xc_ctx->commit_lsn = xlog_assign_lsn(log->l_curr_cycle,
+								log->l_curr_block);
+}
+
+/*
+ * Insert the log item into the CIL and calculate the difference in space
+ * consumed by the item. Add the space to the checkpoint ticket and calculate
+ * if the change requires additional log metadata. If it does, take that space
+ * as well. Remove the amount of space we addded to the checkpoint ticket from
+ * the current transaction ticket so that the accounting works out correctly.
+ *
+ * If this is the first time the item is being placed into the CIL in this
+ * context, pin it so it can't be written to disk until the CIL is flushed to
+ * the iclog and the iclog written to disk.
+ */
+static void
+xlog_cil_insert(
+	struct log		*log,
+	struct xlog_ticket	*ticket,
+	struct xfs_log_item	*item,
+	struct xfs_log_vec	*lv)
+{
+	struct xfs_cil		*cil = log->l_cilp;
+	struct xfs_log_vec	*old = lv->lv_item->li_lv;
+	struct xfs_cil_ctx	*ctx = cil->xc_ctx;
+	int			len;
+	int			diff_iovecs;
+	int			iclog_space;
+
+	if (old) {
+		/* existing lv on log item, space used is a delta */
+		ASSERT(!list_empty(&item->li_cil));
+		ASSERT(old->lv_buf && old->lv_buf_len && old->lv_niovecs);
+
+		len = lv->lv_buf_len - old->lv_buf_len;
+		diff_iovecs = lv->lv_niovecs - old->lv_niovecs;
+		kmem_free(old->lv_buf);
+		kmem_free(old);
+	} else {
+		/* new lv, must pin the log item */
+		ASSERT(!lv->lv_item->li_lv);
+		ASSERT(list_empty(&item->li_cil));
+
+		len = lv->lv_buf_len;
+		diff_iovecs = lv->lv_niovecs;
+		IOP_PIN(lv->lv_item);
+
+	}
+	len += diff_iovecs * sizeof(xlog_op_header_t);
+
+	/* attach new log vector to log item */
+	lv->lv_item->li_lv = lv;
+
+	spin_lock(&cil->xc_cil_lock);
+	list_move_tail(&item->li_cil, &cil->xc_cil);
+	ctx->nvecs += diff_iovecs;
+
+	/*
+	 * If this is the first time the item is being committed to the CIL,
+	 * store the sequence number on the log item so we can tell
+	 * in future commits whether this is the first checkpoint the item is
+	 * being committed into.
+	 */
+	if (!item->li_seq)
+		item->li_seq = ctx->sequence;
+
+	/*
+	 * Now transfer enough transaction reservation to the context ticket
+	 * for the checkpoint. The context ticket is special - the unit
+	 * reservation has to grow as well as the current reservation as we
+	 * steal from tickets so we can correctly determine the space used
+	 * during the transaction commit.
+	 */
+	if (ctx->ticket->t_curr_res == 0) {
+		/* first commit in checkpoint, steal the header reservation */
+		ASSERT(ticket->t_curr_res >= ctx->ticket->t_unit_res + len);
+		ctx->ticket->t_curr_res = ctx->ticket->t_unit_res;
+		ticket->t_curr_res -= ctx->ticket->t_unit_res;
+	}
+
+	/* do we need space for more log record headers? */
+	iclog_space = log->l_iclog_size - log->l_iclog_hsize;
+	if (len > 0 && (ctx->space_used / iclog_space !=
+				(ctx->space_used + len) / iclog_space)) {
+		int hdrs;
+
+		hdrs = (len + iclog_space - 1) / iclog_space;
+		/* need to take into account split region headers, too */
+		hdrs *= log->l_iclog_hsize + sizeof(struct xlog_op_header);
+		ctx->ticket->t_unit_res += hdrs;
+		ctx->ticket->t_curr_res += hdrs;
+		ticket->t_curr_res -= hdrs;
+		ASSERT(ticket->t_curr_res >= len);
+	}
+	ticket->t_curr_res -= len;
+	ctx->space_used += len;
+
+	spin_unlock(&cil->xc_cil_lock);
+}
+
+/*
+ * Format log item into a flat buffers
+ *
+ * For delayed logging, we need to hold a formatted buffer containing all the
+ * changes on the log item. This enables us to relog the item in memory and
+ * write it out asynchronously without needing to relock the object that was
+ * modified at the time it gets written into the iclog.
+ *
+ * This function builds a vector for the changes in each log item in the
+ * transaction. It then works out the length of the buffer needed for each log
+ * item, allocates them and formats the vector for the item into the buffer.
+ * The buffer is then attached to the log item are then inserted into the
+ * Committed Item List for tracking until the next checkpoint is written out.
+ *
+ * We don't set up region headers during this process; we simply copy the
+ * regions into the flat buffer. We can do this because we still have to do a
+ * formatting step to write the regions into the iclog buffer.  Writing the
+ * ophdrs during the iclog write means that we can support splitting large
+ * regions across iclog boundares without needing a change in the format of the
+ * item/region encapsulation.
+ *
+ * Hence what we need to do now is change the rewrite the vector array to point
+ * to the copied region inside the buffer we just allocated. This allows us to
+ * format the regions into the iclog as though they are being formatted
+ * directly out of the objects themselves.
+ */
+static void
+xlog_cil_format_items(
+	struct log		*log,
+	struct xfs_log_vec	*log_vector,
+	struct xlog_ticket	*ticket,
+	xfs_lsn_t		*start_lsn)
+{
+	struct xfs_log_vec *lv;
+
+	if (start_lsn)
+		*start_lsn = log->l_cilp->xc_ctx->sequence;
+
+	ASSERT(log_vector);
+	for (lv = log_vector; lv; lv = lv->lv_next) {
+		void	*ptr;
+		int	index;
+		int	len = 0;
+
+		/* build the vector array and calculate it's length */
+		IOP_FORMAT(lv->lv_item, lv->lv_iovecp);
+		for (index = 0; index < lv->lv_niovecs; index++)
+			len += lv->lv_iovecp[index].i_len;
+
+		lv->lv_buf_len = len;
+		lv->lv_buf = kmem_zalloc(lv->lv_buf_len, KM_SLEEP|KM_NOFS);
+		ptr = lv->lv_buf;
+
+		for (index = 0; index < lv->lv_niovecs; index++) {
+			struct xfs_log_iovec *vec = &lv->lv_iovecp[index];
+
+			memcpy(ptr, vec->i_addr, vec->i_len);
+			vec->i_addr = ptr;
+			ptr += vec->i_len;
+		}
+		ASSERT(ptr == lv->lv_buf + lv->lv_buf_len);
+
+		xlog_cil_insert(log, ticket, lv->lv_item, lv);
+	}
+}
+
+static void
+xlog_cil_free_logvec(
+	struct xfs_log_vec	*log_vector)
+{
+	struct xfs_log_vec	*lv;
+
+	for (lv = log_vector; lv; ) {
+		struct xfs_log_vec *next = lv->lv_next;
+		kmem_free(lv->lv_buf);
+		kmem_free(lv);
+		lv = next;
+	}
+}
+
+/*
+ * Commit a transaction with the given vector to the Committed Item List.
+ *
+ * To do this, we need to format the item, pin it in memory if required and
+ * account for the space used by the transaction. Once we have done that we
+ * need to release the unused reservation for the transaction, attach the
+ * transaction to the checkpoint context so we carry the busy extents through
+ * to checkpoint completion, and then unlock all the items in the transaction.
+ *
+ * For more specific information about the order of operations in
+ * xfs_log_commit_cil() please refer to the comments in
+ * xfs_trans_commit_iclog().
+ *
+ * Called with the context lock already held in read mode to lock out
+ * background commit, returns without it held once background commits are
+ * allowed again.
+ */
+int
+xfs_log_commit_cil(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfs_log_vec	*log_vector,
+	xfs_lsn_t		*commit_lsn,
+	int			flags)
+{
+	struct log		*log = mp->m_log;
+	int			log_flags = 0;
+	int			push = 0;
+
+	if (flags & XFS_TRANS_RELEASE_LOG_RES)
+		log_flags = XFS_LOG_REL_PERM_RESERV;
+
+	if (XLOG_FORCED_SHUTDOWN(log)) {
+		xlog_cil_free_logvec(log_vector);
+		return XFS_ERROR(EIO);
+	}
+
+	/* lock out background commit */
+	down_read(&log->l_cilp->xc_ctx_lock);
+	xlog_cil_format_items(log, log_vector, tp->t_ticket, commit_lsn);
+
+	/* check we didn't blow the reservation */
+	if (tp->t_ticket->t_curr_res < 0)
+		xlog_print_tic_res(log->l_mp, tp->t_ticket);
+
+	/* attach the transaction to the CIL if it has any busy extents */
+	if (!list_empty(&tp->t_busy)) {
+		spin_lock(&log->l_cilp->xc_cil_lock);
+		list_splice_init(&tp->t_busy,
+					&log->l_cilp->xc_ctx->busy_extents);
+		spin_unlock(&log->l_cilp->xc_cil_lock);
+	}
+
+	tp->t_commit_lsn = *commit_lsn;
+	xfs_log_done(mp, tp->t_ticket, NULL, log_flags);
+	xfs_trans_unreserve_and_mod_sb(tp);
+
+	/* check for background commit before unlock */
+	if (log->l_cilp->xc_ctx->space_used > XLOG_CIL_SPACE_LIMIT(log))
+		push = 1;
+	up_read(&log->l_cilp->xc_ctx_lock);
+
+	/*
+	 * We need to push CIL every so often so we don't cache more than we
+	 * can fit in the log. The limit really is that a checkpoint can't be
+	 * more than half the log (the current checkpoint is not allowed to
+	 * overwrite the previous checkpoint), but commit latency and memory
+	 * usage limit this to a smaller size in most cases.
+	 */
+	if (push)
+		xlog_cil_push(log, 0);
+	return 0;
+}
+
+/*
+ * Mark all items committed and clear busy extents. We free the log vector
+ * chains in a separate pass so that we unpin the log items as quickly as
+ * possible.
+ */
+static void
+xlog_cil_committed(
+	void	*args,
+	int	abort)
+{
+	struct xfs_cil_ctx	*ctx = args;
+	struct xfs_log_vec	*lv;
+	int			abortflag = abort ? XFS_LI_ABORTED : 0;
+	struct xfs_busy_extent	*busyp, *n;
+
+	/* unpin all the log items */
+	for (lv = ctx->lv_chain; lv; lv = lv->lv_next ) {
+		xfs_trans_item_committed(lv->lv_item, ctx->start_lsn,
+							abortflag);
+	}
+
+	list_for_each_entry_safe(busyp, n, &ctx->busy_extents, list)
+		xfs_alloc_busy_clear(ctx->cil->xc_log->l_mp, busyp);
+
+	spin_lock(&ctx->cil->xc_cil_lock);
+	list_del(&ctx->committing);
+	spin_unlock(&ctx->cil->xc_cil_lock);
+
+	xlog_cil_free_logvec(ctx->lv_chain);
+	kmem_free(ctx);
+}
+
+/*
+ * Push the Committed Item List to the log. If the push_now flag is not set,
+ * then it is a background flush and so we can chose to ignore it.
+ */
