3 - The only places in the kernel where a file is resized are
4 ntfs_file_write*() and ntfs_truncate() for both of which i_mutex is
5 held. Just have to be careful in read-/writepage and other helpers
6 not running under i_mutex that we play nice. Also need to be careful
7 with initialized_size extension in ntfs_file_write*() and writepage.
8 UPDATE: The only things that need to be checked are the compressed
9 write and the other attribute resize/write cases like index
10 attributes, etc. For now none of these are implemented so are safe.
11 - Implement filling in of holes in aops.c::ntfs_writepage() and its
13 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
14 leave the volume dirty on umount if the final iput(vol->mft_ino)
15 causes a write of any mirrored mft records due to the mft mirror
16 inode having been discarded already. Whether this can actually ever
17 happen is unclear however so it is worth waiting until someone hits
19 - Enable the code for setting the NT4 compatibility flag when we start
20 making NTFS 1.2 specific modifications.
22 2.1.27 - Various bug fixes and cleanups.
24 - Fix two compiler warnings on Alpha. Thanks to Andrew Morton for
26 - Fix an (innocent) off-by-one error in the runlist code.
27 - Fix a buggette in an "should be impossible" case handling where we
28 continued the attribute lookup loop instead of aborting it.
29 - Use buffer_migrate_page() for the ->migratepage function of all ntfs
30 address space operations.
32 2.1.26 - Minor bug fixes and updates.
34 - Fix a potential overflow in file.c where a cast to s64 was missing in
35 a left shift of a page index.
36 - The struct inode has had its i_sem semaphore changed to a mutex named
38 - We have struct kmem_cache now so use it instead of the typedef
39 kmem_cache_t. (Pekka Enberg)
40 - Implement support for sector sizes above 512 bytes (up to the maximum
41 supported by NTFS which is 4096 bytes).
42 - Do more detailed reporting of why we cannot mount read-write by
43 special casing the VOLUME_MODIFIED_BY_CHKDSK flag.
44 - Miscellaneous updates to layout.h.
45 - Cope with attribute list attribute having invalid flags. Windows
46 copes with this and even chkdsk does not detect or fix this so we
47 have to cope with it, too. Thanks to Pawel Kot for reporting the
50 2.1.25 - (Almost) fully implement write(2) and truncate(2).
52 - Change ntfs_map_runlist_nolock(), ntfs_attr_find_vcn_nolock() and
53 {__,}ntfs_cluster_free() to also take an optional attribute search
54 context as argument. This allows calling these functions with the
55 mft record mapped. Update all callers.
56 - Fix potential deadlock in ntfs_mft_data_extend_allocation_nolock()
57 error handling by passing in the active search context when calling
59 - Change ntfs_cluster_alloc() to take an extra boolean parameter
60 specifying whether the cluster are being allocated to extend an
61 attribute or to fill a hole.
62 - Change ntfs_attr_make_non_resident() to call ntfs_cluster_alloc()
63 with @is_extension set to TRUE and remove the runlist terminator
64 fixup code as this is now done by ntfs_cluster_alloc().
65 - Change ntfs_attr_make_non_resident to take the attribute value size
66 as an extra parameter. This is needed since we need to know the size
67 before we can map the mft record and our callers always know it. The
68 reason we cannot simply read the size from the vfs inode i_size is
69 that this is not necessarily uptodate. This happens when
70 ntfs_attr_make_non_resident() is called in the ->truncate call path.
71 - Fix ntfs_attr_make_non_resident() to update the vfs inode i_blocks
72 which is zero for a resident attribute but should no longer be zero
73 once the attribute is non-resident as it then has real clusters
75 - Add fs/ntfs/attrib.[hc]::ntfs_attr_extend_allocation(), a function to
76 extend the allocation of an attributes. Optionally, the data size,
77 but not the initialized size can be extended, too.
78 - Implement fs/ntfs/inode.[hc]::ntfs_truncate(). It only supports
79 uncompressed and unencrypted files and it never creates sparse files
80 at least for the moment (making a file sparse requires us to modify
81 its directory entries and we do not support directory operations at
82 the moment). Also, support for highly fragmented files, i.e. ones
83 whose data attribute is split across multiple extents, is severly
84 limited. When such a case is encountered, EOPNOTSUPP is returned.
85 - Enable ATTR_SIZE attribute changes in ntfs_setattr(). This completes
86 the initial implementation of file truncation. Now both open(2)ing
87 a file with the O_TRUNC flag and the {,f}truncate(2) system calls
88 will resize a file appropriately. The limitations are that only
89 uncompressed and unencrypted files are supported. Also, there is
90 only very limited support for highly fragmented files (the ones whose
91 $DATA attribute is split into multiple attribute extents).
92 - In attrib.c::ntfs_attr_set() call balance_dirty_pages_ratelimited()
93 and cond_resched() in the main loop as we could be dirtying a lot of
94 pages and this ensures we play nice with the VM and the system as a
96 - Implement file operations ->write, ->aio_write, ->writev for regular
97 files. This replaces the old use of generic_file_write(), et al and
98 the address space operations ->prepare_write and ->commit_write.
99 This means that both sparse and non-sparse (unencrypted and
100 uncompressed) files can now be extended using the normal write(2)
101 code path. There are two limitations at present and these are that
102 we never create sparse files and that we only have limited support
103 for highly fragmented files, i.e. ones whose data attribute is split
104 across multiple extents. When such a case is encountered,
105 EOPNOTSUPP is returned.
106 - $EA attributes can be both resident and non-resident.
107 - Use %z for size_t to fix compilation warnings. (Andrew Morton)
108 - Fix compilation warnings with gcc-4.0.2 on SUSE 10.0.
109 - Document extended attribute ($EA) NEED_EA flag. (Based on libntfs
110 patch by Yura Pakhuchiy.)
112 2.1.24 - Lots of bug fixes and support more clean journal states.
114 - Support journals ($LogFile) which have been modified by chkdsk. This
115 means users can boot into Windows after we marked the volume dirty.
116 The Windows boot will run chkdsk and then reboot. The user can then
117 immediately boot into Linux rather than having to do a full Windows
118 boot first before rebooting into Linux and we will recognize such a
119 journal and empty it as it is clean by definition. Note, this only
120 works if chkdsk left the journal in an obviously clean state.
121 - Support journals ($LogFile) with only one restart page as well as
122 journals with two different restart pages. We sanity check both and
123 either use the only sane one or the more recent one of the two in the
124 case that both are valid.
125 - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
126 ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
128 - Use ntfs_malloc_nofs_nofail() in the two critical regions in
129 fs/ntfs/runlist.c::ntfs_runlists_merge(). This means we no longer
130 need to panic() if the allocation fails as it now cannot fail.
131 - Fix two nasty runlist merging bugs that had gone unnoticed so far.
132 Thanks to Stefano Picerno for the bug report.
133 - Remove two bogus BUG_ON()s from fs/ntfs/mft.c.
134 - Fix handling of valid but empty mapping pairs array in
135 fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress().
136 - Report unrepresentable inodes during ntfs_readdir() as KERN_WARNING
137 messages and include the inode number. Thanks to Yura Pakhuchiy for
139 - Change ntfs_rl_truncate_nolock() to throw away the runlist if the new
141 - Add runlist.[hc]::ntfs_rl_punch_nolock() which punches a caller
142 specified hole into a runlist.
143 - Fix a bug in fs/ntfs/index.c::ntfs_index_lookup(). When the returned
144 index entry is in the index root, we forgot to set the @ir pointer in
145 the index context. Thanks to Yura Pakhuchiy for finding this bug.
146 - Remove bogus setting of PageError in ntfs_read_compressed_block().
147 - Add fs/ntfs/attrib.[hc]::ntfs_resident_attr_value_resize().
148 - Fix a bug in ntfs_map_runlist_nolock() where we forgot to protect
149 access to the allocated size in the ntfs inode with the size lock.
150 - Fix ntfs_attr_vcn_to_lcn_nolock() and ntfs_attr_find_vcn_nolock() to
151 return LCN_ENOENT when there is no runlist and the allocated size is
153 - Fix load_attribute_list() to handle the case of a NULL runlist.
154 - Fix handling of sparse attributes in ntfs_attr_make_non_resident().
155 - Add BUG() checks to ntfs_attr_make_non_resident() and ntfs_attr_set()
156 to ensure that these functions are never called for compressed or
157 encrypted attributes.
158 - Fix cluster (de)allocators to work when the runlist is NULL and more
159 importantly to take a locked runlist rather than them locking it
160 which leads to lock reversal.
161 - Truncate {a,c,m}time to the ntfs supported time granularity when
162 updating the times in the inode in ntfs_setattr().
