/*
* utility function that does setup for reiserfs_new_inode.
* dquot_initialize needs lots of credits so it's better to have it
* outside of a transaction, so we had to pull some bits of
* reiserfs_new_inode out into this func.
*/
static int new_inode_init(struct inode *inode, struct inode *dir, umode_t mode)
{
/*
* Make inode invalid - just in case we are going to drop it before
* the initialization happens
*/
INODE_PKEY(inode)->k_objectid = 0;
/*
* the quota init calls have to know who to charge the quota to, so
* we have to set uid and gid here
*/
inode_init_owner(inode, dir, mode);
return dquot_initialize(inode);
}
static int ufs_mkdir(struct inode * dir, struct dentry * dentry, int mode)
{
struct inode * inode;
int err = -EMLINK;
if (dir->i_nlink >= UFS_LINK_MAX)
goto out;
dquot_initialize(dir);
lock_kernel();
inode_inc_link_count(dir);
inode = ufs_new_inode(dir, S_IFDIR|mode);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out_dir;
inode->i_op = &ufs_dir_inode_operations;
inode->i_fop = &ufs_dir_operations;
inode->i_mapping->a_ops = &ufs_aops;
inode_inc_link_count(inode);
err = ufs_make_empty(inode, dir);
if (err)
goto out_fail;
err = ufs_add_link(dentry, inode);
if (err)
goto out_fail;
unlock_kernel();
d_instantiate(dentry, inode);
out:
return err;
out_fail:
inode_dec_link_count(inode);
inode_dec_link_count(inode);
iput (inode);
out_dir:
inode_dec_link_count(dir);
unlock_kernel();
goto out;
}
static int ext2_symlink (struct inode * dir, struct dentry * dentry,
const char * symname)
{
struct super_block * sb = dir->i_sb;
int err = -ENAMETOOLONG;
unsigned l = strlen(symname)+1;
struct inode * inode;
if (l > sb->s_blocksize)
goto out;
dquot_initialize(dir);
inode = ext2_new_inode (dir, S_IFLNK | S_IRWXUGO, &dentry->d_name);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out;
if (l > sizeof (EXT2_I(inode)->i_data)) {
/* slow symlink */
inode->i_op = &ext2_symlink_inode_operations;
if (test_opt(inode->i_sb, NOBH))
inode->i_mapping->a_ops = &ext2_nobh_aops;
else
inode->i_mapping->a_ops = &ext2_aops;
err = page_symlink(inode, symname, l);
if (err)
goto out_fail;
} else {
/* fast symlink */
inode->i_op = &ext2_fast_symlink_inode_operations;
memcpy((char*)(EXT2_I(inode)->i_data),symname,l);
inode->i_size = l-1;
}
mark_inode_dirty(inode);
err = ext2_add_nondir(dentry, inode);
out:
return err;
out_fail:
inode_dec_link_count(inode);
unlock_new_inode(inode);
iput (inode);
goto out;
}
static int ext2_mkdir(struct inode * dir, struct dentry * dentry, umode_t mode)
{
struct inode * inode;
int err;
dquot_initialize(dir);
inode_inc_link_count(dir);
inode = ext2_new_inode(dir, S_IFDIR | mode, &dentry->d_name);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out_dir;
inode->i_op = &ext2_dir_inode_operations;
inode->i_fop = &ext2_dir_operations;
if (test_opt(inode->i_sb, NOBH))
inode->i_mapping->a_ops = &ext2_nobh_aops;
else
inode->i_mapping->a_ops = &ext2_aops;
inode_inc_link_count(inode);
err = ext2_make_empty(inode, dir);
if (err)
goto out_fail;
err = ext2_add_link(dentry, inode);
if (err)
goto out_fail;
unlock_new_inode(inode);
d_instantiate(dentry, inode);
out:
return err;
out_fail:
inode_dec_link_count(inode);
inode_dec_link_count(inode);
unlock_new_inode(inode);
iput(inode);
out_dir:
inode_dec_link_count(dir);
goto out;
}
static int ufs_symlink (struct inode * dir, struct dentry * dentry,
const char * symname)
{
struct super_block * sb = dir->i_sb;
int err = -ENAMETOOLONG;
unsigned l = strlen(symname)+1;
struct inode * inode;
if (l > sb->s_blocksize)
goto out_notlocked;
dquot_initialize(dir);
lock_kernel();
inode = ufs_new_inode(dir, S_IFLNK | S_IRWXUGO);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out;
if (l > UFS_SB(sb)->s_uspi->s_maxsymlinklen) {
/* slow symlink */
inode->i_op = &ufs_symlink_inode_operations;
inode->i_mapping->a_ops = &ufs_aops;
err = page_symlink(inode, symname, l);
if (err)
goto out_fail;
} else {
/* fast symlink */
inode->i_op = &ufs_fast_symlink_inode_operations;
memcpy(UFS_I(inode)->i_u1.i_symlink, symname, l);
inode->i_size = l-1;
}
mark_inode_dirty(inode);
err = ufs_add_nondir(dentry, inode);
out:
unlock_kernel();
out_notlocked:
return err;
out_fail:
inode_dec_link_count(inode);
iput(inode);
goto out;
}
static int f2fs_create(struct inode *dir, struct dentry *dentry, umode_t mode,
bool excl)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
struct inode *inode;
nid_t ino = 0;
int err;
err = dquot_initialize(dir);
if (err)
return err;
inode = f2fs_new_inode(dir, mode);
if (IS_ERR(inode))
return PTR_ERR(inode);
if (!test_opt(sbi, DISABLE_EXT_IDENTIFY))
set_cold_files(sbi, inode, dentry->d_name.name);
inode->i_op = &f2fs_file_inode_operations;
inode->i_fop = &f2fs_file_operations;
inode->i_mapping->a_ops = &f2fs_dblock_aops;
ino = inode->i_ino;
f2fs_lock_op(sbi);
err = f2fs_add_link(dentry, inode);
if (err)
goto out;
f2fs_unlock_op(sbi);
alloc_nid_done(sbi, ino);
d_instantiate(dentry, inode);
unlock_new_inode(inode);
if (IS_DIRSYNC(dir))
f2fs_sync_fs(sbi->sb, 1);
f2fs_balance_fs(sbi, true);
return 0;
out:
handle_failed_inode(inode);
return err;
}
static int f2fs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
struct inode *inode;
int err;
err = dquot_initialize(dir);
if (err)
return err;
inode = f2fs_new_inode(dir, S_IFDIR | mode);
if (IS_ERR(inode))
return PTR_ERR(inode);
inode->i_op = &f2fs_dir_inode_operations;
inode->i_fop = &f2fs_dir_operations;
inode->i_mapping->a_ops = &f2fs_dblock_aops;
mapping_set_gfp_mask(inode->i_mapping, GFP_F2FS_HIGH_ZERO);
set_inode_flag(inode, FI_INC_LINK);
f2fs_lock_op(sbi);
err = f2fs_add_link(dentry, inode);
if (err)
goto out_fail;
f2fs_unlock_op(sbi);
alloc_nid_done(sbi, inode->i_ino);
d_instantiate(dentry, inode);
unlock_new_inode(inode);
if (IS_DIRSYNC(dir))
f2fs_sync_fs(sbi->sb, 1);
f2fs_balance_fs(sbi, true);
return 0;
out_fail:
clear_inode_flag(inode, FI_INC_LINK);
handle_failed_inode(inode);
return err;
}
/*
* Called at the last iput() if i_nlink is zero.
