Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | Read/Write HPFS 2.09 |
2 | 1998-2004, Mikulas Patocka | |
3 | ||
4 | email: mikulas@artax.karlin.mff.cuni.cz | |
5 | homepage: http://artax.karlin.mff.cuni.cz/~mikulas/vyplody/hpfs/index-e.cgi | |
6 | ||
7 | CREDITS: | |
8 | Chris Smith, 1993, original read-only HPFS, some code and hpfs structures file | |
9 | is taken from it | |
10 | Jacques Gelinas, MSDos mmap, Inspired by fs/nfs/mmap.c (Jon Tombs 15 Aug 1993) | |
11 | Werner Almesberger, 1992, 1993, MSDos option parser & CR/LF conversion | |
12 | ||
13 | Mount options | |
14 | ||
15 | uid=xxx,gid=xxx,umask=xxx (default uid=gid=0 umask=default_system_umask) | |
16 | Set owner/group/mode for files that do not have it specified in extended | |
17 | attributes. Mode is inverted umask - for example umask 027 gives owner | |
18 | all permission, group read permission and anybody else no access. Note | |
19 | that for files mode is anded with 0666. If you want files to have 'x' | |
20 | rights, you must use extended attributes. | |
21 | case=lower,asis (default asis) | |
22 | File name lowercasing in readdir. | |
23 | conv=binary,text,auto (default binary) | |
24 | CR/LF -> LF conversion, if auto, decision is made according to extension | |
25 | - there is a list of text extensions (I thing it's better to not convert | |
26 | text file than to damage binary file). If you want to change that list, | |
27 | change it in the source. Original readonly HPFS contained some strange | |
28 | heuristic algorithm that I removed. I thing it's danger to let the | |
29 | computer decide whether file is text or binary. For example, DJGPP | |
30 | binaries contain small text message at the beginning and they could be | |
31 | misidentified and damaged under some circumstances. | |
32 | check=none,normal,strict (default normal) | |
33 | Check level. Selecting none will cause only little speedup and big | |
34 | danger. I tried to write it so that it won't crash if check=normal on | |
35 | corrupted filesystems. check=strict means many superfluous checks - | |
36 | used for debugging (for example it checks if file is allocated in | |
37 | bitmaps when accessing it). | |
38 | errors=continue,remount-ro,panic (default remount-ro) | |
39 | Behaviour when filesystem errors found. | |
40 | chkdsk=no,errors,always (default errors) | |
41 | When to mark filesystem dirty so that OS/2 checks it. | |
42 | eas=no,ro,rw (default rw) | |
43 | What to do with extended attributes. 'no' - ignore them and use always | |
44 | values specified in uid/gid/mode options. 'ro' - read extended | |
45 | attributes but do not create them. 'rw' - create extended attributes | |
46 | when you use chmod/chown/chgrp/mknod/ln -s on the filesystem. | |
47 | timeshift=(-)nnn (default 0) | |
48 | Shifts the time by nnn seconds. For example, if you see under linux | |
49 | one hour more, than under os/2, use timeshift=-3600. | |
50 | ||
51 | ||
52 | File names | |
53 | ||
54 | As in OS/2, filenames are case insensitive. However, shell thinks that names | |
55 | are case sensitive, so for example when you create a file FOO, you can use | |
56 | 'cat FOO', 'cat Foo', 'cat foo' or 'cat F*' but not 'cat f*'. Note, that you | |
57 | also won't be able to compile linux kernel (and maybe other things) on HPFS | |
58 | because kernel creates different files with names like bootsect.S and | |
59 | bootsect.s. When searching for file thats name has characters >= 128, codepages | |
60 | are used - see below. | |
61 | OS/2 ignores dots and spaces at the end of file name, so this driver does as | |
62 | well. If you create 'a. ...', the file 'a' will be created, but you can still | |
63 | access it under names 'a.', 'a..', 'a . . . ' etc. | |
64 | ||
65 | ||
66 | Extended attributes | |
67 | ||
68 | On HPFS partitions, OS/2 can associate to each file a special information called | |
69 | extended attributes. Extended attributes are pairs of (key,value) where key is | |
70 | an ascii string identifying that attribute and value is any string of bytes of | |
71 | variable length. OS/2 stores window and icon positions and file types there. So | |
72 | why not use it for unix-specific info like file owner or access rights? This | |
73 | driver can do it. If you chown/chgrp/chmod on a hpfs partition, extended | |
74 | attributes with keys "UID", "GID" or "MODE" and 2-byte values are created. Only | |
75 | that extended attributes those value differs from defaults specified in mount | |
76 | options are created. Once created, the extended attributes are never deleted, | |
77 | they're just changed. It means that when your default uid=0 and you type | |
78 | something like 'chown luser file; chown root file' the file will contain | |
79 | extended attribute UID=0. And when you umount the fs and mount it again with | |
80 | uid=luser_uid, the file will be still owned by root! If you chmod file to 444, | |
81 | extended attribute "MODE" will not be set, this special case is done by setting | |
82 | read-only flag. When you mknod a block or char device, besides "MODE", the | |
83 | special 4-byte extended attribute "DEV" will be created containing the device | |
84 | number. Currently this driver cannot resize extended attributes - it means | |
85 | that if somebody (I don't know who?) has set "UID", "GID", "MODE" or "DEV" | |
86 | attributes with different sizes, they won't be rewritten and changing these | |
87 | values doesn't work. | |
88 | ||
89 | ||
90 | Symlinks | |
91 | ||
92 | You can do symlinks on HPFS partition, symlinks are achieved by setting extended | |
93 | attribute named "SYMLINK" with symlink value. Like on ext2, you can chown and | |
94 | chgrp symlinks but I don't know what is it good for. chmoding symlink results | |
95 | in chmoding file where symlink points. These symlinks are just for Linux use and | |
96 | incompatible with OS/2. OS/2 PmShell symlinks are not supported because they are | |
97 | stored in very crazy way. They tried to do it so that link changes when file is | |
98 | moved ... sometimes it works. But the link is partly stored in directory | |
99 | extended attributes and partly in OS2SYS.INI. I don't want (and don't know how) | |
100 | to analyze or change OS2SYS.INI. | |
101 | ||
102 | ||
103 | Codepages | |
104 | ||
105 | HPFS can contain several uppercasing tables for several codepages and each | |
a33f3224 | 106 | file has a pointer to codepage its name is in. However OS/2 was created in |
1da177e4 LT |
107 | America where people don't care much about codepages and so multiple codepages |
108 | support is quite buggy. I have Czech OS/2 working in codepage 852 on my disk. | |
109 | Once I booted English OS/2 working in cp 850 and I created a file on my 852 | |
110 | partition. It marked file name codepage as 850 - good. But when I again booted | |
111 | Czech OS/2, the file was completely inaccessible under any name. It seems that | |
112 | OS/2 uppercases the search pattern with its system code page (852) and file | |
113 | name it's comparing to with its code page (850). These could never match. Is it | |
114 | really what IBM developers wanted? But problems continued. When I created in | |
115 | Czech OS/2 another file in that directory, that file was inaccessible too. OS/2 | |
116 | probably uses different uppercasing method when searching where to place a file | |
117 | (note, that files in HPFS directory must be sorted) and when searching for | |
118 | a file. Finally when I opened this directory in PmShell, PmShell crashed (the | |
119 | funny thing was that, when rebooted, PmShell tried to reopen this directory | |
120 | again :-). chkdsk happily ignores these errors and only low-level disk | |
121 | modification saved me. Never mix different language versions of OS/2 on one | |
122 | system although HPFS was designed to allow that. | |
123 | OK, I could implement complex codepage support to this driver but I think it | |
124 | would cause more problems than benefit with such buggy implementation in OS/2. | |
125 | So this driver simply uses first codepage it finds for uppercasing and | |
126 | lowercasing no matter what's file codepage index. Usually all file names are in | |
127 | this codepage - if you don't try to do what I described above :-) | |
128 | ||
129 | ||
130 | Known bugs | |
131 | ||
132 | HPFS386 on OS/2 server is not supported. HPFS386 installed on normal OS/2 client | |
133 | should work. If you have OS/2 server, use only read-only mode. I don't know how | |
134 | to handle some HPFS386 structures like access control list or extended perm | |
135 | list, I don't know how to delete them when file is deleted and how to not | |
136 | overwrite them with extended attributes. Send me some info on these structures | |
137 | and I'll make it. However, this driver should detect presence of HPFS386 | |
138 | structures, remount read-only and not destroy them (I hope). | |
139 | ||
140 | When there's not enough space for extended attributes, they will be truncated | |
141 | and no error is returned. | |
142 | ||
143 | OS/2 can't access files if the path is longer than about 256 chars but this | |
144 | driver allows you to do it. chkdsk ignores such errors. | |
145 | ||
146 | Sometimes you won't be able to delete some files on a very full filesystem | |
147 | (returning error ENOSPC). That's because file in non-leaf node in directory tree | |
148 | (one directory, if it's large, has dirents in tree on HPFS) must be replaced | |
149 | with another node when deleted. And that new file might have larger name than | |
150 | the old one so the new name doesn't fit in directory node (dnode). And that | |
151 | would result in directory tree splitting, that takes disk space. Workaround is | |
152 | to delete other files that are leaf (probability that the file is non-leaf is | |
153 | about 1/50) or to truncate file first to make some space. | |
154 | You encounter this problem only if you have many directories so that | |
155 | preallocated directory band is full i.e. | |
156 | number_of_directories / size_of_filesystem_in_mb > 4. | |
157 | ||
158 | You can't delete open directories. | |
159 | ||
160 | You can't rename over directories (what is it good for?). | |
161 | ||
162 | Renaming files so that only case changes doesn't work. This driver supports it | |
163 | but vfs doesn't. Something like 'mv file FILE' won't work. | |
164 | ||
165 | All atimes and directory mtimes are not updated. That's because of performance | |
166 | reasons. If you extremely wish to update them, let me know, I'll write it (but | |
167 | it will be slow). | |
168 | ||
169 | When the system is out of memory and swap, it may slightly corrupt filesystem | |
170 | (lost files, unbalanced directories). (I guess all filesystem may do it). | |
171 | ||
172 | When compiled, you get warning: function declaration isn't a prototype. Does | |
173 | anybody know what does it mean? | |
174 | ||
175 | ||
176 | What does "unbalanced tree" message mean? | |
177 | ||
178 | Old versions of this driver created sometimes unbalanced dnode trees. OS/2 | |
179 | chkdsk doesn't scream if the tree is unbalanced (and sometimes creates | |
180 | unbalanced trees too :-) but both HPFS and HPFS386 contain bug that it rarely | |
181 | crashes when the tree is not balanced. This driver handles unbalanced trees | |
182 | correctly and writes warning if it finds them. If you see this message, this is | |
183 | probably because of directories created with old version of this driver. | |
184 | Workaround is to move all files from that directory to another and then back | |
185 | again. Do it in Linux, not OS/2! If you see this message in directory that is | |
186 | whole created by this driver, it is BUG - let me know about it. | |
187 | ||
188 | ||
189 | Bugs in OS/2 | |
190 | ||
191 | When you have two (or more) lost directories pointing each to other, chkdsk | |
192 | locks up when repairing filesystem. | |
193 | ||
194 | Sometimes (I think it's random) when you create a file with one-char name under | |
195 | OS/2, OS/2 marks it as 'long'. chkdsk then removes this flag saying "Minor fs | |
196 | error corrected". | |
197 | ||
198 | File names like "a .b" are marked as 'long' by OS/2 but chkdsk "corrects" it and | |
199 | marks them as short (and writes "minor fs error corrected"). This bug is not in | |
200 | HPFS386. | |
201 | ||
202 | Codepage bugs described above. | |
203 | ||
204 | If you don't install fixpacks, there are many, many more... | |
205 | ||
206 | ||
207 | History | |
208 | ||
209 | 0.90 First public release | |
210 | 0.91 Fixed bug that caused shooting to memory when write_inode was called on | |
211 | open inode (rarely happened) | |
212 | 0.92 Fixed a little memory leak in freeing directory inodes | |
213 | 0.93 Fixed bug that locked up the machine when there were too many filenames | |
214 | with first 15 characters same | |
215 | Fixed write_file to zero file when writing behind file end | |
216 | 0.94 Fixed a little memory leak when trying to delete busy file or directory | |
217 | 0.95 Fixed a bug that i_hpfs_parent_dir was not updated when moving files | |
218 | 1.90 First version for 2.1.1xx kernels | |
219 | 1.91 Fixed a bug that chk_sectors failed when sectors were at the end of disk | |
220 | Fixed a race-condition when write_inode is called while deleting file | |
221 | Fixed a bug that could possibly happen (with very low probability) when | |
222 | using 0xff in filenames | |
223 | Rewritten locking to avoid race-conditions | |
224 | Mount option 'eas' now works | |
225 | Fsync no longer returns error | |
226 | Files beginning with '.' are marked hidden | |
227 | Remount support added | |
228 | Alloc is not so slow when filesystem becomes full | |
229 | Atimes are no more updated because it slows down operation | |
230 | Code cleanup (removed all commented debug prints) | |
231 | 1.92 Corrected a bug when sync was called just before closing file | |
232 | 1.93 Modified, so that it works with kernels >= 2.1.131, I don't know if it | |
233 | works with previous versions | |
234 | Fixed a possible problem with disks > 64G (but I don't have one, so I can't | |
235 | test it) | |
236 | Fixed a file overflow at 2G | |
237 | Added new option 'timeshift' | |
238 | Changed behaviour on HPFS386: It is now possible to operate on HPFS386 in | |
239 | read-only mode | |
240 | Fixed a bug that slowed down alloc and prevented allocating 100% space | |
241 | (this bug was not destructive) | |
242 | 1.94 Added workaround for one bug in Linux | |
243 | Fixed one buffer leak | |
244 | Fixed some incompatibilities with large extended attributes (but it's still | |
245 | not 100% ok, I have no info on it and OS/2 doesn't want to create them) | |
246 | Rewritten allocation | |
247 | Fixed a bug with i_blocks (du sometimes didn't display correct values) | |
248 | Directories have no longer archive attribute set (some programs don't like | |
249 | it) | |
250 | Fixed a bug that it set badly one flag in large anode tree (it was not | |
251 | destructive) | |
252 | 1.95 Fixed one buffer leak, that could happen on corrupted filesystem | |
253 | Fixed one bug in allocation in 1.94 | |
254 | 1.96 Added workaround for one bug in OS/2 (HPFS locked up, HPFS386 reported | |
255 | error sometimes when opening directories in PMSHELL) | |
256 | Fixed a possible bitmap race | |
257 | Fixed possible problem on large disks | |
258 | You can now delete open files | |
259 | Fixed a nondestructive race in rename | |
260 | 1.97 Support for HPFS v3 (on large partitions) | |
261 | Fixed a bug that it didn't allow creation of files > 128M (it should be 2G) | |
262 | 1.97.1 Changed names of global symbols | |
263 | Fixed a bug when chmoding or chowning root directory | |
264 | 1.98 Fixed a deadlock when using old_readdir | |
265 | Better directory handling; workaround for "unbalanced tree" bug in OS/2 | |
266 | 1.99 Corrected a possible problem when there's not enough space while deleting | |
267 | file | |
268 | Now it tries to truncate the file if there's not enough space when deleting | |
269 | Removed a lot of redundant code | |
270 | 2.00 Fixed a bug in rename (it was there since 1.96) | |
271 | Better anti-fragmentation strategy | |
272 | 2.01 Fixed problem with directory listing over NFS | |
273 | Directory lseek now checks for proper parameters | |
274 | Fixed race-condition in buffer code - it is in all filesystems in Linux; | |
275 | when reading device (cat /dev/hda) while creating files on it, files | |
276 | could be damaged | |
4ae0edc2 | 277 | 2.02 Workaround for bug in breada in Linux. breada could cause accesses beyond |
1da177e4 LT |
278 | end of partition |
279 | 2.03 Char, block devices and pipes are correctly created | |
280 | Fixed non-crashing race in unlink (Alexander Viro) | |
281 | Now it works with Japanese version of OS/2 | |
282 | 2.04 Fixed error when ftruncate used to extend file | |
283 | 2.05 Fixed crash when got mount parameters without = | |
284 | Fixed crash when allocation of anode failed due to full disk | |
285 | Fixed some crashes when block io or inode allocation failed | |
286 | 2.06 Fixed some crash on corrupted disk structures | |
287 | Better allocation strategy | |
288 | Reschedule points added so that it doesn't lock CPU long time | |
289 | It should work in read-only mode on Warp Server | |
290 | 2.07 More fixes for Warp Server. Now it really works | |
291 | 2.08 Creating new files is not so slow on large disks | |
292 | An attempt to sync deleted file does not generate filesystem error | |
a982ac06 | 293 | 2.09 Fixed error on extremely fragmented files |
1da177e4 LT |
294 | |
295 | ||
296 | vim: set textwidth=80: |