+int
+xlog_cil_push(
+	struct log		*log,
+	int			push_now)
+{
+	struct xfs_cil		*cil = log->l_cilp;
+	struct xfs_log_vec	*lv;
+	struct xfs_cil_ctx	*ctx;
+	struct xfs_cil_ctx	*new_ctx;
+	struct xlog_in_core	*commit_iclog;
+	struct xlog_ticket	*tic;
+	int			num_lv;
+	int			num_iovecs;
+	int			len;
+	int			error = 0;
+	struct xfs_trans_header thdr;
+	struct xfs_log_iovec	lhdr;
+	struct xfs_log_vec	lvhdr = { NULL };
+	xfs_lsn_t		commit_lsn;
+
+	if (!cil)
+		return 0;
+
+	new_ctx = kmem_zalloc(sizeof(*new_ctx), KM_SLEEP|KM_NOFS);
+	new_ctx->ticket = xlog_cil_ticket_alloc(log);
+
+	/* lock out transaction commit, but don't block on background push */
+	if (!down_write_trylock(&cil->xc_ctx_lock)) {
+		if (!push_now)
+			goto out_free_ticket;
+		down_write(&cil->xc_ctx_lock);
+	}
+	ctx = cil->xc_ctx;
+
+	/* check if we've anything to push */
+	if (list_empty(&cil->xc_cil))
+		goto out_skip;
+
+	/* check for spurious background flush */
+	if (!push_now && cil->xc_ctx->space_used < XLOG_CIL_SPACE_LIMIT(log))
+		goto out_skip;
+
+	/*
+	 * pull all the log vectors off the items in the CIL, and
+	 * remove the items from the CIL. We don't need the CIL lock
+	 * here because it's only needed on the transaction commit
+	 * side which is currently locked out by the flush lock.
+	 */
+	lv = NULL;
+	num_lv = 0;
+	num_iovecs = 0;
+	len = 0;
+	while (!list_empty(&cil->xc_cil)) {
+		struct xfs_log_item	*item;
+		int			i;
+
+		item = list_first_entry(&cil->xc_cil,
+					struct xfs_log_item, li_cil);
+		list_del_init(&item->li_cil);
+		if (!ctx->lv_chain)
+			ctx->lv_chain = item->li_lv;
+		else
+			lv->lv_next = item->li_lv;
+		lv = item->li_lv;
+		item->li_lv = NULL;
+
+		num_lv++;
+		num_iovecs += lv->lv_niovecs;
+		for (i = 0; i < lv->lv_niovecs; i++)
+			len += lv->lv_iovecp[i].i_len;
+	}
+
+	/*
+	 * initialise the new context and attach it to the CIL. Then attach
+	 * the current context to the CIL committing lsit so it can be found
+	 * during log forces to extract the commit lsn of the sequence that
+	 * needs to be forced.
+	 */
+	INIT_LIST_HEAD(&new_ctx->committing);
+	INIT_LIST_HEAD(&new_ctx->busy_extents);
+	new_ctx->sequence = ctx->sequence + 1;
+	new_ctx->cil = cil;
+	cil->xc_ctx = new_ctx;
+
+	/*
+	 * The switch is now done, so we can drop the context lock and move out
+	 * of a shared context. We can't just go straight to the commit record,
+	 * though - we need to synchronise with previous and future commits so
+	 * that the commit records are correctly ordered in the log to ensure
+	 * that we process items during log IO completion in the correct order.
+	 *
+	 * For example, if we get an EFI in one checkpoint and the EFD in the
+	 * next (e.g. due to log forces), we do not want the checkpoint with
+	 * the EFD to be committed before the checkpoint with the EFI.  Hence
+	 * we must strictly order the commit records of the checkpoints so
+	 * that: a) the checkpoint callbacks are attached to the iclogs in the
+	 * correct order; and b) the checkpoints are replayed in correct order
+	 * in log recovery.
+	 *
+	 * Hence we need to add this context to the committing context list so
+	 * that higher sequences will wait for us to write out a commit record
+	 * before they do.
+	 */
+	spin_lock(&cil->xc_cil_lock);
+	list_add(&ctx->committing, &cil->xc_committing);
+	spin_unlock(&cil->xc_cil_lock);
+	up_write(&cil->xc_ctx_lock);
+
+	/*
+	 * Build a checkpoint transaction header and write it to the log to
+	 * begin the transaction. We need to account for the space used by the
+	 * transaction header here as it is not accounted for in xlog_write().
+	 *
+	 * The LSN we need to pass to the log items on transaction commit is
+	 * the LSN reported by the first log vector write. If we use the commit
+	 * record lsn then we can move the tail beyond the grant write head.
+	 */
+	tic = ctx->ticket;
+	thdr.th_magic = XFS_TRANS_HEADER_MAGIC;
+	thdr.th_type = XFS_TRANS_CHECKPOINT;
+	thdr.th_tid = tic->t_tid;
+	thdr.th_num_items = num_iovecs;
+	lhdr.i_addr = (xfs_caddr_t)&thdr;
+	lhdr.i_len = sizeof(xfs_trans_header_t);
+	lhdr.i_type = XLOG_REG_TYPE_TRANSHDR;
+	tic->t_curr_res -= lhdr.i_len + sizeof(xlog_op_header_t);
+
+	lvhdr.lv_niovecs = 1;
+	lvhdr.lv_iovecp = &lhdr;
+	lvhdr.lv_next = ctx->lv_chain;
+
+	error = xlog_write(log, &lvhdr, tic, &ctx->start_lsn, NULL, 0);
+	if (error)
+		goto out_abort;
+
+	/*
+	 * now that we've written the checkpoint into the log, strictly
+	 * order the commit records so replay will get them in the right order.
+	 */
+restart:
+	spin_lock(&cil->xc_cil_lock);
+	list_for_each_entry(new_ctx, &cil->xc_committing, committing) {
+		/*
+		 * Higher sequences will wait for this one so skip them.
+		 * Don't wait for own own sequence, either.
+		 */
+		if (new_ctx->sequence >= ctx->sequence)
+			continue;
+		if (!new_ctx->commit_lsn) {
+			/*
+			 * It is still being pushed! Wait for the push to
+			 * complete, then start again from the beginning.
+			 */
+			sv_wait(&cil->xc_commit_wait, 0, &cil->xc_cil_lock, 0);
+			goto restart;
+		}
+	}
+	spin_unlock(&cil->xc_cil_lock);
+
+	commit_lsn = xfs_log_done(log->l_mp, tic, &commit_iclog, 0);
+	if (error || commit_lsn == -1)
+		goto out_abort;
+
+	/* attach all the transactions w/ busy extents to iclog */
+	ctx->log_cb.cb_func = xlog_cil_committed;
+	ctx->log_cb.cb_arg = ctx;
+	error = xfs_log_notify(log->l_mp, commit_iclog, &ctx->log_cb);
+	if (error)
+		goto out_abort;
+
+	/*
+	 * now the checkpoint commit is complete and we've attached the
+	 * callbacks to the iclog we can assign the commit LSN to the context
+	 * and wake up anyone who is waiting for the commit to complete.
+	 */
+	spin_lock(&cil->xc_cil_lock);
+	ctx->commit_lsn = commit_lsn;
+	sv_broadcast(&cil->xc_commit_wait);
+	spin_unlock(&cil->xc_cil_lock);
+
+	/* release the hounds! */
+	return xfs_log_release_iclog(log->l_mp, commit_iclog);
+
+out_skip:
+	up_write(&cil->xc_ctx_lock);
+out_free_ticket:
+	xfs_log_ticket_put(new_ctx->ticket);
+	kmem_free(new_ctx);
+	return 0;
+
+out_abort:
+	xlog_cil_committed(ctx, XFS_LI_ABORTED);
+	return XFS_ERROR(EIO);
+}
+
+/*
+ * Conditionally push the CIL based on the sequence passed in.
+ *
+ * We only need to push if we haven't already pushed the sequence
+ * number given. Hence the only time we will trigger a push here is
+ * if the push sequence is the same as the current context.
+ *
+ * We return the current commit lsn to allow the callers to determine if a
+ * iclog flush is necessary following this call.
+ *
+ * XXX: Initially, just push the CIL unconditionally and return whatever
+ * commit lsn is there. It'll be empty, so this is broken for now.
+ */
+xfs_lsn_t
+xlog_cil_push_lsn(
+	struct log	*log,
+	xfs_lsn_t	push_seq)
+{
+	struct xfs_cil		*cil = log->l_cilp;
+	struct xfs_cil_ctx	*ctx;
+	xfs_lsn_t		commit_lsn = NULLCOMMITLSN;
+
+restart:
+	down_write(&cil->xc_ctx_lock);
+	ASSERT(push_seq <= cil->xc_ctx->sequence);
+
+	/* check to see if we need to force out the current context */
+	if (push_seq == cil->xc_ctx->sequence) {
+		up_write(&cil->xc_ctx_lock);
+		xlog_cil_push(log, 1);
+		goto restart;
+	}
+
+	/*
+	 * See if we can find a previous sequence still committing.
+	 * We can drop the flush lock as soon as we have the cil lock
+	 * because we are now only comparing contexts protected by
+	 * the cil lock.
+	 *
+	 * We need to wait for all previous sequence commits to complete
+	 * before allowing the force of push_seq to go ahead. Hence block
+	 * on commits for those as well.
+	 */
+	spin_lock(&cil->xc_cil_lock);
+	up_write(&cil->xc_ctx_lock);
+	list_for_each_entry(ctx, &cil->xc_committing, committing) {
+		if (ctx->sequence > push_seq)
+			continue;
+		if (!ctx->commit_lsn) {
+			/*
+			 * It is still being pushed! Wait for the push to
+			 * complete, then start again from the beginning.
+			 */
+			sv_wait(&cil->xc_commit_wait, 0, &cil->xc_cil_lock, 0);
+			goto restart;
+		}
+		if (ctx->sequence != push_seq)
+			continue;
+		/* found it! */
+		commit_lsn = ctx->commit_lsn;
+	}
+	spin_unlock(&cil->xc_cil_lock);
+	return commit_lsn;
+}
+
+/*
+ * Check if the current log item was first committed in this sequence.
+ * We can't rely on just the log item being in the CIL, we have to check
+ * the recorded commit sequence number.
+ *
+ * Note: for this to be used in a non-racy manner, it has to be called with
+ * CIL flushing locked out. As a result, it should only be used during the
+ * transaction commit process when deciding what to format into the item.
+ */
+bool
+xfs_log_item_in_current_chkpt(
+	struct xfs_log_item *lip)
+{
+	struct xfs_cil_ctx *ctx;
+
+	if (!(lip->li_mountp->m_flags & XFS_MOUNT_DELAYLOG))
+		return false;
+	if (list_empty(&lip->li_cil))
+		return false;
+
+	ctx = lip->li_mountp->m_log->l_cilp->xc_ctx;
+
+	/*
+	 * li_seq is written on the first commit of a log item to record the
+	 * first checkpoint it is written to. Hence if it is different to the
+	 * current sequence, we're in a new checkpoint.
+	 */
+	if (XFS_LSN_CMP(lip->li_seq, ctx->sequence) != 0)
+		return false;
+	return true;
+}
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index 9cf695154451..8c072618965c 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -152,8 +152,6 @@ static inline uint xlog_get_client_id(__be32 i)
 #define	XLOG_RECOVERY_NEEDED	0x4	/* log was recovered */
 #define XLOG_IO_ERROR		0x8	/* log hit an I/O error, and being
 					   shutdown */
-typedef __uint32_t xlog_tid_t;
-
 
