linux/fs/btrfs/dir-item.c

436 lines
11 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (C) 2007 Oracle. All rights reserved.
*/
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
/*
* insert a name into a directory, doing overflow properly if there is a hash
* collision. data_size indicates how big the item inserted should be. On
* success a struct btrfs_dir_item pointer is returned, otherwise it is
* an ERR_PTR.
*
* The name is not copied into the dir item, you have to do that yourself.
*/
static struct btrfs_dir_item *insert_with_overflow(struct btrfs_trans_handle
*trans,
struct btrfs_root *root,
struct btrfs_path *path,
struct btrfs_key *cpu_key,
u32 data_size,
const char *name,
int name_len)
{
struct btrfs_fs_info *fs_info = root->fs_info;
int ret;
char *ptr;
struct btrfs_item *item;
struct extent_buffer *leaf;
ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size);
if (ret == -EEXIST) {
struct btrfs_dir_item *di;
di = btrfs_match_dir_item_name(fs_info, path, name, name_len);
if (di)
return ERR_PTR(-EEXIST);
btrfs_extend_item(path, data_size);
} else if (ret < 0)
return ERR_PTR(ret);
WARN_ON(ret > 0);
leaf = path->nodes[0];
item = btrfs_item_nr(path->slots[0]);
ptr = btrfs_item_ptr(leaf, path->slots[0], char);
BUG_ON(data_size > btrfs_item_size(leaf, item));
ptr += btrfs_item_size(leaf, item) - data_size;
return (struct btrfs_dir_item *)ptr;
}
/*
* xattrs work a lot like directories, this inserts an xattr item
* into the tree
*/
int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, u64 objectid,
const char *name, u16 name_len,
const void *data, u16 data_len)
{
int ret = 0;
struct btrfs_dir_item *dir_item;
unsigned long name_ptr, data_ptr;
struct btrfs_key key, location;
struct btrfs_disk_key disk_key;
struct extent_buffer *leaf;
u32 data_size;
if (name_len + data_len > BTRFS_MAX_XATTR_SIZE(root->fs_info))
return -ENOSPC;
key.objectid = objectid;
key.type = BTRFS_XATTR_ITEM_KEY;
key.offset = btrfs_name_hash(name, name_len);
data_size = sizeof(*dir_item) + name_len + data_len;
dir_item = insert_with_overflow(trans, root, path, &key, data_size,
name, name_len);
if (IS_ERR(dir_item))
return PTR_ERR(dir_item);
memset(&location, 0, sizeof(location));
leaf = path->nodes[0];
btrfs_cpu_key_to_disk(&disk_key, &location);
btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
btrfs_set_dir_type(leaf, dir_item, BTRFS_FT_XATTR);
btrfs_set_dir_name_len(leaf, dir_item, name_len);
btrfs_set_dir_transid(leaf, dir_item, trans->transid);
btrfs_set_dir_data_len(leaf, dir_item, data_len);
name_ptr = (unsigned long)(dir_item + 1);
data_ptr = (unsigned long)((char *)name_ptr + name_len);
write_extent_buffer(leaf, name, name_ptr, name_len);
write_extent_buffer(leaf, data, data_ptr, data_len);
btrfs_mark_buffer_dirty(path->nodes[0]);
return ret;
}
/*
* insert a directory item in the tree, doing all the magic for
* both indexes. 'dir' indicates which objectid to insert it into,
* 'location' is the key to stuff into the directory item, 'type' is the
* type of the inode we're pointing to, and 'index' is the sequence number
* to use for the second index (if one is created).
