ディレクトリ(ext3)
ディレクトリは作成されるファイル等の領域でなく、ファイルシステム依存のstruct ext3_dir_entry_2ファイル属性のデータテキストファイルに相当します。read()での読み込には .readコールバックで、generic_read_dir()は-EISDIRでできません。readdir()はシステムコールgetdentで.readdirコールされ、ext3_readdir()がstruct ext3_dir_entry_2のテキストファイルをgccのstruct direntに設定します。
ファイル数のディレクトリ毎の制限はありませんが、ディレクトリの係るサイズはブロックサイズ単位の領域で、新規に作成されたディレクトリサイズは1ブロックサイズで、デバイスの有効利用実装のため、ディレクトリとして管理するディレクトリ数はEXT3_LINK_MAX=32000です。これはファイルシステムに依存し、ext2ではディレクトリにかかる制約はありません。
ファイル数のディレクトリ毎の制限はありませんが、ディレクトリの係るサイズはブロックサイズ単位の領域で、新規に作成されたディレクトリサイズは1ブロックサイズで、デバイスの有効利用実装のため、ディレクトリとして管理するディレクトリ数はEXT3_LINK_MAX=32000です。これはファイルシステムに依存し、ext2ではディレクトリにかかる制約はありません。
#define EXT3_NAME_LEN 255 ファイル名長さ
struct ext3_dir_entry_2 {
__le32 inode; /* Inode number */
__le16 rec_len; /* Directory entry length */
__u8 name_len; /* Name length */
__u8 file_type;
char name[EXT3_NAME_LEN]; /* File name */
};
const struct file_operations ext3_dir_operations = {
.llseek = ext3_dir_llseek,
.read = generic_read_dir,
.readdir = ext3_readdir,
.unlocked_ioctl = ext3_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = ext3_compat_ioctl,
#endif
.fsync = ext3_sync_file,
.release = ext3_release_dir,
};
ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t *ppos)
{
return -EISDIR;
}
const struct inode_operations ext3_dir_inode_operations = {
.create = ext3_create,
.lookup = ext3_lookup,
.link = ext3_link,
.unlink = ext3_unlink,
.symlink = ext3_symlink,
.mkdir = ext3_mkdir,
.rmdir = ext3_rmdir,
.mknod = ext3_mknod,
.rename = ext3_rename,
.setattr = ext3_setattr,
#ifdef CONFIG_EXT3_FS_XATTR
.setxattr = generic_setxattr,
.getxattr = generic_getxattr,
.listxattr = ext3_listxattr,
.removexattr = generic_removexattr,
#endif
.get_acl = ext3_get_acl,
};
static int ext3_create (struct inode * dir, struct dentry * dentry, umode_t mode,
bool excl)
{
handle_t *handle;
struct inode * inode;
int err, retries = 0;
dquot_initialize(dir);
retry:
handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS(dir->i_sb) +
EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3 +
EXT3_MAXQUOTAS_INIT_BLOCKS(dir->i_sb));
if (IS_ERR(handle))
return PTR_ERR(handle);
if (IS_DIRSYNC(dir))
handle->h_sync = 1;
inode = ext3_new_inode (handle, dir, &dentry->d_name, mode);
err = PTR_ERR(inode);
if (!IS_ERR(inode)) {
inode->i_op = &ext3_file_inode_operations;
inode->i_fop = &ext3_file_operations;
ext3_set_aops(inode);
err = ext3_add_nondir(handle, dentry, inode);
}
ext3_journal_stop(handle);
if (err == -ENOSPC && ext3_should_retry_alloc(dir->i_sb, &retries))
goto retry;
return err;
}
static int ext3_add_nondir(handle_t *handle,
struct dentry *dentry, struct inode *inode)
{
int err = ext3_add_entry(handle, dentry, inode);
if (!err) {
ext3_mark_inode_dirty(handle, inode);
unlock_new_inode(inode);
d_instantiate(dentry, inode);
return 0;
}
drop_nlink(inode);
unlock_new_inode(inode);
iput(inode);
return err;
}
static int ext3_add_entry (handle_t *handle, struct dentry *dentry,
struct inode *inode)
{
struct inode *dir = dentry->d_parent->d_inode;
struct buffer_head * bh;
struct ext3_dir_entry_2 *de;
struct super_block * sb;
int retval;
int dx_fallback=0;
unsigned blocksize;
u32 block, blocks;
sb = dir->i_sb;
blocksize = sb->s_blocksize;
if (!dentry->d_name.len)
return -EINVAL;
if (is_dx(dir)) {
retval = ext3_dx_add_entry(handle, dentry, inode);
