Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

Pull large btrfs update from Chris Mason:
 "This pull request is very large, and the two main features in here
  have been under testing/devel for quite a while.

  We have subvolume quotas from the strato developers.  This enables
  full tracking of how many blocks are allocated to each subvolume (and
  all snapshots) and you can set limits on a per-subvolume basis.  You
  can also create quota groups and toss multiple subvolumes into a big
  group.  It's everything you need to be a web hosting company and give
  each user their own subvolume.

  The userland side of the quotas is being refreshed, they'll send out
  details on where to grab it soon.

  Next is the kernel side of btrfs send/receive from Alexander Block.
  This leverages the same infrastructure as the quota code to figure out
  relationships between blocks and their owners.  It can then compute
  the difference between two snapshots and sends the diffs in a neutral
  format into userland.

  The basic model:

        create a snapshot
        send that snapshot as the initial backup
        make changes
        create a second snapshot
        send the incremental as a backup
        delete the first snapshot
        (use the second snapshot for the next incremental)

  The receive portion is all in userland, and in the 'next' branch of my
  btrfs-progs repo.

  There's still some work to do in terms of optimizing the send side
  from kernel to userland.  The really important part is figuring out
  how two snapshots are different, and this is where we are
  concentrating right now.  The initial send of a dataset is a little
  slower than tar, but the incremental sends are dramatically faster
  than what rsync can do.

  On top of all of that, we have a nice queue of fixes, cleanups and
  optimizations."

Fix up trivial modify/del conflict in fs/btrfs/ioctl.c

Also fix up semantic conflict in fs/btrfs/send.c: the interface to
dentry_open() changed in commit 765927b2d508 ("switch dentry_open() to
struct path, make it grab references itself"), and since it now grabs
whatever references it needs, we should no longer do the mntget() on the
mnt (and we need to dput() the dentry reference we took).

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (65 commits)
  Btrfs: uninit variable fixes in send/receive
  Btrfs: introduce BTRFS_IOC_SEND for btrfs send/receive
  Btrfs: add btrfs_compare_trees function
  Btrfs: introduce subvol uuids and times
  Btrfs: make iref_to_path non static
  Btrfs: add a barrier before a waitqueue_active check
  Btrfs: call the ordered free operation without any locks held
  Btrfs: Check INCOMPAT flags on remount and add helper function
  Btrfs: add helper for tree enumeration
  btrfs: allow cross-subvolume file clone
  Btrfs: improve multi-thread buffer read
  Btrfs: make btrfs's allocation smoothly with preallocation
  Btrfs: lock the transition from dirty to writeback for an eb
  Btrfs: fix potential race in extent buffer freeing
  Btrfs: don't return true in releasepage unless we actually freed the eb
  Btrfs: suppress printk() if all device I/O stats are zero
  Btrfs: remove unwanted printk() for btrfs device I/O stats
  Btrfs: rewrite BTRFS_SETGET_FUNCS
  Btrfs: zero unused bytes in inode item
  Btrfs: kill free_space pointer from inode structure
  ...

Conflicts:
	fs/btrfs/ioctl.c

34 files changed, 8687 insertions, 631 deletions

diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 0c4fa2befae7..d7fcdba141a2 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -8,7 +8,7 @@ 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
+	   reada.o backref.o ulist.o qgroup.o send.o
 
 btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o
 btrfs-$(CONFIG_BTRFS_FS_CHECK_INTEGRITY) += check-integrity.o
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 42704149b723..58b7d14b08ee 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -206,10 +206,17 @@ static noinline void run_ordered_completions(struct btrfs_workers *workers,
 
 		work->ordered_func(work);
 
-		/* now take the lock again and call the freeing code */
+		/* now take the lock again and drop our item from the list */
 		spin_lock(&workers->order_lock);
 		list_del(&work->order_list);
+		spin_unlock(&workers->order_lock);
+
+		/*
+		 * we don't want to call the ordered free functions
+		 * with the lock held though
+		 */
 		work->ordered_free(work);
+		spin_lock(&workers->order_lock);
 	}
 
 	spin_unlock(&workers->order_lock);
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index a383c18e74e8..a256f3b2a845 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -773,9 +773,8 @@ static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
  */
 static int find_parent_nodes(struct btrfs_trans_handle *trans,
 			     struct btrfs_fs_info *fs_info, u64 bytenr,
-			     u64 delayed_ref_seq, u64 time_seq,
-			     struct ulist *refs, struct ulist *roots,
-			     const u64 *extent_item_pos)
+			     u64 time_seq, struct ulist *refs,
+			     struct ulist *roots, const u64 *extent_item_pos)
 {
 	struct btrfs_key key;
 	struct btrfs_path *path;
@@ -837,7 +836,7 @@ again:
 				btrfs_put_delayed_ref(&head->node);
 				goto again;
 			}
-			ret = __add_delayed_refs(head, delayed_ref_seq,
+			ret = __add_delayed_refs(head, time_seq,
 						 &prefs_delayed);
 			mutex_unlock(&head->mutex);
 			if (ret) {
@@ -981,8 +980,7 @@ static void free_leaf_list(struct ulist *blocks)
  */
 static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
 				struct btrfs_fs_info *fs_info, u64 bytenr,
-				u64 delayed_ref_seq, u64 time_seq,
-				struct ulist **leafs,
+				u64 time_seq, struct ulist **leafs,
 				const u64 *extent_item_pos)
 {
 	struct ulist *tmp;
@@ -997,7 +995,7 @@ static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 	}
 
-	ret = find_parent_nodes(trans, fs_info, bytenr, delayed_ref_seq,
+	ret = find_parent_nodes(trans, fs_info, bytenr,
 				time_seq, *leafs, tmp, extent_item_pos);
 	ulist_free(tmp);
 
@@ -1024,8 +1022,7 @@ static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
  */
 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
 				struct btrfs_fs_info *fs_info, u64 bytenr,
-				u64 delayed_ref_seq, u64 time_seq,
-				struct ulist **roots)
+				u64 time_seq, struct ulist **roots)
 {
 	struct ulist *tmp;
 	struct ulist_node *node = NULL;
@@ -1043,7 +1040,7 @@ int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
 
 	ULIST_ITER_INIT(&uiter);
 	while (1) {
-		ret = find_parent_nodes(trans, fs_info, bytenr, delayed_ref_seq,
+		ret = find_parent_nodes(trans, fs_info, bytenr,
 					time_seq, tmp, *roots, NULL);
 		if (ret < 0 && ret != -ENOENT) {
 			ulist_free(tmp);
@@ -1125,10 +1122,10 @@ static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
  * required for the path to fit into the buffer. in that case, the returned
  * value will be smaller than dest. callers must check this!
  */
-static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
-				struct btrfs_inode_ref *iref,
-				struct extent_buffer *eb_in, u64 parent,
-				char *dest, u32 size)
+char *btrfs_iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
+			 struct btrfs_inode_ref *iref,
+			 struct extent_buffer *eb_in, u64 parent,
+			 char *dest, u32 size)
 {
 	u32 len;
 	int slot;
@@ -1376,11 +1373,9 @@ int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
 	struct ulist *roots = NULL;
 	struct ulist_node *ref_node = NULL;
 	struct ulist_node *root_node = NULL;
-	struct seq_list seq_elem = {};
 	struct seq_list tree_mod_seq_elem = {};
 	struct ulist_iterator ref_uiter;
 	struct ulist_iterator root_uiter;
-	struct btrfs_delayed_ref_root *delayed_refs = NULL;
 
 	pr_debug("resolving all inodes for extent %llu\n",
 			extent_item_objectid);
@@ -1391,16 +1386,11 @@ int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
 		trans = btrfs_join_transaction(fs_info->extent_root);
 		if (IS_ERR(trans))
 			return PTR_ERR(trans);
-
-		delayed_refs = &trans->transaction->delayed_refs;
-		spin_lock(&delayed_refs->lock);
-		btrfs_get_delayed_seq(delayed_refs, &seq_elem);
-		spin_unlock(&delayed_refs->lock);
 		btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
 	}
 