163 - Fixup handling of sparse, compressed, and encrypted attributes in
164 fs/ntfs/inode.c::ntfs_read_locked_{,attr_,index_}inode(),
165 fs/ntfs/aops.c::ntfs_{read,write}page().
166 - Make ntfs_write_block() not instantiate sparse blocks if they contain
168 - Optimize fs/ntfs/aops.c::ntfs_write_block() by extending the page
169 lock protection over the buffer submission for i/o which allows the
170 removal of the get_bh()/put_bh() pairs for each buffer.
171 - Fix fs/ntfs/aops.c::ntfs_{read,write}_block() to handle the case
172 where a concurrent truncate has truncated the runlist under our feet.
173 - Fix page_has_buffers()/page_buffers() handling in fs/ntfs/aops.c.
174 - In fs/ntfs/aops.c::ntfs_end_buffer_async_read(), use a bit spin lock
175 in the first buffer head instead of a driver global spin lock to
177 - Minor fix to error handling and error message display in
178 fs/ntfs/aops.c::ntfs_prepare_nonresident_write().
179 - Change the mount options {u,f,d}mask to always parse the number as
180 an octal number to conform to how chmod(1) works, too. Thanks to
181 Giuseppe Bilotta and Horst von Brand for pointing out the errors of
183 - Fix various bugs in the runlist merging code. (Based on libntfs
184 changes by Richard Russon.)
185 - Fix sparse warnings that have crept in over time.
186 - Change ntfs_cluster_free() to require a write locked runlist on entry
187 since we otherwise get into a lock reversal deadlock if a read locked
188 runlist is passed in. In the process also change it to take an ntfs
189 inode instead of a vfs inode as parameter.
190 - Fix the definition of the CHKD ntfs record magic. It had an off by
191 two error causing it to be CHKB instead of CHKD.
192 - Fix a stupid bug in __ntfs_bitmap_set_bits_in_run() which caused the
193 count to become negative and hence we had a wild memset() scribbling
194 all over the system's ram.
196 2.1.23 - Implement extension of resident files and make writing safe as well as
197 many bug fixes, cleanups, and enhancements...
199 - Add printk rate limiting for ntfs_warning() and ntfs_error() when
200 compiled without debug. This avoids a possible denial of service
201 attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
203 - Fix compilation warnings on ia64. (Randy Dunlap)
204 - Use i_size_{read,write}() instead of reading i_size by hand and cache
205 the value where apropriate.
206 - Add size_lock to the ntfs_inode structure. This is an rw spinlock
207 and it locks against access to the inode sizes. Note, ->size_lock
208 is also accessed from irq context so you must use the _irqsave and
209 _irqrestore lock and unlock functions, respectively. Protect all
210 accesses to allocated_size, initialized_size, and compressed_size.
211 - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
212 - Implement extension of resident files in the regular file write code
213 paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()). At present
214 this only works until the data attribute becomes too big for the mft
215 record after which we abort the write returning -EOPNOTSUPP from
216 ntfs_prepare_write().
217 - Add disable_sparse mount option together with a per volume sparse
218 enable bit which is set appropriately and a per inode sparse disable
219 bit which is preset on some system file inodes as appropriate.
220 - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
221 - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
222 the creation of the unmapped runlist element for the base attribute
224 - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
225 helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
226 This allows us to map runlist fragments with the runlist lock already
227 held without having to drop and reacquire it around the call. Adapt
229 - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
230 runlist. This allows us to find runlist elements with the runlist
231 lock already held without having to drop and reacquire it around the
232 call. Adapt all callers.
233 - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
234 warning in the do_div() call on sparc32. Thanks to Meelis Roos for
235 the report and analysis of the warning.
236 - Fix a nasty runlist merge bug when merging two holes.
237 - Set the ntfs_inode->allocated_size to the real allocated size in the
238 mft record for resident attributes (fs/ntfs/inode.c).
239 - Small readability cleanup to use "a" instead of "ctx->attr"
240 everywhere (fs/ntfs/inode.c).
241 - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
242 definition of ntfs_export_ops from fs/ntfs/super.c to namei.c. Also,
243 declare ntfs_export_ops in fs/ntfs/ntfs.h.
244 - Correct sparse file handling. The compressed values need to be
245 checked and set in the ntfs inode as done for compressed files and
246 the compressed size needs to be used for vfs inode->i_blocks instead
247 of the allocated size, again, as done for compressed files.
248 - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
249 non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
250 - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
252 - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
253 dropping the read lock and taking the write lock we were not checking
254 whether someone else did not already do the work we wanted to do.
255 - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
256 ntfs_attr_find_vcn_nolock() and update all callers.
257 - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
258 - Fix sign of various error return values to be negative in
260 - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
261 handle the case where an attribute is converted from resident to
262 non-resident by a concurrent file write.
263 - Remove checks for NULL before calling kfree() since kfree() does the
264 checking itself. (Jesper Juhl)
265 - Some utilities modify the boot sector but do not update the checksum.
266 Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
267 only emit a warning when the checksum is incorrect rather than
268 refusing the mount. Thanks to Bernd Casimir for pointing this
270 - Update attribute definition handling.
271 - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
272 - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
273 - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
274 better code generation and one less sparse warning in fs/ntfs/aops.c.
275 - Remove spurious void pointer casts from fs/ntfs/. (Pekka Enberg)
276 - Use C99 style structure initialization after memory allocation where
277 possible (fs/ntfs/{attrib.c,index.c,super.c}). Thanks to Al Viro and
279 - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
280 is active on the volume and we are mounting read-write or remounting
281 from read-only to read-write.
282 - Fix a bug in address space operations error recovery code paths where
283 if the runlist was not mapped at all and a mapping error occured we
284 would leave the runlist locked on exit to the function so that the
285 next access to the same file would try to take the lock and deadlock.
286 - Detect the case when Windows has been suspended to disk on the volume
287 to be mounted and if this is the case do not allow (re)mounting
288 read-write. This is done by parsing hiberfil.sys if present.
289 - Fix several occurences of a bug where we would perform 'var & ~const'
290 with a 64-bit variable and a int, i.e. 32-bit, constant. This causes
291 the higher order 32-bits of the 64-bit variable to be zeroed. To fix
292 this cast the 'const' to the same 64-bit type as 'var'.
293 - Change the runlist terminator of the newly allocated cluster(s) to
294 LCN_ENOENT in ntfs_attr_make_non_resident(). Otherwise the runlist
296 - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
297 and ntfs_mapping_pairs_build() to allow the runlist encoding to be
298 partial which is desirable when filling holes in sparse attributes.
300 - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
301 if the requested vcn is inside it. Otherwise we get into problems
302 when we try to map an out of bounds vcn because we then try to map
303 the already mapped runlist fragment which causes
304 ntfs_mapping_pairs_decompress() to fail and return error. Update
305 ntfs_attr_find_vcn_nolock() accordingly.
306 - Fix a nasty deadlock that appeared in recent kernels.
307 The situation: VFS inode X on a mounted ntfs volume is dirty. For
308 same inode X, the ntfs_inode is dirty and thus corresponding on-disk
309 inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
310 to the table of inodes, i.e. $MFT, inode 0.
312 Process 1: sys_sync()/umount()/whatever... calls
313 __sync_single_inode() for $MFT -> do_writepages() -> write_page for
314 the dirty page containing the on-disk inode X, the page is now locked
315 -> ntfs_write_mst_block() which clears PageUptodate() on the page to
316 prevent anyone else getting hold of it whilst it does the write out.
317 This is necessary as the on-disk inode needs "fixups" applied before
318 the write to disk which are removed again after the write and
319 PageUptodate is then set again. It then analyses the page looking
320 for dirty on-disk inodes and when it finds one it calls
321 ntfs_may_write_mft_record() to see if it is safe to write this
322 on-disk inode. This then calls ilookup5() to check if the
323 corresponding VFS inode is in icache(). This in turn calls ifind()
324 which waits on the inode lock via wait_on_inode whilst holding the
326 Process 2: pdflush results in a call to __sync_single_inode for the
327 same VFS inode X on the ntfs volume. This locks the inode (I_LOCK)
328 then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
329 read_cache_page() for the page (in page cache of table of inodes
330 $MFT, inode 0) containing the on-disk inode. This page has
331 PageUptodate() clear because of Process 1 (see above) so
332 read_cache_page() blocks when it tries to take the page lock for the
333 page so it can call ntfs_read_page().
334 Thus Process 1 is holding the page lock on the page containing the
335 on-disk inode X and it is waiting on the inode X to be unlocked in
336 ifind() so it can write the page out and then unlock the page.
337 And Process 2 is holding the inode lock on inode X and is waiting for
338 the page to be unlocked so it can call ntfs_readpage() or discover
339 that Process 1 set PageUptodate() again and use the page.
