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
14 such things. This should probably just flag the ntfs inode such that
15 sops->write_inode(), i.e. ntfs_write_inode(), will copy the times
16 when it is invoked rather than having to update the mft record
18 - In between ntfs_prepare/commit_write, need exclusion between
19 simultaneous file extensions. Need perhaps an NInoResizeUnderway()
20 flag which we can set in ntfs_prepare_write() and clear again in
21 ntfs_commit_write(). Just have to be careful in readpage/writepage,
22 as well as in truncate, that we play nice... We might need to have
23 a data_size field in the ntfs_inode to store the real attribute
24 length. Also need to be careful with initialized_size extention in
25 ntfs_prepare_write. Basically, just be _very_ careful in this code...
26 OTOH, perhaps i_sem, which is held accross generic_file_write is
27 sufficient for synchronisation here. We then just need to make sure
28 ntfs_readpage/writepage/truncate interoperate properly with us.
29 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
30 leave the volume dirty on umount if the final iput(vol->mft_ino)
31 causes a write of any mirrored mft records due to the mft mirror
32 inode having been discarded already. Whether this can actually ever
33 happen is unclear however so it is worth waiting until someone hits
35 - Enable the code for setting the NT4 compatibility flag when we start
36 making NTFS 1.2 specific modifications.
38 2.1.14 - Fix an NFSd caused deadlock reported by several users.
40 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
41 to a buffer so that we can put the search context and unmap the mft
42 record before calling the filldir() callback. We need to do this
43 because of NFSd which calls ->lookup() from its filldir callback()
44 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
45 of the directory and since ntfs_readdir() has got it mapped already
46 ntfs_lookup() deadlocks.
48 2.1.13 - Enable overwriting of resident files and housekeeping of system files.
50 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
51 keeping the mft mirror in sync with the mft when mirrored mft records
52 are written. The functions are write_mft_record{,_nolock}(). The
53 implementation is quite rudimentary for now with lots of things not
54 implemented yet but I am not sure any of them can actually occur so
55 I will wait for people to hit each one and only then implement it.
56 - Commit open system inodes at umount time. This should make it
57 virtually impossible for sync_mft_mirror_umount() to ever be needed.
58 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
59 ntfs super operations. This gives us inode writing via the VFS inode
60 dirty code paths. Note: Access time updates are not implemented yet.
61 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
62 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
63 finally enabling resident file overwrite! (-8 This also includes a
64 placeholder for ->writepage (ntfs_mft_writepage()), which for now
65 just redirties the page and returns. Also, at umount time, we for
66 now throw away all mft data page cache pages after the last call to
67 ntfs_commit_inode() in the hope that all inodes will have been
68 written out by then and hence no dirty (meta)data will be lost. We
69 also check for this case and emit an error message telling the user
71 - Use set_page_writeback() and end_page_writeback() in the resident
72 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
73 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
75 - Implement ntfs_mft_writepage() so it now checks if any of the mft
76 records in the page are dirty and if so redirties the page and
77 returns. Otherwise it just returns (after doing set_page_writeback(),
78 unlock_page(), end_page_writeback() or the radix-tree tag
79 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
80 alowing the VM to do with the page as it pleases. Also, at umount
81 time, now only throw away dirty mft (meta)data pages if dirty inodes
82 are present and ask the user to email us if they see this happening.
83 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
84 information flags (fs/ntfs/super.c).
85 - Mark the volume dirty when (re)mounting read-write and mark it clean
86 when unmounting or remounting read-only. If any volume errors are
87 found, the volume is left marked dirty to force chkdsk to run.
88 - Add code to set the NT4 compatibility flag when (re)mounting
89 read-write for newer NTFS versions but leave it commented out for now
90 since we do not make any modifications that are NTFS 1.2 specific yet
91 and since setting this flag breaks Captive-NTFS which is not nice.
92 This code must be enabled once we start writing NTFS 1.2 specific
93 changes otherwise Windows NTFS driver might crash / cause corruption.
95 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
97 - Add a new address space operations struct, ntfs_mst_aops, for mst
98 protected attributes. This is because the default ntfs_aops do not
99 make sense with mst protected data and were they to write anything to
100 such an attribute they would cause data corruption so we provide
101 ntfs_mst_aops which does not have any write related operations set.
102 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
103 includes an adapted ntfs_commit_inode() and an implementation of
104 ntfs_write_inode() which for now just cleans dirty inodes without
105 writing them (it does emit a warning that this is happening).
106 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
107 entry) as it was only fixing a theoretical bug but at the same time
108 it badly broke the handling of sparse and uncompressed compression
111 2.1.11 - Driver internal cleanups.
113 - Only build logfile.o if building the driver with read-write support.
114 - Really final white space cleanups.
115 - Use generic_ffs() instead of ffs() in logfile.c which allows the
116 log_page_size variable to be optimized by gcc into a constant.
