2007-06-12 21:07:21 +08:00
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
|
|
|
|
2007-03-14 04:47:54 +08:00
|
|
|
#include "ctree.h"
|
2007-06-23 02:16:25 +08:00
|
|
|
#include "transaction.h"
|
2007-03-14 04:47:54 +08:00
|
|
|
#include "disk-io.h"
|
|
|
|
#include "print-tree.h"
|
|
|
|
|
2008-05-09 01:26:18 +08:00
|
|
|
/*
|
2008-09-30 03:18:18 +08:00
|
|
|
* search forward for a root, starting with objectid 'search_start'
|
|
|
|
* if a root key is found, the objectid we find is filled into 'found_objectid'
|
|
|
|
* and 0 is returned. < 0 is returned on error, 1 if there is nothing
|
|
|
|
* left in the tree.
|
2008-05-09 01:26:18 +08:00
|
|
|
*/
|
|
|
|
int btrfs_search_root(struct btrfs_root *root, u64 search_start,
|
|
|
|
u64 *found_objectid)
|
|
|
|
{
|
|
|
|
struct btrfs_path *path;
|
|
|
|
struct btrfs_key search_key;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
root = root->fs_info->tree_root;
|
|
|
|
search_key.objectid = search_start;
|
|
|
|
search_key.type = (u8)-1;
|
|
|
|
search_key.offset = (u64)-1;
|
|
|
|
|
|
|
|
path = btrfs_alloc_path();
|
|
|
|
BUG_ON(!path);
|
|
|
|
again:
|
|
|
|
ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
|
|
|
|
if (ret < 0)
|
|
|
|
goto out;
|
|
|
|
if (ret == 0) {
|
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
|
|
|
|
ret = btrfs_next_leaf(root, path);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
btrfs_item_key_to_cpu(path->nodes[0], &search_key, path->slots[0]);
|
|
|
|
if (search_key.type != BTRFS_ROOT_ITEM_KEY) {
|
|
|
|
search_key.offset++;
|
|
|
|
btrfs_release_path(root, path);
|
|
|
|
goto again;
|
|
|
|
}
|
|
|
|
ret = 0;
|
|
|
|
*found_objectid = search_key.objectid;
|
|
|
|
|
|
|
|
out:
|
|
|
|
btrfs_free_path(path);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-09-30 03:18:18 +08:00
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
2007-03-14 04:47:54 +08:00
|
|
|
int btrfs_find_last_root(struct btrfs_root *root, u64 objectid,
|
|
|
|
struct btrfs_root_item *item, struct btrfs_key *key)
|
|
|
|
{
|
2007-04-02 23:20:42 +08:00
|
|
|
struct btrfs_path *path;
|
2007-03-14 04:47:54 +08:00
|
|
|
struct btrfs_key search_key;
|
2007-10-16 04:14:19 +08:00
|
|
|
struct btrfs_key found_key;
|
|
|
|
struct extent_buffer *l;
|
2007-03-14 04:47:54 +08:00
|
|
|
int ret;
|
|
|
|
int slot;
|
|
|
|
|
|
|
|
search_key.objectid = objectid;
|
2008-11-18 09:37:39 +08:00
|
|
|
search_key.type = BTRFS_ROOT_ITEM_KEY;
|
2007-06-23 02:16:25 +08:00
|
|
|
search_key.offset = (u64)-1;
|
2007-03-14 04:47:54 +08:00
|
|
|
|
2007-04-02 23:20:42 +08:00
|
|
|
path = btrfs_alloc_path();
|
|
|
|
BUG_ON(!path);
|
|
|
|
ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
|
2007-03-14 04:47:54 +08:00
|
|
|
if (ret < 0)
|
|
|
|
goto out;
|
2007-10-16 04:14:19 +08:00
|
|
|
|
2007-03-14 04:47:54 +08:00
|
|
|
BUG_ON(ret == 0);
|
2007-10-16 04:14:19 +08:00
|
|
|
l = path->nodes[0];
|
2007-04-02 23:20:42 +08:00