 #ifdef __KERNEL__
 /*
@@ -379,6 +377,99 @@ typedef struct xlog_in_core {
 } xlog_in_core_t;
 
 /*
+ * The CIL context is used to aggregate per-transaction details as well be
+ * passed to the iclog for checkpoint post-commit processing.  After being
+ * passed to the iclog, another context needs to be allocated for tracking the
+ * next set of transactions to be aggregated into a checkpoint.
+ */
+struct xfs_cil;
+
+struct xfs_cil_ctx {
+	struct xfs_cil		*cil;
+	xfs_lsn_t		sequence;	/* chkpt sequence # */
+	xfs_lsn_t		start_lsn;	/* first LSN of chkpt commit */
+	xfs_lsn_t		commit_lsn;	/* chkpt commit record lsn */
+	struct xlog_ticket	*ticket;	/* chkpt ticket */
+	int			nvecs;		/* number of regions */
+	int			space_used;	/* aggregate size of regions */
+	struct list_head	busy_extents;	/* busy extents in chkpt */
+	struct xfs_log_vec	*lv_chain;	/* logvecs being pushed */
+	xfs_log_callback_t	log_cb;		/* completion callback hook. */
+	struct list_head	committing;	/* ctx committing list */
+};
+
+/*
+ * Committed Item List structure
+ *
+ * This structure is used to track log items that have been committed but not
+ * yet written into the log. It is used only when the delayed logging mount
+ * option is enabled.
+ *
+ * This structure tracks the list of committing checkpoint contexts so
+ * we can avoid the problem of having to hold out new transactions during a
+ * flush until we have a the commit record LSN of the checkpoint. We can
+ * traverse the list of committing contexts in xlog_cil_push_lsn() to find a
+ * sequence match and extract the commit LSN directly from there. If the
+ * checkpoint is still in the process of committing, we can block waiting for
+ * the commit LSN to be determined as well. This should make synchronous
+ * operations almost as efficient as the old logging methods.
+ */
+struct xfs_cil {
+	struct log		*xc_log;
+	struct list_head	xc_cil;
+	spinlock_t		xc_cil_lock;
+	struct xfs_cil_ctx	*xc_ctx;
+	struct rw_semaphore	xc_ctx_lock;
+	struct list_head	xc_committing;
+	sv_t			xc_commit_wait;
+};
+
+/*
+ * The amount of log space we should the CIL to aggregate is difficult to size.
+ * Whatever we chose we have to make we can get a reservation for the log space
+ * effectively, that it is large enough to capture sufficient relogging to
+ * reduce log buffer IO significantly, but it is not too large for the log or
+ * induces too much latency when writing out through the iclogs. We track both
+ * space consumed and the number of vectors in the checkpoint context, so we
+ * need to decide which to use for limiting.
+ *
+ * Every log buffer we write out during a push needs a header reserved, which
+ * is at least one sector and more for v2 logs. Hence we need a reservation of
+ * at least 512 bytes per 32k of log space just for the LR headers. That means
+ * 16KB of reservation per megabyte of delayed logging space we will consume,
+ * plus various headers.  The number of headers will vary based on the num of
+ * io vectors, so limiting on a specific number of vectors is going to result
+ * in transactions of varying size. IOWs, it is more consistent to track and
+ * limit space consumed in the log rather than by the number of objects being
+ * logged in order to prevent checkpoint ticket overruns.
+ *
+ * Further, use of static reservations through the log grant mechanism is
+ * problematic. It introduces a lot of complexity (e.g. reserve grant vs write
+ * grant) and a significant deadlock potential because regranting write space
+ * can block on log pushes. Hence if we have to regrant log space during a log
+ * push, we can deadlock.
+ *
+ * However, we can avoid this by use of a dynamic "reservation stealing"
+ * technique during transaction commit whereby unused reservation space in the
+ * transaction ticket is transferred to the CIL ctx commit ticket to cover the
+ * space needed by the checkpoint transaction. This means that we never need to
+ * specifically reserve space for the CIL checkpoint transaction, nor do we
+ * need to regrant space once the checkpoint completes. This also means the
+ * checkpoint transaction ticket is specific to the checkpoint context, rather
+ * than the CIL itself.
+ *
+ * With dynamic reservations, we can basically make up arbitrary limits for the
+ * checkpoint size so long as they don't violate any other size rules.  Hence
+ * the initial maximum size for the checkpoint transaction will be set to a
+ * quarter of the log or 8MB, which ever is smaller. 8MB is an arbitrary limit
+ * right now based on the latency of writing out a large amount of data through
+ * the circular iclog buffers.
+ */
+
+#define XLOG_CIL_SPACE_LIMIT(log)	\
+	(min((log->l_logsize >> 2), (8 * 1024 * 1024)))
+
+/*
  * The reservation head lsn is not made up of a cycle number and block number.
  * Instead, it uses a cycle number and byte number.  Logs don't expect to
  * overflow 31 bits worth of byte offset, so using a byte number will mean
@@ -388,6 +479,7 @@ typedef struct log {
 	/* The following fields don't need locking */
 	struct xfs_mount	*l_mp;	        /* mount point */
 	struct xfs_ail		*l_ailp;	/* AIL log is working with */
+	struct xfs_cil		*l_cilp;	/* CIL log is working with */
 	struct xfs_buf		*l_xbuf;        /* extra buffer for log
 						 * wrapping */
 	struct xfs_buftarg	*l_targ;        /* buftarg of log */
@@ -438,14 +530,17 @@ typedef struct log {
 
