ChangeLog 90 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639
  1. ToDo/Notes:
  2. - Find and fix bugs.
  3. - The only places in the kernel where a file is resized are
  4. ntfs_file_write*() and ntfs_truncate() for both of which i_sem is
  5. held. Just have to be careful in read-/writepage and other helpers
  6. not running under i_sem that we play nice... Also need to be careful
  7. with initialized_size extension in ntfs_file_write*() and writepage.
  8. UPDATE: The only things that need to be checked are the compressed
  9. write and the other attribute resize/write cases like index
  10. attributes, etc. For now none of these are implemented so are safe.
  11. - Implement filling in of holes in aops.c::ntfs_writepage() and its
  12. helpers.
  13. - Implement mft.c::sync_mft_mirror_umount(). We currently will just
  14. leave the volume dirty on umount if the final iput(vol->mft_ino)
  15. causes a write of any mirrored mft records due to the mft mirror
  16. inode having been discarded already. Whether this can actually ever
  17. happen is unclear however so it is worth waiting until someone hits
  18. the problem.
  19. - Enable the code for setting the NT4 compatibility flag when we start
  20. making NTFS 1.2 specific modifications.
  21. 2.1.25 - (Almost) fully implement write(2) and truncate(2).
  22. - Change ntfs_map_runlist_nolock(), ntfs_attr_find_vcn_nolock() and
  23. {__,}ntfs_cluster_free() to also take an optional attribute search
  24. context as argument. This allows calling these functions with the
  25. mft record mapped. Update all callers.
  26. - Fix potential deadlock in ntfs_mft_data_extend_allocation_nolock()
  27. error handling by passing in the active search context when calling
  28. ntfs_cluster_free().
  29. - Change ntfs_cluster_alloc() to take an extra boolean parameter
  30. specifying whether the cluster are being allocated to extend an
  31. attribute or to fill a hole.
  32. - Change ntfs_attr_make_non_resident() to call ntfs_cluster_alloc()
  33. with @is_extension set to TRUE and remove the runlist terminator
  34. fixup code as this is now done by ntfs_cluster_alloc().
  35. - Change ntfs_attr_make_non_resident to take the attribute value size
  36. as an extra parameter. This is needed since we need to know the size
  37. before we can map the mft record and our callers always know it. The
  38. reason we cannot simply read the size from the vfs inode i_size is
  39. that this is not necessarily uptodate. This happens when
  40. ntfs_attr_make_non_resident() is called in the ->truncate call path.
  41. - Fix ntfs_attr_make_non_resident() to update the vfs inode i_blocks
  42. which is zero for a resident attribute but should no longer be zero
  43. once the attribute is non-resident as it then has real clusters
  44. allocated.
  45. - Add fs/ntfs/attrib.[hc]::ntfs_attr_extend_allocation(), a function to
  46. extend the allocation of an attributes. Optionally, the data size,
  47. but not the initialized size can be extended, too.
  48. - Implement fs/ntfs/inode.[hc]::ntfs_truncate(). It only supports
  49. uncompressed and unencrypted files and it never creates sparse files
  50. at least for the moment (making a file sparse requires us to modify
  51. its directory entries and we do not support directory operations at
  52. the moment). Also, support for highly fragmented files, i.e. ones
  53. whose data attribute is split across multiple extents, is severly
  54. limited. When such a case is encountered, EOPNOTSUPP is returned.
  55. - Enable ATTR_SIZE attribute changes in ntfs_setattr(). This completes
  56. the initial implementation of file truncation. Now both open(2)ing
  57. a file with the O_TRUNC flag and the {,f}truncate(2) system calls
  58. will resize a file appropriately. The limitations are that only
  59. uncompressed and unencrypted files are supported. Also, there is
  60. only very limited support for highly fragmented files (the ones whose
  61. $DATA attribute is split into multiple attribute extents).
  62. - In attrib.c::ntfs_attr_set() call balance_dirty_pages_ratelimited()
  63. and cond_resched() in the main loop as we could be dirtying a lot of
  64. pages and this ensures we play nice with the VM and the system as a
  65. whole.
  66. - Implement file operations ->write, ->aio_write, ->writev for regular
  67. files. This replaces the old use of generic_file_write(), et al and
  68. the address space operations ->prepare_write and ->commit_write.
  69. This means that both sparse and non-sparse (unencrypted and
  70. uncompressed) files can now be extended using the normal write(2)
  71. code path. There are two limitations at present and these are that
  72. we never create sparse files and that we only have limited support
  73. for highly fragmented files, i.e. ones whose data attribute is split
  74. across multiple extents. When such a case is encountered,
  75. EOPNOTSUPP is returned.
  76. - $EA attributes can be both resident and non-resident.
  77. - Use %z for size_t to fix compilation warnings. (Andrew Morton)
  78. - Fix compilation warnings with gcc-4.0.2 on SUSE 10.0.
  79. - Document extended attribute ($EA) NEED_EA flag. (Based on libntfs
  80. patch by Yura Pakhuchiy.)
  81. 2.1.24 - Lots of bug fixes and support more clean journal states.
  82. - Support journals ($LogFile) which have been modified by chkdsk. This
  83. means users can boot into Windows after we marked the volume dirty.
  84. The Windows boot will run chkdsk and then reboot. The user can then
  85. immediately boot into Linux rather than having to do a full Windows
  86. boot first before rebooting into Linux and we will recognize such a
  87. journal and empty it as it is clean by definition. Note, this only
  88. works if chkdsk left the journal in an obviously clean state.
  89. - Support journals ($LogFile) with only one restart page as well as
  90. journals with two different restart pages. We sanity check both and
  91. either use the only sane one or the more recent one of the two in the
  92. case that both are valid.
  93. - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
  94. ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
  95. hence cannot fail.
  96. - Use ntfs_malloc_nofs_nofail() in the two critical regions in
  97. fs/ntfs/runlist.c::ntfs_runlists_merge(). This means we no longer
  98. need to panic() if the allocation fails as it now cannot fail.
  99. - Fix two nasty runlist merging bugs that had gone unnoticed so far.
  100. Thanks to Stefano Picerno for the bug report.
  101. - Remove two bogus BUG_ON()s from fs/ntfs/mft.c.
  102. - Fix handling of valid but empty mapping pairs array in
  103. fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress().
  104. - Report unrepresentable inodes during ntfs_readdir() as KERN_WARNING
  105. messages and include the inode number. Thanks to Yura Pakhuchiy for
  106. pointing this out.
  107. - Change ntfs_rl_truncate_nolock() to throw away the runlist if the new
  108. length is zero.
  109. - Add runlist.[hc]::ntfs_rl_punch_nolock() which punches a caller
  110. specified hole into a runlist.
  111. - Fix a bug in fs/ntfs/index.c::ntfs_index_lookup(). When the returned
  112. index entry is in the index root, we forgot to set the @ir pointer in
  113. the index context. Thanks to Yura Pakhuchiy for finding this bug.
  114. - Remove bogus setting of PageError in ntfs_read_compressed_block().
  115. - Add fs/ntfs/attrib.[hc]::ntfs_resident_attr_value_resize().
  116. - Fix a bug in ntfs_map_runlist_nolock() where we forgot to protect
  117. access to the allocated size in the ntfs inode with the size lock.
  118. - Fix ntfs_attr_vcn_to_lcn_nolock() and ntfs_attr_find_vcn_nolock() to
  119. return LCN_ENOENT when there is no runlist and the allocated size is
  120. zero.
  121. - Fix load_attribute_list() to handle the case of a NULL runlist.
  122. - Fix handling of sparse attributes in ntfs_attr_make_non_resident().
  123. - Add BUG() checks to ntfs_attr_make_non_resident() and ntfs_attr_set()
  124. to ensure that these functions are never called for compressed or
  125. encrypted attributes.
  126. - Fix cluster (de)allocators to work when the runlist is NULL and more
  127. importantly to take a locked runlist rather than them locking it
  128. which leads to lock reversal.
  129. - Truncate {a,c,m}time to the ntfs supported time granularity when
  130. updating the times in the inode in ntfs_setattr().
  131. - Fixup handling of sparse, compressed, and encrypted attributes in
  132. fs/ntfs/inode.c::ntfs_read_locked_{,attr_,index_}inode(),
  133. fs/ntfs/aops.c::ntfs_{read,write}page().
  134. - Make ntfs_write_block() not instantiate sparse blocks if they contain
  135. only zeroes.
  136. - Optimize fs/ntfs/aops.c::ntfs_write_block() by extending the page
  137. lock protection over the buffer submission for i/o which allows the
  138. removal of the get_bh()/put_bh() pairs for each buffer.
  139. - Fix fs/ntfs/aops.c::ntfs_{read,write}_block() to handle the case
  140. where a concurrent truncate has truncated the runlist under our feet.
  141. - Fix page_has_buffers()/page_buffers() handling in fs/ntfs/aops.c.
  142. - In fs/ntfs/aops.c::ntfs_end_buffer_async_read(), use a bit spin lock
  143. in the first buffer head instead of a driver global spin lock to
  144. improve scalability.
  145. - Minor fix to error handling and error message display in
  146. fs/ntfs/aops.c::ntfs_prepare_nonresident_write().
  147. - Change the mount options {u,f,d}mask to always parse the number as
  148. an octal number to conform to how chmod(1) works, too. Thanks to
  149. Giuseppe Bilotta and Horst von Brand for pointing out the errors of
  150. my ways.
  151. - Fix various bugs in the runlist merging code. (Based on libntfs
  152. changes by Richard Russon.)
  153. - Fix sparse warnings that have crept in over time.
