3 - In between ntfs_prepare/commit_write, need exclusion between
4 simultaneous file extensions. This is given to us by holding i_sem
5 on the inode. The only places in the kernel when a file is resized
6 are prepare/commit write and truncate for both of which i_sem is
7 held. Just have to be careful in readpage/writepage and all other
8 helpers not running under i_sem that we play nice...
9 Also need to be careful with initialized_size extention in
10 ntfs_prepare_write. Basically, just be _very_ careful in this code...
11 UPDATE: The only things that need to be checked are read/writepage
12 which do not hold i_sem. Note writepage cannot change i_size but it
13 needs to cope with a concurrent i_size change, just like readpage.
14 Also both need to cope with concurrent changes to the other sizes,
15 i.e. initialized/allocated/compressed size, as well.
16 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
17 leave the volume dirty on umount if the final iput(vol->mft_ino)
18 causes a write of any mirrored mft records due to the mft mirror
19 inode having been discarded already. Whether this can actually ever
20 happen is unclear however so it is worth waiting until someone hits
22 - Enable the code for setting the NT4 compatibility flag when we start
23 making NTFS 1.2 specific modifications.
25 2.1.24 - Lots of bug fixes and support more clean journal states.
27 - Support journals ($LogFile) which have been modified by chkdsk. This
28 means users can boot into Windows after we marked the volume dirty.
29 The Windows boot will run chkdsk and then reboot. The user can then
30 immediately boot into Linux rather than having to do a full Windows
31 boot first before rebooting into Linux and we will recognize such a
32 journal and empty it as it is clean by definition. Note, this only
33 works if chkdsk left the journal in an obviously clean state.
34 - Support journals ($LogFile) with only one restart page as well as
35 journals with two different restart pages. We sanity check both and
36 either use the only sane one or the more recent one of the two in the
37 case that both are valid.
38 - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
39 ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
41 - Use ntfs_malloc_nofs_nofail() in the two critical regions in
42 fs/ntfs/runlist.c::ntfs_runlists_merge(). This means we no longer
43 need to panic() if the allocation fails as it now cannot fail.
44 - Fix two nasty runlist merging bugs that had gone unnoticed so far.
45 Thanks to Stefano Picerno for the bug report.
46 - Remove two bogus BUG_ON()s from fs/ntfs/mft.c.
47 - Fix handling of valid but empty mapping pairs array in
48 fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress().
49 - Report unrepresentable inodes during ntfs_readdir() as KERN_WARNING
50 messages and include the inode number. Thanks to Yura Pakhuchiy for
52 - Change ntfs_rl_truncate_nolock() to throw away the runlist if the new
54 - Add runlist.[hc]::ntfs_rl_punch_nolock() which punches a caller
55 specified hole into a runlist.
56 - Fix a bug in fs/ntfs/index.c::ntfs_index_lookup(). When the returned
57 index entry is in the index root, we forgot to set the @ir pointer in
58 the index context. Thanks to Yura Pakhuchiy for finding this bug.
59 - Remove bogus setting of PageError in ntfs_read_compressed_block().
60 - Add fs/ntfs/attrib.[hc]::ntfs_resident_attr_value_resize().
61 - Fix a bug in ntfs_map_runlist_nolock() where we forgot to protect
62 access to the allocated size in the ntfs inode with the size lock.
63 - Fix ntfs_attr_vcn_to_lcn_nolock() and ntfs_attr_find_vcn_nolock() to
64 return LCN_ENOENT when there is no runlist and the allocated size is
66 - Fix load_attribute_list() to handle the case of a NULL runlist.
67 - Fix handling of sparse attributes in ntfs_attr_make_non_resident().
68 - Add BUG() checks to ntfs_attr_make_non_resident() and ntfs_attr_set()
69 to ensure that these functions are never called for compressed or
71 - Fix cluster (de)allocators to work when the runlist is NULL and more
72 importantly to take a locked runlist rather than them locking it
73 which leads to lock reversal.
74 - Truncate {a,c,m}time to the ntfs supported time granularity when
75 updating the times in the inode in ntfs_setattr().
76 - Fixup handling of sparse, compressed, and encrypted attributes in
77 fs/ntfs/inode.c::ntfs_read_locked_{,attr_,index_}inode(),
78 fs/ntfs/aops.c::ntfs_{read,write}page().
79 - Make ntfs_write_block() not instantiate sparse blocks if they contain
81 - Optimize fs/ntfs/aops.c::ntfs_write_block() by extending the page
82 lock protection over the buffer submission for i/o which allows the
83 removal of the get_bh()/put_bh() pairs for each buffer.
84 - Fix fs/ntfs/aops.c::ntfs_{read,write}_block() to handle the case
85 where a concurrent truncate has truncated the runlist under our feet.
86 - Fix page_has_buffers()/page_buffers() handling in fs/ntfs/aops.c.
87 - In fs/ntfs/aops.c::ntfs_end_buffer_async_read(), use a bit spin lock
88 in the first buffer head instead of a driver global spin lock to
90 - Minor fix to error handling and error message display in
91 fs/ntfs/aops.c::ntfs_prepare_nonresident_write().
92 - Change the mount options {u,f,d}mask to always parse the number as
93 an octal number to conform to how chmod(1) works, too. Thanks to
94 Giuseppe Bilotta and Horst von Brand for pointing out the errors of
96 - Fix various bugs in the runlist merging code. (Based on libntfs
97 changes by Richard Russon.)
98 - Fix sparse warnings that have crept in over time.
99 - Change ntfs_cluster_free() to require a write locked runlist on entry
100 since we otherwise get into a lock reversal deadlock if a read locked
101 runlist is passed in. In the process also change it to take an ntfs
102 inode instead of a vfs inode as parameter.
103 - Fix the definition of the CHKD ntfs record magic. It had an off by
104 two error causing it to be CHKB instead of CHKD.
106 2.1.23 - Implement extension of resident files and make writing safe as well as
107 many bug fixes, cleanups, and enhancements...
109 - Add printk rate limiting for ntfs_warning() and ntfs_error() when
110 compiled without debug. This avoids a possible denial of service
111 attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
113 - Fix compilation warnings on ia64. (Randy Dunlap)
114 - Use i_size_{read,write}() instead of reading i_size by hand and cache
115 the value where apropriate.
116 - Add size_lock to the ntfs_inode structure. This is an rw spinlock
117 and it locks against access to the inode sizes. Note, ->size_lock
118 is also accessed from irq context so you must use the _irqsave and
119 _irqrestore lock and unlock functions, respectively. Protect all
120 accesses to allocated_size, initialized_size, and compressed_size.
121 - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
122 - Implement extension of resident files in the regular file write code
123 paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()). At present
124 this only works until the data attribute becomes too big for the mft
125 record after which we abort the write returning -EOPNOTSUPP from
126 ntfs_prepare_write().
127 - Add disable_sparse mount option together with a per volume sparse
128 enable bit which is set appropriately and a per inode sparse disable
129 bit which is preset on some system file inodes as appropriate.
130 - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
131 - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
132 the creation of the unmapped runlist element for the base attribute
134 - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
135 helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
136 This allows us to map runlist fragments with the runlist lock already
137 held without having to drop and reacquire it around the call. Adapt
139 - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
140 runlist. This allows us to find runlist elements with the runlist
141 lock already held without having to drop and reacquire it around the
142 call. Adapt all callers.
143 - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
144 warning in the do_div() call on sparc32. Thanks to Meelis Roos for
145 the report and analysis of the warning.
146 - Fix a nasty runlist merge bug when merging two holes.
147 - Set the ntfs_inode->allocated_size to the real allocated size in the
148 mft record for resident attributes (fs/ntfs/inode.c).
149 - Small readability cleanup to use "a" instead of "ctx->attr"
150 everywhere (fs/ntfs/inode.c).
151 - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
152 definition of ntfs_export_ops from fs/ntfs/super.c to namei.c. Also,
153 declare ntfs_export_ops in fs/ntfs/ntfs.h.
154 - Correct sparse file handling. The compressed values need to be
155 checked and set in the ntfs inode as done for compressed files and
156 the compressed size needs to be used for vfs inode->i_blocks instead
157 of the allocated size, again, as done for compressed files.
