3 - Implement aops->set_page_dirty() in order to take control of buffer
4 dirtying. Not having it means if page_has_buffers(), all buffers
5 will be dirtied with the page. And if not they won't be. That is
6 fine for the moment but will break once we enable metadata updates.
7 - Implement sops->dirty_inode() to implement {a,m,c} time updates and
9 - Implement sops->write_inode().
10 - In between ntfs_prepare/commit_write, need exclusion between
11 simultaneous file extensions. Need perhaps an NInoResizeUnderway()
12 flag which we can set in ntfs_prepare_write() and clear again in
13 ntfs_commit_write(). Just have to be careful in readpage/writepage,
14 as well as in truncate, that we play nice... We might need to have
15 a data_size field in the ntfs_inode to store the real attribute
16 length. Also need to be careful with initialized_size extention in
17 ntfs_prepare_write. Basically, just be _very_ careful in this code...
18 OTOH, perhaps i_sem, which is held accross generic_file_write is
19 sufficient for synchronisation here. We then just need to make sure
20 ntfs_readpage/writepage/truncate interoperate properly with us.
22 2.1.8 - Handle $MFT mirror and $LogFile, improve time ihandling, and cleanups.
24 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
25 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
26 utc2ntfs() to work with struct timespec instead of time_t on the
27 Linux UTC time side thus preserving the full precision of the NTFS
28 time and only loosing up to 99 nano-seconds in the Linux UTC time.
29 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
31 - Remove unused ntfs_dirty_inode().
32 - Cleanup super operations declaration in fs/ntfs/super.c.
33 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
34 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
35 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
36 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
37 fs/ntfs/inode.h so they can be used elsewhere.
38 - Determine the mft mirror size as the number of mirrored mft records
39 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
40 - Load the mft mirror at mount time and compare the mft records stored
41 in it to the ones in the mft. Force a read-only mount if the two do
42 not match (fs/ntfs/super.c).
43 - Fix type casting related warnings on 64-bit architectures. Thanks
44 to Meelis Roos for reporting them.
45 - Move %L to %ll as %L is floating point and %ll is integer which is
47 - Read the journal ($LogFile) and determine if the volume has been
48 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
49 and fs/ntfs/logfile.c). This is a little bit of a crude check in
50 that we only look at the restart areas and not at the actual log
51 records so that there will be a very small number of cases where we
52 think that a volume is dirty when in fact it is clean. This should
53 only affect volumes that have not been shutdown cleanly and did not
54 have any pending, non-check-pointed i/o.
55 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
56 is requested, empty $LogFile by overwriting it with 0xff bytes to
57 ensure that Windows cannot cause data corruption by replaying a stale
58 journal after Linux has written to the volume.
60 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
62 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
63 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
64 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
65 default doesn't allow inode number 0 which is a valid inode on NTFS
66 and even if it did allow that it uses iget() instead of ntfs_iget()
67 which makes it useless for us.
68 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
69 default just returns -EACCES which is not very useful.
70 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
71 and set them up in the super block at mount time (super.c) this
72 allows mounted NTFS volumes to be exported via NFS.
73 - Add missing return -EOPNOTSUPP; in
74 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
75 - Enforce no atime and no dir atime updates at mount/remount time as
76 they are not implemented yet anyway.
77 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
78 after a NULL check. Thanks to Dave Jones for pointing this out.
80 2.1.6 - Fix minor bug in handling of compressed directories.
82 - Fix bug in handling of compressed directories. A compressed
83 directory is not really compressed so when we set the ->i_blocks
84 field of a compressed directory inode we were setting it from the
85 non-existing field ni->itype.compressed.size which gave random
86 results... For directories we now always use ni->allocated_size.
88 2.1.5 - Fix minor bug in attribute list attribute handling.
90 - Fix bug in attribute list handling. Actually it is not as much a bug
91 as too much protection in that we were not allowing attribute lists
92 which waste space on disk while Windows XP clearly allows it and in
93 fact creates such attribute lists so our driver was failing.
94 - Update NTFS documentation ready for 2.6 kernel release.