 #define XLOG_FORCED_SHUTDOWN(log)	((log)->l_flags & XLOG_IO_ERROR)
 
-
 /* common routines */
 extern xfs_lsn_t xlog_assign_tail_lsn(struct xfs_mount *mp);
 extern int	 xlog_recover(xlog_t *log);
 extern int	 xlog_recover_finish(xlog_t *log);
 extern void	 xlog_pack_data(xlog_t *log, xlog_in_core_t *iclog, int);
 
-extern kmem_zone_t	*xfs_log_ticket_zone;
+extern kmem_zone_t *xfs_log_ticket_zone;
+struct xlog_ticket *xlog_ticket_alloc(struct log *log, int unit_bytes,
+				int count, char client, uint xflags,
+				int alloc_flags);
+
 
 static inline void
 xlog_write_adv_cnt(void **ptr, int *len, int *off, size_t bytes)
@@ -455,6 +550,21 @@ xlog_write_adv_cnt(void **ptr, int *len, int *off, size_t bytes)
 	*off += bytes;
 }
 
+void	xlog_print_tic_res(struct xfs_mount *mp, struct xlog_ticket *ticket);
+int	xlog_write(struct log *log, struct xfs_log_vec *log_vector,
+				struct xlog_ticket *tic, xfs_lsn_t *start_lsn,
+				xlog_in_core_t **commit_iclog, uint flags);
+
+/*
+ * Committed Item List interfaces
+ */
+int	xlog_cil_init(struct log *log);
+void	xlog_cil_init_post_recovery(struct log *log);
+void	xlog_cil_destroy(struct log *log);
+
+int	xlog_cil_push(struct log *log, int push_now);
+xfs_lsn_t xlog_cil_push_lsn(struct log *log, xfs_lsn_t push_sequence);
+
 /*
  * Unmount record type is used as a pseudo transaction type for the ticket.
  * It's value must be outside the range of XFS_TRANS_* values.
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 0de08e366315..14a69aec2c0b 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -1576,7 +1576,7 @@ xlog_recover_reorder_trans(
 
 		switch (ITEM_TYPE(item)) {
 		case XFS_LI_BUF:
-			if (!(buf_f->blf_flags & XFS_BLI_CANCEL)) {
+			if (!(buf_f->blf_flags & XFS_BLF_CANCEL)) {
 				trace_xfs_log_recover_item_reorder_head(log,
 							trans, item, pass);
 				list_move(&item->ri_list, &trans->r_itemq);
@@ -1638,7 +1638,7 @@ xlog_recover_do_buffer_pass1(
 	/*
 	 * If this isn't a cancel buffer item, then just return.
 	 */
-	if (!(flags & XFS_BLI_CANCEL)) {
+	if (!(flags & XFS_BLF_CANCEL)) {
 		trace_xfs_log_recover_buf_not_cancel(log, buf_f);
 		return;
 	}
@@ -1696,7 +1696,7 @@ xlog_recover_do_buffer_pass1(
  * Check to see whether the buffer being recovered has a corresponding
  * entry in the buffer cancel record table.  If it does then return 1
  * so that it will be cancelled, otherwise return 0.  If the buffer is
- * actually a buffer cancel item (XFS_BLI_CANCEL is set), then decrement
+ * actually a buffer cancel item (XFS_BLF_CANCEL is set), then decrement
  * the refcount on the entry in the table and remove it from the table
  * if this is the last reference.
  *
@@ -1721,7 +1721,7 @@ xlog_check_buffer_cancelled(
 		 * There is nothing in the table built in pass one,
 		 * so this buffer must not be cancelled.
 		 */
-		ASSERT(!(flags & XFS_BLI_CANCEL));
+		ASSERT(!(flags & XFS_BLF_CANCEL));
 		return 0;
 	}
 
@@ -1733,7 +1733,7 @@ xlog_check_buffer_cancelled(
 		 * There is no corresponding entry in the table built
 		 * in pass one, so this buffer has not been cancelled.
 		 */
-		ASSERT(!(flags & XFS_BLI_CANCEL));
+		ASSERT(!(flags & XFS_BLF_CANCEL));
 		return 0;
 	}
 
@@ -1752,7 +1752,7 @@ xlog_check_buffer_cancelled(
 			 * one in the table and remove it if this is the
 			 * last reference.
 			 */
-			if (flags & XFS_BLI_CANCEL) {
+			if (flags & XFS_BLF_CANCEL) {
 				bcp->bc_refcount--;
 				if (bcp->bc_refcount == 0) {
 					if (prevp == NULL) {
@@ -1772,7 +1772,7 @@ xlog_check_buffer_cancelled(
 	 * We didn't find a corresponding entry in the table, so
 	 * return 0 so that the buffer is NOT cancelled.
 	 */
-	ASSERT(!(flags & XFS_BLI_CANCEL));
+	ASSERT(!(flags & XFS_BLF_CANCEL));
 	return 0;
 }
 
@@ -1874,8 +1874,8 @@ xlog_recover_do_inode_buffer(
 			nbits = xfs_contig_bits(data_map, map_size,
 							 bit);
 			ASSERT(nbits > 0);
-			reg_buf_offset = bit << XFS_BLI_SHIFT;
-			reg_buf_bytes = nbits << XFS_BLI_SHIFT;
+			reg_buf_offset = bit << XFS_BLF_SHIFT;
+			reg_buf_bytes = nbits << XFS_BLF_SHIFT;
 			item_index++;
 		}
 
@@ -1889,7 +1889,7 @@ xlog_recover_do_inode_buffer(
 		}
 
 		ASSERT(item->ri_buf[item_index].i_addr != NULL);
-		ASSERT((item->ri_buf[item_index].i_len % XFS_BLI_CHUNK) == 0);
+		ASSERT((item->ri_buf[item_index].i_len % XFS_BLF_CHUNK) == 0);
 		ASSERT((reg_buf_offset + reg_buf_bytes) <= XFS_BUF_COUNT(bp));
 
 		/*
@@ -1955,9 +1955,9 @@ xlog_recover_do_reg_buffer(
 		nbits = xfs_contig_bits(data_map, map_size, bit);
 		ASSERT(nbits > 0);
 		ASSERT(item->ri_buf[i].i_addr != NULL);
-		ASSERT(item->ri_buf[i].i_len % XFS_BLI_CHUNK == 0);
+		ASSERT(item->ri_buf[i].i_len % XFS_BLF_CHUNK == 0);
 		ASSERT(XFS_BUF_COUNT(bp) >=
-		       ((uint)bit << XFS_BLI_SHIFT)+(nbits<<XFS_BLI_SHIFT));
+		       ((uint)bit << XFS_BLF_SHIFT)+(nbits<<XFS_BLF_SHIFT));
 