117 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
118 char as defined by POSIX and as found on some systems.
120 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
122 - Finish off the white space cleanups (remove trailing spaces, etc).
123 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
124 the kludges around the first iget(). Instead of (re)setting ->s_op
125 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
126 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
127 kills the need for second super_operations and allows to return error
128 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
130 - Force read-only (re)mounting if any of the following bits are set in
131 the volume information flags:
132 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
133 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
134 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
135 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
136 above bits set so the test is made easy.
138 2.1.9 - Fix two bugs in decompression engine.
140 - Fix a bug where we would not always detect that we have reached the
141 end of a compression block because we were ending at minus one byte
142 which is effectively the same as being at the end. The fix is to
143 check whether the uncompressed buffer has been fully filled and if so
144 we assume we have reached the end of the compression block. A big
145 thank you to Marcin GibuĊa for the bug report, the assistance in
146 tracking down the bug and testing the fix.
147 - Fix a possible bug where when a compressed read is truncated to the
148 end of the file, the offset inside the last page was not truncated.
150 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
152 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
153 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
154 utc2ntfs() to work with struct timespec instead of time_t on the
155 Linux UTC time side thus preserving the full precision of the NTFS
156 time and only loosing up to 99 nano-seconds in the Linux UTC time.
157 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
159 - Remove unused ntfs_dirty_inode().
160 - Cleanup super operations declaration in fs/ntfs/super.c.
161 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
162 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
163 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
164 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
165 fs/ntfs/inode.h so they can be used elsewhere.
166 - Determine the mft mirror size as the number of mirrored mft records
167 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
168 - Load the mft mirror at mount time and compare the mft records stored
169 in it to the ones in the mft. Force a read-only mount if the two do
170 not match (fs/ntfs/super.c).
171 - Fix type casting related warnings on 64-bit architectures. Thanks
172 to Meelis Roos for reporting them.
173 - Move %L to %ll as %L is floating point and %ll is integer which is
175 - Read the journal ($LogFile) and determine if the volume has been
176 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
177 and fs/ntfs/logfile.c). This is a little bit of a crude check in
178 that we only look at the restart areas and not at the actual log
179 records so that there will be a very small number of cases where we
180 think that a volume is dirty when in fact it is clean. This should
181 only affect volumes that have not been shutdown cleanly and did not
182 have any pending, non-check-pointed i/o.
183 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
184 is requested, empty $LogFile by overwriting it with 0xff bytes to
185 ensure that Windows cannot cause data corruption by replaying a stale
186 journal after Linux has written to the volume.
188 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
190 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
191 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
192 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
193 default doesn't allow inode number 0 which is a valid inode on NTFS
194 and even if it did allow that it uses iget() instead of ntfs_iget()
195 which makes it useless for us.
196 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
197 default just returns -EACCES which is not very useful.
198 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
199 and set them up in the super block at mount time (super.c) this
200 allows mounted NTFS volumes to be exported via NFS.
201 - Add missing return -EOPNOTSUPP; in
202 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
203 - Enforce no atime and no dir atime updates at mount/remount time as
204 they are not implemented yet anyway.
205 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
206 after a NULL check. Thanks to Dave Jones for pointing this out.
208 2.1.6 - Fix minor bug in handling of compressed directories.
210 - Fix bug in handling of compressed directories. A compressed
211 directory is not really compressed so when we set the ->i_blocks
212 field of a compressed directory inode we were setting it from the
213 non-existing field ni->itype.compressed.size which gave random
214 results... For directories we now always use ni->allocated_size.
216 2.1.5 - Fix minor bug in attribute list attribute handling.
218 - Fix bug in attribute list handling. Actually it is not as much a bug
219 as too much protection in that we were not allowing attribute lists
220 which waste space on disk while Windows XP clearly allows it and in
221 fact creates such attribute lists so our driver was failing.
222 - Update NTFS documentation ready for 2.6 kernel release.
224 2.1.4 - Reduce compiler requirements.
226 - Remove all uses of unnamed structs and unions in the driver to make
227 old and newer gcc versions happy. Makes it a bit uglier IMO but at
228 least people will stop hassling me about it.
230 2.1.3 - Important bug fixes in corner cases.
232 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
233 clusters. (Philipp Thomas)
234 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
235 multiple of the block_size but not the cluster size. (Szabolcs
236 Szakacsits <szaka@sienet.hu>)
238 2.1.2 - Important bug fixes aleviating the hangs in statfs.
240 - Fix buggy free cluster and free inode determination logic.
242 2.1.1 - Minor updates.
244 - Add handling for initialized_size != data_size in compressed files.
245 - Reduce function local stack usage from 0x3d4 bytes to just noise in
246 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
247 - Remove compiler warnings for newer gcc.
248 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
249 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
250 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
251 kmap_atomic(KM_USER0).
253 2.1.0 - First steps towards write support: implement file overwrite.