|
|
|
BUG_ON(path->slots[0] == 0);
|
|
|
|
slot = path->slots[0] - 1;
|
2007-10-16 04:14:19 +08:00
|
|
|
btrfs_item_key_to_cpu(l, &found_key, slot);
|
|
|
|
if (found_key.objectid != objectid) {
|
2007-03-14 04:47:54 +08:00
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
2007-10-16 04:14:19 +08:00
|
|
|
read_extent_buffer(l, item, btrfs_item_ptr_offset(l, slot),
|
|
|
|
sizeof(*item));
|
|
|
|
memcpy(key, &found_key, sizeof(found_key));
|
2007-03-14 04:47:54 +08:00
|
|
|
ret = 0;
|
|
|
|
out:
|
2007-04-02 23:20:42 +08:00
|
|
|
btrfs_free_path(path);
|
2007-03-14 04:47:54 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-09-30 03:18:18 +08:00
|
|
|
/*
|
|
|
|
* copy the data in 'item' into the btree
|
|
|
|
*/
|
2007-03-17 04:20:31 +08:00
|
|
|
int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
|
|
|
|
*root, struct btrfs_key *key, struct btrfs_root_item
|
|
|
|
*item)
|
2007-03-14 04:47:54 +08:00
|
|
|
{
|
2007-04-02 23:20:42 +08:00
|
|
|
struct btrfs_path *path;
|
2007-10-16 04:14:19 +08:00
|
|
|
struct extent_buffer *l;
|
2007-03-14 04:47:54 +08:00
|
|
|
int ret;
|
|
|
|
int slot;
|
2007-10-16 04:14:19 +08:00
|
|
|
unsigned long ptr;
|
2007-03-14 04:47:54 +08:00
|
|
|
|
2007-04-02 23:20:42 +08:00
|
|
|
path = btrfs_alloc_path();
|
|
|
|
BUG_ON(!path);
|
|
|
|
ret = btrfs_search_slot(trans, root, key, path, 0, 1);
|
2007-03-14 04:47:54 +08:00
|
|
|
if (ret < 0)
|
|
|
|
goto out;
|
2008-01-04 03:51:00 +08:00
|
|
|
|
|
|
|
if (ret != 0) {
|
|
|
|
btrfs_print_leaf(root, path->nodes[0]);
|
2009-01-06 10:25:51 +08:00
|
|
|
printk(KERN_CRIT "unable to update root key %llu %u %llu\n",
|
|
|
|
(unsigned long long)key->objectid, key->type,
|
|
|
|
(unsigned long long)key->offset);
|
2008-01-04 03:51:00 +08:00
|
|
|
BUG_ON(1);
|
|
|
|
}
|
|
|
|
|
2007-10-16 04:14:19 +08:00
|
|
|
l = path->nodes[0];
|
2007-04-02 23:20:42 +08:00
|
|
|
slot = path->slots[0];
|
2007-10-16 04:14:19 +08:00
|
|
|
ptr = btrfs_item_ptr_offset(l, slot);
|
|
|
|
write_extent_buffer(l, item, ptr, sizeof(*item));
|
2007-04-02 23:20:42 +08:00
|
|
|
btrfs_mark_buffer_dirty(path->nodes[0]);
|
2007-03-14 04:47:54 +08:00
|
|
|
out:
|
2007-04-02 23:20:42 +08:00
|
|
|
btrfs_release_path(root, path);
|
|
|
|
btrfs_free_path(path);
|
2007-03-14 04:47:54 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2007-03-17 04:20:31 +08:00
|
|
|
int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root
|
|
|
|
*root, struct btrfs_key *key, struct btrfs_root_item
|
|
|
|
*item)
|
2007-03-14 04:47:54 +08:00
|
|
|
{
|
|
|
|
int ret;
|
2007-03-17 04:20:31 +08:00
|
|
|
ret = btrfs_insert_item(trans, root, key, item, sizeof(*item));
|
2007-03-14 04:47:54 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-09-30 03:18:18 +08:00
|
|
|
/*
|
|
|
|
* at mount time we want to find all the old transaction snapshots that were in
|
2009-01-06 10:25:51 +08:00
|
|
|
* 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.