*/
void ext2_evict_inode(struct inode * inode)
{
struct ext2_block_alloc_info *rsv;
int want_delete = 0;
if (!inode->i_nlink && !is_bad_inode(inode)) {
want_delete = 1;
dquot_initialize(inode);
} else {
dquot_drop(inode);
}
truncate_inode_pages_final(&inode->i_data);
if (want_delete) {
sb_start_intwrite(inode->i_sb);
/* set dtime */
EXT2_I(inode)->i_dtime = get_seconds();
mark_inode_dirty(inode);
__ext2_write_inode(inode, inode_needs_sync(inode));
/* truncate to 0 */
inode->i_size = 0;
if (inode->i_blocks)
ext2_truncate_blocks(inode, 0);
ext2_xattr_delete_inode(inode);
}
invalidate_inode_buffers(inode);
clear_inode(inode);
ext2_discard_reservation(inode);
rsv = EXT2_I(inode)->i_block_alloc_info;
EXT2_I(inode)->i_block_alloc_info = NULL;
if (unlikely(rsv))
kfree(rsv);
if (want_delete) {
ext2_free_inode(inode);
sb_end_intwrite(inode->i_sb);
}
}
static int ext2_link (struct dentry * old_dentry, struct inode * dir,
struct dentry *dentry)
{
struct inode *inode = old_dentry->d_inode;
int err;
dquot_initialize(dir);
inode->i_ctime = CURRENT_TIME_SEC;
inode_inc_link_count(inode);
ihold(inode);
err = ext2_add_link(dentry, inode);
if (!err) {
d_instantiate(dentry, inode);
return 0;
}
inode_dec_link_count(inode);
iput(inode);
return err;
}
/*
* By the time this is called, we already have created
* the directory cache entry for the new file, but it
* is so far negative - it has no inode.
*
* If the create succeeds, we fill in the inode information
* with d_instantiate().
*/
static int ext2_create (struct inode * dir, struct dentry * dentry, umode_t mode, bool excl)
{
struct inode *inode;
dquot_initialize(dir);
inode = ext2_new_inode(dir, mode, &dentry->d_name);
if (IS_ERR(inode))
return PTR_ERR(inode);
inode->i_op = &ext2_file_inode_operations;
if (test_opt(inode->i_sb, NOBH)) {
inode->i_mapping->a_ops = &ext2_nobh_aops;
inode->i_fop = &ext2_file_operations;
} else {
inode->i_mapping->a_ops = &ext2_aops;
inode->i_fop = &ext2_file_operations;
}
mark_inode_dirty(inode);
return ext2_add_nondir(dentry, inode);
}
int jfs_setattr(struct dentry *dentry, struct iattr *iattr)
{
struct inode *inode = d_inode(dentry);
int rc;
rc = setattr_prepare(dentry, iattr);
if (rc)
return rc;
if (is_quota_modification(inode, iattr)) {
rc = dquot_initialize(inode);
if (rc)
return rc;
}
if ((iattr->ia_valid & ATTR_UID && !uid_eq(iattr->ia_uid, inode->i_uid)) ||
(iattr->ia_valid & ATTR_GID && !gid_eq(iattr->ia_gid, inode->i_gid))) {
rc = dquot_transfer(inode, iattr);
if (rc)
return rc;
}
if ((iattr->ia_valid & ATTR_SIZE) &&
iattr->ia_size != i_size_read(inode)) {
inode_dio_wait(inode);
rc = inode_newsize_ok(inode, iattr->ia_size);
if (rc)
return rc;
truncate_setsize(inode, iattr->ia_size);
jfs_truncate(inode);
}
setattr_copy(inode, iattr);
mark_inode_dirty(inode);
if (iattr->ia_valid & ATTR_MODE)
rc = posix_acl_chmod(inode, inode->i_mode);
return rc;
}
static int f2fs_unlink(struct inode *dir, struct dentry *dentry)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
struct inode *inode = d_inode(dentry);
struct f2fs_dir_entry *de;
struct page *page;
int err = -ENOENT;
trace_f2fs_unlink_enter(dir, dentry);
err = dquot_initialize(dir);
if (err)
return err;
de = f2fs_find_entry(dir, &dentry->d_name, &page);
if (!de) {
if (IS_ERR(page))
err = PTR_ERR(page);
goto fail;
}
f2fs_balance_fs(sbi, true);
f2fs_lock_op(sbi);
err = acquire_orphan_inode(sbi);
if (err) {
f2fs_unlock_op(sbi);
f2fs_dentry_kunmap(dir, page);
f2fs_put_page(page, 0);
goto fail;
}
f2fs_delete_entry(de, page, dir, inode);
f2fs_unlock_op(sbi);
if (IS_DIRSYNC(dir))
f2fs_sync_fs(sbi->sb, 1);
fail:
trace_f2fs_unlink_exit(inode, err);
return err;
}
/* utility function that does setup for reiserfs_new_inode.
** dquot_initialize needs lots of credits so it's better to have it
** outside of a transaction, so we had to pull some bits of
** reiserfs_new_inode out into this func.
*/
static int new_inode_init(struct inode *inode, struct inode *dir, int mode)
{
/* the quota init calls have to know who to charge the quota to, so
** we have to set uid and gid here
*/
inode->i_uid = current_fsuid();
inode->i_mode = mode;
/* Make inode invalid - just in case we are going to drop it before
* the initialization happens */
INODE_PKEY(inode)->k_objectid = 0;
if (dir->i_mode & S_ISGID) {
inode->i_gid = dir->i_gid;
if (S_ISDIR(mode))
inode->i_mode |= S_ISGID;
} else {
inode->i_gid = current_fsgid();
}
dquot_initialize(inode);
return 0;
}
static int ufs_link (struct dentry * old_dentry, struct inode * dir,
struct dentry *dentry)
{
struct inode *inode = old_dentry->d_inode;
int error;
lock_kernel();
if (inode->i_nlink >= UFS_LINK_MAX) {
unlock_kernel();
return -EMLINK;
}
dquot_initialize(dir);
inode->i_ctime = CURRENT_TIME_SEC;
inode_inc_link_count(inode);
atomic_inc(&inode->i_count);
error = ufs_add_nondir(dentry, inode);
unlock_kernel();
return error;
}
static int ufs_mknod (struct inode * dir, struct dentry *dentry, int mode, dev_t rdev)
{
struct inode *inode;
int err;
if (!old_valid_dev(rdev))
return -EINVAL;
dquot_initialize(dir);
inode = ufs_new_inode(dir, mode);
err = PTR_ERR(inode);
if (!IS_ERR(inode)) {
init_special_inode(inode, mode, rdev);
ufs_set_inode_dev(inode->i_sb, UFS_I(inode), rdev);
mark_inode_dirty(inode);
lock_kernel();
err = ufs_add_nondir(dentry, inode);
unlock_kernel();
}
return err;
}
static int ext2_mknod (struct inode * dir, struct dentry *dentry, umode_t mode, dev_t rdev)
{
struct inode * inode;
int err;
if (!new_valid_dev(rdev))
return -EINVAL;
dquot_initialize(dir);
inode = ext2_new_inode (dir, mode, &dentry->d_name);
err = PTR_ERR(inode);
if (!IS_ERR(inode)) {
init_special_inode(inode, inode->i_mode, rdev);
#ifdef CONFIG_EXT2_FS_XATTR
inode->i_op = &ext2_special_inode_operations;
#endif
mark_inode_dirty(inode);
err = ext2_add_nondir(dentry, inode);
}
return err;
}
static int f2fs_mknod(struct inode *dir, struct dentry *dentry,
umode_t mode, dev_t rdev)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
struct inode *inode;
int err = 0;
err = dquot_initialize(dir);
if (err)
return err;
inode = f2fs_new_inode(dir, mode);
if (IS_ERR(inode))
return PTR_ERR(inode);
init_special_inode(inode, inode->i_mode, rdev);
inode->i_op = &f2fs_special_inode_operations;
f2fs_lock_op(sbi);
err = f2fs_add_link(dentry, inode);
if (err)
goto out;
f2fs_unlock_op(sbi);
alloc_nid_done(sbi, inode->i_ino);
d_instantiate(dentry, inode);
unlock_new_inode(inode);
if (IS_DIRSYNC(dir))
f2fs_sync_fs(sbi->sb, 1);
f2fs_balance_fs(sbi, true);
return 0;
out:
handle_failed_inode(inode);
return err;
}
static int f2fs_link(struct dentry *old_dentry, struct inode *dir,
struct dentry *dentry)
{
struct inode *inode = d_inode(old_dentry);
struct f2fs_sb_info *sbi = F2FS_I_SB(dir);
int err;
if (f2fs_encrypted_inode(dir) &&
!fscrypt_has_permitted_context(dir, inode))
return -EPERM;
err = dquot_initialize(dir);
if (err)
return err;
f2fs_balance_fs(sbi, true);
inode->i_ctime = current_time(inode);
ihold(inode);
set_inode_flag(inode, FI_INC_LINK);
f2fs_lock_op(sbi);
err = f2fs_add_link(dentry, inode);
if (err)
goto out;
f2fs_unlock_op(sbi);
d_instantiate(dentry, inode);
if (IS_DIRSYNC(dir))
f2fs_sync_fs(sbi->sb, 1);
return 0;
out:
clear_inode_flag(inode, FI_INC_LINK);
iput(inode);
f2fs_unlock_op(sbi);
return err;
}
static int ext2_unlink(struct inode * dir, struct dentry *dentry)
{
struct inode * inode = dentry->d_inode;
struct ext2_dir_entry_2 * de;
struct page * page;
int err = -ENOENT;
dquot_initialize(dir);
de = ext2_find_entry (dir, &dentry->d_name, &page);
if (!de)
goto out;
err = ext2_delete_entry (de, page);
if (err)
goto out;
inode->i_ctime = dir->i_ctime;
inode_dec_link_count(inode);
err = 0;
out:
return err;
}
/*
* NAME: jfs_rmdir(dip, dentry)
*
* FUNCTION: remove a link to child directory
*
* PARAMETER: dip - parent inode
* dentry - child directory dentry
*
* RETURN: -EINVAL - if name is . or ..
* -EINVAL - if . or .. exist but are invalid.
* errors from subroutines
*
* note:
* if other threads have the directory open when the last link
* is removed, the "." and ".." entries, if present, are removed before
* rmdir() returns and no new entries may be created in the directory,
* but the directory is not removed until the last reference to
* the directory is released (cf.unlink() of regular file).
*/
static int jfs_rmdir(struct inode *dip, struct dentry *dentry)
{
int rc;
tid_t tid; /* transaction id */
struct inode *ip = dentry->d_inode;
ino_t ino;
struct component_name dname;
struct inode *iplist[2];
struct tblock *tblk;
jfs_info("jfs_rmdir: dip:0x%p name:%s", dip, dentry->d_name.name);
/* Init inode for quota operations. */
dquot_initialize(dip);
dquot_initialize(ip);
/* directory must be empty to be removed */
if (!dtEmpty(ip)) {
rc = -ENOTEMPTY;
goto out;
}
if ((rc = get_UCSname(&dname, dentry))) {
goto out;
}
tid = txBegin(dip->i_sb, 0);
mutex_lock_nested(&JFS_IP(dip)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
iplist[0] = dip;
iplist[1] = ip;
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = ip;
/*
* delete the entry of target directory from parent directory
*/
ino = ip->i_ino;
if ((rc = dtDelete(tid, dip, &dname, &ino, JFS_REMOVE))) {
jfs_err("jfs_rmdir: dtDelete returned %d", rc);
if (rc == -EIO)
txAbort(tid, 1);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
goto out2;
}
/* update parent directory's link count corresponding
* to ".." entry of the target directory deleted
*/
dip->i_ctime = dip->i_mtime = CURRENT_TIME;
inode_dec_link_count(dip);
/*
* OS/2 could have created EA and/or ACL
*/
/* free EA from both persistent and working map */
if (JFS_IP(ip)->ea.flag & DXD_EXTENT) {
/* free EA pages */
txEA(tid, ip, &JFS_IP(ip)->ea, NULL);
}
JFS_IP(ip)->ea.flag = 0;
/* free ACL from both persistent and working map */
if (JFS_IP(ip)->acl.flag & DXD_EXTENT) {
/* free ACL pages */
txEA(tid, ip, &JFS_IP(ip)->acl, NULL);
}
JFS_IP(ip)->acl.flag = 0;
/* mark the target directory as deleted */
clear_nlink(ip);
mark_inode_dirty(ip);
rc = txCommit(tid, 2, &iplist[0], 0);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
/*
* Truncating the directory index table is not guaranteed. It
* may need to be done iteratively
*/
if (test_cflag(COMMIT_Stale, dip)) {
if (dip->i_size > 1)
jfs_truncate_nolock(dip, 0);
clear_cflag(COMMIT_Stale, dip);
}
out2:
free_UCSname(&dname);
out:
jfs_info("jfs_rmdir: rc:%d", rc);
return rc;
}
static int reiserfs_unlink(struct inode *dir, struct dentry *dentry)
{
int retval, err;
struct inode *inode;
struct reiserfs_dir_entry de;
INITIALIZE_PATH(path);
struct reiserfs_transaction_handle th;
int jbegin_count;
unsigned long savelink;
int depth;
dquot_initialize(dir);
inode = dentry->d_inode;
/* in this transaction we can be doing at max two balancings and update
* two stat datas, we change quotas of the owner of the directory and of
* the owner of the parent directory. The quota structure is possibly
* deleted only on iput => outside of this transaction */
jbegin_count =
JOURNAL_PER_BALANCE_CNT * 2 + 2 +
4 * REISERFS_QUOTA_TRANS_BLOCKS(dir->i_sb);
depth = reiserfs_write_lock_once(dir->i_sb);
retval = journal_begin(&th, dir->i_sb, jbegin_count);
if (retval)
goto out_unlink;
de.de_gen_number_bit_string = NULL;
if ((retval =
reiserfs_find_entry(dir, dentry->d_name.name, dentry->d_name.len,
&path, &de)) == NAME_NOT_FOUND) {
retval = -ENOENT;
goto end_unlink;
} else if (retval == IO_ERROR) {
retval = -EIO;
goto end_unlink;
}
reiserfs_update_inode_transaction(inode);
reiserfs_update_inode_transaction(dir);
if (de.de_objectid != inode->i_ino) {
// FIXME: compare key of an object and a key found in the
// entry
retval = -EIO;
goto end_unlink;
}
if (!inode->i_nlink) {
reiserfs_warning(inode->i_sb, "reiserfs-7042",
"deleting nonexistent file (%lu), %d",
inode->i_ino, inode->i_nlink);
set_nlink(inode, 1);
}
drop_nlink(inode);
/*
* we schedule before doing the add_save_link call, save the link
* count so we don't race
*/
savelink = inode->i_nlink;
retval =
reiserfs_cut_from_item(&th, &path, &(de.de_entry_key), dir, NULL,
0);
if (retval < 0) {
inc_nlink(inode);
goto end_unlink;
}
inode->i_ctime = CURRENT_TIME_SEC;
reiserfs_update_sd(&th, inode);
dir->i_size -= (de.de_entrylen + DEH_SIZE);
dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC;
reiserfs_update_sd(&th, dir);
if (!savelink)
/* prevent file from getting lost */
add_save_link(&th, inode, 0 /* not truncate */ );
retval = journal_end(&th, dir->i_sb, jbegin_count);
reiserfs_check_path(&path);
reiserfs_write_unlock_once(dir->i_sb, depth);
return retval;
end_unlink:
pathrelse(&path);
err = journal_end(&th, dir->i_sb, jbegin_count);
reiserfs_check_path(&path);
if (err)
retval = err;
out_unlink:
reiserfs_write_unlock_once(dir->i_sb, depth);
return retval;
}
static int jfs_symlink(struct inode *dip, struct dentry *dentry,
const char *name)
{
int rc;
tid_t tid;
ino_t ino = 0;
struct component_name dname;
int ssize; /* source pathname size */
struct btstack btstack;
struct inode *ip = dentry->d_inode;
unchar *i_fastsymlink;
s64 xlen = 0;
int bmask = 0, xsize;
s64 extent = 0, xaddr;
struct metapage *mp;
struct super_block *sb;
struct tblock *tblk;
struct inode *iplist[2];
jfs_info("jfs_symlink: dip:0x%p name:%s", dip, name);
dquot_initialize(dip);
ssize = strlen(name) + 1;
/*
* search parent directory for entry/freespace
* (dtSearch() returns parent directory page pinned)
*/
if ((rc = get_UCSname(&dname, dentry)))
goto out1;
/*
* allocate on-disk/in-memory inode for symbolic link:
* (iAlloc() returns new, locked inode)
*/
ip = ialloc(dip, S_IFLNK | 0777);
if (IS_ERR(ip)) {
rc = PTR_ERR(ip);
goto out2;
}
tid = txBegin(dip->i_sb, 0);
mutex_lock_nested(&JFS_IP(dip)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