if (!retval || (retval != ERR_BAD_DX_DIR))
return retval;
EXT3_I(dir)->i_flags &= ~EXT3_INDEX_FL;
dx_fallback++;
ext3_mark_inode_dirty(handle, dir);
}
blocks = dir->i_size >> sb->s_blocksize_bits;
for (block = 0; block < blocks; block++) {
if (!(bh = ext3_dir_bread(handle, dir, block, 0, &retval)))
return retval;
retval = add_dirent_to_buf(handle, dentry, inode, NULL, bh);
if (retval != -ENOSPC)
return retval;
if (blocks == 1 && !dx_fallback &&
EXT3_HAS_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_DIR_INDEX))
return make_indexed_dir(handle, dentry, inode, bh);
brelse(bh);
}
bh = ext3_append(handle, dir, &block, &retval);
if (!bh)
return retval;
de = (struct ext3_dir_entry_2 *) bh->b_data;
de->inode = 0;
de->rec_len = ext3_rec_len_to_disk(blocksize);
return add_dirent_to_buf(handle, dentry, inode, de, bh);
}
static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry,
struct inode *inode, struct ext3_dir_entry_2 *de,
struct buffer_head * bh)
{
struct inode *dir = dentry->d_parent->d_inode;
const char *name = dentry->d_name.name;
int namelen = dentry->d_name.len;
unsigned long offset = 0;
unsigned short reclen;
int nlen, rlen, err;
char *top;
reclen = EXT3_DIR_REC_LEN(namelen);
if (!de) {
de = (struct ext3_dir_entry_2 *)bh->b_data;
top = bh->b_data + dir->i_sb->s_blocksize - reclen;
while ((char *) de <= top) {
if (!ext3_check_dir_entry("ext3_add_entry", dir, de,
bh, offset)) {
brelse (bh);
return -EIO;
}
if (ext3_match (namelen, name, de)) {
brelse (bh);
return -EEXIST;
}
nlen = EXT3_DIR_REC_LEN(de->name_len);
rlen = ext3_rec_len_from_disk(de->rec_len);
if ((de->inode? rlen - nlen: rlen) >= reclen)
break;
de = (struct ext3_dir_entry_2 *)((char *)de + rlen);
offset += rlen;
}
if ((char *) de > top)
return -ENOSPC;
}
BUFFER_TRACE(bh, "get_write_access");
err = ext3_journal_get_write_access(handle, bh);
if (err) {
ext3_std_error(dir->i_sb, err);
brelse(bh);
return err;
}
nlen = EXT3_DIR_REC_LEN(de->name_len);
rlen = ext3_rec_len_from_disk(de->rec_len);
if (de->inode) {
struct ext3_dir_entry_2 *de1 = (struct ext3_dir_entry_2 *)((char *)de + nlen);
de1->rec_len = ext3_rec_len_to_disk(rlen - nlen);
de->rec_len = ext3_rec_len_to_disk(nlen);
de = de1;
}
de->file_type = EXT3_FT_UNKNOWN;
if (inode) {
de->inode = cpu_to_le32(inode->i_ino);
ext3_set_de_type(dir->i_sb, de, inode->i_mode);
} else
de->inode = 0;
de->name_len = namelen;
memcpy (de->name, name, namelen);
dir->i_mtime = dir->i_ctime = CURRENT_TIME_SEC;
ext3_update_dx_flag(dir);
dir->i_version++;
ext3_mark_inode_dirty(handle, dir);
BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
err = ext3_journal_dirty_metadata(handle, bh);
if (err)
ext3_std_error(dir->i_sb, err);
brelse(bh);
return 0;
}
ext3_dir_entry_2の.file_typeはファイルシステム依存で、ext3_readdir()で取得される.file_typeのd_typeは、.file_typeをインデックスとするext3_filetype_table[]で、故にファイルシステムに関係なくDT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNKとなります。
struct ext3_dir_entry_2 {
__le32 inode; /* Inode number */
__le16 rec_len; /* Directory entry length */
__u8 name_len; /* Name length */
__u8 file_type;
char name[EXT3_NAME_LEN]; /* File name */
};
static inline void ext3_set_de_type(struct super_block *sb,
struct ext3_dir_entry_2 *de,
umode_t mode) {
if (EXT3_HAS_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_FILETYPE))
de->file_type = ext3_type_by_mode[(mode & S_IFMT)>>S_SHIFT];
