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

Pull btrfs update from Chris Mason:
 "These are mostly fixes.  The biggest exceptions are Josef's skinny
  extents and Jan Schmidt's code to rebuild our quota indexes if they
  get out of sync (or you enable quotas on an existing filesystem).

  The skinny extents are off by default because they are a new variation
  on the extent allocation tree format.  btrfstune -x enables them, and
  the new format makes the extent allocation tree about 30% smaller.

  I rebased this a few days ago to rework Dave Sterba's crc checks on
  the super block, but almost all of these go back to rc6, since I
  though 3.9 was due any minute.

  The biggest missing fix is the tracepoint bug that was hit late in
  3.9.  I ran into problems with that in overnight testing and I'm still
  tracking it down.  I'll definitely have that fixed for rc2."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (101 commits)
  Btrfs: allow superblock mismatch from older mkfs
  btrfs: enhance superblock checks
  btrfs: fix misleading variable name for flags
  btrfs: use unsigned long type for extent state bits
  Btrfs: improve the loop of scrub_stripe
  btrfs: read entire device info under lock
  btrfs: remove unused gfp mask parameter from release_extent_buffer callchain
  btrfs: handle errors returned from get_tree_block_key
  btrfs: make static code static & remove dead code
  Btrfs: deal with errors in write_dev_supers
  Btrfs: remove almost all of the BUG()'s from tree-log.c
  Btrfs: deal with free space cache errors while replaying log
  Btrfs: automatic rescan after "quota enable" command
  Btrfs: rescan for qgroups
  Btrfs: split btrfs_qgroup_account_ref into four functions
  Btrfs: allocate new chunks if the space is not enough for global rsv
  Btrfs: separate sequence numbers for delayed ref tracking and tree mod log
  btrfs: move leak debug code to functions
  Btrfs: return free space in cow error path
  Btrfs: set UUID in root_item for created trees
  ...

1 files changed, 147 insertions, 235 deletions

diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index ca9d8f1a3bb6..de6de8e60b46 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -37,16 +37,11 @@ static int balance_node_right(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
 			      struct extent_buffer *dst_buf,
 			      struct extent_buffer *src_buf);
-static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		    struct btrfs_path *path, int level, int slot);
+static void del_ptr(struct btrfs_root *root, struct btrfs_path *path,
+		    int level, int slot);
 static void tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
 				 struct extent_buffer *eb);
-struct extent_buffer *read_old_tree_block(struct btrfs_root *root, u64 bytenr,
-					  u32 blocksize, u64 parent_transid,
-					  u64 time_seq);
-struct extent_buffer *btrfs_find_old_tree_block(struct btrfs_root *root,
-						u64 bytenr, u32 blocksize,
-						u64 time_seq);
+static int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path);
 
 struct btrfs_path *btrfs_alloc_path(void)
 {
@@ -208,7 +203,7 @@ struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root)
  * tree until you end up with a lock on the root.  A locked buffer
  * is returned, with a reference held.
  */
-struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root)
+static struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root)
 {
 	struct extent_buffer *eb;
 
@@ -361,6 +356,44 @@ static inline void tree_mod_log_write_unlock(struct btrfs_fs_info *fs_info)
 }
 