|
2008-09-30 03:18:18 +08:00
|
|
|
*/
|
2007-09-11 23:15:39 +08:00
|
|
|
int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid,
|
|
|
|
struct btrfs_root *latest)
|
2007-06-23 02:16:25 +08:00
|
|
|
{
|
|
|
|
struct btrfs_root *dead_root;
|
|
|
|
struct btrfs_item *item;
|
|
|
|
struct btrfs_root_item *ri;
|
|
|
|
struct btrfs_key key;
|
2008-06-26 22:34:20 +08:00
|
|
|
struct btrfs_key found_key;
|
2007-06-23 02:16:25 +08:00
|
|
|
struct btrfs_path *path;
|
|
|
|
int ret;
|
|
|
|
u32 nritems;
|
2007-10-16 04:14:19 +08:00
|
|
|
struct extent_buffer *leaf;
|
2007-06-23 02:16:25 +08:00
|
|
|
int slot;
|
|
|
|
|
2007-09-11 23:15:39 +08:00
|
|
|
key.objectid = objectid;
|
2007-06-23 02:16:25 +08:00
|
|
|
btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
|
|
|
|
key.offset = 0;
|
|
|
|
path = btrfs_alloc_path();
|
|
|
|
if (!path)
|
|
|
|
return -ENOMEM;
|
2008-06-26 22:34:20 +08:00
|
|
|
|
|
|
|
again:
|
2007-06-23 02:16:25 +08:00
|
|
|
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
|
|
|
|
if (ret < 0)
|
|
|
|
goto err;
|
2009-01-06 10:25:51 +08:00
|
|
|
while (1) {
|
2007-10-16 04:14:19 +08:00
|
|
|
leaf = path->nodes[0];
|
|
|
|
nritems = btrfs_header_nritems(leaf);
|
2007-06-23 02:16:25 +08:00
|
|
|
slot = path->slots[0];
|
|
|
|
if (slot >= nritems) {
|
|
|
|
ret = btrfs_next_leaf(root, path);
|
|
|
|
if (ret)
|
|
|
|
break;
|
2007-10-16 04:14:19 +08:00
|
|
|
leaf = path->nodes[0];
|
|
|
|
nritems = btrfs_header_nritems(leaf);
|
2007-06-23 02:16:25 +08:00
|
|
|
slot = path->slots[0];
|
|
|
|
}
|
2007-10-16 04:14:19 +08:00
|
|
|
item = btrfs_item_nr(leaf, slot);
|
|
|
|
btrfs_item_key_to_cpu(leaf, &key, slot);
|
2007-06-23 02:16:25 +08:00
|
|
|
if (btrfs_key_type(&key) != BTRFS_ROOT_ITEM_KEY)
|
|
|
|
goto next;
|
2007-09-11 23:15:39 +08:00
|
|
|
|
|
|
|
if (key.objectid < objectid)
|
|
|
|
goto next;
|
|
|
|
|
|
|
|
if (key.objectid > objectid)
|
|
|
|
break;
|
|
|
|
|
2007-06-23 02:16:25 +08:00
|
|
|
ri = btrfs_item_ptr(leaf, slot, struct btrfs_root_item);
|
2007-10-16 04:14:19 +08:00
|
|
|
if (btrfs_disk_root_refs(leaf, ri) != 0)
|
2007-06-23 02:16:25 +08:00
|
|
|
goto next;
|
2007-09-11 23:15:39 +08:00
|
|
|
|
2008-06-26 22:34:20 +08:00
|
|
|
memcpy(&found_key, &key, sizeof(key));
|
|
|
|
key.offset++;
|
|
|
|
btrfs_release_path(root, path);
|
2008-09-06 04:13:11 +08:00
|
|
|
dead_root =
|
|
|
|
btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
|
|
|
|
&found_key);
|
2007-07-11 22:03:27 +08:00
|
|
|
if (IS_ERR(dead_root)) {
|
|
|
|
ret = PTR_ERR(dead_root);
|
2007-06-23 02:16:25 +08:00
|
|
|
goto err;
|
|
|
|
}
|
2007-09-11 23:15:39 +08:00
|
|
|
|
Btrfs: update space balancing code
This patch updates the space balancing code to utilize the new
backref format. Before, btrfs-vol -b would break any COW links
on data blocks or metadata. This was slow and caused the amount
of space used to explode if a large number of snapshots were present.
