Merge branch 'next' into for-linus

Merge first round of changes for 3.13 merge window.

48 files changed, 6714 insertions, 4254 deletions

diff --git a/fs/btrfs/Kconfig b/fs/btrfs/Kconfig
index 2b3b83296977..398cbd517be2 100644
--- a/fs/btrfs/Kconfig
+++ b/fs/btrfs/Kconfig
@@ -72,3 +72,12 @@ config BTRFS_DEBUG
 	  performance, or export extra information via sysfs.
 
 	  If unsure, say N.
+
+config BTRFS_ASSERT
+	bool "Btrfs assert support"
+	depends on BTRFS_FS
+	help
+	  Enable run-time assertion checking.  This will result in panics if
+	  any of the assertions trip.  This is meant for btrfs developers only.
+
+	  If unsure, say N.
diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 3932224f99e9..a91a6a355cc5 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -8,7 +8,10 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
 	   export.o tree-log.o free-space-cache.o zlib.o lzo.o \
 	   compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \
-	   reada.o backref.o ulist.o qgroup.o send.o dev-replace.o raid56.o
+	   reada.o backref.o ulist.o qgroup.o send.o dev-replace.o raid56.o \
+	   uuid-tree.o
 
 btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o
 btrfs-$(CONFIG_BTRFS_FS_CHECK_INTEGRITY) += check-integrity.o
+
+btrfs-$(CONFIG_BTRFS_FS_RUN_SANITY_TESTS) += tests/free-space-tests.o
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 58b7d14b08ee..08cc08f037a6 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -107,7 +107,8 @@ static void check_idle_worker(struct btrfs_worker_thread *worker)
 		worker->idle = 1;
 
 		/* the list may be empty if the worker is just starting */
-		if (!list_empty(&worker->worker_list)) {
+		if (!list_empty(&worker->worker_list) &&
+		    !worker->workers->stopping) {
 			list_move(&worker->worker_list,
 				 &worker->workers->idle_list);
 		}
@@ -127,7 +128,8 @@ static void check_busy_worker(struct btrfs_worker_thread *worker)
 		spin_lock_irqsave(&worker->workers->lock, flags);
 		worker->idle = 0;
 
-		if (!list_empty(&worker->worker_list)) {
+		if (!list_empty(&worker->worker_list) &&
+		    !worker->workers->stopping) {
 			list_move_tail(&worker->worker_list,
 				      &worker->workers->worker_list);
 		}
@@ -412,6 +414,7 @@ void btrfs_stop_workers(struct btrfs_workers *workers)
 	int can_stop;
 
 	spin_lock_irq(&workers->lock);
+	workers->stopping = 1;
 	list_splice_init(&workers->idle_list, &workers->worker_list);
 	while (!list_empty(&workers->worker_list)) {
 		cur = workers->worker_list.next;
@@ -455,6 +458,7 @@ void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max,
 	workers->ordered = 0;
 	workers->atomic_start_pending = 0;
 	workers->atomic_worker_start = async_helper;
+	workers->stopping = 0;
 }
 
 /*
@@ -480,15 +484,19 @@ static int __btrfs_start_workers(struct btrfs_workers *workers)
 	atomic_set(&worker->num_pending, 0);
 	atomic_set(&worker->refs, 1);
 	worker->workers = workers;
-	worker->task = kthread_run(worker_loop, worker,
-				   "btrfs-%s-%d", workers->name,
-				   workers->num_workers + 1);
+	worker->task = kthread_create(worker_loop, worker,
+				      "btrfs-%s-%d", workers->name,
+				      workers->num_workers + 1);
 	if (IS_ERR(worker->task)) {
 		ret = PTR_ERR(worker->task);
-		kfree(worker);
 		goto fail;
 	}
+
 	spin_lock_irq(&workers->lock);
+	if (workers->stopping) {
+		spin_unlock_irq(&workers->lock);
+		goto fail_kthread;
+	}
 	list_add_tail(&worker->worker_list, &workers->idle_list);
 	worker->idle = 1;
 	workers->num_workers++;
@@ -496,8 +504,13 @@ static int __btrfs_start_workers(struct btrfs_workers *workers)
 	WARN_ON(workers->num_workers_starting < 0);
 	spin_unlock_irq(&workers->lock);
 
+	wake_up_process(worker->task);
 	return 0;
+
+fail_kthread:
+	kthread_stop(worker->task);
 fail:
+	kfree(worker);
 	spin_lock_irq(&workers->lock);
 	workers->num_workers_starting--;
 	spin_unlock_irq(&workers->lock);
diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h
index 063698b90ce2..1f26792683ed 100644
--- a/fs/btrfs/async-thread.h
+++ b/fs/btrfs/async-thread.h
@@ -107,6 +107,8 @@ struct btrfs_workers {
 
 	/* extra name for this worker, used for current->name */
 	char *name;
+
+	int stopping;
 };
 
 void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work);
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 290e347b6db3..0552a599b28f 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -36,16 +36,23 @@ static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb,
 				u64 extent_item_pos,
 				struct extent_inode_elem **eie)
 {
-	u64 data_offset;
-	u64 data_len;
+	u64 offset = 0;
 	struct extent_inode_elem *e;
 
-	data_offset = btrfs_file_extent_offset(eb, fi);
-	data_len = btrfs_file_extent_num_bytes(eb, fi);
+	if (!btrfs_file_extent_compression(eb, fi) &&
+	    !btrfs_file_extent_encryption(eb, fi) &&
+	    !btrfs_file_extent_other_encoding(eb, fi)) {
+		u64 data_offset;
+		u64 data_len;
 
-	if (extent_item_pos < data_offset ||
-	    extent_item_pos >= data_offset + data_len)
-		return 1;
+		data_offset = btrfs_file_extent_offset(eb, fi);
+		data_len = btrfs_file_extent_num_bytes(eb, fi);
+
+		if (extent_item_pos < data_offset ||
+		    extent_item_pos >= data_offset + data_len)
+			return 1;
+		offset = extent_item_pos - data_offset;
+	}
 
 	e = kmalloc(sizeof(*e), GFP_NOFS);
 	if (!e)
@@ -53,7 +60,7 @@ static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb,
 
 	e->next = *eie;
 	e->inum = key->objectid;
-	e->offset = key->offset + (extent_item_pos - data_offset);
+	e->offset = key->offset + offset;
 	*eie = e;
 
 	return 0;
@@ -112,6 +119,26 @@ struct __prelim_ref {
 	u64 wanted_disk_byte;
 };
 
+static struct kmem_cache *btrfs_prelim_ref_cache;
+
+int __init btrfs_prelim_ref_init(void)
+{
+	btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref",
+					sizeof(struct __prelim_ref),
+					0,
+					SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
+					NULL);
+	if (!btrfs_prelim_ref_cache)
+		return -ENOMEM;
+	return 0;
+}
+
+void btrfs_prelim_ref_exit(void)
+{
+	if (btrfs_prelim_ref_cache)
+		kmem_cache_destroy(btrfs_prelim_ref_cache);
+}
+
 /*
  * the rules for all callers of this function are:
  * - obtaining the parent is the goal
@@ -153,12 +180,12 @@ struct __prelim_ref {
 
 static int __add_prelim_ref(struct list_head *head, u64 root_id,
 			    struct btrfs_key *key, int level,
-			    u64 parent, u64 wanted_disk_byte, int count)
+			    u64 parent, u64 wanted_disk_byte, int count,
+			    gfp_t gfp_mask)
 {
 	struct __prelim_ref *ref;
 
-	/* in case we're adding delayed refs, we're holding the refs spinlock */
-	ref = kmalloc(sizeof(*ref), GFP_ATOMIC);
+	ref = kmem_cache_alloc(btrfs_prelim_ref_cache, gfp_mask);
 	if (!ref)
 		return -ENOMEM;
 
@@ -189,7 +216,7 @@ static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
 	struct extent_buffer *eb;
 	struct btrfs_key key;
 	struct btrfs_file_extent_item *fi;
-	struct extent_inode_elem *eie = NULL;
+	struct extent_inode_elem *eie = NULL, *old = NULL;
 	u64 disk_byte;
 
 	if (level != 0) {
@@ -223,6 +250,7 @@ static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
 
 		if (disk_byte == wanted_disk_byte) {
 			eie = NULL;
+			old = NULL;
 			if (extent_item_pos) {
 				ret = check_extent_in_eb(&key, eb, fi,
 						*extent_item_pos,
@@ -230,18 +258,20 @@ static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
 				if (ret < 0)
 					break;
 			}
-			if (!ret) {
-				ret = ulist_add(parents, eb->start,
-						(uintptr_t)eie, GFP_NOFS);
-				if (ret < 0)
-					break;
-				if (!extent_item_pos) {
-					ret = btrfs_next_old_leaf(root, path,
-							time_seq);
-					continue;
-				}
+			if (ret > 0)
+				goto next;
+			ret = ulist_add_merge(parents, eb->start,
+					      (uintptr_t)eie,
+					      (u64 *)&old, GFP_NOFS);
+			if (ret < 0)
+				break;
+			if (!ret && extent_item_pos) {
+				while (old->next)
+					old = old->next;
+				old->next = eie;
 			}
 		}
+next:
 		ret = btrfs_next_old_item(root, path, time_seq);
 	}
 
@@ -255,13 +285,11 @@ static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
  * to a logical address
  */
 static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
-					int search_commit_root,
-					u64 time_seq,
-					struct __prelim_ref *ref,
-					struct ulist *parents,
-					const u64 *extent_item_pos)
+				  struct btrfs_path *path, u64 time_seq,
+				  struct __prelim_ref *ref,
+				  struct ulist *parents,
+				  const u64 *extent_item_pos)
 {
-	struct btrfs_path *path;
 	struct btrfs_root *root;
 	struct btrfs_key root_key;
 	struct extent_buffer *eb;
@@ -269,11 +297,6 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
 	int root_level;
 	int level = ref->level;
 
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
-	path->search_commit_root = !!search_commit_root;
-
 	root_key.objectid = ref->root_id;
 	root_key.type = BTRFS_ROOT_ITEM_KEY;
 	root_key.offset = (u64)-1;
@@ -292,10 +315,9 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
 	ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq);
 	pr_debug("search slot in root %llu (level %d, ref count %d) returned "
 		 "%d for key (%llu %u %llu)\n",
-		 (unsigned long long)ref->root_id, level, ref->count, ret,
-		 (unsigned long long)ref->key_for_search.objectid,
-		 ref->key_for_search.type,
-		 (unsigned long long)ref->key_for_search.offset);
+		 ref->root_id, level, ref->count, ret,
+		 ref->key_for_search.objectid, ref->key_for_search.type,
+		 ref->key_for_search.offset);
 	if (ret < 0)
 		goto out;
 
@@ -314,7 +336,8 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
 				time_seq, ref->wanted_disk_byte,
 				extent_item_pos);
 out:
-	btrfs_free_path(path);
+	path->lowest_level = 0;
+	btrfs_release_path(path);
 	return ret;
 }
 
@@ -322,7 +345,7 @@ out:
  * resolve all indirect backrefs from the list
  */
 static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
-				   int search_commit_root, u64 time_seq,
+				   struct btrfs_path *path, u64 time_seq,
 				   struct list_head *head,
 				   const u64 *extent_item_pos)
 {
@@ -349,9 +372,8 @@ static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
 			continue;
 		if (ref->count == 0)
 			continue;
-		err = __resolve_indirect_ref(fs_info, search_commit_root,
-					     time_seq, ref, parents,
-					     extent_item_pos);
+		err = __resolve_indirect_ref(fs_info, path, time_seq, ref,
+					     parents, extent_item_pos);
 		if (err == -ENOMEM)
 			goto out;
 		if (err)
@@ -362,11 +384,12 @@ static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
 		node = ulist_next(parents, &uiter);
 		ref->parent = node ? node->val : 0;
 		ref->inode_list = node ?
-			(struct extent_inode_elem *)(uintptr_t)node->aux : 0;
+			(struct extent_inode_elem *)(uintptr_t)node->aux : NULL;
 
 		/* additional parents require new refs being added here */
 		while ((node = ulist_next(parents, &uiter))) {
-			new_ref = kmalloc(sizeof(*new_ref), GFP_NOFS);
+			new_ref = kmem_cache_alloc(btrfs_prelim_ref_cache,
+						   GFP_NOFS);
 			if (!new_ref) {
 				ret = -ENOMEM;
 				goto out;
@@ -490,7 +513,7 @@ static void __merge_refs(struct list_head *head, int mode)
 			ref1->count += ref2->count;
 
 			list_del(&ref2->list);
-			kfree(ref2);
+			kmem_cache_free(btrfs_prelim_ref_cache, ref2);
 		}
 
 	}
@@ -545,7 +568,7 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 			ref = btrfs_delayed_node_to_tree_ref(node);
 			ret = __add_prelim_ref(prefs, ref->root, &op_key,
 					       ref->level + 1, 0, node->bytenr,
-					       node->ref_mod * sgn);
+					       node->ref_mod * sgn, GFP_ATOMIC);
 			break;
 		}
 		case BTRFS_SHARED_BLOCK_REF_KEY: {
@@ -555,7 +578,7 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 			ret = __add_prelim_ref(prefs, ref->root, NULL,
 					       ref->level + 1, ref->parent,
 					       node->bytenr,
-					       node->ref_mod * sgn);
+					       node->ref_mod * sgn, GFP_ATOMIC);
 			break;
 		}
 		case BTRFS_EXTENT_DATA_REF_KEY: {
@@ -567,7 +590,7 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 			key.offset = ref->offset;
 			ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0,
 					       node->bytenr,
-					       node->ref_mod * sgn);
+					       node->ref_mod * sgn, GFP_ATOMIC);
 			break;
 		}
 		case BTRFS_SHARED_DATA_REF_KEY: {
@@ -580,7 +603,7 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 			key.offset = ref->offset;
 			ret = __add_prelim_ref(prefs, ref->root, &key, 0,
 					       ref->parent, node->bytenr,
-					       node->ref_mod * sgn);
+					       node->ref_mod * sgn, GFP_ATOMIC);
 			break;
 		}
 		default:
@@ -604,6 +627,7 @@ static int __add_inline_refs(struct btrfs_fs_info *fs_info,
 	int slot;
 	struct extent_buffer *leaf;
 	struct btrfs_key key;
+	struct btrfs_key found_key;
 	unsigned long ptr;
 	unsigned long end;
 	struct btrfs_extent_item *ei;
@@ -621,17 +645,21 @@ static int __add_inline_refs(struct btrfs_fs_info *fs_info,
 
 	ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
 	flags = btrfs_extent_flags(leaf, ei);
+	btrfs_item_key_to_cpu(leaf, &found_key, slot);
 
 	ptr = (unsigned long)(ei + 1);
 	end = (unsigned long)ei + item_size;
 
-	if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+	if (found_key.type == BTRFS_EXTENT_ITEM_KEY &&
+	    flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
 		struct btrfs_tree_block_info *info;
 
 		info = (struct btrfs_tree_block_info *)ptr;
 		*info_level = btrfs_tree_block_level(leaf, info);
 		ptr += sizeof(struct btrfs_tree_block_info);
 		BUG_ON(ptr > end);
+	} else if (found_key.type == BTRFS_METADATA_ITEM_KEY) {
+		*info_level = found_key.offset;
 	} else {
 		BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
 	}
@@ -649,7 +677,7 @@ static int __add_inline_refs(struct btrfs_fs_info *fs_info,
 		case BTRFS_SHARED_BLOCK_REF_KEY:
 			ret = __add_prelim_ref(prefs, 0, NULL,
 						*info_level + 1, offset,
-						bytenr, 1);
+						bytenr, 1, GFP_NOFS);
 			break;
 		case BTRFS_SHARED_DATA_REF_KEY: {
 			struct btrfs_shared_data_ref *sdref;
@@ -658,13 +686,13 @@ static int __add_inline_refs(struct btrfs_fs_info *fs_info,
 			sdref = (struct btrfs_shared_data_ref *)(iref + 1);
 			count = btrfs_shared_data_ref_count(leaf, sdref);
 			ret = __add_prelim_ref(prefs, 0, NULL, 0, offset,
-					       bytenr, count);
+					       bytenr, count, GFP_NOFS);
 			break;
 		}
 		case BTRFS_TREE_BLOCK_REF_KEY:
 			ret = __add_prelim_ref(prefs, offset, NULL,
 					       *info_level + 1, 0,
-					       bytenr, 1);
+					       bytenr, 1, GFP_NOFS);
 			break;
 		case BTRFS_EXTENT_DATA_REF_KEY: {
 			struct btrfs_extent_data_ref *dref;
@@ -679,7 +707,7 @@ static int __add_inline_refs(struct btrfs_fs_info *fs_info,
 			key.offset = btrfs_extent_data_ref_offset(leaf, dref);
 			root = btrfs_extent_data_ref_root(leaf, dref);
 			ret = __add_prelim_ref(prefs, root, &key, 0, 0,
-					       bytenr, count);
+					       bytenr, count, GFP_NOFS);
 			break;
 		}
 		default:
@@ -730,7 +758,7 @@ static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
 		case BTRFS_SHARED_BLOCK_REF_KEY:
 			ret = __add_prelim_ref(prefs, 0, NULL,
 						info_level + 1, key.offset,
-						bytenr, 1);
+						bytenr, 1, GFP_NOFS);
 			break;
 		case BTRFS_SHARED_DATA_REF_KEY: {
 			struct btrfs_shared_data_ref *sdref;
@@ -740,13 +768,13 @@ static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
 					      struct btrfs_shared_data_ref);
 			count = btrfs_shared_data_ref_count(leaf, sdref);
 			ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset,
-						bytenr, count);
+						bytenr, count, GFP_NOFS);
 			break;
 		}
 		case BTRFS_TREE_BLOCK_REF_KEY:
 			ret = __add_prelim_ref(prefs, key.offset, NULL,
 					       info_level + 1, 0,
-					       bytenr, 1);
+					       bytenr, 1, GFP_NOFS);
 			break;
 		case BTRFS_EXTENT_DATA_REF_KEY: {
 			struct btrfs_extent_data_ref *dref;
@@ -762,7 +790,7 @@ static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
 			key.offset = btrfs_extent_data_ref_offset(leaf, dref);
 			root = btrfs_extent_data_ref_root(leaf, dref);
 			ret = __add_prelim_ref(prefs, root, &key, 0, 0,
-					       bytenr, count);
+					       bytenr, count, GFP_NOFS);
 			break;
 		}
 		default:
@@ -795,7 +823,6 @@ static int find_parent_nodes(struct btrfs_trans_handle *trans,
 	struct btrfs_delayed_ref_head *head;
 	int info_level = 0;
 	int ret;
-	int search_commit_root = (trans == BTRFS_BACKREF_SEARCH_COMMIT_ROOT);
 	struct list_head prefs_delayed;
 	struct list_head prefs;
 	struct __prelim_ref *ref;
@@ -804,13 +831,17 @@ static int find_parent_nodes(struct btrfs_trans_handle *trans,
 	INIT_LIST_HEAD(&prefs_delayed);
 
 	key.objectid = bytenr;
-	key.type = BTRFS_EXTENT_ITEM_KEY;
 	key.offset = (u64)-1;
+	if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
+		key.type = BTRFS_METADATA_ITEM_KEY;
+	else
+		key.type = BTRFS_EXTENT_ITEM_KEY;
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
-	path->search_commit_root = !!search_commit_root;
+	if (!trans)
+		path->search_commit_root = 1;
 
 	/*
 	 * grab both a lock on the path and a lock on the delayed ref head.
@@ -825,7 +856,7 @@ again:
 		goto out;
 	BUG_ON(ret == 0);
 
-	if (trans != BTRFS_BACKREF_SEARCH_COMMIT_ROOT) {
+	if (trans) {
 		/*
 		 * look if there are updates for this ref queued and lock the
 		 * head
@@ -869,7 +900,8 @@ again:
 		slot = path->slots[0];
 		btrfs_item_key_to_cpu(leaf, &key, slot);
 		if (key.objectid == bytenr &&
-		    key.type == BTRFS_EXTENT_ITEM_KEY) {
+		    (key.type == BTRFS_EXTENT_ITEM_KEY ||
+		     key.type == BTRFS_METADATA_ITEM_KEY)) {
 			ret = __add_inline_refs(fs_info, path, bytenr,
 						&info_level, &prefs);
 			if (ret)
@@ -890,8 +922,8 @@ again:
 
 	__merge_refs(&prefs, 1);
 
-	ret = __resolve_indirect_refs(fs_info, search_commit_root, time_seq,
-				      &prefs, extent_item_pos);
+	ret = __resolve_indirect_refs(fs_info, path, time_seq, &prefs,
+				      extent_item_pos);
 	if (ret)
 		goto out;
 
@@ -899,7 +931,6 @@ again:
 
 	while (!list_empty(&prefs)) {
 		ref = list_first_entry(&prefs, struct __prelim_ref, list);
-		list_del(&ref->list);
 		WARN_ON(ref->count < 0);
 		if (ref->count && ref->root_id && ref->parent == 0) {
 			/* no parent == root of tree */
@@ -923,8 +954,10 @@ again:
 				}
 				ret = find_extent_in_eb(eb, bytenr,
 							*extent_item_pos, &eie);
-				ref->inode_list = eie;
 				free_extent_buffer(eb);
+				if (ret < 0)
+					goto out;
+				ref->inode_list = eie;
 			}
 			ret = ulist_add_merge(refs, ref->parent,
 					      (uintptr_t)ref->inode_list,
@@ -942,7 +975,8 @@ again:
 				eie->next = ref->inode_list;
 			}
 		}
-		kfree(ref);
+		list_del(&ref->list);
+		kmem_cache_free(btrfs_prelim_ref_cache, ref);
 	}
 
 out:
@@ -950,13 +984,13 @@ out:
 	while (!list_empty(&prefs)) {
 		ref = list_first_entry(&prefs, struct __prelim_ref, list);
 		list_del(&ref->list);
-		kfree(ref);
+		kmem_cache_free(btrfs_prelim_ref_cache, ref);
 	}
 	while (!list_empty(&prefs_delayed)) {
 		ref = list_first_entry(&prefs_delayed, struct __prelim_ref,
 				       list);
 		list_del(&ref->list);
-		kfree(ref);
+		kmem_cache_free(btrfs_prelim_ref_cache, ref);
 	}
 
 	return ret;
@@ -1283,12 +1317,16 @@ int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
 {
 	int ret;
 	u64 flags;
+	u64 size = 0;
 	u32 item_size;
 	struct extent_buffer *eb;
 	struct btrfs_extent_item *ei;
 	struct btrfs_key key;
 
-	key.type = BTRFS_EXTENT_ITEM_KEY;
+	if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
+		key.type = BTRFS_METADATA_ITEM_KEY;
+	else
+		key.type = BTRFS_EXTENT_ITEM_KEY;
 	key.objectid = logical;
 	key.offset = (u64)-1;
 
@@ -1301,11 +1339,16 @@ int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
 		return ret;
 
 	btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]);
-	if (found_key->type != BTRFS_EXTENT_ITEM_KEY ||
+	if (found_key->type == BTRFS_METADATA_ITEM_KEY)
+		size = fs_info->extent_root->leafsize;
+	else if (found_key->type == BTRFS_EXTENT_ITEM_KEY)
+		size = found_key->offset;
+
+	if ((found_key->type != BTRFS_EXTENT_ITEM_KEY &&
+	     found_key->type != BTRFS_METADATA_ITEM_KEY) ||
 	    found_key->objectid > logical ||
-	    found_key->objectid + found_key->offset <= logical) {
-		pr_debug("logical %llu is not within any extent\n",
-			 (unsigned long long)logical);
+	    found_key->objectid + size <= logical) {
+		pr_debug("logical %llu is not within any extent\n", logical);
 		return -ENOENT;
 	}
 
@@ -1318,11 +1361,8 @@ int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
 
 	pr_debug("logical %llu is at position %llu within the extent (%llu "
 		 "EXTENT_ITEM %llu) flags %#llx size %u\n",
-		 (unsigned long long)logical,
-		 (unsigned long long)(logical - found_key->objectid),
-		 (unsigned long long)found_key->objectid,
-		 (unsigned long long)found_key->offset,
-		 (unsigned long long)flags, item_size);
+		 logical, logical - found_key->objectid, found_key->objectid,
+		 found_key->offset, flags, item_size);
 
 	WARN_ON(!flags_ret);
 	if (flags_ret) {
@@ -1459,7 +1499,7 @@ int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
 				iterate_extent_inodes_t *iterate, void *ctx)
 {
 	int ret;
-	struct btrfs_trans_handle *trans;
+	struct btrfs_trans_handle *trans = NULL;
 	struct ulist *refs = NULL;
 	struct ulist *roots = NULL;
 	struct ulist_node *ref_node = NULL;
@@ -1471,9 +1511,7 @@ int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
 	pr_debug("resolving all inodes for extent %llu\n",
 			extent_item_objectid);
 
-	if (search_commit_root) {
-		trans = BTRFS_BACKREF_SEARCH_COMMIT_ROOT;
-	} else {
+	if (!search_commit_root) {
 		trans = btrfs_join_transaction(fs_info->extent_root);
 		if (IS_ERR(trans))
 			return PTR_ERR(trans);
@@ -1496,7 +1534,7 @@ int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
 		while (!ret && (root_node = ulist_next(roots, &root_uiter))) {
 			pr_debug("root %llu references leaf %llu, data list "
 				 "%#llx\n", root_node->val, ref_node->val,
-				 (long long)ref_node->aux);
+				 ref_node->aux);
 			ret = iterate_leaf_refs((struct extent_inode_elem *)
 						(uintptr_t)ref_node->aux,
 						root_node->val,
@@ -1588,9 +1626,8 @@ static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root,
 			name_len = btrfs_inode_ref_name_len(eb, iref);
 			/* path must be released before calling iterate()! */
 			pr_debug("following ref at offset %u for inode %llu in "
-				 "tree %llu\n", cur,
-				 (unsigned long long)found_key.objectid,
-				 (unsigned long long)fs_root->objectid);
+				 "tree %llu\n", cur, found_key.objectid,
+				 fs_root->objectid);
 			ret = iterate(parent, name_len,
 				      (unsigned long)(iref + 1), eb, ctx);
 			if (ret)
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index 0f446d7ca2c0..a910b27a8ad9 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -23,8 +23,6 @@
 #include "ulist.h"
 #include "extent_io.h"
 
-#define BTRFS_BACKREF_SEARCH_COMMIT_ROOT ((struct btrfs_trans_handle *)0)
-
 struct inode_fs_paths {
 	struct btrfs_path		*btrfs_path;
 	struct btrfs_root		*fs_root;
@@ -74,4 +72,6 @@ int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
 			  struct btrfs_inode_extref **ret_extref,
 			  u64 *found_off);
 
+int __init btrfs_prelim_ref_init(void);
+void btrfs_prelim_ref_exit(void);
 #endif
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 08b286b2a2c5..71f074e1870b 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -213,11 +213,35 @@ static inline bool btrfs_is_free_space_inode(struct inode *inode)
 static inline int btrfs_inode_in_log(struct inode *inode, u64 generation)
 {
 	if (BTRFS_I(inode)->logged_trans == generation &&
-	    BTRFS_I(inode)->last_sub_trans <= BTRFS_I(inode)->last_log_commit)
+	    BTRFS_I(inode)->last_sub_trans <=
+	    BTRFS_I(inode)->last_log_commit &&
+	    BTRFS_I(inode)->last_sub_trans <=
+	    BTRFS_I(inode)->root->last_log_commit)
 		return 1;
 	return 0;
 }
 
+struct btrfs_dio_private {
+	struct inode *inode;
+	u64 logical_offset;
+	u64 disk_bytenr;
+	u64 bytes;
+	void *private;
+
+	/* number of bios pending for this dio */
+	atomic_t pending_bios;
+
+	/* IO errors */
+	int errors;
+
+	/* orig_bio is our btrfs_io_bio */
+	struct bio *orig_bio;
+
+	/* dio_bio came from fs/direct-io.c */
+	struct bio *dio_bio;
+	u8 csum[0];
+};
+
 /*
  * Disable DIO read nolock optimization, so new dio readers will be forced
  * to grab i_mutex. It is used to avoid the endless truncate due to
diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c
index 1431a6965017..1c47be187240 100644
--- a/fs/btrfs/check-integrity.c
+++ b/fs/btrfs/check-integrity.c
@@ -701,15 +701,13 @@ static int btrfsic_process_superblock(struct btrfsic_state *state,
 			next_bytenr = btrfs_super_root(selected_super);
 			if (state->print_mask &
 			    BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION)
-				printk(KERN_INFO "root@%llu\n",
-				       (unsigned long long)next_bytenr);
+				printk(KERN_INFO "root@%llu\n", next_bytenr);
 			break;
 		case 1:
 			next_bytenr = btrfs_super_chunk_root(selected_super);
 			if (state->print_mask &
 			    BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION)
-				printk(KERN_INFO "chunk@%llu\n",
-				       (unsigned long long)next_bytenr);
+				printk(KERN_INFO "chunk@%llu\n", next_bytenr);
 			break;
 		case 2:
 			next_bytenr = btrfs_super_log_root(selected_super);
@@ -717,8 +715,7 @@ static int btrfsic_process_superblock(struct btrfsic_state *state,
 				continue;
 			if (state->print_mask &
 			    BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION)
-				printk(KERN_INFO "log@%llu\n",
-				       (unsigned long long)next_bytenr);
+				printk(KERN_INFO "log@%llu\n", next_bytenr);
 			break;
 		}
 
@@ -727,7 +724,7 @@ static int btrfsic_process_superblock(struct btrfsic_state *state,
 				     next_bytenr, state->metablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
-			       (unsigned long long)next_bytenr, num_copies);
+			       next_bytenr, num_copies);
 
 		for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) {
 			struct btrfsic_block *next_block;
@@ -742,8 +739,7 @@ static int btrfsic_process_superblock(struct btrfsic_state *state,
 				printk(KERN_INFO "btrfsic:"
 				       " btrfsic_map_block(root @%llu,"
 				       " mirror %d) failed!\n",
-				       (unsigned long long)next_bytenr,
-				       mirror_num);
+				       next_bytenr, mirror_num);
 				kfree(selected_super);
 				return -1;
 			}
@@ -767,7 +763,6 @@ static int btrfsic_process_superblock(struct btrfsic_state *state,
 			if (ret < (int)PAGE_CACHE_SIZE) {
 				printk(KERN_INFO
 				       "btrfsic: read @logical %llu failed!\n",
-				       (unsigned long long)
 				       tmp_next_block_ctx.start);
 				btrfsic_release_block_ctx(&tmp_next_block_ctx);
 				kfree(selected_super);
@@ -813,7 +808,7 @@ static int btrfsic_process_superblock_dev_mirror(
 	    (bh->b_data + (dev_bytenr & 4095));
 
 	if (btrfs_super_bytenr(super_tmp) != dev_bytenr ||
-	    super_tmp->magic != cpu_to_le64(BTRFS_MAGIC) ||
+	    btrfs_super_magic(super_tmp) != BTRFS_MAGIC ||
 	    memcmp(device->uuid, super_tmp->dev_item.uuid, BTRFS_UUID_SIZE) ||
 	    btrfs_super_nodesize(super_tmp) != state->metablock_size ||
 	    btrfs_super_leafsize(super_tmp) != state->metablock_size ||
@@ -847,10 +842,8 @@ static int btrfsic_process_superblock_dev_mirror(
 			printk_in_rcu(KERN_INFO "New initial S-block (bdev %p, %s)"
 				     " @%llu (%s/%llu/%d)\n",
 				     superblock_bdev,
-				     rcu_str_deref(device->name),
-				     (unsigned long long)dev_bytenr,
-				     dev_state->name,
-				     (unsigned long long)dev_bytenr,
+				     rcu_str_deref(device->name), dev_bytenr,
+				     dev_state->name, dev_bytenr,
 				     superblock_mirror_num);
 		list_add(&superblock_tmp->all_blocks_node,
 			 &state->all_blocks_list);
@@ -880,20 +873,20 @@ static int btrfsic_process_superblock_dev_mirror(
 		tmp_disk_key.offset = 0;
 		switch (pass) {
 		case 0:
-			tmp_disk_key.objectid =
-			    cpu_to_le64(BTRFS_ROOT_TREE_OBJECTID);
+			btrfs_set_disk_key_objectid(&tmp_disk_key,
+						    BTRFS_ROOT_TREE_OBJECTID);
 			additional_string = "initial root ";
 			next_bytenr = btrfs_super_root(super_tmp);
 			break;
 		case 1:
-			tmp_disk_key.objectid =
-			    cpu_to_le64(BTRFS_CHUNK_TREE_OBJECTID);
+			btrfs_set_disk_key_objectid(&tmp_disk_key,
+						    BTRFS_CHUNK_TREE_OBJECTID);
 			additional_string = "initial chunk ";
 			next_bytenr = btrfs_super_chunk_root(super_tmp);
 			break;
 		case 2:
-			tmp_disk_key.objectid =
-			    cpu_to_le64(BTRFS_TREE_LOG_OBJECTID);
+			btrfs_set_disk_key_objectid(&tmp_disk_key,
+						    BTRFS_TREE_LOG_OBJECTID);
 			additional_string = "initial log ";
 			next_bytenr = btrfs_super_log_root(super_tmp);
 			if (0 == next_bytenr)
@@ -906,7 +899,7 @@ static int btrfsic_process_superblock_dev_mirror(
 				     next_bytenr, state->metablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
-			       (unsigned long long)next_bytenr, num_copies);
+			       next_bytenr, num_copies);
 		for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) {
 			struct btrfsic_block *next_block;
 			struct btrfsic_block_data_ctx tmp_next_block_ctx;
@@ -918,8 +911,7 @@ static int btrfsic_process_superblock_dev_mirror(
 					      mirror_num)) {
 				printk(KERN_INFO "btrfsic: btrfsic_map_block("
 				       "bytenr @%llu, mirror %d) failed!\n",
-				       (unsigned long long)next_bytenr,
-				       mirror_num);
+				       next_bytenr, mirror_num);
 				brelse(bh);
 				return -1;
 			}
@@ -1003,19 +995,17 @@ continue_with_new_stack_frame:
 		    (struct btrfs_leaf *)sf->hdr;
 
 		if (-1 == sf->i) {
-			sf->nr = le32_to_cpu(leafhdr->header.nritems);
+			sf->nr = btrfs_stack_header_nritems(&leafhdr->header);
 
 			if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE)
 				printk(KERN_INFO
 				       "leaf %llu items %d generation %llu"
 				       " owner %llu\n",
-				       (unsigned long long)
-				       sf->block_ctx->start,
-				       sf->nr,
-				       (unsigned long long)
-				       le64_to_cpu(leafhdr->header.generation),
-				       (unsigned long long)
-				       le64_to_cpu(leafhdr->header.owner));
+				       sf->block_ctx->start, sf->nr,
+				       btrfs_stack_header_generation(
+					       &leafhdr->header),
+				       btrfs_stack_header_owner(
+					       &leafhdr->header));
 		}
 
 continue_with_current_leaf_stack_frame:
@@ -1047,10 +1037,10 @@ leaf_item_out_of_bounce_error:
 						     &disk_item,
 						     disk_item_offset,
 						     sizeof(struct btrfs_item));
-			item_offset = le32_to_cpu(disk_item.offset);
-			item_size = le32_to_cpu(disk_item.size);
+			item_offset = btrfs_stack_item_offset(&disk_item);
+			item_size = btrfs_stack_item_offset(&disk_item);
 			disk_key = &disk_item.key;
-			type = disk_key->type;
+			type = btrfs_disk_key_type(disk_key);
 
 			if (BTRFS_ROOT_ITEM_KEY == type) {
 				struct btrfs_root_item root_item;
@@ -1066,7 +1056,7 @@ leaf_item_out_of_bounce_error:
 					sf->block_ctx, &root_item,
 					root_item_offset,
 					item_size);
-				next_bytenr = le64_to_cpu(root_item.bytenr);
+				next_bytenr = btrfs_root_bytenr(&root_item);
 
 				sf->error =
 				    btrfsic_create_link_to_next_block(
@@ -1081,8 +1071,8 @@ leaf_item_out_of_bounce_error:
 						&sf->num_copies,
 						&sf->mirror_num,
 						disk_key,
-						le64_to_cpu(root_item.
-						generation));
+						btrfs_root_generation(
+						&root_item));
 				if (sf->error)
 					goto one_stack_frame_backwards;
 
@@ -1130,18 +1120,17 @@ leaf_item_out_of_bounce_error:
 		struct btrfs_node *const nodehdr = (struct btrfs_node *)sf->hdr;
 
 		if (-1 == sf->i) {
-			sf->nr = le32_to_cpu(nodehdr->header.nritems);
+			sf->nr = btrfs_stack_header_nritems(&nodehdr->header);
 
 			if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE)
 				printk(KERN_INFO "node %llu level %d items %d"
 				       " generation %llu owner %llu\n",
-				       (unsigned long long)
 				       sf->block_ctx->start,
 				       nodehdr->header.level, sf->nr,
-				       (unsigned long long)
-				       le64_to_cpu(nodehdr->header.generation),
-				       (unsigned long long)
-				       le64_to_cpu(nodehdr->header.owner));
+				       btrfs_stack_header_generation(
+				       &nodehdr->header),
+				       btrfs_stack_header_owner(
+				       &nodehdr->header));
 		}
 
 continue_with_current_node_stack_frame:
@@ -1168,7 +1157,7 @@ continue_with_current_node_stack_frame:
 			btrfsic_read_from_block_data(
 				sf->block_ctx, &key_ptr, key_ptr_offset,
 				sizeof(struct btrfs_key_ptr));
-			next_bytenr = le64_to_cpu(key_ptr.blockptr);
+			next_bytenr = btrfs_stack_key_blockptr(&key_ptr);
 
 			sf->error = btrfsic_create_link_to_next_block(
 					state,
@@ -1182,7 +1171,7 @@ continue_with_current_node_stack_frame:
 					&sf->num_copies,
 					&sf->mirror_num,
 					&key_ptr.key,
-					le64_to_cpu(key_ptr.generation));
+					btrfs_stack_key_generation(&key_ptr));
 			if (sf->error)
 				goto one_stack_frame_backwards;
 
@@ -1247,8 +1236,7 @@ static void btrfsic_read_from_block_data(
 	unsigned long i = (start_offset + offset) >> PAGE_CACHE_SHIFT;
 
 	WARN_ON(offset + len > block_ctx->len);
-	offset_in_page = (start_offset + offset) &
-			 ((unsigned long)PAGE_CACHE_SIZE - 1);
+	offset_in_page = (start_offset + offset) & (PAGE_CACHE_SIZE - 1);
 
 	while (len > 0) {
 		cur = min(len, ((size_t)PAGE_CACHE_SIZE - offset_in_page));
@@ -1290,7 +1278,7 @@ static int btrfsic_create_link_to_next_block(
 				     next_bytenr, state->metablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
-			       (unsigned long long)next_bytenr, *num_copiesp);
+			       next_bytenr, *num_copiesp);
 		*mirror_nump = 1;
 	}
 
@@ -1307,7 +1295,7 @@ static int btrfsic_create_link_to_next_block(
 	if (ret) {
 		printk(KERN_INFO
 		       "btrfsic: btrfsic_map_block(@%llu, mirror=%d) failed!\n",
-		       (unsigned long long)next_bytenr, *mirror_nump);
+		       next_bytenr, *mirror_nump);
 		btrfsic_release_block_ctx(next_block_ctx);
 		*next_blockp = NULL;
 		return -1;
@@ -1335,20 +1323,16 @@ static int btrfsic_create_link_to_next_block(
 			       "Referenced block @%llu (%s/%llu/%d)"
 			       " found in hash table, %c,"
 			       " bytenr mismatch (!= stored %llu).\n",
-			       (unsigned long long)next_bytenr,
-			       next_block_ctx->dev->name,
-			       (unsigned long long)next_block_ctx->dev_bytenr,
-			       *mirror_nump,
+			       next_bytenr, next_block_ctx->dev->name,
+			       next_block_ctx->dev_bytenr, *mirror_nump,
 			       btrfsic_get_block_type(state, next_block),
-			       (unsigned long long)next_block->logical_bytenr);
+			       next_block->logical_bytenr);
 		} else if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE)
 			printk(KERN_INFO
 			       "Referenced block @%llu (%s/%llu/%d)"
 			       " found in hash table, %c.\n",
-			       (unsigned long long)next_bytenr,
-			       next_block_ctx->dev->name,
-			       (unsigned long long)next_block_ctx->dev_bytenr,
-			       *mirror_nump,
+			       next_bytenr, next_block_ctx->dev->name,
+			       next_block_ctx->dev_bytenr, *mirror_nump,
 			       btrfsic_get_block_type(state, next_block));
 		next_block->logical_bytenr = next_bytenr;
 
@@ -1400,7 +1384,7 @@ static int btrfsic_create_link_to_next_block(
 		if (ret < (int)next_block_ctx->len) {
 			printk(KERN_INFO
 			       "btrfsic: read block @logical %llu failed!\n",
-			       (unsigned long long)next_bytenr);
+			       next_bytenr);
 			btrfsic_release_block_ctx(next_block_ctx);
 			*next_blockp = NULL;
 			return -1;
@@ -1444,12 +1428,12 @@ static int btrfsic_handle_extent_data(
 		file_extent_item_offset,
 		offsetof(struct btrfs_file_extent_item, disk_num_bytes));
 	if (BTRFS_FILE_EXTENT_REG != file_extent_item.type ||
-	    ((u64)0) == le64_to_cpu(file_extent_item.disk_bytenr)) {
+	    btrfs_stack_file_extent_disk_bytenr(&file_extent_item) == 0) {
 		if (state->print_mask & BTRFSIC_PRINT_MASK_VERY_VERBOSE)
 			printk(KERN_INFO "extent_data: type %u, disk_bytenr = %llu\n",
 			       file_extent_item.type,
-			       (unsigned long long)
-			       le64_to_cpu(file_extent_item.disk_bytenr));
+			       btrfs_stack_file_extent_disk_bytenr(
+			       &file_extent_item));
 		return 0;
 	}
 
@@ -1463,20 +1447,19 @@ static int btrfsic_handle_extent_data(
 	btrfsic_read_from_block_data(block_ctx, &file_extent_item,
 				     file_extent_item_offset,
 				     sizeof(struct btrfs_file_extent_item));
-	next_bytenr = le64_to_cpu(file_extent_item.disk_bytenr) +
-		      le64_to_cpu(file_extent_item.offset);
-	generation = le64_to_cpu(file_extent_item.generation);
-	num_bytes = le64_to_cpu(file_extent_item.num_bytes);
-	generation = le64_to_cpu(file_extent_item.generation);
+	next_bytenr = btrfs_stack_file_extent_disk_bytenr(&file_extent_item) +
+		      btrfs_stack_file_extent_offset(&file_extent_item);
+	generation = btrfs_stack_file_extent_generation(&file_extent_item);
+	num_bytes = btrfs_stack_file_extent_num_bytes(&file_extent_item);
+	generation = btrfs_stack_file_extent_generation(&file_extent_item);
 
 	if (state->print_mask & BTRFSIC_PRINT_MASK_VERY_VERBOSE)
 		printk(KERN_INFO "extent_data: type %u, disk_bytenr = %llu,"
 		       " offset = %llu, num_bytes = %llu\n",
 		       file_extent_item.type,
-		       (unsigned long long)
-		       le64_to_cpu(file_extent_item.disk_bytenr),
-		       (unsigned long long)le64_to_cpu(file_extent_item.offset),
-		       (unsigned long long)num_bytes);
+		       btrfs_stack_file_extent_disk_bytenr(&file_extent_item),
+		       btrfs_stack_file_extent_offset(&file_extent_item),
+		       num_bytes);
 	while (num_bytes > 0) {
 		u32 chunk_len;
 		int num_copies;
@@ -1492,7 +1475,7 @@ static int btrfsic_handle_extent_data(
 				     next_bytenr, state->datablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
-			       (unsigned long long)next_bytenr, num_copies);
+			       next_bytenr, num_copies);
 		for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) {
 			struct btrfsic_block_data_ctx next_block_ctx;
 			struct btrfsic_block *next_block;
@@ -1504,8 +1487,7 @@ static int btrfsic_handle_extent_data(
 			if (state->print_mask & BTRFSIC_PRINT_MASK_VERY_VERBOSE)
 				printk(KERN_INFO
 				       "\tdisk_bytenr = %llu, num_bytes %u\n",
-				       (unsigned long long)next_bytenr,
-				       chunk_len);
+				       next_bytenr, chunk_len);
 			ret = btrfsic_map_block(state, next_bytenr,
 						chunk_len, &next_block_ctx,
 						mirror_num);
@@ -1513,8 +1495,7 @@ static int btrfsic_handle_extent_data(
 				printk(KERN_INFO
 				       "btrfsic: btrfsic_map_block(@%llu,"
 				       " mirror=%d) failed!\n",
-				       (unsigned long long)next_bytenr,
-				       mirror_num);
+				       next_bytenr, mirror_num);
 				return -1;
 			}
 
@@ -1543,12 +1524,10 @@ static int btrfsic_handle_extent_data(
 					       " found in hash table, D,"
 					       " bytenr mismatch"
 					       " (!= stored %llu).\n",
-					       (unsigned long long)next_bytenr,
+					       next_bytenr,
 					       next_block_ctx.dev->name,
-					       (unsigned long long)
 					       next_block_ctx.dev_bytenr,
 					       mirror_num,
-					       (unsigned long long)
 					       next_block->logical_bytenr);
 				}
 				next_block->logical_bytenr = next_bytenr;
@@ -1675,7 +1654,7 @@ static int btrfsic_read_block(struct btrfsic_state *state,
 	if (block_ctx->dev_bytenr & ((u64)PAGE_CACHE_SIZE - 1)) {
 		printk(KERN_INFO
 		       "btrfsic: read_block() with unaligned bytenr %llu\n",
-		       (unsigned long long)block_ctx->dev_bytenr);
+		       block_ctx->dev_bytenr);
 		return -1;
 	}
 
@@ -1772,10 +1751,8 @@ static void btrfsic_dump_database(struct btrfsic_state *state)
 
 		printk(KERN_INFO "%c-block @%llu (%s/%llu/%d)\n",
 		       btrfsic_get_block_type(state, b_all),
-		       (unsigned long long)b_all->logical_bytenr,
-		       b_all->dev_state->name,
-		       (unsigned long long)b_all->dev_bytenr,
-		       b_all->mirror_num);
+		       b_all->logical_bytenr, b_all->dev_state->name,
+		       b_all->dev_bytenr, b_all->mirror_num);
 
 		list_for_each(elem_ref_to, &b_all->ref_to_list) {
 			const struct btrfsic_block_link *const l =
@@ -1787,16 +1764,13 @@ static void btrfsic_dump_database(struct btrfsic_state *state)
 			       " refers %u* to"
 			       " %c @%llu (%s/%llu/%d)\n",
 			       btrfsic_get_block_type(state, b_all),
-			       (unsigned long long)b_all->logical_bytenr,
-			       b_all->dev_state->name,
-			       (unsigned long long)b_all->dev_bytenr,
-			       b_all->mirror_num,
+			       b_all->logical_bytenr, b_all->dev_state->name,
+			       b_all->dev_bytenr, b_all->mirror_num,
 			       l->ref_cnt,
 			       btrfsic_get_block_type(state, l->block_ref_to),
-			       (unsigned long long)
 			       l->block_ref_to->logical_bytenr,
 			       l->block_ref_to->dev_state->name,
-			       (unsigned long long)l->block_ref_to->dev_bytenr,
+			       l->block_ref_to->dev_bytenr,
 			       l->block_ref_to->mirror_num);
 		}
 
@@ -1810,16 +1784,12 @@ static void btrfsic_dump_database(struct btrfsic_state *state)
 			       " is ref %u* from"
 			       " %c @%llu (%s/%llu/%d)\n",
 			       btrfsic_get_block_type(state, b_all),
-			       (unsigned long long)b_all->logical_bytenr,
-			       b_all->dev_state->name,
-			       (unsigned long long)b_all->dev_bytenr,
-			       b_all->mirror_num,
+			       b_all->logical_bytenr, b_all->dev_state->name,
+			       b_all->dev_bytenr, b_all->mirror_num,
 			       l->ref_cnt,
 			       btrfsic_get_block_type(state, l->block_ref_from),
-			       (unsigned long long)
 			       l->block_ref_from->logical_bytenr,
 			       l->block_ref_from->dev_state->name,
-			       (unsigned long long)
 			       l->block_ref_from->dev_bytenr,
 			       l->block_ref_from->mirror_num);
 		}
@@ -1896,8 +1866,8 @@ again:
 		struct list_head *tmp_ref_to;
 
 		if (block->is_superblock) {
-			bytenr = le64_to_cpu(((struct btrfs_super_block *)
-					      mapped_datav[0])->bytenr);
+			bytenr = btrfs_super_bytenr((struct btrfs_super_block *)
+						    mapped_datav[0]);
 			if (num_pages * PAGE_CACHE_SIZE <
 			    BTRFS_SUPER_INFO_SIZE) {
 				printk(KERN_INFO
@@ -1923,8 +1893,9 @@ again:
 					return;
 				}
 				processed_len = state->metablock_size;
-				bytenr = le64_to_cpu(((struct btrfs_header *)
-						      mapped_datav[0])->bytenr);
+				bytenr = btrfs_stack_header_bytenr(
+						(struct btrfs_header *)
+						mapped_datav[0]);
 				btrfsic_cmp_log_and_dev_bytenr(state, bytenr,
 							       dev_state,
 							       dev_bytenr);
@@ -1935,12 +1906,9 @@ again:
 				       " found in hash table, %c,"
 				       " bytenr mismatch"
 				       " (!= stored %llu).\n",
-				       (unsigned long long)bytenr,
-				       dev_state->name,
-				       (unsigned long long)dev_bytenr,
+				       bytenr, dev_state->name, dev_bytenr,
 				       block->mirror_num,
 				       btrfsic_get_block_type(state, block),
-				       (unsigned long long)
 				       block->logical_bytenr);
 				block->logical_bytenr = bytenr;
 			} else if (state->print_mask &
@@ -1948,9 +1916,7 @@ again:
 				printk(KERN_INFO
 				       "Written block @%llu (%s/%llu/%d)"
 				       " found in hash table, %c.\n",
-				       (unsigned long long)bytenr,
-				       dev_state->name,
-				       (unsigned long long)dev_bytenr,
+				       bytenr, dev_state->name, dev_bytenr,
 				       block->mirror_num,
 				       btrfsic_get_block_type(state, block));
 		} else {
@@ -1966,9 +1932,7 @@ again:
 				printk(KERN_INFO
 				       "Written block @%llu (%s/%llu/%d)"
 				       " found in hash table, %c.\n",
-				       (unsigned long long)bytenr,
-				       dev_state->name,
-				       (unsigned long long)dev_bytenr,
+				       bytenr, dev_state->name, dev_bytenr,
 				       block->mirror_num,
 				       btrfsic_get_block_type(state, block));
 		}
@@ -1985,21 +1949,14 @@ again:
 			       " new(gen=%llu),"
 			       " which is referenced by most recent superblock"
 			       " (superblockgen=%llu)!\n",
-			       btrfsic_get_block_type(state, block),
-			       (unsigned long long)bytenr,
-			       dev_state->name,
-			       (unsigned long long)dev_bytenr,
-			       block->mirror_num,
-			       (unsigned long long)block->generation,
-			       (unsigned long long)
-			       le64_to_cpu(block->disk_key.objectid),
+			       btrfsic_get_block_type(state, block), bytenr,
+			       dev_state->name, dev_bytenr, block->mirror_num,
+			       block->generation,
+			       btrfs_disk_key_objectid(&block->disk_key),
 			       block->disk_key.type,
-			       (unsigned long long)
-			       le64_to_cpu(block->disk_key.offset),
-			       (unsigned long long)
-			       le64_to_cpu(((struct btrfs_header *)
-					    mapped_datav[0])->generation),
-			       (unsigned long long)
+			       btrfs_disk_key_offset(&block->disk_key),
+			       btrfs_stack_header_generation(
+				       (struct btrfs_header *) mapped_datav[0]),
 			       state->max_superblock_generation);
 			btrfsic_dump_tree(state);
 		}
@@ -2008,15 +1965,12 @@ again:
 			printk(KERN_INFO "btrfs: attempt to overwrite %c-block"
 			       " @%llu (%s/%llu/%d), oldgen=%llu, newgen=%llu,"
 			       " which is not yet iodone!\n",
-			       btrfsic_get_block_type(state, block),
-			       (unsigned long long)bytenr,
-			       dev_state->name,
-			       (unsigned long long)dev_bytenr,
-			       block->mirror_num,
-			       (unsigned long long)block->generation,
-			       (unsigned long long)
-			       le64_to_cpu(((struct btrfs_header *)
-					    mapped_datav[0])->generation));
+			       btrfsic_get_block_type(state, block), bytenr,
+			       dev_state->name, dev_bytenr, block->mirror_num,
+			       block->generation,
+			       btrfs_stack_header_generation(
+				       (struct btrfs_header *)
+				       mapped_datav[0]));
 			/* it would not be safe to go on */
 			btrfsic_dump_tree(state);
 			goto continue_loop;
@@ -2056,7 +2010,7 @@ again:
 		if (ret) {
 			printk(KERN_INFO
 			       "btrfsic: btrfsic_map_block(root @%llu)"
-			       " failed!\n", (unsigned long long)bytenr);
+			       " failed!\n", bytenr);
 			goto continue_loop;
 		}
 		block_ctx.datav = mapped_datav;
@@ -2140,7 +2094,7 @@ again:
 				printk(KERN_INFO
 				       "btrfsic: btrfsic_process_metablock"
 				       "(root @%llu) failed!\n",
-				       (unsigned long long)dev_bytenr);
+				       dev_bytenr);
 		} else {
 			block->is_metadata = 0;
 			block->mirror_num = 0;	/* unknown */
@@ -2168,8 +2122,7 @@ again:
 			if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE)
 				printk(KERN_INFO "Written block (%s/%llu/?)"
 				       " !found in hash table, D.\n",
-				       dev_state->name,
-				       (unsigned long long)dev_bytenr);
+				       dev_state->name, dev_bytenr);
 			if (!state->include_extent_data) {
 				/* ignore that written D block */
 				goto continue_loop;
@@ -2184,17 +2137,16 @@ again:
 			block_ctx.pagev = NULL;
 		} else {
 			processed_len = state->metablock_size;
-			bytenr = le64_to_cpu(((struct btrfs_header *)
-					      mapped_datav[0])->bytenr);
+			bytenr = btrfs_stack_header_bytenr(
+					(struct btrfs_header *)
+					mapped_datav[0]);
 			btrfsic_cmp_log_and_dev_bytenr(state, bytenr, dev_state,
 						       dev_bytenr);
 			if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE)
 				printk(KERN_INFO
 				       "Written block @%llu (%s/%llu/?)"
 				       " !found in hash table, M.\n",
-				       (unsigned long long)bytenr,
-				       dev_state->name,
-				       (unsigned long long)dev_bytenr);
+				       bytenr, dev_state->name, dev_bytenr);
 
 			ret = btrfsic_map_block(state, bytenr, processed_len,
 						&block_ctx, 0);
@@ -2202,7 +2154,7 @@ again:
 				printk(KERN_INFO
 				       "btrfsic: btrfsic_map_block(root @%llu)"
 				       " failed!\n",
-				       (unsigned long long)dev_bytenr);
+				       dev_bytenr);
 				goto continue_loop;
 			}
 		}
@@ -2267,10 +2219,8 @@ again:
 			printk(KERN_INFO
 			       "New written %c-block @%llu (%s/%llu/%d)\n",
 			       is_metadata ? 'M' : 'D',
-			       (unsigned long long)block->logical_bytenr,
-			       block->dev_state->name,
-			       (unsigned long long)block->dev_bytenr,
-			       block->mirror_num);
+			       block->logical_bytenr, block->dev_state->name,
+			       block->dev_bytenr, block->mirror_num);
 		list_add(&block->all_blocks_node, &state->all_blocks_list);
 		btrfsic_block_hashtable_add(block, &state->block_hashtable);
 
@@ -2281,7 +2231,7 @@ again:
 				printk(KERN_INFO
 				       "btrfsic: process_metablock(root @%llu)"
 				       " failed!\n",
-				       (unsigned long long)dev_bytenr);
+				       dev_bytenr);
 		}
 		btrfsic_release_block_ctx(&block_ctx);
 	}
@@ -2319,10 +2269,8 @@ static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status)
 			       "bio_end_io(err=%d) for %c @%llu (%s/%llu/%d)\n",
 			       bio_error_status,
 			       btrfsic_get_block_type(dev_state->state, block),
-			       (unsigned long long)block->logical_bytenr,
-			       dev_state->name,
-			       (unsigned long long)block->dev_bytenr,
-			       block->mirror_num);
+			       block->logical_bytenr, dev_state->name,
+			       block->dev_bytenr, block->mirror_num);
 		next_block = block->next_in_same_bio;
 		block->iodone_w_error = iodone_w_error;
 		if (block->submit_bio_bh_rw & REQ_FLUSH) {
@@ -2332,7 +2280,6 @@ static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status)
 				printk(KERN_INFO
 				       "bio_end_io() new %s flush_gen=%llu\n",
 				       dev_state->name,
-				       (unsigned long long)
 				       dev_state->last_flush_gen);
 		}
 		if (block->submit_bio_bh_rw & REQ_FUA)
@@ -2358,10 +2305,8 @@ static void btrfsic_bh_end_io(struct buffer_head *bh, int uptodate)
 		       "bh_end_io(error=%d) for %c @%llu (%s/%llu/%d)\n",
 		       iodone_w_error,
 		       btrfsic_get_block_type(dev_state->state, block),
-		       (unsigned long long)block->logical_bytenr,
-		       block->dev_state->name,
-		       (unsigned long long)block->dev_bytenr,
-		       block->mirror_num);
+		       block->logical_bytenr, block->dev_state->name,
+		       block->dev_bytenr, block->mirror_num);
 
 	block->iodone_w_error = iodone_w_error;
 	if (block->submit_bio_bh_rw & REQ_FLUSH) {
@@ -2370,8 +2315,7 @@ static void btrfsic_bh_end_io(struct buffer_head *bh, int uptodate)
 		     BTRFSIC_PRINT_MASK_END_IO_BIO_BH))
 			printk(KERN_INFO
 			       "bh_end_io() new %s flush_gen=%llu\n",
-			       dev_state->name,
-			       (unsigned long long)dev_state->last_flush_gen);
+			       dev_state->name, dev_state->last_flush_gen);
 	}
 	if (block->submit_bio_bh_rw & REQ_FUA)
 		block->flush_gen = 0; /* FUA completed means block is on disk */
@@ -2396,26 +2340,20 @@ static int btrfsic_process_written_superblock(
 			printk(KERN_INFO
 			       "btrfsic: superblock @%llu (%s/%llu/%d)"
 			       " with old gen %llu <= %llu\n",
-			       (unsigned long long)superblock->logical_bytenr,
+			       superblock->logical_bytenr,
 			       superblock->dev_state->name,
-			       (unsigned long long)superblock->dev_bytenr,
-			       superblock->mirror_num,
-			       (unsigned long long)
+			       superblock->dev_bytenr, superblock->mirror_num,
 			       btrfs_super_generation(super_hdr),
-			       (unsigned long long)
 			       state->max_superblock_generation);
 	} else {
 		if (state->print_mask & BTRFSIC_PRINT_MASK_SUPERBLOCK_WRITE)
 			printk(KERN_INFO
 			       "btrfsic: got new superblock @%llu (%s/%llu/%d)"
 			       " with new gen %llu > %llu\n",
-			       (unsigned long long)superblock->logical_bytenr,
+			       superblock->logical_bytenr,
 			       superblock->dev_state->name,
-			       (unsigned long long)superblock->dev_bytenr,
-			       superblock->mirror_num,
-			       (unsigned long long)
+			       superblock->dev_bytenr, superblock->mirror_num,
 			       btrfs_super_generation(super_hdr),
-			       (unsigned long long)
 			       state->max_superblock_generation);
 
 		state->max_superblock_generation =
@@ -2432,43 +2370,41 @@ static int btrfsic_process_written_superblock(
 		int num_copies;
 		int mirror_num;
 		const char *additional_string = NULL;
-		struct btrfs_disk_key tmp_disk_key;
+		struct btrfs_disk_key tmp_disk_key = {0};
 
-		tmp_disk_key.type = BTRFS_ROOT_ITEM_KEY;
-		tmp_disk_key.offset = 0;
+		btrfs_set_disk_key_objectid(&tmp_disk_key,
+					    BTRFS_ROOT_ITEM_KEY);
+		btrfs_set_disk_key_objectid(&tmp_disk_key, 0);
 
 		switch (pass) {
 		case 0:
-			tmp_disk_key.objectid =
-			    cpu_to_le64(BTRFS_ROOT_TREE_OBJECTID);
+			btrfs_set_disk_key_objectid(&tmp_disk_key,
+						    BTRFS_ROOT_TREE_OBJECTID);
 			additional_string = "root ";
 			next_bytenr = btrfs_super_root(super_hdr);
 			if (state->print_mask &
 			    BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION)
-				printk(KERN_INFO "root@%llu\n",
-				       (unsigned long long)next_bytenr);
+				printk(KERN_INFO "root@%llu\n", next_bytenr);
 			break;
 		case 1:
-			tmp_disk_key.objectid =
-			    cpu_to_le64(BTRFS_CHUNK_TREE_OBJECTID);
+			btrfs_set_disk_key_objectid(&tmp_disk_key,
+						    BTRFS_CHUNK_TREE_OBJECTID);
 			additional_string = "chunk ";
 			next_bytenr = btrfs_super_chunk_root(super_hdr);
 			if (state->print_mask &
 			    BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION)
-				printk(KERN_INFO "chunk@%llu\n",
-				       (unsigned long long)next_bytenr);
+				printk(KERN_INFO "chunk@%llu\n", next_bytenr);
 			break;
 		case 2:
-			tmp_disk_key.objectid =
-			    cpu_to_le64(BTRFS_TREE_LOG_OBJECTID);
+			btrfs_set_disk_key_objectid(&tmp_disk_key,
+						    BTRFS_TREE_LOG_OBJECTID);
 			additional_string = "log ";
 			next_bytenr = btrfs_super_log_root(super_hdr);
 			if (0 == next_bytenr)
 				continue;
 			if (state->print_mask &
 			    BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION)
-				printk(KERN_INFO "log@%llu\n",
-				       (unsigned long long)next_bytenr);
+				printk(KERN_INFO "log@%llu\n", next_bytenr);
 			break;
 		}
 
@@ -2477,7 +2413,7 @@ static int btrfsic_process_written_superblock(
 				     next_bytenr, BTRFS_SUPER_INFO_SIZE);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
-			       (unsigned long long)next_bytenr, num_copies);
+			       next_bytenr, num_copies);
 		for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) {
 			int was_created;
 
@@ -2493,8 +2429,7 @@ static int btrfsic_process_written_superblock(
 				printk(KERN_INFO
 				       "btrfsic: btrfsic_map_block(@%llu,"
 				       " mirror=%d) failed!\n",
-				       (unsigned long long)next_bytenr,
-				       mirror_num);
+				       next_bytenr, mirror_num);
 				return -1;
 			}
 
@@ -2579,26 +2514,22 @@ static int btrfsic_check_all_ref_blocks(struct btrfsic_state *state,
 			       " %u* refers to %c @%llu (%s/%llu/%d)\n",
 			       recursion_level,
 			       btrfsic_get_block_type(state, block),
-			       (unsigned long long)block->logical_bytenr,
-			       block->dev_state->name,
-			       (unsigned long long)block->dev_bytenr,
-			       block->mirror_num,
+			       block->logical_bytenr, block->dev_state->name,
+			       block->dev_bytenr, block->mirror_num,
 			       l->ref_cnt,
 			       btrfsic_get_block_type(state, l->block_ref_to),
-			       (unsigned long long)
 			       l->block_ref_to->logical_bytenr,
 			       l->block_ref_to->dev_state->name,
-			       (unsigned long long)l->block_ref_to->dev_bytenr,
+			       l->block_ref_to->dev_bytenr,
 			       l->block_ref_to->mirror_num);
 		if (l->block_ref_to->never_written) {
 			printk(KERN_INFO "btrfs: attempt to write superblock"
 			       " which references block %c @%llu (%s/%llu/%d)"
 			       " which is never written!\n",
 			       btrfsic_get_block_type(state, l->block_ref_to),
-			       (unsigned long long)
 			       l->block_ref_to->logical_bytenr,
 			       l->block_ref_to->dev_state->name,
-			       (unsigned long long)l->block_ref_to->dev_bytenr,
+			       l->block_ref_to->dev_bytenr,
 			       l->block_ref_to->mirror_num);
 			ret = -1;
 		} else if (!l->block_ref_to->is_iodone) {
@@ -2606,10 +2537,9 @@ static int btrfsic_check_all_ref_blocks(struct btrfsic_state *state,
 			       " which references block %c @%llu (%s/%llu/%d)"
 			       " which is not yet iodone!\n",
 			       btrfsic_get_block_type(state, l->block_ref_to),
-			       (unsigned long long)
 			       l->block_ref_to->logical_bytenr,
 			       l->block_ref_to->dev_state->name,
-			       (unsigned long long)l->block_ref_to->dev_bytenr,
+			       l->block_ref_to->dev_bytenr,
 			       l->block_ref_to->mirror_num);
 			ret = -1;
 		} else if (l->block_ref_to->iodone_w_error) {
@@ -2617,10 +2547,9 @@ static int btrfsic_check_all_ref_blocks(struct btrfsic_state *state,
 			       " which references block %c @%llu (%s/%llu/%d)"
 			       " which has write error!\n",
 			       btrfsic_get_block_type(state, l->block_ref_to),
-			       (unsigned long long)
 			       l->block_ref_to->logical_bytenr,
 			       l->block_ref_to->dev_state->name,
-			       (unsigned long long)l->block_ref_to->dev_bytenr,
+			       l->block_ref_to->dev_bytenr,
 			       l->block_ref_to->mirror_num);
 			ret = -1;
 		} else if (l->parent_generation !=
@@ -2634,13 +2563,12 @@ static int btrfsic_check_all_ref_blocks(struct btrfsic_state *state,
 			       " with generation %llu !="
 			       " parent generation %llu!\n",
 			       btrfsic_get_block_type(state, l->block_ref_to),
-			       (unsigned long long)
 			       l->block_ref_to->logical_bytenr,
 			       l->block_ref_to->dev_state->name,
-			       (unsigned long long)l->block_ref_to->dev_bytenr,
+			       l->block_ref_to->dev_bytenr,
 			       l->block_ref_to->mirror_num,
-			       (unsigned long long)l->block_ref_to->generation,
-			       (unsigned long long)l->parent_generation);
+			       l->block_ref_to->generation,
+			       l->parent_generation);
 			ret = -1;
 		} else if (l->block_ref_to->flush_gen >
 			   l->block_ref_to->dev_state->last_flush_gen) {
@@ -2650,13 +2578,10 @@ static int btrfsic_check_all_ref_blocks(struct btrfsic_state *state,
 			       " (block flush_gen=%llu,"
 			       " dev->flush_gen=%llu)!\n",
 			       btrfsic_get_block_type(state, l->block_ref_to),
-			       (unsigned long long)
 			       l->block_ref_to->logical_bytenr,
 			       l->block_ref_to->dev_state->name,
-			       (unsigned long long)l->block_ref_to->dev_bytenr,
-			       l->block_ref_to->mirror_num,
-			       (unsigned long long)block->flush_gen,
-			       (unsigned long long)
+			       l->block_ref_to->dev_bytenr,
+			       l->block_ref_to->mirror_num, block->flush_gen,
 			       l->block_ref_to->dev_state->last_flush_gen);
 			ret = -1;
 		} else if (-1 == btrfsic_check_all_ref_blocks(state,
@@ -2701,16 +2626,12 @@ static int btrfsic_is_block_ref_by_superblock(
 			       " is ref %u* from %c @%llu (%s/%llu/%d)\n",
 			       recursion_level,
 			       btrfsic_get_block_type(state, block),
-			       (unsigned long long)block->logical_bytenr,
-			       block->dev_state->name,
-			       (unsigned long long)block->dev_bytenr,
-			       block->mirror_num,
+			       block->logical_bytenr, block->dev_state->name,
+			       block->dev_bytenr, block->mirror_num,
 			       l->ref_cnt,
 			       btrfsic_get_block_type(state, l->block_ref_from),
-			       (unsigned long long)
 			       l->block_ref_from->logical_bytenr,
 			       l->block_ref_from->dev_state->name,
-			       (unsigned long long)
 			       l->block_ref_from->dev_bytenr,
 			       l->block_ref_from->mirror_num);
 		if (l->block_ref_from->is_superblock &&
@@ -2737,14 +2658,12 @@ static void btrfsic_print_add_link(const struct btrfsic_state *state,
 	       " to %c @%llu (%s/%llu/%d).\n",
 	       l->ref_cnt,
 	       btrfsic_get_block_type(state, l->block_ref_from),
-	       (unsigned long long)l->block_ref_from->logical_bytenr,
+	       l->block_ref_from->logical_bytenr,
 	       l->block_ref_from->dev_state->name,
-	       (unsigned long long)l->block_ref_from->dev_bytenr,
-	       l->block_ref_from->mirror_num,
+	       l->block_ref_from->dev_bytenr, l->block_ref_from->mirror_num,
 	       btrfsic_get_block_type(state, l->block_ref_to),
-	       (unsigned long long)l->block_ref_to->logical_bytenr,
-	       l->block_ref_to->dev_state->name,
-	       (unsigned long long)l->block_ref_to->dev_bytenr,
+	       l->block_ref_to->logical_bytenr,
+	       l->block_ref_to->dev_state->name, l->block_ref_to->dev_bytenr,
 	       l->block_ref_to->mirror_num);
 }
 
@@ -2756,14 +2675,12 @@ static void btrfsic_print_rem_link(const struct btrfsic_state *state,
 	       " to %c @%llu (%s/%llu/%d).\n",
 	       l->ref_cnt,
 	       btrfsic_get_block_type(state, l->block_ref_from),
-	       (unsigned long long)l->block_ref_from->logical_bytenr,
+	       l->block_ref_from->logical_bytenr,
 	       l->block_ref_from->dev_state->name,
-	       (unsigned long long)l->block_ref_from->dev_bytenr,
-	       l->block_ref_from->mirror_num,
+	       l->block_ref_from->dev_bytenr, l->block_ref_from->mirror_num,
 	       btrfsic_get_block_type(state, l->block_ref_to),
-	       (unsigned long long)l->block_ref_to->logical_bytenr,
-	       l->block_ref_to->dev_state->name,
-	       (unsigned long long)l->block_ref_to->dev_bytenr,
+	       l->block_ref_to->logical_bytenr,
+	       l->block_ref_to->dev_state->name, l->block_ref_to->dev_bytenr,
 	       l->block_ref_to->mirror_num);
 }
 
@@ -2807,10 +2724,8 @@ static void btrfsic_dump_tree_sub(const struct btrfsic_state *state,
 	 */
 	indent_add = sprintf(buf, "%c-%llu(%s/%llu/%d)",
 			     btrfsic_get_block_type(state, block),
-			     (unsigned long long)block->logical_bytenr,
-			     block->dev_state->name,
-			     (unsigned long long)block->dev_bytenr,
-			     block->mirror_num);
+			     block->logical_bytenr, block->dev_state->name,
+			     block->dev_bytenr, block->mirror_num);
 	if (indent_level + indent_add > BTRFSIC_TREE_DUMP_MAX_INDENT_LEVEL) {
 		printk("[...]\n");
 		return;
@@ -2943,10 +2858,8 @@ static struct btrfsic_block *btrfsic_block_lookup_or_add(
 			       "New %s%c-block @%llu (%s/%llu/%d)\n",
 			       additional_string,
 			       btrfsic_get_block_type(state, block),
-			       (unsigned long long)block->logical_bytenr,
-			       dev_state->name,
-			       (unsigned long long)block->dev_bytenr,
-			       mirror_num);
+			       block->logical_bytenr, dev_state->name,
+			       block->dev_bytenr, mirror_num);
 		list_add(&block->all_blocks_node, &state->all_blocks_list);
 		btrfsic_block_hashtable_add(block, &state->block_hashtable);
 		if (NULL != was_created)
@@ -2980,7 +2893,7 @@ static void btrfsic_cmp_log_and_dev_bytenr(struct btrfsic_state *state,
 			printk(KERN_INFO "btrfsic:"
 			       " btrfsic_map_block(logical @%llu,"
 			       " mirror %d) failed!\n",
-			       (unsigned long long)bytenr, mirror_num);
+			       bytenr, mirror_num);
 			continue;
 		}
 
@@ -2997,8 +2910,7 @@ static void btrfsic_cmp_log_and_dev_bytenr(struct btrfsic_state *state,
 		printk(KERN_INFO "btrfs: attempt to write M-block which contains logical bytenr that doesn't map to dev+physical bytenr of submit_bio,"
 		       " buffer->log_bytenr=%llu, submit_bio(bdev=%s,"
 		       " phys_bytenr=%llu)!\n",
-		       (unsigned long long)bytenr, dev_state->name,
-		       (unsigned long long)dev_bytenr);
+		       bytenr, dev_state->name, dev_bytenr);
 		for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) {
 			ret = btrfsic_map_block(state, bytenr,
 						state->metablock_size,
@@ -3008,10 +2920,8 @@ static void btrfsic_cmp_log_and_dev_bytenr(struct btrfsic_state *state,
 
 			printk(KERN_INFO "Read logical bytenr @%llu maps to"
 			       " (%s/%llu/%d)\n",
-			       (unsigned long long)bytenr,
-			       block_ctx.dev->name,
-			       (unsigned long long)block_ctx.dev_bytenr,
-			       mirror_num);
+			       bytenr, block_ctx.dev->name,
+			       block_ctx.dev_bytenr, mirror_num);
 		}
 		WARN_ON(1);
 	}
@@ -3048,12 +2958,10 @@ int btrfsic_submit_bh(int rw, struct buffer_head *bh)
 		if (dev_state->state->print_mask &
 		    BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH)
 			printk(KERN_INFO
-			       "submit_bh(rw=0x%x, blocknr=%lu (bytenr %llu),"
-			       " size=%lu, data=%p, bdev=%p)\n",
-			       rw, (unsigned long)bh->b_blocknr,
-			       (unsigned long long)dev_bytenr,
-			       (unsigned long)bh->b_size, bh->b_data,
-			       bh->b_bdev);
+			       "submit_bh(rw=0x%x, blocknr=%llu (bytenr %llu),"
+			       " size=%zu, data=%p, bdev=%p)\n",
+			       rw, (unsigned long long)bh->b_blocknr,
+			       dev_bytenr, bh->b_size, bh->b_data, bh->b_bdev);
 		btrfsic_process_written_block(dev_state, dev_bytenr,
 					      &bh->b_data, 1, NULL,
 					      NULL, bh, rw);
@@ -3118,9 +3026,9 @@ void btrfsic_submit_bio(int rw, struct bio *bio)
 		    BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH)
 			printk(KERN_INFO
 			       "submit_bio(rw=0x%x, bi_vcnt=%u,"
-			       " bi_sector=%lu (bytenr %llu), bi_bdev=%p)\n",
-			       rw, bio->bi_vcnt, (unsigned long)bio->bi_sector,
-			       (unsigned long long)dev_bytenr,
+			       " bi_sector=%llu (bytenr %llu), bi_bdev=%p)\n",
+			       rw, bio->bi_vcnt,
+			       (unsigned long long)bio->bi_sector, dev_bytenr,
 			       bio->bi_bdev);
 
 		mapped_datav = kmalloc(sizeof(*mapped_datav) * bio->bi_vcnt,
@@ -3213,19 +3121,19 @@ int btrfsic_mount(struct btrfs_root *root,
 	if (root->nodesize & ((u64)PAGE_CACHE_SIZE - 1)) {
 		printk(KERN_INFO
 		       "btrfsic: cannot handle nodesize %d not being a multiple of PAGE_CACHE_SIZE %ld!\n",
-		       root->nodesize, (unsigned long)PAGE_CACHE_SIZE);
+		       root->nodesize, PAGE_CACHE_SIZE);
 		return -1;
 	}
 	if (root->leafsize & ((u64)PAGE_CACHE_SIZE - 1)) {
 		printk(KERN_INFO
 		       "btrfsic: cannot handle leafsize %d not being a multiple of PAGE_CACHE_SIZE %ld!\n",
-		       root->leafsize, (unsigned long)PAGE_CACHE_SIZE);
+		       root->leafsize, PAGE_CACHE_SIZE);
 		return -1;
 	}
 	if (root->sectorsize & ((u64)PAGE_CACHE_SIZE - 1)) {
 		printk(KERN_INFO
 		       "btrfsic: cannot handle sectorsize %d not being a multiple of PAGE_CACHE_SIZE %ld!\n",
-		       root->sectorsize, (unsigned long)PAGE_CACHE_SIZE);
+		       root->sectorsize, PAGE_CACHE_SIZE);
 		return -1;
 	}
 	state = kzalloc(sizeof(*state), GFP_NOFS);
@@ -3369,10 +3277,8 @@ void btrfsic_unmount(struct btrfs_root *root,
 			       " @%llu (%s/%llu/%d) on umount which is"
 			       " not yet iodone!\n",
 			       btrfsic_get_block_type(state, b_all),
-			       (unsigned long long)b_all->logical_bytenr,
-			       b_all->dev_state->name,
-			       (unsigned long long)b_all->dev_bytenr,
-			       b_all->mirror_num);
+			       b_all->logical_bytenr, b_all->dev_state->name,
+			       b_all->dev_bytenr, b_all->mirror_num);
 	}
 
 	mutex_unlock(&btrfsic_mutex);
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index b189bd1e7a3e..6aad98cb343f 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -132,9 +132,8 @@ static int check_compressed_csum(struct inode *inode,
 			printk(KERN_INFO "btrfs csum failed ino %llu "
 			       "extent %llu csum %u "
 			       "wanted %u mirror %d\n",
-			       (unsigned long long)btrfs_ino(inode),
-			       (unsigned long long)disk_start,
-			       csum, *cb_sum, cb->mirror_num);
+			       btrfs_ino(inode), disk_start, csum, *cb_sum,
+			       cb->mirror_num);
 			ret = -EIO;
 			goto fail;
 		}
@@ -639,7 +638,11 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	faili = nr_pages - 1;
 	cb->nr_pages = nr_pages;
 
-	add_ra_bio_pages(inode, em_start + em_len, cb);
+	/* In the parent-locked case, we only locked the range we are
+	 * interested in.  In all other cases, we can opportunistically
+	 * cache decompressed data that goes beyond the requested range. */
+	if (!(bio_flags & EXTENT_BIO_PARENT_LOCKED))
+		add_ra_bio_pages(inode, em_start + em_len, cb);
 
 	/* include any pages we added in add_ra-bio_pages */
 	uncompressed_len = bio->bi_vcnt * PAGE_CACHE_SIZE;
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 02fae7f7e42c..61b5bcd57b7e 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -274,8 +274,7 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans,
 	else
 		btrfs_set_header_owner(cow, new_root_objectid);
 
-	write_extent_buffer(cow, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(cow),
+	write_extent_buffer(cow, root->fs_info->fsid, btrfs_header_fsid(cow),
 			    BTRFS_FSID_SIZE);
 
 	WARN_ON(btrfs_header_generation(buf) > trans->transid);
@@ -484,8 +483,27 @@ __tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct tree_mod_elem *tm)
 	struct rb_node **new;
 	struct rb_node *parent = NULL;
 	struct tree_mod_elem *cur;
+	int ret = 0;
 
-	BUG_ON(!tm || !tm->seq);
+	BUG_ON(!tm);
+
+	tree_mod_log_write_lock(fs_info);
+	if (list_empty(&fs_info->tree_mod_seq_list)) {
+		tree_mod_log_write_unlock(fs_info);
+		/*
+		 * Ok we no longer care about logging modifications, free up tm
+		 * and return 0.  Any callers shouldn't be using tm after
+		 * calling tree_mod_log_insert, but if they do we can just
+		 * change this to return a special error code to let the callers
+		 * do their own thing.
+		 */
+		kfree(tm);
+		return 0;
+	}
+
+	spin_lock(&fs_info->tree_mod_seq_lock);
+	tm->seq = btrfs_inc_tree_mod_seq_minor(fs_info);
+	spin_unlock(&fs_info->tree_mod_seq_lock);
 
 	tm_root = &fs_info->tree_mod_log;
 	new = &tm_root->rb_node;
@@ -501,14 +519,17 @@ __tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct tree_mod_elem *tm)
 		else if (cur->seq > tm->seq)
 			new = &((*new)->rb_right);
 		else {
+			ret = -EEXIST;
 			kfree(tm);
-			return -EEXIST;
+			goto out;
 		}
 	}
 
 	rb_link_node(&tm->node, parent, new);
 	rb_insert_color(&tm->node, tm_root);
-	return 0;
+out:
+	tree_mod_log_write_unlock(fs_info);
+	return ret;
 }
 
 /*
@@ -524,57 +545,19 @@ static inline int tree_mod_dont_log(struct btrfs_fs_info *fs_info,
 		return 1;
 	if (eb && btrfs_header_level(eb) == 0)
 		return 1;
-
-	tree_mod_log_write_lock(fs_info);
-	if (list_empty(&fs_info->tree_mod_seq_list)) {
-		/*
-		 * someone emptied the list while we were waiting for the lock.
-		 * we must not add to the list when no blocker exists.
-		 */
-		tree_mod_log_write_unlock(fs_info);
-		return 1;
-	}
-
 	return 0;
 }
 
-/*
- * This allocates memory and gets a tree modification sequence number.
- *
- * Returns <0 on error.
- * Returns >0 (the added sequence number) on success.
- */
-static inline int tree_mod_alloc(struct btrfs_fs_info *fs_info, gfp_t flags,
-				 struct tree_mod_elem **tm_ret)
-{
-	struct tree_mod_elem *tm;
-
-	/*
-	 * once we switch from spin locks to something different, we should
-	 * honor the flags parameter here.
-	 */
-	tm = *tm_ret = kzalloc(sizeof(*tm), GFP_ATOMIC);
-	if (!tm)
-		return -ENOMEM;
-
-	spin_lock(&fs_info->tree_mod_seq_lock);
-	tm->seq = btrfs_inc_tree_mod_seq_minor(fs_info);
-	spin_unlock(&fs_info->tree_mod_seq_lock);
-
-	return tm->seq;
-}
-
 static inline int
 __tree_mod_log_insert_key(struct btrfs_fs_info *fs_info,
 			  struct extent_buffer *eb, int slot,
 			  enum mod_log_op op, gfp_t flags)
 {
-	int ret;
 	struct tree_mod_elem *tm;
 
-	ret = tree_mod_alloc(fs_info, flags, &tm);
-	if (ret < 0)
-		return ret;
+	tm = kzalloc(sizeof(*tm), flags);
+	if (!tm)
+		return -ENOMEM;
 
 	tm->index = eb->start >> PAGE_CACHE_SHIFT;
 	if (op != MOD_LOG_KEY_ADD) {
@@ -589,34 +572,14 @@ __tree_mod_log_insert_key(struct btrfs_fs_info *fs_info,
 }
 
 static noinline int
-tree_mod_log_insert_key_mask(struct btrfs_fs_info *fs_info,
-			     struct extent_buffer *eb, int slot,
-			     enum mod_log_op op, gfp_t flags)
+tree_mod_log_insert_key(struct btrfs_fs_info *fs_info,
+			struct extent_buffer *eb, int slot,
+			enum mod_log_op op, gfp_t flags)
 {
-	int ret;
-
 	if (tree_mod_dont_log(fs_info, eb))
 		return 0;
 
-	ret = __tree_mod_log_insert_key(fs_info, eb, slot, op, flags);
-
-	tree_mod_log_write_unlock(fs_info);
-	return ret;
-}
-
-static noinline int
-tree_mod_log_insert_key(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
-			int slot, enum mod_log_op op)
-{
-	return tree_mod_log_insert_key_mask(fs_info, eb, slot, op, GFP_NOFS);
-}
-
-static noinline int
-tree_mod_log_insert_key_locked(struct btrfs_fs_info *fs_info,
-			     struct extent_buffer *eb, int slot,
-			     enum mod_log_op op)
-{
-	return __tree_mod_log_insert_key(fs_info, eb, slot, op, GFP_NOFS);
+	return __tree_mod_log_insert_key(fs_info, eb, slot, op, flags);
 }
 
 static noinline int
@@ -637,14 +600,14 @@ tree_mod_log_insert_move(struct btrfs_fs_info *fs_info,
 	 * buffer, i.e. dst_slot < src_slot.
 	 */
 	for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) {
-		ret = tree_mod_log_insert_key_locked(fs_info, eb, i + dst_slot,
-					      MOD_LOG_KEY_REMOVE_WHILE_MOVING);
+		ret = __tree_mod_log_insert_key(fs_info, eb, i + dst_slot,
+				MOD_LOG_KEY_REMOVE_WHILE_MOVING, GFP_NOFS);
 		BUG_ON(ret < 0);
 	}
 
-	ret = tree_mod_alloc(fs_info, flags, &tm);
-	if (ret < 0)
-		goto out;
+	tm = kzalloc(sizeof(*tm), flags);
+	if (!tm)
+		return -ENOMEM;
 
 	tm->index = eb->start >> PAGE_CACHE_SHIFT;
 	tm->slot = src_slot;
@@ -652,10 +615,7 @@ tree_mod_log_insert_move(struct btrfs_fs_info *fs_info,
 	tm->move.nr_items = nr_items;
 	tm->op = MOD_LOG_MOVE_KEYS;
 
-	ret = __tree_mod_log_insert(fs_info, tm);
-out:
-	tree_mod_log_write_unlock(fs_info);
-	return ret;
+	return __tree_mod_log_insert(fs_info, tm);
 }
 
 static inline void
@@ -670,8 +630,8 @@ __tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
 
 	nritems = btrfs_header_nritems(eb);
 	for (i = nritems - 1; i >= 0; i--) {
-		ret = tree_mod_log_insert_key_locked(fs_info, eb, i,
-					      MOD_LOG_KEY_REMOVE_WHILE_FREEING);
+		ret = __tree_mod_log_insert_key(fs_info, eb, i,
+				MOD_LOG_KEY_REMOVE_WHILE_FREEING, GFP_NOFS);
 		BUG_ON(ret < 0);
 	}
 }
@@ -683,7 +643,6 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 			 int log_removal)
 {
 	struct tree_mod_elem *tm;
-	int ret;
 
 	if (tree_mod_dont_log(fs_info, NULL))
 		return 0;
@@ -691,9 +650,9 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 	if (log_removal)
 		__tree_mod_log_free_eb(fs_info, old_root);
 
-	ret = tree_mod_alloc(fs_info, flags, &tm);
-	if (ret < 0)
-		goto out;
+	tm = kzalloc(sizeof(*tm), flags);
+	if (!tm)
+		return -ENOMEM;
 
 	tm->index = new_root->start >> PAGE_CACHE_SHIFT;
 	tm->old_root.logical = old_root->start;
@@ -701,10 +660,7 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 	tm->generation = btrfs_header_generation(old_root);
 	tm->op = MOD_LOG_ROOT_REPLACE;
 
-	ret = __tree_mod_log_insert(fs_info, tm);
-out:
-	tree_mod_log_write_unlock(fs_info);
-	return ret;
+	return __tree_mod_log_insert(fs_info, tm);
 }
 
 static struct tree_mod_elem *
@@ -784,23 +740,20 @@ tree_mod_log_eb_copy(struct btrfs_fs_info *fs_info, struct extent_buffer *dst,
 	if (tree_mod_dont_log(fs_info, NULL))
 		return;
 
-	if (btrfs_header_level(dst) == 0 && btrfs_header_level(src) == 0) {
-		tree_mod_log_write_unlock(fs_info);
+	if (btrfs_header_level(dst) == 0 && btrfs_header_level(src) == 0)
 		return;
-	}
 
 	for (i = 0; i < nr_items; i++) {
-		ret = tree_mod_log_insert_key_locked(fs_info, src,
+		ret = __tree_mod_log_insert_key(fs_info, src,
 						i + src_offset,
-						MOD_LOG_KEY_REMOVE);
+						MOD_LOG_KEY_REMOVE, GFP_NOFS);
 		BUG_ON(ret < 0);
-		ret = tree_mod_log_insert_key_locked(fs_info, dst,
+		ret = __tree_mod_log_insert_key(fs_info, dst,
 						     i + dst_offset,
-						     MOD_LOG_KEY_ADD);
+						     MOD_LOG_KEY_ADD,
+						     GFP_NOFS);
 		BUG_ON(ret < 0);
 	}
-
-	tree_mod_log_write_unlock(fs_info);
 }
 
 static inline void
@@ -819,9 +772,9 @@ tree_mod_log_set_node_key(struct btrfs_fs_info *fs_info,
 {
 	int ret;
 
-	ret = tree_mod_log_insert_key_mask(fs_info, eb, slot,
-					   MOD_LOG_KEY_REPLACE,
-					   atomic ? GFP_ATOMIC : GFP_NOFS);
+	ret = __tree_mod_log_insert_key(fs_info, eb, slot,
+					MOD_LOG_KEY_REPLACE,
+					atomic ? GFP_ATOMIC : GFP_NOFS);
 	BUG_ON(ret < 0);
 }
 
@@ -830,10 +783,7 @@ tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
 {
 	if (tree_mod_dont_log(fs_info, eb))
 		return;
-
 	__tree_mod_log_free_eb(fs_info, eb);
-
-	tree_mod_log_write_unlock(fs_info);
 }
 
 static noinline void
@@ -1046,8 +996,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 	else
 		btrfs_set_header_owner(cow, root->root_key.objectid);
 
-	write_extent_buffer(cow, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(cow),
+	write_extent_buffer(cow, root->fs_info->fsid, btrfs_header_fsid(cow),
 			    BTRFS_FSID_SIZE);
 
 	ret = update_ref_for_cow(trans, root, buf, cow, &last_ref);
@@ -1056,8 +1005,11 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
-	if (root->ref_cows)
-		btrfs_reloc_cow_block(trans, root, buf, cow);
+	if (root->ref_cows) {
+		ret = btrfs_reloc_cow_block(trans, root, buf, cow);
+		if (ret)
+			return ret;
+	}
 
 	if (buf == root->node) {
 		WARN_ON(parent && parent != buf);
@@ -1083,13 +1035,14 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 
 		WARN_ON(trans->transid != btrfs_header_generation(parent));
 		tree_mod_log_insert_key(root->fs_info, parent, parent_slot,
-					MOD_LOG_KEY_REPLACE);
+					MOD_LOG_KEY_REPLACE, GFP_NOFS);
 		btrfs_set_node_blockptr(parent, parent_slot,
 					cow->start);
 		btrfs_set_node_ptr_generation(parent, parent_slot,
 					      trans->transid);
 		btrfs_mark_buffer_dirty(parent);
-		tree_mod_log_free_eb(root->fs_info, buf);
+		if (last_ref)
+			tree_mod_log_free_eb(root->fs_info, buf);
 		btrfs_free_tree_block(trans, root, buf, parent_start,
 				      last_ref);
 	}
@@ -1115,7 +1068,7 @@ __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info,
 	int looped = 0;
 
 	if (!time_seq)
-		return 0;
+		return NULL;
 
 	/*
 	 * the very last operation that's logged for a root is the replacement
@@ -1126,7 +1079,7 @@ __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info,
 		tm = tree_mod_log_search_oldest(fs_info, root_logical,
 						time_seq);
 		if (!looped && !tm)
-			return 0;
+			return NULL;
 		/*
 		 * if there are no tree operation for the oldest root, we simply
 		 * return it. this should only happen if that (old) root is at
@@ -1161,8 +1114,8 @@ __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info,
  * time_seq).
  */
 static void
-__tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
-		      struct tree_mod_elem *first_tm)
+__tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
+		      u64 time_seq, struct tree_mod_elem *first_tm)
 {
 	u32 n;
 	struct rb_node *next;
@@ -1172,6 +1125,7 @@ __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
 	unsigned long p_size = sizeof(struct btrfs_key_ptr);
 
 	n = btrfs_header_nritems(eb);
+	tree_mod_log_read_lock(fs_info);
 	while (tm && tm->seq >= time_seq) {
 		/*
 		 * all the operations are recorded with the operator used for
@@ -1226,6 +1180,7 @@ __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
 		if (tm->index != first_tm->index)
 			break;
 	}
+	tree_mod_log_read_unlock(fs_info);
 	btrfs_set_header_nritems(eb, n);
 }
 
@@ -1237,8 +1192,8 @@ __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
  * is freed (its refcount is decremented).
  */
 static struct extent_buffer *
-tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
-		    u64 time_seq)
+tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
+		    struct extent_buffer *eb, u64 time_seq)
 {
 	struct extent_buffer *eb_rewin;
 	struct tree_mod_elem *tm;
@@ -1253,11 +1208,18 @@ tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 	if (!tm)
 		return eb;
 
+	btrfs_set_path_blocking(path);
+	btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
+
 	if (tm->op == MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
 		BUG_ON(tm->slot != 0);
 		eb_rewin = alloc_dummy_extent_buffer(eb->start,
 						fs_info->tree_root->nodesize);
-		BUG_ON(!eb_rewin);
+		if (!eb_rewin) {
+			btrfs_tree_read_unlock_blocking(eb);
+			free_extent_buffer(eb);
+			return NULL;
+		}
 		btrfs_set_header_bytenr(eb_rewin, eb->start);
 		btrfs_set_header_backref_rev(eb_rewin,
 					     btrfs_header_backref_rev(eb));
@@ -1265,16 +1227,20 @@ tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 		btrfs_set_header_level(eb_rewin, btrfs_header_level(eb));
 	} else {
 		eb_rewin = btrfs_clone_extent_buffer(eb);
-		BUG_ON(!eb_rewin);
+		if (!eb_rewin) {
+			btrfs_tree_read_unlock_blocking(eb);
+			free_extent_buffer(eb);
+			return NULL;
+		}
 	}
 
-	extent_buffer_get(eb_rewin);
-	btrfs_tree_read_unlock(eb);
+	btrfs_clear_path_blocking(path, NULL, BTRFS_READ_LOCK);
+	btrfs_tree_read_unlock_blocking(eb);
 	free_extent_buffer(eb);
 
 	extent_buffer_get(eb_rewin);
 	btrfs_tree_read_lock(eb_rewin);
-	__tree_mod_log_rewind(eb_rewin, time_seq, tm);
+	__tree_mod_log_rewind(fs_info, eb_rewin, time_seq, tm);
 	WARN_ON(btrfs_header_nritems(eb_rewin) >
 		BTRFS_NODEPTRS_PER_BLOCK(fs_info->tree_root));
 
@@ -1333,8 +1299,9 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 		free_extent_buffer(eb_root);
 		eb = alloc_dummy_extent_buffer(logical, root->nodesize);
 	} else {
+		btrfs_set_lock_blocking_rw(eb_root, BTRFS_READ_LOCK);
 		eb = btrfs_clone_extent_buffer(eb_root);
-		btrfs_tree_read_unlock(eb_root);
+		btrfs_tree_read_unlock_blocking(eb_root);
 		free_extent_buffer(eb_root);
 	}
 
@@ -1350,7 +1317,7 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 		btrfs_set_header_generation(eb, old_generation);
 	}
 	if (tm)
-		__tree_mod_log_rewind(eb, time_seq, tm);
+		__tree_mod_log_rewind(root->fs_info, eb, time_seq, tm);
 	else
 		WARN_ON(btrfs_header_level(eb) != 0);
 	WARN_ON(btrfs_header_nritems(eb) > BTRFS_NODEPTRS_PER_BLOCK(root));
@@ -1417,14 +1384,12 @@ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans,
 
 	if (trans->transaction != root->fs_info->running_transaction)
 		WARN(1, KERN_CRIT "trans %llu running %llu\n",
-		       (unsigned long long)trans->transid,
-		       (unsigned long long)
+		       trans->transid,
 		       root->fs_info->running_transaction->transid);
 
 	if (trans->transid != root->fs_info->generation)
 		WARN(1, KERN_CRIT "trans %llu running %llu\n",
-		       (unsigned long long)trans->transid,
-		       (unsigned long long)root->fs_info->generation);
+		       trans->transid, root->fs_info->generation);
 
 	if (!should_cow_block(trans, root, buf)) {
 		*cow_ret = buf;
@@ -2178,12 +2143,8 @@ static void reada_for_search(struct btrfs_root *root,
 	}
 }
 
-/*
- * returns -EAGAIN if it had to drop the path, or zero if everything was in
- * cache
- */
-static noinline int reada_for_balance(struct btrfs_root *root,
-				      struct btrfs_path *path, int level)
+static noinline void reada_for_balance(struct btrfs_root *root,
+				       struct btrfs_path *path, int level)
 {
 	int slot;
 	int nritems;
@@ -2192,12 +2153,11 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 	u64 gen;
 	u64 block1 = 0;
 	u64 block2 = 0;
-	int ret = 0;
 	int blocksize;
 
 	parent = path->nodes[level + 1];
 	if (!parent)
-		return 0;
+		return;
 
 	nritems = btrfs_header_nritems(parent);
 	slot = path->slots[level + 1];
@@ -2224,28 +2184,11 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 			block2 = 0;
 		free_extent_buffer(eb);
 	}
-	if (block1 || block2) {
-		ret = -EAGAIN;
-
-		/* release the whole path */
-		btrfs_release_path(path);
-
-		/* read the blocks */
-		if (block1)
-			readahead_tree_block(root, block1, blocksize, 0);
-		if (block2)
-			readahead_tree_block(root, block2, blocksize, 0);
 
-		if (block1) {
-			eb = read_tree_block(root, block1, blocksize, 0);
-			free_extent_buffer(eb);
-		}
-		if (block2) {
-			eb = read_tree_block(root, block2, blocksize, 0);
-			free_extent_buffer(eb);
-		}
-	}
-	return ret;
+	if (block1)
+		readahead_tree_block(root, block1, blocksize, 0);
+	if (block2)
+		readahead_tree_block(root, block2, blocksize, 0);
 }
 
 
@@ -2359,35 +2302,28 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 	tmp = btrfs_find_tree_block(root, blocknr, blocksize);
 	if (tmp) {
 		/* first we do an atomic uptodate check */
-		if (btrfs_buffer_uptodate(tmp, 0, 1) > 0) {
-			if (btrfs_buffer_uptodate(tmp, gen, 1) > 0) {
-				/*
-				 * we found an up to date block without
-				 * sleeping, return
-				 * right away
-				 */
-				*eb_ret = tmp;
-				return 0;
-			}
-			/* the pages were up to date, but we failed
-			 * the generation number check.  Do a full
-			 * read for the generation number that is correct.
-			 * We must do this without dropping locks so
-			 * we can trust our generation number
-			 */
-			free_extent_buffer(tmp);
-			btrfs_set_path_blocking(p);
+		if (btrfs_buffer_uptodate(tmp, gen, 1) > 0) {
+			*eb_ret = tmp;
+			return 0;
+		}
 
-			/* now we're allowed to do a blocking uptodate check */
-			tmp = read_tree_block(root, blocknr, blocksize, gen);
-			if (tmp && btrfs_buffer_uptodate(tmp, gen, 0) > 0) {
-				*eb_ret = tmp;
-				return 0;
-			}
-			free_extent_buffer(tmp);
-			btrfs_release_path(p);
-			return -EIO;
+		/* the pages were up to date, but we failed
+		 * the generation number check.  Do a full
+		 * read for the generation number that is correct.
+		 * We must do this without dropping locks so
+		 * we can trust our generation number
+		 */
+		btrfs_set_path_blocking(p);
+
+		/* now we're allowed to do a blocking uptodate check */
+		ret = btrfs_read_buffer(tmp, gen);
+		if (!ret) {
+			*eb_ret = tmp;
+			return 0;
 		}
+		free_extent_buffer(tmp);
+		btrfs_release_path(p);
+		return -EIO;
 	}
 
 	/*
@@ -2448,11 +2384,8 @@ setup_nodes_for_search(struct btrfs_trans_handle *trans,
 			goto again;
 		}
 
-		sret = reada_for_balance(root, p, level);
-		if (sret)
-			goto again;
-
 		btrfs_set_path_blocking(p);
+		reada_for_balance(root, p, level);
 		sret = split_node(trans, root, p, level);
 		btrfs_clear_path_blocking(p, NULL, 0);
 
@@ -2472,11 +2405,8 @@ setup_nodes_for_search(struct btrfs_trans_handle *trans,
 			goto again;
 		}
 
-		sret = reada_for_balance(root, p, level);
-		if (sret)
-			goto again;
-
 		btrfs_set_path_blocking(p);
+		reada_for_balance(root, p, level);
 		sret = balance_level(trans, root, p, level);
 		btrfs_clear_path_blocking(p, NULL, 0);
 
@@ -2499,6 +2429,40 @@ done:
 	return ret;
 }
 
+static void key_search_validate(struct extent_buffer *b,
+				struct btrfs_key *key,
+				int level)
+{
+#ifdef CONFIG_BTRFS_ASSERT
+	struct btrfs_disk_key disk_key;
+
+	btrfs_cpu_key_to_disk(&disk_key, key);
+
+	if (level == 0)
+		ASSERT(!memcmp_extent_buffer(b, &disk_key,
+		    offsetof(struct btrfs_leaf, items[0].key),
+		    sizeof(disk_key)));
+	else
+		ASSERT(!memcmp_extent_buffer(b, &disk_key,
+		    offsetof(struct btrfs_node, ptrs[0].key),
+		    sizeof(disk_key)));
+#endif
+}
+
+static int key_search(struct extent_buffer *b, struct btrfs_key *key,
+		      int level, int *prev_cmp, int *slot)
+{
+	if (*prev_cmp != 0) {
+		*prev_cmp = bin_search(b, key, level, slot);
+		return *prev_cmp;
+	}
+
+	key_search_validate(b, key, level);
+	*slot = 0;
+
+	return 0;
+}
+
 /*
  * look for key in the tree.  path is filled in with nodes along the way
  * if key is found, we return zero and you can find the item in the leaf
@@ -2527,6 +2491,7 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
 	int write_lock_level = 0;
 	u8 lowest_level = 0;
 	int min_write_lock_level;
+	int prev_cmp;
 
 	lowest_level = p->lowest_level;
 	WARN_ON(lowest_level && ins_len > 0);
@@ -2557,6 +2522,7 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
 	min_write_lock_level = write_lock_level;
 
 again:
+	prev_cmp = -1;
 	/*
 	 * we try very hard to do read locks on the root
 	 */
@@ -2657,7 +2623,7 @@ cow_done:
 		if (!cow)
 			btrfs_unlock_up_safe(p, level + 1);
 
-		ret = bin_search(b, key, level, &slot);
+		ret = key_search(b, key, level, &prev_cmp, &slot);
 
 		if (level != 0) {
 			int dec = 0;
@@ -2792,6 +2758,7 @@ int btrfs_search_old_slot(struct btrfs_root *root, struct btrfs_key *key,
 	int level;
 	int lowest_unlock = 1;
 	u8 lowest_level = 0;
+	int prev_cmp;
 
 	lowest_level = p->lowest_level;
 	WARN_ON(p->nodes[0] != NULL);
@@ -2802,6 +2769,7 @@ int btrfs_search_old_slot(struct btrfs_root *root, struct btrfs_key *key,
 	}
 
 again:
+	prev_cmp = -1;
 	b = get_old_root(root, time_seq);
 	level = btrfs_header_level(b);
 	p->locks[level] = BTRFS_READ_LOCK;
@@ -2819,7 +2787,7 @@ again:
 		 */
 		btrfs_unlock_up_safe(p, level + 1);
 
-		ret = bin_search(b, key, level, &slot);
+		ret = key_search(b, key, level, &prev_cmp, &slot);
 
 		if (level != 0) {
 			int dec = 0;
@@ -2853,7 +2821,11 @@ again:
 				btrfs_clear_path_blocking(p, b,
 							  BTRFS_READ_LOCK);
 			}
-			b = tree_mod_log_rewind(root->fs_info, b, time_seq);
+			b = tree_mod_log_rewind(root->fs_info, p, b, time_seq);
+			if (!b) {
+				ret = -ENOMEM;
+				goto done;
+			}
 			p->locks[level] = BTRFS_READ_LOCK;
 			p->nodes[level] = b;
 		} else {
@@ -3143,7 +3115,7 @@ static int balance_node_right(struct btrfs_trans_handle *trans,
  */
 static noinline int insert_new_root(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
-			   struct btrfs_path *path, int level, int log_removal)
+			   struct btrfs_path *path, int level)
 {
 	u64 lower_gen;
 	struct extent_buffer *lower;
@@ -3176,13 +3148,11 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans,
 	btrfs_set_header_backref_rev(c, BTRFS_MIXED_BACKREF_REV);
 	btrfs_set_header_owner(c, root->root_key.objectid);
 
-	write_extent_buffer(c, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(c),
+	write_extent_buffer(c, root->fs_info->fsid, btrfs_header_fsid(c),
 			    BTRFS_FSID_SIZE);
 
 	write_extent_buffer(c, root->fs_info->chunk_tree_uuid,
-			    (unsigned long)btrfs_header_chunk_tree_uuid(c),
-			    BTRFS_UUID_SIZE);
+			    btrfs_header_chunk_tree_uuid(c), BTRFS_UUID_SIZE);
 
 	btrfs_set_node_key(c, &lower_key, 0);
 	btrfs_set_node_blockptr(c, 0, lower->start);
@@ -3194,7 +3164,7 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(c);
 
 	old = root->node;
-	tree_mod_log_set_root_pointer(root, c, log_removal);
+	tree_mod_log_set_root_pointer(root, c, 0);
 	rcu_assign_pointer(root->node, c);
 
 	/* the super has an extra ref to root->node */
@@ -3241,7 +3211,7 @@ static void insert_ptr(struct btrfs_trans_handle *trans,
 	}
 	if (level) {
 		ret = tree_mod_log_insert_key(root->fs_info, lower, slot,
-					      MOD_LOG_KEY_ADD);
+					      MOD_LOG_KEY_ADD, GFP_NOFS);
 		BUG_ON(ret < 0);
 	}
 	btrfs_set_node_key(lower, key, slot);
@@ -3278,14 +3248,14 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 		/*
 		 * trying to split the root, lets make a new one
 		 *
-		 * tree mod log: We pass 0 as log_removal parameter to
+		 * tree mod log: We don't log_removal old root in
 		 * insert_new_root, because that root buffer will be kept as a
 		 * normal node. We are going to log removal of half of the
 		 * elements below with tree_mod_log_eb_copy. We're holding a
 		 * tree lock on the buffer, which is why we cannot race with
 		 * other tree_mod_log users.
 		 */
-		ret = insert_new_root(trans, root, path, level + 1, 0);
+		ret = insert_new_root(trans, root, path, level + 1);
 		if (ret)
 			return ret;
 	} else {
@@ -3317,10 +3287,9 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 	btrfs_set_header_backref_rev(split, BTRFS_MIXED_BACKREF_REV);
 	btrfs_set_header_owner(split, root->root_key.objectid);
 	write_extent_buffer(split, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(split),
-			    BTRFS_FSID_SIZE);
+			    btrfs_header_fsid(split), BTRFS_FSID_SIZE);
 	write_extent_buffer(split, root->fs_info->chunk_tree_uuid,
-			    (unsigned long)btrfs_header_chunk_tree_uuid(split),
+			    btrfs_header_chunk_tree_uuid(split),
 			    BTRFS_UUID_SIZE);
 
 	tree_mod_log_eb_copy(root->fs_info, split, c, 0, mid, c_nritems - mid);
@@ -3986,7 +3955,7 @@ static noinline int split_leaf(struct btrfs_trans_handle *trans,
 		return -EOVERFLOW;
 
 	/* first try to make some room by pushing left and right */
-	if (data_size) {
+	if (data_size && path->nodes[1]) {
 		wret = push_leaf_right(trans, root, path, data_size,
 				       data_size, 0, 0);
 		if (wret < 0)
@@ -4005,7 +3974,7 @@ static noinline int split_leaf(struct btrfs_trans_handle *trans,
 	}
 
 	if (!path->nodes[1]) {
-		ret = insert_new_root(trans, root, path, 1, 1);
+		ret = insert_new_root(trans, root, path, 1);
 		if (ret)
 			return ret;
 	}
@@ -4073,11 +4042,10 @@ again:
 	btrfs_set_header_owner(right, root->root_key.objectid);
 	btrfs_set_header_level(right, 0);
 	write_extent_buffer(right, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(right),
-			    BTRFS_FSID_SIZE);
+			    btrfs_header_fsid(right), BTRFS_FSID_SIZE);
 
 	write_extent_buffer(right, root->fs_info->chunk_tree_uuid,
-			    (unsigned long)btrfs_header_chunk_tree_uuid(right),
+			    btrfs_header_chunk_tree_uuid(right),
 			    BTRFS_UUID_SIZE);
 
 	if (split == 0) {
@@ -4430,7 +4398,7 @@ void btrfs_truncate_item(struct btrfs_root *root, struct btrfs_path *path,
 }
 
 /*
- * make the item pointed to by the path bigger, data_size is the new size.
+ * make the item pointed to by the path bigger, data_size is the added size.
  */
 void btrfs_extend_item(struct btrfs_root *root, struct btrfs_path *path,
 		       u32 data_size)
@@ -4675,7 +4643,7 @@ static void del_ptr(struct btrfs_root *root, struct btrfs_path *path,
 			      (nritems - slot - 1));
 	} else if (level) {
 		ret = tree_mod_log_insert_key(root->fs_info, parent, slot,
-					      MOD_LOG_KEY_REMOVE);
+					      MOD_LOG_KEY_REMOVE, GFP_NOFS);
 		BUG_ON(ret < 0);
 	}
 
@@ -4847,7 +4815,7 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
  * This may release the path, and so you may lose any locks held at the
  * time you call it.
  */
-int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
+static int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
 {
 	struct btrfs_key key;
 	struct btrfs_disk_key found_key;
@@ -5362,19 +5330,20 @@ int btrfs_compare_trees(struct btrfs_root *left_root,
 					goto out;
 				advance_right = ADVANCE;
 			} else {
+				enum btrfs_compare_tree_result cmp;
+
 				WARN_ON(!extent_buffer_uptodate(left_path->nodes[0]));
 				ret = tree_compare_item(left_root, left_path,
 						right_path, tmp_buf);
-				if (ret) {
-					WARN_ON(!extent_buffer_uptodate(left_path->nodes[0]));
-					ret = changed_cb(left_root, right_root,
-						left_path, right_path,
-						&left_key,
-						BTRFS_COMPARE_TREE_CHANGED,
-						ctx);
-					if (ret < 0)
-						goto out;
-				}
+				if (ret)
+					cmp = BTRFS_COMPARE_TREE_CHANGED;
+				else
+					cmp = BTRFS_COMPARE_TREE_SAME;
+				ret = changed_cb(left_root, right_root,
+						 left_path, right_path,
+						 &left_key, cmp, ctx);
+				if (ret < 0)
+					goto out;
 				advance_left = ADVANCE;
 				advance_right = ADVANCE;
 			}
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index d6dd49b51ba8..0506f40ede83 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -23,6 +23,7 @@
 #include <linux/highmem.h>
 #include <linux/fs.h>
 #include <linux/rwsem.h>
+#include <linux/semaphore.h>
 #include <linux/completion.h>
 #include <linux/backing-dev.h>
 #include <linux/wait.h>
@@ -91,6 +92,9 @@ struct btrfs_ordered_sum;
 /* holds quota configuration and tracking */
 #define BTRFS_QUOTA_TREE_OBJECTID 8ULL
 
+/* for storing items that use the BTRFS_UUID_KEY* types */
+#define BTRFS_UUID_TREE_OBJECTID 9ULL
+
 /* for storing balance parameters in the root tree */
 #define BTRFS_BALANCE_OBJECTID -4ULL
 
@@ -142,7 +146,7 @@ struct btrfs_ordered_sum;
 
 #define BTRFS_EMPTY_SUBVOL_DIR_OBJECTID 2
 
-#define BTRFS_DEV_REPLACE_DEVID 0
+#define BTRFS_DEV_REPLACE_DEVID 0ULL
 
 /*
  * the max metadata block size.  This limit is somewhat artificial,
@@ -478,9 +482,10 @@ struct btrfs_super_block {
 	char label[BTRFS_LABEL_SIZE];
 
 	__le64 cache_generation;
+	__le64 uuid_tree_generation;
 
 	/* future expansion */
-	__le64 reserved[31];
+	__le64 reserved[30];
 	u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE];
 	struct btrfs_root_backup super_roots[BTRFS_NUM_BACKUP_ROOTS];
 } __attribute__ ((__packed__));
@@ -961,8 +966,8 @@ struct btrfs_dev_replace_item {
 #define BTRFS_BLOCK_GROUP_RAID1		(1ULL << 4)
 #define BTRFS_BLOCK_GROUP_DUP		(1ULL << 5)
 #define BTRFS_BLOCK_GROUP_RAID10	(1ULL << 6)
-#define BTRFS_BLOCK_GROUP_RAID5    (1 << 7)
-#define BTRFS_BLOCK_GROUP_RAID6    (1 << 8)
+#define BTRFS_BLOCK_GROUP_RAID5         (1ULL << 7)
+#define BTRFS_BLOCK_GROUP_RAID6         (1ULL << 8)
 #define BTRFS_BLOCK_GROUP_RESERVED	BTRFS_AVAIL_ALLOC_BIT_SINGLE
 
 enum btrfs_raid_types {
@@ -1102,13 +1107,16 @@ struct btrfs_space_info {
 				   account */
 
 	/*
-	 * we bump reservation progress every time we decrement
-	 * bytes_reserved.  This way people waiting for reservations
-	 * know something good has happened and they can check
-	 * for progress.  The number here isn't to be trusted, it
-	 * just shows reclaim activity
+	 * bytes_pinned is kept in line with what is actually pinned, as in
+	 * we've called update_block_group and dropped the bytes_used counter
+	 * and increased the bytes_pinned counter.  However this means that
+	 * bytes_pinned does not reflect the bytes that will be pinned once the
+	 * delayed refs are flushed, so this counter is inc'ed everytime we call
+	 * btrfs_free_extent so it is a realtime count of what will be freed
+	 * once the transaction is committed.  It will be zero'ed everytime the
+	 * transaction commits.
 	 */
-	unsigned long reservation_progress;
+	struct percpu_counter total_bytes_pinned;
 
 	unsigned int full:1;	/* indicates that we cannot allocate any more
 				   chunks for this space */
@@ -1176,6 +1184,7 @@ enum btrfs_caching_type {
 	BTRFS_CACHE_STARTED	= 1,
 	BTRFS_CACHE_FAST	= 2,
 	BTRFS_CACHE_FINISHED	= 3,
+	BTRFS_CACHE_ERROR	= 4,
 };
 
 enum btrfs_disk_cache_state {
@@ -1290,6 +1299,7 @@ struct btrfs_fs_info {
 	struct btrfs_root *fs_root;
 	struct btrfs_root *csum_root;
 	struct btrfs_root *quota_root;
+	struct btrfs_root *uuid_root;
 
 	/* the log root tree is a directory of all the other log roots */
 	struct btrfs_root *log_root_tree;
@@ -1338,6 +1348,7 @@ struct btrfs_fs_info {
 	u64 last_trans_log_full_commit;
 	unsigned long mount_opt;
 	unsigned long compress_type:4;
+	int commit_interval;
 	/*
 	 * It is a suggestive number, the read side is safe even it gets a
 	 * wrong number because we will write out the data into a regular
@@ -1399,6 +1410,13 @@ struct btrfs_fs_info {
 	 * before jumping into the main commit.
 	 */
 	struct mutex ordered_operations_mutex;
+
+	/*
+	 * Same as ordered_operations_mutex except this is for ordered extents
+	 * and not the operations.
+	 */
+	struct mutex ordered_extent_flush_mutex;
+
 	struct rw_semaphore extent_commit_sem;
 
 	struct rw_semaphore cleanup_work_sem;
@@ -1437,25 +1455,22 @@ struct btrfs_fs_info {
 	atomic_t open_ioctl_trans;
 
 	/*
-	 * this is used by the balancing code to wait for all the pending
-	 * ordered extents
+	 * this is used to protect the following list -- ordered_roots.
 	 */
-	spinlock_t ordered_extent_lock;
+	spinlock_t ordered_root_lock;
 
 	/*
-	 * all of the data=ordered extents pending writeback
+	 * all fs/file tree roots in which there are data=ordered extents
+	 * pending writeback are added into this list.
+	 *
 	 * these can span multiple transactions and basically include
 	 * every dirty data page that isn't from nodatacow
 	 */
-	struct list_head ordered_extents;
+	struct list_head ordered_roots;
 
-	spinlock_t delalloc_lock;
-	/*
-	 * all of the inodes that have delalloc bytes.  It is possible for
-	 * this list to be empty even when there is still dirty data=ordered
-	 * extents waiting to finish IO.
-	 */
-	struct list_head delalloc_inodes;
+	spinlock_t delalloc_root_lock;
+	/* all fs/file tree roots that have delalloc inodes. */
+	struct list_head delalloc_roots;
 
 	/*
 	 * there is a pool of worker threads for checksumming during writes
@@ -1498,8 +1513,6 @@ struct btrfs_fs_info {
 	int do_barriers;
 	int closing;
 	int log_root_recovering;
-	int enospc_unlink;
-	int trans_no_join;
 
 	u64 total_pinned;
 
@@ -1594,6 +1607,12 @@ struct btrfs_fs_info {
 	struct rb_root qgroup_tree;
 	spinlock_t qgroup_lock;
 
+	/*
+	 * used to avoid frequently calling ulist_alloc()/ulist_free()
+	 * when doing qgroup accounting, it must be protected by qgroup_lock.
+	 */
+	struct ulist *qgroup_ulist;
+
 	/* protect user change for quota operations */
 	struct mutex qgroup_ioctl_lock;
 
@@ -1607,6 +1626,8 @@ struct btrfs_fs_info {
 	struct mutex qgroup_rescan_lock; /* protects the progress item */
 	struct btrfs_key qgroup_rescan_progress;
 	struct btrfs_workers qgroup_rescan_workers;
+	struct completion qgroup_rescan_completion;
+	struct btrfs_work qgroup_rescan_work;
 
 	/* filesystem state */
 	unsigned long fs_state;
@@ -1626,6 +1647,9 @@ struct btrfs_fs_info {
 	struct btrfs_dev_replace dev_replace;
 
 	atomic_t mutually_exclusive_operation_running;
+
+	struct semaphore uuid_tree_rescan_sem;
+	unsigned int update_uuid_tree_gen:1;
 };
 
 /*
@@ -1739,6 +1763,31 @@ struct btrfs_root {
 	int force_cow;
 
 	spinlock_t root_item_lock;
+	atomic_t refs;
+
+	spinlock_t delalloc_lock;
+	/*
+	 * all of the inodes that have delalloc bytes.  It is possible for
+	 * this list to be empty even when there is still dirty data=ordered
+	 * extents waiting to finish IO.
+	 */
+	struct list_head delalloc_inodes;
+	struct list_head delalloc_root;
+	u64 nr_delalloc_inodes;
+	/*
+	 * this is used by the balancing code to wait for all the pending
+	 * ordered extents
+	 */
+	spinlock_t ordered_extent_lock;
+
+	/*
+	 * all of the data=ordered extents pending writeback
+	 * these can span multiple transactions and basically include
+	 * every dirty data page that isn't from nodatacow
+	 */
+	struct list_head ordered_extents;
+	struct list_head ordered_root;
+	u64 nr_ordered_extents;
 };
 
 struct btrfs_ioctl_defrag_range_args {
@@ -1894,6 +1943,19 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_DEV_REPLACE_KEY	250
 
 /*
+ * Stores items that allow to quickly map UUIDs to something else.
+ * These items are part of the filesystem UUID tree.
+ * The key is built like this:
+ * (UUID_upper_64_bits, BTRFS_UUID_KEY*, UUID_lower_64_bits).
+ */
+#if BTRFS_UUID_SIZE != 16
+#error "UUID items require BTRFS_UUID_SIZE == 16!"
+#endif
+#define BTRFS_UUID_KEY_SUBVOL	251	/* for UUIDs assigned to subvols */
+#define BTRFS_UUID_KEY_RECEIVED_SUBVOL	252	/* for UUIDs assigned to
+						 * received subvols */
+
+/*
  * string items are for debugging.  They just store a short string of
  * data in the FS
  */
@@ -1927,6 +1989,9 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_MOUNT_CHECK_INTEGRITY	(1 << 20)
 #define BTRFS_MOUNT_CHECK_INTEGRITY_INCLUDING_EXTENT_DATA (1 << 21)
 #define BTRFS_MOUNT_PANIC_ON_FATAL_ERROR	(1 << 22)
+#define BTRFS_MOUNT_RESCAN_UUID_TREE	(1 << 23)
+
+#define BTRFS_DEFAULT_COMMIT_INTERVAL	(30)
 
 #define btrfs_clear_opt(o, opt)		((o) &= ~BTRFS_MOUNT_##opt)
 #define btrfs_set_opt(o, opt)		((o) |= BTRFS_MOUNT_##opt)
@@ -2090,14 +2155,14 @@ BTRFS_SETGET_STACK_FUNCS(stack_device_bandwidth, struct btrfs_dev_item,
 BTRFS_SETGET_STACK_FUNCS(stack_device_generation, struct btrfs_dev_item,
 			 generation, 64);
 
-static inline char *btrfs_device_uuid(struct btrfs_dev_item *d)
+static inline unsigned long btrfs_device_uuid(struct btrfs_dev_item *d)
 {
-	return (char *)d + offsetof(struct btrfs_dev_item, uuid);
+	return (unsigned long)d + offsetof(struct btrfs_dev_item, uuid);
 }
 
-static inline char *btrfs_device_fsid(struct btrfs_dev_item *d)
+static inline unsigned long btrfs_device_fsid(struct btrfs_dev_item *d)
 {
-	return (char *)d + offsetof(struct btrfs_dev_item, fsid);
+	return (unsigned long)d + offsetof(struct btrfs_dev_item, fsid);
 }
 
 BTRFS_SETGET_FUNCS(chunk_length, struct btrfs_chunk, length, 64);
@@ -2200,6 +2265,23 @@ BTRFS_SETGET_FUNCS(inode_gid, struct btrfs_inode_item, gid, 32);
 BTRFS_SETGET_FUNCS(inode_mode, struct btrfs_inode_item, mode, 32);
 BTRFS_SETGET_FUNCS(inode_rdev, struct btrfs_inode_item, rdev, 64);
 BTRFS_SETGET_FUNCS(inode_flags, struct btrfs_inode_item, flags, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_generation, struct btrfs_inode_item,
+			 generation, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_sequence, struct btrfs_inode_item,
+			 sequence, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_transid, struct btrfs_inode_item,
+			 transid, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_size, struct btrfs_inode_item, size, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_nbytes, struct btrfs_inode_item,
+			 nbytes, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_block_group, struct btrfs_inode_item,
+			 block_group, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_nlink, struct btrfs_inode_item, nlink, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_uid, struct btrfs_inode_item, uid, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_gid, struct btrfs_inode_item, gid, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_mode, struct btrfs_inode_item, mode, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_rdev, struct btrfs_inode_item, rdev, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_flags, struct btrfs_inode_item, flags, 64);
 
 static inline struct btrfs_timespec *
 btrfs_inode_atime(struct btrfs_inode_item *inode_item)
@@ -2227,6 +2309,8 @@ btrfs_inode_ctime(struct btrfs_inode_item *inode_item)
 
 BTRFS_SETGET_FUNCS(timespec_sec, struct btrfs_timespec, sec, 64);
 BTRFS_SETGET_FUNCS(timespec_nsec, struct btrfs_timespec, nsec, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_timespec_sec, struct btrfs_timespec, sec, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_timespec_nsec, struct btrfs_timespec, nsec, 32);
 
 /* struct btrfs_dev_extent */
 BTRFS_SETGET_FUNCS(dev_extent_chunk_tree, struct btrfs_dev_extent,
@@ -2237,10 +2321,10 @@ BTRFS_SETGET_FUNCS(dev_extent_chunk_offset, struct btrfs_dev_extent,
 		   chunk_offset, 64);
 BTRFS_SETGET_FUNCS(dev_extent_length, struct btrfs_dev_extent, length, 64);
 
-static inline u8 *btrfs_dev_extent_chunk_tree_uuid(struct btrfs_dev_extent *dev)
+static inline unsigned long btrfs_dev_extent_chunk_tree_uuid(struct btrfs_dev_extent *dev)
 {
 	unsigned long ptr = offsetof(struct btrfs_dev_extent, chunk_tree_uuid);
-	return (u8 *)((unsigned long)dev + ptr);
+	return (unsigned long)dev + ptr;
 }
 
 BTRFS_SETGET_FUNCS(extent_refs, struct btrfs_extent_item, refs, 64);
@@ -2308,6 +2392,10 @@ BTRFS_SETGET_FUNCS(ref_count_v0, struct btrfs_extent_ref_v0, count, 32);
 /* struct btrfs_node */
 BTRFS_SETGET_FUNCS(key_blockptr, struct btrfs_key_ptr, blockptr, 64);
 BTRFS_SETGET_FUNCS(key_generation, struct btrfs_key_ptr, generation, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_key_blockptr, struct btrfs_key_ptr,
+			 blockptr, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_key_generation, struct btrfs_key_ptr,
+			 generation, 64);
 
 static inline u64 btrfs_node_blockptr(struct extent_buffer *eb, int nr)
 {
@@ -2364,6 +2452,8 @@ static inline void btrfs_set_node_key(struct extent_buffer *eb,
 /* struct btrfs_item */
 BTRFS_SETGET_FUNCS(item_offset, struct btrfs_item, offset, 32);
 BTRFS_SETGET_FUNCS(item_size, struct btrfs_item, size, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_item_offset, struct btrfs_item, offset, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_item_size, struct btrfs_item, size, 32);
 
 static inline unsigned long btrfs_item_nr_offset(int nr)
 {
@@ -2426,6 +2516,13 @@ BTRFS_SETGET_FUNCS(dir_data_len, struct btrfs_dir_item, data_len, 16);
 BTRFS_SETGET_FUNCS(dir_type, struct btrfs_dir_item, type, 8);
 BTRFS_SETGET_FUNCS(dir_name_len, struct btrfs_dir_item, name_len, 16);
 BTRFS_SETGET_FUNCS(dir_transid, struct btrfs_dir_item, transid, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_dir_type, struct btrfs_dir_item, type, 8);
+BTRFS_SETGET_STACK_FUNCS(stack_dir_data_len, struct btrfs_dir_item,
+			 data_len, 16);
+BTRFS_SETGET_STACK_FUNCS(stack_dir_name_len, struct btrfs_dir_item,
+			 name_len, 16);
+BTRFS_SETGET_STACK_FUNCS(stack_dir_transid, struct btrfs_dir_item,
+			 transid, 64);
 
 static inline void btrfs_dir_item_key(struct extent_buffer *eb,
 				      struct btrfs_dir_item *item,
@@ -2528,6 +2625,12 @@ BTRFS_SETGET_HEADER_FUNCS(header_owner, struct btrfs_header, owner, 64);
 BTRFS_SETGET_HEADER_FUNCS(header_nritems, struct btrfs_header, nritems, 32);
 BTRFS_SETGET_HEADER_FUNCS(header_flags, struct btrfs_header, flags, 64);
 BTRFS_SETGET_HEADER_FUNCS(header_level, struct btrfs_header, level, 8);
+BTRFS_SETGET_STACK_FUNCS(stack_header_generation, struct btrfs_header,
+			 generation, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_header_owner, struct btrfs_header, owner, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_header_nritems, struct btrfs_header,
+			 nritems, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_header_bytenr, struct btrfs_header, bytenr, 64);
 
 static inline int btrfs_header_flag(struct extent_buffer *eb, u64 flag)
 {
@@ -2563,16 +2666,14 @@ static inline void btrfs_set_header_backref_rev(struct extent_buffer *eb,
 	btrfs_set_header_flags(eb, flags);
 }
 
-static inline u8 *btrfs_header_fsid(struct extent_buffer *eb)
+static inline unsigned long btrfs_header_fsid(struct extent_buffer *eb)
 {
-	unsigned long ptr = offsetof(struct btrfs_header, fsid);
-	return (u8 *)ptr;
+	return offsetof(struct btrfs_header, fsid);
 }
 
-static inline u8 *btrfs_header_chunk_tree_uuid(struct extent_buffer *eb)
+static inline unsigned long btrfs_header_chunk_tree_uuid(struct extent_buffer *eb)
 {
-	unsigned long ptr = offsetof(struct btrfs_header, chunk_tree_uuid);
-	return (u8 *)ptr;
+	return offsetof(struct btrfs_header, chunk_tree_uuid);
 }
 
 static inline int btrfs_is_leaf(struct extent_buffer *eb)
@@ -2790,6 +2891,9 @@ BTRFS_SETGET_STACK_FUNCS(super_csum_type, struct btrfs_super_block,
 			 csum_type, 16);
 BTRFS_SETGET_STACK_FUNCS(super_cache_generation, struct btrfs_super_block,
 			 cache_generation, 64);
+BTRFS_SETGET_STACK_FUNCS(super_magic, struct btrfs_super_block, magic, 64);
+BTRFS_SETGET_STACK_FUNCS(super_uuid_tree_generation, struct btrfs_super_block,
+			 uuid_tree_generation, 64);
 
 static inline int btrfs_super_csum_size(struct btrfs_super_block *s)
 {
@@ -2807,6 +2911,14 @@ static inline unsigned long btrfs_leaf_data(struct extent_buffer *l)
 
 /* struct btrfs_file_extent_item */
 BTRFS_SETGET_FUNCS(file_extent_type, struct btrfs_file_extent_item, type, 8);
+BTRFS_SETGET_STACK_FUNCS(stack_file_extent_disk_bytenr,
+			 struct btrfs_file_extent_item, disk_bytenr, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_file_extent_offset,
+			 struct btrfs_file_extent_item, offset, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_file_extent_generation,
+			 struct btrfs_file_extent_item, generation, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_file_extent_num_bytes,
+			 struct btrfs_file_extent_item, num_bytes, 64);
 
 static inline unsigned long
 btrfs_file_extent_inline_start(struct btrfs_file_extent_item *e)
@@ -3014,7 +3126,7 @@ static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root,
 						 unsigned num_items)
 {
 	return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) *
-		3 * num_items;
+		2 * num_items;
 }
 
 /*
@@ -3028,6 +3140,8 @@ static inline u64 btrfs_calc_trunc_metadata_size(struct btrfs_root *root,
 		num_items;
 }
 
+int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans,
+				       struct btrfs_root *root);
 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);
@@ -3039,6 +3153,8 @@ int btrfs_pin_extent(struct btrfs_root *root,
 		     u64 bytenr, u64 num, int reserved);
 int btrfs_pin_extent_for_log_replay(struct btrfs_root *root,
 				    u64 bytenr, u64 num_bytes);
+int btrfs_exclude_logged_extents(struct btrfs_root *root,
+				 struct extent_buffer *eb);
 int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root,
 			  u64 objectid, u64 offset, u64 bytenr);
@@ -3063,11 +3179,9 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root,
 				   u64 root_objectid, u64 owner, u64 offset,
 				   struct btrfs_key *ins);
-int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
-				  struct btrfs_root *root,
-				  u64 num_bytes, u64 min_alloc_size,
-				  u64 empty_size, u64 hint_byte,
-				  struct btrfs_key *ins, int is_data);
+int btrfs_reserve_extent(struct btrfs_root *root, u64 num_bytes,
+			 u64 min_alloc_size, u64 empty_size, u64 hint_byte,
+			 struct btrfs_key *ins, int is_data);
 int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		  struct extent_buffer *buf, int full_backref, int for_cow);
 int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
@@ -3131,7 +3245,7 @@ void btrfs_orphan_release_metadata(struct inode *inode);
 int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 				     struct btrfs_block_rsv *rsv,
 				     int nitems,
-				     u64 *qgroup_reserved);
+				     u64 *qgroup_reserved, bool use_global_rsv);
 void btrfs_subvolume_release_metadata(struct btrfs_root *root,
 				      struct btrfs_block_rsv *rsv,
 				      u64 qgroup_reserved);
@@ -3155,6 +3269,9 @@ int btrfs_block_rsv_refill(struct btrfs_root *root,
 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
 			    struct btrfs_block_rsv *dst_rsv,
 			    u64 num_bytes);
+int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info,
+			     struct btrfs_block_rsv *dest, u64 num_bytes,
+			     int min_factor);
 void btrfs_block_rsv_release(struct btrfs_root *root,
 			     struct btrfs_block_rsv *block_rsv,
 			     u64 num_bytes);
@@ -3198,6 +3315,7 @@ enum btrfs_compare_tree_result {
 	BTRFS_COMPARE_TREE_NEW,
 	BTRFS_COMPARE_TREE_DELETED,
 	BTRFS_COMPARE_TREE_CHANGED,
+	BTRFS_COMPARE_TREE_SAME,
 };
 typedef int (*btrfs_changed_cb_t)(struct btrfs_root *left_root,
 				  struct btrfs_root *right_root,
@@ -3311,6 +3429,18 @@ static inline int btrfs_fs_closing(struct btrfs_fs_info *fs_info)
 	smp_mb();
 	return fs_info->closing;
 }
+
+/*
+ * If we remount the fs to be R/O or umount the fs, the cleaner needn't do
+ * anything except sleeping. This function is used to check the status of
+ * the fs.
+ */
+static inline int btrfs_need_cleaner_sleep(struct btrfs_root *root)
+{
+	return (root->fs_info->sb->s_flags & MS_RDONLY ||
+		btrfs_fs_closing(root->fs_info));
+}
+
 static inline void free_fs_info(struct btrfs_fs_info *fs_info)
 {
 	kfree(fs_info->balance_ctl);
@@ -3321,6 +3451,7 @@ static inline void free_fs_info(struct btrfs_fs_info *fs_info)
 	kfree(fs_info->dev_root);
 	kfree(fs_info->csum_root);
 	kfree(fs_info->quota_root);
+	kfree(fs_info->uuid_root);
 	kfree(fs_info->super_copy);
 	kfree(fs_info->super_for_commit);
 	kfree(fs_info);
@@ -3355,11 +3486,9 @@ int __must_check btrfs_update_root(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root,
 				   struct btrfs_key *key,
 				   struct btrfs_root_item *item);
-void btrfs_read_root_item(struct extent_buffer *eb, int slot,
-			  struct btrfs_root_item *item);
-int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct
-			 btrfs_root_item *item, struct btrfs_key *key);
-int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid);
+int btrfs_find_root(struct btrfs_root *root, struct btrfs_key *search_key,
+		    struct btrfs_path *path, struct btrfs_root_item *root_item,
+		    struct btrfs_key *root_key);
 int btrfs_find_orphan_roots(struct btrfs_root *tree_root);
 void btrfs_set_root_node(struct btrfs_root_item *item,
 			 struct extent_buffer *node);
@@ -3367,6 +3496,17 @@ void btrfs_check_and_init_root_item(struct btrfs_root_item *item);
 void btrfs_update_root_times(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root);
 
+/* uuid-tree.c */
+int btrfs_uuid_tree_add(struct btrfs_trans_handle *trans,
+			struct btrfs_root *uuid_root, u8 *uuid, u8 type,
+			u64 subid);
+int btrfs_uuid_tree_rem(struct btrfs_trans_handle *trans,
+			struct btrfs_root *uuid_root, u8 *uuid, u8 type,
+			u64 subid);
+int btrfs_uuid_tree_iterate(struct btrfs_fs_info *fs_info,
+			    int (*check_func)(struct btrfs_fs_info *, u8 *, u8,
+					      u64));
+
 /* dir-item.c */
 int btrfs_check_dir_item_collision(struct btrfs_root *root, u64 dir,
 			  const char *name, int name_len);
@@ -3450,12 +3590,14 @@ int btrfs_find_name_in_ext_backref(struct btrfs_path *path,
 				   struct btrfs_inode_extref **extref_ret);
 
 /* file-item.c */
+struct btrfs_dio_private;
 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);
+			      struct btrfs_dio_private *dip, struct bio *bio,
+			      u64 logical_offset);
 int btrfs_insert_file_extent(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
 			     u64 objectid, u64 pos,
@@ -3493,6 +3635,9 @@ void btrfs_wait_and_free_delalloc_work(struct btrfs_delalloc_work *work);
 struct extent_map *btrfs_get_extent_fiemap(struct inode *inode, struct page *page,
 					   size_t pg_offset, u64 start, u64 len,
 					   int create);
+noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
+			      u64 *orig_start, u64 *orig_block_len,
+			      u64 *ram_bytes);
 
 /* RHEL and EL kernels have a patch that renames PG_checked to FsMisc */
 #if defined(ClearPageFsMisc) && !defined(ClearPageChecked)
@@ -3530,6 +3675,8 @@ 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_all_delalloc_inodes(struct btrfs_fs_info *fs_info,
+				    int delay_iput);
 int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
 			      struct extent_state **cached_state);
 int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
@@ -3578,11 +3725,15 @@ extern const struct dentry_operations btrfs_dentry_operations;
 long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
 void btrfs_update_iflags(struct inode *inode);
 void btrfs_inherit_iflags(struct inode *inode, struct inode *dir);
+int btrfs_is_empty_uuid(u8 *uuid);
 int btrfs_defrag_file(struct inode *inode, struct file *file,
 		      struct btrfs_ioctl_defrag_range_args *range,
 		      u64 newer_than, unsigned long max_pages);
 void btrfs_get_block_group_info(struct list_head *groups_list,
 				struct btrfs_ioctl_space_info *space);
+void update_ioctl_balance_args(struct btrfs_fs_info *fs_info, int lock,
+			       struct btrfs_ioctl_balance_args *bargs);
+
 
 /* file.c */
 int btrfs_auto_defrag_init(void);
@@ -3655,6 +3806,22 @@ void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...)
 #define btrfs_debug(fs_info, fmt, args...) \
 	btrfs_printk(fs_info, KERN_DEBUG fmt, ##args)
 
+#ifdef CONFIG_BTRFS_ASSERT
+
+static inline void assfail(char *expr, char *file, int line)
+{
+	printk(KERN_ERR "BTRFS assertion failed: %s, file: %s, line: %d",
+	       expr, file, line);
+	BUG();
+}
+
+#define ASSERT(expr)	\
+	(likely(expr) ? (void)0 : assfail(#expr, __FILE__, __LINE__))
+#else
+#define ASSERT(expr)	((void)0)
+#endif
+
+#define btrfs_assert()
 __printf(5, 6)
 void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
 		     unsigned int line, int errno, const char *fmt, ...);
@@ -3763,9 +3930,9 @@ 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);
+int 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);
@@ -3814,6 +3981,8 @@ int btrfs_quota_enable(struct btrfs_trans_handle *trans,
 int btrfs_quota_disable(struct btrfs_trans_handle *trans,
 			struct btrfs_fs_info *fs_info);
 int btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info);
+void btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info);
+int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info);
 int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans,
 			      struct btrfs_fs_info *fs_info, u64 src, u64 dst);
 int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index f26f38ccd194..cbd9523ad09c 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -21,6 +21,7 @@
 #include "delayed-inode.h"
 #include "disk-io.h"
 #include "transaction.h"
+#include "ctree.h"
 
 #define BTRFS_DELAYED_WRITEBACK		512
 #define BTRFS_DELAYED_BACKGROUND	128
@@ -535,20 +536,6 @@ static struct btrfs_delayed_item *__btrfs_next_delayed_item(
 	return next;
 }
 
-static inline struct btrfs_root *btrfs_get_fs_root(struct btrfs_root *root,
-						   u64 root_id)
-{
-	struct btrfs_key root_key;
-
-	if (root->objectid == root_id)
-		return root;
-
-	root_key.objectid = root_id;
-	root_key.type = BTRFS_ROOT_ITEM_KEY;
-	root_key.offset = (u64)-1;
-	return btrfs_read_fs_root_no_name(root->fs_info, &root_key);
-}
-
 static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans,
 					       struct btrfs_root *root,
 					       struct btrfs_delayed_item *item)
@@ -1467,10 +1454,10 @@ int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans,
 
 	dir_item = (struct btrfs_dir_item *)delayed_item->data;
 	dir_item->location = *disk_key;
-	dir_item->transid = cpu_to_le64(trans->transid);
-	dir_item->data_len = 0;
-	dir_item->name_len = cpu_to_le16(name_len);
-	dir_item->type = type;
+	btrfs_set_stack_dir_transid(dir_item, trans->transid);
+	btrfs_set_stack_dir_data_len(dir_item, 0);
+	btrfs_set_stack_dir_name_len(dir_item, name_len);
+	btrfs_set_stack_dir_type(dir_item, type);
 	memcpy((char *)(dir_item + 1), name, name_len);
 
 	ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item);
@@ -1484,13 +1471,11 @@ int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans,
 	mutex_lock(&delayed_node->mutex);
 	ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item);
 	if (unlikely(ret)) {
-		printk(KERN_ERR "err add delayed dir index item(name: %s) into "
-				"the insertion tree of the delayed node"
+		printk(KERN_ERR "err add delayed dir index item(name: %.*s) "
+				"into the insertion tree of the delayed node"
 				"(root id: %llu, inode id: %llu, errno: %d)\n",
-				name,
-				(unsigned long long)delayed_node->root->objectid,
-				(unsigned long long)delayed_node->inode_id,
-				ret);
+				name_len, name, delayed_node->root->objectid,
+				delayed_node->inode_id, ret);
 		BUG();
 	}
 	mutex_unlock(&delayed_node->mutex);
@@ -1561,9 +1546,7 @@ int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans,
 		printk(KERN_ERR "err add delayed dir index item(index: %llu) "
 				"into the deletion tree of the delayed node"
 				"(root id: %llu, inode id: %llu, errno: %d)\n",
-				(unsigned long long)index,
-				(unsigned long long)node->root->objectid,
-				(unsigned long long)node->inode_id,
+				index, node->root->objectid, node->inode_id,
 				ret);
 		BUG();
 	}
@@ -1681,8 +1664,7 @@ int btrfs_should_delete_dir_index(struct list_head *del_list,
  * btrfs_readdir_delayed_dir_index - read dir info stored in the delayed tree
  *
  */
-int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent,
-				    filldir_t filldir,
+int btrfs_readdir_delayed_dir_index(struct dir_context *ctx,
 				    struct list_head *ins_list)
 {
 	struct btrfs_dir_item *di;
@@ -1704,22 +1686,22 @@ int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent,
 	list_for_each_entry_safe(curr, next, ins_list, readdir_list) {
 		list_del(&curr->readdir_list);
 
-		if (curr->key.offset < filp->f_pos) {
+		if (curr->key.offset < ctx->pos) {
 			if (atomic_dec_and_test(&curr->refs))
 				kfree(curr);
 			continue;
 		}
 
-		filp->f_pos = curr->key.offset;
+		ctx->pos = curr->key.offset;
 
 		di = (struct btrfs_dir_item *)curr->data;
 		name = (char *)(di + 1);
-		name_len = le16_to_cpu(di->name_len);
+		name_len = btrfs_stack_dir_name_len(di);
 
 		d_type = btrfs_filetype_table[di->type];
 		btrfs_disk_key_to_cpu(&location, &di->location);
 
-		over = filldir(dirent, name, name_len, curr->key.offset,
+		over = !dir_emit(ctx, name, name_len,
 			       location.objectid, d_type);
 
 		if (atomic_dec_and_test(&curr->refs))
@@ -1731,27 +1713,6 @@ int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent,
 	return 0;
 }
 
-BTRFS_SETGET_STACK_FUNCS(stack_inode_generation, struct btrfs_inode_item,
-			 generation, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_sequence, struct btrfs_inode_item,
-			 sequence, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_transid, struct btrfs_inode_item,
-			 transid, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_size, struct btrfs_inode_item, size, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_nbytes, struct btrfs_inode_item,
-			 nbytes, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_block_group, struct btrfs_inode_item,
-			 block_group, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_nlink, struct btrfs_inode_item, nlink, 32);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_uid, struct btrfs_inode_item, uid, 32);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_gid, struct btrfs_inode_item, gid, 32);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_mode, struct btrfs_inode_item, mode, 32);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_rdev, struct btrfs_inode_item, rdev, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_inode_flags, struct btrfs_inode_item, flags, 64);
-
-BTRFS_SETGET_STACK_FUNCS(stack_timespec_sec, struct btrfs_timespec, sec, 64);
-BTRFS_SETGET_STACK_FUNCS(stack_timespec_nsec, struct btrfs_timespec, nsec, 32);
-
 static void fill_stack_inode_item(struct btrfs_trans_handle *trans,
 				  struct btrfs_inode_item *inode_item,
 				  struct inode *inode)
diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h
index 1d5c5f7abe3e..a4b38f934d14 100644
--- a/fs/btrfs/delayed-inode.h
+++ b/fs/btrfs/delayed-inode.h
@@ -139,8 +139,7 @@ void btrfs_put_delayed_items(struct list_head *ins_list,
 			     struct list_head *del_list);
 int btrfs_should_delete_dir_index(struct list_head *del_list,
 				  u64 index);
-int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent,
-				    filldir_t filldir,
+int btrfs_readdir_delayed_dir_index(struct dir_context *ctx,
 				    struct list_head *ins_list);
 
 /* for init */
diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index c219463fb1fd..e4d467be2dd4 100644
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -241,7 +241,7 @@ int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
-static void inline drop_delayed_ref(struct btrfs_trans_handle *trans,
+static inline void drop_delayed_ref(struct btrfs_trans_handle *trans,
 				    struct btrfs_delayed_ref_root *delayed_refs,
 				    struct btrfs_delayed_ref_node *ref)
 {
@@ -600,7 +600,7 @@ static noinline void add_delayed_ref_head(struct btrfs_fs_info *fs_info,
 	INIT_LIST_HEAD(&head_ref->cluster);
 	mutex_init(&head_ref->mutex);
 
-	trace_btrfs_delayed_ref_head(ref, head_ref, action);
+	trace_add_delayed_ref_head(ref, head_ref, action);
 
 	existing = tree_insert(&delayed_refs->root, &ref->rb_node);
 
@@ -661,7 +661,7 @@ static noinline void add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 		ref->type = BTRFS_TREE_BLOCK_REF_KEY;
 	full_ref->level = level;
 
-	trace_btrfs_delayed_tree_ref(ref, full_ref, action);
+	trace_add_delayed_tree_ref(ref, full_ref, action);
 
 	existing = tree_insert(&delayed_refs->root, &ref->rb_node);
 
@@ -722,7 +722,7 @@ static noinline void add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 	full_ref->objectid = owner;
 	full_ref->offset = offset;
 
-	trace_btrfs_delayed_data_ref(ref, full_ref, action);
+	trace_add_delayed_data_ref(ref, full_ref, action);
 
 	existing = tree_insert(&delayed_refs->root, &ref->rb_node);
 
diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index 65241f32d3f8..9efb94e95858 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -148,13 +148,13 @@ no_valid_dev_replace_entry_found:
 		    !btrfs_test_opt(dev_root, DEGRADED)) {
 			ret = -EIO;
 			pr_warn("btrfs: cannot mount because device replace operation is ongoing and\n" "srcdev (devid %llu) is missing, need to run 'btrfs dev scan'?\n",
-				(unsigned long long)src_devid);
+				src_devid);
 		}
 		if (!dev_replace->tgtdev &&
 		    !btrfs_test_opt(dev_root, DEGRADED)) {
 			ret = -EIO;
 			pr_warn("btrfs: cannot mount because device replace operation is ongoing and\n" "tgtdev (devid %llu) is missing, need to run btrfs dev scan?\n",
-				(unsigned long long)BTRFS_DEV_REPLACE_DEVID);
+				BTRFS_DEV_REPLACE_DEVID);
 		}
 		if (dev_replace->tgtdev) {
 			if (dev_replace->srcdev) {
@@ -400,7 +400,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
 	btrfs_dev_replace_unlock(dev_replace);
 
-	btrfs_wait_ordered_extents(root, 0);
+	btrfs_wait_all_ordered_extents(root->fs_info);
 
 	/* force writing the updated state information to disk */
 	trans = btrfs_start_transaction(root, 0);
@@ -470,12 +470,12 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	 * flush all outstanding I/O and inode extent mappings before the
 	 * copy operation is declared as being finished
 	 */
-	ret = btrfs_start_delalloc_inodes(root, 0);
+	ret = btrfs_start_all_delalloc_inodes(root->fs_info, 0);
 	if (ret) {
 		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 		return ret;
 	}
-	btrfs_wait_ordered_extents(root, 0);
+	btrfs_wait_all_ordered_extents(root->fs_info);
 
 	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
@@ -535,10 +535,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list);
 
 	btrfs_rm_dev_replace_srcdev(fs_info, src_device);
-	if (src_device->bdev) {
-		/* zero out the old super */
-		btrfs_scratch_superblock(src_device);
-	}
+
 	/*
 	 * this is again a consistent state where no dev_replace procedure
 	 * is running, the target device is part of the filesystem, the
@@ -747,7 +744,7 @@ int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
 	WARN_ON(atomic_xchg(
 		&fs_info->mutually_exclusive_operation_running, 1));
 	task = kthread_run(btrfs_dev_replace_kthread, fs_info, "btrfs-devrepl");
-	return PTR_RET(task);
+	return PTR_ERR_OR_ZERO(task);
 }
 
 static int btrfs_dev_replace_kthread(void *data)
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index e7b3cb5286a5..4ae17ed13b32 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -31,6 +31,7 @@
 #include <linux/migrate.h>
 #include <linux/ratelimit.h>
 #include <linux/uuid.h>
+#include <linux/semaphore.h>
 #include <asm/unaligned.h>
 #include "compat.h"
 #include "ctree.h"
@@ -156,6 +157,7 @@ static struct btrfs_lockdep_keyset {
 	{ .id = BTRFS_TREE_LOG_OBJECTID,	.name_stem = "log"	},
 	{ .id = BTRFS_TREE_RELOC_OBJECTID,	.name_stem = "treloc"	},
 	{ .id = BTRFS_DATA_RELOC_TREE_OBJECTID,	.name_stem = "dreloc"	},
+	{ .id = BTRFS_UUID_TREE_OBJECTID,	.name_stem = "uuid"	},
 	{ .id = 0,				.name_stem = "tree"	},
 };
 
@@ -302,9 +304,8 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 			printk_ratelimited(KERN_INFO "btrfs: %s checksum verify "
 				       "failed on %llu wanted %X found %X "
 				       "level %d\n",
-				       root->fs_info->sb->s_id,
-				       (unsigned long long)buf->start, val, found,
-				       btrfs_header_level(buf));
+				       root->fs_info->sb->s_id, buf->start,
+				       val, found, btrfs_header_level(buf));
 			if (result != (char *)&inline_result)
 				kfree(result);
 			return 1;
@@ -345,9 +346,7 @@ static int verify_parent_transid(struct extent_io_tree *io_tree,
 	}
 	printk_ratelimited("parent transid verify failed on %llu wanted %llu "
 		       "found %llu\n",
-		       (unsigned long long)eb->start,
-		       (unsigned long long)parent_transid,
-		       (unsigned long long)btrfs_header_generation(eb));
+		       eb->start, parent_transid, btrfs_header_generation(eb));
 	ret = 1;
 	clear_extent_buffer_uptodate(eb);
 out:
@@ -497,8 +496,7 @@ static int check_tree_block_fsid(struct btrfs_root *root,
 	u8 fsid[BTRFS_UUID_SIZE];
 	int ret = 1;
 
-	read_extent_buffer(eb, fsid, (unsigned long)btrfs_header_fsid(eb),
-			   BTRFS_FSID_SIZE);
+	read_extent_buffer(eb, fsid, btrfs_header_fsid(eb), BTRFS_FSID_SIZE);
 	while (fs_devices) {
 		if (!memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE)) {
 			ret = 0;
@@ -512,8 +510,7 @@ static int check_tree_block_fsid(struct btrfs_root *root,
 #define CORRUPT(reason, eb, root, slot)				\
 	printk(KERN_CRIT "btrfs: corrupt leaf, %s: block=%llu,"	\
 	       "root=%llu, slot=%d\n", reason,			\
-	       (unsigned long long)btrfs_header_bytenr(eb),	\
-	       (unsigned long long)root->objectid, slot)
+	       btrfs_header_bytenr(eb),	root->objectid, slot)
 
 static noinline int check_leaf(struct btrfs_root *root,
 			       struct extent_buffer *leaf)
@@ -576,8 +573,9 @@ static noinline int check_leaf(struct btrfs_root *root,
 	return 0;
 }
 
-static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
-			       struct extent_state *state, int mirror)
+static int btree_readpage_end_io_hook(struct btrfs_io_bio *io_bio,
+				      u64 phy_offset, struct page *page,
+				      u64 start, u64 end, int mirror)
 {
 	struct extent_io_tree *tree;
 	u64 found_start;
@@ -612,14 +610,13 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 	if (found_start != eb->start) {
 		printk_ratelimited(KERN_INFO "btrfs bad tree block start "
 			       "%llu %llu\n",
-			       (unsigned long long)found_start,
-			       (unsigned long long)eb->start);
+			       found_start, eb->start);
 		ret = -EIO;
 		goto err;
 	}
 	if (check_tree_block_fsid(root, eb)) {
 		printk_ratelimited(KERN_INFO "btrfs bad fsid on block %llu\n",
-			       (unsigned long long)eb->start);
+			       eb->start);
 		ret = -EIO;
 		goto err;
 	}
@@ -1013,7 +1010,8 @@ static int btree_releasepage(struct page *page, gfp_t gfp_flags)
 	return try_release_extent_buffer(page);
 }
 
-static void btree_invalidatepage(struct page *page, unsigned long offset)
+static void btree_invalidatepage(struct page *page, unsigned int offset,
+				 unsigned int length)
 {
 	struct extent_io_tree *tree;
 	tree = &BTRFS_I(page->mapping->host)->io_tree;
@@ -1147,6 +1145,10 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
 		return NULL;
 
 	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
+	if (ret) {
+		free_extent_buffer(buf);
+		return NULL;
+	}
 	return buf;
 
 }
@@ -1191,6 +1193,8 @@ static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	root->objectid = objectid;
 	root->last_trans = 0;
 	root->highest_objectid = 0;
+	root->nr_delalloc_inodes = 0;
+	root->nr_ordered_extents = 0;
 	root->name = NULL;
 	root->inode_tree = RB_ROOT;
 	INIT_RADIX_TREE(&root->delayed_nodes_tree, GFP_ATOMIC);
@@ -1199,10 +1203,16 @@ static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 
 	INIT_LIST_HEAD(&root->dirty_list);
 	INIT_LIST_HEAD(&root->root_list);
+	INIT_LIST_HEAD(&root->delalloc_inodes);
+	INIT_LIST_HEAD(&root->delalloc_root);
+	INIT_LIST_HEAD(&root->ordered_extents);
+	INIT_LIST_HEAD(&root->ordered_root);
 	INIT_LIST_HEAD(&root->logged_list[0]);
 	INIT_LIST_HEAD(&root->logged_list[1]);
 	spin_lock_init(&root->orphan_lock);
 	spin_lock_init(&root->inode_lock);
+	spin_lock_init(&root->delalloc_lock);
+	spin_lock_init(&root->ordered_extent_lock);
 	spin_lock_init(&root->accounting_lock);
 	spin_lock_init(&root->log_extents_lock[0]);
 	spin_lock_init(&root->log_extents_lock[1]);
@@ -1216,6 +1226,7 @@ static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	atomic_set(&root->log_writers, 0);
 	atomic_set(&root->log_batch, 0);
 	atomic_set(&root->orphan_inodes, 0);
+	atomic_set(&root->refs, 1);
 	root->log_transid = 0;
 	root->last_log_commit = 0;
 	extent_io_tree_init(&root->dirty_log_pages,
@@ -1234,39 +1245,6 @@ static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	spin_lock_init(&root->root_item_lock);
 }
 
-static int __must_check find_and_setup_root(struct btrfs_root *tree_root,
-					    struct btrfs_fs_info *fs_info,
-					    u64 objectid,
-					    struct btrfs_root *root)
-{
-	int ret;
-	u32 blocksize;
-	u64 generation;
-
-	__setup_root(tree_root->nodesize, tree_root->leafsize,
-		     tree_root->sectorsize, tree_root->stripesize,
-		     root, fs_info, objectid);
-	ret = btrfs_find_last_root(tree_root, objectid,
-				   &root->root_item, &root->root_key);
-	if (ret > 0)
-		return -ENOENT;
-	else if (ret < 0)
-		return ret;
-
-	generation = btrfs_root_generation(&root->root_item);
-	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
-	root->commit_root = NULL;
-	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
-				     blocksize, generation);
-	if (!root->node || !btrfs_buffer_uptodate(root->node, generation, 0)) {
-		free_extent_buffer(root->node);
-		root->node = NULL;
-		return -EIO;
-	}
-	root->commit_root = btrfs_root_node(root);
-	return 0;
-}
-
 static struct btrfs_root *btrfs_alloc_root(struct btrfs_fs_info *fs_info)
 {
 	struct btrfs_root *root = kzalloc(sizeof(*root), GFP_NOFS);
@@ -1314,11 +1292,10 @@ struct btrfs_root *btrfs_create_tree(struct btrfs_trans_handle *trans,
 	btrfs_set_header_owner(leaf, objectid);
 	root->node = leaf;
 
-	write_extent_buffer(leaf, fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(leaf),
+	write_extent_buffer(leaf, fs_info->fsid, btrfs_header_fsid(leaf),
 			    BTRFS_FSID_SIZE);
 	write_extent_buffer(leaf, fs_info->chunk_tree_uuid,
-			    (unsigned long)btrfs_header_chunk_tree_uuid(leaf),
+			    btrfs_header_chunk_tree_uuid(leaf),
 			    BTRFS_UUID_SIZE);
 	btrfs_mark_buffer_dirty(leaf);
 
@@ -1402,8 +1379,7 @@ static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans,
 	root->node = leaf;
 
 	write_extent_buffer(root->node, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(root->node),
-			    BTRFS_FSID_SIZE);
+			    btrfs_header_fsid(root->node), BTRFS_FSID_SIZE);
 	btrfs_mark_buffer_dirty(root->node);
 	btrfs_tree_unlock(root->node);
 	return root;
@@ -1436,11 +1412,11 @@ int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
 	log_root->root_key.offset = root->root_key.objectid;
 
 	inode_item = &log_root->root_item.inode;
-	inode_item->generation = cpu_to_le64(1);
-	inode_item->size = cpu_to_le64(3);
-	inode_item->nlink = cpu_to_le32(1);
-	inode_item->nbytes = cpu_to_le64(root->leafsize);
-	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
+	btrfs_set_stack_inode_generation(inode_item, 1);
+	btrfs_set_stack_inode_size(inode_item, 3);
+	btrfs_set_stack_inode_nlink(inode_item, 1);
+	btrfs_set_stack_inode_nbytes(inode_item, root->leafsize);
+	btrfs_set_stack_inode_mode(inode_item, S_IFDIR | 0755);
 
 	btrfs_set_root_node(&log_root->root_item, log_root->node);
 
@@ -1451,70 +1427,73 @@ int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
-struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
-					       struct btrfs_key *location)
+static struct btrfs_root *btrfs_read_tree_root(struct btrfs_root *tree_root,
+					       struct btrfs_key *key)
 {
 	struct btrfs_root *root;
 	struct btrfs_fs_info *fs_info = tree_root->fs_info;
 	struct btrfs_path *path;
-	struct extent_buffer *l;
 	u64 generation;
 	u32 blocksize;
-	int ret = 0;
-	int slot;
+	int ret;
 
-	root = btrfs_alloc_root(fs_info);
-	if (!root)
+	path = btrfs_alloc_path();
+	if (!path)
 		return ERR_PTR(-ENOMEM);
-	if (location->offset == (u64)-1) {
-		ret = find_and_setup_root(tree_root, fs_info,
-					  location->objectid, root);
-		if (ret) {
-			kfree(root);
-			return ERR_PTR(ret);
-		}
-		goto out;
+
+	root = btrfs_alloc_root(fs_info);
+	if (!root) {
+		ret = -ENOMEM;
+		goto alloc_fail;
 	}
 
 	__setup_root(tree_root->nodesize, tree_root->leafsize,
 		     tree_root->sectorsize, tree_root->stripesize,
-		     root, fs_info, location->objectid);
+		     root, fs_info, key->objectid);
 
-	path = btrfs_alloc_path();
-	if (!path) {
-		kfree(root);
-		return ERR_PTR(-ENOMEM);
-	}
-	ret = btrfs_search_slot(NULL, tree_root, location, path, 0, 0);
-	if (ret == 0) {
-		l = path->nodes[0];
-		slot = path->slots[0];
-		btrfs_read_root_item(l, slot, &root->root_item);
-		memcpy(&root->root_key, location, sizeof(*location));
-	}
-	btrfs_free_path(path);
+	ret = btrfs_find_root(tree_root, key, path,
+			      &root->root_item, &root->root_key);
 	if (ret) {
-		kfree(root);
 		if (ret > 0)
 			ret = -ENOENT;
-		return ERR_PTR(ret);
+		goto find_fail;
 	}
 
 	generation = btrfs_root_generation(&root->root_item);
 	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
 	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
 				     blocksize, generation);
-	if (!root->node || !extent_buffer_uptodate(root->node)) {
-		ret = (!root->node) ? -ENOMEM : -EIO;
-
-		free_extent_buffer(root->node);
-		kfree(root);
-		return ERR_PTR(ret);
+	if (!root->node) {
+		ret = -ENOMEM;
+		goto find_fail;
+	} else if (!btrfs_buffer_uptodate(root->node, generation, 0)) {
+		ret = -EIO;
+		goto read_fail;
 	}
-
 	root->commit_root = btrfs_root_node(root);
 out:
-	if (location->objectid != BTRFS_TREE_LOG_OBJECTID) {
+	btrfs_free_path(path);
+	return root;
+
+read_fail:
+	free_extent_buffer(root->node);
+find_fail:
+	kfree(root);
+alloc_fail:
+	root = ERR_PTR(ret);
+	goto out;
+}
+
+struct btrfs_root *btrfs_read_fs_root(struct btrfs_root *tree_root,
+				      struct btrfs_key *location)
+{
+	struct btrfs_root *root;
+
+	root = btrfs_read_tree_root(tree_root, location);
+	if (IS_ERR(root))
+		return root;
+
+	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
 		root->ref_cows = 1;
 		btrfs_check_and_init_root_item(&root->root_item);
 	}
@@ -1522,6 +1501,66 @@ out:
 	return root;
 }
 
+int btrfs_init_fs_root(struct btrfs_root *root)
+{
+	int ret;
+
+	root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS);
+	root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned),
+					GFP_NOFS);
+	if (!root->free_ino_pinned || !root->free_ino_ctl) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	btrfs_init_free_ino_ctl(root);
+	mutex_init(&root->fs_commit_mutex);
+	spin_lock_init(&root->cache_lock);
+	init_waitqueue_head(&root->cache_wait);
+
+	ret = get_anon_bdev(&root->anon_dev);
+	if (ret)
+		goto fail;
+	return 0;
+fail:
+	kfree(root->free_ino_ctl);
+	kfree(root->free_ino_pinned);
+	return ret;
+}
+
+static struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
+					       u64 root_id)
+{
+	struct btrfs_root *root;
+
+	spin_lock(&fs_info->fs_roots_radix_lock);
+	root = radix_tree_lookup(&fs_info->fs_roots_radix,
+				 (unsigned long)root_id);
+	spin_unlock(&fs_info->fs_roots_radix_lock);
+	return root;
+}
+
+int btrfs_insert_fs_root(struct btrfs_fs_info *fs_info,
+			 struct btrfs_root *root)
+{
+	int ret;
+
+	ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
+	if (ret)
+		return ret;
+
+	spin_lock(&fs_info->fs_roots_radix_lock);
+	ret = radix_tree_insert(&fs_info->fs_roots_radix,
+				(unsigned long)root->root_key.objectid,
+				root);
+	if (ret == 0)
+		root->in_radix = 1;
+	spin_unlock(&fs_info->fs_roots_radix_lock);
+	radix_tree_preload_end();
+
+	return ret;
+}
+
 struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
 					      struct btrfs_key *location)
 {
@@ -1541,59 +1580,37 @@ struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
 	if (location->objectid == BTRFS_QUOTA_TREE_OBJECTID)
 		return fs_info->quota_root ? fs_info->quota_root :
 					     ERR_PTR(-ENOENT);
+	if (location->objectid == BTRFS_UUID_TREE_OBJECTID)
+		return fs_info->uuid_root ? fs_info->uuid_root :
+					    ERR_PTR(-ENOENT);
 again:
-	spin_lock(&fs_info->fs_roots_radix_lock);
-	root = radix_tree_lookup(&fs_info->fs_roots_radix,
-				 (unsigned long)location->objectid);
-	spin_unlock(&fs_info->fs_roots_radix_lock);
-	if (root)
+	root = btrfs_lookup_fs_root(fs_info, location->objectid);
+	if (root) {
+		if (btrfs_root_refs(&root->root_item) == 0)
+			return ERR_PTR(-ENOENT);
 		return root;
+	}
 
-	root = btrfs_read_fs_root_no_radix(fs_info->tree_root, location);
+	root = btrfs_read_fs_root(fs_info->tree_root, location);
 	if (IS_ERR(root))
 		return root;
 
-	root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS);
-	root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned),
-					GFP_NOFS);
-	if (!root->free_ino_pinned || !root->free_ino_ctl) {
-		ret = -ENOMEM;
+	if (btrfs_root_refs(&root->root_item) == 0) {
+		ret = -ENOENT;
 		goto fail;
 	}
 
-	btrfs_init_free_ino_ctl(root);
-	mutex_init(&root->fs_commit_mutex);
-	spin_lock_init(&root->cache_lock);
-	init_waitqueue_head(&root->cache_wait);
-
-	ret = get_anon_bdev(&root->anon_dev);
+	ret = btrfs_init_fs_root(root);
 	if (ret)
 		goto fail;
 
-	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;
-
-	spin_lock(&fs_info->fs_roots_radix_lock);
-	ret = radix_tree_insert(&fs_info->fs_roots_radix,
-				(unsigned long)root->root_key.objectid,
-				root);
-	if (ret == 0)
-		root->in_radix = 1;
-
-	spin_unlock(&fs_info->fs_roots_radix_lock);
-	radix_tree_preload_end();
+	ret = btrfs_insert_fs_root(fs_info, root);
 	if (ret) {
 		if (ret == -EEXIST) {
 			free_fs_root(root);
@@ -1601,10 +1618,6 @@ again:
 		}
 		goto fail;
 	}
-
-	ret = btrfs_find_dead_roots(fs_info->tree_root,
-				    root->root_key.objectid);
-	WARN_ON(ret);
 	return root;
 fail:
 	free_fs_root(root);
@@ -1676,21 +1689,37 @@ static void end_workqueue_fn(struct btrfs_work *work)
 static int cleaner_kthread(void *arg)
 {
 	struct btrfs_root *root = arg;
+	int again;
 
 	do {
-		int again = 0;
-
-		if (!(root->fs_info->sb->s_flags & MS_RDONLY) &&
-		    down_read_trylock(&root->fs_info->sb->s_umount)) {
-			if (mutex_trylock(&root->fs_info->cleaner_mutex)) {
-				btrfs_run_delayed_iputs(root);
-				again = btrfs_clean_one_deleted_snapshot(root);
-				mutex_unlock(&root->fs_info->cleaner_mutex);
-			}
-			btrfs_run_defrag_inodes(root->fs_info);
-			up_read(&root->fs_info->sb->s_umount);
+		again = 0;
+
+		/* Make the cleaner go to sleep early. */
+		if (btrfs_need_cleaner_sleep(root))
+			goto sleep;
+
+		if (!mutex_trylock(&root->fs_info->cleaner_mutex))
+			goto sleep;
+
+		/*
+		 * Avoid the problem that we change the status of the fs
+		 * during the above check and trylock.
+		 */
+		if (btrfs_need_cleaner_sleep(root)) {
+			mutex_unlock(&root->fs_info->cleaner_mutex);
+			goto sleep;
 		}
 
+		btrfs_run_delayed_iputs(root);
+		again = btrfs_clean_one_deleted_snapshot(root);
+		mutex_unlock(&root->fs_info->cleaner_mutex);
+
+		/*
+		 * The defragger has dealt with the R/O remount and umount,
+		 * needn't do anything special here.
+		 */
+		btrfs_run_defrag_inodes(root->fs_info);
+sleep:
 		if (!try_to_freeze() && !again) {
 			set_current_state(TASK_INTERRUPTIBLE);
 			if (!kthread_should_stop())
@@ -1713,7 +1742,7 @@ static int transaction_kthread(void *arg)
 
 	do {
 		cannot_commit = false;
-		delay = HZ * 30;
+		delay = HZ * root->fs_info->commit_interval;
 		mutex_lock(&root->fs_info->transaction_kthread_mutex);
 
 		spin_lock(&root->fs_info->trans_lock);
@@ -1724,8 +1753,9 @@ static int transaction_kthread(void *arg)
 		}
 
 		now = get_seconds();
-		if (!cur->blocked &&
-		    (now < cur->start_time || now - cur->start_time < 30)) {
+		if (cur->state < TRANS_STATE_BLOCKED &&
+		    (now < cur->start_time ||
+		     now - cur->start_time < root->fs_info->commit_interval)) {
 			spin_unlock(&root->fs_info->trans_lock);
 			delay = HZ * 5;
 			goto sleep;
@@ -2014,6 +2044,12 @@ static void free_root_pointers(struct btrfs_fs_info *info, int chunk_root)
 		info->quota_root->node = NULL;
 		info->quota_root->commit_root = NULL;
 	}
+	if (info->uuid_root) {
+		free_extent_buffer(info->uuid_root->node);
+		free_extent_buffer(info->uuid_root->commit_root);
+		info->uuid_root->node = NULL;
+		info->uuid_root->commit_root = NULL;
+	}
 	if (chunk_root) {
 		free_extent_buffer(info->chunk_root->node);
 		free_extent_buffer(info->chunk_root->commit_root);
@@ -2034,11 +2070,11 @@ static void del_fs_roots(struct btrfs_fs_info *fs_info)
 		list_del(&gang[0]->root_list);
 
 		if (gang[0]->in_radix) {
-			btrfs_free_fs_root(fs_info, gang[0]);
+			btrfs_drop_and_free_fs_root(fs_info, gang[0]);
 		} else {
 			free_extent_buffer(gang[0]->node);
 			free_extent_buffer(gang[0]->commit_root);
-			kfree(gang[0]);
+			btrfs_put_fs_root(gang[0]);
 		}
 	}
 
@@ -2049,7 +2085,7 @@ static void del_fs_roots(struct btrfs_fs_info *fs_info)
 		if (!ret)
 			break;
 		for (i = 0; i < ret; i++)
-			btrfs_free_fs_root(fs_info, gang[i]);
+			btrfs_drop_and_free_fs_root(fs_info, gang[i]);
 	}
 }
 
@@ -2074,21 +2110,18 @@ int open_ctree(struct super_block *sb,
 	struct btrfs_root *chunk_root;
 	struct btrfs_root *dev_root;
 	struct btrfs_root *quota_root;
+	struct btrfs_root *uuid_root;
 	struct btrfs_root *log_tree_root;
 	int ret;
 	int err = -EINVAL;
 	int num_backups_tried = 0;
 	int backup_index = 0;
+	bool create_uuid_tree;
+	bool check_uuid_tree;
 
 	tree_root = fs_info->tree_root = btrfs_alloc_root(fs_info);
-	extent_root = fs_info->extent_root = btrfs_alloc_root(fs_info);
-	csum_root = fs_info->csum_root = btrfs_alloc_root(fs_info);
 	chunk_root = fs_info->chunk_root = btrfs_alloc_root(fs_info);
-	dev_root = fs_info->dev_root = btrfs_alloc_root(fs_info);
-	quota_root = fs_info->quota_root = btrfs_alloc_root(fs_info);
-
-	if (!tree_root || !extent_root || !csum_root ||
-	    !chunk_root || !dev_root || !quota_root) {
+	if (!tree_root || !chunk_root) {
 		err = -ENOMEM;
 		goto fail;
 	}
@@ -2131,9 +2164,9 @@ int open_ctree(struct super_block *sb,
 	INIT_LIST_HEAD(&fs_info->trans_list);
 	INIT_LIST_HEAD(&fs_info->dead_roots);
 	INIT_LIST_HEAD(&fs_info->delayed_iputs);
-	INIT_LIST_HEAD(&fs_info->delalloc_inodes);
+	INIT_LIST_HEAD(&fs_info->delalloc_roots);
 	INIT_LIST_HEAD(&fs_info->caching_block_groups);
-	spin_lock_init(&fs_info->delalloc_lock);
+	spin_lock_init(&fs_info->delalloc_root_lock);
 	spin_lock_init(&fs_info->trans_lock);
 	spin_lock_init(&fs_info->fs_roots_radix_lock);
 	spin_lock_init(&fs_info->delayed_iput_lock);
@@ -2169,9 +2202,9 @@ int open_ctree(struct super_block *sb,
 	fs_info->max_inline = 8192 * 1024;
 	fs_info->metadata_ratio = 0;
 	fs_info->defrag_inodes = RB_ROOT;
-	fs_info->trans_no_join = 0;
 	fs_info->free_chunk_space = 0;
 	fs_info->tree_mod_log = RB_ROOT;
+	fs_info->commit_interval = BTRFS_DEFAULT_COMMIT_INTERVAL;
 
 	/* readahead state */
 	INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_WAIT);
@@ -2180,8 +2213,8 @@ int open_ctree(struct super_block *sb,
 	fs_info->thread_pool_size = min_t(unsigned long,
 					  num_online_cpus() + 2, 8);
 
-	INIT_LIST_HEAD(&fs_info->ordered_extents);
-	spin_lock_init(&fs_info->ordered_extent_lock);
+	INIT_LIST_HEAD(&fs_info->ordered_roots);
+	spin_lock_init(&fs_info->ordered_root_lock);
 	fs_info->delayed_root = kmalloc(sizeof(struct btrfs_delayed_root),
 					GFP_NOFS);
 	if (!fs_info->delayed_root) {
@@ -2253,6 +2286,7 @@ int open_ctree(struct super_block *sb,
 
 
 	mutex_init(&fs_info->ordered_operations_mutex);
+	mutex_init(&fs_info->ordered_extent_flush_mutex);
 	mutex_init(&fs_info->tree_log_mutex);
 	mutex_init(&fs_info->chunk_mutex);
 	mutex_init(&fs_info->transaction_kthread_mutex);
@@ -2261,6 +2295,7 @@ int open_ctree(struct super_block *sb,
 	init_rwsem(&fs_info->extent_commit_sem);
 	init_rwsem(&fs_info->cleanup_work_sem);
 	init_rwsem(&fs_info->subvol_sem);
+	sema_init(&fs_info->uuid_tree_rescan_sem, 1);
 	fs_info->dev_replace.lock_owner = 0;
 	atomic_set(&fs_info->dev_replace.nesting_level, 0);
 	mutex_init(&fs_info->dev_replace.lock_finishing_cancel_unmount);
@@ -2274,6 +2309,7 @@ int open_ctree(struct super_block *sb,
 	fs_info->qgroup_seq = 1;
 	fs_info->quota_enabled = 0;
 	fs_info->pending_quota_state = 0;
+	fs_info->qgroup_ulist = NULL;
 	mutex_init(&fs_info->qgroup_rescan_lock);
 
 	btrfs_init_free_cluster(&fs_info->meta_alloc_cluster);
@@ -2365,7 +2401,7 @@ int open_ctree(struct super_block *sb,
 	if (features) {
 		printk(KERN_ERR "BTRFS: couldn't mount because of "
 		       "unsupported optional features (%Lx).\n",
-		       (unsigned long long)features);
+		       features);
 		err = -EINVAL;
 		goto fail_alloc;
 	}
@@ -2435,7 +2471,7 @@ int open_ctree(struct super_block *sb,
 	if (!(sb->s_flags & MS_RDONLY) && features) {
 		printk(KERN_ERR "BTRFS: couldn't mount RDWR because of "
 		       "unsupported option features (%Lx).\n",
-		       (unsigned long long)features);
+		       features);
 		err = -EINVAL;
 		goto fail_alloc;
 	}
@@ -2448,20 +2484,17 @@ int open_ctree(struct super_block *sb,
 			   &fs_info->generic_worker);
 
 	btrfs_init_workers(&fs_info->delalloc_workers, "delalloc",
-			   fs_info->thread_pool_size,
-			   &fs_info->generic_worker);
+			   fs_info->thread_pool_size, NULL);
 
 	btrfs_init_workers(&fs_info->flush_workers, "flush_delalloc",
-			   fs_info->thread_pool_size,
-			   &fs_info->generic_worker);
+			   fs_info->thread_pool_size, NULL);
 
 	btrfs_init_workers(&fs_info->submit_workers, "submit",
 			   min_t(u64, fs_devices->num_devices,
-			   fs_info->thread_pool_size),
-			   &fs_info->generic_worker);
+			   fs_info->thread_pool_size), NULL);
 
 	btrfs_init_workers(&fs_info->caching_workers, "cache",
-			   2, &fs_info->generic_worker);
+			   fs_info->thread_pool_size, NULL);
 
 	/* 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
@@ -2557,7 +2590,7 @@ int open_ctree(struct super_block *sb,
 	sb->s_blocksize = sectorsize;
 	sb->s_blocksize_bits = blksize_bits(sectorsize);
 
-	if (disk_super->magic != cpu_to_le64(BTRFS_MAGIC)) {
+	if (btrfs_super_magic(disk_super) != BTRFS_MAGIC) {
 		printk(KERN_INFO "btrfs: valid FS not found on %s\n", sb->s_id);
 		goto fail_sb_buffer;
 	}
@@ -2597,8 +2630,7 @@ int open_ctree(struct super_block *sb,
 	chunk_root->commit_root = btrfs_root_node(chunk_root);
 
 	read_extent_buffer(chunk_root->node, fs_info->chunk_tree_uuid,
-	   (unsigned long)btrfs_header_chunk_tree_uuid(chunk_root->node),
-	   BTRFS_UUID_SIZE);
+	   btrfs_header_chunk_tree_uuid(chunk_root->node), BTRFS_UUID_SIZE);
 
 	ret = btrfs_read_chunk_tree(chunk_root);
 	if (ret) {
@@ -2638,33 +2670,60 @@ retry_root_backup:
 	btrfs_set_root_node(&tree_root->root_item, tree_root->node);
 	tree_root->commit_root = btrfs_root_node(tree_root);
 
-	ret = find_and_setup_root(tree_root, fs_info,
-				  BTRFS_EXTENT_TREE_OBJECTID, extent_root);
-	if (ret)
+	location.objectid = BTRFS_EXTENT_TREE_OBJECTID;
+	location.type = BTRFS_ROOT_ITEM_KEY;
+	location.offset = 0;
+
+	extent_root = btrfs_read_tree_root(tree_root, &location);
+	if (IS_ERR(extent_root)) {
+		ret = PTR_ERR(extent_root);
 		goto recovery_tree_root;
+	}
 	extent_root->track_dirty = 1;
+	fs_info->extent_root = extent_root;
 
-	ret = find_and_setup_root(tree_root, fs_info,
-				  BTRFS_DEV_TREE_OBJECTID, dev_root);
-	if (ret)
+	location.objectid = BTRFS_DEV_TREE_OBJECTID;
+	dev_root = btrfs_read_tree_root(tree_root, &location);
+	if (IS_ERR(dev_root)) {
+		ret = PTR_ERR(dev_root);
 		goto recovery_tree_root;
+	}
 	dev_root->track_dirty = 1;
+	fs_info->dev_root = dev_root;
+	btrfs_init_devices_late(fs_info);
 
-	ret = find_and_setup_root(tree_root, fs_info,
-				  BTRFS_CSUM_TREE_OBJECTID, csum_root);
-	if (ret)
+	location.objectid = BTRFS_CSUM_TREE_OBJECTID;
+	csum_root = btrfs_read_tree_root(tree_root, &location);
+	if (IS_ERR(csum_root)) {
+		ret = PTR_ERR(csum_root);
 		goto recovery_tree_root;
+	}
 	csum_root->track_dirty = 1;
+	fs_info->csum_root = csum_root;
 
-	ret = find_and_setup_root(tree_root, fs_info,
-				  BTRFS_QUOTA_TREE_OBJECTID, quota_root);
-	if (ret) {
-		kfree(quota_root);
-		quota_root = fs_info->quota_root = NULL;
-	} else {
+	location.objectid = BTRFS_QUOTA_TREE_OBJECTID;
+	quota_root = btrfs_read_tree_root(tree_root, &location);
+	if (!IS_ERR(quota_root)) {
 		quota_root->track_dirty = 1;
 		fs_info->quota_enabled = 1;
 		fs_info->pending_quota_state = 1;
+		fs_info->quota_root = quota_root;
+	}
+
+	location.objectid = BTRFS_UUID_TREE_OBJECTID;
+	uuid_root = btrfs_read_tree_root(tree_root, &location);
+	if (IS_ERR(uuid_root)) {
+		ret = PTR_ERR(uuid_root);
+		if (ret != -ENOENT)
+			goto recovery_tree_root;
+		create_uuid_tree = true;
+		check_uuid_tree = false;
+	} else {
+		uuid_root->track_dirty = 1;
+		fs_info->uuid_root = uuid_root;
+		create_uuid_tree = false;
+		check_uuid_tree =
+		    generation != btrfs_super_uuid_tree_generation(disk_super);
 	}
 
 	fs_info->generation = generation;
@@ -2817,11 +2876,9 @@ retry_root_backup:
 
 	location.objectid = BTRFS_FS_TREE_OBJECTID;
 	location.type = BTRFS_ROOT_ITEM_KEY;
-	location.offset = (u64)-1;
+	location.offset = 0;
 
 	fs_info->fs_root = btrfs_read_fs_root_no_name(fs_info, &location);
-	if (!fs_info->fs_root)
-		goto fail_qgroup;
 	if (IS_ERR(fs_info->fs_root)) {
 		err = PTR_ERR(fs_info->fs_root);
 		goto fail_qgroup;
@@ -2853,14 +2910,39 @@ retry_root_backup:
 		return ret;
 	}
 
+	btrfs_qgroup_rescan_resume(fs_info);
+
+	if (create_uuid_tree) {
+		pr_info("btrfs: creating UUID tree\n");
+		ret = btrfs_create_uuid_tree(fs_info);
+		if (ret) {
+			pr_warn("btrfs: failed to create the UUID tree %d\n",
+				ret);
+			close_ctree(tree_root);
+			return ret;
+		}
+	} else if (check_uuid_tree ||
+		   btrfs_test_opt(tree_root, RESCAN_UUID_TREE)) {
+		pr_info("btrfs: checking UUID tree\n");
+		ret = btrfs_check_uuid_tree(fs_info);
+		if (ret) {
+			pr_warn("btrfs: failed to check the UUID tree %d\n",
+				ret);
+			close_ctree(tree_root);
+			return ret;
+		}
+	} else {
+		fs_info->update_uuid_tree_gen = 1;
+	}
+
 	return 0;
 
 fail_qgroup:
 	btrfs_free_qgroup_config(fs_info);
 fail_trans_kthread:
 	kthread_stop(fs_info->transaction_kthread);
-	del_fs_roots(fs_info);
 	btrfs_cleanup_transaction(fs_info->tree_root);
+	del_fs_roots(fs_info);
 fail_cleaner:
 	kthread_stop(fs_info->cleaner_kthread);
 
@@ -2954,15 +3036,17 @@ struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
 	 */
 	for (i = 0; i < 1; i++) {
 		bytenr = btrfs_sb_offset(i);
-		if (bytenr + 4096 >= i_size_read(bdev->bd_inode))
+		if (bytenr + BTRFS_SUPER_INFO_SIZE >=
+					i_size_read(bdev->bd_inode))
 			break;
-		bh = __bread(bdev, bytenr / 4096, 4096);
+		bh = __bread(bdev, bytenr / 4096,
+					BTRFS_SUPER_INFO_SIZE);
 		if (!bh)
 			continue;
 
 		super = (struct btrfs_super_block *)bh->b_data;
 		if (btrfs_super_bytenr(super) != bytenr ||
-		    super->magic != cpu_to_le64(BTRFS_MAGIC)) {
+		    btrfs_super_magic(super) != BTRFS_MAGIC) {
 			brelse(bh);
 			continue;
 		}
@@ -3258,7 +3342,7 @@ int btrfs_calc_num_tolerated_disk_barrier_failures(
 					    BTRFS_BLOCK_GROUP_RAID10)) {
 						num_tolerated_disk_barrier_failures = 1;
 					} else if (flags &
-						   BTRFS_BLOCK_GROUP_RAID5) {
+						   BTRFS_BLOCK_GROUP_RAID6) {
 						num_tolerated_disk_barrier_failures = 2;
 					}
 				}
@@ -3282,7 +3366,6 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors)
 	int total_errors = 0;
 	u64 flags;
 
-	max_errors = btrfs_super_num_devices(root->fs_info->super_copy) - 1;
 	do_barriers = !btrfs_test_opt(root, NOBARRIER);
 	backup_super_roots(root->fs_info);
 
@@ -3291,6 +3374,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors)
 
 	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
 	head = &root->fs_info->fs_devices->devices;
+	max_errors = btrfs_super_num_devices(root->fs_info->super_copy) - 1;
 
 	if (do_barriers) {
 		ret = barrier_all_devices(root->fs_info);
@@ -3332,9 +3416,12 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors)
 	if (total_errors > max_errors) {
 		printk(KERN_ERR "btrfs: %d errors while writing supers\n",
 		       total_errors);
+		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
-		/* This shouldn't happen. FUA is masked off if unsupported */
-		BUG();
+		/* FUA is masked off if unsupported and can't be the reason */
+		btrfs_error(root->fs_info, -EIO,
+			    "%d errors while writing supers", total_errors);
+		return -EIO;
 	}
 
 	total_errors = 0;
@@ -3366,7 +3453,9 @@ int write_ctree_super(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-void btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
+/* Drop a fs root from the radix tree and free it. */
+void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info,
+				  struct btrfs_root *root)
 {
 	spin_lock(&fs_info->fs_roots_radix_lock);
 	radix_tree_delete(&fs_info->fs_roots_radix,
@@ -3390,6 +3479,8 @@ static void free_fs_root(struct btrfs_root *root)
 {
 	iput(root->cache_inode);
 	WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
+	btrfs_free_block_rsv(root, root->orphan_block_rsv);
+	root->orphan_block_rsv = NULL;
 	if (root->anon_dev)
 		free_anon_bdev(root->anon_dev);
 	free_extent_buffer(root->node);
@@ -3397,7 +3488,12 @@ static void free_fs_root(struct btrfs_root *root)
 	kfree(root->free_ino_ctl);
 	kfree(root->free_ino_pinned);
 	kfree(root->name);
-	kfree(root);
+	btrfs_put_fs_root(root);
+}
+
+void btrfs_free_fs_root(struct btrfs_root *root)
+{
+	free_fs_root(root);
 }
 
 int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
@@ -3474,6 +3570,11 @@ int close_ctree(struct btrfs_root *root)
 	fs_info->closing = 1;
 	smp_mb();
 
+	/* wait for the uuid_scan task to finish */
+	down(&fs_info->uuid_tree_rescan_sem);
+	/* avoid complains from lockdep et al., set sem back to initial state */
+	up(&fs_info->uuid_tree_rescan_sem);
+
 	/* pause restriper - we want to resume on mount */
 	btrfs_pause_balance(fs_info);
 
@@ -3512,15 +3613,15 @@ int close_ctree(struct btrfs_root *root)
 		       percpu_counter_sum(&fs_info->delalloc_bytes));
 	}
 
-	free_root_pointers(fs_info, 1);
-
 	btrfs_free_block_groups(fs_info);
 
+	btrfs_stop_all_workers(fs_info);
+
 	del_fs_roots(fs_info);
 
-	iput(fs_info->btree_inode);
+	free_root_pointers(fs_info, 1);
 
-	btrfs_stop_all_workers(fs_info);
+	iput(fs_info->btree_inode);
 
 #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
 	if (btrfs_test_opt(root, CHECK_INTEGRITY))
@@ -3537,6 +3638,9 @@ int close_ctree(struct btrfs_root *root)
 
 	btrfs_free_stripe_hash_table(fs_info);
 
+	btrfs_free_block_rsv(root, root->orphan_block_rsv);
+	root->orphan_block_rsv = NULL;
+
 	return 0;
 }
 
@@ -3572,9 +3676,7 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
 	if (transid != root->fs_info->generation)
 		WARN(1, KERN_CRIT "btrfs transid mismatch buffer %llu, "
 		       "found %llu running %llu\n",
-			(unsigned long long)buf->start,
-			(unsigned long long)transid,
-			(unsigned long long)root->fs_info->generation);
+			buf->start, transid, root->fs_info->generation);
 	was_dirty = set_extent_buffer_dirty(buf);
 	if (!was_dirty)
 		__percpu_counter_add(&root->fs_info->dirty_metadata_bytes,
@@ -3653,7 +3755,7 @@ static void btrfs_destroy_ordered_operations(struct btrfs_transaction *t,
 	INIT_LIST_HEAD(&splice);
 
 	mutex_lock(&root->fs_info->ordered_operations_mutex);
-	spin_lock(&root->fs_info->ordered_extent_lock);
+	spin_lock(&root->fs_info->ordered_root_lock);
 
 	list_splice_init(&t->ordered_operations, &splice);
 	while (!list_empty(&splice)) {
@@ -3661,14 +3763,14 @@ static void btrfs_destroy_ordered_operations(struct btrfs_transaction *t,
 					 ordered_operations);
 
 		list_del_init(&btrfs_inode->ordered_operations);
-		spin_unlock(&root->fs_info->ordered_extent_lock);
+		spin_unlock(&root->fs_info->ordered_root_lock);
 
 		btrfs_invalidate_inodes(btrfs_inode->root);
 
-		spin_lock(&root->fs_info->ordered_extent_lock);
+		spin_lock(&root->fs_info->ordered_root_lock);
 	}
 
-	spin_unlock(&root->fs_info->ordered_extent_lock);
+	spin_unlock(&root->fs_info->ordered_root_lock);
 	mutex_unlock(&root->fs_info->ordered_operations_mutex);
 }
 
@@ -3676,19 +3778,40 @@ static void btrfs_destroy_ordered_extents(struct btrfs_root *root)
 {
 	struct btrfs_ordered_extent *ordered;
 
-	spin_lock(&root->fs_info->ordered_extent_lock);
+	spin_lock(&root->ordered_extent_lock);
 	/*
 	 * This will just short circuit the ordered completion stuff which will
 	 * make sure the ordered extent gets properly cleaned up.
 	 */
-	list_for_each_entry(ordered, &root->fs_info->ordered_extents,
+	list_for_each_entry(ordered, &root->ordered_extents,
 			    root_extent_list)
 		set_bit(BTRFS_ORDERED_IOERR, &ordered->flags);
-	spin_unlock(&root->fs_info->ordered_extent_lock);
+	spin_unlock(&root->ordered_extent_lock);
 }
 
-int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
-			       struct btrfs_root *root)
+static void btrfs_destroy_all_ordered_extents(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_root *root;
+	struct list_head splice;
+
+	INIT_LIST_HEAD(&splice);
+
+	spin_lock(&fs_info->ordered_root_lock);
+	list_splice_init(&fs_info->ordered_roots, &splice);
+	while (!list_empty(&splice)) {
+		root = list_first_entry(&splice, struct btrfs_root,
+					ordered_root);
+		list_del_init(&root->ordered_root);
+
+		btrfs_destroy_ordered_extents(root);
+
+		cond_resched_lock(&fs_info->ordered_root_lock);
+	}
+	spin_unlock(&fs_info->ordered_root_lock);
+}
+
+static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
+				      struct btrfs_root *root)
 {
 	struct rb_node *node;
 	struct btrfs_delayed_ref_root *delayed_refs;
@@ -3706,6 +3829,7 @@ int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 
 	while ((node = rb_first(&delayed_refs->root)) != NULL) {
 		struct btrfs_delayed_ref_head *head = NULL;
+		bool pin_bytes = false;
 
 		ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
 		atomic_set(&ref->refs, 1);
@@ -3726,8 +3850,7 @@ int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 			}
 
 			if (head->must_insert_reserved)
-				btrfs_pin_extent(root, ref->bytenr,
-						 ref->num_bytes, 1);
+				pin_bytes = true;
 			btrfs_free_delayed_extent_op(head->extent_op);
 			delayed_refs->num_heads--;
 			if (list_empty(&head->cluster))
@@ -3738,9 +3861,13 @@ int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 		ref->in_tree = 0;
 		rb_erase(&ref->rb_node, &delayed_refs->root);
 		delayed_refs->num_entries--;
-		if (head)
-			mutex_unlock(&head->mutex);
 		spin_unlock(&delayed_refs->lock);
+		if (head) {
+			if (pin_bytes)
+				btrfs_pin_extent(root, ref->bytenr,
+						 ref->num_bytes, 1);
+			mutex_unlock(&head->mutex);
+		}
 		btrfs_put_delayed_ref(ref);
 
 		cond_resched();
@@ -3777,24 +3904,49 @@ static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 
 	INIT_LIST_HEAD(&splice);
 
-	spin_lock(&root->fs_info->delalloc_lock);
-	list_splice_init(&root->fs_info->delalloc_inodes, &splice);
+	spin_lock(&root->delalloc_lock);
+	list_splice_init(&root->delalloc_inodes, &splice);
 
 	while (!list_empty(&splice)) {
-		btrfs_inode = list_entry(splice.next, struct btrfs_inode,
-				    delalloc_inodes);
+		btrfs_inode = list_first_entry(&splice, struct btrfs_inode,
+					       delalloc_inodes);
 
 		list_del_init(&btrfs_inode->delalloc_inodes);
 		clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
 			  &btrfs_inode->runtime_flags);
-		spin_unlock(&root->fs_info->delalloc_lock);
+		spin_unlock(&root->delalloc_lock);
 
 		btrfs_invalidate_inodes(btrfs_inode->root);
 
-		spin_lock(&root->fs_info->delalloc_lock);
+		spin_lock(&root->delalloc_lock);
 	}
 
-	spin_unlock(&root->fs_info->delalloc_lock);
+	spin_unlock(&root->delalloc_lock);
+}
+
+static void btrfs_destroy_all_delalloc_inodes(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_root *root;
+	struct list_head splice;
+
+	INIT_LIST_HEAD(&splice);
+
+	spin_lock(&fs_info->delalloc_root_lock);
+	list_splice_init(&fs_info->delalloc_roots, &splice);
+	while (!list_empty(&splice)) {
+		root = list_first_entry(&splice, struct btrfs_root,
+					 delalloc_root);
+		list_del_init(&root->delalloc_root);
+		root = btrfs_grab_fs_root(root);
+		BUG_ON(!root);
+		spin_unlock(&fs_info->delalloc_root_lock);
+
+		btrfs_destroy_delalloc_inodes(root);
+		btrfs_put_fs_root(root);
+
+		spin_lock(&fs_info->delalloc_root_lock);
+	}
+	spin_unlock(&fs_info->delalloc_root_lock);
 }
 
 static int btrfs_destroy_marked_extents(struct btrfs_root *root,
@@ -3878,19 +4030,14 @@ void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
 	btrfs_block_rsv_release(root, &root->fs_info->trans_block_rsv,
 				cur_trans->dirty_pages.dirty_bytes);
 
-	/* FIXME: cleanup wait for commit */
-	cur_trans->in_commit = 1;
-	cur_trans->blocked = 1;
+	cur_trans->state = TRANS_STATE_COMMIT_START;
 	wake_up(&root->fs_info->transaction_blocked_wait);
 
 	btrfs_evict_pending_snapshots(cur_trans);
 
-	cur_trans->blocked = 0;
+	cur_trans->state = TRANS_STATE_UNBLOCKED;
 	wake_up(&root->fs_info->transaction_wait);
 
-	cur_trans->commit_done = 1;
-	wake_up(&cur_trans->commit_wait);
-
 	btrfs_destroy_delayed_inodes(root);
 	btrfs_assert_delayed_root_empty(root);
 
@@ -3899,6 +4046,9 @@ void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
 	btrfs_destroy_pinned_extent(root,
 				    root->fs_info->pinned_extents);
 
+	cur_trans->state =TRANS_STATE_COMPLETED;
+	wake_up(&cur_trans->commit_wait);
+
 	/*
 	memset(cur_trans, 0, sizeof(*cur_trans));
 	kmem_cache_free(btrfs_transaction_cachep, cur_trans);
@@ -3914,7 +4064,7 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
 
 	spin_lock(&root->fs_info->trans_lock);
 	list_splice_init(&root->fs_info->trans_list, &list);
-	root->fs_info->trans_no_join = 1;
+	root->fs_info->running_transaction = NULL;
 	spin_unlock(&root->fs_info->trans_lock);
 
 	while (!list_empty(&list)) {
@@ -3922,37 +4072,31 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
 
 		btrfs_destroy_ordered_operations(t, root);
 
-		btrfs_destroy_ordered_extents(root);
+		btrfs_destroy_all_ordered_extents(root->fs_info);
 
 		btrfs_destroy_delayed_refs(t, root);
 
-		/* FIXME: cleanup wait for commit */
-		t->in_commit = 1;
-		t->blocked = 1;
+		/*
+		 *  FIXME: cleanup wait for commit
+		 *  We needn't acquire the lock here, because we are during
+		 *  the umount, there is no other task which will change it.
+		 */
+		t->state = TRANS_STATE_COMMIT_START;
 		smp_mb();
 		if (waitqueue_active(&root->fs_info->transaction_blocked_wait))
 			wake_up(&root->fs_info->transaction_blocked_wait);
 
 		btrfs_evict_pending_snapshots(t);
 
-		t->blocked = 0;
+		t->state = TRANS_STATE_UNBLOCKED;
 		smp_mb();
 		if (waitqueue_active(&root->fs_info->transaction_wait))
 			wake_up(&root->fs_info->transaction_wait);
 
-		t->commit_done = 1;
-		smp_mb();
-		if (waitqueue_active(&t->commit_wait))
-			wake_up(&t->commit_wait);
-
 		btrfs_destroy_delayed_inodes(root);
 		btrfs_assert_delayed_root_empty(root);
 
-		btrfs_destroy_delalloc_inodes(root);
-
-		spin_lock(&root->fs_info->trans_lock);
-		root->fs_info->running_transaction = NULL;
-		spin_unlock(&root->fs_info->trans_lock);
+		btrfs_destroy_all_delalloc_inodes(root->fs_info);
 
 		btrfs_destroy_marked_extents(root, &t->dirty_pages,
 					     EXTENT_DIRTY);
@@ -3960,15 +4104,17 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
 		btrfs_destroy_pinned_extent(root,
 					    root->fs_info->pinned_extents);
 
+		t->state = TRANS_STATE_COMPLETED;
+		smp_mb();
+		if (waitqueue_active(&t->commit_wait))
+			wake_up(&t->commit_wait);
+
 		atomic_set(&t->use_count, 0);
 		list_del_init(&t->list);
 		memset(t, 0, sizeof(*t));
 		kmem_cache_free(btrfs_transaction_cachep, t);
 	}
 
-	spin_lock(&root->fs_info->trans_lock);
-	root->fs_info->trans_no_join = 0;
-	spin_unlock(&root->fs_info->trans_lock);
 	mutex_unlock(&root->fs_info->transaction_kthread_mutex);
 
 	return 0;
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index be69ce1b07a2..b71acd6e1e5b 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -63,14 +63,40 @@ struct buffer_head *btrfs_read_dev_super(struct block_device *bdev);
 int btrfs_commit_super(struct btrfs_root *root);
 struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
 					    u64 bytenr, u32 blocksize);
-struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
-					       struct btrfs_key *location);
+struct btrfs_root *btrfs_read_fs_root(struct btrfs_root *tree_root,
+				      struct btrfs_key *location);
+int btrfs_init_fs_root(struct btrfs_root *root);
+int btrfs_insert_fs_root(struct btrfs_fs_info *fs_info,
+			 struct btrfs_root *root);
 struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
 					      struct btrfs_key *location);
 int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info);
 void btrfs_btree_balance_dirty(struct btrfs_root *root);
 void btrfs_btree_balance_dirty_nodelay(struct btrfs_root *root);
-void btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root);
+void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info,
+				 struct btrfs_root *root);
+void btrfs_free_fs_root(struct btrfs_root *root);
+
+/*
+ * This function is used to grab the root, and avoid it is freed when we
+ * access it. But it doesn't ensure that the tree is not dropped.
+ *
+ * If you want to ensure the whole tree is safe, you should use
+ * 	fs_info->subvol_srcu
+ */
+static inline struct btrfs_root *btrfs_grab_fs_root(struct btrfs_root *root)
+{
+	if (atomic_inc_not_zero(&root->refs))
+		return root;
+	return NULL;
+}
+
+static inline void btrfs_put_fs_root(struct btrfs_root *root)
+{
+	if (atomic_dec_and_test(&root->refs))
+		kfree(root);
+}
+
 void btrfs_mark_buffer_dirty(struct extent_buffer *buf);
 int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid,
 			  int atomic);
diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c
index 81ee29eeb7ca..4b8691607373 100644
--- a/fs/btrfs/export.c
+++ b/fs/btrfs/export.c
@@ -82,11 +82,6 @@ static struct dentry *btrfs_get_dentry(struct super_block *sb, u64 objectid,
 		goto fail;
 	}
 
-	if (btrfs_root_refs(&root->root_item) == 0) {
-		err = -ENOENT;
-		goto fail;
-	}
-
 	key.objectid = objectid;
 	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
 	key.offset = 0;
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index df472ab1b5ac..d58bef130a41 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -24,6 +24,7 @@
 #include <linux/kthread.h>
 #include <linux/slab.h>
 #include <linux/ratelimit.h>
+#include <linux/percpu_counter.h>
 #include "compat.h"
 #include "hash.h"
 #include "ctree.h"
@@ -112,7 +113,8 @@ static noinline int
 block_group_cache_done(struct btrfs_block_group_cache *cache)
 {
 	smp_mb();
-	return cache->cached == BTRFS_CACHE_FINISHED;
+	return cache->cached == BTRFS_CACHE_FINISHED ||
+		cache->cached == BTRFS_CACHE_ERROR;
 }
 
 static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
@@ -388,7 +390,7 @@ static noinline void caching_thread(struct btrfs_work *work)
 	u64 total_found = 0;
 	u64 last = 0;
 	u32 nritems;
-	int ret = 0;
+	int ret = -ENOMEM;
 
 	caching_ctl = container_of(work, struct btrfs_caching_control, work);
 	block_group = caching_ctl->block_group;
@@ -419,6 +421,7 @@ again:
 	/* need to make sure the commit_root doesn't disappear */
 	down_read(&fs_info->extent_commit_sem);
 
+next:
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
 		goto err;
@@ -458,6 +461,16 @@ again:
 			continue;
 		}
 
+		if (key.objectid < last) {
+			key.objectid = last;
+			key.offset = 0;
+			key.type = BTRFS_EXTENT_ITEM_KEY;
+
+			caching_ctl->progress = last;
+			btrfs_release_path(path);
+			goto next;
+		}
+
 		if (key.objectid < block_group->key.objectid) {
 			path->slots[0]++;
 			continue;
@@ -505,6 +518,12 @@ err:
 
 	mutex_unlock(&caching_ctl->mutex);
 out:
+	if (ret) {
+		spin_lock(&block_group->lock);
+		block_group->caching_ctl = NULL;
+		block_group->cached = BTRFS_CACHE_ERROR;
+		spin_unlock(&block_group->lock);
+	}
 	wake_up(&caching_ctl->wait);
 
 	put_caching_control(caching_ctl);
@@ -770,10 +789,23 @@ again:
 		goto out_free;
 
 	if (ret > 0 && metadata && key.type == BTRFS_METADATA_ITEM_KEY) {
-		key.type = BTRFS_EXTENT_ITEM_KEY;
-		key.offset = root->leafsize;
-		btrfs_release_path(path);
-		goto again;
+		metadata = 0;
+		if (path->slots[0]) {
+			path->slots[0]--;
+			btrfs_item_key_to_cpu(path->nodes[0], &key,
+					      path->slots[0]);
+			if (key.objectid == bytenr &&
+			    key.type == BTRFS_EXTENT_ITEM_KEY &&
+			    key.offset == root->leafsize)
+				ret = 0;
+		}
+		if (ret) {
+			key.objectid = bytenr;
+			key.type = BTRFS_EXTENT_ITEM_KEY;
+			key.offset = root->leafsize;
+			btrfs_release_path(path);
+			goto again;
+		}
 	}
 
 	if (ret == 0) {
@@ -2010,6 +2042,8 @@ static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
 	ins.type = BTRFS_EXTENT_ITEM_KEY;
 
 	ref = btrfs_delayed_node_to_data_ref(node);
+	trace_run_delayed_data_ref(node, ref, node->action);
+
 	if (node->type == BTRFS_SHARED_DATA_REF_KEY)
 		parent = ref->parent;
 	else
@@ -2153,6 +2187,8 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
 						 SKINNY_METADATA);
 
 	ref = btrfs_delayed_node_to_tree_ref(node);
+	trace_run_delayed_tree_ref(node, ref, node->action);
+
 	if (node->type == BTRFS_SHARED_BLOCK_REF_KEY)
 		parent = ref->parent;
 	else
@@ -2211,6 +2247,8 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 		 */
 		BUG_ON(extent_op);
 		head = btrfs_delayed_node_to_head(node);
+		trace_run_delayed_ref_head(node, head, node->action);
+
 		if (insert_reserved) {
 			btrfs_pin_extent(root, node->bytenr,
 					 node->num_bytes, 1);
@@ -2402,6 +2440,8 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
 			default:
 				WARN_ON(1);
 			}
+		} else {
+			list_del_init(&locked_ref->cluster);
 		}
 		spin_unlock(&delayed_refs->lock);
 
@@ -2424,7 +2464,6 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
 		 * list before we release it.
 		 */
 		if (btrfs_delayed_ref_is_head(ref)) {
-			list_del_init(&locked_ref->cluster);
 			btrfs_delayed_ref_unlock(locked_ref);
 			locked_ref = NULL;
 		}
@@ -2526,6 +2565,51 @@ static int refs_newer(struct btrfs_delayed_ref_root *delayed_refs, int seq,
 	return 0;
 }
 
+static inline u64 heads_to_leaves(struct btrfs_root *root, u64 heads)
+{
+	u64 num_bytes;
+
+	num_bytes = heads * (sizeof(struct btrfs_extent_item) +
+			     sizeof(struct btrfs_extent_inline_ref));
+	if (!btrfs_fs_incompat(root->fs_info, SKINNY_METADATA))
+		num_bytes += heads * sizeof(struct btrfs_tree_block_info);
+
+	/*
+	 * We don't ever fill up leaves all the way so multiply by 2 just to be
+	 * closer to what we're really going to want to ouse.
+	 */
+	return div64_u64(num_bytes, BTRFS_LEAF_DATA_SIZE(root));
+}
+
+int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans,
+				       struct btrfs_root *root)
+{
+	struct btrfs_block_rsv *global_rsv;
+	u64 num_heads = trans->transaction->delayed_refs.num_heads_ready;
+	u64 num_bytes;
+	int ret = 0;
+
+	num_bytes = btrfs_calc_trans_metadata_size(root, 1);
+	num_heads = heads_to_leaves(root, num_heads);
+	if (num_heads > 1)
+		num_bytes += (num_heads - 1) * root->leafsize;
+	num_bytes <<= 1;
+	global_rsv = &root->fs_info->global_block_rsv;
+
+	/*
+	 * If we can't allocate any more chunks lets make sure we have _lots_ of
+	 * wiggle room since running delayed refs can create more delayed refs.
+	 */
+	if (global_rsv->space_info->full)
+		num_bytes <<= 1;
+
+	spin_lock(&global_rsv->lock);
+	if (global_rsv->reserved <= num_bytes)
+		ret = 1;
+	spin_unlock(&global_rsv->lock);
+	return ret;
+}
+
 /*
  * this starts processing the delayed reference count updates and
  * extent insertions we have queued up so far.  count can be
@@ -2573,7 +2657,8 @@ progress:
 		old = atomic_cmpxchg(&delayed_refs->procs_running_refs, 0, 1);
 		if (old) {
 			DEFINE_WAIT(__wait);
-			if (delayed_refs->num_entries < 16348)
+			if (delayed_refs->flushing ||
+			    !btrfs_should_throttle_delayed_refs(trans, root))
 				return 0;
 
 			prepare_to_wait(&delayed_refs->wait, &__wait,
@@ -2608,7 +2693,7 @@ again:
 
 	while (1) {
 		if (!(run_all || run_most) &&
-		    delayed_refs->num_heads_ready < 64)
+		    !btrfs_should_throttle_delayed_refs(trans, root))
 			break;
 
 		/*
@@ -2629,6 +2714,7 @@ again:
 			spin_unlock(&delayed_refs->lock);
 			btrfs_abort_transaction(trans, root, ret);
 			atomic_dec(&delayed_refs->procs_running_refs);
+			wake_up(&delayed_refs->wait);
 			return ret;
 		}
 
@@ -3310,6 +3396,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	struct btrfs_space_info *found;
 	int i;
 	int factor;
+	int ret;
 
 	if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 |
 		     BTRFS_BLOCK_GROUP_RAID10))
@@ -3333,6 +3420,12 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	if (!found)
 		return -ENOMEM;
 
+	ret = percpu_counter_init(&found->total_bytes_pinned, 0);
+	if (ret) {
+		kfree(found);
+		return ret;
+	}
+
 	for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
 		INIT_LIST_HEAD(&found->block_groups[i]);
 	init_rwsem(&found->groups_sem);
@@ -3565,10 +3658,11 @@ alloc:
 		}
 
 		/*
-		 * If we have less pinned bytes than we want to allocate then
-		 * don't bother committing the transaction, it won't help us.
+		 * If we don't have enough pinned space to deal with this
+		 * allocation don't bother committing the transaction.
 		 */
-		if (data_sinfo->bytes_pinned < bytes)
+		if (percpu_counter_compare(&data_sinfo->total_bytes_pinned,
+					   bytes) < 0)
 			committed = 1;
 		spin_unlock(&data_sinfo->lock);
 
@@ -3577,6 +3671,7 @@ commit_trans:
 		if (!committed &&
 		    !atomic_read(&root->fs_info->open_ioctl_trans)) {
 			committed = 1;
+
 			trans = btrfs_join_transaction(root);
 			if (IS_ERR(trans))
 				return PTR_ERR(trans);
@@ -3609,6 +3704,7 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
 
 	data_sinfo = root->fs_info->data_sinfo;
 	spin_lock(&data_sinfo->lock);
+	WARN_ON(data_sinfo->bytes_may_use < bytes);
 	data_sinfo->bytes_may_use -= bytes;
 	trace_btrfs_space_reservation(root->fs_info, "space_info",
 				      data_sinfo->flags, bytes, 0);
@@ -3741,8 +3837,12 @@ again:
 	if (force < space_info->force_alloc)
 		force = space_info->force_alloc;
 	if (space_info->full) {
+		if (should_alloc_chunk(extent_root, space_info, force))
+			ret = -ENOSPC;
+		else
+			ret = 0;
 		spin_unlock(&space_info->lock);
-		return 0;
+		return ret;
 	}
 
 	if (!should_alloc_chunk(extent_root, space_info, force)) {
@@ -3825,7 +3925,6 @@ static int can_overcommit(struct btrfs_root *root,
 	u64 space_size;
 	u64 avail;
 	u64 used;
-	u64 to_add;
 
 	used = space_info->bytes_used + space_info->bytes_reserved +
 		space_info->bytes_pinned + space_info->bytes_readonly;
@@ -3859,25 +3958,17 @@ static int can_overcommit(struct btrfs_root *root,
 		       BTRFS_BLOCK_GROUP_RAID10))
 		avail >>= 1;
 
-	to_add = space_info->total_bytes;
-
 	/*
 	 * If we aren't flushing all things, let us overcommit up to
 	 * 1/2th of the space. If we can flush, don't let us overcommit
 	 * too much, let it overcommit up to 1/8 of the space.
 	 */
 	if (flush == BTRFS_RESERVE_FLUSH_ALL)
-		to_add >>= 3;
+		avail >>= 3;
 	else
-		to_add >>= 1;
-
-	/*
-	 * Limit the overcommit to the amount of free space we could possibly
-	 * allocate for chunks.
-	 */
-	to_add = min(avail, to_add);
+		avail >>= 1;
 
-	if (used + bytes < space_info->total_bytes + to_add)
+	if (used + bytes < space_info->total_bytes + avail)
 		return 1;
 	return 0;
 }
@@ -3886,12 +3977,11 @@ static void btrfs_writeback_inodes_sb_nr(struct btrfs_root *root,
 					 unsigned long nr_pages)
 {
 	struct super_block *sb = root->fs_info->sb;
-	int started;
 
-	/* If we can not start writeback, just sync all the delalloc file. */
-	started = try_to_writeback_inodes_sb_nr(sb, nr_pages,
-						      WB_REASON_FS_FREE_SPACE);
-	if (!started) {
+	if (down_read_trylock(&sb->s_umount)) {
+		writeback_inodes_sb_nr(sb, nr_pages, WB_REASON_FS_FREE_SPACE);
+		up_read(&sb->s_umount);
+	} else {
 		/*
 		 * We needn't worry the filesystem going from r/w to r/o though
 		 * we don't acquire ->s_umount mutex, because the filesystem
@@ -3899,9 +3989,9 @@ static void btrfs_writeback_inodes_sb_nr(struct btrfs_root *root,
 		 * the filesystem is readonly(all dirty pages are written to
 		 * the disk).
 		 */
-		btrfs_start_delalloc_inodes(root, 0);
+		btrfs_start_all_delalloc_inodes(root->fs_info, 0);
 		if (!current->journal_info)
-			btrfs_wait_ordered_extents(root, 0);
+			btrfs_wait_all_ordered_extents(root->fs_info);
 	}
 }
 
@@ -3931,7 +4021,7 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig,
 	if (delalloc_bytes == 0) {
 		if (trans)
 			return;
-		btrfs_wait_ordered_extents(root, 0);
+		btrfs_wait_all_ordered_extents(root->fs_info);
 		return;
 	}
 
@@ -3959,7 +4049,7 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig,
 
 		loops++;
 		if (wait_ordered && !trans) {
-			btrfs_wait_ordered_extents(root, 0);
+			btrfs_wait_all_ordered_extents(root->fs_info);
 		} else {
 			time_left = schedule_timeout_killable(1);
 			if (time_left)
@@ -3997,7 +4087,8 @@ static int may_commit_transaction(struct btrfs_root *root,
 
 	/* See if there is enough pinned space to make this reservation */
 	spin_lock(&space_info->lock);
-	if (space_info->bytes_pinned >= bytes) {
+	if (percpu_counter_compare(&space_info->total_bytes_pinned,
+				   bytes) >= 0) {
 		spin_unlock(&space_info->lock);
 		goto commit;
 	}
@@ -4012,7 +4103,8 @@ static int may_commit_transaction(struct btrfs_root *root,
 
 	spin_lock(&space_info->lock);
 	spin_lock(&delayed_rsv->lock);
-	if (space_info->bytes_pinned + delayed_rsv->size < bytes) {
+	if (percpu_counter_compare(&space_info->total_bytes_pinned,
+				   bytes - delayed_rsv->size) >= 0) {
 		spin_unlock(&delayed_rsv->lock);
 		spin_unlock(&space_info->lock);
 		return -ENOSPC;
@@ -4261,6 +4353,9 @@ static struct btrfs_block_rsv *get_block_rsv(
 	if (root == root->fs_info->csum_root && trans->adding_csums)
 		block_rsv = trans->block_rsv;
 
+	if (root == root->fs_info->uuid_root)
+		block_rsv = trans->block_rsv;
+
 	if (!block_rsv)
 		block_rsv = root->block_rsv;
 
@@ -4297,6 +4392,31 @@ static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv,
 	spin_unlock(&block_rsv->lock);
 }
 
+int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info,
+			     struct btrfs_block_rsv *dest, u64 num_bytes,
+			     int min_factor)
+{
+	struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv;
+	u64 min_bytes;
+
+	if (global_rsv->space_info != dest->space_info)
+		return -ENOSPC;
+
+	spin_lock(&global_rsv->lock);
+	min_bytes = div_factor(global_rsv->size, min_factor);
+	if (global_rsv->reserved < min_bytes + num_bytes) {
+		spin_unlock(&global_rsv->lock);
+		return -ENOSPC;
+	}
+	global_rsv->reserved -= num_bytes;
+	if (global_rsv->reserved < global_rsv->size)
+		global_rsv->full = 0;
+	spin_unlock(&global_rsv->lock);
+
+	block_rsv_add_bytes(dest, num_bytes, 1);
+	return 0;
+}
+
 static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info,
 				    struct btrfs_block_rsv *block_rsv,
 				    struct btrfs_block_rsv *dest, u64 num_bytes)
@@ -4336,7 +4456,6 @@ static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info,
 			space_info->bytes_may_use -= num_bytes;
 			trace_btrfs_space_reservation(fs_info, "space_info",
 					space_info->flags, num_bytes, 0);
-			space_info->reservation_progress++;
 			spin_unlock(&space_info->lock);
 		}
 	}
@@ -4537,7 +4656,6 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info)
 		sinfo->bytes_may_use -= num_bytes;
 		trace_btrfs_space_reservation(fs_info, "space_info",
 				      sinfo->flags, num_bytes, 0);
-		sinfo->reservation_progress++;
 		block_rsv->reserved = block_rsv->size;
 		block_rsv->full = 1;
 	}
@@ -4645,10 +4763,12 @@ void btrfs_orphan_release_metadata(struct inode *inode)
 int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 				     struct btrfs_block_rsv *rsv,
 				     int items,
-				     u64 *qgroup_reserved)
+				     u64 *qgroup_reserved,
+				     bool use_global_rsv)
 {
 	u64 num_bytes;
 	int ret;
+	struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv;
 
 	if (root->fs_info->quota_enabled) {
 		/* One for parent inode, two for dir entries */
@@ -4667,6 +4787,10 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 					    BTRFS_BLOCK_GROUP_METADATA);
 	ret = btrfs_block_rsv_add(root, rsv, num_bytes,
 				  BTRFS_RESERVE_FLUSH_ALL);
+
+	if (ret == -ENOSPC && use_global_rsv)
+		ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes);
+
 	if (ret) {
 		if (*qgroup_reserved)
 			btrfs_qgroup_free(root, *qgroup_reserved);
@@ -5030,14 +5154,14 @@ static int update_block_group(struct btrfs_root *root,
 	int factor;
 
 	/* block accounting for super block */
-	spin_lock(&info->delalloc_lock);
+	spin_lock(&info->delalloc_root_lock);
 	old_val = btrfs_super_bytes_used(info->super_copy);
 	if (alloc)
 		old_val += num_bytes;
 	else
 		old_val -= num_bytes;
 	btrfs_set_super_bytes_used(info->super_copy, old_val);
-	spin_unlock(&info->delalloc_lock);
+	spin_unlock(&info->delalloc_root_lock);
 
 	while (total) {
 		cache = btrfs_lookup_block_group(info, bytenr);
@@ -5189,6 +5313,80 @@ int btrfs_pin_extent_for_log_replay(struct btrfs_root *root,
 	return ret;
 }
 
+static int __exclude_logged_extent(struct btrfs_root *root, u64 start, u64 num_bytes)
+{
+	int ret;
+	struct btrfs_block_group_cache *block_group;
+	struct btrfs_caching_control *caching_ctl;
+
+	block_group = btrfs_lookup_block_group(root->fs_info, start);
+	if (!block_group)
+		return -EINVAL;
+
+	cache_block_group(block_group, 0);
+	caching_ctl = get_caching_control(block_group);
+
+	if (!caching_ctl) {
+		/* Logic error */
+		BUG_ON(!block_group_cache_done(block_group));
+		ret = btrfs_remove_free_space(block_group, start, num_bytes);
+	} else {
+		mutex_lock(&caching_ctl->mutex);
+
+		if (start >= caching_ctl->progress) {
+			ret = add_excluded_extent(root, start, num_bytes);
+		} else if (start + num_bytes <= caching_ctl->progress) {
+			ret = btrfs_remove_free_space(block_group,
+						      start, num_bytes);
+		} else {
+			num_bytes = caching_ctl->progress - start;
+			ret = btrfs_remove_free_space(block_group,
+						      start, num_bytes);
+			if (ret)
+				goto out_lock;
+
+			num_bytes = (start + num_bytes) -
+				caching_ctl->progress;
+			start = caching_ctl->progress;
+			ret = add_excluded_extent(root, start, num_bytes);
+		}
+out_lock:
+		mutex_unlock(&caching_ctl->mutex);
+		put_caching_control(caching_ctl);
+	}
+	btrfs_put_block_group(block_group);
+	return ret;
+}
+
+int btrfs_exclude_logged_extents(struct btrfs_root *log,
+				 struct extent_buffer *eb)
+{
+	struct btrfs_file_extent_item *item;
+	struct btrfs_key key;
+	int found_type;
+	int i;
+
+	if (!btrfs_fs_incompat(log->fs_info, MIXED_GROUPS))
+		return 0;
+
+	for (i = 0; i < btrfs_header_nritems(eb); i++) {
+		btrfs_item_key_to_cpu(eb, &key, i);
+		if (key.type != BTRFS_EXTENT_DATA_KEY)
+			continue;
+		item = btrfs_item_ptr(eb, i, struct btrfs_file_extent_item);
+		found_type = btrfs_file_extent_type(eb, item);
+		if (found_type == BTRFS_FILE_EXTENT_INLINE)
+			continue;
+		if (btrfs_file_extent_disk_bytenr(eb, item) == 0)
+			continue;
+		key.objectid = btrfs_file_extent_disk_bytenr(eb, item);
+		key.offset = btrfs_file_extent_disk_num_bytes(eb, item);
+		__exclude_logged_extent(log, key.objectid, key.offset);
+	}
+
+	return 0;
+}
+
 /**
  * btrfs_update_reserved_bytes - update the block_group and space info counters
  * @cache:	The cache we are manipulating
@@ -5237,7 +5435,6 @@ static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
 			space_info->bytes_readonly += num_bytes;
 		cache->reserved -= num_bytes;
 		space_info->bytes_reserved -= num_bytes;
-		space_info->reservation_progress++;
 	}
 	spin_unlock(&cache->lock);
 	spin_unlock(&space_info->lock);
@@ -5251,6 +5448,7 @@ void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 	struct btrfs_caching_control *next;
 	struct btrfs_caching_control *caching_ctl;
 	struct btrfs_block_group_cache *cache;
+	struct btrfs_space_info *space_info;
 
 	down_write(&fs_info->extent_commit_sem);
 
@@ -5273,6 +5471,9 @@ void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 
 	up_write(&fs_info->extent_commit_sem);
 
+	list_for_each_entry_rcu(space_info, &fs_info->space_info, list)
+		percpu_counter_set(&space_info->total_bytes_pinned, 0);
+
 	update_global_block_rsv(fs_info);
 }
 
@@ -5370,6 +5571,27 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
+static void add_pinned_bytes(struct btrfs_fs_info *fs_info, u64 num_bytes,
+			     u64 owner, u64 root_objectid)
+{
+	struct btrfs_space_info *space_info;
+	u64 flags;
+
+	if (owner < BTRFS_FIRST_FREE_OBJECTID) {
+		if (root_objectid == BTRFS_CHUNK_TREE_OBJECTID)
+			flags = BTRFS_BLOCK_GROUP_SYSTEM;
+		else
+			flags = BTRFS_BLOCK_GROUP_METADATA;
+	} else {
+		flags = BTRFS_BLOCK_GROUP_DATA;
+	}
+
+	space_info = __find_space_info(fs_info, flags);
+	BUG_ON(!space_info); /* Logic bug */
+	percpu_counter_add(&space_info->total_bytes_pinned, num_bytes);
+}
+
+
 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root,
 				u64 bytenr, u64 num_bytes, u64 parent,
@@ -5485,7 +5707,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 
 			if (ret) {
 				btrfs_err(info, "umm, got %d back from search, was looking for %llu",
-					ret, (unsigned long long)bytenr);
+					ret, bytenr);
 				if (ret > 0)
 					btrfs_print_leaf(extent_root,
 							 path->nodes[0]);
@@ -5501,11 +5723,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 		WARN_ON(1);
 		btrfs_err(info,
 			"unable to find ref byte nr %llu parent %llu root %llu  owner %llu offset %llu",
-			(unsigned long long)bytenr,
-			(unsigned long long)parent,
-			(unsigned long long)root_objectid,
-			(unsigned long long)owner_objectid,
-			(unsigned long long)owner_offset);
+			bytenr, parent, root_objectid, owner_objectid,
+			owner_offset);
 	} else {
 		btrfs_abort_transaction(trans, extent_root, ret);
 		goto out;
@@ -5534,7 +5753,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 					-1, 1);
 		if (ret) {
 			btrfs_err(info, "umm, got %d back from search, was looking for %llu",
-				ret, (unsigned long long)bytenr);
+				ret, bytenr);
 			btrfs_print_leaf(extent_root, path->nodes[0]);
 		}
 		if (ret < 0) {
@@ -5590,6 +5809,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 				goto out;
 			}
 		}
+		add_pinned_bytes(root->fs_info, -num_bytes, owner_objectid,
+				 root_objectid);
 	} else {
 		if (found_extent) {
 			BUG_ON(is_data && refs_to_drop !=
@@ -5713,6 +5934,7 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
 			   u64 parent, int last_ref)
 {
 	struct btrfs_block_group_cache *cache = NULL;
+	int pin = 1;
 	int ret;
 
 	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
@@ -5745,8 +5967,14 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
 
 		btrfs_add_free_space(cache, buf->start, buf->len);
 		btrfs_update_reserved_bytes(cache, buf->len, RESERVE_FREE);
+		pin = 0;
 	}
 out:
+	if (pin)
+		add_pinned_bytes(root->fs_info, buf->len,
+				 btrfs_header_level(buf),
+				 root->root_key.objectid);
+
 	/*
 	 * Deleting the buffer, clear the corrupt flag since it doesn't matter
 	 * anymore.
@@ -5763,6 +5991,8 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	int ret;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
+	add_pinned_bytes(root->fs_info, num_bytes, owner, root_objectid);
+
 	/*
 	 * tree log blocks never actually go into the extent allocation
 	 * tree, just update pinning info and exit early.
@@ -5805,8 +6035,11 @@ static u64 stripe_align(struct btrfs_root *root,
  * for our min num_bytes.  Another option is to have it go ahead
  * and look in the rbtree for a free extent of a given size, but this
  * is a good start.
+ *
+ * Callers of this must check if cache->cached == BTRFS_CACHE_ERROR before using
+ * any of the information in this block group.
  */
-static noinline int
+static noinline void
 wait_block_group_cache_progress(struct btrfs_block_group_cache *cache,
 				u64 num_bytes)
 {
@@ -5814,28 +6047,29 @@ wait_block_group_cache_progress(struct btrfs_block_group_cache *cache,
 
 	caching_ctl = get_caching_control(cache);
 	if (!caching_ctl)
-		return 0;
+		return;
 
 	wait_event(caching_ctl->wait, block_group_cache_done(cache) ||
 		   (cache->free_space_ctl->free_space >= num_bytes));
 
 	put_caching_control(caching_ctl);
-	return 0;
 }
 
 static noinline int
 wait_block_group_cache_done(struct btrfs_block_group_cache *cache)
 {
 	struct btrfs_caching_control *caching_ctl;
+	int ret = 0;
 
 	caching_ctl = get_caching_control(cache);
 	if (!caching_ctl)
-		return 0;
+		return (cache->cached == BTRFS_CACHE_ERROR) ? -EIO : 0;
 
 	wait_event(caching_ctl->wait, block_group_cache_done(cache));
-
+	if (cache->cached == BTRFS_CACHE_ERROR)
+		ret = -EIO;
 	put_caching_control(caching_ctl);
-	return 0;
+	return ret;
 }
 
 int __get_raid_index(u64 flags)
@@ -5871,13 +6105,15 @@ enum btrfs_loop_type {
 /*
  * walks the btree of allocated extents and find a hole of a given size.
  * The key ins is changed to record the hole:
- * ins->objectid == block start
+ * ins->objectid == start position
  * ins->flags = BTRFS_EXTENT_ITEM_KEY
- * ins->offset == number of blocks
+ * ins->offset == the size of the hole.
  * Any available blocks before search_start are skipped.
+ *
+ * If there is no suitable free space, we will record the max size of
+ * the free space extent currently.
  */
-static noinline int find_free_extent(struct btrfs_trans_handle *trans,
-				     struct btrfs_root *orig_root,
+static noinline int find_free_extent(struct btrfs_root *orig_root,
 				     u64 num_bytes, u64 empty_size,
 				     u64 hint_byte, struct btrfs_key *ins,
 				     u64 flags)
@@ -5888,6 +6124,7 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 	struct btrfs_block_group_cache *block_group = NULL;
 	struct btrfs_block_group_cache *used_block_group;
 	u64 search_start = 0;
+	u64 max_extent_size = 0;
 	int empty_cluster = 2 * 1024 * 1024;
 	struct btrfs_space_info *space_info;
 	int loop = 0;
@@ -6018,6 +6255,8 @@ have_block_group:
 			ret = 0;
 		}
 
+		if (unlikely(block_group->cached == BTRFS_CACHE_ERROR))
+			goto loop;
 		if (unlikely(block_group->ro))
 			goto loop;
 
@@ -6045,7 +6284,10 @@ have_block_group:
 				btrfs_get_block_group(used_block_group);
 
 			offset = btrfs_alloc_from_cluster(used_block_group,
-			  last_ptr, num_bytes, used_block_group->key.objectid);
+						last_ptr,
+						num_bytes,
+						used_block_group->key.objectid,
+						&max_extent_size);
 			if (offset) {
 				/* we have a block, we're done */
 				spin_unlock(&last_ptr->refill_lock);
@@ -6098,18 +6340,20 @@ refill_cluster:
 					      block_group->full_stripe_len);
 
 			/* allocate a cluster in this block group */
-			ret = btrfs_find_space_cluster(trans, root,
-					       block_group, last_ptr,
-					       search_start, num_bytes,
-					       aligned_cluster);
+			ret = btrfs_find_space_cluster(root, block_group,
+						       last_ptr, search_start,
+						       num_bytes,
+						       aligned_cluster);
 			if (ret == 0) {
 				/*
 				 * now pull our allocation out of this
 				 * cluster
 				 */
 				offset = btrfs_alloc_from_cluster(block_group,
-						  last_ptr, num_bytes,
-						  search_start);
+							last_ptr,
+							num_bytes,
+							search_start,
+							&max_extent_size);
 				if (offset) {
 					/* we found one, proceed */
 					spin_unlock(&last_ptr->refill_lock);
@@ -6144,13 +6388,18 @@ unclustered_alloc:
 		if (cached &&
 		    block_group->free_space_ctl->free_space <
 		    num_bytes + empty_cluster + empty_size) {
+			if (block_group->free_space_ctl->free_space >
+			    max_extent_size)
+				max_extent_size =
+					block_group->free_space_ctl->free_space;
 			spin_unlock(&block_group->free_space_ctl->tree_lock);
 			goto loop;
 		}
 		spin_unlock(&block_group->free_space_ctl->tree_lock);
 
 		offset = btrfs_find_space_for_alloc(block_group, search_start,
-						    num_bytes, empty_size);
+						    num_bytes, empty_size,
+						    &max_extent_size);
 		/*
 		 * If we didn't find a chunk, and we haven't failed on this
 		 * block group before, and this block group is in the middle of
@@ -6232,17 +6481,28 @@ loop:
 		index = 0;
 		loop++;
 		if (loop == LOOP_ALLOC_CHUNK) {
+			struct btrfs_trans_handle *trans;
+
+			trans = btrfs_join_transaction(root);
+			if (IS_ERR(trans)) {
+				ret = PTR_ERR(trans);
+				goto out;
+			}
+
 			ret = do_chunk_alloc(trans, root, flags,
 					     CHUNK_ALLOC_FORCE);
 			/*
 			 * Do not bail out on ENOSPC since we
 			 * can do more things.
 			 */
-			if (ret < 0 && ret != -ENOSPC) {
+			if (ret < 0 && ret != -ENOSPC)
 				btrfs_abort_transaction(trans,
 							root, ret);
+			else
+				ret = 0;
+			btrfs_end_transaction(trans, root);
+			if (ret)
 				goto out;
-			}
 		}
 
 		if (loop == LOOP_NO_EMPTY_SIZE) {
@@ -6257,7 +6517,8 @@ loop:
 		ret = 0;
 	}
 out:
-
+	if (ret == -ENOSPC)
+		ins->offset = max_extent_size;
 	return ret;
 }
 
@@ -6269,19 +6530,15 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
 
 	spin_lock(&info->lock);
 	printk(KERN_INFO "space_info %llu has %llu free, is %sfull\n",
-	       (unsigned long long)info->flags,
-	       (unsigned long long)(info->total_bytes - info->bytes_used -
-				    info->bytes_pinned - info->bytes_reserved -
-				    info->bytes_readonly),
+	       info->flags,
+	       info->total_bytes - info->bytes_used - info->bytes_pinned -
+	       info->bytes_reserved - info->bytes_readonly,
 	       (info->full) ? "" : "not ");
 	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_reserved,
-	       (unsigned long long)info->bytes_may_use,
-	       (unsigned long long)info->bytes_readonly);
+	       info->total_bytes, info->bytes_used, info->bytes_pinned,
+	       info->bytes_reserved, info->bytes_may_use,
+	       info->bytes_readonly);
 	spin_unlock(&info->lock);
 
 	if (!dump_block_groups)
@@ -6292,12 +6549,9 @@ 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 %s\n",
-		       (unsigned long long)cache->key.objectid,
-		       (unsigned long long)cache->key.offset,
-		       (unsigned long long)btrfs_block_group_used(&cache->item),
-		       (unsigned long long)cache->pinned,
-		       (unsigned long long)cache->reserved,
-		       cache->ro ? "[readonly]" : "");
+		       cache->key.objectid, cache->key.offset,
+		       btrfs_block_group_used(&cache->item), cache->pinned,
+		       cache->reserved, cache->ro ? "[readonly]" : "");
 		btrfs_dump_free_space(cache, bytes);
 		spin_unlock(&cache->lock);
 	}
@@ -6306,8 +6560,7 @@ again:
 	up_read(&info->groups_sem);
 }
 
-int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
-			 struct btrfs_root *root,
+int btrfs_reserve_extent(struct btrfs_root *root,
 			 u64 num_bytes, u64 min_alloc_size,
 			 u64 empty_size, u64 hint_byte,
 			 struct btrfs_key *ins, int is_data)
@@ -6319,12 +6572,12 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 	flags = btrfs_get_alloc_profile(root, is_data);
 again:
 	WARN_ON(num_bytes < root->sectorsize);
-	ret = find_free_extent(trans, root, num_bytes, empty_size,
-			       hint_byte, ins, flags);
+	ret = find_free_extent(root, num_bytes, empty_size, hint_byte, ins,
+			       flags);
 
 	if (ret == -ENOSPC) {
-		if (!final_tried) {
-			num_bytes = num_bytes >> 1;
+		if (!final_tried && ins->offset) {
+			num_bytes = min(num_bytes >> 1, ins->offset);
 			num_bytes = round_down(num_bytes, root->sectorsize);
 			num_bytes = max(num_bytes, min_alloc_size);
 			if (num_bytes == min_alloc_size)
@@ -6335,8 +6588,7 @@ again:
 
 			sinfo = __find_space_info(root->fs_info, flags);
 			btrfs_err(root->fs_info, "allocation failed flags %llu, wanted %llu",
-				(unsigned long long)flags,
-				(unsigned long long)num_bytes);
+				flags, num_bytes);
 			if (sinfo)
 				dump_space_info(sinfo, num_bytes, 1);
 		}
@@ -6356,7 +6608,7 @@ static int __btrfs_free_reserved_extent(struct btrfs_root *root,
 	cache = btrfs_lookup_block_group(root->fs_info, start);
 	if (!cache) {
 		btrfs_err(root->fs_info, "Unable to find block group for %llu",
-			(unsigned long long)start);
+			start);
 		return -ENOSPC;
 	}
 
@@ -6452,8 +6704,7 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 	ret = update_block_group(root, ins->objectid, ins->offset, 1);
 	if (ret) { /* -ENOENT, logic error */
 		btrfs_err(fs_info, "update block group failed for %llu %llu",
-			(unsigned long long)ins->objectid,
-			(unsigned long long)ins->offset);
+			ins->objectid, ins->offset);
 		BUG();
 	}
 	return ret;
@@ -6525,8 +6776,7 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 	ret = update_block_group(root, ins->objectid, root->leafsize, 1);
 	if (ret) { /* -ENOENT, logic error */
 		btrfs_err(fs_info, "update block group failed for %llu %llu",
-			(unsigned long long)ins->objectid,
-			(unsigned long long)ins->offset);
+			ins->objectid, ins->offset);
 		BUG();
 	}
 	return ret;
@@ -6560,52 +6810,26 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
 {
 	int ret;
 	struct btrfs_block_group_cache *block_group;
-	struct btrfs_caching_control *caching_ctl;
-	u64 start = ins->objectid;
-	u64 num_bytes = ins->offset;
 
-	block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
-	cache_block_group(block_group, 0);
-	caching_ctl = get_caching_control(block_group);
-
-	if (!caching_ctl) {
-		BUG_ON(!block_group_cache_done(block_group));
-		ret = btrfs_remove_free_space(block_group, start, num_bytes);
-		if (ret)
-			goto out;
-	} else {
-		mutex_lock(&caching_ctl->mutex);
-
-		if (start >= caching_ctl->progress) {
-			ret = add_excluded_extent(root, start, num_bytes);
-		} else if (start + num_bytes <= caching_ctl->progress) {
-			ret = btrfs_remove_free_space(block_group,
-						      start, num_bytes);
-		} else {
-			num_bytes = caching_ctl->progress - start;
-			ret = btrfs_remove_free_space(block_group,
-						      start, num_bytes);
-			if (ret)
-				goto out_lock;
-
-			start = caching_ctl->progress;
-			num_bytes = ins->objectid + ins->offset -
-				    caching_ctl->progress;
-			ret = add_excluded_extent(root, start, num_bytes);
-		}
-out_lock:
-		mutex_unlock(&caching_ctl->mutex);
-		put_caching_control(caching_ctl);
+	/*
+	 * Mixed block groups will exclude before processing the log so we only
+	 * need to do the exlude dance if this fs isn't mixed.
+	 */
+	if (!btrfs_fs_incompat(root->fs_info, MIXED_GROUPS)) {
+		ret = __exclude_logged_extent(root, ins->objectid, ins->offset);
 		if (ret)
-			goto out;
+			return ret;
 	}
 
+	block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
+	if (!block_group)
+		return -EINVAL;
+
 	ret = btrfs_update_reserved_bytes(block_group, ins->offset,
 					  RESERVE_ALLOC_NO_ACCOUNT);
 	BUG_ON(ret); /* logic error */
 	ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
 					 0, owner, offset, ins, 1);
-out:
 	btrfs_put_block_group(block_group);
 	return ret;
 }
@@ -6734,7 +6958,7 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 	if (IS_ERR(block_rsv))
 		return ERR_CAST(block_rsv);
 
-	ret = btrfs_reserve_extent(trans, root, blocksize, blocksize,
+	ret = btrfs_reserve_extent(root, blocksize, blocksize,
 				   empty_size, hint, &ins, 0);
 	if (ret) {
 		unuse_block_rsv(root->fs_info, block_rsv, blocksize);
@@ -7005,6 +7229,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 		next = btrfs_find_create_tree_block(root, bytenr, blocksize);
 		if (!next)
 			return -ENOMEM;
+		btrfs_set_buffer_lockdep_class(root->root_key.objectid, next,
+					       level - 1);
 		reada = 1;
 	}
 	btrfs_tree_lock(next);
@@ -7298,6 +7524,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 	int err = 0;
 	int ret;
 	int level;
+	bool root_dropped = false;
 
 	path = btrfs_alloc_path();
 	if (!path) {
@@ -7355,6 +7582,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 		while (1) {
 			btrfs_tree_lock(path->nodes[level]);
 			btrfs_set_lock_blocking(path->nodes[level]);
+			path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
 
 			ret = btrfs_lookup_extent_info(trans, root,
 						path->nodes[level]->start,
@@ -7370,6 +7598,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 				break;
 
 			btrfs_tree_unlock(path->nodes[level]);
+			path->locks[level] = 0;
 			WARN_ON(wc->refs[level] != 1);
 			level--;
 		}
@@ -7384,11 +7613,6 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 	wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root);
 
 	while (1) {
-		if (!for_reloc && btrfs_fs_closing(root->fs_info)) {
-			pr_debug("btrfs: drop snapshot early exit\n");
-			err = -EAGAIN;
-			goto out_end_trans;
-		}
 
 		ret = walk_down_tree(trans, root, path, wc);
 		if (ret < 0) {
@@ -7416,7 +7640,8 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 		}
 
 		BUG_ON(wc->level == 0);
-		if (btrfs_should_end_transaction(trans, tree_root)) {
+		if (btrfs_should_end_transaction(trans, tree_root) ||
+		    (!for_reloc && btrfs_need_cleaner_sleep(root))) {
 			ret = btrfs_update_root(trans, tree_root,
 						&root->root_key,
 						root_item);
@@ -7427,6 +7652,12 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 			}
 
 			btrfs_end_transaction_throttle(trans, tree_root);
+			if (!for_reloc && btrfs_need_cleaner_sleep(root)) {
+				pr_debug("btrfs: drop snapshot early exit\n");
+				err = -EAGAIN;
+				goto out_free;
+			}
+
 			trans = btrfs_start_transaction(tree_root, 0);
 			if (IS_ERR(trans)) {
 				err = PTR_ERR(trans);
@@ -7447,8 +7678,8 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 	}
 
 	if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
-		ret = btrfs_find_last_root(tree_root, root->root_key.objectid,
-					   NULL, NULL);
+		ret = btrfs_find_root(tree_root, &root->root_key, path,
+				      NULL, NULL);
 		if (ret < 0) {
 			btrfs_abort_transaction(trans, tree_root, ret);
 			err = ret;
@@ -7465,18 +7696,28 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 	}
 
 	if (root->in_radix) {
-		btrfs_free_fs_root(tree_root->fs_info, root);
+		btrfs_drop_and_free_fs_root(tree_root->fs_info, root);
 	} else {
 		free_extent_buffer(root->node);
 		free_extent_buffer(root->commit_root);
-		kfree(root);
+		btrfs_put_fs_root(root);
 	}
+	root_dropped = true;
 out_end_trans:
 	btrfs_end_transaction_throttle(trans, tree_root);
 out_free:
 	kfree(wc);
 	btrfs_free_path(path);
 out:
+	/*
+	 * So if we need to stop dropping the snapshot for whatever reason we
+	 * need to make sure to add it back to the dead root list so that we
+	 * keep trying to do the work later.  This also cleans up roots if we
+	 * don't have it in the radix (like when we recover after a power fail
+	 * or unmount) so we don't leak memory.
+	 */
+	if (!for_reloc && root_dropped == false)
+		btrfs_add_dead_root(root);
 	if (err)
 		btrfs_std_error(root->fs_info, err);
 	return err;
@@ -7782,6 +8023,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 	struct btrfs_space_info *space_info;
 	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
 	struct btrfs_device *device;
+	struct btrfs_trans_handle *trans;
 	u64 min_free;
 	u64 dev_min = 1;
 	u64 dev_nr = 0;
@@ -7868,6 +8110,13 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 		do_div(min_free, dev_min);
 	}
 
+	/* We need to do this so that we can look at pending chunks */
+	trans = btrfs_join_transaction(root);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		goto out;
+	}
+
 	mutex_lock(&root->fs_info->chunk_mutex);
 	list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) {
 		u64 dev_offset;
@@ -7878,7 +8127,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 		 */
 		if (device->total_bytes > device->bytes_used + min_free &&
 		    !device->is_tgtdev_for_dev_replace) {
-			ret = find_free_dev_extent(device, min_free,
+			ret = find_free_dev_extent(trans, device, min_free,
 						   &dev_offset, NULL);
 			if (!ret)
 				dev_nr++;
@@ -7890,6 +8139,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 		}
 	}
 	mutex_unlock(&root->fs_info->chunk_mutex);
+	btrfs_end_transaction(trans, root);
 out:
 	btrfs_put_block_group(block_group);
 	return ret;
@@ -8000,7 +8250,8 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 		 * We haven't cached this block group, which means we could
 		 * possibly have excluded extents on this block group.
 		 */
-		if (block_group->cached == BTRFS_CACHE_NO)
+		if (block_group->cached == BTRFS_CACHE_NO ||
+		    block_group->cached == BTRFS_CACHE_ERROR)
 			free_excluded_extents(info->extent_root, block_group);
 
 		btrfs_remove_free_space_cache(block_group);
@@ -8032,6 +8283,7 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 				dump_space_info(space_info, 0, 0);
 			}
 		}
+		percpu_counter_destroy(&space_info->total_bytes_pinned);
 		list_del(&space_info->list);
 		kfree(space_info);
 	}
@@ -8216,9 +8468,13 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		 * avoid allocating from un-mirrored block group if there are
 		 * mirrored block groups.
 		 */
-		list_for_each_entry(cache, &space_info->block_groups[3], list)
+		list_for_each_entry(cache,
+				&space_info->block_groups[BTRFS_RAID_RAID0],
+				list)
 			set_block_group_ro(cache, 1);
-		list_for_each_entry(cache, &space_info->block_groups[4], list)
+		list_for_each_entry(cache,
+				&space_info->block_groups[BTRFS_RAID_SINGLE],
+				list)
 			set_block_group_ro(cache, 1);
 	}
 
@@ -8254,6 +8510,10 @@ void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans,
 					sizeof(item));
 		if (ret)
 			btrfs_abort_transaction(trans, extent_root, ret);
+		ret = btrfs_finish_chunk_alloc(trans, extent_root,
+					       key.objectid, key.offset);
+		if (ret)
+			btrfs_abort_transaction(trans, extent_root, ret);
 	}
 }
 
@@ -8591,8 +8851,15 @@ int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range)
 		if (end - start >= range->minlen) {
 			if (!block_group_cache_done(cache)) {
 				ret = cache_block_group(cache, 0);
-				if (!ret)
-					wait_block_group_cache_done(cache);
+				if (ret) {
+					btrfs_put_block_group(cache);
+					break;
+				}
+				ret = wait_block_group_cache_done(cache);
+				if (ret) {
+					btrfs_put_block_group(cache);
+					break;
+				}
 			}
 			ret = btrfs_trim_block_group(cache,
 						     &group_trimmed,
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index e7e7afb4a872..22bda32acb89 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -61,9 +61,8 @@ void btrfs_leak_debug_check(void)
 		state = list_entry(states.next, struct extent_state, leak_list);
 		printk(KERN_ERR "btrfs state leak: start %llu end %llu "
 		       "state %lu in tree %p refs %d\n",
-		       (unsigned long long)state->start,
-		       (unsigned long long)state->end,
-		       state->state, state->tree, atomic_read(&state->refs));
+		       state->start, state->end, state->state, state->tree,
+		       atomic_read(&state->refs));
 		list_del(&state->leak_list);
 		kmem_cache_free(extent_state_cache, state);
 	}
@@ -71,16 +70,31 @@ void btrfs_leak_debug_check(void)
 	while (!list_empty(&buffers)) {
 		eb = list_entry(buffers.next, struct extent_buffer, leak_list);
 		printk(KERN_ERR "btrfs buffer leak start %llu len %lu "
-		       "refs %d\n", (unsigned long long)eb->start,
-		       eb->len, atomic_read(&eb->refs));
+		       "refs %d\n",
+		       eb->start, eb->len, atomic_read(&eb->refs));
 		list_del(&eb->leak_list);
 		kmem_cache_free(extent_buffer_cache, eb);
 	}
 }
+
+#define btrfs_debug_check_extent_io_range(inode, start, end)		\
+	__btrfs_debug_check_extent_io_range(__func__, (inode), (start), (end))
+static inline void __btrfs_debug_check_extent_io_range(const char *caller,
+		struct inode *inode, u64 start, u64 end)
+{
+	u64 isize = i_size_read(inode);
+
+	if (end >= PAGE_SIZE && (end % 2) == 0 && end != isize - 1) {
+		printk_ratelimited(KERN_DEBUG
+		    "btrfs: %s: ino %llu isize %llu odd range [%llu,%llu]\n",
+				caller, btrfs_ino(inode), isize, start, end);
+	}
+}
 #else
 #define btrfs_leak_debug_add(new, head)	do {} while (0)
 #define btrfs_leak_debug_del(entry)	do {} while (0)
 #define btrfs_leak_debug_check()	do {} while (0)
+#define btrfs_debug_check_extent_io_range(c, s, e)	do {} while (0)
 #endif
 
 #define BUFFER_LRU_MAX 64
@@ -131,8 +145,16 @@ int __init extent_io_init(void)
 				     offsetof(struct btrfs_io_bio, bio));
 	if (!btrfs_bioset)
 		goto free_buffer_cache;
+
+	if (bioset_integrity_create(btrfs_bioset, BIO_POOL_SIZE))
+		goto free_bioset;
+
 	return 0;
 
+free_bioset:
+	bioset_free(btrfs_bioset);
+	btrfs_bioset = NULL;
+
 free_buffer_cache:
 	kmem_cache_destroy(extent_buffer_cache);
 	extent_buffer_cache = NULL;
@@ -369,8 +391,7 @@ static int insert_state(struct extent_io_tree *tree,
 
 	if (end < start)
 		WARN(1, KERN_ERR "btrfs end < start %llu %llu\n",
-		       (unsigned long long)end,
-		       (unsigned long long)start);
+		       end, start);
 	state->start = start;
 	state->end = end;
 
@@ -381,9 +402,8 @@ static int insert_state(struct extent_io_tree *tree,
 		struct extent_state *found;
 		found = rb_entry(node, struct extent_state, rb_node);
 		printk(KERN_ERR "btrfs found node %llu %llu on insert of "
-		       "%llu %llu\n", (unsigned long long)found->start,
-		       (unsigned long long)found->end,
-		       (unsigned long long)start, (unsigned long long)end);
+		       "%llu %llu\n",
+		       found->start, found->end, start, end);
 		return -EEXIST;
 	}
 	state->tree = tree;
@@ -522,6 +542,11 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	int err;
 	int clear = 0;
 
+	btrfs_debug_check_extent_io_range(tree->mapping->host, start, end);
+
+	if (bits & EXTENT_DELALLOC)
+		bits |= EXTENT_NORESERVE;
+
 	if (delete)
 		bits |= ~EXTENT_CTLBITS;
 	bits |= EXTENT_FIRST_DELALLOC;
@@ -677,6 +702,8 @@ static void wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	struct extent_state *state;
 	struct rb_node *node;
 
+	btrfs_debug_check_extent_io_range(tree->mapping->host, start, end);
+
 	spin_lock(&tree->lock);
 again:
 	while (1) {
@@ -736,15 +763,6 @@ static void cache_state(struct extent_state *state,
 	}
 }
 
-static void uncache_state(struct extent_state **cached_ptr)
-{
-	if (cached_ptr && (*cached_ptr)) {
-		struct extent_state *state = *cached_ptr;
-		*cached_ptr = NULL;
-		free_extent_state(state);
-	}
-}
-
 /*
  * set some bits on a range in the tree.  This may require allocations or
  * sleeping, so the gfp mask is used to indicate what is allowed.
@@ -769,6 +787,8 @@ __set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	u64 last_start;
 	u64 last_end;
 
+	btrfs_debug_check_extent_io_range(tree->mapping->host, start, end);
+
 	bits |= EXTENT_FIRST_DELALLOC;
 again:
 	if (!prealloc && (mask & __GFP_WAIT)) {
@@ -989,6 +1009,8 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	u64 last_start;
 	u64 last_end;
 
+	btrfs_debug_check_extent_io_range(tree->mapping->host, start, end);
+
 again:
 	if (!prealloc && (mask & __GFP_WAIT)) {
 		prealloc = alloc_extent_state(mask);
@@ -1467,10 +1489,12 @@ static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
 		*end = state->end;
 		cur_start = state->end + 1;
 		node = rb_next(node);
-		if (!node)
-			break;
 		total_bytes += state->end - state->start + 1;
-		if (total_bytes >= max_bytes)
+		if (total_bytes >= max_bytes) {
+			*end = *start + max_bytes - 1;
+			break;
+		}
+		if (!node)
 			break;
 	}
 out:
@@ -1598,7 +1622,7 @@ again:
 		*start = delalloc_start;
 		*end = delalloc_end;
 		free_extent_state(cached_state);
-		return found;
+		return 0;
 	}
 
 	/*
@@ -1657,31 +1681,21 @@ out_failed:
 	return found;
 }
 
-int extent_clear_unlock_delalloc(struct inode *inode,
-				struct extent_io_tree *tree,
-				u64 start, u64 end, struct page *locked_page,
-				unsigned long op)
+int extent_clear_unlock_delalloc(struct inode *inode, u64 start, u64 end,
+				 struct page *locked_page,
+				 unsigned long clear_bits,
+				 unsigned long page_ops)
 {
+	struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree;
 	int ret;
 	struct page *pages[16];
 	unsigned long index = start >> PAGE_CACHE_SHIFT;
 	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
 	unsigned long nr_pages = end_index - index + 1;
 	int i;
-	unsigned long clear_bits = 0;
-
-	if (op & EXTENT_CLEAR_UNLOCK)
-		clear_bits |= EXTENT_LOCKED;
-	if (op & EXTENT_CLEAR_DIRTY)
-		clear_bits |= EXTENT_DIRTY;
-
-	if (op & EXTENT_CLEAR_DELALLOC)
-		clear_bits |= EXTENT_DELALLOC;
 
 	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 |
-		    EXTENT_SET_PRIVATE2)))
+	if (page_ops == 0)
 		return 0;
 
 	while (nr_pages > 0) {
@@ -1690,20 +1704,20 @@ int extent_clear_unlock_delalloc(struct inode *inode,
 				     nr_pages, ARRAY_SIZE(pages)), pages);
 		for (i = 0; i < ret; i++) {
 
-			if (op & EXTENT_SET_PRIVATE2)
+			if (page_ops & PAGE_SET_PRIVATE2)
 				SetPagePrivate2(pages[i]);
 
 			if (pages[i] == locked_page) {
 				page_cache_release(pages[i]);
 				continue;
 			}
-			if (op & EXTENT_CLEAR_DIRTY)
+			if (page_ops & PAGE_CLEAR_DIRTY)
 				clear_page_dirty_for_io(pages[i]);
-			if (op & EXTENT_SET_WRITEBACK)
+			if (page_ops & PAGE_SET_WRITEBACK)
 				set_page_writeback(pages[i]);
-			if (op & EXTENT_END_WRITEBACK)
+			if (page_ops & PAGE_END_WRITEBACK)
 				end_page_writeback(pages[i]);
-			if (op & EXTENT_CLEAR_UNLOCK_PAGE)
+			if (page_ops & PAGE_UNLOCK)
 				unlock_page(pages[i]);
 			page_cache_release(pages[i]);
 		}
@@ -1780,7 +1794,7 @@ out:
  * set the private field for a given byte offset in the tree.  If there isn't
  * an extent_state there already, this does nothing.
  */
-int set_state_private(struct extent_io_tree *tree, u64 start, u64 private)
+static int set_state_private(struct extent_io_tree *tree, u64 start, u64 private)
 {
 	struct rb_node *node;
 	struct extent_state *state;
@@ -1807,64 +1821,6 @@ out:
 	return ret;
 }
 
-void extent_cache_csums_dio(struct extent_io_tree *tree, u64 start, u32 csums[],
-			    int count)
-{
-	struct rb_node *node;
-	struct extent_state *state;
-
-	spin_lock(&tree->lock);
-	/*
-	 * this search will find all the extents that end after
-	 * our range starts.
-	 */
-	node = tree_search(tree, start);
-	BUG_ON(!node);
-
-	state = rb_entry(node, struct extent_state, rb_node);
-	BUG_ON(state->start != start);
-
-	while (count) {
-		state->private = *csums++;
-		count--;
-		state = next_state(state);
-	}
-	spin_unlock(&tree->lock);
-}
-
-static inline u64 __btrfs_get_bio_offset(struct bio *bio, int bio_index)
-{
-	struct bio_vec *bvec = bio->bi_io_vec + bio_index;
-
-	return page_offset(bvec->bv_page) + bvec->bv_offset;
-}
-
-void extent_cache_csums(struct extent_io_tree *tree, struct bio *bio, int bio_index,
-			u32 csums[], int count)
-{
-	struct rb_node *node;
-	struct extent_state *state = NULL;
-	u64 start;
-
-	spin_lock(&tree->lock);
-	do {
-		start = __btrfs_get_bio_offset(bio, bio_index);
-		if (state == NULL || state->start != start) {
-			node = tree_search(tree, start);
-			BUG_ON(!node);
-
-			state = rb_entry(node, struct extent_state, rb_node);
-			BUG_ON(state->start != start);
-		}
-		state->private = *csums++;
-		count--;
-		bio_index++;
-
-		state = next_state(state);
-	} while (count);
-	spin_unlock(&tree->lock);
-}
-
 int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private)
 {
 	struct rb_node *node;
@@ -2143,7 +2099,8 @@ static int clean_io_failure(u64 start, struct page *page)
 					    EXTENT_LOCKED);
 	spin_unlock(&BTRFS_I(inode)->io_tree.lock);
 
-	if (state && state->start == failrec->start) {
+	if (state && state->start <= failrec->start &&
+	    state->end >= failrec->start + failrec->len - 1) {
 		fs_info = BTRFS_I(inode)->root->fs_info;
 		num_copies = btrfs_num_copies(fs_info, failrec->logical,
 					      failrec->len);
@@ -2171,9 +2128,9 @@ out:
  * needed
  */
 
-static int bio_readpage_error(struct bio *failed_bio, struct page *page,
-				u64 start, u64 end, int failed_mirror,
-				struct extent_state *state)
+static int bio_readpage_error(struct bio *failed_bio, u64 phy_offset,
+			      struct page *page, u64 start, u64 end,
+			      int failed_mirror)
 {
 	struct io_failure_record *failrec = NULL;
 	u64 private;
@@ -2183,6 +2140,8 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page,
 	struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree;
 	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
 	struct bio *bio;
+	struct btrfs_io_bio *btrfs_failed_bio;
+	struct btrfs_io_bio *btrfs_bio;
 	int num_copies;
 	int ret;
 	int read_mode;
@@ -2266,23 +2225,12 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page,
 		 * all the retry and error correction code that follows. no
 		 * matter what the error is, it is very likely to persist.
 		 */
-		pr_debug("bio_readpage_error: cannot repair, num_copies == 1. "
-			 "state=%p, num_copies=%d, next_mirror %d, "
-			 "failed_mirror %d\n", state, num_copies,
-			 failrec->this_mirror, failed_mirror);
+		pr_debug("bio_readpage_error: cannot repair, num_copies=%d, next_mirror %d, failed_mirror %d\n",
+			 num_copies, failrec->this_mirror, failed_mirror);
 		free_io_failure(inode, failrec, 0);
 		return -EIO;
 	}
 
-	if (!state) {
-		spin_lock(&tree->lock);
-		state = find_first_extent_bit_state(tree, failrec->start,
-						    EXTENT_LOCKED);
-		if (state && state->start != failrec->start)
-			state = NULL;
-		spin_unlock(&tree->lock);
-	}
-
 	/*
 	 * there are two premises:
 	 *	a) deliver good data to the caller
@@ -2319,9 +2267,8 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page,
 		read_mode = READ_SYNC;
 	}
 
-	if (!state || failrec->this_mirror > num_copies) {
-		pr_debug("bio_readpage_error: (fail) state=%p, num_copies=%d, "
-			 "next_mirror %d, failed_mirror %d\n", state,
+	if (failrec->this_mirror > num_copies) {
+		pr_debug("bio_readpage_error: (fail) num_copies=%d, next_mirror %d, failed_mirror %d\n",
 			 num_copies, failrec->this_mirror, failed_mirror);
 		free_io_failure(inode, failrec, 0);
 		return -EIO;
@@ -2332,12 +2279,24 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page,
 		free_io_failure(inode, failrec, 0);
 		return -EIO;
 	}
-	bio->bi_private = state;
 	bio->bi_end_io = failed_bio->bi_end_io;
 	bio->bi_sector = failrec->logical >> 9;
 	bio->bi_bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
 	bio->bi_size = 0;
 
+	btrfs_failed_bio = btrfs_io_bio(failed_bio);
+	if (btrfs_failed_bio->csum) {
+		struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
+		u16 csum_size = btrfs_super_csum_size(fs_info->super_copy);
+
+		btrfs_bio = btrfs_io_bio(bio);
+		btrfs_bio->csum = btrfs_bio->csum_inline;
+		phy_offset >>= inode->i_sb->s_blocksize_bits;
+		phy_offset *= csum_size;
+		memcpy(btrfs_bio->csum, btrfs_failed_bio->csum + phy_offset,
+		       csum_size);
+	}
+
 	bio_add_page(bio, page, failrec->len, start - page_offset(page));
 
 	pr_debug("bio_readpage_error: submitting new read[%#x] to "
@@ -2420,6 +2379,18 @@ static void end_bio_extent_writepage(struct bio *bio, int err)
 	bio_put(bio);
 }
 
+static void
+endio_readpage_release_extent(struct extent_io_tree *tree, u64 start, u64 len,
+			      int uptodate)
+{
+	struct extent_state *cached = NULL;
+	u64 end = start + len - 1;
+
+	if (uptodate && tree->track_uptodate)
+		set_extent_uptodate(tree, start, end, &cached, GFP_ATOMIC);
+	unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
+}
+
 /*
  * after a readpage IO is done, we need to:
  * clear the uptodate bits on error
@@ -2436,9 +2407,14 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 	int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
 	struct bio_vec *bvec_end = bio->bi_io_vec + bio->bi_vcnt - 1;
 	struct bio_vec *bvec = bio->bi_io_vec;
+	struct btrfs_io_bio *io_bio = btrfs_io_bio(bio);
 	struct extent_io_tree *tree;
+	u64 offset = 0;
 	u64 start;
 	u64 end;
+	u64 len;
+	u64 extent_start = 0;
+	u64 extent_len = 0;
 	int mirror;
 	int ret;
 
@@ -2447,14 +2423,12 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 
 	do {
 		struct page *page = bvec->bv_page;
-		struct extent_state *cached = NULL;
-		struct extent_state *state;
-		struct btrfs_io_bio *io_bio = btrfs_io_bio(bio);
+		struct inode *inode = page->mapping->host;
 
 		pr_debug("end_bio_extent_readpage: bi_sector=%llu, err=%d, "
 			 "mirror=%lu\n", (u64)bio->bi_sector, err,
 			 io_bio->mirror_num);
-		tree = &BTRFS_I(page->mapping->host)->io_tree;
+		tree = &BTRFS_I(inode)->io_tree;
 
 		/* We always issue full-page reads, but if some block
 		 * in a page fails to read, blk_update_request() will
@@ -2469,37 +2443,32 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 
 		start = page_offset(page);
 		end = start + bvec->bv_offset + bvec->bv_len - 1;
+		len = bvec->bv_len;
 
 		if (++bvec <= bvec_end)
 			prefetchw(&bvec->bv_page->flags);
 
-		spin_lock(&tree->lock);
-		state = find_first_extent_bit_state(tree, start, EXTENT_LOCKED);
-		if (state && state->start == start) {
-			/*
-			 * take a reference on the state, unlock will drop
-			 * the ref
-			 */
-			cache_state(state, &cached);
-		}
-		spin_unlock(&tree->lock);
-
 		mirror = io_bio->mirror_num;
-		if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
-			ret = tree->ops->readpage_end_io_hook(page, start, end,
-							      state, mirror);
+		if (likely(uptodate && tree->ops &&
+			   tree->ops->readpage_end_io_hook)) {
+			ret = tree->ops->readpage_end_io_hook(io_bio, offset,
+							      page, start, end,
+							      mirror);
 			if (ret)
 				uptodate = 0;
 			else
 				clean_io_failure(start, page);
 		}
 
-		if (!uptodate && tree->ops && tree->ops->readpage_io_failed_hook) {
+		if (likely(uptodate))
+			goto readpage_ok;
+
+		if (tree->ops && tree->ops->readpage_io_failed_hook) {
 			ret = tree->ops->readpage_io_failed_hook(page, mirror);
 			if (!ret && !err &&
 			    test_bit(BIO_UPTODATE, &bio->bi_flags))
 				uptodate = 1;
-		} else if (!uptodate) {
+		} else {
 			/*
 			 * The generic bio_readpage_error handles errors the
 			 * following way: If possible, new read requests are
@@ -2510,32 +2479,62 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 			 * can't handle the error it will return -EIO and we
 			 * remain responsible for that page.
 			 */
-			ret = bio_readpage_error(bio, page, start, end, mirror, NULL);
+			ret = bio_readpage_error(bio, offset, page, start, end,
+						 mirror);
 			if (ret == 0) {
 				uptodate =
 					test_bit(BIO_UPTODATE, &bio->bi_flags);
 				if (err)
 					uptodate = 0;
-				uncache_state(&cached);
 				continue;
 			}
 		}
-
-		if (uptodate && tree->track_uptodate) {
-			set_extent_uptodate(tree, start, end, &cached,
-					    GFP_ATOMIC);
-		}
-		unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
-
-		if (uptodate) {
+readpage_ok:
+		if (likely(uptodate)) {
+			loff_t i_size = i_size_read(inode);
+			pgoff_t end_index = i_size >> PAGE_CACHE_SHIFT;
+			unsigned offset;
+
+			/* Zero out the end if this page straddles i_size */
+			offset = i_size & (PAGE_CACHE_SIZE-1);
+			if (page->index == end_index && offset)
+				zero_user_segment(page, offset, PAGE_CACHE_SIZE);
 			SetPageUptodate(page);
 		} else {
 			ClearPageUptodate(page);
 			SetPageError(page);
 		}
 		unlock_page(page);
+		offset += len;
+
+		if (unlikely(!uptodate)) {
+			if (extent_len) {
+				endio_readpage_release_extent(tree,
+							      extent_start,
+							      extent_len, 1);
+				extent_start = 0;
+				extent_len = 0;
+			}
+			endio_readpage_release_extent(tree, start,
+						      end - start + 1, 0);
+		} else if (!extent_len) {
+			extent_start = start;
+			extent_len = end + 1 - start;
+		} else if (extent_start + extent_len == start) {
+			extent_len += end + 1 - start;
+		} else {
+			endio_readpage_release_extent(tree, extent_start,
+						      extent_len, uptodate);
+			extent_start = start;
+			extent_len = end + 1 - start;
+		}
 	} while (bvec <= bvec_end);
 
+	if (extent_len)
+		endio_readpage_release_extent(tree, extent_start, extent_len,
+					      uptodate);
+	if (io_bio->end_io)
+		io_bio->end_io(io_bio, err);
 	bio_put(bio);
 }
 
@@ -2547,6 +2546,7 @@ struct bio *
 btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
 		gfp_t gfp_flags)
 {
+	struct btrfs_io_bio *btrfs_bio;
 	struct bio *bio;
 
 	bio = bio_alloc_bioset(gfp_flags, nr_vecs, btrfs_bioset);
@@ -2562,6 +2562,10 @@ btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
 		bio->bi_size = 0;
 		bio->bi_bdev = bdev;
 		bio->bi_sector = first_sector;
+		btrfs_bio = btrfs_io_bio(bio);
+		btrfs_bio->csum = NULL;
+		btrfs_bio->csum_allocated = NULL;
+		btrfs_bio->end_io = NULL;
 	}
 	return bio;
 }
@@ -2575,7 +2579,17 @@ struct bio *btrfs_bio_clone(struct bio *bio, gfp_t gfp_mask)
 /* this also allocates from the btrfs_bioset */
 struct bio *btrfs_io_bio_alloc(gfp_t gfp_mask, unsigned int nr_iovecs)
 {
-	return bio_alloc_bioset(gfp_mask, nr_iovecs, btrfs_bioset);
+	struct btrfs_io_bio *btrfs_bio;
+	struct bio *bio;
+
+	bio = bio_alloc_bioset(gfp_mask, nr_iovecs, btrfs_bioset);
+	if (bio) {
+		btrfs_bio = btrfs_io_bio(bio);
+		btrfs_bio->csum = NULL;
+		btrfs_bio->csum_allocated = NULL;
+		btrfs_bio->end_io = NULL;
+	}
+	return bio;
 }
 
 
@@ -2699,17 +2713,45 @@ void set_page_extent_mapped(struct page *page)
 	}
 }
 
+static struct extent_map *
+__get_extent_map(struct inode *inode, struct page *page, size_t pg_offset,
+		 u64 start, u64 len, get_extent_t *get_extent,
+		 struct extent_map **em_cached)
+{
+	struct extent_map *em;
+
+	if (em_cached && *em_cached) {
+		em = *em_cached;
+		if (em->in_tree && start >= em->start &&
+		    start < extent_map_end(em)) {
+			atomic_inc(&em->refs);
+			return em;
+		}
+
+		free_extent_map(em);
+		*em_cached = NULL;
+	}
+
+	em = get_extent(inode, page, pg_offset, start, len, 0);
+	if (em_cached && !IS_ERR_OR_NULL(em)) {
+		BUG_ON(*em_cached);
+		atomic_inc(&em->refs);
+		*em_cached = em;
+	}
+	return em;
+}
 /*
  * basic readpage implementation.  Locked extent state structs are inserted
  * into the tree that are removed when the IO is done (by the end_io
  * handlers)
  * XXX JDM: This needs looking at to ensure proper page locking
  */
-static int __extent_read_full_page(struct extent_io_tree *tree,
-				   struct page *page,
-				   get_extent_t *get_extent,
-				   struct bio **bio, int mirror_num,
-				   unsigned long *bio_flags, int rw)
+static int __do_readpage(struct extent_io_tree *tree,
+			 struct page *page,
+			 get_extent_t *get_extent,
+			 struct extent_map **em_cached,
+			 struct bio **bio, int mirror_num,
+			 unsigned long *bio_flags, int rw)
 {
 	struct inode *inode = page->mapping->host;
 	u64 start = page_offset(page);
@@ -2723,35 +2765,26 @@ 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;
+	int parent_locked = *bio_flags & EXTENT_BIO_PARENT_LOCKED;
 	size_t pg_offset = 0;
 	size_t iosize;
 	size_t disk_io_size;
 	size_t blocksize = inode->i_sb->s_blocksize;
-	unsigned long this_bio_flag = 0;
+	unsigned long this_bio_flag = *bio_flags & EXTENT_BIO_PARENT_LOCKED;
 
 	set_page_extent_mapped(page);
 
+	end = page_end;
 	if (!PageUptodate(page)) {
 		if (cleancache_get_page(page) == 0) {
 			BUG_ON(blocksize != PAGE_SIZE);
+			unlock_extent(tree, start, end);
 			goto out;
 		}
 	}
 
-	end = page_end;
-	while (1) {
-		lock_extent(tree, start, end);
-		ordered = btrfs_lookup_ordered_extent(inode, start);
-		if (!ordered)
-			break;
-		unlock_extent(tree, start, end);
-		btrfs_start_ordered_extent(inode, ordered, 1);
-		btrfs_put_ordered_extent(ordered);
-	}
-
 	if (page->index == last_byte >> PAGE_CACHE_SHIFT) {
 		char *userpage;
 		size_t zero_offset = last_byte & (PAGE_CACHE_SIZE - 1);
@@ -2778,15 +2811,18 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			kunmap_atomic(userpage);
 			set_extent_uptodate(tree, cur, cur + iosize - 1,
 					    &cached, GFP_NOFS);
-			unlock_extent_cached(tree, cur, cur + iosize - 1,
-					     &cached, GFP_NOFS);
+			if (!parent_locked)
+				unlock_extent_cached(tree, cur,
+						     cur + iosize - 1,
+						     &cached, GFP_NOFS);
 			break;
 		}
-		em = get_extent(inode, page, pg_offset, cur,
-				end - cur + 1, 0);
+		em = __get_extent_map(inode, page, pg_offset, cur,
+				      end - cur + 1, get_extent, em_cached);
 		if (IS_ERR_OR_NULL(em)) {
 			SetPageError(page);
-			unlock_extent(tree, cur, end);
+			if (!parent_locked)
+				unlock_extent(tree, cur, end);
 			break;
 		}
 		extent_offset = cur - em->start;
@@ -2794,7 +2830,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		BUG_ON(end < cur);
 
 		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
-			this_bio_flag = EXTENT_BIO_COMPRESSED;
+			this_bio_flag |= EXTENT_BIO_COMPRESSED;
 			extent_set_compress_type(&this_bio_flag,
 						 em->compress_type);
 		}
@@ -2838,7 +2874,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		if (test_range_bit(tree, cur, cur_end,
 				   EXTENT_UPTODATE, 1, NULL)) {
 			check_page_uptodate(tree, page);
-			unlock_extent(tree, cur, cur + iosize - 1);
+			if (!parent_locked)
+				unlock_extent(tree, cur, cur + iosize - 1);
 			cur = cur + iosize;
 			pg_offset += iosize;
 			continue;
@@ -2848,7 +2885,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		 */
 		if (block_start == EXTENT_MAP_INLINE) {
 			SetPageError(page);
-			unlock_extent(tree, cur, cur + iosize - 1);
+			if (!parent_locked)
+				unlock_extent(tree, cur, cur + iosize - 1);
 			cur = cur + iosize;
 			pg_offset += iosize;
 			continue;
@@ -2866,7 +2904,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			*bio_flags = this_bio_flag;
 		} else {
 			SetPageError(page);
-			unlock_extent(tree, cur, cur + iosize - 1);
+			if (!parent_locked)
+				unlock_extent(tree, cur, cur + iosize - 1);
 		}
 		cur = cur + iosize;
 		pg_offset += iosize;
@@ -2880,6 +2919,104 @@ out:
 	return 0;
 }
 
+static inline void __do_contiguous_readpages(struct extent_io_tree *tree,
+					     struct page *pages[], int nr_pages,
+					     u64 start, u64 end,
+					     get_extent_t *get_extent,
+					     struct extent_map **em_cached,
+					     struct bio **bio, int mirror_num,
+					     unsigned long *bio_flags, int rw)
+{
+	struct inode *inode;
+	struct btrfs_ordered_extent *ordered;
+	int index;
+
+	inode = pages[0]->mapping->host;
+	while (1) {
+		lock_extent(tree, start, end);
+		ordered = btrfs_lookup_ordered_range(inode, start,
+						     end - start + 1);
+		if (!ordered)
+			break;
+		unlock_extent(tree, start, end);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		btrfs_put_ordered_extent(ordered);
+	}
+
+	for (index = 0; index < nr_pages; index++) {
+		__do_readpage(tree, pages[index], get_extent, em_cached, bio,
+			      mirror_num, bio_flags, rw);
+		page_cache_release(pages[index]);
+	}
+}
+
+static void __extent_readpages(struct extent_io_tree *tree,
+			       struct page *pages[],
+			       int nr_pages, get_extent_t *get_extent,
+			       struct extent_map **em_cached,
+			       struct bio **bio, int mirror_num,
+			       unsigned long *bio_flags, int rw)
+{
+	u64 start = 0;
+	u64 end = 0;
+	u64 page_start;
+	int index;
+	int first_index = 0;
+
+	for (index = 0; index < nr_pages; index++) {
+		page_start = page_offset(pages[index]);
+		if (!end) {
+			start = page_start;
+			end = start + PAGE_CACHE_SIZE - 1;
+			first_index = index;
+		} else if (end + 1 == page_start) {
+			end += PAGE_CACHE_SIZE;
+		} else {
+			__do_contiguous_readpages(tree, &pages[first_index],
+						  index - first_index, start,
+						  end, get_extent, em_cached,
+						  bio, mirror_num, bio_flags,
+						  rw);
+			start = page_start;
+			end = start + PAGE_CACHE_SIZE - 1;
+			first_index = index;
+		}
+	}
+
+	if (end)
+		__do_contiguous_readpages(tree, &pages[first_index],
+					  index - first_index, start,
+					  end, get_extent, em_cached, bio,
+					  mirror_num, bio_flags, rw);
+}
+
+static int __extent_read_full_page(struct extent_io_tree *tree,
+				   struct page *page,
+				   get_extent_t *get_extent,
+				   struct bio **bio, int mirror_num,
+				   unsigned long *bio_flags, int rw)
+{
+	struct inode *inode = page->mapping->host;
+	struct btrfs_ordered_extent *ordered;
+	u64 start = page_offset(page);
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+	int ret;
+
+	while (1) {
+		lock_extent(tree, start, end);
+		ordered = btrfs_lookup_ordered_extent(inode, start);
+		if (!ordered)
+			break;
+		unlock_extent(tree, start, end);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		btrfs_put_ordered_extent(ordered);
+	}
+
+	ret = __do_readpage(tree, page, get_extent, NULL, bio, mirror_num,
+			    bio_flags, rw);
+	return ret;
+}
+
 int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
 			    get_extent_t *get_extent, int mirror_num)
 {
@@ -2894,6 +3031,20 @@ int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
 	return ret;
 }
 
+int extent_read_full_page_nolock(struct extent_io_tree *tree, struct page *page,
+				 get_extent_t *get_extent, int mirror_num)
+{
+	struct bio *bio = NULL;
+	unsigned long bio_flags = EXTENT_BIO_PARENT_LOCKED;
+	int ret;
+
+	ret = __do_readpage(tree, page, get_extent, NULL, &bio, mirror_num,
+				      &bio_flags, READ);
+	if (bio)
+		ret = submit_one_bio(READ, bio, mirror_num, bio_flags);
+	return ret;
+}
+
 static noinline void update_nr_written(struct page *page,
 				      struct writeback_control *wbc,
 				      unsigned long nr_written)
@@ -2957,7 +3108,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 	pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
 	if (page->index > end_index ||
 	   (page->index == end_index && !pg_offset)) {
-		page->mapping->a_ops->invalidatepage(page, 0);
+		page->mapping->a_ops->invalidatepage(page, 0, PAGE_CACHE_SIZE);
 		unlock_page(page);
 		return 0;
 	}
@@ -3150,8 +3301,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 			if (!PageWriteback(page)) {
 				printk(KERN_ERR "btrfs warning page %lu not "
 				       "writeback, cur %llu end %llu\n",
-				       page->index, (unsigned long long)cur,
-				       (unsigned long long)end);
+				       page->index, cur, end);
 			}
 
 			ret = submit_extent_page(write_flags, tree, page,
@@ -3730,7 +3880,7 @@ int extent_readpages(struct extent_io_tree *tree,
 	unsigned long bio_flags = 0;
 	struct page *pagepool[16];
 	struct page *page;
-	int i = 0;
+	struct extent_map *em_cached = NULL;
 	int nr = 0;
 
 	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
@@ -3747,18 +3897,16 @@ int extent_readpages(struct extent_io_tree *tree,
 		pagepool[nr++] = page;
 		if (nr < ARRAY_SIZE(pagepool))
 			continue;
-		for (i = 0; i < nr; i++) {
-			__extent_read_full_page(tree, pagepool[i], get_extent,
-					&bio, 0, &bio_flags, READ);
-			page_cache_release(pagepool[i]);
-		}
+		__extent_readpages(tree, pagepool, nr, get_extent, &em_cached,
+				   &bio, 0, &bio_flags, READ);
 		nr = 0;
 	}
-	for (i = 0; i < nr; i++) {
-		__extent_read_full_page(tree, pagepool[i], get_extent,
-					&bio, 0, &bio_flags, READ);
-		page_cache_release(pagepool[i]);
-	}
+	if (nr)
+		__extent_readpages(tree, pagepool, nr, get_extent, &em_cached,
+				   &bio, 0, &bio_flags, READ);
+
+	if (em_cached)
+		free_extent_map(em_cached);
 
 	BUG_ON(!list_empty(pages));
 	if (bio)
@@ -4009,7 +4157,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	}
 
 	while (!end) {
-		u64 offset_in_extent;
+		u64 offset_in_extent = 0;
 
 		/* break if the extent we found is outside the range */
 		if (em->start >= max || extent_map_end(em) < off)
@@ -4025,9 +4173,12 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 
 		/*
 		 * record the offset from the start of the extent
-		 * for adjusting the disk offset below
+		 * for adjusting the disk offset below.  Only do this if the
+		 * extent isn't compressed since our in ram offset may be past
+		 * what we have actually allocated on disk.
 		 */
-		offset_in_extent = em_start - em->start;
+		if (!test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
+			offset_in_extent = em_start - em->start;
 		em_end = extent_map_end(em);
 		em_len = em_end - em_start;
 		emflags = em->flags;
@@ -4094,6 +4245,76 @@ static void __free_extent_buffer(struct extent_buffer *eb)
 	kmem_cache_free(extent_buffer_cache, eb);
 }
 
+static int extent_buffer_under_io(struct extent_buffer *eb)
+{
+	return (atomic_read(&eb->io_pages) ||
+		test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags) ||
+		test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
+}
+
+/*
+ * Helper for releasing extent buffer page.
+ */
+static void btrfs_release_extent_buffer_page(struct extent_buffer *eb,
+						unsigned long start_idx)
+{
+	unsigned long index;
+	unsigned long num_pages;
+	struct page *page;
+	int mapped = !test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags);
+
+	BUG_ON(extent_buffer_under_io(eb));
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+	index = start_idx + num_pages;
+	if (start_idx >= index)
+		return;
+
+	do {
+		index--;
+		page = extent_buffer_page(eb, index);
+		if (page && mapped) {
+			spin_lock(&page->mapping->private_lock);
+			/*
+			 * We do this since we'll remove the pages after we've
+			 * removed the eb from the radix tree, so we could race
+			 * and have this page now attached to the new eb.  So
+			 * only clear page_private if it's still connected to
+			 * this eb.
+			 */
+			if (PagePrivate(page) &&
+			    page->private == (unsigned long)eb) {
+				BUG_ON(test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
+				BUG_ON(PageDirty(page));
+				BUG_ON(PageWriteback(page));
+				/*
+				 * We need to make sure we haven't be attached
+				 * to a new eb.
+				 */
+				ClearPagePrivate(page);
+				set_page_private(page, 0);
+				/* One for the page private */
+				page_cache_release(page);
+			}
+			spin_unlock(&page->mapping->private_lock);
+
+		}
+		if (page) {
+			/* One for when we alloced the page */
+			page_cache_release(page);
+		}
+	} while (index != start_idx);
+}
+
+/*
+ * Helper for releasing the extent buffer.
+ */
+static inline void btrfs_release_extent_buffer(struct extent_buffer *eb)
+{
+	btrfs_release_extent_buffer_page(eb, 0);
+	__free_extent_buffer(eb);
+}
+
 static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 						   u64 start,
 						   unsigned long len,
@@ -4142,13 +4363,16 @@ struct extent_buffer *btrfs_clone_extent_buffer(struct extent_buffer *src)
 	struct extent_buffer *new;
 	unsigned long num_pages = num_extent_pages(src->start, src->len);
 
-	new = __alloc_extent_buffer(NULL, src->start, src->len, GFP_ATOMIC);
+	new = __alloc_extent_buffer(NULL, src->start, src->len, GFP_NOFS);
 	if (new == NULL)
 		return NULL;
 
 	for (i = 0; i < num_pages; i++) {
-		p = alloc_page(GFP_ATOMIC);
-		BUG_ON(!p);
+		p = alloc_page(GFP_NOFS);
+		if (!p) {
+			btrfs_release_extent_buffer(new);
+			return NULL;
+		}
 		attach_extent_buffer_page(new, p);
 		WARN_ON(PageDirty(p));
 		SetPageUptodate(p);
@@ -4168,12 +4392,12 @@ struct extent_buffer *alloc_dummy_extent_buffer(u64 start, unsigned long len)
 	unsigned long num_pages = num_extent_pages(0, len);
 	unsigned long i;
 
-	eb = __alloc_extent_buffer(NULL, start, len, GFP_ATOMIC);
+	eb = __alloc_extent_buffer(NULL, start, len, GFP_NOFS);
 	if (!eb)
 		return NULL;
 
 	for (i = 0; i < num_pages; i++) {
-		eb->pages[i] = alloc_page(GFP_ATOMIC);
+		eb->pages[i] = alloc_page(GFP_NOFS);
 		if (!eb->pages[i])
 			goto err;
 	}
@@ -4189,76 +4413,6 @@ err:
 	return NULL;
 }
 
-static int extent_buffer_under_io(struct extent_buffer *eb)
-{
-	return (atomic_read(&eb->io_pages) ||
-		test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags) ||
-		test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
-}
-
-/*
- * Helper for releasing extent buffer page.
- */
-static void btrfs_release_extent_buffer_page(struct extent_buffer *eb,
-						unsigned long start_idx)
-{
-	unsigned long index;
-	unsigned long num_pages;
-	struct page *page;
-	int mapped = !test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags);
-
-	BUG_ON(extent_buffer_under_io(eb));
-
-	num_pages = num_extent_pages(eb->start, eb->len);
-	index = start_idx + num_pages;
-	if (start_idx >= index)
-		return;
-
-	do {
-		index--;
-		page = extent_buffer_page(eb, index);
-		if (page && mapped) {
-			spin_lock(&page->mapping->private_lock);
-			/*
-			 * We do this since we'll remove the pages after we've
-			 * removed the eb from the radix tree, so we could race
-			 * and have this page now attached to the new eb.  So
-			 * only clear page_private if it's still connected to
-			 * this eb.
-			 */
-			if (PagePrivate(page) &&
-			    page->private == (unsigned long)eb) {
-				BUG_ON(test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
-				BUG_ON(PageDirty(page));
-				BUG_ON(PageWriteback(page));
-				/*
-				 * We need to make sure we haven't be attached
-				 * to a new eb.
-				 */
-				ClearPagePrivate(page);
-				set_page_private(page, 0);
-				/* One for the page private */
-				page_cache_release(page);
-			}
-			spin_unlock(&page->mapping->private_lock);
-
-		}
-		if (page) {
-			/* One for when we alloced the page */
-			page_cache_release(page);
-		}
-	} while (index != start_idx);
-}
-
-/*
- * Helper for releasing the extent buffer.
- */
-static inline void btrfs_release_extent_buffer(struct extent_buffer *eb)
-{
-	btrfs_release_extent_buffer_page(eb, 0);
-	__free_extent_buffer(eb);
-}
-
 static void check_buffer_tree_ref(struct extent_buffer *eb)
 {
 	int refs;
@@ -4729,7 +4883,7 @@ void read_extent_buffer(struct extent_buffer *eb, void *dstv,
 	WARN_ON(start > eb->len);
 	WARN_ON(start + len > eb->start + eb->len);
 
-	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+	offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1);
 
 	while (len > 0) {
 		page = extent_buffer_page(eb, i);
@@ -4771,8 +4925,8 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
 
 	if (start + min_len > eb->len) {
 		WARN(1, KERN_ERR "btrfs bad mapping eb start %llu len %lu, "
-		       "wanted %lu %lu\n", (unsigned long long)eb->start,
-		       eb->len, start, min_len);
+		       "wanted %lu %lu\n",
+		       eb->start, eb->len, start, min_len);
 		return -EINVAL;
 	}
 
@@ -4799,7 +4953,7 @@ int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
 	WARN_ON(start > eb->len);
 	WARN_ON(start + len > eb->start + eb->len);
 
-	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+	offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1);
 
 	while (len > 0) {
 		page = extent_buffer_page(eb, i);
@@ -4833,7 +4987,7 @@ void write_extent_buffer(struct extent_buffer *eb, const void *srcv,
 	WARN_ON(start > eb->len);
 	WARN_ON(start + len > eb->start + eb->len);
 
-	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+	offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1);
 
 	while (len > 0) {
 		page = extent_buffer_page(eb, i);
@@ -4863,7 +5017,7 @@ void memset_extent_buffer(struct extent_buffer *eb, char c,
 	WARN_ON(start > eb->len);
 	WARN_ON(start + len > eb->start + eb->len);
 
-	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+	offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1);
 
 	while (len > 0) {
 		page = extent_buffer_page(eb, i);
@@ -4894,7 +5048,7 @@ void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src,
 	WARN_ON(src->len != dst_len);
 
 	offset = (start_offset + dst_offset) &
-		((unsigned long)PAGE_CACHE_SIZE - 1);
+		(PAGE_CACHE_SIZE - 1);
 
 	while (len > 0) {
 		page = extent_buffer_page(dst, i);
@@ -4980,9 +5134,9 @@ void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 
 	while (len > 0) {
 		dst_off_in_page = (start_offset + dst_offset) &
-			((unsigned long)PAGE_CACHE_SIZE - 1);
+			(PAGE_CACHE_SIZE - 1);
 		src_off_in_page = (start_offset + src_offset) &
-			((unsigned long)PAGE_CACHE_SIZE - 1);
+			(PAGE_CACHE_SIZE - 1);
 
 		dst_i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT;
 		src_i = (start_offset + src_offset) >> PAGE_CACHE_SHIFT;
@@ -5033,9 +5187,9 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 		src_i = (start_offset + src_end) >> PAGE_CACHE_SHIFT;
 
 		dst_off_in_page = (start_offset + dst_end) &
-			((unsigned long)PAGE_CACHE_SIZE - 1);
+			(PAGE_CACHE_SIZE - 1);
 		src_off_in_page = (start_offset + src_end) &
-			((unsigned long)PAGE_CACHE_SIZE - 1);
+			(PAGE_CACHE_SIZE - 1);
 
 		cur = min_t(unsigned long, len, src_off_in_page + 1);
 		cur = min(cur, dst_off_in_page + 1);
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index 41fb81e7ec53..6dbc645f1f3d 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -19,6 +19,7 @@
 #define EXTENT_FIRST_DELALLOC (1 << 12)
 #define EXTENT_NEED_WAIT (1 << 13)
 #define EXTENT_DAMAGED (1 << 14)
+#define EXTENT_NORESERVE (1 << 15)
 #define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK)
 #define EXTENT_CTLBITS (EXTENT_DO_ACCOUNTING | EXTENT_FIRST_DELALLOC)
 
@@ -28,6 +29,7 @@
  */
 #define EXTENT_BIO_COMPRESSED 1
 #define EXTENT_BIO_TREE_LOG 2
+#define EXTENT_BIO_PARENT_LOCKED 4
 #define EXTENT_BIO_FLAG_SHIFT 16
 
 /* these are bit numbers for test/set bit */
@@ -43,14 +45,11 @@
 #define EXTENT_BUFFER_DUMMY 9
 
 /* these are flags for extent_clear_unlock_delalloc */
-#define EXTENT_CLEAR_UNLOCK_PAGE 0x1
-#define EXTENT_CLEAR_UNLOCK	 0x2
-#define EXTENT_CLEAR_DELALLOC	 0x4
-#define EXTENT_CLEAR_DIRTY	 0x8
-#define EXTENT_SET_WRITEBACK	 0x10
-#define EXTENT_END_WRITEBACK	 0x20
-#define EXTENT_SET_PRIVATE2	 0x40
-#define EXTENT_CLEAR_ACCOUNTING  0x80
+#define PAGE_UNLOCK		(1 << 0)
+#define PAGE_CLEAR_DIRTY	(1 << 1)
+#define PAGE_SET_WRITEBACK	(1 << 2)
+#define PAGE_END_WRITEBACK	(1 << 3)
+#define PAGE_SET_PRIVATE2	(1 << 4)
 
 /*
  * page->private values.  Every page that is controlled by the extent
@@ -61,6 +60,7 @@
 
 struct extent_state;
 struct btrfs_root;
+struct btrfs_io_bio;
 
 typedef	int (extent_submit_bio_hook_t)(struct inode *inode, int rw,
 				       struct bio *bio, int mirror_num,
@@ -76,8 +76,9 @@ struct extent_io_ops {
 			      size_t size, struct bio *bio,
 			      unsigned long bio_flags);
 	int (*readpage_io_failed_hook)(struct page *page, int failed_mirror);
-	int (*readpage_end_io_hook)(struct page *page, u64 start, u64 end,
-				    struct extent_state *state, int mirror);
+	int (*readpage_end_io_hook)(struct btrfs_io_bio *io_bio, u64 phy_offset,
+				    struct page *page, u64 start, u64 end,
+				    int mirror);
 	int (*writepage_end_io_hook)(struct page *page, u64 start, u64 end,
 				      struct extent_state *state, int uptodate);
 	void (*set_bit_hook)(struct inode *inode, struct extent_state *state,
@@ -199,6 +200,8 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
 int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end);
 int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
 			  get_extent_t *get_extent, int mirror_num);
+int extent_read_full_page_nolock(struct extent_io_tree *tree, struct page *page,
+				 get_extent_t *get_extent, int mirror_num);
 int __init extent_io_init(void);
 void extent_io_exit(void);
 
@@ -260,11 +263,6 @@ int extent_readpages(struct extent_io_tree *tree,
 		     get_extent_t get_extent);
 int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		__u64 start, __u64 len, get_extent_t *get_extent);
-int set_state_private(struct extent_io_tree *tree, u64 start, u64 private);
-void extent_cache_csums_dio(struct extent_io_tree *tree, u64 start, u32 csums[],
-			    int count);
-void extent_cache_csums(struct extent_io_tree *tree, struct bio *bio,
-			int bvec_index, u32 csums[], int count);
 int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private);
 void set_page_extent_mapped(struct page *page);
 
@@ -329,10 +327,10 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset,
 		      unsigned long *map_len);
 int extent_range_clear_dirty_for_io(struct inode *inode, u64 start, u64 end);
 int extent_range_redirty_for_io(struct inode *inode, u64 start, u64 end);
-int extent_clear_unlock_delalloc(struct inode *inode,
-				struct extent_io_tree *tree,
-				u64 start, u64 end, struct page *locked_page,
-				unsigned long op);
+int extent_clear_unlock_delalloc(struct inode *inode, u64 start, u64 end,
+				 struct page *locked_page,
+				 unsigned long bits_to_clear,
+				 unsigned long page_ops);
 struct bio *
 btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
 		gfp_t gfp_flags);
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index b193bf324a41..4f53159bdb9d 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -23,6 +23,7 @@
 #include "ctree.h"
 #include "disk-io.h"
 #include "transaction.h"
+#include "volumes.h"
 #include "print-tree.h"
 
 #define __MAX_CSUM_ITEMS(r, size) ((unsigned long)(((BTRFS_LEAF_DATA_SIZE(r) - \
@@ -34,8 +35,7 @@
 
 #define MAX_ORDERED_SUM_BYTES(r) ((PAGE_SIZE - \
 				   sizeof(struct btrfs_ordered_sum)) / \
-				   sizeof(struct btrfs_sector_sum) * \
-				   (r)->sectorsize - (r)->sectorsize)
+				   sizeof(u32) * (r)->sectorsize)
 
 int btrfs_insert_file_extent(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
@@ -153,28 +153,54 @@ int btrfs_lookup_file_extent(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
+static void btrfs_io_bio_endio_readpage(struct btrfs_io_bio *bio, int err)
+{
+	kfree(bio->csum_allocated);
+}
+
 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[16];
-	int len;
 	struct bio_vec *bvec = bio->bi_io_vec;
-	int bio_index = 0;
+	struct btrfs_io_bio *btrfs_bio = btrfs_io_bio(bio);
+	struct btrfs_csum_item *item = NULL;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct btrfs_path *path;
+	u8 *csum;
 	u64 offset = 0;
 	u64 item_start_offset = 0;
 	u64 item_last_offset = 0;
 	u64 disk_bytenr;
 	u32 diff;
-	u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
+	int nblocks;
+	int bio_index = 0;
 	int count;
-	struct btrfs_path *path;
-	struct btrfs_csum_item *item = NULL;
-	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+
+	nblocks = bio->bi_size >> inode->i_sb->s_blocksize_bits;
+	if (!dst) {
+		if (nblocks * csum_size > BTRFS_BIO_INLINE_CSUM_SIZE) {
+			btrfs_bio->csum_allocated = kmalloc(nblocks * csum_size,
+							    GFP_NOFS);
+			if (!btrfs_bio->csum_allocated) {
+				btrfs_free_path(path);
+				return -ENOMEM;
+			}
+			btrfs_bio->csum = btrfs_bio->csum_allocated;
+			btrfs_bio->end_io = btrfs_io_bio_endio_readpage;
+		} else {
+			btrfs_bio->csum = btrfs_bio->csum_inline;
+		}
+		csum = btrfs_bio->csum;
+	} else {
+		csum = (u8 *)dst;
+	}
+
 	if (bio->bi_size > PAGE_CACHE_SIZE * 8)
 		path->reada = 2;
 
@@ -195,11 +221,10 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 	if (dio)
 		offset = logical_offset;
 	while (bio_index < bio->bi_vcnt) {
-		len = min_t(int, ARRAY_SIZE(sum), bio->bi_vcnt - bio_index);
 		if (!dio)
 			offset = page_offset(bvec->bv_page) + bvec->bv_offset;
-		count = btrfs_find_ordered_sum(inode, offset, disk_bytenr, sum,
-					       len);
+		count = btrfs_find_ordered_sum(inode, offset, disk_bytenr,
+					       (u32 *)csum, nblocks);
 		if (count)
 			goto found;
 
@@ -214,7 +239,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 						 path, disk_bytenr, 0);
 			if (IS_ERR(item)) {
 				count = 1;
-				sum[0] = 0;
+				memset(csum, 0, csum_size);
 				if (BTRFS_I(inode)->root->root_key.objectid ==
 				    BTRFS_DATA_RELOC_TREE_OBJECTID) {
 					set_extent_bits(io_tree, offset,
@@ -223,9 +248,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 				} else {
 					printk(KERN_INFO "btrfs no csum found "
 					       "for inode %llu start %llu\n",
-					       (unsigned long long)
-					       btrfs_ino(inode),
-					       (unsigned long long)offset);
+					       btrfs_ino(inode), offset);
 				}
 				item = NULL;
 				btrfs_release_path(path);
@@ -250,23 +273,14 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 		diff = disk_bytenr - item_start_offset;
 		diff = diff / root->sectorsize;
 		diff = diff * csum_size;
-		count = min_t(int, len, (item_last_offset - disk_bytenr) >>
-					inode->i_sb->s_blocksize_bits);
-		read_extent_buffer(path->nodes[0], sum,
+		count = min_t(int, nblocks, (item_last_offset - disk_bytenr) >>
+					    inode->i_sb->s_blocksize_bits);
+		read_extent_buffer(path->nodes[0], csum,
 				   ((unsigned long)item) + diff,
 				   csum_size * count);
 found:
-		if (dst) {
-			memcpy(dst, sum, count * csum_size);
-			dst += count;
-		} else {
-			if (dio)
-				extent_cache_csums_dio(io_tree, offset, sum,
-						       count);
-			else
-				extent_cache_csums(io_tree, bio, bio_index, sum,
-					    count);
-		}
+		csum += count * csum_size;
+		nblocks -= count;
 		while (count--) {
 			disk_bytenr += bvec->bv_len;
 			offset += bvec->bv_len;
@@ -285,9 +299,19 @@ int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode,
 }
 
 int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode,
-			      struct bio *bio, u64 offset)
+			      struct btrfs_dio_private *dip, struct bio *bio,
+			      u64 offset)
 {
-	return __btrfs_lookup_bio_sums(root, inode, bio, offset, NULL, 1);
+	int len = (bio->bi_sector << 9) - dip->disk_bytenr;
+	u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
+	int ret;
+
+	len >>= inode->i_sb->s_blocksize_bits;
+	len *= csum_size;
+
+	ret = __btrfs_lookup_bio_sums(root, inode, bio, offset,
+				      (u32 *)(dip->csum + len), 1);
+	return ret;
 }
 
 int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
@@ -297,7 +321,6 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 	struct btrfs_path *path;
 	struct extent_buffer *leaf;
 	struct btrfs_ordered_sum *sums;
-	struct btrfs_sector_sum *sector_sum;
 	struct btrfs_csum_item *item;
 	LIST_HEAD(tmplist);
 	unsigned long offset;
@@ -368,34 +391,28 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 				      struct btrfs_csum_item);
 		while (start < csum_end) {
 			size = min_t(size_t, csum_end - start,
-					MAX_ORDERED_SUM_BYTES(root));
+				     MAX_ORDERED_SUM_BYTES(root));
 			sums = kzalloc(btrfs_ordered_sum_size(root, size),
-					GFP_NOFS);
+				       GFP_NOFS);
 			if (!sums) {
 				ret = -ENOMEM;
 				goto fail;
 			}
 
-			sector_sum = sums->sums;
 			sums->bytenr = start;
-			sums->len = size;
+			sums->len = (int)size;
 
 			offset = (start - key.offset) >>
 				root->fs_info->sb->s_blocksize_bits;
 			offset *= csum_size;
+			size >>= root->fs_info->sb->s_blocksize_bits;
 
-			while (size > 0) {
-				read_extent_buffer(path->nodes[0],
-						&sector_sum->sum,
-						((unsigned long)item) +
-						offset, csum_size);
-				sector_sum->bytenr = start;
-
-				size -= root->sectorsize;
-				start += root->sectorsize;
-				offset += csum_size;
-				sector_sum++;
-			}
+			read_extent_buffer(path->nodes[0],
+					   sums->sums,
+					   ((unsigned long)item) + offset,
+					   csum_size * size);
+
+			start += root->sectorsize * size;
 			list_add_tail(&sums->list, &tmplist);
 		}
 		path->slots[0]++;
@@ -417,23 +434,20 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
 		       struct bio *bio, u64 file_start, int contig)
 {
 	struct btrfs_ordered_sum *sums;
-	struct btrfs_sector_sum *sector_sum;
 	struct btrfs_ordered_extent *ordered;
 	char *data;
 	struct bio_vec *bvec = bio->bi_io_vec;
 	int bio_index = 0;
+	int index;
 	unsigned long total_bytes = 0;
 	unsigned long this_sum_bytes = 0;
 	u64 offset;
-	u64 disk_bytenr;
 
 	WARN_ON(bio->bi_vcnt <= 0);
 	sums = kzalloc(btrfs_ordered_sum_size(root, bio->bi_size), GFP_NOFS);
 	if (!sums)
 		return -ENOMEM;
 
-	sector_sum = sums->sums;
-	disk_bytenr = (u64)bio->bi_sector << 9;
 	sums->len = bio->bi_size;
 	INIT_LIST_HEAD(&sums->list);
 
@@ -444,7 +458,8 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
 
 	ordered = btrfs_lookup_ordered_extent(inode, offset);
 	BUG_ON(!ordered); /* Logic error */
-	sums->bytenr = ordered->start;
+	sums->bytenr = (u64)bio->bi_sector << 9;
+	index = 0;
 
 	while (bio_index < bio->bi_vcnt) {
 		if (!contig)
@@ -463,28 +478,27 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
 			sums = kzalloc(btrfs_ordered_sum_size(root, bytes_left),
 				       GFP_NOFS);
 			BUG_ON(!sums); /* -ENOMEM */
-			sector_sum = sums->sums;
 			sums->len = bytes_left;
 			ordered = btrfs_lookup_ordered_extent(inode, offset);
 			BUG_ON(!ordered); /* Logic error */
-			sums->bytenr = ordered->start;
+			sums->bytenr = ((u64)bio->bi_sector << 9) +
+				       total_bytes;
+			index = 0;
 		}
 
 		data = kmap_atomic(bvec->bv_page);
-		sector_sum->sum = ~(u32)0;
-		sector_sum->sum = btrfs_csum_data(data + bvec->bv_offset,
-						  sector_sum->sum,
-						  bvec->bv_len);
+		sums->sums[index] = ~(u32)0;
+		sums->sums[index] = btrfs_csum_data(data + bvec->bv_offset,
+						    sums->sums[index],
+						    bvec->bv_len);
 		kunmap_atomic(data);
-		btrfs_csum_final(sector_sum->sum,
-				 (char *)&sector_sum->sum);
-		sector_sum->bytenr = disk_bytenr;
+		btrfs_csum_final(sums->sums[index],
+				 (char *)(sums->sums + index));
 
-		sector_sum++;
 		bio_index++;
+		index++;
 		total_bytes += bvec->bv_len;
 		this_sum_bytes += bvec->bv_len;
-		disk_bytenr += bvec->bv_len;
 		offset += bvec->bv_len;
 		bvec++;
 	}
@@ -672,62 +686,46 @@ out:
 	return ret;
 }
 
-static u64 btrfs_sector_sum_left(struct btrfs_ordered_sum *sums,
-				 struct btrfs_sector_sum *sector_sum,
-				 u64 total_bytes, u64 sectorsize)
-{
-	u64 tmp = sectorsize;
-	u64 next_sector = sector_sum->bytenr;
-	struct btrfs_sector_sum *next = sector_sum + 1;
-
-	while ((tmp + total_bytes) < sums->len) {
-		if (next_sector + sectorsize != next->bytenr)
-			break;
-		tmp += sectorsize;
-		next_sector = next->bytenr;
-		next++;
-	}
-	return tmp;
-}
-
 int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
 			   struct btrfs_ordered_sum *sums)
 {
-	u64 bytenr;
-	int ret;
 	struct btrfs_key file_key;
 	struct btrfs_key found_key;
-	u64 next_offset;
-	u64 total_bytes = 0;
-	int found_next;
 	struct btrfs_path *path;
 	struct btrfs_csum_item *item;
 	struct btrfs_csum_item *item_end;
 	struct extent_buffer *leaf = NULL;
+	u64 next_offset;
+	u64 total_bytes = 0;
 	u64 csum_offset;
-	struct btrfs_sector_sum *sector_sum;
+	u64 bytenr;
 	u32 nritems;
 	u32 ins_size;
+	int index = 0;
+	int found_next;
+	int ret;
 	u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
-
-	sector_sum = sums->sums;
 again:
 	next_offset = (u64)-1;
 	found_next = 0;
+	bytenr = sums->bytenr + total_bytes;
 	file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
-	file_key.offset = sector_sum->bytenr;
-	bytenr = sector_sum->bytenr;
+	file_key.offset = bytenr;
 	btrfs_set_key_type(&file_key, BTRFS_EXTENT_CSUM_KEY);
 
-	item = btrfs_lookup_csum(trans, root, path, sector_sum->bytenr, 1);
+	item = btrfs_lookup_csum(trans, root, path, bytenr, 1);
 	if (!IS_ERR(item)) {
-		leaf = path->nodes[0];
 		ret = 0;
+		leaf = path->nodes[0];
+		item_end = btrfs_item_ptr(leaf, path->slots[0],
+					  struct btrfs_csum_item);
+		item_end = (struct btrfs_csum_item *)((char *)item_end +
+			   btrfs_item_size_nr(leaf, path->slots[0]));
 		goto found;
 	}
 	ret = PTR_ERR(item);
@@ -807,8 +805,7 @@ again:
 
 		free_space = btrfs_leaf_free_space(root, leaf) -
 					 sizeof(struct btrfs_item) - csum_size;
-		tmp = btrfs_sector_sum_left(sums, sector_sum, total_bytes,
-					    root->sectorsize);
+		tmp = sums->len - total_bytes;
 		tmp >>= root->fs_info->sb->s_blocksize_bits;
 		WARN_ON(tmp < 1);
 
@@ -822,6 +819,7 @@ again:
 		diff *= csum_size;
 
 		btrfs_extend_item(root, path, diff);
+		ret = 0;
 		goto csum;
 	}
 
@@ -831,8 +829,7 @@ insert:
 	if (found_next) {
 		u64 tmp;
 
-		tmp = btrfs_sector_sum_left(sums, sector_sum, total_bytes,
-					    root->sectorsize);
+		tmp = sums->len - total_bytes;
 		tmp >>= root->fs_info->sb->s_blocksize_bits;
 		tmp = min(tmp, (next_offset - file_key.offset) >>
 					 root->fs_info->sb->s_blocksize_bits);
@@ -853,31 +850,25 @@ insert:
 		WARN_ON(1);
 		goto fail_unlock;
 	}
-csum:
 	leaf = path->nodes[0];
+csum:
 	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
-	ret = 0;
+	item_end = (struct btrfs_csum_item *)((unsigned char *)item +
+				      btrfs_item_size_nr(leaf, path->slots[0]));
 	item = (struct btrfs_csum_item *)((unsigned char *)item +
 					  csum_offset * csum_size);
 found:
-	item_end = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
-	item_end = (struct btrfs_csum_item *)((unsigned char *)item_end +
-				      btrfs_item_size_nr(leaf, path->slots[0]));
-next_sector:
-
-	write_extent_buffer(leaf, &sector_sum->sum, (unsigned long)item, csum_size);
-
-	total_bytes += root->sectorsize;
-	sector_sum++;
-	if (total_bytes < sums->len) {
-		item = (struct btrfs_csum_item *)((char *)item +
-						  csum_size);
-		if (item < item_end && bytenr + PAGE_CACHE_SIZE ==
-		    sector_sum->bytenr) {
-			bytenr = sector_sum->bytenr;
-			goto next_sector;
-		}
-	}
+	ins_size = (u32)(sums->len - total_bytes) >>
+		   root->fs_info->sb->s_blocksize_bits;
+	ins_size *= csum_size;
+	ins_size = min_t(u32, (unsigned long)item_end - (unsigned long)item,
+			      ins_size);
+	write_extent_buffer(leaf, sums->sums + index, (unsigned long)item,
+			    ins_size);
+
+	ins_size /= csum_size;
+	total_bytes += ins_size * root->sectorsize;
+	index += ins_size;
 
 	btrfs_mark_buffer_dirty(path->nodes[0]);
 	if (total_bytes < sums->len) {
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 4205ba752d40..72da4df53c9a 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -309,10 +309,6 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info,
 		ret = PTR_ERR(inode_root);
 		goto cleanup;
 	}
-	if (btrfs_root_refs(&inode_root->root_item) == 0) {
-		ret = -ENOENT;
-		goto cleanup;
-	}
 
 	key.objectid = defrag->ino;
 	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
@@ -600,20 +596,29 @@ void btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 		if (no_splits)
 			goto next;
 
-		if (em->block_start < EXTENT_MAP_LAST_BYTE &&
-		    em->start < start) {
+		if (em->start < start) {
 			split->start = em->start;
 			split->len = start - em->start;
-			split->orig_start = em->orig_start;
-			split->block_start = em->block_start;
 
-			if (compressed)
-				split->block_len = em->block_len;
-			else
-				split->block_len = split->len;
-			split->ram_bytes = em->ram_bytes;
-			split->orig_block_len = max(split->block_len,
-						    em->orig_block_len);
+			if (em->block_start < EXTENT_MAP_LAST_BYTE) {
+				split->orig_start = em->orig_start;
+				split->block_start = em->block_start;
+
+				if (compressed)
+					split->block_len = em->block_len;
+				else
+					split->block_len = split->len;
+				split->orig_block_len = max(split->block_len,
+						em->orig_block_len);
+				split->ram_bytes = em->ram_bytes;
+			} else {
+				split->orig_start = split->start;
+				split->block_len = 0;
+				split->block_start = em->block_start;
+				split->orig_block_len = 0;
+				split->ram_bytes = split->len;
+			}
+
 			split->generation = gen;
 			split->bdev = em->bdev;
 			split->flags = flags;
@@ -624,8 +629,7 @@ void btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 			split = split2;
 			split2 = NULL;
 		}
-		if (em->block_start < EXTENT_MAP_LAST_BYTE &&
-		    testend && em->start + em->len > start + len) {
+		if (testend && em->start + em->len > start + len) {
 			u64 diff = start + len - em->start;
 
 			split->start = start + len;
@@ -634,18 +638,28 @@ void btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 			split->flags = flags;
 			split->compress_type = em->compress_type;
 			split->generation = gen;
-			split->orig_block_len = max(em->block_len,
+
+			if (em->block_start < EXTENT_MAP_LAST_BYTE) {
+				split->orig_block_len = max(em->block_len,
 						    em->orig_block_len);
-			split->ram_bytes = em->ram_bytes;
 
-			if (compressed) {
-				split->block_len = em->block_len;
-				split->block_start = em->block_start;
-				split->orig_start = em->orig_start;
+				split->ram_bytes = em->ram_bytes;
+				if (compressed) {
+					split->block_len = em->block_len;
+					split->block_start = em->block_start;
+					split->orig_start = em->orig_start;
+				} else {
+					split->block_len = split->len;
+					split->block_start = em->block_start
+						+ diff;
+					split->orig_start = em->orig_start;
+				}
 			} else {
-				split->block_len = split->len;
-				split->block_start = em->block_start + diff;
-				split->orig_start = em->orig_start;
+				split->ram_bytes = split->len;
+				split->orig_start = split->start;
+				split->block_len = 0;
+				split->block_start = em->block_start;
+				split->orig_block_len = 0;
 			}
 
 			ret = add_extent_mapping(em_tree, split, modified);
@@ -1317,6 +1331,47 @@ fail:
 
 }
 
+static noinline int check_can_nocow(struct inode *inode, loff_t pos,
+				    size_t *write_bytes)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_ordered_extent *ordered;
+	u64 lockstart, lockend;
+	u64 num_bytes;
+	int ret;
+
+	lockstart = round_down(pos, root->sectorsize);
+	lockend = lockstart + round_up(*write_bytes, root->sectorsize) - 1;
+
+	while (1) {
+		lock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend);
+		ordered = btrfs_lookup_ordered_range(inode, lockstart,
+						     lockend - lockstart + 1);
+		if (!ordered) {
+			break;
+		}
+		unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		btrfs_put_ordered_extent(ordered);
+	}
+
+	num_bytes = lockend - lockstart + 1;
+	ret = can_nocow_extent(inode, lockstart, &num_bytes, NULL, NULL, NULL);
+	if (ret <= 0) {
+		ret = 0;
+	} else {
+		clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend,
+				 EXTENT_DIRTY | EXTENT_DELALLOC |
+				 EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, 0, 0,
+				 NULL, GFP_NOFS);
+		*write_bytes = min_t(size_t, *write_bytes, num_bytes);
+	}
+
+	unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend);
+
+	return ret;
+}
+
 static noinline ssize_t __btrfs_buffered_write(struct file *file,
 					       struct iov_iter *i,
 					       loff_t pos)
@@ -1324,10 +1379,12 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 	struct inode *inode = file_inode(file);
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct page **pages = NULL;
+	u64 release_bytes = 0;
 	unsigned long first_index;
 	size_t num_written = 0;
 	int nrptrs;
 	int ret = 0;
+	bool only_release_metadata = false;
 	bool force_page_uptodate = false;
 
 	nrptrs = min((iov_iter_count(i) + PAGE_CACHE_SIZE - 1) /
@@ -1348,6 +1405,7 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 					 offset);
 		size_t num_pages = (write_bytes + offset +
 				    PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+		size_t reserve_bytes;
 		size_t dirty_pages;
 		size_t copied;
 
@@ -1362,11 +1420,41 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 			break;
 		}
 
-		ret = btrfs_delalloc_reserve_space(inode,
-					num_pages << PAGE_CACHE_SHIFT);
+		reserve_bytes = num_pages << PAGE_CACHE_SHIFT;
+		ret = btrfs_check_data_free_space(inode, reserve_bytes);
+		if (ret == -ENOSPC &&
+		    (BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW |
+					      BTRFS_INODE_PREALLOC))) {
+			ret = check_can_nocow(inode, pos, &write_bytes);
+			if (ret > 0) {
+				only_release_metadata = true;
+				/*
+				 * our prealloc extent may be smaller than
+				 * write_bytes, so scale down.
+				 */
+				num_pages = (write_bytes + offset +
+					     PAGE_CACHE_SIZE - 1) >>
+					PAGE_CACHE_SHIFT;
+				reserve_bytes = num_pages << PAGE_CACHE_SHIFT;
+				ret = 0;
+			} else {
+				ret = -ENOSPC;
+			}
+		}
+
 		if (ret)
 			break;
 
+		ret = btrfs_delalloc_reserve_metadata(inode, reserve_bytes);
+		if (ret) {
+			if (!only_release_metadata)
+				btrfs_free_reserved_data_space(inode,
+							       reserve_bytes);
+			break;
+		}
+
+		release_bytes = reserve_bytes;
+
 		/*
 		 * This is going to setup the pages array with the number of
 		 * pages we want, so we don't really need to worry about the
@@ -1375,11 +1463,8 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 		ret = prepare_pages(root, file, pages, num_pages,
 				    pos, first_index, write_bytes,
 				    force_page_uptodate);
-		if (ret) {
-			btrfs_delalloc_release_space(inode,
-					num_pages << PAGE_CACHE_SHIFT);
+		if (ret)
 			break;
-		}
 
 		copied = btrfs_copy_from_user(pos, num_pages,
 					   write_bytes, pages, i);
@@ -1409,30 +1494,46 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 		 * managed to copy.
 		 */
 		if (num_pages > dirty_pages) {
+			release_bytes = (num_pages - dirty_pages) <<
+				PAGE_CACHE_SHIFT;
 			if (copied > 0) {
 				spin_lock(&BTRFS_I(inode)->lock);
 				BTRFS_I(inode)->outstanding_extents++;
 				spin_unlock(&BTRFS_I(inode)->lock);
 			}
-			btrfs_delalloc_release_space(inode,
-					(num_pages - dirty_pages) <<
-					PAGE_CACHE_SHIFT);
+			if (only_release_metadata)
+				btrfs_delalloc_release_metadata(inode,
+								release_bytes);
+			else
+				btrfs_delalloc_release_space(inode,
+							     release_bytes);
 		}
 
+		release_bytes = dirty_pages << PAGE_CACHE_SHIFT;
 		if (copied > 0) {
 			ret = btrfs_dirty_pages(root, inode, pages,
 						dirty_pages, pos, copied,
 						NULL);
 			if (ret) {
-				btrfs_delalloc_release_space(inode,
-					dirty_pages << PAGE_CACHE_SHIFT);
 				btrfs_drop_pages(pages, num_pages);
 				break;
 			}
 		}
 
+		release_bytes = 0;
 		btrfs_drop_pages(pages, num_pages);
 
+		if (only_release_metadata && copied > 0) {
+			u64 lockstart = round_down(pos, root->sectorsize);
+			u64 lockend = lockstart +
+				(dirty_pages << PAGE_CACHE_SHIFT) - 1;
+
+			set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart,
+				       lockend, EXTENT_NORESERVE, NULL,
+				       NULL, GFP_NOFS);
+			only_release_metadata = false;
+		}
+
 		cond_resched();
 
 		balance_dirty_pages_ratelimited(inode->i_mapping);
@@ -1445,6 +1546,13 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 
 	kfree(pages);
 
+	if (release_bytes) {
+		if (only_release_metadata)
+			btrfs_delalloc_release_metadata(inode, release_bytes);
+		else
+			btrfs_delalloc_release_space(inode, release_bytes);
+	}
+
 	return num_written ? num_written : ret;
 }
 
@@ -1610,7 +1718,7 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb,
 	 */
 	BTRFS_I(inode)->last_trans = root->fs_info->generation + 1;
 	BTRFS_I(inode)->last_sub_trans = root->log_transid;
-	if (num_written > 0 || num_written == -EIOCBQUEUED) {
+	if (num_written > 0) {
 		err = generic_write_sync(file, pos, num_written);
 		if (err < 0 && num_written > 0)
 			num_written = err;
@@ -1751,8 +1859,8 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 
 	ret = btrfs_log_dentry_safe(trans, root, dentry);
 	if (ret < 0) {
-		mutex_unlock(&inode->i_mutex);
-		goto out;
+		/* Fallthrough and commit/free transaction. */
+		ret = 1;
 	}
 
 	/* we've logged all the items and now have a consistent
@@ -2175,12 +2283,6 @@ static long btrfs_fallocate(struct file *file, int mode,
 			goto out_reserve_fail;
 	}
 
-	/*
-	 * wait for ordered IO before we have any locks.  We'll loop again
-	 * below with the locks held.
-	 */
-	btrfs_wait_ordered_range(inode, alloc_start, alloc_end - alloc_start);
-
 	mutex_lock(&inode->i_mutex);
 	ret = inode_newsize_ok(inode, alloc_end);
 	if (ret)
@@ -2191,8 +2293,23 @@ static long btrfs_fallocate(struct file *file, int mode,
 					alloc_start);
 		if (ret)
 			goto out;
+	} else {
+		/*
+		 * If we are fallocating from the end of the file onward we
+		 * need to zero out the end of the page if i_size lands in the
+		 * middle of a page.
+		 */
+		ret = btrfs_truncate_page(inode, inode->i_size, 0, 0);
+		if (ret)
+			goto out;
 	}
 
+	/*
+	 * wait for ordered IO before we have any locks.  We'll loop again
+	 * below with the locks held.
+	 */
+	btrfs_wait_ordered_range(inode, alloc_start, alloc_end - alloc_start);
+
 	locked_end = alloc_end - 1;
 	while (1) {
 		struct btrfs_ordered_extent *ordered;
@@ -2425,20 +2542,7 @@ static loff_t btrfs_file_llseek(struct file *file, loff_t offset, int whence)
 		}
 	}
 
-	if (offset < 0 && !(file->f_mode & FMODE_UNSIGNED_OFFSET)) {
-		offset = -EINVAL;
-		goto out;
-	}
-	if (offset > inode->i_sb->s_maxbytes) {
-		offset = -EINVAL;
-		goto out;
-	}
-
-	/* Special lock needed here? */
-	if (offset != file->f_pos) {
-		file->f_pos = offset;
-		file->f_version = 0;
-	}
+	offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes);
 out:
 	mutex_unlock(&inode->i_mutex);
 	return offset;
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index e53009657f0e..b4f9904c4c6b 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -213,7 +213,7 @@ int btrfs_check_trunc_cache_free_space(struct btrfs_root *root,
 	else
 		ret = 0;
 	spin_unlock(&rsv->lock);
-	return 0;
+	return ret;
 }
 
 int btrfs_truncate_free_space_cache(struct btrfs_root *root,
@@ -221,12 +221,10 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
 				    struct btrfs_path *path,
 				    struct inode *inode)
 {
-	loff_t oldsize;
 	int ret = 0;
 
-	oldsize = i_size_read(inode);
 	btrfs_i_size_write(inode, 0);
-	truncate_pagecache(inode, oldsize, 0);
+	truncate_pagecache(inode, 0);
 
 	/*
 	 * We don't need an orphan item because truncating the free space cache
@@ -308,7 +306,7 @@ static void io_ctl_unmap_page(struct io_ctl *io_ctl)
 
 static void io_ctl_map_page(struct io_ctl *io_ctl, int clear)
 {
-	BUG_ON(io_ctl->index >= io_ctl->num_pages);
+	ASSERT(io_ctl->index < io_ctl->num_pages);
 	io_ctl->page = io_ctl->pages[io_ctl->index++];
 	io_ctl->cur = kmap(io_ctl->page);
 	io_ctl->orig = io_ctl->cur;
@@ -673,8 +671,7 @@ static int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
 		btrfs_err(root->fs_info,
 			"free space inode generation (%llu) "
 			"did not match free space cache generation (%llu)",
-			(unsigned long long)BTRFS_I(inode)->generation,
-			(unsigned long long)generation);
+			BTRFS_I(inode)->generation, generation);
 		return 0;
 	}
 
@@ -729,7 +726,7 @@ static int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
 				goto free_cache;
 			}
 		} else {
-			BUG_ON(!num_bitmaps);
+			ASSERT(num_bitmaps);
 			num_bitmaps--;
 			e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
 			if (!e->bitmap) {
@@ -1029,7 +1026,7 @@ static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
 	leaf = path->nodes[0];
 	if (ret > 0) {
 		struct btrfs_key found_key;
-		BUG_ON(!path->slots[0]);
+		ASSERT(path->slots[0]);
 		path->slots[0]--;
 		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
 		if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
@@ -1117,7 +1114,7 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit,
 					  u64 offset)
 {
-	BUG_ON(offset < bitmap_start);
+	ASSERT(offset >= bitmap_start);
 	offset -= bitmap_start;
 	return (unsigned long)(div_u64(offset, unit));
 }
@@ -1272,7 +1269,7 @@ tree_search_offset(struct btrfs_free_space_ctl *ctl,
 		if (n) {
 			entry = rb_entry(n, struct btrfs_free_space,
 					offset_index);
-			BUG_ON(entry->offset > offset);
+			ASSERT(entry->offset <= offset);
 		} else {
 			if (fuzzy)
 				return entry;
@@ -1336,7 +1333,7 @@ static int link_free_space(struct btrfs_free_space_ctl *ctl,
 {
 	int ret = 0;
 
-	BUG_ON(!info->bitmap && !info->bytes);
+	ASSERT(info->bytes || info->bitmap);
 	ret = tree_insert_offset(&ctl->free_space_offset, info->offset,
 				 &info->offset_index, (info->bitmap != NULL));
 	if (ret)
@@ -1359,7 +1356,7 @@ static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
 
 	max_bitmaps = max(max_bitmaps, 1);
 
-	BUG_ON(ctl->total_bitmaps > max_bitmaps);
+	ASSERT(ctl->total_bitmaps <= max_bitmaps);
 
 	/*
 	 * The goal is to keep the total amount of memory used per 1gb of space
@@ -1403,7 +1400,7 @@ static inline void __bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
 
 	start = offset_to_bit(info->offset, ctl->unit, offset);
 	count = bytes_to_bits(bytes, ctl->unit);
-	BUG_ON(start + count > BITS_PER_BITMAP);
+	ASSERT(start + count <= BITS_PER_BITMAP);
 
 	bitmap_clear(info->bitmap, start, count);
 
@@ -1426,7 +1423,7 @@ static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
 
 	start = offset_to_bit(info->offset, ctl->unit, offset);
 	count = bytes_to_bits(bytes, ctl->unit);
-	BUG_ON(start + count > BITS_PER_BITMAP);
+	ASSERT(start + count <= BITS_PER_BITMAP);
 
 	bitmap_set(info->bitmap, start, count);
 
@@ -1434,13 +1431,19 @@ static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
 	ctl->free_space += bytes;
 }
 
+/*
+ * If we can not find suitable extent, we will use bytes to record
+ * the size of the max extent.
+ */
 static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 			 struct btrfs_free_space *bitmap_info, u64 *offset,
 			 u64 *bytes)
 {
 	unsigned long found_bits = 0;
+	unsigned long max_bits = 0;
 	unsigned long bits, i;
 	unsigned long next_zero;
+	unsigned long extent_bits;
 
 	i = offset_to_bit(bitmap_info->offset, ctl->unit,
 			  max_t(u64, *offset, bitmap_info->offset));
@@ -1449,9 +1452,12 @@ static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 	for_each_set_bit_from(i, bitmap_info->bitmap, BITS_PER_BITMAP) {
 		next_zero = find_next_zero_bit(bitmap_info->bitmap,
 					       BITS_PER_BITMAP, i);
-		if ((next_zero - i) >= bits) {
-			found_bits = next_zero - i;
+		extent_bits = next_zero - i;
+		if (extent_bits >= bits) {
+			found_bits = extent_bits;
 			break;
+		} else if (extent_bits > max_bits) {
+			max_bits = extent_bits;
 		}
 		i = next_zero;
 	}
@@ -1462,38 +1468,41 @@ static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 		return 0;
 	}
 
+	*bytes = (u64)(max_bits) * ctl->unit;
 	return -1;
 }
 
+/* Cache the size of the max extent in bytes */
 static struct btrfs_free_space *
 find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
-		unsigned long align)
+		unsigned long align, u64 *max_extent_size)
 {
 	struct btrfs_free_space *entry;
 	struct rb_node *node;
-	u64 ctl_off;
 	u64 tmp;
 	u64 align_off;
 	int ret;
 
 	if (!ctl->free_space_offset.rb_node)
-		return NULL;
+		goto out;
 
 	entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1);
 	if (!entry)
-		return NULL;
+		goto out;
 
 	for (node = &entry->offset_index; node; node = rb_next(node)) {
 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
-		if (entry->bytes < *bytes)
+		if (entry->bytes < *bytes) {
+			if (entry->bytes > *max_extent_size)
+				*max_extent_size = entry->bytes;
 			continue;
+		}
 
 		/* make sure the space returned is big enough
 		 * to match our requested alignment
 		 */
 		if (*bytes >= align) {
-			ctl_off = entry->offset - ctl->start;
-			tmp = ctl_off + align - 1;;
+			tmp = entry->offset - ctl->start + align - 1;
 			do_div(tmp, align);
 			tmp = tmp * align + ctl->start;
 			align_off = tmp - entry->offset;
@@ -1502,14 +1511,22 @@ find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
 			tmp = entry->offset;
 		}
 
-		if (entry->bytes < *bytes + align_off)
+		if (entry->bytes < *bytes + align_off) {
+			if (entry->bytes > *max_extent_size)
+				*max_extent_size = entry->bytes;
 			continue;
+		}
 
 		if (entry->bitmap) {
-			ret = search_bitmap(ctl, entry, &tmp, bytes);
+			u64 size = *bytes;
+
+			ret = search_bitmap(ctl, entry, &tmp, &size);
 			if (!ret) {
 				*offset = tmp;
+				*bytes = size;
 				return entry;
+			} else if (size > *max_extent_size) {
+				*max_extent_size = size;
 			}
 			continue;
 		}
@@ -1518,7 +1535,7 @@ find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
 		*bytes = entry->bytes - align_off;
 		return entry;
 	}
-
+out:
 	return NULL;
 }
 
@@ -1742,7 +1759,7 @@ no_cluster_bitmap:
 	bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
 					 1, 0);
 	if (!bitmap_info) {
-		BUG_ON(added);
+		ASSERT(added == 0);
 		goto new_bitmap;
 	}
 
@@ -1882,7 +1899,7 @@ out:
 
 	if (ret) {
 		printk(KERN_CRIT "btrfs: unable to add free space :%d\n", ret);
-		BUG_ON(ret == -EEXIST);
+		ASSERT(ret != -EEXIST);
 	}
 
 	return ret;
@@ -1991,8 +2008,7 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 		if (info->bytes >= bytes && !block_group->ro)
 			count++;
 		printk(KERN_CRIT "entry offset %llu, bytes %llu, bitmap %s\n",
-		       (unsigned long long)info->offset,
-		       (unsigned long long)info->bytes,
+		       info->offset, info->bytes,
 		       (info->bitmap) ? "yes" : "no");
 	}
 	printk(KERN_INFO "block group has cluster?: %s\n",
@@ -2120,7 +2136,8 @@ void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
 }
 
 u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
-			       u64 offset, u64 bytes, u64 empty_size)
+			       u64 offset, u64 bytes, u64 empty_size,
+			       u64 *max_extent_size)
 {
 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *entry = NULL;
@@ -2131,7 +2148,7 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 
 	spin_lock(&ctl->tree_lock);
 	entry = find_free_space(ctl, &offset, &bytes_search,
-				block_group->full_stripe_len);
+				block_group->full_stripe_len, max_extent_size);
 	if (!entry)
 		goto out;
 
@@ -2141,7 +2158,6 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 		if (!entry->bytes)
 			free_bitmap(ctl, entry);
 	} else {
-
 		unlink_free_space(ctl, entry);
 		align_gap_len = offset - entry->offset;
 		align_gap = entry->offset;
@@ -2155,7 +2171,6 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 		else
 			link_free_space(ctl, entry);
 	}
-
 out:
 	spin_unlock(&ctl->tree_lock);
 
@@ -2210,7 +2225,8 @@ int btrfs_return_cluster_to_free_space(
 static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
 				   struct btrfs_free_cluster *cluster,
 				   struct btrfs_free_space *entry,
-				   u64 bytes, u64 min_start)
+				   u64 bytes, u64 min_start,
+				   u64 *max_extent_size)
 {
 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	int err;
@@ -2222,8 +2238,11 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
 	search_bytes = bytes;
 
 	err = search_bitmap(ctl, entry, &search_start, &search_bytes);
-	if (err)
+	if (err) {
+		if (search_bytes > *max_extent_size)
+			*max_extent_size = search_bytes;
 		return 0;
+	}
 
 	ret = search_start;
 	__bitmap_clear_bits(ctl, entry, ret, bytes);
@@ -2238,7 +2257,7 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
  */
 u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
 			     struct btrfs_free_cluster *cluster, u64 bytes,
-			     u64 min_start)
+			     u64 min_start, u64 *max_extent_size)
 {
 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *entry = NULL;
@@ -2258,6 +2277,9 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
 
 	entry = rb_entry(node, struct btrfs_free_space, offset_index);
 	while(1) {
+		if (entry->bytes < bytes && entry->bytes > *max_extent_size)
+			*max_extent_size = entry->bytes;
+
 		if (entry->bytes < bytes ||
 		    (!entry->bitmap && entry->offset < min_start)) {
 			node = rb_next(&entry->offset_index);
@@ -2271,7 +2293,8 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
 		if (entry->bitmap) {
 			ret = btrfs_alloc_from_bitmap(block_group,
 						      cluster, entry, bytes,
-						      cluster->window_start);
+						      cluster->window_start,
+						      max_extent_size);
 			if (ret == 0) {
 				node = rb_next(&entry->offset_index);
 				if (!node)
@@ -2371,7 +2394,7 @@ again:
 	rb_erase(&entry->offset_index, &ctl->free_space_offset);
 	ret = tree_insert_offset(&cluster->root, entry->offset,
 				 &entry->offset_index, 1);
-	BUG_ON(ret); /* -EEXIST; Logic error */
+	ASSERT(!ret); /* -EEXIST; Logic error */
 
 	trace_btrfs_setup_cluster(block_group, cluster,
 				  total_found * ctl->unit, 1);
@@ -2464,7 +2487,7 @@ setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
 		ret = tree_insert_offset(&cluster->root, entry->offset,
 					 &entry->offset_index, 0);
 		total_size += entry->bytes;
-		BUG_ON(ret); /* -EEXIST; Logic error */
+		ASSERT(!ret); /* -EEXIST; Logic error */
 	} while (node && entry != last);
 
 	cluster->max_size = max_extent;
@@ -2525,8 +2548,7 @@ setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
  * returns zero and sets up cluster if things worked out, otherwise
  * it returns -enospc
  */
-int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *root,
+int btrfs_find_space_cluster(struct btrfs_root *root,
 			     struct btrfs_block_group_cache *block_group,
 			     struct btrfs_free_cluster *cluster,
 			     u64 offset, u64 bytes, u64 empty_size)
@@ -2856,7 +2878,7 @@ u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
 
 		ret = search_bitmap(ctl, entry, &offset, &count);
 		/* Logic error; Should be empty if it can't find anything */
-		BUG_ON(ret);
+		ASSERT(!ret);
 
 		ino = offset;
 		bitmap_clear_bits(ctl, entry, offset, 1);
@@ -2973,33 +2995,68 @@ int btrfs_write_out_ino_cache(struct btrfs_root *root,
 }
 
 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
-static struct btrfs_block_group_cache *init_test_block_group(void)
+/*
+ * Use this if you need to make a bitmap or extent entry specifically, it
+ * doesn't do any of the merging that add_free_space does, this acts a lot like
+ * how the free space cache loading stuff works, so you can get really weird
+ * configurations.
+ */
+int test_add_free_space_entry(struct btrfs_block_group_cache *cache,
+			      u64 offset, u64 bytes, bool bitmap)
 {
-	struct btrfs_block_group_cache *cache;
+	struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
+	struct btrfs_free_space *info = NULL, *bitmap_info;
+	void *map = NULL;
+	u64 bytes_added;
+	int ret;
 
-	cache = kzalloc(sizeof(*cache), GFP_NOFS);
-	if (!cache)
-		return NULL;
-	cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl),
-					GFP_NOFS);
-	if (!cache->free_space_ctl) {
-		kfree(cache);
-		return NULL;
+again:
+	if (!info) {
+		info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
+		if (!info)
+			return -ENOMEM;
 	}
 
-	cache->key.objectid = 0;
-	cache->key.offset = 1024 * 1024 * 1024;
-	cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
-	cache->sectorsize = 4096;
+	if (!bitmap) {
+		spin_lock(&ctl->tree_lock);
+		info->offset = offset;
+		info->bytes = bytes;
+		ret = link_free_space(ctl, info);
+		spin_unlock(&ctl->tree_lock);
+		if (ret)
+			kmem_cache_free(btrfs_free_space_cachep, info);
+		return ret;
+	}
 
-	spin_lock_init(&cache->lock);
-	INIT_LIST_HEAD(&cache->list);
-	INIT_LIST_HEAD(&cache->cluster_list);
-	INIT_LIST_HEAD(&cache->new_bg_list);
+	if (!map) {
+		map = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
+		if (!map) {
+			kmem_cache_free(btrfs_free_space_cachep, info);
+			return -ENOMEM;
+		}
+	}
 
-	btrfs_init_free_space_ctl(cache);
+	spin_lock(&ctl->tree_lock);
+	bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
+					 1, 0);
+	if (!bitmap_info) {
+		info->bitmap = map;
+		map = NULL;
+		add_new_bitmap(ctl, info, offset);
+		bitmap_info = info;
+	}
 
-	return cache;
+	bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
+	bytes -= bytes_added;
+	offset += bytes_added;
+	spin_unlock(&ctl->tree_lock);
+
+	if (bytes)
+		goto again;
+
+	if (map)
+		kfree(map);
+	return 0;
 }
 
 /*
@@ -3007,8 +3064,8 @@ static struct btrfs_block_group_cache *init_test_block_group(void)
  * just used to check the absence of space, so if there is free space in the
  * range at all we will return 1.
  */
-static int check_exists(struct btrfs_block_group_cache *cache, u64 offset,
-			u64 bytes)
+int test_check_exists(struct btrfs_block_group_cache *cache,
+		      u64 offset, u64 bytes)
 {
 	struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
 	struct btrfs_free_space *info;
@@ -3085,408 +3142,4 @@ out:
 	spin_unlock(&ctl->tree_lock);
 	return ret;
 }
-
-/*
- * Use this if you need to make a bitmap or extent entry specifically, it
- * doesn't do any of the merging that add_free_space does, this acts a lot like
- * how the free space cache loading stuff works, so you can get really weird
- * configurations.
- */
-static int add_free_space_entry(struct btrfs_block_group_cache *cache,
-				u64 offset, u64 bytes, bool bitmap)
-{
-	struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
-	struct btrfs_free_space *info = NULL, *bitmap_info;
-	void *map = NULL;
-	u64 bytes_added;
-	int ret;
-
-again:
-	if (!info) {
-		info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
-		if (!info)
-			return -ENOMEM;
-	}
-
-	if (!bitmap) {
-		spin_lock(&ctl->tree_lock);
-		info->offset = offset;
-		info->bytes = bytes;
-		ret = link_free_space(ctl, info);
-		spin_unlock(&ctl->tree_lock);
-		if (ret)
-			kmem_cache_free(btrfs_free_space_cachep, info);
-		return ret;
-	}
-
-	if (!map) {
-		map = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
-		if (!map) {
-			kmem_cache_free(btrfs_free_space_cachep, info);
-			return -ENOMEM;
-		}
-	}
-
-	spin_lock(&ctl->tree_lock);
-	bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
-					 1, 0);
-	if (!bitmap_info) {
-		info->bitmap = map;
-		map = NULL;
-		add_new_bitmap(ctl, info, offset);
-		bitmap_info = info;
-	}
-
-	bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
-	bytes -= bytes_added;
-	offset += bytes_added;
-	spin_unlock(&ctl->tree_lock);
-
-	if (bytes)
-		goto again;
-
-	if (map)
-		kfree(map);
-	return 0;
-}
-
-/*
- * This test just does basic sanity checking, making sure we can add an exten
- * entry and remove space from either end and the middle, and make sure we can
- * remove space that covers adjacent extent entries.
- */
-static int test_extents(struct btrfs_block_group_cache *cache)
-{
-	int ret = 0;
-
-	printk(KERN_ERR "Running extent only tests\n");
-
-	/* First just make sure we can remove an entire entry */
-	ret = btrfs_add_free_space(cache, 0, 4 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Error adding initial extents %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 0, 4 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Error removing extent %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, 0, 4 * 1024 * 1024)) {
-		printk(KERN_ERR "Full remove left some lingering space\n");
-		return -1;
-	}
-
-	/* Ok edge and middle cases now */
-	ret = btrfs_add_free_space(cache, 0, 4 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Error adding half extent %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 3 * 1024 * 1024, 1 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Error removing tail end %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 0, 1 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Error removing front end %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 2 * 1024 * 1024, 4096);
-	if (ret) {
-		printk(KERN_ERR "Error removing middle peice %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, 0, 1 * 1024 * 1024)) {
-		printk(KERN_ERR "Still have space at the front\n");
-		return -1;
-	}
-
-	if (check_exists(cache, 2 * 1024 * 1024, 4096)) {
-		printk(KERN_ERR "Still have space in the middle\n");
-		return -1;
-	}
-
-	if (check_exists(cache, 3 * 1024 * 1024, 1 * 1024 * 1024)) {
-		printk(KERN_ERR "Still have space at the end\n");
-		return -1;
-	}
-
-	/* Cleanup */
-	__btrfs_remove_free_space_cache(cache->free_space_ctl);
-
-	return 0;
-}
-
-static int test_bitmaps(struct btrfs_block_group_cache *cache)
-{
-	u64 next_bitmap_offset;
-	int ret;
-
-	printk(KERN_ERR "Running bitmap only tests\n");
-
-	ret = add_free_space_entry(cache, 0, 4 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't create a bitmap entry %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 0, 4 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Error removing bitmap full range %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, 0, 4 * 1024 * 1024)) {
-		printk(KERN_ERR "Left some space in bitmap\n");
-		return -1;
-	}
-
-	ret = add_free_space_entry(cache, 0, 4 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add to our bitmap entry %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 1 * 1024 * 1024, 2 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Couldn't remove middle chunk %d\n", ret);
-		return ret;
-	}
-
-	/*
-	 * The first bitmap we have starts at offset 0 so the next one is just
-	 * at the end of the first bitmap.
-	 */
-	next_bitmap_offset = (u64)(BITS_PER_BITMAP * 4096);
-
-	/* Test a bit straddling two bitmaps */
-	ret = add_free_space_entry(cache, next_bitmap_offset -
-				   (2 * 1024 * 1024), 4 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add space that straddles two bitmaps"
-		       " %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, next_bitmap_offset -
-				      (1 * 1024 * 1024), 2 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Couldn't remove overlapping space %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, next_bitmap_offset - (1 * 1024 * 1024),
-			 2 * 1024 * 1024)) {
-		printk(KERN_ERR "Left some space when removing overlapping\n");
-		return -1;
-	}
-
-	__btrfs_remove_free_space_cache(cache->free_space_ctl);
-
-	return 0;
-}
-
-/* This is the high grade jackassery */
-static int test_bitmaps_and_extents(struct btrfs_block_group_cache *cache)
-{
-	u64 bitmap_offset = (u64)(BITS_PER_BITMAP * 4096);
-	int ret;
-
-	printk(KERN_ERR "Running bitmap and extent tests\n");
-
-	/*
-	 * First let's do something simple, an extent at the same offset as the
-	 * bitmap, but the free space completely in the extent and then
-	 * completely in the bitmap.
-	 */
-	ret = add_free_space_entry(cache, 4 * 1024 * 1024, 1 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't create bitmap entry %d\n", ret);
-		return ret;
-	}
-
-	ret = add_free_space_entry(cache, 0, 1 * 1024 * 1024, 0);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add extent entry %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 0, 1 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Couldn't remove extent entry %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, 0, 1 * 1024 * 1024)) {
-		printk(KERN_ERR "Left remnants after our remove\n");
-		return -1;
-	}
-
-	/* Now to add back the extent entry and remove from the bitmap */
-	ret = add_free_space_entry(cache, 0, 1 * 1024 * 1024, 0);
-	if (ret) {
-		printk(KERN_ERR "Couldn't re-add extent entry %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 4 * 1024 * 1024, 1 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Couldn't remove from bitmap %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, 4 * 1024 * 1024, 1 * 1024 * 1024)) {
-		printk(KERN_ERR "Left remnants in the bitmap\n");
-		return -1;
-	}
-
-	/*
-	 * Ok so a little more evil, extent entry and bitmap at the same offset,
-	 * removing an overlapping chunk.
-	 */
-	ret = add_free_space_entry(cache, 1 * 1024 * 1024, 4 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add to a bitmap %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 512 * 1024, 3 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Couldn't remove overlapping space %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, 512 * 1024, 3 * 1024 * 1024)) {
-		printk(KERN_ERR "Left over peices after removing "
-		       "overlapping\n");
-		return -1;
-	}
-
-	__btrfs_remove_free_space_cache(cache->free_space_ctl);
-
-	/* Now with the extent entry offset into the bitmap */
-	ret = add_free_space_entry(cache, 4 * 1024 * 1024, 4 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add space to the bitmap %d\n", ret);
-		return ret;
-	}
-
-	ret = add_free_space_entry(cache, 2 * 1024 * 1024, 2 * 1024 * 1024, 0);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add extent to the cache %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 3 * 1024 * 1024, 4 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Problem removing overlapping space %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, 3 * 1024 * 1024, 4 * 1024 * 1024)) {
-		printk(KERN_ERR "Left something behind when removing space");
-		return -1;
-	}
-
-	/*
-	 * This has blown up in the past, the extent entry starts before the
-	 * bitmap entry, but we're trying to remove an offset that falls
-	 * completely within the bitmap range and is in both the extent entry
-	 * and the bitmap entry, looks like this
-	 *
-	 *   [ extent ]
-	 *      [ bitmap ]
-	 *        [ del ]
-	 */
-	__btrfs_remove_free_space_cache(cache->free_space_ctl);
-	ret = add_free_space_entry(cache, bitmap_offset + 4 * 1024 * 1024,
-				   4 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add bitmap %d\n", ret);
-		return ret;
-	}
-
-	ret = add_free_space_entry(cache, bitmap_offset - 1 * 1024 * 1024,
-				   5 * 1024 * 1024, 0);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add extent entry %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, bitmap_offset + 1 * 1024 * 1024,
-				      5 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Failed to free our space %d\n", ret);
-		return ret;
-	}
-
-	if (check_exists(cache, bitmap_offset + 1 * 1024 * 1024,
-			 5 * 1024 * 1024)) {
-		printk(KERN_ERR "Left stuff over\n");
-		return -1;
-	}
-
-	__btrfs_remove_free_space_cache(cache->free_space_ctl);
-
-	/*
-	 * This blew up before, we have part of the free space in a bitmap and
-	 * then the entirety of the rest of the space in an extent.  This used
-	 * to return -EAGAIN back from btrfs_remove_extent, make sure this
-	 * doesn't happen.
-	 */
-	ret = add_free_space_entry(cache, 1 * 1024 * 1024, 2 * 1024 * 1024, 1);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add bitmap entry %d\n", ret);
-		return ret;
-	}
-
-	ret = add_free_space_entry(cache, 3 * 1024 * 1024, 1 * 1024 * 1024, 0);
-	if (ret) {
-		printk(KERN_ERR "Couldn't add extent entry %d\n", ret);
-		return ret;
-	}
-
-	ret = btrfs_remove_free_space(cache, 1 * 1024 * 1024, 3 * 1024 * 1024);
-	if (ret) {
-		printk(KERN_ERR "Error removing bitmap and extent "
-		       "overlapping %d\n", ret);
-		return ret;
-	}
-
-	__btrfs_remove_free_space_cache(cache->free_space_ctl);
-	return 0;
-}
-
-void btrfs_test_free_space_cache(void)
-{
-	struct btrfs_block_group_cache *cache;
-
-	printk(KERN_ERR "Running btrfs free space cache tests\n");
-
-	cache = init_test_block_group();
-	if (!cache) {
-		printk(KERN_ERR "Couldn't run the tests\n");
-		return;
-	}
-
-	if (test_extents(cache))
-		goto out;
-	if (test_bitmaps(cache))
-		goto out;
-	if (test_bitmaps_and_extents(cache))
-		goto out;
-out:
-	__btrfs_remove_free_space_cache(cache->free_space_ctl);
-	kfree(cache->free_space_ctl);
-	kfree(cache);
-	printk(KERN_ERR "Free space cache tests finished\n");
-}
 #endif /* CONFIG_BTRFS_FS_RUN_SANITY_TESTS */
diff --git a/fs/btrfs/free-space-cache.h b/fs/btrfs/free-space-cache.h
index 8b7f19f44961..e737f92cf6d0 100644
--- a/fs/btrfs/free-space-cache.h
+++ b/fs/btrfs/free-space-cache.h
@@ -94,27 +94,31 @@ void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl);
 void btrfs_remove_free_space_cache(struct btrfs_block_group_cache
 				     *block_group);
 u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
-			       u64 offset, u64 bytes, u64 empty_size);
+			       u64 offset, u64 bytes, u64 empty_size,
+			       u64 *max_extent_size);
 u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root);
 void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 			   u64 bytes);
-int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *root,
+int btrfs_find_space_cluster(struct btrfs_root *root,
 			     struct btrfs_block_group_cache *block_group,
 			     struct btrfs_free_cluster *cluster,
 			     u64 offset, u64 bytes, u64 empty_size);
 void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster);
 u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
 			     struct btrfs_free_cluster *cluster, u64 bytes,
-			     u64 min_start);
+			     u64 min_start, u64 *max_extent_size);
 int btrfs_return_cluster_to_free_space(
 			       struct btrfs_block_group_cache *block_group,
 			       struct btrfs_free_cluster *cluster);
 int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
 			   u64 *trimmed, u64 start, u64 end, u64 minlen);
 
+/* Support functions for runnint our sanity tests */
 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
-void btrfs_test_free_space_cache(void);
+int test_add_free_space_entry(struct btrfs_block_group_cache *cache,
+			      u64 offset, u64 bytes, bool bitmap);
+int test_check_exists(struct btrfs_block_group_cache *cache,
+		      u64 offset, u64 bytes);
 #endif
 
 #endif
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index af978f7682b3..22ebc13b6c99 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -42,6 +42,7 @@
 #include <linux/mount.h>
 #include <linux/btrfs.h>
 #include <linux/blkdev.h>
+#include <linux/posix_acl_xattr.h>
 #include "compat.h"
 #include "ctree.h"
 #include "disk-io.h"
@@ -57,6 +58,7 @@
 #include "free-space-cache.h"
 #include "inode-map.h"
 #include "backref.h"
+#include "hash.h"
 
 struct btrfs_iget_args {
 	u64 ino;
@@ -228,12 +230,13 @@ fail:
  * does the checks required to make sure the data is small enough
  * to fit as an inline extent.
  */
-static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
-				 struct btrfs_root *root,
-				 struct inode *inode, u64 start, u64 end,
-				 size_t compressed_size, int compress_type,
-				 struct page **compressed_pages)
+static noinline int cow_file_range_inline(struct btrfs_root *root,
+					  struct inode *inode, u64 start,
+					  u64 end, size_t compressed_size,
+					  int compress_type,
+					  struct page **compressed_pages)
 {
+	struct btrfs_trans_handle *trans;
 	u64 isize = i_size_read(inode);
 	u64 actual_end = min(end + 1, isize);
 	u64 inline_len = actual_end - start;
@@ -254,9 +257,16 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
 		return 1;
 	}
 
+	trans = btrfs_join_transaction(root);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
+	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
+
 	ret = btrfs_drop_extents(trans, root, inode, start, aligned_end, 1);
-	if (ret)
-		return ret;
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out;
+	}
 
 	if (isize > actual_end)
 		inline_len = min_t(u64, isize, actual_end);
@@ -265,15 +275,18 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
 				   compress_type, compressed_pages);
 	if (ret && ret != -ENOSPC) {
 		btrfs_abort_transaction(trans, root, ret);
-		return ret;
+		goto out;
 	} else if (ret == -ENOSPC) {
-		return 1;
+		ret = 1;
+		goto out;
 	}
 
 	set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags);
 	btrfs_delalloc_release_metadata(inode, end + 1 - start);
 	btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0);
-	return 0;
+out:
+	btrfs_end_transaction(trans, root);
+	return ret;
 }
 
 struct async_extent {
@@ -341,7 +354,6 @@ static noinline int compress_file_range(struct inode *inode,
 					int *num_added)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct btrfs_trans_handle *trans;
 	u64 num_bytes;
 	u64 blocksize = root->sectorsize;
 	u64 actual_end;
@@ -459,45 +471,36 @@ again:
 	}
 cont:
 	if (start == 0) {
-		trans = btrfs_join_transaction(root);
-		if (IS_ERR(trans)) {
-			ret = PTR_ERR(trans);
-			trans = NULL;
-			goto cleanup_and_out;
-		}
-		trans->block_rsv = &root->fs_info->delalloc_block_rsv;
-
 		/* lets try to make an inline extent */
 		if (ret || total_in < (actual_end - start)) {
 			/* we didn't compress the entire range, try
 			 * to make an uncompressed inline extent.
 			 */
-			ret = cow_file_range_inline(trans, root, inode,
-						    start, end, 0, 0, NULL);
+			ret = cow_file_range_inline(root, inode, start, end,
+						    0, 0, NULL);
 		} else {
 			/* try making a compressed inline extent */
-			ret = cow_file_range_inline(trans, root, inode,
-						    start, end,
+			ret = cow_file_range_inline(root, inode, start, end,
 						    total_compressed,
 						    compress_type, pages);
 		}
 		if (ret <= 0) {
+			unsigned long clear_flags = EXTENT_DELALLOC |
+				EXTENT_DEFRAG;
+			clear_flags |= (ret < 0) ? EXTENT_DO_ACCOUNTING : 0;
+
 			/*
 			 * inline extent creation worked or returned error,
 			 * we don't need to create any more async work items.
 			 * Unlock and free up our temp pages.
 			 */
-			extent_clear_unlock_delalloc(inode,
-			     &BTRFS_I(inode)->io_tree,
-			     start, end, NULL,
-			     EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
-			     EXTENT_CLEAR_DELALLOC |
-			     EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK);
-
-			btrfs_end_transaction(trans, root);
+			extent_clear_unlock_delalloc(inode, start, end, NULL,
+						     clear_flags, PAGE_UNLOCK |
+						     PAGE_CLEAR_DIRTY |
+						     PAGE_SET_WRITEBACK |
+						     PAGE_END_WRITEBACK);
 			goto free_pages_out;
 		}
-		btrfs_end_transaction(trans, root);
 	}
 
 	if (will_compress) {
@@ -588,20 +591,6 @@ free_pages_out:
 	kfree(pages);
 
 	goto out;
-
-cleanup_and_out:
-	extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
-				     start, end, NULL,
-				     EXTENT_CLEAR_UNLOCK_PAGE |
-				     EXTENT_CLEAR_DIRTY |
-				     EXTENT_CLEAR_DELALLOC |
-				     EXTENT_SET_WRITEBACK |
-				     EXTENT_END_WRITEBACK);
-	if (!trans || IS_ERR(trans))
-		btrfs_error(root->fs_info, ret, "Failed to join transaction");
-	else
-		btrfs_abort_transaction(trans, root, ret);
-	goto free_pages_out;
 }
 
 /*
@@ -615,7 +604,6 @@ static noinline int submit_compressed_extents(struct inode *inode,
 {
 	struct async_extent *async_extent;
 	u64 alloc_hint = 0;
-	struct btrfs_trans_handle *trans;
 	struct btrfs_key ins;
 	struct extent_map *em;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -676,20 +664,10 @@ retry:
 		lock_extent(io_tree, async_extent->start,
 			    async_extent->start + async_extent->ram_size - 1);
 
-		trans = btrfs_join_transaction(root);
-		if (IS_ERR(trans)) {
-			ret = PTR_ERR(trans);
-		} else {
-			trans->block_rsv = &root->fs_info->delalloc_block_rsv;
-			ret = btrfs_reserve_extent(trans, root,
+		ret = btrfs_reserve_extent(root,
 					   async_extent->compressed_size,
 					   async_extent->compressed_size,
 					   0, alloc_hint, &ins, 1);
-			if (ret && ret != -ENOSPC)
-				btrfs_abort_transaction(trans, root, ret);
-			btrfs_end_transaction(trans, root);
-		}
-
 		if (ret) {
 			int i;
 
@@ -701,8 +679,12 @@ retry:
 			async_extent->nr_pages = 0;
 			async_extent->pages = NULL;
 
-			if (ret == -ENOSPC)
+			if (ret == -ENOSPC) {
+				unlock_extent(io_tree, async_extent->start,
+					      async_extent->start +
+					      async_extent->ram_size - 1);
 				goto retry;
+			}
 			goto out_free;
 		}
 
@@ -764,16 +746,12 @@ retry:
 		/*
 		 * clear dirty, set writeback and unlock the pages.
 		 */
-		extent_clear_unlock_delalloc(inode,
-				&BTRFS_I(inode)->io_tree,
-				async_extent->start,
+		extent_clear_unlock_delalloc(inode, async_extent->start,
 				async_extent->start +
 				async_extent->ram_size - 1,
-				NULL, EXTENT_CLEAR_UNLOCK_PAGE |
-				EXTENT_CLEAR_UNLOCK |
-				EXTENT_CLEAR_DELALLOC |
-				EXTENT_CLEAR_DIRTY | EXTENT_SET_WRITEBACK);
-
+				NULL, EXTENT_LOCKED | EXTENT_DELALLOC,
+				PAGE_UNLOCK | PAGE_CLEAR_DIRTY |
+				PAGE_SET_WRITEBACK);
 		ret = btrfs_submit_compressed_write(inode,
 				    async_extent->start,
 				    async_extent->ram_size,
@@ -792,16 +770,13 @@ out:
 out_free_reserve:
 	btrfs_free_reserved_extent(root, ins.objectid, ins.offset);
 out_free:
-	extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
-				     async_extent->start,
+	extent_clear_unlock_delalloc(inode, async_extent->start,
 				     async_extent->start +
 				     async_extent->ram_size - 1,
-				     NULL, EXTENT_CLEAR_UNLOCK_PAGE |
-				     EXTENT_CLEAR_UNLOCK |
-				     EXTENT_CLEAR_DELALLOC |
-				     EXTENT_CLEAR_DIRTY |
-				     EXTENT_SET_WRITEBACK |
-				     EXTENT_END_WRITEBACK);
+				     NULL, EXTENT_LOCKED | EXTENT_DELALLOC |
+				     EXTENT_DEFRAG | EXTENT_DO_ACCOUNTING,
+				     PAGE_UNLOCK | PAGE_CLEAR_DIRTY |
+				     PAGE_SET_WRITEBACK | PAGE_END_WRITEBACK);
 	kfree(async_extent);
 	goto again;
 }
@@ -851,14 +826,13 @@ static u64 get_extent_allocation_hint(struct inode *inode, u64 start,
  * required to start IO on it.  It may be clean and already done with
  * IO when we return.
  */
-static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
-				     struct inode *inode,
-				     struct btrfs_root *root,
-				     struct page *locked_page,
-				     u64 start, u64 end, int *page_started,
-				     unsigned long *nr_written,
-				     int unlock)
+static noinline int cow_file_range(struct inode *inode,
+				   struct page *locked_page,
+				   u64 start, u64 end, int *page_started,
+				   unsigned long *nr_written,
+				   int unlock)
 {
+	struct btrfs_root *root = BTRFS_I(inode)->root;
 	u64 alloc_hint = 0;
 	u64 num_bytes;
 	unsigned long ram_size;
@@ -879,29 +853,24 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
 	/* if this is a small write inside eof, kick off defrag */
 	if (num_bytes < 64 * 1024 &&
 	    (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size))
-		btrfs_add_inode_defrag(trans, inode);
+		btrfs_add_inode_defrag(NULL, inode);
 
 	if (start == 0) {
 		/* lets try to make an inline extent */
-		ret = cow_file_range_inline(trans, root, inode,
-					    start, end, 0, 0, NULL);
+		ret = cow_file_range_inline(root, inode, start, end, 0, 0,
+					    NULL);
 		if (ret == 0) {
-			extent_clear_unlock_delalloc(inode,
-				     &BTRFS_I(inode)->io_tree,
-				     start, end, NULL,
-				     EXTENT_CLEAR_UNLOCK_PAGE |
-				     EXTENT_CLEAR_UNLOCK |
-				     EXTENT_CLEAR_DELALLOC |
-				     EXTENT_CLEAR_DIRTY |
-				     EXTENT_SET_WRITEBACK |
-				     EXTENT_END_WRITEBACK);
+			extent_clear_unlock_delalloc(inode, start, end, NULL,
+				     EXTENT_LOCKED | EXTENT_DELALLOC |
+				     EXTENT_DEFRAG, PAGE_UNLOCK |
+				     PAGE_CLEAR_DIRTY | PAGE_SET_WRITEBACK |
+				     PAGE_END_WRITEBACK);
 
 			*nr_written = *nr_written +
 			     (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE;
 			*page_started = 1;
 			goto out;
 		} else if (ret < 0) {
-			btrfs_abort_transaction(trans, root, ret);
 			goto out_unlock;
 		}
 	}
@@ -916,13 +885,11 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
 		unsigned long op;
 
 		cur_alloc_size = disk_num_bytes;
-		ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
+		ret = btrfs_reserve_extent(root, cur_alloc_size,
 					   root->sectorsize, 0, alloc_hint,
 					   &ins, 1);
-		if (ret < 0) {
-			btrfs_abort_transaction(trans, root, ret);
+		if (ret < 0)
 			goto out_unlock;
-		}
 
 		em = alloc_extent_map();
 		if (!em) {
@@ -968,10 +935,8 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
 		    BTRFS_DATA_RELOC_TREE_OBJECTID) {
 			ret = btrfs_reloc_clone_csums(inode, start,
 						      cur_alloc_size);
-			if (ret) {
-				btrfs_abort_transaction(trans, root, ret);
+			if (ret)
 				goto out_reserve;
-			}
 		}
 
 		if (disk_num_bytes < cur_alloc_size)
@@ -984,13 +949,13 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
 		 * Do set the Private2 bit so we know this page was properly
 		 * setup for writepage
 		 */
-		op = unlock ? EXTENT_CLEAR_UNLOCK_PAGE : 0;
-		op |= EXTENT_CLEAR_UNLOCK | EXTENT_CLEAR_DELALLOC |
-			EXTENT_SET_PRIVATE2;
+		op = unlock ? PAGE_UNLOCK : 0;
+		op |= PAGE_SET_PRIVATE2;
 
-		extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
-					     start, start + ram_size - 1,
-					     locked_page, op);
+		extent_clear_unlock_delalloc(inode, start,
+					     start + ram_size - 1, locked_page,
+					     EXTENT_LOCKED | EXTENT_DELALLOC,
+					     op);
 		disk_num_bytes -= cur_alloc_size;
 		num_bytes -= cur_alloc_size;
 		alloc_hint = ins.objectid + ins.offset;
@@ -1002,52 +967,14 @@ out:
 out_reserve:
 	btrfs_free_reserved_extent(root, ins.objectid, ins.offset);
 out_unlock:
-	extent_clear_unlock_delalloc(inode,
-		     &BTRFS_I(inode)->io_tree,
-		     start, end, locked_page,
-		     EXTENT_CLEAR_UNLOCK_PAGE |
-		     EXTENT_CLEAR_UNLOCK |
-		     EXTENT_CLEAR_DELALLOC |
-		     EXTENT_CLEAR_DIRTY |
-		     EXTENT_SET_WRITEBACK |
-		     EXTENT_END_WRITEBACK);
-
+	extent_clear_unlock_delalloc(inode, start, end, locked_page,
+				     EXTENT_LOCKED | EXTENT_DO_ACCOUNTING |
+				     EXTENT_DELALLOC | EXTENT_DEFRAG,
+				     PAGE_UNLOCK | PAGE_CLEAR_DIRTY |
+				     PAGE_SET_WRITEBACK | PAGE_END_WRITEBACK);
 	goto out;
 }
 
-static noinline int cow_file_range(struct inode *inode,
-				   struct page *locked_page,
-				   u64 start, u64 end, int *page_started,
-				   unsigned long *nr_written,
-				   int unlock)
-{
-	struct btrfs_trans_handle *trans;
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	int ret;
-
-	trans = btrfs_join_transaction(root);
-	if (IS_ERR(trans)) {
-		extent_clear_unlock_delalloc(inode,
-			     &BTRFS_I(inode)->io_tree,
-			     start, end, locked_page,
-			     EXTENT_CLEAR_UNLOCK_PAGE |
-			     EXTENT_CLEAR_UNLOCK |
-			     EXTENT_CLEAR_DELALLOC |
-			     EXTENT_CLEAR_DIRTY |
-			     EXTENT_SET_WRITEBACK |
-			     EXTENT_END_WRITEBACK);
-		return PTR_ERR(trans);
-	}
-	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
-
-	ret = __cow_file_range(trans, inode, root, locked_page, start, end,
-			       page_started, nr_written, unlock);
-
-	btrfs_end_transaction(trans, root);
-
-	return ret;
-}
-
 /*
  * work queue call back to started compression on a file and pages
  */
@@ -1215,15 +1142,13 @@ static noinline int run_delalloc_nocow(struct inode *inode,
 
 	path = btrfs_alloc_path();
 	if (!path) {
-		extent_clear_unlock_delalloc(inode,
-			     &BTRFS_I(inode)->io_tree,
-			     start, end, locked_page,
-			     EXTENT_CLEAR_UNLOCK_PAGE |
-			     EXTENT_CLEAR_UNLOCK |
-			     EXTENT_CLEAR_DELALLOC |
-			     EXTENT_CLEAR_DIRTY |
-			     EXTENT_SET_WRITEBACK |
-			     EXTENT_END_WRITEBACK);
+		extent_clear_unlock_delalloc(inode, start, end, locked_page,
+					     EXTENT_LOCKED | EXTENT_DELALLOC |
+					     EXTENT_DO_ACCOUNTING |
+					     EXTENT_DEFRAG, PAGE_UNLOCK |
+					     PAGE_CLEAR_DIRTY |
+					     PAGE_SET_WRITEBACK |
+					     PAGE_END_WRITEBACK);
 		return -ENOMEM;
 	}
 
@@ -1235,15 +1160,13 @@ static noinline int run_delalloc_nocow(struct inode *inode,
 		trans = btrfs_join_transaction(root);
 
 	if (IS_ERR(trans)) {
-		extent_clear_unlock_delalloc(inode,
-			     &BTRFS_I(inode)->io_tree,
-			     start, end, locked_page,
-			     EXTENT_CLEAR_UNLOCK_PAGE |
-			     EXTENT_CLEAR_UNLOCK |
-			     EXTENT_CLEAR_DELALLOC |
-			     EXTENT_CLEAR_DIRTY |
-			     EXTENT_SET_WRITEBACK |
-			     EXTENT_END_WRITEBACK);
+		extent_clear_unlock_delalloc(inode, start, end, locked_page,
+					     EXTENT_LOCKED | EXTENT_DELALLOC |
+					     EXTENT_DO_ACCOUNTING |
+					     EXTENT_DEFRAG, PAGE_UNLOCK |
+					     PAGE_CLEAR_DIRTY |
+					     PAGE_SET_WRITEBACK |
+					     PAGE_END_WRITEBACK);
 		btrfs_free_path(path);
 		return PTR_ERR(trans);
 	}
@@ -1363,9 +1286,9 @@ out_check:
 
 		btrfs_release_path(path);
 		if (cow_start != (u64)-1) {
-			ret = __cow_file_range(trans, inode, root, locked_page,
-					       cow_start, found_key.offset - 1,
-					       page_started, nr_written, 1);
+			ret = cow_file_range(inode, locked_page,
+					     cow_start, found_key.offset - 1,
+					     page_started, nr_written, 1);
 			if (ret) {
 				btrfs_abort_transaction(trans, root, ret);
 				goto error;
@@ -1422,11 +1345,11 @@ out_check:
 			}
 		}
 
-		extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
-				cur_offset, cur_offset + num_bytes - 1,
-				locked_page, EXTENT_CLEAR_UNLOCK_PAGE |
-				EXTENT_CLEAR_UNLOCK | EXTENT_CLEAR_DELALLOC |
-				EXTENT_SET_PRIVATE2);
+		extent_clear_unlock_delalloc(inode, cur_offset,
+					     cur_offset + num_bytes - 1,
+					     locked_page, EXTENT_LOCKED |
+					     EXTENT_DELALLOC, PAGE_UNLOCK |
+					     PAGE_SET_PRIVATE2);
 		cur_offset = extent_end;
 		if (cur_offset > end)
 			break;
@@ -1439,9 +1362,8 @@ out_check:
 	}
 
 	if (cow_start != (u64)-1) {
-		ret = __cow_file_range(trans, inode, root, locked_page,
-				       cow_start, end,
-				       page_started, nr_written, 1);
+		ret = cow_file_range(inode, locked_page, cow_start, end,
+				     page_started, nr_written, 1);
 		if (ret) {
 			btrfs_abort_transaction(trans, root, ret);
 			goto error;
@@ -1454,16 +1376,13 @@ error:
 		ret = err;
 
 	if (ret && cur_offset < end)
-		extent_clear_unlock_delalloc(inode,
-			     &BTRFS_I(inode)->io_tree,
-			     cur_offset, end, locked_page,
-			     EXTENT_CLEAR_UNLOCK_PAGE |
-			     EXTENT_CLEAR_UNLOCK |
-			     EXTENT_CLEAR_DELALLOC |
-			     EXTENT_CLEAR_DIRTY |
-			     EXTENT_SET_WRITEBACK |
-			     EXTENT_END_WRITEBACK);
-
+		extent_clear_unlock_delalloc(inode, cur_offset, end,
+					     locked_page, EXTENT_LOCKED |
+					     EXTENT_DELALLOC | EXTENT_DEFRAG |
+					     EXTENT_DO_ACCOUNTING, PAGE_UNLOCK |
+					     PAGE_CLEAR_DIRTY |
+					     PAGE_SET_WRITEBACK |
+					     PAGE_END_WRITEBACK);
 	btrfs_free_path(path);
 	return ret;
 }
@@ -1529,6 +1448,46 @@ static void btrfs_merge_extent_hook(struct inode *inode,
 	spin_unlock(&BTRFS_I(inode)->lock);
 }
 
+static void btrfs_add_delalloc_inodes(struct btrfs_root *root,
+				      struct inode *inode)
+{
+	spin_lock(&root->delalloc_lock);
+	if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+		list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
+			      &root->delalloc_inodes);
+		set_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+			&BTRFS_I(inode)->runtime_flags);
+		root->nr_delalloc_inodes++;
+		if (root->nr_delalloc_inodes == 1) {
+			spin_lock(&root->fs_info->delalloc_root_lock);
+			BUG_ON(!list_empty(&root->delalloc_root));
+			list_add_tail(&root->delalloc_root,
+				      &root->fs_info->delalloc_roots);
+			spin_unlock(&root->fs_info->delalloc_root_lock);
+		}
+	}
+	spin_unlock(&root->delalloc_lock);
+}
+
+static void btrfs_del_delalloc_inode(struct btrfs_root *root,
+				     struct inode *inode)
+{
+	spin_lock(&root->delalloc_lock);
+	if (!list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+		list_del_init(&BTRFS_I(inode)->delalloc_inodes);
+		clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+			  &BTRFS_I(inode)->runtime_flags);
+		root->nr_delalloc_inodes--;
+		if (!root->nr_delalloc_inodes) {
+			spin_lock(&root->fs_info->delalloc_root_lock);
+			BUG_ON(list_empty(&root->delalloc_root));
+			list_del_init(&root->delalloc_root);
+			spin_unlock(&root->fs_info->delalloc_root_lock);
+		}
+	}
+	spin_unlock(&root->delalloc_lock);
+}
+
 /*
  * extent_io.c set_bit_hook, used to track delayed allocation
  * bytes in this file, and to maintain the list of inodes that
@@ -1561,16 +1520,8 @@ static void btrfs_set_bit_hook(struct inode *inode,
 		spin_lock(&BTRFS_I(inode)->lock);
 		BTRFS_I(inode)->delalloc_bytes += len;
 		if (do_list && !test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-					 &BTRFS_I(inode)->runtime_flags)) {
-			spin_lock(&root->fs_info->delalloc_lock);
-			if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
-				list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
-					      &root->fs_info->delalloc_inodes);
-				set_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-					&BTRFS_I(inode)->runtime_flags);
-			}
-			spin_unlock(&root->fs_info->delalloc_lock);
-		}
+					 &BTRFS_I(inode)->runtime_flags))
+			btrfs_add_delalloc_inodes(root, inode);
 		spin_unlock(&BTRFS_I(inode)->lock);
 	}
 }
@@ -1604,7 +1555,7 @@ static void btrfs_clear_bit_hook(struct inode *inode,
 			btrfs_delalloc_release_metadata(inode, len);
 
 		if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID
-		    && do_list)
+		    && do_list && !(state->state & EXTENT_NORESERVE))
 			btrfs_free_reserved_data_space(inode, len);
 
 		__percpu_counter_add(&root->fs_info->delalloc_bytes, -len,
@@ -1613,15 +1564,8 @@ static void btrfs_clear_bit_hook(struct inode *inode,
 		BTRFS_I(inode)->delalloc_bytes -= len;
 		if (do_list && BTRFS_I(inode)->delalloc_bytes == 0 &&
 		    test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-			     &BTRFS_I(inode)->runtime_flags)) {
-			spin_lock(&root->fs_info->delalloc_lock);
-			if (!list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
-				list_del_init(&BTRFS_I(inode)->delalloc_inodes);
-				clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-					  &BTRFS_I(inode)->runtime_flags);
-			}
-			spin_unlock(&root->fs_info->delalloc_lock);
-		}
+			     &BTRFS_I(inode)->runtime_flags))
+			btrfs_del_delalloc_inode(root, inode);
 		spin_unlock(&BTRFS_I(inode)->lock);
 	}
 }
@@ -2101,6 +2045,7 @@ static noinline int record_one_backref(u64 inum, u64 offset, u64 root_id,
 		WARN_ON(1);
 		return ret;
 	}
+	ret = 0;
 
 	while (1) {
 		cond_resched();
@@ -2135,16 +2080,21 @@ static noinline int record_one_backref(u64 inum, u64 offset, u64 root_id,
 		if (btrfs_file_extent_disk_bytenr(leaf, extent) != old->bytenr)
 			continue;
 
-		extent_offset = btrfs_file_extent_offset(leaf, extent);
-		if (key.offset - extent_offset != offset)
+		/*
+		 * 'offset' refers to the exact key.offset,
+		 * NOT the 'offset' field in btrfs_extent_data_ref, ie.
+		 * (key.offset - extent_offset).
+		 */
+		if (key.offset != offset)
 			continue;
 
+		extent_offset = btrfs_file_extent_offset(leaf, extent);
 		num_bytes = btrfs_file_extent_num_bytes(leaf, extent);
+
 		if (extent_offset >= old->extent_offset + old->offset +
 		    old->len || extent_offset + num_bytes <=
 		    old->extent_offset + old->offset)
 			continue;
-
 		break;
 	}
 
@@ -2156,7 +2106,7 @@ static noinline int record_one_backref(u64 inum, u64 offset, u64 root_id,
 
 	backref->root_id = root_id;
 	backref->inum = inum;
-	backref->file_pos = offset + extent_offset;
+	backref->file_pos = offset;
 	backref->num_bytes = num_bytes;
 	backref->extent_offset = extent_offset;
 	backref->generation = btrfs_file_extent_generation(leaf, extent);
@@ -2179,7 +2129,8 @@ static noinline bool record_extent_backrefs(struct btrfs_path *path,
 	new->path = path;
 
 	list_for_each_entry_safe(old, tmp, &new->head, list) {
-		ret = iterate_inodes_from_logical(old->bytenr, fs_info,
+		ret = iterate_inodes_from_logical(old->bytenr +
+						  old->extent_offset, fs_info,
 						  path, record_one_backref,
 						  old);
 		BUG_ON(ret < 0 && ret != -ENOENT);
@@ -2199,16 +2150,18 @@ static noinline bool record_extent_backrefs(struct btrfs_path *path,
 
 static int relink_is_mergable(struct extent_buffer *leaf,
 			      struct btrfs_file_extent_item *fi,
-			      u64 disk_bytenr)
+			      struct new_sa_defrag_extent *new)
 {
-	if (btrfs_file_extent_disk_bytenr(leaf, fi) != disk_bytenr)
+	if (btrfs_file_extent_disk_bytenr(leaf, fi) != new->bytenr)
 		return 0;
 
 	if (btrfs_file_extent_type(leaf, fi) != BTRFS_FILE_EXTENT_REG)
 		return 0;
 
-	if (btrfs_file_extent_compression(leaf, fi) ||
-	    btrfs_file_extent_encryption(leaf, fi) ||
+	if (btrfs_file_extent_compression(leaf, fi) != new->compress_type)
+		return 0;
+
+	if (btrfs_file_extent_encryption(leaf, fi) ||
 	    btrfs_file_extent_other_encoding(leaf, fi))
 		return 0;
 
@@ -2263,11 +2216,6 @@ static noinline int relink_extent_backref(struct btrfs_path *path,
 			return 0;
 		return PTR_ERR(root);
 	}
-	if (btrfs_root_refs(&root->root_item) == 0) {
-		srcu_read_unlock(&fs_info->subvol_srcu, index);
-		/* parse ENOENT to 0 */
-		return 0;
-	}
 
 	/* step 2: get inode */
 	key.objectid = backref->inum;
@@ -2357,8 +2305,8 @@ again:
 				    struct btrfs_file_extent_item);
 		extent_len = btrfs_file_extent_num_bytes(leaf, fi);
 
-		if (relink_is_mergable(leaf, fi, new->bytenr) &&
-		    extent_len + found_key.offset == start) {
+		if (extent_len + found_key.offset == start &&
+		    relink_is_mergable(leaf, fi, new)) {
 			btrfs_set_file_extent_num_bytes(leaf, fi,
 							extent_len + len);
 			btrfs_mark_buffer_dirty(leaf);
@@ -2614,8 +2562,10 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent)
 	struct extent_state *cached_state = NULL;
 	struct new_sa_defrag_extent *new = NULL;
 	int compress_type = 0;
-	int ret;
+	int ret = 0;
+	u64 logical_len = ordered_extent->len;
 	bool nolock;
+	bool truncated = false;
 
 	nolock = btrfs_is_free_space_inode(inode);
 
@@ -2624,6 +2574,14 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent)
 		goto out;
 	}
 
+	if (test_bit(BTRFS_ORDERED_TRUNCATED, &ordered_extent->flags)) {
+		truncated = true;
+		logical_len = ordered_extent->truncated_len;
+		/* Truncated the entire extent, don't bother adding */
+		if (!logical_len)
+			goto out;
+	}
+
 	if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
 		BUG_ON(!list_empty(&ordered_extent->list)); /* Logic error */
 		btrfs_ordered_update_i_size(inode, 0, ordered_extent);
@@ -2679,15 +2637,14 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent)
 		ret = btrfs_mark_extent_written(trans, inode,
 						ordered_extent->file_offset,
 						ordered_extent->file_offset +
-						ordered_extent->len);
+						logical_len);
 	} else {
 		BUG_ON(root == root->fs_info->tree_root);
 		ret = insert_reserved_file_extent(trans, inode,
 						ordered_extent->file_offset,
 						ordered_extent->start,
 						ordered_extent->disk_len,
-						ordered_extent->len,
-						ordered_extent->len,
+						logical_len, logical_len,
 						compress_type, 0, 0,
 						BTRFS_FILE_EXTENT_REG);
 	}
@@ -2719,17 +2676,27 @@ out:
 	if (trans)
 		btrfs_end_transaction(trans, root);
 
-	if (ret) {
-		clear_extent_uptodate(io_tree, ordered_extent->file_offset,
-				      ordered_extent->file_offset +
-				      ordered_extent->len - 1, NULL, GFP_NOFS);
+	if (ret || truncated) {
+		u64 start, end;
+
+		if (truncated)
+			start = ordered_extent->file_offset + logical_len;
+		else
+			start = ordered_extent->file_offset;
+		end = ordered_extent->file_offset + ordered_extent->len - 1;
+		clear_extent_uptodate(io_tree, start, end, NULL, GFP_NOFS);
+
+		/* Drop the cache for the part of the extent we didn't write. */
+		btrfs_drop_extent_cache(inode, start, end, 0);
 
 		/*
 		 * If the ordered extent had an IOERR or something else went
 		 * wrong we need to return the space for this ordered extent
-		 * back to the allocator.
+		 * back to the allocator.  We only free the extent in the
+		 * truncated case if we didn't write out the extent at all.
 		 */
-		if (!test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags) &&
+		if ((ret || !logical_len) &&
+		    !test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags) &&
 		    !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags))
 			btrfs_free_reserved_extent(root, ordered_extent->start,
 						   ordered_extent->disk_len);
@@ -2793,16 +2760,16 @@ static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
  * if there's a match, we allow the bio to finish.  If not, the code in
  * extent_io.c will try to find good copies for us.
  */
-static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
-			       struct extent_state *state, int mirror)
+static int btrfs_readpage_end_io_hook(struct btrfs_io_bio *io_bio,
+				      u64 phy_offset, struct page *page,
+				      u64 start, u64 end, int mirror)
 {
 	size_t offset = start - page_offset(page);
 	struct inode *inode = page->mapping->host;
 	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
 	char *kaddr;
-	u64 private = ~(u32)0;
-	int ret;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u32 csum_expected;
 	u32 csum = ~(u32)0;
 	static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL,
 	                              DEFAULT_RATELIMIT_BURST);
@@ -2822,19 +2789,13 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		return 0;
 	}
 
-	if (state && state->start == start) {
-		private = state->private;
-		ret = 0;
-	} else {
-		ret = get_state_private(io_tree, start, &private);
-	}
-	kaddr = kmap_atomic(page);
-	if (ret)
-		goto zeroit;
+	phy_offset >>= inode->i_sb->s_blocksize_bits;
+	csum_expected = *(((u32 *)io_bio->csum) + phy_offset);
 
+	kaddr = kmap_atomic(page);
 	csum = btrfs_csum_data(kaddr + offset, csum,  end - start + 1);
 	btrfs_csum_final(csum, (char *)&csum);
-	if (csum != private)
+	if (csum != csum_expected)
 		goto zeroit;
 
 	kunmap_atomic(kaddr);
@@ -2843,14 +2804,12 @@ good:
 
 zeroit:
 	if (__ratelimit(&_rs))
-		btrfs_info(root->fs_info, "csum failed ino %llu off %llu csum %u private %llu",
-			(unsigned long long)btrfs_ino(page->mapping->host),
-			(unsigned long long)start, csum,
-			(unsigned long long)private);
+		btrfs_info(root->fs_info, "csum failed ino %llu off %llu csum %u expected csum %u",
+			btrfs_ino(page->mapping->host), start, csum, csum_expected);
 	memset(kaddr + offset, 1, end - start + 1);
 	flush_dcache_page(page);
 	kunmap_atomic(kaddr);
-	if (private == 0)
+	if (csum_expected == 0)
 		return 0;
 	return -EIO;
 }
@@ -2937,8 +2896,10 @@ void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans,
 	    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 (ret)
+			btrfs_abort_transaction(trans, root, ret);
+		else
+			root->orphan_item_inserted = 0;
 	}
 
 	if (block_rsv) {
@@ -3007,11 +2968,18 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
 	/* insert an orphan item to track this unlinked/truncated file */
 	if (insert >= 1) {
 		ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode));
-		if (ret && ret != -EEXIST) {
-			clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM,
-				  &BTRFS_I(inode)->runtime_flags);
-			btrfs_abort_transaction(trans, root, ret);
-			return ret;
+		if (ret) {
+			if (reserve) {
+				clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED,
+					  &BTRFS_I(inode)->runtime_flags);
+				btrfs_orphan_release_metadata(inode);
+			}
+			if (ret != -EEXIST) {
+				clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM,
+					  &BTRFS_I(inode)->runtime_flags);
+				btrfs_abort_transaction(trans, root, ret);
+				return ret;
+			}
 		}
 		ret = 0;
 	}
@@ -3050,17 +3018,15 @@ static int btrfs_orphan_del(struct btrfs_trans_handle *trans,
 		release_rsv = 1;
 	spin_unlock(&root->orphan_lock);
 
-	if (trans && delete_item) {
+	if (trans && delete_item)
 		ret = btrfs_del_orphan_item(trans, root, btrfs_ino(inode));
-		BUG_ON(ret); /* -ENOMEM or corruption (JDM: Recheck) */
-	}
 
 	if (release_rsv) {
 		btrfs_orphan_release_metadata(inode);
 		atomic_dec(&root->orphan_inodes);
 	}
 
-	return 0;
+	return ret;
 }
 
 /*
@@ -3140,7 +3106,7 @@ int 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);
-		ret = PTR_RET(inode);
+		ret = PTR_ERR_OR_ZERO(inode);
 		if (ret && ret != -ESTALE)
 			goto out;
 
@@ -3190,8 +3156,9 @@ int btrfs_orphan_cleanup(struct btrfs_root *root)
 				found_key.objectid);
 			ret = btrfs_del_orphan_item(trans, root,
 						    found_key.objectid);
-			BUG_ON(ret); /* -ENOMEM or corruption (JDM: Recheck) */
 			btrfs_end_transaction(trans, root);
+			if (ret)
+				goto out;
 			continue;
 		}
 
@@ -3215,13 +3182,16 @@ int btrfs_orphan_cleanup(struct btrfs_root *root)
 			/* 1 for the orphan item deletion. */
 			trans = btrfs_start_transaction(root, 1);
 			if (IS_ERR(trans)) {
+				iput(inode);
 				ret = PTR_ERR(trans);
 				goto out;
 			}
 			ret = btrfs_orphan_add(trans, inode);
 			btrfs_end_transaction(trans, root);
-			if (ret)
+			if (ret) {
+				iput(inode);
 				goto out;
+			}
 
 			ret = btrfs_truncate(inode);
 			if (ret)
@@ -3274,8 +3244,17 @@ static noinline int acls_after_inode_item(struct extent_buffer *leaf,
 {
 	u32 nritems = btrfs_header_nritems(leaf);
 	struct btrfs_key found_key;
+	static u64 xattr_access = 0;
+	static u64 xattr_default = 0;
 	int scanned = 0;
 
+	if (!xattr_access) {
+		xattr_access = btrfs_name_hash(POSIX_ACL_XATTR_ACCESS,
+					strlen(POSIX_ACL_XATTR_ACCESS));
+		xattr_default = btrfs_name_hash(POSIX_ACL_XATTR_DEFAULT,
+					strlen(POSIX_ACL_XATTR_DEFAULT));
+	}
+
 	slot++;
 	while (slot < nritems) {
 		btrfs_item_key_to_cpu(leaf, &found_key, slot);
@@ -3285,8 +3264,11 @@ static noinline int acls_after_inode_item(struct extent_buffer *leaf,
 			return 0;
 
 		/* we found an xattr, assume we've got an acl */
-		if (found_key.type == BTRFS_XATTR_ITEM_KEY)
-			return 1;
+		if (found_key.type == BTRFS_XATTR_ITEM_KEY) {
+			if (found_key.offset == xattr_access ||
+			    found_key.offset == xattr_default)
+				return 1;
+		}
 
 		/*
 		 * we found a key greater than an xattr key, there can't
@@ -3608,8 +3590,7 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 	if (ret) {
 		btrfs_info(root->fs_info,
 			"failed to delete reference to %.*s, inode %llu parent %llu",
-			name_len, name,
-			(unsigned long long)ino, (unsigned long long)dir_ino);
+			name_len, name, ino, dir_ino);
 		btrfs_abort_transaction(trans, root, ret);
 		goto err;
 	}
@@ -3660,53 +3641,20 @@ int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 	}
 	return ret;
 }
-		
-
-/* 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;
-	u64 refs = 1;
-
-	for (level = 0; level < BTRFS_MAX_LEVEL; level++) {
-		int ret;
-
-		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, level, 1,
-					       &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.
+ * unlink and rmdir are special in btrfs, they do not always free space, so
+ * if we cannot make our reservations the normal way try and see if there is
+ * plenty of slack room in the global reserve to migrate, otherwise we cannot
+ * allow the unlink to occur.
  */
-static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
-						       struct dentry *dentry)
+static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir)
 {
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(dir)->root;
-	struct btrfs_path *path;
-	struct btrfs_dir_item *di;
-	struct inode *inode = dentry->d_inode;
-	u64 index;
-	int check_link = 1;
-	int err = -ENOSPC;
 	int ret;
-	u64 ino = btrfs_ino(inode);
-	u64 dir_ino = btrfs_ino(dir);
 
 	/*
 	 * 1 for the possible orphan item
@@ -3719,158 +3667,23 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 	if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC)
 		return trans;
 
-	if (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);
-
-	path = btrfs_alloc_path();
-	if (!path) {
-		root->fs_info->enospc_unlink = 0;
-		return ERR_PTR(-ENOMEM);
-	}
-
-	/* 1 for the orphan item */
-	trans = btrfs_start_transaction(root, 1);
-	if (IS_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(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(path);
+	if (PTR_ERR(trans) == -ENOSPC) {
+		u64 num_bytes = btrfs_calc_trans_metadata_size(root, 5);
 
-	if (ret == 0 && S_ISREG(inode->i_mode)) {
-		ret = btrfs_lookup_file_extent(trans, root, path,
-					       ino, (u64)-1, 0);
-		if (ret < 0) {
-			err = ret;
-			goto out;
+		trans = btrfs_start_transaction(root, 0);
+		if (IS_ERR(trans))
+			return trans;
+		ret = btrfs_cond_migrate_bytes(root->fs_info,
+					       &root->fs_info->trans_block_rsv,
+					       num_bytes, 5);
+		if (ret) {
+			btrfs_end_transaction(trans, root);
+			return ERR_PTR(ret);
 		}
-		BUG_ON(ret == 0); /* Corruption */
-		if (check_path_shared(root, path))
-			goto out;
-		btrfs_release_path(path);
-	}
-
-	if (!check_link) {
-		err = 0;
-		goto out;
-	}
-
-	di = btrfs_lookup_dir_item(trans, root, path, dir_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(path);
-
-	ret = btrfs_get_inode_ref_index(trans, root, path, dentry->d_name.name,
-					dentry->d_name.len, ino, dir_ino, 0,
-					&index);
-	if (ret) {
-		err = ret;
-		goto out;
-	}
-
-	if (check_path_shared(root, path))
-		goto out;
-
-	btrfs_release_path(path);
-
-	/*
-	 * This is a commit root search, if we can lookup inode item and other
-	 * relative items in the commit root, it means the transaction of
-	 * dir/file creation has been committed, and the dir index item that we
-	 * delay to insert has also been inserted into the commit root. So
-	 * we needn't worry about the delayed insertion of the dir index item
-	 * here.
-	 */
-	di = btrfs_lookup_dir_index_item(trans, root, path, dir_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);
-	/* Migrate the orphan reservation over */
-	if (!err)
-		err = btrfs_block_rsv_migrate(trans->block_rsv,
-				&root->fs_info->global_block_rsv,
-				trans->bytes_reserved);
-
-	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->type == BTRFS_BLOCK_RSV_GLOBAL) {
-		btrfs_block_rsv_release(root, trans->block_rsv,
-					trans->bytes_reserved);
 		trans->block_rsv = &root->fs_info->trans_block_rsv;
-		BUG_ON(!root->fs_info->enospc_unlink);
-		root->fs_info->enospc_unlink = 0;
+		trans->bytes_reserved = num_bytes;
 	}
-	btrfs_end_transaction(trans, root);
+	return trans;
 }
 
 static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
@@ -3880,7 +3693,7 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 	struct inode *inode = dentry->d_inode;
 	int ret;
 
-	trans = __unlink_start_trans(dir, dentry);
+	trans = __unlink_start_trans(dir);
 	if (IS_ERR(trans))
 		return PTR_ERR(trans);
 
@@ -3898,7 +3711,7 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 	}
 
 out:
-	__unlink_end_trans(trans, root);
+	btrfs_end_transaction(trans, root);
 	btrfs_btree_balance_dirty(root);
 	return ret;
 }
@@ -3995,7 +3808,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 	if (btrfs_ino(inode) == BTRFS_FIRST_FREE_OBJECTID)
 		return -EPERM;
 
-	trans = __unlink_start_trans(dir, dentry);
+	trans = __unlink_start_trans(dir);
 	if (IS_ERR(trans))
 		return PTR_ERR(trans);
 
@@ -4017,7 +3830,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 	if (!err)
 		btrfs_i_size_write(inode, 0);
 out:
-	__unlink_end_trans(trans, root);
+	btrfs_end_transaction(trans, root);
 	btrfs_btree_balance_dirty(root);
 
 	return err;
@@ -4048,6 +3861,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	u64 extent_num_bytes = 0;
 	u64 extent_offset = 0;
 	u64 item_end = 0;
+	u64 last_size = (u64)-1;
 	u32 found_type = (u8)-1;
 	int found_extent;
 	int del_item;
@@ -4145,6 +3959,11 @@ search_again:
 		if (found_type != BTRFS_EXTENT_DATA_KEY)
 			goto delete;
 
+		if (del_item)
+			last_size = found_key.offset;
+		else
+			last_size = new_size;
+
 		if (extent_type != BTRFS_FILE_EXTENT_INLINE) {
 			u64 num_dec;
 			extent_start = btrfs_file_extent_disk_bytenr(leaf, fi);
@@ -4256,6 +4075,8 @@ out:
 			btrfs_abort_transaction(trans, root, ret);
 	}
 error:
+	if (last_size != (u64)-1)
+		btrfs_ordered_update_i_size(inode, last_size, NULL);
 	btrfs_free_path(path);
 	return err;
 }
@@ -4395,6 +4216,15 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
 	u64 hole_size;
 	int err = 0;
 
+	/*
+	 * If our size started in the middle of a page we need to zero out the
+	 * rest of the page before we expand the i_size, otherwise we could
+	 * expose stale data.
+	 */
+	err = btrfs_truncate_page(inode, oldsize, 0, 0);
+	if (err)
+		return err;
+
 	if (size <= hole_start)
 		return 0;
 
@@ -4509,9 +4339,6 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr)
 	int mask = attr->ia_valid;
 	int ret;
 
-	if (newsize == oldsize)
-		return 0;
-
 	/*
 	 * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
 	 * special case where we need to update the times despite not having
@@ -4522,7 +4349,7 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr)
 		inode->i_ctime = inode->i_mtime = current_fs_time(inode->i_sb);
 
 	if (newsize > oldsize) {
-		truncate_pagecache(inode, oldsize, newsize);
+		truncate_pagecache(inode, newsize);
 		ret = btrfs_cont_expand(inode, oldsize, newsize);
 		if (ret)
 			return ret;
@@ -4578,8 +4405,26 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr)
 		btrfs_inode_resume_unlocked_dio(inode);
 
 		ret = btrfs_truncate(inode);
-		if (ret && inode->i_nlink)
-			btrfs_orphan_del(NULL, inode);
+		if (ret && inode->i_nlink) {
+			int err;
+
+			/*
+			 * failed to truncate, disk_i_size is only adjusted down
+			 * as we remove extents, so it should represent the true
+			 * size of the inode, so reset the in memory size and
+			 * delete our orphan entry.
+			 */
+			trans = btrfs_join_transaction(root);
+			if (IS_ERR(trans)) {
+				btrfs_orphan_del(NULL, inode);
+				return ret;
+			}
+			i_size_write(inode, BTRFS_I(inode)->disk_i_size);
+			err = btrfs_orphan_del(trans, inode);
+			if (err)
+				btrfs_abort_transaction(trans, root, err);
+			btrfs_end_transaction(trans, root);
+		}
 	}
 
 	return ret;
@@ -4714,10 +4559,15 @@ void btrfs_evict_inode(struct inode *inode)
 
 	btrfs_free_block_rsv(root, rsv);
 
+	/*
+	 * Errors here aren't a big deal, it just means we leave orphan items
+	 * in the tree.  They will be cleaned up on the next mount.
+	 */
 	if (ret == 0) {
 		trans->block_rsv = root->orphan_block_rsv;
-		ret = btrfs_orphan_del(trans, inode);
-		BUG_ON(ret);
+		btrfs_orphan_del(trans, inode);
+	} else {
+		btrfs_orphan_del(NULL, inode);
 	}
 
 	trans->block_rsv = &root->fs_info->trans_block_rsv;
@@ -4822,11 +4672,6 @@ static int fixup_tree_root_location(struct btrfs_root *root,
 		goto out;
 	}
 
-	if (btrfs_root_refs(&new_root->root_item) == 0) {
-		err = -ENOENT;
-		goto out;
-	}
-
 	*sub_root = new_root;
 	location->objectid = btrfs_root_dirid(&new_root->root_item);
 	location->type = BTRFS_INODE_ITEM_KEY;
@@ -4843,11 +4688,11 @@ static void inode_tree_add(struct inode *inode)
 	struct btrfs_inode *entry;
 	struct rb_node **p;
 	struct rb_node *parent;
+	struct rb_node *new = &BTRFS_I(inode)->rb_node;
 	u64 ino = btrfs_ino(inode);
 
 	if (inode_unhashed(inode))
 		return;
-again:
 	parent = NULL;
 	spin_lock(&root->inode_lock);
 	p = &root->inode_tree.rb_node;
@@ -4862,14 +4707,14 @@ again:
 		else {
 			WARN_ON(!(entry->vfs_inode.i_state &
 				  (I_WILL_FREE | I_FREEING)));
-			rb_erase(parent, &root->inode_tree);
+			rb_replace_node(parent, new, &root->inode_tree);
 			RB_CLEAR_NODE(parent);
 			spin_unlock(&root->inode_lock);
-			goto again;
+			return;
 		}
 	}
-	rb_link_node(&BTRFS_I(inode)->rb_node, parent, p);
-	rb_insert_color(&BTRFS_I(inode)->rb_node, &root->inode_tree);
+	rb_link_node(new, parent, p);
+	rb_insert_color(new, &root->inode_tree);
 	spin_unlock(&root->inode_lock);
 }
 
@@ -5092,8 +4937,10 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
 		if (!(inode->i_sb->s_flags & MS_RDONLY))
 			ret = btrfs_orphan_cleanup(sub_root);
 		up_read(&root->fs_info->cleanup_work_sem);
-		if (ret)
+		if (ret) {
+			iput(inode);
 			inode = ERR_PTR(ret);
+		}
 	}
 
 	return inode;
@@ -5137,10 +4984,9 @@ unsigned char btrfs_filetype_table[] = {
 	DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
 };
 
-static int btrfs_real_readdir(struct file *filp, void *dirent,
-			      filldir_t filldir)
+static int btrfs_real_readdir(struct file *file, struct dir_context *ctx)
 {
-	struct inode *inode = file_inode(filp);
+	struct inode *inode = file_inode(file);
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_item *item;
 	struct btrfs_dir_item *di;
@@ -5161,29 +5007,15 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 	char tmp_name[32];
 	char *name_ptr;
 	int name_len;
-	int is_curr = 0;	/* filp->f_pos points to the current index? */
+	int is_curr = 0;	/* ctx->pos points to the current index? */
 
 	/* FIXME, use a real flag for deciding about the key type */
 	if (root->fs_info->tree_root == root)
 		key_type = BTRFS_DIR_ITEM_KEY;
 
-	/* special case for "." */
-	if (filp->f_pos == 0) {
-		over = filldir(dirent, ".", 1,
-			       filp->f_pos, btrfs_ino(inode), DT_DIR);
-		if (over)
-			return 0;
-		filp->f_pos = 1;
-	}
-	/* special case for .., just use the back ref */
-	if (filp->f_pos == 1) {
-		u64 pino = parent_ino(filp->f_path.dentry);
-		over = filldir(dirent, "..", 2,
-			       filp->f_pos, pino, DT_DIR);
-		if (over)
-			return 0;
-		filp->f_pos = 2;
-	}
+	if (!dir_emit_dots(file, ctx))
+		return 0;
+
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -5197,7 +5029,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 	}
 
 	btrfs_set_key_type(&key, key_type);
-	key.offset = filp->f_pos;
+	key.offset = ctx->pos;
 	key.objectid = btrfs_ino(inode);
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
@@ -5223,14 +5055,14 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 			break;
 		if (btrfs_key_type(&found_key) != key_type)
 			break;
-		if (found_key.offset < filp->f_pos)
+		if (found_key.offset < ctx->pos)
 			goto next;
 		if (key_type == BTRFS_DIR_INDEX_KEY &&
 		    btrfs_should_delete_dir_index(&del_list,
 						  found_key.offset))
 			goto next;
 
-		filp->f_pos = found_key.offset;
+		ctx->pos = found_key.offset;
 		is_curr = 1;
 
 		di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item);
@@ -5274,9 +5106,8 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 				over = 0;
 				goto skip;
 			}
-			over = filldir(dirent, name_ptr, name_len,
-				       found_key.offset, location.objectid,
-				       d_type);
+			over = !dir_emit(ctx, name_ptr, name_len,
+				       location.objectid, d_type);
 
 skip:
 			if (name_ptr != tmp_name)
@@ -5295,22 +5126,38 @@ next:
 
 	if (key_type == BTRFS_DIR_INDEX_KEY) {
 		if (is_curr)
-			filp->f_pos++;
-		ret = btrfs_readdir_delayed_dir_index(filp, dirent, filldir,
-						      &ins_list);
+			ctx->pos++;
+		ret = btrfs_readdir_delayed_dir_index(ctx, &ins_list);
 		if (ret)
 			goto nopos;
 	}
 
 	/* Reached end of directory/root. Bump pos past the last item. */
-	if (key_type == BTRFS_DIR_INDEX_KEY)
-		/*
-		 * 32-bit glibc will use getdents64, but then strtol -
-		 * so the last number we can serve is this.
-		 */
-		filp->f_pos = 0x7fffffff;
-	else
-		filp->f_pos++;
+	ctx->pos++;
+
+	/*
+	 * Stop new entries from being returned after we return the last
+	 * entry.
+	 *
+	 * New directory entries are assigned a strictly increasing
+	 * offset.  This means that new entries created during readdir
+	 * are *guaranteed* to be seen in the future by that readdir.
+	 * This has broken buggy programs which operate on names as
+	 * they're returned by readdir.  Until we re-use freed offsets
+	 * we have this hack to stop new entries from being returned
+	 * under the assumption that they'll never reach this huge
+	 * offset.
+	 *
+	 * This is being careful not to overflow 32bit loff_t unless the
+	 * last entry requires it because doing so has broken 32bit apps
+	 * in the past.
+	 */
+	if (key_type == BTRFS_DIR_INDEX_KEY) {
+		if (ctx->pos >= INT_MAX)
+			ctx->pos = LLONG_MAX;
+		else
+			ctx->pos = INT_MAX;
+	}
 nopos:
 	ret = 0;
 err:
@@ -6277,10 +6124,7 @@ insert:
 	btrfs_release_path(path);
 	if (em->start > start || extent_map_end(em) <= start) {
 		btrfs_err(root->fs_info, "bad extent! em: [%llu %llu] passed [%llu %llu]",
-			(unsigned long long)em->start,
-			(unsigned long long)em->len,
-			(unsigned long long)start,
-			(unsigned long long)len);
+			em->start, em->len, start, len);
 		err = -EIO;
 		goto out;
 	}
@@ -6478,39 +6322,32 @@ 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 btrfs_key ins;
 	u64 alloc_hint;
 	int ret;
 
-	trans = btrfs_join_transaction(root);
-	if (IS_ERR(trans))
-		return ERR_CAST(trans);
-
-	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,
+	ret = btrfs_reserve_extent(root, len, root->sectorsize, 0,
 				   alloc_hint, &ins, 1);
-	if (ret) {
-		em = ERR_PTR(ret);
-		goto out;
-	}
+	if (ret)
+		return ERR_PTR(ret);
 
 	em = create_pinned_em(inode, start, ins.offset, start, ins.objectid,
 			      ins.offset, ins.offset, ins.offset, 0);
-	if (IS_ERR(em))
-		goto out;
+	if (IS_ERR(em)) {
+		btrfs_free_reserved_extent(root, ins.objectid, ins.offset);
+		return em;
+	}
 
 	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);
+		free_extent_map(em);
+		return ERR_PTR(ret);
 	}
-out:
-	btrfs_end_transaction(trans, root);
+
 	return em;
 }
 
@@ -6518,11 +6355,11 @@ out:
  * 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,
-				      u64 *orig_start, u64 *orig_block_len,
-				      u64 *ram_bytes)
+noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
+			      u64 *orig_start, u64 *orig_block_len,
+			      u64 *ram_bytes)
 {
+	struct btrfs_trans_handle *trans;
 	struct btrfs_path *path;
 	int ret;
 	struct extent_buffer *leaf;
@@ -6535,12 +6372,12 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans,
 	u64 num_bytes;
 	int slot;
 	int found_type;
-
+	bool nocow = (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW);
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
-	ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode),
+	ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(inode),
 				       offset, 0);
 	if (ret < 0)
 		goto out;
@@ -6575,18 +6412,28 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans,
 		/* not a regular extent, must cow */
 		goto out;
 	}
+
+	if (!nocow && found_type == BTRFS_FILE_EXTENT_REG)
+		goto out;
+
 	disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+	if (disk_bytenr == 0)
+		goto out;
+
+	if (btrfs_file_extent_compression(leaf, fi) ||
+	    btrfs_file_extent_encryption(leaf, fi) ||
+	    btrfs_file_extent_other_encoding(leaf, fi))
+		goto out;
+
 	backref_offset = btrfs_file_extent_offset(leaf, fi);
 
-	*orig_start = key.offset - backref_offset;
-	*orig_block_len = btrfs_file_extent_disk_num_bytes(leaf, fi);
-	*ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
+	if (orig_start) {
+		*orig_start = key.offset - backref_offset;
+		*orig_block_len = btrfs_file_extent_disk_num_bytes(leaf, fi);
+		*ram_bytes = btrfs_file_extent_ram_bytes(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;
@@ -6595,9 +6442,19 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans,
 	 * look for other files referencing this extent, if we
 	 * find any we must cow
 	 */
-	if (btrfs_cross_ref_exist(trans, root, btrfs_ino(inode),
-				  key.offset - backref_offset, disk_bytenr))
+	trans = btrfs_join_transaction(root);
+	if (IS_ERR(trans)) {
+		ret = 0;
 		goto out;
+	}
+
+	ret = btrfs_cross_ref_exist(trans, root, btrfs_ino(inode),
+				    key.offset - backref_offset, disk_bytenr);
+	btrfs_end_transaction(trans, root);
+	if (ret) {
+		ret = 0;
+		goto out;
+	}
 
 	/*
 	 * adjust disk_bytenr and num_bytes to cover just the bytes
@@ -6739,7 +6596,6 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 	u64 start = iblock << inode->i_blkbits;
 	u64 lockstart, lockend;
 	u64 len = bh_result->b_size;
-	struct btrfs_trans_handle *trans;
 	int unlock_bits = EXTENT_LOCKED;
 	int ret = 0;
 
@@ -6821,17 +6677,8 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 		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);
-		if (IS_ERR(trans))
-			goto must_cow;
-
-		if (can_nocow_odirect(trans, inode, start, &len, &orig_start,
-				      &orig_block_len, &ram_bytes) == 1) {
+		if (can_nocow_extent(inode, start, &len, &orig_start,
+				     &orig_block_len, &ram_bytes) == 1) {
 			if (type == BTRFS_ORDERED_PREALLOC) {
 				free_extent_map(em);
 				em = create_pinned_em(inode, start, len,
@@ -6839,24 +6686,20 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 						       block_start, len,
 						       orig_block_len,
 						       ram_bytes, type);
-				if (IS_ERR(em)) {
-					btrfs_end_transaction(trans, root);
+				if (IS_ERR(em))
 					goto unlock_err;
-				}
 			}
 
 			ret = btrfs_add_ordered_extent_dio(inode, start,
 					   block_start, len, len, type);
-			btrfs_end_transaction(trans, root);
 			if (ret) {
 				free_extent_map(em);
 				goto unlock_err;
 			}
 			goto unlock;
 		}
-		btrfs_end_transaction(trans, root);
 	}
-must_cow:
+
 	/*
 	 * this will cow the extent, reset the len in case we changed
 	 * it above
@@ -6919,26 +6762,6 @@ unlock_err:
 	return ret;
 }
 
-struct btrfs_dio_private {
-	struct inode *inode;
-	u64 logical_offset;
-	u64 disk_bytenr;
-	u64 bytes;
-	void *private;
-
-	/* number of bios pending for this dio */
-	atomic_t pending_bios;
-
-	/* IO errors */
-	int errors;
-
-	/* orig_bio is our btrfs_io_bio */
-	struct bio *orig_bio;
-
-	/* dio_bio came from fs/direct-io.c */
-	struct bio *dio_bio;
-};
-
 static void btrfs_endio_direct_read(struct bio *bio, int err)
 {
 	struct btrfs_dio_private *dip = bio->bi_private;
@@ -6947,6 +6770,8 @@ static void btrfs_endio_direct_read(struct bio *bio, int err)
 	struct inode *inode = dip->inode;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct bio *dio_bio;
+	u32 *csums = (u32 *)dip->csum;
+	int index = 0;
 	u64 start;
 
 	start = dip->logical_offset;
@@ -6955,12 +6780,8 @@ static void btrfs_endio_direct_read(struct bio *bio, int err)
 			struct page *page = bvec->bv_page;
 			char *kaddr;
 			u32 csum = ~(u32)0;
-			u64 private = ~(u32)0;
 			unsigned long flags;
 
-			if (get_state_private(&BTRFS_I(inode)->io_tree,
-					      start, &private))
-				goto failed;
 			local_irq_save(flags);
 			kaddr = kmap_atomic(page);
 			csum = btrfs_csum_data(kaddr + bvec->bv_offset,
@@ -6970,18 +6791,17 @@ static void btrfs_endio_direct_read(struct bio *bio, int err)
 			local_irq_restore(flags);
 
 			flush_dcache_page(bvec->bv_page);
-			if (csum != private) {
-failed:
-				btrfs_err(root->fs_info, "csum failed ino %llu off %llu csum %u private %u",
-					(unsigned long long)btrfs_ino(inode),
-					(unsigned long long)start,
-					csum, (unsigned)private);
+			if (csum != csums[index]) {
+				btrfs_err(root->fs_info, "csum failed ino %llu off %llu csum %u expected csum %u",
+					  btrfs_ino(inode), start, csum,
+					  csums[index]);
 				err = -EIO;
 			}
 		}
 
 		start += bvec->bv_len;
 		bvec++;
+		index++;
 	} while (bvec <= bvec_end);
 
 	unlock_extent(&BTRFS_I(inode)->io_tree, dip->logical_offset,
@@ -7062,7 +6882,7 @@ static void btrfs_end_dio_bio(struct bio *bio, int err)
 	if (err) {
 		printk(KERN_ERR "btrfs direct IO failed ino %llu rw %lu "
 		      "sector %#Lx len %u err no %d\n",
-		      (unsigned long long)btrfs_ino(dip->inode), bio->bi_rw,
+		      btrfs_ino(dip->inode), bio->bi_rw,
 		      (unsigned long long)bio->bi_sector, bio->bi_size, err);
 		dip->errors = 1;
 
@@ -7098,6 +6918,7 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode,
 					 int rw, u64 file_offset, int skip_sum,
 					 int async_submit)
 {
+	struct btrfs_dio_private *dip = bio->bi_private;
 	int write = rw & REQ_WRITE;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret;
@@ -7132,7 +6953,8 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode,
 		if (ret)
 			goto err;
 	} else if (!skip_sum) {
-		ret = btrfs_lookup_bio_sums_dio(root, inode, bio, file_offset);
+		ret = btrfs_lookup_bio_sums_dio(root, inode, dip, bio,
+						file_offset);
 		if (ret)
 			goto err;
 	}
@@ -7167,6 +6989,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip,
 		bio_put(orig_bio);
 		return -EIO;
 	}
+
 	if (map_length >= orig_bio->bi_size) {
 		bio = orig_bio;
 		goto submit;
@@ -7260,38 +7083,39 @@ static void btrfs_submit_direct(int rw, struct bio *dio_bio,
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_dio_private *dip;
-	struct bio_vec *bvec = dio_bio->bi_io_vec;
 	struct bio *io_bio;
 	int skip_sum;
+	int sum_len;
 	int write = rw & REQ_WRITE;
 	int ret = 0;
+	u16 csum_size;
 
 	skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
 
 	io_bio = btrfs_bio_clone(dio_bio, GFP_NOFS);
-
 	if (!io_bio) {
 		ret = -ENOMEM;
 		goto free_ordered;
 	}
 
-	dip = kmalloc(sizeof(*dip), GFP_NOFS);
+	if (!skip_sum && !write) {
+		csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
+		sum_len = dio_bio->bi_size >> inode->i_sb->s_blocksize_bits;
+		sum_len *= csum_size;
+	} else {
+		sum_len = 0;
+	}
+
+	dip = kmalloc(sizeof(*dip) + sum_len, GFP_NOFS);
 	if (!dip) {
 		ret = -ENOMEM;
 		goto free_io_bio;
 	}
 
 	dip->private = dio_bio->bi_private;
-	io_bio->bi_private = dio_bio->bi_private;
 	dip->inode = inode;
 	dip->logical_offset = file_offset;
-
-	dip->bytes = 0;
-	do {
-		dip->bytes += bvec->bv_len;
-		bvec++;
-	} while (bvec <= (dio_bio->bi_io_vec + dio_bio->bi_vcnt - 1));
-
+	dip->bytes = dio_bio->bi_size;
 	dip->disk_bytenr = (u64)dio_bio->bi_sector << 9;
 	io_bio->bi_private = dip;
 	dip->errors = 0;
@@ -7390,8 +7214,16 @@ static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
 	atomic_inc(&inode->i_dio_count);
 	smp_mb__after_atomic_inc();
 
+	/*
+	 * The generic stuff only does filemap_write_and_wait_range, which isn't
+	 * enough if we've written compressed pages to this area, so we need to
+	 * call btrfs_wait_ordered_range to make absolutely sure that any
+	 * outstanding dirty pages are on disk.
+	 */
+	count = iov_length(iov, nr_segs);
+	btrfs_wait_ordered_range(inode, offset, count);
+
 	if (rw & WRITE) {
-		count = iov_length(iov, nr_segs);
 		/*
 		 * If the write DIO is beyond the EOF, we need update
 		 * the isize, but it is protected by i_mutex. So we can
@@ -7510,7 +7342,8 @@ static int btrfs_releasepage(struct page *page, gfp_t gfp_flags)
 	return __btrfs_releasepage(page, gfp_flags & GFP_NOFS);
 }
 
-static void btrfs_invalidatepage(struct page *page, unsigned long offset)
+static void btrfs_invalidatepage(struct page *page, unsigned int offset,
+				 unsigned int length)
 {
 	struct inode *inode = page->mapping->host;
 	struct extent_io_tree *tree;
@@ -7548,10 +7381,23 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
 		 * whoever cleared the private bit is responsible
 		 * for the finish_ordered_io
 		 */
-		if (TestClearPagePrivate2(page) &&
-		    btrfs_dec_test_ordered_pending(inode, &ordered, page_start,
-						   PAGE_CACHE_SIZE, 1)) {
-			btrfs_finish_ordered_io(ordered);
+		if (TestClearPagePrivate2(page)) {
+			struct btrfs_ordered_inode_tree *tree;
+			u64 new_len;
+
+			tree = &BTRFS_I(inode)->ordered_tree;
+
+			spin_lock_irq(&tree->lock);
+			set_bit(BTRFS_ORDERED_TRUNCATED, &ordered->flags);
+			new_len = page_start - ordered->file_offset;
+			if (new_len < ordered->truncated_len)
+				ordered->truncated_len = new_len;
+			spin_unlock_irq(&tree->lock);
+
+			if (btrfs_dec_test_ordered_pending(inode, &ordered,
+							   page_start,
+							   PAGE_CACHE_SIZE, 1))
+				btrfs_finish_ordered_io(ordered);
 		}
 		btrfs_put_ordered_extent(ordered);
 		cached_state = NULL;
@@ -7710,18 +7556,13 @@ static int btrfs_truncate(struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_block_rsv *rsv;
-	int ret;
+	int ret = 0;
 	int err = 0;
 	struct btrfs_trans_handle *trans;
 	u64 mask = root->sectorsize - 1;
 	u64 min_size = btrfs_calc_trunc_metadata_size(root, 1);
 
-	ret = btrfs_truncate_page(inode, inode->i_size, 0, 0);
-	if (ret)
-		return ret;
-
 	btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
-	btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
 
 	/*
 	 * Yes ladies and gentelment, this is indeed ugly.  The fact is we have
@@ -7977,15 +7818,15 @@ void btrfs_destroy_inode(struct inode *inode)
 	 */
 	smp_mb();
 	if (!list_empty(&BTRFS_I(inode)->ordered_operations)) {
-		spin_lock(&root->fs_info->ordered_extent_lock);
+		spin_lock(&root->fs_info->ordered_root_lock);
 		list_del_init(&BTRFS_I(inode)->ordered_operations);
-		spin_unlock(&root->fs_info->ordered_extent_lock);
+		spin_unlock(&root->fs_info->ordered_root_lock);
 	}
 
 	if (test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM,
 		     &BTRFS_I(inode)->runtime_flags)) {
 		btrfs_info(root->fs_info, "inode %llu still on the orphan list",
-			(unsigned long long)btrfs_ino(inode));
+			btrfs_ino(inode));
 		atomic_dec(&root->orphan_inodes);
 	}
 
@@ -7995,8 +7836,7 @@ void btrfs_destroy_inode(struct inode *inode)
 			break;
 		else {
 			btrfs_err(root->fs_info, "found ordered extent %llu %llu on inode cleanup",
-				(unsigned long long)ordered->file_offset,
-				(unsigned long long)ordered->len);
+				ordered->file_offset, ordered->len);
 			btrfs_remove_ordered_extent(inode, ordered);
 			btrfs_put_ordered_extent(ordered);
 			btrfs_put_ordered_extent(ordered);
@@ -8012,6 +7852,9 @@ int btrfs_drop_inode(struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 
+	if (root == NULL)
+		return 1;
+
 	/* the snap/subvol tree is on deleting */
 	if (btrfs_root_refs(&root->root_item) == 0 &&
 	    root != root->fs_info->tree_root)
@@ -8267,10 +8110,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 						 new_dentry->d_name.name,
 						 new_dentry->d_name.len);
 		}
-		if (!ret && new_inode->i_nlink == 0) {
+		if (!ret && new_inode->i_nlink == 0)
 			ret = btrfs_orphan_add(trans, new_dentry->d_inode);
-			BUG_ON(ret);
-		}
 		if (ret) {
 			btrfs_abort_transaction(trans, root, ret);
 			goto out_fail;
@@ -8346,7 +8187,7 @@ void btrfs_wait_and_free_delalloc_work(struct btrfs_delalloc_work *work)
  * some fairly slow code that needs optimization. This walks the list
  * of all the inodes with pending delalloc and forces them to disk.
  */
-int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
+static int __start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
 {
 	struct btrfs_inode *binode;
 	struct inode *inode;
@@ -8355,33 +8196,30 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
 	struct list_head splice;
 	int ret = 0;
 
-	if (root->fs_info->sb->s_flags & MS_RDONLY)
-		return -EROFS;
-
 	INIT_LIST_HEAD(&works);
 	INIT_LIST_HEAD(&splice);
 
-	spin_lock(&root->fs_info->delalloc_lock);
-	list_splice_init(&root->fs_info->delalloc_inodes, &splice);
+	spin_lock(&root->delalloc_lock);
+	list_splice_init(&root->delalloc_inodes, &splice);
 	while (!list_empty(&splice)) {
 		binode = list_entry(splice.next, struct btrfs_inode,
 				    delalloc_inodes);
 
-		list_del_init(&binode->delalloc_inodes);
-
+		list_move_tail(&binode->delalloc_inodes,
+			       &root->delalloc_inodes);
 		inode = igrab(&binode->vfs_inode);
 		if (!inode) {
-			clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
-				  &binode->runtime_flags);
+			cond_resched_lock(&root->delalloc_lock);
 			continue;
 		}
-
-		list_add_tail(&binode->delalloc_inodes,
-			      &root->fs_info->delalloc_inodes);
-		spin_unlock(&root->fs_info->delalloc_lock);
+		spin_unlock(&root->delalloc_lock);
 
 		work = btrfs_alloc_delalloc_work(inode, 0, delay_iput);
 		if (unlikely(!work)) {
+			if (delay_iput)
+				btrfs_add_delayed_iput(inode);
+			else
+				iput(inode);
 			ret = -ENOMEM;
 			goto out;
 		}
@@ -8390,16 +8228,39 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
 				   &work->work);
 
 		cond_resched();
-		spin_lock(&root->fs_info->delalloc_lock);
+		spin_lock(&root->delalloc_lock);
 	}
-	spin_unlock(&root->fs_info->delalloc_lock);
+	spin_unlock(&root->delalloc_lock);
 
 	list_for_each_entry_safe(work, next, &works, list) {
 		list_del_init(&work->list);
 		btrfs_wait_and_free_delalloc_work(work);
 	}
+	return 0;
+out:
+	list_for_each_entry_safe(work, next, &works, list) {
+		list_del_init(&work->list);
+		btrfs_wait_and_free_delalloc_work(work);
+	}
+
+	if (!list_empty_careful(&splice)) {
+		spin_lock(&root->delalloc_lock);
+		list_splice_tail(&splice, &root->delalloc_inodes);
+		spin_unlock(&root->delalloc_lock);
+	}
+	return ret;
+}
+
+int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
+{
+	int ret;
+
+	if (root->fs_info->sb->s_flags & MS_RDONLY)
+		return -EROFS;
 
-	/* the filemap_flush will queue IO into the worker threads, but
+	ret = __start_delalloc_inodes(root, delay_iput);
+	/*
+	 * the filemap_flush will queue IO into the worker threads, but
 	 * we have to make sure the IO is actually started and that
 	 * ordered extents get created before we return
 	 */
@@ -8411,17 +8272,55 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
 		    atomic_read(&root->fs_info->async_delalloc_pages) == 0));
 	}
 	atomic_dec(&root->fs_info->async_submit_draining);
-	return 0;
-out:
-	list_for_each_entry_safe(work, next, &works, list) {
-		list_del_init(&work->list);
-		btrfs_wait_and_free_delalloc_work(work);
+	return ret;
+}
+
+int btrfs_start_all_delalloc_inodes(struct btrfs_fs_info *fs_info,
+				    int delay_iput)
+{
+	struct btrfs_root *root;
+	struct list_head splice;
+	int ret;
+
+	if (fs_info->sb->s_flags & MS_RDONLY)
+		return -EROFS;
+
+	INIT_LIST_HEAD(&splice);
+
+	spin_lock(&fs_info->delalloc_root_lock);
+	list_splice_init(&fs_info->delalloc_roots, &splice);
+	while (!list_empty(&splice)) {
+		root = list_first_entry(&splice, struct btrfs_root,
+					delalloc_root);
+		root = btrfs_grab_fs_root(root);
+		BUG_ON(!root);
+		list_move_tail(&root->delalloc_root,
+			       &fs_info->delalloc_roots);
+		spin_unlock(&fs_info->delalloc_root_lock);
+
+		ret = __start_delalloc_inodes(root, delay_iput);
+		btrfs_put_fs_root(root);
+		if (ret)
+			goto out;
+
+		spin_lock(&fs_info->delalloc_root_lock);
 	}
+	spin_unlock(&fs_info->delalloc_root_lock);
 
+	atomic_inc(&fs_info->async_submit_draining);
+	while (atomic_read(&fs_info->nr_async_submits) ||
+	      atomic_read(&fs_info->async_delalloc_pages)) {
+		wait_event(fs_info->async_submit_wait,
+		   (atomic_read(&fs_info->nr_async_submits) == 0 &&
+		    atomic_read(&fs_info->async_delalloc_pages) == 0));
+	}
+	atomic_dec(&fs_info->async_submit_draining);
+	return 0;
+out:
 	if (!list_empty_careful(&splice)) {
-		spin_lock(&root->fs_info->delalloc_lock);
-		list_splice_tail(&splice, &root->fs_info->delalloc_inodes);
-		spin_unlock(&root->fs_info->delalloc_lock);
+		spin_lock(&fs_info->delalloc_root_lock);
+		list_splice_tail(&splice, &fs_info->delalloc_roots);
+		spin_unlock(&fs_info->delalloc_root_lock);
 	}
 	return ret;
 }
@@ -8577,8 +8476,8 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
 
 		cur_bytes = min(num_bytes, 256ULL * 1024 * 1024);
 		cur_bytes = max(cur_bytes, min_size);
-		ret = btrfs_reserve_extent(trans, root, cur_bytes,
-					   min_size, 0, *alloc_hint, &ins, 1);
+		ret = btrfs_reserve_extent(root, cur_bytes, min_size, 0,
+					   *alloc_hint, &ins, 1);
 		if (ret) {
 			if (own_trans)
 				btrfs_end_transaction(trans, root);
@@ -8718,17 +8617,19 @@ static const struct inode_operations btrfs_dir_inode_operations = {
 	.removexattr	= btrfs_removexattr,
 	.permission	= btrfs_permission,
 	.get_acl	= btrfs_get_acl,
+	.update_time	= btrfs_update_time,
 };
 static const struct inode_operations btrfs_dir_ro_inode_operations = {
 	.lookup		= btrfs_lookup,
 	.permission	= btrfs_permission,
 	.get_acl	= btrfs_get_acl,
+	.update_time	= btrfs_update_time,
 };
 
 static const struct file_operations btrfs_dir_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
-	.readdir	= btrfs_real_readdir,
+	.iterate	= btrfs_real_readdir,
 	.unlocked_ioctl	= btrfs_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= btrfs_ioctl,
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 0f81d67cdc8d..9d46f60cb943 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -43,6 +43,7 @@
 #include <linux/blkdev.h>
 #include <linux/uuid.h>
 #include <linux/btrfs.h>
+#include <linux/uaccess.h>
 #include "compat.h"
 #include "ctree.h"
 #include "disk-io.h"
@@ -57,6 +58,9 @@
 #include "send.h"
 #include "dev-replace.h"
 
+static int btrfs_clone(struct inode *src, struct inode *inode,
+		       u64 off, u64 olen, u64 olen_aligned, u64 destoff);
+
 /* Mask out flags that are inappropriate for the given type of inode. */
 static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags)
 {
@@ -363,6 +367,13 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg)
 	return 0;
 }
 
+int btrfs_is_empty_uuid(u8 *uuid)
+{
+	static char empty_uuid[BTRFS_UUID_SIZE] = {0};
+
+	return !memcmp(uuid, empty_uuid, BTRFS_UUID_SIZE);
+}
+
 static noinline int create_subvol(struct inode *dir,
 				  struct dentry *dentry,
 				  char *name, int namelen,
@@ -396,7 +407,7 @@ static noinline int create_subvol(struct inode *dir,
 	 * of create_snapshot().
 	 */
 	ret = btrfs_subvolume_reserve_metadata(root, &block_rsv,
-					       7, &qgroup_reserved);
+					       8, &qgroup_reserved, false);
 	if (ret)
 		return ret;
 
@@ -425,26 +436,25 @@ static noinline int create_subvol(struct inode *dir,
 	btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
 	btrfs_set_header_owner(leaf, objectid);
 
-	write_extent_buffer(leaf, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(leaf),
+	write_extent_buffer(leaf, root->fs_info->fsid, btrfs_header_fsid(leaf),
 			    BTRFS_FSID_SIZE);
 	write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
-			    (unsigned long)btrfs_header_chunk_tree_uuid(leaf),
+			    btrfs_header_chunk_tree_uuid(leaf),
 			    BTRFS_UUID_SIZE);
 	btrfs_mark_buffer_dirty(leaf);
 
 	memset(&root_item, 0, sizeof(root_item));
 
 	inode_item = &root_item.inode;
-	inode_item->generation = cpu_to_le64(1);
-	inode_item->size = cpu_to_le64(3);
-	inode_item->nlink = cpu_to_le32(1);
-	inode_item->nbytes = cpu_to_le64(root->leafsize);
-	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
+	btrfs_set_stack_inode_generation(inode_item, 1);
+	btrfs_set_stack_inode_size(inode_item, 3);
+	btrfs_set_stack_inode_nlink(inode_item, 1);
+	btrfs_set_stack_inode_nbytes(inode_item, root->leafsize);
+	btrfs_set_stack_inode_mode(inode_item, S_IFDIR | 0755);
 
-	root_item.flags = 0;
-	root_item.byte_limit = 0;
-	inode_item->flags = cpu_to_le64(BTRFS_INODE_ROOT_ITEM_INIT);
+	btrfs_set_root_flags(&root_item, 0);
+	btrfs_set_root_limit(&root_item, 0);
+	btrfs_set_stack_inode_flags(inode_item, BTRFS_INODE_ROOT_ITEM_INIT);
 
 	btrfs_set_root_bytenr(&root_item, leaf->start);
 	btrfs_set_root_generation(&root_item, trans->transid);
@@ -457,8 +467,8 @@ static noinline int create_subvol(struct inode *dir,
 			btrfs_root_generation(&root_item));
 	uuid_le_gen(&new_uuid);
 	memcpy(root_item.uuid, new_uuid.b, BTRFS_UUID_SIZE);
-	root_item.otime.sec = cpu_to_le64(cur_time.tv_sec);
-	root_item.otime.nsec = cpu_to_le32(cur_time.tv_nsec);
+	btrfs_set_stack_timespec_sec(&root_item.otime, cur_time.tv_sec);
+	btrfs_set_stack_timespec_nsec(&root_item.otime, cur_time.tv_nsec);
 	root_item.ctime = root_item.otime;
 	btrfs_set_root_ctransid(&root_item, trans->transid);
 	btrfs_set_root_otransid(&root_item, trans->transid);
@@ -518,9 +528,14 @@ static noinline int create_subvol(struct inode *dir,
 	ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
 				 objectid, root->root_key.objectid,
 				 btrfs_ino(dir), index, name, namelen);
-
 	BUG_ON(ret);
 
+	ret = btrfs_uuid_tree_add(trans, root->fs_info->uuid_root,
+				  root_item.uuid, BTRFS_UUID_KEY_SUBVOL,
+				  objectid);
+	if (ret)
+		btrfs_abort_transaction(trans, root, ret);
+
 fail:
 	trans->block_rsv = NULL;
 	trans->bytes_reserved = 0;
@@ -555,6 +570,12 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 	if (!root->ref_cows)
 		return -EINVAL;
 
+	ret = btrfs_start_delalloc_inodes(root, 0);
+	if (ret)
+		return ret;
+
+	btrfs_wait_ordered_extents(root);
+
 	pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_NOFS);
 	if (!pending_snapshot)
 		return -ENOMEM;
@@ -567,10 +588,12 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 	 * 1 - root item
 	 * 2 - root ref/backref
 	 * 1 - root of snapshot
+	 * 1 - UUID item
 	 */
 	ret = btrfs_subvolume_reserve_metadata(BTRFS_I(dir)->root,
-					&pending_snapshot->block_rsv, 7,
-					&pending_snapshot->qgroup_reserved);
+					&pending_snapshot->block_rsv, 8,
+					&pending_snapshot->qgroup_reserved,
+					false);
 	if (ret)
 		goto out;
 
@@ -1261,9 +1284,6 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 			cluster = max_cluster;
 		}
 
-		if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)
-			BTRFS_I(inode)->force_compress = compress_type;
-
 		if (i + cluster > ra_index) {
 			ra_index = max(i, ra_index);
 			btrfs_force_ra(inode->i_mapping, ra, file, ra_index,
@@ -1272,6 +1292,8 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 		}
 
 		mutex_lock(&inode->i_mutex);
+		if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)
+			BTRFS_I(inode)->force_compress = compress_type;
 		ret = cluster_pages_for_defrag(inode, pages, i, cluster);
 		if (ret < 0) {
 			mutex_unlock(&inode->i_mutex);
@@ -1328,10 +1350,6 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 			    atomic_read(&root->fs_info->async_delalloc_pages) == 0));
 		}
 		atomic_dec(&root->fs_info->async_submit_draining);
-
-		mutex_lock(&inode->i_mutex);
-		BTRFS_I(inode)->force_compress = BTRFS_COMPRESS_NONE;
-		mutex_unlock(&inode->i_mutex);
 	}
 
 	if (range->compress_type == BTRFS_COMPRESS_LZO) {
@@ -1341,6 +1359,11 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 	ret = defrag_count;
 
 out_ra:
+	if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) {
+		mutex_lock(&inode->i_mutex);
+		BTRFS_I(inode)->force_compress = BTRFS_COMPRESS_NONE;
+		mutex_unlock(&inode->i_mutex);
+	}
 	if (!file)
 		kfree(ra);
 	kfree(pages);
@@ -1371,9 +1394,8 @@ static noinline int btrfs_ioctl_resize(struct file *file,
 
 	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
 			1)) {
-		pr_info("btrfs: dev add/delete/balance/replace/resize operation in progress\n");
 		mnt_drop_write_file(file);
-		return -EINVAL;
+		return BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
 	}
 
 	mutex_lock(&root->fs_info->volume_mutex);
@@ -1397,14 +1419,13 @@ static noinline int btrfs_ioctl_resize(struct file *file,
 			ret = -EINVAL;
 			goto out_free;
 		}
-		printk(KERN_INFO "btrfs: resizing devid %llu\n",
-		       (unsigned long long)devid);
+		printk(KERN_INFO "btrfs: resizing devid %llu\n", devid);
 	}
 
 	device = btrfs_find_device(root->fs_info, devid, NULL, NULL);
 	if (!device) {
 		printk(KERN_INFO "btrfs: resizer unable to find device %llu\n",
-		       (unsigned long long)devid);
+		       devid);
 		ret = -ENODEV;
 		goto out_free;
 	}
@@ -1412,7 +1433,7 @@ static noinline int btrfs_ioctl_resize(struct file *file,
 	if (!device->writeable) {
 		printk(KERN_INFO "btrfs: resizer unable to apply on "
 		       "readonly device %llu\n",
-		       (unsigned long long)devid);
+		       devid);
 		ret = -EPERM;
 		goto out_free;
 	}
@@ -1464,8 +1485,7 @@ static noinline int btrfs_ioctl_resize(struct file *file,
 	new_size *= root->sectorsize;
 
 	printk_in_rcu(KERN_INFO "btrfs: new size for %s is %llu\n",
-		      rcu_str_deref(device->name),
-		      (unsigned long long)new_size);
+		      rcu_str_deref(device->name), new_size);
 
 	if (new_size > old_size) {
 		trans = btrfs_start_transaction(root, 0);
@@ -1715,13 +1735,28 @@ out:
 static noinline int may_destroy_subvol(struct btrfs_root *root)
 {
 	struct btrfs_path *path;
+	struct btrfs_dir_item *di;
 	struct btrfs_key key;
+	u64 dir_id;
 	int ret;
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
+	/* Make sure this root isn't set as the default subvol */
+	dir_id = btrfs_super_root_dir(root->fs_info->super_copy);
+	di = btrfs_lookup_dir_item(NULL, root->fs_info->tree_root, path,
+				   dir_id, "default", 7, 0);
+	if (di && !IS_ERR(di)) {
+		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &key);
+		if (key.objectid == root->root_key.objectid) {
+			ret = -ENOTEMPTY;
+			goto out;
+		}
+		btrfs_release_path(path);
+	}
+
 	key.objectid = root->root_key.objectid;
 	key.type = BTRFS_ROOT_REF_KEY;
 	key.offset = (u64)-1;
@@ -1987,25 +2022,29 @@ static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info,
 		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 		if (ret < 0)
 			goto out;
+		else if (ret > 0) {
+			ret = btrfs_previous_item(root, path, dirid,
+						  BTRFS_INODE_REF_KEY);
+			if (ret < 0)
+				goto out;
+			else if (ret > 0) {
+				ret = -ENOENT;
+				goto out;
+			}
+		}
 
 		l = path->nodes[0];
 		slot = path->slots[0];
-		if (ret > 0 && slot > 0)
-			slot--;
 		btrfs_item_key_to_cpu(l, &key, slot);
 
-		if (ret > 0 && (key.objectid != dirid ||
-				key.type != BTRFS_INODE_REF_KEY)) {
-			ret = -ENOENT;
-			goto out;
-		}
-
 		iref = btrfs_item_ptr(l, slot, struct btrfs_inode_ref);
 		len = btrfs_inode_ref_name_len(l, iref);
 		ptr -= len + 1;
 		total_len += len + 1;
-		if (ptr < name)
+		if (ptr < name) {
+			ret = -ENAMETOOLONG;
 			goto out;
+		}
 
 		*(ptr + len) = '/';
 		read_extent_buffer(l, ptr,(unsigned long)(iref + 1), len);
@@ -2018,8 +2057,6 @@ static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info,
 		key.offset = (u64)-1;
 		dirid = key.objectid;
 	}
-	if (ptr < name)
-		goto out;
 	memmove(name, ptr, total_len);
 	name[total_len]='\0';
 	ret = 0;
@@ -2168,7 +2205,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 	 * ref/backref.
 	 */
 	err = btrfs_subvolume_reserve_metadata(root, &block_rsv,
-					       5, &qgroup_reserved);
+					       5, &qgroup_reserved, true);
 	if (err)
 		goto out_up_write;
 
@@ -2207,6 +2244,27 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 			goto out_end_trans;
 		}
 	}
+
+	ret = btrfs_uuid_tree_rem(trans, root->fs_info->uuid_root,
+				  dest->root_item.uuid, BTRFS_UUID_KEY_SUBVOL,
+				  dest->root_key.objectid);
+	if (ret && ret != -ENOENT) {
+		btrfs_abort_transaction(trans, root, ret);
+		err = ret;
+		goto out_end_trans;
+	}
+	if (!btrfs_is_empty_uuid(dest->root_item.received_uuid)) {
+		ret = btrfs_uuid_tree_rem(trans, root->fs_info->uuid_root,
+					  dest->root_item.received_uuid,
+					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
+					  dest->root_key.objectid);
+		if (ret && ret != -ENOENT) {
+			btrfs_abort_transaction(trans, root, ret);
+			err = ret;
+			goto out_end_trans;
+		}
+	}
+
 out_end_trans:
 	trans->block_rsv = NULL;
 	trans->bytes_reserved = 0;
@@ -2320,8 +2378,7 @@ static long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg)
 
 	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
 			1)) {
-		pr_info("btrfs: dev add/delete/balance/replace/resize operation in progress\n");
-		return -EINVAL;
+		return BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
 	}
 
 	mutex_lock(&root->fs_info->volume_mutex);
@@ -2354,14 +2411,6 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	if (ret)
 		return ret;
 
-	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
-			1)) {
-		pr_info("btrfs: dev add/delete/balance/replace/resize operation in progress\n");
-		mnt_drop_write_file(file);
-		return -EINVAL;
-	}
-
-	mutex_lock(&root->fs_info->volume_mutex);
 	vol_args = memdup_user(arg, sizeof(*vol_args));
 	if (IS_ERR(vol_args)) {
 		ret = PTR_ERR(vol_args);
@@ -2369,12 +2418,20 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	}
 
 	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
-	ret = btrfs_rm_device(root, vol_args->name);
 
-	kfree(vol_args);
-out:
+	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
+			1)) {
+		ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
+		goto out;
+	}
+
+	mutex_lock(&root->fs_info->volume_mutex);
+	ret = btrfs_rm_device(root, vol_args->name);
 	mutex_unlock(&root->fs_info->volume_mutex);
 	atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0);
+
+out:
+	kfree(vol_args);
 	mnt_drop_write_file(file);
 	return ret;
 }
@@ -2394,10 +2451,10 @@ static long btrfs_ioctl_fs_info(struct btrfs_root *root, void __user *arg)
 	if (!fi_args)
 		return -ENOMEM;
 
+	mutex_lock(&fs_devices->device_list_mutex);
 	fi_args->num_devices = fs_devices->num_devices;
 	memcpy(&fi_args->fsid, root->fs_info->fsid, sizeof(fi_args->fsid));
 
-	mutex_lock(&fs_devices->device_list_mutex);
 	list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
 		if (device->devid > fi_args->max_id)
 			fi_args->max_id = device->devid;
@@ -2418,7 +2475,6 @@ static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg)
 	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
 	int ret = 0;
 	char *s_uuid = NULL;
-	char empty_uuid[BTRFS_UUID_SIZE] = {0};
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
@@ -2427,7 +2483,7 @@ static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg)
 	if (IS_ERR(di_args))
 		return PTR_ERR(di_args);
 
-	if (memcmp(empty_uuid, di_args->uuid, BTRFS_UUID_SIZE) != 0)
+	if (!btrfs_is_empty_uuid(di_args->uuid))
 		s_uuid = di_args->uuid;
 
 	mutex_lock(&fs_devices->device_list_mutex);
@@ -2463,139 +2519,350 @@ out:
 	return ret;
 }
 
-static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
-				       u64 off, u64 olen, u64 destoff)
+static struct page *extent_same_get_page(struct inode *inode, u64 off)
+{
+	struct page *page;
+	pgoff_t index;
+	struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree;
+
+	index = off >> PAGE_CACHE_SHIFT;
+
+	page = grab_cache_page(inode->i_mapping, index);
+	if (!page)
+		return NULL;
+
+	if (!PageUptodate(page)) {
+		if (extent_read_full_page_nolock(tree, page, btrfs_get_extent,
+						 0))
+			return NULL;
+		lock_page(page);
+		if (!PageUptodate(page)) {
+			unlock_page(page);
+			page_cache_release(page);
+			return NULL;
+		}
+	}
+	unlock_page(page);
+
+	return page;
+}
+
+static inline void lock_extent_range(struct inode *inode, u64 off, u64 len)
+{
+	/* do any pending delalloc/csum calc on src, one way or
+	   another, and lock file content */
+	while (1) {
+		struct btrfs_ordered_extent *ordered;
+		lock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
+		ordered = btrfs_lookup_first_ordered_extent(inode,
+							    off + len - 1);
+		if (!ordered &&
+		    !test_range_bit(&BTRFS_I(inode)->io_tree, off,
+				    off + len - 1, EXTENT_DELALLOC, 0, NULL))
+			break;
+		unlock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
+		if (ordered)
+			btrfs_put_ordered_extent(ordered);
+		btrfs_wait_ordered_range(inode, off, len);
+	}
+}
+
+static void btrfs_double_unlock(struct inode *inode1, u64 loff1,
+				struct inode *inode2, u64 loff2, u64 len)
+{
+	unlock_extent(&BTRFS_I(inode1)->io_tree, loff1, loff1 + len - 1);
+	unlock_extent(&BTRFS_I(inode2)->io_tree, loff2, loff2 + len - 1);
+
+	mutex_unlock(&inode1->i_mutex);
+	mutex_unlock(&inode2->i_mutex);
+}
+
+static void btrfs_double_lock(struct inode *inode1, u64 loff1,
+			      struct inode *inode2, u64 loff2, u64 len)
+{
+	if (inode1 < inode2) {
+		swap(inode1, inode2);
+		swap(loff1, loff2);
+	}
+
+	mutex_lock_nested(&inode1->i_mutex, I_MUTEX_PARENT);
+	lock_extent_range(inode1, loff1, len);
+	if (inode1 != inode2) {
+		mutex_lock_nested(&inode2->i_mutex, I_MUTEX_CHILD);
+		lock_extent_range(inode2, loff2, len);
+	}
+}
+
+static int btrfs_cmp_data(struct inode *src, u64 loff, struct inode *dst,
+			  u64 dst_loff, u64 len)
+{
+	int ret = 0;
+	struct page *src_page, *dst_page;
+	unsigned int cmp_len = PAGE_CACHE_SIZE;
+	void *addr, *dst_addr;
+
+	while (len) {
+		if (len < PAGE_CACHE_SIZE)
+			cmp_len = len;
+
+		src_page = extent_same_get_page(src, loff);
+		if (!src_page)
+			return -EINVAL;
+		dst_page = extent_same_get_page(dst, dst_loff);
+		if (!dst_page) {
+			page_cache_release(src_page);
+			return -EINVAL;
+		}
+		addr = kmap_atomic(src_page);
+		dst_addr = kmap_atomic(dst_page);
+
+		flush_dcache_page(src_page);
+		flush_dcache_page(dst_page);
+
+		if (memcmp(addr, dst_addr, cmp_len))
+			ret = BTRFS_SAME_DATA_DIFFERS;
+
+		kunmap_atomic(addr);
+		kunmap_atomic(dst_addr);
+		page_cache_release(src_page);
+		page_cache_release(dst_page);
+
+		if (ret)
+			break;
+
+		loff += cmp_len;
+		dst_loff += cmp_len;
+		len -= cmp_len;
+	}
+
+	return ret;
+}
+
+static int extent_same_check_offsets(struct inode *inode, u64 off, u64 len)
+{
+	u64 bs = BTRFS_I(inode)->root->fs_info->sb->s_blocksize;
+
+	if (off + len > inode->i_size || off + len < off)
+		return -EINVAL;
+	/* Check that we are block aligned - btrfs_clone() requires this */
+	if (!IS_ALIGNED(off, bs) || !IS_ALIGNED(off + len, bs))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int btrfs_extent_same(struct inode *src, u64 loff, u64 len,
+			     struct inode *dst, u64 dst_loff)
 {
-	struct inode *inode = file_inode(file);
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct fd src_file;
-	struct inode *src;
-	struct btrfs_trans_handle *trans;
-	struct btrfs_path *path;
-	struct extent_buffer *leaf;
-	char *buf;
-	struct btrfs_key key;
-	u32 nritems;
-	int slot;
 	int ret;
-	u64 len = olen;
-	u64 bs = root->fs_info->sb->s_blocksize;
 
 	/*
-	 * TODO:
-	 * - split compressed inline extents.  annoying: we need to
-	 *   decompress into destination's address_space (the file offset
-	 *   may change, so source mapping won't do), then recompress (or
-	 *   otherwise reinsert) a subrange.
-	 * - allow ranges within the same file to be cloned (provided
-	 *   they don't overlap)?
+	 * btrfs_clone() can't handle extents in the same file
+	 * yet. Once that works, we can drop this check and replace it
+	 * with a check for the same inode, but overlapping extents.
 	 */
-
-	/* the destination must be opened for writing */
-	if (!(file->f_mode & FMODE_WRITE) || (file->f_flags & O_APPEND))
+	if (src == dst)
 		return -EINVAL;
 
-	if (btrfs_root_readonly(root))
-		return -EROFS;
+	btrfs_double_lock(src, loff, dst, dst_loff, len);
+
+	ret = extent_same_check_offsets(src, loff, len);
+	if (ret)
+		goto out_unlock;
+
+	ret = extent_same_check_offsets(dst, dst_loff, len);
+	if (ret)
+		goto out_unlock;
+
+	/* don't make the dst file partly checksummed */
+	if ((BTRFS_I(src)->flags & BTRFS_INODE_NODATASUM) !=
+	    (BTRFS_I(dst)->flags & BTRFS_INODE_NODATASUM)) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	ret = btrfs_cmp_data(src, loff, dst, dst_loff, len);
+	if (ret == 0)
+		ret = btrfs_clone(src, dst, loff, len, len, dst_loff);
+
+out_unlock:
+	btrfs_double_unlock(src, loff, dst, dst_loff, len);
+
+	return ret;
+}
+
+#define BTRFS_MAX_DEDUPE_LEN	(16 * 1024 * 1024)
+
+static long btrfs_ioctl_file_extent_same(struct file *file,
+					 void __user *argp)
+{
+	struct btrfs_ioctl_same_args tmp;
+	struct btrfs_ioctl_same_args *same;
+	struct btrfs_ioctl_same_extent_info *info;
+	struct inode *src = file->f_dentry->d_inode;
+	struct file *dst_file = NULL;
+	struct inode *dst;
+	u64 off;
+	u64 len;
+	int i;
+	int ret;
+	unsigned long size;
+	u64 bs = BTRFS_I(src)->root->fs_info->sb->s_blocksize;
+	bool is_admin = capable(CAP_SYS_ADMIN);
+
+	if (!(file->f_mode & FMODE_READ))
+		return -EINVAL;
 
 	ret = mnt_want_write_file(file);
 	if (ret)
 		return ret;
 
-	src_file = fdget(srcfd);
-	if (!src_file.file) {
-		ret = -EBADF;
-		goto out_drop_write;
+	if (copy_from_user(&tmp,
+			   (struct btrfs_ioctl_same_args __user *)argp,
+			   sizeof(tmp))) {
+		ret = -EFAULT;
+		goto out;
 	}
 
-	ret = -EXDEV;
-	if (src_file.file->f_path.mnt != file->f_path.mnt)
-		goto out_fput;
+	size = sizeof(tmp) +
+		tmp.dest_count * sizeof(struct btrfs_ioctl_same_extent_info);
 
-	src = file_inode(src_file.file);
+	same = kmalloc(size, GFP_NOFS);
+	if (!same) {
+		ret = -EFAULT;
+		goto out;
+	}
 
-	ret = -EINVAL;
-	if (src == inode)
-		goto out_fput;
+	if (copy_from_user(same,
+			   (struct btrfs_ioctl_same_args __user *)argp, size)) {
+		ret = -EFAULT;
+		goto out;
+	}
 
-	/* the src must be open for reading */
-	if (!(src_file.file->f_mode & FMODE_READ))
-		goto out_fput;
+	off = same->logical_offset;
+	len = same->length;
 
-	/* don't make the dst file partly checksummed */
-	if ((BTRFS_I(src)->flags & BTRFS_INODE_NODATASUM) !=
-	    (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM))
-		goto out_fput;
+	/*
+	 * Limit the total length we will dedupe for each operation.
+	 * This is intended to bound the total time spent in this
+	 * ioctl to something sane.
+	 */
+	if (len > BTRFS_MAX_DEDUPE_LEN)
+		len = BTRFS_MAX_DEDUPE_LEN;
 
-	ret = -EISDIR;
-	if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode))
-		goto out_fput;
+	if (WARN_ON_ONCE(bs < PAGE_CACHE_SIZE)) {
+		/*
+		 * Btrfs does not support blocksize < page_size. As a
+		 * result, btrfs_cmp_data() won't correctly handle
+		 * this situation without an update.
+		 */
+		ret = -EINVAL;
+		goto out;
+	}
 
-	ret = -EXDEV;
-	if (src->i_sb != inode->i_sb)
-		goto out_fput;
+	ret = -EISDIR;
+	if (S_ISDIR(src->i_mode))
+		goto out;
 
-	ret = -ENOMEM;
-	buf = vmalloc(btrfs_level_size(root, 0));
-	if (!buf)
-		goto out_fput;
+	ret = -EACCES;
+	if (!S_ISREG(src->i_mode))
+		goto out;
 
-	path = btrfs_alloc_path();
-	if (!path) {
-		vfree(buf);
-		goto out_fput;
+	/* pre-format output fields to sane values */
+	for (i = 0; i < same->dest_count; i++) {
+		same->info[i].bytes_deduped = 0ULL;
+		same->info[i].status = 0;
 	}
-	path->reada = 2;
 
-	if (inode < src) {
-		mutex_lock_nested(&inode->i_mutex, I_MUTEX_PARENT);
-		mutex_lock_nested(&src->i_mutex, I_MUTEX_CHILD);
-	} else {
-		mutex_lock_nested(&src->i_mutex, I_MUTEX_PARENT);
-		mutex_lock_nested(&inode->i_mutex, I_MUTEX_CHILD);
-	}
+	ret = 0;
+	for (i = 0; i < same->dest_count; i++) {
+		info = &same->info[i];
 
-	/* determine range to clone */
-	ret = -EINVAL;
-	if (off + len > src->i_size || off + len < off)
-		goto out_unlock;
-	if (len == 0)
-		olen = len = src->i_size - off;
-	/* if we extend to eof, continue to block boundary */
-	if (off + len == src->i_size)
-		len = ALIGN(src->i_size, bs) - off;
+		dst_file = fget(info->fd);
+		if (!dst_file) {
+			info->status = -EBADF;
+			goto next;
+		}
 
-	/* verify the end result is block aligned */
-	if (!IS_ALIGNED(off, bs) || !IS_ALIGNED(off + len, bs) ||
-	    !IS_ALIGNED(destoff, bs))
-		goto out_unlock;
+		if (!(is_admin || (dst_file->f_mode & FMODE_WRITE))) {
+			info->status = -EINVAL;
+			goto next;
+		}
 
-	if (destoff > inode->i_size) {
-		ret = btrfs_cont_expand(inode, inode->i_size, destoff);
-		if (ret)
-			goto out_unlock;
+		info->status = -EXDEV;
+		if (file->f_path.mnt != dst_file->f_path.mnt)
+			goto next;
+
+		dst = dst_file->f_dentry->d_inode;
+		if (src->i_sb != dst->i_sb)
+			goto next;
+
+		if (S_ISDIR(dst->i_mode)) {
+			info->status = -EISDIR;
+			goto next;
+		}
+
+		if (!S_ISREG(dst->i_mode)) {
+			info->status = -EACCES;
+			goto next;
+		}
+
+		info->status = btrfs_extent_same(src, off, len, dst,
+						info->logical_offset);
+		if (info->status == 0)
+			info->bytes_deduped += len;
+
+next:
+		if (dst_file)
+			fput(dst_file);
 	}
 
-	/* truncate page cache pages from target inode range */
-	truncate_inode_pages_range(&inode->i_data, destoff,
-				   PAGE_CACHE_ALIGN(destoff + len) - 1);
+	ret = copy_to_user(argp, same, size);
+	if (ret)
+		ret = -EFAULT;
 
-	/* do any pending delalloc/csum calc on src, one way or
-	   another, and lock file content */
-	while (1) {
-		struct btrfs_ordered_extent *ordered;
-		lock_extent(&BTRFS_I(src)->io_tree, off, off + len - 1);
-		ordered = btrfs_lookup_first_ordered_extent(src, off + len - 1);
-		if (!ordered &&
-		    !test_range_bit(&BTRFS_I(src)->io_tree, off, off + len - 1,
-				    EXTENT_DELALLOC, 0, NULL))
-			break;
-		unlock_extent(&BTRFS_I(src)->io_tree, off, off + len - 1);
-		if (ordered)
-			btrfs_put_ordered_extent(ordered);
-		btrfs_wait_ordered_range(src, off, len);
+out:
+	mnt_drop_write_file(file);
+	return ret;
+}
+
+/**
+ * btrfs_clone() - clone a range from inode file to another
+ *
+ * @src: Inode to clone from
+ * @inode: Inode to clone to
+ * @off: Offset within source to start clone from
+ * @olen: Original length, passed by user, of range to clone
+ * @olen_aligned: Block-aligned value of olen, extent_same uses
+ *               identical values here
+ * @destoff: Offset within @inode to start clone
+ */
+static int btrfs_clone(struct inode *src, struct inode *inode,
+		       u64 off, u64 olen, u64 olen_aligned, u64 destoff)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_path *path = NULL;
+	struct extent_buffer *leaf;
+	struct btrfs_trans_handle *trans;
+	char *buf = NULL;
+	struct btrfs_key key;
+	u32 nritems;
+	int slot;
+	int ret;
+	u64 len = olen_aligned;
+
+	ret = -ENOMEM;
+	buf = vmalloc(btrfs_level_size(root, 0));
+	if (!buf)
+		return ret;
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		vfree(buf);
+		return ret;
 	}
 
+	path->reada = 2;
 	/* clone data */
 	key.objectid = btrfs_ino(src);
 	key.type = BTRFS_EXTENT_DATA_KEY;
@@ -2841,14 +3108,132 @@ next:
 		key.offset++;
 	}
 	ret = 0;
+
 out:
 	btrfs_release_path(path);
+	btrfs_free_path(path);
+	vfree(buf);
+	return ret;
+}
+
+static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
+				       u64 off, u64 olen, u64 destoff)
+{
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct fd src_file;
+	struct inode *src;
+	int ret;
+	u64 len = olen;
+	u64 bs = root->fs_info->sb->s_blocksize;
+	int same_inode = 0;
+
+	/*
+	 * TODO:
+	 * - split compressed inline extents.  annoying: we need to
+	 *   decompress into destination's address_space (the file offset
+	 *   may change, so source mapping won't do), then recompress (or
+	 *   otherwise reinsert) a subrange.
+	 * - allow ranges within the same file to be cloned (provided
+	 *   they don't overlap)?
+	 */
+
+	/* the destination must be opened for writing */
+	if (!(file->f_mode & FMODE_WRITE) || (file->f_flags & O_APPEND))
+		return -EINVAL;
+
+	if (btrfs_root_readonly(root))
+		return -EROFS;
+
+	ret = mnt_want_write_file(file);
+	if (ret)
+		return ret;
+
+	src_file = fdget(srcfd);
+	if (!src_file.file) {
+		ret = -EBADF;
+		goto out_drop_write;
+	}
+
+	ret = -EXDEV;
+	if (src_file.file->f_path.mnt != file->f_path.mnt)
+		goto out_fput;
+
+	src = file_inode(src_file.file);
+
+	ret = -EINVAL;
+	if (src == inode)
+		same_inode = 1;
+
+	/* the src must be open for reading */
+	if (!(src_file.file->f_mode & FMODE_READ))
+		goto out_fput;
+
+	/* don't make the dst file partly checksummed */
+	if ((BTRFS_I(src)->flags & BTRFS_INODE_NODATASUM) !=
+	    (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM))
+		goto out_fput;
+
+	ret = -EISDIR;
+	if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode))
+		goto out_fput;
+
+	ret = -EXDEV;
+	if (src->i_sb != inode->i_sb)
+		goto out_fput;
+
+	if (!same_inode) {
+		if (inode < src) {
+			mutex_lock_nested(&inode->i_mutex, I_MUTEX_PARENT);
+			mutex_lock_nested(&src->i_mutex, I_MUTEX_CHILD);
+		} else {
+			mutex_lock_nested(&src->i_mutex, I_MUTEX_PARENT);
+			mutex_lock_nested(&inode->i_mutex, I_MUTEX_CHILD);
+		}
+	} else {
+		mutex_lock(&src->i_mutex);
+	}
+
+	/* determine range to clone */
+	ret = -EINVAL;
+	if (off + len > src->i_size || off + len < off)
+		goto out_unlock;
+	if (len == 0)
+		olen = len = src->i_size - off;
+	/* if we extend to eof, continue to block boundary */
+	if (off + len == src->i_size)
+		len = ALIGN(src->i_size, bs) - off;
+
+	/* verify the end result is block aligned */
+	if (!IS_ALIGNED(off, bs) || !IS_ALIGNED(off + len, bs) ||
+	    !IS_ALIGNED(destoff, bs))
+		goto out_unlock;
+
+	/* verify if ranges are overlapped within the same file */
+	if (same_inode) {
+		if (destoff + len > off && destoff < off + len)
+			goto out_unlock;
+	}
+
+	if (destoff > inode->i_size) {
+		ret = btrfs_cont_expand(inode, inode->i_size, destoff);
+		if (ret)
+			goto out_unlock;
+	}
+
+	/* truncate page cache pages from target inode range */
+	truncate_inode_pages_range(&inode->i_data, destoff,
+				   PAGE_CACHE_ALIGN(destoff + len) - 1);
+
+	lock_extent_range(src, off, len);
+
+	ret = btrfs_clone(src, inode, off, olen, len, destoff);
+
 	unlock_extent(&BTRFS_I(src)->io_tree, off, off + len - 1);
 out_unlock:
 	mutex_unlock(&src->i_mutex);
-	mutex_unlock(&inode->i_mutex);
-	vfree(buf);
-	btrfs_free_path(path);
+	if (!same_inode)
+		mutex_unlock(&inode->i_mutex);
 out_fput:
 	fdput(src_file);
 out_drop_write:
@@ -2939,7 +3324,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp)
 	}
 
 	if (!objectid)
-		objectid = root->root_key.objectid;
+		objectid = BTRFS_FS_TREE_OBJECTID;
 
 	location.objectid = objectid;
 	location.type = BTRFS_ROOT_ITEM_KEY;
@@ -2951,11 +3336,6 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp)
 		goto out;
 	}
 
-	if (btrfs_root_refs(&new_root->root_item) == 0) {
-		ret = -ENOENT;
-		goto out;
-	}
-
 	path = btrfs_alloc_path();
 	if (!path) {
 		ret = -ENOMEM;
@@ -3299,11 +3679,13 @@ static long btrfs_ioctl_dev_replace(struct btrfs_root *root, void __user *arg)
 
 	switch (p->cmd) {
 	case BTRFS_IOCTL_DEV_REPLACE_CMD_START:
+		if (root->fs_info->sb->s_flags & MS_RDONLY)
+			return -EROFS;
+
 		if (atomic_xchg(
 			&root->fs_info->mutually_exclusive_operation_running,
 			1)) {
-			pr_info("btrfs: dev add/delete/balance/replace/resize operation in progress\n");
-			ret = -EINPROGRESS;
+			ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
 		} else {
 			ret = btrfs_dev_replace_start(root, p);
 			atomic_set(
@@ -3547,8 +3929,7 @@ again:
 	} else {
 		/* this is (1) */
 		mutex_unlock(&fs_info->balance_mutex);
-		pr_info("btrfs: dev add/delete/balance/replace/resize operation in progress\n");
-		ret = -EINVAL;
+		ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
 		goto out;
 	}
 
@@ -3719,9 +4100,6 @@ static long btrfs_ioctl_quota_ctl(struct file *file, void __user *arg)
 		break;
 	}
 
-	if (copy_to_user(arg, sa, sizeof(*sa)))
-		ret = -EFAULT;
-
 	err = btrfs_commit_transaction(trans, root->fs_info->tree_root);
 	if (err && !ret)
 		ret = err;
@@ -3881,7 +4259,7 @@ drop_write:
 
 static long btrfs_ioctl_quota_rescan(struct file *file, void __user *arg)
 {
-	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
+	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
 	struct btrfs_ioctl_quota_rescan_args *qsa;
 	int ret;
 
@@ -3914,7 +4292,7 @@ drop_write:
 
 static long btrfs_ioctl_quota_rescan_status(struct file *file, void __user *arg)
 {
-	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
+	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
 	struct btrfs_ioctl_quota_rescan_args *qsa;
 	int ret = 0;
 
@@ -3937,6 +4315,16 @@ static long btrfs_ioctl_quota_rescan_status(struct file *file, void __user *arg)
 	return ret;
 }
 
+static long btrfs_ioctl_quota_rescan_wait(struct file *file, void __user *arg)
+{
+	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	return btrfs_qgroup_wait_for_completion(root->fs_info);
+}
+
 static long btrfs_ioctl_set_received_subvol(struct file *file,
 					    void __user *arg)
 {
@@ -3947,6 +4335,7 @@ static long btrfs_ioctl_set_received_subvol(struct file *file,
 	struct btrfs_trans_handle *trans;
 	struct timespec ct = CURRENT_TIME;
 	int ret = 0;
+	int received_uuid_changed;
 
 	ret = mnt_want_write_file(file);
 	if (ret < 0)
@@ -3976,7 +4365,11 @@ static long btrfs_ioctl_set_received_subvol(struct file *file,
 		goto out;
 	}
 
-	trans = btrfs_start_transaction(root, 1);
+	/*
+	 * 1 - root item
+	 * 2 - uuid items (received uuid + subvol uuid)
+	 */
+	trans = btrfs_start_transaction(root, 3);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
 		trans = NULL;
@@ -3987,24 +4380,42 @@ static long btrfs_ioctl_set_received_subvol(struct file *file,
 	sa->rtime.sec = ct.tv_sec;
 	sa->rtime.nsec = ct.tv_nsec;
 
+	received_uuid_changed = memcmp(root_item->received_uuid, sa->uuid,
+				       BTRFS_UUID_SIZE);
+	if (received_uuid_changed &&
+	    !btrfs_is_empty_uuid(root_item->received_uuid))
+		btrfs_uuid_tree_rem(trans, root->fs_info->uuid_root,
+				    root_item->received_uuid,
+				    BTRFS_UUID_KEY_RECEIVED_SUBVOL,
+				    root->root_key.objectid);
 	memcpy(root_item->received_uuid, sa->uuid, BTRFS_UUID_SIZE);
 	btrfs_set_root_stransid(root_item, sa->stransid);
 	btrfs_set_root_rtransid(root_item, sa->rtransid);
-	root_item->stime.sec = cpu_to_le64(sa->stime.sec);
-	root_item->stime.nsec = cpu_to_le32(sa->stime.nsec);
-	root_item->rtime.sec = cpu_to_le64(sa->rtime.sec);
-	root_item->rtime.nsec = cpu_to_le32(sa->rtime.nsec);
+	btrfs_set_stack_timespec_sec(&root_item->stime, sa->stime.sec);
+	btrfs_set_stack_timespec_nsec(&root_item->stime, sa->stime.nsec);
+	btrfs_set_stack_timespec_sec(&root_item->rtime, sa->rtime.sec);
+	btrfs_set_stack_timespec_nsec(&root_item->rtime, sa->rtime.nsec);
 
 	ret = btrfs_update_root(trans, root->fs_info->tree_root,
 				&root->root_key, &root->root_item);
 	if (ret < 0) {
 		btrfs_end_transaction(trans, root);
-		trans = NULL;
 		goto out;
-	} else {
-		ret = btrfs_commit_transaction(trans, root);
-		if (ret < 0)
+	}
+	if (received_uuid_changed && !btrfs_is_empty_uuid(sa->uuid)) {
+		ret = btrfs_uuid_tree_add(trans, root->fs_info->uuid_root,
+					  sa->uuid,
+					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
+					  root->root_key.objectid);
+		if (ret < 0 && ret != -EEXIST) {
+			btrfs_abort_transaction(trans, root, ret);
 			goto out;
+		}
+	}
+	ret = btrfs_commit_transaction(trans, root);
+	if (ret < 0) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out;
 	}
 
 	ret = copy_to_user(arg, sa, sizeof(*sa));
@@ -4020,26 +4431,30 @@ out:
 
 static int btrfs_ioctl_get_fslabel(struct file *file, void __user *arg)
 {
-	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
-	const char *label = root->fs_info->super_copy->label;
-	size_t len = strnlen(label, BTRFS_LABEL_SIZE);
+	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
+	size_t len;
 	int ret;
+	char label[BTRFS_LABEL_SIZE];
+
+	spin_lock(&root->fs_info->super_lock);
+	memcpy(label, root->fs_info->super_copy->label, BTRFS_LABEL_SIZE);
+	spin_unlock(&root->fs_info->super_lock);
+
+	len = strnlen(label, BTRFS_LABEL_SIZE);
 
 	if (len == BTRFS_LABEL_SIZE) {
 		pr_warn("btrfs: label is too long, return the first %zu bytes\n",
 			--len);
 	}
 
-	mutex_lock(&root->fs_info->volume_mutex);
 	ret = copy_to_user(arg, label, len);
-	mutex_unlock(&root->fs_info->volume_mutex);
 
 	return ret ? -EFAULT : 0;
 }
 
 static int btrfs_ioctl_set_fslabel(struct file *file, void __user *arg)
 {
-	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
+	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
 	struct btrfs_super_block *super_block = root->fs_info->super_copy;
 	struct btrfs_trans_handle *trans;
 	char label[BTRFS_LABEL_SIZE];
@@ -4061,18 +4476,18 @@ static int btrfs_ioctl_set_fslabel(struct file *file, void __user *arg)
 	if (ret)
 		return ret;
 
-	mutex_lock(&root->fs_info->volume_mutex);
 	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
 		goto out_unlock;
 	}
 
+	spin_lock(&root->fs_info->super_lock);
 	strcpy(super_block->label, label);
+	spin_unlock(&root->fs_info->super_lock);
 	ret = btrfs_end_transaction(trans, root);
 
 out_unlock:
-	mutex_unlock(&root->fs_info->volume_mutex);
 	mnt_drop_write_file(file);
 	return ret;
 }
@@ -4179,12 +4594,16 @@ long btrfs_ioctl(struct file *file, unsigned int
 		return btrfs_ioctl_quota_rescan(file, argp);
 	case BTRFS_IOC_QUOTA_RESCAN_STATUS:
 		return btrfs_ioctl_quota_rescan_status(file, argp);
+	case BTRFS_IOC_QUOTA_RESCAN_WAIT:
+		return btrfs_ioctl_quota_rescan_wait(file, argp);
 	case BTRFS_IOC_DEV_REPLACE:
 		return btrfs_ioctl_dev_replace(root, argp);
 	case BTRFS_IOC_GET_FSLABEL:
 		return btrfs_ioctl_get_fslabel(file, argp);
 	case BTRFS_IOC_SET_FSLABEL:
 		return btrfs_ioctl_set_fslabel(file, argp);
+	case BTRFS_IOC_FILE_EXTENT_SAME:
+		return btrfs_ioctl_file_extent_same(file, argp);
 	}
 
 	return -ENOTTY;
diff --git a/fs/btrfs/lzo.c b/fs/btrfs/lzo.c
index 743b86fa4fcb..b6a6f07c5ce2 100644
--- a/fs/btrfs/lzo.c
+++ b/fs/btrfs/lzo.c
@@ -31,8 +31,8 @@
 
 struct workspace {
 	void *mem;
-	void *buf;	/* where compressed data goes */
-	void *cbuf;	/* where decompressed data goes */
+	void *buf;	/* where decompressed data goes */
+	void *cbuf;	/* where compressed data goes */
 	struct list_head list;
 };
 
@@ -207,8 +207,10 @@ static int lzo_compress_pages(struct list_head *ws,
 		}
 
 		/* we're making it bigger, give up */
-		if (tot_in > 8192 && tot_in < tot_out)
+		if (tot_in > 8192 && tot_in < tot_out) {
+			ret = -1;
 			goto out;
+		}
 
 		/* we're all done */
 		if (tot_in >= len)
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index 1ddd728541ee..c702cb62f78a 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -24,6 +24,7 @@
 #include "transaction.h"
 #include "btrfs_inode.h"
 #include "extent_io.h"
+#include "disk-io.h"
 
 static struct kmem_cache *btrfs_ordered_extent_cache;
 
@@ -66,7 +67,7 @@ static void ordered_data_tree_panic(struct inode *inode, int errno,
 {
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
 	btrfs_panic(fs_info, errno, "Inconsistency in ordered tree at offset "
-		    "%llu\n", (unsigned long long)offset);
+		    "%llu\n", offset);
 }
 
 /*
@@ -184,6 +185,7 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 				      u64 start, u64 len, u64 disk_len,
 				      int type, int dio, int compress_type)
 {
+	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_ordered_inode_tree *tree;
 	struct rb_node *node;
 	struct btrfs_ordered_extent *entry;
@@ -203,6 +205,7 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 	entry->bytes_left = len;
 	entry->inode = igrab(inode);
 	entry->compress_type = compress_type;
+	entry->truncated_len = (u64)-1;
 	if (type != BTRFS_ORDERED_IO_DONE && type != BTRFS_ORDERED_COMPLETE)
 		set_bit(type, &entry->flags);
 
@@ -227,10 +230,18 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 		ordered_data_tree_panic(inode, -EEXIST, file_offset);
 	spin_unlock_irq(&tree->lock);
 
-	spin_lock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
+	spin_lock(&root->ordered_extent_lock);
 	list_add_tail(&entry->root_extent_list,
-		      &BTRFS_I(inode)->root->fs_info->ordered_extents);
-	spin_unlock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
+		      &root->ordered_extents);
+	root->nr_ordered_extents++;
+	if (root->nr_ordered_extents == 1) {
+		spin_lock(&root->fs_info->ordered_root_lock);
+		BUG_ON(!list_empty(&root->ordered_root));
+		list_add_tail(&root->ordered_root,
+			      &root->fs_info->ordered_roots);
+		spin_unlock(&root->fs_info->ordered_root_lock);
+	}
+	spin_unlock(&root->ordered_extent_lock);
 
 	return 0;
 }
@@ -326,14 +337,12 @@ int btrfs_dec_test_first_ordered_pending(struct inode *inode,
 	*file_offset = dec_end;
 	if (dec_start > dec_end) {
 		printk(KERN_CRIT "bad ordering dec_start %llu end %llu\n",
-		       (unsigned long long)dec_start,
-		       (unsigned long long)dec_end);
+		       dec_start, dec_end);
 	}
 	to_dec = dec_end - dec_start;
 	if (to_dec > entry->bytes_left) {
 		printk(KERN_CRIT "bad ordered accounting left %llu size %llu\n",
-		       (unsigned long long)entry->bytes_left,
-		       (unsigned long long)to_dec);
+		       entry->bytes_left, to_dec);
 	}
 	entry->bytes_left -= to_dec;
 	if (!uptodate)
@@ -393,8 +402,7 @@ have_entry:
 
 	if (io_size > entry->bytes_left) {
 		printk(KERN_CRIT "bad ordered accounting left %llu size %llu\n",
-		       (unsigned long long)entry->bytes_left,
-		       (unsigned long long)io_size);
+		       entry->bytes_left, io_size);
 	}
 	entry->bytes_left -= io_size;
 	if (!uptodate)
@@ -516,8 +524,9 @@ void btrfs_remove_ordered_extent(struct inode *inode,
 	set_bit(BTRFS_ORDERED_COMPLETE, &entry->flags);
 	spin_unlock_irq(&tree->lock);
 
-	spin_lock(&root->fs_info->ordered_extent_lock);
+	spin_lock(&root->ordered_extent_lock);
 	list_del_init(&entry->root_extent_list);
+	root->nr_ordered_extents--;
 
 	trace_btrfs_ordered_extent_remove(inode, entry);
 
@@ -530,7 +539,14 @@ void btrfs_remove_ordered_extent(struct inode *inode,
 	    !mapping_tagged(inode->i_mapping, PAGECACHE_TAG_DIRTY)) {
 		list_del_init(&BTRFS_I(inode)->ordered_operations);
 	}
-	spin_unlock(&root->fs_info->ordered_extent_lock);
+
+	if (!root->nr_ordered_extents) {
+		spin_lock(&root->fs_info->ordered_root_lock);
+		BUG_ON(list_empty(&root->ordered_root));
+		list_del_init(&root->ordered_root);
+		spin_unlock(&root->fs_info->ordered_root_lock);
+	}
+	spin_unlock(&root->ordered_extent_lock);
 	wake_up(&entry->wait);
 }
 
@@ -547,63 +563,70 @@ static void btrfs_run_ordered_extent_work(struct btrfs_work *work)
  * wait for all the ordered extents in a root.  This is done when balancing
  * space between drives.
  */
-void btrfs_wait_ordered_extents(struct btrfs_root *root, int delay_iput)
+void btrfs_wait_ordered_extents(struct btrfs_root *root)
 {
 	struct list_head splice, works;
-	struct list_head *cur;
 	struct btrfs_ordered_extent *ordered, *next;
-	struct inode *inode;
 
 	INIT_LIST_HEAD(&splice);
 	INIT_LIST_HEAD(&works);
 
 	mutex_lock(&root->fs_info->ordered_operations_mutex);
-	spin_lock(&root->fs_info->ordered_extent_lock);
-	list_splice_init(&root->fs_info->ordered_extents, &splice);
+	spin_lock(&root->ordered_extent_lock);
+	list_splice_init(&root->ordered_extents, &splice);
 	while (!list_empty(&splice)) {
-		cur = splice.next;
-		ordered = list_entry(cur, struct btrfs_ordered_extent,
-				     root_extent_list);
-		list_del_init(&ordered->root_extent_list);
+		ordered = list_first_entry(&splice, struct btrfs_ordered_extent,
+					   root_extent_list);
+		list_move_tail(&ordered->root_extent_list,
+			       &root->ordered_extents);
 		atomic_inc(&ordered->refs);
+		spin_unlock(&root->ordered_extent_lock);
 
-		/*
-		 * the inode may be getting freed (in sys_unlink path).
-		 */
-		inode = igrab(ordered->inode);
-
-		spin_unlock(&root->fs_info->ordered_extent_lock);
-
-		if (inode) {
-			ordered->flush_work.func = btrfs_run_ordered_extent_work;
-			list_add_tail(&ordered->work_list, &works);
-			btrfs_queue_worker(&root->fs_info->flush_workers,
-					   &ordered->flush_work);
-		} else {
-			btrfs_put_ordered_extent(ordered);
-		}
+		ordered->flush_work.func = btrfs_run_ordered_extent_work;
+		list_add_tail(&ordered->work_list, &works);
+		btrfs_queue_worker(&root->fs_info->flush_workers,
+				   &ordered->flush_work);
 
 		cond_resched();
-		spin_lock(&root->fs_info->ordered_extent_lock);
+		spin_lock(&root->ordered_extent_lock);
 	}
-	spin_unlock(&root->fs_info->ordered_extent_lock);
+	spin_unlock(&root->ordered_extent_lock);
 
 	list_for_each_entry_safe(ordered, next, &works, work_list) {
 		list_del_init(&ordered->work_list);
 		wait_for_completion(&ordered->completion);
-
-		inode = ordered->inode;
 		btrfs_put_ordered_extent(ordered);
-		if (delay_iput)
-			btrfs_add_delayed_iput(inode);
-		else
-			iput(inode);
-
 		cond_resched();
 	}
 	mutex_unlock(&root->fs_info->ordered_operations_mutex);
 }
 
+void btrfs_wait_all_ordered_extents(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_root *root;
+	struct list_head splice;
+
+	INIT_LIST_HEAD(&splice);
+
+	spin_lock(&fs_info->ordered_root_lock);
+	list_splice_init(&fs_info->ordered_roots, &splice);
+	while (!list_empty(&splice)) {
+		root = list_first_entry(&splice, struct btrfs_root,
+					ordered_root);
+		root = btrfs_grab_fs_root(root);
+		BUG_ON(!root);
+		list_move_tail(&root->ordered_root,
+			       &fs_info->ordered_roots);
+		spin_unlock(&fs_info->ordered_root_lock);
+
+		btrfs_wait_ordered_extents(root);
+		btrfs_put_fs_root(root);
+
+		spin_lock(&fs_info->ordered_root_lock);
+	}
+	spin_unlock(&fs_info->ordered_root_lock);
+}
+
 /*
  * this is used during transaction commit to write all the inodes
  * added to the ordered operation list.  These files must be fully on
@@ -628,8 +651,8 @@ int btrfs_run_ordered_operations(struct btrfs_trans_handle *trans,
 	INIT_LIST_HEAD(&splice);
 	INIT_LIST_HEAD(&works);
 
-	mutex_lock(&root->fs_info->ordered_operations_mutex);
-	spin_lock(&root->fs_info->ordered_extent_lock);
+	mutex_lock(&root->fs_info->ordered_extent_flush_mutex);
+	spin_lock(&root->fs_info->ordered_root_lock);
 	list_splice_init(&cur_trans->ordered_operations, &splice);
 	while (!list_empty(&splice)) {
 		btrfs_inode = list_entry(splice.next, struct btrfs_inode,
@@ -648,17 +671,17 @@ int btrfs_run_ordered_operations(struct btrfs_trans_handle *trans,
 		if (!wait)
 			list_add_tail(&BTRFS_I(inode)->ordered_operations,
 				      &cur_trans->ordered_operations);
-		spin_unlock(&root->fs_info->ordered_extent_lock);
+		spin_unlock(&root->fs_info->ordered_root_lock);
 
 		work = btrfs_alloc_delalloc_work(inode, wait, 1);
 		if (!work) {
-			spin_lock(&root->fs_info->ordered_extent_lock);
+			spin_lock(&root->fs_info->ordered_root_lock);
 			if (list_empty(&BTRFS_I(inode)->ordered_operations))
 				list_add_tail(&btrfs_inode->ordered_operations,
 					      &splice);
 			list_splice_tail(&splice,
 					 &cur_trans->ordered_operations);
-			spin_unlock(&root->fs_info->ordered_extent_lock);
+			spin_unlock(&root->fs_info->ordered_root_lock);
 			ret = -ENOMEM;
 			goto out;
 		}
@@ -667,15 +690,15 @@ int btrfs_run_ordered_operations(struct btrfs_trans_handle *trans,
 				   &work->work);
 
 		cond_resched();
-		spin_lock(&root->fs_info->ordered_extent_lock);
+		spin_lock(&root->fs_info->ordered_root_lock);
 	}
-	spin_unlock(&root->fs_info->ordered_extent_lock);
+	spin_unlock(&root->fs_info->ordered_root_lock);
 out:
 	list_for_each_entry_safe(work, next, &works, list) {
 		list_del_init(&work->list);
 		btrfs_wait_and_free_delalloc_work(work);
 	}
-	mutex_unlock(&root->fs_info->ordered_operations_mutex);
+	mutex_unlock(&root->fs_info->ordered_extent_flush_mutex);
 	return ret;
 }
 
@@ -880,12 +903,16 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset,
 	struct btrfs_ordered_extent *test;
 	int ret = 1;
 
-	if (ordered)
+	spin_lock_irq(&tree->lock);
+	if (ordered) {
 		offset = entry_end(ordered);
-	else
+		if (test_bit(BTRFS_ORDERED_TRUNCATED, &ordered->flags))
+			offset = min(offset,
+				     ordered->file_offset +
+				     ordered->truncated_len);
+	} else {
 		offset = ALIGN(offset, BTRFS_I(inode)->root->sectorsize);
-
-	spin_lock_irq(&tree->lock);
+	}
 	disk_i_size = BTRFS_I(inode)->disk_i_size;
 
 	/* truncate file */
@@ -989,7 +1016,6 @@ int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
 			   u32 *sum, int len)
 {
 	struct btrfs_ordered_sum *ordered_sum;
-	struct btrfs_sector_sum *sector_sums;
 	struct btrfs_ordered_extent *ordered;
 	struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree;
 	unsigned long num_sectors;
@@ -1007,18 +1033,16 @@ int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
 		    disk_bytenr < ordered_sum->bytenr + ordered_sum->len) {
 			i = (disk_bytenr - ordered_sum->bytenr) >>
 			    inode->i_sb->s_blocksize_bits;
-			sector_sums = ordered_sum->sums + i;
 			num_sectors = ordered_sum->len >>
 				      inode->i_sb->s_blocksize_bits;
-			for (; i < num_sectors; i++) {
-				if (sector_sums[i].bytenr == disk_bytenr) {
-					sum[index] = sector_sums[i].sum;
-					index++;
-					if (index == len)
-						goto out;
-					disk_bytenr += sectorsize;
-				}
-			}
+			num_sectors = min_t(int, len - index, num_sectors - i);
+			memcpy(sum + index, ordered_sum->sums + i,
+			       num_sectors);
+
+			index += (int)num_sectors;
+			if (index == len)
+				goto out;
+			disk_bytenr += num_sectors * sectorsize;
 		}
 	}
 out:
@@ -1055,12 +1079,12 @@ void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
 	if (last_mod < root->fs_info->last_trans_committed)
 		return;
 
-	spin_lock(&root->fs_info->ordered_extent_lock);
+	spin_lock(&root->fs_info->ordered_root_lock);
 	if (list_empty(&BTRFS_I(inode)->ordered_operations)) {
 		list_add_tail(&BTRFS_I(inode)->ordered_operations,
 			      &cur_trans->ordered_operations);
 	}
-	spin_unlock(&root->fs_info->ordered_extent_lock);
+	spin_unlock(&root->fs_info->ordered_root_lock);
 }
 
 int __init ordered_data_init(void)
diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h
index 58b0e3b0ebad..0c0b35612d7a 100644
--- a/fs/btrfs/ordered-data.h
+++ b/fs/btrfs/ordered-data.h
@@ -26,18 +26,6 @@ struct btrfs_ordered_inode_tree {
 	struct rb_node *last;
 };
 
-/*
- * these are used to collect checksums done just before bios submission.
- * They are attached via a list into the ordered extent, and
- * checksum items are inserted into the tree after all the blocks in
- * the ordered extent are on disk
- */
-struct btrfs_sector_sum {
-	/* bytenr on disk */
-	u64 bytenr;
-	u32 sum;
-};
-
 struct btrfs_ordered_sum {
 	/* bytenr is the start of this extent on disk */
 	u64 bytenr;
@@ -45,10 +33,10 @@ struct btrfs_ordered_sum {
 	/*
 	 * this is the length in bytes covered by the sums array below.
 	 */
-	unsigned long len;
+	int len;
 	struct list_head list;
-	/* last field is a variable length array of btrfs_sector_sums */
-	struct btrfs_sector_sum sums[];
+	/* last field is a variable length array of csums */
+	u32 sums[];
 };
 
 /*
@@ -81,6 +69,7 @@ struct btrfs_ordered_sum {
 				       * the isize. */
 #define BTRFS_ORDERED_LOGGED_CSUM 8 /* We've logged the csums on this ordered
 				       ordered extent */
+#define BTRFS_ORDERED_TRUNCATED 9 /* Set when we have to truncate an extent */
 
 struct btrfs_ordered_extent {
 	/* logical offset in the file */
@@ -108,6 +97,12 @@ struct btrfs_ordered_extent {
 	 */
 	u64 outstanding_isize;
 
+	/*
+	 * If we get truncated we need to adjust the file extent we enter for
+	 * this ordered extent so that we do not expose stale data.
+	 */
+	u64 truncated_len;
+
 	/* flags (described above) */
 	unsigned long flags;
 
@@ -149,11 +144,8 @@ struct btrfs_ordered_extent {
 static inline int btrfs_ordered_sum_size(struct btrfs_root *root,
 					 unsigned long bytes)
 {
-	unsigned long num_sectors = (bytes + root->sectorsize - 1) /
-		root->sectorsize;
-	num_sectors++;
-	return sizeof(struct btrfs_ordered_sum) +
-		num_sectors * sizeof(struct btrfs_sector_sum);
+	int num_sectors = (int)DIV_ROUND_UP(bytes, root->sectorsize);
+	return sizeof(struct btrfs_ordered_sum) + num_sectors * sizeof(u32);
 }
 
 static inline void
@@ -203,7 +195,8 @@ int btrfs_run_ordered_operations(struct btrfs_trans_handle *trans,
 void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
 				 struct btrfs_root *root,
 				 struct inode *inode);
-void btrfs_wait_ordered_extents(struct btrfs_root *root, int delay_iput);
+void btrfs_wait_ordered_extents(struct btrfs_root *root);
+void btrfs_wait_all_ordered_extents(struct btrfs_fs_info *fs_info);
 void btrfs_get_logged_extents(struct btrfs_root *log, struct inode *inode);
 void btrfs_wait_logged_extents(struct btrfs_root *log, u64 transid);
 void btrfs_free_logged_extents(struct btrfs_root *log, u64 transid);
diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c
index dc0024f17c1f..0088bedc8631 100644
--- a/fs/btrfs/print-tree.c
+++ b/fs/btrfs/print-tree.c
@@ -26,14 +26,12 @@ static void print_chunk(struct extent_buffer *eb, struct btrfs_chunk *chunk)
 	int i;
 	printk(KERN_INFO "\t\tchunk length %llu owner %llu type %llu "
 	       "num_stripes %d\n",
-	       (unsigned long long)btrfs_chunk_length(eb, chunk),
-	       (unsigned long long)btrfs_chunk_owner(eb, chunk),
-	       (unsigned long long)btrfs_chunk_type(eb, chunk),
-	       num_stripes);
+	       btrfs_chunk_length(eb, chunk), btrfs_chunk_owner(eb, chunk),
+	       btrfs_chunk_type(eb, chunk), num_stripes);
 	for (i = 0 ; i < num_stripes ; i++) {
 		printk(KERN_INFO "\t\t\tstripe %d devid %llu offset %llu\n", i,
-		      (unsigned long long)btrfs_stripe_devid_nr(eb, chunk, i),
-		      (unsigned long long)btrfs_stripe_offset_nr(eb, chunk, i));
+		      btrfs_stripe_devid_nr(eb, chunk, i),
+		      btrfs_stripe_offset_nr(eb, chunk, i));
 	}
 }
 static void print_dev_item(struct extent_buffer *eb,
@@ -41,18 +39,18 @@ static void print_dev_item(struct extent_buffer *eb,
 {
 	printk(KERN_INFO "\t\tdev item devid %llu "
 	       "total_bytes %llu bytes used %llu\n",
-	       (unsigned long long)btrfs_device_id(eb, dev_item),
-	       (unsigned long long)btrfs_device_total_bytes(eb, dev_item),
-	       (unsigned long long)btrfs_device_bytes_used(eb, dev_item));
+	       btrfs_device_id(eb, dev_item),
+	       btrfs_device_total_bytes(eb, dev_item),
+	       btrfs_device_bytes_used(eb, dev_item));
 }
 static void print_extent_data_ref(struct extent_buffer *eb,
 				  struct btrfs_extent_data_ref *ref)
 {
 	printk(KERN_INFO "\t\textent data backref root %llu "
 	       "objectid %llu offset %llu count %u\n",
-	       (unsigned long long)btrfs_extent_data_ref_root(eb, ref),
-	       (unsigned long long)btrfs_extent_data_ref_objectid(eb, ref),
-	       (unsigned long long)btrfs_extent_data_ref_offset(eb, ref),
+	       btrfs_extent_data_ref_root(eb, ref),
+	       btrfs_extent_data_ref_objectid(eb, ref),
+	       btrfs_extent_data_ref_offset(eb, ref),
 	       btrfs_extent_data_ref_count(eb, ref));
 }
 
@@ -87,19 +85,17 @@ static void print_extent_item(struct extent_buffer *eb, int slot)
 	flags = btrfs_extent_flags(eb, ei);
 
 	printk(KERN_INFO "\t\textent refs %llu gen %llu flags %llu\n",
-	       (unsigned long long)btrfs_extent_refs(eb, ei),
-	       (unsigned long long)btrfs_extent_generation(eb, ei),
-	       (unsigned long long)flags);
+	       btrfs_extent_refs(eb, ei), btrfs_extent_generation(eb, ei),
+	       flags);
 
 	if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
 		struct btrfs_tree_block_info *info;
 		info = (struct btrfs_tree_block_info *)(ei + 1);
 		btrfs_tree_block_key(eb, info, &key);
-		printk(KERN_INFO "\t\ttree block key (%llu %x %llu) "
+		printk(KERN_INFO "\t\ttree block key (%llu %u %llu) "
 		       "level %d\n",
-		       (unsigned long long)btrfs_disk_key_objectid(&key),
-		       key.type,
-		       (unsigned long long)btrfs_disk_key_offset(&key),
+		       btrfs_disk_key_objectid(&key), key.type,
+		       btrfs_disk_key_offset(&key),
 		       btrfs_tree_block_level(eb, info));
 		iref = (struct btrfs_extent_inline_ref *)(info + 1);
 	} else {
@@ -115,11 +111,11 @@ static void print_extent_item(struct extent_buffer *eb, int slot)
 		switch (type) {
 		case BTRFS_TREE_BLOCK_REF_KEY:
 			printk(KERN_INFO "\t\ttree block backref "
-				"root %llu\n", (unsigned long long)offset);
+				"root %llu\n", offset);
 			break;
 		case BTRFS_SHARED_BLOCK_REF_KEY:
 			printk(KERN_INFO "\t\tshared block backref "
-				"parent %llu\n", (unsigned long long)offset);
+				"parent %llu\n", offset);
 			break;
 		case BTRFS_EXTENT_DATA_REF_KEY:
 			dref = (struct btrfs_extent_data_ref *)(&iref->offset);
@@ -129,8 +125,7 @@ static void print_extent_item(struct extent_buffer *eb, int slot)
 			sref = (struct btrfs_shared_data_ref *)(iref + 1);
 			printk(KERN_INFO "\t\tshared data backref "
 			       "parent %llu count %u\n",
-			       (unsigned long long)offset,
-			       btrfs_shared_data_ref_count(eb, sref));
+			       offset, btrfs_shared_data_ref_count(eb, sref));
 			break;
 		default:
 			BUG();
@@ -148,13 +143,32 @@ static void print_extent_ref_v0(struct extent_buffer *eb, int slot)
 	ref0 = btrfs_item_ptr(eb, slot, struct btrfs_extent_ref_v0);
 	printk("\t\textent back ref root %llu gen %llu "
 		"owner %llu num_refs %lu\n",
-		(unsigned long long)btrfs_ref_root_v0(eb, ref0),
-		(unsigned long long)btrfs_ref_generation_v0(eb, ref0),
-		(unsigned long long)btrfs_ref_objectid_v0(eb, ref0),
+		btrfs_ref_root_v0(eb, ref0),
+		btrfs_ref_generation_v0(eb, ref0),
+		btrfs_ref_objectid_v0(eb, ref0),
 		(unsigned long)btrfs_ref_count_v0(eb, ref0));
 }
 #endif
 
+static void print_uuid_item(struct extent_buffer *l, unsigned long offset,
+			    u32 item_size)
+{
+	if (!IS_ALIGNED(item_size, sizeof(u64))) {
+		pr_warn("btrfs: uuid item with illegal size %lu!\n",
+			(unsigned long)item_size);
+		return;
+	}
+	while (item_size) {
+		__le64 subvol_id;
+
+		read_extent_buffer(l, &subvol_id, offset, sizeof(subvol_id));
+		printk(KERN_INFO "\t\tsubvol_id %llu\n",
+		       (unsigned long long)le64_to_cpu(subvol_id));
+		item_size -= sizeof(u64);
+		offset += sizeof(u64);
+	}
+}
+
 void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 {
 	int i;
@@ -177,39 +191,34 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 	nr = btrfs_header_nritems(l);
 
 	btrfs_info(root->fs_info, "leaf %llu total ptrs %d free space %d",
-		(unsigned long long)btrfs_header_bytenr(l), nr,
-		btrfs_leaf_free_space(root, l));
+		   btrfs_header_bytenr(l), nr, btrfs_leaf_free_space(root, l));
 	for (i = 0 ; i < nr ; i++) {
 		item = btrfs_item_nr(l, i);
 		btrfs_item_key_to_cpu(l, &key, i);
 		type = btrfs_key_type(&key);
-		printk(KERN_INFO "\titem %d key (%llu %x %llu) itemoff %d "
+		printk(KERN_INFO "\titem %d key (%llu %u %llu) itemoff %d "
 		       "itemsize %d\n",
-			i,
-			(unsigned long long)key.objectid, type,
-			(unsigned long long)key.offset,
+			i, key.objectid, type, key.offset,
 			btrfs_item_offset(l, item), btrfs_item_size(l, item));
 		switch (type) {
 		case BTRFS_INODE_ITEM_KEY:
 			ii = btrfs_item_ptr(l, i, struct btrfs_inode_item);
 			printk(KERN_INFO "\t\tinode generation %llu size %llu "
 			       "mode %o\n",
-			       (unsigned long long)
 			       btrfs_inode_generation(l, ii),
-			      (unsigned long long)btrfs_inode_size(l, ii),
+			       btrfs_inode_size(l, ii),
 			       btrfs_inode_mode(l, ii));
 			break;
 		case BTRFS_DIR_ITEM_KEY:
 			di = btrfs_item_ptr(l, i, struct btrfs_dir_item);
 			btrfs_dir_item_key_to_cpu(l, di, &found_key);
 			printk(KERN_INFO "\t\tdir oid %llu type %u\n",
-				(unsigned long long)found_key.objectid,
+				found_key.objectid,
 				btrfs_dir_type(l, di));
 			break;
 		case BTRFS_ROOT_ITEM_KEY:
 			ri = btrfs_item_ptr(l, i, struct btrfs_root_item);
 			printk(KERN_INFO "\t\troot data bytenr %llu refs %u\n",
-				(unsigned long long)
 				btrfs_disk_root_bytenr(l, ri),
 				btrfs_disk_root_refs(l, ri));
 			break;
@@ -245,17 +254,12 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 			}
 			printk(KERN_INFO "\t\textent data disk bytenr %llu "
 			       "nr %llu\n",
-			       (unsigned long long)
 			       btrfs_file_extent_disk_bytenr(l, fi),
-			       (unsigned long long)
 			       btrfs_file_extent_disk_num_bytes(l, fi));
 			printk(KERN_INFO "\t\textent data offset %llu "
 			       "nr %llu ram %llu\n",
-			       (unsigned long long)
 			       btrfs_file_extent_offset(l, fi),
-			       (unsigned long long)
 			       btrfs_file_extent_num_bytes(l, fi),
-			       (unsigned long long)
 			       btrfs_file_extent_ram_bytes(l, fi));
 			break;
 		case BTRFS_EXTENT_REF_V0_KEY:
@@ -269,7 +273,6 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 			bi = btrfs_item_ptr(l, i,
 					    struct btrfs_block_group_item);
 			printk(KERN_INFO "\t\tblock group used %llu\n",
-			       (unsigned long long)
 			       btrfs_disk_block_group_used(l, bi));
 			break;
 		case BTRFS_CHUNK_ITEM_KEY:
@@ -286,13 +289,9 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 			printk(KERN_INFO "\t\tdev extent chunk_tree %llu\n"
 			       "\t\tchunk objectid %llu chunk offset %llu "
 			       "length %llu\n",
-			       (unsigned long long)
 			       btrfs_dev_extent_chunk_tree(l, dev_extent),
-			       (unsigned long long)
 			       btrfs_dev_extent_chunk_objectid(l, dev_extent),
-			       (unsigned long long)
 			       btrfs_dev_extent_chunk_offset(l, dev_extent),
-			       (unsigned long long)
 			       btrfs_dev_extent_length(l, dev_extent));
 			break;
 		case BTRFS_DEV_STATS_KEY:
@@ -301,6 +300,11 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 		case BTRFS_DEV_REPLACE_KEY:
 			printk(KERN_INFO "\t\tdev replace\n");
 			break;
+		case BTRFS_UUID_KEY_SUBVOL:
+		case BTRFS_UUID_KEY_RECEIVED_SUBVOL:
+			print_uuid_item(l, btrfs_item_ptr_offset(l, i),
+					btrfs_item_size_nr(l, i));
+			break;
 		};
 	}
 }
@@ -320,16 +324,13 @@ void btrfs_print_tree(struct btrfs_root *root, struct extent_buffer *c)
 		return;
 	}
 	btrfs_info(root->fs_info, "node %llu level %d total ptrs %d free spc %u",
-		(unsigned long long)btrfs_header_bytenr(c),
-		level, nr, (u32)BTRFS_NODEPTRS_PER_BLOCK(root) - nr);
+		btrfs_header_bytenr(c), level, nr,
+		(u32)BTRFS_NODEPTRS_PER_BLOCK(root) - nr);
 	for (i = 0; i < nr; i++) {
 		btrfs_node_key_to_cpu(c, &key, i);
 		printk(KERN_INFO "\tkey %d (%llu %u %llu) block %llu\n",
-		       i,
-		       (unsigned long long)key.objectid,
-		       key.type,
-		       (unsigned long long)key.offset,
-		       (unsigned long long)btrfs_node_blockptr(c, i));
+		       i, key.objectid, key.type, key.offset,
+		       btrfs_node_blockptr(c, i));
 	}
 	for (i = 0; i < nr; i++) {
 		struct extent_buffer *next = read_tree_block(root,
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index 9d49c586995a..4e6ef490619e 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -98,13 +98,10 @@ struct btrfs_qgroup_list {
 	struct btrfs_qgroup *member;
 };
 
-struct qgroup_rescan {
-	struct btrfs_work	work;
-	struct btrfs_fs_info	*fs_info;
-};
-
-static void qgroup_rescan_start(struct btrfs_fs_info *fs_info,
-				struct qgroup_rescan *qscan);
+static int
+qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
+		   int init_flags);
+static void qgroup_rescan_zero_tracking(struct btrfs_fs_info *fs_info);
 
 /* must be called with qgroup_ioctl_lock held */
 static struct btrfs_qgroup *find_qgroup_rb(struct btrfs_fs_info *fs_info,
@@ -160,18 +157,11 @@ static struct btrfs_qgroup *add_qgroup_rb(struct btrfs_fs_info *fs_info,
 	return qgroup;
 }
 
-/* must be called with qgroup_lock held */
-static int del_qgroup_rb(struct btrfs_fs_info *fs_info, u64 qgroupid)
+static void __del_qgroup_rb(struct btrfs_qgroup *qgroup)
 {
-	struct btrfs_qgroup *qgroup = find_qgroup_rb(fs_info, qgroupid);
 	struct btrfs_qgroup_list *list;
 
-	if (!qgroup)
-		return -ENOENT;
-
-	rb_erase(&qgroup->node, &fs_info->qgroup_tree);
 	list_del(&qgroup->dirty);
-
 	while (!list_empty(&qgroup->groups)) {
 		list = list_first_entry(&qgroup->groups,
 					struct btrfs_qgroup_list, next_group);
@@ -188,7 +178,18 @@ static int del_qgroup_rb(struct btrfs_fs_info *fs_info, u64 qgroupid)
 		kfree(list);
 	}
 	kfree(qgroup);
+}
+
+/* must be called with qgroup_lock held */
+static int del_qgroup_rb(struct btrfs_fs_info *fs_info, u64 qgroupid)
+{
+	struct btrfs_qgroup *qgroup = find_qgroup_rb(fs_info, qgroupid);
+
+	if (!qgroup)
+		return -ENOENT;
 
+	rb_erase(&qgroup->node, &fs_info->qgroup_tree);
+	__del_qgroup_rb(qgroup);
 	return 0;
 }
 
@@ -255,10 +256,17 @@ int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info)
 	int slot;
 	int ret = 0;
 	u64 flags = 0;
+	u64 rescan_progress = 0;
 
 	if (!fs_info->quota_enabled)
 		return 0;
 
+	fs_info->qgroup_ulist = ulist_alloc(GFP_NOFS);
+	if (!fs_info->qgroup_ulist) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
 	path = btrfs_alloc_path();
 	if (!path) {
 		ret = -ENOMEM;
@@ -306,20 +314,7 @@ int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info)
 			}
 			fs_info->qgroup_flags = btrfs_qgroup_status_flags(l,
 									  ptr);
-			fs_info->qgroup_rescan_progress.objectid =
-					btrfs_qgroup_status_rescan(l, ptr);
-			if (fs_info->qgroup_flags &
-			    BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
-				struct qgroup_rescan *qscan =
-					kmalloc(sizeof(*qscan), GFP_NOFS);
-				if (!qscan) {
-					ret = -ENOMEM;
-					goto out;
-				}
-				fs_info->qgroup_rescan_progress.type = 0;
-				fs_info->qgroup_rescan_progress.offset = 0;
-				qgroup_rescan_start(fs_info, qscan);
-			}
+			rescan_progress = btrfs_qgroup_status_rescan(l, ptr);
 			goto next1;
 		}
 
@@ -403,8 +398,7 @@ next1:
 		if (ret == -ENOENT) {
 			printk(KERN_WARNING
 				"btrfs: orphan qgroup relation 0x%llx->0x%llx\n",
-				(unsigned long long)found_key.objectid,
-				(unsigned long long)found_key.offset);
+				found_key.objectid, found_key.offset);
 			ret = 0;	/* ignore the error */
 		}
 		if (ret)
@@ -421,45 +415,44 @@ out:
 	if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON)) {
 		fs_info->quota_enabled = 0;
 		fs_info->pending_quota_state = 0;
+	} else if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN &&
+		   ret >= 0) {
+		ret = qgroup_rescan_init(fs_info, rescan_progress, 0);
 	}
 	btrfs_free_path(path);
 
+	if (ret < 0) {
+		ulist_free(fs_info->qgroup_ulist);
+		fs_info->qgroup_ulist = NULL;
+		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
+	}
+
 	return ret < 0 ? ret : 0;
 }
 
 /*
- * This is only called from close_ctree() or open_ctree(), both in single-
- * treaded paths. Clean up the in-memory structures. No locking needed.
+ * This is called from close_ctree() or open_ctree() or btrfs_quota_disable(),
+ * first two are in single-threaded paths.And for the third one, we have set
+ * quota_root to be null with qgroup_lock held before, so it is safe to clean
+ * up the in-memory structures without qgroup_lock held.
  */
 void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info)
 {
 	struct rb_node *n;
 	struct btrfs_qgroup *qgroup;
-	struct btrfs_qgroup_list *list;
 
 	while ((n = rb_first(&fs_info->qgroup_tree))) {
 		qgroup = rb_entry(n, struct btrfs_qgroup, node);
 		rb_erase(n, &fs_info->qgroup_tree);
-
-		while (!list_empty(&qgroup->groups)) {
-			list = list_first_entry(&qgroup->groups,
-						struct btrfs_qgroup_list,
-						next_group);
-			list_del(&list->next_group);
-			list_del(&list->next_member);
-			kfree(list);
-		}
-
-		while (!list_empty(&qgroup->members)) {
-			list = list_first_entry(&qgroup->members,
-						struct btrfs_qgroup_list,
-						next_member);
-			list_del(&list->next_group);
-			list_del(&list->next_member);
-			kfree(list);
-		}
-		kfree(qgroup);
+		__del_qgroup_rb(qgroup);
 	}
+	/*
+	 * we call btrfs_free_qgroup_config() when umounting
+	 * filesystem and disabling quota, so we set qgroup_ulit
+	 * to be null here to avoid double free.
+	 */
+	ulist_free(fs_info->qgroup_ulist);
+	fs_info->qgroup_ulist = NULL;
 }
 
 static int add_qgroup_relation_item(struct btrfs_trans_handle *trans,
@@ -819,6 +812,12 @@ int btrfs_quota_enable(struct btrfs_trans_handle *trans,
 		goto out;
 	}
 
+	fs_info->qgroup_ulist = ulist_alloc(GFP_NOFS);
+	if (!fs_info->qgroup_ulist) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
 	/*
 	 * initially create the quota tree
 	 */
@@ -916,6 +915,10 @@ out_free_root:
 		kfree(quota_root);
 	}
 out:
+	if (ret) {
+		ulist_free(fs_info->qgroup_ulist);
+		fs_info->qgroup_ulist = NULL;
+	}
 	mutex_unlock(&fs_info->qgroup_ioctl_lock);
 	return ret;
 }
@@ -935,13 +938,9 @@ int btrfs_quota_disable(struct btrfs_trans_handle *trans,
 	fs_info->pending_quota_state = 0;
 	quota_root = fs_info->quota_root;
 	fs_info->quota_root = NULL;
-	btrfs_free_qgroup_config(fs_info);
 	spin_unlock(&fs_info->qgroup_lock);
 
-	if (!quota_root) {
-		ret = -EINVAL;
-		goto out;
-	}
+	btrfs_free_qgroup_config(fs_info);
 
 	ret = btrfs_clean_quota_tree(trans, quota_root);
 	if (ret)
@@ -1163,7 +1162,7 @@ int btrfs_limit_qgroup(struct btrfs_trans_handle *trans,
 	if (ret) {
 		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
 		printk(KERN_INFO "unable to update quota limit for %llu\n",
-		       (unsigned long long)qgroupid);
+		       qgroupid);
 	}
 
 	spin_lock(&fs_info->qgroup_lock);
@@ -1355,7 +1354,6 @@ int btrfs_qgroup_account_ref(struct btrfs_trans_handle *trans,
 	u64 ref_root;
 	struct btrfs_qgroup *qgroup;
 	struct ulist *roots = NULL;
-	struct ulist *tmp = NULL;
 	u64 seq;
 	int ret = 0;
 	int sgn;
@@ -1428,14 +1426,7 @@ int btrfs_qgroup_account_ref(struct btrfs_trans_handle *trans,
 	if (ret < 0)
 		return ret;
 
-	mutex_lock(&fs_info->qgroup_rescan_lock);
 	spin_lock(&fs_info->qgroup_lock);
-	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
-		if (fs_info->qgroup_rescan_progress.objectid <= node->bytenr) {
-			ret = 0;
-			goto unlock;
-		}
-	}
 
 	quota_root = fs_info->quota_root;
 	if (!quota_root)
@@ -1448,39 +1439,34 @@ int btrfs_qgroup_account_ref(struct btrfs_trans_handle *trans,
 	/*
 	 * step 1: for each old ref, visit all nodes once and inc refcnt
 	 */
-	tmp = ulist_alloc(GFP_ATOMIC);
-	if (!tmp) {
-		ret = -ENOMEM;
-		goto unlock;
-	}
+	ulist_reinit(fs_info->qgroup_ulist);
 	seq = fs_info->qgroup_seq;
 	fs_info->qgroup_seq += roots->nnodes + 1; /* max refcnt */
 
-	ret = qgroup_account_ref_step1(fs_info, roots, tmp, seq);
+	ret = qgroup_account_ref_step1(fs_info, roots, fs_info->qgroup_ulist,
+				       seq);
 	if (ret)
 		goto unlock;
 
 	/*
 	 * step 2: walk from the new root
 	 */
-	ret = qgroup_account_ref_step2(fs_info, roots, tmp, seq, sgn,
-				       node->num_bytes, qgroup);
+	ret = qgroup_account_ref_step2(fs_info, roots, fs_info->qgroup_ulist,
+				       seq, sgn, node->num_bytes, qgroup);
 	if (ret)
 		goto unlock;
 
 	/*
 	 * step 3: walk again from old refs
 	 */
-	ret = qgroup_account_ref_step3(fs_info, roots, tmp, seq, sgn,
-				       node->num_bytes);
+	ret = qgroup_account_ref_step3(fs_info, roots, fs_info->qgroup_ulist,
+				       seq, sgn, node->num_bytes);
 	if (ret)
 		goto unlock;
 
 unlock:
 	spin_unlock(&fs_info->qgroup_lock);
-	mutex_unlock(&fs_info->qgroup_rescan_lock);
 	ulist_free(roots);
-	ulist_free(tmp);
 
 	return ret;
 }
@@ -1527,9 +1513,12 @@ int btrfs_run_qgroups(struct btrfs_trans_handle *trans,
 		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
 
 	if (!ret && start_rescan_worker) {
-		ret = btrfs_qgroup_rescan(fs_info);
-		if (ret)
-			pr_err("btrfs: start rescan quota failed: %d\n", ret);
+		ret = qgroup_rescan_init(fs_info, 0, 1);
+		if (!ret) {
+			qgroup_rescan_zero_tracking(fs_info);
+			btrfs_queue_worker(&fs_info->qgroup_rescan_workers,
+					   &fs_info->qgroup_rescan_work);
+		}
 		ret = 0;
 	}
 
@@ -1720,7 +1709,6 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	u64 ref_root = root->root_key.objectid;
 	int ret = 0;
-	struct ulist *ulist = NULL;
 	struct ulist_node *unode;
 	struct ulist_iterator uiter;
 
@@ -1743,17 +1731,13 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
 	 * in a first step, we check all affected qgroups if any limits would
 	 * be exceeded
 	 */
-	ulist = ulist_alloc(GFP_ATOMIC);
-	if (!ulist) {
-		ret = -ENOMEM;
-		goto out;
-	}
-	ret = ulist_add(ulist, qgroup->qgroupid,
+	ulist_reinit(fs_info->qgroup_ulist);
+	ret = ulist_add(fs_info->qgroup_ulist, qgroup->qgroupid,
 			(uintptr_t)qgroup, GFP_ATOMIC);
 	if (ret < 0)
 		goto out;
 	ULIST_ITER_INIT(&uiter);
-	while ((unode = ulist_next(ulist, &uiter))) {
+	while ((unode = ulist_next(fs_info->qgroup_ulist, &uiter))) {
 		struct btrfs_qgroup *qg;
 		struct btrfs_qgroup_list *glist;
 
@@ -1774,7 +1758,8 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
 		}
 
 		list_for_each_entry(glist, &qg->groups, next_group) {
-			ret = ulist_add(ulist, glist->group->qgroupid,
+			ret = ulist_add(fs_info->qgroup_ulist,
+					glist->group->qgroupid,
 					(uintptr_t)glist->group, GFP_ATOMIC);
 			if (ret < 0)
 				goto out;
@@ -1785,7 +1770,7 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
 	 * no limits exceeded, now record the reservation into all qgroups
 	 */
 	ULIST_ITER_INIT(&uiter);
-	while ((unode = ulist_next(ulist, &uiter))) {
+	while ((unode = ulist_next(fs_info->qgroup_ulist, &uiter))) {
 		struct btrfs_qgroup *qg;
 
 		qg = (struct btrfs_qgroup *)(uintptr_t)unode->aux;
@@ -1795,8 +1780,6 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
 
 out:
 	spin_unlock(&fs_info->qgroup_lock);
-	ulist_free(ulist);
-
 	return ret;
 }
 
@@ -1805,7 +1788,6 @@ void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes)
 	struct btrfs_root *quota_root;
 	struct btrfs_qgroup *qgroup;
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct ulist *ulist = NULL;
 	struct ulist_node *unode;
 	struct ulist_iterator uiter;
 	u64 ref_root = root->root_key.objectid;
@@ -1827,17 +1809,13 @@ void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes)
 	if (!qgroup)
 		goto out;
 
-	ulist = ulist_alloc(GFP_ATOMIC);
-	if (!ulist) {
-		btrfs_std_error(fs_info, -ENOMEM);
-		goto out;
-	}
-	ret = ulist_add(ulist, qgroup->qgroupid,
+	ulist_reinit(fs_info->qgroup_ulist);
+	ret = ulist_add(fs_info->qgroup_ulist, qgroup->qgroupid,
 			(uintptr_t)qgroup, GFP_ATOMIC);
 	if (ret < 0)
 		goto out;
 	ULIST_ITER_INIT(&uiter);
-	while ((unode = ulist_next(ulist, &uiter))) {
+	while ((unode = ulist_next(fs_info->qgroup_ulist, &uiter))) {
 		struct btrfs_qgroup *qg;
 		struct btrfs_qgroup_list *glist;
 
@@ -1846,7 +1824,8 @@ void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes)
 		qg->reserved -= num_bytes;
 
 		list_for_each_entry(glist, &qg->groups, next_group) {
-			ret = ulist_add(ulist, glist->group->qgroupid,
+			ret = ulist_add(fs_info->qgroup_ulist,
+					glist->group->qgroupid,
 					(uintptr_t)glist->group, GFP_ATOMIC);
 			if (ret < 0)
 				goto out;
@@ -1855,7 +1834,6 @@ void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes)
 
 out:
 	spin_unlock(&fs_info->qgroup_lock);
-	ulist_free(ulist);
 }
 
 void assert_qgroups_uptodate(struct btrfs_trans_handle *trans)
@@ -1874,12 +1852,11 @@ void assert_qgroups_uptodate(struct btrfs_trans_handle *trans)
  * returns 1 when done, 2 when done and FLAG_INCONSISTENT was cleared.
  */
 static int
-qgroup_rescan_leaf(struct qgroup_rescan *qscan, struct btrfs_path *path,
+qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
 		   struct btrfs_trans_handle *trans, struct ulist *tmp,
 		   struct extent_buffer *scratch_leaf)
 {
 	struct btrfs_key found;
-	struct btrfs_fs_info *fs_info = qscan->fs_info;
 	struct ulist *roots = NULL;
 	struct ulist_node *unode;
 	struct ulist_iterator uiter;
@@ -1895,10 +1872,9 @@ qgroup_rescan_leaf(struct qgroup_rescan *qscan, struct btrfs_path *path,
 					 path, 1, 0);
 
 	pr_debug("current progress key (%llu %u %llu), search_slot ret %d\n",
-		 (unsigned long long)fs_info->qgroup_rescan_progress.objectid,
+		 fs_info->qgroup_rescan_progress.objectid,
 		 fs_info->qgroup_rescan_progress.type,
-		 (unsigned long long)fs_info->qgroup_rescan_progress.offset,
-		 ret);
+		 fs_info->qgroup_rescan_progress.offset, ret);
 
 	if (ret) {
 		/*
@@ -2007,11 +1983,10 @@ out:
 
 static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 {
-	struct qgroup_rescan *qscan = container_of(work, struct qgroup_rescan,
-						   work);
+	struct btrfs_fs_info *fs_info = container_of(work, struct btrfs_fs_info,
+						     qgroup_rescan_work);
 	struct btrfs_path *path;
 	struct btrfs_trans_handle *trans = NULL;
-	struct btrfs_fs_info *fs_info = qscan->fs_info;
 	struct ulist *tmp = NULL;
 	struct extent_buffer *scratch_leaf = NULL;
 	int err = -ENOMEM;
@@ -2036,7 +2011,7 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 		if (!fs_info->quota_enabled) {
 			err = -EINTR;
 		} else {
-			err = qgroup_rescan_leaf(qscan, path, trans,
+			err = qgroup_rescan_leaf(fs_info, path, trans,
 						 tmp, scratch_leaf);
 		}
 		if (err > 0)
@@ -2049,7 +2024,6 @@ out:
 	kfree(scratch_leaf);
 	ulist_free(tmp);
 	btrfs_free_path(path);
-	kfree(qscan);
 
 	mutex_lock(&fs_info->qgroup_rescan_lock);
 	fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
@@ -2068,47 +2042,74 @@ out:
 	} else {
 		pr_err("btrfs: qgroup scan failed with %d\n", err);
 	}
-}
-
-static void
-qgroup_rescan_start(struct btrfs_fs_info *fs_info, struct qgroup_rescan *qscan)
-{
-	memset(&qscan->work, 0, sizeof(qscan->work));
-	qscan->work.func = btrfs_qgroup_rescan_worker;
-	qscan->fs_info = fs_info;
 
-	pr_info("btrfs: qgroup scan started\n");
-	btrfs_queue_worker(&fs_info->qgroup_rescan_workers, &qscan->work);
+	complete_all(&fs_info->qgroup_rescan_completion);
 }
 
-int
-btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info)
+/*
+ * Checks that (a) no rescan is running and (b) quota is enabled. Allocates all
+ * memory required for the rescan context.
+ */
+static int
+qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
+		   int init_flags)
 {
 	int ret = 0;
-	struct rb_node *n;
-	struct btrfs_qgroup *qgroup;
-	struct qgroup_rescan *qscan = kmalloc(sizeof(*qscan), GFP_NOFS);
 
-	if (!qscan)
-		return -ENOMEM;
+	if (!init_flags &&
+	    (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) ||
+	     !(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON))) {
+		ret = -EINVAL;
+		goto err;
+	}
 
 	mutex_lock(&fs_info->qgroup_rescan_lock);
 	spin_lock(&fs_info->qgroup_lock);
-	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN)
-		ret = -EINPROGRESS;
-	else if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON))
-		ret = -EINVAL;
-	if (ret) {
-		spin_unlock(&fs_info->qgroup_lock);
-		mutex_unlock(&fs_info->qgroup_rescan_lock);
-		kfree(qscan);
-		return ret;
+
+	if (init_flags) {
+		if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN)
+			ret = -EINPROGRESS;
+		else if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON))
+			ret = -EINVAL;
+
+		if (ret) {
+			spin_unlock(&fs_info->qgroup_lock);
+			mutex_unlock(&fs_info->qgroup_rescan_lock);
+			goto err;
+		}
+
+		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_RESCAN;
 	}
 
-	fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_RESCAN;
 	memset(&fs_info->qgroup_rescan_progress, 0,
 		sizeof(fs_info->qgroup_rescan_progress));
+	fs_info->qgroup_rescan_progress.objectid = progress_objectid;
+
+	spin_unlock(&fs_info->qgroup_lock);
+	mutex_unlock(&fs_info->qgroup_rescan_lock);
+
+	init_completion(&fs_info->qgroup_rescan_completion);
+
+	memset(&fs_info->qgroup_rescan_work, 0,
+	       sizeof(fs_info->qgroup_rescan_work));
+	fs_info->qgroup_rescan_work.func = btrfs_qgroup_rescan_worker;
+
+	if (ret) {
+err:
+		pr_info("btrfs: qgroup_rescan_init failed with %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
 
+static void
+qgroup_rescan_zero_tracking(struct btrfs_fs_info *fs_info)
+{
+	struct rb_node *n;
+	struct btrfs_qgroup *qgroup;
+
+	spin_lock(&fs_info->qgroup_lock);
 	/* clear all current qgroup tracking information */
 	for (n = rb_first(&fs_info->qgroup_tree); n; n = rb_next(n)) {
 		qgroup = rb_entry(n, struct btrfs_qgroup, node);
@@ -2118,9 +2119,74 @@ btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info)
 		qgroup->excl_cmpr = 0;
 	}
 	spin_unlock(&fs_info->qgroup_lock);
-	mutex_unlock(&fs_info->qgroup_rescan_lock);
+}
+
+int
+btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info)
+{
+	int ret = 0;
+	struct btrfs_trans_handle *trans;
+
+	ret = qgroup_rescan_init(fs_info, 0, 1);
+	if (ret)
+		return ret;
+
+	/*
+	 * We have set the rescan_progress to 0, which means no more
+	 * delayed refs will be accounted by btrfs_qgroup_account_ref.
+	 * However, btrfs_qgroup_account_ref may be right after its call
+	 * to btrfs_find_all_roots, in which case it would still do the
+	 * accounting.
+	 * To solve this, we're committing the transaction, which will
+	 * ensure we run all delayed refs and only after that, we are
+	 * going to clear all tracking information for a clean start.
+	 */
+
+	trans = btrfs_join_transaction(fs_info->fs_root);
+	if (IS_ERR(trans)) {
+		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
+		return PTR_ERR(trans);
+	}
+	ret = btrfs_commit_transaction(trans, fs_info->fs_root);
+	if (ret) {
+		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
+		return ret;
+	}
 
-	qgroup_rescan_start(fs_info, qscan);
+	qgroup_rescan_zero_tracking(fs_info);
+
+	btrfs_queue_worker(&fs_info->qgroup_rescan_workers,
+			   &fs_info->qgroup_rescan_work);
 
 	return 0;
 }
+
+int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info)
+{
+	int running;
+	int ret = 0;
+
+	mutex_lock(&fs_info->qgroup_rescan_lock);
+	spin_lock(&fs_info->qgroup_lock);
+	running = fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN;
+	spin_unlock(&fs_info->qgroup_lock);
+	mutex_unlock(&fs_info->qgroup_rescan_lock);
+
+	if (running)
+		ret = wait_for_completion_interruptible(
+					&fs_info->qgroup_rescan_completion);
+
+	return ret;
+}
+
+/*
+ * this is only called from open_ctree where we're still single threaded, thus
+ * locking is omitted here.
+ */
+void
+btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info)
+{
+	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN)
+		btrfs_queue_worker(&fs_info->qgroup_rescan_workers,
+				   &fs_info->qgroup_rescan_work);
+}
diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index 0525e1389f5b..d0ecfbd9cc9f 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -1540,8 +1540,10 @@ static int full_stripe_write(struct btrfs_raid_bio *rbio)
 	int ret;
 
 	ret = alloc_rbio_parity_pages(rbio);
-	if (ret)
+	if (ret) {
+		__free_raid_bio(rbio);
 		return ret;
+	}
 
 	ret = lock_stripe_add(rbio);
 	if (ret == 0)
@@ -1687,11 +1689,8 @@ int raid56_parity_write(struct btrfs_root *root, struct bio *bio,
 	struct blk_plug_cb *cb;
 
 	rbio = alloc_rbio(root, bbio, raid_map, stripe_len);
-	if (IS_ERR(rbio)) {
-		kfree(raid_map);
-		kfree(bbio);
+	if (IS_ERR(rbio))
 		return PTR_ERR(rbio);
-	}
 	bio_list_add(&rbio->bio_list, bio);
 	rbio->bio_list_bytes = bio->bi_size;
 
@@ -2041,9 +2040,8 @@ int raid56_parity_recover(struct btrfs_root *root, struct bio *bio,
 	int ret;
 
 	rbio = alloc_rbio(root, bbio, raid_map, stripe_len);
-	if (IS_ERR(rbio)) {
+	if (IS_ERR(rbio))
 		return PTR_ERR(rbio);
-	}
 
 	rbio->read_rebuild = 1;
 	bio_list_add(&rbio->bio_list, bio);
@@ -2052,6 +2050,8 @@ int raid56_parity_recover(struct btrfs_root *root, struct bio *bio,
 	rbio->faila = find_logical_bio_stripe(rbio, bio);
 	if (rbio->faila == -1) {
 		BUG();
+		kfree(raid_map);
+		kfree(bbio);
 		kfree(rbio);
 		return -EIO;
 	}
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 395b82031a42..a5a26320503f 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -335,7 +335,7 @@ static void backref_tree_panic(struct rb_node *rb_node, int errno, u64 bytenr)
 	if (bnode->root)
 		fs_info = bnode->root->fs_info;
 	btrfs_panic(fs_info, errno, "Inconsistency in backref cache "
-		    "found at offset %llu\n", (unsigned long long)bytenr);
+		    "found at offset %llu\n", bytenr);
 }
 
 /*
@@ -641,6 +641,11 @@ int find_inline_backref(struct extent_buffer *leaf, int slot,
 		WARN_ON(item_size < sizeof(*ei) + sizeof(*bi));
 		return 1;
 	}
+	if (key.type == BTRFS_METADATA_ITEM_KEY &&
+	    item_size <= sizeof(*ei)) {
+		WARN_ON(item_size < sizeof(*ei));
+		return 1;
+	}
 
 	if (key.type == BTRFS_EXTENT_ITEM_KEY) {
 		bi = (struct btrfs_tree_block_info *)(ei + 1);
@@ -691,6 +696,7 @@ struct backref_node *build_backref_tree(struct reloc_control *rc,
 	int cowonly;
 	int ret;
 	int err = 0;
+	bool need_check = true;
 
 	path1 = btrfs_alloc_path();
 	path2 = btrfs_alloc_path();
@@ -914,6 +920,7 @@ again:
 			cur->bytenr);
 
 		lower = cur;
+		need_check = true;
 		for (; level < BTRFS_MAX_LEVEL; level++) {
 			if (!path2->nodes[level]) {
 				BUG_ON(btrfs_root_bytenr(&root->root_item) !=
@@ -957,14 +964,12 @@ again:
 
 				/*
 				 * add the block to pending list if we
-				 * need check its backrefs. only block
-				 * at 'cur->level + 1' is added to the
-				 * tail of pending list. this guarantees
-				 * we check backrefs from lower level
-				 * blocks to upper level blocks.
+				 * need check its backrefs, we only do this once
+				 * while walking up a tree as we will catch
+				 * anything else later on.
 				 */
-				if (!upper->checked &&
-				    level == cur->level + 1) {
+				if (!upper->checked && need_check) {
+					need_check = false;
 					list_add_tail(&edge->list[UPPER],
 						      &list);
 				} else
@@ -1305,6 +1310,7 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 	struct extent_buffer *eb;
 	struct btrfs_root_item *root_item;
 	struct btrfs_key root_key;
+	u64 last_snap = 0;
 	int ret;
 
 	root_item = kmalloc(sizeof(*root_item), GFP_NOFS);
@@ -1320,6 +1326,7 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 				      BTRFS_TREE_RELOC_OBJECTID);
 		BUG_ON(ret);
 
+		last_snap = btrfs_root_last_snapshot(&root->root_item);
 		btrfs_set_root_last_snapshot(&root->root_item,
 					     trans->transid - 1);
 	} else {
@@ -1345,6 +1352,12 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 		memset(&root_item->drop_progress, 0,
 		       sizeof(struct btrfs_disk_key));
 		root_item->drop_level = 0;
+		/*
+		 * abuse rtransid, it is safe because it is impossible to
+		 * receive data into a relocation tree.
+		 */
+		btrfs_set_root_rtransid(root_item, last_snap);
+		btrfs_set_root_otransid(root_item, trans->transid);
 	}
 
 	btrfs_tree_unlock(eb);
@@ -1355,8 +1368,7 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 	BUG_ON(ret);
 	kfree(root_item);
 
-	reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
-						 &root_key);
+	reloc_root = btrfs_read_fs_root(root->fs_info->tree_root, &root_key);
 	BUG_ON(IS_ERR(reloc_root));
 	reloc_root->last_trans = trans->transid;
 	return reloc_root;
@@ -1536,7 +1548,7 @@ static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr,
 	       btrfs_file_extent_other_encoding(leaf, fi));
 
 	if (num_bytes != btrfs_file_extent_disk_num_bytes(leaf, fi)) {
-		ret = 1;
+		ret = -EINVAL;
 		goto out;
 	}
 
@@ -1567,7 +1579,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 	u64 end;
 	u32 nritems;
 	u32 i;
-	int ret;
+	int ret = 0;
 	int first = 1;
 	int dirty = 0;
 
@@ -1630,11 +1642,13 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 
 		ret = get_new_location(rc->data_inode, &new_bytenr,
 				       bytenr, num_bytes);
-		if (ret > 0) {
-			WARN_ON(1);
-			continue;
+		if (ret) {
+			/*
+			 * Don't have to abort since we've not changed anything
+			 * in the file extent yet.
+			 */
+			break;
 		}
-		BUG_ON(ret < 0);
 
 		btrfs_set_file_extent_disk_bytenr(leaf, fi, new_bytenr);
 		dirty = 1;
@@ -1644,18 +1658,24 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 					   num_bytes, parent,
 					   btrfs_header_owner(leaf),
 					   key.objectid, key.offset, 1);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
+			break;
+		}
 
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					parent, btrfs_header_owner(leaf),
 					key.objectid, key.offset, 1);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
+			break;
+		}
 	}
 	if (dirty)
 		btrfs_mark_buffer_dirty(leaf);
 	if (inode)
 		btrfs_add_delayed_iput(inode);
-	return 0;
+	return ret;
 }
 
 static noinline_for_stack
@@ -2273,8 +2293,12 @@ void free_reloc_roots(struct list_head *list)
 static noinline_for_stack
 int merge_reloc_roots(struct reloc_control *rc)
 {
+	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root;
 	struct btrfs_root *reloc_root;
+	u64 last_snap;
+	u64 otransid;
+	u64 objectid;
 	LIST_HEAD(reloc_roots);
 	int found = 0;
 	int ret = 0;
@@ -2303,17 +2327,51 @@ again:
 			BUG_ON(root->reloc_root != reloc_root);
 
 			ret = merge_reloc_root(rc, root);
-			if (ret)
+			if (ret) {
+				__update_reloc_root(reloc_root, 1);
+				free_extent_buffer(reloc_root->node);
+				free_extent_buffer(reloc_root->commit_root);
+				kfree(reloc_root);
 				goto out;
+			}
 		} else {
 			list_del_init(&reloc_root->root_list);
 		}
+
+		/*
+		 * we keep the old last snapshod transid in rtranid when we
+		 * created the relocation tree.
+		 */
+		last_snap = btrfs_root_rtransid(&reloc_root->root_item);
+		otransid = btrfs_root_otransid(&reloc_root->root_item);
+		objectid = reloc_root->root_key.offset;
+
 		ret = btrfs_drop_snapshot(reloc_root, rc->block_rsv, 0, 1);
 		if (ret < 0) {
 			if (list_empty(&reloc_root->root_list))
 				list_add_tail(&reloc_root->root_list,
 					      &reloc_roots);
 			goto out;
+		} else if (!ret) {
+			/*
+			 * recover the last snapshot tranid to avoid
+			 * the space balance break NOCOW.
+			 */
+			root = read_fs_root(rc->extent_root->fs_info,
+					    objectid);
+			if (IS_ERR(root))
+				continue;
+
+			trans = btrfs_join_transaction(root);
+			BUG_ON(IS_ERR(trans));
+
+			/* Check if the fs/file tree was snapshoted or not. */
+			if (btrfs_root_last_snapshot(&root->root_item) ==
+			    otransid - 1)
+				btrfs_set_root_last_snapshot(&root->root_item,
+							     last_snap);
+				
+			btrfs_end_transaction(trans, root);
 		}
 	}
 
@@ -3266,6 +3324,8 @@ static int __add_tree_block(struct reloc_control *rc,
 	struct btrfs_path *path;
 	struct btrfs_key key;
 	int ret;
+	bool skinny = btrfs_fs_incompat(rc->extent_root->fs_info,
+					SKINNY_METADATA);
 
 	if (tree_block_processed(bytenr, blocksize, rc))
 		return 0;
@@ -3276,10 +3336,15 @@ static int __add_tree_block(struct reloc_control *rc,
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
-
+again:
 	key.objectid = bytenr;
-	key.type = BTRFS_EXTENT_ITEM_KEY;
-	key.offset = blocksize;
+	if (skinny) {
+		key.type = BTRFS_METADATA_ITEM_KEY;
+		key.offset = (u64)-1;
+	} else {
+		key.type = BTRFS_EXTENT_ITEM_KEY;
+		key.offset = blocksize;
+	}
 
 	path->search_commit_root = 1;
 	path->skip_locking = 1;
@@ -3287,11 +3352,23 @@ static int __add_tree_block(struct reloc_control *rc,
 	if (ret < 0)
 		goto out;
 
-	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
-	if (ret > 0) {
-		if (key.objectid == bytenr &&
-		    key.type == BTRFS_METADATA_ITEM_KEY)
-			ret = 0;
+	if (ret > 0 && skinny) {
+		if (path->slots[0]) {
+			path->slots[0]--;
+			btrfs_item_key_to_cpu(path->nodes[0], &key,
+					      path->slots[0]);
+			if (key.objectid == bytenr &&
+			    (key.type == BTRFS_METADATA_ITEM_KEY ||
+			     (key.type == BTRFS_EXTENT_ITEM_KEY &&
+			      key.offset == blocksize)))
+				ret = 0;
+		}
+
+		if (ret) {
+			skinny = false;
+			btrfs_release_path(path);
+			goto again;
+		}
 	}
 	BUG_ON(ret);
 
@@ -3566,7 +3643,7 @@ int add_data_references(struct reloc_control *rc,
 	unsigned long ptr;
 	unsigned long end;
 	u32 blocksize = btrfs_level_size(rc->extent_root, 0);
-	int ret;
+	int ret = 0;
 	int err = 0;
 
 	eb = path->nodes[0];
@@ -3593,6 +3670,10 @@ int add_data_references(struct reloc_control *rc,
 		} else {
 			BUG();
 		}
+		if (ret) {
+			err = ret;
+			goto out;
+		}
 		ptr += btrfs_extent_inline_ref_size(key.type);
 	}
 	WARN_ON(ptr > end);
@@ -3638,6 +3719,7 @@ int add_data_references(struct reloc_control *rc,
 		}
 		path->slots[0]++;
 	}
+out:
 	btrfs_release_path(path);
 	if (err)
 		free_block_list(blocks);
@@ -4082,7 +4164,7 @@ out:
 	return inode;
 }
 
-static struct reloc_control *alloc_reloc_control(void)
+static struct reloc_control *alloc_reloc_control(struct btrfs_fs_info *fs_info)
 {
 	struct reloc_control *rc;
 
@@ -4093,7 +4175,8 @@ static struct reloc_control *alloc_reloc_control(void)
 	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);
+	extent_io_tree_init(&rc->processed_blocks,
+			    fs_info->btree_inode->i_mapping);
 	return rc;
 }
 
@@ -4110,7 +4193,7 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 	int rw = 0;
 	int err = 0;
 
-	rc = alloc_reloc_control();
+	rc = alloc_reloc_control(fs_info);
 	if (!rc)
 		return -ENOMEM;
 
@@ -4156,15 +4239,14 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 	}
 
 	printk(KERN_INFO "btrfs: relocating block group %llu flags %llu\n",
-	       (unsigned long long)rc->block_group->key.objectid,
-	       (unsigned long long)rc->block_group->flags);
+	       rc->block_group->key.objectid, rc->block_group->flags);
 
-	ret = btrfs_start_delalloc_inodes(fs_info->tree_root, 0);
+	ret = btrfs_start_all_delalloc_inodes(fs_info, 0);
 	if (ret < 0) {
 		err = ret;
 		goto out;
 	}
-	btrfs_wait_ordered_extents(fs_info->tree_root, 0);
+	btrfs_wait_all_ordered_extents(fs_info);
 
 	while (1) {
 		mutex_lock(&fs_info->cleaner_mutex);
@@ -4179,7 +4261,7 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 			break;
 
 		printk(KERN_INFO "btrfs: found %llu extents\n",
-			(unsigned long long)rc->extents_found);
+			rc->extents_found);
 
 		if (rc->stage == MOVE_DATA_EXTENTS && rc->found_file_extent) {
 			btrfs_wait_ordered_range(rc->data_inode, 0, (u64)-1);
@@ -4276,7 +4358,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 		    key.type != BTRFS_ROOT_ITEM_KEY)
 			break;
 
-		reloc_root = btrfs_read_fs_root_no_radix(root, &key);
+		reloc_root = btrfs_read_fs_root(root, &key);
 		if (IS_ERR(reloc_root)) {
 			err = PTR_ERR(reloc_root);
 			goto out;
@@ -4311,7 +4393,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 	if (list_empty(&reloc_roots))
 		goto out;
 
-	rc = alloc_reloc_control();
+	rc = alloc_reloc_control(root->fs_info);
 	if (!rc) {
 		err = -ENOMEM;
 		goto out;
@@ -4395,10 +4477,8 @@ out:
 int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 {
 	struct btrfs_ordered_sum *sums;
-	struct btrfs_sector_sum *sector_sum;
 	struct btrfs_ordered_extent *ordered;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-	size_t offset;
 	int ret;
 	u64 disk_bytenr;
 	LIST_HEAD(list);
@@ -4412,19 +4492,13 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 	if (ret)
 		goto out;
 
+	disk_bytenr = ordered->start;
 	while (!list_empty(&list)) {
 		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
 		list_del_init(&sums->list);
 
-		sector_sum = sums->sums;
-		sums->bytenr = ordered->start;
-
-		offset = 0;
-		while (offset < sums->len) {
-			sector_sum->bytenr += ordered->start - disk_bytenr;
-			sector_sum++;
-			offset += root->sectorsize;
-		}
+		sums->bytenr = disk_bytenr;
+		disk_bytenr += sums->len;
 
 		btrfs_add_ordered_sum(inode, ordered, sums);
 	}
@@ -4433,19 +4507,19 @@ out:
 	return ret;
 }
 
-void btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
-			   struct btrfs_root *root, struct extent_buffer *buf,
-			   struct extent_buffer *cow)
+int 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;
+	int ret = 0;
 
 	rc = root->fs_info->reloc_ctl;
 	if (!rc)
-		return;
+		return 0;
 
 	BUG_ON(rc->stage == UPDATE_DATA_PTRS &&
 	       root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID);
@@ -4481,10 +4555,9 @@ void btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
 			rc->nodes_relocated += buf->len;
 	}
 
-	if (level == 0 && first_cow && rc->stage == UPDATE_DATA_PTRS) {
+	if (level == 0 && first_cow && rc->stage == UPDATE_DATA_PTRS)
 		ret = replace_file_extents(trans, rc, root, cow);
-		BUG_ON(ret);
-	}
+	return ret;
 }
 
 /*
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index 5bf1ed57f178..0b1f4ef8db98 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -29,8 +29,8 @@
  * generation numbers as then we know the root was once mounted with an older
  * kernel that was not aware of the root item structure change.
  */
-void btrfs_read_root_item(struct extent_buffer *eb, int slot,
-			  struct btrfs_root_item *item)
+static void btrfs_read_root_item(struct extent_buffer *eb, int slot,
+				struct btrfs_root_item *item)
 {
 	uuid_le uuid;
 	int len;
@@ -64,52 +64,59 @@ void btrfs_read_root_item(struct extent_buffer *eb, int slot,
 }
 
 /*
- * lookup the root with the highest offset for a given objectid.  The key we do
- * find is copied into 'key'.  If we find something return 0, otherwise 1, < 0
- * on error.
+ * btrfs_find_root - lookup the root by the key.
+ * root: the root of the root tree
+ * search_key: the key to search
+ * path: the path we search
+ * root_item: the root item of the tree we look for
+ * root_key: the reak key of the tree we look for
+ *
+ * If ->offset of 'seach_key' is -1ULL, it means we are not sure the offset
+ * of the search key, just lookup the root with the highest offset for a
+ * given objectid.
+ *
+ * If we find something return 0, otherwise > 0, < 0 on error.
  */
-int btrfs_find_last_root(struct btrfs_root *root, u64 objectid,
-			struct btrfs_root_item *item, struct btrfs_key *key)
+int btrfs_find_root(struct btrfs_root *root, struct btrfs_key *search_key,
+		    struct btrfs_path *path, struct btrfs_root_item *root_item,
+		    struct btrfs_key *root_key)
 {
-	struct btrfs_path *path;
-	struct btrfs_key search_key;
 	struct btrfs_key found_key;
 	struct extent_buffer *l;
 	int ret;
 	int slot;
 
-	search_key.objectid = objectid;
-	search_key.type = BTRFS_ROOT_ITEM_KEY;
-	search_key.offset = (u64)-1;
-
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
-	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
+	ret = btrfs_search_slot(NULL, root, search_key, path, 0, 0);
 	if (ret < 0)
-		goto out;
+		return ret;
 
-	BUG_ON(ret == 0);
-	if (path->slots[0] == 0) {
-		ret = 1;
-		goto out;
+	if (search_key->offset != -1ULL) {	/* the search key is exact */
+		if (ret > 0)
+			goto out;
+	} else {
+		BUG_ON(ret == 0);		/* Logical error */
+		if (path->slots[0] == 0)
+			goto out;
+		path->slots[0]--;
+		ret = 0;
 	}
+
 	l = path->nodes[0];
-	slot = path->slots[0] - 1;
+	slot = path->slots[0];
+
 	btrfs_item_key_to_cpu(l, &found_key, slot);
-	if (found_key.objectid != objectid ||
+	if (found_key.objectid != search_key->objectid ||
 	    found_key.type != BTRFS_ROOT_ITEM_KEY) {
 		ret = 1;
 		goto out;
 	}
-	if (item)
-		btrfs_read_root_item(l, slot, item);
-	if (key)
-		memcpy(key, &found_key, sizeof(found_key));
 
-	ret = 0;
+	if (root_item)
+		btrfs_read_root_item(l, slot, root_item);
+	if (root_key)
+		memcpy(root_key, &found_key, sizeof(found_key));
 out:
-	btrfs_free_path(path);
+	btrfs_release_path(path);
 	return ret;
 }
 
@@ -148,8 +155,7 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (ret != 0) {
 		btrfs_print_leaf(root, path->nodes[0]);
 		printk(KERN_CRIT "unable to update root key %llu %u %llu\n",
-		       (unsigned long long)key->objectid, key->type,
-		       (unsigned long long)key->offset);
+		       key->objectid, key->type, key->offset);
 		BUG_ON(1);
 	}
 
@@ -212,86 +218,6 @@ int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	return btrfs_insert_item(trans, root, key, item, sizeof(*item));
 }
 
-/*
- * at mount time we want to find all the old transaction snapshots that were in
- * the process of being deleted if we crashed.  This is any root item with an
- * offset lower than the latest root.  They need to be queued for deletion to
- * finish what was happening when we crashed.
- */
-int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid)
-{
-	struct btrfs_root *dead_root;
-	struct btrfs_root_item *ri;
-	struct btrfs_key key;
-	struct btrfs_key found_key;
-	struct btrfs_path *path;
-	int ret;
-	u32 nritems;
-	struct extent_buffer *leaf;
-	int slot;
-
-	key.objectid = objectid;
-	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
-	key.offset = 0;
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
-
-again:
-	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-	if (ret < 0)
-		goto err;
-	while (1) {
-		leaf = path->nodes[0];
-		nritems = btrfs_header_nritems(leaf);
-		slot = path->slots[0];
-		if (slot >= nritems) {
-			ret = btrfs_next_leaf(root, path);
-			if (ret)
-				break;
-			leaf = path->nodes[0];
-			nritems = btrfs_header_nritems(leaf);
-			slot = path->slots[0];
-		}
-		btrfs_item_key_to_cpu(leaf, &key, slot);
-		if (btrfs_key_type(&key) != BTRFS_ROOT_ITEM_KEY)
-			goto next;
-
-		if (key.objectid < objectid)
-			goto next;
-
-		if (key.objectid > objectid)
-			break;
-
-		ri = btrfs_item_ptr(leaf, slot, struct btrfs_root_item);
-		if (btrfs_disk_root_refs(leaf, ri) != 0)
-			goto next;
-
-		memcpy(&found_key, &key, sizeof(key));
-		key.offset++;
-		btrfs_release_path(path);
-		dead_root =
-			btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
-						    &found_key);
-		if (IS_ERR(dead_root)) {
-			ret = PTR_ERR(dead_root);
-			goto err;
-		}
-
-		ret = btrfs_add_dead_root(dead_root);
-		if (ret)
-			goto err;
-		goto again;
-next:
-		slot++;
-		path->slots[0]++;
-	}
-	ret = 0;
-err:
-	btrfs_free_path(path);
-	return ret;
-}
-
 int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 {
 	struct extent_buffer *leaf;
@@ -301,6 +227,10 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 	struct btrfs_root *root;
 	int err = 0;
 	int ret;
+	bool can_recover = true;
+
+	if (tree_root->fs_info->sb->s_flags & MS_RDONLY)
+		can_recover = false;
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -340,20 +270,52 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 		root_key.objectid = key.offset;
 		key.offset++;
 
-		root = btrfs_read_fs_root_no_name(tree_root->fs_info,
-						  &root_key);
-		if (!IS_ERR(root))
+		root = btrfs_read_fs_root(tree_root, &root_key);
+		err = PTR_RET(root);
+		if (err && err != -ENOENT) {
+			break;
+		} else if (err == -ENOENT) {
+			struct btrfs_trans_handle *trans;
+
+			btrfs_release_path(path);
+
+			trans = btrfs_join_transaction(tree_root);
+			if (IS_ERR(trans)) {
+				err = PTR_ERR(trans);
+				btrfs_error(tree_root->fs_info, err,
+					    "Failed to start trans to delete "
+					    "orphan item");
+				break;
+			}
+			err = btrfs_del_orphan_item(trans, tree_root,
+						    root_key.objectid);
+			btrfs_end_transaction(trans, tree_root);
+			if (err) {
+				btrfs_error(tree_root->fs_info, err,
+					    "Failed to delete root orphan "
+					    "item");
+				break;
+			}
 			continue;
+		}
 
-		ret = PTR_ERR(root);
-		if (ret != -ENOENT) {
-			err = ret;
+		if (btrfs_root_refs(&root->root_item) == 0) {
+			btrfs_add_dead_root(root);
+			continue;
+		}
+
+		err = btrfs_init_fs_root(root);
+		if (err) {
+			btrfs_free_fs_root(root);
 			break;
 		}
 
-		ret = btrfs_find_dead_roots(tree_root, root_key.objectid);
-		if (ret) {
-			err = ret;
+		root->orphan_item_inserted = 1;
+
+		err = btrfs_insert_fs_root(root->fs_info, root);
+		if (err) {
+			BUG_ON(err == -EEXIST);
+			btrfs_free_fs_root(root);
 			break;
 		}
 	}
@@ -368,8 +330,6 @@ int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 {
 	struct btrfs_path *path;
 	int ret;
-	struct btrfs_root_item *ri;
-	struct extent_buffer *leaf;
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -379,8 +339,6 @@ int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		goto out;
 
 	BUG_ON(ret != 0);
-	leaf = path->nodes[0];
-	ri = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_item);
 
 	ret = btrfs_del_item(trans, root, path);
 out:
@@ -531,13 +489,13 @@ again:
  */
 void btrfs_check_and_init_root_item(struct btrfs_root_item *root_item)
 {
-	u64 inode_flags = le64_to_cpu(root_item->inode.flags);
+	u64 inode_flags = btrfs_stack_inode_flags(&root_item->inode);
 
 	if (!(inode_flags & BTRFS_INODE_ROOT_ITEM_INIT)) {
 		inode_flags |= BTRFS_INODE_ROOT_ITEM_INIT;
-		root_item->inode.flags = cpu_to_le64(inode_flags);
-		root_item->flags = 0;
-		root_item->byte_limit = 0;
+		btrfs_set_stack_inode_flags(&root_item->inode, inode_flags);
+		btrfs_set_root_flags(root_item, 0);
+		btrfs_set_root_limit(root_item, 0);
 	}
 }
 
@@ -548,8 +506,8 @@ void btrfs_update_root_times(struct btrfs_trans_handle *trans,
 	struct timespec ct = CURRENT_TIME;
 
 	spin_lock(&root->root_item_lock);
-	item->ctransid = cpu_to_le64(trans->transid);
-	item->ctime.sec = cpu_to_le64(ct.tv_sec);
-	item->ctime.nsec = cpu_to_le32(ct.tv_nsec);
+	btrfs_set_root_ctransid(item, trans->transid);
+	btrfs_set_stack_timespec_sec(&item->ctime, ct.tv_sec);
+	btrfs_set_stack_timespec_nsec(&item->ctime, ct.tv_nsec);
 	spin_unlock(&root->root_item_lock);
 }
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index 79bd479317cb..a18e0e23f6a6 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -158,12 +158,20 @@ struct scrub_fixup_nodatasum {
 	int			mirror_num;
 };
 
+struct scrub_nocow_inode {
+	u64			inum;
+	u64			offset;
+	u64			root;
+	struct list_head	list;
+};
+
 struct scrub_copy_nocow_ctx {
 	struct scrub_ctx	*sctx;
 	u64			logical;
 	u64			len;
 	int			mirror_num;
 	u64			physical_for_dev_replace;
+	struct list_head	inodes;
 	struct btrfs_work	work;
 };
 
@@ -245,7 +253,7 @@ static void scrub_wr_bio_end_io_worker(struct btrfs_work *work);
 static int write_page_nocow(struct scrub_ctx *sctx,
 			    u64 physical_for_dev_replace, struct page *page);
 static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root,
-				      void *ctx);
+				      struct scrub_copy_nocow_ctx *ctx);
 static int copy_nocow_pages(struct scrub_ctx *sctx, u64 logical, u64 len,
 			    int mirror_num, u64 physical_for_dev_replace);
 static void copy_nocow_pages_worker(struct btrfs_work *work);
@@ -754,8 +762,7 @@ out:
 			num_uncorrectable_read_errors);
 		printk_ratelimited_in_rcu(KERN_ERR
 			"btrfs: unable to fixup (nodatasum) error at logical %llu on dev %s\n",
-			(unsigned long long)fixup->logical,
-			rcu_str_deref(fixup->dev->name));
+			fixup->logical, rcu_str_deref(fixup->dev->name));
 	}
 
 	btrfs_free_path(path);
@@ -1154,8 +1161,7 @@ corrected_error:
 			spin_unlock(&sctx->stat_lock);
 			printk_ratelimited_in_rcu(KERN_ERR
 				"btrfs: fixed up error at logical %llu on dev %s\n",
-				(unsigned long long)logical,
-				rcu_str_deref(dev->name));
+				logical, rcu_str_deref(dev->name));
 		}
 	} else {
 did_not_correct_error:
@@ -1164,8 +1170,7 @@ did_not_correct_error:
 		spin_unlock(&sctx->stat_lock);
 		printk_ratelimited_in_rcu(KERN_ERR
 			"btrfs: unable to fixup (regular) error at logical %llu on dev %s\n",
-			(unsigned long long)logical,
-			rcu_str_deref(dev->name));
+			logical, rcu_str_deref(dev->name));
 	}
 
 out:
@@ -1345,12 +1350,12 @@ static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info,
 		mapped_buffer = kmap_atomic(sblock->pagev[0]->page);
 		h = (struct btrfs_header *)mapped_buffer;
 
-		if (sblock->pagev[0]->logical != le64_to_cpu(h->bytenr) ||
+		if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h) ||
 		    memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE) ||
 		    memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid,
 			   BTRFS_UUID_SIZE)) {
 			sblock->header_error = 1;
-		} else if (generation != le64_to_cpu(h->generation)) {
+		} else if (generation != btrfs_stack_header_generation(h)) {
 			sblock->header_error = 1;
 			sblock->generation_error = 1;
 		}
@@ -1720,10 +1725,10 @@ static int scrub_checksum_tree_block(struct scrub_block *sblock)
 	 * b) the page is already kmapped
 	 */
 
-	if (sblock->pagev[0]->logical != le64_to_cpu(h->bytenr))
+	if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h))
 		++fail;
 
-	if (sblock->pagev[0]->generation != le64_to_cpu(h->generation))
+	if (sblock->pagev[0]->generation != btrfs_stack_header_generation(h))
 		++fail;
 
 	if (memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
@@ -1786,10 +1791,10 @@ static int scrub_checksum_super(struct scrub_block *sblock)
 	s = (struct btrfs_super_block *)mapped_buffer;
 	memcpy(on_disk_csum, s->csum, sctx->csum_size);
 
-	if (sblock->pagev[0]->logical != le64_to_cpu(s->bytenr))
+	if (sblock->pagev[0]->logical != btrfs_super_bytenr(s))
 		++fail_cor;
 
-	if (sblock->pagev[0]->generation != le64_to_cpu(s->generation))
+	if (sblock->pagev[0]->generation != btrfs_super_generation(s))
 		++fail_gen;
 
 	if (memcmp(s->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
@@ -2126,8 +2131,7 @@ static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u64 len,
 			   u8 *csum)
 {
 	struct btrfs_ordered_sum *sum = NULL;
-	int ret = 0;
-	unsigned long i;
+	unsigned long index;
 	unsigned long num_sectors;
 
 	while (!list_empty(&sctx->csum_list)) {
@@ -2146,19 +2150,14 @@ static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u64 len,
 	if (!sum)
 		return 0;
 
+	index = ((u32)(logical - sum->bytenr)) / sctx->sectorsize;
 	num_sectors = sum->len / sctx->sectorsize;
-	for (i = 0; i < num_sectors; ++i) {
-		if (sum->sums[i].bytenr == logical) {
-			memcpy(csum, &sum->sums[i].sum, sctx->csum_size);
-			ret = 1;
-			break;
-		}
-	}
-	if (ret && i == num_sectors - 1) {
+	memcpy(csum, sum->sums + index, sctx->csum_size);
+	if (index == num_sectors - 1) {
 		list_del(&sum->list);
 		kfree(sum);
 	}
-	return ret;
+	return 1;
 }
 
 /* scrub extent tries to collect up to 64 kB for each bio */
@@ -2461,8 +2460,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 				printk(KERN_ERR
 				       "btrfs scrub: tree block %llu spanning "
 				       "stripes, ignored. logical=%llu\n",
-				       (unsigned long long)key.objectid,
-				       (unsigned long long)logical);
+				       key.objectid, logical);
 				goto next;
 			}
 
@@ -2501,10 +2499,11 @@ again:
 			ret = scrub_extent(sctx, extent_logical, extent_len,
 					   extent_physical, extent_dev, flags,
 					   generation, extent_mirror_num,
-					   extent_physical);
+					   extent_logical - logical + physical);
 			if (ret)
 				goto out;
 
+			scrub_free_csums(sctx);
 			if (extent_logical + extent_len <
 			    key.objectid + bytes) {
 				logical += increment;
@@ -2868,9 +2867,8 @@ int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start,
 	if (fs_info->chunk_root->sectorsize != PAGE_SIZE) {
 		/* not supported for data w/o checksums */
 		printk(KERN_ERR
-		       "btrfs_scrub: size assumption sectorsize != PAGE_SIZE (%d != %lld) fails\n",
-		       fs_info->chunk_root->sectorsize,
-		       (unsigned long long)PAGE_SIZE);
+		       "btrfs_scrub: size assumption sectorsize != PAGE_SIZE (%d != %lu) fails\n",
+		       fs_info->chunk_root->sectorsize, PAGE_SIZE);
 		return -EINVAL;
 	}
 
@@ -3136,12 +3134,30 @@ static int copy_nocow_pages(struct scrub_ctx *sctx, u64 logical, u64 len,
 	nocow_ctx->mirror_num = mirror_num;
 	nocow_ctx->physical_for_dev_replace = physical_for_dev_replace;
 	nocow_ctx->work.func = copy_nocow_pages_worker;
+	INIT_LIST_HEAD(&nocow_ctx->inodes);
 	btrfs_queue_worker(&fs_info->scrub_nocow_workers,
 			   &nocow_ctx->work);
 
 	return 0;
 }
 
+static int record_inode_for_nocow(u64 inum, u64 offset, u64 root, void *ctx)
+{
+	struct scrub_copy_nocow_ctx *nocow_ctx = ctx;
+	struct scrub_nocow_inode *nocow_inode;
+
+	nocow_inode = kzalloc(sizeof(*nocow_inode), GFP_NOFS);
+	if (!nocow_inode)
+		return -ENOMEM;
+	nocow_inode->inum = inum;
+	nocow_inode->offset = offset;
+	nocow_inode->root = root;
+	list_add_tail(&nocow_inode->list, &nocow_ctx->inodes);
+	return 0;
+}
+
+#define COPY_COMPLETE 1
+
 static void copy_nocow_pages_worker(struct btrfs_work *work)
 {
 	struct scrub_copy_nocow_ctx *nocow_ctx =
@@ -3177,19 +3193,42 @@ static void copy_nocow_pages_worker(struct btrfs_work *work)
 	}
 
 	ret = iterate_inodes_from_logical(logical, fs_info, path,
-					  copy_nocow_pages_for_inode,
-					  nocow_ctx);
+					  record_inode_for_nocow, nocow_ctx);
 	if (ret != 0 && ret != -ENOENT) {
-		pr_warn("iterate_inodes_from_logical() failed: log %llu, phys %llu, len %llu, mir %llu, ret %d\n",
-			(unsigned long long)logical,
-			(unsigned long long)physical_for_dev_replace,
-			(unsigned long long)len,
-			(unsigned long long)mirror_num, ret);
+		pr_warn("iterate_inodes_from_logical() failed: log %llu, phys %llu, len %llu, mir %u, ret %d\n",
+			logical, physical_for_dev_replace, len, mirror_num,
+			ret);
 		not_written = 1;
 		goto out;
 	}
 
+	btrfs_end_transaction(trans, root);
+	trans = NULL;
+	while (!list_empty(&nocow_ctx->inodes)) {
+		struct scrub_nocow_inode *entry;
+		entry = list_first_entry(&nocow_ctx->inodes,
+					 struct scrub_nocow_inode,
+					 list);
+		list_del_init(&entry->list);
+		ret = copy_nocow_pages_for_inode(entry->inum, entry->offset,
+						 entry->root, nocow_ctx);
+		kfree(entry);
+		if (ret == COPY_COMPLETE) {
+			ret = 0;
+			break;
+		} else if (ret) {
+			break;
+		}
+	}
 out:
+	while (!list_empty(&nocow_ctx->inodes)) {
+		struct scrub_nocow_inode *entry;
+		entry = list_first_entry(&nocow_ctx->inodes,
+					 struct scrub_nocow_inode,
+					 list);
+		list_del_init(&entry->list);
+		kfree(entry);
+	}
 	if (trans && !IS_ERR(trans))
 		btrfs_end_transaction(trans, root);
 	if (not_written)
@@ -3202,18 +3241,25 @@ out:
 	scrub_pending_trans_workers_dec(sctx);
 }
 
-static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, void *ctx)
+static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root,
+				      struct scrub_copy_nocow_ctx *nocow_ctx)
 {
-	unsigned long index;
-	struct scrub_copy_nocow_ctx *nocow_ctx = ctx;
-	int ret = 0;
+	struct btrfs_fs_info *fs_info = nocow_ctx->sctx->dev_root->fs_info;
 	struct btrfs_key key;
-	struct inode *inode = NULL;
+	struct inode *inode;
+	struct page *page;
 	struct btrfs_root *local_root;
+	struct btrfs_ordered_extent *ordered;
+	struct extent_map *em;
+	struct extent_state *cached_state = NULL;
+	struct extent_io_tree *io_tree;
 	u64 physical_for_dev_replace;
-	u64 len;
-	struct btrfs_fs_info *fs_info = nocow_ctx->sctx->dev_root->fs_info;
+	u64 len = nocow_ctx->len;
+	u64 lockstart = offset, lockend = offset + len - 1;
+	unsigned long index;
 	int srcu_index;
+	int ret = 0;
+	int err = 0;
 
 	key.objectid = root;
 	key.type = BTRFS_ROOT_ITEM_KEY;
@@ -3235,19 +3281,45 @@ static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, void *ctx)
 	if (IS_ERR(inode))
 		return PTR_ERR(inode);
 
+	/* Avoid truncate/dio/punch hole.. */
+	mutex_lock(&inode->i_mutex);
+	inode_dio_wait(inode);
+
 	physical_for_dev_replace = nocow_ctx->physical_for_dev_replace;
-	len = nocow_ctx->len;
-	while (len >= PAGE_CACHE_SIZE) {
-		struct page *page = NULL;
-		int ret_sub;
+	io_tree = &BTRFS_I(inode)->io_tree;
 
-		index = offset >> PAGE_CACHE_SHIFT;
+	lock_extent_bits(io_tree, lockstart, lockend, 0, &cached_state);
+	ordered = btrfs_lookup_ordered_range(inode, lockstart, len);
+	if (ordered) {
+		btrfs_put_ordered_extent(ordered);
+		goto out_unlock;
+	}
+
+	em = btrfs_get_extent(inode, NULL, 0, lockstart, len, 0);
+	if (IS_ERR(em)) {
+		ret = PTR_ERR(em);
+		goto out_unlock;
+	}
 
+	/*
+	 * This extent does not actually cover the logical extent anymore,
+	 * move on to the next inode.
+	 */
+	if (em->block_start > nocow_ctx->logical ||
+	    em->block_start + em->block_len < nocow_ctx->logical + len) {
+		free_extent_map(em);
+		goto out_unlock;
+	}
+	free_extent_map(em);
+
+	while (len >= PAGE_CACHE_SIZE) {
+		index = offset >> PAGE_CACHE_SHIFT;
+again:
 		page = find_or_create_page(inode->i_mapping, index, GFP_NOFS);
 		if (!page) {
 			pr_err("find_or_create_page() failed\n");
 			ret = -ENOMEM;
-			goto next_page;
+			goto out;
 		}
 
 		if (PageUptodate(page)) {
@@ -3255,39 +3327,53 @@ static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, void *ctx)
 				goto next_page;
 		} else {
 			ClearPageError(page);
-			ret_sub = extent_read_full_page(&BTRFS_I(inode)->
-							 io_tree,
-							page, btrfs_get_extent,
-							nocow_ctx->mirror_num);
-			if (ret_sub) {
-				ret = ret_sub;
+			err = extent_read_full_page_nolock(io_tree, page,
+							   btrfs_get_extent,
+							   nocow_ctx->mirror_num);
+			if (err) {
+				ret = err;
 				goto next_page;
 			}
-			wait_on_page_locked(page);
+
+			lock_page(page);
+			/*
+			 * If the page has been remove from the page cache,
+			 * the data on it is meaningless, because it may be
+			 * old one, the new data may be written into the new
+			 * page in the page cache.
+			 */
+			if (page->mapping != inode->i_mapping) {
+				unlock_page(page);
+				page_cache_release(page);
+				goto again;
+			}
 			if (!PageUptodate(page)) {
 				ret = -EIO;
 				goto next_page;
 			}
 		}
-		ret_sub = write_page_nocow(nocow_ctx->sctx,
-					   physical_for_dev_replace, page);
-		if (ret_sub) {
-			ret = ret_sub;
-			goto next_page;
-		}
-
+		err = write_page_nocow(nocow_ctx->sctx,
+				       physical_for_dev_replace, page);
+		if (err)
+			ret = err;
 next_page:
-		if (page) {
-			unlock_page(page);
-			put_page(page);
-		}
+		unlock_page(page);
+		page_cache_release(page);
+
+		if (ret)
+			break;
+
 		offset += PAGE_CACHE_SIZE;
 		physical_for_dev_replace += PAGE_CACHE_SIZE;
 		len -= PAGE_CACHE_SIZE;
 	}
-
-	if (inode)
-		iput(inode);
+	ret = COPY_COMPLETE;
+out_unlock:
+	unlock_extent_cached(io_tree, lockstart, lockend, &cached_state,
+			     GFP_NOFS);
+out:
+	mutex_unlock(&inode->i_mutex);
+	iput(inode);
 	return ret;
 }
 
diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index ff40f1c00ce3..e46e0ed74925 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -26,6 +26,7 @@
 #include <linux/radix-tree.h>
 #include <linux/crc32c.h>
 #include <linux/vmalloc.h>
+#include <linux/string.h>
 
 #include "send.h"
 #include "backref.h"
@@ -54,8 +55,8 @@ struct fs_path {
 
 			char *buf;
 			int buf_len;
-			int reversed:1;
-			int virtual_mem:1;
+			unsigned int reversed:1;
+			unsigned int virtual_mem:1;
 			char inline_buf[];
 		};
 		char pad[PAGE_SIZE];
@@ -158,7 +159,7 @@ static void fs_path_reset(struct fs_path *p)
 	}
 }
 
-static struct fs_path *fs_path_alloc(struct send_ctx *sctx)
+static struct fs_path *fs_path_alloc(void)
 {
 	struct fs_path *p;
 
@@ -173,11 +174,11 @@ static struct fs_path *fs_path_alloc(struct send_ctx *sctx)
 	return p;
 }
 
-static struct fs_path *fs_path_alloc_reversed(struct send_ctx *sctx)
+static struct fs_path *fs_path_alloc_reversed(void)
 {
 	struct fs_path *p;
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return NULL;
 	p->reversed = 1;
@@ -185,7 +186,7 @@ static struct fs_path *fs_path_alloc_reversed(struct send_ctx *sctx)
 	return p;
 }
 
-static void fs_path_free(struct send_ctx *sctx, struct fs_path *p)
+static void fs_path_free(struct fs_path *p)
 {
 	if (!p)
 		return;
@@ -219,7 +220,7 @@ static int fs_path_ensure_buf(struct fs_path *p, int len)
 	len = PAGE_ALIGN(len);
 
 	if (p->buf == p->inline_buf) {
-		tmp_buf = kmalloc(len, GFP_NOFS);
+		tmp_buf = kmalloc(len, GFP_NOFS | __GFP_NOWARN);
 		if (!tmp_buf) {
 			tmp_buf = vmalloc(len);
 			if (!tmp_buf)
@@ -753,8 +754,7 @@ typedef int (*iterate_inode_ref_t)(int num, u64 dir, int index,
  *
  * path must point to the INODE_REF or INODE_EXTREF when called.
  */
-static int iterate_inode_ref(struct send_ctx *sctx,
-			     struct btrfs_root *root, struct btrfs_path *path,
+static int iterate_inode_ref(struct btrfs_root *root, struct btrfs_path *path,
 			     struct btrfs_key *found_key, int resolve,
 			     iterate_inode_ref_t iterate, void *ctx)
 {
@@ -777,13 +777,13 @@ static int iterate_inode_ref(struct send_ctx *sctx,
 	unsigned long elem_size;
 	unsigned long ptr;
 
-	p = fs_path_alloc_reversed(sctx);
+	p = fs_path_alloc_reversed();
 	if (!p)
 		return -ENOMEM;
 
 	tmp_path = alloc_path_for_send();
 	if (!tmp_path) {
-		fs_path_free(sctx, p);
+		fs_path_free(p);
 		return -ENOMEM;
 	}
 
@@ -858,7 +858,7 @@ static int iterate_inode_ref(struct send_ctx *sctx,
 
 out:
 	btrfs_free_path(tmp_path);
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -874,8 +874,7 @@ typedef int (*iterate_dir_item_t)(int num, struct btrfs_key *di_key,
  *
  * path must point to the dir item when called.
  */
-static int iterate_dir_item(struct send_ctx *sctx,
-			    struct btrfs_root *root, struct btrfs_path *path,
+static int iterate_dir_item(struct btrfs_root *root, struct btrfs_path *path,
 			    struct btrfs_key *found_key,
 			    iterate_dir_item_t iterate, void *ctx)
 {
@@ -990,7 +989,7 @@ static int __copy_first_ref(int num, u64 dir, int index,
  * Retrieve the first path of an inode. If an inode has more then one
  * ref/hardlink, this is ignored.
  */
-static int get_inode_path(struct send_ctx *sctx, struct btrfs_root *root,
+static int get_inode_path(struct btrfs_root *root,
 			  u64 ino, struct fs_path *path)
 {
 	int ret;
@@ -1022,8 +1021,8 @@ static int get_inode_path(struct send_ctx *sctx, struct btrfs_root *root,
 		goto out;
 	}
 
-	ret = iterate_inode_ref(sctx, root, p, &found_key, 1,
-			__copy_first_ref, path);
+	ret = iterate_inode_ref(root, p, &found_key, 1,
+				__copy_first_ref, path);
 	if (ret < 0)
 		goto out;
 	ret = 0;
@@ -1314,8 +1313,7 @@ out:
 	return ret;
 }
 
-static int read_symlink(struct send_ctx *sctx,
-			struct btrfs_root *root,
+static int read_symlink(struct btrfs_root *root,
 			u64 ino,
 			struct fs_path *dest)
 {
@@ -1562,8 +1560,7 @@ out:
  * Looks up the first btrfs_inode_ref of a given ino. It returns the parent dir,
  * generation of the parent dir and the name of the dir entry.
  */
-static int get_first_ref(struct send_ctx *sctx,
-			 struct btrfs_root *root, u64 ino,
+static int get_first_ref(struct btrfs_root *root, u64 ino,
 			 u64 *dir, u64 *dir_gen, struct fs_path *name)
 {
 	int ret;
@@ -1628,8 +1625,7 @@ out:
 	return ret;
 }
 
-static int is_first_ref(struct send_ctx *sctx,
-			struct btrfs_root *root,
+static int is_first_ref(struct btrfs_root *root,
 			u64 ino, u64 dir,
 			const char *name, int name_len)
 {
@@ -1638,11 +1634,11 @@ static int is_first_ref(struct send_ctx *sctx,
 	u64 tmp_dir;
 	u64 tmp_dir_gen;
 
-	tmp_name = fs_path_alloc(sctx);
+	tmp_name = fs_path_alloc();
 	if (!tmp_name)
 		return -ENOMEM;
 
-	ret = get_first_ref(sctx, root, ino, &tmp_dir, &tmp_dir_gen, tmp_name);
+	ret = get_first_ref(root, ino, &tmp_dir, &tmp_dir_gen, tmp_name);
 	if (ret < 0)
 		goto out;
 
@@ -1654,7 +1650,7 @@ static int is_first_ref(struct send_ctx *sctx,
 	ret = !memcmp(tmp_name->start, name, name_len);
 
 out:
-	fs_path_free(sctx, tmp_name);
+	fs_path_free(tmp_name);
 	return ret;
 }
 
@@ -1673,6 +1669,7 @@ static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 			      u64 *who_ino, u64 *who_gen)
 {
 	int ret = 0;
+	u64 gen;
 	u64 other_inode = 0;
 	u8 other_type = 0;
 
@@ -1683,6 +1680,24 @@ static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen,
 	if (ret <= 0)
 		goto out;
 
+	/*
+	 * If we have a parent root we need to verify that the parent dir was
+	 * not delted and then re-created, if it was then we have no overwrite
+	 * and we can just unlink this entry.
+	 */
+	if (sctx->parent_root) {
+		ret = get_inode_info(sctx->parent_root, dir, NULL, &gen, NULL,
+				     NULL, NULL, NULL);
+		if (ret < 0 && ret != -ENOENT)
+			goto out;
+		if (ret) {
+			ret = 0;
+			goto out;
+		}
+		if (gen != dir_gen)
+			goto out;
+	}
+
 	ret = lookup_dir_item_inode(sctx->parent_root, dir, name, name_len,
 			&other_inode, &other_type);
 	if (ret < 0 && ret != -ENOENT)
@@ -1783,11 +1798,11 @@ static int did_overwrite_first_ref(struct send_ctx *sctx, u64 ino, u64 gen)
 	if (!sctx->parent_root)
 		goto out;
 
-	name = fs_path_alloc(sctx);
+	name = fs_path_alloc();
 	if (!name)
 		return -ENOMEM;
 
-	ret = get_first_ref(sctx, sctx->parent_root, ino, &dir, &dir_gen, name);
+	ret = get_first_ref(sctx->parent_root, ino, &dir, &dir_gen, name);
 	if (ret < 0)
 		goto out;
 
@@ -1795,7 +1810,7 @@ static int did_overwrite_first_ref(struct send_ctx *sctx, u64 ino, u64 gen)
 			name->start, fs_path_len(name));
 
 out:
-	fs_path_free(sctx, name);
+	fs_path_free(name);
 	return ret;
 }
 
@@ -1979,11 +1994,11 @@ static int __get_cur_name_and_parent(struct send_ctx *sctx,
 	 * send_root or parent_root for ref lookup.
 	 */
 	if (ino < sctx->send_progress)
-		ret = get_first_ref(sctx, sctx->send_root, ino,
-				parent_ino, parent_gen, dest);
+		ret = get_first_ref(sctx->send_root, ino,
+				    parent_ino, parent_gen, dest);
 	else
-		ret = get_first_ref(sctx, sctx->parent_root, ino,
-				parent_ino, parent_gen, dest);
+		ret = get_first_ref(sctx->parent_root, ino,
+				    parent_ino, parent_gen, dest);
 	if (ret < 0)
 		goto out;
 
@@ -2070,7 +2085,7 @@ static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen,
 	u64 parent_gen = 0;
 	int stop = 0;
 
-	name = fs_path_alloc(sctx);
+	name = fs_path_alloc();
 	if (!name) {
 		ret = -ENOMEM;
 		goto out;
@@ -2098,7 +2113,7 @@ static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen,
 	}
 
 out:
-	fs_path_free(sctx, name);
+	fs_path_free(name);
 	if (!ret)
 		fs_path_unreverse(dest);
 	return ret;
@@ -2263,7 +2278,7 @@ static int send_truncate(struct send_ctx *sctx, u64 ino, u64 gen, u64 size)
 
 verbose_printk("btrfs: send_truncate %llu size=%llu\n", ino, size);
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -2281,7 +2296,7 @@ verbose_printk("btrfs: send_truncate %llu size=%llu\n", ino, size);
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -2292,7 +2307,7 @@ static int send_chmod(struct send_ctx *sctx, u64 ino, u64 gen, u64 mode)
 
 verbose_printk("btrfs: send_chmod %llu mode=%llu\n", ino, mode);
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -2310,7 +2325,7 @@ verbose_printk("btrfs: send_chmod %llu mode=%llu\n", ino, mode);
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -2321,7 +2336,7 @@ static int send_chown(struct send_ctx *sctx, u64 ino, u64 gen, u64 uid, u64 gid)
 
 verbose_printk("btrfs: send_chown %llu uid=%llu, gid=%llu\n", ino, uid, gid);
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -2340,7 +2355,7 @@ verbose_printk("btrfs: send_chown %llu uid=%llu, gid=%llu\n", ino, uid, gid);
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -2356,7 +2371,7 @@ static int send_utimes(struct send_ctx *sctx, u64 ino, u64 gen)
 
 verbose_printk("btrfs: send_utimes %llu\n", ino);
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -2397,7 +2412,7 @@ verbose_printk("btrfs: send_utimes %llu\n", ino);
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	btrfs_free_path(path);
 	return ret;
 }
@@ -2418,7 +2433,7 @@ static int send_create_inode(struct send_ctx *sctx, u64 ino)
 
 verbose_printk("btrfs: send_create_inode %llu\n", ino);
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -2459,7 +2474,7 @@ verbose_printk("btrfs: send_create_inode %llu\n", ino);
 
 	if (S_ISLNK(mode)) {
 		fs_path_reset(p);
-		ret = read_symlink(sctx, sctx->send_root, ino, p);
+		ret = read_symlink(sctx->send_root, ino, p);
 		if (ret < 0)
 			goto out;
 		TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_LINK, p);
@@ -2476,7 +2491,7 @@ verbose_printk("btrfs: send_create_inode %llu\n", ino);
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -2524,7 +2539,8 @@ static int did_create_dir(struct send_ctx *sctx, u64 dir)
 		di = btrfs_item_ptr(eb, slot, struct btrfs_dir_item);
 		btrfs_dir_item_key_to_cpu(eb, di, &di_key);
 
-		if (di_key.objectid < sctx->send_progress) {
+		if (di_key.type != BTRFS_ROOT_ITEM_KEY &&
+		    di_key.objectid < sctx->send_progress) {
 			ret = 1;
 			goto out;
 		}
@@ -2586,7 +2602,6 @@ static int record_ref(struct list_head *head, u64 dir,
 		      u64 dir_gen, struct fs_path *path)
 {
 	struct recorded_ref *ref;
-	char *tmp;
 
 	ref = kmalloc(sizeof(*ref), GFP_NOFS);
 	if (!ref)
@@ -2596,32 +2611,42 @@ static int record_ref(struct list_head *head, u64 dir,
 	ref->dir_gen = dir_gen;
 	ref->full_path = path;
 
-	tmp = strrchr(ref->full_path->start, '/');
-	if (!tmp) {
-		ref->name_len = ref->full_path->end - ref->full_path->start;
-		ref->name = ref->full_path->start;
+	ref->name = (char *)kbasename(ref->full_path->start);
+	ref->name_len = ref->full_path->end - ref->name;
+	ref->dir_path = ref->full_path->start;
+	if (ref->name == ref->full_path->start)
 		ref->dir_path_len = 0;
-		ref->dir_path = ref->full_path->start;
-	} else {
-		tmp++;
-		ref->name_len = ref->full_path->end - tmp;
-		ref->name = tmp;
-		ref->dir_path = ref->full_path->start;
+	else
 		ref->dir_path_len = ref->full_path->end -
 				ref->full_path->start - 1 - ref->name_len;
-	}
 
 	list_add_tail(&ref->list, head);
 	return 0;
 }
 
-static void __free_recorded_refs(struct send_ctx *sctx, struct list_head *head)
+static int dup_ref(struct recorded_ref *ref, struct list_head *list)
+{
+	struct recorded_ref *new;
+
+	new = kmalloc(sizeof(*ref), GFP_NOFS);
+	if (!new)
+		return -ENOMEM;
+
+	new->dir = ref->dir;
+	new->dir_gen = ref->dir_gen;
+	new->full_path = NULL;
+	INIT_LIST_HEAD(&new->list);
+	list_add_tail(&new->list, list);
+	return 0;
+}
+
+static void __free_recorded_refs(struct list_head *head)
 {
 	struct recorded_ref *cur;
 
 	while (!list_empty(head)) {
 		cur = list_entry(head->next, struct recorded_ref, list);
-		fs_path_free(sctx, cur->full_path);
+		fs_path_free(cur->full_path);
 		list_del(&cur->list);
 		kfree(cur);
 	}
@@ -2629,8 +2654,8 @@ static void __free_recorded_refs(struct send_ctx *sctx, struct list_head *head)
 
 static void free_recorded_refs(struct send_ctx *sctx)
 {
-	__free_recorded_refs(sctx, &sctx->new_refs);
-	__free_recorded_refs(sctx, &sctx->deleted_refs);
+	__free_recorded_refs(&sctx->new_refs);
+	__free_recorded_refs(&sctx->deleted_refs);
 }
 
 /*
@@ -2644,7 +2669,7 @@ static int orphanize_inode(struct send_ctx *sctx, u64 ino, u64 gen,
 	int ret;
 	struct fs_path *orphan;
 
-	orphan = fs_path_alloc(sctx);
+	orphan = fs_path_alloc();
 	if (!orphan)
 		return -ENOMEM;
 
@@ -2655,7 +2680,7 @@ static int orphanize_inode(struct send_ctx *sctx, u64 ino, u64 gen,
 	ret = send_rename(sctx, path, orphan);
 
 out:
-	fs_path_free(sctx, orphan);
+	fs_path_free(orphan);
 	return ret;
 }
 
@@ -2729,9 +2754,7 @@ static int process_recorded_refs(struct send_ctx *sctx)
 	int ret = 0;
 	struct recorded_ref *cur;
 	struct recorded_ref *cur2;
-	struct ulist *check_dirs = NULL;
-	struct ulist_iterator uit;
-	struct ulist_node *un;
+	struct list_head check_dirs;
 	struct fs_path *valid_path = NULL;
 	u64 ow_inode = 0;
 	u64 ow_gen;
@@ -2745,19 +2768,14 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 	 * which is always '..'
 	 */
 	BUG_ON(sctx->cur_ino <= BTRFS_FIRST_FREE_OBJECTID);
+	INIT_LIST_HEAD(&check_dirs);
 
-	valid_path = fs_path_alloc(sctx);
+	valid_path = fs_path_alloc();
 	if (!valid_path) {
 		ret = -ENOMEM;
 		goto out;
 	}
 
-	check_dirs = ulist_alloc(GFP_NOFS);
-	if (!check_dirs) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
 	/*
 	 * First, check if the first ref of the current inode was overwritten
 	 * before. If yes, we know that the current inode was already orphanized
@@ -2843,9 +2861,9 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 		if (ret < 0)
 			goto out;
 		if (ret) {
-			ret = is_first_ref(sctx, sctx->parent_root,
-					ow_inode, cur->dir, cur->name,
-					cur->name_len);
+			ret = is_first_ref(sctx->parent_root,
+					   ow_inode, cur->dir, cur->name,
+					   cur->name_len);
 			if (ret < 0)
 				goto out;
 			if (ret) {
@@ -2894,8 +2912,7 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 					goto out;
 			}
 		}
-		ret = ulist_add(check_dirs, cur->dir, cur->dir_gen,
-				GFP_NOFS);
+		ret = dup_ref(cur, &check_dirs);
 		if (ret < 0)
 			goto out;
 	}
@@ -2923,8 +2940,7 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 		}
 
 		list_for_each_entry(cur, &sctx->deleted_refs, list) {
-			ret = ulist_add(check_dirs, cur->dir, cur->dir_gen,
-					GFP_NOFS);
+			ret = dup_ref(cur, &check_dirs);
 			if (ret < 0)
 				goto out;
 		}
@@ -2935,8 +2951,7 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 		 */
 		cur = list_entry(sctx->deleted_refs.next, struct recorded_ref,
 				list);
-		ret = ulist_add(check_dirs, cur->dir, cur->dir_gen,
-				GFP_NOFS);
+		ret = dup_ref(cur, &check_dirs);
 		if (ret < 0)
 			goto out;
 	} else if (!S_ISDIR(sctx->cur_inode_mode)) {
@@ -2956,12 +2971,10 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 				if (ret < 0)
 					goto out;
 			}
-			ret = ulist_add(check_dirs, cur->dir, cur->dir_gen,
-					GFP_NOFS);
+			ret = dup_ref(cur, &check_dirs);
 			if (ret < 0)
 				goto out;
 		}
-
 		/*
 		 * If the inode is still orphan, unlink the orphan. This may
 		 * happen when a previous inode did overwrite the first ref
@@ -2983,33 +2996,32 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 	 * deletion and if it's finally possible to perform the rmdir now.
 	 * We also update the inode stats of the parent dirs here.
 	 */
-	ULIST_ITER_INIT(&uit);
-	while ((un = ulist_next(check_dirs, &uit))) {
+	list_for_each_entry(cur, &check_dirs, list) {
 		/*
 		 * In case we had refs into dirs that were not processed yet,
 		 * we don't need to do the utime and rmdir logic for these dirs.
 		 * The dir will be processed later.
 		 */
-		if (un->val > sctx->cur_ino)
+		if (cur->dir > sctx->cur_ino)
 			continue;
 
-		ret = get_cur_inode_state(sctx, un->val, un->aux);
+		ret = get_cur_inode_state(sctx, cur->dir, cur->dir_gen);
 		if (ret < 0)
 			goto out;
 
 		if (ret == inode_state_did_create ||
 		    ret == inode_state_no_change) {
 			/* TODO delayed utimes */
-			ret = send_utimes(sctx, un->val, un->aux);
+			ret = send_utimes(sctx, cur->dir, cur->dir_gen);
 			if (ret < 0)
 				goto out;
 		} else if (ret == inode_state_did_delete) {
-			ret = can_rmdir(sctx, un->val, sctx->cur_ino);
+			ret = can_rmdir(sctx, cur->dir, sctx->cur_ino);
 			if (ret < 0)
 				goto out;
 			if (ret) {
-				ret = get_cur_path(sctx, un->val, un->aux,
-						valid_path);
+				ret = get_cur_path(sctx, cur->dir,
+						   cur->dir_gen, valid_path);
 				if (ret < 0)
 					goto out;
 				ret = send_rmdir(sctx, valid_path);
@@ -3022,9 +3034,9 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 	ret = 0;
 
 out:
+	__free_recorded_refs(&check_dirs);
 	free_recorded_refs(sctx);
-	ulist_free(check_dirs);
-	fs_path_free(sctx, valid_path);
+	fs_path_free(valid_path);
 	return ret;
 }
 
@@ -3037,7 +3049,7 @@ static int __record_new_ref(int num, u64 dir, int index,
 	struct fs_path *p;
 	u64 gen;
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -3057,7 +3069,7 @@ static int __record_new_ref(int num, u64 dir, int index,
 
 out:
 	if (ret)
-		fs_path_free(sctx, p);
+		fs_path_free(p);
 	return ret;
 }
 
@@ -3070,7 +3082,7 @@ static int __record_deleted_ref(int num, u64 dir, int index,
 	struct fs_path *p;
 	u64 gen;
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -3090,7 +3102,7 @@ static int __record_deleted_ref(int num, u64 dir, int index,
 
 out:
 	if (ret)
-		fs_path_free(sctx, p);
+		fs_path_free(p);
 	return ret;
 }
 
@@ -3098,8 +3110,8 @@ static int record_new_ref(struct send_ctx *sctx)
 {
 	int ret;
 
-	ret = iterate_inode_ref(sctx, sctx->send_root, sctx->left_path,
-			sctx->cmp_key, 0, __record_new_ref, sctx);
+	ret = iterate_inode_ref(sctx->send_root, sctx->left_path,
+				sctx->cmp_key, 0, __record_new_ref, sctx);
 	if (ret < 0)
 		goto out;
 	ret = 0;
@@ -3112,8 +3124,8 @@ static int record_deleted_ref(struct send_ctx *sctx)
 {
 	int ret;
 
-	ret = iterate_inode_ref(sctx, sctx->parent_root, sctx->right_path,
-			sctx->cmp_key, 0, __record_deleted_ref, sctx);
+	ret = iterate_inode_ref(sctx->parent_root, sctx->right_path,
+				sctx->cmp_key, 0, __record_deleted_ref, sctx);
 	if (ret < 0)
 		goto out;
 	ret = 0;
@@ -3124,6 +3136,8 @@ out:
 
 struct find_ref_ctx {
 	u64 dir;
+	u64 dir_gen;
+	struct btrfs_root *root;
 	struct fs_path *name;
 	int found_idx;
 };
@@ -3133,29 +3147,42 @@ static int __find_iref(int num, u64 dir, int index,
 		       void *ctx_)
 {
 	struct find_ref_ctx *ctx = ctx_;
+	u64 dir_gen;
+	int ret;
 
 	if (dir == ctx->dir && fs_path_len(name) == fs_path_len(ctx->name) &&
 	    strncmp(name->start, ctx->name->start, fs_path_len(name)) == 0) {
+		/*
+		 * To avoid doing extra lookups we'll only do this if everything
+		 * else matches.
+		 */
+		ret = get_inode_info(ctx->root, dir, NULL, &dir_gen, NULL,
+				     NULL, NULL, NULL);
+		if (ret)
+			return ret;
+		if (dir_gen != ctx->dir_gen)
+			return 0;
 		ctx->found_idx = num;
 		return 1;
 	}
 	return 0;
 }
 
-static int find_iref(struct send_ctx *sctx,
-		     struct btrfs_root *root,
+static int find_iref(struct btrfs_root *root,
 		     struct btrfs_path *path,
 		     struct btrfs_key *key,
-		     u64 dir, struct fs_path *name)
+		     u64 dir, u64 dir_gen, struct fs_path *name)
 {
 	int ret;
 	struct find_ref_ctx ctx;
 
 	ctx.dir = dir;
 	ctx.name = name;
+	ctx.dir_gen = dir_gen;
 	ctx.found_idx = -1;
+	ctx.root = root;
 
-	ret = iterate_inode_ref(sctx, root, path, key, 0, __find_iref, &ctx);
+	ret = iterate_inode_ref(root, path, key, 0, __find_iref, &ctx);
 	if (ret < 0)
 		return ret;
 
@@ -3169,11 +3196,17 @@ static int __record_changed_new_ref(int num, u64 dir, int index,
 				    struct fs_path *name,
 				    void *ctx)
 {
+	u64 dir_gen;
 	int ret;
 	struct send_ctx *sctx = ctx;
 
-	ret = find_iref(sctx, sctx->parent_root, sctx->right_path,
-			sctx->cmp_key, dir, name);
+	ret = get_inode_info(sctx->send_root, dir, NULL, &dir_gen, NULL,
+			     NULL, NULL, NULL);
+	if (ret)
+		return ret;
+
+	ret = find_iref(sctx->parent_root, sctx->right_path,
+			sctx->cmp_key, dir, dir_gen, name);
 	if (ret == -ENOENT)
 		ret = __record_new_ref(num, dir, index, name, sctx);
 	else if (ret > 0)
@@ -3186,11 +3219,17 @@ static int __record_changed_deleted_ref(int num, u64 dir, int index,
 					struct fs_path *name,
 					void *ctx)
 {
+	u64 dir_gen;
 	int ret;
 	struct send_ctx *sctx = ctx;
 
-	ret = find_iref(sctx, sctx->send_root, sctx->left_path, sctx->cmp_key,
-			dir, name);
+	ret = get_inode_info(sctx->parent_root, dir, NULL, &dir_gen, NULL,
+			     NULL, NULL, NULL);
+	if (ret)
+		return ret;
+
+	ret = find_iref(sctx->send_root, sctx->left_path, sctx->cmp_key,
+			dir, dir_gen, name);
 	if (ret == -ENOENT)
 		ret = __record_deleted_ref(num, dir, index, name, sctx);
 	else if (ret > 0)
@@ -3203,11 +3242,11 @@ static int record_changed_ref(struct send_ctx *sctx)
 {
 	int ret = 0;
 
-	ret = iterate_inode_ref(sctx, sctx->send_root, sctx->left_path,
+	ret = iterate_inode_ref(sctx->send_root, sctx->left_path,
 			sctx->cmp_key, 0, __record_changed_new_ref, sctx);
 	if (ret < 0)
 		goto out;
-	ret = iterate_inode_ref(sctx, sctx->parent_root, sctx->right_path,
+	ret = iterate_inode_ref(sctx->parent_root, sctx->right_path,
 			sctx->cmp_key, 0, __record_changed_deleted_ref, sctx);
 	if (ret < 0)
 		goto out;
@@ -3266,8 +3305,7 @@ static int process_all_refs(struct send_ctx *sctx,
 		     found_key.type != BTRFS_INODE_EXTREF_KEY))
 			break;
 
-		ret = iterate_inode_ref(sctx, root, path, &found_key, 0, cb,
-				sctx);
+		ret = iterate_inode_ref(root, path, &found_key, 0, cb, sctx);
 		btrfs_release_path(path);
 		if (ret < 0)
 			goto out;
@@ -3335,7 +3373,7 @@ static int __process_new_xattr(int num, struct btrfs_key *di_key,
 	struct fs_path *p;
 	posix_acl_xattr_header dummy_acl;
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -3362,7 +3400,7 @@ static int __process_new_xattr(int num, struct btrfs_key *di_key,
 	ret = send_set_xattr(sctx, p, name, name_len, data, data_len);
 
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -3375,7 +3413,7 @@ static int __process_deleted_xattr(int num, struct btrfs_key *di_key,
 	struct send_ctx *sctx = ctx;
 	struct fs_path *p;
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -3386,7 +3424,7 @@ static int __process_deleted_xattr(int num, struct btrfs_key *di_key,
 	ret = send_remove_xattr(sctx, p, name, name_len);
 
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -3394,8 +3432,8 @@ static int process_new_xattr(struct send_ctx *sctx)
 {
 	int ret = 0;
 
-	ret = iterate_dir_item(sctx, sctx->send_root, sctx->left_path,
-			sctx->cmp_key, __process_new_xattr, sctx);
+	ret = iterate_dir_item(sctx->send_root, sctx->left_path,
+			       sctx->cmp_key, __process_new_xattr, sctx);
 
 	return ret;
 }
@@ -3404,8 +3442,8 @@ static int process_deleted_xattr(struct send_ctx *sctx)
 {
 	int ret;
 
-	ret = iterate_dir_item(sctx, sctx->parent_root, sctx->right_path,
-			sctx->cmp_key, __process_deleted_xattr, sctx);
+	ret = iterate_dir_item(sctx->parent_root, sctx->right_path,
+			       sctx->cmp_key, __process_deleted_xattr, sctx);
 
 	return ret;
 }
@@ -3429,17 +3467,15 @@ static int __find_xattr(int num, struct btrfs_key *di_key,
 	    strncmp(name, ctx->name, name_len) == 0) {
 		ctx->found_idx = num;
 		ctx->found_data_len = data_len;
-		ctx->found_data = kmalloc(data_len, GFP_NOFS);
+		ctx->found_data = kmemdup(data, data_len, GFP_NOFS);
 		if (!ctx->found_data)
 			return -ENOMEM;
-		memcpy(ctx->found_data, data, data_len);
 		return 1;
 	}
 	return 0;
 }
 
-static int find_xattr(struct send_ctx *sctx,
-		      struct btrfs_root *root,
+static int find_xattr(struct btrfs_root *root,
 		      struct btrfs_path *path,
 		      struct btrfs_key *key,
 		      const char *name, int name_len,
@@ -3454,7 +3490,7 @@ static int find_xattr(struct send_ctx *sctx,
 	ctx.found_data = NULL;
 	ctx.found_data_len = 0;
 
-	ret = iterate_dir_item(sctx, root, path, key, __find_xattr, &ctx);
+	ret = iterate_dir_item(root, path, key, __find_xattr, &ctx);
 	if (ret < 0)
 		return ret;
 
@@ -3480,9 +3516,9 @@ static int __process_changed_new_xattr(int num, struct btrfs_key *di_key,
 	char *found_data = NULL;
 	int found_data_len  = 0;
 
-	ret = find_xattr(sctx, sctx->parent_root, sctx->right_path,
-			sctx->cmp_key, name, name_len, &found_data,
-			&found_data_len);
+	ret = find_xattr(sctx->parent_root, sctx->right_path,
+			 sctx->cmp_key, name, name_len, &found_data,
+			 &found_data_len);
 	if (ret == -ENOENT) {
 		ret = __process_new_xattr(num, di_key, name, name_len, data,
 				data_len, type, ctx);
@@ -3508,8 +3544,8 @@ static int __process_changed_deleted_xattr(int num, struct btrfs_key *di_key,
 	int ret;
 	struct send_ctx *sctx = ctx;
 
-	ret = find_xattr(sctx, sctx->send_root, sctx->left_path, sctx->cmp_key,
-			name, name_len, NULL, NULL);
+	ret = find_xattr(sctx->send_root, sctx->left_path, sctx->cmp_key,
+			 name, name_len, NULL, NULL);
 	if (ret == -ENOENT)
 		ret = __process_deleted_xattr(num, di_key, name, name_len, data,
 				data_len, type, ctx);
@@ -3523,11 +3559,11 @@ static int process_changed_xattr(struct send_ctx *sctx)
 {
 	int ret = 0;
 
-	ret = iterate_dir_item(sctx, sctx->send_root, sctx->left_path,
+	ret = iterate_dir_item(sctx->send_root, sctx->left_path,
 			sctx->cmp_key, __process_changed_new_xattr, sctx);
 	if (ret < 0)
 		goto out;
-	ret = iterate_dir_item(sctx, sctx->parent_root, sctx->right_path,
+	ret = iterate_dir_item(sctx->parent_root, sctx->right_path,
 			sctx->cmp_key, __process_changed_deleted_xattr, sctx);
 
 out:
@@ -3572,8 +3608,8 @@ static int process_all_new_xattrs(struct send_ctx *sctx)
 			goto out;
 		}
 
-		ret = iterate_dir_item(sctx, root, path, &found_key,
-				__process_new_xattr, sctx);
+		ret = iterate_dir_item(root, path, &found_key,
+				       __process_new_xattr, sctx);
 		if (ret < 0)
 			goto out;
 
@@ -3598,7 +3634,7 @@ static int send_write(struct send_ctx *sctx, u64 offset, u32 len)
 	int num_read = 0;
 	mm_segment_t old_fs;
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -3640,7 +3676,7 @@ verbose_printk("btrfs: send_write offset=%llu, len=%d\n", offset, len);
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	set_fs(old_fs);
 	if (ret < 0)
 		return ret;
@@ -3663,7 +3699,7 @@ verbose_printk("btrfs: send_clone offset=%llu, len=%d, clone_root=%llu, "
 		clone_root->root->objectid, clone_root->ino,
 		clone_root->offset);
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -3686,8 +3722,7 @@ verbose_printk("btrfs: send_clone offset=%llu, len=%d, clone_root=%llu, "
 			goto out;
 		ret = get_cur_path(sctx, clone_root->ino, gen, p);
 	} else {
-		ret = get_inode_path(sctx, clone_root->root,
-				clone_root->ino, p);
+		ret = get_inode_path(clone_root->root, clone_root->ino, p);
 	}
 	if (ret < 0)
 		goto out;
@@ -3704,7 +3739,7 @@ verbose_printk("btrfs: send_clone offset=%llu, len=%d, clone_root=%llu, "
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -3717,7 +3752,7 @@ static int send_update_extent(struct send_ctx *sctx,
 	int ret = 0;
 	struct fs_path *p;
 
-	p = fs_path_alloc(sctx);
+	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
 
@@ -3737,7 +3772,7 @@ static int send_update_extent(struct send_ctx *sctx,
 
 tlv_put_failure:
 out:
-	fs_path_free(sctx, p);
+	fs_path_free(p);
 	return ret;
 }
 
@@ -3879,7 +3914,8 @@ static int is_extent_unchanged(struct send_ctx *sctx,
 	btrfs_item_key_to_cpu(eb, &found_key, slot);
 	if (found_key.objectid != key.objectid ||
 	    found_key.type != key.type) {
-		ret = 0;
+		/* If we're a hole then just pretend nothing changed */
+		ret = (left_disknr) ? 0 : 1;
 		goto out;
 	}
 
@@ -3905,7 +3941,8 @@ static int is_extent_unchanged(struct send_ctx *sctx,
 		 * This may only happen on the first iteration.
 		 */
 		if (found_key.offset + right_len <= ekey->offset) {
-			ret = 0;
+			/* If we're a hole just pretend nothing changed */
+			ret = (left_disknr) ? 0 : 1;
 			goto out;
 		}
 
@@ -3970,8 +4007,8 @@ static int process_extent(struct send_ctx *sctx,
 			  struct btrfs_path *path,
 			  struct btrfs_key *key)
 {
-	int ret = 0;
 	struct clone_root *found_clone = NULL;
+	int ret = 0;
 
 	if (S_ISLNK(sctx->cur_inode_mode))
 		return 0;
@@ -3984,6 +4021,32 @@ static int process_extent(struct send_ctx *sctx,
 			ret = 0;
 			goto out;
 		}
+	} else {
+		struct btrfs_file_extent_item *ei;
+		u8 type;
+
+		ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
+				    struct btrfs_file_extent_item);
+		type = btrfs_file_extent_type(path->nodes[0], ei);
+		if (type == BTRFS_FILE_EXTENT_PREALLOC ||
+		    type == BTRFS_FILE_EXTENT_REG) {
+			/*
+			 * The send spec does not have a prealloc command yet,
+			 * so just leave a hole for prealloc'ed extents until
+			 * we have enough commands queued up to justify rev'ing
+			 * the send spec.
+			 */
+			if (type == BTRFS_FILE_EXTENT_PREALLOC) {
+				ret = 0;
+				goto out;
+			}
+
+			/* Have a hole, just skip it. */
+			if (btrfs_file_extent_disk_bytenr(path->nodes[0], ei) == 0) {
+				ret = 0;
+				goto out;
+			}
+		}
 	}
 
 	ret = find_extent_clone(sctx, path, key->objectid, key->offset,
@@ -4371,6 +4434,64 @@ static int changed_extent(struct send_ctx *sctx,
 	return ret;
 }
 
+static int dir_changed(struct send_ctx *sctx, u64 dir)
+{
+	u64 orig_gen, new_gen;
+	int ret;
+
+	ret = get_inode_info(sctx->send_root, dir, NULL, &new_gen, NULL, NULL,
+			     NULL, NULL);
+	if (ret)
+		return ret;
+
+	ret = get_inode_info(sctx->parent_root, dir, NULL, &orig_gen, NULL,
+			     NULL, NULL, NULL);
+	if (ret)
+		return ret;
+
+	return (orig_gen != new_gen) ? 1 : 0;
+}
+
+static int compare_refs(struct send_ctx *sctx, struct btrfs_path *path,
+			struct btrfs_key *key)
+{
+	struct btrfs_inode_extref *extref;
+	struct extent_buffer *leaf;
+	u64 dirid = 0, last_dirid = 0;
+	unsigned long ptr;
+	u32 item_size;
+	u32 cur_offset = 0;
+	int ref_name_len;
+	int ret = 0;
+
+	/* Easy case, just check this one dirid */
+	if (key->type == BTRFS_INODE_REF_KEY) {
+		dirid = key->offset;
+
+		ret = dir_changed(sctx, dirid);
+		goto out;
+	}
+
+	leaf = path->nodes[0];
+	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+	while (cur_offset < item_size) {
+		extref = (struct btrfs_inode_extref *)(ptr +
+						       cur_offset);
+		dirid = btrfs_inode_extref_parent(leaf, extref);
+		ref_name_len = btrfs_inode_extref_name_len(leaf, extref);
+		cur_offset += ref_name_len + sizeof(*extref);
+		if (dirid == last_dirid)
+			continue;
+		ret = dir_changed(sctx, dirid);
+		if (ret)
+			break;
+		last_dirid = dirid;
+	}
+out:
+	return ret;
+}
+
 /*
  * Updates compare related fields in sctx and simply forwards to the actual
  * changed_xxx functions.
@@ -4386,6 +4507,19 @@ static int changed_cb(struct btrfs_root *left_root,
 	int ret = 0;
 	struct send_ctx *sctx = ctx;
 
+	if (result == BTRFS_COMPARE_TREE_SAME) {
+		if (key->type != BTRFS_INODE_REF_KEY &&
+		    key->type != BTRFS_INODE_EXTREF_KEY)
+			return 0;
+		ret = compare_refs(sctx, left_path, key);
+		if (!ret)
+			return 0;
+		if (ret < 0)
+			return ret;
+		result = BTRFS_COMPARE_TREE_CHANGED;
+		ret = 0;
+	}
+
 	sctx->left_path = left_path;
 	sctx->right_path = right_path;
 	sctx->cmp_key = key;
@@ -4579,6 +4713,41 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
 	send_root = BTRFS_I(file_inode(mnt_file))->root;
 	fs_info = send_root->fs_info;
 
+	/*
+	 * This is done when we lookup the root, it should already be complete
+	 * by the time we get here.
+	 */
+	WARN_ON(send_root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE);
+
+	/*
+	 * If we just created this root we need to make sure that the orphan
+	 * cleanup has been done and committed since we search the commit root,
+	 * so check its commit root transid with our otransid and if they match
+	 * commit the transaction to make sure everything is updated.
+	 */
+	down_read(&send_root->fs_info->extent_commit_sem);
+	if (btrfs_header_generation(send_root->commit_root) ==
+	    btrfs_root_otransid(&send_root->root_item)) {
+		struct btrfs_trans_handle *trans;
+
+		up_read(&send_root->fs_info->extent_commit_sem);
+
+		trans = btrfs_attach_transaction_barrier(send_root);
+		if (IS_ERR(trans)) {
+			if (PTR_ERR(trans) != -ENOENT) {
+				ret = PTR_ERR(trans);
+				goto out;
+			}
+			/* ENOENT means theres no transaction */
+		} else {
+			ret = btrfs_commit_transaction(trans, send_root);
+			if (ret)
+				goto out;
+		}
+	} else {
+		up_read(&send_root->fs_info->extent_commit_sem);
+	}
+
 	arg = memdup_user(arg_, sizeof(*arg));
 	if (IS_ERR(arg)) {
 		ret = PTR_ERR(arg);
@@ -4663,10 +4832,6 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
 			key.type = BTRFS_ROOT_ITEM_KEY;
 			key.offset = (u64)-1;
 			clone_root = btrfs_read_fs_root_no_name(fs_info, &key);
-			if (!clone_root) {
-				ret = -EINVAL;
-				goto out;
-			}
 			if (IS_ERR(clone_root)) {
 				ret = PTR_ERR(clone_root);
 				goto out;
@@ -4682,8 +4847,8 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
 		key.type = BTRFS_ROOT_ITEM_KEY;
 		key.offset = (u64)-1;
 		sctx->parent_root = btrfs_read_fs_root_no_name(fs_info, &key);
-		if (!sctx->parent_root) {
-			ret = -EINVAL;
+		if (IS_ERR(sctx->parent_root)) {
+			ret = PTR_ERR(sctx->parent_root);
 			goto out;
 		}
 	}
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index f0857e092a3c..e913328d0f2a 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -51,12 +51,13 @@
 #include "print-tree.h"
 #include "xattr.h"
 #include "volumes.h"
-#include "version.h"
 #include "export.h"
 #include "compression.h"
 #include "rcu-string.h"
 #include "dev-replace.h"
 #include "free-space-cache.h"
+#include "backref.h"
+#include "tests/btrfs-tests.h"
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/btrfs.h>
@@ -266,6 +267,9 @@ void __btrfs_abort_transaction(struct btrfs_trans_handle *trans,
 		return;
 	}
 	ACCESS_ONCE(trans->transaction->aborted) = errno;
+	/* Wake up anybody who may be waiting on this transaction */
+	wake_up(&root->fs_info->transaction_wait);
+	wake_up(&root->fs_info->transaction_blocked_wait);
 	__btrfs_std_error(root->fs_info, function, line, errno, NULL);
 }
 /*
@@ -318,14 +322,15 @@ enum {
 	Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, Opt_inode_cache,
 	Opt_no_space_cache, Opt_recovery, Opt_skip_balance,
 	Opt_check_integrity, Opt_check_integrity_including_extent_data,
-	Opt_check_integrity_print_mask, Opt_fatal_errors,
+	Opt_check_integrity_print_mask, Opt_fatal_errors, Opt_rescan_uuid_tree,
+	Opt_commit_interval,
 	Opt_err,
 };
 
 static match_table_t tokens = {
 	{Opt_degraded, "degraded"},
 	{Opt_subvol, "subvol=%s"},
-	{Opt_subvolid, "subvolid=%d"},
+	{Opt_subvolid, "subvolid=%s"},
 	{Opt_device, "device=%s"},
 	{Opt_nodatasum, "nodatasum"},
 	{Opt_nodatacow, "nodatacow"},
@@ -358,7 +363,9 @@ static match_table_t tokens = {
 	{Opt_check_integrity, "check_int"},
 	{Opt_check_integrity_including_extent_data, "check_int_data"},
 	{Opt_check_integrity_print_mask, "check_int_print_mask=%d"},
+	{Opt_rescan_uuid_tree, "rescan_uuid_tree"},
 	{Opt_fatal_errors, "fatal_errors=%s"},
+	{Opt_commit_interval, "commit=%d"},
 	{Opt_err, NULL},
 };
 
@@ -494,10 +501,15 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 			btrfs_set_opt(info->mount_opt, NOBARRIER);
 			break;
 		case Opt_thread_pool:
-			intarg = 0;
-			match_int(&args[0], &intarg);
-			if (intarg)
+			ret = match_int(&args[0], &intarg);
+			if (ret) {
+				goto out;
+			} else if (intarg > 0) {
 				info->thread_pool_size = intarg;
+			} else {
+				ret = -EINVAL;
+				goto out;
+			}
 			break;
 		case Opt_max_inline:
 			num = match_strdup(&args[0]);
@@ -511,7 +523,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 						root->sectorsize);
 				}
 				printk(KERN_INFO "btrfs: max_inline at %llu\n",
-					(unsigned long long)info->max_inline);
+					info->max_inline);
+			} else {
+				ret = -ENOMEM;
+				goto out;
 			}
 			break;
 		case Opt_alloc_start:
@@ -523,7 +538,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 				kfree(num);
 				printk(KERN_INFO
 					"btrfs: allocations start at %llu\n",
-					(unsigned long long)info->alloc_start);
+					info->alloc_start);
+			} else {
+				ret = -ENOMEM;
+				goto out;
 			}
 			break;
 		case Opt_noacl:
@@ -538,12 +556,16 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 			btrfs_set_opt(info->mount_opt, FLUSHONCOMMIT);
 			break;
 		case Opt_ratio:
-			intarg = 0;
-			match_int(&args[0], &intarg);
-			if (intarg) {
+			ret = match_int(&args[0], &intarg);
+			if (ret) {
+				goto out;
+			} else if (intarg >= 0) {
 				info->metadata_ratio = intarg;
 				printk(KERN_INFO "btrfs: metadata ratio %d\n",
 				       info->metadata_ratio);
+			} else {
+				ret = -EINVAL;
+				goto out;
 			}
 			break;
 		case Opt_discard:
@@ -552,6 +574,9 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 		case Opt_space_cache:
 			btrfs_set_opt(info->mount_opt, SPACE_CACHE);
 			break;
+		case Opt_rescan_uuid_tree:
+			btrfs_set_opt(info->mount_opt, RESCAN_UUID_TREE);
+			break;
 		case Opt_no_space_cache:
 			printk(KERN_INFO "btrfs: disabling disk space caching\n");
 			btrfs_clear_opt(info->mount_opt, SPACE_CACHE);
@@ -594,13 +619,17 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY);
 			break;
 		case Opt_check_integrity_print_mask:
-			intarg = 0;
-			match_int(&args[0], &intarg);
-			if (intarg) {
+			ret = match_int(&args[0], &intarg);
+			if (ret) {
+				goto out;
+			} else if (intarg >= 0) {
 				info->check_integrity_print_mask = intarg;
 				printk(KERN_INFO "btrfs:"
 				       " check_integrity_print_mask 0x%x\n",
 				       info->check_integrity_print_mask);
+			} else {
+				ret = -EINVAL;
+				goto out;
 			}
 			break;
 #else
@@ -624,6 +653,29 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 				goto out;
 			}
 			break;
+		case Opt_commit_interval:
+			intarg = 0;
+			ret = match_int(&args[0], &intarg);
+			if (ret < 0) {
+				printk(KERN_ERR
+					"btrfs: invalid commit interval\n");
+				ret = -EINVAL;
+				goto out;
+			}
+			if (intarg > 0) {
+				if (intarg > 300) {
+					printk(KERN_WARNING
+					    "btrfs: excessive commit interval %d\n",
+							intarg);
+				}
+				info->commit_interval = intarg;
+			} else {
+				printk(KERN_INFO
+				    "btrfs: using default commit interval %ds\n",
+				    BTRFS_DEFAULT_COMMIT_INTERVAL);
+				info->commit_interval = BTRFS_DEFAULT_COMMIT_INTERVAL;
+			}
+			break;
 		case Opt_err:
 			printk(KERN_INFO "btrfs: unrecognized mount option "
 			       "'%s'\n", p);
@@ -652,8 +704,8 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags,
 {
 	substring_t args[MAX_OPT_ARGS];
 	char *device_name, *opts, *orig, *p;
+	char *num = NULL;
 	int error = 0;
-	int intarg;
 
 	if (!options)
 		return 0;
@@ -677,17 +729,23 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags,
 		case Opt_subvol:
 			kfree(*subvol_name);
 			*subvol_name = match_strdup(&args[0]);
+			if (!*subvol_name) {
+				error = -ENOMEM;
+				goto out;
+			}
 			break;
 		case Opt_subvolid:
-			intarg = 0;
-			error = match_int(&args[0], &intarg);
-			if (!error) {
+			num = match_strdup(&args[0]);
+			if (num) {
+				*subvol_objectid = memparse(num, NULL);
+				kfree(num);
 				/* we want the original fs_tree */
-				if (!intarg)
+				if (!*subvol_objectid)
 					*subvol_objectid =
 						BTRFS_FS_TREE_OBJECTID;
-				else
-					*subvol_objectid = intarg;
+			} else {
+				error = -EINVAL;
+				goto out;
 			}
 			break;
 		case Opt_subvolrootid:
@@ -776,9 +834,6 @@ find_root:
 	if (IS_ERR(new_root))
 		return ERR_CAST(new_root);
 
-	if (btrfs_root_refs(&new_root->root_item) == 0)
-		return ERR_PTR(-ENOENT);
-
 	dir_id = btrfs_root_dirid(&new_root->root_item);
 setup_root:
 	location.objectid = dir_id;
@@ -866,7 +921,7 @@ int btrfs_sync_fs(struct super_block *sb, int wait)
 		return 0;
 	}
 
-	btrfs_wait_ordered_extents(root, 1);
+	btrfs_wait_all_ordered_extents(fs_info);
 
 	trans = btrfs_attach_transaction_barrier(root);
 	if (IS_ERR(trans)) {
@@ -893,11 +948,9 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
 	if (btrfs_test_opt(root, NOBARRIER))
 		seq_puts(seq, ",nobarrier");
 	if (info->max_inline != 8192 * 1024)
-		seq_printf(seq, ",max_inline=%llu",
-			   (unsigned long long)info->max_inline);
+		seq_printf(seq, ",max_inline=%llu", info->max_inline);
 	if (info->alloc_start != 0)
-		seq_printf(seq, ",alloc_start=%llu",
-			   (unsigned long long)info->alloc_start);
+		seq_printf(seq, ",alloc_start=%llu", info->alloc_start);
 	if (info->thread_pool_size !=  min_t(unsigned long,
 					     num_online_cpus() + 2, 8))
 		seq_printf(seq, ",thread_pool=%d", info->thread_pool_size);
@@ -929,6 +982,8 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
 		seq_puts(seq, ",space_cache");
 	else
 		seq_puts(seq, ",nospace_cache");
+	if (btrfs_test_opt(root, RESCAN_UUID_TREE))
+		seq_puts(seq, ",rescan_uuid_tree");
 	if (btrfs_test_opt(root, CLEAR_CACHE))
 		seq_puts(seq, ",clear_cache");
 	if (btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED))
@@ -941,8 +996,24 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
 		seq_puts(seq, ",inode_cache");
 	if (btrfs_test_opt(root, SKIP_BALANCE))
 		seq_puts(seq, ",skip_balance");
+	if (btrfs_test_opt(root, RECOVERY))
+		seq_puts(seq, ",recovery");
+#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
+	if (btrfs_test_opt(root, CHECK_INTEGRITY_INCLUDING_EXTENT_DATA))
+		seq_puts(seq, ",check_int_data");
+	else if (btrfs_test_opt(root, CHECK_INTEGRITY))
+		seq_puts(seq, ",check_int");
+	if (info->check_integrity_print_mask)
+		seq_printf(seq, ",check_int_print_mask=%d",
+				info->check_integrity_print_mask);
+#endif
+	if (info->metadata_ratio)
+		seq_printf(seq, ",metadata_ratio=%d",
+				info->metadata_ratio);
 	if (btrfs_test_opt(root, PANIC_ON_FATAL_ERROR))
 		seq_puts(seq, ",fatal_errors=panic");
+	if (info->commit_interval != BTRFS_DEFAULT_COMMIT_INTERVAL)
+		seq_printf(seq, ",commit=%d", info->commit_interval);
 	return 0;
 }
 
@@ -1269,6 +1340,12 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 		if (ret)
 			goto restore;
 	} else {
+		if (test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) {
+			btrfs_err(fs_info,
+				"Remounting read-write after error is not allowed\n");
+			ret = -EINVAL;
+			goto restore;
+		}
 		if (fs_info->fs_devices->rw_devices == 0) {
 			ret = -EACCES;
 			goto restore;
@@ -1306,6 +1383,16 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 			pr_warn("btrfs: failed to resume dev_replace\n");
 			goto restore;
 		}
+
+		if (!fs_info->uuid_root) {
+			pr_info("btrfs: creating UUID tree\n");
+			ret = btrfs_create_uuid_tree(fs_info);
+			if (ret) {
+				pr_warn("btrfs: failed to create the uuid tree"
+					"%d\n", ret);
+				goto restore;
+			}
+		}
 		sb->s_flags &= ~MS_RDONLY;
 	}
 out:
@@ -1685,6 +1772,26 @@ static void btrfs_interface_exit(void)
 		printk(KERN_INFO "btrfs: misc_deregister failed for control device\n");
 }
 
+static void btrfs_print_info(void)
+{
+	printk(KERN_INFO "Btrfs loaded"
+#ifdef CONFIG_BTRFS_DEBUG
+			", debug=on"
+#endif
+#ifdef CONFIG_BTRFS_ASSERT
+			", assert=on"
+#endif
+#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
+			", integrity-checker=on"
+#endif
+			"\n");
+}
+
+static int btrfs_run_sanity_tests(void)
+{
+	return btrfs_test_free_space_cache();
+}
+
 static int __init init_btrfs_fs(void)
 {
 	int err;
@@ -1723,25 +1830,32 @@ static int __init init_btrfs_fs(void)
 	if (err)
 		goto free_auto_defrag;
 
+	err = btrfs_prelim_ref_init();
+	if (err)
+		goto free_prelim_ref;
+
 	err = btrfs_interface_init();
 	if (err)
 		goto free_delayed_ref;
 
-	err = register_filesystem(&btrfs_fs_type);
+	btrfs_init_lockdep();
+
+	btrfs_print_info();
+
+	err = btrfs_run_sanity_tests();
 	if (err)
 		goto unregister_ioctl;
 
-	btrfs_init_lockdep();
-
-#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
-	btrfs_test_free_space_cache();
-#endif
+	err = register_filesystem(&btrfs_fs_type);
+	if (err)
+		goto unregister_ioctl;
 
-	printk(KERN_INFO "%s loaded\n", BTRFS_BUILD_VERSION);
 	return 0;
 
 unregister_ioctl:
 	btrfs_interface_exit();
+free_prelim_ref:
+	btrfs_prelim_ref_exit();
 free_delayed_ref:
 	btrfs_delayed_ref_exit();
 free_auto_defrag:
@@ -1768,6 +1882,7 @@ static void __exit exit_btrfs_fs(void)
 	btrfs_delayed_ref_exit();
 	btrfs_auto_defrag_exit();
 	btrfs_delayed_inode_exit();
+	btrfs_prelim_ref_exit();
 	ordered_data_exit();
 	extent_map_exit();
 	extent_io_exit();
diff --git a/fs/btrfs/tests/btrfs-tests.h b/fs/btrfs/tests/btrfs-tests.h
new file mode 100644
index 000000000000..580877625776
--- /dev/null
+++ b/fs/btrfs/tests/btrfs-tests.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright (C) 2013 Fusion IO.  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 v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_TESTS
+#define __BTRFS_TESTS
+
+#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
+
+#define test_msg(fmt, ...) pr_info("btrfs: selftest: " fmt, ##__VA_ARGS__)
+
+int btrfs_test_free_space_cache(void);
+#else
+static inline int btrfs_test_free_space_cache(void)
+{
+	return 0;
+}
+#endif
+
+#endif
diff --git a/fs/btrfs/tests/free-space-tests.c b/fs/btrfs/tests/free-space-tests.c
new file mode 100644
index 000000000000..6fc82010dc15
--- /dev/null
+++ b/fs/btrfs/tests/free-space-tests.c
@@ -0,0 +1,395 @@
+/*
+ * Copyright (C) 2013 Fusion IO.  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 v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/slab.h>
+#include "btrfs-tests.h"
+#include "../ctree.h"
+#include "../free-space-cache.h"
+
+#define BITS_PER_BITMAP		(PAGE_CACHE_SIZE * 8)
+static struct btrfs_block_group_cache *init_test_block_group(void)
+{
+	struct btrfs_block_group_cache *cache;
+
+	cache = kzalloc(sizeof(*cache), GFP_NOFS);
+	if (!cache)
+		return NULL;
+	cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl),
+					GFP_NOFS);
+	if (!cache->free_space_ctl) {
+		kfree(cache);
+		return NULL;
+	}
+
+	cache->key.objectid = 0;
+	cache->key.offset = 1024 * 1024 * 1024;
+	cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
+	cache->sectorsize = 4096;
+
+	spin_lock_init(&cache->lock);
+	INIT_LIST_HEAD(&cache->list);
+	INIT_LIST_HEAD(&cache->cluster_list);
+	INIT_LIST_HEAD(&cache->new_bg_list);
+
+	btrfs_init_free_space_ctl(cache);
+
+	return cache;
+}
+
+/*
+ * This test just does basic sanity checking, making sure we can add an exten
+ * entry and remove space from either end and the middle, and make sure we can
+ * remove space that covers adjacent extent entries.
+ */
+static int test_extents(struct btrfs_block_group_cache *cache)
+{
+	int ret = 0;
+
+	test_msg("Running extent only tests\n");
+
+	/* First just make sure we can remove an entire entry */
+	ret = btrfs_add_free_space(cache, 0, 4 * 1024 * 1024);
+	if (ret) {
+		test_msg("Error adding initial extents %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 0, 4 * 1024 * 1024);
+	if (ret) {
+		test_msg("Error removing extent %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, 0, 4 * 1024 * 1024)) {
+		test_msg("Full remove left some lingering space\n");
+		return -1;
+	}
+
+	/* Ok edge and middle cases now */
+	ret = btrfs_add_free_space(cache, 0, 4 * 1024 * 1024);
+	if (ret) {
+		test_msg("Error adding half extent %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 3 * 1024 * 1024, 1 * 1024 * 1024);
+	if (ret) {
+		test_msg("Error removing tail end %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 0, 1 * 1024 * 1024);
+	if (ret) {
+		test_msg("Error removing front end %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 2 * 1024 * 1024, 4096);
+	if (ret) {
+		test_msg("Error removing middle peice %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, 0, 1 * 1024 * 1024)) {
+		test_msg("Still have space at the front\n");
+		return -1;
+	}
+
+	if (test_check_exists(cache, 2 * 1024 * 1024, 4096)) {
+		test_msg("Still have space in the middle\n");
+		return -1;
+	}
+
+	if (test_check_exists(cache, 3 * 1024 * 1024, 1 * 1024 * 1024)) {
+		test_msg("Still have space at the end\n");
+		return -1;
+	}
+
+	/* Cleanup */
+	__btrfs_remove_free_space_cache(cache->free_space_ctl);
+
+	return 0;
+}
+
+static int test_bitmaps(struct btrfs_block_group_cache *cache)
+{
+	u64 next_bitmap_offset;
+	int ret;
+
+	test_msg("Running bitmap only tests\n");
+
+	ret = test_add_free_space_entry(cache, 0, 4 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't create a bitmap entry %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 0, 4 * 1024 * 1024);
+	if (ret) {
+		test_msg("Error removing bitmap full range %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, 0, 4 * 1024 * 1024)) {
+		test_msg("Left some space in bitmap\n");
+		return -1;
+	}
+
+	ret = test_add_free_space_entry(cache, 0, 4 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't add to our bitmap entry %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 1 * 1024 * 1024, 2 * 1024 * 1024);
+	if (ret) {
+		test_msg("Couldn't remove middle chunk %d\n", ret);
+		return ret;
+	}
+
+	/*
+	 * The first bitmap we have starts at offset 0 so the next one is just
+	 * at the end of the first bitmap.
+	 */
+	next_bitmap_offset = (u64)(BITS_PER_BITMAP * 4096);
+
+	/* Test a bit straddling two bitmaps */
+	ret = test_add_free_space_entry(cache, next_bitmap_offset -
+				   (2 * 1024 * 1024), 4 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't add space that straddles two bitmaps %d\n",
+				ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, next_bitmap_offset -
+				      (1 * 1024 * 1024), 2 * 1024 * 1024);
+	if (ret) {
+		test_msg("Couldn't remove overlapping space %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, next_bitmap_offset - (1 * 1024 * 1024),
+			 2 * 1024 * 1024)) {
+		test_msg("Left some space when removing overlapping\n");
+		return -1;
+	}
+
+	__btrfs_remove_free_space_cache(cache->free_space_ctl);
+
+	return 0;
+}
+
+/* This is the high grade jackassery */
+static int test_bitmaps_and_extents(struct btrfs_block_group_cache *cache)
+{
+	u64 bitmap_offset = (u64)(BITS_PER_BITMAP * 4096);
+	int ret;
+
+	test_msg("Running bitmap and extent tests\n");
+
+	/*
+	 * First let's do something simple, an extent at the same offset as the
+	 * bitmap, but the free space completely in the extent and then
+	 * completely in the bitmap.
+	 */
+	ret = test_add_free_space_entry(cache, 4 * 1024 * 1024, 1 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't create bitmap entry %d\n", ret);
+		return ret;
+	}
+
+	ret = test_add_free_space_entry(cache, 0, 1 * 1024 * 1024, 0);
+	if (ret) {
+		test_msg("Couldn't add extent entry %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 0, 1 * 1024 * 1024);
+	if (ret) {
+		test_msg("Couldn't remove extent entry %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, 0, 1 * 1024 * 1024)) {
+		test_msg("Left remnants after our remove\n");
+		return -1;
+	}
+
+	/* Now to add back the extent entry and remove from the bitmap */
+	ret = test_add_free_space_entry(cache, 0, 1 * 1024 * 1024, 0);
+	if (ret) {
+		test_msg("Couldn't re-add extent entry %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 4 * 1024 * 1024, 1 * 1024 * 1024);
+	if (ret) {
+		test_msg("Couldn't remove from bitmap %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, 4 * 1024 * 1024, 1 * 1024 * 1024)) {
+		test_msg("Left remnants in the bitmap\n");
+		return -1;
+	}
+
+	/*
+	 * Ok so a little more evil, extent entry and bitmap at the same offset,
+	 * removing an overlapping chunk.
+	 */
+	ret = test_add_free_space_entry(cache, 1 * 1024 * 1024, 4 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't add to a bitmap %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 512 * 1024, 3 * 1024 * 1024);
+	if (ret) {
+		test_msg("Couldn't remove overlapping space %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, 512 * 1024, 3 * 1024 * 1024)) {
+		test_msg("Left over peices after removing overlapping\n");
+		return -1;
+	}
+
+	__btrfs_remove_free_space_cache(cache->free_space_ctl);
+
+	/* Now with the extent entry offset into the bitmap */
+	ret = test_add_free_space_entry(cache, 4 * 1024 * 1024, 4 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't add space to the bitmap %d\n", ret);
+		return ret;
+	}
+
+	ret = test_add_free_space_entry(cache, 2 * 1024 * 1024, 2 * 1024 * 1024, 0);
+	if (ret) {
+		test_msg("Couldn't add extent to the cache %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 3 * 1024 * 1024, 4 * 1024 * 1024);
+	if (ret) {
+		test_msg("Problem removing overlapping space %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, 3 * 1024 * 1024, 4 * 1024 * 1024)) {
+		test_msg("Left something behind when removing space");
+		return -1;
+	}
+
+	/*
+	 * This has blown up in the past, the extent entry starts before the
+	 * bitmap entry, but we're trying to remove an offset that falls
+	 * completely within the bitmap range and is in both the extent entry
+	 * and the bitmap entry, looks like this
+	 *
+	 *   [ extent ]
+	 *      [ bitmap ]
+	 *        [ del ]
+	 */
+	__btrfs_remove_free_space_cache(cache->free_space_ctl);
+	ret = test_add_free_space_entry(cache, bitmap_offset + 4 * 1024 * 1024,
+				   4 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't add bitmap %d\n", ret);
+		return ret;
+	}
+
+	ret = test_add_free_space_entry(cache, bitmap_offset - 1 * 1024 * 1024,
+				   5 * 1024 * 1024, 0);
+	if (ret) {
+		test_msg("Couldn't add extent entry %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, bitmap_offset + 1 * 1024 * 1024,
+				      5 * 1024 * 1024);
+	if (ret) {
+		test_msg("Failed to free our space %d\n", ret);
+		return ret;
+	}
+
+	if (test_check_exists(cache, bitmap_offset + 1 * 1024 * 1024,
+			 5 * 1024 * 1024)) {
+		test_msg("Left stuff over\n");
+		return -1;
+	}
+
+	__btrfs_remove_free_space_cache(cache->free_space_ctl);
+
+	/*
+	 * This blew up before, we have part of the free space in a bitmap and
+	 * then the entirety of the rest of the space in an extent.  This used
+	 * to return -EAGAIN back from btrfs_remove_extent, make sure this
+	 * doesn't happen.
+	 */
+	ret = test_add_free_space_entry(cache, 1 * 1024 * 1024, 2 * 1024 * 1024, 1);
+	if (ret) {
+		test_msg("Couldn't add bitmap entry %d\n", ret);
+		return ret;
+	}
+
+	ret = test_add_free_space_entry(cache, 3 * 1024 * 1024, 1 * 1024 * 1024, 0);
+	if (ret) {
+		test_msg("Couldn't add extent entry %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_remove_free_space(cache, 1 * 1024 * 1024, 3 * 1024 * 1024);
+	if (ret) {
+		test_msg("Error removing bitmap and extent overlapping %d\n", ret);
+		return ret;
+	}
+
+	__btrfs_remove_free_space_cache(cache->free_space_ctl);
+	return 0;
+}
+
+int btrfs_test_free_space_cache(void)
+{
+	struct btrfs_block_group_cache *cache;
+	int ret;
+
+	test_msg("Running btrfs free space cache tests\n");
+
+	cache = init_test_block_group();
+	if (!cache) {
+		test_msg("Couldn't run the tests\n");
+		return 0;
+	}
+
+	ret = test_extents(cache);
+	if (ret)
+		goto out;
+	ret = test_bitmaps(cache);
+	if (ret)
+		goto out;
+	ret = test_bitmaps_and_extents(cache);
+	if (ret)
+		goto out;
+out:
+	__btrfs_remove_free_space_cache(cache->free_space_ctl);
+	kfree(cache->free_space_ctl);
+	kfree(cache);
+	test_msg("Free space cache tests finished\n");
+	return ret;
+}
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 0544587d74f4..8c81bdc1ef9b 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -34,12 +34,43 @@
 
 #define BTRFS_ROOT_TRANS_TAG 0
 
+static unsigned int btrfs_blocked_trans_types[TRANS_STATE_MAX] = {
+	[TRANS_STATE_RUNNING]		= 0U,
+	[TRANS_STATE_BLOCKED]		= (__TRANS_USERSPACE |
+					   __TRANS_START),
+	[TRANS_STATE_COMMIT_START]	= (__TRANS_USERSPACE |
+					   __TRANS_START |
+					   __TRANS_ATTACH),
+	[TRANS_STATE_COMMIT_DOING]	= (__TRANS_USERSPACE |
+					   __TRANS_START |
+					   __TRANS_ATTACH |
+					   __TRANS_JOIN),
+	[TRANS_STATE_UNBLOCKED]		= (__TRANS_USERSPACE |
+					   __TRANS_START |
+					   __TRANS_ATTACH |
+					   __TRANS_JOIN |
+					   __TRANS_JOIN_NOLOCK),
+	[TRANS_STATE_COMPLETED]		= (__TRANS_USERSPACE |
+					   __TRANS_START |
+					   __TRANS_ATTACH |
+					   __TRANS_JOIN |
+					   __TRANS_JOIN_NOLOCK),
+};
+
 static void put_transaction(struct btrfs_transaction *transaction)
 {
 	WARN_ON(atomic_read(&transaction->use_count) == 0);
 	if (atomic_dec_and_test(&transaction->use_count)) {
 		BUG_ON(!list_empty(&transaction->list));
 		WARN_ON(transaction->delayed_refs.root.rb_node);
+		while (!list_empty(&transaction->pending_chunks)) {
+			struct extent_map *em;
+
+			em = list_first_entry(&transaction->pending_chunks,
+					      struct extent_map, list);
+			list_del_init(&em->list);
+			free_extent_map(em);
+		}
 		kmem_cache_free(btrfs_transaction_cachep, transaction);
 	}
 }
@@ -50,18 +81,35 @@ static noinline void switch_commit_root(struct btrfs_root *root)
 	root->commit_root = btrfs_root_node(root);
 }
 
-static inline int can_join_transaction(struct btrfs_transaction *trans,
-				       int type)
+static inline void extwriter_counter_inc(struct btrfs_transaction *trans,
+					 unsigned int type)
+{
+	if (type & TRANS_EXTWRITERS)
+		atomic_inc(&trans->num_extwriters);
+}
+
+static inline void extwriter_counter_dec(struct btrfs_transaction *trans,
+					 unsigned int type)
 {
-	return !(trans->in_commit &&
-		 type != TRANS_JOIN &&
-		 type != TRANS_JOIN_NOLOCK);
+	if (type & TRANS_EXTWRITERS)
+		atomic_dec(&trans->num_extwriters);
+}
+
+static inline void extwriter_counter_init(struct btrfs_transaction *trans,
+					  unsigned int type)
+{
+	atomic_set(&trans->num_extwriters, ((type & TRANS_EXTWRITERS) ? 1 : 0));
+}
+
+static inline int extwriter_counter_read(struct btrfs_transaction *trans)
+{
+	return atomic_read(&trans->num_extwriters);
 }
 
 /*
  * either allocate a new transaction or hop into the existing one
  */
-static noinline int join_transaction(struct btrfs_root *root, int type)
+static noinline int join_transaction(struct btrfs_root *root, unsigned int type)
 {
 	struct btrfs_transaction *cur_trans;
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -74,32 +122,19 @@ loop:
 		return -EROFS;
 	}
 
-	if (fs_info->trans_no_join) {
-		/* 
-		 * If we are JOIN_NOLOCK we're already committing a current
-		 * transaction, we just need a handle to deal with something
-		 * when committing the transaction, such as inode cache and
-		 * space cache. It is a special case.
-		 */
-		if (type != TRANS_JOIN_NOLOCK) {
-			spin_unlock(&fs_info->trans_lock);
-			return -EBUSY;
-		}
-	}
-
 	cur_trans = fs_info->running_transaction;
 	if (cur_trans) {
 		if (cur_trans->aborted) {
 			spin_unlock(&fs_info->trans_lock);
 			return cur_trans->aborted;
 		}
-		if (!can_join_transaction(cur_trans, type)) {
+		if (btrfs_blocked_trans_types[cur_trans->state] & type) {
 			spin_unlock(&fs_info->trans_lock);
 			return -EBUSY;
 		}
 		atomic_inc(&cur_trans->use_count);
 		atomic_inc(&cur_trans->num_writers);
-		cur_trans->num_joined++;
+		extwriter_counter_inc(cur_trans, type);
 		spin_unlock(&fs_info->trans_lock);
 		return 0;
 	}
@@ -112,6 +147,12 @@ loop:
 	if (type == TRANS_ATTACH)
 		return -ENOENT;
 
+	/*
+	 * JOIN_NOLOCK only happens during the transaction commit, so
+	 * it is impossible that ->running_transaction is NULL
+	 */
+	BUG_ON(type == TRANS_JOIN_NOLOCK);
+
 	cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS);
 	if (!cur_trans)
 		return -ENOMEM;
@@ -120,7 +161,7 @@ loop:
 	if (fs_info->running_transaction) {
 		/*
 		 * someone started a transaction after we unlocked.  Make sure
-		 * to redo the trans_no_join checks above
+		 * to redo the checks above
 		 */
 		kmem_cache_free(btrfs_transaction_cachep, cur_trans);
 		goto loop;
@@ -131,17 +172,15 @@ loop:
 	}
 
 	atomic_set(&cur_trans->num_writers, 1);
-	cur_trans->num_joined = 0;
+	extwriter_counter_init(cur_trans, type);
 	init_waitqueue_head(&cur_trans->writer_wait);
 	init_waitqueue_head(&cur_trans->commit_wait);
-	cur_trans->in_commit = 0;
-	cur_trans->blocked = 0;
+	cur_trans->state = TRANS_STATE_RUNNING;
 	/*
 	 * One for this trans handle, one so it will live on until we
 	 * commit the transaction.
 	 */
 	atomic_set(&cur_trans->use_count, 2);
-	cur_trans->commit_done = 0;
 	cur_trans->start_time = get_seconds();
 
 	cur_trans->delayed_refs.root = RB_ROOT;
@@ -164,7 +203,6 @@ loop:
 			"creating a fresh transaction\n");
 	atomic64_set(&fs_info->tree_mod_seq, 0);
 
-	spin_lock_init(&cur_trans->commit_lock);
 	spin_lock_init(&cur_trans->delayed_refs.lock);
 	atomic_set(&cur_trans->delayed_refs.procs_running_refs, 0);
 	atomic_set(&cur_trans->delayed_refs.ref_seq, 0);
@@ -172,6 +210,7 @@ loop:
 
 	INIT_LIST_HEAD(&cur_trans->pending_snapshots);
 	INIT_LIST_HEAD(&cur_trans->ordered_operations);
+	INIT_LIST_HEAD(&cur_trans->pending_chunks);
 	list_add_tail(&cur_trans->list, &fs_info->trans_list);
 	extent_io_tree_init(&cur_trans->dirty_pages,
 			     fs_info->btree_inode->i_mapping);
@@ -269,6 +308,13 @@ int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
+static inline int is_transaction_blocked(struct btrfs_transaction *trans)
+{
+	return (trans->state >= TRANS_STATE_BLOCKED &&
+		trans->state < TRANS_STATE_UNBLOCKED &&
+		!trans->aborted);
+}
+
 /* wait for commit against the current transaction to become unblocked
  * when this is done, it is safe to start a new transaction, but the current
  * transaction might not be fully on disk.
@@ -279,12 +325,13 @@ static void wait_current_trans(struct btrfs_root *root)
 
 	spin_lock(&root->fs_info->trans_lock);
 	cur_trans = root->fs_info->running_transaction;
-	if (cur_trans && cur_trans->blocked) {
+	if (cur_trans && is_transaction_blocked(cur_trans)) {
 		atomic_inc(&cur_trans->use_count);
 		spin_unlock(&root->fs_info->trans_lock);
 
 		wait_event(root->fs_info->transaction_wait,
-			   !cur_trans->blocked);
+			   cur_trans->state >= TRANS_STATE_UNBLOCKED ||
+			   cur_trans->aborted);
 		put_transaction(cur_trans);
 	} else {
 		spin_unlock(&root->fs_info->trans_lock);
@@ -307,7 +354,7 @@ static int may_wait_transaction(struct btrfs_root *root, int type)
 }
 
 static struct btrfs_trans_handle *
-start_transaction(struct btrfs_root *root, u64 num_items, int type,
+start_transaction(struct btrfs_root *root, u64 num_items, unsigned int type,
 		  enum btrfs_reserve_flush_enum flush)
 {
 	struct btrfs_trans_handle *h;
@@ -320,7 +367,7 @@ start_transaction(struct btrfs_root *root, u64 num_items, int type,
 		return ERR_PTR(-EROFS);
 
 	if (current->journal_info) {
-		WARN_ON(type != TRANS_JOIN && type != TRANS_JOIN_NOLOCK);
+		WARN_ON(type & TRANS_EXTWRITERS);
 		h = current->journal_info;
 		h->use_count++;
 		WARN_ON(h->use_count > 2);
@@ -366,7 +413,7 @@ again:
 	 * If we are ATTACH, it means we just want to catch the current
 	 * transaction and commit it, so we needn't do sb_start_intwrite(). 
 	 */
-	if (type < TRANS_JOIN_NOLOCK)
+	if (type & __TRANS_FREEZABLE)
 		sb_start_intwrite(root->fs_info->sb);
 
 	if (may_wait_transaction(root, type))
@@ -408,7 +455,8 @@ again:
 	INIT_LIST_HEAD(&h->new_bgs);
 
 	smp_mb();
-	if (cur_trans->blocked && may_wait_transaction(root, type)) {
+	if (cur_trans->state >= TRANS_STATE_BLOCKED &&
+	    may_wait_transaction(root, type)) {
 		btrfs_commit_transaction(h, root);
 		goto again;
 	}
@@ -429,7 +477,7 @@ got_it:
 	return h;
 
 join_fail:
-	if (type < TRANS_JOIN_NOLOCK)
+	if (type & __TRANS_FREEZABLE)
 		sb_end_intwrite(root->fs_info->sb);
 	kmem_cache_free(btrfs_trans_handle_cachep, h);
 alloc_fail:
@@ -490,7 +538,7 @@ struct btrfs_trans_handle *btrfs_attach_transaction(struct btrfs_root *root)
 }
 
 /*
- * btrfs_attach_transaction() - catch the running transaction
+ * btrfs_attach_transaction_barrier() - catch the running transaction
  *
  * It is similar to the above function, the differentia is this one
  * will wait for all the inactive transactions until they fully
@@ -512,7 +560,7 @@ btrfs_attach_transaction_barrier(struct btrfs_root *root)
 static noinline void wait_for_commit(struct btrfs_root *root,
 				    struct btrfs_transaction *commit)
 {
-	wait_event(commit->commit_wait, commit->commit_done);
+	wait_event(commit->commit_wait, commit->state == TRANS_STATE_COMPLETED);
 }
 
 int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid)
@@ -548,8 +596,8 @@ int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid)
 		spin_lock(&root->fs_info->trans_lock);
 		list_for_each_entry_reverse(t, &root->fs_info->trans_list,
 					    list) {
-			if (t->in_commit) {
-				if (t->commit_done)
+			if (t->state >= TRANS_STATE_COMMIT_START) {
+				if (t->state == TRANS_STATE_COMPLETED)
 					break;
 				cur_trans = t;
 				atomic_inc(&cur_trans->use_count);
@@ -576,10 +624,11 @@ void btrfs_throttle(struct btrfs_root *root)
 static int should_end_transaction(struct btrfs_trans_handle *trans,
 				  struct btrfs_root *root)
 {
-	int ret;
+	if (root->fs_info->global_block_rsv.space_info->full &&
+	    btrfs_should_throttle_delayed_refs(trans, root))
+		return 1;
 
-	ret = btrfs_block_rsv_check(root, &root->fs_info->global_block_rsv, 5);
-	return ret ? 1 : 0;
+	return !!btrfs_block_rsv_check(root, &root->fs_info->global_block_rsv, 5);
 }
 
 int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
@@ -590,7 +639,8 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
 	int err;
 
 	smp_mb();
-	if (cur_trans->blocked || cur_trans->delayed_refs.flushing)
+	if (cur_trans->state >= TRANS_STATE_BLOCKED ||
+	    cur_trans->delayed_refs.flushing)
 		return 1;
 
 	updates = trans->delayed_ref_updates;
@@ -609,7 +659,7 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_transaction *cur_trans = trans->transaction;
 	struct btrfs_fs_info *info = root->fs_info;
-	int count = 0;
+	unsigned long cur = trans->delayed_ref_updates;
 	int lock = (trans->type != TRANS_JOIN_NOLOCK);
 	int err = 0;
 
@@ -638,17 +688,11 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 	if (!list_empty(&trans->new_bgs))
 		btrfs_create_pending_block_groups(trans, root);
 
-	while (count < 1) {
-		unsigned long cur = trans->delayed_ref_updates;
+	trans->delayed_ref_updates = 0;
+	if (btrfs_should_throttle_delayed_refs(trans, root)) {
+		cur = max_t(unsigned long, cur, 1);
 		trans->delayed_ref_updates = 0;
-		if (cur &&
-		    trans->transaction->delayed_refs.num_heads_ready > 64) {
-			trans->delayed_ref_updates = 0;
-			btrfs_run_delayed_refs(trans, root, cur);
-		} else {
-			break;
-		}
-		count++;
+		btrfs_run_delayed_refs(trans, root, cur);
 	}
 
 	btrfs_trans_release_metadata(trans, root);
@@ -658,12 +702,15 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 		btrfs_create_pending_block_groups(trans, root);
 
 	if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) &&
-	    should_end_transaction(trans, root)) {
-		trans->transaction->blocked = 1;
-		smp_wmb();
+	    should_end_transaction(trans, root) &&
+	    ACCESS_ONCE(cur_trans->state) == TRANS_STATE_RUNNING) {
+		spin_lock(&info->trans_lock);
+		if (cur_trans->state == TRANS_STATE_RUNNING)
+			cur_trans->state = TRANS_STATE_BLOCKED;
+		spin_unlock(&info->trans_lock);
 	}
 
-	if (lock && cur_trans->blocked && !cur_trans->in_commit) {
+	if (lock && ACCESS_ONCE(cur_trans->state) == TRANS_STATE_BLOCKED) {
 		if (throttle) {
 			/*
 			 * We may race with somebody else here so end up having
@@ -677,12 +724,13 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 		}
 	}
 
-	if (trans->type < TRANS_JOIN_NOLOCK)
+	if (trans->type & __TRANS_FREEZABLE)
 		sb_end_intwrite(root->fs_info->sb);
 
 	WARN_ON(cur_trans != info->running_transaction);
 	WARN_ON(atomic_read(&cur_trans->num_writers) < 1);
 	atomic_dec(&cur_trans->num_writers);
+	extwriter_counter_dec(cur_trans, trans->type);
 
 	smp_mb();
 	if (waitqueue_active(&cur_trans->writer_wait))
@@ -736,9 +784,7 @@ int btrfs_write_marked_extents(struct btrfs_root *root,
 	struct extent_state *cached_state = NULL;
 	u64 start = 0;
 	u64 end;
-	struct blk_plug plug;
 
-	blk_start_plug(&plug);
 	while (!find_first_extent_bit(dirty_pages, start, &start, &end,
 				      mark, &cached_state)) {
 		convert_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT,
@@ -752,7 +798,6 @@ int btrfs_write_marked_extents(struct btrfs_root *root,
 	}
 	if (err)
 		werr = err;
-	blk_finish_plug(&plug);
 	return werr;
 }
 
@@ -792,13 +837,16 @@ int btrfs_wait_marked_extents(struct btrfs_root *root,
  * them in one of two extent_io trees.  This is used to make sure all of
  * those extents are on disk for transaction or log commit
  */
-int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
+static int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
 				struct extent_io_tree *dirty_pages, int mark)
 {
 	int ret;
 	int ret2;
+	struct blk_plug plug;
 
+	blk_start_plug(&plug);
 	ret = btrfs_write_marked_extents(root, dirty_pages, mark);
+	blk_finish_plug(&plug);
 	ret2 = btrfs_wait_marked_extents(root, dirty_pages, mark);
 
 	if (ret)
@@ -935,12 +983,12 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans,
  * a dirty root struct and adds it into the list of dead roots that need to
  * be deleted
  */
-int btrfs_add_dead_root(struct btrfs_root *root)
+void btrfs_add_dead_root(struct btrfs_root *root)
 {
 	spin_lock(&root->fs_info->trans_lock);
-	list_add_tail(&root->root_list, &root->fs_info->dead_roots);
+	if (list_empty(&root->root_list))
+		list_add_tail(&root->root_list, &root->fs_info->dead_roots);
 	spin_unlock(&root->fs_info->trans_lock);
-	return 0;
 }
 
 /*
@@ -1177,8 +1225,8 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 		btrfs_set_root_stransid(new_root_item, 0);
 		btrfs_set_root_rtransid(new_root_item, 0);
 	}
-	new_root_item->otime.sec = cpu_to_le64(cur_time.tv_sec);
-	new_root_item->otime.nsec = cpu_to_le32(cur_time.tv_nsec);
+	btrfs_set_stack_timespec_sec(&new_root_item->otime, cur_time.tv_sec);
+	btrfs_set_stack_timespec_nsec(&new_root_item->otime, cur_time.tv_nsec);
 	btrfs_set_root_otransid(new_root_item, trans->transid);
 
 	old = btrfs_lock_root_node(root);
@@ -1263,8 +1311,26 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 					 dentry->d_name.len * 2);
 	parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
 	ret = btrfs_update_inode_fallback(trans, parent_root, parent_inode);
-	if (ret)
+	if (ret) {
 		btrfs_abort_transaction(trans, root, ret);
+		goto fail;
+	}
+	ret = btrfs_uuid_tree_add(trans, fs_info->uuid_root, new_uuid.b,
+				  BTRFS_UUID_KEY_SUBVOL, objectid);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto fail;
+	}
+	if (!btrfs_is_empty_uuid(new_root_item->received_uuid)) {
+		ret = btrfs_uuid_tree_add(trans, fs_info->uuid_root,
+					  new_root_item->received_uuid,
+					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
+					  objectid);
+		if (ret && ret != -EEXIST) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto fail;
+		}
+	}
 fail:
 	pending->error = ret;
 dir_item_existed:
@@ -1314,24 +1380,32 @@ static void update_super_roots(struct btrfs_root *root)
 	super->root_level = root_item->level;
 	if (btrfs_test_opt(root, SPACE_CACHE))
 		super->cache_generation = root_item->generation;
+	if (root->fs_info->update_uuid_tree_gen)
+		super->uuid_tree_generation = root_item->generation;
 }
 
 int btrfs_transaction_in_commit(struct btrfs_fs_info *info)
 {
+	struct btrfs_transaction *trans;
 	int ret = 0;
+
 	spin_lock(&info->trans_lock);
-	if (info->running_transaction)
-		ret = info->running_transaction->in_commit;
+	trans = info->running_transaction;
+	if (trans)
+		ret = (trans->state >= TRANS_STATE_COMMIT_START);
 	spin_unlock(&info->trans_lock);
 	return ret;
 }
 
 int btrfs_transaction_blocked(struct btrfs_fs_info *info)
 {
+	struct btrfs_transaction *trans;
 	int ret = 0;
+
 	spin_lock(&info->trans_lock);
-	if (info->running_transaction)
-		ret = info->running_transaction->blocked;
+	trans = info->running_transaction;
+	if (trans)
+		ret = is_transaction_blocked(trans);
 	spin_unlock(&info->trans_lock);
 	return ret;
 }
@@ -1343,7 +1417,9 @@ int btrfs_transaction_blocked(struct btrfs_fs_info *info)
 static void wait_current_trans_commit_start(struct btrfs_root *root,
 					    struct btrfs_transaction *trans)
 {
-	wait_event(root->fs_info->transaction_blocked_wait, trans->in_commit);
+	wait_event(root->fs_info->transaction_blocked_wait,
+		   trans->state >= TRANS_STATE_COMMIT_START ||
+		   trans->aborted);
 }
 
 /*
@@ -1354,7 +1430,8 @@ static void wait_current_trans_commit_start_and_unblock(struct btrfs_root *root,
 					 struct btrfs_transaction *trans)
 {
 	wait_event(root->fs_info->transaction_wait,
-		   trans->commit_done || (trans->in_commit && !trans->blocked));
+		   trans->state >= TRANS_STATE_UNBLOCKED ||
+		   trans->aborted);
 }
 
 /*
@@ -1450,26 +1527,31 @@ static void cleanup_transaction(struct btrfs_trans_handle *trans,
 
 	spin_lock(&root->fs_info->trans_lock);
 
-	if (list_empty(&cur_trans->list)) {
-		spin_unlock(&root->fs_info->trans_lock);
-		btrfs_end_transaction(trans, root);
-		return;
-	}
+	/*
+	 * If the transaction is removed from the list, it means this
+	 * transaction has been committed successfully, so it is impossible
+	 * to call the cleanup function.
+	 */
+	BUG_ON(list_empty(&cur_trans->list));
 
 	list_del_init(&cur_trans->list);
 	if (cur_trans == root->fs_info->running_transaction) {
-		root->fs_info->trans_no_join = 1;
+		cur_trans->state = TRANS_STATE_COMMIT_DOING;
 		spin_unlock(&root->fs_info->trans_lock);
 		wait_event(cur_trans->writer_wait,
 			   atomic_read(&cur_trans->num_writers) == 1);
 
 		spin_lock(&root->fs_info->trans_lock);
-		root->fs_info->running_transaction = NULL;
 	}
 	spin_unlock(&root->fs_info->trans_lock);
 
 	btrfs_cleanup_one_transaction(trans->transaction, root);
 
+	spin_lock(&root->fs_info->trans_lock);
+	if (cur_trans == root->fs_info->running_transaction)
+		root->fs_info->running_transaction = NULL;
+	spin_unlock(&root->fs_info->trans_lock);
+
 	put_transaction(cur_trans);
 	put_transaction(cur_trans);
 
@@ -1481,33 +1563,13 @@ static void cleanup_transaction(struct btrfs_trans_handle *trans,
 		current->journal_info = NULL;
 
 	kmem_cache_free(btrfs_trans_handle_cachep, trans);
-
-	spin_lock(&root->fs_info->trans_lock);
-	root->fs_info->trans_no_join = 0;
-	spin_unlock(&root->fs_info->trans_lock);
 }
 
 static int btrfs_flush_all_pending_stuffs(struct btrfs_trans_handle *trans,
 					  struct btrfs_root *root)
 {
-	int flush_on_commit = btrfs_test_opt(root, FLUSHONCOMMIT);
-	int snap_pending = 0;
 	int ret;
 
-	if (!flush_on_commit) {
-		spin_lock(&root->fs_info->trans_lock);
-		if (!list_empty(&trans->transaction->pending_snapshots))
-			snap_pending = 1;
-		spin_unlock(&root->fs_info->trans_lock);
-	}
-
-	if (flush_on_commit || snap_pending) {
-		ret = btrfs_start_delalloc_inodes(root, 1);
-		if (ret)
-			return ret;
-		btrfs_wait_ordered_extents(root, 1);
-	}
-
 	ret = btrfs_run_delayed_items(trans, root);
 	if (ret)
 		return ret;
@@ -1531,23 +1593,25 @@ static int btrfs_flush_all_pending_stuffs(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-/*
- * btrfs_transaction state sequence:
- *    in_commit = 0, blocked = 0  (initial)
- *    in_commit = 1, blocked = 1
- *    blocked = 0
- *    commit_done = 1
- */
+static inline int btrfs_start_delalloc_flush(struct btrfs_fs_info *fs_info)
+{
+	if (btrfs_test_opt(fs_info->tree_root, FLUSHONCOMMIT))
+		return btrfs_start_all_delalloc_inodes(fs_info, 1);
+	return 0;
+}
+
+static inline void btrfs_wait_delalloc_flush(struct btrfs_fs_info *fs_info)
+{
+	if (btrfs_test_opt(fs_info->tree_root, FLUSHONCOMMIT))
+		btrfs_wait_all_ordered_extents(fs_info);
+}
+
 int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root)
 {
-	unsigned long joined = 0;
 	struct btrfs_transaction *cur_trans = trans->transaction;
 	struct btrfs_transaction *prev_trans = NULL;
-	DEFINE_WAIT(wait);
 	int ret;
-	int should_grow = 0;
-	unsigned long now = get_seconds();
 
 	ret = btrfs_run_ordered_operations(trans, root, 0);
 	if (ret) {
@@ -1586,6 +1650,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	 * start sending their work down.
 	 */
 	cur_trans->delayed_refs.flushing = 1;
+	smp_wmb();
 
 	if (!list_empty(&trans->new_bgs))
 		btrfs_create_pending_block_groups(trans, root);
@@ -1596,9 +1661,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
-	spin_lock(&cur_trans->commit_lock);
-	if (cur_trans->in_commit) {
-		spin_unlock(&cur_trans->commit_lock);
+	spin_lock(&root->fs_info->trans_lock);
+	if (cur_trans->state >= TRANS_STATE_COMMIT_START) {
+		spin_unlock(&root->fs_info->trans_lock);
 		atomic_inc(&cur_trans->use_count);
 		ret = btrfs_end_transaction(trans, root);
 
@@ -1609,16 +1674,13 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
-	trans->transaction->in_commit = 1;
-	trans->transaction->blocked = 1;
-	spin_unlock(&cur_trans->commit_lock);
+	cur_trans->state = TRANS_STATE_COMMIT_START;
 	wake_up(&root->fs_info->transaction_blocked_wait);
 
-	spin_lock(&root->fs_info->trans_lock);
 	if (cur_trans->list.prev != &root->fs_info->trans_list) {
 		prev_trans = list_entry(cur_trans->list.prev,
 					struct btrfs_transaction, list);
-		if (!prev_trans->commit_done) {
+		if (prev_trans->state != TRANS_STATE_COMPLETED) {
 			atomic_inc(&prev_trans->use_count);
 			spin_unlock(&root->fs_info->trans_lock);
 
@@ -1632,42 +1694,32 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		spin_unlock(&root->fs_info->trans_lock);
 	}
 
-	if (!btrfs_test_opt(root, SSD) &&
-	    (now < cur_trans->start_time || now - cur_trans->start_time < 1))
-		should_grow = 1;
-
-	do {
-		joined = cur_trans->num_joined;
-
-		WARN_ON(cur_trans != trans->transaction);
-
-		ret = btrfs_flush_all_pending_stuffs(trans, root);
-		if (ret)
-			goto cleanup_transaction;
+	extwriter_counter_dec(cur_trans, trans->type);
 
-		prepare_to_wait(&cur_trans->writer_wait, &wait,
-				TASK_UNINTERRUPTIBLE);
+	ret = btrfs_start_delalloc_flush(root->fs_info);
+	if (ret)
+		goto cleanup_transaction;
 
-		if (atomic_read(&cur_trans->num_writers) > 1)
-			schedule_timeout(MAX_SCHEDULE_TIMEOUT);
-		else if (should_grow)
-			schedule_timeout(1);
+	ret = btrfs_flush_all_pending_stuffs(trans, root);
+	if (ret)
+		goto cleanup_transaction;
 
-		finish_wait(&cur_trans->writer_wait, &wait);
-	} while (atomic_read(&cur_trans->num_writers) > 1 ||
-		 (should_grow && cur_trans->num_joined != joined));
+	wait_event(cur_trans->writer_wait,
+		   extwriter_counter_read(cur_trans) == 0);
 
+	/* some pending stuffs might be added after the previous flush. */
 	ret = btrfs_flush_all_pending_stuffs(trans, root);
 	if (ret)
 		goto cleanup_transaction;
 
+	btrfs_wait_delalloc_flush(root->fs_info);
 	/*
 	 * Ok now we need to make sure to block out any other joins while we
 	 * commit the transaction.  We could have started a join before setting
-	 * no_join so make sure to wait for num_writers to == 1 again.
+	 * COMMIT_DOING so make sure to wait for num_writers to == 1 again.
 	 */
 	spin_lock(&root->fs_info->trans_lock);
-	root->fs_info->trans_no_join = 1;
+	cur_trans->state = TRANS_STATE_COMMIT_DOING;
 	spin_unlock(&root->fs_info->trans_lock);
 	wait_event(cur_trans->writer_wait,
 		   atomic_read(&cur_trans->num_writers) == 1);
@@ -1786,18 +1838,14 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	assert_qgroups_uptodate(trans);
 	update_super_roots(root);
 
-	if (!root->fs_info->log_root_recovering) {
-		btrfs_set_super_log_root(root->fs_info->super_copy, 0);
-		btrfs_set_super_log_root_level(root->fs_info->super_copy, 0);
-	}
-
+	btrfs_set_super_log_root(root->fs_info->super_copy, 0);
+	btrfs_set_super_log_root_level(root->fs_info->super_copy, 0);
 	memcpy(root->fs_info->super_for_commit, root->fs_info->super_copy,
 	       sizeof(*root->fs_info->super_copy));
 
-	trans->transaction->blocked = 0;
 	spin_lock(&root->fs_info->trans_lock);
+	cur_trans->state = TRANS_STATE_UNBLOCKED;
 	root->fs_info->running_transaction = NULL;
-	root->fs_info->trans_no_join = 0;
 	spin_unlock(&root->fs_info->trans_lock);
 	mutex_unlock(&root->fs_info->reloc_mutex);
 
@@ -1825,10 +1873,12 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	btrfs_finish_extent_commit(trans, root);
 
-	cur_trans->commit_done = 1;
-
 	root->fs_info->last_trans_committed = cur_trans->transid;
-
+	/*
+	 * We needn't acquire the lock here because there is no other task
+	 * which can change it.
+	 */
+	cur_trans->state = TRANS_STATE_COMPLETED;
 	wake_up(&cur_trans->commit_wait);
 
 	spin_lock(&root->fs_info->trans_lock);
@@ -1838,7 +1888,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	put_transaction(cur_trans);
 	put_transaction(cur_trans);
 
-	if (trans->type < TRANS_JOIN_NOLOCK)
+	if (trans->type & __TRANS_FREEZABLE)
 		sb_end_intwrite(root->fs_info->sb);
 
 	trace_btrfs_transaction_commit(root);
@@ -1885,11 +1935,6 @@ int btrfs_clean_one_deleted_snapshot(struct btrfs_root *root)
 	int ret;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	if (fs_info->sb->s_flags & MS_RDONLY) {
-		pr_debug("btrfs: cleaner called for RO fs!\n");
-		return 0;
-	}
-
 	spin_lock(&fs_info->trans_lock);
 	if (list_empty(&fs_info->dead_roots)) {
 		spin_unlock(&fs_info->trans_lock);
@@ -1897,11 +1942,10 @@ int btrfs_clean_one_deleted_snapshot(struct btrfs_root *root)
 	}
 	root = list_first_entry(&fs_info->dead_roots,
 			struct btrfs_root, root_list);
-	list_del(&root->root_list);
+	list_del_init(&root->root_list);
 	spin_unlock(&fs_info->trans_lock);
 
-	pr_debug("btrfs: cleaner removing %llu\n",
-			(unsigned long long)root->objectid);
+	pr_debug("btrfs: cleaner removing %llu\n", root->objectid);
 
 	btrfs_kill_all_delayed_nodes(root);
 
@@ -1914,6 +1958,5 @@ int btrfs_clean_one_deleted_snapshot(struct btrfs_root *root)
 	 * If we encounter a transaction abort during snapshot cleaning, we
 	 * don't want to crash here
 	 */
-	BUG_ON(ret < 0 && ret != -EAGAIN && ret != -EROFS);
-	return 1;
+	return (ret < 0) ? 0 : 1;
 }
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index 24c97335a59f..5c2af8491621 100644
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -22,21 +22,33 @@
 #include "delayed-ref.h"
 #include "ctree.h"
 
+enum btrfs_trans_state {
+	TRANS_STATE_RUNNING		= 0,
+	TRANS_STATE_BLOCKED		= 1,
+	TRANS_STATE_COMMIT_START	= 2,
+	TRANS_STATE_COMMIT_DOING	= 3,
+	TRANS_STATE_UNBLOCKED		= 4,
+	TRANS_STATE_COMPLETED		= 5,
+	TRANS_STATE_MAX			= 6,
+};
+
 struct btrfs_transaction {
 	u64 transid;
 	/*
+	 * total external writers(USERSPACE/START/ATTACH) in this
+	 * transaction, it must be zero before the transaction is
+	 * being committed
+	 */
+	atomic_t num_extwriters;
+	/*
 	 * total writers in this transaction, it must be zero before the
 	 * transaction can end
 	 */
 	atomic_t num_writers;
 	atomic_t use_count;
 
-	unsigned long num_joined;
-
-	spinlock_t commit_lock;
-	int in_commit;
-	int commit_done;
-	int blocked;
+	/* Be protected by fs_info->trans_lock when we want to change it. */
+	enum btrfs_trans_state state;
 	struct list_head list;
 	struct extent_io_tree dirty_pages;
 	unsigned long start_time;
@@ -44,17 +56,27 @@ struct btrfs_transaction {
 	wait_queue_head_t commit_wait;
 	struct list_head pending_snapshots;
 	struct list_head ordered_operations;
+	struct list_head pending_chunks;
 	struct btrfs_delayed_ref_root delayed_refs;
 	int aborted;
 };
 
-enum btrfs_trans_type {
-	TRANS_START,
-	TRANS_JOIN,
-	TRANS_USERSPACE,
-	TRANS_JOIN_NOLOCK,
-	TRANS_ATTACH,
-};
+#define __TRANS_FREEZABLE	(1U << 0)
+
+#define __TRANS_USERSPACE	(1U << 8)
+#define __TRANS_START		(1U << 9)
+#define __TRANS_ATTACH		(1U << 10)
+#define __TRANS_JOIN		(1U << 11)
+#define __TRANS_JOIN_NOLOCK	(1U << 12)
+
+#define TRANS_USERSPACE		(__TRANS_USERSPACE | __TRANS_FREEZABLE)
+#define TRANS_START		(__TRANS_START | __TRANS_FREEZABLE)
+#define TRANS_ATTACH		(__TRANS_ATTACH)
+#define TRANS_JOIN		(__TRANS_JOIN | __TRANS_FREEZABLE)
+#define TRANS_JOIN_NOLOCK	(__TRANS_JOIN_NOLOCK)
+
+#define TRANS_EXTWRITERS	(__TRANS_USERSPACE | __TRANS_START |	\
+				 __TRANS_ATTACH)
 
 struct btrfs_trans_handle {
 	u64 transid;
@@ -70,7 +92,7 @@ struct btrfs_trans_handle {
 	short aborted;
 	short adding_csums;
 	bool allocating_chunk;
-	enum btrfs_trans_type type;
+	unsigned int type;
 	/*
 	 * this root is only needed to validate that the root passed to
 	 * start_transaction is the same as the one passed to end_transaction.
@@ -121,7 +143,7 @@ int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid);
 int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root);
 
-int btrfs_add_dead_root(struct btrfs_root *root);
+void btrfs_add_dead_root(struct btrfs_root *root);
 int btrfs_defrag_root(struct btrfs_root *root);
 int btrfs_clean_one_deleted_snapshot(struct btrfs_root *root);
 int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
@@ -138,8 +160,6 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
 void btrfs_throttle(struct btrfs_root *root);
 int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root);
-int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
-				struct extent_io_tree *dirty_pages, int mark);
 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,
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index c276ac9a0ec3..79f057c0619a 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -18,6 +18,7 @@
 
 #include <linux/sched.h>
 #include <linux/slab.h>
+#include <linux/blkdev.h>
 #include <linux/list_sort.h>
 #include "ctree.h"
 #include "transaction.h"
@@ -92,7 +93,8 @@
  */
 #define LOG_WALK_PIN_ONLY 0
 #define LOG_WALK_REPLAY_INODES 1
-#define LOG_WALK_REPLAY_ALL 2
+#define LOG_WALK_REPLAY_DIR_INDEX 2
+#define LOG_WALK_REPLAY_ALL 3
 
 static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root, struct inode *inode,
@@ -279,11 +281,23 @@ static int process_one_buffer(struct btrfs_root *log,
 {
 	int ret = 0;
 
+	/*
+	 * If this fs is mixed then we need to be able to process the leaves to
+	 * pin down any logged extents, so we have to read the block.
+	 */
+	if (btrfs_fs_incompat(log->fs_info, MIXED_GROUPS)) {
+		ret = btrfs_read_buffer(eb, gen);
+		if (ret)
+			return ret;
+	}
+
 	if (wc->pin)
 		ret = btrfs_pin_extent_for_log_replay(log->fs_info->extent_root,
 						      eb->start, eb->len);
 
 	if (!ret && btrfs_buffer_uptodate(eb, gen, 0)) {
+		if (wc->pin && btrfs_header_level(eb) == 0)
+			ret = btrfs_exclude_logged_extents(log, eb);
 		if (wc->write)
 			btrfs_write_tree_block(eb);
 		if (wc->wait)
@@ -380,6 +394,7 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans,
 		if (inode_item) {
 			struct btrfs_inode_item *item;
 			u64 nbytes;
+			u32 mode;
 
 			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
 					      struct btrfs_inode_item);
@@ -387,9 +402,19 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans,
 			item = btrfs_item_ptr(eb, slot,
 					      struct btrfs_inode_item);
 			btrfs_set_inode_nbytes(eb, item, nbytes);
+
+			/*
+			 * If this is a directory we need to reset the i_size to
+			 * 0 so that we can set it up properly when replaying
+			 * the rest of the items in this log.
+			 */
+			mode = btrfs_inode_mode(eb, item);
+			if (S_ISDIR(mode))
+				btrfs_set_inode_size(eb, item, 0);
 		}
 	} else if (inode_item) {
 		struct btrfs_inode_item *item;
+		u32 mode;
 
 		/*
 		 * New inode, set nbytes to 0 so that the nbytes comes out
@@ -397,6 +422,15 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans,
 		 */
 		item = btrfs_item_ptr(eb, slot, struct btrfs_inode_item);
 		btrfs_set_inode_nbytes(eb, item, 0);
+
+		/*
+		 * If this is a directory we need to reset the i_size to 0 so
+		 * that we can set it up properly when replaying the rest of
+		 * the items in this log.
+		 */
+		mode = btrfs_inode_mode(eb, item);
+		if (S_ISDIR(mode))
+			btrfs_set_inode_size(eb, item, 0);
 	}
 insert:
 	btrfs_release_path(path);
@@ -734,7 +768,8 @@ static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
 	ret = btrfs_unlink_inode(trans, root, dir, inode, name, name_len);
 	if (ret)
 		goto out;
-	btrfs_run_delayed_items(trans, root);
+	else
+		ret = btrfs_run_delayed_items(trans, root);
 out:
 	kfree(name);
 	iput(inode);
@@ -910,7 +945,9 @@ again:
 				kfree(victim_name);
 				if (ret)
 					return ret;
-				btrfs_run_delayed_items(trans, root);
+				ret = btrfs_run_delayed_items(trans, root);
+				if (ret)
+					return ret;
 				*search_done = 1;
 				goto again;
 			}
@@ -977,7 +1014,9 @@ again:
 								 inode,
 								 victim_name,
 								 victim_name_len);
-					btrfs_run_delayed_items(trans, root);
+					if (!ret)
+						ret = btrfs_run_delayed_items(
+								  trans, root);
 				}
 				iput(victim_parent);
 				kfree(victim_name);
@@ -1478,6 +1517,7 @@ static noinline int insert_one_name(struct btrfs_trans_handle *trans,
 		iput(inode);
 		return -EIO;
 	}
+
 	ret = btrfs_add_link(trans, dir, inode, name, name_len, 1, index);
 
 	/* FIXME, put inode into FIXUP list */
@@ -1516,6 +1556,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 	u8 log_type;
 	int exists;
 	int ret = 0;
+	bool update_size = (key->type == BTRFS_DIR_INDEX_KEY);
 
 	dir = read_one_inode(root, key->objectid);
 	if (!dir)
@@ -1523,8 +1564,10 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 
 	name_len = btrfs_dir_name_len(eb, di);
 	name = kmalloc(name_len, GFP_NOFS);
-	if (!name)
-		return -ENOMEM;
+	if (!name) {
+		ret = -ENOMEM;
+		goto out;
+	}
 
 	log_type = btrfs_dir_type(eb, di);
 	read_extent_buffer(eb, name, (unsigned long)(di + 1),
@@ -1584,6 +1627,10 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 		goto insert;
 out:
 	btrfs_release_path(path);
+	if (!ret && update_size) {
+		btrfs_i_size_write(dir, dir->i_size + name_len * 2);
+		ret = btrfs_update_inode(trans, root, dir);
+	}
 	kfree(name);
 	iput(dir);
 	return ret;
@@ -1594,6 +1641,7 @@ insert:
 			      name, name_len, log_type, &log_key);
 	if (ret && ret != -ENOENT)
 		goto out;
+	update_size = false;
 	ret = 0;
 	goto out;
 }
@@ -1797,7 +1845,7 @@ again:
 			ret = btrfs_unlink_inode(trans, root, dir, inode,
 						 name, name_len);
 			if (!ret)
-				btrfs_run_delayed_items(trans, root);
+				ret = btrfs_run_delayed_items(trans, root);
 			kfree(name);
 			iput(inode);
 			if (ret)
@@ -2007,6 +2055,15 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
 			if (ret)
 				break;
 		}
+
+		if (key.type == BTRFS_DIR_INDEX_KEY &&
+		    wc->stage == LOG_WALK_REPLAY_DIR_INDEX) {
+			ret = replay_one_dir_item(wc->trans, root, path,
+						  eb, i, &key);
+			if (ret)
+				break;
+		}
+
 		if (wc->stage < LOG_WALK_REPLAY_ALL)
 			continue;
 
@@ -2016,13 +2073,8 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
 					     eb, i, &key);
 			if (ret)
 				break;
-		} else if (key.type == BTRFS_INODE_REF_KEY) {
-			ret = add_inode_ref(wc->trans, root, log, path,
-					    eb, i, &key);
-			if (ret && ret != -ENOENT)
-				break;
-			ret = 0;
-		} else if (key.type == BTRFS_INODE_EXTREF_KEY) {
+		} else if (key.type == BTRFS_INODE_REF_KEY ||
+			   key.type == BTRFS_INODE_EXTREF_KEY) {
 			ret = add_inode_ref(wc->trans, root, log, path,
 					    eb, i, &key);
 			if (ret && ret != -ENOENT)
@@ -2033,8 +2085,7 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
 						eb, i, &key);
 			if (ret)
 				break;
-		} else if (key.type == BTRFS_DIR_ITEM_KEY ||
-			   key.type == BTRFS_DIR_INDEX_KEY) {
+		} else if (key.type == BTRFS_DIR_ITEM_KEY) {
 			ret = replay_one_dir_item(wc->trans, root, path,
 						  eb, i, &key);
 			if (ret)
@@ -2358,6 +2409,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	struct btrfs_root *log = root->log_root;
 	struct btrfs_root *log_root_tree = root->fs_info->log_root_tree;
 	unsigned long log_transid = 0;
+	struct blk_plug plug;
 
 	mutex_lock(&root->log_mutex);
 	log_transid = root->log_transid;
@@ -2401,8 +2453,10 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	/* we start IO on  all the marked extents here, but we don't actually
 	 * wait for them until later.
 	 */
+	blk_start_plug(&plug);
 	ret = btrfs_write_marked_extents(log, &log->dirty_log_pages, mark);
 	if (ret) {
+		blk_finish_plug(&plug);
 		btrfs_abort_transaction(trans, root, ret);
 		btrfs_free_logged_extents(log, log_transid);
 		mutex_unlock(&root->log_mutex);
@@ -2437,6 +2491,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	}
 
 	if (ret) {
+		blk_finish_plug(&plug);
 		if (ret != -ENOSPC) {
 			btrfs_abort_transaction(trans, root, ret);
 			mutex_unlock(&log_root_tree->log_mutex);
@@ -2452,6 +2507,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 
 	index2 = log_root_tree->log_transid % 2;
 	if (atomic_read(&log_root_tree->log_commit[index2])) {
+		blk_finish_plug(&plug);
 		btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
 		wait_log_commit(trans, log_root_tree,
 				log_root_tree->log_transid);
@@ -2474,6 +2530,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	 * check the full commit flag again
 	 */
 	if (root->fs_info->last_trans_log_full_commit == trans->transid) {
+		blk_finish_plug(&plug);
 		btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
 		btrfs_free_logged_extents(log, log_transid);
 		mutex_unlock(&log_root_tree->log_mutex);
@@ -2481,9 +2538,10 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 		goto out_wake_log_root;
 	}
 
-	ret = btrfs_write_and_wait_marked_extents(log_root_tree,
-				&log_root_tree->dirty_log_pages,
-				EXTENT_DIRTY | EXTENT_NEW);
+	ret = btrfs_write_marked_extents(log_root_tree,
+					 &log_root_tree->dirty_log_pages,
+					 EXTENT_DIRTY | EXTENT_NEW);
+	blk_finish_plug(&plug);
 	if (ret) {
 		btrfs_abort_transaction(trans, root, ret);
 		btrfs_free_logged_extents(log, log_transid);
@@ -2491,6 +2549,9 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 		goto out_wake_log_root;
 	}
 	btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
+	btrfs_wait_marked_extents(log_root_tree,
+				  &log_root_tree->dirty_log_pages,
+				  EXTENT_NEW | EXTENT_DIRTY);
 	btrfs_wait_logged_extents(log, log_transid);
 
 	btrfs_set_super_log_root(root->fs_info->super_for_commit,
@@ -3728,8 +3789,9 @@ next_slot:
 	}
 
 log_extents:
+	btrfs_release_path(path);
+	btrfs_release_path(dst_path);
 	if (fast_search) {
-		btrfs_release_path(dst_path);
 		ret = btrfs_log_changed_extents(trans, root, inode, dst_path);
 		if (ret) {
 			err = ret;
@@ -3746,8 +3808,6 @@ log_extents:
 	}
 
 	if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {
-		btrfs_release_path(path);
-		btrfs_release_path(dst_path);
 		ret = log_directory_changes(trans, root, inode, path, dst_path);
 		if (ret) {
 			err = ret;
@@ -3781,6 +3841,7 @@ static noinline int check_parent_dirs_for_sync(struct btrfs_trans_handle *trans,
 	int ret = 0;
 	struct btrfs_root *root;
 	struct dentry *old_parent = NULL;
+	struct inode *orig_inode = inode;
 
 	/*
 	 * for regular files, if its inode is already on disk, we don't
@@ -3800,7 +3861,14 @@ static noinline int check_parent_dirs_for_sync(struct btrfs_trans_handle *trans,
 	}
 
 	while (1) {
-		BTRFS_I(inode)->logged_trans = trans->transid;
+		/*
+		 * If we are logging a directory then we start with our inode,
+		 * not our parents inode, so we need to skipp setting the
+		 * logged_trans so that further down in the log code we don't
+		 * think this inode has already been logged.
+		 */
+		if (inode != orig_inode)
+			BTRFS_I(inode)->logged_trans = trans->transid;
 		smp_mb();
 
 		if (BTRFS_I(inode)->last_unlink_trans > last_committed) {
@@ -4016,8 +4084,7 @@ again:
 		if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
 			break;
 
-		log = btrfs_read_fs_root_no_radix(log_root_tree,
-						  &found_key);
+		log = btrfs_read_fs_root(log_root_tree, &found_key);
 		if (IS_ERR(log)) {
 			ret = PTR_ERR(log);
 			btrfs_error(fs_info, ret,
diff --git a/fs/btrfs/ulist.c b/fs/btrfs/ulist.c
index 7b417e20efe2..b0a523b2c60e 100644
--- a/fs/btrfs/ulist.c
+++ b/fs/btrfs/ulist.c
@@ -205,6 +205,10 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
 		u64 new_alloced = ulist->nodes_alloced + 128;
 		struct ulist_node *new_nodes;
 		void *old = NULL;
+		int i;
+
+		for (i = 0; i < ulist->nnodes; i++)
+			rb_erase(&ulist->nodes[i].rb_node, &ulist->root);
 
 		/*
 		 * if nodes_alloced == ULIST_SIZE no memory has been allocated
@@ -224,6 +228,17 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
 
 		ulist->nodes = new_nodes;
 		ulist->nodes_alloced = new_alloced;
+
+		/*
+		 * krealloc actually uses memcpy, which does not copy rb_node
+		 * pointers, so we have to do it ourselves.  Otherwise we may
+		 * be bitten by crashes.
+		 */
+		for (i = 0; i < ulist->nnodes; i++) {
+			ret = ulist_rbtree_insert(ulist, &ulist->nodes[i]);
+			if (ret < 0)
+				return ret;
+		}
 	}
 	ulist->nodes[ulist->nnodes].val = val;
 	ulist->nodes[ulist->nnodes].aux = aux;
diff --git a/fs/btrfs/uuid-tree.c b/fs/btrfs/uuid-tree.c
new file mode 100644
index 000000000000..dd0dea3766f7
--- /dev/null
+++ b/fs/btrfs/uuid-tree.c
@@ -0,0 +1,358 @@
+/*
+ * Copyright (C) STRATO AG 2013.  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 v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+#include <linux/uuid.h>
+#include <asm/unaligned.h>
+#include "ctree.h"
+#include "transaction.h"
+#include "disk-io.h"
+#include "print-tree.h"
+
+
+static void btrfs_uuid_to_key(u8 *uuid, u8 type, struct btrfs_key *key)
+{
+	key->type = type;
+	key->objectid = get_unaligned_le64(uuid);
+	key->offset = get_unaligned_le64(uuid + sizeof(u64));
+}
+
+/* return -ENOENT for !found, < 0 for errors, or 0 if an item was found */
+static int btrfs_uuid_tree_lookup(struct btrfs_root *uuid_root, u8 *uuid,
+				  u8 type, u64 subid)
+{
+	int ret;
+	struct btrfs_path *path = NULL;
+	struct extent_buffer *eb;
+	int slot;
+	u32 item_size;
+	unsigned long offset;
+	struct btrfs_key key;
+
+	if (WARN_ON_ONCE(!uuid_root)) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	btrfs_uuid_to_key(uuid, type, &key);
+	ret = btrfs_search_slot(NULL, uuid_root, &key, path, 0, 0);
+	if (ret < 0) {
+		goto out;
+	} else if (ret > 0) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	eb = path->nodes[0];
+	slot = path->slots[0];
+	item_size = btrfs_item_size_nr(eb, slot);
+	offset = btrfs_item_ptr_offset(eb, slot);
+	ret = -ENOENT;
+
+	if (!IS_ALIGNED(item_size, sizeof(u64))) {
+		pr_warn("btrfs: uuid item with illegal size %lu!\n",
+			(unsigned long)item_size);
+		goto out;
+	}
+	while (item_size) {
+		__le64 data;
+
+		read_extent_buffer(eb, &data, offset, sizeof(data));
+		if (le64_to_cpu(data) == subid) {
+			ret = 0;
+			break;
+		}
+		offset += sizeof(data);
+		item_size -= sizeof(data);
+	}
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_uuid_tree_add(struct btrfs_trans_handle *trans,
+			struct btrfs_root *uuid_root, u8 *uuid, u8 type,
+			u64 subid_cpu)
+{
+	int ret;
+	struct btrfs_path *path = NULL;
+	struct btrfs_key key;
+	struct extent_buffer *eb;
+	int slot;
+	unsigned long offset;
+	__le64 subid_le;
+
+	ret = btrfs_uuid_tree_lookup(uuid_root, uuid, type, subid_cpu);
+	if (ret != -ENOENT)
+		return ret;
+
+	if (WARN_ON_ONCE(!uuid_root)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	btrfs_uuid_to_key(uuid, type, &key);
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = btrfs_insert_empty_item(trans, uuid_root, path, &key,
+				      sizeof(subid_le));
+	if (ret >= 0) {
+		/* Add an item for the type for the first time */
+		eb = path->nodes[0];
+		slot = path->slots[0];
+		offset = btrfs_item_ptr_offset(eb, slot);
+	} else if (ret == -EEXIST) {
+		/*
+		 * An item with that type already exists.
+		 * Extend the item and store the new subid at the end.
+		 */
+		btrfs_extend_item(uuid_root, path, sizeof(subid_le));
+		eb = path->nodes[0];
+		slot = path->slots[0];
+		offset = btrfs_item_ptr_offset(eb, slot);
+		offset += btrfs_item_size_nr(eb, slot) - sizeof(subid_le);
+	} else if (ret < 0) {
+		pr_warn("btrfs: insert uuid item failed %d (0x%016llx, 0x%016llx) type %u!\n",
+			ret, (unsigned long long)key.objectid,
+			(unsigned long long)key.offset, type);
+		goto out;
+	}
+
+	ret = 0;
+	subid_le = cpu_to_le64(subid_cpu);
+	write_extent_buffer(eb, &subid_le, offset, sizeof(subid_le));
+	btrfs_mark_buffer_dirty(eb);
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_uuid_tree_rem(struct btrfs_trans_handle *trans,
+			struct btrfs_root *uuid_root, u8 *uuid, u8 type,
+			u64 subid)
+{
+	int ret;
+	struct btrfs_path *path = NULL;
+	struct btrfs_key key;
+	struct extent_buffer *eb;
+	int slot;
+	unsigned long offset;
+	u32 item_size;
+	unsigned long move_dst;
+	unsigned long move_src;
+	unsigned long move_len;
+
+	if (WARN_ON_ONCE(!uuid_root)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	btrfs_uuid_to_key(uuid, type, &key);
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = btrfs_search_slot(trans, uuid_root, &key, path, -1, 1);
+	if (ret < 0) {
+		pr_warn("btrfs: error %d while searching for uuid item!\n",
+			ret);
+		goto out;
+	}
+	if (ret > 0) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	eb = path->nodes[0];
+	slot = path->slots[0];
+	offset = btrfs_item_ptr_offset(eb, slot);
+	item_size = btrfs_item_size_nr(eb, slot);
+	if (!IS_ALIGNED(item_size, sizeof(u64))) {
+		pr_warn("btrfs: uuid item with illegal size %lu!\n",
+			(unsigned long)item_size);
+		ret = -ENOENT;
+		goto out;
+	}
+	while (item_size) {
+		__le64 read_subid;
+
+		read_extent_buffer(eb, &read_subid, offset, sizeof(read_subid));
+		if (le64_to_cpu(read_subid) == subid)
+			break;
+		offset += sizeof(read_subid);
+		item_size -= sizeof(read_subid);
+	}
+
+	if (!item_size) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	item_size = btrfs_item_size_nr(eb, slot);
+	if (item_size == sizeof(subid)) {
+		ret = btrfs_del_item(trans, uuid_root, path);
+		goto out;
+	}
+
+	move_dst = offset;
+	move_src = offset + sizeof(subid);
+	move_len = item_size - (move_src - btrfs_item_ptr_offset(eb, slot));
+	memmove_extent_buffer(eb, move_dst, move_src, move_len);
+	btrfs_truncate_item(uuid_root, path, item_size - sizeof(subid), 1);
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_uuid_iter_rem(struct btrfs_root *uuid_root, u8 *uuid, u8 type,
+			       u64 subid)
+{
+	struct btrfs_trans_handle *trans;
+	int ret;
+
+	/* 1 - for the uuid item */
+	trans = btrfs_start_transaction(uuid_root, 1);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		goto out;
+	}
+
+	ret = btrfs_uuid_tree_rem(trans, uuid_root, uuid, type, subid);
+	btrfs_end_transaction(trans, uuid_root);
+
+out:
+	return ret;
+}
+
+int btrfs_uuid_tree_iterate(struct btrfs_fs_info *fs_info,
+			    int (*check_func)(struct btrfs_fs_info *, u8 *, u8,
+					      u64))
+{
+	struct btrfs_root *root = fs_info->uuid_root;
+	struct btrfs_key key;
+	struct btrfs_key max_key;
+	struct btrfs_path *path;
+	int ret = 0;
+	struct extent_buffer *leaf;
+	int slot;
+	u32 item_size;
+	unsigned long offset;
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	key.objectid = 0;
+	key.type = 0;
+	key.offset = 0;
+	max_key.objectid = (u64)-1;
+	max_key.type = (u8)-1;
+	max_key.offset = (u64)-1;
+
+again_search_slot:
+	path->keep_locks = 1;
+	ret = btrfs_search_forward(root, &key, &max_key, path, 0);
+	if (ret) {
+		if (ret > 0)
+			ret = 0;
+		goto out;
+	}
+
+	while (1) {
+		cond_resched();
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+
+		if (key.type != BTRFS_UUID_KEY_SUBVOL &&
+		    key.type != BTRFS_UUID_KEY_RECEIVED_SUBVOL)
+			goto skip;
+
+		offset = btrfs_item_ptr_offset(leaf, slot);
+		item_size = btrfs_item_size_nr(leaf, slot);
+		if (!IS_ALIGNED(item_size, sizeof(u64))) {
+			pr_warn("btrfs: uuid item with illegal size %lu!\n",
+				(unsigned long)item_size);
+			goto skip;
+		}
+		while (item_size) {
+			u8 uuid[BTRFS_UUID_SIZE];
+			__le64 subid_le;
+			u64 subid_cpu;
+
+			put_unaligned_le64(key.objectid, uuid);
+			put_unaligned_le64(key.offset, uuid + sizeof(u64));
+			read_extent_buffer(leaf, &subid_le, offset,
+					   sizeof(subid_le));
+			subid_cpu = le64_to_cpu(subid_le);
+			ret = check_func(fs_info, uuid, key.type, subid_cpu);
+			if (ret < 0)
+				goto out;
+			if (ret > 0) {
+				btrfs_release_path(path);
+				ret = btrfs_uuid_iter_rem(root, uuid, key.type,
+							  subid_cpu);
+				if (ret == 0) {
+					/*
+					 * this might look inefficient, but the
+					 * justification is that it is an
+					 * exception that check_func returns 1,
+					 * and that in the regular case only one
+					 * entry per UUID exists.
+					 */
+					goto again_search_slot;
+				}
+				if (ret < 0 && ret != -ENOENT)
+					goto out;
+			}
+			item_size -= sizeof(subid_le);
+			offset += sizeof(subid_le);
+		}
+
+skip:
+		ret = btrfs_next_item(root, path);
+		if (ret == 0)
+			continue;
+		else if (ret > 0)
+			ret = 0;
+		break;
+	}
+
+out:
+	btrfs_free_path(path);
+	if (ret)
+		pr_warn("btrfs: btrfs_uuid_tree_iterate failed %d\n", ret);
+	return 0;
+}
diff --git a/fs/btrfs/version.h b/fs/btrfs/version.h
deleted file mode 100644
index 9bf3946d5ef2..000000000000
--- a/fs/btrfs/version.h
+++ /dev/null
@@ -1,4 +0,0 @@
-#ifndef __BTRFS_VERSION_H
-#define __BTRFS_VERSION_H
-#define BTRFS_BUILD_VERSION "Btrfs"
-#endif
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 8bffb9174afb..043b215769c2 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -26,6 +26,7 @@
 #include <linux/ratelimit.h>
 #include <linux/kthread.h>
 #include <linux/raid/pq.h>
+#include <linux/semaphore.h>
 #include <asm/div64.h>
 #include "compat.h"
 #include "ctree.h"
@@ -62,6 +63,48 @@ static void unlock_chunks(struct btrfs_root *root)
 	mutex_unlock(&root->fs_info->chunk_mutex);
 }
 
+static struct btrfs_fs_devices *__alloc_fs_devices(void)
+{
+	struct btrfs_fs_devices *fs_devs;
+
+	fs_devs = kzalloc(sizeof(*fs_devs), GFP_NOFS);
+	if (!fs_devs)
+		return ERR_PTR(-ENOMEM);
+
+	mutex_init(&fs_devs->device_list_mutex);
+
+	INIT_LIST_HEAD(&fs_devs->devices);
+	INIT_LIST_HEAD(&fs_devs->alloc_list);
+	INIT_LIST_HEAD(&fs_devs->list);
+
+	return fs_devs;
+}
+
+/**
+ * alloc_fs_devices - allocate struct btrfs_fs_devices
+ * @fsid:	a pointer to UUID for this FS.  If NULL a new UUID is
+ *		generated.
+ *
+ * Return: a pointer to a new &struct btrfs_fs_devices on success;
+ * ERR_PTR() on error.  Returned struct is not linked onto any lists and
+ * can be destroyed with kfree() right away.
+ */
+static struct btrfs_fs_devices *alloc_fs_devices(const u8 *fsid)
+{
+	struct btrfs_fs_devices *fs_devs;
+
+	fs_devs = __alloc_fs_devices();
+	if (IS_ERR(fs_devs))
+		return fs_devs;
+
+	if (fsid)
+		memcpy(fs_devs->fsid, fsid, BTRFS_FSID_SIZE);
+	else
+		generate_random_uuid(fs_devs->fsid);
+
+	return fs_devs;
+}
+
 static void free_fs_devices(struct btrfs_fs_devices *fs_devices)
 {
 	struct btrfs_device *device;
@@ -101,6 +144,27 @@ void btrfs_cleanup_fs_uuids(void)
 	}
 }
 
+static struct btrfs_device *__alloc_device(void)
+{
+	struct btrfs_device *dev;
+
+	dev = kzalloc(sizeof(*dev), GFP_NOFS);
+	if (!dev)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&dev->dev_list);
+	INIT_LIST_HEAD(&dev->dev_alloc_list);
+
+	spin_lock_init(&dev->io_lock);
+
+	spin_lock_init(&dev->reada_lock);
+	atomic_set(&dev->reada_in_flight, 0);
+	INIT_RADIX_TREE(&dev->reada_zones, GFP_NOFS & ~__GFP_WAIT);
+	INIT_RADIX_TREE(&dev->reada_extents, GFP_NOFS & ~__GFP_WAIT);
+
+	return dev;
+}
+
 static noinline struct btrfs_device *__find_device(struct list_head *head,
 						   u64 devid, u8 *uuid)
 {
@@ -395,16 +459,14 @@ static noinline int device_list_add(const char *path,
 
 	fs_devices = find_fsid(disk_super->fsid);
 	if (!fs_devices) {
-		fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
-		if (!fs_devices)
-			return -ENOMEM;
-		INIT_LIST_HEAD(&fs_devices->devices);
-		INIT_LIST_HEAD(&fs_devices->alloc_list);
+		fs_devices = alloc_fs_devices(disk_super->fsid);
+		if (IS_ERR(fs_devices))
+			return PTR_ERR(fs_devices);
+
 		list_add(&fs_devices->list, &fs_uuids);
-		memcpy(fs_devices->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
 		fs_devices->latest_devid = devid;
 		fs_devices->latest_trans = found_transid;
-		mutex_init(&fs_devices->device_list_mutex);
+
 		device = NULL;
 	} else {
 		device = __find_device(&fs_devices->devices, devid,
@@ -414,17 +476,12 @@ static noinline int device_list_add(const char *path,
 		if (fs_devices->opened)
 			return -EBUSY;
 
-		device = kzalloc(sizeof(*device), GFP_NOFS);
-		if (!device) {
+		device = btrfs_alloc_device(NULL, &devid,
+					    disk_super->dev_item.uuid);
+		if (IS_ERR(device)) {
 			/* we can safely leave the fs_devices entry around */
-			return -ENOMEM;
+			return PTR_ERR(device);
 		}
-		device->devid = devid;
-		device->dev_stats_valid = 0;
-		device->work.func = pending_bios_fn;
-		memcpy(device->uuid, disk_super->dev_item.uuid,
-		       BTRFS_UUID_SIZE);
-		spin_lock_init(&device->io_lock);
 
 		name = rcu_string_strdup(path, GFP_NOFS);
 		if (!name) {
@@ -432,22 +489,13 @@ static noinline int device_list_add(const char *path,
 			return -ENOMEM;
 		}
 		rcu_assign_pointer(device->name, name);
-		INIT_LIST_HEAD(&device->dev_alloc_list);
-
-		/* init readahead state */
-		spin_lock_init(&device->reada_lock);
-		device->reada_curr_zone = NULL;
-		atomic_set(&device->reada_in_flight, 0);
-		device->reada_next = 0;
-		INIT_RADIX_TREE(&device->reada_zones, GFP_NOFS & ~__GFP_WAIT);
-		INIT_RADIX_TREE(&device->reada_extents, GFP_NOFS & ~__GFP_WAIT);
 
 		mutex_lock(&fs_devices->device_list_mutex);
 		list_add_rcu(&device->dev_list, &fs_devices->devices);
+		fs_devices->num_devices++;
 		mutex_unlock(&fs_devices->device_list_mutex);
 
 		device->fs_devices = fs_devices;
-		fs_devices->num_devices++;
 	} else if (!device->name || strcmp(device->name->str, path)) {
 		name = rcu_string_strdup(path, GFP_NOFS);
 		if (!name)
@@ -474,25 +522,21 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
 	struct btrfs_device *device;
 	struct btrfs_device *orig_dev;
 
-	fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
-	if (!fs_devices)
-		return ERR_PTR(-ENOMEM);
+	fs_devices = alloc_fs_devices(orig->fsid);
+	if (IS_ERR(fs_devices))
+		return fs_devices;
 
-	INIT_LIST_HEAD(&fs_devices->devices);
-	INIT_LIST_HEAD(&fs_devices->alloc_list);
-	INIT_LIST_HEAD(&fs_devices->list);
-	mutex_init(&fs_devices->device_list_mutex);
 	fs_devices->latest_devid = orig->latest_devid;
 	fs_devices->latest_trans = orig->latest_trans;
 	fs_devices->total_devices = orig->total_devices;
-	memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid));
 
 	/* We have held the volume lock, it is safe to get the devices. */
 	list_for_each_entry(orig_dev, &orig->devices, dev_list) {
 		struct rcu_string *name;
 
-		device = kzalloc(sizeof(*device), GFP_NOFS);
-		if (!device)
+		device = btrfs_alloc_device(NULL, &orig_dev->devid,
+					    orig_dev->uuid);
+		if (IS_ERR(device))
 			goto error;
 
 		/*
@@ -506,13 +550,6 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
 		}
 		rcu_assign_pointer(device->name, name);
 
-		device->devid = orig_dev->devid;
-		device->work.func = pending_bios_fn;
-		memcpy(device->uuid, orig_dev->uuid, sizeof(device->uuid));
-		spin_lock_init(&device->io_lock);
-		INIT_LIST_HEAD(&device->dev_list);
-		INIT_LIST_HEAD(&device->dev_alloc_list);
-
 		list_add(&device->dev_list, &fs_devices->devices);
 		device->fs_devices = fs_devices;
 		fs_devices->num_devices++;
@@ -636,23 +673,22 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
 
 		if (device->can_discard)
 			fs_devices->num_can_discard--;
+		if (device->missing)
+			fs_devices->missing_devices--;
 
-		new_device = kmalloc(sizeof(*new_device), GFP_NOFS);
-		BUG_ON(!new_device); /* -ENOMEM */
-		memcpy(new_device, device, sizeof(*new_device));
+		new_device = btrfs_alloc_device(NULL, &device->devid,
+						device->uuid);
+		BUG_ON(IS_ERR(new_device)); /* -ENOMEM */
 
 		/* Safe because we are under uuid_mutex */
 		if (device->name) {
 			name = rcu_string_strdup(device->name->str, GFP_NOFS);
-			BUG_ON(device->name && !name); /* -ENOMEM */
+			BUG_ON(!name); /* -ENOMEM */
 			rcu_assign_pointer(new_device->name, name);
 		}
-		new_device->bdev = NULL;
-		new_device->writeable = 0;
-		new_device->in_fs_metadata = 0;
-		new_device->can_discard = 0;
-		spin_lock_init(&new_device->io_lock);
+
 		list_replace_rcu(&device->dev_list, &new_device->dev_list);
+		new_device->fs_devices = device->fs_devices;
 
 		call_rcu(&device->rcu, free_device);
 	}
@@ -760,7 +796,8 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
 			fs_devices->rotating = 1;
 
 		fs_devices->open_devices++;
-		if (device->writeable && !device->is_tgtdev_for_dev_replace) {
+		if (device->writeable &&
+		    device->devid != BTRFS_DEV_REPLACE_DEVID) {
 			fs_devices->rw_devices++;
 			list_add(&device->dev_alloc_list,
 				 &fs_devices->alloc_list);
@@ -865,7 +902,7 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 	disk_super = p + (bytenr & ~PAGE_CACHE_MASK);
 
 	if (btrfs_super_bytenr(disk_super) != bytenr ||
-	    disk_super->magic != cpu_to_le64(BTRFS_MAGIC))
+	    btrfs_super_magic(disk_super) != BTRFS_MAGIC)
 		goto error_unmap;
 
 	devid = btrfs_stack_device_id(&disk_super->dev_item);
@@ -875,13 +912,12 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 	if (disk_super->label[0]) {
 		if (disk_super->label[BTRFS_LABEL_SIZE - 1])
 			disk_super->label[BTRFS_LABEL_SIZE - 1] = '\0';
-		printk(KERN_INFO "device label %s ", disk_super->label);
+		printk(KERN_INFO "btrfs: device label %s ", disk_super->label);
 	} else {
-		printk(KERN_INFO "device fsid %pU ", disk_super->fsid);
+		printk(KERN_INFO "btrfs: device fsid %pU ", disk_super->fsid);
 	}
 
-	printk(KERN_CONT "devid %llu transid %llu %s\n",
-	       (unsigned long long)devid, (unsigned long long)transid, path);
+	printk(KERN_CONT "devid %llu transid %llu %s\n", devid, transid, path);
 
 	ret = device_list_add(path, disk_super, devid, fs_devices_ret);
 	if (!ret && fs_devices_ret)
@@ -982,6 +1018,35 @@ out:
 	return ret;
 }
 
+static int contains_pending_extent(struct btrfs_trans_handle *trans,
+				   struct btrfs_device *device,
+				   u64 *start, u64 len)
+{
+	struct extent_map *em;
+	int ret = 0;
+
+	list_for_each_entry(em, &trans->transaction->pending_chunks, list) {
+		struct map_lookup *map;
+		int i;
+
+		map = (struct map_lookup *)em->bdev;
+		for (i = 0; i < map->num_stripes; i++) {
+			if (map->stripes[i].dev != device)
+				continue;
+			if (map->stripes[i].physical >= *start + len ||
+			    map->stripes[i].physical + em->orig_block_len <=
+			    *start)
+				continue;
+			*start = map->stripes[i].physical +
+				em->orig_block_len;
+			ret = 1;
+		}
+	}
+
+	return ret;
+}
+
+
 /*
  * find_free_dev_extent - find free space in the specified device
  * @device:	the device which we search the free space in
@@ -1002,7 +1067,8 @@ out:
  * But if we don't find suitable free space, it is used to store the size of
  * the max free space.
  */
-int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
+int find_free_dev_extent(struct btrfs_trans_handle *trans,
+			 struct btrfs_device *device, u64 num_bytes,
 			 u64 *start, u64 *len)
 {
 	struct btrfs_key key;
@@ -1026,21 +1092,22 @@ int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
 	 */
 	search_start = max(root->fs_info->alloc_start, 1024ull * 1024);
 
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+again:
 	max_hole_start = search_start;
 	max_hole_size = 0;
 	hole_size = 0;
 
 	if (search_start >= search_end || device->is_tgtdev_for_dev_replace) {
 		ret = -ENOSPC;
-		goto error;
+		goto out;
 	}
 
-	path = btrfs_alloc_path();
-	if (!path) {
-		ret = -ENOMEM;
-		goto error;
-	}
 	path->reada = 2;
+	path->search_commit_root = 1;
+	path->skip_locking = 1;
 
 	key.objectid = device->devid;
 	key.offset = search_start;
@@ -1081,6 +1148,15 @@ int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
 		if (key.offset > search_start) {
 			hole_size = key.offset - search_start;
 
+			/*
+			 * Have to check before we set max_hole_start, otherwise
+			 * we could end up sending back this offset anyway.
+			 */
+			if (contains_pending_extent(trans, device,
+						    &search_start,
+						    hole_size))
+				hole_size = 0;
+
 			if (hole_size > max_hole_size) {
 				max_hole_start = search_start;
 				max_hole_size = hole_size;
@@ -1124,6 +1200,11 @@ next:
 		max_hole_size = hole_size;
 	}
 
+	if (contains_pending_extent(trans, device, &search_start, hole_size)) {
+		btrfs_release_path(path);
+		goto again;
+	}
+
 	/* See above. */
 	if (hole_size < num_bytes)
 		ret = -ENOSPC;
@@ -1132,7 +1213,6 @@ next:
 
 out:
 	btrfs_free_path(path);
-error:
 	*start = max_hole_start;
 	if (len)
 		*len = max_hole_size;
@@ -1234,8 +1314,7 @@ static int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
 	btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);
 
 	write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
-		    (unsigned long)btrfs_dev_extent_chunk_tree_uuid(extent),
-		    BTRFS_UUID_SIZE);
+		    btrfs_dev_extent_chunk_tree_uuid(extent), BTRFS_UUID_SIZE);
 
 	btrfs_set_dev_extent_length(leaf, extent, num_bytes);
 	btrfs_mark_buffer_dirty(leaf);
@@ -1244,59 +1323,33 @@ out:
 	return ret;
 }
 
-static noinline int find_next_chunk(struct btrfs_root *root,
-				    u64 objectid, u64 *offset)
+static u64 find_next_chunk(struct btrfs_fs_info *fs_info)
 {
-	struct btrfs_path *path;
-	int ret;
-	struct btrfs_key key;
-	struct btrfs_chunk *chunk;
-	struct btrfs_key found_key;
-
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
-
-	key.objectid = objectid;
-	key.offset = (u64)-1;
-	key.type = BTRFS_CHUNK_ITEM_KEY;
-
-	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-	if (ret < 0)
-		goto error;
-
-	BUG_ON(ret == 0); /* Corruption */
+	struct extent_map_tree *em_tree;
+	struct extent_map *em;
+	struct rb_node *n;
+	u64 ret = 0;
 
-	ret = btrfs_previous_item(root, path, 0, BTRFS_CHUNK_ITEM_KEY);
-	if (ret) {
-		*offset = 0;
-	} else {
-		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
-				      path->slots[0]);
-		if (found_key.objectid != objectid)
-			*offset = 0;
-		else {
-			chunk = btrfs_item_ptr(path->nodes[0], path->slots[0],
-					       struct btrfs_chunk);
-			*offset = found_key.offset +
-				btrfs_chunk_length(path->nodes[0], chunk);
-		}
+	em_tree = &fs_info->mapping_tree.map_tree;
+	read_lock(&em_tree->lock);
+	n = rb_last(&em_tree->map);
+	if (n) {
+		em = rb_entry(n, struct extent_map, rb_node);
+		ret = em->start + em->len;
 	}
-	ret = 0;
-error:
-	btrfs_free_path(path);
+	read_unlock(&em_tree->lock);
+
 	return ret;
 }
 
-static noinline int find_next_devid(struct btrfs_root *root, u64 *objectid)
+static noinline int find_next_devid(struct btrfs_fs_info *fs_info,
+				    u64 *devid_ret)
 {
 	int ret;
 	struct btrfs_key key;
 	struct btrfs_key found_key;
 	struct btrfs_path *path;
 
-	root = root->fs_info->chunk_root;
-
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -1305,20 +1358,21 @@ static noinline int find_next_devid(struct btrfs_root *root, u64 *objectid)
 	key.type = BTRFS_DEV_ITEM_KEY;
 	key.offset = (u64)-1;
 
-	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	ret = btrfs_search_slot(NULL, fs_info->chunk_root, &key, path, 0, 0);
 	if (ret < 0)
 		goto error;
 
 	BUG_ON(ret == 0); /* Corruption */
 
-	ret = btrfs_previous_item(root, path, BTRFS_DEV_ITEMS_OBJECTID,
+	ret = btrfs_previous_item(fs_info->chunk_root, path,
+				  BTRFS_DEV_ITEMS_OBJECTID,
 				  BTRFS_DEV_ITEM_KEY);
 	if (ret) {
-		*objectid = 1;
+		*devid_ret = 1;
 	} else {
 		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
 				      path->slots[0]);
-		*objectid = found_key.offset + 1;
+		*devid_ret = found_key.offset + 1;
 	}
 	ret = 0;
 error:
@@ -1372,9 +1426,9 @@ static int btrfs_add_device(struct btrfs_trans_handle *trans,
 	btrfs_set_device_bandwidth(leaf, dev_item, 0);
 	btrfs_set_device_start_offset(leaf, dev_item, 0);
 
-	ptr = (unsigned long)btrfs_device_uuid(dev_item);
+	ptr = btrfs_device_uuid(dev_item);
 	write_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
-	ptr = (unsigned long)btrfs_device_fsid(dev_item);
+	ptr = btrfs_device_fsid(dev_item);
 	write_extent_buffer(leaf, root->fs_info->fsid, ptr, BTRFS_UUID_SIZE);
 	btrfs_mark_buffer_dirty(leaf);
 
@@ -1462,31 +1516,23 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 	btrfs_dev_replace_unlock(&root->fs_info->dev_replace);
 
 	if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices <= 4) {
-		printk(KERN_ERR "btrfs: unable to go below four devices "
-		       "on raid10\n");
-		ret = -EINVAL;
+		ret = BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET;
 		goto out;
 	}
 
 	if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices <= 2) {
-		printk(KERN_ERR "btrfs: unable to go below two "
-		       "devices on raid1\n");
-		ret = -EINVAL;
+		ret = BTRFS_ERROR_DEV_RAID1_MIN_NOT_MET;
 		goto out;
 	}
 
 	if ((all_avail & BTRFS_BLOCK_GROUP_RAID5) &&
 	    root->fs_info->fs_devices->rw_devices <= 2) {
-		printk(KERN_ERR "btrfs: unable to go below two "
-		       "devices on raid5\n");
-		ret = -EINVAL;
+		ret = BTRFS_ERROR_DEV_RAID5_MIN_NOT_MET;
 		goto out;
 	}
 	if ((all_avail & BTRFS_BLOCK_GROUP_RAID6) &&
 	    root->fs_info->fs_devices->rw_devices <= 3) {
-		printk(KERN_ERR "btrfs: unable to go below three "
-		       "devices on raid6\n");
-		ret = -EINVAL;
+		ret = BTRFS_ERROR_DEV_RAID6_MIN_NOT_MET;
 		goto out;
 	}
 
@@ -1512,8 +1558,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 		bh = NULL;
 		disk_super = NULL;
 		if (!device) {
-			printk(KERN_ERR "btrfs: no missing devices found to "
-			       "remove\n");
+			ret = BTRFS_ERROR_DEV_MISSING_NOT_FOUND;
 			goto out;
 		}
 	} else {
@@ -1535,15 +1580,12 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 	}
 
 	if (device->is_tgtdev_for_dev_replace) {
-		pr_err("btrfs: unable to remove the dev_replace target dev\n");
-		ret = -EINVAL;
+		ret = BTRFS_ERROR_DEV_TGT_REPLACE;
 		goto error_brelse;
 	}
 
 	if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) {
-		printk(KERN_ERR "btrfs: unable to remove the only writeable "
-		       "device\n");
-		ret = -EINVAL;
+		ret = BTRFS_ERROR_DEV_ONLY_WRITABLE;
 		goto error_brelse;
 	}
 
@@ -1555,7 +1597,9 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 		clear_super = true;
 	}
 
+	mutex_unlock(&uuid_mutex);
 	ret = btrfs_shrink_device(device, 0);
+	mutex_lock(&uuid_mutex);
 	if (ret)
 		goto error_undo;
 
@@ -1579,7 +1623,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 	/*
 	 * the device list mutex makes sure that we don't change
 	 * the device list while someone else is writing out all
-	 * the device supers.
+	 * the device supers. Whoever is writing all supers, should
+	 * lock the device list mutex before getting the number of
+	 * devices in the super block (super_copy). Conversely,
+	 * whoever updates the number of devices in the super block
+	 * (super_copy) should hold the device list mutex.
 	 */
 
 	cur_devices = device->fs_devices;
@@ -1603,10 +1651,10 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 		device->fs_devices->open_devices--;
 
 	call_rcu(&device->rcu, free_device);
-	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
 	num_devices = btrfs_super_num_devices(root->fs_info->super_copy) - 1;
 	btrfs_set_super_num_devices(root->fs_info->super_copy, num_devices);
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
 	if (cur_devices->open_devices == 0) {
 		struct btrfs_fs_devices *fs_devices;
@@ -1668,6 +1716,7 @@ void btrfs_rm_dev_replace_srcdev(struct btrfs_fs_info *fs_info,
 				 struct btrfs_device *srcdev)
 {
 	WARN_ON(!mutex_is_locked(&fs_info->fs_devices->device_list_mutex));
+
 	list_del_rcu(&srcdev->dev_list);
 	list_del_rcu(&srcdev->dev_alloc_list);
 	fs_info->fs_devices->num_devices--;
@@ -1677,9 +1726,13 @@ void btrfs_rm_dev_replace_srcdev(struct btrfs_fs_info *fs_info,
 	}
 	if (srcdev->can_discard)
 		fs_info->fs_devices->num_can_discard--;
-	if (srcdev->bdev)
+	if (srcdev->bdev) {
 		fs_info->fs_devices->open_devices--;
 
+		/* zero out the old super */
+		btrfs_scratch_superblock(srcdev);
+	}
+
 	call_rcu(&srcdev->rcu, free_device);
 }
 
@@ -1786,9 +1839,9 @@ static int btrfs_prepare_sprout(struct btrfs_root *root)
 	if (!fs_devices->seeding)
 		return -EINVAL;
 
-	seed_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
-	if (!seed_devices)
-		return -ENOMEM;
+	seed_devices = __alloc_fs_devices();
+	if (IS_ERR(seed_devices))
+		return PTR_ERR(seed_devices);
 
 	old_devices = clone_fs_devices(fs_devices);
 	if (IS_ERR(old_devices)) {
@@ -1807,7 +1860,6 @@ static int btrfs_prepare_sprout(struct btrfs_root *root)
 	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
 	list_splice_init_rcu(&fs_devices->devices, &seed_devices->devices,
 			      synchronize_rcu);
-	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
 	list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list);
 	list_for_each_entry(device, &seed_devices->devices, dev_list) {
@@ -1823,6 +1875,8 @@ static int btrfs_prepare_sprout(struct btrfs_root *root)
 	generate_random_uuid(fs_devices->fsid);
 	memcpy(root->fs_info->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
 	memcpy(disk_super->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+
 	super_flags = btrfs_super_flags(disk_super) &
 		      ~BTRFS_SUPER_FLAG_SEEDING;
 	btrfs_set_super_flags(disk_super, super_flags);
@@ -1882,11 +1936,9 @@ next_slot:
 		dev_item = btrfs_item_ptr(leaf, path->slots[0],
 					  struct btrfs_dev_item);
 		devid = btrfs_device_id(leaf, dev_item);
-		read_extent_buffer(leaf, dev_uuid,
-				   (unsigned long)btrfs_device_uuid(dev_item),
+		read_extent_buffer(leaf, dev_uuid, btrfs_device_uuid(dev_item),
 				   BTRFS_UUID_SIZE);
-		read_extent_buffer(leaf, fs_uuid,
-				   (unsigned long)btrfs_device_fsid(dev_item),
+		read_extent_buffer(leaf, fs_uuid, btrfs_device_fsid(dev_item),
 				   BTRFS_UUID_SIZE);
 		device = btrfs_find_device(root->fs_info, devid, dev_uuid,
 					   fs_uuid);
@@ -1949,10 +2001,10 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 	}
 	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
-	device = kzalloc(sizeof(*device), GFP_NOFS);
-	if (!device) {
+	device = btrfs_alloc_device(root->fs_info, NULL, NULL);
+	if (IS_ERR(device)) {
 		/* we can safely leave the fs_devices entry around */
-		ret = -ENOMEM;
+		ret = PTR_ERR(device);
 		goto error;
 	}
 
@@ -1964,13 +2016,6 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 	}
 	rcu_assign_pointer(device->name, name);
 
-	ret = find_next_devid(root, &device->devid);
-	if (ret) {
-		rcu_string_free(device->name);
-		kfree(device);
-		goto error;
-	}
-
 	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
 		rcu_string_free(device->name);
@@ -1985,9 +2030,6 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 	if (blk_queue_discard(q))
 		device->can_discard = 1;
 	device->writeable = 1;
-	device->work.func = pending_bios_fn;
-	generate_random_uuid(device->uuid);
-	spin_lock_init(&device->io_lock);
 	device->generation = trans->transid;
 	device->io_width = root->sectorsize;
 	device->io_align = root->sectorsize;
@@ -2114,6 +2156,7 @@ int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path,
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct list_head *devices;
 	struct rcu_string *name;
+	u64 devid = BTRFS_DEV_REPLACE_DEVID;
 	int ret = 0;
 
 	*device_out = NULL;
@@ -2135,9 +2178,9 @@ int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path,
 		}
 	}
 
-	device = kzalloc(sizeof(*device), GFP_NOFS);
-	if (!device) {
-		ret = -ENOMEM;
+	device = btrfs_alloc_device(NULL, &devid, NULL);
+	if (IS_ERR(device)) {
+		ret = PTR_ERR(device);
 		goto error;
 	}
 
@@ -2154,10 +2197,6 @@ int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path,
 		device->can_discard = 1;
 	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
 	device->writeable = 1;
-	device->work.func = pending_bios_fn;
-	generate_random_uuid(device->uuid);
-	device->devid = BTRFS_DEV_REPLACE_DEVID;
-	spin_lock_init(&device->io_lock);
 	device->generation = 0;
 	device->io_width = root->sectorsize;
 	device->io_align = root->sectorsize;
@@ -2964,10 +3003,6 @@ again:
 		if (found_key.objectid != key.objectid)
 			break;
 
-		/* chunk zero is special */
-		if (found_key.offset == 0)
-			break;
-
 		chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
 
 		if (!counting) {
@@ -3003,6 +3038,8 @@ again:
 			spin_unlock(&fs_info->balance_lock);
 		}
 loop:
+		if (found_key.offset == 0)
+			break;
 		key.offset = found_key.offset - 1;
 	}
 
@@ -3067,9 +3104,6 @@ static void __cancel_balance(struct btrfs_fs_info *fs_info)
 	atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
 }
 
-void update_ioctl_balance_args(struct btrfs_fs_info *fs_info, int lock,
-			       struct btrfs_ioctl_balance_args *bargs);
-
 /*
  * Should be called with both balance and volume mutexes held
  */
@@ -3132,7 +3166,7 @@ int btrfs_balance(struct btrfs_balance_control *bctl,
 	     (bctl->data.target & ~allowed))) {
 		printk(KERN_ERR "btrfs: unable to start balance with target "
 		       "data profile %llu\n",
-		       (unsigned long long)bctl->data.target);
+		       bctl->data.target);
 		ret = -EINVAL;
 		goto out;
 	}
@@ -3141,7 +3175,7 @@ int btrfs_balance(struct btrfs_balance_control *bctl,
 	     (bctl->meta.target & ~allowed))) {
 		printk(KERN_ERR "btrfs: unable to start balance with target "
 		       "metadata profile %llu\n",
-		       (unsigned long long)bctl->meta.target);
+		       bctl->meta.target);
 		ret = -EINVAL;
 		goto out;
 	}
@@ -3150,7 +3184,7 @@ int btrfs_balance(struct btrfs_balance_control *bctl,
 	     (bctl->sys.target & ~allowed))) {
 		printk(KERN_ERR "btrfs: unable to start balance with target "
 		       "system profile %llu\n",
-		       (unsigned long long)bctl->sys.target);
+		       bctl->sys.target);
 		ret = -EINVAL;
 		goto out;
 	}
@@ -3295,10 +3329,7 @@ int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info)
 	}
 
 	tsk = kthread_run(balance_kthread, fs_info, "btrfs-balance");
-	if (IS_ERR(tsk))
-		return PTR_ERR(tsk);
-
-	return 0;
+	return PTR_ERR_OR_ZERO(tsk);
 }
 
 int btrfs_recover_balance(struct btrfs_fs_info *fs_info)
@@ -3426,6 +3457,264 @@ int btrfs_cancel_balance(struct btrfs_fs_info *fs_info)
 	return 0;
 }
 
+static int btrfs_uuid_scan_kthread(void *data)
+{
+	struct btrfs_fs_info *fs_info = data;
+	struct btrfs_root *root = fs_info->tree_root;
+	struct btrfs_key key;
+	struct btrfs_key max_key;
+	struct btrfs_path *path = NULL;
+	int ret = 0;
+	struct extent_buffer *eb;
+	int slot;
+	struct btrfs_root_item root_item;
+	u32 item_size;
+	struct btrfs_trans_handle *trans = NULL;
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	key.objectid = 0;
+	key.type = BTRFS_ROOT_ITEM_KEY;
+	key.offset = 0;
+
+	max_key.objectid = (u64)-1;
+	max_key.type = BTRFS_ROOT_ITEM_KEY;
+	max_key.offset = (u64)-1;
+
+	path->keep_locks = 1;
+
+	while (1) {
+		ret = btrfs_search_forward(root, &key, &max_key, path, 0);
+		if (ret) {
+			if (ret > 0)
+				ret = 0;
+			break;
+		}
+
+		if (key.type != BTRFS_ROOT_ITEM_KEY ||
+		    (key.objectid < BTRFS_FIRST_FREE_OBJECTID &&
+		     key.objectid != BTRFS_FS_TREE_OBJECTID) ||
+		    key.objectid > BTRFS_LAST_FREE_OBJECTID)
+			goto skip;
+
+		eb = path->nodes[0];
+		slot = path->slots[0];
+		item_size = btrfs_item_size_nr(eb, slot);
+		if (item_size < sizeof(root_item))
+			goto skip;
+
+		read_extent_buffer(eb, &root_item,
+				   btrfs_item_ptr_offset(eb, slot),
+				   (int)sizeof(root_item));
+		if (btrfs_root_refs(&root_item) == 0)
+			goto skip;
+
+		if (!btrfs_is_empty_uuid(root_item.uuid) ||
+		    !btrfs_is_empty_uuid(root_item.received_uuid)) {
+			if (trans)
+				goto update_tree;
+
+			btrfs_release_path(path);
+			/*
+			 * 1 - subvol uuid item
+			 * 1 - received_subvol uuid item
+			 */
+			trans = btrfs_start_transaction(fs_info->uuid_root, 2);
+			if (IS_ERR(trans)) {
+				ret = PTR_ERR(trans);
+				break;
+			}
+			continue;
+		} else {
+			goto skip;
+		}
+update_tree:
+		if (!btrfs_is_empty_uuid(root_item.uuid)) {
+			ret = btrfs_uuid_tree_add(trans, fs_info->uuid_root,
+						  root_item.uuid,
+						  BTRFS_UUID_KEY_SUBVOL,
+						  key.objectid);
+			if (ret < 0) {
+				pr_warn("btrfs: uuid_tree_add failed %d\n",
+					ret);
+				break;
+			}
+		}
+
+		if (!btrfs_is_empty_uuid(root_item.received_uuid)) {
+			ret = btrfs_uuid_tree_add(trans, fs_info->uuid_root,
+						  root_item.received_uuid,
+						 BTRFS_UUID_KEY_RECEIVED_SUBVOL,
+						  key.objectid);
+			if (ret < 0) {
+				pr_warn("btrfs: uuid_tree_add failed %d\n",
+					ret);
+				break;
+			}
+		}
+
+skip:
+		if (trans) {
+			ret = btrfs_end_transaction(trans, fs_info->uuid_root);
+			trans = NULL;
+			if (ret)
+				break;
+		}
+
+		btrfs_release_path(path);
+		if (key.offset < (u64)-1) {
+			key.offset++;
+		} else if (key.type < BTRFS_ROOT_ITEM_KEY) {
+			key.offset = 0;
+			key.type = BTRFS_ROOT_ITEM_KEY;
+		} else if (key.objectid < (u64)-1) {
+			key.offset = 0;
+			key.type = BTRFS_ROOT_ITEM_KEY;
+			key.objectid++;
+		} else {
+			break;
+		}
+		cond_resched();
+	}
+
+out:
+	btrfs_free_path(path);
+	if (trans && !IS_ERR(trans))
+		btrfs_end_transaction(trans, fs_info->uuid_root);
+	if (ret)
+		pr_warn("btrfs: btrfs_uuid_scan_kthread failed %d\n", ret);
+	else
+		fs_info->update_uuid_tree_gen = 1;
+	up(&fs_info->uuid_tree_rescan_sem);
+	return 0;
+}
+
+/*
+ * Callback for btrfs_uuid_tree_iterate().
+ * returns:
+ * 0	check succeeded, the entry is not outdated.
+ * < 0	if an error occured.
+ * > 0	if the check failed, which means the caller shall remove the entry.
+ */
+static int btrfs_check_uuid_tree_entry(struct btrfs_fs_info *fs_info,
+				       u8 *uuid, u8 type, u64 subid)
+{
+	struct btrfs_key key;
+	int ret = 0;
+	struct btrfs_root *subvol_root;
+
+	if (type != BTRFS_UUID_KEY_SUBVOL &&
+	    type != BTRFS_UUID_KEY_RECEIVED_SUBVOL)
+		goto out;
+
+	key.objectid = subid;
+	key.type = BTRFS_ROOT_ITEM_KEY;
+	key.offset = (u64)-1;
+	subvol_root = btrfs_read_fs_root_no_name(fs_info, &key);
+	if (IS_ERR(subvol_root)) {
+		ret = PTR_ERR(subvol_root);
+		if (ret == -ENOENT)
+			ret = 1;
+		goto out;
+	}
+
+	switch (type) {
+	case BTRFS_UUID_KEY_SUBVOL:
+		if (memcmp(uuid, subvol_root->root_item.uuid, BTRFS_UUID_SIZE))
+			ret = 1;
+		break;
+	case BTRFS_UUID_KEY_RECEIVED_SUBVOL:
+		if (memcmp(uuid, subvol_root->root_item.received_uuid,
+			   BTRFS_UUID_SIZE))
+			ret = 1;
+		break;
+	}
+
+out:
+	return ret;
+}
+
+static int btrfs_uuid_rescan_kthread(void *data)
+{
+	struct btrfs_fs_info *fs_info = (struct btrfs_fs_info *)data;
+	int ret;
+
+	/*
+	 * 1st step is to iterate through the existing UUID tree and
+	 * to delete all entries that contain outdated data.
+	 * 2nd step is to add all missing entries to the UUID tree.
+	 */
+	ret = btrfs_uuid_tree_iterate(fs_info, btrfs_check_uuid_tree_entry);
+	if (ret < 0) {
+		pr_warn("btrfs: iterating uuid_tree failed %d\n", ret);
+		up(&fs_info->uuid_tree_rescan_sem);
+		return ret;
+	}
+	return btrfs_uuid_scan_kthread(data);
+}
+
+int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *tree_root = fs_info->tree_root;
+	struct btrfs_root *uuid_root;
+	struct task_struct *task;
+	int ret;
+
+	/*
+	 * 1 - root node
+	 * 1 - root item
+	 */
+	trans = btrfs_start_transaction(tree_root, 2);
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
+
+	uuid_root = btrfs_create_tree(trans, fs_info,
+				      BTRFS_UUID_TREE_OBJECTID);
+	if (IS_ERR(uuid_root)) {
+		btrfs_abort_transaction(trans, tree_root,
+					PTR_ERR(uuid_root));
+		return PTR_ERR(uuid_root);
+	}
+
+	fs_info->uuid_root = uuid_root;
+
+	ret = btrfs_commit_transaction(trans, tree_root);
+	if (ret)
+		return ret;
+
+	down(&fs_info->uuid_tree_rescan_sem);
+	task = kthread_run(btrfs_uuid_scan_kthread, fs_info, "btrfs-uuid");
+	if (IS_ERR(task)) {
+		/* fs_info->update_uuid_tree_gen remains 0 in all error case */
+		pr_warn("btrfs: failed to start uuid_scan task\n");
+		up(&fs_info->uuid_tree_rescan_sem);
+		return PTR_ERR(task);
+	}
+
+	return 0;
+}
+
+int btrfs_check_uuid_tree(struct btrfs_fs_info *fs_info)
+{
+	struct task_struct *task;
+
+	down(&fs_info->uuid_tree_rescan_sem);
+	task = kthread_run(btrfs_uuid_rescan_kthread, fs_info, "btrfs-uuid");
+	if (IS_ERR(task)) {
+		/* fs_info->update_uuid_tree_gen remains 0 in all error case */
+		pr_warn("btrfs: failed to start uuid_rescan task\n");
+		up(&fs_info->uuid_tree_rescan_sem);
+		return PTR_ERR(task);
+	}
+
+	return 0;
+}
+
 /*
  * shrinking a device means finding all of the device extents past
  * the new size, and then following the back refs to the chunks.
@@ -3681,10 +3970,8 @@ static void check_raid56_incompat_flag(struct btrfs_fs_info *info, u64 type)
 }
 
 static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
-			       struct btrfs_root *extent_root,
-			       struct map_lookup **map_ret,
-			       u64 *num_bytes_out, u64 *stripe_size_out,
-			       u64 start, u64 type)
+			       struct btrfs_root *extent_root, u64 start,
+			       u64 type)
 {
 	struct btrfs_fs_info *info = extent_root->fs_info;
 	struct btrfs_fs_devices *fs_devices = info->fs_devices;
@@ -3791,7 +4078,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 		if (total_avail == 0)
 			continue;
 
-		ret = find_free_dev_extent(device,
+		ret = find_free_dev_extent(trans, device,
 					   max_stripe_size * dev_stripes,
 					   &dev_offset, &max_avail);
 		if (ret && ret != -ENOSPC)
@@ -3903,12 +4190,8 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 	map->type = type;
 	map->sub_stripes = sub_stripes;
 
-	*map_ret = map;
 	num_bytes = stripe_size * data_stripes;
 
-	*stripe_size_out = stripe_size;
-	*num_bytes_out = num_bytes;
-
 	trace_btrfs_chunk_alloc(info->chunk_root, map, start, num_bytes);
 
 	em = alloc_extent_map();
@@ -3921,38 +4204,26 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 	em->len = num_bytes;
 	em->block_start = 0;
 	em->block_len = em->len;
+	em->orig_block_len = stripe_size;
 
 	em_tree = &extent_root->fs_info->mapping_tree.map_tree;
 	write_lock(&em_tree->lock);
 	ret = add_extent_mapping(em_tree, em, 0);
+	if (!ret) {
+		list_add_tail(&em->list, &trans->transaction->pending_chunks);
+		atomic_inc(&em->refs);
+	}
 	write_unlock(&em_tree->lock);
 	if (ret) {
 		free_extent_map(em);
 		goto error;
 	}
 
-	for (i = 0; i < map->num_stripes; ++i) {
-		struct btrfs_device *device;
-		u64 dev_offset;
-
-		device = map->stripes[i].dev;
-		dev_offset = map->stripes[i].physical;
-
-		ret = btrfs_alloc_dev_extent(trans, device,
-				info->chunk_root->root_key.objectid,
-				BTRFS_FIRST_CHUNK_TREE_OBJECTID,
-				start, dev_offset, stripe_size);
-		if (ret)
-			goto error_dev_extent;
-	}
-
 	ret = btrfs_make_block_group(trans, extent_root, 0, type,
 				     BTRFS_FIRST_CHUNK_TREE_OBJECTID,
 				     start, num_bytes);
-	if (ret) {
-		i = map->num_stripes - 1;
-		goto error_dev_extent;
-	}
+	if (ret)
+		goto error_del_extent;
 
 	free_extent_map(em);
 	check_raid56_incompat_flag(extent_root->fs_info, type);
@@ -3960,18 +4231,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 	kfree(devices_info);
 	return 0;
 
-error_dev_extent:
-	for (; i >= 0; i--) {
-		struct btrfs_device *device;
-		int err;
-
-		device = map->stripes[i].dev;
-		err = btrfs_free_dev_extent(trans, device, start);
-		if (err) {
-			btrfs_abort_transaction(trans, extent_root, err);
-			break;
-		}
-	}
+error_del_extent:
 	write_lock(&em_tree->lock);
 	remove_extent_mapping(em_tree, em);
 	write_unlock(&em_tree->lock);
@@ -3986,33 +4246,68 @@ error:
 	return ret;
 }
 
-static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
+int btrfs_finish_chunk_alloc(struct btrfs_trans_handle *trans,
 				struct btrfs_root *extent_root,
-				struct map_lookup *map, u64 chunk_offset,
-				u64 chunk_size, u64 stripe_size)
+				u64 chunk_offset, u64 chunk_size)
 {
-	u64 dev_offset;
 	struct btrfs_key key;
 	struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
 	struct btrfs_device *device;
 	struct btrfs_chunk *chunk;
 	struct btrfs_stripe *stripe;
-	size_t item_size = btrfs_chunk_item_size(map->num_stripes);
-	int index = 0;
+	struct extent_map_tree *em_tree;
+	struct extent_map *em;
+	struct map_lookup *map;
+	size_t item_size;
+	u64 dev_offset;
+	u64 stripe_size;
+	int i = 0;
 	int ret;
 
+	em_tree = &extent_root->fs_info->mapping_tree.map_tree;
+	read_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, chunk_offset, chunk_size);
+	read_unlock(&em_tree->lock);
+
+	if (!em) {
+		btrfs_crit(extent_root->fs_info, "unable to find logical "
+			   "%Lu len %Lu", chunk_offset, chunk_size);
+		return -EINVAL;
+	}
+
+	if (em->start != chunk_offset || em->len != chunk_size) {
+		btrfs_crit(extent_root->fs_info, "found a bad mapping, wanted"
+			  " %Lu-%Lu, found %Lu-%Lu\n", chunk_offset,
+			  chunk_size, em->start, em->len);
+		free_extent_map(em);
+		return -EINVAL;
+	}
+
+	map = (struct map_lookup *)em->bdev;
+	item_size = btrfs_chunk_item_size(map->num_stripes);
+	stripe_size = em->orig_block_len;
+
 	chunk = kzalloc(item_size, GFP_NOFS);
-	if (!chunk)
-		return -ENOMEM;
+	if (!chunk) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	for (i = 0; i < map->num_stripes; i++) {
+		device = map->stripes[i].dev;
+		dev_offset = map->stripes[i].physical;
 
-	index = 0;
-	while (index < map->num_stripes) {
-		device = map->stripes[index].dev;
 		device->bytes_used += stripe_size;
 		ret = btrfs_update_device(trans, device);
 		if (ret)
-			goto out_free;
-		index++;
+			goto out;
+		ret = btrfs_alloc_dev_extent(trans, device,
+					     chunk_root->root_key.objectid,
+					     BTRFS_FIRST_CHUNK_TREE_OBJECTID,
+					     chunk_offset, dev_offset,
+					     stripe_size);
+		if (ret)
+			goto out;
 	}
 
 	spin_lock(&extent_root->fs_info->free_chunk_lock);
@@ -4020,17 +4315,15 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
 						   map->num_stripes);
 	spin_unlock(&extent_root->fs_info->free_chunk_lock);
 
-	index = 0;
 	stripe = &chunk->stripe;
-	while (index < map->num_stripes) {
-		device = map->stripes[index].dev;
-		dev_offset = map->stripes[index].physical;
+	for (i = 0; i < map->num_stripes; i++) {
+		device = map->stripes[i].dev;
+		dev_offset = map->stripes[i].physical;
 
 		btrfs_set_stack_stripe_devid(stripe, device->devid);
 		btrfs_set_stack_stripe_offset(stripe, dev_offset);
 		memcpy(stripe->dev_uuid, device->uuid, BTRFS_UUID_SIZE);
 		stripe++;
-		index++;
 	}
 
 	btrfs_set_stack_chunk_length(chunk, chunk_size);
@@ -4048,7 +4341,6 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
 	key.offset = chunk_offset;
 
 	ret = btrfs_insert_item(trans, chunk_root, &key, chunk, item_size);
-
 	if (ret == 0 && map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
 		/*
 		 * TODO: Cleanup of inserted chunk root in case of
@@ -4058,8 +4350,9 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
 					     item_size);
 	}
 
-out_free:
+out:
 	kfree(chunk);
+	free_extent_map(em);
 	return ret;
 }
 
@@ -4074,27 +4367,9 @@ int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 		      struct btrfs_root *extent_root, u64 type)
 {
 	u64 chunk_offset;
-	u64 chunk_size;
-	u64 stripe_size;
-	struct map_lookup *map;
-	struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
-	int ret;
 
-	ret = find_next_chunk(chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID,
-			      &chunk_offset);
-	if (ret)
-		return ret;
-
-	ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
-				  &stripe_size, chunk_offset, type);
-	if (ret)
-		return ret;
-
-	ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
-				   chunk_size, stripe_size);
-	if (ret)
-		return ret;
-	return 0;
+	chunk_offset = find_next_chunk(extent_root->fs_info);
+	return __btrfs_alloc_chunk(trans, extent_root, chunk_offset, type);
 }
 
 static noinline int init_first_rw_device(struct btrfs_trans_handle *trans,
@@ -4103,66 +4378,31 @@ static noinline int init_first_rw_device(struct btrfs_trans_handle *trans,
 {
 	u64 chunk_offset;
 	u64 sys_chunk_offset;
-	u64 chunk_size;
-	u64 sys_chunk_size;
-	u64 stripe_size;
-	u64 sys_stripe_size;
 	u64 alloc_profile;
-	struct map_lookup *map;
-	struct map_lookup *sys_map;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_root *extent_root = fs_info->extent_root;
 	int ret;
 
-	ret = find_next_chunk(fs_info->chunk_root,
-			      BTRFS_FIRST_CHUNK_TREE_OBJECTID, &chunk_offset);
-	if (ret)
-		return ret;
-
+	chunk_offset = find_next_chunk(fs_info);
 	alloc_profile = btrfs_get_alloc_profile(extent_root, 0);
-	ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
-				  &stripe_size, chunk_offset, alloc_profile);
+	ret = __btrfs_alloc_chunk(trans, extent_root, chunk_offset,
+				  alloc_profile);
 	if (ret)
 		return ret;
 
-	sys_chunk_offset = chunk_offset + chunk_size;
-
+	sys_chunk_offset = find_next_chunk(root->fs_info);
 	alloc_profile = btrfs_get_alloc_profile(fs_info->chunk_root, 0);
-	ret = __btrfs_alloc_chunk(trans, extent_root, &sys_map,
-				  &sys_chunk_size, &sys_stripe_size,
-				  sys_chunk_offset, alloc_profile);
+	ret = __btrfs_alloc_chunk(trans, extent_root, sys_chunk_offset,
+				  alloc_profile);
 	if (ret) {
 		btrfs_abort_transaction(trans, root, ret);
 		goto out;
 	}
 
 	ret = btrfs_add_device(trans, fs_info->chunk_root, device);
-	if (ret) {
-		btrfs_abort_transaction(trans, root, ret);
-		goto out;
-	}
-
-	/*
-	 * Modifying chunk tree needs allocating new blocks from both
-	 * system block group and metadata block group. So we only can
-	 * do operations require modifying the chunk tree after both
-	 * block groups were created.
-	 */
-	ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
-				   chunk_size, stripe_size);
-	if (ret) {
-		btrfs_abort_transaction(trans, root, ret);
-		goto out;
-	}
-
-	ret = __finish_chunk_alloc(trans, extent_root, sys_map,
-				   sys_chunk_offset, sys_chunk_size,
-				   sys_stripe_size);
 	if (ret)
 		btrfs_abort_transaction(trans, root, ret);
-
 out:
-
 	return ret;
 }
 
@@ -4239,13 +4479,13 @@ int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len)
 	 * and exit, so return 1 so the callers don't try to use other copies.
 	 */
 	if (!em) {
-		btrfs_emerg(fs_info, "No mapping for %Lu-%Lu\n", logical,
+		btrfs_crit(fs_info, "No mapping for %Lu-%Lu\n", logical,
 			    logical+len);
 		return 1;
 	}
 
 	if (em->start > logical || em->start + em->len < logical) {
-		btrfs_emerg(fs_info, "Invalid mapping for %Lu-%Lu, got "
+		btrfs_crit(fs_info, "Invalid mapping for %Lu-%Lu, got "
 			    "%Lu-%Lu\n", logical, logical+len, em->start,
 			    em->start + em->len);
 		return 1;
@@ -4420,8 +4660,7 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, int rw,
 
 	if (!em) {
 		btrfs_crit(fs_info, "unable to find logical %llu len %llu",
-			(unsigned long long)logical,
-			(unsigned long long)*length);
+			logical, *length);
 		return -EINVAL;
 	}
 
@@ -4435,9 +4674,6 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, int rw,
 	map = (struct map_lookup *)em->bdev;
 	offset = logical - em->start;
 
-	if (mirror_num > map->num_stripes)
-		mirror_num = 0;
-
 	stripe_len = map->stripe_len;
 	stripe_nr = offset;
 	/*
@@ -4719,6 +4955,7 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, int rw,
 	}
 	bbio = kzalloc(btrfs_bio_size(num_alloc_stripes), GFP_NOFS);
 	if (!bbio) {
+		kfree(raid_map);
 		ret = -ENOMEM;
 		goto out;
 	}
@@ -5294,9 +5531,7 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
 
 	if (map_length < length) {
 		btrfs_crit(root->fs_info, "mapping failed logical %llu bio len %llu len %llu",
-			(unsigned long long)logical,
-			(unsigned long long)length,
-			(unsigned long long)map_length);
+			logical, length, map_length);
 		BUG();
 	}
 
@@ -5362,24 +5597,72 @@ static struct btrfs_device *add_missing_dev(struct btrfs_root *root,
 	struct btrfs_device *device;
 	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
 
-	device = kzalloc(sizeof(*device), GFP_NOFS);
-	if (!device)
+	device = btrfs_alloc_device(NULL, &devid, dev_uuid);
+	if (IS_ERR(device))
 		return NULL;
-	list_add(&device->dev_list,
-		 &fs_devices->devices);
-	device->dev_root = root->fs_info->dev_root;
-	device->devid = devid;
-	device->work.func = pending_bios_fn;
+
+	list_add(&device->dev_list, &fs_devices->devices);
 	device->fs_devices = fs_devices;
-	device->missing = 1;
 	fs_devices->num_devices++;
+
+	device->missing = 1;
 	fs_devices->missing_devices++;
-	spin_lock_init(&device->io_lock);
-	INIT_LIST_HEAD(&device->dev_alloc_list);
-	memcpy(device->uuid, dev_uuid, BTRFS_UUID_SIZE);
+
 	return device;
 }
 
+/**
+ * btrfs_alloc_device - allocate struct btrfs_device
+ * @fs_info:	used only for generating a new devid, can be NULL if
+ *		devid is provided (i.e. @devid != NULL).
+ * @devid:	a pointer to devid for this device.  If NULL a new devid
+ *		is generated.
+ * @uuid:	a pointer to UUID for this device.  If NULL a new UUID
+ *		is generated.
+ *
+ * Return: a pointer to a new &struct btrfs_device on success; ERR_PTR()
+ * on error.  Returned struct is not linked onto any lists and can be
+ * destroyed with kfree() right away.
+ */
+struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info,
+					const u64 *devid,
+					const u8 *uuid)
+{
+	struct btrfs_device *dev;
+	u64 tmp;
+
+	if (!devid && !fs_info) {
+		WARN_ON(1);
+		return ERR_PTR(-EINVAL);
+	}
+
+	dev = __alloc_device();
+	if (IS_ERR(dev))
+		return dev;
+
+	if (devid)
+		tmp = *devid;
+	else {
+		int ret;
+
+		ret = find_next_devid(fs_info, &tmp);
+		if (ret) {
+			kfree(dev);
+			return ERR_PTR(ret);
+		}
+	}
+	dev->devid = tmp;
+
+	if (uuid)
+		memcpy(dev->uuid, uuid, BTRFS_UUID_SIZE);
+	else
+		generate_random_uuid(dev->uuid);
+
+	dev->work.func = pending_bios_fn;
+
+	return dev;
+}
+
 static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
 			  struct extent_buffer *leaf,
 			  struct btrfs_chunk *chunk)
@@ -5486,7 +5769,7 @@ static void fill_device_from_item(struct extent_buffer *leaf,
 	WARN_ON(device->devid == BTRFS_DEV_REPLACE_DEVID);
 	device->is_tgtdev_for_dev_replace = 0;
 
-	ptr = (unsigned long)btrfs_device_uuid(dev_item);
+	ptr = btrfs_device_uuid(dev_item);
 	read_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
 }
 
@@ -5549,11 +5832,9 @@ static int read_one_dev(struct btrfs_root *root,
 	u8 dev_uuid[BTRFS_UUID_SIZE];
 
 	devid = btrfs_device_id(leaf, dev_item);
-	read_extent_buffer(leaf, dev_uuid,
-			   (unsigned long)btrfs_device_uuid(dev_item),
+	read_extent_buffer(leaf, dev_uuid, btrfs_device_uuid(dev_item),
 			   BTRFS_UUID_SIZE);
-	read_extent_buffer(leaf, fs_uuid,
-			   (unsigned long)btrfs_device_fsid(dev_item),
+	read_extent_buffer(leaf, fs_uuid, btrfs_device_fsid(dev_item),
 			   BTRFS_UUID_SIZE);
 
 	if (memcmp(fs_uuid, root->fs_info->fsid, BTRFS_UUID_SIZE)) {
@@ -5568,8 +5849,7 @@ static int read_one_dev(struct btrfs_root *root,
 			return -EIO;
 
 		if (!device) {
-			btrfs_warn(root->fs_info, "devid %llu missing",
-				(unsigned long long)devid);
+			btrfs_warn(root->fs_info, "devid %llu missing", devid);
 			device = add_missing_dev(root, devid, dev_uuid);
 			if (!device)
 				return -ENOMEM;
@@ -5593,7 +5873,6 @@ static int read_one_dev(struct btrfs_root *root,
 	}
 
 	fill_device_from_item(leaf, dev_item, device);
-	device->dev_root = root->fs_info->dev_root;
 	device->in_fs_metadata = 1;
 	if (device->writeable && !device->is_tgtdev_for_dev_replace) {
 		device->fs_devices->total_rw_bytes += device->total_bytes;
@@ -5694,14 +5973,15 @@ int btrfs_read_chunk_tree(struct btrfs_root *root)
 	mutex_lock(&uuid_mutex);
 	lock_chunks(root);
 
-	/* first we search for all of the device items, and then we
-	 * read in all of the chunk items.  This way we can create chunk
-	 * mappings that reference all of the devices that are afound
+	/*
+	 * Read all device items, and then all the chunk items. All
+	 * device items are found before any chunk item (their object id
+	 * is smaller than the lowest possible object id for a chunk
+	 * item - BTRFS_FIRST_CHUNK_TREE_OBJECTID).
 	 */
 	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
 	key.offset = 0;
 	key.type = 0;
-again:
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
 		goto error;
@@ -5717,17 +5997,13 @@ again:
 			break;
 		}
 		btrfs_item_key_to_cpu(leaf, &found_key, slot);
-		if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
-			if (found_key.objectid != BTRFS_DEV_ITEMS_OBJECTID)
-				break;
-			if (found_key.type == BTRFS_DEV_ITEM_KEY) {
-				struct btrfs_dev_item *dev_item;
-				dev_item = btrfs_item_ptr(leaf, slot,
+		if (found_key.type == BTRFS_DEV_ITEM_KEY) {
+			struct btrfs_dev_item *dev_item;
+			dev_item = btrfs_item_ptr(leaf, slot,
 						  struct btrfs_dev_item);
-				ret = read_one_dev(root, leaf, dev_item);
-				if (ret)
-					goto error;
-			}
+			ret = read_one_dev(root, leaf, dev_item);
+			if (ret)
+				goto error;
 		} else if (found_key.type == BTRFS_CHUNK_ITEM_KEY) {
 			struct btrfs_chunk *chunk;
 			chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
@@ -5737,11 +6013,6 @@ again:
 		}
 		path->slots[0]++;
 	}
-	if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
-		key.objectid = 0;
-		btrfs_release_path(path);
-		goto again;
-	}
 	ret = 0;
 error:
 	unlock_chunks(root);
@@ -5751,6 +6022,17 @@ error:
 	return ret;
 }
 
+void btrfs_init_devices_late(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+	struct btrfs_device *device;
+
+	mutex_lock(&fs_devices->device_list_mutex);
+	list_for_each_entry(device, &fs_devices->devices, dev_list)
+		device->dev_root = fs_info->dev_root;
+	mutex_unlock(&fs_devices->device_list_mutex);
+}
+
 static void __btrfs_reset_dev_stats(struct btrfs_device *dev)
 {
 	int i;
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index f6247e2a47f7..b72f540c8b29 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -152,6 +152,8 @@ struct btrfs_fs_devices {
 	int rotating;
 };
 
+#define BTRFS_BIO_INLINE_CSUM_SIZE	64
+
 /*
  * we need the mirror number and stripe index to be passed around
  * the call chain while we are processing end_io (especially errors).
@@ -161,9 +163,14 @@ struct btrfs_fs_devices {
  * we allocate are actually btrfs_io_bios.  We'll cram as much of
  * struct btrfs_bio as we can into this over time.
  */
+typedef void (btrfs_io_bio_end_io_t) (struct btrfs_io_bio *bio, int err);
 struct btrfs_io_bio {
 	unsigned long mirror_num;
 	unsigned long stripe_index;
+	u8 *csum;
+	u8 csum_inline[BTRFS_BIO_INLINE_CSUM_SIZE];
+	u8 *csum_allocated;
+	btrfs_io_bio_end_io_t *end_io;
 	struct bio bio;
 };
 
@@ -298,6 +305,9 @@ void btrfs_close_extra_devices(struct btrfs_fs_info *fs_info,
 int btrfs_find_device_missing_or_by_path(struct btrfs_root *root,
 					 char *device_path,
 					 struct btrfs_device **device);
+struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info,
+					const u64 *devid,
+					const u8 *uuid);
 int btrfs_rm_device(struct btrfs_root *root, char *device_path);
 void btrfs_cleanup_fs_uuids(void);
 int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len);
@@ -315,12 +325,16 @@ int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info);
 int btrfs_recover_balance(struct btrfs_fs_info *fs_info);
 int btrfs_pause_balance(struct btrfs_fs_info *fs_info);
 int btrfs_cancel_balance(struct btrfs_fs_info *fs_info);
+int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info);
+int btrfs_check_uuid_tree(struct btrfs_fs_info *fs_info);
 int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset);
-int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
+int find_free_dev_extent(struct btrfs_trans_handle *trans,
+			 struct btrfs_device *device, u64 num_bytes,
 			 u64 *start, u64 *max_avail);
 void btrfs_dev_stat_inc_and_print(struct btrfs_device *dev, int index);
 int btrfs_get_dev_stats(struct btrfs_root *root,
 			struct btrfs_ioctl_get_dev_stats *stats);
+void btrfs_init_devices_late(struct btrfs_fs_info *fs_info);
 int btrfs_init_dev_stats(struct btrfs_fs_info *fs_info);
 int btrfs_run_dev_stats(struct btrfs_trans_handle *trans,
 			struct btrfs_fs_info *fs_info);
@@ -336,6 +350,9 @@ int btrfs_is_parity_mirror(struct btrfs_mapping_tree *map_tree,
 unsigned long btrfs_full_stripe_len(struct btrfs_root *root,
 				    struct btrfs_mapping_tree *map_tree,
 				    u64 logical);
+int btrfs_finish_chunk_alloc(struct btrfs_trans_handle *trans,
+				struct btrfs_root *extent_root,
+				u64 chunk_offset, u64 chunk_size);
 static inline void btrfs_dev_stat_inc(struct btrfs_device *dev,
 				      int index)
 {