3 - In between ntfs_prepare/commit_write, need exclusion between
4 simultaneous file extensions. This is given to us by holding i_sem
5 on the inode. The only places in the kernel when a file is resized
6 are prepare/commit write and truncate for both of which i_sem is
7 held. Just have to be careful in readpage/writepage and all other
8 helpers not running under i_sem that we play nice...
9 Also need to be careful with initialized_size extention in
10 ntfs_prepare_write. Basically, just be _very_ careful in this code...
11 UPDATE: The only things that need to be checked are read/writepage
12 which do not hold i_sem. Note writepage cannot change i_size but it
13 needs to cope with a concurrent i_size change, just like readpage.
14 Also both need to cope with concurrent changes to the other sizes,
15 i.e. initialized/allocated/compressed size, as well.
16 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
17 leave the volume dirty on umount if the final iput(vol->mft_ino)
18 causes a write of any mirrored mft records due to the mft mirror
19 inode having been discarded already. Whether this can actually ever
20 happen is unclear however so it is worth waiting until someone hits
22 - Enable the code for setting the NT4 compatibility flag when we start
23 making NTFS 1.2 specific modifications.
27 - Support journals ($LogFile) which have been modified by chkdsk. This
28 means users can boot into Windows after we marked the volume dirty.
29 The Windows boot will run chkdsk and then reboot. The user can then
30 immediately boot into Linux rather than having to do a full Windows
31 boot first before rebooting into Linux and we will recognize such a
32 journal and empty it as it is clean by definition.
33 - Support journals ($LogFile) with only one restart page as well as
34 journals with two different restart pages. We sanity check both and
35 either use the only sane one or the more recent one of the two in the
36 case that both are valid.
37 - Modify fs/ntfs/malloc.h::ntfs_malloc_nofs() to do the kmalloc() based
38 allocations with __GFP_HIGHMEM, analogous to how the vmalloc() based
40 - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
41 ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
43 - Use ntfs_malloc_nofs_nofail() in the two critical regions in
44 fs/ntfs/runlist.c::ntfs_runlists_merge(). This means we no longer
45 need to panic() if the allocation fails as it now cannot fail.
46 - Fix two nasty runlist merging bugs that had gone unnoticed so far.
47 Thanks to Stefano Picerno for the bug report.
48 - Remove two bogus BUG_ON()s from fs/ntfs/mft.c.
50 2.1.23 - Implement extension of resident files and make writing safe as well as
51 many bug fixes, cleanups, and enhancements...
53 - Add printk rate limiting for ntfs_warning() and ntfs_error() when
54 compiled without debug. This avoids a possible denial of service
55 attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
57 - Fix compilation warnings on ia64. (Randy Dunlap)
58 - Use i_size_{read,write}() instead of reading i_size by hand and cache
59 the value where apropriate.
60 - Add size_lock to the ntfs_inode structure. This is an rw spinlock
61 and it locks against access to the inode sizes. Note, ->size_lock
62 is also accessed from irq context so you must use the _irqsave and
63 _irqrestore lock and unlock functions, respectively. Protect all
64 accesses to allocated_size, initialized_size, and compressed_size.
65 - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
66 - Implement extension of resident files in the regular file write code
67 paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()). At present
68 this only works until the data attribute becomes too big for the mft
69 record after which we abort the write returning -EOPNOTSUPP from
71 - Add disable_sparse mount option together with a per volume sparse
72 enable bit which is set appropriately and a per inode sparse disable
73 bit which is preset on some system file inodes as appropriate.
74 - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
75 - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
76 the creation of the unmapped runlist element for the base attribute
78 - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
79 helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
80 This allows us to map runlist fragments with the runlist lock already
81 held without having to drop and reacquire it around the call. Adapt
83 - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
84 runlist. This allows us to find runlist elements with the runlist
85 lock already held without having to drop and reacquire it around the
86 call. Adapt all callers.
87 - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
88 warning in the do_div() call on sparc32. Thanks to Meelis Roos for
89 the report and analysis of the warning.
90 - Fix a nasty runlist merge bug when merging two holes.
91 - Set the ntfs_inode->allocated_size to the real allocated size in the
92 mft record for resident attributes (fs/ntfs/inode.c).
93 - Small readability cleanup to use "a" instead of "ctx->attr"
94 everywhere (fs/ntfs/inode.c).
95 - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
96 definition of ntfs_export_ops from fs/ntfs/super.c to namei.c. Also,
97 declare ntfs_export_ops in fs/ntfs/ntfs.h.
98 - Correct sparse file handling. The compressed values need to be
99 checked and set in the ntfs inode as done for compressed files and
100 the compressed size needs to be used for vfs inode->i_blocks instead
101 of the allocated size, again, as done for compressed files.
102 - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
103 non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
104 - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
106 - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
107 dropping the read lock and taking the write lock we were not checking
108 whether someone else did not already do the work we wanted to do.
109 - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
110 ntfs_attr_find_vcn_nolock() and update all callers.
111 - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
112 - Fix sign of various error return values to be negative in
114 - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
115 handle the case where an attribute is converted from resident to
116 non-resident by a concurrent file write.
117 - Remove checks for NULL before calling kfree() since kfree() does the
118 checking itself. (Jesper Juhl)
119 - Some utilities modify the boot sector but do not update the checksum.
120 Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
121 only emit a warning when the checksum is incorrect rather than
122 refusing the mount. Thanks to Bernd Casimir for pointing this
124 - Update attribute definition handling.
125 - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
126 - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
127 - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
128 better code generation and one less sparse warning in fs/ntfs/aops.c.
129 - Remove spurious void pointer casts from fs/ntfs/. (Pekka Enberg)
130 - Use C99 style structure initialization after memory allocation where
131 possible (fs/ntfs/{attrib.c,index.c,super.c}). Thanks to Al Viro and
133 - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
134 is active on the volume and we are mounting read-write or remounting
135 from read-only to read-write.
136 - Fix a bug in address space operations error recovery code paths where
137 if the runlist was not mapped at all and a mapping error occured we
138 would leave the runlist locked on exit to the function so that the
139 next access to the same file would try to take the lock and deadlock.
140 - Detect the case when Windows has been suspended to disk on the volume
141 to be mounted and if this is the case do not allow (re)mounting
142 read-write. This is done by parsing hiberfil.sys if present.
143 - Fix several occurences of a bug where we would perform 'var & ~const'
144 with a 64-bit variable and a int, i.e. 32-bit, constant. This causes
145 the higher order 32-bits of the 64-bit variable to be zeroed. To fix
146 this cast the 'const' to the same 64-bit type as 'var'.
147 - Change the runlist terminator of the newly allocated cluster(s) to
148 LCN_ENOENT in ntfs_attr_make_non_resident(). Otherwise the runlist
150 - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
151 and ntfs_mapping_pairs_build() to allow the runlist encoding to be
152 partial which is desirable when filling holes in sparse attributes.
154 - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
155 if the requested vcn is inside it. Otherwise we get into problems
156 when we try to map an out of bounds vcn because we then try to map
157 the already mapped runlist fragment which causes
158 ntfs_mapping_pairs_decompress() to fail and return error. Update
159 ntfs_attr_find_vcn_nolock() accordingly.
160 - Fix a nasty deadlock that appeared in recent kernels.
161 The situation: VFS inode X on a mounted ntfs volume is dirty. For
162 same inode X, the ntfs_inode is dirty and thus corresponding on-disk
163 inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
164 to the table of inodes, i.e. $MFT, inode 0.