 		/*
 		 * Do a sanity check if this is a dquot buffer. Just checking
@@ -1966,7 +1966,7 @@ xlog_recover_do_reg_buffer(
 		 */
 		error = 0;
 		if (buf_f->blf_flags &
-		   (XFS_BLI_UDQUOT_BUF|XFS_BLI_PDQUOT_BUF|XFS_BLI_GDQUOT_BUF)) {
+		   (XFS_BLF_UDQUOT_BUF|XFS_BLF_PDQUOT_BUF|XFS_BLF_GDQUOT_BUF)) {
 			if (item->ri_buf[i].i_addr == NULL) {
 				cmn_err(CE_ALERT,
 					"XFS: NULL dquot in %s.", __func__);
@@ -1987,9 +1987,9 @@ xlog_recover_do_reg_buffer(
 		}
 
 		memcpy(xfs_buf_offset(bp,
-			(uint)bit << XFS_BLI_SHIFT),	/* dest */
+			(uint)bit << XFS_BLF_SHIFT),	/* dest */
 			item->ri_buf[i].i_addr,		/* source */
-			nbits<<XFS_BLI_SHIFT);		/* length */
+			nbits<<XFS_BLF_SHIFT);		/* length */
  next:
 		i++;
 		bit += nbits;
@@ -2148,11 +2148,11 @@ xlog_recover_do_dquot_buffer(
 	}
 
 	type = 0;
-	if (buf_f->blf_flags & XFS_BLI_UDQUOT_BUF)
+	if (buf_f->blf_flags & XFS_BLF_UDQUOT_BUF)
 		type |= XFS_DQ_USER;
-	if (buf_f->blf_flags & XFS_BLI_PDQUOT_BUF)
+	if (buf_f->blf_flags & XFS_BLF_PDQUOT_BUF)
 		type |= XFS_DQ_PROJ;
-	if (buf_f->blf_flags & XFS_BLI_GDQUOT_BUF)
+	if (buf_f->blf_flags & XFS_BLF_GDQUOT_BUF)
 		type |= XFS_DQ_GROUP;
 	/*
 	 * This type of quotas was turned off, so ignore this buffer
@@ -2173,7 +2173,7 @@ xlog_recover_do_dquot_buffer(
  * here which overlaps that may be stale.
  *
  * When meta-data buffers are freed at run time we log a buffer item
- * with the XFS_BLI_CANCEL bit set to indicate that previous copies
+ * with the XFS_BLF_CANCEL bit set to indicate that previous copies
  * of the buffer in the log should not be replayed at recovery time.
  * This is so that if the blocks covered by the buffer are reused for
  * file data before we crash we don't end up replaying old, freed
@@ -2207,7 +2207,7 @@ xlog_recover_do_buffer_trans(
 	if (pass == XLOG_RECOVER_PASS1) {
 		/*
 		 * In this pass we're only looking for buf items
-		 * with the XFS_BLI_CANCEL bit set.
+		 * with the XFS_BLF_CANCEL bit set.
 		 */
 		xlog_recover_do_buffer_pass1(log, buf_f);
 		return 0;
@@ -2244,7 +2244,7 @@ xlog_recover_do_buffer_trans(
 
 	mp = log->l_mp;
 	buf_flags = XBF_LOCK;
-	if (!(flags & XFS_BLI_INODE_BUF))
+	if (!(flags & XFS_BLF_INODE_BUF))
 		buf_flags |= XBF_MAPPED;
 
 	bp = xfs_buf_read(mp->m_ddev_targp, blkno, len, buf_flags);
@@ -2257,10 +2257,10 @@ xlog_recover_do_buffer_trans(
 	}
 
 	error = 0;
-	if (flags & XFS_BLI_INODE_BUF) {
+	if (flags & XFS_BLF_INODE_BUF) {
 		error = xlog_recover_do_inode_buffer(mp, item, bp, buf_f);
 	} else if (flags &
-		  (XFS_BLI_UDQUOT_BUF|XFS_BLI_PDQUOT_BUF|XFS_BLI_GDQUOT_BUF)) {
+		  (XFS_BLF_UDQUOT_BUF|XFS_BLF_PDQUOT_BUF|XFS_BLF_GDQUOT_BUF)) {
 		xlog_recover_do_dquot_buffer(mp, log, item, bp, buf_f);
 	} else {
 		xlog_recover_do_reg_buffer(mp, item, bp, buf_f);
diff --git a/fs/xfs/xfs_log_recover.h b/fs/xfs/xfs_log_recover.h
index 75d749207258..1c55ccbb379d 100644
--- a/fs/xfs/xfs_log_recover.h
+++ b/fs/xfs/xfs_log_recover.h
@@ -28,7 +28,7 @@
 #define XLOG_RHASH(tid)	\
 	((((__uint32_t)tid)>>XLOG_RHASH_SHIFT) & (XLOG_RHASH_SIZE-1))
 
-#define XLOG_MAX_REGIONS_IN_ITEM   (XFS_MAX_BLOCKSIZE / XFS_BLI_CHUNK / 2 + 1)
+#define XLOG_MAX_REGIONS_IN_ITEM   (XFS_MAX_BLOCKSIZE / XFS_BLF_CHUNK / 2 + 1)
 
 
 /*
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 9ff48a16a7ee..1d2c7eed4eda 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -268,6 +268,7 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_WSYNC		(1ULL << 0)	/* for nfs - all metadata ops
 						   must be synchronous except
 						   for space allocations */
+#define XFS_MOUNT_DELAYLOG	(1ULL << 1)	/* delayed logging is enabled */
 #define XFS_MOUNT_DMAPI		(1ULL << 2)	/* dmapi is enabled */
 #define XFS_MOUNT_WAS_CLEAN	(1ULL << 3)
 #define XFS_MOUNT_FS_SHUTDOWN	(1ULL << 4)	/* atomic stop of all filesystem
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index be578ecb4af2..ce558efa2ea0 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -44,6 +44,7 @@
 #include "xfs_trans_priv.h"
 #include "xfs_trans_space.h"
 #include "xfs_inode_item.h"
+#include "xfs_trace.h"
 
 kmem_zone_t	*xfs_trans_zone;
 
@@ -243,9 +244,8 @@ _xfs_trans_alloc(
 	tp->t_type = type;
 	tp->t_mountp = mp;
 	tp->t_items_free = XFS_LIC_NUM_SLOTS;
-	tp->t_busy_free = XFS_LBC_NUM_SLOTS;
 	xfs_lic_init(&(tp->t_items));
-	XFS_LBC_INIT(&(tp->t_busy));
+	INIT_LIST_HEAD(&tp->t_busy);
 	return tp;
 }
 
@@ -255,8 +255,13 @@ _xfs_trans_alloc(
  */
 STATIC void
 xfs_trans_free(
-	xfs_trans_t	*tp)
+	struct xfs_trans	*tp)
 {
+	struct xfs_busy_extent	*busyp, *n;
+
+	list_for_each_entry_safe(busyp, n, &tp->t_busy, list)
+		xfs_alloc_busy_clear(tp->t_mountp, busyp);
+
 	atomic_dec(&tp->t_mountp->m_active_trans);
 	xfs_trans_free_dqinfo(tp);
 	kmem_zone_free(xfs_trans_zone, tp);
@@ -285,9 +290,8 @@ xfs_trans_dup(
 	ntp->t_type = tp->t_type;
 	ntp->t_mountp = tp->t_mountp;
 	ntp->t_items_free = XFS_LIC_NUM_SLOTS;
-	ntp->t_busy_free = XFS_LBC_NUM_SLOTS;
 	xfs_lic_init(&(ntp->t_items));
-	XFS_LBC_INIT(&(ntp->t_busy));
+	INIT_LIST_HEAD(&ntp->t_busy);
 
 	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
 	ASSERT(tp->t_ticket != NULL);
@@ -423,7 +427,6 @@ undo_blocks:
 	return error;
 }
 
-
 /*
  * Record the indicated change to the given field for application
  * to the file system's superblock when the transaction commits.
@@ -652,7 +655,7 @@ xfs_trans_apply_sb_deltas(
  * XFS_TRANS_SB_DIRTY will not be set when the transaction is updated but we
  * still need to update the incore superblock with the changes.
  */
-STATIC void
+void
 xfs_trans_unreserve_and_mod_sb(
 	xfs_trans_t	*tp)
 {
@@ -880,7 +883,7 @@ xfs_trans_fill_vecs(
  * they could be immediately flushed and we'd have to race with the flusher
  * trying to pull the item from the AIL as we add it.
  */
-static void
+void
 xfs_trans_item_committed(
 	struct xfs_log_item	*lip,
 	xfs_lsn_t		commit_lsn,
@@ -930,26 +933,6 @@ xfs_trans_item_committed(
 	IOP_UNPIN(lip);
 }
 