 	ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid,
-				   seq_elem.seq, tree_mod_seq_elem.seq, &refs,
+				   tree_mod_seq_elem.seq, &refs,
 				   &extent_item_pos);
 	if (ret)
 		goto out;
@@ -1408,8 +1398,7 @@ int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
 	ULIST_ITER_INIT(&ref_uiter);
 	while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) {
 		ret = btrfs_find_all_roots(trans, fs_info, ref_node->val,
-						seq_elem.seq,
-						tree_mod_seq_elem.seq, &roots);
+					   tree_mod_seq_elem.seq, &roots);
 		if (ret)
 			break;
 		ULIST_ITER_INIT(&root_uiter);
@@ -1431,7 +1420,6 @@ int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
 out:
 	if (!search_commit_root) {
 		btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
-		btrfs_put_delayed_seq(delayed_refs, &seq_elem);
 		btrfs_end_transaction(trans, fs_info->extent_root);
 	}
 
@@ -1543,7 +1531,7 @@ static int inode_to_path(u64 inum, struct btrfs_inode_ref *iref,
 					ipath->fspath->bytes_left - s_ptr : 0;
 
 	fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
-	fspath = iref_to_path(ipath->fs_root, ipath->btrfs_path, iref, eb,
+	fspath = btrfs_iref_to_path(ipath->fs_root, ipath->btrfs_path, iref, eb,
 				inum, fspath_min, bytes_left);
 	if (IS_ERR(fspath))
 		return PTR_ERR(fspath);
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index c18d8ac7b795..032f4dc7eab8 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -21,6 +21,7 @@
 
 #include "ioctl.h"
 #include "ulist.h"
+#include "extent_io.h"
 
 #define BTRFS_BACKREF_SEARCH_COMMIT_ROOT ((struct btrfs_trans_handle *)0)
 
@@ -58,8 +59,10 @@ int paths_from_inode(u64 inum, struct inode_fs_paths *ipath);
 
 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
 				struct btrfs_fs_info *fs_info, u64 bytenr,
-				u64 delayed_ref_seq, u64 time_seq,
-				struct ulist **roots);
+				u64 time_seq, struct ulist **roots);
+char *btrfs_iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
+			 struct btrfs_inode_ref *iref, struct extent_buffer *eb,
+			 u64 parent, char *dest, u32 size);
 
 struct btrfs_data_container *init_data_container(u32 total_bytes);
 struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 12394a90d60f..5b2ad6bc4fe7 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -87,9 +87,6 @@ struct btrfs_inode {
 	/* node for the red-black tree that links inodes in subvolume root */
 	struct rb_node rb_node;
 
-	/* the space_info for where this inode's data allocations are done */
-	struct btrfs_space_info *space_info;
-
 	unsigned long runtime_flags;
 
 	/* full 64 bit generation number, struct vfs_inode doesn't have a big
@@ -191,11 +188,14 @@ static inline void btrfs_i_size_write(struct inode *inode, u64 size)
 	BTRFS_I(inode)->disk_i_size = size;
 }
 
-static inline bool btrfs_is_free_space_inode(struct btrfs_root *root,
-				       struct inode *inode)
+static inline bool btrfs_is_free_space_inode(struct inode *inode)
 {
-	if (root == root->fs_info->tree_root ||
-	    BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID)
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+
+	if (root == root->fs_info->tree_root &&
+	    btrfs_ino(inode) != BTRFS_BTREE_INODE_OBJECTID)
+		return true;
+	if (BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID)
 		return true;
 	return false;
 }
diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c
index da6e9364a5e3..9197e2e33407 100644
--- a/fs/btrfs/check-integrity.c
+++ b/fs/btrfs/check-integrity.c
@@ -1032,6 +1032,7 @@ continue_with_current_leaf_stack_frame:
 			struct btrfs_disk_key *disk_key;
 			u8 type;
 			u32 item_offset;
+			u32 item_size;
 
 			if (disk_item_offset + sizeof(struct btrfs_item) >
 			    sf->block_ctx->len) {
@@ -1047,6 +1048,7 @@ leaf_item_out_of_bounce_error:
 						     disk_item_offset,
 						     sizeof(struct btrfs_item));
 			item_offset = le32_to_cpu(disk_item.offset);
+			item_size = le32_to_cpu(disk_item.size);
 			disk_key = &disk_item.key;
 			type = disk_key->type;
 
@@ -1057,14 +1059,13 @@ leaf_item_out_of_bounce_error:
 
 				root_item_offset = item_offset +
 					offsetof(struct btrfs_leaf, items);
-				if (root_item_offset +
-				    sizeof(struct btrfs_root_item) >
+				if (root_item_offset + item_size >
 				    sf->block_ctx->len)
 					goto leaf_item_out_of_bounce_error;
 				btrfsic_read_from_block_data(
 					sf->block_ctx, &root_item,
 					root_item_offset,
-					sizeof(struct btrfs_root_item));
+					item_size);
 				next_bytenr = le64_to_cpu(root_item.bytenr);
 
 				sf->error =
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 8206b3900587..9d7621f271ff 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -321,7 +321,7 @@ struct tree_mod_root {
 struct tree_mod_elem {
 	struct rb_node node;
 	u64 index;		/* shifted logical */
-	struct seq_list elem;
+	u64 seq;
 	enum mod_log_op op;
 
 	/* this is used for MOD_LOG_KEY_* and MOD_LOG_MOVE_KEYS operations */
@@ -341,20 +341,50 @@ struct tree_mod_elem {
 	struct tree_mod_root old_root;
 };
 
-static inline void
-__get_tree_mod_seq(struct btrfs_fs_info *fs_info, struct seq_list *elem)
+static inline void tree_mod_log_read_lock(struct btrfs_fs_info *fs_info)
 {
-	elem->seq = atomic_inc_return(&fs_info->tree_mod_seq);
-	list_add_tail(&elem->list, &fs_info->tree_mod_seq_list);
+	read_lock(&fs_info->tree_mod_log_lock);
 }
 
-void btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info,
-			    struct seq_list *elem)
+static inline void tree_mod_log_read_unlock(struct btrfs_fs_info *fs_info)
+{
+	read_unlock(&fs_info->tree_mod_log_lock);
+}
+
+static inline void tree_mod_log_write_lock(struct btrfs_fs_info *fs_info)
+{
+	write_lock(&fs_info->tree_mod_log_lock);
+}
+
+static inline void tree_mod_log_write_unlock(struct btrfs_fs_info *fs_info)
+{
+	write_unlock(&fs_info->tree_mod_log_lock);
+}
+
+/*
+ * This adds a new blocker to the tree mod log's blocker list if the @elem
+ * passed does not already have a sequence number set. So when a caller expects
+ * to record tree modifications, it should ensure to set elem->seq to zero
+ * before calling btrfs_get_tree_mod_seq.
+ * Returns a fresh, unused tree log modification sequence number, even if no new
+ * blocker was added.
+ */
+u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info,
+			   struct seq_list *elem)
 {
-	elem->flags = 1;
+	u64 seq;
+
+	tree_mod_log_write_lock(fs_info);
 	spin_lock(&fs_info->tree_mod_seq_lock);
-	__get_tree_mod_seq(fs_info, elem);
+	if (!elem->seq) {
+		elem->seq = btrfs_inc_tree_mod_seq(fs_info);
+		list_add_tail(&elem->list, &fs_info->tree_mod_seq_list);
+	}
+	seq = btrfs_inc_tree_mod_seq(fs_info);
 	spin_unlock(&fs_info->tree_mod_seq_lock);
+	tree_mod_log_write_unlock(fs_info);
+
+	return seq;
 }
 
 void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info,
@@ -371,41 +401,46 @@ void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info,
 	if (!seq_putting)
 		return;
 
