1ioctl_iflags(2) System Calls Manual ioctl_iflags(2)
2
3
4
6 ioctl_iflags - ioctl() operations for inode flags
7
9 Various Linux filesystems support the notion of inode flags—attributes
10 that modify the semantics of files and directories. These flags can be
11 retrieved and modified using two ioctl(2) operations:
12
13 int attr;
14 fd = open("pathname", ...);
15
16 ioctl(fd, FS_IOC_GETFLAGS, &attr); /* Place current flags
17 in 'attr' */
18 attr |= FS_NOATIME_FL; /* Tweak returned bit mask */
19 ioctl(fd, FS_IOC_SETFLAGS, &attr); /* Update flags for inode
20 referred to by 'fd' */
21
22 The lsattr(1) and chattr(1) shell commands provide interfaces to these
23 two operations, allowing a user to view and modify the inode flags as‐
24 sociated with a file.
25
26 The following flags are supported (shown along with the corresponding
27 letter used to indicate the flag by lsattr(1) and chattr(1)):
28
29 FS_APPEND_FL 'a'
30 The file can be opened only with the O_APPEND flag. (This re‐
31 striction applies even to the superuser.) Only a privileged
32 process (CAP_LINUX_IMMUTABLE) can set or clear this attribute.
33
34 FS_COMPR_FL 'c'
35 Store the file in a compressed format on disk. This flag is not
36 supported by most of the mainstream filesystem implementations;
37 one exception is btrfs(5).
38
39 FS_DIRSYNC_FL 'D' (since Linux 2.6.0)
40 Write directory changes synchronously to disk. This flag pro‐
41 vides semantics equivalent to the mount(2) MS_DIRSYNC option,
42 but on a per-directory basis. This flag can be applied only to
43 directories.
44
45 FS_IMMUTABLE_FL 'i'
46 The file is immutable: no changes are permitted to the file con‐
47 tents or metadata (permissions, timestamps, ownership, link
48 count, and so on). (This restriction applies even to the supe‐
49 ruser.) Only a privileged process (CAP_LINUX_IMMUTABLE) can set
50 or clear this attribute.
51
52 FS_JOURNAL_DATA_FL 'j'
53 Enable journaling of file data on ext3(5) and ext4(5) filesys‐
54 tems. On a filesystem that is journaling in ordered or write‐
55 back mode, a privileged (CAP_SYS_RESOURCE) process can set this
56 flag to enable journaling of data updates on a per-file basis.
57
58 FS_NOATIME_FL 'A'
59 Don't update the file last access time when the file is ac‐
60 cessed. This can provide I/O performance benefits for applica‐
61 tions that do not care about the accuracy of this timestamp.
62 This flag provides functionality similar to the mount(2) MS_NOA‐
63 TIME flag, but on a per-file basis.
64
65 FS_NOCOW_FL 'C' (since Linux 2.6.39)
66 The file will not be subject to copy-on-write updates. This
67 flag has an effect only on filesystems that support copy-on-
68 write semantics, such as Btrfs. See chattr(1) and btrfs(5).
69
70 FS_NODUMP_FL 'd'
71 Don't include this file in backups made using dump(8).
72
73 FS_NOTAIL_FL 't'
74 This flag is supported only on Reiserfs. It disables the Reis‐
75 erfs tail-packing feature, which tries to pack small files (and
76 the final fragment of larger files) into the same disk block as
77 the file metadata.
78
79 FS_PROJINHERIT_FL 'P' (since Linux 4.5)
80 Inherit the quota project ID. Files and subdirectories will in‐
81 herit the project ID of the directory. This flag can be applied
82 only to directories.
83
84 FS_SECRM_FL 's'
85 Mark the file for secure deletion. This feature is not imple‐
86 mented by any filesystem, since the task of securely erasing a
87 file from a recording medium is surprisingly difficult.
88
89 FS_SYNC_FL 'S'
90 Make file updates synchronous. For files, this makes all writes
91 synchronous (as though all opens of the file were with the
92 O_SYNC flag). For directories, this has the same effect as the
93 FS_DIRSYNC_FL flag.
94
95 FS_TOPDIR_FL 'T'
96 Mark a directory for special treatment under the Orlov block-al‐
97 location strategy. See chattr(1) for details. This flag can be
98 applied only to directories and has an effect only for ext2,
99 ext3, and ext4.
100
101 FS_UNRM_FL 'u'
102 Allow the file to be undeleted if it is deleted. This feature
103 is not implemented by any filesystem, since it is possible to
104 implement file-recovery mechanisms outside the kernel.
105
106 In most cases, when any of the above flags is set on a directory, the
107 flag is inherited by files and subdirectories created inside that di‐
108 rectory. Exceptions include FS_TOPDIR_FL, which is not inheritable,
109 and FS_DIRSYNC_FL, which is inherited only by subdirectories.
110
112 Linux.
113
115 In order to change the inode flags of a file using the FS_IOC_SETFLAGS
116 operation, the effective user ID of the caller must match the owner of
117 the file, or the caller must have the CAP_FOWNER capability.
118
119 The type of the argument given to the FS_IOC_GETFLAGS and FS_IOC_SET‐
120 FLAGS operations is int *, notwithstanding the implication in the ker‐
121 nel source file include/uapi/linux/fs.h that the argument is long *.
122
124 chattr(1), lsattr(1), mount(2), btrfs(5), ext4(5), xfs(5), xattr(7),
125 mount(8)
126
127
128
129Linux man-pages 6.05 2023-05-03 ioctl_iflags(2)