The new code can keeps the sharing of all data extents and
most of the tree blocks.
To maintain the sharing of data extents, the space balance code uses
a seperate inode hold data extent pointers, then updates the references
to point to the new location.
To maintain the sharing of tree blocks, the space balance code uses
reloc trees to relocate tree blocks in reference counted roots.
There is one reloc tree for each subvol, and all reloc trees share
same root key objectid. Reloc trees are snapshots of the latest
committed roots of subvols (root->commit_root).
To relocate a tree block referenced by a subvol, there are two steps.
COW the block through subvol's reloc tree, then update block pointer in
the subvol to point to the new block. Since all reloc trees share
same root key objectid, doing special handing for tree blocks
owned by them is easy. Once a tree block has been COWed in one
reloc tree, we can use the resulting new block directly when the
same block is required to COW again through other reloc trees.
In this way, relocated tree blocks are shared between reloc trees,
so they are also shared between subvols.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-26 22:09:34 +08:00
|
|
|
if (objectid == BTRFS_TREE_RELOC_OBJECTID)
|
|
|
|
ret = btrfs_add_dead_reloc_root(dead_root);
|
|
|
|
else
|
|
|
|
ret = btrfs_add_dead_root(dead_root, latest);
|
2007-06-23 02:16:25 +08:00
|
|
|
if (ret)
|
|
|
|
goto err;
|
2008-06-26 22:34:20 +08:00
|
|
|
goto again;
|
2007-06-23 02:16:25 +08:00
|
|
|
next:
|
|
|
|
slot++;
|
|
|
|
path->slots[0]++;
|
|
|
|
}
|
|
|
|
ret = 0;
|
|
|
|
err:
|
|
|
|
btrfs_free_path(path);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-09-30 03:18:18 +08:00
|
|
|
/* drop the root item for 'key' from 'root' */
|
2007-03-17 04:20:31 +08:00
|
|
|
int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
|
|
|
|
struct btrfs_key *key)
|
2007-03-14 04:47:54 +08:00
|
|
|
{
|
2007-04-02 23:20:42 +08:00
|
|
|
struct btrfs_path *path;
|
2007-03-14 04:47:54 +08:00
|
|
|
int ret;
|
2007-04-10 21:27:04 +08:00
|
|
|
u32 refs;
|
|
|
|
struct btrfs_root_item *ri;
|
2007-10-16 04:14:19 +08:00
|
|
|
struct extent_buffer *leaf;
|
2007-03-14 04:47:54 +08:00
|
|
|
|
2007-04-02 23:20:42 +08:00
|
|
|
path = btrfs_alloc_path();
|
|
|
|
BUG_ON(!path);
|
|
|
|
ret = btrfs_search_slot(trans, root, key, path, -1, 1);
|
2007-03-14 04:47:54 +08:00
|
|
|
if (ret < 0)
|
|
|
|
goto out;
|
2007-12-22 05:27:24 +08:00
|
|
|
|
2007-03-14 04:47:54 +08:00
|
|
|
BUG_ON(ret != 0);
|
2007-10-16 04:14:19 +08:00
|
|
|
leaf = path->nodes[0];
|
|
|
|
ri = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_item);
|
2007-04-10 21:27:04 +08:00
|
|
|
|
2007-10-16 04:14:19 +08:00
|
|
|
refs = btrfs_disk_root_refs(leaf, ri);
|
2007-06-23 02:16:25 +08:00
|
|
|
BUG_ON(refs != 0);
|
|
|
|
ret = btrfs_del_item(trans, root, path);
|
2007-03-14 04:47:54 +08:00
|
|
|
out:
|
2007-04-02 23:20:42 +08:00
|
|
|
btrfs_release_path(root, path);
|
|
|
|
btrfs_free_path(path);
|
2007-03-14 04:47:54 +08:00
|
|
|
return ret;
|