-/* Clear all the per-AG busy list items listed in this transaction */
-static void
-xfs_trans_clear_busy_extents(
-	struct xfs_trans	*tp)
-{
-	xfs_log_busy_chunk_t	*lbcp;
-	xfs_log_busy_slot_t	*lbsp;
-	int			i;
-
-	for (lbcp = &tp->t_busy; lbcp != NULL; lbcp = lbcp->lbc_next) {
-		i = 0;
-		for (lbsp = lbcp->lbc_busy; i < lbcp->lbc_unused; i++, lbsp++) {
-			if (XFS_LBC_ISFREE(lbcp, i))
-				continue;
-			xfs_alloc_clear_busy(tp, lbsp->lbc_ag, lbsp->lbc_idx);
-		}
-	}
-	xfs_trans_free_busy(tp);
-}
-
 /*
  * This is typically called by the LM when a transaction has been fully
  * committed to disk.  It needs to unpin the items which have
@@ -984,7 +967,6 @@ xfs_trans_committed(
 		kmem_free(licp);
 	}
 
-	xfs_trans_clear_busy_extents(tp);
 	xfs_trans_free(tp);
 }
 
@@ -1012,8 +994,7 @@ xfs_trans_uncommit(
 	xfs_trans_unreserve_and_mod_sb(tp);
 	xfs_trans_unreserve_and_mod_dquots(tp);
 
-	xfs_trans_free_items(tp, flags);
-	xfs_trans_free_busy(tp);
+	xfs_trans_free_items(tp, NULLCOMMITLSN, flags);
 	xfs_trans_free(tp);
 }
 
@@ -1075,6 +1056,8 @@ xfs_trans_commit_iclog(
 	*commit_lsn = xfs_log_done(mp, tp->t_ticket, &commit_iclog, log_flags);
 
 	tp->t_commit_lsn = *commit_lsn;
+	trace_xfs_trans_commit_lsn(tp);
+
 	if (nvec > XFS_TRANS_LOGVEC_COUNT)
 		kmem_free(log_vector);
 
@@ -1161,6 +1144,93 @@ xfs_trans_commit_iclog(
 	return xfs_log_release_iclog(mp, commit_iclog);
 }
 
+/*
+ * Walk the log items and allocate log vector structures for
+ * each item large enough to fit all the vectors they require.
+ * Note that this format differs from the old log vector format in
+ * that there is no transaction header in these log vectors.
+ */
+STATIC struct xfs_log_vec *
+xfs_trans_alloc_log_vecs(
+	xfs_trans_t	*tp)
+{
+	xfs_log_item_desc_t	*lidp;
+	struct xfs_log_vec	*lv = NULL;
+	struct xfs_log_vec	*ret_lv = NULL;
+
+	lidp = xfs_trans_first_item(tp);
+
+	/* Bail out if we didn't find a log item.  */
+	if (!lidp) {
+		ASSERT(0);
+		return NULL;
+	}
+
+	while (lidp != NULL) {
+		struct xfs_log_vec *new_lv;
+
+		/* Skip items which aren't dirty in this transaction. */
+		if (!(lidp->lid_flags & XFS_LID_DIRTY)) {
+			lidp = xfs_trans_next_item(tp, lidp);
+			continue;
+		}
+
+		/* Skip items that do not have any vectors for writing */
+		lidp->lid_size = IOP_SIZE(lidp->lid_item);
+		if (!lidp->lid_size) {
+			lidp = xfs_trans_next_item(tp, lidp);
+			continue;
+		}
+
+		new_lv = kmem_zalloc(sizeof(*new_lv) +
+				lidp->lid_size * sizeof(struct xfs_log_iovec),
+				KM_SLEEP);
+
+		/* The allocated iovec region lies beyond the log vector. */
+		new_lv->lv_iovecp = (struct xfs_log_iovec *)&new_lv[1];
+		new_lv->lv_niovecs = lidp->lid_size;
+		new_lv->lv_item = lidp->lid_item;
+		if (!ret_lv)
+			ret_lv = new_lv;
+		else
+			lv->lv_next = new_lv;
+		lv = new_lv;
+		lidp = xfs_trans_next_item(tp, lidp);
+	}
+
+	return ret_lv;
+}
+
+static int
+xfs_trans_commit_cil(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_lsn_t		*commit_lsn,
+	int			flags)
+{
+	struct xfs_log_vec	*log_vector;
+	int			error;
+
+	/*
+	 * Get each log item to allocate a vector structure for
+	 * the log item to to pass to the log write code. The
+	 * CIL commit code will format the vector and save it away.
+	 */
+	log_vector = xfs_trans_alloc_log_vecs(tp);
+	if (!log_vector)
+		return ENOMEM;
+
+	error = xfs_log_commit_cil(mp, tp, log_vector, commit_lsn, flags);
+	if (error)
+		return error;
+
+	current_restore_flags_nested(&tp->t_pflags, PF_FSTRANS);
+
+	/* xfs_trans_free_items() unlocks them first */
+	xfs_trans_free_items(tp, *commit_lsn, 0);
+	xfs_trans_free(tp);
+	return 0;
+}
 
 /*
  * xfs_trans_commit
@@ -1221,7 +1291,11 @@ _xfs_trans_commit(
 		xfs_trans_apply_sb_deltas(tp);
 	xfs_trans_apply_dquot_deltas(tp);
 
-	error = xfs_trans_commit_iclog(mp, tp, &commit_lsn, flags);
+	if (mp->m_flags & XFS_MOUNT_DELAYLOG)
+		error = xfs_trans_commit_cil(mp, tp, &commit_lsn, flags);
+	else
+		error = xfs_trans_commit_iclog(mp, tp, &commit_lsn, flags);
+
 	if (error == ENOMEM) {
 		xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
 		error = XFS_ERROR(EIO);
@@ -1259,8 +1333,7 @@ out_unreserve:
 			error = XFS_ERROR(EIO);
 	}
 	current_restore_flags_nested(&tp->t_pflags, PF_FSTRANS);
-	xfs_trans_free_items(tp, error ? XFS_TRANS_ABORT : 0);
-	xfs_trans_free_busy(tp);
+	xfs_trans_free_items(tp, NULLCOMMITLSN, error ? XFS_TRANS_ABORT : 0);
 	xfs_trans_free(tp);
 
 	XFS_STATS_INC(xs_trans_empty);
@@ -1338,8 +1411,7 @@ xfs_trans_cancel(
 	/* mark this thread as no longer being in a transaction */
 	current_restore_flags_nested(&tp->t_pflags, PF_FSTRANS);
 
-	xfs_trans_free_items(tp, flags);
-	xfs_trans_free_busy(tp);
+	xfs_trans_free_items(tp, NULLCOMMITLSN, flags);
 	xfs_trans_free(tp);
 }
 
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index c62beee0921e..8c69e7824f68 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -106,7 +106,8 @@ typedef struct xfs_trans_header {
 #define	XFS_TRANS_GROWFSRT_FREE		39
 #define	XFS_TRANS_SWAPEXT		40
 #define	XFS_TRANS_SB_COUNT		41
-#define	XFS_TRANS_TYPE_MAX		41
+#define	XFS_TRANS_CHECKPOINT		42
+#define	XFS_TRANS_TYPE_MAX		42
 /* new transaction types need to be reflected in xfs_logprint(8) */
 
 #define XFS_TRANS_TYPES \
@@ -148,6 +149,7 @@ typedef struct xfs_trans_header {
 	{ XFS_TRANS_GROWFSRT_FREE,	"GROWFSRT_FREE" }, \
 	{ XFS_TRANS_SWAPEXT,		"SWAPEXT" }, \
 	{ XFS_TRANS_SB_COUNT,		"SB_COUNT" }, \
+	{ XFS_TRANS_CHECKPOINT,		"CHECKPOINT" }, \
 	{ XFS_TRANS_DUMMY1,		"DUMMY1" }, \
 	{ XFS_TRANS_DUMMY2,		"DUMMY2" }, \
 	{ XLOG_UNMOUNT_REC_TYPE,	"UNMOUNT" }
@@ -813,6 +815,7 @@ struct xfs_log_item_desc;
 struct xfs_mount;
 struct xfs_trans;
 struct xfs_dquot_acct;
+struct xfs_busy_extent;
 
 typedef struct xfs_log_item {
 	struct list_head		li_ail;		/* AIL pointers */
@@ -828,6 +831,11 @@ typedef struct xfs_log_item {
 							/* buffer item iodone */
 							/* callback func */
 	struct xfs_item_ops		*li_ops;	/* function list */
+
+	/* delayed logging */
+	struct list_head		li_cil;		/* CIL pointers */
+	struct xfs_log_vec		*li_lv;		/* active log vector */
+	xfs_lsn_t			li_seq;		/* CIL commit seq */
 } xfs_log_item_t;
 
