blob: c5bbedff5ae5fc534c15d1df0bacafd0d051d8b9 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001ToDo/Notes:
2 - Find and fix bugs.
Linus Torvalds1da177e2005-04-16 15:20:36 -07003 - In between ntfs_prepare/commit_write, need exclusion between
Anton Altaparmakov66129f82004-11-11 12:34:00 +00004 simultaneous file extensions. This is given to us by holding i_sem
5 on the inode. The only places in the kernel when a file is resized
6 are prepare/commit write and truncate for both of which i_sem is
7 held. Just have to be careful in readpage/writepage and all other
8 helpers not running under i_sem that we play nice...
9 Also need to be careful with initialized_size extention in
Linus Torvalds1da177e2005-04-16 15:20:36 -070010 ntfs_prepare_write. Basically, just be _very_ careful in this code...
Anton Altaparmakov66129f82004-11-11 12:34:00 +000011 UPDATE: The only things that need to be checked are read/writepage
12 which do not hold i_sem. Note writepage cannot change i_size but it
13 needs to cope with a concurrent i_size change, just like readpage.
14 Also both need to cope with concurrent changes to the other sizes,
15 i.e. initialized/allocated/compressed size, as well.
Linus Torvalds1da177e2005-04-16 15:20:36 -070016 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
17 leave the volume dirty on umount if the final iput(vol->mft_ino)
18 causes a write of any mirrored mft records due to the mft mirror
19 inode having been discarded already. Whether this can actually ever
20 happen is unclear however so it is worth waiting until someone hits
21 the problem.
22 - Enable the code for setting the NT4 compatibility flag when we start
23 making NTFS 1.2 specific modifications.
24
Anton Altaparmakove7a10332005-09-08 16:12:28 +0100252.1.24-WIP
26
27 - Support journals ($LogFile) which have been modified by chkdsk. This
28 means users can boot into Windows after we marked the volume dirty.
29 The Windows boot will run chkdsk and then reboot. The user can then
30 immediately boot into Linux rather than having to do a full Windows
31 boot first before rebooting into Linux and we will recognize such a
32 journal and empty it as it is clean by definition.
33 - Support journals ($LogFile) with only one restart page as well as
34 journals with two different restart pages. We sanity check both and
35 either use the only sane one or the more recent one of the two in the
36 case that both are valid.
Anton Altaparmakov06d0e3c2005-09-08 16:28:25 +010037 - Modify fs/ntfs/malloc.h::ntfs_malloc_nofs() to do the kmalloc() based
38 allocations with __GFP_HIGHMEM, analogous to how the vmalloc() based
39 allocations are done.
40 - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
41 ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
42 hence cannot fail.
Anton Altaparmakov9529d462005-09-08 16:33:12 +010043 - Use ntfs_malloc_nofs_nofail() in the two critical regions in
44 fs/ntfs/runlist.c::ntfs_runlists_merge(). This means we no longer
45 need to panic() if the allocation fails as it now cannot fail.
Anton Altaparmakove7a10332005-09-08 16:12:28 +010046
Anton Altaparmakovaf859a42005-06-25 21:07:27 +0100472.1.23 - Implement extension of resident files and make writing safe as well as
48 many bug fixes, cleanups, and enhancements...
Linus Torvalds1da177e2005-04-16 15:20:36 -070049
50 - Add printk rate limiting for ntfs_warning() and ntfs_error() when
51 compiled without debug. This avoids a possible denial of service
52 attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
53 out.
Randy Dunlap89075472005-03-03 11:19:53 +000054 - Fix compilation warnings on ia64. (Randy Dunlap)
Anton Altaparmakovaf859a42005-06-25 21:07:27 +010055 - Use i_size_{read,write}() instead of reading i_size by hand and cache
56 the value where apropriate.
Anton Altaparmakov36763672004-11-18 13:46:45 +000057 - Add size_lock to the ntfs_inode structure. This is an rw spinlock
58 and it locks against access to the inode sizes. Note, ->size_lock
59 is also accessed from irq context so you must use the _irqsave and
Anton Altaparmakovaf859a42005-06-25 21:07:27 +010060 _irqrestore lock and unlock functions, respectively. Protect all
61 accesses to allocated_size, initialized_size, and compressed_size.
Anton Altaparmakov218357f2004-11-18 20:34:59 +000062 - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
Anton Altaparmakov946929d2005-01-13 15:26:29 +000063 - Implement extension of resident files in the regular file write code
64 paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()). At present
65 this only works until the data attribute becomes too big for the mft
66 record after which we abort the write returning -EOPNOTSUPP from
67 ntfs_prepare_write().
Anton Altaparmakovc002f422005-02-03 12:02:56 +000068 - Add disable_sparse mount option together with a per volume sparse
69 enable bit which is set appropriately and a per inode sparse disable
70 bit which is preset on some system file inodes as appropriate.
71 - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
Anton Altaparmakov1a0df152005-02-03 12:04:36 +000072 - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
73 the creation of the unmapped runlist element for the base attribute
74 extent.
Anton Altaparmakovb6ad6c52005-02-15 10:08:43 +000075 - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
76 helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
77 This allows us to map runlist fragments with the runlist lock already
78 held without having to drop and reacquire it around the call. Adapt
79 all callers.
80 - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
81 runlist. This allows us to find runlist elements with the runlist
82 lock already held without having to drop and reacquire it around the
83 call. Adapt all callers.
Anton Altaparmakov37e4c132005-02-18 10:03:13 +000084 - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
Anton Altaparmakovd8ec7852005-02-18 09:23:39 +000085 warning in the do_div() call on sparc32. Thanks to Meelis Roos for
86 the report and analysis of the warning.
Anton Altaparmakov37e4c132005-02-18 10:03:13 +000087 - Fix a nasty runlist merge bug when merging two holes.
Anton Altaparmakov5ae9fcf2005-03-02 17:03:24 +000088 - Set the ntfs_inode->allocated_size to the real allocated size in the
89 mft record for resident attributes (fs/ntfs/inode.c).
90 - Small readability cleanup to use "a" instead of "ctx->attr"
91 everywhere (fs/ntfs/inode.c).
Anton Altaparmakov41382682005-03-03 13:44:15 +000092 - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
93 definition of ntfs_export_ops from fs/ntfs/super.c to namei.c. Also,
94 declare ntfs_export_ops in fs/ntfs/ntfs.h.
Anton Altaparmakov9451f852005-03-03 14:43:43 +000095 - Correct sparse file handling. The compressed values need to be
96 checked and set in the ntfs inode as done for compressed files and
97 the compressed size needs to be used for vfs inode->i_blocks instead
98 of the allocated size, again, as done for compressed files.
Anton Altaparmakov271849a2005-03-07 21:36:18 +000099 - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
100 non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
101 - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
102 write code.
Anton Altaparmakovc0c1cc02005-03-07 21:43:38 +0000103 - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
104 dropping the read lock and taking the write lock we were not checking
105 whether someone else did not already do the work we wanted to do.
106 - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
107 ntfs_attr_find_vcn_nolock() and update all callers.
Anton Altaparmakov2bfb4ff2005-03-09 15:15:06 +0000108 - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
Anton Altaparmakov43b01fd2005-03-09 15:18:43 +0000109 - Fix sign of various error return values to be negative in
110 fs/ntfs/lcnalloc.c.
Anton Altaparmakov905685f2005-03-10 11:06:19 +0000111 - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
112 handle the case where an attribute is converted from resident to
113 non-resident by a concurrent file write.
Jesper Juhl251c8422005-04-04 14:59:56 +0100114 - Remove checks for NULL before calling kfree() since kfree() does the
115 checking itself. (Jesper Juhl)
Anton Altaparmakovb0d23742005-04-04 16:20:14 +0100116 - Some utilities modify the boot sector but do not update the checksum.
117 Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
118 only emit a warning when the checksum is incorrect rather than
119 refusing the mount. Thanks to Bernd Casimir for pointing this
120 problem out.
Anton Altaparmakovbb3cf332005-04-06 13:34:31 +0100121 - Update attribute definition handling.
Anton Altaparmakov7fafb8b2005-04-06 16:09:21 +0100122 - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
123 - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
Anton Altaparmakovd53ee322005-04-06 16:11:20 +0100124 - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
125 better code generation and one less sparse warning in fs/ntfs/aops.c.
Pekka Enberg2fb21db2005-05-25 21:15:34 +0300126 - Remove spurious void pointer casts from fs/ntfs/. (Pekka Enberg)
Anton Altaparmakov442d2072005-05-27 16:42:56 +0100127 - Use C99 style structure initialization after memory allocation where
128 possible (fs/ntfs/{attrib.c,index.c,super.c}). Thanks to Al Viro and
129 Pekka Enberg.
Anton Altaparmakov3f2faef2005-06-25 15:28:56 +0100130 - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
131 is active on the volume and we are mounting read-write or remounting
132 from read-only to read-write.
Anton Altaparmakov9f993fe2005-06-25 16:15:36 +0100133 - Fix a bug in address space operations error recovery code paths where
134 if the runlist was not mapped at all and a mapping error occured we
135 would leave the runlist locked on exit to the function so that the
136 next access to the same file would try to take the lock and deadlock.
Anton Altaparmakovca8fd7a2005-06-25 16:31:27 +0100137 - Detect the case when Windows has been suspended to disk on the volume
138 to be mounted and if this is the case do not allow (re)mounting
139 read-write. This is done by parsing hiberfil.sys if present.
Anton Altaparmakov3bd1f4a2005-06-25 16:51:58 +0100140 - Fix several occurences of a bug where we would perform 'var & ~const'
141 with a 64-bit variable and a int, i.e. 32-bit, constant. This causes
142 the higher order 32-bits of the 64-bit variable to be zeroed. To fix
143 this cast the 'const' to the same 64-bit type as 'var'.
Anton Altaparmakov1d58b272005-06-25 17:04:55 +0100144 - Change the runlist terminator of the newly allocated cluster(s) to
145 LCN_ENOENT in ntfs_attr_make_non_resident(). Otherwise the runlist
146 code gets confused.
Anton Altaparmakovfa3be922005-06-25 17:15:36 +0100147 - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
148 and ntfs_mapping_pairs_build() to allow the runlist encoding to be
149 partial which is desirable when filling holes in sparse attributes.
150 Update all callers.
Anton Altaparmakov4757d7df2005-06-25 17:24:08 +0100151 - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
152 if the requested vcn is inside it. Otherwise we get into problems
153 when we try to map an out of bounds vcn because we then try to map
154 the already mapped runlist fragment which causes
155 ntfs_mapping_pairs_decompress() to fail and return error. Update
156 ntfs_attr_find_vcn_nolock() accordingly.
Anton Altaparmakovba6d2372005-06-26 22:12:02 +0100157 - Fix a nasty deadlock that appeared in recent kernels.
158 The situation: VFS inode X on a mounted ntfs volume is dirty. For
159 same inode X, the ntfs_inode is dirty and thus corresponding on-disk
160 inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
161 to the table of inodes, i.e. $MFT, inode 0.
162 What happens:
163 Process 1: sys_sync()/umount()/whatever... calls
164 __sync_single_inode() for $MFT -> do_writepages() -> write_page for
165 the dirty page containing the on-disk inode X, the page is now locked
166 -> ntfs_write_mst_block() which clears PageUptodate() on the page to
167 prevent anyone else getting hold of it whilst it does the write out.
168 This is necessary as the on-disk inode needs "fixups" applied before
169 the write to disk which are removed again after the write and
170 PageUptodate is then set again. It then analyses the page looking
171 for dirty on-disk inodes and when it finds one it calls
172 ntfs_may_write_mft_record() to see if it is safe to write this
173 on-disk inode. This then calls ilookup5() to check if the
174 corresponding VFS inode is in icache(). This in turn calls ifind()
175 which waits on the inode lock via wait_on_inode whilst holding the
176 global inode_lock.
177 Process 2: pdflush results in a call to __sync_single_inode for the
178 same VFS inode X on the ntfs volume. This locks the inode (I_LOCK)
179 then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
180 read_cache_page() for the page (in page cache of table of inodes
181 $MFT, inode 0) containing the on-disk inode. This page has
182 PageUptodate() clear because of Process 1 (see above) so
183 read_cache_page() blocks when it tries to take the page lock for the
184 page so it can call ntfs_read_page().
185 Thus Process 1 is holding the page lock on the page containing the
186 on-disk inode X and it is waiting on the inode X to be unlocked in
187 ifind() so it can write the page out and then unlock the page.
188 And Process 2 is holding the inode lock on inode X and is waiting for
189 the page to be unlocked so it can call ntfs_readpage() or discover
190 that Process 1 set PageUptodate() again and use the page.
191 Thus we have a deadlock due to ifind() waiting on the inode lock.
192 The solution: The fix is to use the newly introduced
193 ilookup5_nowait() which does not wait on the inode's lock and hence
194 avoids the deadlock. This is safe as we do not care about the VFS
195 inode and only use the fact that it is in the VFS inode cache and the
196 fact that the vfs and ntfs inodes are one struct in memory to find
197 the ntfs inode in memory if present. Also, the ntfs inode has its
198 own locking so it does not matter if the vfs inode is locked.
Anton Altaparmakove74589a2005-08-16 16:38:28 +0100199 - Fix bug in mft record writing where we forgot to set the device in
Anton Altaparmakov481d0372005-08-16 19:42:56 +0100200 the buffers when mapping them after the VM had discarded them.
Anton Altaparmakove74589a2005-08-16 16:38:28 +0100201 Thanks to Martin MOKREJÅ  for the bug report.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700202
2032.1.22 - Many bug and race fixes and error handling improvements.
204
205 - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
206 - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
207 instead of void and provide a helper ntfs_truncate_vfs() for the
208 vfs ->truncate method.
209 - Add a new ntfs inode flag NInoTruncateFailed() and modify
210 fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
211 - Fix min_size and max_size definitions in ATTR_DEF structure in
212 fs/ntfs/layout.h to be signed.
213 - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
214 ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
215 ntfs_attr_can_be_resident(), which in turn use the new private helper
216 ntfs_attr_find_in_attrdef().
217 - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
218 mapping->private_lock around the dirtying of the buffer heads
219 analagous to the way it is done in __set_page_dirty_buffers().
220 - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
221 mount time as this cannot work with the current implementation.
222 - Check for location of attribute name and improve error handling in
223 general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
224 - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
225 i_size, i.e. race with truncate, invalidate the buffers on the page
226 so that they become freeable and hence the page does not leak.
227 - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
228 Bunk)
229 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
230 a NULL pointer dereference in the error code path when a corrupt
231 attribute was found. (Thanks to Domen Puncer for the bug report.)
232 - Add MODULE_VERSION() to fs/ntfs/super.c.
233 - Make several functions and variables static. (Adrian Bunk)
234 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
235 buffers for the page if they are not present and then marks the
236 buffers belonging to the ntfs record dirty. This causes the buffers
237 to become busy and hence they are safe from removal until the page
238 has been written out.
239 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
240 error handling code path that resulted in a BUG() due to trying to
241 unmap an extent mft record when the mapping of it had failed and it
242 thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
243 - Drop the runlist lock after the vcn has been read in
244 fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
245 - Rewrite handling of multi sector transfer errors. We now do not set
246 PageError() when such errors are detected in the async i/o handler
247 fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
248 protected attributes now check the magic of each ntfs record as they
249 use it and act appropriately. This has the effect of making errors
250 granular per ntfs record rather than per page which solves the case
251 where we cannot access any of the ntfs records in a page when a
252 single one of them had an mst error. (Thanks to Ken MacFerrin for
253 the bug report.)
254 - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
255 where we failed to release i_sem on the $Quota/$Q attribute inode.
256 - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
257 - Add mapping of unmapped buffers to all remaining code paths, i.e.
258 fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
259 and write_mft_record_nolock(). From now on we require that the
260 complete runlist for the mft mirror is always mapped into memory.
261 - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
262 - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
263 - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
264 resident attribute will be smaller than a page which makes the code
265 simpler. Also make the code more tolerant to concurrent ->truncate.
266
2672.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
268
269 - Implement extent mft record deallocation
270 fs/ntfs/mft.c::ntfs_extent_mft_record_free().
271 - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
272 - Add vol->mft_data_pos and initialize it at mount time.
273 - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
274 ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
275 ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
276 ntfs_runlists_merge() and adapt all callers.
277 - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
278 ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
279 and ntfs_mapping_pairs_build(), adapted from libntfs.
280 - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
281 static and add a declaration for it to lcnalloc.h.
282 - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
283 inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
284 cluster bitmap lock for the duration of the call.
285 - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
286 - Implement the equivalent of memset() for an ntfs attribute in
287 fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
288 fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
289 - Remove unnecessary casts from LCN_* constants.
290 - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
291 - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
292 change MFT_RECORD to contain the NTFS 3.1+ specific fields.
293 - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
294 marks all buffers belonging to an ntfs record dirty, followed by
295 marking the page the ntfs record is in dirty and also marking the vfs
296 inode containing the ntfs record dirty (I_DIRTY_PAGES).
297 - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
298 new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
299 longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
300 - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
301 include errors.
302 - Move the typedefs for runlist_element and runlist from types.h to
303 runlist.h and fix resulting include errors.
304 - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
305 - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
306 mark_ntfs_record_dirty() which also changes the behaviour in that we
307 now set the buffers belonging to the mft record dirty as well as the
308 page itself.
309 - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
310 to cope with the fact that there now are dirty buffers in mft pages.
311 - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
312 mark_ntfs_record_dirty() and thus to set the buffers belonging to the
313 mft record dirty as well as the page itself.
314 - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
315 slightly modified by me)
316 - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
317 the mft record is already locked and otherwise behaves the same way
318 as fs/ntfs/mft.c::map_mft_record().
319 - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
320 writes the mft record if the buffers belonging to it are dirty.
321 Otherwise we assume that it was written out by other means already.
322 - Attempting to write outside initialized size is _not_ a bug so remove
323 the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
324 fact required to write outside initialized size when preparing to
325 extend the initialized size.
326 - Map the page instead of using page_address() before writing to it in
327 fs/ntfs/aops.c::ntfs_mft_writepage().
328 - Provide exclusion between opening an inode / mapping an mft record
329 and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
330 by setting the page not uptodate throughout ntfs_mft_writepage().
331 - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
332 to ensure noone can see the page whilst the mst fixups are applied.
333 - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
334 checks if an mft record may be written out safely obtaining any
335 necessary locks in the process. This is used by
336 fs/ntfs/aops.c::ntfs_write_mst_block().
337 - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
338 writing mft records and improve its error handling in the process.
339 Now if any of the records in the page fail to be written out, all
340 other records will be written out instead of aborting completely.
341 - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
342 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
343 ntfs_mst_aops for all inodes which are NInoMstProtected() and
344 ntfs_aops for all other inodes.
345 - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
346 ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
347 no longer require an ntfs inode to be present. Update all callers.
348 - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
349 - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
350 to ensure noone can see the page whilst the mst fixups are applied.
351 - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
352 fs/ntfs/mft.c::try_map_mft_record().
353 - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
354 with the ntfs inode which contains the page rather than the ntfs
355 inode the mft record of which is in the page.
356 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
357 index inode bitmap inode release code from there to
358 fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
359 Hellwig for spotting this.)
360 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
361 inode semaphore around the code that sets ni->itype.index.bmp_ino to
362 NULL and reorganize the code to optimize it a bit. (Thanks to
363 Christoph Hellwig for spotting this.)
364 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
365 ntfs inode as a parameter as this is confusing and misleading and the
366 needed ntfs inode is available via NTFS_I(page->mapping->host).
367 Adapt all callers to this change.
368 - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
369 fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
370 of the first buffer in a record and to take this as the ntfs record
371 dirty state. We cannot look at the dirty state for subsequent
372 buffers because we might be racing with
373 fs/ntfs/aops.c::mark_ntfs_record_dirty().
374 - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
375 inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
376 add a declaration for it to inode.h. Fix some compilation issues
377 that resulted due to #includes and header file interdependencies.
378 - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
379 - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
380 - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
381 record sequence number if it is specified (i.e. not zero).
382 - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
383 functions used by it.
384 - Update Documentation/filesystems/ntfs.txt with instructions on how to
385 use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
386 the linear raid problem with the Software RAID / MD driver when one
387 or more of the devices has an odd number of sectors.
388
3892.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
390
391 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
392 where we did not clear ctx->al_entry but it was still set due to
393 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
394 particular.
395 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
396 where we forgot to unmap the extent mft record when we had finished
397 enumerating an attribute which caused a bug check to trigger when the
398 VFS calls ->clear_inode.
399
4002.1.19 - Many cleanups, improvements, and a minor bug fix.
401
402 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
403 change the uid, gid, and mode of an inode as we do not support NTFS
404 ACLs yet.
405 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
406 kernel.
407 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
408 and ntfs_filldir() as per suggestion from Al Viro.
409 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
410 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
411 inode size has changed and to only output an error if so.
412 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
413 - Add le{16,32,64} as well as sle{16,32,64} data types to
414 fs/ntfs/types.h.
415 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
416 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
417 respectively, to fs/ntfs/types.h.
418 - Update endianness conversion macros in fs/ntfs/endian.h to use the
419 new types as appropriate.
420 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
421 and index.c.
422 - Add leMFT_REF data type to fs/ntfs/layout.h.
423 - Update all NTFS header files with the new little endian data types.
424 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
425 - Do proper type casting when using ntfs_is_*_recordp() in
426 fs/ntfs/logfile.c, mft.c, and super.c.
427 - Fix all the sparse bitwise warnings. Had to change all the typedef
428 enums storing little endian values to simple enums plus a typedef for
429 the datatype to make sparse happy.
430 - Fix a bug found by the new sparse bitwise warnings where the default
431 upcase table was defined as a pointer to wchar_t rather than ntfschar
432 in fs/ntfs/ntfs.h and super.c.
433 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
434
4352.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
436
437 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
438 the calculation of total/free inodes as used by statfs().
439 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
440 because the code itself is using the ntfs_lock semaphore which
441 provides safe locking. (Ingo Molnar)
442 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
443 could occur in the future for when we start closing/freeing extent
444 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
445 we free it.
446 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
447 find_external_attr() to ntfs_external_attr_find() to cleanup the
448 namespace a bit and to be more consistent with libntfs.
449 - Rename {{re,}init,get,put}_attr_search_ctx() to
450 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
451 attr_search_context to ntfs_attr_search_ctx.
452 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
453 for the attribute list attribute itself.
454 - Fix endianness bug in ntfs_external_attr_find().
455 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
456 if the attribute is not found, and -EIO on real error. In the case
457 of -ENOENT, the search context is updated to describe the attribute
458 before which the attribute being searched for would need to be
459 inserted if such an action were to be desired and in the case of
460 ntfs_external_attr_find() the search context is also updated to
461 indicate the attribute list entry before which the attribute list
462 entry of the attribute being searched for would need to be inserted
463 if such an action were to be desired. Also make ntfs_find_attr()
464 static and remove its prototype from attrib.h as it is not used
465 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
466 callers of ntfs_{external,}attr_{find,lookup}() for the new return
467 values.
468 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
469
4702.1.17 - Fix bugs in mount time error code paths and other updates.
471
472 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
473 includes functions to set/clear a single bit or a run of bits.
474 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
475 runlist element containing a particular vcn. It also takes care of
476 mapping any needed runlist fragments.
477 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
478 - Load attribute definition table from $AttrDef at mount time.
479 - Fix bugs in mount time error code paths involving (de)allocation of
480 the default and volume upcase tables.
481 - Remove ntfs_nr_mounts as it is no longer used.
482
4832.1.16 - Implement access time updates, file sync, async io, and read/writev.
484
485 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
486 This is done by setting the appropriate file operations pointers to
487 the generic helper functions provided by mm/filemap.c.
488 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
489 and directories (fs/ntfs/dir.c).
490 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
491 Note, except for the root directory and any other system files opened
492 by the user, the system files will not have their access times
493 updated as they are only accessed at the inode level an hence the
494 file level functions which cause the times to be updated are never
495 invoked.
496
4972.1.15 - Invalidate quotas when (re)mounting read-write.
498
499 - Add new element itype.index.collation_rule to the ntfs inode
500 structure and set it appropriately in ntfs_read_locked_inode().
501 - Implement a new inode type "index" to allow efficient access to the
502 indices found in various system files and adapt inode handling
503 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
504 attribute inode (NInoAttr() is true) with an attribute type of
505 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
506 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
507 there would be no way to distinguish between normal attribute inodes
508 and index inodes. The function to obtain an index inode is
509 ntfs_index_iget() and it uses the helper function
510 ntfs_read_locked_index_inode(). Note, we do not overload
511 ntfs_attr_iget() as indices consist of multiple attributes so using
512 ntfs_attr_iget() to obtain an index inode would be confusing.
513 - Ensure that there is no overflow when doing page->index <<
514 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
515 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
516 and ntfs_read_block().
517 - Use case sensitive attribute lookups instead of case insensitive ones.
518 - Lock all page cache pages belonging to mst protected attributes while
519 accessing them to ensure we never see corrupt data while the page is
520 under writeout.
521 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
522 We have ntfs_is_collation_rule_supported() to check if the collation
523 rule you want to use is supported and ntfs_collation() which actually
524 collates two data items. We currently only support COLLATION_BINARY
525 and COLLATION_NTOFS_ULONG but support for other collation rules will
526 be added as the need arises.
527 - Add a new type, ntfs_index_context, to allow retrieval of an index
528 entry using the corresponding index key. To get an index context,
529 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
530 This also adds a new slab cache for the index contexts. To lookup a
531 key in an index inode, use ntfs_index_lookup(). After modifying an
532 index entry, call ntfs_index_entry_flush_dcache_page() followed by
533 ntfs_index_entry_mark_dirty() to ensure the changes are written out
534 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
535 an index entry is in the index allocation attribute rather than the
536 index root attribute it will not be written out (you will get a
537 warning message about discarded changes instead).
538 - Load the quota file ($Quota) and check if quota tracking is enabled
539 and if so, mark the quotas out of date. This causes windows to
540 rescan the volume on boot and update all quota entries.
541 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
542 It is simply set to __set_page_dirty_nobuffers() to make sure that
543 running set_page_dirty() on a page containing mft/ntfs records will
544 not affect the dirty state of the page buffers.
545 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
546 buffers that are inside the ntfs record in the page dirty after which
547 it sets the page dirty. This allows ->writepage to only write the
548 dirty index records rather than having to write all the records in
549 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
550 use this rather than __set_page_dirty_nobuffers().
551 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
552 writing of page cache pages belonging to mst protected attributes
553 like the index allocation attribute in directory indices and other
554 indices like $Quota/$Q, etc. This means that the quota is now marked
555 out of date on all volumes rather than only on ones where the quota
556 defaults entry is in the index root attribute of the $Quota/$Q index.
557
5582.1.14 - Fix an NFSd caused deadlock reported by several users.
559
560 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
561 to a buffer so that we can put the search context and unmap the mft
562 record before calling the filldir() callback. We need to do this
563 because of NFSd which calls ->lookup() from its filldir callback()
564 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
565 of the directory and since ntfs_readdir() has got it mapped already
566 ntfs_lookup() deadlocks.
567
5682.1.13 - Enable overwriting of resident files and housekeeping of system files.
569
570 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
571 keeping the mft mirror in sync with the mft when mirrored mft records
572 are written. The functions are write_mft_record{,_nolock}(). The
573 implementation is quite rudimentary for now with lots of things not
574 implemented yet but I am not sure any of them can actually occur so
575 I will wait for people to hit each one and only then implement it.
576 - Commit open system inodes at umount time. This should make it
577 virtually impossible for sync_mft_mirror_umount() to ever be needed.
578 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
579 ntfs super operations. This gives us inode writing via the VFS inode
580 dirty code paths. Note: Access time updates are not implemented yet.
581 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
582 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
583 finally enabling resident file overwrite! (-8 This also includes a
584 placeholder for ->writepage (ntfs_mft_writepage()), which for now
585 just redirties the page and returns. Also, at umount time, we for
586 now throw away all mft data page cache pages after the last call to
587 ntfs_commit_inode() in the hope that all inodes will have been
588 written out by then and hence no dirty (meta)data will be lost. We
589 also check for this case and emit an error message telling the user
590 to run chkdsk.
591 - Use set_page_writeback() and end_page_writeback() in the resident
592 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
593 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
594 page is clean.
595 - Implement ntfs_mft_writepage() so it now checks if any of the mft
596 records in the page are dirty and if so redirties the page and
597 returns. Otherwise it just returns (after doing set_page_writeback(),
598 unlock_page(), end_page_writeback() or the radix-tree tag
599 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
600 alowing the VM to do with the page as it pleases. Also, at umount
601 time, now only throw away dirty mft (meta)data pages if dirty inodes
602 are present and ask the user to email us if they see this happening.
603 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
604 information flags (fs/ntfs/super.c).
605 - Mark the volume dirty when (re)mounting read-write and mark it clean
606 when unmounting or remounting read-only. If any volume errors are
607 found, the volume is left marked dirty to force chkdsk to run.
608 - Add code to set the NT4 compatibility flag when (re)mounting
609 read-write for newer NTFS versions but leave it commented out for now
610 since we do not make any modifications that are NTFS 1.2 specific yet
611 and since setting this flag breaks Captive-NTFS which is not nice.
612 This code must be enabled once we start writing NTFS 1.2 specific
613 changes otherwise Windows NTFS driver might crash / cause corruption.
614
6152.1.12 - Fix the second fix to the decompression engine and some cleanups.
616
617 - Add a new address space operations struct, ntfs_mst_aops, for mst
618 protected attributes. This is because the default ntfs_aops do not
619 make sense with mst protected data and were they to write anything to
620 such an attribute they would cause data corruption so we provide
621 ntfs_mst_aops which does not have any write related operations set.
622 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
623 includes an adapted ntfs_commit_inode() and an implementation of
624 ntfs_write_inode() which for now just cleans dirty inodes without
625 writing them (it does emit a warning that this is happening).
626 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
627 entry) as it was only fixing a theoretical bug but at the same time
628 it badly broke the handling of sparse and uncompressed compression
629 blocks.
630
6312.1.11 - Driver internal cleanups.
632
633 - Only build logfile.o if building the driver with read-write support.
634 - Really final white space cleanups.
635 - Use generic_ffs() instead of ffs() in logfile.c which allows the
636 log_page_size variable to be optimized by gcc into a constant.
637 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
638 char as defined by POSIX and as found on some systems.
639
6402.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
641
642 - Finish off the white space cleanups (remove trailing spaces, etc).
643 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
644 the kludges around the first iget(). Instead of (re)setting ->s_op
645 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
646 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
647 kills the need for second super_operations and allows to return error
648 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
649 tricks. (Al Viro)
650 - Force read-only (re)mounting if any of the following bits are set in
651 the volume information flags:
652 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
653 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
654 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
655 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
656 above bits set so the test is made easy.
657
6582.1.9 - Fix two bugs in decompression engine.
659
660 - Fix a bug where we would not always detect that we have reached the
661 end of a compression block because we were ending at minus one byte
662 which is effectively the same as being at the end. The fix is to
663 check whether the uncompressed buffer has been fully filled and if so
664 we assume we have reached the end of the compression block. A big
665 thank you to Marcin Gibuła for the bug report, the assistance in
666 tracking down the bug and testing the fix.
667 - Fix a possible bug where when a compressed read is truncated to the
668 end of the file, the offset inside the last page was not truncated.
669
6702.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
671
672 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
673 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
674 utc2ntfs() to work with struct timespec instead of time_t on the
675 Linux UTC time side thus preserving the full precision of the NTFS
676 time and only loosing up to 99 nano-seconds in the Linux UTC time.
677 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
678 static inline.
679 - Remove unused ntfs_dirty_inode().
680 - Cleanup super operations declaration in fs/ntfs/super.c.
681 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
682 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
683 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
684 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
685 fs/ntfs/inode.h so they can be used elsewhere.
686 - Determine the mft mirror size as the number of mirrored mft records
687 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
688 - Load the mft mirror at mount time and compare the mft records stored
689 in it to the ones in the mft. Force a read-only mount if the two do
690 not match (fs/ntfs/super.c).
691 - Fix type casting related warnings on 64-bit architectures. Thanks
692 to Meelis Roos for reporting them.
693 - Move %L to %ll as %L is floating point and %ll is integer which is
694 what we want.
695 - Read the journal ($LogFile) and determine if the volume has been
696 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
697 and fs/ntfs/logfile.c). This is a little bit of a crude check in
698 that we only look at the restart areas and not at the actual log
699 records so that there will be a very small number of cases where we
700 think that a volume is dirty when in fact it is clean. This should
701 only affect volumes that have not been shutdown cleanly and did not
702 have any pending, non-check-pointed i/o.
703 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
704 is requested, empty $LogFile by overwriting it with 0xff bytes to
705 ensure that Windows cannot cause data corruption by replaying a stale
706 journal after Linux has written to the volume.
707
7082.1.7 - Enable NFS exporting of mounted NTFS volumes.
709
710 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
711 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
712 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
713 default doesn't allow inode number 0 which is a valid inode on NTFS
714 and even if it did allow that it uses iget() instead of ntfs_iget()
715 which makes it useless for us.
716 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
717 default just returns -EACCES which is not very useful.
718 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
719 and set them up in the super block at mount time (super.c) this
720 allows mounted NTFS volumes to be exported via NFS.
721 - Add missing return -EOPNOTSUPP; in
722 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
723 - Enforce no atime and no dir atime updates at mount/remount time as
724 they are not implemented yet anyway.
725 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
726 after a NULL check. Thanks to Dave Jones for pointing this out.
727
7282.1.6 - Fix minor bug in handling of compressed directories.
729
730 - Fix bug in handling of compressed directories. A compressed
731 directory is not really compressed so when we set the ->i_blocks
732 field of a compressed directory inode we were setting it from the
733 non-existing field ni->itype.compressed.size which gave random
734 results... For directories we now always use ni->allocated_size.
735
7362.1.5 - Fix minor bug in attribute list attribute handling.
737
738 - Fix bug in attribute list handling. Actually it is not as much a bug
739 as too much protection in that we were not allowing attribute lists
740 which waste space on disk while Windows XP clearly allows it and in
741 fact creates such attribute lists so our driver was failing.
742 - Update NTFS documentation ready for 2.6 kernel release.
743
7442.1.4 - Reduce compiler requirements.
745
746 - Remove all uses of unnamed structs and unions in the driver to make
747 old and newer gcc versions happy. Makes it a bit uglier IMO but at
748 least people will stop hassling me about it.
749
7502.1.3 - Important bug fixes in corner cases.
751
752 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
753 clusters. (Philipp Thomas)
754 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
755 multiple of the block_size but not the cluster size. (Szabolcs
756 Szakacsits <szaka@sienet.hu>)
757
7582.1.2 - Important bug fixes aleviating the hangs in statfs.
759
760 - Fix buggy free cluster and free inode determination logic.
761
7622.1.1 - Minor updates.
763
764 - Add handling for initialized_size != data_size in compressed files.
765 - Reduce function local stack usage from 0x3d4 bytes to just noise in
766 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
767 - Remove compiler warnings for newer gcc.
768 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
769 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
770 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
771 kmap_atomic(KM_USER0).
772
7732.1.0 - First steps towards write support: implement file overwrite.
774
775 - Add configuration option for developmental write support with an
776 appropriately scary configuration help text.
777 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
778 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
779 overwriting of existing files on ntfs. Note: Resident files are
780 only written into memory, and not written out to disk at present, so
781 avoid writing to files smaller than about 1kiB.
782 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
783 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
784 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
785 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
786 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
787 This enables write(2) based overwriting of existing files on ntfs.
788 Note: As with mmap(2) based overwriting, resident files are only
789 written into memory, and not written out to disk at present, so avoid
790 writing to files smaller than about 1kiB.
791 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
792 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
793 files with the purpose of intercepting and aborting all i_size
794 changes which we do not support yet. ntfs_truncate() actually only
795 emits a warning message but AFAICS our interception of i_size changes
796 elsewhere means ntfs_truncate() never gets called for i_size changes.
797 It is only called from generic_file_write() when we fail in
798 ntfs_prepare_{,nonresident_}write() in order to discard any
799 instantiated buffers beyond i_size. Thus i_size is not actually
800 changed so our warning message is enough. Unfortunately it is not
801 possible to easily determine if i_size is being changed or not hence
802 we just emit an appropriately worded error message.
803
8042.0.25 - Small bug fixes and cleanups.
805
806 - Unlock the page in an out of memory error code path in
807 fs/ntfs/aops.c::ntfs_read_block().
808 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
809 just unlock the page and return. (This can happen due to ->writepage
810 clearing PageUptodate() during write out of MstProtected()
811 attributes.
812 - Remove leaked write code again.
813
8142.0.24 - Cleanups.
815
816 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
817 inside BUG_ON(). (Adam J. Richter)
818 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
819 calls for improved debugging. (Adam J. Richter)
820 - Add errors flag to the ntfs volume state, accessed via
821 NVol{,Set,Clear}Errors(vol).
822 - Do not allow read-write remounts of read-only volumes with errors.
823 - Clarify comment for ntfs file operation sendfile which was added by
824 Christoph Hellwig a while ago (just using generic_file_sendfile())
825 to say that ntfs ->sendfile is only used for the case where the
826 source data is on the ntfs partition and the destination is
827 somewhere else, i.e. nothing we need to concern ourselves with.
828 - Add generic_file_write() as our ntfs file write operation.
829
8302.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
831
832 - Massive internal locking changes to mft record locking. Fixes lock
833 recursion and replaces the mrec_lock read/write semaphore with a
834 mutex. Also removes the now superfluous mft_count. This fixes several
835 race conditions and deadlocks, especially in the future write code.
836 - Fix ntfs over loopback for compressed files by adding an
837 optimization barrier. (gcc was screwing up otherwise ?)
838 - Miscellaneous cleanups all over the code and a fix or two in error
839 handling code paths.
840 Thanks go to Christoph Hellwig for pointing out the following two:
841 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
842 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
843
8442.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
845
846 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
847 at entry/exit respectively.
848 - Use C99 initializers for structures.
849 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
850
8512.0.21 - Check for, and refuse to work with too large files/directories/volumes.
852
853 - Limit volume size at mount time to 2TiB on architectures where
854 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
855 This is the most we can do without overflowing the 32-bit limit of
856 the block device size imposed on us by sb_bread() and sb_getblk()
857 for the time being.
858 - Limit file/directory size at open() time to 16TiB on architectures
859 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
860 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
861 overflowing the page cache page index.
862
8632.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
864
865 - Move the directory index bitmap to use an attribute inode instead of
866 having special fields for it inside the ntfs inode structure. This
867 means that the index bitmaps now use the page cache for i/o, too,
868 and also as a side effect we get support for non-resident index
869 bitmaps for free.
870 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
871 fix a page leak that manifested itself in some cases.
872 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
873 index bitmap inode on the final iput().
874
8752.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
876
877 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
878 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
879 - Drop the "file" from ntfs_file_read_compressed_block().
880 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
881 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
882 - Update ntfs_end_buffer_async_read() with the improved logic from
883 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
884 further logic improvements to better determine when we set PageError.
885 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
886 check for the buffers being uptodate first in line with the updated
887 fs/buffer.c::block_read_full_page(). This plugs a small race
888 condition.
889
8902.0.18 - Fix race condition in reading of compressed files.
891
892 - There was a narrow window between checking a buffer head for being
893 uptodate and locking it in ntfs_file_read_compressed_block(). We now
894 lock the buffer and then check whether it is uptodate or not.
895
8962.0.17 - Cleanups and optimizations - shrinking the ToDo list.
897
898 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
899 code and update callers, i.e. ntfs_iget(), to pass that error code
900 up instead of just using -EIO.
901 - Modifications to super.c to ensure that both mount and remount
902 cannot set any write related options when the driver is compiled
903 read-only.
904 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
905 cache the current runlist element. This should improve performance
906 when reading very large and/or very fragmented data.
907
9082.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
909
910 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
911 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
912 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
913 - Convert $MFT/$BITMAP access to attribute inode API and remove all
914 remnants of the ugly mftbmp address space and operations hack. This
915 means we finally have only one readpage function as well as only one
916 async io completion handler. Yey! The mft bitmap is now just an
917 attribute inode and is accessed from vol->mftbmp_ino just as if it
918 were a normal file. Fake inodes rule. (-:
919
9202.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
921
922 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
923 remounts to fail when the partition had an entry in /etc/fstab and
924 the entry specified the nls= option.
925 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
926 expand all the helper functions NVolFoo(), NVolSetFoo(), and
927 NVolClearFoo().
928 - Move copyright statement from driver initialisation message to
929 module description (fs/super.c). This makes the initialisation
930 message fit on one line and fits in better with rest of kernel.
931 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
932 attribute inodes, and both for files and directories.
933 - Implement fake attribute inodes allowing all attribute i/o to go via
934 the page cache and to use all the normal vfs/mm functionality:
935 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
936 to fs/ntfs/inode.c.
937 - Add needed cleanup code to ntfs_clear_big_inode().
938 - Merge address space operations for files and directories (aops.c),
939 now just have ntfs_aops:
940 - Rename:
941 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
942 ntfs_attr_read_block() -> ntfs_read_block(),
943 ntfs_file_read_page() -> ntfs_readpage().
944 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
945 attribute inodes, and both for files and directories.
946 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
947
9482.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
949
950 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
951 the locking out of super.c::get_nr_free_mft_records() and taking and
952 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
953 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
954 current userspace ntfs library code. This means that if a merge
955 fails the original runlists are always left unmodified instead of
956 being silently corrupted.
957 - Misc typo fixes.
958
9592.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
960
961 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
962 from ntfs_volume structure.
963 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
964 from ntfs_volume structure.
965 - Use iget5_locked() and friends instead of conventional iget(). Wrap
966 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
967 to use ntfs_iget(). Leave only one iget() call at mount time so we
968 don't need an ntfs_iget_mount().
969 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
970 additional argument.
971
9722.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
973
974 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
975 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
976 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
977 to determine whether to apply mst fixups or not.
978 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
979 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
980 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
981 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
982 the VFS readpage function prototype to the ntfs_attr_read_block()
983 function prototype.
984
9852.0.11 - Initial preparations for fake inode based attribute i/o.
986
987 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
988 do some macro magic (adapted from include/linux/buffer_head.h) to
989 expand all the helper functions NInoFoo(), NInoSetFoo(), and
990 NInoClearFoo().
991 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
992 - Add new fields to ntfs_inode structure to allow use of fake inodes
993 for attribute i/o: type, name, name_len. Also add new state bits:
994 NI_Attr, which, if set, indicates the inode is a fake inode, and
995 NI_MstProtected, which, if set, indicates the attribute uses multi
996 sector transfer protection, i.e. fixups need to be applied after
997 reads and before/after writes.
998 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
999 ntfs_{new,clear,destroy}_extent_inode() and update callers.
1000 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
1001 instead of ntfs_destroy_extent_inode().
1002 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
1003 - Make all operations on ntfs inode state bits use the NIno* functions.
1004 - Set up the new ntfs inode fields and state bits in
1005 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
1006 allocated memory to __ntfs_clear_inode().
1007 - Cleanup ntfs_inode structure a bit for better ordering of elements
1008 w.r.t. their size to allow better packing of the structure in memory.
1009
10102.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
1011
1012 - Add check at mount time to verify that the number of inodes on the
1013 volume does not exceed 2^32 - 1, which is the maximum allowed for
1014 NTFS according to Microsoft.
1015 - Change mft_no member of ntfs_inode structure to be unsigned long.
1016 Update all users. This makes ntfs_inode->mft_no just a copy of struct
1017 inode->i_ino. But we can't just always use struct inode->i_ino and
1018 remove mft_no because extent inodes do not have an attached struct
1019 inode.
1020
10212.0.9 - Decompression engine now uses a single buffer and other cleanups.
1022
1023 - Change decompression engine to use a single buffer protected by a
1024 spin lock instead of per-CPU buffers. (Rusty Russell)
1025 - Do not update cb_pos when handling a partial final page during
1026 decompression of a sparse compression block, as the value is later
1027 reset without being read/used. (Rusty Russell)
1028 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
1029 Morton)
1030 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
1031 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
1032 it also makes everything safer so it is a good thing.
1033 - Miscellaneous minor cleanups to comments.
1034
10352.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
1036
1037 Big thanks go to Al Viro and other inhabitants of #kernel for investing
1038 their time to discuss the case sensitivity and dcache aliasing issues.
1039
1040 - Remove unused source file fs/ntfs/attraops.c.
1041 - Remove show_inodes mount option(s), thus dropping support for
1042 displaying of short file names.
1043 - Remove deprecated mount option posix.
1044 - Restore show_sys_files mount option.
1045 - Add new mount option case_sensitive, to determine if the driver
1046 treats file names as case sensitive or not. If case sensitive, create
1047 file names in the POSIX namespace. Otherwise create file names in the
1048 LONG/WIN32 namespace. Note, files remain accessible via their short
1049 file name, if it exists.
1050 - Remove really dumb logic bug in boot sector recovery code.
1051 - Fix dcache aliasing issues wrt short/long file names via changes
1052 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
1053 fs/ntfs/namei.c::ntfs_lookup():
1054 - Add additional argument to ntfs_lookup_inode_by_name() in which we
1055 return information about the matching file name if the case is not
1056 matching or the match is a short file name. See comments above the
1057 function definition for details.
1058 - Change ntfs_lookup() to only create dcache entries for the correctly
1059 cased file name and only for the WIN32 namespace counterpart of DOS
1060 namespace file names. This ensures we have only one dentry per
1061 directory and also removes all dcache aliasing issues between short
1062 and long file names once we add write support. See comments above
1063 function for details.
1064 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
1065
10662.0.7 - Minor cleanups and updates for changes in core kernel code.
1067
1068 - Remove much of the NULL struct element initializers.
1069 - Various updates to make compatible with recent kernels.
1070 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
1071 in fs/ntfs/ntfs.h instead.
1072 - Remove no longer needed KERNEL_VERSION checks. We are now in the
1073 kernel proper so they are no longer needed.
1074
10752.0.6 - Major bugfix to make compatible with other kernel changes.
1076
1077 - Initialize the mftbmp address space properly now that there are more
1078 fields in the struct address_space. This was leading to hangs and
1079 oopses on umount since 2.5.12 because of changes to other parts of
1080 the kernel. We probably want a kernel generic init_address_space()
1081 function...
1082 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
1083 only caller of ->readdir() is vfs_readdir() which holds i_sem during
1084 the call, and i_sem is sufficient protection against changes in the
1085 directory inode (including ->i_size).
1086 - Use generic_file_llseek() for directories (as opposed to
1087 default_llseek()) as this downs i_sem instead of the BKL which is
1088 what we now need for exclusion against ->f_pos changes considering we
1089 no longer take the BKL in ntfs_readdir().
1090
10912.0.5 - Major bugfix. Buffer overflow in extent inode handling.
1092
1093 - No need to set old blocksize in super.c::ntfs_fill_super() as the
1094 VFS does so via invocation of deactivate_super() calling
1095 fs->fill_super() calling block_kill_super() which does it.
1096 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
1097 -> Do we really need it? I don't think so as we have exclusion on
1098 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
1099 move the ->f_pos accesses under the mrec_lock though. Check this...
1100 - Fix really, really, really stupid buffer overflow in extent inode
1101 handling in mft.c::map_extent_mft_record().
1102
11032.0.4 - Cleanups and updates for kernel 2.5.11.
1104
1105 - Add documentation on how to use the MD driver to be able to use NTFS
1106 stripe and volume sets in Linux and generally cleanup documentation
1107 a bit.
1108 Remove all uses of kdev_t in favour of struct block_device *:
1109 - Change compress.c::ntfs_file_read_compressed_block() to use
1110 sb_getblk() instead of getblk().
1111 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
1112 of get_hardsect_size().
1113 - No need to get old blocksize in super.c::ntfs_fill_super() as
1114 fs/super.c::get_sb_bdev() already does this.
1115 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
1116
11172.0.3 - Small bug fixes, cleanups, and performance improvements.
1118
1119 - Remove some dead code from mft.c.
1120 - Optimize readpage and read_block functions throughout aops.c so that
1121 only initialized blocks are read. Non-initialized ones have their
1122 buffer head mapped, zeroed, and set up to date, without scheduling
1123 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
1124 Thanks go to Andrew Morton for spotting the below:
1125 - Fix buglet in allocate_compression_buffers() error code path.
1126 - Call flush_dcache_page() after modifying page cache page contents in
1127 ntfs_file_readpage().
1128 - Check for existence of page buffers throughout aops.c before calling
1129 create_empty_buffers(). This happens when an I/O error occurs and the
1130 read is retried. (It also happens once writing is implemented so that
1131 needed doing anyway but I had left it for later...)
1132 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
1133 readpage and read_block functions. Reasoning same as above (i.e. I/O
1134 error retries and future write code paths.)
1135
11362.0.2 - Minor updates and cleanups.
1137
1138 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
1139 and cleanup the code a bit, removing the unused size parameter.
1140 - Change default fmask to 0177 and update documentation.
1141 - Change attrib.c::get_attr_search_ctx() to return the search context
1142 directly instead of taking the address of a pointer. A return value
1143 of NULL means the allocation failed. Updated all callers
1144 appropriately.
1145 - Update to 2.5.9 kernel (preserving backwards compatibility) by
1146 replacing all occurences of page->buffers with page_buffers(page).
1147 - Fix minor bugs in runlist merging, also minor cleanup.
1148 - Updates to bootsector layout and mft mirror contents descriptions.
1149 - Small bug fix in error detection in unistr.c and some cleanups.
1150 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
1151 bytes.
1152
11532.0.1 - Minor updates.
1154
1155 - Make default umask correspond to documentation.
1156 - Improve documentation.
1157 - Set default mode to include execute bit. The {u,f,d}mask can be used
1158 to take it away if desired. This allows binaries to be executed from
1159 a mounted ntfs partition.
1160
11612.0.0 - New version number. Remove TNG from the name. Now in the kernel.
1162
1163 - Add kill_super, just keeping up with the vfs changes in the kernel.
1164 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
1165 from the CVS import into BitKeeper.
1166 - Begin to implement proper handling of allocated_size vs
1167 initialized_size vs data_size (i.e. i_size). Done are
1168 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
1169 and attrib.c::load_attribute_list().
1170 - Lock the runlist in attrib.c::load_attribute_list() while using it.
1171 - Fix memory leak in ntfs_file_read_compressed_block() and generally
1172 clean up compress.c a little, removing some uncommented/unused debug
1173 code.
1174 - Tidy up dir.c a little bit.
1175 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
1176 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
1177 creating aops.c::ntfs_mst_readpage(), improving the handling of
1178 holes and overflow in the process and implementing the correct
1179 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
1180 I am aiming for correctness at the moment. Modularisation can come
1181 later.
1182 - Rename aops.c::end_buffer_read_index_async() to
1183 end_buffer_read_mst_async() and optimize the overflow checking and
1184 handling.
1185 - Use the host of the mftbmp address space mapping to hold the ntfs
1186 volume. This is needed so the async i/o completion handler can
1187 retrieve a pointer to the volume. Hopefully this will not cause
1188 problems elsewhere in the kernel... Otherwise will need to use a
1189 fake inode.
1190 - Complete implementation of proper handling of allocated_size vs
1191 initialized_size vs data_size (i.e. i_size) in whole driver.
1192 Basically aops.c is now completely rewritten.
1193 - Change NTFS driver name to just NTFS and set version number to 2.0.0
1194 to make a clear distinction from the old driver which is still on
1195 version 1.1.22.
1196
1197tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
1198
1199 - Replace bdevname(sb->s_dev) with sb->s_id.
1200 - Remove now superfluous new-line characters in all callers of
1201 ntfs_debug().
1202 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
1203 directories. Without this the "find" utility gets very upset which is
1204 fair enough as Linux/Unix do not support directory hard links.
1205 - Further runlist merging work. (Richard Russon)
1206 - Backwards compatibility for gcc-2.95. (Richard Russon)
1207 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001208 - Convert to new filesystem declaration using ->ntfs_get_sb() and
Linus Torvalds1da177e2005-04-16 15:20:36 -07001209 replacing ntfs_read_super() with ntfs_fill_super().
1210 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
1211 overflow on 32-bit architectures.
1212 - Cleanup upcase loading code to use ntfs_(un)map_page().
1213 - Disable/reenable preemtion in critical sections of compession engine.
1214 - Replace device size determination in ntfs_fill_super() with
1215 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
1216 function super.c::get_nr_blocks().
1217 - Implement a mount time option (show_inodes) allowing choice of which
1218 types of inode names readdir() returns and modify ntfs_filldir()
1219 accordingly. There are several parameters to show_inodes:
1220 system: system files
1221 win32: long file names (including POSIX file names) [DEFAULT]
1222 long: same as win32
1223 dos: short file names only (excluding POSIX file names)
1224 short: same as dos
1225 posix: same as both win32 and dos
1226 all: all file names
1227 Note that the options are additive, i.e. specifying:
1228 -o show_inodes=system,show_inodes=win32,show_inodes=dos
1229 is the same as specifying:
1230 -o show_inodes=all
1231 Note that the "posix" and "all" options will show all directory
1232 names, BUT the link count on each directory inode entry is set to 1,
1233 due to Linux not supporting directory hard links. This may well
1234 confuse some userspace applications, since the directory names will
1235 have the same inode numbers. Thus it is NOT advisable to use the
1236 "posix" or "all" options. We provide them only for completeness sake.
1237 - Add copies of allocated_size, initialized_size, and compressed_size to
1238 the ntfs inode structure and set them up in
1239 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
1240 for files and the index allocation attribute for directories.
1241 - Add copies of allocated_size and initialized_size to ntfs inode for
1242 $BITMAP attribute of large directories and set them up in
1243 inode.c::ntfs_read_inode().
1244 - Add copies of allocated_size and initialized_size to ntfs volume for
1245 $BITMAP attribute of $MFT and set them up in
1246 super.c::load_system_files().
1247 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
1248 posix, and utf8) and tell user what the new options to use are. Note
1249 we still do support them but they will be removed with kernel 2.7.x.
1250 - Change all occurences of integer long long printf formatting to hex
1251 as printk() will not support long long integer format if/when the
1252 div64 patch goes into the kernel.
1253 - Make slab caches have stable names and change the names to what they
1254 were intended to be. These changes are required/made possible by the
1255 new slab cache name handling which removes the length limitation by
1256 requiring the caller of kmem_cache_create() to supply a stable name
1257 which is then referenced but not copied.
1258 - Rename run_list structure to run_list_element and create a new
1259 run_list structure containing a pointer to a run_list_element
1260 structure and a read/write semaphore. Adapt all users of runlists
1261 to new scheme and take and release the lock as needed. This fixes a
1262 nasty race as the run_list changes even when inodes are locked for
1263 reading and even when the inode isn't locked at all, so we really
1264 needed the serialization. We use a semaphore rather than a spinlock
1265 as memory allocations can sleep and doing everything GFP_ATOMIC
1266 would be silly.
1267 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
1268 This can never happen due to the nature of lookup_attr() and how we
1269 support attribute lists. If it did happen it would imply the inode
1270 being corrupt.
1271 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
1272 bad if found.
1273 - Update to 2.5.6-pre2 changes in struct address_space.
1274 - Use parent_ino() when accessing d_parent inode number in dir.c.
1275 - Import Sourceforge CVS repository into BitKeeper repository:
1276 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
1277 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
1278 Documentation/filesystems/ntfs.txt for NTFS TNG.
1279 - Create kernel configuration option controlling whether debugging
1280 is enabled or not.
1281 - Add the required export of end_buffer_io_sync() from the patches
1282 directory to the kernel code.
1283 - Update inode.c::ntfs_show_options() with show_inodes mount option.
1284 - Update errors mount option.
1285
1286tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
1287
1288 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
1289 bug in mapping of extent inodes. Update all the comments to fit all
1290 the recent code changes.
1291 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
1292 - Cleanups in compress.c, mostly comments and folding help.
1293 - Implement attrib.c::map_run_list() as a generic helper.
1294 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
1295 thus making code shorter and enabling attribute list support.
1296 - Cleanup incorrect use of [su]64 with %L printf format specifier in
1297 all source files. Type casts to [unsigned] long long added to correct
1298 the mismatches (important for architectures which have long long not
1299 being 64 bits).
1300 - Merge async io completion handlers for directory indexes and $MFT
1301 data into one by setting the index_block_size{_bits} of the ntfs
1302 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
1303 - Cleanup aops.c, update comments.
1304 - Make ntfs_file_get_block() use map_run_list() so all files now
1305 support attribute lists.
1306 - Make ntfs_dir_readpage() almost verbatim copy of
1307 block_read_full_page() by using ntfs_file_get_block() with only real
1308 difference being the use of our own async io completion handler
1309 rather than the default one, thus reducing the amount of code and
1310 automatically enabling attribute list support for directory indices.
1311 - Fix bug in load_attribute_list() - forgot to call brelse in error
1312 code path.
1313 - Change parameters to find_attr() and lookup_attr(). We no longer
1314 pass in the upcase table and its length. These can be gotten from
1315 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
1316 - Cleanups in attrib.c.
1317 - Implement merging of runlists, attrib.c::merge_run_lists() and its
1318 helpers. (Richard Russon)
1319 - Attribute lists part 2, attribute extents and multi part runlists:
1320 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
1321 further runlist parts via attrib.c::map_run_list().
1322 - Tiny endianness bug fix in decompress_mapping_pairs().
1323
1324tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
1325
1326 - Enable encrypted directories. (Their index root is marked encrypted
1327 to indicate that new files in that directory should be created
1328 encrypted.)
1329 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
1330 - Enable $Extend system directory. Most (if not all) extended system
1331 files do not have unnamed data attributes so ntfs_read_inode() had to
1332 special case them but that is ok, as the special casing recovery
1333 happens inside an error code path so there is zero slow down in the
1334 normal fast path. The special casing is done by introducing a new
1335 function inode.c::ntfs_is_extended_system_file() which checks if any
1336 of the hard links in the inode point to $Extend as being their parent
1337 directory and if they do we assume this is an extended system file.
1338 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
1339 when compiled with -DDEBUG. Default is debug messages to be disabled.
1340 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
1341 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
1342 interface is enabled). Inspired by old ntfs driver.
1343 - Add debug_msgs insmod/kernel boot parameter to set whether debug
1344 messages are {dis,en}abled. This is useful to enable debug messages
1345 during ntfs initialization and is the only way to activate debugging
1346 when the sysctl interface is not enabled.
1347 - Cleanup debug output in various places.
1348 - Remove all dollar signs ($) from the source (except comments) to
1349 enable compilation on architectures whose gcc compiler does not
1350 support dollar signs in the names of variables/constants. Attribute
1351 types now start with AT_ instead of $ and $I30 is now just I30.
1352 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
1353 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
1354 access functions. This means we now cope with $MFT/$BITMAP being
1355 spread accross several mft records.
1356 - Disable modification of mft_zone_multiplier on remount. We can always
1357 reenable this later on if we really want to, but we will need to make
1358 sure we readjust the mft_zone size / layout accordingly.
1359
1360tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
1361
1362 - Use sb_set_blocksize() instead of set_blocksize() and verify the
1363 return value.
1364 - Use sb_bread() instead of bread() throughout.
1365 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1366 of a directory index block vcn. Apply resulting simplifications in
1367 dir.c everywhere.
1368 - Fix a small bug somewhere (but forgot what it was).
1369 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1370 on the printf-format parameter list and fix bugs reported by gcc
1371 as a result. (Richard Russon)
1372 - Move inode allocation strategy to Al's new stuff but maintain the
1373 divorce of ntfs_inode from struct inode. To achieve this we have two
1374 separate slab caches, one for big ntfs inodes containing a struct
1375 inode and pure ntfs inodes and at the same time fix some faulty
1376 error code paths in ntfs_read_inode().
1377 - Show mount options in proc (inode.c::ntfs_show_options()).
1378
1379tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1380
1381 - Modified (un)map_mft_record functions to be common for read and write
1382 case. To specify which is which, added extra parameter at front of
1383 parameter list. Pass either READ or WRITE to this, each has the
1384 obvious meaning.
1385 - General cleanups to allow for easier folding in vi.
1386 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1387 argument, and invokes attrib.c::merge_run_lists() to merge the old
1388 and the new runlists.
1389 - Removed attrib.c::find_first_attr().
1390 - Implemented loading of attribute list and complete runlist for $MFT.
1391 This means we now cope with $MFT being spread across several mft
1392 records.
1393 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1394 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1395 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1396 cache. There are too little ntfs_volume structures at any one time
1397 to justify a private slab cache.
1398 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1399 Use KM_BIO_IRQ on advice from IRC/kernel...
1400 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1401 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1402 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1403 removed the now superfluous exports from the kernel core patch.
1404 - Fix a bug where kfree() was used insted of ntfs_free().
1405 - Change map_mft_record() to take ntfs_inode as argument instead of
1406 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1407 - Add pointer to ntfs_volume to ntfs_inode.
1408 - Add mft record number and sequence number to ntfs_inode. Stop using
1409 i_ino and i_generation for in-driver purposes.
1410 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1411 - Remove use of proper inodes by extent inodes. Move i_ino and
1412 i_generation to ntfs_inode to do this. Apply simplifications that
1413 result and remove iget_no_wait(), etc.
1414 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1415 - Add reference counting in ntfs_inode for the ntfs inode itself and
1416 for the mapped mft record.
1417 - Extend mft record mapping so we can (un)map extent mft records (new
1418 functions (un)map_extent_mft_record), and so mappings are reference
1419 counted and don't have to happen twice if already mapped - just ref
1420 count increases.
1421 - Add -o iocharset as alias to -o nls for backwards compatibility.
1422 - The latest core patch is now tiny. In fact just a single additional
1423 export is necessary over the base kernel.
1424
1425tng-0.0.3 - Cleanups, enhancements, bug fixes.
1426
1427 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1428 and setup the runlist appropriately using knowledge provided by the
1429 sizes in the base attribute record.
1430 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1431 any more.
1432 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1433 page or use vmalloc depending on the amount of memory requested.
1434 - Cleanup error output. The __FUNCTION__ "(): " is now added
1435 automatically. Introduced a new header file debug.h to support this
1436 and also moved ntfs_debug() function into it.
1437 - Make reading of compressed files more intelligent and especially get
1438 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1439 buffers (allocated at first mount with cluster size <= 4kiB and
1440 deallocated on last umount with cluster size <= 4kiB), and
1441 asynchronous io for the compressed data using a list of buffer heads.
1442 Er, we use synchronous io as async io only works on whole pages
1443 covered by buffers and not on individual buffer heads...
1444 - Bug fix for reading compressed files with sparse compression blocks.
1445
1446tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1447
1448 - Fixed handling of directories when cluster size exceeds index block
1449 size.
1450 - Hide DOS only name space directory entries from readdir() but allow
1451 them in lookup(). This should fix the problem that Linux doesn't
1452 support directory hard links, while still allowing access to entries
1453 via their short file name. This also has the benefit of mimicking
1454 what Windows users are used to, so it is the ideal solution.
1455 - Implemented sync_page everywhere so no more hangs in D state when
1456 waiting for a page.
1457 - Stop using bforget() in favour of brelse().
1458 - Stop locking buffers unnecessarily.
1459 - Implemented compressed files (inode->mapping contains uncompressed
1460 data, raw compressed data is currently bread() into a vmalloc()ed
1461 memory buffer).
1462 - Enable compressed directories. (Their index root is marked compressed
1463 to indicate that new files in that directory should be created
1464 compressed.)
1465 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1466 functions. (Thanks to Will Dyson for pointing this out.)
1467 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1468 ntfs_sb_info) out of the common inode and super_block structures and
1469 started using the generic_ip and generic_sbp pointers instead. This
1470 makes ntfs entirely private with respect to the kernel tree.
1471 - Detect compiler version and abort with error message if gcc less than
1472 2.96 is used.
1473 - Fix bug in name comparison function in unistr.c.
1474 - Implement attribute lists part 1, the infrastructure: search contexts
1475 and operations, find_external_attr(), lookup_attr()) and make the
1476 code use the infrastructure.
1477 - Fix stupid buffer overflow bug that became apparent on larger run
1478 list containing attributes.
1479 - Fix bugs in readdir() that became apparent on larger directories.
1480
1481 The driver is now really useful and survives the test
1482 find . -type f -exec md5sum "{}" \;
1483 without any error messages on a over 1GiB sized partition with >16k
1484 files on it, including compressed files and directories and many files
1485 and directories with attribute lists.
1486
1487tng-0.0.1 - The first useful version.
1488
1489 - Added ntfs_lookup().
1490 - Added default upcase generation and handling.
1491 - Added compile options to be shown on module init.
1492 - Many bug fixes that were "hidden" before.
1493 - Update to latest kernel.
1494 - Added ntfs_readdir().
1495 - Added file operations for mmap(), read(), open() and llseek(). We just
1496 use the generic ones. The whole point of going through implementing
1497 readpage() methods and where possible get_block() call backs is that
1498 this allows us to make use of the generic high level methods provided
1499 by the kernel.
1500
1501 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1502 though and it doesn't implement accesssing compressed files yet. Also,
1503 accessing files with attribute list attributes is not implemented yet
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001504 either. But for small or simple filesystems it should work and allow
Linus Torvalds1da177e2005-04-16 15:20:36 -07001505 you to list directories, use stat on directory entries and the file
1506 system, open, read, mmap and llseek around in files. A big mile stone
1507 has been reached!
1508
1509tng-0.0.0 - Initial version tag.
1510
1511 Initial driver implementation. The driver can mount and umount simple
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001512 NTFS filesystems (i.e. ones without attribute lists in the system
Linus Torvalds1da177e2005-04-16 15:20:36 -07001513 files). If the mount fails there might be problems in the error handling
1514 code paths, so be warned. Otherwise it seems to be loading the system
1515 files nicely and the mft record read mapping/unmapping seems to be
1516 working nicely, too. Proof of inode metadata in the page cache and non-
1517 resident file unnamed stream data in the page cache concepts is thus
1518 complete.