}
#define DT_UNKNOWN 0
#define DT_FIFO 1
#define DT_CHR 2
#define DT_DIR 4
#define DT_BLK 6
#define DT_REG 8
#define DT_LNK 10
#define DT_SOCK 12
#define DT_WHT 14
static unsigned char ext3_filetype_table[] = {
DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
};
#define S_SHIFT 12
#define S_IFMT 00170000
#define S_IFSOCK 0140000
#define S_IFLNK 0120000
#define S_IFREG 0100000
#define S_IFBLK 0060000
#define S_IFDIR 0040000
#define S_IFCHR 0020000
#define S_IFIFO 0010000
#define S_ISUID 0004000
#define S_ISGID 0002000
#define S_ISVTX 0001000
#define EXT3_FT_UNKNOWN 0
#define EXT3_FT_REG_FILE 1
#define EXT3_FT_DIR 2
#define EXT3_FT_CHRDEV 3
#define EXT3_FT_BLKDEV 4
#define EXT3_FT_FIFO 5
#define EXT3_FT_SOCK 6
#define EXT3_FT_SYMLINK 7
static unsigned char ext3_type_by_mode[S_IFMT >> S_SHIFT] = {
[S_IFREG >> S_SHIFT] = EXT3_FT_REG_FILE,
[S_IFDIR >> S_SHIFT] = EXT3_FT_DIR,
[S_IFCHR >> S_SHIFT] = EXT3_FT_CHRDEV,
[S_IFBLK >> S_SHIFT] = EXT3_FT_BLKDEV,
[S_IFIFO >> S_SHIFT] = EXT3_FT_FIFO,
[S_IFSOCK >> S_SHIFT] = EXT3_FT_SOCK,
[S_IFLNK >> S_SHIFT] = EXT3_FT_SYMLINK,
};
static int ext3_readdir(struct file * filp,
void * dirent, filldir_t filldir)
{
int error = 0;
unsigned long offset;
int i, stored;
struct ext3_dir_entry_2 *de;
int err;
struct inode *inode = filp->f_path.dentry->d_inode;
struct super_block *sb = inode->i_sb;
int ret = 0;
int dir_has_error = 0;
if (is_dx_dir(inode)) {
err = ext3_dx_readdir(filp, dirent, filldir);
if (err != ERR_BAD_DX_DIR) {
ret = err;
goto out;
}
EXT3_I(filp->f_path.dentry->d_inode)->i_flags &= ~EXT3_INDEX_FL;
}
stored = 0;
offset = filp->f_pos & (sb->s_blocksize - 1);
while (!error && !stored && filp->f_pos < inode->i_size) {
unsigned long blk = filp->f_pos >> EXT3_BLOCK_SIZE_BITS(sb);
struct buffer_head map_bh;
struct buffer_head *bh = NULL;
map_bh.b_state = 0;
err = ext3_get_blocks_handle(NULL, inode, blk, 1, &map_bh, 0);
if (err > 0) {
pgoff_t index = map_bh.b_blocknr >>
(PAGE_CACHE_SHIFT - inode->i_blkbits);
if (!ra_has_index(&filp->f_ra, index))
page_cache_sync_readahead(
sb->s_bdev->bd_inode->i_mapping,
&filp->f_ra, filp,
index, 1);
filp->f_ra.prev_pos = (loff_t)index << PAGE_CACHE_SHIFT;
bh = ext3_bread(NULL, inode, blk, 0, &err);
}
if (!bh) {
if (!dir_has_error) {
ext3_error(sb, __func__, "directory #%lu "
"contains a hole at offset %lld",
inode->i_ino, filp->f_pos);
dir_has_error = 1;
}
/* corrupt size? Maybe no more blocks to read */
if (filp->f_pos > inode->i_blocks << 9)
break;
filp->f_pos += sb->s_blocksize - offset;
continue;
}
revalidate:
if (filp->f_version != inode->i_version) {
for (i = 0; i < sb->s_blocksize && i < offset; ) {
de = (struct ext3_dir_entry_2 *)
(bh->b_data + i);
if (ext3_rec_len_from_disk(de->rec_len) <
EXT3_DIR_REC_LEN(1))
break;
i += ext3_rec_len_from_disk(de->rec_len);
}
offset = i;
filp->f_pos = (filp->f_pos & ~(sb->s_blocksize - 1))
| offset;
filp->f_version = inode->i_version;
}
while (!error && filp->f_pos < inode->i_size
&& offset < sb->s_blocksize) {
de = (struct ext3_dir_entry_2 *) (bh->b_data + offset);
if (!ext3_check_dir_entry ("ext3_readdir", inode, de,
bh, offset)) {
filp->f_pos = (filp->f_pos |
(sb->s_blocksize - 1)) + 1;
brelse (bh);
ret = stored;
goto out;
}
offset += ext3_rec_len_from_disk(de->rec_len);
if (le32_to_cpu(de->inode)) {
u64 version = filp->f_version;