 /*
+ * Increment the upper half of tree_mod_seq, set lower half zero.
+ *
+ * Must be called with fs_info->tree_mod_seq_lock held.
+ */
+static inline u64 btrfs_inc_tree_mod_seq_major(struct btrfs_fs_info *fs_info)
+{
+	u64 seq = atomic64_read(&fs_info->tree_mod_seq);
+	seq &= 0xffffffff00000000ull;
+	seq += 1ull << 32;
+	atomic64_set(&fs_info->tree_mod_seq, seq);
+	return seq;
+}
+
+/*
+ * Increment the lower half of tree_mod_seq.
+ *
+ * Must be called with fs_info->tree_mod_seq_lock held. The way major numbers
+ * are generated should not technically require a spin lock here. (Rationale:
+ * incrementing the minor while incrementing the major seq number is between its
+ * atomic64_read and atomic64_set calls doesn't duplicate sequence numbers, it
+ * just returns a unique sequence number as usual.) We have decided to leave
+ * that requirement in here and rethink it once we notice it really imposes a
+ * problem on some workload.
+ */
+static inline u64 btrfs_inc_tree_mod_seq_minor(struct btrfs_fs_info *fs_info)
+{
+	return atomic64_inc_return(&fs_info->tree_mod_seq);
+}
+
+/*
+ * return the last minor in the previous major tree_mod_seq number
+ */
+u64 btrfs_tree_mod_seq_prev(u64 seq)
+{
+	return (seq & 0xffffffff00000000ull) - 1ull;
+}
+
+/*
  * 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
@@ -376,10 +409,10 @@ u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info,
 	tree_mod_log_write_lock(fs_info);
 	spin_lock(&fs_info->tree_mod_seq_lock);
 	if (!elem->seq) {
-		elem->seq = btrfs_inc_tree_mod_seq(fs_info);
+		elem->seq = btrfs_inc_tree_mod_seq_major(fs_info);
 		list_add_tail(&elem->list, &fs_info->tree_mod_seq_list);
 	}
-	seq = btrfs_inc_tree_mod_seq(fs_info);
+	seq = btrfs_inc_tree_mod_seq_minor(fs_info);
 	spin_unlock(&fs_info->tree_mod_seq_lock);
 	tree_mod_log_write_unlock(fs_info);
 
@@ -524,7 +557,10 @@ static inline int tree_mod_alloc(struct btrfs_fs_info *fs_info, gfp_t flags,
 	if (!tm)
 		return -ENOMEM;
 
-	tm->seq = btrfs_inc_tree_mod_seq(fs_info);
+	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;
 }
 
@@ -643,7 +679,8 @@ __tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
 static noinline int
 tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 			 struct extent_buffer *old_root,
-			 struct extent_buffer *new_root, gfp_t flags)
+			 struct extent_buffer *new_root, gfp_t flags,
+			 int log_removal)
 {
 	struct tree_mod_elem *tm;
 	int ret;
@@ -651,7 +688,8 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 	if (tree_mod_dont_log(fs_info, NULL))
 		return 0;
 
-	__tree_mod_log_free_eb(fs_info, old_root);
+	if (log_removal)
+		__tree_mod_log_free_eb(fs_info, old_root);
 
 	ret = tree_mod_alloc(fs_info, flags, &tm);
 	if (ret < 0)
@@ -738,7 +776,7 @@ tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq)
 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, int log_removal)
+		     unsigned long src_offset, int nr_items)
 {
 	int ret;
 	int i;
@@ -752,12 +790,10 @@ tree_mod_log_eb_copy(struct btrfs_fs_info *fs_info, struct extent_buffer *dst,
 	}
 
 	for (i = 0; i < nr_items; i++) {
-		if (log_removal) {
-			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_locked(fs_info, src,
+						i + src_offset,
+						MOD_LOG_KEY_REMOVE);
+		BUG_ON(ret < 0);
 		ret = tree_mod_log_insert_key_locked(fs_info, dst,
 						     i + dst_offset,
 						     MOD_LOG_KEY_ADD);
@@ -802,11 +838,12 @@ tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
 
 static noinline void
 tree_mod_log_set_root_pointer(struct btrfs_root *root,
-			      struct extent_buffer *new_root_node)
+			      struct extent_buffer *new_root_node,
+			      int log_removal)
 {
 	int ret;
 	ret = tree_mod_log_insert_root(root->fs_info, root->node,
-				       new_root_node, GFP_NOFS);
+				       new_root_node, GFP_NOFS, log_removal);
 	BUG_ON(ret < 0);
 }
 
