spi-topcliff-pch: add recovery processing in case wait-event timeout
[zen-stable.git] / Documentation / ABI / testing / sysfs-fs-ext4
blobf22ac0872ae8de41a44406edfab71384d54b830a
1 What:           /sys/fs/ext4/<disk>/mb_stats
2 Date:           March 2008
3 Contact:        "Theodore Ts'o" <tytso@mit.edu>
4 Description:
5                  Controls whether the multiblock allocator should
6                  collect statistics, which are shown during the unmount.
7                  1 means to collect statistics, 0 means not to collect
8                  statistics
10 What:           /sys/fs/ext4/<disk>/mb_group_prealloc
11 Date:           March 2008
12 Contact:        "Theodore Ts'o" <tytso@mit.edu>
13 Description:
14                 The multiblock allocator will round up allocation
15                 requests to a multiple of this tuning parameter if the
16                 stripe size is not set in the ext4 superblock
18 What:           /sys/fs/ext4/<disk>/mb_max_to_scan
19 Date:           March 2008
20 Contact:        "Theodore Ts'o" <tytso@mit.edu>
21 Description:
22                 The maximum number of extents the multiblock allocator
23                 will search to find the best extent
25 What:           /sys/fs/ext4/<disk>/mb_min_to_scan
26 Date:           March 2008
27 Contact:        "Theodore Ts'o" <tytso@mit.edu>
28 Description:
29                 The minimum number of extents the multiblock allocator
30                 will search to find the best extent
32 What:           /sys/fs/ext4/<disk>/mb_order2_req
33 Date:           March 2008
34 Contact:        "Theodore Ts'o" <tytso@mit.edu>
35 Description:
36                 Tuning parameter which controls the minimum size for 
37                 requests (as a power of 2) where the buddy cache is
38                 used
40 What:           /sys/fs/ext4/<disk>/mb_stream_req
41 Date:           March 2008
42 Contact:        "Theodore Ts'o" <tytso@mit.edu>
43 Description:
44                 Files which have fewer blocks than this tunable
45                 parameter will have their blocks allocated out of a
46                 block group specific preallocation pool, so that small
47                 files are packed closely together.  Each large file
48                  will have its blocks allocated out of its own unique
49                  preallocation pool.
51 What:           /sys/fs/ext4/<disk>/inode_readahead_blks
52 Date:           March 2008
53 Contact:        "Theodore Ts'o" <tytso@mit.edu>
54 Description:
55                 Tuning parameter which controls the maximum number of
56                 inode table blocks that ext4's inode table readahead
57                 algorithm will pre-read into the buffer cache
59 What:           /sys/fs/ext4/<disk>/delayed_allocation_blocks
60 Date:           March 2008
61 Contact:        "Theodore Ts'o" <tytso@mit.edu>
62 Description:
63                 This file is read-only and shows the number of blocks
64                 that are dirty in the page cache, but which do not
65                 have their location in the filesystem allocated yet.
67 What:           /sys/fs/ext4/<disk>/lifetime_write_kbytes
68 Date:           March 2008
69 Contact:        "Theodore Ts'o" <tytso@mit.edu>
70 Description:
71                 This file is read-only and shows the number of kilobytes
72                 of data that have been written to this filesystem since it was
73                 created.
75 What:           /sys/fs/ext4/<disk>/session_write_kbytes
76 Date:           March 2008
77 Contact:        "Theodore Ts'o" <tytso@mit.edu>
78 Description:
79                 This file is read-only and shows the number of
80                 kilobytes of data that have been written to this
81                 filesystem since it was mounted.
83 What:           /sys/fs/ext4/<disk>/inode_goal
84 Date:           June 2008
85 Contact:        "Theodore Ts'o" <tytso@mit.edu>
86 Description:
87                 Tuning parameter which (if non-zero) controls the goal
88                 inode used by the inode allocator in preference to
89                 all other allocation heuristics.  This is intended for
90                 debugging use only, and should be 0 on production
91                 systems.
93 What:           /sys/fs/ext4/<disk>/max_writeback_mb_bump
94 Date:           September 2009
95 Contact:        "Theodore Ts'o" <tytso@mit.edu>
96 Description:
97                 The maximum number of megabytes the writeback code will
98                 try to write out before move on to another inode.