255 - Add configuration option for developmental write support with an
256 appropriately scary configuration help text.
257 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
258 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
259 overwriting of existing files on ntfs. Note: Resident files are
260 only written into memory, and not written out to disk at present, so
261 avoid writing to files smaller than about 1kiB.
262 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
263 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
264 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
265 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
266 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
267 This enables write(2) based overwriting of existing files on ntfs.
268 Note: As with mmap(2) based overwriting, resident files are only
269 written into memory, and not written out to disk at present, so avoid
270 writing to files smaller than about 1kiB.
271 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
272 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
273 files with the purpose of intercepting and aborting all i_size
274 changes which we do not support yet. ntfs_truncate() actually only
275 emits a warning message but AFAICS our interception of i_size changes
276 elsewhere means ntfs_truncate() never gets called for i_size changes.
277 It is only called from generic_file_write() when we fail in
278 ntfs_prepare_{,nonresident_}write() in order to discard any
279 instantiated buffers beyond i_size. Thus i_size is not actually
280 changed so our warning message is enough. Unfortunately it is not
281 possible to easily determine if i_size is being changed or not hence
282 we just emit an appropriately worded error message.
284 2.0.25 - Small bug fixes and cleanups.
286 - Unlock the page in an out of memory error code path in
287 fs/ntfs/aops.c::ntfs_read_block().
288 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
289 just unlock the page and return. (This can happen due to ->writepage
290 clearing PageUptodate() during write out of MstProtected()
292 - Remove leaked write code again.
296 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
297 inside BUG_ON(). (Adam J. Richter)
298 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
299 calls for improved debugging. (Adam J. Richter)
300 - Add errors flag to the ntfs volume state, accessed via
301 NVol{,Set,Clear}Errors(vol).
302 - Do not allow read-write remounts of read-only volumes with errors.
303 - Clarify comment for ntfs file operation sendfile which was added by
304 Christoph Hellwig a while ago (just using generic_file_sendfile())
305 to say that ntfs ->sendfile is only used for the case where the
306 source data is on the ntfs partition and the destination is
307 somewhere else, i.e. nothing we need to concern ourselves with.
308 - Add generic_file_write() as our ntfs file write operation.
310 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
312 - Massive internal locking changes to mft record locking. Fixes lock
313 recursion and replaces the mrec_lock read/write semaphore with a
314 mutex. Also removes the now superfluous mft_count. This fixes several
315 race conditions and deadlocks, especially in the future write code.
316 - Fix ntfs over loopback for compressed files by adding an
317 optimization barrier. (gcc was screwing up otherwise ?)
318 - Miscellaneous cleanups all over the code and a fix or two in error
320 Thanks go to Christoph Hellwig for pointing out the following two:
321 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
322 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
324 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
326 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
327 at entry/exit respectively.
328 - Use C99 initializers for structures.
329 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
331 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
333 - Limit volume size at mount time to 2TiB on architectures where
334 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
335 This is the most we can do without overflowing the 32-bit limit of
336 the block device size imposed on us by sb_bread() and sb_getblk()
338 - Limit file/directory size at open() time to 16TiB on architectures
339 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
340 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
341 overflowing the page cache page index.
343 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
345 - Move the directory index bitmap to use an attribute inode instead of
346 having special fields for it inside the ntfs inode structure. This
347 means that the index bitmaps now use the page cache for i/o, too,
348 and also as a side effect we get support for non-resident index
350 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
351 fix a page leak that manifested itself in some cases.
352 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
353 index bitmap inode on the final iput().
355 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
357 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
358 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
359 - Drop the "file" from ntfs_file_read_compressed_block().
360 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
361 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
362 - Update ntfs_end_buffer_async_read() with the improved logic from
363 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
364 further logic improvements to better determine when we set PageError.
365 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
366 check for the buffers being uptodate first in line with the updated
367 fs/buffer.c::block_read_full_page(). This plugs a small race
370 2.0.18 - Fix race condition in reading of compressed files.
372 - There was a narrow window between checking a buffer head for being
373 uptodate and locking it in ntfs_file_read_compressed_block(). We now
374 lock the buffer and then check whether it is uptodate or not.
376 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
378 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
379 code and update callers, i.e. ntfs_iget(), to pass that error code
380 up instead of just using -EIO.
381 - Modifications to super.c to ensure that both mount and remount
382 cannot set any write related options when the driver is compiled
384 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
385 cache the current run list element. This should improve performance
386 when reading very large and/or very fragmented data.
388 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
390 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
391 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
392 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
393 - Convert $MFT/$BITMAP access to attribute inode API and remove all
394 remnants of the ugly mftbmp address space and operations hack. This
395 means we finally have only one readpage function as well as only one
396 async io completion handler. Yey! The mft bitmap is now just an
397 attribute inode and is accessed from vol->mftbmp_ino just as if it
398 were a normal file. Fake inodes rule. (-:
400 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
402 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
403 remounts to fail when the partition had an entry in /etc/fstab and
404 the entry specified the nls= option.
