3 - Checkpoint or disable the user space journal ($UsnJrnl).
4 - In between ntfs_prepare/commit_write, need exclusion between
5 simultaneous file extensions. Need perhaps an NInoResizeUnderway()
6 flag which we can set in ntfs_prepare_write() and clear again in
7 ntfs_commit_write(). Just have to be careful in readpage/writepage,
8 as well as in truncate, that we play nice... We might need to have
9 a data_size field in the ntfs_inode to store the real attribute
10 length. Also need to be careful with initialized_size extention in
11 ntfs_prepare_write. Basically, just be _very_ careful in this code...
12 OTOH, perhaps i_sem, which is held accross generic_file_write is
13 sufficient for synchronisation here. We then just need to make sure
14 ntfs_readpage/writepage/truncate interoperate properly with us.
15 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
16 leave the volume dirty on umount if the final iput(vol->mft_ino)
17 causes a write of any mirrored mft records due to the mft mirror
18 inode having been discarded already. Whether this can actually ever
19 happen is unclear however so it is worth waiting until someone hits
21 - Enable the code for setting the NT4 compatibility flag when we start
22 making NTFS 1.2 specific modifications.
24 2.1.20 - Fix a stupid bug in ntfs_attr_reinit_search_ctx().
26 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
27 where we did not clear ctx->al_entry but it was still set due to
28 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
30 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
31 where we forgot to unmap the extent mft record when we had finished
32 enumerating an attribute which caused a bug check to trigger when the
33 VFS calls ->clear_inode.
35 2.1.19 - Many cleanups, improvements, and a minor bug fix.
37 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
38 change the uid, gid, and mode of an inode as we do not support NTFS
40 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
42 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
43 and ntfs_filldir() as per suggestion from Al Viro.
44 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
45 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
46 inode size has changed and to only output an error if so.
47 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
48 - Add le{16,32,64} as well as sle{16,32,64} data types to
50 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
51 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
52 respectively, to fs/ntfs/types.h.
53 - Update endianness conversion macros in fs/ntfs/endian.h to use the
54 new types as appropriate.
55 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
57 - Add leMFT_REF data type to fs/ntfs/layout.h.
58 - Update all NTFS header files with the new little endian data types.
59 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
60 - Do proper type casting when using ntfs_is_*_recordp() in
61 fs/ntfs/logfile.c, mft.c, and super.c.
62 - Fix all the sparse bitwise warnings. Had to change all the typedef
63 enums storing little endian values to simple enums plus a typedef for
64 the datatype to make sparse happy.
65 - Fix a bug found by the new sparse bitwise warnings where the default
66 upcase table was defined as a pointer to wchar_t rather than ntfschar
67 in fs/ntfs/ntfs.h and super.c.
68 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
70 2.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
72 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
73 the calculation of total/free inodes as used by statfs().
74 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
75 because the code itself is using the ntfs_lock semaphore which
76 provides safe locking. (Ingo Molnar)
77 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
78 could occur in the future for when we start closing/freeing extent
79 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
81 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
82 find_external_attr() to ntfs_external_attr_find() to cleanup the
83 namespace a bit and to be more consistent with libntfs.
84 - Rename {{re,}init,get,put}_attr_search_ctx() to
85 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
86 attr_search_context to ntfs_attr_search_ctx.
87 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
88 for the attribute list attribute itself.
89 - Fix endianness bug in ntfs_external_attr_find().
90 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
91 if the attribute is not found, and -EIO on real error. In the case
92 of -ENOENT, the search context is updated to describe the attribute
93 before which the attribute being searched for would need to be
94 inserted if such an action were to be desired and in the case of
95 ntfs_external_attr_find() the search context is also updated to
96 indicate the attribute list entry before which the attribute list
97 entry of the attribute being searched for would need to be inserted
98 if such an action were to be desired. Also make ntfs_find_attr()
99 static and remove its prototype from attrib.h as it is not used
100 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
101 callers of ntfs_{external,}attr_{find,lookup}() for the new return
103 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
105 2.1.17 - Fix bugs in mount time error code paths and other updates.
107 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
108 includes functions to set/clear a single bit or a run of bits.
109 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
110 runlist element containing a particular vcn. It also takes care of
111 mapping any needed runlist fragments.
112 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
113 - Load attribute definition table from $AttrDef at mount time.
114 - Fix bugs in mount time error code paths involving (de)allocation of
115 the default and volume upcase tables.
116 - Remove ntfs_nr_mounts as it is no longer used.
118 2.1.16 - Implement access time updates, file sync, async io, and read/writev.
120 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
121 This is done by setting the appropriate file operations pointers to
122 the generic helper functions provided by mm/filemap.c.
123 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
124 and directories (fs/ntfs/dir.c).
125 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
126 Note, except for the root directory and any other system files opened
127 by the user, the system files will not have their access times
128 updated as they are only accessed at the inode level an hence the
129 file level functions which cause the times to be updated are never
132 2.1.15 - Invalidate quotas when (re)mounting read-write.
134 - Add new element itype.index.collation_rule to the ntfs inode
135 structure and set it appropriately in ntfs_read_locked_inode().
136 - Implement a new inode type "index" to allow efficient access to the
137 indices found in various system files and adapt inode handling
138 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
139 attribute inode (NInoAttr() is true) with an attribute type of
140 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
141 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
142 there would be no way to distinguish between normal attribute inodes
143 and index inodes. The function to obtain an index inode is
144 ntfs_index_iget() and it uses the helper function
145 ntfs_read_locked_index_inode(). Note, we do not overload
146 ntfs_attr_iget() as indices consist of multiple attributes so using
147 ntfs_attr_iget() to obtain an index inode would be confusing.
148 - Ensure that there is no overflow when doing page->index <<
149 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
150 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
151 and ntfs_read_block().
152 - Use case sensitive attribute lookups instead of case insensitive ones.
153 - Lock all page cache pages belonging to mst protected attributes while
154 accessing them to ensure we never see corrupt data while the page is
156 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
157 We have ntfs_is_collation_rule_supported() to check if the collation
158 rule you want to use is supported and ntfs_collation() which actually
159 collates two data items. We currently only support COLLATION_BINARY
160 and COLLATION_NTOFS_ULONG but support for other collation rules will
161 be added as the need arises.
162 - Add a new type, ntfs_index_context, to allow retrieval of an index
163 entry using the corresponding index key. To get an index context,
164 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
165 This also adds a new slab cache for the index contexts. To lookup a
166 key in an index inode, use ntfs_index_lookup(). After modifying an
167 index entry, call ntfs_index_entry_flush_dcache_page() followed by
168 ntfs_index_entry_mark_dirty() to ensure the changes are written out
169 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
170 an index entry is in the index allocation attribute rather than the
171 index root attribute it will not be written out (you will get a
172 warning message about discarded changes instead).
173 - Load the quota file ($Quota) and check if quota tracking is enabled
174 and if so, mark the quotas out of date. This causes windows to
175 rescan the volume on boot and update all quota entries.
176 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
177 It is simply set to __set_page_dirty_nobuffers() to make sure that
178 running set_page_dirty() on a page containing mft/ntfs records will
179 not affect the dirty state of the page buffers.
180 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
181 buffers that are inside the ntfs record in the page dirty after which
182 it sets the page dirty. This allows ->writepage to only write the
183 dirty index records rather than having to write all the records in
184 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
185 use this rather than __set_page_dirty_nobuffers().
186 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
187 writing of page cache pages belonging to mst protected attributes
188 like the index allocation attribute in directory indices and other
189 indices like $Quota/$Q, etc. This means that the quota is now marked
190 out of date on all volumes rather than only on ones where the quota
191 defaults entry is in the index root attribute of the $Quota/$Q index.
193 2.1.14 - Fix an NFSd caused deadlock reported by several users.
195 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
196 to a buffer so that we can put the search context and unmap the mft
197 record before calling the filldir() callback. We need to do this
198 because of NFSd which calls ->lookup() from its filldir callback()
199 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
200 of the directory and since ntfs_readdir() has got it mapped already
201 ntfs_lookup() deadlocks.
203 2.1.13 - Enable overwriting of resident files and housekeeping of system files.
205 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
206 keeping the mft mirror in sync with the mft when mirrored mft records
207 are written. The functions are write_mft_record{,_nolock}(). The
208 implementation is quite rudimentary for now with lots of things not
209 implemented yet but I am not sure any of them can actually occur so
210 I will wait for people to hit each one and only then implement it.
211 - Commit open system inodes at umount time. This should make it
212 virtually impossible for sync_mft_mirror_umount() to ever be needed.
213 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
214 ntfs super operations. This gives us inode writing via the VFS inode
215 dirty code paths. Note: Access time updates are not implemented yet.
216 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
217 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
218 finally enabling resident file overwrite! (-8 This also includes a
219 placeholder for ->writepage (ntfs_mft_writepage()), which for now
220 just redirties the page and returns. Also, at umount time, we for
221 now throw away all mft data page cache pages after the last call to
222 ntfs_commit_inode() in the hope that all inodes will have been
223 written out by then and hence no dirty (meta)data will be lost. We
224 also check for this case and emit an error message telling the user
226 - Use set_page_writeback() and end_page_writeback() in the resident
227 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
228 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
230 - Implement ntfs_mft_writepage() so it now checks if any of the mft
231 records in the page are dirty and if so redirties the page and
232 returns. Otherwise it just returns (after doing set_page_writeback(),
233 unlock_page(), end_page_writeback() or the radix-tree tag
234 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
235 alowing the VM to do with the page as it pleases. Also, at umount
236 time, now only throw away dirty mft (meta)data pages if dirty inodes
237 are present and ask the user to email us if they see this happening.
238 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
239 information flags (fs/ntfs/super.c).
240 - Mark the volume dirty when (re)mounting read-write and mark it clean
241 when unmounting or remounting read-only. If any volume errors are
242 found, the volume is left marked dirty to force chkdsk to run.
243 - Add code to set the NT4 compatibility flag when (re)mounting
244 read-write for newer NTFS versions but leave it commented out for now
245 since we do not make any modifications that are NTFS 1.2 specific yet
246 and since setting this flag breaks Captive-NTFS which is not nice.
247 This code must be enabled once we start writing NTFS 1.2 specific
248 changes otherwise Windows NTFS driver might crash / cause corruption.
250 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
252 - Add a new address space operations struct, ntfs_mst_aops, for mst
253 protected attributes. This is because the default ntfs_aops do not
254 make sense with mst protected data and were they to write anything to
255 such an attribute they would cause data corruption so we provide
256 ntfs_mst_aops which does not have any write related operations set.
257 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
258 includes an adapted ntfs_commit_inode() and an implementation of
259 ntfs_write_inode() which for now just cleans dirty inodes without
260 writing them (it does emit a warning that this is happening).
261 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
262 entry) as it was only fixing a theoretical bug but at the same time
263 it badly broke the handling of sparse and uncompressed compression
266 2.1.11 - Driver internal cleanups.
268 - Only build logfile.o if building the driver with read-write support.
269 - Really final white space cleanups.
270 - Use generic_ffs() instead of ffs() in logfile.c which allows the
271 log_page_size variable to be optimized by gcc into a constant.
272 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
273 char as defined by POSIX and as found on some systems.
275 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
277 - Finish off the white space cleanups (remove trailing spaces, etc).
278 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
279 the kludges around the first iget(). Instead of (re)setting ->s_op
280 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
281 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
282 kills the need for second super_operations and allows to return error
283 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
285 - Force read-only (re)mounting if any of the following bits are set in
286 the volume information flags:
287 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
288 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
289 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
290 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
291 above bits set so the test is made easy.
293 2.1.9 - Fix two bugs in decompression engine.
295 - Fix a bug where we would not always detect that we have reached the
296 end of a compression block because we were ending at minus one byte
297 which is effectively the same as being at the end. The fix is to
298 check whether the uncompressed buffer has been fully filled and if so
299 we assume we have reached the end of the compression block. A big
300 thank you to Marcin Gibuła for the bug report, the assistance in
301 tracking down the bug and testing the fix.
302 - Fix a possible bug where when a compressed read is truncated to the
303 end of the file, the offset inside the last page was not truncated.
305 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
307 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
308 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
309 utc2ntfs() to work with struct timespec instead of time_t on the
310 Linux UTC time side thus preserving the full precision of the NTFS
311 time and only loosing up to 99 nano-seconds in the Linux UTC time.
312 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
314 - Remove unused ntfs_dirty_inode().
315 - Cleanup super operations declaration in fs/ntfs/super.c.
316 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
317 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
318 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
319 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
320 fs/ntfs/inode.h so they can be used elsewhere.
321 - Determine the mft mirror size as the number of mirrored mft records
322 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
323 - Load the mft mirror at mount time and compare the mft records stored
324 in it to the ones in the mft. Force a read-only mount if the two do
325 not match (fs/ntfs/super.c).
326 - Fix type casting related warnings on 64-bit architectures. Thanks
327 to Meelis Roos for reporting them.
328 - Move %L to %ll as %L is floating point and %ll is integer which is
330 - Read the journal ($LogFile) and determine if the volume has been
331 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
332 and fs/ntfs/logfile.c). This is a little bit of a crude check in
333 that we only look at the restart areas and not at the actual log
334 records so that there will be a very small number of cases where we
335 think that a volume is dirty when in fact it is clean. This should
336 only affect volumes that have not been shutdown cleanly and did not
337 have any pending, non-check-pointed i/o.
338 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
339 is requested, empty $LogFile by overwriting it with 0xff bytes to
340 ensure that Windows cannot cause data corruption by replaying a stale
341 journal after Linux has written to the volume.
343 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
345 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
346 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
347 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
348 default doesn't allow inode number 0 which is a valid inode on NTFS
349 and even if it did allow that it uses iget() instead of ntfs_iget()
350 which makes it useless for us.
351 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
352 default just returns -EACCES which is not very useful.
353 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
354 and set them up in the super block at mount time (super.c) this
355 allows mounted NTFS volumes to be exported via NFS.
356 - Add missing return -EOPNOTSUPP; in
357 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
358 - Enforce no atime and no dir atime updates at mount/remount time as
359 they are not implemented yet anyway.
360 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
361 after a NULL check. Thanks to Dave Jones for pointing this out.
363 2.1.6 - Fix minor bug in handling of compressed directories.
365 - Fix bug in handling of compressed directories. A compressed
366 directory is not really compressed so when we set the ->i_blocks
367 field of a compressed directory inode we were setting it from the
368 non-existing field ni->itype.compressed.size which gave random
369 results... For directories we now always use ni->allocated_size.
371 2.1.5 - Fix minor bug in attribute list attribute handling.
373 - Fix bug in attribute list handling. Actually it is not as much a bug
374 as too much protection in that we were not allowing attribute lists
375 which waste space on disk while Windows XP clearly allows it and in
376 fact creates such attribute lists so our driver was failing.
377 - Update NTFS documentation ready for 2.6 kernel release.
379 2.1.4 - Reduce compiler requirements.
381 - Remove all uses of unnamed structs and unions in the driver to make
382 old and newer gcc versions happy. Makes it a bit uglier IMO but at
383 least people will stop hassling me about it.
385 2.1.3 - Important bug fixes in corner cases.
387 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
388 clusters. (Philipp Thomas)
389 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
390 multiple of the block_size but not the cluster size. (Szabolcs
391 Szakacsits <szaka@sienet.hu>)
393 2.1.2 - Important bug fixes aleviating the hangs in statfs.
395 - Fix buggy free cluster and free inode determination logic.
397 2.1.1 - Minor updates.
399 - Add handling for initialized_size != data_size in compressed files.
400 - Reduce function local stack usage from 0x3d4 bytes to just noise in
401 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
402 - Remove compiler warnings for newer gcc.
403 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
404 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
405 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
406 kmap_atomic(KM_USER0).
408 2.1.0 - First steps towards write support: implement file overwrite.
410 - Add configuration option for developmental write support with an
411 appropriately scary configuration help text.
412 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
413 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
414 overwriting of existing files on ntfs. Note: Resident files are
415 only written into memory, and not written out to disk at present, so
416 avoid writing to files smaller than about 1kiB.
417 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
418 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
419 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
420 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
421 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
422 This enables write(2) based overwriting of existing files on ntfs.
423 Note: As with mmap(2) based overwriting, resident files are only
424 written into memory, and not written out to disk at present, so avoid
425 writing to files smaller than about 1kiB.
426 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
427 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
428 files with the purpose of intercepting and aborting all i_size
429 changes which we do not support yet. ntfs_truncate() actually only
430 emits a warning message but AFAICS our interception of i_size changes
431 elsewhere means ntfs_truncate() never gets called for i_size changes.
432 It is only called from generic_file_write() when we fail in
433 ntfs_prepare_{,nonresident_}write() in order to discard any
434 instantiated buffers beyond i_size. Thus i_size is not actually
435 changed so our warning message is enough. Unfortunately it is not
436 possible to easily determine if i_size is being changed or not hence
437 we just emit an appropriately worded error message.
439 2.0.25 - Small bug fixes and cleanups.
441 - Unlock the page in an out of memory error code path in
442 fs/ntfs/aops.c::ntfs_read_block().
443 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
444 just unlock the page and return. (This can happen due to ->writepage
445 clearing PageUptodate() during write out of MstProtected()
447 - Remove leaked write code again.
451 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
452 inside BUG_ON(). (Adam J. Richter)
453 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
454 calls for improved debugging. (Adam J. Richter)
455 - Add errors flag to the ntfs volume state, accessed via
456 NVol{,Set,Clear}Errors(vol).
457 - Do not allow read-write remounts of read-only volumes with errors.
458 - Clarify comment for ntfs file operation sendfile which was added by
459 Christoph Hellwig a while ago (just using generic_file_sendfile())
460 to say that ntfs ->sendfile is only used for the case where the
461 source data is on the ntfs partition and the destination is
462 somewhere else, i.e. nothing we need to concern ourselves with.
463 - Add generic_file_write() as our ntfs file write operation.
465 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
467 - Massive internal locking changes to mft record locking. Fixes lock
468 recursion and replaces the mrec_lock read/write semaphore with a
469 mutex. Also removes the now superfluous mft_count. This fixes several
470 race conditions and deadlocks, especially in the future write code.
471 - Fix ntfs over loopback for compressed files by adding an
472 optimization barrier. (gcc was screwing up otherwise ?)
473 - Miscellaneous cleanups all over the code and a fix or two in error
475 Thanks go to Christoph Hellwig for pointing out the following two:
476 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
477 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
479 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
481 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
482 at entry/exit respectively.
483 - Use C99 initializers for structures.
484 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
486 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
488 - Limit volume size at mount time to 2TiB on architectures where
489 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
490 This is the most we can do without overflowing the 32-bit limit of
491 the block device size imposed on us by sb_bread() and sb_getblk()
493 - Limit file/directory size at open() time to 16TiB on architectures
494 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
495 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
496 overflowing the page cache page index.
498 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
500 - Move the directory index bitmap to use an attribute inode instead of
501 having special fields for it inside the ntfs inode structure. This
502 means that the index bitmaps now use the page cache for i/o, too,
503 and also as a side effect we get support for non-resident index
505 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
506 fix a page leak that manifested itself in some cases.
507 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
508 index bitmap inode on the final iput().
510 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
512 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
513 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
514 - Drop the "file" from ntfs_file_read_compressed_block().
515 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
516 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
517 - Update ntfs_end_buffer_async_read() with the improved logic from
518 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
519 further logic improvements to better determine when we set PageError.
520 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
521 check for the buffers being uptodate first in line with the updated
522 fs/buffer.c::block_read_full_page(). This plugs a small race
525 2.0.18 - Fix race condition in reading of compressed files.
527 - There was a narrow window between checking a buffer head for being
528 uptodate and locking it in ntfs_file_read_compressed_block(). We now
529 lock the buffer and then check whether it is uptodate or not.
531 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
533 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
534 code and update callers, i.e. ntfs_iget(), to pass that error code
535 up instead of just using -EIO.
536 - Modifications to super.c to ensure that both mount and remount
537 cannot set any write related options when the driver is compiled
539 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
540 cache the current runlist element. This should improve performance
541 when reading very large and/or very fragmented data.
543 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
545 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
546 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
547 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
548 - Convert $MFT/$BITMAP access to attribute inode API and remove all
549 remnants of the ugly mftbmp address space and operations hack. This
550 means we finally have only one readpage function as well as only one
551 async io completion handler. Yey! The mft bitmap is now just an
552 attribute inode and is accessed from vol->mftbmp_ino just as if it
553 were a normal file. Fake inodes rule. (-:
555 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
557 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
558 remounts to fail when the partition had an entry in /etc/fstab and
559 the entry specified the nls= option.
560 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
561 expand all the helper functions NVolFoo(), NVolSetFoo(), and
563 - Move copyright statement from driver initialisation message to
564 module description (fs/super.c). This makes the initialisation
565 message fit on one line and fits in better with rest of kernel.
566 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
567 attribute inodes, and both for files and directories.
568 - Implement fake attribute inodes allowing all attribute i/o to go via
569 the page cache and to use all the normal vfs/mm functionality:
570 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
572 - Add needed cleanup code to ntfs_clear_big_inode().
573 - Merge address space operations for files and directories (aops.c),
574 now just have ntfs_aops:
576 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
577 ntfs_attr_read_block() -> ntfs_read_block(),
578 ntfs_file_read_page() -> ntfs_readpage().
579 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
580 attribute inodes, and both for files and directories.
581 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
583 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
585 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
586 the locking out of super.c::get_nr_free_mft_records() and taking and
587 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
588 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
589 current userspace ntfs library code. This means that if a merge
590 fails the original runlists are always left unmodified instead of
591 being silently corrupted.
594 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
596 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
597 from ntfs_volume structure.
598 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
599 from ntfs_volume structure.
600 - Use iget5_locked() and friends instead of conventional iget(). Wrap
601 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
602 to use ntfs_iget(). Leave only one iget() call at mount time so we
603 don't need an ntfs_iget_mount().
604 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
607 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
609 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
610 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
611 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
612 to determine whether to apply mst fixups or not.
613 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
614 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
615 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
616 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
617 the VFS readpage function prototype to the ntfs_attr_read_block()
620 2.0.11 - Initial preparations for fake inode based attribute i/o.
622 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
623 do some macro magic (adapted from include/linux/buffer_head.h) to
624 expand all the helper functions NInoFoo(), NInoSetFoo(), and
626 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
627 - Add new fields to ntfs_inode structure to allow use of fake inodes
628 for attribute i/o: type, name, name_len. Also add new state bits:
629 NI_Attr, which, if set, indicates the inode is a fake inode, and
630 NI_MstProtected, which, if set, indicates the attribute uses multi
631 sector transfer protection, i.e. fixups need to be applied after
632 reads and before/after writes.
633 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
634 ntfs_{new,clear,destroy}_extent_inode() and update callers.
635 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
636 instead of ntfs_destroy_extent_inode().
637 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
638 - Make all operations on ntfs inode state bits use the NIno* functions.
639 - Set up the new ntfs inode fields and state bits in
640 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
641 allocated memory to __ntfs_clear_inode().
642 - Cleanup ntfs_inode structure a bit for better ordering of elements
643 w.r.t. their size to allow better packing of the structure in memory.
645 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
647 - Add check at mount time to verify that the number of inodes on the
648 volume does not exceed 2^32 - 1, which is the maximum allowed for
649 NTFS according to Microsoft.
650 - Change mft_no member of ntfs_inode structure to be unsigned long.
651 Update all users. This makes ntfs_inode->mft_no just a copy of struct
652 inode->i_ino. But we can't just always use struct inode->i_ino and
653 remove mft_no because extent inodes do not have an attached struct
656 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
658 - Change decompression engine to use a single buffer protected by a
659 spin lock instead of per-CPU buffers. (Rusty Russell)
660 - Do not update cb_pos when handling a partial final page during
661 decompression of a sparse compression block, as the value is later
662 reset without being read/used. (Rusty Russell)
663 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
665 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
666 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
667 it also makes everything safer so it is a good thing.
668 - Miscellaneous minor cleanups to comments.
670 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
672 Big thanks go to Al Viro and other inhabitants of #kernel for investing
673 their time to discuss the case sensitivity and dcache aliasing issues.
675 - Remove unused source file fs/ntfs/attraops.c.
676 - Remove show_inodes mount option(s), thus dropping support for
677 displaying of short file names.
678 - Remove deprecated mount option posix.
679 - Restore show_sys_files mount option.
680 - Add new mount option case_sensitive, to determine if the driver
681 treats file names as case sensitive or not. If case sensitive, create
682 file names in the POSIX namespace. Otherwise create file names in the
683 LONG/WIN32 namespace. Note, files remain accessible via their short
684 file name, if it exists.
685 - Remove really dumb logic bug in boot sector recovery code.
686 - Fix dcache aliasing issues wrt short/long file names via changes
687 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
688 fs/ntfs/namei.c::ntfs_lookup():
689 - Add additional argument to ntfs_lookup_inode_by_name() in which we
690 return information about the matching file name if the case is not
691 matching or the match is a short file name. See comments above the
692 function definition for details.
693 - Change ntfs_lookup() to only create dcache entries for the correctly
694 cased file name and only for the WIN32 namespace counterpart of DOS
695 namespace file names. This ensures we have only one dentry per
696 directory and also removes all dcache aliasing issues between short
697 and long file names once we add write support. See comments above
698 function for details.
699 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
701 2.0.7 - Minor cleanups and updates for changes in core kernel code.
703 - Remove much of the NULL struct element initializers.
704 - Various updates to make compatible with recent kernels.
705 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
706 in fs/ntfs/ntfs.h instead.
707 - Remove no longer needed KERNEL_VERSION checks. We are now in the
708 kernel proper so they are no longer needed.
710 2.0.6 - Major bugfix to make compatible with other kernel changes.
712 - Initialize the mftbmp address space properly now that there are more
713 fields in the struct address_space. This was leading to hangs and
714 oopses on umount since 2.5.12 because of changes to other parts of
715 the kernel. We probably want a kernel generic init_address_space()
717 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
718 only caller of ->readdir() is vfs_readdir() which holds i_sem during
719 the call, and i_sem is sufficient protection against changes in the
720 directory inode (including ->i_size).
721 - Use generic_file_llseek() for directories (as opposed to
722 default_llseek()) as this downs i_sem instead of the BKL which is
723 what we now need for exclusion against ->f_pos changes considering we
724 no longer take the BKL in ntfs_readdir().
726 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
728 - No need to set old blocksize in super.c::ntfs_fill_super() as the
729 VFS does so via invocation of deactivate_super() calling
730 fs->fill_super() calling block_kill_super() which does it.
731 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
732 -> Do we really need it? I don't think so as we have exclusion on
733 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
734 move the ->f_pos accesses under the mrec_lock though. Check this...
735 - Fix really, really, really stupid buffer overflow in extent inode
736 handling in mft.c::map_extent_mft_record().
738 2.0.4 - Cleanups and updates for kernel 2.5.11.
740 - Add documentation on how to use the MD driver to be able to use NTFS
741 stripe and volume sets in Linux and generally cleanup documentation
743 Remove all uses of kdev_t in favour of struct block_device *:
744 - Change compress.c::ntfs_file_read_compressed_block() to use
745 sb_getblk() instead of getblk().
746 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
747 of get_hardsect_size().
748 - No need to get old blocksize in super.c::ntfs_fill_super() as
749 fs/super.c::get_sb_bdev() already does this.
750 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
752 2.0.3 - Small bug fixes, cleanups, and performance improvements.
754 - Remove some dead code from mft.c.
755 - Optimize readpage and read_block functions throughout aops.c so that
756 only initialized blocks are read. Non-initialized ones have their
757 buffer head mapped, zeroed, and set up to date, without scheduling
758 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
759 Thanks go to Andrew Morton for spotting the below:
760 - Fix buglet in allocate_compression_buffers() error code path.
761 - Call flush_dcache_page() after modifying page cache page contents in
762 ntfs_file_readpage().
763 - Check for existence of page buffers throughout aops.c before calling
764 create_empty_buffers(). This happens when an I/O error occurs and the
765 read is retried. (It also happens once writing is implemented so that
766 needed doing anyway but I had left it for later...)
767 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
768 readpage and read_block functions. Reasoning same as above (i.e. I/O
769 error retries and future write code paths.)
771 2.0.2 - Minor updates and cleanups.
773 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
774 and cleanup the code a bit, removing the unused size parameter.
775 - Change default fmask to 0177 and update documentation.
776 - Change attrib.c::get_attr_search_ctx() to return the search context
777 directly instead of taking the address of a pointer. A return value
778 of NULL means the allocation failed. Updated all callers
780 - Update to 2.5.9 kernel (preserving backwards compatibility) by
781 replacing all occurences of page->buffers with page_buffers(page).
782 - Fix minor bugs in runlist merging, also minor cleanup.
783 - Updates to bootsector layout and mft mirror contents descriptions.
784 - Small bug fix in error detection in unistr.c and some cleanups.
785 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
788 2.0.1 - Minor updates.
790 - Make default umask correspond to documentation.
791 - Improve documentation.
792 - Set default mode to include execute bit. The {u,f,d}mask can be used
793 to take it away if desired. This allows binaries to be executed from
794 a mounted ntfs partition.
796 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
798 - Add kill_super, just keeping up with the vfs changes in the kernel.
799 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
800 from the CVS import into BitKeeper.
801 - Begin to implement proper handling of allocated_size vs
802 initialized_size vs data_size (i.e. i_size). Done are
803 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
804 and attrib.c::load_attribute_list().
805 - Lock the runlist in attrib.c::load_attribute_list() while using it.
806 - Fix memory leak in ntfs_file_read_compressed_block() and generally
807 clean up compress.c a little, removing some uncommented/unused debug
809 - Tidy up dir.c a little bit.
810 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
811 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
812 creating aops.c::ntfs_mst_readpage(), improving the handling of
813 holes and overflow in the process and implementing the correct
814 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
815 I am aiming for correctness at the moment. Modularisation can come
817 - Rename aops.c::end_buffer_read_index_async() to
818 end_buffer_read_mst_async() and optimize the overflow checking and
820 - Use the host of the mftbmp address space mapping to hold the ntfs
821 volume. This is needed so the async i/o completion handler can
822 retrieve a pointer to the volume. Hopefully this will not cause
823 problems elsewhere in the kernel... Otherwise will need to use a
825 - Complete implementation of proper handling of allocated_size vs
826 initialized_size vs data_size (i.e. i_size) in whole driver.
827 Basically aops.c is now completely rewritten.
828 - Change NTFS driver name to just NTFS and set version number to 2.0.0
829 to make a clear distinction from the old driver which is still on
832 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
834 - Replace bdevname(sb->s_dev) with sb->s_id.
835 - Remove now superfluous new-line characters in all callers of
837 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
838 directories. Without this the "find" utility gets very upset which is
839 fair enough as Linux/Unix do not support directory hard links.
840 - Further runlist merging work. (Richard Russon)
841 - Backwards compatibility for gcc-2.95. (Richard Russon)
842 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
843 - Convert to new file system declaration using ->ntfs_get_sb() and
844 replacing ntfs_read_super() with ntfs_fill_super().
845 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
846 overflow on 32-bit architectures.
847 - Cleanup upcase loading code to use ntfs_(un)map_page().
848 - Disable/reenable preemtion in critical sections of compession engine.
849 - Replace device size determination in ntfs_fill_super() with
850 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
851 function super.c::get_nr_blocks().
852 - Implement a mount time option (show_inodes) allowing choice of which
853 types of inode names readdir() returns and modify ntfs_filldir()
854 accordingly. There are several parameters to show_inodes:
856 win32: long file names (including POSIX file names) [DEFAULT]
858 dos: short file names only (excluding POSIX file names)
860 posix: same as both win32 and dos
862 Note that the options are additive, i.e. specifying:
863 -o show_inodes=system,show_inodes=win32,show_inodes=dos
864 is the same as specifying:
866 Note that the "posix" and "all" options will show all directory
867 names, BUT the link count on each directory inode entry is set to 1,
868 due to Linux not supporting directory hard links. This may well
869 confuse some userspace applications, since the directory names will
870 have the same inode numbers. Thus it is NOT advisable to use the
871 "posix" or "all" options. We provide them only for completeness sake.
872 - Add copies of allocated_size, initialized_size, and compressed_size to
873 the ntfs inode structure and set them up in
874 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
875 for files and the index allocation attribute for directories.
876 - Add copies of allocated_size and initialized_size to ntfs inode for
877 $BITMAP attribute of large directories and set them up in
878 inode.c::ntfs_read_inode().
879 - Add copies of allocated_size and initialized_size to ntfs volume for
880 $BITMAP attribute of $MFT and set them up in
881 super.c::load_system_files().
882 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
883 posix, and utf8) and tell user what the new options to use are. Note
884 we still do support them but they will be removed with kernel 2.7.x.
885 - Change all occurences of integer long long printf formatting to hex
886 as printk() will not support long long integer format if/when the
887 div64 patch goes into the kernel.
888 - Make slab caches have stable names and change the names to what they
889 were intended to be. These changes are required/made possible by the
890 new slab cache name handling which removes the length limitation by
891 requiring the caller of kmem_cache_create() to supply a stable name
892 which is then referenced but not copied.
893 - Rename run_list structure to run_list_element and create a new
894 run_list structure containing a pointer to a run_list_element
895 structure and a read/write semaphore. Adapt all users of runlists
896 to new scheme and take and release the lock as needed. This fixes a
897 nasty race as the run_list changes even when inodes are locked for
898 reading and even when the inode isn't locked at all, so we really
899 needed the serialization. We use a semaphore rather than a spinlock
900 as memory allocations can sleep and doing everything GFP_ATOMIC
902 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
903 This can never happen due to the nature of lookup_attr() and how we
904 support attribute lists. If it did happen it would imply the inode
906 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
908 - Update to 2.5.6-pre2 changes in struct address_space.
909 - Use parent_ino() when accessing d_parent inode number in dir.c.
910 - Import Sourceforge CVS repository into BitKeeper repository:
911 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
912 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
913 Documentation/filesystems/ntfs.txt for NTFS TNG.
914 - Create kernel configuration option controlling whether debugging
916 - Add the required export of end_buffer_io_sync() from the patches
917 directory to the kernel code.
918 - Update inode.c::ntfs_show_options() with show_inodes mount option.
919 - Update errors mount option.
921 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
923 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
924 bug in mapping of extent inodes. Update all the comments to fit all
925 the recent code changes.
926 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
927 - Cleanups in compress.c, mostly comments and folding help.
928 - Implement attrib.c::map_run_list() as a generic helper.
929 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
930 thus making code shorter and enabling attribute list support.
931 - Cleanup incorrect use of [su]64 with %L printf format specifier in
932 all source files. Type casts to [unsigned] long long added to correct
933 the mismatches (important for architectures which have long long not
935 - Merge async io completion handlers for directory indexes and $MFT
936 data into one by setting the index_block_size{_bits} of the ntfs
937 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
938 - Cleanup aops.c, update comments.
939 - Make ntfs_file_get_block() use map_run_list() so all files now
940 support attribute lists.
941 - Make ntfs_dir_readpage() almost verbatim copy of
942 block_read_full_page() by using ntfs_file_get_block() with only real
943 difference being the use of our own async io completion handler
944 rather than the default one, thus reducing the amount of code and
945 automatically enabling attribute list support for directory indices.
946 - Fix bug in load_attribute_list() - forgot to call brelse in error
948 - Change parameters to find_attr() and lookup_attr(). We no longer
949 pass in the upcase table and its length. These can be gotten from
950 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
951 - Cleanups in attrib.c.
952 - Implement merging of runlists, attrib.c::merge_run_lists() and its
953 helpers. (Richard Russon)
954 - Attribute lists part 2, attribute extents and multi part runlists:
955 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
956 further runlist parts via attrib.c::map_run_list().
957 - Tiny endianness bug fix in decompress_mapping_pairs().
959 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
961 - Enable encrypted directories. (Their index root is marked encrypted
962 to indicate that new files in that directory should be created
964 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
965 - Enable $Extend system directory. Most (if not all) extended system
966 files do not have unnamed data attributes so ntfs_read_inode() had to
967 special case them but that is ok, as the special casing recovery
968 happens inside an error code path so there is zero slow down in the
969 normal fast path. The special casing is done by introducing a new
970 function inode.c::ntfs_is_extended_system_file() which checks if any
971 of the hard links in the inode point to $Extend as being their parent
972 directory and if they do we assume this is an extended system file.
973 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
974 when compiled with -DDEBUG. Default is debug messages to be disabled.
975 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
976 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
977 interface is enabled). Inspired by old ntfs driver.
978 - Add debug_msgs insmod/kernel boot parameter to set whether debug
979 messages are {dis,en}abled. This is useful to enable debug messages
980 during ntfs initialization and is the only way to activate debugging
981 when the sysctl interface is not enabled.
982 - Cleanup debug output in various places.
983 - Remove all dollar signs ($) from the source (except comments) to
984 enable compilation on architectures whose gcc compiler does not
985 support dollar signs in the names of variables/constants. Attribute
986 types now start with AT_ instead of $ and $I30 is now just I30.
987 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
988 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
989 access functions. This means we now cope with $MFT/$BITMAP being
990 spread accross several mft records.
991 - Disable modification of mft_zone_multiplier on remount. We can always
992 reenable this later on if we really want to, but we will need to make
993 sure we readjust the mft_zone size / layout accordingly.
995 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
997 - Use sb_set_blocksize() instead of set_blocksize() and verify the
999 - Use sb_bread() instead of bread() throughout.
1000 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1001 of a directory index block vcn. Apply resulting simplifications in
1003 - Fix a small bug somewhere (but forgot what it was).
1004 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1005 on the printf-format parameter list and fix bugs reported by gcc
1006 as a result. (Richard Russon)
1007 - Move inode allocation strategy to Al's new stuff but maintain the
1008 divorce of ntfs_inode from struct inode. To achieve this we have two
1009 separate slab caches, one for big ntfs inodes containing a struct
1010 inode and pure ntfs inodes and at the same time fix some faulty
1011 error code paths in ntfs_read_inode().
1012 - Show mount options in proc (inode.c::ntfs_show_options()).
1014 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1016 - Modified (un)map_mft_record functions to be common for read and write
1017 case. To specify which is which, added extra parameter at front of
1018 parameter list. Pass either READ or WRITE to this, each has the
1020 - General cleanups to allow for easier folding in vi.
1021 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1022 argument, and invokes attrib.c::merge_run_lists() to merge the old
1023 and the new runlists.
1024 - Removed attrib.c::find_first_attr().
1025 - Implemented loading of attribute list and complete runlist for $MFT.
1026 This means we now cope with $MFT being spread across several mft
1028 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1029 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1030 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1031 cache. There are too little ntfs_volume structures at any one time
1032 to justify a private slab cache.
1033 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1034 Use KM_BIO_IRQ on advice from IRC/kernel...
1035 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1036 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1037 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1038 removed the now superfluous exports from the kernel core patch.
1039 - Fix a bug where kfree() was used insted of ntfs_free().
1040 - Change map_mft_record() to take ntfs_inode as argument instead of
1041 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1042 - Add pointer to ntfs_volume to ntfs_inode.
1043 - Add mft record number and sequence number to ntfs_inode. Stop using
1044 i_ino and i_generation for in-driver purposes.
1045 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1046 - Remove use of proper inodes by extent inodes. Move i_ino and
1047 i_generation to ntfs_inode to do this. Apply simplifications that
1048 result and remove iget_no_wait(), etc.
1049 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1050 - Add reference counting in ntfs_inode for the ntfs inode itself and
1051 for the mapped mft record.
1052 - Extend mft record mapping so we can (un)map extent mft records (new
1053 functions (un)map_extent_mft_record), and so mappings are reference
1054 counted and don't have to happen twice if already mapped - just ref
1056 - Add -o iocharset as alias to -o nls for backwards compatibility.
1057 - The latest core patch is now tiny. In fact just a single additional
1058 export is necessary over the base kernel.
1060 tng-0.0.3 - Cleanups, enhancements, bug fixes.
1062 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1063 and setup the runlist appropriately using knowledge provided by the
1064 sizes in the base attribute record.
1065 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1067 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1068 page or use vmalloc depending on the amount of memory requested.
1069 - Cleanup error output. The __FUNCTION__ "(): " is now added
1070 automatically. Introduced a new header file debug.h to support this
1071 and also moved ntfs_debug() function into it.
1072 - Make reading of compressed files more intelligent and especially get
1073 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1074 buffers (allocated at first mount with cluster size <= 4kiB and
1075 deallocated on last umount with cluster size <= 4kiB), and
1076 asynchronous io for the compressed data using a list of buffer heads.
1077 Er, we use synchronous io as async io only works on whole pages
1078 covered by buffers and not on individual buffer heads...
1079 - Bug fix for reading compressed files with sparse compression blocks.
1081 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1083 - Fixed handling of directories when cluster size exceeds index block
1085 - Hide DOS only name space directory entries from readdir() but allow
1086 them in lookup(). This should fix the problem that Linux doesn't
1087 support directory hard links, while still allowing access to entries
1088 via their short file name. This also has the benefit of mimicking
1089 what Windows users are used to, so it is the ideal solution.
1090 - Implemented sync_page everywhere so no more hangs in D state when
1092 - Stop using bforget() in favour of brelse().
1093 - Stop locking buffers unnecessarily.
1094 - Implemented compressed files (inode->mapping contains uncompressed
1095 data, raw compressed data is currently bread() into a vmalloc()ed
1097 - Enable compressed directories. (Their index root is marked compressed
1098 to indicate that new files in that directory should be created
1100 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1101 functions. (Thanks to Will Dyson for pointing this out.)
1102 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1103 ntfs_sb_info) out of the common inode and super_block structures and
1104 started using the generic_ip and generic_sbp pointers instead. This
1105 makes ntfs entirely private with respect to the kernel tree.
1106 - Detect compiler version and abort with error message if gcc less than
1108 - Fix bug in name comparison function in unistr.c.
1109 - Implement attribute lists part 1, the infrastructure: search contexts
1110 and operations, find_external_attr(), lookup_attr()) and make the
1111 code use the infrastructure.
1112 - Fix stupid buffer overflow bug that became apparent on larger run
1113 list containing attributes.
1114 - Fix bugs in readdir() that became apparent on larger directories.
1116 The driver is now really useful and survives the test
1117 find . -type f -exec md5sum "{}" \;
1118 without any error messages on a over 1GiB sized partition with >16k
1119 files on it, including compressed files and directories and many files
1120 and directories with attribute lists.
1122 tng-0.0.1 - The first useful version.
1124 - Added ntfs_lookup().
1125 - Added default upcase generation and handling.
1126 - Added compile options to be shown on module init.
1127 - Many bug fixes that were "hidden" before.
1128 - Update to latest kernel.
1129 - Added ntfs_readdir().
1130 - Added file operations for mmap(), read(), open() and llseek(). We just
1131 use the generic ones. The whole point of going through implementing
1132 readpage() methods and where possible get_block() call backs is that
1133 this allows us to make use of the generic high level methods provided
1136 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1137 though and it doesn't implement accesssing compressed files yet. Also,
1138 accessing files with attribute list attributes is not implemented yet
1139 either. But for small or simple file systems it should work and allow
1140 you to list directories, use stat on directory entries and the file
1141 system, open, read, mmap and llseek around in files. A big mile stone
1144 tng-0.0.0 - Initial version tag.
1146 Initial driver implementation. The driver can mount and umount simple
1147 NTFS file systems (i.e. ones without attribute lists in the system
1148 files). If the mount fails there might be problems in the error handling
1149 code paths, so be warned. Otherwise it seems to be loading the system
1150 files nicely and the mft record read mapping/unmapping seems to be
1151 working nicely, too. Proof of inode metadata in the page cache and non-
1152 resident file unnamed stream data in the page cache concepts is thus