340 Thus we have a deadlock due to ifind() waiting on the inode lock.
341 The solution: The fix is to use the newly introduced
342 ilookup5_nowait() which does not wait on the inode's lock and hence
343 avoids the deadlock. This is safe as we do not care about the VFS
344 inode and only use the fact that it is in the VFS inode cache and the
345 fact that the vfs and ntfs inodes are one struct in memory to find
346 the ntfs inode in memory if present. Also, the ntfs inode has its
347 own locking so it does not matter if the vfs inode is locked.
348 - Fix bug in mft record writing where we forgot to set the device in
349 the buffers when mapping them after the VM had discarded them.
350 Thanks to Martin MOKREJŠ for the bug report.
352 2.1.22 - Many bug and race fixes and error handling improvements.
354 - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
355 - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
356 instead of void and provide a helper ntfs_truncate_vfs() for the
357 vfs ->truncate method.
358 - Add a new ntfs inode flag NInoTruncateFailed() and modify
359 fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
360 - Fix min_size and max_size definitions in ATTR_DEF structure in
361 fs/ntfs/layout.h to be signed.
362 - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
363 ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
364 ntfs_attr_can_be_resident(), which in turn use the new private helper
365 ntfs_attr_find_in_attrdef().
366 - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
367 mapping->private_lock around the dirtying of the buffer heads
368 analagous to the way it is done in __set_page_dirty_buffers().
369 - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
370 mount time as this cannot work with the current implementation.
371 - Check for location of attribute name and improve error handling in
372 general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
373 - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
374 i_size, i.e. race with truncate, invalidate the buffers on the page
375 so that they become freeable and hence the page does not leak.
376 - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
378 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
379 a NULL pointer dereference in the error code path when a corrupt
380 attribute was found. (Thanks to Domen Puncer for the bug report.)
381 - Add MODULE_VERSION() to fs/ntfs/super.c.
382 - Make several functions and variables static. (Adrian Bunk)
383 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
384 buffers for the page if they are not present and then marks the
385 buffers belonging to the ntfs record dirty. This causes the buffers
386 to become busy and hence they are safe from removal until the page
387 has been written out.
388 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
389 error handling code path that resulted in a BUG() due to trying to
390 unmap an extent mft record when the mapping of it had failed and it
391 thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
392 - Drop the runlist lock after the vcn has been read in
393 fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
394 - Rewrite handling of multi sector transfer errors. We now do not set
395 PageError() when such errors are detected in the async i/o handler
396 fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
397 protected attributes now check the magic of each ntfs record as they
398 use it and act appropriately. This has the effect of making errors
399 granular per ntfs record rather than per page which solves the case
400 where we cannot access any of the ntfs records in a page when a
401 single one of them had an mst error. (Thanks to Ken MacFerrin for
403 - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
404 where we failed to release i_mutex on the $Quota/$Q attribute inode.
405 - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
406 - Add mapping of unmapped buffers to all remaining code paths, i.e.
407 fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
408 and write_mft_record_nolock(). From now on we require that the
409 complete runlist for the mft mirror is always mapped into memory.
410 - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
411 - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
412 - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
413 resident attribute will be smaller than a page which makes the code
414 simpler. Also make the code more tolerant to concurrent ->truncate.
416 2.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
418 - Implement extent mft record deallocation
419 fs/ntfs/mft.c::ntfs_extent_mft_record_free().
420 - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
421 - Add vol->mft_data_pos and initialize it at mount time.
422 - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
423 ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
424 ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
425 ntfs_runlists_merge() and adapt all callers.
426 - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
427 ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
428 and ntfs_mapping_pairs_build(), adapted from libntfs.
429 - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
430 static and add a declaration for it to lcnalloc.h.
431 - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
432 inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
433 cluster bitmap lock for the duration of the call.
434 - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
435 - Implement the equivalent of memset() for an ntfs attribute in
436 fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
437 fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
438 - Remove unnecessary casts from LCN_* constants.
439 - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
440 - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
441 change MFT_RECORD to contain the NTFS 3.1+ specific fields.
442 - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
443 marks all buffers belonging to an ntfs record dirty, followed by
444 marking the page the ntfs record is in dirty and also marking the vfs
445 inode containing the ntfs record dirty (I_DIRTY_PAGES).
446 - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
447 new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
448 longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
449 - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
451 - Move the typedefs for runlist_element and runlist from types.h to
452 runlist.h and fix resulting include errors.
453 - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
454 - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
455 mark_ntfs_record_dirty() which also changes the behaviour in that we
456 now set the buffers belonging to the mft record dirty as well as the
458 - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
459 to cope with the fact that there now are dirty buffers in mft pages.
460 - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
461 mark_ntfs_record_dirty() and thus to set the buffers belonging to the
462 mft record dirty as well as the page itself.
463 - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
464 slightly modified by me)
465 - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
466 the mft record is already locked and otherwise behaves the same way
467 as fs/ntfs/mft.c::map_mft_record().
468 - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
469 writes the mft record if the buffers belonging to it are dirty.
470 Otherwise we assume that it was written out by other means already.
471 - Attempting to write outside initialized size is _not_ a bug so remove
472 the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
473 fact required to write outside initialized size when preparing to
474 extend the initialized size.
475 - Map the page instead of using page_address() before writing to it in
476 fs/ntfs/aops.c::ntfs_mft_writepage().
477 - Provide exclusion between opening an inode / mapping an mft record
478 and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
479 by setting the page not uptodate throughout ntfs_mft_writepage().
480 - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
481 to ensure noone can see the page whilst the mst fixups are applied.
482 - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
483 checks if an mft record may be written out safely obtaining any
484 necessary locks in the process. This is used by
485 fs/ntfs/aops.c::ntfs_write_mst_block().
486 - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
487 writing mft records and improve its error handling in the process.
488 Now if any of the records in the page fail to be written out, all
489 other records will be written out instead of aborting completely.
490 - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
491 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
492 ntfs_mst_aops for all inodes which are NInoMstProtected() and
493 ntfs_aops for all other inodes.
494 - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
495 ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
496 no longer require an ntfs inode to be present. Update all callers.
497 - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
498 - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
499 to ensure noone can see the page whilst the mst fixups are applied.
500 - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
501 fs/ntfs/mft.c::try_map_mft_record().
502 - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
503 with the ntfs inode which contains the page rather than the ntfs
504 inode the mft record of which is in the page.
505 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
506 index inode bitmap inode release code from there to
507 fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
508 Hellwig for spotting this.)
509 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
510 inode semaphore around the code that sets ni->itype.index.bmp_ino to
511 NULL and reorganize the code to optimize it a bit. (Thanks to
512 Christoph Hellwig for spotting this.)
513 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
514 ntfs inode as a parameter as this is confusing and misleading and the
515 needed ntfs inode is available via NTFS_I(page->mapping->host).
516 Adapt all callers to this change.
517 - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
518 fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
519 of the first buffer in a record and to take this as the ntfs record
520 dirty state. We cannot look at the dirty state for subsequent
521 buffers because we might be racing with
522 fs/ntfs/aops.c::mark_ntfs_record_dirty().
523 - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
524 inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
525 add a declaration for it to inode.h. Fix some compilation issues
526 that resulted due to #includes and header file interdependencies.
527 - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
528 - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
529 - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
530 record sequence number if it is specified (i.e. not zero).
531 - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
532 functions used by it.
533 - Update Documentation/filesystems/ntfs.txt with instructions on how to
534 use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
535 the linear raid problem with the Software RAID / MD driver when one
536 or more of the devices has an odd number of sectors.
538 2.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
540 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
541 where we did not clear ctx->al_entry but it was still set due to
542 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
544 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
545 where we forgot to unmap the extent mft record when we had finished
546 enumerating an attribute which caused a bug check to trigger when the
547 VFS calls ->clear_inode.
549 2.1.19 - Many cleanups, improvements, and a minor bug fix.
551 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
552 change the uid, gid, and mode of an inode as we do not support NTFS
554 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
556 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
557 and ntfs_filldir() as per suggestion from Al Viro.
558 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
559 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
560 inode size has changed and to only output an error if so.
561 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
562 - Add le{16,32,64} as well as sle{16,32,64} data types to
564 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
565 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
566 respectively, to fs/ntfs/types.h.
567 - Update endianness conversion macros in fs/ntfs/endian.h to use the
568 new types as appropriate.
569 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
571 - Add leMFT_REF data type to fs/ntfs/layout.h.
572 - Update all NTFS header files with the new little endian data types.
573 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
574 - Do proper type casting when using ntfs_is_*_recordp() in
575 fs/ntfs/logfile.c, mft.c, and super.c.
576 - Fix all the sparse bitwise warnings. Had to change all the typedef
577 enums storing little endian values to simple enums plus a typedef for
578 the datatype to make sparse happy.
579 - Fix a bug found by the new sparse bitwise warnings where the default
580 upcase table was defined as a pointer to wchar_t rather than ntfschar
581 in fs/ntfs/ntfs.h and super.c.
582 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
584 2.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
586 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
587 the calculation of total/free inodes as used by statfs().
588 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
589 because the code itself is using the ntfs_lock semaphore which
590 provides safe locking. (Ingo Molnar)
591 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
592 could occur in the future for when we start closing/freeing extent
593 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
595 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
596 find_external_attr() to ntfs_external_attr_find() to cleanup the
597 namespace a bit and to be more consistent with libntfs.
598 - Rename {{re,}init,get,put}_attr_search_ctx() to
599 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
600 attr_search_context to ntfs_attr_search_ctx.
601 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
602 for the attribute list attribute itself.
603 - Fix endianness bug in ntfs_external_attr_find().
604 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
605 if the attribute is not found, and -EIO on real error. In the case
606 of -ENOENT, the search context is updated to describe the attribute
607 before which the attribute being searched for would need to be
608 inserted if such an action were to be desired and in the case of
609 ntfs_external_attr_find() the search context is also updated to
610 indicate the attribute list entry before which the attribute list
611 entry of the attribute being searched for would need to be inserted
612 if such an action were to be desired. Also make ntfs_find_attr()
613 static and remove its prototype from attrib.h as it is not used
614 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
615 callers of ntfs_{external,}attr_{find,lookup}() for the new return
617 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
619 2.1.17 - Fix bugs in mount time error code paths and other updates.
621 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
622 includes functions to set/clear a single bit or a run of bits.
623 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
624 runlist element containing a particular vcn. It also takes care of
625 mapping any needed runlist fragments.
626 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
627 - Load attribute definition table from $AttrDef at mount time.
628 - Fix bugs in mount time error code paths involving (de)allocation of
629 the default and volume upcase tables.
630 - Remove ntfs_nr_mounts as it is no longer used.
632 2.1.16 - Implement access time updates, file sync, async io, and read/writev.
634 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
635 This is done by setting the appropriate file operations pointers to
636 the generic helper functions provided by mm/filemap.c.
637 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
638 and directories (fs/ntfs/dir.c).
639 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
640 Note, except for the root directory and any other system files opened
641 by the user, the system files will not have their access times
642 updated as they are only accessed at the inode level an hence the
643 file level functions which cause the times to be updated are never
646 2.1.15 - Invalidate quotas when (re)mounting read-write.
648 - Add new element itype.index.collation_rule to the ntfs inode
649 structure and set it appropriately in ntfs_read_locked_inode().
650 - Implement a new inode type "index" to allow efficient access to the
651 indices found in various system files and adapt inode handling
652 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
653 attribute inode (NInoAttr() is true) with an attribute type of
654 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
655 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
656 there would be no way to distinguish between normal attribute inodes
657 and index inodes. The function to obtain an index inode is
658 ntfs_index_iget() and it uses the helper function
659 ntfs_read_locked_index_inode(). Note, we do not overload
660 ntfs_attr_iget() as indices consist of multiple attributes so using
661 ntfs_attr_iget() to obtain an index inode would be confusing.
662 - Ensure that there is no overflow when doing page->index <<
663 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
664 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
665 and ntfs_read_block().
666 - Use case sensitive attribute lookups instead of case insensitive ones.
667 - Lock all page cache pages belonging to mst protected attributes while
668 accessing them to ensure we never see corrupt data while the page is
670 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
671 We have ntfs_is_collation_rule_supported() to check if the collation
672 rule you want to use is supported and ntfs_collation() which actually
673 collates two data items. We currently only support COLLATION_BINARY
674 and COLLATION_NTOFS_ULONG but support for other collation rules will
675 be added as the need arises.
676 - Add a new type, ntfs_index_context, to allow retrieval of an index
677 entry using the corresponding index key. To get an index context,
678 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
679 This also adds a new slab cache for the index contexts. To lookup a
680 key in an index inode, use ntfs_index_lookup(). After modifying an
681 index entry, call ntfs_index_entry_flush_dcache_page() followed by
682 ntfs_index_entry_mark_dirty() to ensure the changes are written out
683 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
684 an index entry is in the index allocation attribute rather than the
685 index root attribute it will not be written out (you will get a
686 warning message about discarded changes instead).
687 - Load the quota file ($Quota) and check if quota tracking is enabled
688 and if so, mark the quotas out of date. This causes windows to
689 rescan the volume on boot and update all quota entries.
690 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
691 It is simply set to __set_page_dirty_nobuffers() to make sure that
692 running set_page_dirty() on a page containing mft/ntfs records will
693 not affect the dirty state of the page buffers.
694 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
695 buffers that are inside the ntfs record in the page dirty after which
696 it sets the page dirty. This allows ->writepage to only write the
697 dirty index records rather than having to write all the records in
698 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
699 use this rather than __set_page_dirty_nobuffers().
700 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
701 writing of page cache pages belonging to mst protected attributes
702 like the index allocation attribute in directory indices and other
703 indices like $Quota/$Q, etc. This means that the quota is now marked
704 out of date on all volumes rather than only on ones where the quota
705 defaults entry is in the index root attribute of the $Quota/$Q index.
707 2.1.14 - Fix an NFSd caused deadlock reported by several users.
709 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
710 to a buffer so that we can put the search context and unmap the mft
711 record before calling the filldir() callback. We need to do this
712 because of NFSd which calls ->lookup() from its filldir callback()
713 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
714 of the directory and since ntfs_readdir() has got it mapped already
715 ntfs_lookup() deadlocks.
717 2.1.13 - Enable overwriting of resident files and housekeeping of system files.
719 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
720 keeping the mft mirror in sync with the mft when mirrored mft records
721 are written. The functions are write_mft_record{,_nolock}(). The
722 implementation is quite rudimentary for now with lots of things not
723 implemented yet but I am not sure any of them can actually occur so
724 I will wait for people to hit each one and only then implement it.
725 - Commit open system inodes at umount time. This should make it
726 virtually impossible for sync_mft_mirror_umount() to ever be needed.
727 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
728 ntfs super operations. This gives us inode writing via the VFS inode
729 dirty code paths. Note: Access time updates are not implemented yet.
730 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
731 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
732 finally enabling resident file overwrite! (-8 This also includes a
733 placeholder for ->writepage (ntfs_mft_writepage()), which for now
734 just redirties the page and returns. Also, at umount time, we for
735 now throw away all mft data page cache pages after the last call to
736 ntfs_commit_inode() in the hope that all inodes will have been
737 written out by then and hence no dirty (meta)data will be lost. We
738 also check for this case and emit an error message telling the user
740 - Use set_page_writeback() and end_page_writeback() in the resident
741 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
742 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
744 - Implement ntfs_mft_writepage() so it now checks if any of the mft
745 records in the page are dirty and if so redirties the page and
746 returns. Otherwise it just returns (after doing set_page_writeback(),
747 unlock_page(), end_page_writeback() or the radix-tree tag
748 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
749 alowing the VM to do with the page as it pleases. Also, at umount
750 time, now only throw away dirty mft (meta)data pages if dirty inodes
751 are present and ask the user to email us if they see this happening.
752 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
753 information flags (fs/ntfs/super.c).
754 - Mark the volume dirty when (re)mounting read-write and mark it clean
755 when unmounting or remounting read-only. If any volume errors are
756 found, the volume is left marked dirty to force chkdsk to run.
757 - Add code to set the NT4 compatibility flag when (re)mounting
758 read-write for newer NTFS versions but leave it commented out for now
759 since we do not make any modifications that are NTFS 1.2 specific yet
760 and since setting this flag breaks Captive-NTFS which is not nice.
761 This code must be enabled once we start writing NTFS 1.2 specific
762 changes otherwise Windows NTFS driver might crash / cause corruption.
764 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
766 - Add a new address space operations struct, ntfs_mst_aops, for mst
767 protected attributes. This is because the default ntfs_aops do not
768 make sense with mst protected data and were they to write anything to
769 such an attribute they would cause data corruption so we provide
770 ntfs_mst_aops which does not have any write related operations set.
771 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
772 includes an adapted ntfs_commit_inode() and an implementation of
773 ntfs_write_inode() which for now just cleans dirty inodes without
774 writing them (it does emit a warning that this is happening).
775 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
776 entry) as it was only fixing a theoretical bug but at the same time
777 it badly broke the handling of sparse and uncompressed compression
780 2.1.11 - Driver internal cleanups.
782 - Only build logfile.o if building the driver with read-write support.
783 - Really final white space cleanups.
784 - Use generic_ffs() instead of ffs() in logfile.c which allows the
785 log_page_size variable to be optimized by gcc into a constant.
786 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
787 char as defined by POSIX and as found on some systems.
789 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
791 - Finish off the white space cleanups (remove trailing spaces, etc).
792 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
793 the kludges around the first iget(). Instead of (re)setting ->s_op
794 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
795 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
796 kills the need for second super_operations and allows to return error
797 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
799 - Force read-only (re)mounting if any of the following bits are set in
800 the volume information flags:
801 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
802 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
803 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
804 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
805 above bits set so the test is made easy.
807 2.1.9 - Fix two bugs in decompression engine.
809 - Fix a bug where we would not always detect that we have reached the
810 end of a compression block because we were ending at minus one byte
811 which is effectively the same as being at the end. The fix is to
812 check whether the uncompressed buffer has been fully filled and if so
813 we assume we have reached the end of the compression block. A big
814 thank you to Marcin Gibuła for the bug report, the assistance in
815 tracking down the bug and testing the fix.
816 - Fix a possible bug where when a compressed read is truncated to the
817 end of the file, the offset inside the last page was not truncated.
819 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
821 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
822 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
823 utc2ntfs() to work with struct timespec instead of time_t on the
824 Linux UTC time side thus preserving the full precision of the NTFS
825 time and only loosing up to 99 nano-seconds in the Linux UTC time.
826 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
828 - Remove unused ntfs_dirty_inode().
829 - Cleanup super operations declaration in fs/ntfs/super.c.
830 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
831 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
832 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
833 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
834 fs/ntfs/inode.h so they can be used elsewhere.
835 - Determine the mft mirror size as the number of mirrored mft records
836 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
837 - Load the mft mirror at mount time and compare the mft records stored
838 in it to the ones in the mft. Force a read-only mount if the two do
839 not match (fs/ntfs/super.c).
840 - Fix type casting related warnings on 64-bit architectures. Thanks
841 to Meelis Roos for reporting them.
842 - Move %L to %ll as %L is floating point and %ll is integer which is
844 - Read the journal ($LogFile) and determine if the volume has been
845 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
846 and fs/ntfs/logfile.c). This is a little bit of a crude check in
847 that we only look at the restart areas and not at the actual log
848 records so that there will be a very small number of cases where we
849 think that a volume is dirty when in fact it is clean. This should
850 only affect volumes that have not been shutdown cleanly and did not
851 have any pending, non-check-pointed i/o.
852 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
853 is requested, empty $LogFile by overwriting it with 0xff bytes to
854 ensure that Windows cannot cause data corruption by replaying a stale
855 journal after Linux has written to the volume.
857 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
859 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
860 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
861 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
862 default doesn't allow inode number 0 which is a valid inode on NTFS
863 and even if it did allow that it uses iget() instead of ntfs_iget()
864 which makes it useless for us.
865 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
866 default just returns -EACCES which is not very useful.
867 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
868 and set them up in the super block at mount time (super.c) this
869 allows mounted NTFS volumes to be exported via NFS.
870 - Add missing return -EOPNOTSUPP; in
871 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
872 - Enforce no atime and no dir atime updates at mount/remount time as
873 they are not implemented yet anyway.
874 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
875 after a NULL check. Thanks to Dave Jones for pointing this out.
877 2.1.6 - Fix minor bug in handling of compressed directories.
879 - Fix bug in handling of compressed directories. A compressed
880 directory is not really compressed so when we set the ->i_blocks
881 field of a compressed directory inode we were setting it from the
882 non-existing field ni->itype.compressed.size which gave random
883 results... For directories we now always use ni->allocated_size.
885 2.1.5 - Fix minor bug in attribute list attribute handling.
887 - Fix bug in attribute list handling. Actually it is not as much a bug
888 as too much protection in that we were not allowing attribute lists
889 which waste space on disk while Windows XP clearly allows it and in
890 fact creates such attribute lists so our driver was failing.
891 - Update NTFS documentation ready for 2.6 kernel release.
893 2.1.4 - Reduce compiler requirements.
895 - Remove all uses of unnamed structs and unions in the driver to make
896 old and newer gcc versions happy. Makes it a bit uglier IMO but at
897 least people will stop hassling me about it.
899 2.1.3 - Important bug fixes in corner cases.
901 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
902 clusters. (Philipp Thomas)
903 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
904 multiple of the block_size but not the cluster size. (Szabolcs
907 2.1.2 - Important bug fixes aleviating the hangs in statfs.
909 - Fix buggy free cluster and free inode determination logic.
911 2.1.1 - Minor updates.
913 - Add handling for initialized_size != data_size in compressed files.
914 - Reduce function local stack usage from 0x3d4 bytes to just noise in
915 fs/ntfs/upcase.c. (Randy Dunlap)
916 - Remove compiler warnings for newer gcc.
917 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
918 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
919 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
920 kmap_atomic(KM_USER0).
922 2.1.0 - First steps towards write support: implement file overwrite.
924 - Add configuration option for developmental write support with an
925 appropriately scary configuration help text.
926 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
927 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
928 overwriting of existing files on ntfs. Note: Resident files are
929 only written into memory, and not written out to disk at present, so
930 avoid writing to files smaller than about 1kiB.
931 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
932 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
933 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
934 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
935 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
936 This enables write(2) based overwriting of existing files on ntfs.
937 Note: As with mmap(2) based overwriting, resident files are only
938 written into memory, and not written out to disk at present, so avoid
939 writing to files smaller than about 1kiB.
940 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
941 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
942 files with the purpose of intercepting and aborting all i_size
943 changes which we do not support yet. ntfs_truncate() actually only
944 emits a warning message but AFAICS our interception of i_size changes
945 elsewhere means ntfs_truncate() never gets called for i_size changes.
946 It is only called from generic_file_write() when we fail in
947 ntfs_prepare_{,nonresident_}write() in order to discard any
948 instantiated buffers beyond i_size. Thus i_size is not actually
949 changed so our warning message is enough. Unfortunately it is not
950 possible to easily determine if i_size is being changed or not hence
951 we just emit an appropriately worded error message.
953 2.0.25 - Small bug fixes and cleanups.
955 - Unlock the page in an out of memory error code path in
956 fs/ntfs/aops.c::ntfs_read_block().
957 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
958 just unlock the page and return. (This can happen due to ->writepage
959 clearing PageUptodate() during write out of MstProtected()
961 - Remove leaked write code again.
965 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
966 inside BUG_ON(). (Adam J. Richter)
967 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
968 calls for improved debugging. (Adam J. Richter)
969 - Add errors flag to the ntfs volume state, accessed via
970 NVol{,Set,Clear}Errors(vol).
971 - Do not allow read-write remounts of read-only volumes with errors.
972 - Clarify comment for ntfs file operation sendfile which was added by
973 Christoph Hellwig a while ago (just using generic_file_sendfile())
974 to say that ntfs ->sendfile is only used for the case where the
975 source data is on the ntfs partition and the destination is
976 somewhere else, i.e. nothing we need to concern ourselves with.
977 - Add generic_file_write() as our ntfs file write operation.
979 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
981 - Massive internal locking changes to mft record locking. Fixes lock
982 recursion and replaces the mrec_lock read/write semaphore with a
983 mutex. Also removes the now superfluous mft_count. This fixes several
984 race conditions and deadlocks, especially in the future write code.
985 - Fix ntfs over loopback for compressed files by adding an
986 optimization barrier. (gcc was screwing up otherwise ?)
987 - Miscellaneous cleanups all over the code and a fix or two in error
989 Thanks go to Christoph Hellwig for pointing out the following two:
990 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
991 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
993 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
995 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
996 at entry/exit respectively.
997 - Use C99 initializers for structures.
998 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
1000 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
1002 - Limit volume size at mount time to 2TiB on architectures where
1003 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
1004 This is the most we can do without overflowing the 32-bit limit of
1005 the block device size imposed on us by sb_bread() and sb_getblk()
1007 - Limit file/directory size at open() time to 16TiB on architectures
1008 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
1009 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
1010 overflowing the page cache page index.
1012 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
1014 - Move the directory index bitmap to use an attribute inode instead of
1015 having special fields for it inside the ntfs inode structure. This
1016 means that the index bitmaps now use the page cache for i/o, too,
1017 and also as a side effect we get support for non-resident index
1019 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
1020 fix a page leak that manifested itself in some cases.
1021 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
1022 index bitmap inode on the final iput().
1024 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
1026 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
1027 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
1028 - Drop the "file" from ntfs_file_read_compressed_block().
1029 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
1030 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
1031 - Update ntfs_end_buffer_async_read() with the improved logic from
1032 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
1033 further logic improvements to better determine when we set PageError.
1034 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
1035 check for the buffers being uptodate first in line with the updated
1036 fs/buffer.c::block_read_full_page(). This plugs a small race
1039 2.0.18 - Fix race condition in reading of compressed files.
1041 - There was a narrow window between checking a buffer head for being
1042 uptodate and locking it in ntfs_file_read_compressed_block(). We now
1043 lock the buffer and then check whether it is uptodate or not.
1045 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
1047 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
1048 code and update callers, i.e. ntfs_iget(), to pass that error code
1049 up instead of just using -EIO.
1050 - Modifications to super.c to ensure that both mount and remount
1051 cannot set any write related options when the driver is compiled
1053 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
1054 cache the current runlist element. This should improve performance
1055 when reading very large and/or very fragmented data.
1057 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
1059 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
1060 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
1061 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
1062 - Convert $MFT/$BITMAP access to attribute inode API and remove all
1063 remnants of the ugly mftbmp address space and operations hack. This
1064 means we finally have only one readpage function as well as only one
1065 async io completion handler. Yey! The mft bitmap is now just an
1066 attribute inode and is accessed from vol->mftbmp_ino just as if it
1067 were a normal file. Fake inodes rule. (-:
1069 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
1071 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
1072 remounts to fail when the partition had an entry in /etc/fstab and
1073 the entry specified the nls= option.
1074 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
1075 expand all the helper functions NVolFoo(), NVolSetFoo(), and
1077 - Move copyright statement from driver initialisation message to
1078 module description (fs/super.c). This makes the initialisation
1079 message fit on one line and fits in better with rest of kernel.
1080 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
1081 attribute inodes, and both for files and directories.
1082 - Implement fake attribute inodes allowing all attribute i/o to go via
1083 the page cache and to use all the normal vfs/mm functionality:
1084 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
1086 - Add needed cleanup code to ntfs_clear_big_inode().
1087 - Merge address space operations for files and directories (aops.c),
1088 now just have ntfs_aops:
1090 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
1091 ntfs_attr_read_block() -> ntfs_read_block(),
1092 ntfs_file_read_page() -> ntfs_readpage().
1093 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
1094 attribute inodes, and both for files and directories.
1095 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
1097 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
1099 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
1100 the locking out of super.c::get_nr_free_mft_records() and taking and
1101 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
1102 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
1103 current userspace ntfs library code. This means that if a merge
1104 fails the original runlists are always left unmodified instead of
1105 being silently corrupted.
1108 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
1110 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
1111 from ntfs_volume structure.
1112 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
1113 from ntfs_volume structure.
1114 - Use iget5_locked() and friends instead of conventional iget(). Wrap
1115 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
1116 to use ntfs_iget(). Leave only one iget() call at mount time so we
1117 don't need an ntfs_iget_mount().
1118 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
1119 additional argument.
1121 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
1123 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
1124 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
1125 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
1126 to determine whether to apply mst fixups or not.
1127 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
1128 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
1129 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
1130 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
1131 the VFS readpage function prototype to the ntfs_attr_read_block()
1134 2.0.11 - Initial preparations for fake inode based attribute i/o.
1136 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
1137 do some macro magic (adapted from include/linux/buffer_head.h) to
1138 expand all the helper functions NInoFoo(), NInoSetFoo(), and
1140 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
1141 - Add new fields to ntfs_inode structure to allow use of fake inodes
1142 for attribute i/o: type, name, name_len. Also add new state bits:
1143 NI_Attr, which, if set, indicates the inode is a fake inode, and
1144 NI_MstProtected, which, if set, indicates the attribute uses multi
1145 sector transfer protection, i.e. fixups need to be applied after
1146 reads and before/after writes.
1147 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
1148 ntfs_{new,clear,destroy}_extent_inode() and update callers.
1149 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
1150 instead of ntfs_destroy_extent_inode().
1151 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
1152 - Make all operations on ntfs inode state bits use the NIno* functions.
1153 - Set up the new ntfs inode fields and state bits in
1154 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
1155 allocated memory to __ntfs_clear_inode().
1156 - Cleanup ntfs_inode structure a bit for better ordering of elements
1157 w.r.t. their size to allow better packing of the structure in memory.
1159 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
1161 - Add check at mount time to verify that the number of inodes on the
1162 volume does not exceed 2^32 - 1, which is the maximum allowed for
1163 NTFS according to Microsoft.
1164 - Change mft_no member of ntfs_inode structure to be unsigned long.
1165 Update all users. This makes ntfs_inode->mft_no just a copy of struct
1166 inode->i_ino. But we can't just always use struct inode->i_ino and
1167 remove mft_no because extent inodes do not have an attached struct
1170 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
1172 - Change decompression engine to use a single buffer protected by a
1173 spin lock instead of per-CPU buffers. (Rusty Russell)
1174 - Do not update cb_pos when handling a partial final page during
1175 decompression of a sparse compression block, as the value is later
1176 reset without being read/used. (Rusty Russell)
1177 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
1179 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
1180 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
1181 it also makes everything safer so it is a good thing.
1182 - Miscellaneous minor cleanups to comments.
1184 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
1186 Big thanks go to Al Viro and other inhabitants of #kernel for investing
1187 their time to discuss the case sensitivity and dcache aliasing issues.
1189 - Remove unused source file fs/ntfs/attraops.c.
1190 - Remove show_inodes mount option(s), thus dropping support for
1191 displaying of short file names.
1192 - Remove deprecated mount option posix.
1193 - Restore show_sys_files mount option.
1194 - Add new mount option case_sensitive, to determine if the driver
1195 treats file names as case sensitive or not. If case sensitive, create
1196 file names in the POSIX namespace. Otherwise create file names in the
1197 LONG/WIN32 namespace. Note, files remain accessible via their short
1198 file name, if it exists.
1199 - Remove really dumb logic bug in boot sector recovery code.
1200 - Fix dcache aliasing issues wrt short/long file names via changes
1201 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
1202 fs/ntfs/namei.c::ntfs_lookup():
1203 - Add additional argument to ntfs_lookup_inode_by_name() in which we
1204 return information about the matching file name if the case is not
1205 matching or the match is a short file name. See comments above the
1206 function definition for details.
1207 - Change ntfs_lookup() to only create dcache entries for the correctly
1208 cased file name and only for the WIN32 namespace counterpart of DOS
1209 namespace file names. This ensures we have only one dentry per
1210 directory and also removes all dcache aliasing issues between short
1211 and long file names once we add write support. See comments above
1212 function for details.
1213 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
1215 2.0.7 - Minor cleanups and updates for changes in core kernel code.
1217 - Remove much of the NULL struct element initializers.
1218 - Various updates to make compatible with recent kernels.
1219 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
1220 in fs/ntfs/ntfs.h instead.
1221 - Remove no longer needed KERNEL_VERSION checks. We are now in the
1222 kernel proper so they are no longer needed.
1224 2.0.6 - Major bugfix to make compatible with other kernel changes.
1226 - Initialize the mftbmp address space properly now that there are more
1227 fields in the struct address_space. This was leading to hangs and
1228 oopses on umount since 2.5.12 because of changes to other parts of
1229 the kernel. We probably want a kernel generic init_address_space()
1231 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
1232 only caller of ->readdir() is vfs_readdir() which holds i_mutex
1233 during the call, and i_mutex is sufficient protection against changes
1234 in the directory inode (including ->i_size).
1235 - Use generic_file_llseek() for directories (as opposed to
1236 default_llseek()) as this downs i_mutex instead of the BKL which is
1237 what we now need for exclusion against ->f_pos changes considering we
1238 no longer take the BKL in ntfs_readdir().
1240 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
1242 - No need to set old blocksize in super.c::ntfs_fill_super() as the
1243 VFS does so via invocation of deactivate_super() calling
1244 fs->fill_super() calling block_kill_super() which does it.
1245 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
1246 -> Do we really need it? I don't think so as we have exclusion on
1247 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
1248 move the ->f_pos accesses under the mrec_lock though. Check this...
1249 - Fix really, really, really stupid buffer overflow in extent inode
1250 handling in mft.c::map_extent_mft_record().
1252 2.0.4 - Cleanups and updates for kernel 2.5.11.
1254 - Add documentation on how to use the MD driver to be able to use NTFS
1255 stripe and volume sets in Linux and generally cleanup documentation
1257 Remove all uses of kdev_t in favour of struct block_device *:
1258 - Change compress.c::ntfs_file_read_compressed_block() to use
1259 sb_getblk() instead of getblk().
1260 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
1261 of get_hardsect_size().
1262 - No need to get old blocksize in super.c::ntfs_fill_super() as
1263 fs/super.c::get_sb_bdev() already does this.
1264 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
1266 2.0.3 - Small bug fixes, cleanups, and performance improvements.
1268 - Remove some dead code from mft.c.
1269 - Optimize readpage and read_block functions throughout aops.c so that
1270 only initialized blocks are read. Non-initialized ones have their
1271 buffer head mapped, zeroed, and set up to date, without scheduling
1272 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
1273 Thanks go to Andrew Morton for spotting the below:
1274 - Fix buglet in allocate_compression_buffers() error code path.
1275 - Call flush_dcache_page() after modifying page cache page contents in
1276 ntfs_file_readpage().
1277 - Check for existence of page buffers throughout aops.c before calling
1278 create_empty_buffers(). This happens when an I/O error occurs and the
1279 read is retried. (It also happens once writing is implemented so that
1280 needed doing anyway but I had left it for later...)
1281 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
1282 readpage and read_block functions. Reasoning same as above (i.e. I/O
1283 error retries and future write code paths.)
1285 2.0.2 - Minor updates and cleanups.
1287 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
1288 and cleanup the code a bit, removing the unused size parameter.
1289 - Change default fmask to 0177 and update documentation.
1290 - Change attrib.c::get_attr_search_ctx() to return the search context
1291 directly instead of taking the address of a pointer. A return value
1292 of NULL means the allocation failed. Updated all callers
1294 - Update to 2.5.9 kernel (preserving backwards compatibility) by
1295 replacing all occurences of page->buffers with page_buffers(page).
1296 - Fix minor bugs in runlist merging, also minor cleanup.
1297 - Updates to bootsector layout and mft mirror contents descriptions.
1298 - Small bug fix in error detection in unistr.c and some cleanups.
1299 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
1302 2.0.1 - Minor updates.
1304 - Make default umask correspond to documentation.
1305 - Improve documentation.
1306 - Set default mode to include execute bit. The {u,f,d}mask can be used
1307 to take it away if desired. This allows binaries to be executed from
1308 a mounted ntfs partition.
1310 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
1312 - Add kill_super, just keeping up with the vfs changes in the kernel.
1313 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
1314 from the CVS import into BitKeeper.
1315 - Begin to implement proper handling of allocated_size vs
1316 initialized_size vs data_size (i.e. i_size). Done are
1317 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
1318 and attrib.c::load_attribute_list().
1319 - Lock the runlist in attrib.c::load_attribute_list() while using it.
1320 - Fix memory leak in ntfs_file_read_compressed_block() and generally
1321 clean up compress.c a little, removing some uncommented/unused debug
1323 - Tidy up dir.c a little bit.
1324 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
1325 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
1326 creating aops.c::ntfs_mst_readpage(), improving the handling of
1327 holes and overflow in the process and implementing the correct
1328 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
1329 I am aiming for correctness at the moment. Modularisation can come
1331 - Rename aops.c::end_buffer_read_index_async() to
1332 end_buffer_read_mst_async() and optimize the overflow checking and
1334 - Use the host of the mftbmp address space mapping to hold the ntfs
1335 volume. This is needed so the async i/o completion handler can
1336 retrieve a pointer to the volume. Hopefully this will not cause
1337 problems elsewhere in the kernel... Otherwise will need to use a
1339 - Complete implementation of proper handling of allocated_size vs
1340 initialized_size vs data_size (i.e. i_size) in whole driver.
1341 Basically aops.c is now completely rewritten.
1342 - Change NTFS driver name to just NTFS and set version number to 2.0.0
1343 to make a clear distinction from the old driver which is still on
1346 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
1348 - Replace bdevname(sb->s_dev) with sb->s_id.
1349 - Remove now superfluous new-line characters in all callers of
1351 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
1352 directories. Without this the "find" utility gets very upset which is
1353 fair enough as Linux/Unix do not support directory hard links.
1354 - Further runlist merging work. (Richard Russon)
1355 - Backwards compatibility for gcc-2.95. (Richard Russon)
1356 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
1357 - Convert to new filesystem declaration using ->ntfs_get_sb() and
1358 replacing ntfs_read_super() with ntfs_fill_super().
1359 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
1360 overflow on 32-bit architectures.
1361 - Cleanup upcase loading code to use ntfs_(un)map_page().
1362 - Disable/reenable preemtion in critical sections of compession engine.
1363 - Replace device size determination in ntfs_fill_super() with
1364 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
1365 function super.c::get_nr_blocks().
1366 - Implement a mount time option (show_inodes) allowing choice of which
1367 types of inode names readdir() returns and modify ntfs_filldir()
1368 accordingly. There are several parameters to show_inodes:
1369 system: system files
1370 win32: long file names (including POSIX file names) [DEFAULT]
1372 dos: short file names only (excluding POSIX file names)
1374 posix: same as both win32 and dos
1376 Note that the options are additive, i.e. specifying:
1377 -o show_inodes=system,show_inodes=win32,show_inodes=dos
1378 is the same as specifying:
1380 Note that the "posix" and "all" options will show all directory
1381 names, BUT the link count on each directory inode entry is set to 1,
1382 due to Linux not supporting directory hard links. This may well
1383 confuse some userspace applications, since the directory names will
1384 have the same inode numbers. Thus it is NOT advisable to use the
1385 "posix" or "all" options. We provide them only for completeness sake.
1386 - Add copies of allocated_size, initialized_size, and compressed_size to
1387 the ntfs inode structure and set them up in
1388 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
1389 for files and the index allocation attribute for directories.
1390 - Add copies of allocated_size and initialized_size to ntfs inode for
1391 $BITMAP attribute of large directories and set them up in
1392 inode.c::ntfs_read_inode().
1393 - Add copies of allocated_size and initialized_size to ntfs volume for
1394 $BITMAP attribute of $MFT and set them up in
1395 super.c::load_system_files().
1396 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
1397 posix, and utf8) and tell user what the new options to use are. Note
1398 we still do support them but they will be removed with kernel 2.7.x.
1399 - Change all occurences of integer long long printf formatting to hex
1400 as printk() will not support long long integer format if/when the
1401 div64 patch goes into the kernel.
1402 - Make slab caches have stable names and change the names to what they
1403 were intended to be. These changes are required/made possible by the
1404 new slab cache name handling which removes the length limitation by
1405 requiring the caller of kmem_cache_create() to supply a stable name
1406 which is then referenced but not copied.
1407 - Rename run_list structure to run_list_element and create a new
1408 run_list structure containing a pointer to a run_list_element
1409 structure and a read/write semaphore. Adapt all users of runlists
1410 to new scheme and take and release the lock as needed. This fixes a
1411 nasty race as the run_list changes even when inodes are locked for
1412 reading and even when the inode isn't locked at all, so we really
1413 needed the serialization. We use a semaphore rather than a spinlock
1414 as memory allocations can sleep and doing everything GFP_ATOMIC
1416 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
1417 This can never happen due to the nature of lookup_attr() and how we
1418 support attribute lists. If it did happen it would imply the inode
1420 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
1422 - Update to 2.5.6-pre2 changes in struct address_space.
1423 - Use parent_ino() when accessing d_parent inode number in dir.c.
1424 - Import Sourceforge CVS repository into BitKeeper repository:
1425 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
1426 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
1427 Documentation/filesystems/ntfs.txt for NTFS TNG.
1428 - Create kernel configuration option controlling whether debugging
1430 - Add the required export of end_buffer_io_sync() from the patches
1431 directory to the kernel code.
1432 - Update inode.c::ntfs_show_options() with show_inodes mount option.
1433 - Update errors mount option.
1435 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
1437 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
1438 bug in mapping of extent inodes. Update all the comments to fit all
1439 the recent code changes.
1440 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
1441 - Cleanups in compress.c, mostly comments and folding help.
1442 - Implement attrib.c::map_run_list() as a generic helper.
1443 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
1444 thus making code shorter and enabling attribute list support.
1445 - Cleanup incorrect use of [su]64 with %L printf format specifier in
1446 all source files. Type casts to [unsigned] long long added to correct
1447 the mismatches (important for architectures which have long long not
1449 - Merge async io completion handlers for directory indexes and $MFT
1450 data into one by setting the index_block_size{_bits} of the ntfs
1451 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
1452 - Cleanup aops.c, update comments.
1453 - Make ntfs_file_get_block() use map_run_list() so all files now
1454 support attribute lists.
1455 - Make ntfs_dir_readpage() almost verbatim copy of
1456 block_read_full_page() by using ntfs_file_get_block() with only real
1457 difference being the use of our own async io completion handler
1458 rather than the default one, thus reducing the amount of code and
1459 automatically enabling attribute list support for directory indices.
1460 - Fix bug in load_attribute_list() - forgot to call brelse in error
1462 - Change parameters to find_attr() and lookup_attr(). We no longer
1463 pass in the upcase table and its length. These can be gotten from
1464 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
1465 - Cleanups in attrib.c.