405 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
406 expand all the helper functions NVolFoo(), NVolSetFoo(), and
408 - Move copyright statement from driver initialisation message to
409 module description (fs/super.c). This makes the initialisation
410 message fit on one line and fits in better with rest of kernel.
411 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
412 attribute inodes, and both for files and directories.
413 - Implement fake attribute inodes allowing all attribute i/o to go via
414 the page cache and to use all the normal vfs/mm functionality:
415 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
417 - Add needed cleanup code to ntfs_clear_big_inode().
418 - Merge address space operations for files and directories (aops.c),
419 now just have ntfs_aops:
421 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
422 ntfs_attr_read_block() -> ntfs_read_block(),
423 ntfs_file_read_page() -> ntfs_readpage().
424 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
425 attribute inodes, and both for files and directories.
426 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
428 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
430 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
431 the locking out of super.c::get_nr_free_mft_records() and taking and
432 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
433 - Bring attribute run list merging code (fs/ntfs/attrib.c) in sync with
434 current userspace ntfs library code. This means that if a merge
435 fails the original run lists are always left unmodified instead of
436 being silently corrupted.
439 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
441 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
442 from ntfs_volume structure.
443 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
444 from ntfs_volume structure.
445 - Use iget5_locked() and friends instead of conventional iget(). Wrap
446 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
447 to use ntfs_iget(). Leave only one iget() call at mount time so we
448 don't need an ntfs_iget_mount().
449 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
452 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
454 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
455 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
456 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
457 to determine whether to apply mst fixups or not.
458 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
459 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
460 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
461 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
462 the VFS readpage function prototype to the ntfs_attr_read_block()
465 2.0.11 - Initial preparations for fake inode based attribute i/o.
467 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
468 do some macro magic (adapted from include/linux/buffer_head.h) to
469 expand all the helper functions NInoFoo(), NInoSetFoo(), and
471 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
472 - Add new fields to ntfs_inode structure to allow use of fake inodes
473 for attribute i/o: type, name, name_len. Also add new state bits:
474 NI_Attr, which, if set, indicates the inode is a fake inode, and
475 NI_MstProtected, which, if set, indicates the attribute uses multi
476 sector transfer protection, i.e. fixups need to be applied after
477 reads and before/after writes.
478 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
479 ntfs_{new,clear,destroy}_extent_inode() and update callers.
480 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
481 instead of ntfs_destroy_extent_inode().
482 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
483 - Make all operations on ntfs inode state bits use the NIno* functions.
484 - Set up the new ntfs inode fields and state bits in
485 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
486 allocated memory to __ntfs_clear_inode().
487 - Cleanup ntfs_inode structure a bit for better ordering of elements
488 w.r.t. their size to allow better packing of the structure in memory.
490 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
492 - Add check at mount time to verify that the number of inodes on the
493 volume does not exceed 2^32 - 1, which is the maximum allowed for
494 NTFS according to Microsoft.
495 - Change mft_no member of ntfs_inode structure to be unsigned long.
496 Update all users. This makes ntfs_inode->mft_no just a copy of struct
497 inode->i_ino. But we can't just always use struct inode->i_ino and
498 remove mft_no because extent inodes do not have an attached struct
501 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
503 - Change decompression engine to use a single buffer protected by a
504 spin lock instead of per-CPU buffers. (Rusty Russell)
505 - Do not update cb_pos when handling a partial final page during
506 decompression of a sparse compression block, as the value is later
507 reset without being read/used. (Rusty Russell)
508 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
510 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
511 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
512 it also makes everything safer so it is a good thing.
513 - Miscellaneous minor cleanups to comments.
515 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
517 Big thanks go to Al Viro and other inhabitants of #kernel for investing
518 their time to discuss the case sensitivity and dcache aliasing issues.
520 - Remove unused source file fs/ntfs/attraops.c.
521 - Remove show_inodes mount option(s), thus dropping support for
522 displaying of short file names.
523 - Remove deprecated mount option posix.
524 - Restore show_sys_files mount option.
525 - Add new mount option case_sensitive, to determine if the driver
526 treats file names as case sensitive or not. If case sensitive, create
527 file names in the POSIX namespace. Otherwise create file names in the
528 LONG/WIN32 namespace. Note, files remain accessible via their short
529 file name, if it exists.
530 - Remove really dumb logic bug in boot sector recovery code.
531 - Fix dcache aliasing issues wrt short/long file names via changes
532 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
533 fs/ntfs/namei.c::ntfs_lookup():
534 - Add additional argument to ntfs_lookup_inode_by_name() in which we
535 return information about the matching file name if the case is not
536 matching or the match is a short file name. See comments above the
537 function definition for details.
