hpfs.txt 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296
  1. Read/Write HPFS 2.09
  2. 1998-2004, Mikulas Patocka
  3. email: mikulas@artax.karlin.mff.cuni.cz
  4. homepage: http://artax.karlin.mff.cuni.cz/~mikulas/vyplody/hpfs/index-e.cgi
  5. CREDITS:
  6. Chris Smith, 1993, original read-only HPFS, some code and hpfs structures file
  7. is taken from it
  8. Jacques Gelinas, MSDos mmap, Inspired by fs/nfs/mmap.c (Jon Tombs 15 Aug 1993)
  9. Werner Almesberger, 1992, 1993, MSDos option parser & CR/LF conversion
  10. Mount options
  11. uid=xxx,gid=xxx,umask=xxx (default uid=gid=0 umask=default_system_umask)
  12. Set owner/group/mode for files that do not have it specified in extended
  13. attributes. Mode is inverted umask - for example umask 027 gives owner
  14. all permission, group read permission and anybody else no access. Note
  15. that for files mode is anded with 0666. If you want files to have 'x'
  16. rights, you must use extended attributes.
  17. case=lower,asis (default asis)
  18. File name lowercasing in readdir.
  19. conv=binary,text,auto (default binary)
  20. CR/LF -> LF conversion, if auto, decision is made according to extension
  21. - there is a list of text extensions (I thing it's better to not convert
  22. text file than to damage binary file). If you want to change that list,
  23. change it in the source. Original readonly HPFS contained some strange
  24. heuristic algorithm that I removed. I thing it's danger to let the
  25. computer decide whether file is text or binary. For example, DJGPP
  26. binaries contain small text message at the beginning and they could be
  27. misidentified and damaged under some circumstances.
  28. check=none,normal,strict (default normal)
  29. Check level. Selecting none will cause only little speedup and big
  30. danger. I tried to write it so that it won't crash if check=normal on
  31. corrupted filesystems. check=strict means many superfluous checks -
  32. used for debugging (for example it checks if file is allocated in
  33. bitmaps when accessing it).
  34. errors=continue,remount-ro,panic (default remount-ro)
  35. Behaviour when filesystem errors found.
  36. chkdsk=no,errors,always (default errors)
  37. When to mark filesystem dirty so that OS/2 checks it.
  38. eas=no,ro,rw (default rw)
  39. What to do with extended attributes. 'no' - ignore them and use always
  40. values specified in uid/gid/mode options. 'ro' - read extended
  41. attributes but do not create them. 'rw' - create extended attributes
  42. when you use chmod/chown/chgrp/mknod/ln -s on the filesystem.
  43. timeshift=(-)nnn (default 0)
  44. Shifts the time by nnn seconds. For example, if you see under linux
  45. one hour more, than under os/2, use timeshift=-3600.
  46. File names
  47. As in OS/2, filenames are case insensitive. However, shell thinks that names
  48. are case sensitive, so for example when you create a file FOO, you can use
  49. 'cat FOO', 'cat Foo', 'cat foo' or 'cat F*' but not 'cat f*'. Note, that you
  50. also won't be able to compile linux kernel (and maybe other things) on HPFS
  51. because kernel creates different files with names like bootsect.S and
  52. bootsect.s. When searching for file thats name has characters >= 128, codepages
  53. are used - see below.
  54. OS/2 ignores dots and spaces at the end of file name, so this driver does as
  55. well. If you create 'a. ...', the file 'a' will be created, but you can still
  56. access it under names 'a.', 'a..', 'a . . . ' etc.
  57. Extended attributes
  58. On HPFS partitions, OS/2 can associate to each file a special information called
  59. extended attributes. Extended attributes are pairs of (key,value) where key is
  60. an ascii string identifying that attribute and value is any string of bytes of
  61. variable length. OS/2 stores window and icon positions and file types there. So
  62. why not use it for unix-specific info like file owner or access rights? This
  63. driver can do it. If you chown/chgrp/chmod on a hpfs partition, extended
  64. attributes with keys "UID", "GID" or "MODE" and 2-byte values are created. Only
  65. that extended attributes those value differs from defaults specified in mount
  66. options are created. Once created, the extended attributes are never deleted,
  67. they're just changed. It means that when your default uid=0 and you type
  68. something like 'chown luser file; chown root file' the file will contain
  69. extended attribute UID=0. And when you umount the fs and mount it again with
  70. uid=luser_uid, the file will be still owned by root! If you chmod file to 444,
  71. extended attribute "MODE" will not be set, this special case is done by setting
  72. read-only flag. When you mknod a block or char device, besides "MODE", the
  73. special 4-byte extended attribute "DEV" will be created containing the device
  74. number. Currently this driver cannot resize extended attributes - it means
  75. that if somebody (I don't know who?) has set "UID", "GID", "MODE" or "DEV"
  76. attributes with different sizes, they won't be rewritten and changing these
  77. values doesn't work.