* Will return 0 or -ENOMEM
*/
int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, const char *name,
int name_len, struct btrfs_inode *dir,
struct btrfs_key *location, u8 type, u64 index)
{
int ret = 0;
int ret2 = 0;
struct btrfs_root *root = dir->root;
struct btrfs_path *path;
struct btrfs_dir_item *dir_item;
struct extent_buffer *leaf;
unsigned long name_ptr;
struct btrfs_key key;
struct btrfs_disk_key disk_key;
u32 data_size;
key.objectid = btrfs_ino(dir);
key.type = BTRFS_DIR_ITEM_KEY;
key.offset = btrfs_name_hash(name, name_len);
path = btrfs_alloc_path();
btrfs: implement delayed inode items operation Changelog V5 -> V6: - Fix oom when the memory load is high, by storing the delayed nodes into the root's radix tree, and letting btrfs inodes go. Changelog V4 -> V5: - Fix the race on adding the delayed node to the inode, which is spotted by Chris Mason. - Merge Chris Mason's incremental patch into this patch. - Fix deadlock between readdir() and memory fault, which is reported by Itaru Kitayama. Changelog V3 -> V4: - Fix nested lock, which is reported by Itaru Kitayama, by updating space cache inode in time. Changelog V2 -> V3: - Fix the race between the delayed worker and the task which does delayed items balance, which is reported by Tsutomu Itoh. - Modify the patch address David Sterba's comment. - Fix the bug of the cpu recursion spinlock, reported by Chris Mason Changelog V1 -> V2: - break up the global rb-tree, use a list to manage the delayed nodes, which is created for every directory and file, and used to manage the delayed directory name index items and the delayed inode item. - introduce a worker to deal with the delayed nodes. Compare with Ext3/4, the performance of file creation and deletion on btrfs is very poor. the reason is that btrfs must do a lot of b+ tree insertions, such as inode item, directory name item, directory name index and so on. If we can do some delayed b+ tree insertion or deletion, we can improve the performance, so we made this patch which implemented delayed directory name index insertion/deletion and delayed inode update. Implementation: - introduce a delayed root object into the filesystem, that use two lists to manage the delayed nodes which are created for every file/directory. One is used to manage all the delayed nodes that have delayed items. And the other is used to manage the delayed nodes which is waiting to be dealt with by the work thread. - Every delayed node has two rb-tree, one is used to manage the directory name index which is going to be inserted into b+ tree, and the other is used to manage the directory name index which is going to be deleted from b+ tree. - introduce a worker to deal with the delayed operation. This worker is used to deal with the works of the delayed directory name index items insertion and deletion and the delayed inode update. When the delayed items is beyond the lower limit, we create works for some delayed nodes and insert them into the work queue of the worker, and then go back. When the delayed items is beyond the upper bound, we create works for all the delayed nodes that haven't been dealt with, and insert them into the work queue of the worker, and then wait for that the untreated items is below some threshold value. - When we want to insert a directory name index into b+ tree, we just add the information into the delayed inserting rb-tree. And then we check the number of the delayed items and do delayed items balance. (The balance policy is above.) - When we want to delete a directory name index from the b+ tree, we search it in the inserting rb-tree at first. If we look it up, just drop it. If not, add the key of it into the delayed deleting rb-tree. Similar to the delayed inserting rb-tree, we also check the number of the delayed items and do delayed items balance. (The same to inserting manipulation) - When we want to update the metadata of some inode, we cached the data of the inode into the delayed node. the worker will flush it into the b+ tree after dealing with the delayed insertion and deletion. - We will move the delayed node to the tail of the list after we access the delayed node, By this way, we can cache more delayed items and merge more inode updates. - If we want to commit transaction, we will deal with all the delayed node. - the delayed node will be freed when we free the btrfs inode. - Before we log the inode items, we commit all the directory name index items and the delayed inode update. I did a quick test by the benchmark tool[1] and found we can improve the performance of file creation by ~15%, and file deletion by ~20%. Before applying this patch: Create files: Total files: 50000 Total time: 1.096108 Average time: 0.000022 Delete files: Total files: 50000 Total time: 1.510403 Average time: 0.000030 After applying this patch: Create files: Total files: 50000 Total time: 0.932899 Average time: 0.000019 Delete files: Total files: 50000 Total time: 1.215732 Average time: 0.000024 [1] http://marc.info/?l=linux-btrfs&m=128212635122920&q=p3 Many thanks for Kitayama-san's help! Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Reviewed-by: David Sterba <dave@jikos.cz> Tested-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com> Tested-by: Itaru Kitayama <kitayama@cl.bb4u.ne.jp> Signed-off-by: Chris Mason <chris.mason@oracle.com>
2011-04-22 10:12:22 +00:00
if (!path)
return -ENOMEM;
path->leave_spinning = 1;
btrfs: implement delayed inode items operation Changelog V5 -> V6: - Fix oom when the memory load is high, by storing the delayed nodes into the root's radix tree, and letting btrfs inodes go. Changelog V4 -> V5: - Fix the race on adding the delayed node to the inode, which is spotted by Chris Mason. - Merge Chris Mason's incremental patch into this patch. - Fix deadlock between readdir() and memory fault, which is reported by Itaru Kitayama. Changelog V3 -> V4: - Fix nested lock, which is reported by Itaru Kitayama, by updating space cache inode in time. Changelog V2 -> V3: - Fix the race between the delayed worker and the task which does delayed items balance, which is reported by Tsutomu Itoh. - Modify the patch address David Sterba's comment. - Fix the bug of the cpu recursion spinlock, reported by Chris Mason Changelog V1 -> V2: - break up the global rb-tree, use a list to manage the delayed nodes, which is created for every directory and file, and used to manage the delayed directory name index items and the delayed inode item. - introduce a worker to deal with the delayed nodes. Compare with Ext3/4, the performance of file creation and deletion on btrfs is very poor. the reason is that btrfs must do a lot of b+ tree insertions, such as inode item, directory name item, directory name index and so on. If we can do some delayed b+ tree insertion or deletion, we can improve the performance, so we made this patch which implemented delayed directory name index insertion/deletion and delayed inode update. Implementation: - introduce a delayed root object into the filesystem, that use two lists to manage the delayed nodes which are created for every file/directory. One is used to manage all the delayed nodes that have delayed items. And the other is used to manage the delayed nodes which is waiting to be dealt with by the work thread. - Every delayed node has two rb-tree, one is used to manage the directory name index which is going to be inserted into b+ tree, and the other is used to manage the directory name index which is going to be deleted from b+ tree. - introduce a worker to deal with the delayed operation. This worker is used to deal with the works of the delayed directory name index items insertion and deletion and the delayed inode update. When the delayed items is beyond the lower limit, we create works for some delayed nodes and insert them into the work queue of the worker, and then go back. When the delayed items is beyond the upper bound, we create works for all the delayed nodes that haven't been dealt with, and insert them into the work queue of the worker, and then wait for that the untreated items is below some threshold value. - When we want to insert a directory name index into b+ tree, we just add the information into the delayed inserting rb-tree. And then we check the number of the delayed items and do delayed items balance. (The balance policy is above.) - When we want to delete a directory name index from the b+ tree, we search it in the inserting rb-tree at first. If we look it up, just drop it. If not, add the key of it into the delayed deleting rb-tree. Similar to the delayed inserting rb-tree, we also check the number of the delayed items and do delayed items balance. (The same to inserting manipulation) - When we want to update the metadata of some inode, we cached the data of the inode into the delayed node. the worker will flush it into the b+ tree after dealing with