538 - Change ntfs_lookup() to only create dcache entries for the correctly
539 cased file name and only for the WIN32 namespace counterpart of DOS
540 namespace file names. This ensures we have only one dentry per
541 directory and also removes all dcache aliasing issues between short
542 and long file names once we add write support. See comments above
543 function for details.
544 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
546 2.0.7 - Minor cleanups and updates for changes in core kernel code.
548 - Remove much of the NULL struct element initializers.
549 - Various updates to make compatible with recent kernels.
550 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
551 in fs/ntfs/ntfs.h instead.
552 - Remove no longer needed KERNEL_VERSION checks. We are now in the
553 kernel proper so they are no longer needed.
555 2.0.6 - Major bugfix to make compatible with other kernel changes.
557 - Initialize the mftbmp address space properly now that there are more
558 fields in the struct address_space. This was leading to hangs and
559 oopses on umount since 2.5.12 because of changes to other parts of
560 the kernel. We probably want a kernel generic init_address_space()
562 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
563 only caller of ->readdir() is vfs_readdir() which holds i_sem during
564 the call, and i_sem is sufficient protection against changes in the
565 directory inode (including ->i_size).
566 - Use generic_file_llseek() for directories (as opposed to
567 default_llseek()) as this downs i_sem instead of the BKL which is
568 what we now need for exclusion against ->f_pos changes considering we
569 no longer take the BKL in ntfs_readdir().
571 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
573 - No need to set old blocksize in super.c::ntfs_fill_super() as the
574 VFS does so via invocation of deactivate_super() calling
575 fs->fill_super() calling block_kill_super() which does it.
576 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
577 -> Do we really need it? I don't think so as we have exclusion on
578 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
579 move the ->f_pos accesses under the mrec_lock though. Check this...
580 - Fix really, really, really stupid buffer overflow in extent inode
581 handling in mft.c::map_extent_mft_record().
583 2.0.4 - Cleanups and updates for kernel 2.5.11.
585 - Add documentation on how to use the MD driver to be able to use NTFS
586 stripe and volume sets in Linux and generally cleanup documentation
588 Remove all uses of kdev_t in favour of struct block_device *:
589 - Change compress.c::ntfs_file_read_compressed_block() to use
590 sb_getblk() instead of getblk().
591 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
592 of get_hardsect_size().
593 - No need to get old blocksize in super.c::ntfs_fill_super() as
594 fs/super.c::get_sb_bdev() already does this.
595 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
597 2.0.3 - Small bug fixes, cleanups, and performance improvements.
599 - Remove some dead code from mft.c.
600 - Optimize readpage and read_block functions throughout aops.c so that
601 only initialized blocks are read. Non-initialized ones have their
602 buffer head mapped, zeroed, and set up to date, without scheduling
603 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
604 Thanks go to Andrew Morton for spotting the below:
605 - Fix buglet in allocate_compression_buffers() error code path.
606 - Call flush_dcache_page() after modifying page cache page contents in
607 ntfs_file_readpage().
608 - Check for existence of page buffers throughout aops.c before calling
609 create_empty_buffers(). This happens when an I/O error occurs and the
610 read is retried. (It also happens once writing is implemented so that
611 needed doing anyway but I had left it for later...)
612 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
613 readpage and read_block functions. Reasoning same as above (i.e. I/O
614 error retries and future write code paths.)
616 2.0.2 - Minor updates and cleanups.
618 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
619 and cleanup the code a bit, removing the unused size parameter.
620 - Change default fmask to 0177 and update documentation.
621 - Change attrib.c::get_attr_search_ctx() to return the search context
622 directly instead of taking the address of a pointer. A return value
623 of NULL means the allocation failed. Updated all callers
625 - Update to 2.5.9 kernel (preserving backwards compatibility) by
626 replacing all occurences of page->buffers with page_buffers(page).
627 - Fix minor bugs in run list merging, also minor cleanup.
628 - Updates to bootsector layout and mft mirror contents descriptions.
629 - Small bug fix in error detection in unistr.c and some cleanups.
630 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
633 2.0.1 - Minor updates.
635 - Make default umask correspond to documentation.
636 - Improve documentation.
637 - Set default mode to include execute bit. The {u,f,d}mask can be used
638 to take it away if desired. This allows binaries to be executed from
639 a mounted ntfs partition.
641 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
643 - Add kill_super, just keeping up with the vfs changes in the kernel.
644 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
645 from the CVS import into BitKeeper.
646 - Begin to implement proper handling of allocated_size vs
647 initialized_size vs data_size (i.e. i_size). Done are
648 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
649 and attrib.c::load_attribute_list().
650 - Lock the run list in attrib.c::load_attribute_list() while using it.
651 - Fix memory leak in ntfs_file_read_compressed_block() and generally
652 clean up compress.c a little, removing some uncommented/unused debug
654 - Tidy up dir.c a little bit.
655 - Don't bother getting the run list in inode.c::ntfs_read_inode().
