extent-tree.c

/*
 * Copyright (C) 2007 Oracle.  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/sched.h>
#include <linux/pagemap.h>
#include <linux/writeback.h>
#include <linux/blkdev.h>
#include "hash.h"
#include "crc32c.h"
#include "ctree.h"
#include "disk-io.h"
#include "print-tree.h"
#include "transaction.h"
#include "volumes.h"
#include "locking.h"
#include "ref-cache.h"

#define PENDING_EXTENT_INSERT 0
#define PENDING_EXTENT_DELETE 1
#define PENDING_BACKREF_UPDATE 2

struct pending_extent_op {
	int type;
	u64 bytenr;
	u64 num_bytes;
	u64 parent;
	u64 orig_parent;
	u64 generation;
	u64 orig_generation;
	int level;
};

static int finish_current_insert(struct btrfs_trans_handle *trans, struct
				 btrfs_root *extent_root);
static int del_pending_extents(struct btrfs_trans_handle *trans, struct
			       btrfs_root *extent_root);
static struct btrfs_block_group_cache *
__btrfs_find_block_group(struct btrfs_root *root,
			 struct btrfs_block_group_cache *hint,
			 u64 search_start, int data, int owner);

void maybe_lock_mutex(struct btrfs_root *root)
{
	if (root != root->fs_info->extent_root &&
	    root != root->fs_info->chunk_root &&
	    root != root->fs_info->dev_root) {
		mutex_lock(&root->fs_info->alloc_mutex);
	}
}

void maybe_unlock_mutex(struct btrfs_root *root)
{
	if (root != root->fs_info->extent_root &&
	    root != root->fs_info->chunk_root &&
	    root != root->fs_info->dev_root) {
		mutex_unlock(&root->fs_info->alloc_mutex);
	}
}

static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
{
	return (cache->flags & bits) == bits;
}

/*
 * this adds the block group to the fs_info rb tree for the block group
 * cache
 */
int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
				struct btrfs_block_group_cache *block_group)
{
	struct rb_node **p;
	struct rb_node *parent = NULL;
	struct btrfs_block_group_cache *cache;

	spin_lock(&info->block_group_cache_lock);
	p = &info->block_group_cache_tree.rb_node;

	while (*p) {
		parent = *p;
		cache = rb_entry(parent, struct btrfs_block_group_cache,
				 cache_node);
		if (block_group->key.objectid < cache->key.objectid) {
			p = &(*p)->rb_left;
		} else if (block_group->key.objectid > cache->key.objectid) {
			p = &(*p)->rb_right;
		} else {
			spin_unlock(&info->block_group_cache_lock);
			return -EEXIST;
		}
	}

	rb_link_node(&block_group->cache_node, parent, p);
	rb_insert_color(&block_group->cache_node,
			&info->block_group_cache_tree);
	spin_unlock(&info->block_group_cache_lock);

	return 0;
}

/*
 * This will return the block group at or after bytenr if contains is 0, else
 * it will return the block group that contains the bytenr
 */
static struct btrfs_block_group_cache *
block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr,
			      int contains)
{
	struct btrfs_block_group_cache *cache, *ret = NULL;
	struct rb_node *n;
	u64 end, start;

	spin_lock(&info->block_group_cache_lock);
	n = info->block_group_cache_tree.rb_node;

	while (n) {
		cache = rb_entry(n, struct btrfs_block_group_cache,
				 cache_node);
		end = cache->key.objectid + cache->key.offset - 1;
		start = cache->key.objectid;

		if (bytenr < start) {
			if (!contains && (!ret || start < ret->key.objectid))
				ret = cache;
			n = n->rb_left;
		} else if (bytenr > start) {
			if (contains && bytenr <= end) {
				ret = cache;
				break;
			}
			n = n->rb_right;
		} else {
			ret = cache;
			break;
		}
	}
	spin_unlock(&info->block_group_cache_lock);

	return ret;
}

/*
 * this is only called by cache_block_group, since we could have freed extents
 * we need to check the pinned_extents for any extents that can't be used yet
 * since their free space will be released as soon as the transaction commits.
 */
static int add_new_free_space(struct btrfs_block_group_cache *block_group,
			      struct btrfs_fs_info *info, u64 start, u64 end)
{
	u64 extent_start, extent_end, size;
	int ret;