166 Process 1: sys_sync()/umount()/whatever... calls
167 __sync_single_inode() for $MFT -> do_writepages() -> write_page for
168 the dirty page containing the on-disk inode X, the page is now locked
169 -> ntfs_write_mst_block() which clears PageUptodate() on the page to
170 prevent anyone else getting hold of it whilst it does the write out.
171 This is necessary as the on-disk inode needs "fixups" applied before
172 the write to disk which are removed again after the write and
173 PageUptodate is then set again. It then analyses the page looking
174 for dirty on-disk inodes and when it finds one it calls
175 ntfs_may_write_mft_record() to see if it is safe to write this
176 on-disk inode. This then calls ilookup5() to check if the
177 corresponding VFS inode is in icache(). This in turn calls ifind()
178 which waits on the inode lock via wait_on_inode whilst holding the
180 Process 2: pdflush results in a call to __sync_single_inode for the
181 same VFS inode X on the ntfs volume. This locks the inode (I_LOCK)
182 then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
183 read_cache_page() for the page (in page cache of table of inodes
184 $MFT, inode 0) containing the on-disk inode. This page has
185 PageUptodate() clear because of Process 1 (see above) so
186 read_cache_page() blocks when it tries to take the page lock for the
187 page so it can call ntfs_read_page().
188 Thus Process 1 is holding the page lock on the page containing the
189 on-disk inode X and it is waiting on the inode X to be unlocked in
190 ifind() so it can write the page out and then unlock the page.
191 And Process 2 is holding the inode lock on inode X and is waiting for
192 the page to be unlocked so it can call ntfs_readpage() or discover
193 that Process 1 set PageUptodate() again and use the page.
194 Thus we have a deadlock due to ifind() waiting on the inode lock.
195 The solution: The fix is to use the newly introduced
196 ilookup5_nowait() which does not wait on the inode's lock and hence
197 avoids the deadlock. This is safe as we do not care about the VFS
198 inode and only use the fact that it is in the VFS inode cache and the
199 fact that the vfs and ntfs inodes are one struct in memory to find
200 the ntfs inode in memory if present. Also, the ntfs inode has its
201 own locking so it does not matter if the vfs inode is locked.
202 - Fix bug in mft record writing where we forgot to set the device in
203 the buffers when mapping them after the VM had discarded them.
204 Thanks to Martin MOKREJÅ for the bug report.
206 2.1.22 - Many bug and race fixes and error handling improvements.
208 - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
209 - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
210 instead of void and provide a helper ntfs_truncate_vfs() for the
211 vfs ->truncate method.
212 - Add a new ntfs inode flag NInoTruncateFailed() and modify
213 fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
214 - Fix min_size and max_size definitions in ATTR_DEF structure in
215 fs/ntfs/layout.h to be signed.
216 - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
217 ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
218 ntfs_attr_can_be_resident(), which in turn use the new private helper
219 ntfs_attr_find_in_attrdef().
220 - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
221 mapping->private_lock around the dirtying of the buffer heads
222 analagous to the way it is done in __set_page_dirty_buffers().
223 - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
224 mount time as this cannot work with the current implementation.
225 - Check for location of attribute name and improve error handling in
226 general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
227 - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
228 i_size, i.e. race with truncate, invalidate the buffers on the page
229 so that they become freeable and hence the page does not leak.
230 - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
232 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
233 a NULL pointer dereference in the error code path when a corrupt
234 attribute was found. (Thanks to Domen Puncer for the bug report.)
235 - Add MODULE_VERSION() to fs/ntfs/super.c.
236 - Make several functions and variables static. (Adrian Bunk)
237 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
238 buffers for the page if they are not present and then marks the
239 buffers belonging to the ntfs record dirty. This causes the buffers
240 to become busy and hence they are safe from removal until the page
241 has been written out.
242 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
243 error handling code path that resulted in a BUG() due to trying to
244 unmap an extent mft record when the mapping of it had failed and it
245 thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
246 - Drop the runlist lock after the vcn has been read in
247 fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
248 - Rewrite handling of multi sector transfer errors. We now do not set
249 PageError() when such errors are detected in the async i/o handler
250 fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
251 protected attributes now check the magic of each ntfs record as they
252 use it and act appropriately. This has the effect of making errors
253 granular per ntfs record rather than per page which solves the case
254 where we cannot access any of the ntfs records in a page when a
255 single one of them had an mst error. (Thanks to Ken MacFerrin for
257 - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
258 where we failed to release i_sem on the $Quota/$Q attribute inode.
259 - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
260 - Add mapping of unmapped buffers to all remaining code paths, i.e.
261 fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
262 and write_mft_record_nolock(). From now on we require that the
263 complete runlist for the mft mirror is always mapped into memory.
264 - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
265 - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
266 - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
267 resident attribute will be smaller than a page which makes the code
268 simpler. Also make the code more tolerant to concurrent ->truncate.
270 2.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
272 - Implement extent mft record deallocation
273 fs/ntfs/mft.c::ntfs_extent_mft_record_free().
274 - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
275 - Add vol->mft_data_pos and initialize it at mount time.
276 - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
277 ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
278 ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
279 ntfs_runlists_merge() and adapt all callers.
280 - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
281 ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
282 and ntfs_mapping_pairs_build(), adapted from libntfs.
283 - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
284 static and add a declaration for it to lcnalloc.h.
285 - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
286 inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
287 cluster bitmap lock for the duration of the call.
288 - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
289 - Implement the equivalent of memset() for an ntfs attribute in
290 fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
291 fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
292 - Remove unnecessary casts from LCN_* constants.
293 - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
294 - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
295 change MFT_RECORD to contain the NTFS 3.1+ specific fields.
296 - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
297 marks all buffers belonging to an ntfs record dirty, followed by
298 marking the page the ntfs record is in dirty and also marking the vfs
299 inode containing the ntfs record dirty (I_DIRTY_PAGES).
300 - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
301 new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
302 longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
303 - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
305 - Move the typedefs for runlist_element and runlist from types.h to
306 runlist.h and fix resulting include errors.
307 - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
308 - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
309 mark_ntfs_record_dirty() which also changes the behaviour in that we
310 now set the buffers belonging to the mft record dirty as well as the
312 - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
313 to cope with the fact that there now are dirty buffers in mft pages.
314 - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
315 mark_ntfs_record_dirty() and thus to set the buffers belonging to the
316 mft record dirty as well as the page itself.
317 - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
318 slightly modified by me)
319 - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
320 the mft record is already locked and otherwise behaves the same way
321 as fs/ntfs/mft.c::map_mft_record().
322 - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
323 writes the mft record if the buffers belonging to it are dirty.
324 Otherwise we assume that it was written out by other means already.
325 - Attempting to write outside initialized size is _not_ a bug so remove
326 the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
327 fact required to write outside initialized size when preparing to
328 extend the initialized size.
329 - Map the page instead of using page_address() before writing to it in
330 fs/ntfs/aops.c::ntfs_mft_writepage().
331 - Provide exclusion between opening an inode / mapping an mft record
332 and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
333 by setting the page not uptodate throughout ntfs_mft_writepage().
334 - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
335 to ensure noone can see the page whilst the mst fixups are applied.
336 - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
337 checks if an mft record may be written out safely obtaining any
338 necessary locks in the process. This is used by
339 fs/ntfs/aops.c::ntfs_write_mst_block().
340 - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
341 writing mft records and improve its error handling in the process.
342 Now if any of the records in the page fail to be written out, all
343 other records will be written out instead of aborting completely.
344 - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
345 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
346 ntfs_mst_aops for all inodes which are NInoMstProtected() and
347 ntfs_aops for all other inodes.
348 - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
349 ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
350 no longer require an ntfs inode to be present. Update all callers.
351 - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
352 - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
353 to ensure noone can see the page whilst the mst fixups are applied.