 #define	XFS_LI_IN_AIL	0x1
@@ -872,34 +880,6 @@ typedef struct xfs_item_ops {
 #define XFS_ITEM_PUSHBUF	3
 
 /*
- * This structure is used to maintain a list of block ranges that have been
- * freed in the transaction.  The ranges are listed in the perag[] busy list
- * between when they're freed and the transaction is committed to disk.
- */
-
-typedef struct xfs_log_busy_slot {
-	xfs_agnumber_t		lbc_ag;
-	ushort			lbc_idx;	/* index in perag.busy[] */
-} xfs_log_busy_slot_t;
-
-#define XFS_LBC_NUM_SLOTS	31
-typedef struct xfs_log_busy_chunk {
-	struct xfs_log_busy_chunk	*lbc_next;
-	uint				lbc_free;	/* free slots bitmask */
-	ushort				lbc_unused;	/* first unused */
-	xfs_log_busy_slot_t		lbc_busy[XFS_LBC_NUM_SLOTS];
-} xfs_log_busy_chunk_t;
-
-#define	XFS_LBC_MAX_SLOT	(XFS_LBC_NUM_SLOTS - 1)
-#define	XFS_LBC_FREEMASK	((1U << XFS_LBC_NUM_SLOTS) - 1)
-
-#define	XFS_LBC_INIT(cp)	((cp)->lbc_free = XFS_LBC_FREEMASK)
-#define	XFS_LBC_CLAIM(cp, slot)	((cp)->lbc_free &= ~(1 << (slot)))
-#define	XFS_LBC_SLOT(cp, slot)	(&((cp)->lbc_busy[(slot)]))
-#define	XFS_LBC_VACANCY(cp)	(((cp)->lbc_free) & XFS_LBC_FREEMASK)
-#define	XFS_LBC_ISFREE(cp, slot) ((cp)->lbc_free & (1 << (slot)))
-
-/*
  * This is the type of function which can be given to xfs_trans_callback()
  * to be called upon the transaction's commit to disk.
  */
@@ -950,8 +930,7 @@ typedef struct xfs_trans {
 	unsigned int		t_items_free;	/* log item descs free */
 	xfs_log_item_chunk_t	t_items;	/* first log item desc chunk */
 	xfs_trans_header_t	t_header;	/* header for in-log trans */
-	unsigned int		t_busy_free;	/* busy descs free */
-	xfs_log_busy_chunk_t	t_busy;		/* busy/async free blocks */
+	struct list_head	t_busy;		/* list of busy extents */
 	unsigned long		t_pflags;	/* saved process flags state */
 } xfs_trans_t;
 
@@ -1025,9 +1004,6 @@ int		_xfs_trans_commit(xfs_trans_t *,
 void		xfs_trans_cancel(xfs_trans_t *, int);
 int		xfs_trans_ail_init(struct xfs_mount *);
 void		xfs_trans_ail_destroy(struct xfs_mount *);
-xfs_log_busy_slot_t *xfs_trans_add_busy(xfs_trans_t *tp,
-					xfs_agnumber_t ag,
-					xfs_extlen_t idx);
 
 extern kmem_zone_t	*xfs_trans_zone;
 
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 9cd809025f3a..63d81a22f4fd 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -114,7 +114,7 @@ _xfs_trans_bjoin(
 	xfs_buf_item_init(bp, tp->t_mountp);
 	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
 	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-	ASSERT(!(bip->bli_format.blf_flags & XFS_BLI_CANCEL));
+	ASSERT(!(bip->bli_format.blf_flags & XFS_BLF_CANCEL));
 	ASSERT(!(bip->bli_flags & XFS_BLI_LOGGED));
 	if (reset_recur)
 		bip->bli_recur = 0;
@@ -511,7 +511,7 @@ xfs_trans_brelse(xfs_trans_t	*tp,
 	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
 	ASSERT(bip->bli_item.li_type == XFS_LI_BUF);
 	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-	ASSERT(!(bip->bli_format.blf_flags & XFS_BLI_CANCEL));
+	ASSERT(!(bip->bli_format.blf_flags & XFS_BLF_CANCEL));
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 
 	/*
@@ -619,7 +619,7 @@ xfs_trans_bhold(xfs_trans_t	*tp,
 
 	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
 	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-	ASSERT(!(bip->bli_format.blf_flags & XFS_BLI_CANCEL));
+	ASSERT(!(bip->bli_format.blf_flags & XFS_BLF_CANCEL));
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 	bip->bli_flags |= XFS_BLI_HOLD;
 	trace_xfs_trans_bhold(bip);
@@ -641,7 +641,7 @@ xfs_trans_bhold_release(xfs_trans_t	*tp,
 
 	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
 	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
-	ASSERT(!(bip->bli_format.blf_flags & XFS_BLI_CANCEL));
+	ASSERT(!(bip->bli_format.blf_flags & XFS_BLF_CANCEL));
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 	ASSERT(bip->bli_flags & XFS_BLI_HOLD);
 	bip->bli_flags &= ~XFS_BLI_HOLD;
@@ -704,7 +704,7 @@ xfs_trans_log_buf(xfs_trans_t	*tp,
 		bip->bli_flags &= ~XFS_BLI_STALE;
 		ASSERT(XFS_BUF_ISSTALE(bp));
 		XFS_BUF_UNSTALE(bp);
-		bip->bli_format.blf_flags &= ~XFS_BLI_CANCEL;
+		bip->bli_format.blf_flags &= ~XFS_BLF_CANCEL;
 	}
 
 	lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)bip);
@@ -762,8 +762,8 @@ xfs_trans_binval(
 		ASSERT(!(XFS_BUF_ISDELAYWRITE(bp)));
 		ASSERT(XFS_BUF_ISSTALE(bp));
 		ASSERT(!(bip->bli_flags & (XFS_BLI_LOGGED | XFS_BLI_DIRTY)));
-		ASSERT(!(bip->bli_format.blf_flags & XFS_BLI_INODE_BUF));
-		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
+		ASSERT(!(bip->bli_format.blf_flags & XFS_BLF_INODE_BUF));
+		ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
 		ASSERT(lidp->lid_flags & XFS_LID_DIRTY);
 		ASSERT(tp->t_flags & XFS_TRANS_DIRTY);
 		return;
@@ -774,7 +774,7 @@ xfs_trans_binval(
 	 * in the buf log item.  The STALE flag will be used in
 	 * xfs_buf_item_unpin() to determine if it should clean up
 	 * when the last reference to the buf item is given up.
-	 * We set the XFS_BLI_CANCEL flag in the buf log format structure
+	 * We set the XFS_BLF_CANCEL flag in the buf log format structure
 	 * and log the buf item.  This will be used at recovery time
 	 * to determine that copies of the buffer in the log before
 	 * this should not be replayed.
@@ -792,9 +792,9 @@ xfs_trans_binval(
 	XFS_BUF_UNDELAYWRITE(bp);
 	XFS_BUF_STALE(bp);
 	bip->bli_flags |= XFS_BLI_STALE;
-	bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_DIRTY);
-	bip->bli_format.blf_flags &= ~XFS_BLI_INODE_BUF;
-	bip->bli_format.blf_flags |= XFS_BLI_CANCEL;
+	bip->bli_flags &= ~(XFS_BLI_INODE_BUF | XFS_BLI_LOGGED | XFS_BLI_DIRTY);
+	bip->bli_format.blf_flags &= ~XFS_BLF_INODE_BUF;
+	bip->bli_format.blf_flags |= XFS_BLF_CANCEL;
 	memset((char *)(bip->bli_format.blf_data_map), 0,
 	      (bip->bli_format.blf_map_size * sizeof(uint)));
 	lidp->lid_flags |= XFS_LID_DIRTY;
@@ -802,16 +802,16 @@ xfs_trans_binval(
 }
 
 /*
- * This call is used to indicate that the buffer contains on-disk
- * inodes which must be handled specially during recovery.  They
- * require special handling because only the di_next_unlinked from
- * the inodes in the buffer should be recovered.  The rest of the
- * data in the buffer is logged via the inodes themselves.
+ * This call is used to indicate that the buffer contains on-disk inodes which
+ * must be handled specially during recovery.  They require special handling
+ * because only the di_next_unlinked from the inodes in the buffer should be
+ * recovered.  The rest of the data in the buffer is logged via the inodes
+ * themselves.
  *
- * All we do is set the XFS_BLI_INODE_BUF flag in the buffer's log
- * format structure so that we'll know what to do at recovery time.
+ * All we do is set the XFS_BLI_INODE_BUF flag in the items flags so it can be
+ * transferred to the buffer's log format structure so that we'll know what to
+ * do at recovery time.
  */
-/* ARGSUSED */
 void
 xfs_trans_inode_buf(
 	xfs_trans_t	*tp,
@@ -826,7 +826,7 @@ xfs_trans_inode_buf(
 	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 
-	bip->bli_format.blf_flags |= XFS_BLI_INODE_BUF;
+	bip->bli_flags |= XFS_BLI_INODE_BUF;
 }
 