	while (start < end) {
		ret = find_first_extent_bit(&info->pinned_extents, start,
					    &extent_start, &extent_end,
					    EXTENT_DIRTY);
		if (ret)
			break;

		if (extent_start == start) {
			start = extent_end + 1;
		} else if (extent_start > start && extent_start < end) {
			size = extent_start - start;
			ret = btrfs_add_free_space(block_group, start, size);
			BUG_ON(ret);
			start = extent_end + 1;
		} else {
			break;
		}
	}

	if (start < end) {
		size = end - start;
		ret = btrfs_add_free_space(block_group, start, size);
		BUG_ON(ret);
	}

	return 0;
}

static int cache_block_group(struct btrfs_root *root,
			     struct btrfs_block_group_cache *block_group)
{
	struct btrfs_path *path;
	int ret = 0;
	struct btrfs_key key;
	struct extent_buffer *leaf;
	int slot;
	u64 last = 0;
	u64 first_free;
	int found = 0;

	if (!block_group)
		return 0;

	root = root->fs_info->extent_root;

	if (block_group->cached)
		return 0;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	path->reada = 2;
	/*
	 * we get into deadlocks with paths held by callers of this function.
	 * since the alloc_mutex is protecting things right now, just
	 * skip the locking here
	 */
	path->skip_locking = 1;
	first_free = max_t(u64, block_group->key.objectid,
			   BTRFS_SUPER_INFO_OFFSET + BTRFS_SUPER_INFO_SIZE);
	key.objectid = block_group->key.objectid;
	key.offset = 0;
	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto err;
	ret = btrfs_previous_item(root, path, 0, BTRFS_EXTENT_ITEM_KEY);
	if (ret < 0)
		goto err;
	if (ret == 0) {
		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
		if (key.objectid + key.offset > first_free)
			first_free = key.objectid + key.offset;
	}
	while(1) {
		leaf = path->nodes[0];
		slot = path->slots[0];
		if (slot >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto err;
			if (ret == 0)
				continue;
			else
				break;
		}
		btrfs_item_key_to_cpu(leaf, &key, slot);
		if (key.objectid < block_group->key.objectid)
			goto next;

		if (key.objectid >= block_group->key.objectid +
		    block_group->key.offset)
			break;

		if (btrfs_key_type(&key) == BTRFS_EXTENT_ITEM_KEY) {
			if (!found) {
				last = first_free;
				found = 1;
			}

			add_new_free_space(block_group, root->fs_info, last,
					   key.objectid);

			last = key.objectid + key.offset;
		}
next:
		path->slots[0]++;
	}

	if (!found)
		last = first_free;

	add_new_free_space(block_group, root->fs_info, last,
			   block_group->key.objectid +
			   block_group->key.offset);

	block_group->cached = 1;
	ret = 0;
err:
	btrfs_free_path(path);
	return ret;
}

/*
 * return the block group that starts at or after bytenr
 */
struct btrfs_block_group_cache *btrfs_lookup_first_block_group(struct
						       btrfs_fs_info *info,
							 u64 bytenr)
{
	struct btrfs_block_group_cache *cache;

	cache = block_group_cache_tree_search(info, bytenr, 0);

	return cache;
}

/*
 * return the block group that contains teh given bytenr
 */
struct btrfs_block_group_cache *btrfs_lookup_block_group(struct
							 btrfs_fs_info *info,
							 u64 bytenr)
{
	struct btrfs_block_group_cache *cache;

	cache = block_group_cache_tree_search(info, bytenr, 1);

	return cache;
}

static int noinline find_free_space(struct btrfs_root *root,
				    struct btrfs_block_group_cache **cache_ret,
				    u64 *start_ret, u64 num, int data)
{
	int ret;
	struct btrfs_block_group_cache *cache = *cache_ret;
	struct btrfs_free_space *info = NULL;
	u64 last;
	u64 total_fs_bytes;
	u64 search_start = *start_ret;