354 - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
355 fs/ntfs/mft.c::try_map_mft_record().
356 - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
357 with the ntfs inode which contains the page rather than the ntfs
358 inode the mft record of which is in the page.
359 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
360 index inode bitmap inode release code from there to
361 fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
362 Hellwig for spotting this.)
363 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
364 inode semaphore around the code that sets ni->itype.index.bmp_ino to
365 NULL and reorganize the code to optimize it a bit. (Thanks to
366 Christoph Hellwig for spotting this.)
367 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
368 ntfs inode as a parameter as this is confusing and misleading and the
369 needed ntfs inode is available via NTFS_I(page->mapping->host).
370 Adapt all callers to this change.
371 - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
372 fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
373 of the first buffer in a record and to take this as the ntfs record
374 dirty state. We cannot look at the dirty state for subsequent
375 buffers because we might be racing with
376 fs/ntfs/aops.c::mark_ntfs_record_dirty().
377 - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
378 inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
379 add a declaration for it to inode.h. Fix some compilation issues
380 that resulted due to #includes and header file interdependencies.
381 - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
382 - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
383 - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
384 record sequence number if it is specified (i.e. not zero).
385 - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
386 functions used by it.
387 - Update Documentation/filesystems/ntfs.txt with instructions on how to
388 use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
389 the linear raid problem with the Software RAID / MD driver when one
390 or more of the devices has an odd number of sectors.
392 2.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
394 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
395 where we did not clear ctx->al_entry but it was still set due to
396 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
398 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
399 where we forgot to unmap the extent mft record when we had finished
400 enumerating an attribute which caused a bug check to trigger when the
401 VFS calls ->clear_inode.
403 2.1.19 - Many cleanups, improvements, and a minor bug fix.
405 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
406 change the uid, gid, and mode of an inode as we do not support NTFS
408 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
410 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
411 and ntfs_filldir() as per suggestion from Al Viro.
412 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
413 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
414 inode size has changed and to only output an error if so.
415 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
416 - Add le{16,32,64} as well as sle{16,32,64} data types to
418 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
419 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
420 respectively, to fs/ntfs/types.h.
421 - Update endianness conversion macros in fs/ntfs/endian.h to use the
422 new types as appropriate.
423 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
425 - Add leMFT_REF data type to fs/ntfs/layout.h.
426 - Update all NTFS header files with the new little endian data types.
427 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
428 - Do proper type casting when using ntfs_is_*_recordp() in
429 fs/ntfs/logfile.c, mft.c, and super.c.
430 - Fix all the sparse bitwise warnings. Had to change all the typedef
431 enums storing little endian values to simple enums plus a typedef for
432 the datatype to make sparse happy.
433 - Fix a bug found by the new sparse bitwise warnings where the default
434 upcase table was defined as a pointer to wchar_t rather than ntfschar
435 in fs/ntfs/ntfs.h and super.c.
436 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
438 2.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
440 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
441 the calculation of total/free inodes as used by statfs().
442 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
443 because the code itself is using the ntfs_lock semaphore which
444 provides safe locking. (Ingo Molnar)
445 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
446 could occur in the future for when we start closing/freeing extent
447 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
449 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
450 find_external_attr() to ntfs_external_attr_find() to cleanup the
451 namespace a bit and to be more consistent with libntfs.
452 - Rename {{re,}init,get,put}_attr_search_ctx() to
453 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
454 attr_search_context to ntfs_attr_search_ctx.
455 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
456 for the attribute list attribute itself.
457 - Fix endianness bug in ntfs_external_attr_find().
458 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
459 if the attribute is not found, and -EIO on real error. In the case
460 of -ENOENT, the search context is updated to describe the attribute
461 before which the attribute being searched for would need to be
462 inserted if such an action were to be desired and in the case of
463 ntfs_external_attr_find() the search context is also updated to
464 indicate the attribute list entry before which the attribute list
465 entry of the attribute being searched for would need to be inserted
466 if such an action were to be desired. Also make ntfs_find_attr()
467 static and remove its prototype from attrib.h as it is not used
468 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
469 callers of ntfs_{external,}attr_{find,lookup}() for the new return
471 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
473 2.1.17 - Fix bugs in mount time error code paths and other updates.
475 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
476 includes functions to set/clear a single bit or a run of bits.
477 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
478 runlist element containing a particular vcn. It also takes care of
479 mapping any needed runlist fragments.
480 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
481 - Load attribute definition table from $AttrDef at mount time.
482 - Fix bugs in mount time error code paths involving (de)allocation of
483 the default and volume upcase tables.
484 - Remove ntfs_nr_mounts as it is no longer used.
486 2.1.16 - Implement access time updates, file sync, async io, and read/writev.
488 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
489 This is done by setting the appropriate file operations pointers to
490 the generic helper functions provided by mm/filemap.c.
491 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
492 and directories (fs/ntfs/dir.c).
493 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
494 Note, except for the root directory and any other system files opened
495 by the user, the system files will not have their access times
496 updated as they are only accessed at the inode level an hence the
497 file level functions which cause the times to be updated are never
500 2.1.15 - Invalidate quotas when (re)mounting read-write.
502 - Add new element itype.index.collation_rule to the ntfs inode
503 structure and set it appropriately in ntfs_read_locked_inode().
504 - Implement a new inode type "index" to allow efficient access to the
505 indices found in various system files and adapt inode handling
506 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
507 attribute inode (NInoAttr() is true) with an attribute type of
508 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
509 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
510 there would be no way to distinguish between normal attribute inodes
511 and index inodes. The function to obtain an index inode is
512 ntfs_index_iget() and it uses the helper function
513 ntfs_read_locked_index_inode(). Note, we do not overload
514 ntfs_attr_iget() as indices consist of multiple attributes so using
515 ntfs_attr_iget() to obtain an index inode would be confusing.
516 - Ensure that there is no overflow when doing page->index <<
517 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
518 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
519 and ntfs_read_block().
520 - Use case sensitive attribute lookups instead of case insensitive ones.
521 - Lock all page cache pages belonging to mst protected attributes while
522 accessing them to ensure we never see corrupt data while the page is
524 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
525 We have ntfs_is_collation_rule_supported() to check if the collation
526 rule you want to use is supported and ntfs_collation() which actually
527 collates two data items. We currently only support COLLATION_BINARY
528 and COLLATION_NTOFS_ULONG but support for other collation rules will
529 be added as the need arises.
530 - Add a new type, ntfs_index_context, to allow retrieval of an index
531 entry using the corresponding index key. To get an index context,
532 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
533 This also adds a new slab cache for the index contexts. To lookup a
534 key in an index inode, use ntfs_index_lookup(). After modifying an
535 index entry, call ntfs_index_entry_flush_dcache_page() followed by
536 ntfs_index_entry_mark_dirty() to ensure the changes are written out
537 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
538 an index entry is in the index allocation attribute rather than the
539 index root attribute it will not be written out (you will get a
540 warning message about discarded changes instead).
541 - Load the quota file ($Quota) and check if quota tracking is enabled
542 and if so, mark the quotas out of date. This causes windows to
543 rescan the volume on boot and update all quota entries.
544 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
545 It is simply set to __set_page_dirty_nobuffers() to make sure that
546 running set_page_dirty() on a page containing mft/ntfs records will
547 not affect the dirty state of the page buffers.
548 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
549 buffers that are inside the ntfs record in the page dirty after which
550 it sets the page dirty. This allows ->writepage to only write the
551 dirty index records rather than having to write all the records in
552 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
553 use this rather than __set_page_dirty_nobuffers().
554 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
555 writing of page cache pages belonging to mst protected attributes
556 like the index allocation attribute in directory indices and other
557 indices like $Quota/$Q, etc. This means that the quota is now marked
558 out of date on all volumes rather than only on ones where the quota
559 defaults entry is in the index root attribute of the $Quota/$Q index.
561 2.1.14 - Fix an NFSd caused deadlock reported by several users.
563 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
564 to a buffer so that we can put the search context and unmap the mft
565 record before calling the filldir() callback. We need to do this
566 because of NFSd which calls ->lookup() from its filldir callback()
567 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
568 of the directory and since ntfs_readdir() has got it mapped already
569 ntfs_lookup() deadlocks.
571 2.1.13 - Enable overwriting of resident files and housekeeping of system files.
573 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
574 keeping the mft mirror in sync with the mft when mirrored mft records
575 are written. The functions are write_mft_record{,_nolock}(). The
576 implementation is quite rudimentary for now with lots of things not
577 implemented yet but I am not sure any of them can actually occur so
578 I will wait for people to hit each one and only then implement it.
579 - Commit open system inodes at umount time. This should make it
580 virtually impossible for sync_mft_mirror_umount() to ever be needed.
581 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
582 ntfs super operations. This gives us inode writing via the VFS inode
583 dirty code paths. Note: Access time updates are not implemented yet.
584 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
585 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
586 finally enabling resident file overwrite! (-8 This also includes a
587 placeholder for ->writepage (ntfs_mft_writepage()), which for now
588 just redirties the page and returns. Also, at umount time, we for
589 now throw away all mft data page cache pages after the last call to
590 ntfs_commit_inode() in the hope that all inodes will have been
591 written out by then and hence no dirty (meta)data will be lost. We
592 also check for this case and emit an error message telling the user
594 - Use set_page_writeback() and end_page_writeback() in the resident
595 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
596 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
598 - Implement ntfs_mft_writepage() so it now checks if any of the mft
599 records in the page are dirty and if so redirties the page and
600 returns. Otherwise it just returns (after doing set_page_writeback(),
601 unlock_page(), end_page_writeback() or the radix-tree tag
602 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
603 alowing the VM to do with the page as it pleases. Also, at umount
604 time, now only throw away dirty mft (meta)data pages if dirty inodes
605 are present and ask the user to email us if they see this happening.
606 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
607 information flags (fs/ntfs/super.c).
608 - Mark the volume dirty when (re)mounting read-write and mark it clean
609 when unmounting or remounting read-only. If any volume errors are
610 found, the volume is left marked dirty to force chkdsk to run.
611 - Add code to set the NT4 compatibility flag when (re)mounting
612 read-write for newer NTFS versions but leave it commented out for now
613 since we do not make any modifications that are NTFS 1.2 specific yet
614 and since setting this flag breaks Captive-NTFS which is not nice.
615 This code must be enabled once we start writing NTFS 1.2 specific
616 changes otherwise Windows NTFS driver might crash / cause corruption.
618 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
620 - Add a new address space operations struct, ntfs_mst_aops, for mst
621 protected attributes. This is because the default ntfs_aops do not
622 make sense with mst protected data and were they to write anything to
623 such an attribute they would cause data corruption so we provide
624 ntfs_mst_aops which does not have any write related operations set.
625 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
626 includes an adapted ntfs_commit_inode() and an implementation of
627 ntfs_write_inode() which for now just cleans dirty inodes without
628 writing them (it does emit a warning that this is happening).
629 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
630 entry) as it was only fixing a theoretical bug but at the same time
631 it badly broke the handling of sparse and uncompressed compression
634 2.1.11 - Driver internal cleanups.
636 - Only build logfile.o if building the driver with read-write support.
637 - Really final white space cleanups.
638 - Use generic_ffs() instead of ffs() in logfile.c which allows the
639 log_page_size variable to be optimized by gcc into a constant.
640 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
641 char as defined by POSIX and as found on some systems.
643 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
645 - Finish off the white space cleanups (remove trailing spaces, etc).
646 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
647 the kludges around the first iget(). Instead of (re)setting ->s_op
648 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
649 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
650 kills the need for second super_operations and allows to return error
651 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
653 - Force read-only (re)mounting if any of the following bits are set in
654 the volume information flags:
655 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
656 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
657 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
658 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
659 above bits set so the test is made easy.
661 2.1.9 - Fix two bugs in decompression engine.
663 - Fix a bug where we would not always detect that we have reached the
664 end of a compression block because we were ending at minus one byte
665 which is effectively the same as being at the end. The fix is to
666 check whether the uncompressed buffer has been fully filled and if so
667 we assume we have reached the end of the compression block. A big
668 thank you to Marcin Gibuła for the bug report, the assistance in
669 tracking down the bug and testing the fix.
670 - Fix a possible bug where when a compressed read is truncated to the
671 end of the file, the offset inside the last page was not truncated.
673 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
675 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
676 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
677 utc2ntfs() to work with struct timespec instead of time_t on the
678 Linux UTC time side thus preserving the full precision of the NTFS
679 time and only loosing up to 99 nano-seconds in the Linux UTC time.
680 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
682 - Remove unused ntfs_dirty_inode().
683 - Cleanup super operations declaration in fs/ntfs/super.c.
684 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
685 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
686 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
687 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
688 fs/ntfs/inode.h so they can be used elsewhere.
689 - Determine the mft mirror size as the number of mirrored mft records
690 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
691 - Load the mft mirror at mount time and compare the mft records stored
692 in it to the ones in the mft. Force a read-only mount if the two do
693 not match (fs/ntfs/super.c).
694 - Fix type casting related warnings on 64-bit architectures. Thanks
695 to Meelis Roos for reporting them.
696 - Move %L to %ll as %L is floating point and %ll is integer which is
698 - Read the journal ($LogFile) and determine if the volume has been
699 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
700 and fs/ntfs/logfile.c). This is a little bit of a crude check in
701 that we only look at the restart areas and not at the actual log
702 records so that there will be a very small number of cases where we
703 think that a volume is dirty when in fact it is clean. This should
704 only affect volumes that have not been shutdown cleanly and did not
705 have any pending, non-check-pointed i/o.
706 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
707 is requested, empty $LogFile by overwriting it with 0xff bytes to
708 ensure that Windows cannot cause data corruption by replaying a stale
709 journal after Linux has written to the volume.
711 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
713 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
714 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
715 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
716 default doesn't allow inode number 0 which is a valid inode on NTFS
717 and even if it did allow that it uses iget() instead of ntfs_iget()
718 which makes it useless for us.
719 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
720 default just returns -EACCES which is not very useful.
721 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
722 and set them up in the super block at mount time (super.c) this
723 allows mounted NTFS volumes to be exported via NFS.
724 - Add missing return -EOPNOTSUPP; in
725 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
726 - Enforce no atime and no dir atime updates at mount/remount time as
727 they are not implemented yet anyway.
728 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
729 after a NULL check. Thanks to Dave Jones for pointing this out.
731 2.1.6 - Fix minor bug in handling of compressed directories.
733 - Fix bug in handling of compressed directories. A compressed
734 directory is not really compressed so when we set the ->i_blocks
735 field of a compressed directory inode we were setting it from the
736 non-existing field ni->itype.compressed.size which gave random
737 results... For directories we now always use ni->allocated_size.
739 2.1.5 - Fix minor bug in attribute list attribute handling.
741 - Fix bug in attribute list handling. Actually it is not as much a bug
742 as too much protection in that we were not allowing attribute lists
743 which waste space on disk while Windows XP clearly allows it and in
744 fact creates such attribute lists so our driver was failing.
745 - Update NTFS documentation ready for 2.6 kernel release.
747 2.1.4 - Reduce compiler requirements.
749 - Remove all uses of unnamed structs and unions in the driver to make
750 old and newer gcc versions happy. Makes it a bit uglier IMO but at
751 least people will stop hassling me about it.
753 2.1.3 - Important bug fixes in corner cases.
755 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
756 clusters. (Philipp Thomas)
757 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
758 multiple of the block_size but not the cluster size. (Szabolcs
759 Szakacsits <szaka@sienet.hu>)
761 2.1.2 - Important bug fixes aleviating the hangs in statfs.
763 - Fix buggy free cluster and free inode determination logic.
765 2.1.1 - Minor updates.
767 - Add handling for initialized_size != data_size in compressed files.
768 - Reduce function local stack usage from 0x3d4 bytes to just noise in
769 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
770 - Remove compiler warnings for newer gcc.
771 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
772 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
773 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
774 kmap_atomic(KM_USER0).
776 2.1.0 - First steps towards write support: implement file overwrite.
778 - Add configuration option for developmental write support with an
779 appropriately scary configuration help text.
780 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
781 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
782 overwriting of existing files on ntfs. Note: Resident files are
783 only written into memory, and not written out to disk at present, so
784 avoid writing to files smaller than about 1kiB.
785 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
786 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
787 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
788 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
789 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
790 This enables write(2) based overwriting of existing files on ntfs.
791 Note: As with mmap(2) based overwriting, resident files are only
792 written into memory, and not written out to disk at present, so avoid
793 writing to files smaller than about 1kiB.
794 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
795 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
796 files with the purpose of intercepting and aborting all i_size
797 changes which we do not support yet. ntfs_truncate() actually only
798 emits a warning message but AFAICS our interception of i_size changes
799 elsewhere means ntfs_truncate() never gets called for i_size changes.
800 It is only called from generic_file_write() when we fail in
801 ntfs_prepare_{,nonresident_}write() in order to discard any
802 instantiated buffers beyond i_size. Thus i_size is not actually
803 changed so our warning message is enough. Unfortunately it is not
804 possible to easily determine if i_size is being changed or not hence
805 we just emit an appropriately worded error message.
807 2.0.25 - Small bug fixes and cleanups.
809 - Unlock the page in an out of memory error code path in
810 fs/ntfs/aops.c::ntfs_read_block().
811 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
812 just unlock the page and return. (This can happen due to ->writepage
813 clearing PageUptodate() during write out of MstProtected()
815 - Remove leaked write code again.
819 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
820 inside BUG_ON(). (Adam J. Richter)
821 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
822 calls for improved debugging. (Adam J. Richter)
823 - Add errors flag to the ntfs volume state, accessed via
824 NVol{,Set,Clear}Errors(vol).
825 - Do not allow read-write remounts of read-only volumes with errors.
826 - Clarify comment for ntfs file operation sendfile which was added by
827 Christoph Hellwig a while ago (just using generic_file_sendfile())
828 to say that ntfs ->sendfile is only used for the case where the
829 source data is on the ntfs partition and the destination is
830 somewhere else, i.e. nothing we need to concern ourselves with.
831 - Add generic_file_write() as our ntfs file write operation.
833 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
835 - Massive internal locking changes to mft record locking. Fixes lock
836 recursion and replaces the mrec_lock read/write semaphore with a
837 mutex. Also removes the now superfluous mft_count. This fixes several
838 race conditions and deadlocks, especially in the future write code.
839 - Fix ntfs over loopback for compressed files by adding an
840 optimization barrier. (gcc was screwing up otherwise ?)
841 - Miscellaneous cleanups all over the code and a fix or two in error
843 Thanks go to Christoph Hellwig for pointing out the following two:
844 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
845 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
847 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
849 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
850 at entry/exit respectively.
851 - Use C99 initializers for structures.
852 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
854 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
856 - Limit volume size at mount time to 2TiB on architectures where
857 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
858 This is the most we can do without overflowing the 32-bit limit of
859 the block device size imposed on us by sb_bread() and sb_getblk()
861 - Limit file/directory size at open() time to 16TiB on architectures
862 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
863 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
864 overflowing the page cache page index.
866 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
868 - Move the directory index bitmap to use an attribute inode instead of
869 having special fields for it inside the ntfs inode structure. This
870 means that the index bitmaps now use the page cache for i/o, too,
871 and also as a side effect we get support for non-resident index
873 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
874 fix a page leak that manifested itself in some cases.
875 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
876 index bitmap inode on the final iput().
878 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
880 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
881 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
882 - Drop the "file" from ntfs_file_read_compressed_block().
883 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
884 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
885 - Update ntfs_end_buffer_async_read() with the improved logic from
886 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
887 further logic improvements to better determine when we set PageError.
888 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
889 check for the buffers being uptodate first in line with the updated
890 fs/buffer.c::block_read_full_page(). This plugs a small race
893 2.0.18 - Fix race condition in reading of compressed files.
895 - There was a narrow window between checking a buffer head for being
896 uptodate and locking it in ntfs_file_read_compressed_block(). We now
897 lock the buffer and then check whether it is uptodate or not.
899 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
901 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
902 code and update callers, i.e. ntfs_iget(), to pass that error code
903 up instead of just using -EIO.
904 - Modifications to super.c to ensure that both mount and remount
905 cannot set any write related options when the driver is compiled
907 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
908 cache the current runlist element. This should improve performance
909 when reading very large and/or very fragmented data.
911 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
913 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
914 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
915 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
916 - Convert $MFT/$BITMAP access to attribute inode API and remove all
917 remnants of the ugly mftbmp address space and operations hack. This
918 means we finally have only one readpage function as well as only one
919 async io completion handler. Yey! The mft bitmap is now just an
920 attribute inode and is accessed from vol->mftbmp_ino just as if it
921 were a normal file. Fake inodes rule. (-:
923 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
925 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
926 remounts to fail when the partition had an entry in /etc/fstab and
927 the entry specified the nls= option.
928 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
929 expand all the helper functions NVolFoo(), NVolSetFoo(), and
931 - Move copyright statement from driver initialisation message to
932 module description (fs/super.c). This makes the initialisation
933 message fit on one line and fits in better with rest of kernel.
934 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
935 attribute inodes, and both for files and directories.
936 - Implement fake attribute inodes allowing all attribute i/o to go via
937 the page cache and to use all the normal vfs/mm functionality:
938 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
940 - Add needed cleanup code to ntfs_clear_big_inode().
941 - Merge address space operations for files and directories (aops.c),
942 now just have ntfs_aops:
944 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
945 ntfs_attr_read_block() -> ntfs_read_block(),
946 ntfs_file_read_page() -> ntfs_readpage().
947 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
948 attribute inodes, and both for files and directories.
949 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
951 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
953 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
954 the locking out of super.c::get_nr_free_mft_records() and taking and
955 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
956 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
957 current userspace ntfs library code. This means that if a merge
958 fails the original runlists are always left unmodified instead of
959 being silently corrupted.
962 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
964 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
965 from ntfs_volume structure.
966 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
967 from ntfs_volume structure.
968 - Use iget5_locked() and friends instead of conventional iget(). Wrap
969 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
970 to use ntfs_iget(). Leave only one iget() call at mount time so we
971 don't need an ntfs_iget_mount().
972 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
975 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
977 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
978 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
979 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
980 to determine whether to apply mst fixups or not.
981 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
982 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
983 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
984 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
985 the VFS readpage function prototype to the ntfs_attr_read_block()
988 2.0.11 - Initial preparations for fake inode based attribute i/o.
990 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
991 do some macro magic (adapted from include/linux/buffer_head.h) to
992 expand all the helper functions NInoFoo(), NInoSetFoo(), and
994 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
995 - Add new fields to ntfs_inode structure to allow use of fake inodes
996 for attribute i/o: type, name, name_len. Also add new state bits:
997 NI_Attr, which, if set, indicates the inode is a fake inode, and
998 NI_MstProtected, which, if set, indicates the attribute uses multi
999 sector transfer protection, i.e. fixups need to be applied after
1000 reads and before/after writes.
1001 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
1002 ntfs_{new,clear,destroy}_extent_inode() and update callers.
1003 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
1004 instead of ntfs_destroy_extent_inode().
1005 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
1006 - Make all operations on ntfs inode state bits use the NIno* functions.
1007 - Set up the new ntfs inode fields and state bits in
1008 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
1009 allocated memory to __ntfs_clear_inode().
1010 - Cleanup ntfs_inode structure a bit for better ordering of elements
1011 w.r.t. their size to allow better packing of the structure in memory.
1013 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
1015 - Add check at mount time to verify that the number of inodes on the
1016 volume does not exceed 2^32 - 1, which is the maximum allowed for
1017 NTFS according to Microsoft.
1018 - Change mft_no member of ntfs_inode structure to be unsigned long.
1019 Update all users. This makes ntfs_inode->mft_no just a copy of struct
1020 inode->i_ino. But we can't just always use struct inode->i_ino and
1021 remove mft_no because extent inodes do not have an attached struct
1024 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
1026 - Change decompression engine to use a single buffer protected by a
1027 spin lock instead of per-CPU buffers. (Rusty Russell)
1028 - Do not update cb_pos when handling a partial final page during
1029 decompression of a sparse compression block, as the value is later
1030 reset without being read/used. (Rusty Russell)
1031 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
1033 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
1034 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
1035 it also makes everything safer so it is a good thing.
1036 - Miscellaneous minor cleanups to comments.
1038 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
1040 Big thanks go to Al Viro and other inhabitants of #kernel for investing
1041 their time to discuss the case sensitivity and dcache aliasing issues.
1043 - Remove unused source file fs/ntfs/attraops.c.
1044 - Remove show_inodes mount option(s), thus dropping support for
1045 displaying of short file names.
1046 - Remove deprecated mount option posix.
1047 - Restore show_sys_files mount option.
1048 - Add new mount option case_sensitive, to determine if the driver
1049 treats file names as case sensitive or not. If case sensitive, create
1050 file names in the POSIX namespace. Otherwise create file names in the
1051 LONG/WIN32 namespace. Note, files remain accessible via their short
1052 file name, if it exists.
1053 - Remove really dumb logic bug in boot sector recovery code.
1054 - Fix dcache aliasing issues wrt short/long file names via changes
1055 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
1056 fs/ntfs/namei.c::ntfs_lookup():
1057 - Add additional argument to ntfs_lookup_inode_by_name() in which we
1058 return information about the matching file name if the case is not
1059 matching or the match is a short file name. See comments above the
1060 function definition for details.
1061 - Change ntfs_lookup() to only create dcache entries for the correctly
1062 cased file name and only for the WIN32 namespace counterpart of DOS
1063 namespace file names. This ensures we have only one dentry per
1064 directory and also removes all dcache aliasing issues between short
1065 and long file names once we add write support. See comments above
1066 function for details.
1067 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
1069 2.0.7 - Minor cleanups and updates for changes in core kernel code.
1071 - Remove much of the NULL struct element initializers.
1072 - Various updates to make compatible with recent kernels.
1073 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
1074 in fs/ntfs/ntfs.h instead.
1075 - Remove no longer needed KERNEL_VERSION checks. We are now in the
1076 kernel proper so they are no longer needed.
1078 2.0.6 - Major bugfix to make compatible with other kernel changes.
1080 - Initialize the mftbmp address space properly now that there are more
1081 fields in the struct address_space. This was leading to hangs and
1082 oopses on umount since 2.5.12 because of changes to other parts of
1083 the kernel. We probably want a kernel generic init_address_space()
1085 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
1086 only caller of ->readdir() is vfs_readdir() which holds i_sem during
1087 the call, and i_sem is sufficient protection against changes in the
1088 directory inode (including ->i_size).
1089 - Use generic_file_llseek() for directories (as opposed to
1090 default_llseek()) as this downs i_sem instead of the BKL which is
1091 what we now need for exclusion against ->f_pos changes considering we
1092 no longer take the BKL in ntfs_readdir().
1094 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
1096 - No need to set old blocksize in super.c::ntfs_fill_super() as the
1097 VFS does so via invocation of deactivate_super() calling
1098 fs->fill_super() calling block_kill_super() which does it.
1099 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
1100 -> Do we really need it? I don't think so as we have exclusion on
1101 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
1102 move the ->f_pos accesses under the mrec_lock though. Check this...
1103 - Fix really, really, really stupid buffer overflow in extent inode
1104 handling in mft.c::map_extent_mft_record().
1106 2.0.4 - Cleanups and updates for kernel 2.5.11.
1108 - Add documentation on how to use the MD driver to be able to use NTFS
1109 stripe and volume sets in Linux and generally cleanup documentation
1111 Remove all uses of kdev_t in favour of struct block_device *:
1112 - Change compress.c::ntfs_file_read_compressed_block() to use
1113 sb_getblk() instead of getblk().
1114 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
1115 of get_hardsect_size().
1116 - No need to get old blocksize in super.c::ntfs_fill_super() as
1117 fs/super.c::get_sb_bdev() already does this.
1118 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
1120 2.0.3 - Small bug fixes, cleanups, and performance improvements.
1122 - Remove some dead code from mft.c.
1123 - Optimize readpage and read_block functions throughout aops.c so that
1124 only initialized blocks are read. Non-initialized ones have their
1125 buffer head mapped, zeroed, and set up to date, without scheduling
1126 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
1127 Thanks go to Andrew Morton for spotting the below:
1128 - Fix buglet in allocate_compression_buffers() error code path.
1129 - Call flush_dcache_page() after modifying page cache page contents in
1130 ntfs_file_readpage().
1131 - Check for existence of page buffers throughout aops.c before calling
1132 create_empty_buffers(). This happens when an I/O error occurs and the
1133 read is retried. (It also happens once writing is implemented so that
1134 needed doing anyway but I had left it for later...)
1135 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
1136 readpage and read_block functions. Reasoning same as above (i.e. I/O
1137 error retries and future write code paths.)
1139 2.0.2 - Minor updates and cleanups.
1141 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
1142 and cleanup the code a bit, removing the unused size parameter.
1143 - Change default fmask to 0177 and update documentation.
1144 - Change attrib.c::get_attr_search_ctx() to return the search context
1145 directly instead of taking the address of a pointer. A return value
1146 of NULL means the allocation failed. Updated all callers
1148 - Update to 2.5.9 kernel (preserving backwards compatibility) by
1149 replacing all occurences of page->buffers with page_buffers(page).
1150 - Fix minor bugs in runlist merging, also minor cleanup.
1151 - Updates to bootsector layout and mft mirror contents descriptions.
1152 - Small bug fix in error detection in unistr.c and some cleanups.
1153 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
1156 2.0.1 - Minor updates.
1158 - Make default umask correspond to documentation.
1159 - Improve documentation.
1160 - Set default mode to include execute bit. The {u,f,d}mask can be used
1161 to take it away if desired. This allows binaries to be executed from
1162 a mounted ntfs partition.
1164 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
1166 - Add kill_super, just keeping up with the vfs changes in the kernel.
1167 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
1168 from the CVS import into BitKeeper.
1169 - Begin to implement proper handling of allocated_size vs
1170 initialized_size vs data_size (i.e. i_size). Done are
1171 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
1172 and attrib.c::load_attribute_list().
1173 - Lock the runlist in attrib.c::load_attribute_list() while using it.
1174 - Fix memory leak in ntfs_file_read_compressed_block() and generally
1175 clean up compress.c a little, removing some uncommented/unused debug
1177 - Tidy up dir.c a little bit.
1178 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
1179 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
1180 creating aops.c::ntfs_mst_readpage(), improving the handling of
1181 holes and overflow in the process and implementing the correct
1182 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
1183 I am aiming for correctness at the moment. Modularisation can come
1185 - Rename aops.c::end_buffer_read_index_async() to
1186 end_buffer_read_mst_async() and optimize the overflow checking and
1188 - Use the host of the mftbmp address space mapping to hold the ntfs
1189 volume. This is needed so the async i/o completion handler can
1190 retrieve a pointer to the volume. Hopefully this will not cause
1191 problems elsewhere in the kernel... Otherwise will need to use a
1193 - Complete implementation of proper handling of allocated_size vs
1194 initialized_size vs data_size (i.e. i_size) in whole driver.
1195 Basically aops.c is now completely rewritten.
1196 - Change NTFS driver name to just NTFS and set version number to 2.0.0
1197 to make a clear distinction from the old driver which is still on
1200 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
1202 - Replace bdevname(sb->s_dev) with sb->s_id.
1203 - Remove now superfluous new-line characters in all callers of
1205 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
1206 directories. Without this the "find" utility gets very upset which is
1207 fair enough as Linux/Unix do not support directory hard links.
1208 - Further runlist merging work. (Richard Russon)
1209 - Backwards compatibility for gcc-2.95. (Richard Russon)
1210 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
1211 - Convert to new filesystem declaration using ->ntfs_get_sb() and
1212 replacing ntfs_read_super() with ntfs_fill_super().
1213 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
1214 overflow on 32-bit architectures.
1215 - Cleanup upcase loading code to use ntfs_(un)map_page().
1216 - Disable/reenable preemtion in critical sections of compession engine.
1217 - Replace device size determination in ntfs_fill_super() with
1218 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
1219 function super.c::get_nr_blocks().
1220 - Implement a mount time option (show_inodes) allowing choice of which
1221 types of inode names readdir() returns and modify ntfs_filldir()
1222 accordingly. There are several parameters to show_inodes:
1223 system: system files
1224 win32: long file names (including POSIX file names) [DEFAULT]
1226 dos: short file names only (excluding POSIX file names)
1228 posix: same as both win32 and dos
1230 Note that the options are additive, i.e. specifying:
1231 -o show_inodes=system,show_inodes=win32,show_inodes=dos
1232 is the same as specifying:
1234 Note that the "posix" and "all" options will show all directory
1235 names, BUT the link count on each directory inode entry is set to 1,
1236 due to Linux not supporting directory hard links. This may well
1237 confuse some userspace applications, since the directory names will
1238 have the same inode numbers. Thus it is NOT advisable to use the
1239 "posix" or "all" options. We provide them only for completeness sake.
1240 - Add copies of allocated_size, initialized_size, and compressed_size to
1241 the ntfs inode structure and set them up in
1242 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
1243 for files and the index allocation attribute for directories.
1244 - Add copies of allocated_size and initialized_size to ntfs inode for
1245 $BITMAP attribute of large directories and set them up in
1246 inode.c::ntfs_read_inode().
1247 - Add copies of allocated_size and initialized_size to ntfs volume for
1248 $BITMAP attribute of $MFT and set them up in
1249 super.c::load_system_files().
1250 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
1251 posix, and utf8) and tell user what the new options to use are. Note
1252 we still do support them but they will be removed with kernel 2.7.x.
1253 - Change all occurences of integer long long printf formatting to hex
1254 as printk() will not support long long integer format if/when the
1255 div64 patch goes into the kernel.
1256 - Make slab caches have stable names and change the names to what they
1257 were intended to be. These changes are required/made possible by the
1258 new slab cache name handling which removes the length limitation by
1259 requiring the caller of kmem_cache_create() to supply a stable name
1260 which is then referenced but not copied.
1261 - Rename run_list structure to run_list_element and create a new
1262 run_list structure containing a pointer to a run_list_element
1263 structure and a read/write semaphore. Adapt all users of runlists
1264 to new scheme and take and release the lock as needed. This fixes a
1265 nasty race as the run_list changes even when inodes are locked for
1266 reading and even when the inode isn't locked at all, so we really
1267 needed the serialization. We use a semaphore rather than a spinlock
1268 as memory allocations can sleep and doing everything GFP_ATOMIC
1270 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
1271 This can never happen due to the nature of lookup_attr() and how we
1272 support attribute lists. If it did happen it would imply the inode
1274 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
1276 - Update to 2.5.6-pre2 changes in struct address_space.
1277 - Use parent_ino() when accessing d_parent inode number in dir.c.
1278 - Import Sourceforge CVS repository into BitKeeper repository:
1279 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
1280 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
1281 Documentation/filesystems/ntfs.txt for NTFS TNG.
1282 - Create kernel configuration option controlling whether debugging
1284 - Add the required export of end_buffer_io_sync() from the patches
1285 directory to the kernel code.
1286 - Update inode.c::ntfs_show_options() with show_inodes mount option.
1287 - Update errors mount option.
1289 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
1291 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
1292 bug in mapping of extent inodes. Update all the comments to fit all
1293 the recent code changes.
1294 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
1295 - Cleanups in compress.c, mostly comments and folding help.
1296 - Implement attrib.c::map_run_list() as a generic helper.
1297 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
1298 thus making code shorter and enabling attribute list support.
1299 - Cleanup incorrect use of [su]64 with %L printf format specifier in
1300 all source files. Type casts to [unsigned] long long added to correct
1301 the mismatches (important for architectures which have long long not
1303 - Merge async io completion handlers for directory indexes and $MFT
1304 data into one by setting the index_block_size{_bits} of the ntfs
1305 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
1306 - Cleanup aops.c, update comments.
1307 - Make ntfs_file_get_block() use map_run_list() so all files now
1308 support attribute lists.
1309 - Make ntfs_dir_readpage() almost verbatim copy of
1310 block_read_full_page() by using ntfs_file_get_block() with only real
1311 difference being the use of our own async io completion handler
1312 rather than the default one, thus reducing the amount of code and
1313 automatically enabling attribute list support for directory indices.
1314 - Fix bug in load_attribute_list() - forgot to call brelse in error
1316 - Change parameters to find_attr() and lookup_attr(). We no longer
1317 pass in the upcase table and its length. These can be gotten from
1318 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
1319 - Cleanups in attrib.c.
1320 - Implement merging of runlists, attrib.c::merge_run_lists() and its
1321 helpers. (Richard Russon)
1322 - Attribute lists part 2, attribute extents and multi part runlists:
1323 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
1324 further runlist parts via attrib.c::map_run_list().
1325 - Tiny endianness bug fix in decompress_mapping_pairs().
1327 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
1329 - Enable encrypted directories. (Their index root is marked encrypted
1330 to indicate that new files in that directory should be created
1332 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
1333 - Enable $Extend system directory. Most (if not all) extended system
1334 files do not have unnamed data attributes so ntfs_read_inode() had to
1335 special case them but that is ok, as the special casing recovery
1336 happens inside an error code path so there is zero slow down in the
1337 normal fast path. The special casing is done by introducing a new
1338 function inode.c::ntfs_is_extended_system_file() which checks if any
1339 of the hard links in the inode point to $Extend as being their parent
1340 directory and if they do we assume this is an extended system file.