@@ -867,7 +904,8 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 
 	if (btrfs_block_can_be_shared(root, buf)) {
 		ret = btrfs_lookup_extent_info(trans, root, buf->start,
-					       buf->len, &refs, &flags);
+					       btrfs_header_level(buf), 1,
+					       &refs, &flags);
 		if (ret)
 			return ret;
 		if (refs == 0) {
@@ -1028,7 +1066,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 			parent_start = 0;
 
 		extent_buffer_get(cow);
-		tree_mod_log_set_root_pointer(root, cow);
+		tree_mod_log_set_root_pointer(root, cow, 1);
 		rcu_assign_pointer(root->node, cow);
 
 		btrfs_free_tree_block(trans, root, buf, parent_start,
@@ -1067,11 +1105,11 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
  */
 static struct tree_mod_elem *
 __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info,
-			   struct btrfs_root *root, u64 time_seq)
+			   struct extent_buffer *eb_root, u64 time_seq)
 {
 	struct tree_mod_elem *tm;
 	struct tree_mod_elem *found = NULL;
-	u64 root_logical = root->node->start;
+	u64 root_logical = eb_root->start;
 	int looped = 0;
 
 	if (!time_seq)
@@ -1105,7 +1143,6 @@ __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info,
 
 		found = tm;
 		root_logical = tm->old_root.logical;
-		BUG_ON(root_logical == root->node->start);
 		looped = 1;
 	}
 
@@ -1190,6 +1227,13 @@ __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
 	btrfs_set_header_nritems(eb, n);
 }
 
+/*
+ * Called with eb read locked. If the buffer cannot be rewinded, the same buffer
+ * is returned. If rewind operations happen, a fresh buffer is returned. The
+ * returned buffer is always read-locked. If the returned buffer is not the
+ * input buffer, the lock on the input buffer is released and the input buffer
+ * 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)
@@ -1223,8 +1267,11 @@ tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 	}
 
 	extent_buffer_get(eb_rewin);
+	btrfs_tree_read_unlock(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);
 	WARN_ON(btrfs_header_nritems(eb_rewin) >
 		BTRFS_NODEPTRS_PER_BLOCK(fs_info->tree_root));
@@ -1243,33 +1290,35 @@ static inline struct extent_buffer *
 get_old_root(struct btrfs_root *root, u64 time_seq)
 {
 	struct tree_mod_elem *tm;
-	struct extent_buffer *eb;
+	struct extent_buffer *eb = NULL;
+	struct extent_buffer *eb_root;
 	struct extent_buffer *old;
 	struct tree_mod_root *old_root = NULL;
 	u64 old_generation = 0;
 	u64 logical;
 	u32 blocksize;
 
-	eb = btrfs_read_lock_root_node(root);
-	tm = __tree_mod_log_oldest_root(root->fs_info, root, time_seq);
+	eb_root = btrfs_read_lock_root_node(root);
+	tm = __tree_mod_log_oldest_root(root->fs_info, eb_root, time_seq);
 	if (!tm)
-		return root->node;
+		return eb_root;
 
 	if (tm->op == MOD_LOG_ROOT_REPLACE) {
 		old_root = &tm->old_root;
 		old_generation = tm->generation;
 		logical = old_root->logical;
 	} else {
-		logical = root->node->start;
+		logical = eb_root->start;
 	}
 
 	tm = tree_mod_log_search(root->fs_info, logical, time_seq);
 	if (old_root && tm && tm->op != MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
-		btrfs_tree_read_unlock(root->node);
-		free_extent_buffer(root->node);
+		btrfs_tree_read_unlock(eb_root);
+		free_extent_buffer(eb_root);
 		blocksize = btrfs_level_size(root, old_root->level);
 		old = read_tree_block(root, logical, blocksize, 0);
-		if (!old) {
+		if (!old || !extent_buffer_uptodate(old)) {
+			free_extent_buffer(old);
 			pr_warn("btrfs: failed to read tree block %llu from get_old_root\n",
 				logical);
 			WARN_ON(1);
@@ -1278,13 +1327,13 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 			free_extent_buffer(old);
 		}
 	} else if (old_root) {
-		btrfs_tree_read_unlock(root->node);
-		free_extent_buffer(root->node);
+		btrfs_tree_read_unlock(eb_root);
+		free_extent_buffer(eb_root);
 		eb = alloc_dummy_extent_buffer(logical, root->nodesize);
 	} else {
-		eb = btrfs_clone_extent_buffer(root->node);
-		btrfs_tree_read_unlock(root->node);
-		free_extent_buffer(root->node);
+		eb = btrfs_clone_extent_buffer(eb_root);
+		btrfs_tree_read_unlock(eb_root);
+		free_extent_buffer(eb_root);
 	}
 
