1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
|
ciopfs - case insensitive on purpose filesystem
ciopfs is a stackable or overlay linux userspace file system (implemented with
FUSE) which mounts a normal directory on a regular file system in case
insensitive fashion.
The commands below should illustrate it’s function:
mkdir -p ~/tmp/ciopfs/{.data,case-insensitive}
ciopfs ~/tmp/ciopfs/.data ~/tmp/ciopfs/case-insensitive
cd ~/tmp/ciopfs
mkdir -p case-insensitive/DeMo/SubFolder
echo demo >> case-insensitive/DEMO/subFolder/MyFile
At this point your file system should look like this:
case-insensitive
`-- DeMo
`-- SubFolder
`-- MyFile
.data
`-- demo
`-- subfolder
`-- myfile
To avoid any conflicts you should not manipulate the data directory directly,
any change should be done over the mount point. All filenames in the data
directory which aren’t all lower case are ignored.
If you want to mount the file system automatically at boot time add a line like
the one below to your /etc/fstab.
/ciopfs/data /ciopfs/mnt ciopfs allow_other,default_permissions,use_ino,attr_timeout=0 0 0
Note that ciopfs is primarily designed for single user mode. It was originally
developed to mount the wine program folder and provide faster case insensitive
file access. If you want to give multiple users write access to the same file
system, then you have to mount it as root. However, in order to avoid security
problems ciopfs will force fuse into single threaded mode and thus hurt
performance.
News
* ciopfs-0.4 released (18.06.2011)
* Bugfix in symlink creation
* Better errno handling
* ciopfs-0.3 released (25.09.2010)
* Security improvements: ciopfs forces single threaded mode if the file
system is mounted by root and accessible for others
* ASCII mode should now work (an off by one error which caused a segfault
was fixed)
* Various bug fixes
* ciopfs-0.2 released (30.06.2008)
* Unicode support based on glib
* Better error handling in out of memory situations
* Various code cleanups
* ciopfs-0.1 released (24.05.2008)
How it works
ciopfs works by translating every path element to lower case before further
operations take place. On file or directory creation the original file name is
stored in an extended attribute which is later returned upon request.
This is illustrated below:
getfattr -dR .data
# file: .data/demo
user.filename="DeMo"
# file: .data/demo/subfolder
user.filename="SubFolder"
# file: .data/demo/subfolder/myfile
user.filename="MyFile"
Runtime Requirements
If you want the file system to preserve case information you have to make sure
that the underlying file system supports extended attributes (for example for
ext{2,3} you need a kernel with CONFIG_EXT{2,3}_FS_XATTR enabled). You probably
also want to mount the underlying filesystem with the user_xattr option which
allows non root users to create extended attributes.
Build Requirements
In order to compile ciopfs you will need the fuse development files, libattr and
if you plan to use Unicode characters within file names you will either need
glib which is the default or alternatively libicu.
If you want to use neither of those the file system will fall back to libc’s
tolower(3) function which is only defined for [a-zA-Z]. Hence, it will only work
case insensitively for ASCII file names.
For ease of use the following 3 Makefile targets are supported:
* unicode-glib (default)
* unicode-icu
* ascii
Running one of those followed by sudo make install should do everything that is
needed.
Alternatively, you can also use one of the distribution provided binary
packages.
POSIX Compliance
ciopfs passes all test of a slightly patched POSIX file system test suite when
mounted as root user with the following options:
allow_other,use_ino,attr_timeout=0,entry_timeout=0
and $fs set to "ciopfs" in the test suite configuration file. This was last
tested with pjd-fstest-20090130-RC.tgz and ext3 as the underlying file system.
Stability and Speed
ciopfs just passes every requested operation to the underlying file system, so
in theory it shouldn’t have a negative impact on stability. However, if you find
a bug then send me an email with the instruction to reproduce it.
As far as speed is of concern, I didn’t really benchmark or optimize it so far.
There is the usual overhead associated with user / kernel space context
switches. Furthermore, ciopfs in it’s current implementation uses libc’s
malloc/free quite extensively, maybe this could be a bottleneck.
Development
You can always fetch the current code base from the git repository located at
Github or Sourcehut.
If you have comments, suggestions, ideas, a bug report, a patch or something
else related to ciopfs then don’t hesitate to write me an email.
License
ciopfs is licensed under the GNU GPL v2.
|