158 - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
159 non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
160 - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
162 - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
163 dropping the read lock and taking the write lock we were not checking
164 whether someone else did not already do the work we wanted to do.
165 - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
166 ntfs_attr_find_vcn_nolock() and update all callers.
167 - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
168 - Fix sign of various error return values to be negative in
170 - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
171 handle the case where an attribute is converted from resident to
172 non-resident by a concurrent file write.
173 - Remove checks for NULL before calling kfree() since kfree() does the
174 checking itself. (Jesper Juhl)
175 - Some utilities modify the boot sector but do not update the checksum.
176 Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
177 only emit a warning when the checksum is incorrect rather than
178 refusing the mount. Thanks to Bernd Casimir for pointing this
180 - Update attribute definition handling.
181 - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
182 - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
183 - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
184 better code generation and one less sparse warning in fs/ntfs/aops.c.
185 - Remove spurious void pointer casts from fs/ntfs/. (Pekka Enberg)
186 - Use C99 style structure initialization after memory allocation where
187 possible (fs/ntfs/{attrib.c,index.c,super.c}). Thanks to Al Viro and
189 - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
190 is active on the volume and we are mounting read-write or remounting
191 from read-only to read-write.
192 - Fix a bug in address space operations error recovery code paths where
193 if the runlist was not mapped at all and a mapping error occured we
194 would leave the runlist locked on exit to the function so that the
195 next access to the same file would try to take the lock and deadlock.
196 - Detect the case when Windows has been suspended to disk on the volume
197 to be mounted and if this is the case do not allow (re)mounting
198 read-write. This is done by parsing hiberfil.sys if present.
199 - Fix several occurences of a bug where we would perform 'var & ~const'
200 with a 64-bit variable and a int, i.e. 32-bit, constant. This causes
201 the higher order 32-bits of the 64-bit variable to be zeroed. To fix
202 this cast the 'const' to the same 64-bit type as 'var'.
203 - Change the runlist terminator of the newly allocated cluster(s) to
204 LCN_ENOENT in ntfs_attr_make_non_resident(). Otherwise the runlist
206 - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
207 and ntfs_mapping_pairs_build() to allow the runlist encoding to be
208 partial which is desirable when filling holes in sparse attributes.
210 - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
211 if the requested vcn is inside it. Otherwise we get into problems
212 when we try to map an out of bounds vcn because we then try to map
213 the already mapped runlist fragment which causes
214 ntfs_mapping_pairs_decompress() to fail and return error. Update
215 ntfs_attr_find_vcn_nolock() accordingly.
216 - Fix a nasty deadlock that appeared in recent kernels.
217 The situation: VFS inode X on a mounted ntfs volume is dirty. For
218 same inode X, the ntfs_inode is dirty and thus corresponding on-disk
219 inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
220 to the table of inodes, i.e. $MFT, inode 0.
222 Process 1: sys_sync()/umount()/whatever... calls
223 __sync_single_inode() for $MFT -> do_writepages() -> write_page for
224 the dirty page containing the on-disk inode X, the page is now locked
225 -> ntfs_write_mst_block() which clears PageUptodate() on the page to
226 prevent anyone else getting hold of it whilst it does the write out.
227 This is necessary as the on-disk inode needs "fixups" applied before
228 the write to disk which are removed again after the write and
229 PageUptodate is then set again. It then analyses the page looking
230 for dirty on-disk inodes and when it finds one it calls
231 ntfs_may_write_mft_record() to see if it is safe to write this
232 on-disk inode. This then calls ilookup5() to check if the
233 corresponding VFS inode is in icache(). This in turn calls ifind()
234 which waits on the inode lock via wait_on_inode whilst holding the
236 Process 2: pdflush results in a call to __sync_single_inode for the
237 same VFS inode X on the ntfs volume. This locks the inode (I_LOCK)
238 then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
239 read_cache_page() for the page (in page cache of table of inodes
240 $MFT, inode 0) containing the on-disk inode. This page has
241 PageUptodate() clear because of Process 1 (see above) so
242 read_cache_page() blocks when it tries to take the page lock for the
243 page so it can call ntfs_read_page().
244 Thus Process 1 is holding the page lock on the page containing the
245 on-disk inode X and it is waiting on the inode X to be unlocked in
246 ifind() so it can write the page out and then unlock the page.
247 And Process 2 is holding the inode lock on inode X and is waiting for
248 the page to be unlocked so it can call ntfs_readpage() or discover
249 that Process 1 set PageUptodate() again and use the page.
250 Thus we have a deadlock due to ifind() waiting on the inode lock.
251 The solution: The fix is to use the newly introduced
252 ilookup5_nowait() which does not wait on the inode's lock and hence
253 avoids the deadlock. This is safe as we do not care about the VFS
254 inode and only use the fact that it is in the VFS inode cache and the
255 fact that the vfs and ntfs inodes are one struct in memory to find
256 the ntfs inode in memory if present. Also, the ntfs inode has its
257 own locking so it does not matter if the vfs inode is locked.
258 - Fix bug in mft record writing where we forgot to set the device in
259 the buffers when mapping them after the VM had discarded them.
260 Thanks to Martin MOKREJŠ for the bug report.
262 2.1.22 - Many bug and race fixes and error handling improvements.
264 - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
265 - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
266 instead of void and provide a helper ntfs_truncate_vfs() for the
267 vfs ->truncate method.
268 - Add a new ntfs inode flag NInoTruncateFailed() and modify
269 fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
270 - Fix min_size and max_size definitions in ATTR_DEF structure in
271 fs/ntfs/layout.h to be signed.
272 - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
273 ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
274 ntfs_attr_can_be_resident(), which in turn use the new private helper
275 ntfs_attr_find_in_attrdef().
276 - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
277 mapping->private_lock around the dirtying of the buffer heads
278 analagous to the way it is done in __set_page_dirty_buffers().
279 - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
280 mount time as this cannot work with the current implementation.
281 - Check for location of attribute name and improve error handling in
282 general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
283 - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
284 i_size, i.e. race with truncate, invalidate the buffers on the page
285 so that they become freeable and hence the page does not leak.
286 - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
288 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
289 a NULL pointer dereference in the error code path when a corrupt
290 attribute was found. (Thanks to Domen Puncer for the bug report.)
291 - Add MODULE_VERSION() to fs/ntfs/super.c.
292 - Make several functions and variables static. (Adrian Bunk)
293 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
294 buffers for the page if they are not present and then marks the
295 buffers belonging to the ntfs record dirty. This causes the buffers
296 to become busy and hence they are safe from removal until the page
297 has been written out.
298 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
299 error handling code path that resulted in a BUG() due to trying to
300 unmap an extent mft record when the mapping of it had failed and it
301 thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
302 - Drop the runlist lock after the vcn has been read in
303 fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
304 - Rewrite handling of multi sector transfer errors. We now do not set
305 PageError() when such errors are detected in the async i/o handler
306 fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
307 protected attributes now check the magic of each ntfs record as they
308 use it and act appropriately. This has the effect of making errors
309 granular per ntfs record rather than per page which solves the case
310 where we cannot access any of the ntfs records in a page when a
311 single one of them had an mst error. (Thanks to Ken MacFerrin for
313 - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
314 where we failed to release i_sem on the $Quota/$Q attribute inode.
315 - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
316 - Add mapping of unmapped buffers to all remaining code paths, i.e.
317 fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
318 and write_mft_record_nolock(). From now on we require that the
319 complete runlist for the mft mirror is always mapped into memory.
320 - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
321 - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
322 - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
323 resident attribute will be smaller than a page which makes the code
324 simpler. Also make the code more tolerant to concurrent ->truncate.
326 2.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
328 - Implement extent mft record deallocation
329 fs/ntfs/mft.c::ntfs_extent_mft_record_free().
330 - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
331 - Add vol->mft_data_pos and initialize it at mount time.
332 - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
333 ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
334 ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
335 ntfs_runlists_merge() and adapt all callers.
336 - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
337 ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
338 and ntfs_mapping_pairs_build(), adapted from libntfs.
339 - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
340 static and add a declaration for it to lcnalloc.h.
341 - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
342 inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
343 cluster bitmap lock for the duration of the call.
344 - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
345 - Implement the equivalent of memset() for an ntfs attribute in
346 fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
347 fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
348 - Remove unnecessary casts from LCN_* constants.
349 - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
350 - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
351 change MFT_RECORD to contain the NTFS 3.1+ specific fields.
352 - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
353 marks all buffers belonging to an ntfs record dirty, followed by
354 marking the page the ntfs record is in dirty and also marking the vfs
355 inode containing the ntfs record dirty (I_DIRTY_PAGES).
356 - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
357 new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
358 longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
359 - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
361 - Move the typedefs for runlist_element and runlist from types.h to
362 runlist.h and fix resulting include errors.
363 - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
364 - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
365 mark_ntfs_record_dirty() which also changes the behaviour in that we
366 now set the buffers belonging to the mft record dirty as well as the
368 - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
369 to cope with the fact that there now are dirty buffers in mft pages.
370 - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
371 mark_ntfs_record_dirty() and thus to set the buffers belonging to the
372 mft record dirty as well as the page itself.
373 - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
374 slightly modified by me)
375 - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
376 the mft record is already locked and otherwise behaves the same way
377 as fs/ntfs/mft.c::map_mft_record().
378 - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
379 writes the mft record if the buffers belonging to it are dirty.
380 Otherwise we assume that it was written out by other means already.
381 - Attempting to write outside initialized size is _not_ a bug so remove
382 the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
383 fact required to write outside initialized size when preparing to
384 extend the initialized size.
385 - Map the page instead of using page_address() before writing to it in
386 fs/ntfs/aops.c::ntfs_mft_writepage().
387 - Provide exclusion between opening an inode / mapping an mft record
388 and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
389 by setting the page not uptodate throughout ntfs_mft_writepage().
390 - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
391 to ensure noone can see the page whilst the mst fixups are applied.
392 - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
393 checks if an mft record may be written out safely obtaining any
394 necessary locks in the process. This is used by
395 fs/ntfs/aops.c::ntfs_write_mst_block().
396 - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
397 writing mft records and improve its error handling in the process.
398 Now if any of the records in the page fail to be written out, all
399 other records will be written out instead of aborting completely.
400 - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
401 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
402 ntfs_mst_aops for all inodes which are NInoMstProtected() and
403 ntfs_aops for all other inodes.
404 - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
405 ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
406 no longer require an ntfs inode to be present. Update all callers.
407 - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
408 - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
409 to ensure noone can see the page whilst the mst fixups are applied.
410 - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
411 fs/ntfs/mft.c::try_map_mft_record().
412 - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
413 with the ntfs inode which contains the page rather than the ntfs
414 inode the mft record of which is in the page.
415 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
416 index inode bitmap inode release code from there to
417 fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
418 Hellwig for spotting this.)
419 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
420 inode semaphore around the code that sets ni->itype.index.bmp_ino to
421 NULL and reorganize the code to optimize it a bit. (Thanks to
422 Christoph Hellwig for spotting this.)
423 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
424 ntfs inode as a parameter as this is confusing and misleading and the
425 needed ntfs inode is available via NTFS_I(page->mapping->host).
426 Adapt all callers to this change.
427 - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
428 fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
429 of the first buffer in a record and to take this as the ntfs record
430 dirty state. We cannot look at the dirty state for subsequent
431 buffers because we might be racing with
432 fs/ntfs/aops.c::mark_ntfs_record_dirty().
433 - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
434 inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
435 add a declaration for it to inode.h. Fix some compilation issues
436 that resulted due to #includes and header file interdependencies.
437 - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
438 - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
439 - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
440 record sequence number if it is specified (i.e. not zero).
441 - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
442 functions used by it.
443 - Update Documentation/filesystems/ntfs.txt with instructions on how to
444 use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
445 the linear raid problem with the Software RAID / MD driver when one
446 or more of the devices has an odd number of sectors.
448 2.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
450 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
451 where we did not clear ctx->al_entry but it was still set due to
452 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
454 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
455 where we forgot to unmap the extent mft record when we had finished
456 enumerating an attribute which caused a bug check to trigger when the
457 VFS calls ->clear_inode.
459 2.1.19 - Many cleanups, improvements, and a minor bug fix.
461 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
462 change the uid, gid, and mode of an inode as we do not support NTFS
464 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
466 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
467 and ntfs_filldir() as per suggestion from Al Viro.
468 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
469 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
470 inode size has changed and to only output an error if so.
471 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
472 - Add le{16,32,64} as well as sle{16,32,64} data types to
474 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
475 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
476 respectively, to fs/ntfs/types.h.
477 - Update endianness conversion macros in fs/ntfs/endian.h to use the
478 new types as appropriate.
479 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
481 - Add leMFT_REF data type to fs/ntfs/layout.h.
482 - Update all NTFS header files with the new little endian data types.
483 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
484 - Do proper type casting when using ntfs_is_*_recordp() in
485 fs/ntfs/logfile.c, mft.c, and super.c.
486 - Fix all the sparse bitwise warnings. Had to change all the typedef
487 enums storing little endian values to simple enums plus a typedef for
488 the datatype to make sparse happy.
489 - Fix a bug found by the new sparse bitwise warnings where the default
490 upcase table was defined as a pointer to wchar_t rather than ntfschar
491 in fs/ntfs/ntfs.h and super.c.
492 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
494 2.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
496 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
497 the calculation of total/free inodes as used by statfs().
498 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
499 because the code itself is using the ntfs_lock semaphore which
500 provides safe locking. (Ingo Molnar)
501 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
502 could occur in the future for when we start closing/freeing extent
503 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
505 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
506 find_external_attr() to ntfs_external_attr_find() to cleanup the
507 namespace a bit and to be more consistent with libntfs.
508 - Rename {{re,}init,get,put}_attr_search_ctx() to
509 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
510 attr_search_context to ntfs_attr_search_ctx.
511 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
512 for the attribute list attribute itself.
513 - Fix endianness bug in ntfs_external_attr_find().
514 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
515 if the attribute is not found, and -EIO on real error. In the case
516 of -ENOENT, the search context is updated to describe the attribute
517 before which the attribute being searched for would need to be
518 inserted if such an action were to be desired and in the case of
519 ntfs_external_attr_find() the search context is also updated to
520 indicate the attribute list entry before which the attribute list
521 entry of the attribute being searched for would need to be inserted
522 if such an action were to be desired. Also make ntfs_find_attr()
523 static and remove its prototype from attrib.h as it is not used
524 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
525 callers of ntfs_{external,}attr_{find,lookup}() for the new return
527 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
529 2.1.17 - Fix bugs in mount time error code paths and other updates.
531 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
532 includes functions to set/clear a single bit or a run of bits.
533 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
534 runlist element containing a particular vcn. It also takes care of
535 mapping any needed runlist fragments.
536 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
537 - Load attribute definition table from $AttrDef at mount time.
538 - Fix bugs in mount time error code paths involving (de)allocation of
539 the default and volume upcase tables.
540 - Remove ntfs_nr_mounts as it is no longer used.
542 2.1.16 - Implement access time updates, file sync, async io, and read/writev.
544 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
545 This is done by setting the appropriate file operations pointers to
546 the generic helper functions provided by mm/filemap.c.
547 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
548 and directories (fs/ntfs/dir.c).
549 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
550 Note, except for the root directory and any other system files opened
551 by the user, the system files will not have their access times
552 updated as they are only accessed at the inode level an hence the
553 file level functions which cause the times to be updated are never
556 2.1.15 - Invalidate quotas when (re)mounting read-write.
558 - Add new element itype.index.collation_rule to the ntfs inode
559 structure and set it appropriately in ntfs_read_locked_inode().
560 - Implement a new inode type "index" to allow efficient access to the
561 indices found in various system files and adapt inode handling
562 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
563 attribute inode (NInoAttr() is true) with an attribute type of
564 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
565 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
566 there would be no way to distinguish between normal attribute inodes
567 and index inodes. The function to obtain an index inode is
568 ntfs_index_iget() and it uses the helper function
569 ntfs_read_locked_index_inode(). Note, we do not overload
570 ntfs_attr_iget() as indices consist of multiple attributes so using
571 ntfs_attr_iget() to obtain an index inode would be confusing.
572 - Ensure that there is no overflow when doing page->index <<
573 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
574 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
575 and ntfs_read_block().
576 - Use case sensitive attribute lookups instead of case insensitive ones.
577 - Lock all page cache pages belonging to mst protected attributes while
578 accessing them to ensure we never see corrupt data while the page is
580 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
581 We have ntfs_is_collation_rule_supported() to check if the collation
582 rule you want to use is supported and ntfs_collation() which actually
583 collates two data items. We currently only support COLLATION_BINARY
584 and COLLATION_NTOFS_ULONG but support for other collation rules will
585 be added as the need arises.
586 - Add a new type, ntfs_index_context, to allow retrieval of an index
587 entry using the corresponding index key. To get an index context,
588 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
589 This also adds a new slab cache for the index contexts. To lookup a
590 key in an index inode, use ntfs_index_lookup(). After modifying an
591 index entry, call ntfs_index_entry_flush_dcache_page() followed by
592 ntfs_index_entry_mark_dirty() to ensure the changes are written out
593 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
594 an index entry is in the index allocation attribute rather than the
595 index root attribute it will not be written out (you will get a
596 warning message about discarded changes instead).
597 - Load the quota file ($Quota) and check if quota tracking is enabled
598 and if so, mark the quotas out of date. This causes windows to
599 rescan the volume on boot and update all quota entries.
600 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
601 It is simply set to __set_page_dirty_nobuffers() to make sure that
602 running set_page_dirty() on a page containing mft/ntfs records will
603 not affect the dirty state of the page buffers.
604 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
605 buffers that are inside the ntfs record in the page dirty after which
606 it sets the page dirty. This allows ->writepage to only write the
607 dirty index records rather than having to write all the records in
608 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
609 use this rather than __set_page_dirty_nobuffers().
610 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
611 writing of page cache pages belonging to mst protected attributes
612 like the index allocation attribute in directory indices and other
613 indices like $Quota/$Q, etc. This means that the quota is now marked
614 out of date on all volumes rather than only on ones where the quota
615 defaults entry is in the index root attribute of the $Quota/$Q index.
617 2.1.14 - Fix an NFSd caused deadlock reported by several users.
619 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
620 to a buffer so that we can put the search context and unmap the mft
621 record before calling the filldir() callback. We need to do this
622 because of NFSd which calls ->lookup() from its filldir callback()
623 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
624 of the directory and since ntfs_readdir() has got it mapped already
625 ntfs_lookup() deadlocks.
627 2.1.13 - Enable overwriting of resident files and housekeeping of system files.
629 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
630 keeping the mft mirror in sync with the mft when mirrored mft records
631 are written. The functions are write_mft_record{,_nolock}(). The
632 implementation is quite rudimentary for now with lots of things not
633 implemented yet but I am not sure any of them can actually occur so
634 I will wait for people to hit each one and only then implement it.
635 - Commit open system inodes at umount time. This should make it
636 virtually impossible for sync_mft_mirror_umount() to ever be needed.
637 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
638 ntfs super operations. This gives us inode writing via the VFS inode
639 dirty code paths. Note: Access time updates are not implemented yet.
640 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
641 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
642 finally enabling resident file overwrite! (-8 This also includes a
643 placeholder for ->writepage (ntfs_mft_writepage()), which for now
644 just redirties the page and returns. Also, at umount time, we for
645 now throw away all mft data page cache pages after the last call to
646 ntfs_commit_inode() in the hope that all inodes will have been
647 written out by then and hence no dirty (meta)data will be lost. We
648 also check for this case and emit an error message telling the user
650 - Use set_page_writeback() and end_page_writeback() in the resident
651 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
652 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
654 - Implement ntfs_mft_writepage() so it now checks if any of the mft
655 records in the page are dirty and if so redirties the page and
656 returns. Otherwise it just returns (after doing set_page_writeback(),
657 unlock_page(), end_page_writeback() or the radix-tree tag
658 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
659 alowing the VM to do with the page as it pleases. Also, at umount
660 time, now only throw away dirty mft (meta)data pages if dirty inodes
661 are present and ask the user to email us if they see this happening.
662 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
663 information flags (fs/ntfs/super.c).
664 - Mark the volume dirty when (re)mounting read-write and mark it clean
665 when unmounting or remounting read-only. If any volume errors are
666 found, the volume is left marked dirty to force chkdsk to run.
667 - Add code to set the NT4 compatibility flag when (re)mounting
668 read-write for newer NTFS versions but leave it commented out for now
669 since we do not make any modifications that are NTFS 1.2 specific yet
670 and since setting this flag breaks Captive-NTFS which is not nice.
671 This code must be enabled once we start writing NTFS 1.2 specific
672 changes otherwise Windows NTFS driver might crash / cause corruption.
674 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
676 - Add a new address space operations struct, ntfs_mst_aops, for mst
677 protected attributes. This is because the default ntfs_aops do not
678 make sense with mst protected data and were they to write anything to
679 such an attribute they would cause data corruption so we provide
680 ntfs_mst_aops which does not have any write related operations set.
681 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
682 includes an adapted ntfs_commit_inode() and an implementation of
683 ntfs_write_inode() which for now just cleans dirty inodes without
684 writing them (it does emit a warning that this is happening).
685 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
686 entry) as it was only fixing a theoretical bug but at the same time
687 it badly broke the handling of sparse and uncompressed compression
690 2.1.11 - Driver internal cleanups.
692 - Only build logfile.o if building the driver with read-write support.
693 - Really final white space cleanups.
694 - Use generic_ffs() instead of ffs() in logfile.c which allows the
695 log_page_size variable to be optimized by gcc into a constant.
696 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
697 char as defined by POSIX and as found on some systems.
699 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
701 - Finish off the white space cleanups (remove trailing spaces, etc).
702 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
703 the kludges around the first iget(). Instead of (re)setting ->s_op
704 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
705 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
706 kills the need for second super_operations and allows to return error
707 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
709 - Force read-only (re)mounting if any of the following bits are set in
710 the volume information flags:
711 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
712 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
713 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
714 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
715 above bits set so the test is made easy.
717 2.1.9 - Fix two bugs in decompression engine.
719 - Fix a bug where we would not always detect that we have reached the
720 end of a compression block because we were ending at minus one byte
721 which is effectively the same as being at the end. The fix is to
722 check whether the uncompressed buffer has been fully filled and if so
723 we assume we have reached the end of the compression block. A big
724 thank you to Marcin Gibuła for the bug report, the assistance in
725 tracking down the bug and testing the fix.
726 - Fix a possible bug where when a compressed read is truncated to the
727 end of the file, the offset inside the last page was not truncated.
729 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
731 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
732 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
733 utc2ntfs() to work with struct timespec instead of time_t on the
734 Linux UTC time side thus preserving the full precision of the NTFS
735 time and only loosing up to 99 nano-seconds in the Linux UTC time.
736 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
738 - Remove unused ntfs_dirty_inode().
739 - Cleanup super operations declaration in fs/ntfs/super.c.
740 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
741 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
742 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
743 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
744 fs/ntfs/inode.h so they can be used elsewhere.
745 - Determine the mft mirror size as the number of mirrored mft records
746 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
747 - Load the mft mirror at mount time and compare the mft records stored
748 in it to the ones in the mft. Force a read-only mount if the two do
749 not match (fs/ntfs/super.c).
750 - Fix type casting related warnings on 64-bit architectures. Thanks
751 to Meelis Roos for reporting them.
752 - Move %L to %ll as %L is floating point and %ll is integer which is
754 - Read the journal ($LogFile) and determine if the volume has been
755 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
756 and fs/ntfs/logfile.c). This is a little bit of a crude check in
757 that we only look at the restart areas and not at the actual log
758 records so that there will be a very small number of cases where we
759 think that a volume is dirty when in fact it is clean. This should
760 only affect volumes that have not been shutdown cleanly and did not
761 have any pending, non-check-pointed i/o.
762 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
763 is requested, empty $LogFile by overwriting it with 0xff bytes to
764 ensure that Windows cannot cause data corruption by replaying a stale
765 journal after Linux has written to the volume.
767 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
769 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
770 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
771 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
772 default doesn't allow inode number 0 which is a valid inode on NTFS
773 and even if it did allow that it uses iget() instead of ntfs_iget()
774 which makes it useless for us.
775 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
776 default just returns -EACCES which is not very useful.
777 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
778 and set them up in the super block at mount time (super.c) this
779 allows mounted NTFS volumes to be exported via NFS.
780 - Add missing return -EOPNOTSUPP; in
781 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
782 - Enforce no atime and no dir atime updates at mount/remount time as
783 they are not implemented yet anyway.
784 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
785 after a NULL check. Thanks to Dave Jones for pointing this out.
787 2.1.6 - Fix minor bug in handling of compressed directories.
789 - Fix bug in handling of compressed directories. A compressed
790 directory is not really compressed so when we set the ->i_blocks
791 field of a compressed directory inode we were setting it from the
792 non-existing field ni->itype.compressed.size which gave random
793 results... For directories we now always use ni->allocated_size.
795 2.1.5 - Fix minor bug in attribute list attribute handling.
797 - Fix bug in attribute list handling. Actually it is not as much a bug
798 as too much protection in that we were not allowing attribute lists
799 which waste space on disk while Windows XP clearly allows it and in
800 fact creates such attribute lists so our driver was failing.
801 - Update NTFS documentation ready for 2.6 kernel release.
803 2.1.4 - Reduce compiler requirements.
805 - Remove all uses of unnamed structs and unions in the driver to make
806 old and newer gcc versions happy. Makes it a bit uglier IMO but at
807 least people will stop hassling me about it.
809 2.1.3 - Important bug fixes in corner cases.
811 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
812 clusters. (Philipp Thomas)
813 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
814 multiple of the block_size but not the cluster size. (Szabolcs
815 Szakacsits <szaka@sienet.hu>)
817 2.1.2 - Important bug fixes aleviating the hangs in statfs.
819 - Fix buggy free cluster and free inode determination logic.
821 2.1.1 - Minor updates.
823 - Add handling for initialized_size != data_size in compressed files.
824 - Reduce function local stack usage from 0x3d4 bytes to just noise in
825 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
826 - Remove compiler warnings for newer gcc.
827 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
828 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
829 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
830 kmap_atomic(KM_USER0).
832 2.1.0 - First steps towards write support: implement file overwrite.
834 - Add configuration option for developmental write support with an
835 appropriately scary configuration help text.
836 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
837 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
838 overwriting of existing files on ntfs. Note: Resident files are
839 only written into memory, and not written out to disk at present, so
840 avoid writing to files smaller than about 1kiB.
841 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
842 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
843 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
844 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
845 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
846 This enables write(2) based overwriting of existing files on ntfs.
847 Note: As with mmap(2) based overwriting, resident files are only
848 written into memory, and not written out to disk at present, so avoid
849 writing to files smaller than about 1kiB.
850 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
851 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
852 files with the purpose of intercepting and aborting all i_size
853 changes which we do not support yet. ntfs_truncate() actually only
854 emits a warning message but AFAICS our interception of i_size changes
855 elsewhere means ntfs_truncate() never gets called for i_size changes.
856 It is only called from generic_file_write() when we fail in
857 ntfs_prepare_{,nonresident_}write() in order to discard any
858 instantiated buffers beyond i_size. Thus i_size is not actually
859 changed so our warning message is enough. Unfortunately it is not
860 possible to easily determine if i_size is being changed or not hence
861 we just emit an appropriately worded error message.
863 2.0.25 - Small bug fixes and cleanups.
865 - Unlock the page in an out of memory error code path in
866 fs/ntfs/aops.c::ntfs_read_block().
867 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
868 just unlock the page and return. (This can happen due to ->writepage
869 clearing PageUptodate() during write out of MstProtected()
871 - Remove leaked write code again.
875 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
876 inside BUG_ON(). (Adam J. Richter)
877 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
878 calls for improved debugging. (Adam J. Richter)
879 - Add errors flag to the ntfs volume state, accessed via
880 NVol{,Set,Clear}Errors(vol).
881 - Do not allow read-write remounts of read-only volumes with errors.
882 - Clarify comment for ntfs file operation sendfile which was added by
883 Christoph Hellwig a while ago (just using generic_file_sendfile())
884 to say that ntfs ->sendfile is only used for the case where the
885 source data is on the ntfs partition and the destination is
886 somewhere else, i.e. nothing we need to concern ourselves with.
887 - Add generic_file_write() as our ntfs file write operation.
889 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
891 - Massive internal locking changes to mft record locking. Fixes lock
892 recursion and replaces the mrec_lock read/write semaphore with a
893 mutex. Also removes the now superfluous mft_count. This fixes several
894 race conditions and deadlocks, especially in the future write code.
895 - Fix ntfs over loopback for compressed files by adding an
896 optimization barrier. (gcc was screwing up otherwise ?)
897 - Miscellaneous cleanups all over the code and a fix or two in error
899 Thanks go to Christoph Hellwig for pointing out the following two:
900 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
901 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
903 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
905 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
906 at entry/exit respectively.
907 - Use C99 initializers for structures.
908 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
910 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
912 - Limit volume size at mount time to 2TiB on architectures where
913 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
914 This is the most we can do without overflowing the 32-bit limit of
915 the block device size imposed on us by sb_bread() and sb_getblk()
917 - Limit file/directory size at open() time to 16TiB on architectures
918 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
919 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
920 overflowing the page cache page index.
922 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
924 - Move the directory index bitmap to use an attribute inode instead of
925 having special fields for it inside the ntfs inode structure. This
926 means that the index bitmaps now use the page cache for i/o, too,
927 and also as a side effect we get support for non-resident index
929 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
930 fix a page leak that manifested itself in some cases.
931 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
932 index bitmap inode on the final iput().
934 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
936 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
937 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
938 - Drop the "file" from ntfs_file_read_compressed_block().
939 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
940 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
941 - Update ntfs_end_buffer_async_read() with the improved logic from
942 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
943 further logic improvements to better determine when we set PageError.
944 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
945 check for the buffers being uptodate first in line with the updated
946 fs/buffer.c::block_read_full_page(). This plugs a small race
949 2.0.18 - Fix race condition in reading of compressed files.
951 - There was a narrow window between checking a buffer head for being
952 uptodate and locking it in ntfs_file_read_compressed_block(). We now
953 lock the buffer and then check whether it is uptodate or not.
955 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
957 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
958 code and update callers, i.e. ntfs_iget(), to pass that error code
959 up instead of just using -EIO.
960 - Modifications to super.c to ensure that both mount and remount
961 cannot set any write related options when the driver is compiled
963 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
964 cache the current runlist element. This should improve performance
965 when reading very large and/or very fragmented data.
967 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
969 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
970 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
971 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
972 - Convert $MFT/$BITMAP access to attribute inode API and remove all
973 remnants of the ugly mftbmp address space and operations hack. This
974 means we finally have only one readpage function as well as only one
975 async io completion handler. Yey! The mft bitmap is now just an
976 attribute inode and is accessed from vol->mftbmp_ino just as if it
977 were a normal file. Fake inodes rule. (-:
979 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
981 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
982 remounts to fail when the partition had an entry in /etc/fstab and
983 the entry specified the nls= option.
984 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
985 expand all the helper functions NVolFoo(), NVolSetFoo(), and
987 - Move copyright statement from driver initialisation message to
988 module description (fs/super.c). This makes the initialisation
989 message fit on one line and fits in better with rest of kernel.
990 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
991 attribute inodes, and both for files and directories.
992 - Implement fake attribute inodes allowing all attribute i/o to go via
993 the page cache and to use all the normal vfs/mm functionality:
994 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
996 - Add needed cleanup code to ntfs_clear_big_inode().
997 - Merge address space operations for files and directories (aops.c),
998 now just have ntfs_aops:
1000 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
1001 ntfs_attr_read_block() -> ntfs_read_block(),
1002 ntfs_file_read_page() -> ntfs_readpage().
1003 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
1004 attribute inodes, and both for files and directories.
1005 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
1007 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
1009 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
1010 the locking out of super.c::get_nr_free_mft_records() and taking and
1011 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
1012 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
1013 current userspace ntfs library code. This means that if a merge
1014 fails the original runlists are always left unmodified instead of
1015 being silently corrupted.
1018 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
1020 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
1021 from ntfs_volume structure.
1022 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
1023 from ntfs_volume structure.
1024 - Use iget5_locked() and friends instead of conventional iget(). Wrap
1025 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
1026 to use ntfs_iget(). Leave only one iget() call at mount time so we
1027 don't need an ntfs_iget_mount().
1028 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
1029 additional argument.
1031 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
1033 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
1034 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
1035 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
1036 to determine whether to apply mst fixups or not.
1037 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
1038 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
1039 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
1040 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
1041 the VFS readpage function prototype to the ntfs_attr_read_block()
1044 2.0.11 - Initial preparations for fake inode based attribute i/o.
1046 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
1047 do some macro magic (adapted from include/linux/buffer_head.h) to
1048 expand all the helper functions NInoFoo(), NInoSetFoo(), and
1050 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
1051 - Add new fields to ntfs_inode structure to allow use of fake inodes
1052 for attribute i/o: type, name, name_len. Also add new state bits:
1053 NI_Attr, which, if set, indicates the inode is a fake inode, and
1054 NI_MstProtected, which, if set, indicates the attribute uses multi
1055 sector transfer protection, i.e. fixups need to be applied after
1056 reads and before/after writes.
1057 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
1058 ntfs_{new,clear,destroy}_extent_inode() and update callers.
1059 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
1060 instead of ntfs_destroy_extent_inode().
1061 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
1062 - Make all operations on ntfs inode state bits use the NIno* functions.
1063 - Set up the new ntfs inode fields and state bits in
1064 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
1065 allocated memory to __ntfs_clear_inode().
1066 - Cleanup ntfs_inode structure a bit for better ordering of elements
1067 w.r.t. their size to allow better packing of the structure in memory.
1069 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
1071 - Add check at mount time to verify that the number of inodes on the
1072 volume does not exceed 2^32 - 1, which is the maximum allowed for
1073 NTFS according to Microsoft.
1074 - Change mft_no member of ntfs_inode structure to be unsigned long.
1075 Update all users. This makes ntfs_inode->mft_no just a copy of struct
1076 inode->i_ino. But we can't just always use struct inode->i_ino and
1077 remove mft_no because extent inodes do not have an attached struct
1080 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
1082 - Change decompression engine to use a single buffer protected by a
1083 spin lock instead of per-CPU buffers. (Rusty Russell)
1084 - Do not update cb_pos when handling a partial final page during
1085 decompression of a sparse compression block, as the value is later
1086 reset without being read/used. (Rusty Russell)
1087 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
1089 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
1090 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
1091 it also makes everything safer so it is a good thing.
1092 - Miscellaneous minor cleanups to comments.
1094 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
1096 Big thanks go to Al Viro and other inhabitants of #kernel for investing
1097 their time to discuss the case sensitivity and dcache aliasing issues.
1099 - Remove unused source file fs/ntfs/attraops.c.
1100 - Remove show_inodes mount option(s), thus dropping support for
1101 displaying of short file names.
1102 - Remove deprecated mount option posix.
1103 - Restore show_sys_files mount option.
1104 - Add new mount option case_sensitive, to determine if the driver
1105 treats file names as case sensitive or not. If case sensitive, create
1106 file names in the POSIX namespace. Otherwise create file names in the
1107 LONG/WIN32 namespace. Note, files remain accessible via their short
1108 file name, if it exists.
1109 - Remove really dumb logic bug in boot sector recovery code.
1110 - Fix dcache aliasing issues wrt short/long file names via changes
1111 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
1112 fs/ntfs/namei.c::ntfs_lookup():
1113 - Add additional argument to ntfs_lookup_inode_by_name() in which we
1114 return information about the matching file name if the case is not
1115 matching or the match is a short file name. See comments above the
1116 function definition for details.
1117 - Change ntfs_lookup() to only create dcache entries for the correctly
1118 cased file name and only for the WIN32 namespace counterpart of DOS
1119 namespace file names. This ensures we have only one dentry per
1120 directory and also removes all dcache aliasing issues between short
1121 and long file names once we add write support. See comments above
1122 function for details.
1123 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
1125 2.0.7 - Minor cleanups and updates for changes in core kernel code.
1127 - Remove much of the NULL struct element initializers.
1128 - Various updates to make compatible with recent kernels.
1129 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
1130 in fs/ntfs/ntfs.h instead.
1131 - Remove no longer needed KERNEL_VERSION checks. We are now in the
1132 kernel proper so they are no longer needed.
1134 2.0.6 - Major bugfix to make compatible with other kernel changes.
1136 - Initialize the mftbmp address space properly now that there are more
1137 fields in the struct address_space. This was leading to hangs and
1138 oopses on umount since 2.5.12 because of changes to other parts of
1139 the kernel. We probably want a kernel generic init_address_space()
1141 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
1142 only caller of ->readdir() is vfs_readdir() which holds i_sem during
1143 the call, and i_sem is sufficient protection against changes in the
1144 directory inode (including ->i_size).
1145 - Use generic_file_llseek() for directories (as opposed to
1146 default_llseek()) as this downs i_sem instead of the BKL which is
1147 what we now need for exclusion against ->f_pos changes considering we
1148 no longer take the BKL in ntfs_readdir().
1150 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
1152 - No need to set old blocksize in super.c::ntfs_fill_super() as the
1153 VFS does so via invocation of deactivate_super() calling
1154 fs->fill_super() calling block_kill_super() which does it.
1155 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
1156 -> Do we really need it? I don't think so as we have exclusion on
1157 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
1158 move the ->f_pos accesses under the mrec_lock though. Check this...
1159 - Fix really, really, really stupid buffer overflow in extent inode
1160 handling in mft.c::map_extent_mft_record().
1162 2.0.4 - Cleanups and updates for kernel 2.5.11.
1164 - Add documentation on how to use the MD driver to be able to use NTFS
1165 stripe and volume sets in Linux and generally cleanup documentation
1167 Remove all uses of kdev_t in favour of struct block_device *:
1168 - Change compress.c::ntfs_file_read_compressed_block() to use
1169 sb_getblk() instead of getblk().
1170 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
1171 of get_hardsect_size().
1172 - No need to get old blocksize in super.c::ntfs_fill_super() as
1173 fs/super.c::get_sb_bdev() already does this.
1174 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
1176 2.0.3 - Small bug fixes, cleanups, and performance improvements.
1178 - Remove some dead code from mft.c.
1179 - Optimize readpage and read_block functions throughout aops.c so that
1180 only initialized blocks are read. Non-initialized ones have their
1181 buffer head mapped, zeroed, and set up to date, without scheduling
1182 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
1183 Thanks go to Andrew Morton for spotting the below:
1184 - Fix buglet in allocate_compression_buffers() error code path.
1185 - Call flush_dcache_page() after modifying page cache page contents in
1186 ntfs_file_readpage().
1187 - Check for existence of page buffers throughout aops.c before calling
1188 create_empty_buffers(). This happens when an I/O error occurs and the
1189 read is retried. (It also happens once writing is implemented so that
1190 needed doing anyway but I had left it for later...)
1191 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
1192 readpage and read_block functions. Reasoning same as above (i.e. I/O
1193 error retries and future write code paths.)
1195 2.0.2 - Minor updates and cleanups.
1197 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
1198 and cleanup the code a bit, removing the unused size parameter.
1199 - Change default fmask to 0177 and update documentation.
1200 - Change attrib.c::get_attr_search_ctx() to return the search context
1201 directly instead of taking the address of a pointer. A return value
1202 of NULL means the allocation failed. Updated all callers
1204 - Update to 2.5.9 kernel (preserving backwards compatibility) by
1205 replacing all occurences of page->buffers with page_buffers(page).
1206 - Fix minor bugs in runlist merging, also minor cleanup.
1207 - Updates to bootsector layout and mft mirror contents descriptions.
1208 - Small bug fix in error detection in unistr.c and some cleanups.
1209 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
1212 2.0.1 - Minor updates.
1214 - Make default umask correspond to documentation.
1215 - Improve documentation.
1216 - Set default mode to include execute bit. The {u,f,d}mask can be used
1217 to take it away if desired. This allows binaries to be executed from
1218 a mounted ntfs partition.
1220 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
1222 - Add kill_super, just keeping up with the vfs changes in the kernel.
1223 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
1224 from the CVS import into BitKeeper.
1225 - Begin to implement proper handling of allocated_size vs
1226 initialized_size vs data_size (i.e. i_size). Done are
1227 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
1228 and attrib.c::load_attribute_list().
1229 - Lock the runlist in attrib.c::load_attribute_list() while using it.
1230 - Fix memory leak in ntfs_file_read_compressed_block() and generally
1231 clean up compress.c a little, removing some uncommented/unused debug
1233 - Tidy up dir.c a little bit.
1234 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
1235 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
1236 creating aops.c::ntfs_mst_readpage(), improving the handling of
1237 holes and overflow in the process and implementing the correct
1238 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
1239 I am aiming for correctness at the moment. Modularisation can come
1241 - Rename aops.c::end_buffer_read_index_async() to
1242 end_buffer_read_mst_async() and optimize the overflow checking and
1244 - Use the host of the mftbmp address space mapping to hold the ntfs
1245 volume. This is needed so the async i/o completion handler can
1246 retrieve a pointer to the volume. Hopefully this will not cause
1247 problems elsewhere in the kernel... Otherwise will need to use a
1249 - Complete implementation of proper handling of allocated_size vs
1250 initialized_size vs data_size (i.e. i_size) in whole driver.
1251 Basically aops.c is now completely rewritten.
1252 - Change NTFS driver name to just NTFS and set version number to 2.0.0
1253 to make a clear distinction from the old driver which is still on
1256 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
1258 - Replace bdevname(sb->s_dev) with sb->s_id.
1259 - Remove now superfluous new-line characters in all callers of
1261 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
1262 directories. Without this the "find" utility gets very upset which is
1263 fair enough as Linux/Unix do not support directory hard links.
1264 - Further runlist merging work. (Richard Russon)
1265 - Backwards compatibility for gcc-2.95. (Richard Russon)
1266 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
1267 - Convert to new filesystem declaration using ->ntfs_get_sb() and
1268 replacing ntfs_read_super() with ntfs_fill_super().
1269 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
1270 overflow on 32-bit architectures.
1271 - Cleanup upcase loading code to use ntfs_(un)map_page().
1272 - Disable/reenable preemtion in critical sections of compession engine.
1273 - Replace device size determination in ntfs_fill_super() with
1274 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
1275 function super.c::get_nr_blocks().
1276 - Implement a mount time option (show_inodes) allowing choice of which
1277 types of inode names readdir() returns and modify ntfs_filldir()
1278 accordingly. There are several parameters to show_inodes:
1279 system: system files
1280 win32: long file names (including POSIX file names) [DEFAULT]
1282 dos: short file names only (excluding POSIX file names)
1284 posix: same as both win32 and dos
1286 Note that the options are additive, i.e. specifying:
1287 -o show_inodes=system,show_inodes=win32,show_inodes=dos
1288 is the same as specifying:
1290 Note that the "posix" and "all" options will show all directory
1291 names, BUT the link count on each directory inode entry is set to 1,
1292 due to Linux not supporting directory hard links. This may well
1293 confuse some userspace applications, since the directory names will
1294 have the same inode numbers. Thus it is NOT advisable to use the
1295 "posix" or "all" options. We provide them only for completeness sake.
1296 - Add copies of allocated_size, initialized_size, and compressed_size to
1297 the ntfs inode structure and set them up in
1298 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
1299 for files and the index allocation attribute for directories.
1300 - Add copies of allocated_size and initialized_size to ntfs inode for
1301 $BITMAP attribute of large directories and set them up in
1302 inode.c::ntfs_read_inode().
1303 - Add copies of allocated_size and initialized_size to ntfs volume for
1304 $BITMAP attribute of $MFT and set them up in
1305 super.c::load_system_files().
1306 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
1307 posix, and utf8) and tell user what the new options to use are. Note
1308 we still do support them but they will be removed with kernel 2.7.x.
1309 - Change all occurences of integer long long printf formatting to hex
1310 as printk() will not support long long integer format if/when the
1311 div64 patch goes into the kernel.
1312 - Make slab caches have stable names and change the names to what they
1313 were intended to be. These changes are required/made possible by the
1314 new slab cache name handling which removes the length limitation by
1315 requiring the caller of kmem_cache_create() to supply a stable name
1316 which is then referenced but not copied.
1317 - Rename run_list structure to run_list_element and create a new
1318 run_list structure containing a pointer to a run_list_element
1319 structure and a read/write semaphore. Adapt all users of runlists
1320 to new scheme and take and release the lock as needed. This fixes a
1321 nasty race as the run_list changes even when inodes are locked for
1322 reading and even when the inode isn't locked at all, so we really
1323 needed the serialization. We use a semaphore rather than a spinlock
1324 as memory allocations can sleep and doing everything GFP_ATOMIC
1326 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
1327 This can never happen due to the nature of lookup_attr() and how we
1328 support attribute lists. If it did happen it would imply the inode
1330 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
1332 - Update to 2.5.6-pre2 changes in struct address_space.
1333 - Use parent_ino() when accessing d_parent inode number in dir.c.
1334 - Import Sourceforge CVS repository into BitKeeper repository:
1335 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
1336 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
1337 Documentation/filesystems/ntfs.txt for NTFS TNG.
1338 - Create kernel configuration option controlling whether debugging
1340 - Add the required export of end_buffer_io_sync() from the patches
1341 directory to the kernel code.
1342 - Update inode.c::ntfs_show_options() with show_inodes mount option.
1343 - Update errors mount option.
1345 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
1347 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
1348 bug in mapping of extent inodes. Update all the comments to fit all
1349 the recent code changes.
1350 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
1351 - Cleanups in compress.c, mostly comments and folding help.
1352 - Implement attrib.c::map_run_list() as a generic helper.
1353 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
1354 thus making code shorter and enabling attribute list support.
1355 - Cleanup incorrect use of [su]64 with %L printf format specifier in
1356 all source files. Type casts to [unsigned] long long added to correct
1357 the mismatches (important for architectures which have long long not
1359 - Merge async io completion handlers for directory indexes and $MFT
1360 data into one by setting the index_block_size{_bits} of the ntfs
1361 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
1362 - Cleanup aops.c, update comments.
1363 - Make ntfs_file_get_block() use map_run_list() so all files now
1364 support attribute lists.
1365 - Make ntfs_dir_readpage() almost verbatim copy of
1366 block_read_full_page() by using ntfs_file_get_block() with only real
1367 difference being the use of our own async io completion handler
1368 rather than the default one, thus reducing the amount of code and
1369 automatically enabling attribute list support for directory indices.
1370 - Fix bug in load_attribute_list() - forgot to call brelse in error
1372 - Change parameters to find_attr() and lookup_attr(). We no longer
1373 pass in the upcase table and its length. These can be gotten from
1374 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
1375 - Cleanups in attrib.c.
1376 - Implement merging of runlists, attrib.c::merge_run_lists() and its
1377 helpers. (Richard Russon)
1378 - Attribute lists part 2, attribute extents and multi part runlists:
1379 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
1380 further runlist parts via attrib.c::map_run_list().
1381 - Tiny endianness bug fix in decompress_mapping_pairs().
1383 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
1385 - Enable encrypted directories. (Their index root is marked encrypted
1386 to indicate that new files in that directory should be created
1388 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
1389 - Enable $Extend system directory. Most (if not all) extended system
1390 files do not have unnamed data attributes so ntfs_read_inode() had to
1391 special case them but that is ok, as the special casing recovery
1392 happens inside an error code path so there is zero slow down in the
1393 normal fast path. The special casing is done by introducing a new
1394 function inode.c::ntfs_is_extended_system_file() which checks if any
1395 of the hard links in the inode point to $Extend as being their parent
1396 directory and if they do we assume this is an extended system file.
1397 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
1398 when compiled with -DDEBUG. Default is debug messages to be disabled.
1399 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
1400 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
1401 interface is enabled). Inspired by old ntfs driver.
1402 - Add debug_msgs insmod/kernel boot parameter to set whether debug
1403 messages are {dis,en}abled. This is useful to enable debug messages
1404 during ntfs initialization and is the only way to activate debugging
1405 when the sysctl interface is not enabled.
1406 - Cleanup debug output in various places.
1407 - Remove all dollar signs ($) from the source (except comments) to
1408 enable compilation on architectures whose gcc compiler does not
1409 support dollar signs in the names of variables/constants. Attribute
1410 types now start with AT_ instead of $ and $I30 is now just I30.
1411 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
1412 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
1413 access functions. This means we now cope with $MFT/$BITMAP being
1414 spread accross several mft records.
1415 - Disable modification of mft_zone_multiplier on remount. We can always
1416 reenable this later on if we really want to, but we will need to make
1417 sure we readjust the mft_zone size / layout accordingly.
1419 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
1421 - Use sb_set_blocksize() instead of set_blocksize() and verify the
1423 - Use sb_bread() instead of bread() throughout.
1424 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1425 of a directory index block vcn. Apply resulting simplifications in
1427 - Fix a small bug somewhere (but forgot what it was).
1428 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1429 on the printf-format parameter list and fix bugs reported by gcc
1430 as a result. (Richard Russon)
1431 - Move inode allocation strategy to Al's new stuff but maintain the
1432 divorce of ntfs_inode from struct inode. To achieve this we have two
1433 separate slab caches, one for big ntfs inodes containing a struct
1434 inode and pure ntfs inodes and at the same time fix some faulty
1435 error code paths in ntfs_read_inode().
1436 - Show mount options in proc (inode.c::ntfs_show_options()).
1438 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1440 - Modified (un)map_mft_record functions to be common for read and write
1441 case. To specify which is which, added extra parameter at front of
1442 parameter list. Pass either READ or WRITE to this, each has the
1444 - General cleanups to allow for easier folding in vi.
1445 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1446 argument, and invokes attrib.c::merge_run_lists() to merge the old
1447 and the new runlists.
1448 - Removed attrib.c::find_first_attr().
1449 - Implemented loading of attribute list and complete runlist for $MFT.
1450 This means we now cope with $MFT being spread across several mft
1452 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1453 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1454 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1455 cache. There are too little ntfs_volume structures at any one time
1456 to justify a private slab cache.
1457 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1458 Use KM_BIO_IRQ on advice from IRC/kernel...
1459 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1460 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1461 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1462 removed the now superfluous exports from the kernel core patch.
1463 - Fix a bug where kfree() was used insted of ntfs_free().
1464 - Change map_mft_record() to take ntfs_inode as argument instead of
1465 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1466 - Add pointer to ntfs_volume to ntfs_inode.
1467 - Add mft record number and sequence number to ntfs_inode. Stop using
1468 i_ino and i_generation for in-driver purposes.
1469 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1470 - Remove use of proper inodes by extent inodes. Move i_ino and
1471 i_generation to ntfs_inode to do this. Apply simplifications that
1472 result and remove iget_no_wait(), etc.
1473 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1474 - Add reference counting in ntfs_inode for the ntfs inode itself and
1475 for the mapped mft record.
1476 - Extend mft record mapping so we can (un)map extent mft records (new
1477 functions (un)map_extent_mft_record), and so mappings are reference
1478 counted and don't have to happen twice if already mapped - just ref
1480 - Add -o iocharset as alias to -o nls for backwards compatibility.
1481 - The latest core patch is now tiny. In fact just a single additional
1482 export is necessary over the base kernel.
1484 tng-0.0.3 - Cleanups, enhancements, bug fixes.
1486 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1487 and setup the runlist appropriately using knowledge provided by the
1488 sizes in the base attribute record.
1489 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1491 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1492 page or use vmalloc depending on the amount of memory requested.
1493 - Cleanup error output. The __FUNCTION__ "(): " is now added
1494 automatically. Introduced a new header file debug.h to support this
1495 and also moved ntfs_debug() function into it.
1496 - Make reading of compressed files more intelligent and especially get
1497 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1498 buffers (allocated at first mount with cluster size <= 4kiB and
1499 deallocated on last umount with cluster size <= 4kiB), and
1500 asynchronous io for the compressed data using a list of buffer heads.
1501 Er, we use synchronous io as async io only works on whole pages
1502 covered by buffers and not on individual buffer heads...
1503 - Bug fix for reading compressed files with sparse compression blocks.
1505 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1507 - Fixed handling of directories when cluster size exceeds index block
1509 - Hide DOS only name space directory entries from readdir() but allow
1510 them in lookup(). This should fix the problem that Linux doesn't
1511 support directory hard links, while still allowing access to entries
1512 via their short file name. This also has the benefit of mimicking
1513 what Windows users are used to, so it is the ideal solution.
1514 - Implemented sync_page everywhere so no more hangs in D state when
1516 - Stop using bforget() in favour of brelse().
1517 - Stop locking buffers unnecessarily.
1518 - Implemented compressed files (inode->mapping contains uncompressed
1519 data, raw compressed data is currently bread() into a vmalloc()ed
1521 - Enable compressed directories. (Their index root is marked compressed
1522 to indicate that new files in that directory should be created
1524 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1525 functions. (Thanks to Will Dyson for pointing this out.)
1526 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1527 ntfs_sb_info) out of the common inode and super_block structures and
1528 started using the generic_ip and generic_sbp pointers instead. This
1529 makes ntfs entirely private with respect to the kernel tree.
1530 - Detect compiler version and abort with error message if gcc less than
1532 - Fix bug in name comparison function in unistr.c.
1533 - Implement attribute lists part 1, the infrastructure: search contexts
1534 and operations, find_external_attr(), lookup_attr()) and make the
1535 code use the infrastructure.
1536 - Fix stupid buffer overflow bug that became apparent on larger run
1537 list containing attributes.
1538 - Fix bugs in readdir() that became apparent on larger directories.
1540 The driver is now really useful and survives the test
1541 find . -type f -exec md5sum "{}" \;
1542 without any error messages on a over 1GiB sized partition with >16k
1543 files on it, including compressed files and directories and many files
1544 and directories with attribute lists.
1546 tng-0.0.1 - The first useful version.
1548 - Added ntfs_lookup().
1549 - Added default upcase generation and handling.
1550 - Added compile options to be shown on module init.
1551 - Many bug fixes that were "hidden" before.
1552 - Update to latest kernel.
1553 - Added ntfs_readdir().
1554 - Added file operations for mmap(), read(), open() and llseek(). We just
1555 use the generic ones. The whole point of going through implementing
1556 readpage() methods and where possible get_block() call backs is that
1557 this allows us to make use of the generic high level methods provided
1560 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1561 though and it doesn't implement accesssing compressed files yet. Also,
1562 accessing files with attribute list attributes is not implemented yet
1563 either. But for small or simple filesystems it should work and allow
1564 you to list directories, use stat on directory entries and the file
1565 system, open, read, mmap and llseek around in files. A big mile stone
1568 tng-0.0.0 - Initial version tag.
1570 Initial driver implementation. The driver can mount and umount simple
1571 NTFS filesystems (i.e. ones without attribute lists in the system
1572 files). If the mount fails there might be problems in the error handling
1573 code paths, so be warned. Otherwise it seems to be loading the system
1574 files nicely and the mft record read mapping/unmapping seems to be
1575 working nicely, too. Proof of inode metadata in the page cache and non-
1576 resident file unnamed stream data in the page cache concepts is thus
Code Review / linux-2.6.git / blob