the delayed insertion and deletion. - We will move the delayed node to the tail of the list after we access the delayed node, By this way, we can cache more delayed items and merge more inode updates. - If we want to commit transaction, we will deal with all the delayed node. - the delayed node will be freed when we free the btrfs inode. - Before we log the inode items, we commit all the directory name index items and the delayed inode update. I did a quick test by the benchmark tool[1] and found we can improve the performance of file creation by ~15%, and file deletion by ~20%. Before applying this patch: Create files: Total files: 50000 Total time: 1.096108 Average time: 0.000022 Delete files: Total files: 50000 Total time: 1.510403 Average time: 0.000030 After applying this patch: Create files: Total files: 50000 Total time: 0.932899 Average time: 0.000019 Delete files: Total files: 50000 Total time: 1.215732 Average time: 0.000024 [1] http://marc.info/?l=linux-btrfs&m=128212635122920&q=p3 Many thanks for Kitayama-san's help! Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Reviewed-by: David Sterba <dave@jikos.cz> Tested-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com> Tested-by: Itaru Kitayama <kitayama@cl.bb4u.ne.jp> Signed-off-by: Chris Mason <chris.mason@oracle.com>
2011-04-22 10:12:22 +00:00
btrfs_cpu_key_to_disk(&disk_key, location);
data_size = sizeof(*dir_item) + name_len;
dir_item = insert_with_overflow(trans, root, path, &key, data_size,
name, name_len);
if (IS_ERR(dir_item)) {
ret = PTR_ERR(dir_item);
if (ret == -EEXIST)
goto second_insert;
goto out_free;
}
leaf = path->nodes[0];
btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
btrfs_set_dir_type(leaf, dir_item, type);
btrfs_set_dir_data_len(leaf, dir_item, 0);
btrfs_set_dir_name_len(leaf, dir_item, name_len);
btrfs_set_dir_transid(leaf, dir_item, trans->transid);
name_ptr = (unsigned long)(dir_item + 1);
write_extent_buffer(leaf, name, name_ptr, name_len);
btrfs_mark_buffer_dirty(leaf);
second_insert:
/* FIXME, use some real flag for selecting the extra index */
if (root == root->fs_info->tree_root) {
ret = 0;
goto out_free;
}
btrfs_release_path(path);
ret2 = btrfs_insert_delayed_dir_index(trans, name, name_len, dir,
&disk_key, type, index);
out_free:
btrfs_free_path(path);
if (ret)
return ret;
if (ret2)
return ret2;
return 0;
}
/*
* lookup a directory item based on name. 'dir' is the objectid
* we're searching in, and 'mod' tells us if you plan on deleting the
* item (use mod < 0) or changing the options (use mod > 0)
*/
struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, u64 dir,
const char *name, int name_len,
int mod)
{
int ret;
struct btrfs_key key;
int ins_len = mod < 0 ? -1 : 0;
int cow = mod != 0;
key.objectid = dir;
key.type = BTRFS_DIR_ITEM_KEY;
key.offset = btrfs_name_hash(name, name_len);
ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
if (ret < 0)
return ERR_PTR(ret);
if (ret > 0)
return NULL;
return btrfs_match_dir_item_name(root->fs_info, path, name, name_len);
}
int btrfs_check_dir_item_collision(struct btrfs_root *root, u64 dir,
const char *name, int name_len)
{
int ret;
struct btrfs_key key;
struct btrfs_dir_item *di;
int data_size;
struct extent_buffer *leaf;
int slot;
struct btrfs_path *path;
path = btrfs_alloc_path();
if (!path)
return -ENOMEM;
key.objectid = dir;
key.type = BTRFS_DIR_ITEM_KEY;
key.offset = btrfs_name_hash(name, name_len);
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
/* return back any errors */
if (ret < 0)
goto out;
/* nothing found, we're safe */
if (ret > 0) {
ret = 0;
goto out;
}
/* we found an item, look for our name in the item */
di = btrfs_match_dir_item_name(root->fs_info, path, name, name_len);
if (di) {
/* our exact name was found */
ret = -EEXIST;
goto out;
}
/*
* see if there is room in the item to insert this
* name
*/
data_size = sizeof(*di) + name_len;
leaf = path->nodes[0];
slot = path->slots[0];
if (data_size + btrfs_item_size_nr(leaf, slot) +
sizeof(struct btrfs_item) > BTRFS_LEAF_DATA_SIZE(root->fs_info)) {
ret = -EOVERFLOW;
} else {
/* plenty of insertion room */
ret = 0;
}
out:
btrfs_free_path(path);
return ret;
}
/*
* lookup a directory item based on index. 'dir' is the objectid
* we're searching in, and 'mod' tells us if you plan on deleting the
* item (use mod < 0) or changing the options (use mod > 0)
*
* The name is used to make sure the index really points to the name you were
* looking for.