-	BUG_ON(!(elem->flags & 1));
 	spin_lock(&fs_info->tree_mod_seq_lock);
 	list_del(&elem->list);
+	elem->seq = 0;
 
 	list_for_each_entry(cur_elem, &fs_info->tree_mod_seq_list, list) {
-		if ((cur_elem->flags & 1) && cur_elem->seq < min_seq) {
+		if (cur_elem->seq < min_seq) {
 			if (seq_putting > cur_elem->seq) {
 				/*
 				 * blocker with lower sequence number exists, we
 				 * cannot remove anything from the log
 				 */
-				goto out;
+				spin_unlock(&fs_info->tree_mod_seq_lock);
+				return;
 			}
 			min_seq = cur_elem->seq;
 		}
 	}
+	spin_unlock(&fs_info->tree_mod_seq_lock);
+
+	/*
+	 * we removed the lowest blocker from the blocker list, so there may be
+	 * more processible delayed refs.
+	 */
+	wake_up(&fs_info->tree_mod_seq_wait);
 
 	/*
 	 * anything that's lower than the lowest existing (read: blocked)
 	 * sequence number can be removed from the tree.
 	 */
-	write_lock(&fs_info->tree_mod_log_lock);
+	tree_mod_log_write_lock(fs_info);
 	tm_root = &fs_info->tree_mod_log;
 	for (node = rb_first(tm_root); node; node = next) {
 		next = rb_next(node);
 		tm = container_of(node, struct tree_mod_elem, node);
-		if (tm->elem.seq > min_seq)
+		if (tm->seq > min_seq)
 			continue;
 		rb_erase(node, tm_root);
-		list_del(&tm->elem.list);
 		kfree(tm);
 	}
-	write_unlock(&fs_info->tree_mod_log_lock);
-out:
-	spin_unlock(&fs_info->tree_mod_seq_lock);
+	tree_mod_log_write_unlock(fs_info);
 }
 
 /*
@@ -423,11 +458,9 @@ __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->elem.seq);
+	BUG_ON(!tm || !tm->seq);
 
-	write_lock(&fs_info->tree_mod_log_lock);
 	tm_root = &fs_info->tree_mod_log;
 	new = &tm_root->rb_node;
 	while (*new) {
@@ -437,88 +470,81 @@ __tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct tree_mod_elem *tm)
 			new = &((*new)->rb_left);
 		else if (cur->index > tm->index)
 			new = &((*new)->rb_right);
-		else if (cur->elem.seq < tm->elem.seq)
+		else if (cur->seq < tm->seq)
 			new = &((*new)->rb_left);
-		else if (cur->elem.seq > tm->elem.seq)
+		else if (cur->seq > tm->seq)
 			new = &((*new)->rb_right);
 		else {
 			kfree(tm);
-			ret = -EEXIST;
-			goto unlock;
+			return -EEXIST;
 		}
 	}
 
 	rb_link_node(&tm->node, parent, new);
 	rb_insert_color(&tm->node, tm_root);
-unlock:
-	write_unlock(&fs_info->tree_mod_log_lock);
-	return ret;
+	return 0;
 }
 
+/*
+ * Determines if logging can be omitted. Returns 1 if it can. Otherwise, it
+ * returns zero with the tree_mod_log_lock acquired. The caller must hold
+ * this until all tree mod log insertions are recorded in the rb tree and then
+ * call tree_mod_log_write_unlock() to release.
+ */
 static inline int tree_mod_dont_log(struct btrfs_fs_info *fs_info,
 				    struct extent_buffer *eb) {
 	smp_mb();
 	if (list_empty(&(fs_info)->tree_mod_seq_list))
 		return 1;
-	if (!eb)
-		return 0;
-	if (btrfs_header_level(eb) == 0)
+	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 when
- * needed.
+ * This allocates memory and gets a tree modification sequence number.
  *
- * Returns 0 when no sequence number is needed, < 0 on error.
- * Returns 1 when a sequence number was added. In this case,
- * fs_info->tree_mod_seq_lock was acquired and must be released by the caller
- * after inserting into the rb tree.
+ * 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;
-	int seq;
 