	WARN_ON(!mutex_is_locked(&root->fs_info->alloc_mutex));
	total_fs_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);

	if (!cache)
		goto out;

	last = max(search_start, cache->key.objectid);

again:
	ret = cache_block_group(root, cache);
	if (ret)
		goto out;

	if (cache->ro || !block_group_bits(cache, data))
		goto new_group;

	info = btrfs_find_free_space(cache, last, num);
	if (info) {
		*start_ret = info->offset;
		return 0;
	}

new_group:
	last = cache->key.objectid + cache->key.offset;

	cache = btrfs_lookup_first_block_group(root->fs_info, last);
	if (!cache || cache->key.objectid >= total_fs_bytes)
		goto out;

	*cache_ret = cache;
	goto again;

out:
	return -ENOSPC;
}

static u64 div_factor(u64 num, int factor)
{
	if (factor == 10)
		return num;
	num *= factor;
	do_div(num, 10);
	return num;
}

static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
						  u64 flags)
{
	struct list_head *head = &info->space_info;
	struct list_head *cur;
	struct btrfs_space_info *found;
	list_for_each(cur, head) {
		found = list_entry(cur, struct btrfs_space_info, list);
		if (found->flags == flags)
			return found;
	}
	return NULL;

}

static struct btrfs_block_group_cache *
__btrfs_find_block_group(struct btrfs_root *root,
			 struct btrfs_block_group_cache *hint,
			 u64 search_start, int data, int owner)
{
	struct btrfs_block_group_cache *cache;
	struct btrfs_block_group_cache *found_group = NULL;
	struct btrfs_fs_info *info = root->fs_info;
	struct btrfs_space_info *sinfo;
	u64 used;
	u64 last = 0;
	u64 free_check;
	int full_search = 0;
	int factor = 10;
	int wrapped = 0;

	if (data & BTRFS_BLOCK_GROUP_METADATA)
		factor = 9;

	if (search_start) {
		struct btrfs_block_group_cache *shint;
		shint = btrfs_lookup_first_block_group(info, search_start);
		if (shint && block_group_bits(shint, data) && !shint->ro) {
			spin_lock(&shint->lock);
			used = btrfs_block_group_used(&shint->item);
			if (used + shint->pinned <
			    div_factor(shint->key.offset, factor)) {
				spin_unlock(&shint->lock);
				return shint;
			}
			spin_unlock(&shint->lock);
		}
	}
	if (hint && !hint->ro && block_group_bits(hint, data)) {
		spin_lock(&hint->lock);
		used = btrfs_block_group_used(&hint->item);
		if (used + hint->pinned <
		    div_factor(hint->key.offset, factor)) {
			spin_unlock(&hint->lock);
			return hint;
		}
		spin_unlock(&hint->lock);
		last = hint->key.objectid + hint->key.offset;
	} else {
		if (hint)
			last = max(hint->key.objectid, search_start);
		else
			last = search_start;
	}
	sinfo = __find_space_info(root->fs_info, data);
	if (!sinfo)
		goto found;
again:
	while(1) {
		struct list_head *l;

		cache = NULL;

		spin_lock(&sinfo->lock);
		list_for_each(l, &sinfo->block_groups) {
			struct btrfs_block_group_cache *entry;
			entry = list_entry(l, struct btrfs_block_group_cache,
					   list);
			if ((entry->key.objectid >= last) &&
			    (!cache || (entry->key.objectid <
					cache->key.objectid)))
				cache = entry;
		}
		spin_unlock(&sinfo->lock);

		if (!cache)
			break;

		spin_lock(&cache->lock);
		last = cache->key.objectid + cache->key.offset;
		used = btrfs_block_group_used(&cache->item);

		if (!cache->ro && block_group_bits(cache, data)) {
			free_check = div_factor(cache->key.offset, factor);
			if (used + cache->pinned < free_check) {
				found_group = cache;
				spin_unlock(&cache->lock);
				goto found;
			}
		}
		spin_unlock(&cache->lock);
		cond_resched();
	}
	if (!wrapped) {
		last = search_start;
		wrapped = 1;
		goto again;
	}
	if (!full_search && factor < 10) {
		last = search_start;
		full_search = 1;
		factor = 10;
		goto again;
	}
found:
	return found_group;
}

struct btrfs_block_group_cache *btrfs_find_block_group(struct btrfs_root *root,
						 struct btrfs_block_group_cache
						 *hint, u64 search_start,
						 int data, int owner)
{

	struct btrfs_block_group_cache *ret;
	ret = __btrfs_find_block_group(root, hint, search_start, data, owner);
	return ret;
}

/* simple helper to search for an existing extent at a given offset */
int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_path *path;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	maybe_lock_mutex(root);
	key.objectid = start;
	key.offset = len;
	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
	ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path,
				0, 0);
	maybe_unlock_mutex(root);
	btrfs_free_path(path);
	return ret;
}

/*
 * Back reference rules.  Back refs have three main goals:
 *
 * 1) differentiate between all holders of references to an extent so that
 *    when a reference is dropped we can make sure it was a valid reference
 *    before freeing the extent.
 *
 * 2) Provide enough information to quickly find the holders of an extent
 *    if we notice a given block is corrupted or bad.
 *
 * 3) Make it easy to migrate blocks for FS shrinking or storage pool
 *    maintenance.  This is actually the same as #2, but with a slightly
 *    different use case.
 *
 * File extents can be referenced by:
 *
 * - multiple snapshots, subvolumes, or different generations in one subvol
 * - different files inside a single subvolume
 * - different offsets inside a file (bookend extents in file.c)
 *
 * The extent ref structure has fields for:
 *
 * - Objectid of the subvolume root
 * - Generation number of the tree holding the reference
 * - objectid of the file holding the reference
 * - offset in the file corresponding to the key holding the reference
 * - number of references holding by parent node (alway 1 for tree blocks)
 *
 * Btree leaf may hold multiple references to a file extent. In most cases,
 * these references are from same file and the corresponding offsets inside
 * the file are close together. So inode objectid and offset in file are
 * just hints, they provide hints about where in the btree the references
 * can be found and when we can stop searching.
 *
 * When a file extent is allocated the fields are filled in:
 *     (root_key.objectid, trans->transid, inode objectid, offset in file, 1)
 *
 * When a leaf is cow'd new references are added for every file extent found
 * in the leaf.  It looks similar to the create case, but trans->transid will
 * be different when the block is cow'd.
 *
 *     (root_key.objectid, trans->transid, inode objectid, offset in file,
 *      number of references in the leaf)
 *
 * Because inode objectid and offset in file are just hints, they are not
 * used when backrefs are deleted. When a file extent is removed either
 * during snapshot deletion or file truncation, we find the corresponding
 * back back reference and check the following fields.
 *
 *     (btrfs_header_owner(leaf), btrfs_header_generation(leaf))
 *
 * Btree extents can be referenced by:
 *
 * - Different subvolumes
 * - Different generations of the same subvolume
 *
 * When a tree block is created, back references are inserted:
 *
 * (root->root_key.objectid, trans->transid, level, 0, 1)
 *
 * When a tree block is cow'd, new back references are added for all the
 * blocks it points to. If the tree block isn't in reference counted root,
 * the old back references are removed. These new back references are of
 * the form (trans->transid will have increased since creation):
 *
 * (root->root_key.objectid, trans->transid, level, 0, 1)
 *
 * When a backref is in deleting, the following fields are checked:
 *
 * if backref was for a tree root:
 *     (btrfs_header_owner(itself), btrfs_header_generation(itself))
 * else
 *     (btrfs_header_owner(parent), btrfs_header_generation(parent))
 *
 * Back Reference Key composing:
 *
 * The key objectid corresponds to the first byte in the extent, the key
 * type is set to BTRFS_EXTENT_REF_KEY, and the key offset is the first
 * byte of parent extent. If a extent is tree root, the key offset is set
 * to the key objectid.
 */

static int noinline lookup_extent_backref(struct btrfs_trans_handle *trans,
					  struct btrfs_root *root,
					  struct btrfs_path *path, u64 bytenr,
					  u64 parent, u64 ref_root,
					  u64 ref_generation, int del)
{
	struct btrfs_key key;
	struct btrfs_extent_ref *ref;
	struct extent_buffer *leaf;
	int ret;

	key.objectid = bytenr;
	key.type = BTRFS_EXTENT_REF_KEY;
	key.offset = parent;