error = filldir(dirent, de->name,
de->name_len,
filp->f_pos,
le32_to_cpu(de->inode),
get_dtype(sb, de->file_type));
if (error)
break;
if (version != filp->f_version)
goto revalidate;
stored ++;
}
filp->f_pos += ext3_rec_len_from_disk(de->rec_len);
}
offset = 0;
brelse (bh);
}
out:
return ret;
}
static unsigned char get_dtype(struct super_block *sb, int filetype)
{
if (!EXT3_HAS_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_FILETYPE) ||
(filetype >= EXT3_FT_MAX))
return DT_UNKNOWN;
return (ext3_filetype_table[filetype]);
}
検証
[root@localhost north]# cat dir.c
#include <stdio.h>
#include <stdlib.h>
#include <dirent.h>
void main(int argc, char *argv[])
{
DIR *dir;
struct dirent *dp;
dir=opendir(argv[1]);
printf ("%7s %10s %3s %4s %s\n", "inode", "off", "len", "type", "name");
while (1) {
dp = readdir(dir);
if (!dp) {
break;
}
printf ("%07d %10d %03d %04d %s\n",
dp->d_ino,
dp->d_off,
dp->d_reclen,
dp->d_type,
dp->d_name);
free(dp);
}
}
[root@localhost north]# touch babakaka/text
[root@localhost north]# mkdir babakaka/dir
[root@localhost north]# mkfifo babakaka/fifo
[root@localhost north]# mknod babakaka/blockdev b 10 1
[root@localhost north]# mknod babakaka/chardev c 11 1
[root@localhost north]# ln -s babakaka/text babakaka/symlink
[root@localhost north]# ln babakaka/text babakaka/hardlink
[root@localhost north]# ./dir babakaka/
inode off len type name
1802243 698206621 016 0008 text DT_REG : 8
1802242 942110034 016 0004 dir DT_DIR : 4
0000002 1106706763 016 0004 .. DT_DIR : 4
1802247 1162905680 020 0010 symlink DT_LNK :10
1802245 1470500359 020 0006 blockdev DT_BLK : 6
1802243 1475777499 020 0008 hardlink DT_REG : 8 hardlinkはinodeを共有での実装です。
1802241 1852366747 016 0004 . DT_DIR : 4
1802246 1936624415 020 0002 chardev DT_CHR : 2
1802244 2147483647 016 0001 fifo DT_FIFO: 1
[root@localhost north]# ls -lai babakaka/
合計 12
1802241 drwxr-xr-x 3 root root 4096 7月 7 19:46 .
2 drwxr-xr-x 5 root root 4096 7月 7 20:23 ..
1802245 brw-r--r-- 1 root root 10, 1 7月 7 19:17 blockdev
1802246 crw-r--r-- 1 root root 11, 1 7月 7 19:46 chardev
1802242 drwxr-xr-x 2 root root 4096 7月 7 19:15 dir
1802244 prw-r--r-- 1 root root 0 7月 7 19:16 fifo
1802243 -rw-r--r-- 2 root root 0 7月 7 19:15 hardlink
1802247 lrwxrwxrwx 1 root root 13 7月 7 19:18 symlink -> babakaka/text
1802243 -rw-r--r-- 2 root root 0 7月 7 19:15 text
struct linux_dirent {
unsigned long d_ino;
unsigned long d_off;
unsigned short d_reclen;
char d_name[1];
};
SYSCALL_DEFINE3(getdents, unsigned int, fd,
struct linux_dirent __user *, dirent, unsigned int, count)
{
struct fd f;
struct linux_dirent __user * lastdirent;
struct getdents_callback buf;
int error;
if (!access_ok(VERIFY_WRITE, dirent, count))
return -EFAULT;
f = fdget(fd);
if (!f.file)
return -EBADF;
buf.current_dir = dirent;
buf.previous = NULL;
buf.count = count;
buf.error = 0;
error = vfs_readdir(f.file, filldir, &buf);
if (error >= 0)
error = buf.error;
lastdirent = buf.previous;
if (lastdirent) {
if (put_user(f.file->f_pos, &lastdirent->d_off))
error = -EFAULT;
else
error = count - buf.count;
}
fdput(f);
return error;
}
作成できるディレクトリ下のファイルの制約はありませんが、ディレクトリはEXT3_LINK_MAX=32000までです。ディレクトリはファイルと異り、ブロックサイズの領域を有する故の実装かと。
struct dirent
{
#ifndef __USE_FILE_OFFSET64
__ino_t d_ino;
__off_t d_off;
#else
__ino64_t d_ino;
__off64_t d_off;
#endif
unsigned short int d_reclen;
unsigned char d_type;
char d_name[256]; /* We must not include limits.h! */
};
#define EXT3_LINK_MAX 32000
static int ext3_mkdir(struct inode * dir, struct dentry * dentry, umode_t mode)
{
handle_t *handle;
struct inode * inode;
struct buffer_head * dir_block = NULL;
struct ext3_dir_entry_2 * de;
int err, retries = 0;
if (dir->i_nlink >= EXT3_LINK_MAX)
return -EMLINK;
dquot_initialize(dir);
retry:
handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS(dir->i_sb) +
EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3 +
EXT3_MAXQUOTAS_INIT_BLOCKS(dir->i_sb));
if (IS_ERR(handle))
return PTR_ERR(handle);
if (IS_DIRSYNC(dir))
handle->h_sync = 1;
inode = ext3_new_inode (handle, dir, &dentry->d_name, S_IFDIR | mode);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out_stop;
inode->i_op = &ext3_dir_inode_operations;
inode->i_fop = &ext3_dir_operations;
inode->i_size = EXT3_I(inode)->i_disksize = inode->i_sb->s_blocksize;
if (!(dir_block = ext3_dir_bread(handle, inode, 0, 1, &err)))
goto out_clear_inode;
BUFFER_TRACE(dir_block, "get_write_access");
err = ext3_journal_get_write_access(handle, dir_block);
if (err)
goto out_clear_inode;
de = (struct ext3_dir_entry_2 *) dir_block->b_data;
de->inode = cpu_to_le32(inode->i_ino);
de->name_len = 1;
de->rec_len = ext3_rec_len_to_disk(EXT3_DIR_REC_LEN(de->name_len));
strcpy (de->name, ".");
ext3_set_de_type(dir->i_sb, de, S_IFDIR);
de = ext3_next_entry(de);
de->inode = cpu_to_le32(dir->i_ino);
de->rec_len = ext3_rec_len_to_disk(inode->i_sb->s_blocksize -
EXT3_DIR_REC_LEN(1));
de->name_len = 2;
strcpy (de->name, "..");
ext3_set_de_type(dir->i_sb, de, S_IFDIR);
set_nlink(inode, 2);
BUFFER_TRACE(dir_block, "call ext3_journal_dirty_metadata");
err = ext3_journal_dirty_metadata(handle, dir_block);
if (err)
goto out_clear_inode;
err = ext3_mark_inode_dirty(handle, inode);
if (!err)
err = ext3_add_entry (handle, dentry, inode);
if (err) {
out_clear_inode:
clear_nlink(inode);
unlock_new_inode(inode);
ext3_mark_inode_dirty(handle, inode);
iput (inode);
goto out_stop;
}
inc_nlink(dir);
ext3_update_dx_flag(dir);
err = ext3_mark_inode_dirty(handle, dir);
if (err)
goto out_clear_inode;
unlock_new_inode(inode);
d_instantiate(dentry, inode);
out_stop:
brelse(dir_block);
ext3_journal_stop(handle);
if (err == -ENOSPC && ext3_should_retry_alloc(dir->i_sb, &retries))
goto retry;
return err;
}
void inc_nlink(struct inode *inode)
{
if (WARN_ON(inode->i_nlink == 0))
atomic_long_dec(&inode->i_sb->s_remove_count);
inode->__i_nlink++;
}
cdコマンドはdirに変移するのでなく、current->fs->pwdにdirを設定して、パス捜査はcurrent->fs->pwdのdirファイルを読み込むことで運用されます。
SYSCALL_DEFINE1(chdir, const char __user *, filename)
{
struct path path;
int error;
error = user_path_dir(filename, &path);
if (error)
goto out;
error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_CHDIR);
if (error)
goto dput_and_out;
set_fs_pwd(current->fs, &path);
dput_and_out:
path_put(&path);
out:
return error;
}
void set_fs_pwd(struct fs_struct *fs, struct path *path)
{
struct path old_pwd;
path_get(path);
spin_lock(&fs->lock);
write_seqcount_begin(&fs->seq);
old_pwd = fs->pwd;
fs->pwd = *path;
write_seqcount_end(&fs->seq);
spin_unlock(&fs->lock);
if (old_pwd.dentry)
path_put(&old_pwd);
}
追記
ext2ファイルシステムは、dir->i_nlinkかかる制約はありません。
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;
}