 	if (!eb)
@@ -1294,7 +1343,7 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 	if (old_root) {
 		btrfs_set_header_bytenr(eb, eb->start);
 		btrfs_set_header_backref_rev(eb, BTRFS_MIXED_BACKREF_REV);
-		btrfs_set_header_owner(eb, root->root_key.objectid);
+		btrfs_set_header_owner(eb, btrfs_header_owner(eb_root));
 		btrfs_set_header_level(eb, old_root->level);
 		btrfs_set_header_generation(eb, old_generation);
 	}
@@ -1311,15 +1360,15 @@ int btrfs_old_root_level(struct btrfs_root *root, u64 time_seq)
 {
 	struct tree_mod_elem *tm;
 	int level;
+	struct extent_buffer *eb_root = btrfs_root_node(root);
 
-	tm = __tree_mod_log_oldest_root(root->fs_info, root, time_seq);
+	tm = __tree_mod_log_oldest_root(root->fs_info, eb_root, time_seq);
 	if (tm && tm->op == MOD_LOG_ROOT_REPLACE) {
 		level = tm->old_root.level;
 	} else {
-		rcu_read_lock();
-		level = btrfs_header_level(root->node);
-		rcu_read_unlock();
+		level = btrfs_header_level(eb_root);
 	}
+	free_extent_buffer(eb_root);
 
 	return level;
 }
@@ -1514,8 +1563,10 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 			if (!cur) {
 				cur = read_tree_block(root, blocknr,
 							 blocksize, gen);
-				if (!cur)
+				if (!cur || !extent_buffer_uptodate(cur)) {
+					free_extent_buffer(cur);
 					return -EIO;
+				}
 			} else if (!uptodate) {
 				err = btrfs_read_buffer(cur, gen);
 				if (err) {
@@ -1680,6 +1731,8 @@ static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
 				   struct extent_buffer *parent, int slot)
 {
 	int level = btrfs_header_level(parent);
+	struct extent_buffer *eb;
+
 	if (slot < 0)
 		return NULL;
 	if (slot >= btrfs_header_nritems(parent))
@@ -1687,9 +1740,15 @@ static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
 
 	BUG_ON(level == 0);
 
-	return read_tree_block(root, btrfs_node_blockptr(parent, slot),
-		       btrfs_level_size(root, level - 1),
-		       btrfs_node_ptr_generation(parent, slot));
+	eb = read_tree_block(root, btrfs_node_blockptr(parent, slot),
+			     btrfs_level_size(root, level - 1),
+			     btrfs_node_ptr_generation(parent, slot));
+	if (eb && !extent_buffer_uptodate(eb)) {
+		free_extent_buffer(eb);
+		eb = NULL;
+	}
+
+	return eb;
 }
 
 /*
@@ -1754,7 +1813,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 			goto enospc;
 		}
 
-		tree_mod_log_set_root_pointer(root, child);
+		tree_mod_log_set_root_pointer(root, child, 1);
 		rcu_assign_pointer(root->node, child);
 
 		add_root_to_dirty_list(root);
@@ -1818,7 +1877,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		if (btrfs_header_nritems(right) == 0) {
 			clean_tree_block(trans, root, right);
 			btrfs_tree_unlock(right);
-			del_ptr(trans, root, path, level + 1, pslot + 1);
+			del_ptr(root, path, level + 1, pslot + 1);
 			root_sub_used(root, right->len);
 			btrfs_free_tree_block(trans, root, right, 0, 1);
 			free_extent_buffer_stale(right);
@@ -1862,7 +1921,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 	if (btrfs_header_nritems(mid) == 0) {
 		clean_tree_block(trans, root, mid);
 		btrfs_tree_unlock(mid);
-		del_ptr(trans, root, path, level + 1, pslot);
+		del_ptr(root, path, level + 1, pslot);
 		root_sub_used(root, mid->len);
 		btrfs_free_tree_block(trans, root, mid, 0, 1);
 		free_extent_buffer_stale(mid);
@@ -2210,9 +2269,6 @@ static noinline void unlock_up(struct btrfs_path *path, int level,
 	int no_skips = 0;
 	struct extent_buffer *t;
 
-	if (path->really_keep_locks)
-		return;
-
 	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
 		if (!path->nodes[i])
 			break;
@@ -2260,7 +2316,7 @@ noinline void btrfs_unlock_up_safe(struct btrfs_path *path, int level)
 {
 	int i;
 
-	if (path->keep_locks || path->really_keep_locks)
+	if (path->keep_locks)
 		return;
 
 	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
@@ -2493,7 +2549,7 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (!cow)
 		write_lock_level = -1;
 
-	if (cow && (p->really_keep_locks || p->keep_locks || p->lowest_level))
+	if (cow && (p->keep_locks || p->lowest_level))
 		write_lock_level = BTRFS_MAX_LEVEL;
 
 	min_write_lock_level = write_lock_level;
@@ -2795,15 +2851,9 @@ again:
 				btrfs_clear_path_blocking(p, b,
 							  BTRFS_READ_LOCK);
 			}
+			b = tree_mod_log_rewind(root->fs_info, b, time_seq);
 			p->locks[level] = BTRFS_READ_LOCK;
 			p->nodes[level] = b;
-			b = tree_mod_log_rewind(root->fs_info, b, time_seq);
-			if (b != p->nodes[level]) {
-				btrfs_tree_unlock_rw(p->nodes[level],
-						     p->locks[level]);
-				p->locks[level] = 0;
-				p->nodes[level] = b;
-			}
 		} else {
 			p->slots[level] = slot;
 			unlock_up(p, level, lowest_unlock, 0, NULL);
@@ -2902,8 +2952,7 @@ again:
  * higher levels
  *
  */
-static void fixup_low_keys(struct btrfs_trans_handle *trans,
-			   struct btrfs_root *root, struct btrfs_path *path,
+static void fixup_low_keys(struct btrfs_root *root, struct btrfs_path *path,
 			   struct btrfs_disk_key *key, int level)
 {
 	int i;
@@ -2928,8 +2977,7 @@ static void fixup_low_keys(struct btrfs_trans_handle *trans,
  * This function isn't completely safe. It's the caller's responsibility
  * that the new key won't break the order
  */
-void btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *root, struct btrfs_path *path,
+void btrfs_set_item_key_safe(struct btrfs_root *root, struct btrfs_path *path,
 			     struct btrfs_key *new_key)
 {
 	struct btrfs_disk_key disk_key;
@@ -2951,7 +2999,7 @@ void btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
 	btrfs_set_item_key(eb, &disk_key, slot);
 	btrfs_mark_buffer_dirty(eb);
 	if (slot == 0)
-		fixup_low_keys(trans, root, path, &disk_key, 1);
+		fixup_low_keys(root, path, &disk_key, 1);
 }
 
 /*
@@ -2998,7 +3046,7 @@ static int push_node_left(struct btrfs_trans_handle *trans,
 		push_items = min(src_nritems - 8, push_items);
 
 	tree_mod_log_eb_copy(root->fs_info, dst, src, dst_nritems, 0,
-			     push_items, 1);
+			     push_items);
 	copy_extent_buffer(dst, src,
 			   btrfs_node_key_ptr_offset(dst_nritems),
 			   btrfs_node_key_ptr_offset(0),
@@ -3069,7 +3117,7 @@ static int balance_node_right(struct btrfs_trans_handle *trans,
 				      sizeof(struct btrfs_key_ptr));
 