	ret = btrfs_search_slot(trans, root, &key, path, del ? -1 : 0, 1);
	if (ret < 0)
		goto out;
	if (ret > 0) {
		ret = -ENOENT;
		goto out;
	}

	leaf = path->nodes[0];
	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_ref);
	if (btrfs_ref_root(leaf, ref) != ref_root ||
	    btrfs_ref_generation(leaf, ref) != ref_generation) {
		ret = -EIO;
		WARN_ON(1);
		goto out;
	}
	ret = 0;
out:
	return ret;
}

static int noinline insert_extent_backref(struct btrfs_trans_handle *trans,
					  struct btrfs_root *root,
					  struct btrfs_path *path,
					  u64 bytenr, u64 parent,
					  u64 ref_root, u64 ref_generation,
					  u64 owner_objectid, u64 owner_offset)
{
	struct btrfs_key key;
	struct extent_buffer *leaf;
	struct btrfs_extent_ref *ref;
	u32 num_refs;
	int ret;

	key.objectid = bytenr;
	key.type = BTRFS_EXTENT_REF_KEY;
	key.offset = parent;

	ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*ref));
	if (ret == 0) {
		leaf = path->nodes[0];
		ref = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_extent_ref);
		btrfs_set_ref_root(leaf, ref, ref_root);
		btrfs_set_ref_generation(leaf, ref, ref_generation);
		btrfs_set_ref_objectid(leaf, ref, owner_objectid);
		btrfs_set_ref_offset(leaf, ref, owner_offset);
		btrfs_set_ref_num_refs(leaf, ref, 1);
	} else if (ret == -EEXIST) {
		u64 existing_owner;
		BUG_ON(owner_objectid < BTRFS_FIRST_FREE_OBJECTID);
		leaf = path->nodes[0];
		ref = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_extent_ref);
		if (btrfs_ref_root(leaf, ref) != ref_root ||
		    btrfs_ref_generation(leaf, ref) != ref_generation) {
			ret = -EIO;
			WARN_ON(1);
			goto out;
		}

		num_refs = btrfs_ref_num_refs(leaf, ref);
		BUG_ON(num_refs == 0);
		btrfs_set_ref_num_refs(leaf, ref, num_refs + 1);

		existing_owner = btrfs_ref_objectid(leaf, ref);
		if (existing_owner == owner_objectid &&
		    btrfs_ref_offset(leaf, ref) > owner_offset) {
			btrfs_set_ref_offset(leaf, ref, owner_offset);
		} else if (existing_owner != owner_objectid &&
			   existing_owner != BTRFS_MULTIPLE_OBJECTIDS) {
			btrfs_set_ref_objectid(leaf, ref,
					BTRFS_MULTIPLE_OBJECTIDS);
			btrfs_set_ref_offset(leaf, ref, 0);
		}
		ret = 0;
	} else {
		goto out;
	}
	btrfs_mark_buffer_dirty(path->nodes[0]);
out:
	btrfs_release_path(root, path);
	return ret;
}

static int noinline remove_extent_backref(struct btrfs_trans_handle *trans,
					  struct btrfs_root *root,
					  struct btrfs_path *path)
{
	struct extent_buffer *leaf;
	struct btrfs_extent_ref *ref;
	u32 num_refs;
	int ret = 0;

	leaf = path->nodes[0];
	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_ref);
	num_refs = btrfs_ref_num_refs(leaf, ref);
	BUG_ON(num_refs == 0);
	num_refs -= 1;
	if (num_refs == 0) {
		ret = btrfs_del_item(trans, root, path);
	} else {
		btrfs_set_ref_num_refs(leaf, ref, num_refs);
		btrfs_mark_buffer_dirty(leaf);
	}
	btrfs_release_path(root, path);
	return ret;
}

static int __btrfs_update_extent_ref(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root, u64 bytenr,
				     u64 orig_parent, u64 parent,
				     u64 orig_root, u64 ref_root,
				     u64 orig_generation, u64 ref_generation,
				     u64 owner_objectid, u64 owner_offset)
{
	int ret;
	struct btrfs_root *extent_root = root->fs_info->extent_root;
	struct btrfs_path *path;

	if (root == root->fs_info->extent_root) {
		struct pending_extent_op *extent_op;
		u64 num_bytes;