656 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
657 creating aops.c::ntfs_mst_readpage(), improving the handling of
658 holes and overflow in the process and implementing the correct
659 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
660 I am aiming for correctness at the moment. Modularisation can come
662 - Rename aops.c::end_buffer_read_index_async() to
663 end_buffer_read_mst_async() and optimize the overflow checking and
665 - Use the host of the mftbmp address space mapping to hold the ntfs
666 volume. This is needed so the async i/o completion handler can
667 retrieve a pointer to the volume. Hopefully this will not cause
668 problems elsewhere in the kernel... Otherwise will need to use a
670 - Complete implementation of proper handling of allocated_size vs
671 initialized_size vs data_size (i.e. i_size) in whole driver.
672 Basically aops.c is now completely rewritten.
673 - Change NTFS driver name to just NTFS and set version number to 2.0.0
674 to make a clear distinction from the old driver which is still on
677 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
679 - Replace bdevname(sb->s_dev) with sb->s_id.
680 - Remove now superfluous new-line characters in all callers of
682 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
683 directories. Without this the "find" utility gets very upset which is
684 fair enough as Linux/Unix do not support directory hard links.
685 - Further run list merging work. (Richard Russon)
686 - Backwards compatibility for gcc-2.95. (Richard Russon)
687 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
688 - Convert to new file system declaration using ->ntfs_get_sb() and
689 replacing ntfs_read_super() with ntfs_fill_super().
690 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
691 overflow on 32-bit architectures.
692 - Cleanup upcase loading code to use ntfs_(un)map_page().
693 - Disable/reenable preemtion in critical sections of compession engine.
694 - Replace device size determination in ntfs_fill_super() with
695 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
696 function super.c::get_nr_blocks().
697 - Implement a mount time option (show_inodes) allowing choice of which
698 types of inode names readdir() returns and modify ntfs_filldir()
699 accordingly. There are several parameters to show_inodes:
701 win32: long file names (including POSIX file names) [DEFAULT]
703 dos: short file names only (excluding POSIX file names)
705 posix: same as both win32 and dos
707 Note that the options are additive, i.e. specifying:
708 -o show_inodes=system,show_inodes=win32,show_inodes=dos
709 is the same as specifying:
711 Note that the "posix" and "all" options will show all directory
712 names, BUT the link count on each directory inode entry is set to 1,
713 due to Linux not supporting directory hard links. This may well
714 confuse some userspace applications, since the directory names will
715 have the same inode numbers. Thus it is NOT advisable to use the
716 "posix" or "all" options. We provide them only for completeness sake.
717 - Add copies of allocated_size, initialized_size, and compressed_size to
718 the ntfs inode structure and set them up in
719 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
720 for files and the index allocation attribute for directories.
721 - Add copies of allocated_size and initialized_size to ntfs inode for
722 $BITMAP attribute of large directories and set them up in
723 inode.c::ntfs_read_inode().
724 - Add copies of allocated_size and initialized_size to ntfs volume for
725 $BITMAP attribute of $MFT and set them up in
726 super.c::load_system_files().
727 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
728 posix, and utf8) and tell user what the new options to use are. Note
729 we still do support them but they will be removed with kernel 2.7.x.
730 - Change all occurences of integer long long printf formatting to hex
731 as printk() will not support long long integer format if/when the
732 div64 patch goes into the kernel.
733 - Make slab caches have stable names and change the names to what they
734 were intended to be. These changes are required/made possible by the
735 new slab cache name handling which removes the length limitation by
736 requiring the caller of kmem_cache_create() to supply a stable name
737 which is then referenced but not copied.
738 - Rename run_list structure to run_list_element and create a new
739 run_list structure containing a pointer to a run_list_element
740 structure and a read/write semaphore. Adapt all users of run lists
741 to new scheme and take and release the lock as needed. This fixes a
742 nasty race as the run_list changes even when inodes are locked for
743 reading and even when the inode isn't locked at all, so we really
744 needed the serialization. We use a semaphore rather than a spinlock
745 as memory allocations can sleep and doing everything GFP_ATOMIC
747 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
748 This can never happen due to the nature of lookup_attr() and how we
749 support attribute lists. If it did happen it would imply the inode
751 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
753 - Update to 2.5.6-pre2 changes in struct address_space.
754 - Use parent_ino() when accessing d_parent inode number in dir.c.
755 - Import Sourceforge CVS repository into BitKeeper repository:
756 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
757 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
758 Documentation/filesystems/ntfs.txt for NTFS TNG.
759 - Create kernel configuration option controlling whether debugging
761 - Add the required export of end_buffer_io_sync() from the patches
762 directory to the kernel code.
763 - Update inode.c::ntfs_show_options() with show_inodes mount option.
764 - Update errors mount option.
766 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
768 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
769 bug in mapping of extent inodes. Update all the comments to fit all
770 the recent code changes.
771 - Modify vcn_to_lcn() to cope with entirely unmapped run lists.
