int __btrfs_setxattr(struct btrfs_trans_handle *trans,
struct inode *inode, const char *name,
const void *value, size_t size, int flags)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
int ret;
if (trans)
return do_setxattr(trans, inode, name, value, size, flags);
trans = btrfs_start_transaction(root, 2);
if (IS_ERR(trans))
return PTR_ERR(trans);
ret = do_setxattr(trans, inode, name, value, size, flags);
if (ret)
goto out;
inode->i_ctime = CURRENT_TIME;
ret = btrfs_update_inode(trans, root, inode);
BUG_ON(ret);
out:
btrfs_end_transaction(trans, root);
return ret;
}
/*
* after copy_from_user, pages need to be dirtied and we need to make
* sure holes are created between the current EOF and the start of
* any next extents (if required).
*
* this also makes the decision about creating an inline extent vs
* doing real data extents, marking pages dirty and delalloc as required.
*/
static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct file *file,
struct page **pages,
size_t num_pages,
loff_t pos,
size_t write_bytes)
{
int err = 0;
int i;
struct inode *inode = fdentry(file)->d_inode;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
u64 hint_byte;
u64 num_bytes;
u64 start_pos;
u64 end_of_last_block;
u64 end_pos = pos + write_bytes;
loff_t isize = i_size_read(inode);
start_pos = pos & ~((u64)root->sectorsize - 1);
num_bytes = (write_bytes + pos - start_pos +
root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
end_of_last_block = start_pos + num_bytes - 1;
lock_extent(io_tree, start_pos, end_of_last_block, GFP_NOFS);
trans = btrfs_join_transaction(root, 1);
if (!trans) {
err = -ENOMEM;
goto out_unlock;
}
btrfs_set_trans_block_group(trans, inode);
hint_byte = 0;
set_extent_uptodate(io_tree, start_pos, end_of_last_block, GFP_NOFS);
/* check for reserved extents on each page, we don't want
* to reset the delalloc bit on things that already have
* extents reserved.
*/
btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block);
for (i = 0; i < num_pages; i++) {
struct page *p = pages[i];
SetPageUptodate(p);
ClearPageChecked(p);
set_page_dirty(p);
}
if (end_pos > isize) {
i_size_write(inode, end_pos);
/* we've only changed i_size in ram, and we haven't updated
* the disk i_size. There is no need to log the inode
* at this time.
*/
}
err = btrfs_end_transaction(trans, root);
out_unlock:
unlock_extent(io_tree, start_pos, end_of_last_block, GFP_NOFS);
return err;
}
static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
u64 num_items, int type)
{
struct btrfs_trans_handle *h;
struct btrfs_transaction *cur_trans;
int retries = 0;
int ret;
again:
h = kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
if (!h)
return ERR_PTR(-ENOMEM);
mutex_lock(&root->fs_info->trans_mutex);
if (may_wait_transaction(root, type))
wait_current_trans(root);
ret = join_transaction(root);
BUG_ON(ret);
cur_trans = root->fs_info->running_transaction;
cur_trans->use_count++;
mutex_unlock(&root->fs_info->trans_mutex);
h->transid = cur_trans->transid;
h->transaction = cur_trans;
h->blocks_used = 0;
h->block_group = 0;
h->bytes_reserved = 0;
h->delayed_ref_updates = 0;
h->block_rsv = NULL;
smp_mb();
if (cur_trans->blocked && may_wait_transaction(root, type)) {
btrfs_commit_transaction(h, root);
goto again;
}
if (num_items > 0) {
ret = btrfs_trans_reserve_metadata(h, root, num_items,
&retries);
if (ret == -EAGAIN) {
btrfs_commit_transaction(h, root);
goto again;
}
if (ret < 0) {
btrfs_end_transaction(h, root);
return ERR_PTR(ret);
}
}
mutex_lock(&root->fs_info->trans_mutex);
record_root_in_trans(h, root);
mutex_unlock(&root->fs_info->trans_mutex);
if (!current->journal_info && type != TRANS_USERSPACE)
current->journal_info = h;
return h;
}
int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info)
{
struct btrfs_trans_handle *trans;
struct btrfs_root *tree_root = fs_info->tree_root;
struct btrfs_root *free_space_root;
struct btrfs_block_group_cache *block_group;
struct rb_node *node;
int ret;
trans = btrfs_start_transaction(tree_root, 0);
if (IS_ERR(trans))
return PTR_ERR(trans);
fs_info->creating_free_space_tree = 1;
free_space_root = btrfs_create_tree(trans, fs_info,
BTRFS_FREE_SPACE_TREE_OBJECTID);
if (IS_ERR(free_space_root)) {
ret = PTR_ERR(free_space_root);
goto abort;
}
fs_info->free_space_root = free_space_root;
node = rb_first(&fs_info->block_group_cache_tree);
while (node) {
block_group = rb_entry(node, struct btrfs_block_group_cache,
cache_node);
ret = populate_free_space_tree(trans, fs_info, block_group);
if (ret)
goto abort;
node = rb_next(node);
}
btrfs_set_fs_compat_ro(fs_info, FREE_SPACE_TREE);
fs_info->creating_free_space_tree = 0;
ret = btrfs_commit_transaction(trans, tree_root);
if (ret)
return ret;
return 0;
abort:
fs_info->creating_free_space_tree = 0;
btrfs_abort_transaction(trans, ret);
btrfs_end_transaction(trans, tree_root);
return ret;
}
static noinline int finish_pending_snapshot(struct btrfs_fs_info *fs_info,
struct btrfs_pending_snapshot *pending)
{
int ret;
int namelen;
u64 index = 0;
struct btrfs_trans_handle *trans;
struct inode *parent_inode;
struct inode *inode;
struct btrfs_root *parent_root;
parent_inode = pending->dentry->d_parent->d_inode;
parent_root = BTRFS_I(parent_inode)->root;
trans = btrfs_join_transaction(parent_root, 1);
/*
* insert the directory item
*/
namelen = strlen(pending->name);
ret = btrfs_set_inode_index(parent_inode, &index);
ret = btrfs_insert_dir_item(trans, parent_root,
pending->name, namelen,
parent_inode->i_ino,
&pending->root_key, BTRFS_FT_DIR, index);
if (ret)
goto fail;
btrfs_i_size_write(parent_inode, parent_inode->i_size + namelen * 2);
ret = btrfs_update_inode(trans, parent_root, parent_inode);
BUG_ON(ret);
ret = btrfs_add_root_ref(trans, parent_root->fs_info->tree_root,
pending->root_key.objectid,
parent_root->root_key.objectid,
parent_inode->i_ino, index, pending->name,
namelen);
BUG_ON(ret);
inode = btrfs_lookup_dentry(parent_inode, pending->dentry);
d_instantiate(pending->dentry, inode);
fail:
btrfs_end_transaction(trans, fs_info->fs_root);
return ret;
}
int btrfs_clear_free_space_tree(struct btrfs_fs_info *fs_info)
{
struct btrfs_trans_handle *trans;
struct btrfs_root *tree_root = fs_info->tree_root;
struct btrfs_root *free_space_root = fs_info->free_space_root;
int ret;
trans = btrfs_start_transaction(tree_root, 0);
if (IS_ERR(trans))
return PTR_ERR(trans);
btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE);
fs_info->free_space_root = NULL;
ret = clear_free_space_tree(trans, free_space_root);
if (ret)
goto abort;
ret = btrfs_del_root(trans, tree_root, &free_space_root->root_key);
if (ret)
goto abort;
list_del(&free_space_root->dirty_list);
btrfs_tree_lock(free_space_root->node);
clean_tree_block(trans, tree_root->fs_info, free_space_root->node);
btrfs_tree_unlock(free_space_root->node);
btrfs_free_tree_block(trans, free_space_root, free_space_root->node,