 	tree_mod_log_eb_copy(root->fs_info, dst, src, 0,
-			     src_nritems - push_items, push_items, 1);
+			     src_nritems - push_items, push_items);
 	copy_extent_buffer(dst, src,
 			   btrfs_node_key_ptr_offset(0),
 			   btrfs_node_key_ptr_offset(src_nritems - push_items),
@@ -3093,7 +3141,7 @@ 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)
+			   struct btrfs_path *path, int level, int log_removal)
 {
 	u64 lower_gen;
 	struct extent_buffer *lower;
@@ -3144,7 +3192,7 @@ 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);
+	tree_mod_log_set_root_pointer(root, c, log_removal);
 	rcu_assign_pointer(root->node, c);
 
 	/* the super has an extra ref to root->node */
@@ -3221,18 +3269,21 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 	int mid;
 	int ret;
 	u32 c_nritems;
-	int tree_mod_log_removal = 1;
 
 	c = path->nodes[level];
 	WARN_ON(btrfs_header_generation(c) != trans->transid);
 	if (c == root->node) {
-		/* trying to split the root, lets make a new one */
-		ret = insert_new_root(trans, root, path, level + 1);
 		/*
-		 * removal of root nodes has been logged by
-		 * tree_mod_log_set_root_pointer due to locking
+		 * trying to split the root, lets make a new one
+		 *
+		 * tree mod log: We pass 0 as log_removal parameter to
+		 * 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.
 		 */
-		tree_mod_log_removal = 0;
+		ret = insert_new_root(trans, root, path, level + 1, 0);
 		if (ret)
 			return ret;
 	} else {
@@ -3270,8 +3321,7 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 			    (unsigned long)btrfs_header_chunk_tree_uuid(split),
 			    BTRFS_UUID_SIZE);
 
-	tree_mod_log_eb_copy(root->fs_info, split, c, 0, mid, c_nritems - mid,
-			     tree_mod_log_removal);
+	tree_mod_log_eb_copy(root->fs_info, split, c, 0, mid, c_nritems - mid);
 	copy_extent_buffer(split, c,
 			   btrfs_node_key_ptr_offset(0),
 			   btrfs_node_key_ptr_offset(mid),
@@ -3687,7 +3737,7 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 		clean_tree_block(trans, root, right);
 
 	btrfs_item_key(right, &disk_key, 0);
-	fixup_low_keys(trans, root, path, &disk_key, 1);
+	fixup_low_keys(root, path, &disk_key, 1);
 
 	/* then fixup the leaf pointer in the path */
 	if (path->slots[0] < push_items) {
@@ -3953,7 +4003,7 @@ static noinline int split_leaf(struct btrfs_trans_handle *trans,
 	}
 
 	if (!path->nodes[1]) {
-		ret = insert_new_root(trans, root, path, 1);
+		ret = insert_new_root(trans, root, path, 1, 1);
 		if (ret)
 			return ret;
 	}
@@ -4047,8 +4097,7 @@ again:
 			path->nodes[0] = right;
 			path->slots[0] = 0;
 			if (path->slots[1] == 0)
-				fixup_low_keys(trans, root, path,
-					       &disk_key, 1);
+				fixup_low_keys(root, path, &disk_key, 1);
 		}
 		btrfs_mark_buffer_dirty(right);
 		return ret;
@@ -4264,7 +4313,7 @@ int btrfs_duplicate_item(struct btrfs_trans_handle *trans,
 		return ret;
 
 	path->slots[0]++;
-	setup_items_for_insert(trans, root, path, new_key, &item_size,
+	setup_items_for_insert(root, path, new_key, &item_size,
 			       item_size, item_size +
 			       sizeof(struct btrfs_item), 1);
 	leaf = path->nodes[0];
@@ -4281,9 +4330,7 @@ int btrfs_duplicate_item(struct btrfs_trans_handle *trans,
  * off the end of the item or if we shift the item to chop bytes off
  * the front.
  */
-void btrfs_truncate_item(struct btrfs_trans_handle *trans,
-			 struct btrfs_root *root,
-			 struct btrfs_path *path,
+void btrfs_truncate_item(struct btrfs_root *root, struct btrfs_path *path,
 			 u32 new_size, int from_end)
 {
 	int slot;
@@ -4367,7 +4414,7 @@ void btrfs_truncate_item(struct btrfs_trans_handle *trans,
 		btrfs_set_disk_key_offset(&disk_key, offset + size_diff);
 		btrfs_set_item_key(leaf, &disk_key, slot);
 		if (slot == 0)
-			fixup_low_keys(trans, root, path, &disk_key, 1);
+			fixup_low_keys(root, path, &disk_key, 1);
 	}
 
