1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 Orphan file 3 Orphan file
4 ----------- 4 -----------
5 5
6 In unix there can inodes that are unlinked fro 6 In unix there can inodes that are unlinked from directory hierarchy but that
7 are still alive because they are open. In case 7 are still alive because they are open. In case of crash the filesystem has to
8 clean up these inodes as otherwise they (and t 8 clean up these inodes as otherwise they (and the blocks referenced from them)
9 would leak. Similarly if we truncate or extend 9 would leak. Similarly if we truncate or extend the file, we need not be able
10 to perform the operation in a single journalli 10 to perform the operation in a single journalling transaction. In such case we
11 track the inode as orphan so that in case of c 11 track the inode as orphan so that in case of crash extra blocks allocated to
12 the file get truncated. 12 the file get truncated.
13 13
14 Traditionally ext4 tracks orphan inodes in a f 14 Traditionally ext4 tracks orphan inodes in a form of single linked list where
15 superblock contains the inode number of the la 15 superblock contains the inode number of the last orphan inode (s_last_orphan
16 field) and then each inode contains inode numb 16 field) and then each inode contains inode number of the previously orphaned
17 inode (we overload i_dtime inode field for thi 17 inode (we overload i_dtime inode field for this). However this filesystem
18 global single linked list is a scalability bot 18 global single linked list is a scalability bottleneck for workloads that result
19 in heavy creation of orphan inodes. When orpha 19 in heavy creation of orphan inodes. When orphan file feature
20 (COMPAT_ORPHAN_FILE) is enabled, the filesyste 20 (COMPAT_ORPHAN_FILE) is enabled, the filesystem has a special inode
21 (referenced from the superblock through s_orph 21 (referenced from the superblock through s_orphan_file_inum) with several
22 blocks. Each of these blocks has a structure: 22 blocks. Each of these blocks has a structure:
23 23
24 ============= ================ =============== 24 ============= ================ =============== ===============================
25 Offset Type Name 25 Offset Type Name Description
26 ============= ================ =============== 26 ============= ================ =============== ===============================
27 0x0 Array of Orphan inode 27 0x0 Array of Orphan inode Each __le32 entry is either
28 __le32 entries entries 28 __le32 entries entries empty (0) or it contains
29 29 inode number of an orphan
30 30 inode.
31 blocksize-8 __le32 ob_magic 31 blocksize-8 __le32 ob_magic Magic value stored in orphan
32 32 block tail (0x0b10ca04)
33 blocksize-4 __le32 ob_checksum 33 blocksize-4 __le32 ob_checksum Checksum of the orphan block.
34 ============= ================ =============== 34 ============= ================ =============== ===============================
35 35
36 When a filesystem with orphan file feature is 36 When a filesystem with orphan file feature is writeably mounted, we set
37 RO_COMPAT_ORPHAN_PRESENT feature in the superb 37 RO_COMPAT_ORPHAN_PRESENT feature in the superblock to indicate there may
38 be valid orphan entries. In case we see this f 38 be valid orphan entries. In case we see this feature when mounting the
39 filesystem, we read the whole orphan file and 39 filesystem, we read the whole orphan file and process all orphan inodes found
40 there as usual. When cleanly unmounting the fi 40 there as usual. When cleanly unmounting the filesystem we remove the
41 RO_COMPAT_ORPHAN_PRESENT feature to avoid unne 41 RO_COMPAT_ORPHAN_PRESENT feature to avoid unnecessary scanning of the orphan
42 file and also make the filesystem fully compat 42 file and also make the filesystem fully compatible with older kernels.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.