*/
struct btrfs_dir_item *
btrfs_lookup_dir_index_item(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, u64 dir,
u64 objectid, const char *name, int name_len,
int mod)
{
int ret;
struct btrfs_key key;
int ins_len = mod < 0 ? -1 : 0;
int cow = mod != 0;
key.objectid = dir;
key.type = BTRFS_DIR_INDEX_KEY;
key.offset = objectid;
ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
if (ret < 0)
return ERR_PTR(ret);
if (ret > 0)
return ERR_PTR(-ENOENT);
return btrfs_match_dir_item_name(root->fs_info, path, name, name_len);
}
struct btrfs_dir_item *
btrfs_search_dir_index_item(struct btrfs_root *root,
struct btrfs_path *path, u64 dirid,
const char *name, int name_len)
{
struct extent_buffer *leaf;
struct btrfs_dir_item *di;
struct btrfs_key key;
u32 nritems;
int ret;
key.objectid = dirid;
key.type = BTRFS_DIR_INDEX_KEY;
key.offset = 0;
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
if (ret < 0)
return ERR_PTR(ret);
leaf = path->nodes[0];
nritems = btrfs_header_nritems(leaf);
while (1) {
if (path->slots[0] >= nritems) {
ret = btrfs_next_leaf(root, path);
if (ret < 0)
return ERR_PTR(ret);
if (ret > 0)
break;
leaf = path->nodes[0];
nritems = btrfs_header_nritems(leaf);
continue;
}
btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
if (key.objectid != dirid || key.type != BTRFS_DIR_INDEX_KEY)
break;
di = btrfs_match_dir_item_name(root->fs_info, path,
name, name_len);
if (di)
return di;
path->slots[0]++;
}
return NULL;
}
struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, u64 dir,
const char *name, u16 name_len,
int mod)
{
int ret;
struct btrfs_key key;
int ins_len = mod < 0 ? -1 : 0;
int cow = mod != 0;
key.objectid = dir;
key.type = BTRFS_XATTR_ITEM_KEY;
key.offset = btrfs_name_hash(name, name_len);
ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
if (ret < 0)
return ERR_PTR(ret);
if (ret > 0)
return NULL;
return btrfs_match_dir_item_name(root->fs_info, path, name, name_len);
}
/*
* helper function to look at the directory item pointed to by 'path'
* this walks through all the entries in a dir item and finds one
* for a specific name.
*/
struct btrfs_dir_item *btrfs_match_dir_item_name(struct btrfs_fs_info *fs_info,
struct btrfs_path *path,
const char *name, int name_len)
{
struct btrfs_dir_item *dir_item;
unsigned long name_ptr;
u32 total_len;
u32 cur = 0;
u32 this_len;
struct extent_buffer *leaf;
leaf = path->nodes[0];
dir_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dir_item);
total_len = btrfs_item_size_nr(leaf, path->slots[0]);
while (cur < total_len) {
this_len = sizeof(*dir_item) +
btrfs_dir_name_len(leaf, dir_item) +
btrfs_dir_data_len(leaf, dir_item);
name_ptr = (unsigned long)(dir_item + 1);
if (btrfs_dir_name_len(leaf, dir_item) == name_len &&
memcmp_extent_buffer(leaf, name, name_ptr, name_len) == 0)
return dir_item;
cur += this_len;
dir_item = (struct btrfs_dir_item *)((char *)dir_item +
this_len);
}
return NULL;
}
/*
* given a pointer into a directory item, delete it. This
* handles items that have more than one entry in them.
*/
int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path,
struct btrfs_dir_item *di)
{
struct extent_buffer *leaf;
u32 sub_item_len;
u32 item_len;
int ret = 0;
leaf = path->nodes[0];
sub_item_len = sizeof(*di) + btrfs_dir_name_len(leaf, di) +
btrfs_dir_data_len(leaf, di);
item_len = btrfs_item_size_nr(leaf, path->slots[0]);
if (sub_item_len == item_len) {
ret = btrfs_del_item(trans, root, path);
} else {
/* MARKER */
unsigned long ptr = (unsigned long)di;
unsigned long start;
start = btrfs_item_ptr_offset(leaf, path->slots[0]);
memmove_extent_buffer(leaf, ptr, ptr + sub_item_len,
item_len - (ptr + sub_item_len - start));
btrfs_truncate_item(path, item_len - sub_item_len, 1);
}
return ret;
}