rc = jfs_init_security(tid, ip, dip, &dentry->d_name);
if (rc)
goto out3;
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_CREATE;
tblk->ino = ip->i_ino;
tblk->u.ixpxd = JFS_IP(ip)->ixpxd;
/* fix symlink access permission
* (dir_create() ANDs in the u.u_cmask,
* but symlinks really need to be 777 access)
*/
ip->i_mode |= 0777;
/*
* write symbolic link target path name
*/
xtInitRoot(tid, ip);
/*
* write source path name inline in on-disk inode (fast symbolic link)
*/
if (ssize <= IDATASIZE) {
ip->i_op = &jfs_fast_symlink_inode_operations;
i_fastsymlink = JFS_IP(ip)->i_inline;
memcpy(i_fastsymlink, name, ssize);
ip->i_size = ssize - 1;
/*
* if symlink is > 128 bytes, we don't have the space to
* store inline extended attributes
*/
if (ssize > sizeof (JFS_IP(ip)->i_inline))
JFS_IP(ip)->mode2 &= ~INLINEEA;
jfs_info("jfs_symlink: fast symlink added ssize:%d name:%s ",
ssize, name);
}
/*
* write source path name in a single extent
*/
else {
jfs_info("jfs_symlink: allocate extent ip:0x%p", ip);
ip->i_op = &jfs_symlink_inode_operations;
ip->i_mapping->a_ops = &jfs_aops;
/*
* even though the data of symlink object (source
* path name) is treated as non-journaled user data,
* it is read/written thru buffer cache for performance.
*/
sb = ip->i_sb;
bmask = JFS_SBI(sb)->bsize - 1;
xsize = (ssize + bmask) & ~bmask;
xaddr = 0;
xlen = xsize >> JFS_SBI(sb)->l2bsize;
if ((rc = xtInsert(tid, ip, 0, 0, xlen, &xaddr, 0))) {
txAbort(tid, 0);
goto out3;
}
extent = xaddr;
ip->i_size = ssize - 1;
while (ssize) {
/* This is kind of silly since PATH_MAX == 4K */
int copy_size = min(ssize, PSIZE);
mp = get_metapage(ip, xaddr, PSIZE, 1);
if (mp == NULL) {
xtTruncate(tid, ip, 0, COMMIT_PWMAP);
rc = -EIO;
txAbort(tid, 0);
goto out3;
}
memcpy(mp->data, name, copy_size);
flush_metapage(mp);
ssize -= copy_size;
name += copy_size;
xaddr += JFS_SBI(sb)->nbperpage;
}
}
/*
* create entry for symbolic link in parent directory
*/
rc = dtSearch(dip, &dname, &ino, &btstack, JFS_CREATE);
if (rc == 0) {
ino = ip->i_ino;
rc = dtInsert(tid, dip, &dname, &ino, &btstack);
}
if (rc) {
if (xlen)
xtTruncate(tid, ip, 0, COMMIT_PWMAP);
txAbort(tid, 0);
/* discard new inode */
goto out3;
}
mark_inode_dirty(ip);
dip->i_ctime = dip->i_mtime = CURRENT_TIME;
mark_inode_dirty(dip);
/*
* commit update of parent directory and link object
*/
iplist[0] = dip;
iplist[1] = ip;
rc = txCommit(tid, 2, &iplist[0], 0);
out3:
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
if (rc) {
free_ea_wmap(ip);
ip->i_nlink = 0;
unlock_new_inode(ip);
iput(ip);
} else {
d_instantiate(dentry, ip);
unlock_new_inode(ip);
}
out2:
free_UCSname(&dname);
out1:
jfs_info("jfs_symlink: rc:%d", rc);
return rc;
}
/*
* NAME: jfs_link(vp, dvp, name, crp)
*
* FUNCTION: create a link to <vp> by the name = <name>
* in the parent directory <dvp>
*
* PARAMETER: vp - target object
* dvp - parent directory of new link
* name - name of new link to target object
* crp - credential
*
* RETURN: Errors from subroutines
*
* note:
* JFS does NOT support link() on directories (to prevent circular
* path in the directory hierarchy);
* EPERM: the target object is a directory, and either the caller
* does not have appropriate privileges or the implementation prohibits
* using link() on directories [XPG4.2].
*
* JFS does NOT support links between file systems:
* EXDEV: target object and new link are on different file systems and
* implementation does not support links between file systems [XPG4.2].
*/
static int jfs_link(struct dentry *old_dentry,
struct inode *dir, struct dentry *dentry)
{
int rc;
tid_t tid;
struct inode *ip = old_dentry->d_inode;
ino_t ino;
struct component_name dname;
struct btstack btstack;
struct inode *iplist[2];
jfs_info("jfs_link: %s %s", old_dentry->d_name.name,
dentry->d_name.name);
if (ip->i_nlink == JFS_LINK_MAX)
return -EMLINK;
dquot_initialize(dir);
tid = txBegin(ip->i_sb, 0);
mutex_lock_nested(&JFS_IP(dir)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
/*
* scan parent directory for entry/freespace
*/
if ((rc = get_UCSname(&dname, dentry)))
goto out;
if ((rc = dtSearch(dir, &dname, &ino, &btstack, JFS_CREATE)))
goto free_dname;
/*
* create entry for new link in parent directory
*/
ino = ip->i_ino;
if ((rc = dtInsert(tid, dir, &dname, &ino, &btstack)))
goto free_dname;
/* update object inode */
inc_nlink(ip); /* for new link */
ip->i_ctime = CURRENT_TIME;
dir->i_ctime = dir->i_mtime = CURRENT_TIME;
mark_inode_dirty(dir);
ihold(ip);
iplist[0] = ip;
iplist[1] = dir;
rc = txCommit(tid, 2, &iplist[0], 0);
if (rc) {
ip->i_nlink--; /* never instantiated */
iput(ip);
} else
d_instantiate(dentry, ip);
free_dname:
free_UCSname(&dname);
out:
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dir)->commit_mutex);
jfs_info("jfs_link: rc:%d", rc);
return rc;
}
static void ocfs2_delete_inode(struct inode *inode)
{
int wipe, status;
sigset_t oldset;
struct buffer_head *di_bh = NULL;
trace_ocfs2_delete_inode(inode->i_ino,
(unsigned long long)OCFS2_I(inode)->ip_blkno,
is_bad_inode(inode));
/* When we fail in read_inode() we mark inode as bad. The second test
* catches the case when inode allocation fails before allocating
* a block for inode. */
if (is_bad_inode(inode) || !OCFS2_I(inode)->ip_blkno)
goto bail;
dquot_initialize(inode);
if (!ocfs2_inode_is_valid_to_delete(inode)) {
/* It's probably not necessary to truncate_inode_pages
* here but we do it for safety anyway (it will most
* likely be a no-op anyway) */
ocfs2_cleanup_delete_inode(inode, 0);
goto bail;
}
/* We want to block signals in delete_inode as the lock and
* messaging paths may return us -ERESTARTSYS. Which would
* cause us to exit early, resulting in inodes being orphaned
* forever. */
ocfs2_block_signals(&oldset);
/*
* Synchronize us against ocfs2_get_dentry. We take this in
* shared mode so that all nodes can still concurrently
* process deletes.