-	if (tree_mod_dont_log(fs_info, NULL))
-		return 0;
-
-	tm = *tm_ret = kzalloc(sizeof(*tm), flags);
+	/*
+	 * 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;
 
-	tm->elem.flags = 0;
-	spin_lock(&fs_info->tree_mod_seq_lock);
-	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, because no blocker exists. items
-		 * are removed from the list only when the existing blocker is
-		 * removed from the list.
-		 */
-		kfree(tm);
-		seq = 0;
-		spin_unlock(&fs_info->tree_mod_seq_lock);
-	} else {
-		__get_tree_mod_seq(fs_info, &tm->elem);
-		seq = tm->elem.seq;
-	}
-
-	return seq;
+	tm->seq = btrfs_inc_tree_mod_seq(fs_info);
+	return tm->seq;
 }
 
-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)
+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)
 {
-	struct tree_mod_elem *tm;
 	int ret;
+	struct tree_mod_elem *tm;
 
 	ret = tree_mod_alloc(fs_info, flags, &tm);
-	if (ret <= 0)
+	if (ret < 0)
 		return ret;
 
 	tm->index = eb->start >> PAGE_CACHE_SHIFT;
@@ -530,8 +556,22 @@ tree_mod_log_insert_key_mask(struct btrfs_fs_info *fs_info,
 	tm->slot = slot;
 	tm->generation = btrfs_node_ptr_generation(eb, slot);
 
-	ret = __tree_mod_log_insert(fs_info, tm);
-	spin_unlock(&fs_info->tree_mod_seq_lock);
+	return __tree_mod_log_insert(fs_info, tm);
+}
+
+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)
+{
+	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;
 }
 
@@ -543,6 +583,14 @@ tree_mod_log_insert_key(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 }
 
 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);
+}
+
+static noinline int
 tree_mod_log_insert_move(struct btrfs_fs_info *fs_info,
 			 struct extent_buffer *eb, int dst_slot, int src_slot,
 			 int nr_items, gfp_t flags)
@@ -555,14 +603,14 @@ tree_mod_log_insert_move(struct btrfs_fs_info *fs_info,
 		return 0;
 
 	for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) {
-		ret = tree_mod_log_insert_key(fs_info, eb, i + dst_slot,
+		ret = tree_mod_log_insert_key_locked(fs_info, eb, i + dst_slot,
 					      MOD_LOG_KEY_REMOVE_WHILE_MOVING);
 		BUG_ON(ret < 0);
 	}
 
 	ret = tree_mod_alloc(fs_info, flags, &tm);
-	if (ret <= 0)
-		return ret;
+	if (ret < 0)
+		goto out;
 
 	tm->index = eb->start >> PAGE_CACHE_SHIFT;
 	tm->slot = src_slot;
@@ -571,10 +619,26 @@ tree_mod_log_insert_move(struct btrfs_fs_info *fs_info,
 	tm->op = MOD_LOG_MOVE_KEYS;
 
 	ret = __tree_mod_log_insert(fs_info, tm);
-	spin_unlock(&fs_info->tree_mod_seq_lock);
+out:
+	tree_mod_log_write_unlock(fs_info);
 	return ret;
 }
 
+static inline void
+__tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
+{
+	int i;
+	u32 nritems;
+	int ret;
+
+	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);
+		BUG_ON(ret < 0);
+	}
+}
+
 static noinline int
 tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 			 struct extent_buffer *old_root,
@@ -583,9 +647,14 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 	struct tree_mod_elem *tm;
 	int ret;
 
+	if (tree_mod_dont_log(fs_info, NULL))
+		return 0;
+
+	__tree_mod_log_free_eb(fs_info, old_root);
+
 	ret = tree_mod_alloc(fs_info, flags, &tm);
-	if (ret <= 0)
-		return ret;
+	if (ret < 0)
+		goto out;
 
 	tm->index = new_root->start >> PAGE_CACHE_SHIFT;
 	tm->old_root.logical = old_root->start;
@@ -594,7 +663,8 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 	tm->op = MOD_LOG_ROOT_REPLACE;
 
 	ret = __tree_mod_log_insert(fs_info, tm);
-	spin_unlock(&fs_info->tree_mod_seq_lock);
+out:
+	tree_mod_log_write_unlock(fs_info);
 	return ret;
 }
 
