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