772 - Cleanups in compress.c, mostly comments and folding help.
773 - Implement attrib.c::map_run_list() as a generic helper.
774 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
775 thus making code shorter and enabling attribute list support.
776 - Cleanup incorrect use of [su]64 with %L printf format specifier in
777 all source files. Type casts to [unsigned] long long added to correct
778 the mismatches (important for architectures which have long long not
780 - Merge async io completion handlers for directory indexes and $MFT
781 data into one by setting the index_block_size{_bits} of the ntfs
782 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
783 - Cleanup aops.c, update comments.
784 - Make ntfs_file_get_block() use map_run_list() so all files now
785 support attribute lists.
786 - Make ntfs_dir_readpage() almost verbatim copy of
787 block_read_full_page() by using ntfs_file_get_block() with only real
788 difference being the use of our own async io completion handler
789 rather than the default one, thus reducing the amount of code and
790 automatically enabling attribute list support for directory indices.
791 - Fix bug in load_attribute_list() - forgot to call brelse in error
793 - Change parameters to find_attr() and lookup_attr(). We no longer
794 pass in the upcase table and its length. These can be gotten from
795 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
796 - Cleanups in attrib.c.
797 - Implement merging of run lists, attrib.c::merge_run_lists() and its
798 helpers. (Richard Russon)
799 - Attribute lists part 2, attribute extents and multi part run lists:
800 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
801 further run list parts via attrib.c::map_run_list().
802 - Tiny endianness bug fix in decompress_mapping_pairs().
804 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
806 - Enable encrypted directories. (Their index root is marked encrypted
807 to indicate that new files in that directory should be created
809 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
810 - Enable $Extend system directory. Most (if not all) extended system
811 files do not have unnamed data attributes so ntfs_read_inode() had to
812 special case them but that is ok, as the special casing recovery
813 happens inside an error code path so there is zero slow down in the
814 normal fast path. The special casing is done by introducing a new
815 function inode.c::ntfs_is_extended_system_file() which checks if any
816 of the hard links in the inode point to $Extend as being their parent
817 directory and if they do we assume this is an extended system file.
818 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
819 when compiled with -DDEBUG. Default is debug messages to be disabled.
820 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
821 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
822 interface is enabled). Inspired by old ntfs driver.
823 - Add debug_msgs insmod/kernel boot parameter to set whether debug
824 messages are {dis,en}abled. This is useful to enable debug messages
825 during ntfs initialization and is the only way to activate debugging
826 when the sysctl interface is not enabled.
827 - Cleanup debug output in various places.
828 - Remove all dollar signs ($) from the source (except comments) to
829 enable compilation on architectures whose gcc compiler does not
830 support dollar signs in the names of variables/constants. Attribute
831 types now start with AT_ instead of $ and $I30 is now just I30.
832 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
833 - Load complete run list for $MFT/$BITMAP during mount and cleanup
834 access functions. This means we now cope with $MFT/$BITMAP being
835 spread accross several mft records.
836 - Disable modification of mft_zone_multiplier on remount. We can always
837 reenable this later on if we really want to, but we will need to make
838 sure we readjust the mft_zone size / layout accordingly.
840 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
842 - Use sb_set_blocksize() instead of set_blocksize() and verify the
844 - Use sb_bread() instead of bread() throughout.
845 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
846 of a directory index block vcn. Apply resulting simplifications in
848 - Fix a small bug somewhere (but forgot what it was).
849 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
850 on the printf-format parameter list and fix bugs reported by gcc
851 as a result. (Richard Russon)
852 - Move inode allocation strategy to Al's new stuff but maintain the
853 divorce of ntfs_inode from struct inode. To achieve this we have two
854 separate slab caches, one for big ntfs inodes containing a struct
855 inode and pure ntfs inodes and at the same time fix some faulty
856 error code paths in ntfs_read_inode().
857 - Show mount options in proc (inode.c::ntfs_show_options()).
859 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
861 - Modified (un)map_mft_record functions to be common for read and write
862 case. To specify which is which, added extra parameter at front of
863 parameter list. Pass either READ or WRITE to this, each has the
865 - General cleanups to allow for easier folding in vi.
866 - attrib.c::decompress_mapping_pairs() now accepts the old run list
867 argument, and invokes attrib.c::merge_run_lists() to merge the old
868 and the new run lists.
869 - Removed attrib.c::find_first_attr().
870 - Implemented loading of attribute list and complete run list for $MFT.
871 This means we now cope with $MFT being spread across several mft
873 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
874 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
875 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
876 cache. There are too little ntfs_volume structures at any one time
877 to justify a private slab cache.
878 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
879 Use KM_BIO_IRQ on advice from IRC/kernel...
880 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
881 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
882 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
883 removed the now superfluous exports from the kernel core patch.
884 - Fix a bug where kfree() was used insted of ntfs_free().