@@ -608,7 +678,7 @@ __tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq,
 	struct tree_mod_elem *found = NULL;
 	u64 index = start >> PAGE_CACHE_SHIFT;
 
-	read_lock(&fs_info->tree_mod_log_lock);
+	tree_mod_log_read_lock(fs_info);
 	tm_root = &fs_info->tree_mod_log;
 	node = tm_root->rb_node;
 	while (node) {
@@ -617,18 +687,18 @@ __tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq,
 			node = node->rb_left;
 		} else if (cur->index > index) {
 			node = node->rb_right;
-		} else if (cur->elem.seq < min_seq) {
+		} else if (cur->seq < min_seq) {
 			node = node->rb_left;
 		} else if (!smallest) {
 			/* we want the node with the highest seq */
 			if (found)
-				BUG_ON(found->elem.seq > cur->elem.seq);
+				BUG_ON(found->seq > cur->seq);
 			found = cur;
 			node = node->rb_left;
-		} else if (cur->elem.seq > min_seq) {
+		} else if (cur->seq > min_seq) {
 			/* we want the node with the smallest seq */
 			if (found)
-				BUG_ON(found->elem.seq < cur->elem.seq);
+				BUG_ON(found->seq < cur->seq);
 			found = cur;
 			node = node->rb_right;
 		} else {
@@ -636,7 +706,7 @@ __tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq,
 			break;
 		}
 	}
-	read_unlock(&fs_info->tree_mod_log_lock);
+	tree_mod_log_read_unlock(fs_info);
 
 	return found;
 }
@@ -664,7 +734,7 @@ tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq)
 	return __tree_mod_log_search(fs_info, start, min_seq, 0);
 }
 
-static inline void
+static noinline void
 tree_mod_log_eb_copy(struct btrfs_fs_info *fs_info, struct extent_buffer *dst,
 		     struct extent_buffer *src, unsigned long dst_offset,
 		     unsigned long src_offset, int nr_items)
@@ -675,18 +745,23 @@ 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)
+	if (btrfs_header_level(dst) == 0 && btrfs_header_level(src) == 0) {
+		tree_mod_log_write_unlock(fs_info);
 		return;
+	}
 
-	/* speed this up by single seq for all operations? */
 	for (i = 0; i < nr_items; i++) {
-		ret = tree_mod_log_insert_key(fs_info, src, i + src_offset,
-					      MOD_LOG_KEY_REMOVE);
+		ret = tree_mod_log_insert_key_locked(fs_info, src,
+						     i + src_offset,
+						     MOD_LOG_KEY_REMOVE);
 		BUG_ON(ret < 0);
-		ret = tree_mod_log_insert_key(fs_info, dst, i + dst_offset,
-					      MOD_LOG_KEY_ADD);
+		ret = tree_mod_log_insert_key_locked(fs_info, dst,
+						     i + dst_offset,
+						     MOD_LOG_KEY_ADD);
 		BUG_ON(ret < 0);
 	}
+
+	tree_mod_log_write_unlock(fs_info);
 }
 
 static inline void
@@ -699,7 +774,7 @@ tree_mod_log_eb_move(struct btrfs_fs_info *fs_info, struct extent_buffer *dst,
 	BUG_ON(ret < 0);
 }
 
-static inline void
+static noinline void
 tree_mod_log_set_node_key(struct btrfs_fs_info *fs_info,
 			  struct extent_buffer *eb,
 			  struct btrfs_disk_key *disk_key, int slot, int atomic)
@@ -712,30 +787,22 @@ tree_mod_log_set_node_key(struct btrfs_fs_info *fs_info,
 	BUG_ON(ret < 0);
 }
 
-static void tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
-				 struct extent_buffer *eb)
+static noinline void
+tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
 {
-	int i;
-	int ret;