|
|
|
}
|
2008-11-18 09:37:39 +08:00
|
|
|
|
2008-12-02 22:54:17 +08:00
|
|
|
#if 0 /* this will get used when snapshot deletion is implemented */
|
2008-11-18 09:37:39 +08:00
|
|
|
int btrfs_del_root_ref(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *tree_root,
|
|
|
|
u64 root_id, u8 type, u64 ref_id)
|
|
|
|
{
|
|
|
|
struct btrfs_key key;
|
|
|
|
int ret;
|
|
|
|
struct btrfs_path *path;
|
|
|
|
|
|
|
|
path = btrfs_alloc_path();
|
|
|
|
|
|
|
|
key.objectid = root_id;
|
|
|
|
key.type = type;
|
|
|
|
key.offset = ref_id;
|
|
|
|
|
|
|
|
ret = btrfs_search_slot(trans, tree_root, &key, path, -1, 1);
|
|
|
|
BUG_ON(ret);
|
|
|
|
|
|
|
|
ret = btrfs_del_item(trans, tree_root, path);
|
|
|
|
BUG_ON(ret);
|
|
|
|
|
|
|
|
btrfs_free_path(path);
|
|
|
|
return ret;
|
|
|
|
}
|
2008-12-02 22:54:17 +08:00
|
|
|
#endif
|
2008-11-18 09:37:39 +08:00
|
|
|
|
2008-11-18 10:14:24 +08:00
|
|
|
int btrfs_find_root_ref(struct btrfs_root *tree_root,
|
|
|
|
struct btrfs_path *path,
|
|
|
|
u64 root_id, u64 ref_id)
|
|
|
|
{
|
|
|
|
struct btrfs_key key;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
key.objectid = root_id;
|
|
|
|
key.type = BTRFS_ROOT_REF_KEY;
|
|
|
|
key.offset = ref_id;
|
|
|
|
|
|
|
|
ret = btrfs_search_slot(NULL, tree_root, &key, path, 0, 0);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2008-11-18 09:37:39 +08:00
|
|
|
/*
|
|
|
|
* add a btrfs_root_ref item. type is either BTRFS_ROOT_REF_KEY
|
|
|
|
* or BTRFS_ROOT_BACKREF_KEY.
|
|
|
|
*
|
|
|
|
* The dirid, sequence, name and name_len refer to the directory entry
|
|
|
|
* that is referencing the root.
|
|
|
|
*
|
|
|
|
* For a forward ref, the root_id is the id of the tree referencing
|
|
|
|
* the root and ref_id is the id of the subvol or snapshot.
|
|
|
|
*
|
|
|
|
* For a back ref the root_id is the id of the subvol or snapshot and
|
|
|
|
* ref_id is the id of the tree referencing it.
|
|
|
|
*/
|
|
|
|
int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *tree_root,
|
|
|
|
u64 root_id, u8 type, u64 ref_id,
|
|
|
|
u64 dirid, u64 sequence,
|
|
|
|
const char *name, int name_len)
|
|
|
|
{
|
|
|
|
struct btrfs_key key;
|
|
|
|
int ret;
|
|
|
|
struct btrfs_path *path;
|
|
|
|
struct btrfs_root_ref *ref;
|
|
|
|
struct extent_buffer *leaf;
|
|
|
|
unsigned long ptr;
|
|
|
|
|
|
|
|
|
|
|
|
path = btrfs_alloc_path();
|
|
|
|
|
|
|
|
key.objectid = root_id;
|
|
|
|
key.type = type;
|
|
|
|
key.offset = ref_id;
|
|
|
|
|
|
|
|
ret = btrfs_insert_empty_item(trans, tree_root, path, &key,
|
|
|
|
sizeof(*ref) + name_len);
|
|
|
|
BUG_ON(ret);
|
|
|
|
|
|
|
|
leaf = path->nodes[0];
|
|
|
|
ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
|
|
|
|
btrfs_set_root_ref_dirid(leaf, ref, dirid);
|
|
|
|
btrfs_set_root_ref_sequence(leaf, ref, sequence);
|
|
|
|
btrfs_set_root_ref_name_len(leaf, ref, name_len);
|
|
|
|
ptr = (unsigned long)(ref + 1);
|
|
|
|
write_extent_buffer(leaf, name, ptr, name_len);
|
|
|
|
btrfs_mark_buffer_dirty(leaf);
|
|
|
|
|
|
|
|
btrfs_free_path(path);
|
|
|
|
return ret;
|
|
|
|
}
|