*/
status = ocfs2_nfs_sync_lock(OCFS2_SB(inode->i_sb), 0);
if (status < 0) {
mlog(ML_ERROR, "getting nfs sync lock(PR) failed %d\n", status);
ocfs2_cleanup_delete_inode(inode, 0);
goto bail_unblock;
}
/* Lock down the inode. This gives us an up to date view of
* it's metadata (for verification), and allows us to
* serialize delete_inode on multiple nodes.
*
* Even though we might be doing a truncate, we don't take the
* allocation lock here as it won't be needed - nobody will
* have the file open.
*/
status = ocfs2_inode_lock(inode, &di_bh, 1);
if (status < 0) {
if (status != -ENOENT)
mlog_errno(status);
ocfs2_cleanup_delete_inode(inode, 0);
goto bail_unlock_nfs_sync;
}
/* Query the cluster. This will be the final decision made
* before we go ahead and wipe the inode. */
status = ocfs2_query_inode_wipe(inode, di_bh, &wipe);
if (!wipe || status < 0) {
/* Error and remote inode busy both mean we won't be
* removing the inode, so they take almost the same
* path. */
if (status < 0)
mlog_errno(status);
/* Someone in the cluster has disallowed a wipe of
* this inode, or it was never completely
* orphaned. Write out the pages and exit now. */
ocfs2_cleanup_delete_inode(inode, 1);
goto bail_unlock_inode;
}
ocfs2_cleanup_delete_inode(inode, 0);
status = ocfs2_wipe_inode(inode, di_bh);
if (status < 0) {
if (status != -EDEADLK)
mlog_errno(status);
goto bail_unlock_inode;
}
/*
* Mark the inode as successfully deleted.
*
* This is important for ocfs2_clear_inode() as it will check
* this flag and skip any checkpointing work
*
* ocfs2_stuff_meta_lvb() also uses this flag to invalidate
* the LVB for other nodes.
*/
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_DELETED;
bail_unlock_inode:
ocfs2_inode_unlock(inode, 1);
brelse(di_bh);
bail_unlock_nfs_sync:
ocfs2_nfs_sync_unlock(OCFS2_SB(inode->i_sb), 0);
bail_unblock:
ocfs2_unblock_signals(&oldset);
bail:
return;
}
/*
* NAME: jfs_mknod
*
* FUNCTION: Create a special file (device)
*/
static int jfs_mknod(struct inode *dir, struct dentry *dentry,
int mode, dev_t rdev)
{
struct jfs_inode_info *jfs_ip;
struct btstack btstack;
struct component_name dname;
ino_t ino;
struct inode *ip;
struct inode *iplist[2];
int rc;
tid_t tid;
struct tblock *tblk;
if (!new_valid_dev(rdev))
return -EINVAL;
jfs_info("jfs_mknod: %s", dentry->d_name.name);
dquot_initialize(dir);
if ((rc = get_UCSname(&dname, dentry)))
goto out;
ip = ialloc(dir, mode);
if (IS_ERR(ip)) {
rc = PTR_ERR(ip);
goto out1;
}
jfs_ip = JFS_IP(ip);
tid = txBegin(dir->i_sb, 0);
mutex_lock_nested(&JFS_IP(dir)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
rc = jfs_init_acl(tid, ip, dir);
if (rc)
goto out3;
rc = jfs_init_security(tid, ip, dir, &dentry->d_name);
if (rc) {
txAbort(tid, 0);
goto out3;
}
if ((rc = dtSearch(dir, &dname, &ino, &btstack, JFS_CREATE))) {
txAbort(tid, 0);
goto out3;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_CREATE;
tblk->ino = ip->i_ino;
tblk->u.ixpxd = JFS_IP(ip)->ixpxd;
ino = ip->i_ino;
if ((rc = dtInsert(tid, dir, &dname, &ino, &btstack))) {
txAbort(tid, 0);
goto out3;
}
ip->i_op = &jfs_file_inode_operations;
jfs_ip->dev = new_encode_dev(rdev);
init_special_inode(ip, ip->i_mode, rdev);
mark_inode_dirty(ip);
dir->i_ctime = dir->i_mtime = CURRENT_TIME;
mark_inode_dirty(dir);
iplist[0] = dir;
iplist[1] = ip;
rc = txCommit(tid, 2, iplist, 0);
out3:
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dir)->commit_mutex);
if (rc) {
free_ea_wmap(ip);
ip->i_nlink = 0;
unlock_new_inode(ip);
iput(ip);
} else {
d_instantiate(dentry, ip);
unlock_new_inode(ip);
}
out1:
free_UCSname(&dname);
out:
jfs_info("jfs_mknod: returning %d", rc);
return rc;
}
/*
* There are two policies for allocating an inode. If the new inode is
* a directory, then a forward search is made for a block group with both
* free space and a low directory-to-inode ratio; if that fails, then of
* the groups with above-average free space, that group with the fewest
* directories already is chosen.
*
* For other inodes, search forward from the parent directory's block
* group to find a free inode.
*/
struct inode *ext3_new_inode(handle_t *handle, struct inode * dir,
const struct qstr *qstr, int mode)
{
struct super_block *sb;
struct buffer_head *bitmap_bh = NULL;
struct buffer_head *bh2;
int group;
unsigned long ino = 0;
struct inode * inode;
struct ext3_group_desc * gdp = NULL;
struct ext3_super_block * es;
struct ext3_inode_info *ei;
struct ext3_sb_info *sbi;
int err = 0;
struct inode *ret;
int i;
/* Cannot create files in a deleted directory */
if (!dir || !dir->i_nlink)
return ERR_PTR(-EPERM);
sb = dir->i_sb;
trace_ext3_request_inode(dir, mode);
inode = new_inode(sb);
if (!inode)
return ERR_PTR(-ENOMEM);
ei = EXT3_I(inode);
sbi = EXT3_SB(sb);
es = sbi->s_es;
if (S_ISDIR(mode)) {
if (test_opt (sb, OLDALLOC))
group = find_group_dir(sb, dir);
else
group = find_group_orlov(sb, dir);
} else
group = find_group_other(sb, dir);
err = -ENOSPC;
if (group == -1)
goto out;
for (i = 0; i < sbi->s_groups_count; i++) {
err = -EIO;
gdp = ext3_get_group_desc(sb, group, &bh2);
if (!gdp)
goto fail;
brelse(bitmap_bh);
bitmap_bh = read_inode_bitmap(sb, group);
if (!bitmap_bh)
goto fail;
ino = 0;
repeat_in_this_group:
ino = ext3_find_next_zero_bit((unsigned long *)
bitmap_bh->b_data, EXT3_INODES_PER_GROUP(sb), ino);
if (ino < EXT3_INODES_PER_GROUP(sb)) {
BUFFER_TRACE(bitmap_bh, "get_write_access");
err = ext3_journal_get_write_access(handle, bitmap_bh);
if (err)
goto fail;
if (!ext3_set_bit_atomic(sb_bgl_lock(sbi, group),
ino, bitmap_bh->b_data)) {
/* we won it */
BUFFER_TRACE(bitmap_bh,
"call ext3_journal_dirty_metadata");
err = ext3_journal_dirty_metadata(handle,
bitmap_bh);
if (err)
goto fail;
goto got;
}
/* we lost it */
journal_release_buffer(handle, bitmap_bh);
if (++ino < EXT3_INODES_PER_GROUP(sb))
goto repeat_in_this_group;
}
/*
* This case is possible in concurrent environment. It is very
* rare. We cannot repeat the find_group_xxx() call because
* that will simply return the same blockgroup, because the
* group descriptor metadata has not yet been updated.