 /*
@@ -908,9 +908,9 @@ xfs_trans_dquot_buf(
 	ASSERT(XFS_BUF_ISBUSY(bp));
 	ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
 	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);
-	ASSERT(type == XFS_BLI_UDQUOT_BUF ||
-	       type == XFS_BLI_PDQUOT_BUF ||
-	       type == XFS_BLI_GDQUOT_BUF);
+	ASSERT(type == XFS_BLF_UDQUOT_BUF ||
+	       type == XFS_BLF_PDQUOT_BUF ||
+	       type == XFS_BLF_GDQUOT_BUF);
 
 	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
diff --git a/fs/xfs/xfs_trans_item.c b/fs/xfs/xfs_trans_item.c
index eb3fc57f9eef..f11d37d06dcc 100644
--- a/fs/xfs/xfs_trans_item.c
+++ b/fs/xfs/xfs_trans_item.c
@@ -299,6 +299,7 @@ xfs_trans_next_item(xfs_trans_t *tp, xfs_log_item_desc_t *lidp)
 void
 xfs_trans_free_items(
 	xfs_trans_t	*tp,
+	xfs_lsn_t	commit_lsn,
 	int		flags)
 {
 	xfs_log_item_chunk_t	*licp;
@@ -311,7 +312,7 @@ xfs_trans_free_items(
 	 * Special case the embedded chunk so we don't free it below.
 	 */
 	if (!xfs_lic_are_all_free(licp)) {
-		(void) xfs_trans_unlock_chunk(licp, 1, abort, NULLCOMMITLSN);
+		(void) xfs_trans_unlock_chunk(licp, 1, abort, commit_lsn);
 		xfs_lic_all_free(licp);
 		licp->lic_unused = 0;
 	}
@@ -322,7 +323,7 @@ xfs_trans_free_items(
 	 */
 	while (licp != NULL) {
 		ASSERT(!xfs_lic_are_all_free(licp));
-		(void) xfs_trans_unlock_chunk(licp, 1, abort, NULLCOMMITLSN);
+		(void) xfs_trans_unlock_chunk(licp, 1, abort, commit_lsn);
 		next_licp = licp->lic_next;
 		kmem_free(licp);
 		licp = next_licp;
@@ -438,112 +439,3 @@ xfs_trans_unlock_chunk(
 
 	return freed;
 }
-
-
-/*
- * This is called to add the given busy item to the transaction's
- * list of busy items.  It must find a free busy item descriptor
- * or allocate a new one and add the item to that descriptor.
- * The function returns a pointer to busy descriptor used to point
- * to the new busy entry.  The log busy entry will now point to its new
- * descriptor with its ???? field.
- */
-xfs_log_busy_slot_t *
-xfs_trans_add_busy(xfs_trans_t *tp, xfs_agnumber_t ag, xfs_extlen_t idx)
-{
-	xfs_log_busy_chunk_t	*lbcp;
-	xfs_log_busy_slot_t	*lbsp;
-	int			i=0;
-
-	/*
-	 * If there are no free descriptors, allocate a new chunk
-	 * of them and put it at the front of the chunk list.
-	 */
-	if (tp->t_busy_free == 0) {
-		lbcp = (xfs_log_busy_chunk_t*)
-		       kmem_alloc(sizeof(xfs_log_busy_chunk_t), KM_SLEEP);
-		ASSERT(lbcp != NULL);
-		/*
-		 * Initialize the chunk, and then
-		 * claim the first slot in the newly allocated chunk.
-		 */
-		XFS_LBC_INIT(lbcp);
-		XFS_LBC_CLAIM(lbcp, 0);
-		lbcp->lbc_unused = 1;
-		lbsp = XFS_LBC_SLOT(lbcp, 0);
-
-		/*
-		 * Link in the new chunk and update the free count.
-		 */
-		lbcp->lbc_next = tp->t_busy.lbc_next;
-		tp->t_busy.lbc_next = lbcp;
-		tp->t_busy_free = XFS_LIC_NUM_SLOTS - 1;
-
-		/*
-		 * Initialize the descriptor and the generic portion
-		 * of the log item.
-		 *
-		 * Point the new slot at this item and return it.
-		 * Also point the log item at its currently active
-		 * descriptor and set the item's mount pointer.
-		 */
-		lbsp->lbc_ag = ag;
-		lbsp->lbc_idx = idx;
-		return lbsp;
-	}
-
-	/*
-	 * Find the free descriptor. It is somewhere in the chunklist
-	 * of descriptors.
-	 */
-	lbcp = &tp->t_busy;
-	while (lbcp != NULL) {
-		if (XFS_LBC_VACANCY(lbcp)) {
-			if (lbcp->lbc_unused <= XFS_LBC_MAX_SLOT) {
-				i = lbcp->lbc_unused;
-				break;
-			} else {
-				/* out-of-order vacancy */
-				cmn_err(CE_DEBUG, "OOO vacancy lbcp 0x%p\n", lbcp);
-				ASSERT(0);
-			}
-		}
-		lbcp = lbcp->lbc_next;
-	}
-	ASSERT(lbcp != NULL);
-	/*
-	 * If we find a free descriptor, claim it,
-	 * initialize it, and return it.
-	 */
-	XFS_LBC_CLAIM(lbcp, i);
-	if (lbcp->lbc_unused <= i) {
-		lbcp->lbc_unused = i + 1;
-	}
-	lbsp = XFS_LBC_SLOT(lbcp, i);
-	tp->t_busy_free--;
-	lbsp->lbc_ag = ag;
-	lbsp->lbc_idx = idx;
-	return lbsp;
-}
-
-
-/*
- * xfs_trans_free_busy
- * Free all of the busy lists from a transaction
- */
-void
-xfs_trans_free_busy(xfs_trans_t *tp)
-{
-	xfs_log_busy_chunk_t	*lbcp;
-	xfs_log_busy_chunk_t	*lbcq;
-
-	lbcp = tp->t_busy.lbc_next;
-	while (lbcp != NULL) {
-		lbcq = lbcp->lbc_next;
-		kmem_free(lbcp);
-		lbcp = lbcq;
-	}
-
-	XFS_LBC_INIT(&tp->t_busy);
-	tp->t_busy.lbc_unused = 0;
-}
diff --git a/fs/xfs/xfs_trans_priv.h b/fs/xfs/xfs_trans_priv.h
index 73e2ad397432..c6e4f2c8de6e 100644
--- a/fs/xfs/xfs_trans_priv.h
+++ b/fs/xfs/xfs_trans_priv.h
@@ -35,13 +35,14 @@ struct xfs_log_item_desc	*xfs_trans_find_item(struct xfs_trans *,
 struct xfs_log_item_desc	*xfs_trans_first_item(struct xfs_trans *);
 struct xfs_log_item_desc	*xfs_trans_next_item(struct xfs_trans *,
 					     struct xfs_log_item_desc *);
-void				xfs_trans_free_items(struct xfs_trans *, int);
-void				xfs_trans_unlock_items(struct xfs_trans *,
-							xfs_lsn_t);
-void				xfs_trans_free_busy(xfs_trans_t *tp);
-xfs_log_busy_slot_t		*xfs_trans_add_busy(xfs_trans_t *tp,
-						    xfs_agnumber_t ag,
-						    xfs_extlen_t idx);
+
+void	xfs_trans_unlock_items(struct xfs_trans *tp, xfs_lsn_t commit_lsn);
+void	xfs_trans_free_items(struct xfs_trans *tp, xfs_lsn_t commit_lsn,
+				int flags);
+
+void	xfs_trans_item_committed(struct xfs_log_item *lip,
+				xfs_lsn_t commit_lsn, int aborted);
+void	xfs_trans_unreserve_and_mod_sb(struct xfs_trans *tp);
 
 /*
  * AIL traversal cursor.
diff --git a/fs/xfs/xfs_types.h b/fs/xfs/xfs_types.h
index b09904555d07..320775295e32 100644
--- a/fs/xfs/xfs_types.h
+++ b/fs/xfs/xfs_types.h
@@ -75,6 +75,8 @@ typedef	__uint32_t	xfs_dahash_t;	/* dir/attr hash value */
 
 typedef __uint16_t	xfs_prid_t;	/* prid_t truncated to 16bits in XFS */
 
+typedef __uint32_t	xlog_tid_t;	/* transaction ID type */
+
 /*
  * These types are 64 bits on disk but are either 32 or 64 bits in memory.
  * Disk based types: