blob: 1d2ad15f1533add09bc6ff58ce666c6bf6fe305f [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001ToDo/Notes:
2 - Find and fix bugs.
3 - Checkpoint or disable the user space journal ($UsnJrnl).
4 - In between ntfs_prepare/commit_write, need exclusion between
5 simultaneous file extensions. Need perhaps an NInoResizeUnderway()
6 flag which we can set in ntfs_prepare_write() and clear again in
7 ntfs_commit_write(). Just have to be careful in readpage/writepage,
8 as well as in truncate, that we play nice... We might need to have
9 a data_size field in the ntfs_inode to store the real attribute
10 length. Also need to be careful with initialized_size extention in
11 ntfs_prepare_write. Basically, just be _very_ careful in this code...
12 OTOH, perhaps i_sem, which is held accross generic_file_write is
13 sufficient for synchronisation here. We then just need to make sure
14 ntfs_readpage/writepage/truncate interoperate properly with us.
15 UPDATE: The above is all ok as it is due to i_sem held. The only
16 thing that needs to be checked is ntfs_writepage() which does not
17 hold i_sem. It cannot change i_size but it needs to cope with a
18 concurrent i_size change.
19 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
20 leave the volume dirty on umount if the final iput(vol->mft_ino)
21 causes a write of any mirrored mft records due to the mft mirror
22 inode having been discarded already. Whether this can actually ever
23 happen is unclear however so it is worth waiting until someone hits
24 the problem.
25 - Enable the code for setting the NT4 compatibility flag when we start
26 making NTFS 1.2 specific modifications.
27
282.1.23-WIP
29
30 - Add printk rate limiting for ntfs_warning() and ntfs_error() when
31 compiled without debug. This avoids a possible denial of service
32 attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
33 out.
34
352.1.22 - Many bug and race fixes and error handling improvements.
36
37 - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
38 - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
39 instead of void and provide a helper ntfs_truncate_vfs() for the
40 vfs ->truncate method.
41 - Add a new ntfs inode flag NInoTruncateFailed() and modify
42 fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
43 - Fix min_size and max_size definitions in ATTR_DEF structure in
44 fs/ntfs/layout.h to be signed.
45 - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
46 ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
47 ntfs_attr_can_be_resident(), which in turn use the new private helper
48 ntfs_attr_find_in_attrdef().
49 - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
50 mapping->private_lock around the dirtying of the buffer heads
51 analagous to the way it is done in __set_page_dirty_buffers().
52 - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
53 mount time as this cannot work with the current implementation.
54 - Check for location of attribute name and improve error handling in
55 general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
56 - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
57 i_size, i.e. race with truncate, invalidate the buffers on the page
58 so that they become freeable and hence the page does not leak.
59 - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
60 Bunk)
61 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
62 a NULL pointer dereference in the error code path when a corrupt
63 attribute was found. (Thanks to Domen Puncer for the bug report.)
64 - Add MODULE_VERSION() to fs/ntfs/super.c.
65 - Make several functions and variables static. (Adrian Bunk)
66 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
67 buffers for the page if they are not present and then marks the
68 buffers belonging to the ntfs record dirty. This causes the buffers
69 to become busy and hence they are safe from removal until the page
70 has been written out.
71 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
72 error handling code path that resulted in a BUG() due to trying to
73 unmap an extent mft record when the mapping of it had failed and it
74 thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
75 - Drop the runlist lock after the vcn has been read in
76 fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
77 - Rewrite handling of multi sector transfer errors. We now do not set
78 PageError() when such errors are detected in the async i/o handler
79 fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
80 protected attributes now check the magic of each ntfs record as they
81 use it and act appropriately. This has the effect of making errors
82 granular per ntfs record rather than per page which solves the case
83 where we cannot access any of the ntfs records in a page when a
84 single one of them had an mst error. (Thanks to Ken MacFerrin for
85 the bug report.)
86 - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
87 where we failed to release i_sem on the $Quota/$Q attribute inode.
88 - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
89 - Add mapping of unmapped buffers to all remaining code paths, i.e.
90 fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
91 and write_mft_record_nolock(). From now on we require that the
92 complete runlist for the mft mirror is always mapped into memory.
93 - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
94 - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
95 - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
96 resident attribute will be smaller than a page which makes the code
97 simpler. Also make the code more tolerant to concurrent ->truncate.
98
992.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
100
101 - Implement extent mft record deallocation
102 fs/ntfs/mft.c::ntfs_extent_mft_record_free().
103 - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
104 - Add vol->mft_data_pos and initialize it at mount time.
105 - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
106 ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
107 ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
108 ntfs_runlists_merge() and adapt all callers.
109 - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
110 ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
111 and ntfs_mapping_pairs_build(), adapted from libntfs.
112 - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
113 static and add a declaration for it to lcnalloc.h.
114 - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
115 inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
116 cluster bitmap lock for the duration of the call.
117 - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
118 - Implement the equivalent of memset() for an ntfs attribute in
119 fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
120 fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
121 - Remove unnecessary casts from LCN_* constants.
122 - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
123 - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
124 change MFT_RECORD to contain the NTFS 3.1+ specific fields.
125 - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
126 marks all buffers belonging to an ntfs record dirty, followed by
127 marking the page the ntfs record is in dirty and also marking the vfs
128 inode containing the ntfs record dirty (I_DIRTY_PAGES).
129 - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
130 new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
131 longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
132 - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
133 include errors.
134 - Move the typedefs for runlist_element and runlist from types.h to
135 runlist.h and fix resulting include errors.
136 - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
137 - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
138 mark_ntfs_record_dirty() which also changes the behaviour in that we
139 now set the buffers belonging to the mft record dirty as well as the
140 page itself.
141 - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
142 to cope with the fact that there now are dirty buffers in mft pages.
143 - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
144 mark_ntfs_record_dirty() and thus to set the buffers belonging to the
145 mft record dirty as well as the page itself.
146 - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
147 slightly modified by me)
148 - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
149 the mft record is already locked and otherwise behaves the same way
150 as fs/ntfs/mft.c::map_mft_record().
151 - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
152 writes the mft record if the buffers belonging to it are dirty.
153 Otherwise we assume that it was written out by other means already.
154 - Attempting to write outside initialized size is _not_ a bug so remove
155 the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
156 fact required to write outside initialized size when preparing to
157 extend the initialized size.
158 - Map the page instead of using page_address() before writing to it in
159 fs/ntfs/aops.c::ntfs_mft_writepage().
160 - Provide exclusion between opening an inode / mapping an mft record
161 and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
162 by setting the page not uptodate throughout ntfs_mft_writepage().
163 - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
164 to ensure noone can see the page whilst the mst fixups are applied.
165 - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
166 checks if an mft record may be written out safely obtaining any
167 necessary locks in the process. This is used by
168 fs/ntfs/aops.c::ntfs_write_mst_block().
169 - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
170 writing mft records and improve its error handling in the process.
171 Now if any of the records in the page fail to be written out, all
172 other records will be written out instead of aborting completely.
173 - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
174 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
175 ntfs_mst_aops for all inodes which are NInoMstProtected() and
176 ntfs_aops for all other inodes.
177 - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
178 ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
179 no longer require an ntfs inode to be present. Update all callers.
180 - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
181 - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
182 to ensure noone can see the page whilst the mst fixups are applied.
183 - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
184 fs/ntfs/mft.c::try_map_mft_record().
185 - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
186 with the ntfs inode which contains the page rather than the ntfs
187 inode the mft record of which is in the page.
188 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
189 index inode bitmap inode release code from there to
190 fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
191 Hellwig for spotting this.)
192 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
193 inode semaphore around the code that sets ni->itype.index.bmp_ino to
194 NULL and reorganize the code to optimize it a bit. (Thanks to
195 Christoph Hellwig for spotting this.)
196 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
197 ntfs inode as a parameter as this is confusing and misleading and the
198 needed ntfs inode is available via NTFS_I(page->mapping->host).
199 Adapt all callers to this change.
200 - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
201 fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
202 of the first buffer in a record and to take this as the ntfs record
203 dirty state. We cannot look at the dirty state for subsequent
204 buffers because we might be racing with
205 fs/ntfs/aops.c::mark_ntfs_record_dirty().
206 - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
207 inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
208 add a declaration for it to inode.h. Fix some compilation issues
209 that resulted due to #includes and header file interdependencies.
210 - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
211 - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
212 - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
213 record sequence number if it is specified (i.e. not zero).
214 - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
215 functions used by it.
216 - Update Documentation/filesystems/ntfs.txt with instructions on how to
217 use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
218 the linear raid problem with the Software RAID / MD driver when one
219 or more of the devices has an odd number of sectors.
220
2212.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
222
223 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
224 where we did not clear ctx->al_entry but it was still set due to
225 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
226 particular.
227 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
228 where we forgot to unmap the extent mft record when we had finished
229 enumerating an attribute which caused a bug check to trigger when the
230 VFS calls ->clear_inode.
231
2322.1.19 - Many cleanups, improvements, and a minor bug fix.
233
234 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
235 change the uid, gid, and mode of an inode as we do not support NTFS
236 ACLs yet.
237 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
238 kernel.
239 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
240 and ntfs_filldir() as per suggestion from Al Viro.
241 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
242 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
243 inode size has changed and to only output an error if so.
244 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
245 - Add le{16,32,64} as well as sle{16,32,64} data types to
246 fs/ntfs/types.h.
247 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
248 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
249 respectively, to fs/ntfs/types.h.
250 - Update endianness conversion macros in fs/ntfs/endian.h to use the
251 new types as appropriate.
252 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
253 and index.c.
254 - Add leMFT_REF data type to fs/ntfs/layout.h.
255 - Update all NTFS header files with the new little endian data types.
256 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
257 - Do proper type casting when using ntfs_is_*_recordp() in
258 fs/ntfs/logfile.c, mft.c, and super.c.
259 - Fix all the sparse bitwise warnings. Had to change all the typedef
260 enums storing little endian values to simple enums plus a typedef for
261 the datatype to make sparse happy.
262 - Fix a bug found by the new sparse bitwise warnings where the default
263 upcase table was defined as a pointer to wchar_t rather than ntfschar
264 in fs/ntfs/ntfs.h and super.c.
265 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
266
2672.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
268
269 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
270 the calculation of total/free inodes as used by statfs().
271 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
272 because the code itself is using the ntfs_lock semaphore which
273 provides safe locking. (Ingo Molnar)
274 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
275 could occur in the future for when we start closing/freeing extent
276 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
277 we free it.
278 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
279 find_external_attr() to ntfs_external_attr_find() to cleanup the
280 namespace a bit and to be more consistent with libntfs.
281 - Rename {{re,}init,get,put}_attr_search_ctx() to
282 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
283 attr_search_context to ntfs_attr_search_ctx.
284 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
285 for the attribute list attribute itself.
286 - Fix endianness bug in ntfs_external_attr_find().
287 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
288 if the attribute is not found, and -EIO on real error. In the case
289 of -ENOENT, the search context is updated to describe the attribute
290 before which the attribute being searched for would need to be
291 inserted if such an action were to be desired and in the case of
292 ntfs_external_attr_find() the search context is also updated to
293 indicate the attribute list entry before which the attribute list
294 entry of the attribute being searched for would need to be inserted
295 if such an action were to be desired. Also make ntfs_find_attr()
296 static and remove its prototype from attrib.h as it is not used
297 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
298 callers of ntfs_{external,}attr_{find,lookup}() for the new return
299 values.
300 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
301
3022.1.17 - Fix bugs in mount time error code paths and other updates.
303
304 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
305 includes functions to set/clear a single bit or a run of bits.
306 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
307 runlist element containing a particular vcn. It also takes care of
308 mapping any needed runlist fragments.
309 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
310 - Load attribute definition table from $AttrDef at mount time.
311 - Fix bugs in mount time error code paths involving (de)allocation of
312 the default and volume upcase tables.
313 - Remove ntfs_nr_mounts as it is no longer used.
314
3152.1.16 - Implement access time updates, file sync, async io, and read/writev.
316
317 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
318 This is done by setting the appropriate file operations pointers to
319 the generic helper functions provided by mm/filemap.c.
320 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
321 and directories (fs/ntfs/dir.c).
322 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
323 Note, except for the root directory and any other system files opened
324 by the user, the system files will not have their access times
325 updated as they are only accessed at the inode level an hence the
326 file level functions which cause the times to be updated are never
327 invoked.
328
3292.1.15 - Invalidate quotas when (re)mounting read-write.
330
331 - Add new element itype.index.collation_rule to the ntfs inode
332 structure and set it appropriately in ntfs_read_locked_inode().
333 - Implement a new inode type "index" to allow efficient access to the
334 indices found in various system files and adapt inode handling
335 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
336 attribute inode (NInoAttr() is true) with an attribute type of
337 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
338 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
339 there would be no way to distinguish between normal attribute inodes
340 and index inodes. The function to obtain an index inode is
341 ntfs_index_iget() and it uses the helper function
342 ntfs_read_locked_index_inode(). Note, we do not overload
343 ntfs_attr_iget() as indices consist of multiple attributes so using
344 ntfs_attr_iget() to obtain an index inode would be confusing.
345 - Ensure that there is no overflow when doing page->index <<
346 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
347 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
348 and ntfs_read_block().
349 - Use case sensitive attribute lookups instead of case insensitive ones.
350 - Lock all page cache pages belonging to mst protected attributes while
351 accessing them to ensure we never see corrupt data while the page is
352 under writeout.
353 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
354 We have ntfs_is_collation_rule_supported() to check if the collation
355 rule you want to use is supported and ntfs_collation() which actually
356 collates two data items. We currently only support COLLATION_BINARY
357 and COLLATION_NTOFS_ULONG but support for other collation rules will
358 be added as the need arises.
359 - Add a new type, ntfs_index_context, to allow retrieval of an index
360 entry using the corresponding index key. To get an index context,
361 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
362 This also adds a new slab cache for the index contexts. To lookup a
363 key in an index inode, use ntfs_index_lookup(). After modifying an
364 index entry, call ntfs_index_entry_flush_dcache_page() followed by
365 ntfs_index_entry_mark_dirty() to ensure the changes are written out
366 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
367 an index entry is in the index allocation attribute rather than the
368 index root attribute it will not be written out (you will get a
369 warning message about discarded changes instead).
370 - Load the quota file ($Quota) and check if quota tracking is enabled
371 and if so, mark the quotas out of date. This causes windows to
372 rescan the volume on boot and update all quota entries.
373 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
374 It is simply set to __set_page_dirty_nobuffers() to make sure that
375 running set_page_dirty() on a page containing mft/ntfs records will
376 not affect the dirty state of the page buffers.
377 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
378 buffers that are inside the ntfs record in the page dirty after which
379 it sets the page dirty. This allows ->writepage to only write the
380 dirty index records rather than having to write all the records in
381 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
382 use this rather than __set_page_dirty_nobuffers().
383 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
384 writing of page cache pages belonging to mst protected attributes
385 like the index allocation attribute in directory indices and other
386 indices like $Quota/$Q, etc. This means that the quota is now marked
387 out of date on all volumes rather than only on ones where the quota
388 defaults entry is in the index root attribute of the $Quota/$Q index.
389
3902.1.14 - Fix an NFSd caused deadlock reported by several users.
391
392 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
393 to a buffer so that we can put the search context and unmap the mft
394 record before calling the filldir() callback. We need to do this
395 because of NFSd which calls ->lookup() from its filldir callback()
396 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
397 of the directory and since ntfs_readdir() has got it mapped already
398 ntfs_lookup() deadlocks.
399
4002.1.13 - Enable overwriting of resident files and housekeeping of system files.
401
402 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
403 keeping the mft mirror in sync with the mft when mirrored mft records
404 are written. The functions are write_mft_record{,_nolock}(). The
405 implementation is quite rudimentary for now with lots of things not
406 implemented yet but I am not sure any of them can actually occur so
407 I will wait for people to hit each one and only then implement it.
408 - Commit open system inodes at umount time. This should make it
409 virtually impossible for sync_mft_mirror_umount() to ever be needed.
410 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
411 ntfs super operations. This gives us inode writing via the VFS inode
412 dirty code paths. Note: Access time updates are not implemented yet.
413 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
414 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
415 finally enabling resident file overwrite! (-8 This also includes a
416 placeholder for ->writepage (ntfs_mft_writepage()), which for now
417 just redirties the page and returns. Also, at umount time, we for
418 now throw away all mft data page cache pages after the last call to
419 ntfs_commit_inode() in the hope that all inodes will have been
420 written out by then and hence no dirty (meta)data will be lost. We
421 also check for this case and emit an error message telling the user
422 to run chkdsk.
423 - Use set_page_writeback() and end_page_writeback() in the resident
424 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
425 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
426 page is clean.
427 - Implement ntfs_mft_writepage() so it now checks if any of the mft
428 records in the page are dirty and if so redirties the page and
429 returns. Otherwise it just returns (after doing set_page_writeback(),
430 unlock_page(), end_page_writeback() or the radix-tree tag
431 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
432 alowing the VM to do with the page as it pleases. Also, at umount
433 time, now only throw away dirty mft (meta)data pages if dirty inodes
434 are present and ask the user to email us if they see this happening.
435 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
436 information flags (fs/ntfs/super.c).
437 - Mark the volume dirty when (re)mounting read-write and mark it clean
438 when unmounting or remounting read-only. If any volume errors are
439 found, the volume is left marked dirty to force chkdsk to run.
440 - Add code to set the NT4 compatibility flag when (re)mounting
441 read-write for newer NTFS versions but leave it commented out for now
442 since we do not make any modifications that are NTFS 1.2 specific yet
443 and since setting this flag breaks Captive-NTFS which is not nice.
444 This code must be enabled once we start writing NTFS 1.2 specific
445 changes otherwise Windows NTFS driver might crash / cause corruption.
446
4472.1.12 - Fix the second fix to the decompression engine and some cleanups.
448
449 - Add a new address space operations struct, ntfs_mst_aops, for mst
450 protected attributes. This is because the default ntfs_aops do not
451 make sense with mst protected data and were they to write anything to
452 such an attribute they would cause data corruption so we provide
453 ntfs_mst_aops which does not have any write related operations set.
454 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
455 includes an adapted ntfs_commit_inode() and an implementation of
456 ntfs_write_inode() which for now just cleans dirty inodes without
457 writing them (it does emit a warning that this is happening).
458 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
459 entry) as it was only fixing a theoretical bug but at the same time
460 it badly broke the handling of sparse and uncompressed compression
461 blocks.
462
4632.1.11 - Driver internal cleanups.
464
465 - Only build logfile.o if building the driver with read-write support.
466 - Really final white space cleanups.
467 - Use generic_ffs() instead of ffs() in logfile.c which allows the
468 log_page_size variable to be optimized by gcc into a constant.
469 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
470 char as defined by POSIX and as found on some systems.
471
4722.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
473
474 - Finish off the white space cleanups (remove trailing spaces, etc).
475 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
476 the kludges around the first iget(). Instead of (re)setting ->s_op
477 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
478 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
479 kills the need for second super_operations and allows to return error
480 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
481 tricks. (Al Viro)
482 - Force read-only (re)mounting if any of the following bits are set in
483 the volume information flags:
484 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
485 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
486 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
487 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
488 above bits set so the test is made easy.
489
4902.1.9 - Fix two bugs in decompression engine.
491
492 - Fix a bug where we would not always detect that we have reached the
493 end of a compression block because we were ending at minus one byte
494 which is effectively the same as being at the end. The fix is to
495 check whether the uncompressed buffer has been fully filled and if so
496 we assume we have reached the end of the compression block. A big
497 thank you to Marcin Gibuła for the bug report, the assistance in
498 tracking down the bug and testing the fix.
499 - Fix a possible bug where when a compressed read is truncated to the
500 end of the file, the offset inside the last page was not truncated.
501
5022.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
503
504 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
505 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
506 utc2ntfs() to work with struct timespec instead of time_t on the
507 Linux UTC time side thus preserving the full precision of the NTFS
508 time and only loosing up to 99 nano-seconds in the Linux UTC time.
509 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
510 static inline.
511 - Remove unused ntfs_dirty_inode().
512 - Cleanup super operations declaration in fs/ntfs/super.c.
513 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
514 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
515 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
516 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
517 fs/ntfs/inode.h so they can be used elsewhere.
518 - Determine the mft mirror size as the number of mirrored mft records
519 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
520 - Load the mft mirror at mount time and compare the mft records stored
521 in it to the ones in the mft. Force a read-only mount if the two do
522 not match (fs/ntfs/super.c).
523 - Fix type casting related warnings on 64-bit architectures. Thanks
524 to Meelis Roos for reporting them.
525 - Move %L to %ll as %L is floating point and %ll is integer which is
526 what we want.
527 - Read the journal ($LogFile) and determine if the volume has been
528 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
529 and fs/ntfs/logfile.c). This is a little bit of a crude check in
530 that we only look at the restart areas and not at the actual log
531 records so that there will be a very small number of cases where we
532 think that a volume is dirty when in fact it is clean. This should
533 only affect volumes that have not been shutdown cleanly and did not
534 have any pending, non-check-pointed i/o.
535 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
536 is requested, empty $LogFile by overwriting it with 0xff bytes to
537 ensure that Windows cannot cause data corruption by replaying a stale
538 journal after Linux has written to the volume.
539
5402.1.7 - Enable NFS exporting of mounted NTFS volumes.
541
542 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
543 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
544 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
545 default doesn't allow inode number 0 which is a valid inode on NTFS
546 and even if it did allow that it uses iget() instead of ntfs_iget()
547 which makes it useless for us.
548 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
549 default just returns -EACCES which is not very useful.
550 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
551 and set them up in the super block at mount time (super.c) this
552 allows mounted NTFS volumes to be exported via NFS.
553 - Add missing return -EOPNOTSUPP; in
554 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
555 - Enforce no atime and no dir atime updates at mount/remount time as
556 they are not implemented yet anyway.
557 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
558 after a NULL check. Thanks to Dave Jones for pointing this out.
559
5602.1.6 - Fix minor bug in handling of compressed directories.
561
562 - Fix bug in handling of compressed directories. A compressed
563 directory is not really compressed so when we set the ->i_blocks
564 field of a compressed directory inode we were setting it from the
565 non-existing field ni->itype.compressed.size which gave random
566 results... For directories we now always use ni->allocated_size.
567
5682.1.5 - Fix minor bug in attribute list attribute handling.
569
570 - Fix bug in attribute list handling. Actually it is not as much a bug
571 as too much protection in that we were not allowing attribute lists
572 which waste space on disk while Windows XP clearly allows it and in
573 fact creates such attribute lists so our driver was failing.
574 - Update NTFS documentation ready for 2.6 kernel release.
575
5762.1.4 - Reduce compiler requirements.
577
578 - Remove all uses of unnamed structs and unions in the driver to make
579 old and newer gcc versions happy. Makes it a bit uglier IMO but at
580 least people will stop hassling me about it.
581
5822.1.3 - Important bug fixes in corner cases.
583
584 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
585 clusters. (Philipp Thomas)
586 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
587 multiple of the block_size but not the cluster size. (Szabolcs
588 Szakacsits <szaka@sienet.hu>)
589
5902.1.2 - Important bug fixes aleviating the hangs in statfs.
591
592 - Fix buggy free cluster and free inode determination logic.
593
5942.1.1 - Minor updates.
595
596 - Add handling for initialized_size != data_size in compressed files.
597 - Reduce function local stack usage from 0x3d4 bytes to just noise in
598 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
599 - Remove compiler warnings for newer gcc.
600 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
601 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
602 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
603 kmap_atomic(KM_USER0).
604
6052.1.0 - First steps towards write support: implement file overwrite.
606
607 - Add configuration option for developmental write support with an
608 appropriately scary configuration help text.
609 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
610 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
611 overwriting of existing files on ntfs. Note: Resident files are
612 only written into memory, and not written out to disk at present, so
613 avoid writing to files smaller than about 1kiB.
614 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
615 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
616 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
617 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
618 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
619 This enables write(2) based overwriting of existing files on ntfs.
620 Note: As with mmap(2) based overwriting, resident files are only
621 written into memory, and not written out to disk at present, so avoid
622 writing to files smaller than about 1kiB.
623 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
624 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
625 files with the purpose of intercepting and aborting all i_size
626 changes which we do not support yet. ntfs_truncate() actually only
627 emits a warning message but AFAICS our interception of i_size changes
628 elsewhere means ntfs_truncate() never gets called for i_size changes.
629 It is only called from generic_file_write() when we fail in
630 ntfs_prepare_{,nonresident_}write() in order to discard any
631 instantiated buffers beyond i_size. Thus i_size is not actually
632 changed so our warning message is enough. Unfortunately it is not
633 possible to easily determine if i_size is being changed or not hence
634 we just emit an appropriately worded error message.
635
6362.0.25 - Small bug fixes and cleanups.
637
638 - Unlock the page in an out of memory error code path in
639 fs/ntfs/aops.c::ntfs_read_block().
640 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
641 just unlock the page and return. (This can happen due to ->writepage
642 clearing PageUptodate() during write out of MstProtected()
643 attributes.
644 - Remove leaked write code again.
645
6462.0.24 - Cleanups.
647
648 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
649 inside BUG_ON(). (Adam J. Richter)
650 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
651 calls for improved debugging. (Adam J. Richter)
652 - Add errors flag to the ntfs volume state, accessed via
653 NVol{,Set,Clear}Errors(vol).
654 - Do not allow read-write remounts of read-only volumes with errors.
655 - Clarify comment for ntfs file operation sendfile which was added by
656 Christoph Hellwig a while ago (just using generic_file_sendfile())
657 to say that ntfs ->sendfile is only used for the case where the
658 source data is on the ntfs partition and the destination is
659 somewhere else, i.e. nothing we need to concern ourselves with.
660 - Add generic_file_write() as our ntfs file write operation.
661
6622.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
663
664 - Massive internal locking changes to mft record locking. Fixes lock
665 recursion and replaces the mrec_lock read/write semaphore with a
666 mutex. Also removes the now superfluous mft_count. This fixes several
667 race conditions and deadlocks, especially in the future write code.
668 - Fix ntfs over loopback for compressed files by adding an
669 optimization barrier. (gcc was screwing up otherwise ?)
670 - Miscellaneous cleanups all over the code and a fix or two in error
671 handling code paths.
672 Thanks go to Christoph Hellwig for pointing out the following two:
673 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
674 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
675
6762.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
677
678 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
679 at entry/exit respectively.
680 - Use C99 initializers for structures.
681 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
682
6832.0.21 - Check for, and refuse to work with too large files/directories/volumes.
684
685 - Limit volume size at mount time to 2TiB on architectures where
686 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
687 This is the most we can do without overflowing the 32-bit limit of
688 the block device size imposed on us by sb_bread() and sb_getblk()
689 for the time being.
690 - Limit file/directory size at open() time to 16TiB on architectures
691 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
692 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
693 overflowing the page cache page index.
694
6952.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
696
697 - Move the directory index bitmap to use an attribute inode instead of
698 having special fields for it inside the ntfs inode structure. This
699 means that the index bitmaps now use the page cache for i/o, too,
700 and also as a side effect we get support for non-resident index
701 bitmaps for free.
702 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
703 fix a page leak that manifested itself in some cases.
704 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
705 index bitmap inode on the final iput().
706
7072.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
708
709 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
710 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
711 - Drop the "file" from ntfs_file_read_compressed_block().
712 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
713 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
714 - Update ntfs_end_buffer_async_read() with the improved logic from
715 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
716 further logic improvements to better determine when we set PageError.
717 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
718 check for the buffers being uptodate first in line with the updated
719 fs/buffer.c::block_read_full_page(). This plugs a small race
720 condition.
721
7222.0.18 - Fix race condition in reading of compressed files.
723
724 - There was a narrow window between checking a buffer head for being
725 uptodate and locking it in ntfs_file_read_compressed_block(). We now
726 lock the buffer and then check whether it is uptodate or not.
727
7282.0.17 - Cleanups and optimizations - shrinking the ToDo list.
729
730 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
731 code and update callers, i.e. ntfs_iget(), to pass that error code
732 up instead of just using -EIO.
733 - Modifications to super.c to ensure that both mount and remount
734 cannot set any write related options when the driver is compiled
735 read-only.
736 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
737 cache the current runlist element. This should improve performance
738 when reading very large and/or very fragmented data.
739
7402.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
741
742 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
743 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
744 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
745 - Convert $MFT/$BITMAP access to attribute inode API and remove all
746 remnants of the ugly mftbmp address space and operations hack. This
747 means we finally have only one readpage function as well as only one
748 async io completion handler. Yey! The mft bitmap is now just an
749 attribute inode and is accessed from vol->mftbmp_ino just as if it
750 were a normal file. Fake inodes rule. (-:
751
7522.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
753
754 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
755 remounts to fail when the partition had an entry in /etc/fstab and
756 the entry specified the nls= option.
757 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
758 expand all the helper functions NVolFoo(), NVolSetFoo(), and
759 NVolClearFoo().
760 - Move copyright statement from driver initialisation message to
761 module description (fs/super.c). This makes the initialisation
762 message fit on one line and fits in better with rest of kernel.
763 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
764 attribute inodes, and both for files and directories.
765 - Implement fake attribute inodes allowing all attribute i/o to go via
766 the page cache and to use all the normal vfs/mm functionality:
767 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
768 to fs/ntfs/inode.c.
769 - Add needed cleanup code to ntfs_clear_big_inode().
770 - Merge address space operations for files and directories (aops.c),
771 now just have ntfs_aops:
772 - Rename:
773 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
774 ntfs_attr_read_block() -> ntfs_read_block(),
775 ntfs_file_read_page() -> ntfs_readpage().
776 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
777 attribute inodes, and both for files and directories.
778 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
779
7802.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
781
782 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
783 the locking out of super.c::get_nr_free_mft_records() and taking and
784 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
785 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
786 current userspace ntfs library code. This means that if a merge
787 fails the original runlists are always left unmodified instead of
788 being silently corrupted.
789 - Misc typo fixes.
790
7912.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
792
793 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
794 from ntfs_volume structure.
795 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
796 from ntfs_volume structure.
797 - Use iget5_locked() and friends instead of conventional iget(). Wrap
798 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
799 to use ntfs_iget(). Leave only one iget() call at mount time so we
800 don't need an ntfs_iget_mount().
801 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
802 additional argument.
803
8042.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
805
806 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
807 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
808 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
809 to determine whether to apply mst fixups or not.
810 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
811 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
812 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
813 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
814 the VFS readpage function prototype to the ntfs_attr_read_block()
815 function prototype.
816
8172.0.11 - Initial preparations for fake inode based attribute i/o.
818
819 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
820 do some macro magic (adapted from include/linux/buffer_head.h) to
821 expand all the helper functions NInoFoo(), NInoSetFoo(), and
822 NInoClearFoo().
823 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
824 - Add new fields to ntfs_inode structure to allow use of fake inodes
825 for attribute i/o: type, name, name_len. Also add new state bits:
826 NI_Attr, which, if set, indicates the inode is a fake inode, and
827 NI_MstProtected, which, if set, indicates the attribute uses multi
828 sector transfer protection, i.e. fixups need to be applied after
829 reads and before/after writes.
830 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
831 ntfs_{new,clear,destroy}_extent_inode() and update callers.
832 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
833 instead of ntfs_destroy_extent_inode().
834 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
835 - Make all operations on ntfs inode state bits use the NIno* functions.
836 - Set up the new ntfs inode fields and state bits in
837 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
838 allocated memory to __ntfs_clear_inode().
839 - Cleanup ntfs_inode structure a bit for better ordering of elements
840 w.r.t. their size to allow better packing of the structure in memory.
841
8422.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
843
844 - Add check at mount time to verify that the number of inodes on the
845 volume does not exceed 2^32 - 1, which is the maximum allowed for
846 NTFS according to Microsoft.
847 - Change mft_no member of ntfs_inode structure to be unsigned long.
848 Update all users. This makes ntfs_inode->mft_no just a copy of struct
849 inode->i_ino. But we can't just always use struct inode->i_ino and
850 remove mft_no because extent inodes do not have an attached struct
851 inode.
852
8532.0.9 - Decompression engine now uses a single buffer and other cleanups.
854
855 - Change decompression engine to use a single buffer protected by a
856 spin lock instead of per-CPU buffers. (Rusty Russell)
857 - Do not update cb_pos when handling a partial final page during
858 decompression of a sparse compression block, as the value is later
859 reset without being read/used. (Rusty Russell)
860 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
861 Morton)
862 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
863 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
864 it also makes everything safer so it is a good thing.
865 - Miscellaneous minor cleanups to comments.
866
8672.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
868
869 Big thanks go to Al Viro and other inhabitants of #kernel for investing
870 their time to discuss the case sensitivity and dcache aliasing issues.
871
872 - Remove unused source file fs/ntfs/attraops.c.
873 - Remove show_inodes mount option(s), thus dropping support for
874 displaying of short file names.
875 - Remove deprecated mount option posix.
876 - Restore show_sys_files mount option.
877 - Add new mount option case_sensitive, to determine if the driver
878 treats file names as case sensitive or not. If case sensitive, create
879 file names in the POSIX namespace. Otherwise create file names in the
880 LONG/WIN32 namespace. Note, files remain accessible via their short
881 file name, if it exists.
882 - Remove really dumb logic bug in boot sector recovery code.
883 - Fix dcache aliasing issues wrt short/long file names via changes
884 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
885 fs/ntfs/namei.c::ntfs_lookup():
886 - Add additional argument to ntfs_lookup_inode_by_name() in which we
887 return information about the matching file name if the case is not
888 matching or the match is a short file name. See comments above the
889 function definition for details.
890 - Change ntfs_lookup() to only create dcache entries for the correctly
891 cased file name and only for the WIN32 namespace counterpart of DOS
892 namespace file names. This ensures we have only one dentry per
893 directory and also removes all dcache aliasing issues between short
894 and long file names once we add write support. See comments above
895 function for details.
896 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
897
8982.0.7 - Minor cleanups and updates for changes in core kernel code.
899
900 - Remove much of the NULL struct element initializers.
901 - Various updates to make compatible with recent kernels.
902 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
903 in fs/ntfs/ntfs.h instead.
904 - Remove no longer needed KERNEL_VERSION checks. We are now in the
905 kernel proper so they are no longer needed.
906
9072.0.6 - Major bugfix to make compatible with other kernel changes.
908
909 - Initialize the mftbmp address space properly now that there are more
910 fields in the struct address_space. This was leading to hangs and
911 oopses on umount since 2.5.12 because of changes to other parts of
912 the kernel. We probably want a kernel generic init_address_space()
913 function...
914 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
915 only caller of ->readdir() is vfs_readdir() which holds i_sem during
916 the call, and i_sem is sufficient protection against changes in the
917 directory inode (including ->i_size).
918 - Use generic_file_llseek() for directories (as opposed to
919 default_llseek()) as this downs i_sem instead of the BKL which is
920 what we now need for exclusion against ->f_pos changes considering we
921 no longer take the BKL in ntfs_readdir().
922
9232.0.5 - Major bugfix. Buffer overflow in extent inode handling.
924
925 - No need to set old blocksize in super.c::ntfs_fill_super() as the
926 VFS does so via invocation of deactivate_super() calling
927 fs->fill_super() calling block_kill_super() which does it.
928 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
929 -> Do we really need it? I don't think so as we have exclusion on
930 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
931 move the ->f_pos accesses under the mrec_lock though. Check this...
932 - Fix really, really, really stupid buffer overflow in extent inode
933 handling in mft.c::map_extent_mft_record().
934
9352.0.4 - Cleanups and updates for kernel 2.5.11.
936
937 - Add documentation on how to use the MD driver to be able to use NTFS
938 stripe and volume sets in Linux and generally cleanup documentation
939 a bit.
940 Remove all uses of kdev_t in favour of struct block_device *:
941 - Change compress.c::ntfs_file_read_compressed_block() to use
942 sb_getblk() instead of getblk().
943 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
944 of get_hardsect_size().
945 - No need to get old blocksize in super.c::ntfs_fill_super() as
946 fs/super.c::get_sb_bdev() already does this.
947 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
948
9492.0.3 - Small bug fixes, cleanups, and performance improvements.
950
951 - Remove some dead code from mft.c.
952 - Optimize readpage and read_block functions throughout aops.c so that
953 only initialized blocks are read. Non-initialized ones have their
954 buffer head mapped, zeroed, and set up to date, without scheduling
955 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
956 Thanks go to Andrew Morton for spotting the below:
957 - Fix buglet in allocate_compression_buffers() error code path.
958 - Call flush_dcache_page() after modifying page cache page contents in
959 ntfs_file_readpage().
960 - Check for existence of page buffers throughout aops.c before calling
961 create_empty_buffers(). This happens when an I/O error occurs and the
962 read is retried. (It also happens once writing is implemented so that
963 needed doing anyway but I had left it for later...)
964 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
965 readpage and read_block functions. Reasoning same as above (i.e. I/O
966 error retries and future write code paths.)
967
9682.0.2 - Minor updates and cleanups.
969
970 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
971 and cleanup the code a bit, removing the unused size parameter.
972 - Change default fmask to 0177 and update documentation.
973 - Change attrib.c::get_attr_search_ctx() to return the search context
974 directly instead of taking the address of a pointer. A return value
975 of NULL means the allocation failed. Updated all callers
976 appropriately.
977 - Update to 2.5.9 kernel (preserving backwards compatibility) by
978 replacing all occurences of page->buffers with page_buffers(page).
979 - Fix minor bugs in runlist merging, also minor cleanup.
980 - Updates to bootsector layout and mft mirror contents descriptions.
981 - Small bug fix in error detection in unistr.c and some cleanups.
982 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
983 bytes.
984
9852.0.1 - Minor updates.
986
987 - Make default umask correspond to documentation.
988 - Improve documentation.
989 - Set default mode to include execute bit. The {u,f,d}mask can be used
990 to take it away if desired. This allows binaries to be executed from
991 a mounted ntfs partition.
992
9932.0.0 - New version number. Remove TNG from the name. Now in the kernel.
994
995 - Add kill_super, just keeping up with the vfs changes in the kernel.
996 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
997 from the CVS import into BitKeeper.
998 - Begin to implement proper handling of allocated_size vs
999 initialized_size vs data_size (i.e. i_size). Done are
1000 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
1001 and attrib.c::load_attribute_list().
1002 - Lock the runlist in attrib.c::load_attribute_list() while using it.
1003 - Fix memory leak in ntfs_file_read_compressed_block() and generally
1004 clean up compress.c a little, removing some uncommented/unused debug
1005 code.
1006 - Tidy up dir.c a little bit.
1007 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
1008 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
1009 creating aops.c::ntfs_mst_readpage(), improving the handling of
1010 holes and overflow in the process and implementing the correct
1011 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
1012 I am aiming for correctness at the moment. Modularisation can come
1013 later.
1014 - Rename aops.c::end_buffer_read_index_async() to
1015 end_buffer_read_mst_async() and optimize the overflow checking and
1016 handling.
1017 - Use the host of the mftbmp address space mapping to hold the ntfs
1018 volume. This is needed so the async i/o completion handler can
1019 retrieve a pointer to the volume. Hopefully this will not cause
1020 problems elsewhere in the kernel... Otherwise will need to use a
1021 fake inode.
1022 - Complete implementation of proper handling of allocated_size vs
1023 initialized_size vs data_size (i.e. i_size) in whole driver.
1024 Basically aops.c is now completely rewritten.
1025 - Change NTFS driver name to just NTFS and set version number to 2.0.0
1026 to make a clear distinction from the old driver which is still on
1027 version 1.1.22.
1028
1029tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
1030
1031 - Replace bdevname(sb->s_dev) with sb->s_id.
1032 - Remove now superfluous new-line characters in all callers of
1033 ntfs_debug().
1034 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
1035 directories. Without this the "find" utility gets very upset which is
1036 fair enough as Linux/Unix do not support directory hard links.
1037 - Further runlist merging work. (Richard Russon)
1038 - Backwards compatibility for gcc-2.95. (Richard Russon)
1039 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
1040 - Convert to new file system declaration using ->ntfs_get_sb() and
1041 replacing ntfs_read_super() with ntfs_fill_super().
1042 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
1043 overflow on 32-bit architectures.
1044 - Cleanup upcase loading code to use ntfs_(un)map_page().
1045 - Disable/reenable preemtion in critical sections of compession engine.
1046 - Replace device size determination in ntfs_fill_super() with
1047 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
1048 function super.c::get_nr_blocks().
1049 - Implement a mount time option (show_inodes) allowing choice of which
1050 types of inode names readdir() returns and modify ntfs_filldir()
1051 accordingly. There are several parameters to show_inodes:
1052 system: system files
1053 win32: long file names (including POSIX file names) [DEFAULT]
1054 long: same as win32
1055 dos: short file names only (excluding POSIX file names)
1056 short: same as dos
1057 posix: same as both win32 and dos
1058 all: all file names
1059 Note that the options are additive, i.e. specifying:
1060 -o show_inodes=system,show_inodes=win32,show_inodes=dos
1061 is the same as specifying:
1062 -o show_inodes=all
1063 Note that the "posix" and "all" options will show all directory
1064 names, BUT the link count on each directory inode entry is set to 1,
1065 due to Linux not supporting directory hard links. This may well
1066 confuse some userspace applications, since the directory names will
1067 have the same inode numbers. Thus it is NOT advisable to use the
1068 "posix" or "all" options. We provide them only for completeness sake.
1069 - Add copies of allocated_size, initialized_size, and compressed_size to
1070 the ntfs inode structure and set them up in
1071 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
1072 for files and the index allocation attribute for directories.
1073 - Add copies of allocated_size and initialized_size to ntfs inode for
1074 $BITMAP attribute of large directories and set them up in
1075 inode.c::ntfs_read_inode().
1076 - Add copies of allocated_size and initialized_size to ntfs volume for
1077 $BITMAP attribute of $MFT and set them up in
1078 super.c::load_system_files().
1079 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
1080 posix, and utf8) and tell user what the new options to use are. Note
1081 we still do support them but they will be removed with kernel 2.7.x.
1082 - Change all occurences of integer long long printf formatting to hex
1083 as printk() will not support long long integer format if/when the
1084 div64 patch goes into the kernel.
1085 - Make slab caches have stable names and change the names to what they
1086 were intended to be. These changes are required/made possible by the
1087 new slab cache name handling which removes the length limitation by
1088 requiring the caller of kmem_cache_create() to supply a stable name
1089 which is then referenced but not copied.
1090 - Rename run_list structure to run_list_element and create a new
1091 run_list structure containing a pointer to a run_list_element
1092 structure and a read/write semaphore. Adapt all users of runlists
1093 to new scheme and take and release the lock as needed. This fixes a
1094 nasty race as the run_list changes even when inodes are locked for
1095 reading and even when the inode isn't locked at all, so we really
1096 needed the serialization. We use a semaphore rather than a spinlock
1097 as memory allocations can sleep and doing everything GFP_ATOMIC
1098 would be silly.
1099 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
1100 This can never happen due to the nature of lookup_attr() and how we
1101 support attribute lists. If it did happen it would imply the inode
1102 being corrupt.
1103 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
1104 bad if found.
1105 - Update to 2.5.6-pre2 changes in struct address_space.
1106 - Use parent_ino() when accessing d_parent inode number in dir.c.
1107 - Import Sourceforge CVS repository into BitKeeper repository:
1108 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
1109 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
1110 Documentation/filesystems/ntfs.txt for NTFS TNG.
1111 - Create kernel configuration option controlling whether debugging
1112 is enabled or not.
1113 - Add the required export of end_buffer_io_sync() from the patches
1114 directory to the kernel code.
1115 - Update inode.c::ntfs_show_options() with show_inodes mount option.
1116 - Update errors mount option.
1117
1118tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
1119
1120 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
1121 bug in mapping of extent inodes. Update all the comments to fit all
1122 the recent code changes.
1123 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
1124 - Cleanups in compress.c, mostly comments and folding help.
1125 - Implement attrib.c::map_run_list() as a generic helper.
1126 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
1127 thus making code shorter and enabling attribute list support.
1128 - Cleanup incorrect use of [su]64 with %L printf format specifier in
1129 all source files. Type casts to [unsigned] long long added to correct
1130 the mismatches (important for architectures which have long long not
1131 being 64 bits).
1132 - Merge async io completion handlers for directory indexes and $MFT
1133 data into one by setting the index_block_size{_bits} of the ntfs
1134 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
1135 - Cleanup aops.c, update comments.
1136 - Make ntfs_file_get_block() use map_run_list() so all files now
1137 support attribute lists.
1138 - Make ntfs_dir_readpage() almost verbatim copy of
1139 block_read_full_page() by using ntfs_file_get_block() with only real
1140 difference being the use of our own async io completion handler
1141 rather than the default one, thus reducing the amount of code and
1142 automatically enabling attribute list support for directory indices.
1143 - Fix bug in load_attribute_list() - forgot to call brelse in error
1144 code path.
1145 - Change parameters to find_attr() and lookup_attr(). We no longer
1146 pass in the upcase table and its length. These can be gotten from
1147 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
1148 - Cleanups in attrib.c.
1149 - Implement merging of runlists, attrib.c::merge_run_lists() and its
1150 helpers. (Richard Russon)
1151 - Attribute lists part 2, attribute extents and multi part runlists:
1152 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
1153 further runlist parts via attrib.c::map_run_list().
1154 - Tiny endianness bug fix in decompress_mapping_pairs().
1155
1156tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
1157
1158 - Enable encrypted directories. (Their index root is marked encrypted
1159 to indicate that new files in that directory should be created
1160 encrypted.)
1161 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
1162 - Enable $Extend system directory. Most (if not all) extended system
1163 files do not have unnamed data attributes so ntfs_read_inode() had to
1164 special case them but that is ok, as the special casing recovery
1165 happens inside an error code path so there is zero slow down in the
1166 normal fast path. The special casing is done by introducing a new
1167 function inode.c::ntfs_is_extended_system_file() which checks if any
1168 of the hard links in the inode point to $Extend as being their parent
1169 directory and if they do we assume this is an extended system file.
1170 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
1171 when compiled with -DDEBUG. Default is debug messages to be disabled.
1172 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
1173 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
1174 interface is enabled). Inspired by old ntfs driver.
1175 - Add debug_msgs insmod/kernel boot parameter to set whether debug
1176 messages are {dis,en}abled. This is useful to enable debug messages
1177 during ntfs initialization and is the only way to activate debugging
1178 when the sysctl interface is not enabled.
1179 - Cleanup debug output in various places.
1180 - Remove all dollar signs ($) from the source (except comments) to
1181 enable compilation on architectures whose gcc compiler does not
1182 support dollar signs in the names of variables/constants. Attribute
1183 types now start with AT_ instead of $ and $I30 is now just I30.
1184 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
1185 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
1186 access functions. This means we now cope with $MFT/$BITMAP being
1187 spread accross several mft records.
1188 - Disable modification of mft_zone_multiplier on remount. We can always
1189 reenable this later on if we really want to, but we will need to make
1190 sure we readjust the mft_zone size / layout accordingly.
1191
1192tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
1193
1194 - Use sb_set_blocksize() instead of set_blocksize() and verify the
1195 return value.
1196 - Use sb_bread() instead of bread() throughout.
1197 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1198 of a directory index block vcn. Apply resulting simplifications in
1199 dir.c everywhere.
1200 - Fix a small bug somewhere (but forgot what it was).
1201 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1202 on the printf-format parameter list and fix bugs reported by gcc
1203 as a result. (Richard Russon)
1204 - Move inode allocation strategy to Al's new stuff but maintain the
1205 divorce of ntfs_inode from struct inode. To achieve this we have two
1206 separate slab caches, one for big ntfs inodes containing a struct
1207 inode and pure ntfs inodes and at the same time fix some faulty
1208 error code paths in ntfs_read_inode().
1209 - Show mount options in proc (inode.c::ntfs_show_options()).
1210
1211tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1212
1213 - Modified (un)map_mft_record functions to be common for read and write
1214 case. To specify which is which, added extra parameter at front of
1215 parameter list. Pass either READ or WRITE to this, each has the
1216 obvious meaning.
1217 - General cleanups to allow for easier folding in vi.
1218 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1219 argument, and invokes attrib.c::merge_run_lists() to merge the old
1220 and the new runlists.
1221 - Removed attrib.c::find_first_attr().
1222 - Implemented loading of attribute list and complete runlist for $MFT.
1223 This means we now cope with $MFT being spread across several mft
1224 records.
1225 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1226 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1227 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1228 cache. There are too little ntfs_volume structures at any one time
1229 to justify a private slab cache.
1230 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1231 Use KM_BIO_IRQ on advice from IRC/kernel...
1232 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1233 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1234 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1235 removed the now superfluous exports from the kernel core patch.
1236 - Fix a bug where kfree() was used insted of ntfs_free().
1237 - Change map_mft_record() to take ntfs_inode as argument instead of
1238 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1239 - Add pointer to ntfs_volume to ntfs_inode.
1240 - Add mft record number and sequence number to ntfs_inode. Stop using
1241 i_ino and i_generation for in-driver purposes.
1242 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1243 - Remove use of proper inodes by extent inodes. Move i_ino and
1244 i_generation to ntfs_inode to do this. Apply simplifications that
1245 result and remove iget_no_wait(), etc.
1246 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1247 - Add reference counting in ntfs_inode for the ntfs inode itself and
1248 for the mapped mft record.
1249 - Extend mft record mapping so we can (un)map extent mft records (new
1250 functions (un)map_extent_mft_record), and so mappings are reference
1251 counted and don't have to happen twice if already mapped - just ref
1252 count increases.
1253 - Add -o iocharset as alias to -o nls for backwards compatibility.
1254 - The latest core patch is now tiny. In fact just a single additional
1255 export is necessary over the base kernel.
1256
1257tng-0.0.3 - Cleanups, enhancements, bug fixes.
1258
1259 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1260 and setup the runlist appropriately using knowledge provided by the
1261 sizes in the base attribute record.
1262 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1263 any more.
1264 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1265 page or use vmalloc depending on the amount of memory requested.
1266 - Cleanup error output. The __FUNCTION__ "(): " is now added
1267 automatically. Introduced a new header file debug.h to support this
1268 and also moved ntfs_debug() function into it.
1269 - Make reading of compressed files more intelligent and especially get
1270 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1271 buffers (allocated at first mount with cluster size <= 4kiB and
1272 deallocated on last umount with cluster size <= 4kiB), and
1273 asynchronous io for the compressed data using a list of buffer heads.
1274 Er, we use synchronous io as async io only works on whole pages
1275 covered by buffers and not on individual buffer heads...
1276 - Bug fix for reading compressed files with sparse compression blocks.
1277
1278tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1279
1280 - Fixed handling of directories when cluster size exceeds index block
1281 size.
1282 - Hide DOS only name space directory entries from readdir() but allow
1283 them in lookup(). This should fix the problem that Linux doesn't
1284 support directory hard links, while still allowing access to entries
1285 via their short file name. This also has the benefit of mimicking
1286 what Windows users are used to, so it is the ideal solution.
1287 - Implemented sync_page everywhere so no more hangs in D state when
1288 waiting for a page.
1289 - Stop using bforget() in favour of brelse().
1290 - Stop locking buffers unnecessarily.
1291 - Implemented compressed files (inode->mapping contains uncompressed
1292 data, raw compressed data is currently bread() into a vmalloc()ed
1293 memory buffer).
1294 - Enable compressed directories. (Their index root is marked compressed
1295 to indicate that new files in that directory should be created
1296 compressed.)
1297 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1298 functions. (Thanks to Will Dyson for pointing this out.)
1299 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1300 ntfs_sb_info) out of the common inode and super_block structures and
1301 started using the generic_ip and generic_sbp pointers instead. This
1302 makes ntfs entirely private with respect to the kernel tree.
1303 - Detect compiler version and abort with error message if gcc less than
1304 2.96 is used.
1305 - Fix bug in name comparison function in unistr.c.
1306 - Implement attribute lists part 1, the infrastructure: search contexts
1307 and operations, find_external_attr(), lookup_attr()) and make the
1308 code use the infrastructure.
1309 - Fix stupid buffer overflow bug that became apparent on larger run
1310 list containing attributes.
1311 - Fix bugs in readdir() that became apparent on larger directories.
1312
1313 The driver is now really useful and survives the test
1314 find . -type f -exec md5sum "{}" \;
1315 without any error messages on a over 1GiB sized partition with >16k
1316 files on it, including compressed files and directories and many files
1317 and directories with attribute lists.
1318
1319tng-0.0.1 - The first useful version.
1320
1321 - Added ntfs_lookup().
1322 - Added default upcase generation and handling.
1323 - Added compile options to be shown on module init.
1324 - Many bug fixes that were "hidden" before.
1325 - Update to latest kernel.
1326 - Added ntfs_readdir().
1327 - Added file operations for mmap(), read(), open() and llseek(). We just
1328 use the generic ones. The whole point of going through implementing
1329 readpage() methods and where possible get_block() call backs is that
1330 this allows us to make use of the generic high level methods provided
1331 by the kernel.
1332
1333 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1334 though and it doesn't implement accesssing compressed files yet. Also,
1335 accessing files with attribute list attributes is not implemented yet
1336 either. But for small or simple file systems it should work and allow
1337 you to list directories, use stat on directory entries and the file
1338 system, open, read, mmap and llseek around in files. A big mile stone
1339 has been reached!
1340
1341tng-0.0.0 - Initial version tag.
1342
1343 Initial driver implementation. The driver can mount and umount simple
1344 NTFS file systems (i.e. ones without attribute lists in the system
1345 files). If the mount fails there might be problems in the error handling
1346 code paths, so be warned. Otherwise it seems to be loading the system
1347 files nicely and the mft record read mapping/unmapping seems to be
1348 working nicely, too. Proof of inode metadata in the page cache and non-
1349 resident file unnamed stream data in the page cache concepts is thus
1350 complete.