* So we just go onto the next blockgroup.
*/
if (++group == sbi->s_groups_count)
group = 0;
}
err = -ENOSPC;
goto out;
got:
ino += group * EXT3_INODES_PER_GROUP(sb) + 1;
if (ino < EXT3_FIRST_INO(sb) || ino > le32_to_cpu(es->s_inodes_count)) {
ext3_error (sb, "ext3_new_inode",
"reserved inode or inode > inodes count - "
"block_group = %d, inode=%lu", group, ino);
err = -EIO;
goto fail;
}
BUFFER_TRACE(bh2, "get_write_access");
err = ext3_journal_get_write_access(handle, bh2);
if (err) goto fail;
spin_lock(sb_bgl_lock(sbi, group));
le16_add_cpu(&gdp->bg_free_inodes_count, -1);
if (S_ISDIR(mode)) {
le16_add_cpu(&gdp->bg_used_dirs_count, 1);
}
spin_unlock(sb_bgl_lock(sbi, group));
BUFFER_TRACE(bh2, "call ext3_journal_dirty_metadata");
err = ext3_journal_dirty_metadata(handle, bh2);
if (err) goto fail;
percpu_counter_dec(&sbi->s_freeinodes_counter);
if (S_ISDIR(mode))
percpu_counter_inc(&sbi->s_dirs_counter);
if (test_opt(sb, GRPID)) {
inode->i_mode = mode;
inode->i_uid = current_fsuid();
inode->i_gid = dir->i_gid;
} else
inode_init_owner(inode, dir, mode);
inode->i_ino = ino;
/* This is the optimal IO size (for stat), not the fs block size */
inode->i_blocks = 0;
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC;
memset(ei->i_data, 0, sizeof(ei->i_data));
ei->i_dir_start_lookup = 0;
ei->i_disksize = 0;
ei->i_flags =
ext3_mask_flags(mode, EXT3_I(dir)->i_flags & EXT3_FL_INHERITED);
#ifdef EXT3_FRAGMENTS
ei->i_faddr = 0;
ei->i_frag_no = 0;
ei->i_frag_size = 0;
#endif
ei->i_file_acl = 0;
ei->i_dir_acl = 0;
ei->i_dtime = 0;
ei->i_block_alloc_info = NULL;
ei->i_block_group = group;
ext3_set_inode_flags(inode);
if (IS_DIRSYNC(inode))
handle->h_sync = 1;
if (insert_inode_locked(inode) < 0) {
err = -EINVAL;
goto fail_drop;
}
spin_lock(&sbi->s_next_gen_lock);
inode->i_generation = sbi->s_next_generation++;
spin_unlock(&sbi->s_next_gen_lock);
ei->i_state_flags = 0;
ext3_set_inode_state(inode, EXT3_STATE_NEW);
/* See comment in ext3_iget for explanation */
if (ino >= EXT3_FIRST_INO(sb) + 1 &&
EXT3_INODE_SIZE(sb) > EXT3_GOOD_OLD_INODE_SIZE) {
ei->i_extra_isize =
sizeof(struct ext3_inode) - EXT3_GOOD_OLD_INODE_SIZE;
} else {
ei->i_extra_isize = 0;
}
ret = inode;
dquot_initialize(inode);
err = dquot_alloc_inode(inode);
if (err)
goto fail_drop;
err = ext3_init_acl(handle, inode, dir);
if (err)
goto fail_free_drop;
err = ext3_init_security(handle, inode, dir, qstr);
if (err)
goto fail_free_drop;
err = ext3_mark_inode_dirty(handle, inode);
if (err) {
ext3_std_error(sb, err);
goto fail_free_drop;
}
ext3_debug("allocating inode %lu\n", inode->i_ino);
trace_ext3_allocate_inode(inode, dir, mode);
goto really_out;
fail:
ext3_std_error(sb, err);
out:
iput(inode);
ret = ERR_PTR(err);
really_out:
brelse(bitmap_bh);
return ret;
fail_free_drop:
dquot_free_inode(inode);
fail_drop:
dquot_drop(inode);
inode->i_flags |= S_NOQUOTA;
inode->i_nlink = 0;
unlock_new_inode(inode);
iput(inode);
brelse(bitmap_bh);
return ERR_PTR(err);
}
static int ext2_rename (struct inode * old_dir, struct dentry * old_dentry,
struct inode * new_dir, struct dentry * new_dentry )
{
struct inode * old_inode = old_dentry->d_inode;
struct inode * new_inode = new_dentry->d_inode;
struct page * dir_page = NULL;
struct ext2_dir_entry_2 * dir_de = NULL;
struct page * old_page;
struct ext2_dir_entry_2 * old_de;
int err = -ENOENT;
dquot_initialize(old_dir);
dquot_initialize(new_dir);
old_de = ext2_find_entry (old_dir, &old_dentry->d_name, &old_page);
if (!old_de)
goto out;
if (S_ISDIR(old_inode->i_mode)) {
err = -EIO;
dir_de = ext2_dotdot(old_inode, &dir_page);
if (!dir_de)
goto out_old;
}
if (new_inode) {
struct page *new_page;
struct ext2_dir_entry_2 *new_de;
err = -ENOTEMPTY;
if (dir_de && !ext2_empty_dir (new_inode))
goto out_dir;
err = -ENOENT;
new_de = ext2_find_entry (new_dir, &new_dentry->d_name, &new_page);
if (!new_de)
goto out_dir;
ext2_set_link(new_dir, new_de, new_page, old_inode, 1);
new_inode->i_ctime = CURRENT_TIME_SEC;
if (dir_de)
drop_nlink(new_inode);
inode_dec_link_count(new_inode);
} else {
err = ext2_add_link(new_dentry, old_inode);
if (err)
goto out_dir;
if (dir_de)
inode_inc_link_count(new_dir);
}
/*
* Like most other Unix systems, set the ctime for inodes on a
* rename.
*/
old_inode->i_ctime = CURRENT_TIME_SEC;
mark_inode_dirty(old_inode);
ext2_delete_entry (old_de, old_page);
if (dir_de) {
if (old_dir != new_dir)
ext2_set_link(old_inode, dir_de, dir_page, new_dir, 0);
else {
kunmap(dir_page);
page_cache_release(dir_page);
}
inode_dec_link_count(old_dir);
}
return 0;
out_dir:
if (dir_de) {
kunmap(dir_page);
page_cache_release(dir_page);
}
out_old:
kunmap(old_page);
page_cache_release(old_page);
out:
return err;
}
/*
* NAME: jfs_unlink(dip, dentry)
*
* FUNCTION: remove a link to object <vp> named by <name>
* from parent directory <dvp>
*
* PARAMETER: dip - inode of parent directory
* dentry - dentry of object to be removed
*
* RETURN: errors from subroutines
*
* note:
* temporary file: if one or more processes have the file open
* when the last link is removed, the link will be removed before
* unlink() returns, but the removal of the file contents will be
* postponed until all references to the files are closed.
*
* JFS does NOT support unlink() on directories.
*
*/
static int jfs_unlink(struct inode *dip, struct dentry *dentry)
{
int rc;
tid_t tid; /* transaction id */
struct inode *ip = dentry->d_inode;
ino_t ino;
struct component_name dname; /* object name */
struct inode *iplist[2];
struct tblock *tblk;
s64 new_size = 0;
int commit_flag;
jfs_info("jfs_unlink: dip:0x%p name:%s", dip, dentry->d_name.name);
/* Init inode for quota operations. */
dquot_initialize(dip);
dquot_initialize(ip);
if ((rc = get_UCSname(&dname, dentry)))
goto out;
IWRITE_LOCK(ip, RDWRLOCK_NORMAL);
tid = txBegin(dip->i_sb, 0);
mutex_lock_nested(&JFS_IP(dip)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
iplist[0] = dip;
iplist[1] = ip;
/*
* delete the entry of target file from parent directory
*/
ino = ip->i_ino;
if ((rc = dtDelete(tid, dip, &dname, &ino, JFS_REMOVE))) {
jfs_err("jfs_unlink: dtDelete returned %d", rc);
if (rc == -EIO)
txAbort(tid, 1); /* Marks FS Dirty */
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
IWRITE_UNLOCK(ip);
goto out1;
}
ASSERT(ip->i_nlink);
ip->i_ctime = dip->i_ctime = dip->i_mtime = CURRENT_TIME;
mark_inode_dirty(dip);
/* update target's inode */
inode_dec_link_count(ip);
/*
* commit zero link count object
*/
if (ip->i_nlink == 0) {
assert(!test_cflag(COMMIT_Nolink, ip));
/* free block resources */
if ((new_size = commitZeroLink(tid, ip)) < 0) {
txAbort(tid, 1); /* Marks FS Dirty */
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
IWRITE_UNLOCK(ip);
rc = new_size;
goto out1;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = ip;
}
/*
* Incomplete truncate of file data can
* result in timing problems unless we synchronously commit the
* transaction.
*/
if (new_size)
commit_flag = COMMIT_SYNC;
else
commit_flag = 0;
/*
* If xtTruncate was incomplete, commit synchronously to avoid
* timing complications
*/
rc = txCommit(tid, 2, &iplist[0], commit_flag);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
while (new_size && (rc == 0)) {
tid = txBegin(dip->i_sb, 0);
mutex_lock(&JFS_IP(ip)->commit_mutex);
new_size = xtTruncate_pmap(tid, ip, new_size);
if (new_size < 0) {
txAbort(tid, 1); /* Marks FS Dirty */
rc = new_size;
} else
rc = txCommit(tid, 2, &iplist[0], COMMIT_SYNC);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
}
if (ip->i_nlink == 0)
set_cflag(COMMIT_Nolink, ip);
IWRITE_UNLOCK(ip);
/*
* Truncating the directory index table is not guaranteed. It
* may need to be done iteratively
*/
if (test_cflag(COMMIT_Stale, dip)) {
if (dip->i_size > 1)
jfs_truncate_nolock(dip, 0);
clear_cflag(COMMIT_Stale, dip);
}
out1:
free_UCSname(&dname);
out:
jfs_info("jfs_unlink: rc:%d", rc);
return rc;
}
/*
* NAME: jfs_rename
*
* FUNCTION: rename a file or directory
*/
static int jfs_rename(struct inode *old_dir, struct dentry *old_dentry,
struct inode *new_dir, struct dentry *new_dentry)
{
struct btstack btstack;
ino_t ino;
struct component_name new_dname;
struct inode *new_ip;
struct component_name old_dname;
struct inode *old_ip;
int rc;
tid_t tid;
struct tlock *tlck;
struct dt_lock *dtlck;
struct lv *lv;
int ipcount;
struct inode *iplist[4];
struct tblock *tblk;
s64 new_size = 0;
int commit_flag;
jfs_info("jfs_rename: %s %s", old_dentry->d_name.name,
new_dentry->d_name.name);
dquot_initialize(old_dir);
dquot_initialize(new_dir);
old_ip = old_dentry->d_inode;
new_ip = new_dentry->d_inode;
if ((rc = get_UCSname(&old_dname, old_dentry)))
goto out1;
if ((rc = get_UCSname(&new_dname, new_dentry)))
goto out2;
/*
* Make sure source inode number is what we think it is
*/
rc = dtSearch(old_dir, &old_dname, &ino, &btstack, JFS_LOOKUP);
if (rc || (ino != old_ip->i_ino)) {
rc = -ENOENT;
goto out3;
}
/*
* Make sure dest inode number (if any) is what we think it is
*/
rc = dtSearch(new_dir, &new_dname, &ino, &btstack, JFS_LOOKUP);
if (!rc) {
if ((!new_ip) || (ino != new_ip->i_ino)) {
rc = -ESTALE;
goto out3;
}
} else if (rc != -ENOENT)
goto out3;
else if (new_ip) {
/* no entry exists, but one was expected */
rc = -ESTALE;
goto out3;
}
if (S_ISDIR(old_ip->i_mode)) {
if (new_ip) {
if (!dtEmpty(new_ip)) {
rc = -ENOTEMPTY;
goto out3;
}
} else if ((new_dir != old_dir) &&
(new_dir->i_nlink == JFS_LINK_MAX)) {
rc = -EMLINK;
goto out3;
}
} else if (new_ip) {
IWRITE_LOCK(new_ip, RDWRLOCK_NORMAL);
/* Init inode for quota operations. */
dquot_initialize(new_ip);
}
/*
* The real work starts here
*/
tid = txBegin(new_dir->i_sb, 0);
/*
* How do we know the locking is safe from deadlocks?
* The vfs does the hard part for us. Any time we are taking nested
* commit_mutexes, the vfs already has i_mutex held on the parent.
* Here, the vfs has already taken i_mutex on both old_dir and new_dir.
*/
mutex_lock_nested(&JFS_IP(new_dir)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(old_ip)->commit_mutex, COMMIT_MUTEX_CHILD);
if (old_dir != new_dir)
mutex_lock_nested(&JFS_IP(old_dir)->commit_mutex,
COMMIT_MUTEX_SECOND_PARENT);
if (new_ip) {
mutex_lock_nested(&JFS_IP(new_ip)->commit_mutex,
COMMIT_MUTEX_VICTIM);
/*
* Change existing directory entry to new inode number
*/
ino = new_ip->i_ino;
rc = dtModify(tid, new_dir, &new_dname, &ino,
old_ip->i_ino, JFS_RENAME);
if (rc)
goto out4;
drop_nlink(new_ip);
if (S_ISDIR(new_ip->i_mode)) {
drop_nlink(new_ip);
if (new_ip->i_nlink) {
mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
if (old_dir != new_dir)
mutex_unlock(&JFS_IP(old_dir)->commit_mutex);
mutex_unlock(&JFS_IP(old_ip)->commit_mutex);
mutex_unlock(&JFS_IP(new_dir)->commit_mutex);
if (!S_ISDIR(old_ip->i_mode) && new_ip)
IWRITE_UNLOCK(new_ip);
jfs_error(new_ip->i_sb,
"jfs_rename: new_ip->i_nlink != 0");
return -EIO;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = new_ip;
} else if (new_ip->i_nlink == 0) {
assert(!test_cflag(COMMIT_Nolink, new_ip));
/* free block resources */
if ((new_size = commitZeroLink(tid, new_ip)) < 0) {
txAbort(tid, 1); /* Marks FS Dirty */
rc = new_size;
goto out4;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = new_ip;
} else {
new_ip->i_ctime = CURRENT_TIME;
mark_inode_dirty(new_ip);
}
} else {
/*
* Add new directory entry
*/
rc = dtSearch(new_dir, &new_dname, &ino, &btstack,
JFS_CREATE);
if (rc) {
jfs_err("jfs_rename didn't expect dtSearch to fail "
"w/rc = %d", rc);
goto out4;
}
ino = old_ip->i_ino;
rc = dtInsert(tid, new_dir, &new_dname, &ino, &btstack);
if (rc) {
if (rc == -EIO)
jfs_err("jfs_rename: dtInsert returned -EIO");
goto out4;
}
if (S_ISDIR(old_ip->i_mode))
inc_nlink(new_dir);
}
/*
* Remove old directory entry
*/
ino = old_ip->i_ino;
rc = dtDelete(tid, old_dir, &old_dname, &ino, JFS_REMOVE);
if (rc) {
jfs_err("jfs_rename did not expect dtDelete to return rc = %d",
rc);
txAbort(tid, 1); /* Marks Filesystem dirty */
goto out4;
}
if (S_ISDIR(old_ip->i_mode)) {
drop_nlink(old_dir);
if (old_dir != new_dir) {
/*
* Change inode number of parent for moved directory
*/
JFS_IP(old_ip)->i_dtroot.header.idotdot =
cpu_to_le32(new_dir->i_ino);
/* Linelock header of dtree */
tlck = txLock(tid, old_ip,
(struct metapage *) &JFS_IP(old_ip)->bxflag,
tlckDTREE | tlckBTROOT | tlckRELINK);
dtlck = (struct dt_lock *) & tlck->lock;
ASSERT(dtlck->index == 0);
lv = & dtlck->lv[0];
lv->offset = 0;
lv->length = 1;
dtlck->index++;
}
}
/*
* Update ctime on changed/moved inodes & mark dirty
*/
old_ip->i_ctime = CURRENT_TIME;
mark_inode_dirty(old_ip);
new_dir->i_ctime = new_dir->i_mtime = current_fs_time(new_dir->i_sb);
mark_inode_dirty(new_dir);
/* Build list of inodes modified by this transaction */
ipcount = 0;
iplist[ipcount++] = old_ip;
if (new_ip)
iplist[ipcount++] = new_ip;
iplist[ipcount++] = old_dir;
if (old_dir != new_dir) {
iplist[ipcount++] = new_dir;
old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME;
mark_inode_dirty(old_dir);
}
/*
* Incomplete truncate of file data can
* result in timing problems unless we synchronously commit the
* transaction.
*/
if (new_size)
commit_flag = COMMIT_SYNC;
else
commit_flag = 0;
rc = txCommit(tid, ipcount, iplist, commit_flag);
out4:
txEnd(tid);
if (new_ip)
mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
if (old_dir != new_dir)
mutex_unlock(&JFS_IP(old_dir)->commit_mutex);
mutex_unlock(&JFS_IP(old_ip)->commit_mutex);
mutex_unlock(&JFS_IP(new_dir)->commit_mutex);
while (new_size && (rc == 0)) {
tid = txBegin(new_ip->i_sb, 0);
mutex_lock(&JFS_IP(new_ip)->commit_mutex);
new_size = xtTruncate_pmap(tid, new_ip, new_size);
if (new_size < 0) {
txAbort(tid, 1);
rc = new_size;
} else
rc = txCommit(tid, 1, &new_ip, COMMIT_SYNC);
txEnd(tid);
mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
}
if (new_ip && (new_ip->i_nlink == 0))
set_cflag(COMMIT_Nolink, new_ip);
out3:
free_UCSname(&new_dname);
out2:
free_UCSname(&old_dname);
out1:
if (new_ip && !S_ISDIR(new_ip->i_mode))
IWRITE_UNLOCK(new_ip);
/*
* Truncating the directory index table is not guaranteed. It
* may need to be done iteratively
*/
if (test_cflag(COMMIT_Stale, old_dir)) {
if (old_dir->i_size > 1)
jfs_truncate_nolock(old_dir, 0);
clear_cflag(COMMIT_Stale, old_dir);
}
jfs_info("jfs_rename: returning %d", rc);
return rc;
}
/*
* NAME: jfs_create(dip, dentry, mode)
*
* FUNCTION: create a regular file in the parent directory <dip>
* with name = <from dentry> and mode = <mode>
*
* PARAMETER: dip - parent directory vnode
* dentry - dentry of new file
* mode - create mode (rwxrwxrwx).
* nd- nd struct
*
* RETURN: Errors from subroutines
*
*/
static int jfs_create(struct inode *dip, struct dentry *dentry, int mode,
struct nameidata *nd)
{
int rc = 0;
tid_t tid; /* transaction id */
struct inode *ip = NULL; /* child directory inode */
ino_t ino;
struct component_name dname; /* child directory name */
struct btstack btstack;
struct inode *iplist[2];
struct tblock *tblk;
jfs_info("jfs_create: dip:0x%p name:%s", dip, dentry->d_name.name);
dquot_initialize(dip);
/*
* search parent directory for entry/freespace
* (dtSearch() returns parent directory page pinned)
*/
if ((rc = get_UCSname(&dname, dentry)))
goto out1;
/*
* Either iAlloc() or txBegin() may block. Deadlock can occur if we
* block there while holding dtree page, so we allocate the inode &
* begin the transaction before we search the directory.
*/
ip = ialloc(dip, mode);
if (IS_ERR(ip)) {
rc = PTR_ERR(ip);
goto out2;
}
tid = txBegin(dip->i_sb, 0);
mutex_lock_nested(&JFS_IP(dip)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
rc = jfs_init_acl(tid, ip, dip);
if (rc)
goto out3;
rc = jfs_init_security(tid, ip, dip, &dentry->d_name);
if (rc) {
txAbort(tid, 0);
goto out3;
}
if ((rc = dtSearch(dip, &dname, &ino, &btstack, JFS_CREATE))) {
jfs_err("jfs_create: dtSearch returned %d", rc);
txAbort(tid, 0);
goto out3;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_CREATE;
tblk->ino = ip->i_ino;
tblk->u.ixpxd = JFS_IP(ip)->ixpxd;
iplist[0] = dip;
iplist[1] = ip;
/*
* initialize the child XAD tree root in-line in inode
*/
xtInitRoot(tid, ip);
/*
* create entry in parent directory for child directory
* (dtInsert() releases parent directory page)
*/
ino = ip->i_ino;
if ((rc = dtInsert(tid, dip, &dname, &ino, &btstack))) {
if (rc == -EIO) {
jfs_err("jfs_create: dtInsert returned -EIO");
txAbort(tid, 1); /* Marks Filesystem dirty */
} else
txAbort(tid, 0); /* Filesystem full */
goto out3;
}
ip->i_op = &jfs_file_inode_operations;
ip->i_fop = &jfs_file_operations;
ip->i_mapping->a_ops = &jfs_aops;
mark_inode_dirty(ip);
dip->i_ctime = dip->i_mtime = CURRENT_TIME;
mark_inode_dirty(dip);
rc = txCommit(tid, 2, &iplist[0], 0);
out3:
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
if (rc) {
free_ea_wmap(ip);
ip->i_nlink = 0;
unlock_new_inode(ip);
iput(ip);
} else {
d_instantiate(dentry, ip);
unlock_new_inode(ip);
}
out2:
free_UCSname(&dname);
out1:
jfs_info("jfs_create: rc:%d", rc);
return rc;
}