  78. Symlinks
  79. You can do symlinks on HPFS partition, symlinks are achieved by setting extended
  80. attribute named "SYMLINK" with symlink value. Like on ext2, you can chown and
  81. chgrp symlinks but I don't know what is it good for. chmoding symlink results
  82. in chmoding file where symlink points. These symlinks are just for Linux use and
  83. incompatible with OS/2. OS/2 PmShell symlinks are not supported because they are
  84. stored in very crazy way. They tried to do it so that link changes when file is
  85. moved ... sometimes it works. But the link is partly stored in directory
  86. extended attributes and partly in OS2SYS.INI. I don't want (and don't know how)
  87. to analyze or change OS2SYS.INI.
  88. Codepages
  89. HPFS can contain several uppercasing tables for several codepages and each
  90. file has a pointer to codepage its name is in. However OS/2 was created in
  91. America where people don't care much about codepages and so multiple codepages
  92. support is quite buggy. I have Czech OS/2 working in codepage 852 on my disk.
  93. Once I booted English OS/2 working in cp 850 and I created a file on my 852
  94. partition. It marked file name codepage as 850 - good. But when I again booted
  95. Czech OS/2, the file was completely inaccessible under any name. It seems that
  96. OS/2 uppercases the search pattern with its system code page (852) and file
  97. name it's comparing to with its code page (850). These could never match. Is it
  98. really what IBM developers wanted? But problems continued. When I created in
  99. Czech OS/2 another file in that directory, that file was inaccessible too. OS/2
  100. probably uses different uppercasing method when searching where to place a file
  101. (note, that files in HPFS directory must be sorted) and when searching for
  102. a file. Finally when I opened this directory in PmShell, PmShell crashed (the
  103. funny thing was that, when rebooted, PmShell tried to reopen this directory
  104. again :-). chkdsk happily ignores these errors and only low-level disk
  105. modification saved me. Never mix different language versions of OS/2 on one
  106. system although HPFS was designed to allow that.
  107. OK, I could implement complex codepage support to this driver but I think it
  108. would cause more problems than benefit with such buggy implementation in OS/2.
  109. So this driver simply uses first codepage it finds for uppercasing and
  110. lowercasing no matter what's file codepage index. Usually all file names are in
  111. this codepage - if you don't try to do what I described above :-)
  112. Known bugs
  113. HPFS386 on OS/2 server is not supported. HPFS386 installed on normal OS/2 client
  114. should work. If you have OS/2 server, use only read-only mode. I don't know how
  115. to handle some HPFS386 structures like access control list or extended perm
  116. list, I don't know how to delete them when file is deleted and how to not
  117. overwrite them with extended attributes. Send me some info on these structures
  118. and I'll make it. However, this driver should detect presence of HPFS386
  119. structures, remount read-only and not destroy them (I hope).
  120. When there's not enough space for extended attributes, they will be truncated
  121. and no error is returned.
  122. OS/2 can't access files if the path is longer than about 256 chars but this
  123. driver allows you to do it. chkdsk ignores such errors.
  124. Sometimes you won't be able to delete some files on a very full filesystem
  125. (returning error ENOSPC). That's because file in non-leaf node in directory tree
  126. (one directory, if it's large, has dirents in tree on HPFS) must be replaced
  127. with another node when deleted. And that new file might have larger name than
  128. the old one so the new name doesn't fit in directory node (dnode). And that
  129. would result in directory tree splitting, that takes disk space. Workaround is
  130. to delete other files that are leaf (probability that the file is non-leaf is
  131. about 1/50) or to truncate file first to make some space.
  132. You encounter this problem only if you have many directories so that
  133. preallocated directory band is full i.e.
  134. number_of_directories / size_of_filesystem_in_mb > 4.
  135. You can't delete open directories.
  136. You can't rename over directories (what is it good for?).
  137. Renaming files so that only case changes doesn't work. This driver supports it
  138. but vfs doesn't. Something like 'mv file FILE' won't work.
  139. All atimes and directory mtimes are not updated. That's because of performance
  140. reasons. If you extremely wish to update them, let me know, I'll write it (but
  141. it will be slow).
  142. When the system is out of memory and swap, it may slightly corrupt filesystem
  143. (lost files, unbalanced directories). (I guess all filesystem may do it).
  144. When compiled, you get warning: function declaration isn't a prototype. Does
  145. anybody know what does it mean?
  146. What does "unbalanced tree" message mean?
  147. Old versions of this driver created sometimes unbalanced dnode trees. OS/2
  148. chkdsk doesn't scream if the tree is unbalanced (and sometimes creates
  149. unbalanced trees too :-) but both HPFS and HPFS386 contain bug that it rarely
  150. crashes when the tree is not balanced. This driver handles unbalanced trees
  151. correctly and writes warning if it finds them. If you see this message, this is
  152. probably because of directories created with old version of this driver.
  153. Workaround is to move all files from that directory to another and then back
  154. again. Do it in Linux, not OS/2! If you see this message in directory that is
  155. whole created by this driver, it is BUG - let me know about it.
  156. Bugs in OS/2
  157. When you have two (or more) lost directories pointing each to other, chkdsk
  158. locks up when repairing filesystem.
  159. Sometimes (I think it's random) when you create a file with one-char name under
  160. OS/2, OS/2 marks it as 'long'. chkdsk then removes this flag saying "Minor fs
  161. error corrected".
  162. File names like "a .b" are marked as 'long' by OS/2 but chkdsk "corrects" it and
  163. marks them as short (and writes "minor fs error corrected"). This bug is not in
  164. HPFS386.
  165. Codepage bugs described above.
  166. If you don't install fixpacks, there are many, many more...
  167. History
  168. 0.90 First public release
  169. 0.91 Fixed bug that caused shooting to memory when write_inode was called on
  170. open inode (rarely happened)
  171. 0.92 Fixed a little memory leak in freeing directory inodes
  172. 0.93 Fixed bug that locked up the machine when there were too many filenames
  173. with first 15 characters same
  174. Fixed write_file to zero file when writing behind file end
  175. 0.94 Fixed a little memory leak when trying to delete busy file or directory
  176. 0.95 Fixed a bug that i_hpfs_parent_dir was not updated when moving files
  177. 1.90 First version for 2.1.1xx kernels
  178. 1.91 Fixed a bug that chk_sectors failed when sectors were at the end of disk
  179. Fixed a race-condition when write_inode is called while deleting file
  180. Fixed a bug that could possibly happen (with very low probability) when
  181. using 0xff in filenames
  182. Rewritten locking to avoid race-conditions
  183. Mount option 'eas' now works
  184. Fsync no longer returns error
  185. Files beginning with '.' are marked hidden
  186. Remount support added
  187. Alloc is not so slow when filesystem becomes full
  188. Atimes are no more updated because it slows down operation
  189. Code cleanup (removed all commented debug prints)
  190. 1.92 Corrected a bug when sync was called just before closing file
  191. 1.93 Modified, so that it works with kernels >= 2.1.131, I don't know if it
  192. works with previous versions
  193. Fixed a possible problem with disks > 64G (but I don't have one, so I can't
  194. test it)
  195. Fixed a file overflow at 2G
  196. Added new option 'timeshift'
  197. Changed behaviour on HPFS386: It is now possible to operate on HPFS386 in
  198. read-only mode
  199. Fixed a bug that slowed down alloc and prevented allocating 100% space
  200. (this bug was not destructive)
  201. 1.94 Added workaround for one bug in Linux
  202. Fixed one buffer leak
  203. Fixed some incompatibilities with large extended attributes (but it's still
  204. not 100% ok, I have no info on it and OS/2 doesn't want to create them)
  205. Rewritten allocation
  206. Fixed a bug with i_blocks (du sometimes didn't display correct values)
  207. Directories have no longer archive attribute set (some programs don't like
  208. it)
  209. Fixed a bug that it set badly one flag in large anode tree (it was not
  210. destructive)
  211. 1.95 Fixed one buffer leak, that could happen on corrupted filesystem
  212. Fixed one bug in allocation in 1.94
  213. 1.96 Added workaround for one bug in OS/2 (HPFS locked up, HPFS386 reported
  214. error sometimes when opening directories in PMSHELL)
  215. Fixed a possible bitmap race
  216. Fixed possible problem on large disks
  217. You can now delete open files
  218. Fixed a nondestructive race in rename
  219. 1.97 Support for HPFS v3 (on large partitions)
  220. Fixed a bug that it didn't allow creation of files > 128M (it should be 2G)
  221. 1.97.1 Changed names of global symbols
  222. Fixed a bug when chmoding or chowning root directory
  223. 1.98 Fixed a deadlock when using old_readdir
  224. Better directory handling; workaround for "unbalanced tree" bug in OS/2
  225. 1.99 Corrected a possible problem when there's not enough space while deleting
  226. file
  227. Now it tries to truncate the file if there's not enough space when deleting
  228. Removed a lot of redundant code
  229. 2.00 Fixed a bug in rename (it was there since 1.96)
  230. Better anti-fragmentation strategy
  231. 2.01 Fixed problem with directory listing over NFS
  232. Directory lseek now checks for proper parameters
  233. Fixed race-condition in buffer code - it is in all filesystems in Linux;
  234. when reading device (cat /dev/hda) while creating files on it, files
  235. could be damaged
  236. 2.02 Workaround for bug in breada in Linux. breada could cause accesses beyond
  237. end of partition
  238. 2.03 Char, block devices and pipes are correctly created
  239. Fixed non-crashing race in unlink (Alexander Viro)
  240. Now it works with Japanese version of OS/2
  241. 2.04 Fixed error when ftruncate used to extend file
  242. 2.05 Fixed crash when got mount parameters without =
  243. Fixed crash when allocation of anode failed due to full disk
  244. Fixed some crashes when block io or inode allocation failed
  245. 2.06 Fixed some crash on corrupted disk structures
  246. Better allocation strategy
  247. Reschedule points added so that it doesn't lock CPU long time
  248. It should work in read-only mode on Warp Server
  249. 2.07 More fixes for Warp Server. Now it really works
  250. 2.08 Creating new files is not so slow on large disks
  251. An attempt to sync deleted file does not generate filesystem error
  252. 2.09 Fixed error on extremely fragmented files
  253. vim: set textwidth=80: