3 - Either invalidate quotas or update the quota charges on NTFS 3.x
4 volumes with quota tracking enabled ($Quota).
5 - Checkpoint or disable the user space journal ($UsnJrnl).
6 - Implement aops->set_page_dirty() in order to take control of buffer
7 dirtying. Not having it means if page_has_buffers(), all buffers
8 will be dirtied with the page. And if not they won't be. That is
9 fine for the moment but will break once we enable metadata updates.
10 For now just always using __set_page_dirty_nobuffers() for metadata
11 pages as nothing can dirty a page other than ourselves. Should this
12 change, we will really need to roll our own ->set_page_dirty().
13 - Implement sops->dirty_inode() to implement {a,m,c}time updates and
15 - Implement sops->write_inode().
16 - In between ntfs_prepare/commit_write, need exclusion between
17 simultaneous file extensions. Need perhaps an NInoResizeUnderway()
18 flag which we can set in ntfs_prepare_write() and clear again in
19 ntfs_commit_write(). Just have to be careful in readpage/writepage,
20 as well as in truncate, that we play nice... We might need to have
21 a data_size field in the ntfs_inode to store the real attribute
22 length. Also need to be careful with initialized_size extention in
23 ntfs_prepare_write. Basically, just be _very_ careful in this code...
24 OTOH, perhaps i_sem, which is held accross generic_file_write is
25 sufficient for synchronisation here. We then just need to make sure
26 ntfs_readpage/writepage/truncate interoperate properly with us.
28 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
30 - Add a new address space operations struct, ntfs_mst_aops, for mst
31 protected attributes. This is because the default ntfs_aops do not
32 make sense with mst protected data and were they to write anything to
33 such an attribute they would cause data corruption so we provide
34 ntfs_mst_aops which does not have any write related operations set.
35 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
36 includes an adapted ntfs_commit_inode() and an implementation of
37 ntfs_write_inode() which for now just cleans dirty inodes without
38 writing them (it does emit a warning that this is happening).
39 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
40 entry) as it was only fixing a theoretical bug but at the same time
41 it badly broke the handling of sparse and uncompressed compression
44 2.1.11 - Driver internal cleanups.
46 - Only build logfile.o if building the driver with read-write support.
47 - Really final white space cleanups.
48 - Use generic_ffs() instead of ffs() in logfile.c which allows the
49 log_page_size variable to be optimized by gcc into a constant.
50 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
51 char as defined by POSIX and as found on some systems.
53 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
55 - Finish off the white space cleanups (remove trailing spaces, etc).
56 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
57 the kludges around the first iget(). Instead of (re)setting ->s_op
58 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
59 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
60 kills the need for second super_operations and allows to return error
61 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
63 - Force read-only (re)mounting if any of the following bits are set in
64 the volume information flags:
65 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
66 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
67 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
68 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
69 above bits set so the test is made easy.
71 2.1.9 - Fix two bugs in decompression engine.
73 - Fix a bug where we would not always detect that we have reached the
74 end of a compression block because we were ending at minus one byte
75 which is effectively the same as being at the end. The fix is to
76 check whether the uncompressed buffer has been fully filled and if so
77 we assume we have reached the end of the compression block. A big
78 thank you to Marcin GibuĊa for the bug report, the assistance in
79 tracking down the bug and testing the fix.
80 - Fix a possible bug where when a compressed read is truncated to the
81 end of the file, the offset inside the last page was not truncated.
83 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
85 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
86 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
87 utc2ntfs() to work with struct timespec instead of time_t on the
88 Linux UTC time side thus preserving the full precision of the NTFS
89 time and only loosing up to 99 nano-seconds in the Linux UTC time.
90 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
92 - Remove unused ntfs_dirty_inode().
93 - Cleanup super operations declaration in fs/ntfs/super.c.
94 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
95 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
96 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
97 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
98 fs/ntfs/inode.h so they can be used elsewhere.
99 - Determine the mft mirror size as the number of mirrored mft records
100 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
101 - Load the mft mirror at mount time and compare the mft records stored
102 in it to the ones in the mft. Force a read-only mount if the two do
103 not match (fs/ntfs/super.c).
104 - Fix type casting related warnings on 64-bit architectures. Thanks
105 to Meelis Roos for reporting them.
106 - Move %L to %ll as %L is floating point and %ll is integer which is
108 - Read the journal ($LogFile) and determine if the volume has been
109 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
110 and fs/ntfs/logfile.c). This is a little bit of a crude check in
111 that we only look at the restart areas and not at the actual log
112 records so that there will be a very small number of cases where we
113 think that a volume is dirty when in fact it is clean. This should
114 only affect volumes that have not been shutdown cleanly and did not
115 have any pending, non-check-pointed i/o.
116 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
117 is requested, empty $LogFile by overwriting it with 0xff bytes to
118 ensure that Windows cannot cause data corruption by replaying a stale
119 journal after Linux has written to the volume.
121 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
123 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
124 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
125 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
126 default doesn't allow inode number 0 which is a valid inode on NTFS
127 and even if it did allow that it uses iget() instead of ntfs_iget()
128 which makes it useless for us.
129 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
130 default just returns -EACCES which is not very useful.
131 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
132 and set them up in the super block at mount time (super.c) this
133 allows mounted NTFS volumes to be exported via NFS.
134 - Add missing return -EOPNOTSUPP; in
135 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
136 - Enforce no atime and no dir atime updates at mount/remount time as
137 they are not implemented yet anyway.
138 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
139 after a NULL check. Thanks to Dave Jones for pointing this out.
141 2.1.6 - Fix minor bug in handling of compressed directories.
143 - Fix bug in handling of compressed directories. A compressed
144 directory is not really compressed so when we set the ->i_blocks
145 field of a compressed directory inode we were setting it from the
146 non-existing field ni->itype.compressed.size which gave random
147 results... For directories we now always use ni->allocated_size.
149 2.1.5 - Fix minor bug in attribute list attribute handling.
151 - Fix bug in attribute list handling. Actually it is not as much a bug
152 as too much protection in that we were not allowing attribute lists
153 which waste space on disk while Windows XP clearly allows it and in
154 fact creates such attribute lists so our driver was failing.
155 - Update NTFS documentation ready for 2.6 kernel release.
157 2.1.4 - Reduce compiler requirements.
159 - Remove all uses of unnamed structs and unions in the driver to make
160 old and newer gcc versions happy. Makes it a bit uglier IMO but at
161 least people will stop hassling me about it.
163 2.1.3 - Important bug fixes in corner cases.
165 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
166 clusters. (Philipp Thomas)
167 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
168 multiple of the block_size but not the cluster size. (Szabolcs
169 Szakacsits <szaka@sienet.hu>)
171 2.1.2 - Important bug fixes aleviating the hangs in statfs.
173 - Fix buggy free cluster and free inode determination logic.
175 2.1.1 - Minor updates.
177 - Add handling for initialized_size != data_size in compressed files.
178 - Reduce function local stack usage from 0x3d4 bytes to just noise in
179 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
180 - Remove compiler warnings for newer gcc.
181 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
182 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
183 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
184 kmap_atomic(KM_USER0).
186 2.1.0 - First steps towards write support: implement file overwrite.
188 - Add configuration option for developmental write support with an
189 appropriately scary configuration help text.
190 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
191 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
192 overwriting of existing files on ntfs. Note: Resident files are
193 only written into memory, and not written out to disk at present, so
194 avoid writing to files smaller than about 1kiB.
195 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
196 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
197 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
198 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
199 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
200 This enables write(2) based overwriting of existing files on ntfs.
201 Note: As with mmap(2) based overwriting, resident files are only
202 written into memory, and not written out to disk at present, so avoid
203 writing to files smaller than about 1kiB.
204 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
205 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
206 files with the purpose of intercepting and aborting all i_size
207 changes which we do not support yet. ntfs_truncate() actually only
208 emits a warning message but AFAICS our interception of i_size changes
209 elsewhere means ntfs_truncate() never gets called for i_size changes.
210 It is only called from generic_file_write() when we fail in
211 ntfs_prepare_{,nonresident_}write() in order to discard any
212 instantiated buffers beyond i_size. Thus i_size is not actually
213 changed so our warning message is enough. Unfortunately it is not
214 possible to easily determine if i_size is being changed or not hence
215 we just emit an appropriately worded error message.
217 2.0.25 - Small bug fixes and cleanups.
219 - Unlock the page in an out of memory error code path in
220 fs/ntfs/aops.c::ntfs_read_block().
221 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
222 just unlock the page and return. (This can happen due to ->writepage
223 clearing PageUptodate() during write out of MstProtected()
225 - Remove leaked write code again.
229 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
230 inside BUG_ON(). (Adam J. Richter)
231 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
232 calls for improved debugging. (Adam J. Richter)
233 - Add errors flag to the ntfs volume state, accessed via
234 NVol{,Set,Clear}Errors(vol).
235 - Do not allow read-write remounts of read-only volumes with errors.
236 - Clarify comment for ntfs file operation sendfile which was added by
237 Christoph Hellwig a while ago (just using generic_file_sendfile())
238 to say that ntfs ->sendfile is only used for the case where the
239 source data is on the ntfs partition and the destination is
240 somewhere else, i.e. nothing we need to concern ourselves with.
241 - Add generic_file_write() as our ntfs file write operation.
243 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
245 - Massive internal locking changes to mft record locking. Fixes lock
246 recursion and replaces the mrec_lock read/write semaphore with a
247 mutex. Also removes the now superfluous mft_count. This fixes several
248 race conditions and deadlocks, especially in the future write code.
249 - Fix ntfs over loopback for compressed files by adding an
250 optimization barrier. (gcc was screwing up otherwise ?)
251 - Miscellaneous cleanups all over the code and a fix or two in error
253 Thanks go to Christoph Hellwig for pointing out the following two:
254 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
255 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
257 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
259 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
260 at entry/exit respectively.
261 - Use C99 initializers for structures.
262 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
264 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
266 - Limit volume size at mount time to 2TiB on architectures where
267 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
268 This is the most we can do without overflowing the 32-bit limit of
269 the block device size imposed on us by sb_bread() and sb_getblk()
271 - Limit file/directory size at open() time to 16TiB on architectures
272 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
273 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
274 overflowing the page cache page index.
276 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
278 - Move the directory index bitmap to use an attribute inode instead of
279 having special fields for it inside the ntfs inode structure. This
280 means that the index bitmaps now use the page cache for i/o, too,
281 and also as a side effect we get support for non-resident index
283 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
284 fix a page leak that manifested itself in some cases.
285 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
286 index bitmap inode on the final iput().
288 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
290 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
291 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
292 - Drop the "file" from ntfs_file_read_compressed_block().
293 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
294 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
295 - Update ntfs_end_buffer_async_read() with the improved logic from
296 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
297 further logic improvements to better determine when we set PageError.
298 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
299 check for the buffers being uptodate first in line with the updated
300 fs/buffer.c::block_read_full_page(). This plugs a small race
303 2.0.18 - Fix race condition in reading of compressed files.
305 - There was a narrow window between checking a buffer head for being
306 uptodate and locking it in ntfs_file_read_compressed_block(). We now
307 lock the buffer and then check whether it is uptodate or not.
309 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
311 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
312 code and update callers, i.e. ntfs_iget(), to pass that error code
313 up instead of just using -EIO.
314 - Modifications to super.c to ensure that both mount and remount
315 cannot set any write related options when the driver is compiled
317 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
318 cache the current run list element. This should improve performance
319 when reading very large and/or very fragmented data.
321 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
323 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
324 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
325 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
326 - Convert $MFT/$BITMAP access to attribute inode API and remove all
327 remnants of the ugly mftbmp address space and operations hack. This
328 means we finally have only one readpage function as well as only one
329 async io completion handler. Yey! The mft bitmap is now just an
330 attribute inode and is accessed from vol->mftbmp_ino just as if it
331 were a normal file. Fake inodes rule. (-:
333 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
335 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
336 remounts to fail when the partition had an entry in /etc/fstab and
337 the entry specified the nls= option.
338 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
339 expand all the helper functions NVolFoo(), NVolSetFoo(), and
341 - Move copyright statement from driver initialisation message to
342 module description (fs/super.c). This makes the initialisation
343 message fit on one line and fits in better with rest of kernel.
344 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
345 attribute inodes, and both for files and directories.
346 - Implement fake attribute inodes allowing all attribute i/o to go via
347 the page cache and to use all the normal vfs/mm functionality:
348 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
350 - Add needed cleanup code to ntfs_clear_big_inode().
351 - Merge address space operations for files and directories (aops.c),
352 now just have ntfs_aops:
354 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
355 ntfs_attr_read_block() -> ntfs_read_block(),
356 ntfs_file_read_page() -> ntfs_readpage().
357 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
358 attribute inodes, and both for files and directories.
359 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
361 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
363 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
364 the locking out of super.c::get_nr_free_mft_records() and taking and
365 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
366 - Bring attribute run list merging code (fs/ntfs/attrib.c) in sync with
367 current userspace ntfs library code. This means that if a merge
368 fails the original run lists are always left unmodified instead of
369 being silently corrupted.
372 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
374 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
375 from ntfs_volume structure.
376 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
377 from ntfs_volume structure.
378 - Use iget5_locked() and friends instead of conventional iget(). Wrap
379 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
380 to use ntfs_iget(). Leave only one iget() call at mount time so we
381 don't need an ntfs_iget_mount().
382 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
385 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
387 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
388 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
389 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
390 to determine whether to apply mst fixups or not.
391 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
392 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
393 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
394 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
395 the VFS readpage function prototype to the ntfs_attr_read_block()
398 2.0.11 - Initial preparations for fake inode based attribute i/o.
400 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
401 do some macro magic (adapted from include/linux/buffer_head.h) to
402 expand all the helper functions NInoFoo(), NInoSetFoo(), and
404 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
405 - Add new fields to ntfs_inode structure to allow use of fake inodes
406 for attribute i/o: type, name, name_len. Also add new state bits:
407 NI_Attr, which, if set, indicates the inode is a fake inode, and
408 NI_MstProtected, which, if set, indicates the attribute uses multi
409 sector transfer protection, i.e. fixups need to be applied after
410 reads and before/after writes.
411 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
412 ntfs_{new,clear,destroy}_extent_inode() and update callers.
413 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
414 instead of ntfs_destroy_extent_inode().
415 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
416 - Make all operations on ntfs inode state bits use the NIno* functions.
417 - Set up the new ntfs inode fields and state bits in
418 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
419 allocated memory to __ntfs_clear_inode().
420 - Cleanup ntfs_inode structure a bit for better ordering of elements
421 w.r.t. their size to allow better packing of the structure in memory.
423 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
425 - Add check at mount time to verify that the number of inodes on the
426 volume does not exceed 2^32 - 1, which is the maximum allowed for
427 NTFS according to Microsoft.
428 - Change mft_no member of ntfs_inode structure to be unsigned long.
429 Update all users. This makes ntfs_inode->mft_no just a copy of struct
430 inode->i_ino. But we can't just always use struct inode->i_ino and
431 remove mft_no because extent inodes do not have an attached struct
434 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
436 - Change decompression engine to use a single buffer protected by a
437 spin lock instead of per-CPU buffers. (Rusty Russell)
438 - Do not update cb_pos when handling a partial final page during
439 decompression of a sparse compression block, as the value is later
440 reset without being read/used. (Rusty Russell)
441 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
443 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
444 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
445 it also makes everything safer so it is a good thing.
446 - Miscellaneous minor cleanups to comments.
448 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
450 Big thanks go to Al Viro and other inhabitants of #kernel for investing
451 their time to discuss the case sensitivity and dcache aliasing issues.
453 - Remove unused source file fs/ntfs/attraops.c.
454 - Remove show_inodes mount option(s), thus dropping support for
455 displaying of short file names.
456 - Remove deprecated mount option posix.
457 - Restore show_sys_files mount option.
458 - Add new mount option case_sensitive, to determine if the driver
459 treats file names as case sensitive or not. If case sensitive, create
460 file names in the POSIX namespace. Otherwise create file names in the
461 LONG/WIN32 namespace. Note, files remain accessible via their short
462 file name, if it exists.
463 - Remove really dumb logic bug in boot sector recovery code.
464 - Fix dcache aliasing issues wrt short/long file names via changes
465 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
466 fs/ntfs/namei.c::ntfs_lookup():
467 - Add additional argument to ntfs_lookup_inode_by_name() in which we
468 return information about the matching file name if the case is not
469 matching or the match is a short file name. See comments above the
470 function definition for details.
471 - Change ntfs_lookup() to only create dcache entries for the correctly
472 cased file name and only for the WIN32 namespace counterpart of DOS
473 namespace file names. This ensures we have only one dentry per
474 directory and also removes all dcache aliasing issues between short
475 and long file names once we add write support. See comments above
476 function for details.
477 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
479 2.0.7 - Minor cleanups and updates for changes in core kernel code.
481 - Remove much of the NULL struct element initializers.
482 - Various updates to make compatible with recent kernels.
483 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
484 in fs/ntfs/ntfs.h instead.
485 - Remove no longer needed KERNEL_VERSION checks. We are now in the
486 kernel proper so they are no longer needed.
488 2.0.6 - Major bugfix to make compatible with other kernel changes.
490 - Initialize the mftbmp address space properly now that there are more
491 fields in the struct address_space. This was leading to hangs and
492 oopses on umount since 2.5.12 because of changes to other parts of
493 the kernel. We probably want a kernel generic init_address_space()
495 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
496 only caller of ->readdir() is vfs_readdir() which holds i_sem during
497 the call, and i_sem is sufficient protection against changes in the
498 directory inode (including ->i_size).
499 - Use generic_file_llseek() for directories (as opposed to
500 default_llseek()) as this downs i_sem instead of the BKL which is
501 what we now need for exclusion against ->f_pos changes considering we
502 no longer take the BKL in ntfs_readdir().
504 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
506 - No need to set old blocksize in super.c::ntfs_fill_super() as the
507 VFS does so via invocation of deactivate_super() calling
508 fs->fill_super() calling block_kill_super() which does it.
509 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
510 -> Do we really need it? I don't think so as we have exclusion on
511 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
512 move the ->f_pos accesses under the mrec_lock though. Check this...
513 - Fix really, really, really stupid buffer overflow in extent inode
514 handling in mft.c::map_extent_mft_record().
516 2.0.4 - Cleanups and updates for kernel 2.5.11.
518 - Add documentation on how to use the MD driver to be able to use NTFS
519 stripe and volume sets in Linux and generally cleanup documentation
521 Remove all uses of kdev_t in favour of struct block_device *:
522 - Change compress.c::ntfs_file_read_compressed_block() to use
523 sb_getblk() instead of getblk().
524 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
525 of get_hardsect_size().
526 - No need to get old blocksize in super.c::ntfs_fill_super() as
527 fs/super.c::get_sb_bdev() already does this.
528 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
530 2.0.3 - Small bug fixes, cleanups, and performance improvements.
532 - Remove some dead code from mft.c.
533 - Optimize readpage and read_block functions throughout aops.c so that
534 only initialized blocks are read. Non-initialized ones have their
535 buffer head mapped, zeroed, and set up to date, without scheduling
536 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
537 Thanks go to Andrew Morton for spotting the below:
538 - Fix buglet in allocate_compression_buffers() error code path.
539 - Call flush_dcache_page() after modifying page cache page contents in
540 ntfs_file_readpage().
541 - Check for existence of page buffers throughout aops.c before calling
542 create_empty_buffers(). This happens when an I/O error occurs and the
543 read is retried. (It also happens once writing is implemented so that
544 needed doing anyway but I had left it for later...)
545 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
546 readpage and read_block functions. Reasoning same as above (i.e. I/O
547 error retries and future write code paths.)
549 2.0.2 - Minor updates and cleanups.
551 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
552 and cleanup the code a bit, removing the unused size parameter.
553 - Change default fmask to 0177 and update documentation.
554 - Change attrib.c::get_attr_search_ctx() to return the search context
555 directly instead of taking the address of a pointer. A return value
556 of NULL means the allocation failed. Updated all callers
558 - Update to 2.5.9 kernel (preserving backwards compatibility) by
559 replacing all occurences of page->buffers with page_buffers(page).
560 - Fix minor bugs in run list merging, also minor cleanup.
561 - Updates to bootsector layout and mft mirror contents descriptions.
562 - Small bug fix in error detection in unistr.c and some cleanups.
563 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
566 2.0.1 - Minor updates.
568 - Make default umask correspond to documentation.
569 - Improve documentation.
570 - Set default mode to include execute bit. The {u,f,d}mask can be used
571 to take it away if desired. This allows binaries to be executed from
572 a mounted ntfs partition.
574 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
576 - Add kill_super, just keeping up with the vfs changes in the kernel.
577 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
578 from the CVS import into BitKeeper.
579 - Begin to implement proper handling of allocated_size vs
580 initialized_size vs data_size (i.e. i_size). Done are
581 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
582 and attrib.c::load_attribute_list().
583 - Lock the run list in attrib.c::load_attribute_list() while using it.
584 - Fix memory leak in ntfs_file_read_compressed_block() and generally
585 clean up compress.c a little, removing some uncommented/unused debug
587 - Tidy up dir.c a little bit.
588 - Don't bother getting the run list in inode.c::ntfs_read_inode().
589 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
590 creating aops.c::ntfs_mst_readpage(), improving the handling of
591 holes and overflow in the process and implementing the correct
592 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
593 I am aiming for correctness at the moment. Modularisation can come
595 - Rename aops.c::end_buffer_read_index_async() to
596 end_buffer_read_mst_async() and optimize the overflow checking and
598 - Use the host of the mftbmp address space mapping to hold the ntfs
599 volume. This is needed so the async i/o completion handler can
600 retrieve a pointer to the volume. Hopefully this will not cause
601 problems elsewhere in the kernel... Otherwise will need to use a
603 - Complete implementation of proper handling of allocated_size vs
604 initialized_size vs data_size (i.e. i_size) in whole driver.
605 Basically aops.c is now completely rewritten.
606 - Change NTFS driver name to just NTFS and set version number to 2.0.0
607 to make a clear distinction from the old driver which is still on
610 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
612 - Replace bdevname(sb->s_dev) with sb->s_id.
613 - Remove now superfluous new-line characters in all callers of
615 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
616 directories. Without this the "find" utility gets very upset which is
617 fair enough as Linux/Unix do not support directory hard links.
618 - Further run list merging work. (Richard Russon)
619 - Backwards compatibility for gcc-2.95. (Richard Russon)
620 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
621 - Convert to new file system declaration using ->ntfs_get_sb() and
622 replacing ntfs_read_super() with ntfs_fill_super().
623 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
624 overflow on 32-bit architectures.
625 - Cleanup upcase loading code to use ntfs_(un)map_page().
626 - Disable/reenable preemtion in critical sections of compession engine.
627 - Replace device size determination in ntfs_fill_super() with
628 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
629 function super.c::get_nr_blocks().
630 - Implement a mount time option (show_inodes) allowing choice of which
631 types of inode names readdir() returns and modify ntfs_filldir()
632 accordingly. There are several parameters to show_inodes:
634 win32: long file names (including POSIX file names) [DEFAULT]
636 dos: short file names only (excluding POSIX file names)
638 posix: same as both win32 and dos
640 Note that the options are additive, i.e. specifying:
641 -o show_inodes=system,show_inodes=win32,show_inodes=dos
642 is the same as specifying:
644 Note that the "posix" and "all" options will show all directory
645 names, BUT the link count on each directory inode entry is set to 1,
646 due to Linux not supporting directory hard links. This may well
647 confuse some userspace applications, since the directory names will
648 have the same inode numbers. Thus it is NOT advisable to use the
649 "posix" or "all" options. We provide them only for completeness sake.
650 - Add copies of allocated_size, initialized_size, and compressed_size to
651 the ntfs inode structure and set them up in
652 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
653 for files and the index allocation attribute for directories.
654 - Add copies of allocated_size and initialized_size to ntfs inode for
655 $BITMAP attribute of large directories and set them up in
656 inode.c::ntfs_read_inode().
657 - Add copies of allocated_size and initialized_size to ntfs volume for
658 $BITMAP attribute of $MFT and set them up in
659 super.c::load_system_files().
660 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
661 posix, and utf8) and tell user what the new options to use are. Note
662 we still do support them but they will be removed with kernel 2.7.x.
663 - Change all occurences of integer long long printf formatting to hex
664 as printk() will not support long long integer format if/when the
665 div64 patch goes into the kernel.
666 - Make slab caches have stable names and change the names to what they
667 were intended to be. These changes are required/made possible by the
668 new slab cache name handling which removes the length limitation by
669 requiring the caller of kmem_cache_create() to supply a stable name
670 which is then referenced but not copied.
671 - Rename run_list structure to run_list_element and create a new
672 run_list structure containing a pointer to a run_list_element
673 structure and a read/write semaphore. Adapt all users of run lists
674 to new scheme and take and release the lock as needed. This fixes a
675 nasty race as the run_list changes even when inodes are locked for
676 reading and even when the inode isn't locked at all, so we really
677 needed the serialization. We use a semaphore rather than a spinlock
678 as memory allocations can sleep and doing everything GFP_ATOMIC
680 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
681 This can never happen due to the nature of lookup_attr() and how we
682 support attribute lists. If it did happen it would imply the inode
684 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
686 - Update to 2.5.6-pre2 changes in struct address_space.
687 - Use parent_ino() when accessing d_parent inode number in dir.c.
688 - Import Sourceforge CVS repository into BitKeeper repository:
689 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
690 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
691 Documentation/filesystems/ntfs.txt for NTFS TNG.
692 - Create kernel configuration option controlling whether debugging
694 - Add the required export of end_buffer_io_sync() from the patches
695 directory to the kernel code.
696 - Update inode.c::ntfs_show_options() with show_inodes mount option.
697 - Update errors mount option.
699 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
701 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
702 bug in mapping of extent inodes. Update all the comments to fit all
703 the recent code changes.
704 - Modify vcn_to_lcn() to cope with entirely unmapped run lists.
705 - Cleanups in compress.c, mostly comments and folding help.
706 - Implement attrib.c::map_run_list() as a generic helper.
707 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
708 thus making code shorter and enabling attribute list support.
709 - Cleanup incorrect use of [su]64 with %L printf format specifier in
710 all source files. Type casts to [unsigned] long long added to correct
711 the mismatches (important for architectures which have long long not
713 - Merge async io completion handlers for directory indexes and $MFT
714 data into one by setting the index_block_size{_bits} of the ntfs
715 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
716 - Cleanup aops.c, update comments.
717 - Make ntfs_file_get_block() use map_run_list() so all files now
718 support attribute lists.
719 - Make ntfs_dir_readpage() almost verbatim copy of
720 block_read_full_page() by using ntfs_file_get_block() with only real
721 difference being the use of our own async io completion handler
722 rather than the default one, thus reducing the amount of code and
723 automatically enabling attribute list support for directory indices.
724 - Fix bug in load_attribute_list() - forgot to call brelse in error
726 - Change parameters to find_attr() and lookup_attr(). We no longer
727 pass in the upcase table and its length. These can be gotten from
728 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
729 - Cleanups in attrib.c.
730 - Implement merging of run lists, attrib.c::merge_run_lists() and its
731 helpers. (Richard Russon)
732 - Attribute lists part 2, attribute extents and multi part run lists:
733 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
734 further run list parts via attrib.c::map_run_list().
735 - Tiny endianness bug fix in decompress_mapping_pairs().
737 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
739 - Enable encrypted directories. (Their index root is marked encrypted
740 to indicate that new files in that directory should be created
742 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
743 - Enable $Extend system directory. Most (if not all) extended system
744 files do not have unnamed data attributes so ntfs_read_inode() had to
745 special case them but that is ok, as the special casing recovery
746 happens inside an error code path so there is zero slow down in the
747 normal fast path. The special casing is done by introducing a new
748 function inode.c::ntfs_is_extended_system_file() which checks if any
749 of the hard links in the inode point to $Extend as being their parent
750 directory and if they do we assume this is an extended system file.
751 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
752 when compiled with -DDEBUG. Default is debug messages to be disabled.
753 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
754 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
755 interface is enabled). Inspired by old ntfs driver.
756 - Add debug_msgs insmod/kernel boot parameter to set whether debug
757 messages are {dis,en}abled. This is useful to enable debug messages
758 during ntfs initialization and is the only way to activate debugging
759 when the sysctl interface is not enabled.
760 - Cleanup debug output in various places.
761 - Remove all dollar signs ($) from the source (except comments) to
762 enable compilation on architectures whose gcc compiler does not
763 support dollar signs in the names of variables/constants. Attribute
764 types now start with AT_ instead of $ and $I30 is now just I30.
765 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
766 - Load complete run list for $MFT/$BITMAP during mount and cleanup
767 access functions. This means we now cope with $MFT/$BITMAP being
768 spread accross several mft records.
769 - Disable modification of mft_zone_multiplier on remount. We can always
770 reenable this later on if we really want to, but we will need to make
771 sure we readjust the mft_zone size / layout accordingly.
773 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
775 - Use sb_set_blocksize() instead of set_blocksize() and verify the
777 - Use sb_bread() instead of bread() throughout.
778 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
779 of a directory index block vcn. Apply resulting simplifications in
781 - Fix a small bug somewhere (but forgot what it was).
782 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
783 on the printf-format parameter list and fix bugs reported by gcc
784 as a result. (Richard Russon)
785 - Move inode allocation strategy to Al's new stuff but maintain the
786 divorce of ntfs_inode from struct inode. To achieve this we have two
787 separate slab caches, one for big ntfs inodes containing a struct
788 inode and pure ntfs inodes and at the same time fix some faulty
789 error code paths in ntfs_read_inode().
790 - Show mount options in proc (inode.c::ntfs_show_options()).
792 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
794 - Modified (un)map_mft_record functions to be common for read and write
795 case. To specify which is which, added extra parameter at front of
796 parameter list. Pass either READ or WRITE to this, each has the
798 - General cleanups to allow for easier folding in vi.
799 - attrib.c::decompress_mapping_pairs() now accepts the old run list
800 argument, and invokes attrib.c::merge_run_lists() to merge the old
801 and the new run lists.
802 - Removed attrib.c::find_first_attr().
803 - Implemented loading of attribute list and complete run list for $MFT.
804 This means we now cope with $MFT being spread across several mft
806 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
807 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
808 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
809 cache. There are too little ntfs_volume structures at any one time
810 to justify a private slab cache.
811 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
812 Use KM_BIO_IRQ on advice from IRC/kernel...
813 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
814 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
815 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
816 removed the now superfluous exports from the kernel core patch.
817 - Fix a bug where kfree() was used insted of ntfs_free().
818 - Change map_mft_record() to take ntfs_inode as argument instead of
819 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
820 - Add pointer to ntfs_volume to ntfs_inode.
821 - Add mft record number and sequence number to ntfs_inode. Stop using
822 i_ino and i_generation for in-driver purposes.