885 - Change map_mft_record() to take ntfs_inode as argument instead of
886 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
887 - Add pointer to ntfs_volume to ntfs_inode.
888 - Add mft record number and sequence number to ntfs_inode. Stop using
889 i_ino and i_generation for in-driver purposes.
890 - Implement attrib.c::merge_run_lists(). (Richard Russon)
891 - Remove use of proper inodes by extent inodes. Move i_ino and
892 i_generation to ntfs_inode to do this. Apply simplifications that
893 result and remove iget_no_wait(), etc.
894 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
895 - Add reference counting in ntfs_inode for the ntfs inode itself and
896 for the mapped mft record.
897 - Extend mft record mapping so we can (un)map extent mft records (new
898 functions (un)map_extent_mft_record), and so mappings are reference
899 counted and don't have to happen twice if already mapped - just ref
901 - Add -o iocharset as alias to -o nls for backwards compatibility.
902 - The latest core patch is now tiny. In fact just a single additional
903 export is necessary over the base kernel.
905 tng-0.0.3 - Cleanups, enhancements, bug fixes.
907 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
908 and setup the run list appropriately using knowledge provided by the
909 sizes in the base attribute record.
910 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
912 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
913 page or use vmalloc depending on the amount of memory requested.
914 - Cleanup error output. The __FUNCTION__ "(): " is now added
915 automatically. Introduced a new header file debug.h to support this
916 and also moved ntfs_debug() function into it.
917 - Make reading of compressed files more intelligent and especially get
918 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
919 buffers (allocated at first mount with cluster size <= 4kiB and
920 deallocated on last umount with cluster size <= 4kiB), and
921 asynchronous io for the compressed data using a list of buffer heads.
922 Er, we use synchronous io as async io only works on whole pages
923 covered by buffers and not on individual buffer heads...
924 - Bug fix for reading compressed files with sparse compression blocks.
926 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
928 - Fixed handling of directories when cluster size exceeds index block
930 - Hide DOS only name space directory entries from readdir() but allow
931 them in lookup(). This should fix the problem that Linux doesn't
932 support directory hard links, while still allowing access to entries
933 via their short file name. This also has the benefit of mimicking
934 what Windows users are used to, so it is the ideal solution.
935 - Implemented sync_page everywhere so no more hangs in D state when
937 - Stop using bforget() in favour of brelse().
938 - Stop locking buffers unnecessarily.
939 - Implemented compressed files (inode->mapping contains uncompressed
940 data, raw compressed data is currently bread() into a vmalloc()ed
942 - Enable compressed directories. (Their index root is marked compressed
943 to indicate that new files in that directory should be created
945 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
946 functions. (Thanks to Will Dyson for pointing this out.)
947 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
948 ntfs_sb_info) out of the common inode and super_block structures and
949 started using the generic_ip and generic_sbp pointers instead. This
950 makes ntfs entirely private with respect to the kernel tree.
951 - Detect compiler version and abort with error message if gcc less than
953 - Fix bug in name comparison function in unistr.c.
954 - Implement attribute lists part 1, the infrastructure: search contexts
955 and operations, find_external_attr(), lookup_attr()) and make the
956 code use the infrastructure.
957 - Fix stupid buffer overflow bug that became apparent on larger run
958 list containing attributes.
959 - Fix bugs in readdir() that became apparent on larger directories.
961 The driver is now really useful and survives the test
962 find . -type f -exec md5sum "{}" \;
963 without any error messages on a over 1GiB sized partition with >16k
964 files on it, including compressed files and directories and many files
965 and directories with attribute lists.
967 tng-0.0.1 - The first useful version.
969 - Added ntfs_lookup().
970 - Added default upcase generation and handling.
971 - Added compile options to be shown on module init.
972 - Many bug fixes that were "hidden" before.
973 - Update to latest kernel.
974 - Added ntfs_readdir().
975 - Added file operations for mmap(), read(), open() and llseek(). We just
976 use the generic ones. The whole point of going through implementing
977 readpage() methods and where possible get_block() call backs is that
978 this allows us to make use of the generic high level methods provided
981 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
982 though and it doesn't implement accesssing compressed files yet. Also,
983 accessing files with attribute list attributes is not implemented yet
984 either. But for small or simple file systems it should work and allow
985 you to list directories, use stat on directory entries and the file
986 system, open, read, mmap and llseek around in files. A big mile stone
989 tng-0.0.0 - Initial version tag.
991 Initial driver implementation. The driver can mount and umount simple
992 NTFS file systems (i.e. ones without attribute lists in the system
993 files). If the mount fails there might be problems in the error handling
994 code paths, so be warned. Otherwise it seems to be loading the system
995 files nicely and the mft record read mapping/unmapping seems to be
996 working nicely, too. Proof of inode metadata in the page cache and non-
997 resident file unnamed stream data in the page cache concepts is thus
git://git.onelab.eu / linux-2.6.git / blob