1341 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
1342 when compiled with -DDEBUG. Default is debug messages to be disabled.
1343 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
1344 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
1345 interface is enabled). Inspired by old ntfs driver.
1346 - Add debug_msgs insmod/kernel boot parameter to set whether debug
1347 messages are {dis,en}abled. This is useful to enable debug messages
1348 during ntfs initialization and is the only way to activate debugging
1349 when the sysctl interface is not enabled.
1350 - Cleanup debug output in various places.
1351 - Remove all dollar signs ($) from the source (except comments) to
1352 enable compilation on architectures whose gcc compiler does not
1353 support dollar signs in the names of variables/constants. Attribute
1354 types now start with AT_ instead of $ and $I30 is now just I30.
1355 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
1356 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
1357 access functions. This means we now cope with $MFT/$BITMAP being
1358 spread accross several mft records.
1359 - Disable modification of mft_zone_multiplier on remount. We can always
1360 reenable this later on if we really want to, but we will need to make
1361 sure we readjust the mft_zone size / layout accordingly.
1363 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
1365 - Use sb_set_blocksize() instead of set_blocksize() and verify the
1367 - Use sb_bread() instead of bread() throughout.
1368 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1369 of a directory index block vcn. Apply resulting simplifications in
1371 - Fix a small bug somewhere (but forgot what it was).
1372 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1373 on the printf-format parameter list and fix bugs reported by gcc
1374 as a result. (Richard Russon)
1375 - Move inode allocation strategy to Al's new stuff but maintain the
1376 divorce of ntfs_inode from struct inode. To achieve this we have two
1377 separate slab caches, one for big ntfs inodes containing a struct
1378 inode and pure ntfs inodes and at the same time fix some faulty
1379 error code paths in ntfs_read_inode().
1380 - Show mount options in proc (inode.c::ntfs_show_options()).
1382 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1384 - Modified (un)map_mft_record functions to be common for read and write
1385 case. To specify which is which, added extra parameter at front of
1386 parameter list. Pass either READ or WRITE to this, each has the
1388 - General cleanups to allow for easier folding in vi.
1389 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1390 argument, and invokes attrib.c::merge_run_lists() to merge the old
1391 and the new runlists.
1392 - Removed attrib.c::find_first_attr().
1393 - Implemented loading of attribute list and complete runlist for $MFT.
1394 This means we now cope with $MFT being spread across several mft
1396 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1397 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1398 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1399 cache. There are too little ntfs_volume structures at any one time
1400 to justify a private slab cache.
1401 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1402 Use KM_BIO_IRQ on advice from IRC/kernel...
1403 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1404 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1405 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1406 removed the now superfluous exports from the kernel core patch.
1407 - Fix a bug where kfree() was used insted of ntfs_free().
1408 - Change map_mft_record() to take ntfs_inode as argument instead of
1409 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1410 - Add pointer to ntfs_volume to ntfs_inode.
1411 - Add mft record number and sequence number to ntfs_inode. Stop using
1412 i_ino and i_generation for in-driver purposes.
1413 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1414 - Remove use of proper inodes by extent inodes. Move i_ino and
1415 i_generation to ntfs_inode to do this. Apply simplifications that
1416 result and remove iget_no_wait(), etc.
1417 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1418 - Add reference counting in ntfs_inode for the ntfs inode itself and
1419 for the mapped mft record.
1420 - Extend mft record mapping so we can (un)map extent mft records (new
1421 functions (un)map_extent_mft_record), and so mappings are reference
1422 counted and don't have to happen twice if already mapped - just ref
1424 - Add -o iocharset as alias to -o nls for backwards compatibility.
1425 - The latest core patch is now tiny. In fact just a single additional
1426 export is necessary over the base kernel.
1428 tng-0.0.3 - Cleanups, enhancements, bug fixes.
1430 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1431 and setup the runlist appropriately using knowledge provided by the
1432 sizes in the base attribute record.
1433 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1435 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1436 page or use vmalloc depending on the amount of memory requested.
1437 - Cleanup error output. The __FUNCTION__ "(): " is now added
1438 automatically. Introduced a new header file debug.h to support this
1439 and also moved ntfs_debug() function into it.
1440 - Make reading of compressed files more intelligent and especially get
1441 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1442 buffers (allocated at first mount with cluster size <= 4kiB and
1443 deallocated on last umount with cluster size <= 4kiB), and
1444 asynchronous io for the compressed data using a list of buffer heads.
1445 Er, we use synchronous io as async io only works on whole pages
1446 covered by buffers and not on individual buffer heads...
1447 - Bug fix for reading compressed files with sparse compression blocks.
1449 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1451 - Fixed handling of directories when cluster size exceeds index block
1453 - Hide DOS only name space directory entries from readdir() but allow
1454 them in lookup(). This should fix the problem that Linux doesn't
1455 support directory hard links, while still allowing access to entries
1456 via their short file name. This also has the benefit of mimicking
1457 what Windows users are used to, so it is the ideal solution.
1458 - Implemented sync_page everywhere so no more hangs in D state when
1460 - Stop using bforget() in favour of brelse().
1461 - Stop locking buffers unnecessarily.
1462 - Implemented compressed files (inode->mapping contains uncompressed
1463 data, raw compressed data is currently bread() into a vmalloc()ed
1465 - Enable compressed directories. (Their index root is marked compressed
1466 to indicate that new files in that directory should be created
1468 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1469 functions. (Thanks to Will Dyson for pointing this out.)
1470 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1471 ntfs_sb_info) out of the common inode and super_block structures and
1472 started using the generic_ip and generic_sbp pointers instead. This
1473 makes ntfs entirely private with respect to the kernel tree.
1474 - Detect compiler version and abort with error message if gcc less than
1476 - Fix bug in name comparison function in unistr.c.
1477 - Implement attribute lists part 1, the infrastructure: search contexts
1478 and operations, find_external_attr(), lookup_attr()) and make the
1479 code use the infrastructure.
1480 - Fix stupid buffer overflow bug that became apparent on larger run
1481 list containing attributes.
1482 - Fix bugs in readdir() that became apparent on larger directories.
1484 The driver is now really useful and survives the test
1485 find . -type f -exec md5sum "{}" \;
1486 without any error messages on a over 1GiB sized partition with >16k
1487 files on it, including compressed files and directories and many files
1488 and directories with attribute lists.
1490 tng-0.0.1 - The first useful version.
1492 - Added ntfs_lookup().
1493 - Added default upcase generation and handling.
1494 - Added compile options to be shown on module init.
1495 - Many bug fixes that were "hidden" before.
1496 - Update to latest kernel.
1497 - Added ntfs_readdir().
1498 - Added file operations for mmap(), read(), open() and llseek(). We just
1499 use the generic ones. The whole point of going through implementing
1500 readpage() methods and where possible get_block() call backs is that
1501 this allows us to make use of the generic high level methods provided
1504 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1505 though and it doesn't implement accesssing compressed files yet. Also,
1506 accessing files with attribute list attributes is not implemented yet
1507 either. But for small or simple filesystems it should work and allow
1508 you to list directories, use stat on directory entries and the file
1509 system, open, read, mmap and llseek around in files. A big mile stone
1512 tng-0.0.0 - Initial version tag.
1514 Initial driver implementation. The driver can mount and umount simple
1515 NTFS filesystems (i.e. ones without attribute lists in the system
1516 files). If the mount fails there might be problems in the error handling
1517 code paths, so be warned. Otherwise it seems to be loading the system
1518 files nicely and the mft record read mapping/unmapping seems to be
1519 working nicely, too. Proof of inode metadata in the page cache and non-
1520 resident file unnamed stream data in the page cache concepts is thus