 	item = btrfs_item_nr(leaf, slot);
@@ -4383,8 +4430,7 @@ void btrfs_truncate_item(struct btrfs_trans_handle *trans,
 /*
  * make the item pointed to by the path bigger, data_size is the new size.
  */
-void btrfs_extend_item(struct btrfs_trans_handle *trans,
-		       struct btrfs_root *root, struct btrfs_path *path,
+void btrfs_extend_item(struct btrfs_root *root, struct btrfs_path *path,
 		       u32 data_size)
 {
 	int slot;
@@ -4454,8 +4500,7 @@ void btrfs_extend_item(struct btrfs_trans_handle *trans,
  * to save stack depth by doing the bulk of the work in a function
  * that doesn't call btrfs_search_slot
  */
-void setup_items_for_insert(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root, struct btrfs_path *path,
+void setup_items_for_insert(struct btrfs_root *root, struct btrfs_path *path,
 			    struct btrfs_key *cpu_key, u32 *data_size,
 			    u32 total_data, u32 total_size, int nr)
 {
@@ -4531,7 +4576,7 @@ void setup_items_for_insert(struct btrfs_trans_handle *trans,
 
 	if (slot == 0) {
 		btrfs_cpu_key_to_disk(&disk_key, cpu_key);
-		fixup_low_keys(trans, root, path, &disk_key, 1);
+		fixup_low_keys(root, path, &disk_key, 1);
 	}
 	btrfs_unlock_up_safe(path, 1);
 	btrfs_mark_buffer_dirty(leaf);
@@ -4571,7 +4616,7 @@ int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
 	slot = path->slots[0];
 	BUG_ON(slot < 0);
 
-	setup_items_for_insert(trans, root, path, cpu_key, data_size,
+	setup_items_for_insert(root, path, cpu_key, data_size,
 			       total_data, total_size, nr);
 	return 0;
 }
@@ -4609,8 +4654,8 @@ int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
  * the tree should have been previously balanced so the deletion does not
  * empty a node.
  */
-static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		    struct btrfs_path *path, int level, int slot)
+static void del_ptr(struct btrfs_root *root, struct btrfs_path *path,
+		    int level, int slot)
 {
 	struct extent_buffer *parent = path->nodes[level];
 	u32 nritems;
@@ -4642,7 +4687,7 @@ static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		struct btrfs_disk_key disk_key;
 
 		btrfs_node_key(parent, &disk_key, 0);
-		fixup_low_keys(trans, root, path, &disk_key, level + 1);
+		fixup_low_keys(root, path, &disk_key, level + 1);
 	}
 	btrfs_mark_buffer_dirty(parent);
 }
@@ -4663,7 +4708,7 @@ static noinline void btrfs_del_leaf(struct btrfs_trans_handle *trans,
 				    struct extent_buffer *leaf)
 {
 	WARN_ON(btrfs_header_generation(leaf) != trans->transid);
-	del_ptr(trans, root, path, 1, path->slots[1]);
+	del_ptr(root, path, 1, path->slots[1]);
 
 	/*
 	 * btrfs_free_extent is expensive, we want to make sure we
@@ -4744,7 +4789,7 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 			struct btrfs_disk_key disk_key;
 
 			btrfs_item_key(leaf, &disk_key, 0);
-			fixup_low_keys(trans, root, path, &disk_key, 1);
+			fixup_low_keys(root, path, &disk_key, 1);
 		}
 
 		/* delete the leaf if it is mostly empty */
@@ -5464,139 +5509,6 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
 	return btrfs_next_old_leaf(root, path, 0);
 }
 
-/* Release the path up to but not including the given level */
-static void btrfs_release_level(struct btrfs_path *path, int level)
-{
-	int i;
-
-	for (i = 0; i < level; i++) {
-		path->slots[i] = 0;
-		if (!path->nodes[i])
-			continue;
-		if (path->locks[i]) {
-			btrfs_tree_unlock_rw(path->nodes[i], path->locks[i]);
-			path->locks[i] = 0;
-		}
-		free_extent_buffer(path->nodes[i]);
-		path->nodes[i] = NULL;
-	}
-}
-
-/*
- * This function assumes 2 things
- *
- * 1) You are using path->keep_locks
- * 2) You are not inserting items.
- *
- * If either of these are not true do not use this function. If you need a next
- * leaf with either of these not being true then this function can be easily
- * adapted to do that, but at the moment these are the limitations.
- */
-int btrfs_next_leaf_write(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root, struct btrfs_path *path,
-			  int del)
-{
-	struct extent_buffer *b;
-	struct btrfs_key key;
-	u32 nritems;
-	int level = 1;
-	int slot;
-	int ret = 1;
-	int write_lock_level = BTRFS_MAX_LEVEL;
-	int ins_len = del ? -1 : 0;
-
-	WARN_ON(!(path->keep_locks || path->really_keep_locks));
-
-	nritems = btrfs_header_nritems(path->nodes[0]);
-	btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1);
-
-	while (path->nodes[level]) {
-		nritems = btrfs_header_nritems(path->nodes[level]);
-		if (!(path->locks[level] & BTRFS_WRITE_LOCK)) {
-search:
-			btrfs_release_path(path);
-			ret = btrfs_search_slot(trans, root, &key, path,
-						ins_len, 1);
-			if (ret < 0)
-				goto out;
-			level = 1;
-			continue;
-		}
-
-		if (path->slots[level] >= nritems - 1) {
-			level++;
-			continue;
-		}
-
-		btrfs_release_level(path, level);
-		break;
-	}
-
-	if (!path->nodes[level]) {
-		ret = 1;
-		goto out;
-	}
-
-	path->slots[level]++;
-	b = path->nodes[level];
-
-	while (b) {
-		level = btrfs_header_level(b);
-
-		if (!should_cow_block(trans, root, b))
-			goto cow_done;
-
-		btrfs_set_path_blocking(path);
-		ret = btrfs_cow_block(trans, root, b,
-				      path->nodes[level + 1],
-				      path->slots[level + 1], &b);
-		if (ret)
-			goto out;
-cow_done:
-		path->nodes[level] = b;
-		btrfs_clear_path_blocking(path, NULL, 0);
-		if (level != 0) {
-			ret = setup_nodes_for_search(trans, root, path, b,
-						     level, ins_len,
-						     &write_lock_level);
-			if (ret == -EAGAIN)
-				goto search;
-			if (ret)
-				goto out;
-
-			b = path->nodes[level];
-			slot = path->slots[level];
-
-			ret = read_block_for_search(trans, root, path,
-						    &b, level, slot, &key, 0);
-			if (ret == -EAGAIN)
-				goto search;
-			if (ret)
-				goto out;
-			level = btrfs_header_level(b);
-			if (!btrfs_try_tree_write_lock(b)) {
-				btrfs_set_path_blocking(path);
-				btrfs_tree_lock(b);
-				btrfs_clear_path_blocking(path, b,
-							  BTRFS_WRITE_LOCK);
-			}
-			path->locks[level] = BTRFS_WRITE_LOCK;
-			path->nodes[level] = b;
-			path->slots[level] = 0;
-		} else {
-			path->slots[level] = 0;
-			ret = 0;
-			break;
-		}
-	}
-
-out:
-	if (ret)
-		btrfs_release_path(path);
-
-	return ret;
-}
-
 int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path,
 			u64 time_seq)
 {