823 - Implement attrib.c::merge_run_lists(). (Richard Russon)
824 - Remove use of proper inodes by extent inodes. Move i_ino and
825 i_generation to ntfs_inode to do this. Apply simplifications that
826 result and remove iget_no_wait(), etc.
827 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
828 - Add reference counting in ntfs_inode for the ntfs inode itself and
829 for the mapped mft record.
830 - Extend mft record mapping so we can (un)map extent mft records (new
831 functions (un)map_extent_mft_record), and so mappings are reference
832 counted and don't have to happen twice if already mapped - just ref
834 - Add -o iocharset as alias to -o nls for backwards compatibility.
835 - The latest core patch is now tiny. In fact just a single additional
836 export is necessary over the base kernel.
838 tng-0.0.3 - Cleanups, enhancements, bug fixes.
840 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
841 and setup the run list appropriately using knowledge provided by the
842 sizes in the base attribute record.
843 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
845 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
846 page or use vmalloc depending on the amount of memory requested.
847 - Cleanup error output. The __FUNCTION__ "(): " is now added
848 automatically. Introduced a new header file debug.h to support this
849 and also moved ntfs_debug() function into it.
850 - Make reading of compressed files more intelligent and especially get
851 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
852 buffers (allocated at first mount with cluster size <= 4kiB and
853 deallocated on last umount with cluster size <= 4kiB), and
854 asynchronous io for the compressed data using a list of buffer heads.
855 Er, we use synchronous io as async io only works on whole pages
856 covered by buffers and not on individual buffer heads...
857 - Bug fix for reading compressed files with sparse compression blocks.
859 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
861 - Fixed handling of directories when cluster size exceeds index block
863 - Hide DOS only name space directory entries from readdir() but allow
864 them in lookup(). This should fix the problem that Linux doesn't
865 support directory hard links, while still allowing access to entries
866 via their short file name. This also has the benefit of mimicking
867 what Windows users are used to, so it is the ideal solution.
868 - Implemented sync_page everywhere so no more hangs in D state when
870 - Stop using bforget() in favour of brelse().
871 - Stop locking buffers unnecessarily.
872 - Implemented compressed files (inode->mapping contains uncompressed
873 data, raw compressed data is currently bread() into a vmalloc()ed
875 - Enable compressed directories. (Their index root is marked compressed
876 to indicate that new files in that directory should be created
878 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
879 functions. (Thanks to Will Dyson for pointing this out.)
880 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
881 ntfs_sb_info) out of the common inode and super_block structures and
882 started using the generic_ip and generic_sbp pointers instead. This
883 makes ntfs entirely private with respect to the kernel tree.
884 - Detect compiler version and abort with error message if gcc less than
886 - Fix bug in name comparison function in unistr.c.
887 - Implement attribute lists part 1, the infrastructure: search contexts
888 and operations, find_external_attr(), lookup_attr()) and make the
889 code use the infrastructure.
890 - Fix stupid buffer overflow bug that became apparent on larger run
891 list containing attributes.
892 - Fix bugs in readdir() that became apparent on larger directories.
894 The driver is now really useful and survives the test
895 find . -type f -exec md5sum "{}" \;
896 without any error messages on a over 1GiB sized partition with >16k
897 files on it, including compressed files and directories and many files
898 and directories with attribute lists.
900 tng-0.0.1 - The first useful version.
902 - Added ntfs_lookup().
903 - Added default upcase generation and handling.
904 - Added compile options to be shown on module init.
905 - Many bug fixes that were "hidden" before.
906 - Update to latest kernel.
907 - Added ntfs_readdir().
908 - Added file operations for mmap(), read(), open() and llseek(). We just
909 use the generic ones. The whole point of going through implementing
910 readpage() methods and where possible get_block() call backs is that
911 this allows us to make use of the generic high level methods provided
914 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
915 though and it doesn't implement accesssing compressed files yet. Also,
916 accessing files with attribute list attributes is not implemented yet
917 either. But for small or simple file systems it should work and allow
918 you to list directories, use stat on directory entries and the file
919 system, open, read, mmap and llseek around in files. A big mile stone
922 tng-0.0.0 - Initial version tag.
924 Initial driver implementation. The driver can mount and umount simple
925 NTFS file systems (i.e. ones without attribute lists in the system
926 files). If the mount fails there might be problems in the error handling
927 code paths, so be warned. Otherwise it seems to be loading the system
928 files nicely and the mft record read mapping/unmapping seems to be
929 working nicely, too. Proof of inode metadata in the page cache and non-
930 resident file unnamed stream data in the page cache concepts is thus