		BUG_ON(owner_objectid >= BTRFS_MAX_LEVEL);
		num_bytes = btrfs_level_size(root, (int)owner_objectid);
		if (test_range_bit(&root->fs_info->extent_ins, bytenr,
				bytenr + num_bytes - 1, EXTENT_LOCKED, 0)) {
			u64 priv;
			ret = get_state_private(&root->fs_info->extent_ins,
						bytenr, &priv);
			BUG_ON(ret);
			extent_op = (struct pending_extent_op *)
							(unsigned long)priv;
			BUG_ON(extent_op->parent != orig_parent);
			BUG_ON(extent_op->generation != orig_generation);
			extent_op->parent = parent;
			extent_op->generation = ref_generation;
		} else {
			extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
			BUG_ON(!extent_op);

			extent_op->type = PENDING_BACKREF_UPDATE;
			extent_op->bytenr = bytenr;
			extent_op->num_bytes = num_bytes;
			extent_op->parent = parent;
			extent_op->orig_parent = orig_parent;
			extent_op->generation = ref_generation;
			extent_op->orig_generation = orig_generation;
			extent_op->level = (int)owner_objectid;

			set_extent_bits(&root->fs_info->extent_ins,
					bytenr, bytenr + num_bytes - 1,
					EXTENT_LOCKED, GFP_NOFS);
			set_state_private(&root->fs_info->extent_ins,
					  bytenr, (unsigned long)extent_op);
		}
		return 0;
	}

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;
	ret = lookup_extent_backref(trans, extent_root, path,
				    bytenr, orig_parent, orig_root,
				    orig_generation, 1);
	if (ret)
		goto out;
	ret = remove_extent_backref(trans, extent_root, path);
	if (ret)
		goto out;
	ret = insert_extent_backref(trans, extent_root, path, bytenr,
				    parent, ref_root, ref_generation,
				    owner_objectid, owner_offset);
	BUG_ON(ret);
	finish_current_insert(trans, extent_root);
	del_pending_extents(trans, extent_root);
out:
	btrfs_free_path(path);
	return ret;
}

int btrfs_update_extent_ref(struct btrfs_trans_handle *trans,
			    struct btrfs_root *root, u64 bytenr,
			    u64 orig_parent, u64 parent,
			    u64 ref_root, u64 ref_generation,
			    u64 owner_objectid, u64 owner_offset)
{
	int ret;
	if (ref_root == BTRFS_TREE_LOG_OBJECTID &&
	    owner_objectid < BTRFS_FIRST_FREE_OBJECTID)
		return 0;
	maybe_lock_mutex(root);
	ret = __btrfs_update_extent_ref(trans, root, bytenr, orig_parent,
					parent, ref_root, ref_root,
					ref_generation, ref_generation,
					owner_objectid, owner_offset);
	maybe_unlock_mutex(root);
	return ret;
}

static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root, u64 bytenr,
				  u64 orig_parent, u64 parent,
				  u64 orig_root, u64 ref_root,
				  u64 orig_generation, u64 ref_generation,
				  u64 owner_objectid, u64 owner_offset)
{
	struct btrfs_path *path;
	int ret;
	struct btrfs_key key;
	struct extent_buffer *l;
	struct btrfs_extent_item *item;
	u32 refs;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	path->reada = 1;
	key.objectid = bytenr;
	key.type = BTRFS_EXTENT_ITEM_KEY;
	key.offset = (u64)-1;

	ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, path,
				0, 1);
	if (ret < 0)
		return ret;
	BUG_ON(ret == 0 || path->slots[0] == 0);

	path->slots[0]--;
	l = path->nodes[0];

	btrfs_item_key_to_cpu(l, &key, path->slots[0]);
	BUG_ON(key.objectid != bytenr);
	BUG_ON(key.type != BTRFS_EXTENT_ITEM_KEY);

	item = btrfs_item_ptr(l, path->slots[0], struct btrfs_extent_item);
	refs = btrfs_extent_refs(l, item);
	btrfs_set_extent_refs(l, item, refs + 1);
	btrfs_mark_buffer_dirty(path->nodes[0]);

	btrfs_release_path(root->fs_info->extent_root, path);

	path->reada = 1;
	ret = insert_extent_backref(trans, root->fs_info->extent_root,
				    path, bytenr, parent,
				    ref_root, ref_generation,
				    owner_objectid, owner_offset);
	BUG_ON(ret);
	finish_current_insert(trans, root->fs_info->extent_root);
	del_pending_extents(trans, root->fs_info->extent_root);

	btrfs_free_path(path);
	return 0;
}

int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
			 struct btrfs_root *root,
			 u64 bytenr, u64 num_bytes, u64 parent,
			 u64 ref_root, u64 ref_generation,
			 u64 owner_objectid, u64 owner_offset)
{
	int ret;
	if (ref_root == BTRFS_TREE_LOG_OBJECTID &&
	    owner_objectid < BTRFS_FIRST_FREE_OBJECTID)
		return 0;
	maybe_lock_mutex(root);
	ret = __btrfs_inc_extent_ref(trans, root, bytenr, 0, parent,
				     0, ref_root, 0, ref_generation,
				     owner_objectid, owner_offset);
	maybe_unlock_mutex(root);
	return ret;
}

int btrfs_extent_post_op(struct btrfs_trans_handle *trans,
			 struct btrfs_root *root)
{
	finish_current_insert(trans, root->fs_info->extent_root);
	del_pending_extents(trans, root->fs_info->extent_root);
	return 0;
}

int btrfs_lookup_extent_ref(struct btrfs_trans_handle *trans,
			    struct btrfs_root *root, u64 bytenr,
			    u64 num_bytes, u32 *refs)
{
	struct btrfs_path *path;
	int ret;
	struct btrfs_key key;
	struct extent_buffer *l;
	struct btrfs_extent_item *item;

	WARN_ON(num_bytes < root->sectorsize);
	path = btrfs_alloc_path();
	path->reada = 1;
	key.objectid = bytenr;
	key.offset = num_bytes;
	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
	ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, path,
				0, 0);
	if (ret < 0)
		goto out;
	if (ret != 0) {
		btrfs_print_leaf(root, path->nodes[0]);
		printk("failed to find block number %Lu\n", bytenr);
		BUG();
	}
	l = path->nodes[0];
	item = btrfs_item_ptr(l, path->slots[0], struct btrfs_extent_item);
	*refs = btrfs_extent_refs(l, item);
out:
	btrfs_free_path(path);
	return 0;
}

static int get_reference_status(struct btrfs_root *root, u64 bytenr,
				u64 parent_gen, u64 ref_objectid,
			        u64 *min_generation, u32 *ref_count)
{
	struct btrfs_root *extent_root = root->fs_info->extent_root;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_extent_ref *ref_item;
	struct btrfs_key key;
	struct btrfs_key found_key;
	u64 root_objectid = root->root_key.objectid;
	u64 ref_generation;
	u32 nritems;
	int ret;

	key.objectid = bytenr;
	key.offset = (u64)-1;
	key.type = BTRFS_EXTENT_ITEM_KEY;

	path = btrfs_alloc_path();
	mutex_lock(&root->fs_info->alloc_mutex);
	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
	BUG_ON(ret == 0);
	if (ret < 0 || path->slots[0] == 0)
		goto out;

	path->slots[0]--;
	leaf = path->nodes[0];
	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);

	if (found_key.objectid != bytenr ||
	    found_key.type != BTRFS_EXTENT_ITEM_KEY) {
		ret = 1;
		goto out;
	}

	*ref_count = 0;
	*min_generation = (u64)-1;

	while (1) {
		leaf = path->nodes[0];
		nritems = btrfs_header_nritems(leaf);
		if (path->slots[0] >= nritems) {
			ret = btrfs_next_leaf(extent_root, path);
			if (ret < 0)
				goto out;
			if (ret == 0)
				continue;
			break;
		}
		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
		if (found_key.objectid != bytenr)
			break;

		if (found_key.type != BTRFS_EXTENT_REF_KEY) {
			path->slots[0]++;
			continue;
		}

		ref_item = btrfs_item_ptr(leaf, path->slots[0],
					  struct btrfs_extent_ref);