1466 - Implement merging of runlists, attrib.c::merge_run_lists() and its
1467 helpers. (Richard Russon)
1468 - Attribute lists part 2, attribute extents and multi part runlists:
1469 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
1470 further runlist parts via attrib.c::map_run_list().
1471 - Tiny endianness bug fix in decompress_mapping_pairs().
1473 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
1475 - Enable encrypted directories. (Their index root is marked encrypted
1476 to indicate that new files in that directory should be created
1478 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
1479 - Enable $Extend system directory. Most (if not all) extended system
1480 files do not have unnamed data attributes so ntfs_read_inode() had to
1481 special case them but that is ok, as the special casing recovery
1482 happens inside an error code path so there is zero slow down in the
1483 normal fast path. The special casing is done by introducing a new
1484 function inode.c::ntfs_is_extended_system_file() which checks if any
1485 of the hard links in the inode point to $Extend as being their parent
1486 directory and if they do we assume this is an extended system file.
1487 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
1488 when compiled with -DDEBUG. Default is debug messages to be disabled.
1489 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
1490 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
1491 interface is enabled). Inspired by old ntfs driver.
1492 - Add debug_msgs insmod/kernel boot parameter to set whether debug
1493 messages are {dis,en}abled. This is useful to enable debug messages
1494 during ntfs initialization and is the only way to activate debugging
1495 when the sysctl interface is not enabled.
1496 - Cleanup debug output in various places.
1497 - Remove all dollar signs ($) from the source (except comments) to
1498 enable compilation on architectures whose gcc compiler does not
1499 support dollar signs in the names of variables/constants. Attribute
1500 types now start with AT_ instead of $ and $I30 is now just I30.
1501 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
1502 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
1503 access functions. This means we now cope with $MFT/$BITMAP being
1504 spread accross several mft records.
1505 - Disable modification of mft_zone_multiplier on remount. We can always
1506 reenable this later on if we really want to, but we will need to make
1507 sure we readjust the mft_zone size / layout accordingly.
1509 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
1511 - Use sb_set_blocksize() instead of set_blocksize() and verify the
1513 - Use sb_bread() instead of bread() throughout.
1514 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1515 of a directory index block vcn. Apply resulting simplifications in
1517 - Fix a small bug somewhere (but forgot what it was).
1518 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1519 on the printf-format parameter list and fix bugs reported by gcc
1520 as a result. (Richard Russon)
1521 - Move inode allocation strategy to Al's new stuff but maintain the
1522 divorce of ntfs_inode from struct inode. To achieve this we have two
1523 separate slab caches, one for big ntfs inodes containing a struct
1524 inode and pure ntfs inodes and at the same time fix some faulty
1525 error code paths in ntfs_read_inode().
1526 - Show mount options in proc (inode.c::ntfs_show_options()).
1528 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1530 - Modified (un)map_mft_record functions to be common for read and write
1531 case. To specify which is which, added extra parameter at front of
1532 parameter list. Pass either READ or WRITE to this, each has the
1534 - General cleanups to allow for easier folding in vi.
1535 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1536 argument, and invokes attrib.c::merge_run_lists() to merge the old
1537 and the new runlists.
1538 - Removed attrib.c::find_first_attr().
1539 - Implemented loading of attribute list and complete runlist for $MFT.
1540 This means we now cope with $MFT being spread across several mft
1542 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1543 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1544 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1545 cache. There are too little ntfs_volume structures at any one time
1546 to justify a private slab cache.
1547 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1548 Use KM_BIO_IRQ on advice from IRC/kernel...
1549 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1550 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1551 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1552 removed the now superfluous exports from the kernel core patch.
1553 - Fix a bug where kfree() was used insted of ntfs_free().
1554 - Change map_mft_record() to take ntfs_inode as argument instead of
1555 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1556 - Add pointer to ntfs_volume to ntfs_inode.
1557 - Add mft record number and sequence number to ntfs_inode. Stop using
1558 i_ino and i_generation for in-driver purposes.
1559 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1560 - Remove use of proper inodes by extent inodes. Move i_ino and
1561 i_generation to ntfs_inode to do this. Apply simplifications that
1562 result and remove iget_no_wait(), etc.
1563 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1564 - Add reference counting in ntfs_inode for the ntfs inode itself and
1565 for the mapped mft record.
1566 - Extend mft record mapping so we can (un)map extent mft records (new
1567 functions (un)map_extent_mft_record), and so mappings are reference
1568 counted and don't have to happen twice if already mapped - just ref
1570 - Add -o iocharset as alias to -o nls for backwards compatibility.
1571 - The latest core patch is now tiny. In fact just a single additional
1572 export is necessary over the base kernel.
1574 tng-0.0.3 - Cleanups, enhancements, bug fixes.
1576 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1577 and setup the runlist appropriately using knowledge provided by the
1578 sizes in the base attribute record.
1579 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1581 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1582 page or use vmalloc depending on the amount of memory requested.
1583 - Cleanup error output. The __FUNCTION__ "(): " is now added
1584 automatically. Introduced a new header file debug.h to support this
1585 and also moved ntfs_debug() function into it.
1586 - Make reading of compressed files more intelligent and especially get
1587 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1588 buffers (allocated at first mount with cluster size <= 4kiB and
1589 deallocated on last umount with cluster size <= 4kiB), and
1590 asynchronous io for the compressed data using a list of buffer heads.
1591 Er, we use synchronous io as async io only works on whole pages
1592 covered by buffers and not on individual buffer heads...
1593 - Bug fix for reading compressed files with sparse compression blocks.
1595 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1597 - Fixed handling of directories when cluster size exceeds index block
1599 - Hide DOS only name space directory entries from readdir() but allow
1600 them in lookup(). This should fix the problem that Linux doesn't
1601 support directory hard links, while still allowing access to entries
1602 via their short file name. This also has the benefit of mimicking
1603 what Windows users are used to, so it is the ideal solution.
1604 - Implemented sync_page everywhere so no more hangs in D state when
1606 - Stop using bforget() in favour of brelse().
1607 - Stop locking buffers unnecessarily.
1608 - Implemented compressed files (inode->mapping contains uncompressed
1609 data, raw compressed data is currently bread() into a vmalloc()ed
1611 - Enable compressed directories. (Their index root is marked compressed
1612 to indicate that new files in that directory should be created
1614 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1615 functions. (Thanks to Will Dyson for pointing this out.)
1616 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1617 ntfs_sb_info) out of the common inode and super_block structures and
1618 started using the generic_ip and generic_sbp pointers instead. This
1619 makes ntfs entirely private with respect to the kernel tree.
1620 - Detect compiler version and abort with error message if gcc less than
1622 - Fix bug in name comparison function in unistr.c.
1623 - Implement attribute lists part 1, the infrastructure: search contexts
1624 and operations, find_external_attr(), lookup_attr()) and make the
1625 code use the infrastructure.
1626 - Fix stupid buffer overflow bug that became apparent on larger run
1627 list containing attributes.
1628 - Fix bugs in readdir() that became apparent on larger directories.
1630 The driver is now really useful and survives the test
1631 find . -type f -exec md5sum "{}" \;
1632 without any error messages on a over 1GiB sized partition with >16k
1633 files on it, including compressed files and directories and many files
1634 and directories with attribute lists.
1636 tng-0.0.1 - The first useful version.
1638 - Added ntfs_lookup().
1639 - Added default upcase generation and handling.
1640 - Added compile options to be shown on module init.
1641 - Many bug fixes that were "hidden" before.
1642 - Update to latest kernel.
1643 - Added ntfs_readdir().
1644 - Added file operations for mmap(), read(), open() and llseek(). We just
1645 use the generic ones. The whole point of going through implementing
1646 readpage() methods and where possible get_block() call backs is that
1647 this allows us to make use of the generic high level methods provided
1650 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1651 though and it doesn't implement accesssing compressed files yet. Also,
1652 accessing files with attribute list attributes is not implemented yet
1653 either. But for small or simple filesystems it should work and allow
1654 you to list directories, use stat on directory entries and the file
1655 system, open, read, mmap and llseek around in files. A big mile stone
1658 tng-0.0.0 - Initial version tag.
1660 Initial driver implementation. The driver can mount and umount simple
1661 NTFS filesystems (i.e. ones without attribute lists in the system
1662 files). If the mount fails there might be problems in the error handling
1663 code paths, so be warned. Otherwise it seems to be loading the system
1664 files nicely and the mft record read mapping/unmapping seems to be
1665 working nicely, too. Proof of inode metadata in the page cache and non-
1666 resident file unnamed stream data in the page cache concepts is thus
projects / linux-2.6-omap-h63xx.git / blob