  154. - Change ntfs_cluster_free() to require a write locked runlist on entry
  155. since we otherwise get into a lock reversal deadlock if a read locked
  156. runlist is passed in. In the process also change it to take an ntfs
  157. inode instead of a vfs inode as parameter.
  158. - Fix the definition of the CHKD ntfs record magic. It had an off by
  159. two error causing it to be CHKB instead of CHKD.
  160. - Fix a stupid bug in __ntfs_bitmap_set_bits_in_run() which caused the
  161. count to become negative and hence we had a wild memset() scribbling
  162. all over the system's ram.
  163. 2.1.23 - Implement extension of resident files and make writing safe as well as
  164. many bug fixes, cleanups, and enhancements...
  165. - Add printk rate limiting for ntfs_warning() and ntfs_error() when
  166. compiled without debug. This avoids a possible denial of service
  167. attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
  168. out.
  169. - Fix compilation warnings on ia64. (Randy Dunlap)
  170. - Use i_size_{read,write}() instead of reading i_size by hand and cache
  171. the value where apropriate.
  172. - Add size_lock to the ntfs_inode structure. This is an rw spinlock
  173. and it locks against access to the inode sizes. Note, ->size_lock
  174. is also accessed from irq context so you must use the _irqsave and
  175. _irqrestore lock and unlock functions, respectively. Protect all
  176. accesses to allocated_size, initialized_size, and compressed_size.
  177. - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
  178. - Implement extension of resident files in the regular file write code
  179. paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()). At present
  180. this only works until the data attribute becomes too big for the mft
  181. record after which we abort the write returning -EOPNOTSUPP from
  182. ntfs_prepare_write().
  183. - Add disable_sparse mount option together with a per volume sparse
  184. enable bit which is set appropriately and a per inode sparse disable
  185. bit which is preset on some system file inodes as appropriate.
  186. - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
  187. - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
  188. the creation of the unmapped runlist element for the base attribute
  189. extent.
  190. - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
  191. helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
  192. This allows us to map runlist fragments with the runlist lock already
  193. held without having to drop and reacquire it around the call. Adapt
  194. all callers.
  195. - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
  196. runlist. This allows us to find runlist elements with the runlist
  197. lock already held without having to drop and reacquire it around the
  198. call. Adapt all callers.
  199. - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
  200. warning in the do_div() call on sparc32. Thanks to Meelis Roos for
  201. the report and analysis of the warning.
  202. - Fix a nasty runlist merge bug when merging two holes.
  203. - Set the ntfs_inode->allocated_size to the real allocated size in the
  204. mft record for resident attributes (fs/ntfs/inode.c).
  205. - Small readability cleanup to use "a" instead of "ctx->attr"
  206. everywhere (fs/ntfs/inode.c).
  207. - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
  208. definition of ntfs_export_ops from fs/ntfs/super.c to namei.c. Also,
  209. declare ntfs_export_ops in fs/ntfs/ntfs.h.
  210. - Correct sparse file handling. The compressed values need to be
  211. checked and set in the ntfs inode as done for compressed files and
  212. the compressed size needs to be used for vfs inode->i_blocks instead
  213. of the allocated size, again, as done for compressed files.
  214. - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
  215. non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
  216. - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
  217. write code.
  218. - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
  219. dropping the read lock and taking the write lock we were not checking
  220. whether someone else did not already do the work we wanted to do.
  221. - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
  222. ntfs_attr_find_vcn_nolock() and update all callers.
  223. - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
  224. - Fix sign of various error return values to be negative in
  225. fs/ntfs/lcnalloc.c.
  226. - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
  227. handle the case where an attribute is converted from resident to
  228. non-resident by a concurrent file write.
  229. - Remove checks for NULL before calling kfree() since kfree() does the
  230. checking itself. (Jesper Juhl)
  231. - Some utilities modify the boot sector but do not update the checksum.
  232. Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
  233. only emit a warning when the checksum is incorrect rather than
  234. refusing the mount. Thanks to Bernd Casimir for pointing this
  235. problem out.
  236. - Update attribute definition handling.
  237. - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
  238. - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
  239. - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
  240. better code generation and one less sparse warning in fs/ntfs/aops.c.
  241. - Remove spurious void pointer casts from fs/ntfs/. (Pekka Enberg)
  242. - Use C99 style structure initialization after memory allocation where
  243. possible (fs/ntfs/{attrib.c,index.c,super.c}). Thanks to Al Viro and
  244. Pekka Enberg.
  245. - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
  246. is active on the volume and we are mounting read-write or remounting
  247. from read-only to read-write.
  248. - Fix a bug in address space operations error recovery code paths where
  249. if the runlist was not mapped at all and a mapping error occured we
  250. would leave the runlist locked on exit to the function so that the
  251. next access to the same file would try to take the lock and deadlock.
  252. - Detect the case when Windows has been suspended to disk on the volume
  253. to be mounted and if this is the case do not allow (re)mounting
  254. read-write. This is done by parsing hiberfil.sys if present.
  255. - Fix several occurences of a bug where we would perform 'var & ~const'
  256. with a 64-bit variable and a int, i.e. 32-bit, constant. This causes
  257. the higher order 32-bits of the 64-bit variable to be zeroed. To fix
  258. this cast the 'const' to the same 64-bit type as 'var'.
  259. - Change the runlist terminator of the newly allocated cluster(s) to
  260. LCN_ENOENT in ntfs_attr_make_non_resident(). Otherwise the runlist
  261. code gets confused.
  262. - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
  263. and ntfs_mapping_pairs_build() to allow the runlist encoding to be
  264. partial which is desirable when filling holes in sparse attributes.
  265. Update all callers.
  266. - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
  267. if the requested vcn is inside it. Otherwise we get into problems
  268. when we try to map an out of bounds vcn because we then try to map
  269. the already mapped runlist fragment which causes
  270. ntfs_mapping_pairs_decompress() to fail and return error. Update
  271. ntfs_attr_find_vcn_nolock() accordingly.
  272. - Fix a nasty deadlock that appeared in recent kernels.
  273. The situation: VFS inode X on a mounted ntfs volume is dirty. For
  274. same inode X, the ntfs_inode is dirty and thus corresponding on-disk
  275. inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
  276. to the table of inodes, i.e. $MFT, inode 0.
  277. What happens:
  278. Process 1: sys_sync()/umount()/whatever... calls
  279. __sync_single_inode() for $MFT -> do_writepages() -> write_page for
  280. the dirty page containing the on-disk inode X, the page is now locked
  281. -> ntfs_write_mst_block() which clears PageUptodate() on the page to
  282. prevent anyone else getting hold of it whilst it does the write out.
  283. This is necessary as the on-disk inode needs "fixups" applied before
  284. the write to disk which are removed again after the write and
  285. PageUptodate is then set again. It then analyses the page looking
  286. for dirty on-disk inodes and when it finds one it calls
  287. ntfs_may_write_mft_record() to see if it is safe to write this
  288. on-disk inode. This then calls ilookup5() to check if the
  289. corresponding VFS inode is in icache(). This in turn calls ifind()
  290. which waits on the inode lock via wait_on_inode whilst holding the
  291. global inode_lock.
  292. Process 2: pdflush results in a call to __sync_single_inode for the
  293. same VFS inode X on the ntfs volume. This locks the inode (I_LOCK)
  294. then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
  295. read_cache_page() for the page (in page cache of table of inodes
  296. $MFT, inode 0) containing the on-disk inode. This page has
  297. PageUptodate() clear because of Process 1 (see above) so
  298. read_cache_page() blocks when it tries to take the page lock for the
  299. page so it can call ntfs_read_page().
  300. Thus Process 1 is holding the page lock on the page containing the
  301. on-disk inode X and it is waiting on the inode X to be unlocked in
  302. ifind() so it can write the page out and then unlock the page.
  303. And Process 2 is holding the inode lock on inode X and is waiting for
  304. the page to be unlocked so it can call ntfs_readpage() or discover
  305. that Process 1 set PageUptodate() again and use the page.
  306. Thus we have a deadlock due to ifind() waiting on the inode lock.
  307. The solution: The fix is to use the newly introduced
  308. ilookup5_nowait() which does not wait on the inode's lock and hence
  309. avoids the deadlock. This is safe as we do not care about the VFS
  310. inode and only use the fact that it is in the VFS inode cache and the
  311. fact that the vfs and ntfs inodes are one struct in memory to find
  312. the ntfs inode in memory if present. Also, the ntfs inode has its
  313. own locking so it does not matter if the vfs inode is locked.
  314. - Fix bug in mft record writing where we forgot to set the device in
  315. the buffers when mapping them after the VM had discarded them.
  316. Thanks to Martin MOKREJŠ for the bug report.
  317. 2.1.22 - Many bug and race fixes and error handling improvements.
  318. - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
  319. - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
  320. instead of void and provide a helper ntfs_truncate_vfs() for the
  321. vfs ->truncate method.
  322. - Add a new ntfs inode flag NInoTruncateFailed() and modify
  323. fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
  324. - Fix min_size and max_size definitions in ATTR_DEF structure in
  325. fs/ntfs/layout.h to be signed.
  326. - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
  327. ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
  328. ntfs_attr_can_be_resident(), which in turn use the new private helper
  329. ntfs_attr_find_in_attrdef().
  330. - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
  331. mapping->private_lock around the dirtying of the buffer heads
  332. analagous to the way it is done in __set_page_dirty_buffers().
  333. - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
  334. mount time as this cannot work with the current implementation.
  335. - Check for location of attribute name and improve error handling in
  336. general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
  337. - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
  338. i_size, i.e. race with truncate, invalidate the buffers on the page
  339. so that they become freeable and hence the page does not leak.
  340. - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
  341. Bunk)
  342. - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
  343. a NULL pointer dereference in the error code path when a corrupt
  344. attribute was found. (Thanks to Domen Puncer for the bug report.)
  345. - Add MODULE_VERSION() to fs/ntfs/super.c.
  346. - Make several functions and variables static. (Adrian Bunk)
  347. - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
  348. buffers for the page if they are not present and then marks the
  349. buffers belonging to the ntfs record dirty. This causes the buffers
  350. to become busy and hence they are safe from removal until the page
  351. has been written out.
  352. - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
  353. error handling code path that resulted in a BUG() due to trying to
  354. unmap an extent mft record when the mapping of it had failed and it
  355. thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
  356. - Drop the runlist lock after the vcn has been read in
  357. fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
  358. - Rewrite handling of multi sector transfer errors. We now do not set
  359. PageError() when such errors are detected in the async i/o handler
  360. fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
  361. protected attributes now check the magic of each ntfs record as they
  362. use it and act appropriately. This has the effect of making errors
  363. granular per ntfs record rather than per page which solves the case
  364. where we cannot access any of the ntfs records in a page when a
  365. single one of them had an mst error. (Thanks to Ken MacFerrin for
  366. the bug report.)
  367. - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
  368. where we failed to release i_sem on the $Quota/$Q attribute inode.
  369. - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
  370. - Add mapping of unmapped buffers to all remaining code paths, i.e.
  371. fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
  372. and write_mft_record_nolock(). From now on we require that the
  373. complete runlist for the mft mirror is always mapped into memory.
  374. - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
  375. - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
  376. - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
  377. resident attribute will be smaller than a page which makes the code
  378. simpler. Also make the code more tolerant to concurrent ->truncate.
  379. 2.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
  380. - Implement extent mft record deallocation
  381. fs/ntfs/mft.c::ntfs_extent_mft_record_free().
  382. - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
  383. - Add vol->mft_data_pos and initialize it at mount time.
  384. - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
  385. ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
  386. ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
  387. ntfs_runlists_merge() and adapt all callers.
  388. - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
  389. ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
  390. and ntfs_mapping_pairs_build(), adapted from libntfs.
  391. - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
  392. static and add a declaration for it to lcnalloc.h.
  393. - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
  394. inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
  395. cluster bitmap lock for the duration of the call.
  396. - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
  397. - Implement the equivalent of memset() for an ntfs attribute in
  398. fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
  399. fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
  400. - Remove unnecessary casts from LCN_* constants.
  401. - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
  402. - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
  403. change MFT_RECORD to contain the NTFS 3.1+ specific fields.
  404. - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
  405. marks all buffers belonging to an ntfs record dirty, followed by
  406. marking the page the ntfs record is in dirty and also marking the vfs
  407. inode containing the ntfs record dirty (I_DIRTY_PAGES).
  408. - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
  409. new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
  410. longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
  411. - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
  412. include errors.
  413. - Move the typedefs for runlist_element and runlist from types.h to
  414. runlist.h and fix resulting include errors.
  415. - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
  416. - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
  417. mark_ntfs_record_dirty() which also changes the behaviour in that we
  418. now set the buffers belonging to the mft record dirty as well as the
  419. page itself.
  420. - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
  421. to cope with the fact that there now are dirty buffers in mft pages.
  422. - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
  423. mark_ntfs_record_dirty() and thus to set the buffers belonging to the
  424. mft record dirty as well as the page itself.
  425. - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
  426. slightly modified by me)
  427. - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
  428. the mft record is already locked and otherwise behaves the same way
  429. as fs/ntfs/mft.c::map_mft_record().
  430. - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
  431. writes the mft record if the buffers belonging to it are dirty.
  432. Otherwise we assume that it was written out by other means already.
  433. - Attempting to write outside initialized size is _not_ a bug so remove
  434. the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
  435. fact required to write outside initialized size when preparing to
  436. extend the initialized size.
  437. - Map the page instead of using page_address() before writing to it in
  438. fs/ntfs/aops.c::ntfs_mft_writepage().
  439. - Provide exclusion between opening an inode / mapping an mft record
  440. and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
  441. by setting the page not uptodate throughout ntfs_mft_writepage().
  442. - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
  443. to ensure noone can see the page whilst the mst fixups are applied.
  444. - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
  445. checks if an mft record may be written out safely obtaining any
  446. necessary locks in the process. This is used by
  447. fs/ntfs/aops.c::ntfs_write_mst_block().
  448. - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
  449. writing mft records and improve its error handling in the process.
  450. Now if any of the records in the page fail to be written out, all
  451. other records will be written out instead of aborting completely.
  452. - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
  453. - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
  454. ntfs_mst_aops for all inodes which are NInoMstProtected() and
  455. ntfs_aops for all other inodes.
  456. - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
  457. ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
  458. no longer require an ntfs inode to be present. Update all callers.
  459. - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
  460. - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
  461. to ensure noone can see the page whilst the mst fixups are applied.
  462. - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
  463. fs/ntfs/mft.c::try_map_mft_record().
  464. - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
  465. with the ntfs inode which contains the page rather than the ntfs
  466. inode the mft record of which is in the page.
  467. - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
  468. index inode bitmap inode release code from there to
  469. fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
  470. Hellwig for spotting this.)
  471. - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
  472. inode semaphore around the code that sets ni->itype.index.bmp_ino to
  473. NULL and reorganize the code to optimize it a bit. (Thanks to
  474. Christoph Hellwig for spotting this.)
  475. - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
  476. ntfs inode as a parameter as this is confusing and misleading and the
  477. needed ntfs inode is available via NTFS_I(page->mapping->host).
  478. Adapt all callers to this change.
  479. - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
  480. fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
  481. of the first buffer in a record and to take this as the ntfs record
  482. dirty state. We cannot look at the dirty state for subsequent
  483. buffers because we might be racing with
  484. fs/ntfs/aops.c::mark_ntfs_record_dirty().
  485. - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
  486. inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
  487. add a declaration for it to inode.h. Fix some compilation issues
  488. that resulted due to #includes and header file interdependencies.
  489. - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
  490. - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
  491. - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
  492. record sequence number if it is specified (i.e. not zero).
  493. - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
  494. functions used by it.
  495. - Update Documentation/filesystems/ntfs.txt with instructions on how to
  496. use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
  497. the linear raid problem with the Software RAID / MD driver when one
  498. or more of the devices has an odd number of sectors.
  499. 2.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
  500. - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
  501. where we did not clear ctx->al_entry but it was still set due to
  502. changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
  503. particular.
  504. - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
  505. where we forgot to unmap the extent mft record when we had finished
  506. enumerating an attribute which caused a bug check to trigger when the
  507. VFS calls ->clear_inode.
  508. 2.1.19 - Many cleanups, improvements, and a minor bug fix.
  509. - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
  510. change the uid, gid, and mode of an inode as we do not support NTFS
  511. ACLs yet.
  512. - Remove BKL use from ntfs_setattr() syncing up with the rest of the
  513. kernel.
  514. - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
  515. and ntfs_filldir() as per suggestion from Al Viro.
  516. - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
  517. - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
  518. inode size has changed and to only output an error if so.
  519. - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
  520. - Add le{16,32,64} as well as sle{16,32,64} data types to
  521. fs/ntfs/types.h.
  522. - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
  523. - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
  524. respectively, to fs/ntfs/types.h.
  525. - Update endianness conversion macros in fs/ntfs/endian.h to use the
  526. new types as appropriate.
  527. - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
  528. and index.c.
  529. - Add leMFT_REF data type to fs/ntfs/layout.h.
  530. - Update all NTFS header files with the new little endian data types.
  531. Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
  532. - Do proper type casting when using ntfs_is_*_recordp() in
  533. fs/ntfs/logfile.c, mft.c, and super.c.
  534. - Fix all the sparse bitwise warnings. Had to change all the typedef
  535. enums storing little endian values to simple enums plus a typedef for
  536. the datatype to make sparse happy.
  537. - Fix a bug found by the new sparse bitwise warnings where the default
  538. upcase table was defined as a pointer to wchar_t rather than ntfschar
  539. in fs/ntfs/ntfs.h and super.c.
  540. - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
  541. 2.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
  542. - Remove vol->nr_mft_records as it was pretty meaningless and optimize
  543. the calculation of total/free inodes as used by statfs().
  544. - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
  545. because the code itself is using the ntfs_lock semaphore which
  546. provides safe locking. (Ingo Molnar)
  547. - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
  548. could occur in the future for when we start closing/freeing extent
  549. inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
  550. we free it.
  551. - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
  552. find_external_attr() to ntfs_external_attr_find() to cleanup the
  553. namespace a bit and to be more consistent with libntfs.
  554. - Rename {{re,}init,get,put}_attr_search_ctx() to
  555. ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
  556. attr_search_context to ntfs_attr_search_ctx.
  557. - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
  558. for the attribute list attribute itself.
  559. - Fix endianness bug in ntfs_external_attr_find().
  560. - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
  561. if the attribute is not found, and -EIO on real error. In the case
  562. of -ENOENT, the search context is updated to describe the attribute
  563. before which the attribute being searched for would need to be
  564. inserted if such an action were to be desired and in the case of
  565. ntfs_external_attr_find() the search context is also updated to
  566. indicate the attribute list entry before which the attribute list
  567. entry of the attribute being searched for would need to be inserted
  568. if such an action were to be desired. Also make ntfs_find_attr()
  569. static and remove its prototype from attrib.h as it is not used
  570. anywhere other than attrib.c. Update ntfs_attr_lookup() and all
  571. callers of ntfs_{external,}attr_{find,lookup}() for the new return
  572. values.
  573. - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
  574. 2.1.17 - Fix bugs in mount time error code paths and other updates.
  575. - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
  576. includes functions to set/clear a single bit or a run of bits.
  577. - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
  578. runlist element containing a particular vcn. It also takes care of
  579. mapping any needed runlist fragments.
  580. - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
  581. - Load attribute definition table from $AttrDef at mount time.
  582. - Fix bugs in mount time error code paths involving (de)allocation of
  583. the default and volume upcase tables.
  584. - Remove ntfs_nr_mounts as it is no longer used.
  585. 2.1.16 - Implement access time updates, file sync, async io, and read/writev.
  586. - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
  587. This is done by setting the appropriate file operations pointers to
  588. the generic helper functions provided by mm/filemap.c.
  589. - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
  590. and directories (fs/ntfs/dir.c).
  591. - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
  592. Note, except for the root directory and any other system files opened
  593. by the user, the system files will not have their access times
  594. updated as they are only accessed at the inode level an hence the
  595. file level functions which cause the times to be updated are never
  596. invoked.
  597. 2.1.15 - Invalidate quotas when (re)mounting read-write.
  598. - Add new element itype.index.collation_rule to the ntfs inode
  599. structure and set it appropriately in ntfs_read_locked_inode().
  600. - Implement a new inode type "index" to allow efficient access to the
  601. indices found in various system files and adapt inode handling
  602. accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
  603. attribute inode (NInoAttr() is true) with an attribute type of
  604. AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
  605. ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
  606. there would be no way to distinguish between normal attribute inodes
  607. and index inodes. The function to obtain an index inode is
  608. ntfs_index_iget() and it uses the helper function
  609. ntfs_read_locked_index_inode(). Note, we do not overload
  610. ntfs_attr_iget() as indices consist of multiple attributes so using
  611. ntfs_attr_iget() to obtain an index inode would be confusing.
  612. - Ensure that there is no overflow when doing page->index <<
  613. PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
  614. - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
  615. and ntfs_read_block().
  616. - Use case sensitive attribute lookups instead of case insensitive ones.
  617. - Lock all page cache pages belonging to mst protected attributes while
  618. accessing them to ensure we never see corrupt data while the page is
  619. under writeout.
  620. - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
  621. We have ntfs_is_collation_rule_supported() to check if the collation
  622. rule you want to use is supported and ntfs_collation() which actually
  623. collates two data items. We currently only support COLLATION_BINARY
  624. and COLLATION_NTOFS_ULONG but support for other collation rules will
  625. be added as the need arises.
  626. - Add a new type, ntfs_index_context, to allow retrieval of an index
  627. entry using the corresponding index key. To get an index context,
  628. use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
  629. This also adds a new slab cache for the index contexts. To lookup a
  630. key in an index inode, use ntfs_index_lookup(). After modifying an
  631. index entry, call ntfs_index_entry_flush_dcache_page() followed by
  632. ntfs_index_entry_mark_dirty() to ensure the changes are written out
  633. to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
  634. an index entry is in the index allocation attribute rather than the
  635. index root attribute it will not be written out (you will get a
  636. warning message about discarded changes instead).
  637. - Load the quota file ($Quota) and check if quota tracking is enabled
  638. and if so, mark the quotas out of date. This causes windows to
  639. rescan the volume on boot and update all quota entries.
  640. - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
  641. It is simply set to __set_page_dirty_nobuffers() to make sure that
  642. running set_page_dirty() on a page containing mft/ntfs records will
  643. not affect the dirty state of the page buffers.
  644. - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
  645. buffers that are inside the ntfs record in the page dirty after which
  646. it sets the page dirty. This allows ->writepage to only write the
  647. dirty index records rather than having to write all the records in
  648. the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
  649. use this rather than __set_page_dirty_nobuffers().
  650. - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
  651. writing of page cache pages belonging to mst protected attributes
  652. like the index allocation attribute in directory indices and other
  653. indices like $Quota/$Q, etc. This means that the quota is now marked
  654. out of date on all volumes rather than only on ones where the quota
  655. defaults entry is in the index root attribute of the $Quota/$Q index.
  656. 2.1.14 - Fix an NFSd caused deadlock reported by several users.
  657. - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
  658. to a buffer so that we can put the search context and unmap the mft
  659. record before calling the filldir() callback. We need to do this
  660. because of NFSd which calls ->lookup() from its filldir callback()
  661. and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
  662. of the directory and since ntfs_readdir() has got it mapped already
  663. ntfs_lookup() deadlocks.
  664. 2.1.13 - Enable overwriting of resident files and housekeeping of system files.
  665. - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
  666. keeping the mft mirror in sync with the mft when mirrored mft records
  667. are written. The functions are write_mft_record{,_nolock}(). The
  668. implementation is quite rudimentary for now with lots of things not
  669. implemented yet but I am not sure any of them can actually occur so
  670. I will wait for people to hit each one and only then implement it.
  671. - Commit open system inodes at umount time. This should make it
  672. virtually impossible for sync_mft_mirror_umount() to ever be needed.
  673. - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
  674. ntfs super operations. This gives us inode writing via the VFS inode
  675. dirty code paths. Note: Access time updates are not implemented yet.
  676. - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
  677. fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
  678. finally enabling resident file overwrite! (-8 This also includes a
  679. placeholder for ->writepage (ntfs_mft_writepage()), which for now
  680. just redirties the page and returns. Also, at umount time, we for
  681. now throw away all mft data page cache pages after the last call to
  682. ntfs_commit_inode() in the hope that all inodes will have been
  683. written out by then and hence no dirty (meta)data will be lost. We
  684. also check for this case and emit an error message telling the user
  685. to run chkdsk.
  686. - Use set_page_writeback() and end_page_writeback() in the resident
  687. attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
  688. the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
  689. page is clean.
  690. - Implement ntfs_mft_writepage() so it now checks if any of the mft
  691. records in the page are dirty and if so redirties the page and
  692. returns. Otherwise it just returns (after doing set_page_writeback(),
  693. unlock_page(), end_page_writeback() or the radix-tree tag
  694. PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
  695. alowing the VM to do with the page as it pleases. Also, at umount
  696. time, now only throw away dirty mft (meta)data pages if dirty inodes
  697. are present and ask the user to email us if they see this happening.
  698. - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
  699. information flags (fs/ntfs/super.c).
  700. - Mark the volume dirty when (re)mounting read-write and mark it clean
  701. when unmounting or remounting read-only. If any volume errors are
  702. found, the volume is left marked dirty to force chkdsk to run.
  703. - Add code to set the NT4 compatibility flag when (re)mounting
  704. read-write for newer NTFS versions but leave it commented out for now
  705. since we do not make any modifications that are NTFS 1.2 specific yet
  706. and since setting this flag breaks Captive-NTFS which is not nice.
  707. This code must be enabled once we start writing NTFS 1.2 specific
  708. changes otherwise Windows NTFS driver might crash / cause corruption.
  709. 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
  710. - Add a new address space operations struct, ntfs_mst_aops, for mst
  711. protected attributes. This is because the default ntfs_aops do not
  712. make sense with mst protected data and were they to write anything to
  713. such an attribute they would cause data corruption so we provide
  714. ntfs_mst_aops which does not have any write related operations set.
  715. - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
  716. includes an adapted ntfs_commit_inode() and an implementation of
  717. ntfs_write_inode() which for now just cleans dirty inodes without
  718. writing them (it does emit a warning that this is happening).
  719. - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
  720. entry) as it was only fixing a theoretical bug but at the same time
  721. it badly broke the handling of sparse and uncompressed compression
  722. blocks.
  723. 2.1.11 - Driver internal cleanups.
  724. - Only build logfile.o if building the driver with read-write support.
  725. - Really final white space cleanups.
  726. - Use generic_ffs() instead of ffs() in logfile.c which allows the
  727. log_page_size variable to be optimized by gcc into a constant.
  728. - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
  729. char as defined by POSIX and as found on some systems.
  730. 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
  731. - Finish off the white space cleanups (remove trailing spaces, etc).
  732. - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
  733. the kludges around the first iget(). Instead of (re)setting ->s_op
  734. we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
  735. insert_inode_hash() / call ntfs_read_inode_mount() directly. This
  736. kills the need for second super_operations and allows to return error
  737. from ntfs_read_inode_mount() without resorting to ugly "poisoning"
  738. tricks. (Al Viro)
  739. - Force read-only (re)mounting if any of the following bits are set in
  740. the volume information flags:
  741. VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
  742. VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
  743. VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
  744. To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
  745. above bits set so the test is made easy.
  746. 2.1.9 - Fix two bugs in decompression engine.
  747. - Fix a bug where we would not always detect that we have reached the
  748. end of a compression block because we were ending at minus one byte
  749. which is effectively the same as being at the end. The fix is to
  750. check whether the uncompressed buffer has been fully filled and if so
  751. we assume we have reached the end of the compression block. A big
  752. thank you to Marcin Gibuła for the bug report, the assistance in
  753. tracking down the bug and testing the fix.
  754. - Fix a possible bug where when a compressed read is truncated to the
  755. end of the file, the offset inside the last page was not truncated.
  756. 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
  757. - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
  758. - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
  759. utc2ntfs() to work with struct timespec instead of time_t on the
  760. Linux UTC time side thus preserving the full precision of the NTFS
  761. time and only loosing up to 99 nano-seconds in the Linux UTC time.
  762. - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
  763. static inline.
  764. - Remove unused ntfs_dirty_inode().
  765. - Cleanup super operations declaration in fs/ntfs/super.c.
  766. - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
  767. - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
  768. fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
  769. - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
  770. fs/ntfs/inode.h so they can be used elsewhere.
  771. - Determine the mft mirror size as the number of mirrored mft records
  772. and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
  773. - Load the mft mirror at mount time and compare the mft records stored
  774. in it to the ones in the mft. Force a read-only mount if the two do
  775. not match (fs/ntfs/super.c).
  776. - Fix type casting related warnings on 64-bit architectures. Thanks
  777. to Meelis Roos for reporting them.
  778. - Move %L to %ll as %L is floating point and %ll is integer which is
  779. what we want.
  780. - Read the journal ($LogFile) and determine if the volume has been
  781. shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
  782. and fs/ntfs/logfile.c). This is a little bit of a crude check in
  783. that we only look at the restart areas and not at the actual log
  784. records so that there will be a very small number of cases where we
  785. think that a volume is dirty when in fact it is clean. This should
  786. only affect volumes that have not been shutdown cleanly and did not
  787. have any pending, non-check-pointed i/o.
  788. - If the $LogFile indicates a clean shutdown and a read-write (re)mount
  789. is requested, empty $LogFile by overwriting it with 0xff bytes to
  790. ensure that Windows cannot cause data corruption by replaying a stale
  791. journal after Linux has written to the volume.
  792. 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
  793. - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
  794. - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
  795. - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
  796. default doesn't allow inode number 0 which is a valid inode on NTFS
  797. and even if it did allow that it uses iget() instead of ntfs_iget()
  798. which makes it useless for us.
  799. - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
  800. default just returns -EACCES which is not very useful.
  801. - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
  802. and set them up in the super block at mount time (super.c) this
  803. allows mounted NTFS volumes to be exported via NFS.
  804. - Add missing return -EOPNOTSUPP; in
  805. fs/ntfs/aops.c::ntfs_commit_nonresident_write().
  806. - Enforce no atime and no dir atime updates at mount/remount time as
  807. they are not implemented yet anyway.
  808. - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
  809. after a NULL check. Thanks to Dave Jones for pointing this out.
  810. 2.1.6 - Fix minor bug in handling of compressed directories.
  811. - Fix bug in handling of compressed directories. A compressed
  812. directory is not really compressed so when we set the ->i_blocks
  813. field of a compressed directory inode we were setting it from the
  814. non-existing field ni->itype.compressed.size which gave random
  815. results... For directories we now always use ni->allocated_size.
  816. 2.1.5 - Fix minor bug in attribute list attribute handling.
  817. - Fix bug in attribute list handling. Actually it is not as much a bug
  818. as too much protection in that we were not allowing attribute lists
  819. which waste space on disk while Windows XP clearly allows it and in
  820. fact creates such attribute lists so our driver was failing.
  821. - Update NTFS documentation ready for 2.6 kernel release.
  822. 2.1.4 - Reduce compiler requirements.
  823. - Remove all uses of unnamed structs and unions in the driver to make
  824. old and newer gcc versions happy. Makes it a bit uglier IMO but at
  825. least people will stop hassling me about it.
  826. 2.1.3 - Important bug fixes in corner cases.
  827. - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
  828. clusters. (Philipp Thomas)
  829. - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
  830. multiple of the block_size but not the cluster size. (Szabolcs
  831. Szakacsits <szaka@sienet.hu>)
  832. 2.1.2 - Important bug fixes aleviating the hangs in statfs.
  833. - Fix buggy free cluster and free inode determination logic.
  834. 2.1.1 - Minor updates.
  835. - Add handling for initialized_size != data_size in compressed files.
  836. - Reduce function local stack usage from 0x3d4 bytes to just noise in
  837. fs/ntfs/upcase.c. (Randy Dunlap <rdunlap@xenotime.net>)
  838. - Remove compiler warnings for newer gcc.
  839. - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
  840. around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
  841. in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
  842. kmap_atomic(KM_USER0).
  843. 2.1.0 - First steps towards write support: implement file overwrite.
  844. - Add configuration option for developmental write support with an
  845. appropriately scary configuration help text.
  846. - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
  847. helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
  848. overwriting of existing files on ntfs. Note: Resident files are
  849. only written into memory, and not written out to disk at present, so
  850. avoid writing to files smaller than about 1kiB.
  851. - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
  852. helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
  853. counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
  854. fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
  855. add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
  856. This enables write(2) based overwriting of existing files on ntfs.
  857. Note: As with mmap(2) based overwriting, resident files are only
  858. written into memory, and not written out to disk at present, so avoid
  859. writing to files smaller than about 1kiB.
  860. - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
  861. ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
  862. files with the purpose of intercepting and aborting all i_size
  863. changes which we do not support yet. ntfs_truncate() actually only
  864. emits a warning message but AFAICS our interception of i_size changes
  865. elsewhere means ntfs_truncate() never gets called for i_size changes.
  866. It is only called from generic_file_write() when we fail in
  867. ntfs_prepare_{,nonresident_}write() in order to discard any
  868. instantiated buffers beyond i_size. Thus i_size is not actually
  869. changed so our warning message is enough. Unfortunately it is not
  870. possible to easily determine if i_size is being changed or not hence
  871. we just emit an appropriately worded error message.
  872. 2.0.25 - Small bug fixes and cleanups.
  873. - Unlock the page in an out of memory error code path in
  874. fs/ntfs/aops.c::ntfs_read_block().
  875. - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
  876. just unlock the page and return. (This can happen due to ->writepage
  877. clearing PageUptodate() during write out of MstProtected()
  878. attributes.
  879. - Remove leaked write code again.
  880. 2.0.24 - Cleanups.
  881. - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
  882. inside BUG_ON(). (Adam J. Richter)
  883. - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
  884. calls for improved debugging. (Adam J. Richter)
  885. - Add errors flag to the ntfs volume state, accessed via
  886. NVol{,Set,Clear}Errors(vol).
  887. - Do not allow read-write remounts of read-only volumes with errors.
  888. - Clarify comment for ntfs file operation sendfile which was added by
  889. Christoph Hellwig a while ago (just using generic_file_sendfile())
  890. to say that ntfs ->sendfile is only used for the case where the
  891. source data is on the ntfs partition and the destination is
  892. somewhere else, i.e. nothing we need to concern ourselves with.
  893. - Add generic_file_write() as our ntfs file write operation.
  894. 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
  895. - Massive internal locking changes to mft record locking. Fixes lock
  896. recursion and replaces the mrec_lock read/write semaphore with a
  897. mutex. Also removes the now superfluous mft_count. This fixes several
  898. race conditions and deadlocks, especially in the future write code.
  899. - Fix ntfs over loopback for compressed files by adding an
  900. optimization barrier. (gcc was screwing up otherwise ?)
  901. - Miscellaneous cleanups all over the code and a fix or two in error
  902. handling code paths.
  903. Thanks go to Christoph Hellwig for pointing out the following two:
  904. - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
  905. - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
  906. 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
  907. - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
  908. at entry/exit respectively.
  909. - Use C99 initializers for structures.
  910. - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
  911. 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
  912. - Limit volume size at mount time to 2TiB on architectures where
  913. unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
  914. This is the most we can do without overflowing the 32-bit limit of
  915. the block device size imposed on us by sb_bread() and sb_getblk()
  916. for the time being.
  917. - Limit file/directory size at open() time to 16TiB on architectures
  918. where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
  919. fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
  920. overflowing the page cache page index.
  921. 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
  922. - Move the directory index bitmap to use an attribute inode instead of
  923. having special fields for it inside the ntfs inode structure. This
  924. means that the index bitmaps now use the page cache for i/o, too,
  925. and also as a side effect we get support for non-resident index
  926. bitmaps for free.
  927. - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
  928. fix a page leak that manifested itself in some cases.
  929. - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
  930. index bitmap inode on the final iput().
  931. 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
  932. - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
  933. to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
  934. - Drop the "file" from ntfs_file_read_compressed_block().
  935. - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
  936. ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
  937. - Update ntfs_end_buffer_async_read() with the improved logic from
  938. its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
  939. further logic improvements to better determine when we set PageError.
  940. - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
  941. check for the buffers being uptodate first in line with the updated
  942. fs/buffer.c::block_read_full_page(). This plugs a small race
  943. condition.
  944. 2.0.18 - Fix race condition in reading of compressed files.
  945. - There was a narrow window between checking a buffer head for being
  946. uptodate and locking it in ntfs_file_read_compressed_block(). We now
  947. lock the buffer and then check whether it is uptodate or not.
  948. 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
  949. - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
  950. code and update callers, i.e. ntfs_iget(), to pass that error code
  951. up instead of just using -EIO.
  952. - Modifications to super.c to ensure that both mount and remount
  953. cannot set any write related options when the driver is compiled
  954. read-only.
  955. - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
  956. cache the current runlist element. This should improve performance
  957. when reading very large and/or very fragmented data.
  958. 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
  959. - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
  960. wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
  961. - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
  962. - Convert $MFT/$BITMAP access to attribute inode API and remove all
  963. remnants of the ugly mftbmp address space and operations hack. This
  964. means we finally have only one readpage function as well as only one
  965. async io completion handler. Yey! The mft bitmap is now just an
  966. attribute inode and is accessed from vol->mftbmp_ino just as if it
  967. were a normal file. Fake inodes rule. (-:
  968. 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
  969. - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
  970. remounts to fail when the partition had an entry in /etc/fstab and
  971. the entry specified the nls= option.
  972. - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
  973. expand all the helper functions NVolFoo(), NVolSetFoo(), and
  974. NVolClearFoo().
  975. - Move copyright statement from driver initialisation message to
  976. module description (fs/super.c). This makes the initialisation
  977. message fit on one line and fits in better with rest of kernel.
  978. - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
  979. attribute inodes, and both for files and directories.
  980. - Implement fake attribute inodes allowing all attribute i/o to go via
  981. the page cache and to use all the normal vfs/mm functionality:
  982. - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
  983. to fs/ntfs/inode.c.
  984. - Add needed cleanup code to ntfs_clear_big_inode().
  985. - Merge address space operations for files and directories (aops.c),
  986. now just have ntfs_aops:
  987. - Rename:
  988. end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
  989. ntfs_attr_read_block() -> ntfs_read_block(),
  990. ntfs_file_read_page() -> ntfs_readpage().
  991. - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
  992. attribute inodes, and both for files and directories.
  993. - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
  994. 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
  995. - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
  996. the locking out of super.c::get_nr_free_mft_records() and taking and
  997. dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
  998. - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
  999. current userspace ntfs library code. This means that if a merge
  1000. fails the original runlists are always left unmodified instead of
  1001. being silently corrupted.
  1002. - Misc typo fixes.
  1003. 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
  1004. - Remove nr_mft_bits and the now superfluous union with nr_mft_records
  1005. from ntfs_volume structure.
  1006. - Remove nr_lcn_bits and the now superfluous union with nr_clusters
  1007. from ntfs_volume structure.
  1008. - Use iget5_locked() and friends instead of conventional iget(). Wrap
  1009. the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
  1010. to use ntfs_iget(). Leave only one iget() call at mount time so we
  1011. don't need an ntfs_iget_mount().
  1012. - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
  1013. additional argument.
  1014. 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
  1015. - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
  1016. fs/ntfs/aops.c::end_buffer_read_file_async() into one function
  1017. fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
  1018. to determine whether to apply mst fixups or not.
  1019. - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
  1020. and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
  1021. fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
  1022. fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
  1023. the VFS readpage function prototype to the ntfs_attr_read_block()
  1024. function prototype.
  1025. 2.0.11 - Initial preparations for fake inode based attribute i/o.
  1026. - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
  1027. do some macro magic (adapted from include/linux/buffer_head.h) to
  1028. expand all the helper functions NInoFoo(), NInoSetFoo(), and
  1029. NInoClearFoo().
  1030. - Add new flag to ntfs_inode_state_bits: NI_Sparse.
  1031. - Add new fields to ntfs_inode structure to allow use of fake inodes
  1032. for attribute i/o: type, name, name_len. Also add new state bits:
  1033. NI_Attr, which, if set, indicates the inode is a fake inode, and
  1034. NI_MstProtected, which, if set, indicates the attribute uses multi
  1035. sector transfer protection, i.e. fixups need to be applied after
  1036. reads and before/after writes.
  1037. - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
  1038. ntfs_{new,clear,destroy}_extent_inode() and update callers.
  1039. - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
  1040. instead of ntfs_destroy_extent_inode().
  1041. - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
  1042. - Make all operations on ntfs inode state bits use the NIno* functions.
  1043. - Set up the new ntfs inode fields and state bits in
  1044. fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
  1045. allocated memory to __ntfs_clear_inode().
  1046. - Cleanup ntfs_inode structure a bit for better ordering of elements
  1047. w.r.t. their size to allow better packing of the structure in memory.
  1048. 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
  1049. - Add check at mount time to verify that the number of inodes on the
  1050. volume does not exceed 2^32 - 1, which is the maximum allowed for
  1051. NTFS according to Microsoft.
  1052. - Change mft_no member of ntfs_inode structure to be unsigned long.
  1053. Update all users. This makes ntfs_inode->mft_no just a copy of struct
  1054. inode->i_ino. But we can't just always use struct inode->i_ino and
  1055. remove mft_no because extent inodes do not have an attached struct
  1056. inode.
  1057. 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
  1058. - Change decompression engine to use a single buffer protected by a
  1059. spin lock instead of per-CPU buffers. (Rusty Russell)
  1060. - Do not update cb_pos when handling a partial final page during
  1061. decompression of a sparse compression block, as the value is later
  1062. reset without being read/used. (Rusty Russell)
  1063. - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
  1064. Morton)
  1065. - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
  1066. NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
  1067. it also makes everything safer so it is a good thing.
  1068. - Miscellaneous minor cleanups to comments.
  1069. 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
  1070. Big thanks go to Al Viro and other inhabitants of #kernel for investing
  1071. their time to discuss the case sensitivity and dcache aliasing issues.
  1072. - Remove unused source file fs/ntfs/attraops.c.
  1073. - Remove show_inodes mount option(s), thus dropping support for
  1074. displaying of short file names.
  1075. - Remove deprecated mount option posix.
  1076. - Restore show_sys_files mount option.
  1077. - Add new mount option case_sensitive, to determine if the driver
  1078. treats file names as case sensitive or not. If case sensitive, create
  1079. file names in the POSIX namespace. Otherwise create file names in the
  1080. LONG/WIN32 namespace. Note, files remain accessible via their short
  1081. file name, if it exists.
  1082. - Remove really dumb logic bug in boot sector recovery code.
  1083. - Fix dcache aliasing issues wrt short/long file names via changes
  1084. to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
  1085. fs/ntfs/namei.c::ntfs_lookup():
  1086. - Add additional argument to ntfs_lookup_inode_by_name() in which we
  1087. return information about the matching file name if the case is not
  1088. matching or the match is a short file name. See comments above the
  1089. function definition for details.
  1090. - Change ntfs_lookup() to only create dcache entries for the correctly
  1091. cased file name and only for the WIN32 namespace counterpart of DOS
  1092. namespace file names. This ensures we have only one dentry per
  1093. directory and also removes all dcache aliasing issues between short
  1094. and long file names once we add write support. See comments above
  1095. function for details.
  1096. - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
  1097. 2.0.7 - Minor cleanups and updates for changes in core kernel code.
  1098. - Remove much of the NULL struct element initializers.
  1099. - Various updates to make compatible with recent kernels.
  1100. - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
  1101. in fs/ntfs/ntfs.h instead.
  1102. - Remove no longer needed KERNEL_VERSION checks. We are now in the
  1103. kernel proper so they are no longer needed.
  1104. 2.0.6 - Major bugfix to make compatible with other kernel changes.
  1105. - Initialize the mftbmp address space properly now that there are more
  1106. fields in the struct address_space. This was leading to hangs and
  1107. oopses on umount since 2.5.12 because of changes to other parts of
  1108. the kernel. We probably want a kernel generic init_address_space()
  1109. function...
  1110. - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
  1111. only caller of ->readdir() is vfs_readdir() which holds i_sem during
  1112. the call, and i_sem is sufficient protection against changes in the
  1113. directory inode (including ->i_size).
  1114. - Use generic_file_llseek() for directories (as opposed to
  1115. default_llseek()) as this downs i_sem instead of the BKL which is
  1116. what we now need for exclusion against ->f_pos changes considering we
  1117. no longer take the BKL in ntfs_readdir().
  1118. 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
  1119. - No need to set old blocksize in super.c::ntfs_fill_super() as the
  1120. VFS does so via invocation of deactivate_super() calling
  1121. fs->fill_super() calling block_kill_super() which does it.
  1122. - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
  1123. -> Do we really need it? I don't think so as we have exclusion on
  1124. the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
  1125. move the ->f_pos accesses under the mrec_lock though. Check this...
  1126. - Fix really, really, really stupid buffer overflow in extent inode
  1127. handling in mft.c::map_extent_mft_record().
  1128. 2.0.4 - Cleanups and updates for kernel 2.5.11.
  1129. - Add documentation on how to use the MD driver to be able to use NTFS
  1130. stripe and volume sets in Linux and generally cleanup documentation
  1131. a bit.
  1132. Remove all uses of kdev_t in favour of struct block_device *:
  1133. - Change compress.c::ntfs_file_read_compressed_block() to use
  1134. sb_getblk() instead of getblk().
  1135. - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
  1136. of get_hardsect_size().
  1137. - No need to get old blocksize in super.c::ntfs_fill_super() as
  1138. fs/super.c::get_sb_bdev() already does this.
  1139. - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
  1140. 2.0.3 - Small bug fixes, cleanups, and performance improvements.
  1141. - Remove some dead code from mft.c.
  1142. - Optimize readpage and read_block functions throughout aops.c so that
  1143. only initialized blocks are read. Non-initialized ones have their
  1144. buffer head mapped, zeroed, and set up to date, without scheduling
  1145. any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
  1146. Thanks go to Andrew Morton for spotting the below:
  1147. - Fix buglet in allocate_compression_buffers() error code path.
  1148. - Call flush_dcache_page() after modifying page cache page contents in
  1149. ntfs_file_readpage().
  1150. - Check for existence of page buffers throughout aops.c before calling
  1151. create_empty_buffers(). This happens when an I/O error occurs and the
  1152. read is retried. (It also happens once writing is implemented so that
  1153. needed doing anyway but I had left it for later...)
  1154. - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
  1155. readpage and read_block functions. Reasoning same as above (i.e. I/O
  1156. error retries and future write code paths.)
  1157. 2.0.2 - Minor updates and cleanups.
  1158. - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
  1159. and cleanup the code a bit, removing the unused size parameter.
  1160. - Change default fmask to 0177 and update documentation.
  1161. - Change attrib.c::get_attr_search_ctx() to return the search context
  1162. directly instead of taking the address of a pointer. A return value
  1163. of NULL means the allocation failed. Updated all callers
  1164. appropriately.
  1165. - Update to 2.5.9 kernel (preserving backwards compatibility) by
  1166. replacing all occurences of page->buffers with page_buffers(page).
  1167. - Fix minor bugs in runlist merging, also minor cleanup.
  1168. - Updates to bootsector layout and mft mirror contents descriptions.
  1169. - Small bug fix in error detection in unistr.c and some cleanups.
  1170. - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
  1171. bytes.
  1172. 2.0.1 - Minor updates.
  1173. - Make default umask correspond to documentation.
  1174. - Improve documentation.
  1175. - Set default mode to include execute bit. The {u,f,d}mask can be used
  1176. to take it away if desired. This allows binaries to be executed from
  1177. a mounted ntfs partition.
  1178. 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
  1179. - Add kill_super, just keeping up with the vfs changes in the kernel.
  1180. - Repeat some changes from tng-0.0.8 that somehow got lost on the way
  1181. from the CVS import into BitKeeper.
  1182. - Begin to implement proper handling of allocated_size vs
  1183. initialized_size vs data_size (i.e. i_size). Done are
  1184. mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
  1185. and attrib.c::load_attribute_list().
  1186. - Lock the runlist in attrib.c::load_attribute_list() while using it.
  1187. - Fix memory leak in ntfs_file_read_compressed_block() and generally
  1188. clean up compress.c a little, removing some uncommented/unused debug
  1189. code.
  1190. - Tidy up dir.c a little bit.
  1191. - Don't bother getting the runlist in inode.c::ntfs_read_inode().
  1192. - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
  1193. creating aops.c::ntfs_mst_readpage(), improving the handling of
  1194. holes and overflow in the process and implementing the correct
  1195. equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
  1196. I am aiming for correctness at the moment. Modularisation can come
  1197. later.
  1198. - Rename aops.c::end_buffer_read_index_async() to
  1199. end_buffer_read_mst_async() and optimize the overflow checking and
  1200. handling.
  1201. - Use the host of the mftbmp address space mapping to hold the ntfs
  1202. volume. This is needed so the async i/o completion handler can
  1203. retrieve a pointer to the volume. Hopefully this will not cause
  1204. problems elsewhere in the kernel... Otherwise will need to use a
  1205. fake inode.
  1206. - Complete implementation of proper handling of allocated_size vs
  1207. initialized_size vs data_size (i.e. i_size) in whole driver.
  1208. Basically aops.c is now completely rewritten.
  1209. - Change NTFS driver name to just NTFS and set version number to 2.0.0
  1210. to make a clear distinction from the old driver which is still on
  1211. version 1.1.22.
  1212. tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
  1213. - Replace bdevname(sb->s_dev) with sb->s_id.
  1214. - Remove now superfluous new-line characters in all callers of
  1215. ntfs_debug().
  1216. - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
  1217. directories. Without this the "find" utility gets very upset which is
  1218. fair enough as Linux/Unix do not support directory hard links.
  1219. - Further runlist merging work. (Richard Russon)
  1220. - Backwards compatibility for gcc-2.95. (Richard Russon)
  1221. - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
  1222. - Convert to new filesystem declaration using ->ntfs_get_sb() and
  1223. replacing ntfs_read_super() with ntfs_fill_super().
  1224. - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
  1225. overflow on 32-bit architectures.
  1226. - Cleanup upcase loading code to use ntfs_(un)map_page().
  1227. - Disable/reenable preemtion in critical sections of compession engine.
  1228. - Replace device size determination in ntfs_fill_super() with
  1229. sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
  1230. function super.c::get_nr_blocks().
  1231. - Implement a mount time option (show_inodes) allowing choice of which
  1232. types of inode names readdir() returns and modify ntfs_filldir()
  1233. accordingly. There are several parameters to show_inodes:
  1234. system: system files
  1235. win32: long file names (including POSIX file names) [DEFAULT]
  1236. long: same as win32
  1237. dos: short file names only (excluding POSIX file names)
  1238. short: same as dos
  1239. posix: same as both win32 and dos
  1240. all: all file names
  1241. Note that the options are additive, i.e. specifying:
  1242. -o show_inodes=system,show_inodes=win32,show_inodes=dos
  1243. is the same as specifying:
  1244. -o show_inodes=all
  1245. Note that the "posix" and "all" options will show all directory
  1246. names, BUT the link count on each directory inode entry is set to 1,
  1247. due to Linux not supporting directory hard links. This may well
  1248. confuse some userspace applications, since the directory names will
  1249. have the same inode numbers. Thus it is NOT advisable to use the
  1250. "posix" or "all" options. We provide them only for completeness sake.
  1251. - Add copies of allocated_size, initialized_size, and compressed_size to
  1252. the ntfs inode structure and set them up in
  1253. inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
  1254. for files and the index allocation attribute for directories.
  1255. - Add copies of allocated_size and initialized_size to ntfs inode for
  1256. $BITMAP attribute of large directories and set them up in
  1257. inode.c::ntfs_read_inode().
  1258. - Add copies of allocated_size and initialized_size to ntfs volume for
  1259. $BITMAP attribute of $MFT and set them up in
  1260. super.c::load_system_files().
  1261. - Parse deprecated ntfs driver options (iocharset, show_sys_files,
  1262. posix, and utf8) and tell user what the new options to use are. Note
  1263. we still do support them but they will be removed with kernel 2.7.x.
  1264. - Change all occurences of integer long long printf formatting to hex
  1265. as printk() will not support long long integer format if/when the
  1266. div64 patch goes into the kernel.
  1267. - Make slab caches have stable names and change the names to what they
  1268. were intended to be. These changes are required/made possible by the
  1269. new slab cache name handling which removes the length limitation by
  1270. requiring the caller of kmem_cache_create() to supply a stable name
  1271. which is then referenced but not copied.
  1272. - Rename run_list structure to run_list_element and create a new
  1273. run_list structure containing a pointer to a run_list_element
  1274. structure and a read/write semaphore. Adapt all users of runlists
  1275. to new scheme and take and release the lock as needed. This fixes a
  1276. nasty race as the run_list changes even when inodes are locked for
  1277. reading and even when the inode isn't locked at all, so we really
  1278. needed the serialization. We use a semaphore rather than a spinlock
  1279. as memory allocations can sleep and doing everything GFP_ATOMIC
  1280. would be silly.
  1281. - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
  1282. This can never happen due to the nature of lookup_attr() and how we
  1283. support attribute lists. If it did happen it would imply the inode
  1284. being corrupt.
  1285. - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
  1286. bad if found.
  1287. - Update to 2.5.6-pre2 changes in struct address_space.
  1288. - Use parent_ino() when accessing d_parent inode number in dir.c.
  1289. - Import Sourceforge CVS repository into BitKeeper repository:
  1290. http://linux-ntfs.bkbits.net/ntfs-tng-2.5
  1291. - Update fs/Makefile, fs/Config.help, fs/Config.in, and
  1292. Documentation/filesystems/ntfs.txt for NTFS TNG.
  1293. - Create kernel configuration option controlling whether debugging
  1294. is enabled or not.
  1295. - Add the required export of end_buffer_io_sync() from the patches
  1296. directory to the kernel code.
  1297. - Update inode.c::ntfs_show_options() with show_inodes mount option.
  1298. - Update errors mount option.
  1299. tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
  1300. - Cleanup mft.c and it's debug/error output in particular. Fix a minor
  1301. bug in mapping of extent inodes. Update all the comments to fit all
  1302. the recent code changes.
  1303. - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
  1304. - Cleanups in compress.c, mostly comments and folding help.
  1305. - Implement attrib.c::map_run_list() as a generic helper.
  1306. - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
  1307. thus making code shorter and enabling attribute list support.
  1308. - Cleanup incorrect use of [su]64 with %L printf format specifier in
  1309. all source files. Type casts to [unsigned] long long added to correct
  1310. the mismatches (important for architectures which have long long not
  1311. being 64 bits).
  1312. - Merge async io completion handlers for directory indexes and $MFT
  1313. data into one by setting the index_block_size{_bits} of the ntfs
  1314. inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
  1315. - Cleanup aops.c, update comments.
  1316. - Make ntfs_file_get_block() use map_run_list() so all files now
  1317. support attribute lists.
  1318. - Make ntfs_dir_readpage() almost verbatim copy of
  1319. block_read_full_page() by using ntfs_file_get_block() with only real
  1320. difference being the use of our own async io completion handler
  1321. rather than the default one, thus reducing the amount of code and
  1322. automatically enabling attribute list support for directory indices.
  1323. - Fix bug in load_attribute_list() - forgot to call brelse in error
  1324. code path.
  1325. - Change parameters to find_attr() and lookup_attr(). We no longer
  1326. pass in the upcase table and its length. These can be gotten from
  1327. ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
  1328. - Cleanups in attrib.c.
  1329. - Implement merging of runlists, attrib.c::merge_run_lists() and its
  1330. helpers. (Richard Russon)
  1331. - Attribute lists part 2, attribute extents and multi part runlists:
  1332. enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
  1333. further runlist parts via attrib.c::map_run_list().
  1334. - Tiny endianness bug fix in decompress_mapping_pairs().
  1335. tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
  1336. - Enable encrypted directories. (Their index root is marked encrypted
  1337. to indicate that new files in that directory should be created
  1338. encrypted.)
  1339. - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
  1340. - Enable $Extend system directory. Most (if not all) extended system
  1341. files do not have unnamed data attributes so ntfs_read_inode() had to
  1342. special case them but that is ok, as the special casing recovery
  1343. happens inside an error code path so there is zero slow down in the
  1344. normal fast path. The special casing is done by introducing a new
  1345. function inode.c::ntfs_is_extended_system_file() which checks if any
  1346. of the hard links in the inode point to $Extend as being their parent
  1347. directory and if they do we assume this is an extended system file.
  1348. - Create a sysctl/proc interface to allow {dis,en}abling of debug output
  1349. when compiled with -DDEBUG. Default is debug messages to be disabled.
  1350. To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
  1351. (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
  1352. interface is enabled). Inspired by old ntfs driver.
  1353. - Add debug_msgs insmod/kernel boot parameter to set whether debug
  1354. messages are {dis,en}abled. This is useful to enable debug messages
  1355. during ntfs initialization and is the only way to activate debugging
  1356. when the sysctl interface is not enabled.
  1357. - Cleanup debug output in various places.
  1358. - Remove all dollar signs ($) from the source (except comments) to
  1359. enable compilation on architectures whose gcc compiler does not
  1360. support dollar signs in the names of variables/constants. Attribute
  1361. types now start with AT_ instead of $ and $I30 is now just I30.
  1362. - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
  1363. - Load complete runlist for $MFT/$BITMAP during mount and cleanup
  1364. access functions. This means we now cope with $MFT/$BITMAP being
  1365. spread accross several mft records.
  1366. - Disable modification of mft_zone_multiplier on remount. We can always
  1367. reenable this later on if we really want to, but we will need to make
  1368. sure we readjust the mft_zone size / layout accordingly.
  1369. tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
  1370. - Use sb_set_blocksize() instead of set_blocksize() and verify the
  1371. return value.
  1372. - Use sb_bread() instead of bread() throughout.
  1373. - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
  1374. of a directory index block vcn. Apply resulting simplifications in
  1375. dir.c everywhere.
  1376. - Fix a small bug somewhere (but forgot what it was).
  1377. - Change ntfs_{debug,error,warning} to enable gcc to do type checking
  1378. on the printf-format parameter list and fix bugs reported by gcc
  1379. as a result. (Richard Russon)
  1380. - Move inode allocation strategy to Al's new stuff but maintain the
  1381. divorce of ntfs_inode from struct inode. To achieve this we have two
  1382. separate slab caches, one for big ntfs inodes containing a struct
  1383. inode and pure ntfs inodes and at the same time fix some faulty
  1384. error code paths in ntfs_read_inode().
  1385. - Show mount options in proc (inode.c::ntfs_show_options()).
  1386. tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
  1387. - Modified (un)map_mft_record functions to be common for read and write
  1388. case. To specify which is which, added extra parameter at front of
  1389. parameter list. Pass either READ or WRITE to this, each has the
  1390. obvious meaning.
  1391. - General cleanups to allow for easier folding in vi.
  1392. - attrib.c::decompress_mapping_pairs() now accepts the old runlist
  1393. argument, and invokes attrib.c::merge_run_lists() to merge the old
  1394. and the new runlists.
  1395. - Removed attrib.c::find_first_attr().
  1396. - Implemented loading of attribute list and complete runlist for $MFT.
  1397. This means we now cope with $MFT being spread across several mft
  1398. records.
  1399. - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
  1400. - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
  1401. - Make ntfs_volume be allocated via kmalloc() instead of using a slab
  1402. cache. There are too little ntfs_volume structures at any one time
  1403. to justify a private slab cache.
  1404. - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
  1405. Use KM_BIO_IRQ on advice from IRC/kernel...
  1406. - Use ntfs_map_page() in map_mft_record() and create ->readpage method
  1407. for reading $MFT (ntfs_mft_readpage). In the process create dedicated
  1408. address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
  1409. removed the now superfluous exports from the kernel core patch.
  1410. - Fix a bug where kfree() was used insted of ntfs_free().
  1411. - Change map_mft_record() to take ntfs_inode as argument instead of
  1412. vfs inode. Dito for unmap_mft_record(). Adapt all callers.
  1413. - Add pointer to ntfs_volume to ntfs_inode.
  1414. - Add mft record number and sequence number to ntfs_inode. Stop using
  1415. i_ino and i_generation for in-driver purposes.
  1416. - Implement attrib.c::merge_run_lists(). (Richard Russon)
  1417. - Remove use of proper inodes by extent inodes. Move i_ino and
  1418. i_generation to ntfs_inode to do this. Apply simplifications that
  1419. result and remove iget_no_wait(), etc.
  1420. - Pass ntfs_inode everywhere in the driver (used to be struct inode).
  1421. - Add reference counting in ntfs_inode for the ntfs inode itself and
  1422. for the mapped mft record.
  1423. - Extend mft record mapping so we can (un)map extent mft records (new
  1424. functions (un)map_extent_mft_record), and so mappings are reference
  1425. counted and don't have to happen twice if already mapped - just ref
  1426. count increases.
  1427. - Add -o iocharset as alias to -o nls for backwards compatibility.
  1428. - The latest core patch is now tiny. In fact just a single additional
  1429. export is necessary over the base kernel.
  1430. tng-0.0.3 - Cleanups, enhancements, bug fixes.
  1431. - Work on attrib.c::decompress_mapping_pairs() to detect base extents
  1432. and setup the runlist appropriately using knowledge provided by the
  1433. sizes in the base attribute record.
  1434. - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
  1435. any more.
  1436. - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
  1437. page or use vmalloc depending on the amount of memory requested.
  1438. - Cleanup error output. The __FUNCTION__ "(): " is now added
  1439. automatically. Introduced a new header file debug.h to support this
  1440. and also moved ntfs_debug() function into it.
  1441. - Make reading of compressed files more intelligent and especially get
  1442. rid of the vmalloc_nofs() from readpage(). This now uses per CPU
  1443. buffers (allocated at first mount with cluster size <= 4kiB and
  1444. deallocated on last umount with cluster size <= 4kiB), and
  1445. asynchronous io for the compressed data using a list of buffer heads.
  1446. Er, we use synchronous io as async io only works on whole pages
  1447. covered by buffers and not on individual buffer heads...
  1448. - Bug fix for reading compressed files with sparse compression blocks.
  1449. tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
  1450. - Fixed handling of directories when cluster size exceeds index block
  1451. size.
  1452. - Hide DOS only name space directory entries from readdir() but allow
  1453. them in lookup(). This should fix the problem that Linux doesn't
  1454. support directory hard links, while still allowing access to entries
  1455. via their short file name. This also has the benefit of mimicking
  1456. what Windows users are used to, so it is the ideal solution.
  1457. - Implemented sync_page everywhere so no more hangs in D state when
  1458. waiting for a page.
  1459. - Stop using bforget() in favour of brelse().
  1460. - Stop locking buffers unnecessarily.
  1461. - Implemented compressed files (inode->mapping contains uncompressed
  1462. data, raw compressed data is currently bread() into a vmalloc()ed
  1463. memory buffer).
  1464. - Enable compressed directories. (Their index root is marked compressed
  1465. to indicate that new files in that directory should be created
  1466. compressed.)
  1467. - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
  1468. functions. (Thanks to Will Dyson for pointing this out.)
  1469. - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
  1470. ntfs_sb_info) out of the common inode and super_block structures and
  1471. started using the generic_ip and generic_sbp pointers instead. This
  1472. makes ntfs entirely private with respect to the kernel tree.
  1473. - Detect compiler version and abort with error message if gcc less than
  1474. 2.96 is used.
  1475. - Fix bug in name comparison function in unistr.c.
  1476. - Implement attribute lists part 1, the infrastructure: search contexts
  1477. and operations, find_external_attr(), lookup_attr()) and make the
  1478. code use the infrastructure.
  1479. - Fix stupid buffer overflow bug that became apparent on larger run
  1480. list containing attributes.
  1481. - Fix bugs in readdir() that became apparent on larger directories.
  1482. The driver is now really useful and survives the test
  1483. find . -type f -exec md5sum "{}" \;
  1484. without any error messages on a over 1GiB sized partition with >16k
  1485. files on it, including compressed files and directories and many files
  1486. and directories with attribute lists.
  1487. tng-0.0.1 - The first useful version.
  1488. - Added ntfs_lookup().
  1489. - Added default upcase generation and handling.
  1490. - Added compile options to be shown on module init.
  1491. - Many bug fixes that were "hidden" before.
  1492. - Update to latest kernel.
  1493. - Added ntfs_readdir().
  1494. - Added file operations for mmap(), read(), open() and llseek(). We just
  1495. use the generic ones. The whole point of going through implementing
  1496. readpage() methods and where possible get_block() call backs is that
  1497. this allows us to make use of the generic high level methods provided
  1498. by the kernel.
  1499. The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
  1500. though and it doesn't implement accesssing compressed files yet. Also,
  1501. accessing files with attribute list attributes is not implemented yet
  1502. either. But for small or simple filesystems it should work and allow
  1503. you to list directories, use stat on directory entries and the file
  1504. system, open, read, mmap and llseek around in files. A big mile stone
  1505. has been reached!
  1506. tng-0.0.0 - Initial version tag.
  1507. Initial driver implementation. The driver can mount and umount simple
  1508. NTFS filesystems (i.e. ones without attribute lists in the system
  1509. files). If the mount fails there might be problems in the error handling
  1510. code paths, so be warned. Otherwise it seems to be loading the system
  1511. files nicely and the mft record read mapping/unmapping seems to be
  1512. working nicely, too. Proof of inode metadata in the page cache and non-
  1513. resident file unnamed stream data in the page cache concepts is thus
  1514. complete.