0, 1);
free_extent_buffer(free_space_root->node);
free_extent_buffer(free_space_root->commit_root);
kfree(free_space_root);
ret = btrfs_commit_transaction(trans, tree_root);
if (ret)
return ret;
return 0;
abort:
btrfs_abort_transaction(trans, ret);
btrfs_end_transaction(trans, tree_root);
return ret;
}
static int btrfs_uuid_iter_rem(struct btrfs_root *uuid_root, u8 *uuid, u8 type,
u64 subid)
{
struct btrfs_trans_handle *trans;
int ret;
/* 1 - for the uuid item */
trans = btrfs_start_transaction(uuid_root, 1);
if (IS_ERR(trans)) {
ret = PTR_ERR(trans);
goto out;
}
ret = btrfs_uuid_tree_remove(trans, uuid, type, subid);
btrfs_end_transaction(trans);
out:
return ret;
}
static int generic_search(struct inode *inode, u64 file_pos,
struct btrfs_dedup_hash *hash)
{
int ret;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_fs_info *fs_info = root->fs_info;
struct btrfs_trans_handle *trans;
struct btrfs_delayed_ref_root *delayed_refs;
struct btrfs_delayed_ref_head *head;
struct btrfs_dedup_info *dedup_info = fs_info->dedup_info;
u64 bytenr;
u64 tmp_bytenr;
u32 num_bytes;
trans = btrfs_join_transaction(root);
if (IS_ERR(trans))
return PTR_ERR(trans);
again:
mutex_lock(&dedup_info->lock);
ret = generic_search_hash(dedup_info, hash->hash, &bytenr, &num_bytes);
if (ret <= 0)
goto out;
delayed_refs = &trans->transaction->delayed_refs;
spin_lock(&delayed_refs->lock);
head = btrfs_find_delayed_ref_head(trans, bytenr);
if (!head) {
/*
* We can safely insert a new delayed_ref as long as we
* hold delayed_refs->lock.
* Only need to use atomic inc_extent_ref()
*/
ret = btrfs_inc_extent_ref_atomic(trans, root, bytenr,
num_bytes, 0, root->root_key.objectid,
btrfs_ino(inode), file_pos);
spin_unlock(&delayed_refs->lock);
if (ret == 0) {
hash->bytenr = bytenr;
hash->num_bytes = num_bytes;
ret = 1;
}
goto out;
}
/*
* We can't lock ref head with dedup_info->lock hold or we will cause
* ABBA dead lock.
*/
mutex_unlock(&dedup_info->lock);
ret = btrfs_delayed_ref_lock(trans, head);
spin_unlock(&delayed_refs->lock);
if (ret == -EAGAIN)
goto again;
mutex_lock(&dedup_info->lock);
/*
* Search again to ensure the hash is still here and bytenr didn't
* change
*/
ret = generic_search_hash(dedup_info, hash->hash, &tmp_bytenr,
&num_bytes);
if (ret <= 0) {
mutex_unlock(&head->mutex);
goto out;
}
if (tmp_bytenr != bytenr) {
mutex_unlock(&head->mutex);
mutex_unlock(&dedup_info->lock);
goto again;
}
hash->bytenr = bytenr;
hash->num_bytes = num_bytes;
/*
* Increase the extent ref right now, to avoid delayed ref run
* Or we may increase ref on non-exist extent.
*/
btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0,
root->root_key.objectid,
btrfs_ino(inode), file_pos);
mutex_unlock(&head->mutex);
out:
mutex_unlock(&dedup_info->lock);
btrfs_end_transaction(trans, root);
return ret;
}
static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct btrfs_inode *ip = BTRFS_I(inode);
struct btrfs_root *root = ip->root;
struct btrfs_trans_handle *trans;
unsigned int flags, oldflags;
int ret;
if (copy_from_user(&flags, arg, sizeof(flags)))
return -EFAULT;
if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
FS_NOATIME_FL | FS_NODUMP_FL | \
FS_SYNC_FL | FS_DIRSYNC_FL))
return -EOPNOTSUPP;
if (!is_owner_or_cap(inode))
return -EACCES;
mutex_lock(&inode->i_mutex);
flags = btrfs_mask_flags(inode->i_mode, flags);
oldflags = btrfs_flags_to_ioctl(ip->flags);
if ((flags ^ oldflags) & (FS_APPEND_FL | FS_IMMUTABLE_FL)) {
if (!capable(CAP_LINUX_IMMUTABLE)) {
ret = -EPERM;
goto out_unlock;
}
}
ret = mnt_want_write(file->f_path.mnt);
if (ret)
goto out_unlock;
if (flags & FS_SYNC_FL)
ip->flags |= BTRFS_INODE_SYNC;
else
ip->flags &= ~BTRFS_INODE_SYNC;
if (flags & FS_IMMUTABLE_FL)
ip->flags |= BTRFS_INODE_IMMUTABLE;
else
ip->flags &= ~BTRFS_INODE_IMMUTABLE;
if (flags & FS_APPEND_FL)
ip->flags |= BTRFS_INODE_APPEND;
else
ip->flags &= ~BTRFS_INODE_APPEND;
if (flags & FS_NODUMP_FL)
ip->flags |= BTRFS_INODE_NODUMP;
else
ip->flags &= ~BTRFS_INODE_NODUMP;
if (flags & FS_NOATIME_FL)
ip->flags |= BTRFS_INODE_NOATIME;
else
ip->flags &= ~BTRFS_INODE_NOATIME;
if (flags & FS_DIRSYNC_FL)
ip->flags |= BTRFS_INODE_DIRSYNC;
else
ip->flags &= ~BTRFS_INODE_DIRSYNC;
trans = btrfs_join_transaction(root, 1);
BUG_ON(!trans);
ret = btrfs_update_inode(trans, root, inode);
BUG_ON(ret);
btrfs_update_iflags(inode);
inode->i_ctime = CURRENT_TIME;
btrfs_end_transaction(trans, root);
mnt_drop_write(file->f_path.mnt);
out_unlock:
mutex_unlock(&inode->i_mutex);
return 0;
}
/*
* Given a list of roots that need to be deleted, call btrfs_drop_snapshot on
* all of them
*/
static noinline int drop_dirty_roots(struct btrfs_root *tree_root,
struct list_head *list)
{
struct btrfs_dirty_root *dirty;
struct btrfs_trans_handle *trans;
unsigned long nr;
u64 num_bytes;
u64 bytes_used;
u64 max_useless;
int ret = 0;
int err;
while (!list_empty(list)) {
struct btrfs_root *root;
dirty = list_entry(list->prev, struct btrfs_dirty_root, list);
list_del_init(&dirty->list);
num_bytes = btrfs_root_used(&dirty->root->root_item);
root = dirty->latest_root;
atomic_inc(&root->fs_info->throttles);
while (1) {
/*
* we don't want to jump in and create a bunch of
* delayed refs if the transaction is starting to close
*/
wait_transaction_pre_flush(tree_root->fs_info);
trans = btrfs_start_transaction(tree_root, 1);
/*
* we've joined a transaction, make sure it isn't
* closing right now
*/
if (trans->transaction->delayed_refs.flushing) {
btrfs_end_transaction(trans, tree_root);
continue;
}
mutex_lock(&root->fs_info->drop_mutex);
ret = btrfs_drop_snapshot(trans, dirty->root);
if (ret != -EAGAIN)
break;
mutex_unlock(&root->fs_info->drop_mutex);
err = btrfs_update_root(trans,
tree_root,
&dirty->root->root_key,
&dirty->root->root_item);
if (err)
ret = err;
nr = trans->blocks_used;
ret = btrfs_end_transaction(trans, tree_root);
BUG_ON(ret);
btrfs_btree_balance_dirty(tree_root, nr);
cond_resched();
}
BUG_ON(ret);
atomic_dec(&root->fs_info->throttles);
wake_up(&root->fs_info->transaction_throttle);
num_bytes -= btrfs_root_used(&dirty->root->root_item);
bytes_used = btrfs_root_used(&root->root_item);
if (num_bytes) {
mutex_lock(&root->fs_info->trans_mutex);
btrfs_record_root_in_trans(root);
mutex_unlock(&root->fs_info->trans_mutex);
btrfs_set_root_used(&root->root_item,
bytes_used - num_bytes);
}
ret = btrfs_del_root(trans, tree_root, &dirty->root->root_key);
if (ret) {
BUG();
break;
}
mutex_unlock(&root->fs_info->drop_mutex);
spin_lock(&root->list_lock);
list_del_init(&dirty->root->dead_list);
if (!list_empty(&root->dead_list)) {
struct btrfs_root *oldest;
oldest = list_entry(root->dead_list.prev,
struct btrfs_root, dead_list);
max_useless = oldest->root_key.offset - 1;
} else {
max_useless = root->root_key.offset - 1;
}
spin_unlock(&root->list_lock);
nr = trans->blocks_used;
ret = btrfs_end_transaction(trans, tree_root);
BUG_ON(ret);
ret = btrfs_remove_leaf_refs(root, max_useless, 0);
BUG_ON(ret);
free_extent_buffer(dirty->root->node);
kfree(dirty->root);
kfree(dirty);
btrfs_btree_balance_dirty(tree_root, nr);
cond_resched();
}
int __btrfs_setxattr(struct inode *inode, const char *name,
const void *value, size_t size, int flags)
{
struct btrfs_dir_item *di;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
struct btrfs_path *path;
int ret = 0, mod = 0;
path = btrfs_alloc_path();
if (!path)
return -ENOMEM;
trans = btrfs_join_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
/* first lets see if we already have this xattr */
di = btrfs_lookup_xattr(trans, root, path, inode->i_ino, name,
strlen(name), -1);
if (IS_ERR(di)) {
ret = PTR_ERR(di);
goto out;
}
/* ok we already have this xattr, lets remove it */
if (di) {
/* if we want create only exit */
if (flags & XATTR_CREATE) {
ret = -EEXIST;
goto out;
}
ret = btrfs_delete_one_dir_name(trans, root, path, di);
if (ret)
goto out;
btrfs_release_path(root, path);
/* if we don't have a value then we are removing the xattr */
if (!value) {
mod = 1;
goto out;
}
} else {
btrfs_release_path(root, path);
if (flags & XATTR_REPLACE) {
/* we couldn't find the attr to replace */
ret = -ENODATA;
goto out;
}
}
/* ok we have to create a completely new xattr */
ret = btrfs_insert_xattr_item(trans, root, name, strlen(name),
value, size, inode->i_ino);
if (ret)
goto out;
mod = 1;
out:
if (mod) {
inode->i_ctime = CURRENT_TIME;
ret = btrfs_update_inode(trans, root, inode);
}
btrfs_end_transaction(trans, root);
btrfs_free_path(path);
return ret;
}
/*
* calls iterate() for every inode that references the extent identified by
* the given parameters.
* when the iterator function returns a non-zero value, iteration stops.
*/
int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
u64 extent_item_objectid, u64 extent_item_pos,
int search_commit_root,
iterate_extent_inodes_t *iterate, void *ctx)
{
int ret;
struct btrfs_trans_handle *trans = NULL;
struct ulist *refs = NULL;
struct ulist *roots = NULL;
struct ulist_node *ref_node = NULL;
struct ulist_node *root_node = NULL;
struct seq_list tree_mod_seq_elem = {};
struct ulist_iterator ref_uiter;
struct ulist_iterator root_uiter;
pr_debug("resolving all inodes for extent %llu\n",
extent_item_objectid);
if (!search_commit_root) {
trans = btrfs_join_transaction(fs_info->extent_root);
if (IS_ERR(trans))
return PTR_ERR(trans);
btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
} else {
down_read(&fs_info->commit_root_sem);
}
ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid,
tree_mod_seq_elem.seq, &refs,
&extent_item_pos);
if (ret)
goto out;
ULIST_ITER_INIT(&ref_uiter);
while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) {
ret = __btrfs_find_all_roots(trans, fs_info, ref_node->val,
tree_mod_seq_elem.seq, &roots);
if (ret)
break;
ULIST_ITER_INIT(&root_uiter);
while (!ret && (root_node = ulist_next(roots, &root_uiter))) {
pr_debug("root %llu references leaf %llu, data list "
"%#llx\n", root_node->val, ref_node->val,
ref_node->aux);
ret = iterate_leaf_refs((struct extent_inode_elem *)
(uintptr_t)ref_node->aux,
root_node->val,
extent_item_objectid,
iterate, ctx);
}
ulist_free(roots);
}
free_leaf_list(refs);
out:
if (!search_commit_root) {
btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
btrfs_end_transaction(trans, fs_info->extent_root);
} else {
up_read(&fs_info->commit_root_sem);
}
return ret;
}