96 2.1.4 - Reduce compiler requirements.
98 - Remove all uses of unnamed structs and unions in the driver to make
99 old and newer gcc versions happy. Makes it a bit uglier IMO but at
100 least people will stop hassling me about it.
102 2.1.3 - Important bug fixes in corner cases.
104 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
105 clusters. (Philipp Thomas)
106 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
107 multiple of the block_size but not the cluster size. (Szabolcs
108 Szakacsits <szaka@sienet.hu>)
110 2.1.2 - Important bug fixes aleviating the hangs in statfs.
112 - Fix buggy free cluster and free inode determination logic.
114 2.1.1 - Minor updates.
116 - Add handling for initialized_size != data_size in compressed files.
117 - Reduce function local stack usage from 0x3d4 bytes to just noise in
118 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
119 - Remove compiler warnings for newer gcc.
120 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
121 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
122 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
123 kmap_atomic(KM_USER0).
125 2.1.0 - First steps towards write support: implement file overwrite.
127 - Add configuration option for developmental write support with an
128 appropriately scary configuration help text.
129 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
130 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
131 overwriting of existing files on ntfs. Note: Resident files are
132 only written into memory, and not written out to disk at present, so
133 avoid writing to files smaller than about 1kiB.
134 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
135 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
136 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
137 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
138 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
139 This enables write(2) based overwriting of existing files on ntfs.
140 Note: As with mmap(2) based overwriting, resident files are only
141 written into memory, and not written out to disk at present, so avoid
142 writing to files smaller than about 1kiB.
143 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
144 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
145 files with the purpose of intercepting and aborting all i_size
146 changes which we do not support yet. ntfs_truncate() actually only
147 emits a warning message but AFAICS our interception of i_size changes
148 elsewhere means ntfs_truncate() never gets called for i_size changes.
149 It is only called from generic_file_write() when we fail in
150 ntfs_prepare_{,nonresident_}write() in order to discard any
151 instantiated buffers beyond i_size. Thus i_size is not actually
152 changed so our warning message is enough. Unfortunately it is not
153 possible to easily determine if i_size is being changed or not hence
154 we just emit an appropriately worded error message.
156 2.0.25 - Small bug fixes and cleanups.
158 - Unlock the page in an out of memory error code path in
159 fs/ntfs/aops.c::ntfs_read_block().
160 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
161 just unlock the page and return. (This can happen due to ->writepage
162 clearing PageUptodate() during write out of MstProtected()
164 - Remove leaked write code again.
168 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
169 inside BUG_ON(). (Adam J. Richter)
170 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
171 calls for improved debugging. (Adam J. Richter)
172 - Add errors flag to the ntfs volume state, accessed via
173 NVol{,Set,Clear}Errors(vol).
174 - Do not allow read-write remounts of read-only volumes with errors.
175 - Clarify comment for ntfs file operation sendfile which was added by
176 Christoph Hellwig a while ago (just using generic_file_sendfile())
177 to say that ntfs ->sendfile is only used for the case where the
178 source data is on the ntfs partition and the destination is
179 somewhere else, i.e. nothing we need to concern ourselves with.
180 - Add generic_file_write() as our ntfs file write operation.
182 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
184 - Massive internal locking changes to mft record locking. Fixes lock
185 recursion and replaces the mrec_lock read/write semaphore with a
186 mutex. Also removes the now superfluous mft_count. This fixes several
187 race conditions and deadlocks, especially in the future write code.
188 - Fix ntfs over loopback for compressed files by adding an
189 optimization barrier. (gcc was screwing up otherwise ?)
190 - Miscellaneous cleanups all over the code and a fix or two in error
192 Thanks go to Christoph Hellwig for pointing out the following two:
193 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
194 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
196 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
198 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
199 at entry/exit respectively.
200 - Use C99 initializers for structures.
201 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
203 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
205 - Limit volume size at mount time to 2TiB on architectures where
206 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
207 This is the most we can do without overflowing the 32-bit limit of
208 the block device size imposed on us by sb_bread() and sb_getblk()
210 - Limit file/directory size at open() time to 16TiB on architectures
211 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
212 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
213 overflowing the page cache page index.
215 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
217 - Move the directory index bitmap to use an attribute inode instead of
218 having special fields for it inside the ntfs inode structure. This
219 means that the index bitmaps now use the page cache for i/o, too,
220 and also as a side effect we get support for non-resident index
222 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
223 fix a page leak that manifested itself in some cases.
224 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
225 index bitmap inode on the final iput().
227 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
229 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
230 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
231 - Drop the "file" from ntfs_file_read_compressed_block().
232 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
233 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
234 - Update ntfs_end_buffer_async_read() with the improved logic from
235 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
236 further logic improvements to better determine when we set PageError.
237 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
238 check for the buffers being uptodate first in line with the updated
239 fs/buffer.c::block_read_full_page(). This plugs a small race
242 2.0.18 - Fix race condition in reading of compressed files.
244 - There was a narrow window between checking a buffer head for being
245 uptodate and locking it in ntfs_file_read_compressed_block(). We now
246 lock the buffer and then check whether it is uptodate or not.
248 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
250 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
251 code and update callers, i.e. ntfs_iget(), to pass that error code
252 up instead of just using -EIO.
253 - Modifications to super.c to ensure that both mount and remount
254 cannot set any write related options when the driver is compiled
256 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
257 cache the current run list element. This should improve performance
258 when reading very large and/or very fragmented data.
260 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
262 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
263 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
264 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
265 - Convert $MFT/$BITMAP access to attribute inode API and remove all
266 remnants of the ugly mftbmp address space and operations hack. This
267 means we finally have only one readpage function as well as only one
268 async io completion handler. Yey! The mft bitmap is now just an
269 attribute inode and is accessed from vol->mftbmp_ino just as if it
270 were a normal file. Fake inodes rule. (-:
272 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
274 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
275 remounts to fail when the partition had an entry in /etc/fstab and
276 the entry specified the nls= option.
277 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
278 expand all the helper functions NVolFoo(), NVolSetFoo(), and
280 - Move copyright statement from driver initialisation message to
281 module description (fs/super.c). This makes the initialisation
282 message fit on one line and fits in better with rest of kernel.
283 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
284 attribute inodes, and both for files and directories.
285 - Implement fake attribute inodes allowing all attribute i/o to go via
286 the page cache and to use all the normal vfs/mm functionality:
287 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
289 - Add needed cleanup code to ntfs_clear_big_inode().
290 - Merge address space operations for files and directories (aops.c),
291 now just have ntfs_aops:
293 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
294 ntfs_attr_read_block() -> ntfs_read_block(),
295 ntfs_file_read_page() -> ntfs_readpage().
296 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
297 attribute inodes, and both for files and directories.
298 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
300 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
302 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
303 the locking out of super.c::get_nr_free_mft_records() and taking and
304 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
305 - Bring attribute run list merging code (fs/ntfs/attrib.c) in sync with
306 current userspace ntfs library code. This means that if a merge
307 fails the original run lists are always left unmodified instead of
308 being silently corrupted.
311 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
313 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
314 from ntfs_volume structure.
315 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
316 from ntfs_volume structure.
317 - Use iget5_locked() and friends instead of conventional iget(). Wrap
318 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
319 to use ntfs_iget(). Leave only one iget() call at mount time so we
320 don't need an ntfs_iget_mount().
321 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
324 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
326 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
327 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
328 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
329 to determine whether to apply mst fixups or not.
330 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
331 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
332 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
333 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
334 the VFS readpage function prototype to the ntfs_attr_read_block()
337 2.0.11 - Initial preparations for fake inode based attribute i/o.
339 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
340 do some macro magic (adapted from include/linux/buffer_head.h) to
341 expand all the helper functions NInoFoo(), NInoSetFoo(), and
343 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
344 - Add new fields to ntfs_inode structure to allow use of fake inodes
345 for attribute i/o: type, name, name_len. Also add new state bits:
346 NI_Attr, which, if set, indicates the inode is a fake inode, and
347 NI_MstProtected, which, if set, indicates the attribute uses multi
348 sector transfer protection, i.e. fixups need to be applied after
349 reads and before/after writes.
350 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
351 ntfs_{new,clear,destroy}_extent_inode() and update callers.
352 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
353 instead of ntfs_destroy_extent_inode().
354 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
355 - Make all operations on ntfs inode state bits use the NIno* functions.
356 - Set up the new ntfs inode fields and state bits in
357 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
358 allocated memory to __ntfs_clear_inode().
359 - Cleanup ntfs_inode structure a bit for better ordering of elements
360 w.r.t. their size to allow better packing of the structure in memory.
362 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
364 - Add check at mount time to verify that the number of inodes on the
365 volume does not exceed 2^32 - 1, which is the maximum allowed for
366 NTFS according to Microsoft.
367 - Change mft_no member of ntfs_inode structure to be unsigned long.
368 Update all users. This makes ntfs_inode->mft_no just a copy of struct
369 inode->i_ino. But we can't just always use struct inode->i_ino and
370 remove mft_no because extent inodes do not have an attached struct
373 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
375 - Change decompression engine to use a single buffer protected by a
376 spin lock instead of per-CPU buffers. (Rusty Russell)
377 - Do not update cb_pos when handling a partial final page during
378 decompression of a sparse compression block, as the value is later
379 reset without being read/used. (Rusty Russell)
380 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
382 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
383 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
384 it also makes everything safer so it is a good thing.
385 - Miscellaneous minor cleanups to comments.
387 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
389 Big thanks go to Al Viro and other inhabitants of #kernel for investing
390 their time to discuss the case sensitivity and dcache aliasing issues.
392 - Remove unused source file fs/ntfs/attraops.c.
393 - Remove show_inodes mount option(s), thus dropping support for
394 displaying of short file names.
395 - Remove deprecated mount option posix.
396 - Restore show_sys_files mount option.
397 - Add new mount option case_sensitive, to determine if the driver
398 treats file names as case sensitive or not. If case sensitive, create
399 file names in the POSIX namespace. Otherwise create file names in the
400 LONG/WIN32 namespace. Note, files remain accessible via their short
401 file name, if it exists.
402 - Remove really dumb logic bug in boot sector recovery code.
403 - Fix dcache aliasing issues wrt short/long file names via changes
404 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
405 fs/ntfs/namei.c::ntfs_lookup():
406 - Add additional argument to ntfs_lookup_inode_by_name() in which we
407 return information about the matching file name if the case is not
408 matching or the match is a short file name. See comments above the
409 function definition for details.
410 - Change ntfs_lookup() to only create dcache entries for the correctly
411 cased file name and only for the WIN32 namespace counterpart of DOS
412 namespace file names. This ensures we have only one dentry per
413 directory and also removes all dcache aliasing issues between short
414 and long file names once we add write support. See comments above
415 function for details.
416 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
418 2.0.7 - Minor cleanups and updates for changes in core kernel code.
420 - Remove much of the NULL struct element initializers.
421 - Various updates to make compatible with recent kernels.
422 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
423 in fs/ntfs/ntfs.h instead.
424 - Remove no longer needed KERNEL_VERSION checks. We are now in the
425 kernel proper so they are no longer needed.
427 2.0.6 - Major bugfix to make compatible with other kernel changes.
429 - Initialize the mftbmp address space properly now that there are more
430 fields in the struct address_space. This was leading to hangs and
431 oopses on umount since 2.5.12 because of changes to other parts of
432 the kernel. We probably want a kernel generic init_address_space()
434 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
435 only caller of ->readdir() is vfs_readdir() which holds i_sem during
436 the call, and i_sem is sufficient protection against changes in the
437 directory inode (including ->i_size).
438 - Use generic_file_llseek() for directories (as opposed to
439 default_llseek()) as this downs i_sem instead of the BKL which is
440 what we now need for exclusion against ->f_pos changes considering we
441 no longer take the BKL in ntfs_readdir().
443 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
445 - No need to set old blocksize in super.c::ntfs_fill_super() as the
446 VFS does so via invocation of deactivate_super() calling
447 fs->fill_super() calling block_kill_super() which does it.
448 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
449 -> Do we really need it? I don't think so as we have exclusion on
450 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
451 move the ->f_pos accesses under the mrec_lock though. Check this...
452 - Fix really, really, really stupid buffer overflow in extent inode
453 handling in mft.c::map_extent_mft_record().
455 2.0.4 - Cleanups and updates for kernel 2.5.11.
457 - Add documentation on how to use the MD driver to be able to use NTFS
458 stripe and volume sets in Linux and generally cleanup documentation
460 Remove all uses of kdev_t in favour of struct block_device *:
461 - Change compress.c::ntfs_file_read_compressed_block() to use
462 sb_getblk() instead of getblk().
463 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
464 of get_hardsect_size().
465 - No need to get old blocksize in super.c::ntfs_fill_super() as
466 fs/super.c::get_sb_bdev() already does this.
467 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
469 2.0.3 - Small bug fixes, cleanups, and performance improvements.
471 - Remove some dead code from mft.c.
472 - Optimize readpage and read_block functions throughout aops.c so that
473 only initialized blocks are read. Non-initialized ones have their
474 buffer head mapped, zeroed, and set up to date, without scheduling
475 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
476 Thanks go to Andrew Morton for spotting the below:
477 - Fix buglet in allocate_compression_buffers() error code path.
478 - Call flush_dcache_page() after modifying page cache page contents in
479 ntfs_file_readpage().
480 - Check for existence of page buffers throughout aops.c before calling
481 create_empty_buffers(). This happens when an I/O error occurs and the
482 read is retried. (It also happens once writing is implemented so that
483 needed doing anyway but I had left it for later...)
484 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
485 readpage and read_block functions. Reasoning same as above (i.e. I/O
486 error retries and future write code paths.)
488 2.0.2 - Minor updates and cleanups.
490 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
491 and cleanup the code a bit, removing the unused size parameter.
492 - Change default fmask to 0177 and update documentation.
493 - Change attrib.c::get_attr_search_ctx() to return the search context
494 directly instead of taking the address of a pointer. A return value
495 of NULL means the allocation failed. Updated all callers
497 - Update to 2.5.9 kernel (preserving backwards compatibility) by
498 replacing all occurences of page->buffers with page_buffers(page).
499 - Fix minor bugs in run list merging, also minor cleanup.
500 - Updates to bootsector layout and mft mirror contents descriptions.
501 - Small bug fix in error detection in unistr.c and some cleanups.
502 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
505 2.0.1 - Minor updates.
507 - Make default umask correspond to documentation.
508 - Improve documentation.
509 - Set default mode to include execute bit. The {u,f,d}mask can be used
510 to take it away if desired. This allows binaries to be executed from
511 a mounted ntfs partition.
513 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
515 - Add kill_super, just keeping up with the vfs changes in the kernel.
516 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
517 from the CVS import into BitKeeper.
518 - Begin to implement proper handling of allocated_size vs
519 initialized_size vs data_size (i.e. i_size). Done are
520 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
521 and attrib.c::load_attribute_list().
522 - Lock the run list in attrib.c::load_attribute_list() while using it.
523 - Fix memory leak in ntfs_file_read_compressed_block() and generally
524 clean up compress.c a little, removing some uncommented/unused debug
526 - Tidy up dir.c a little bit.
527 - Don't bother getting the run list in inode.c::ntfs_read_inode().
528 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
529 creating aops.c::ntfs_mst_readpage(), improving the handling of
530 holes and overflow in the process and implementing the correct
531 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
532 I am aiming for correctness at the moment. Modularisation can come
534 - Rename aops.c::end_buffer_read_index_async() to
535 end_buffer_read_mst_async() and optimize the overflow checking and
537 - Use the host of the mftbmp address space mapping to hold the ntfs
538 volume. This is needed so the async i/o completion handler can
539 retrieve a pointer to the volume. Hopefully this will not cause
540 problems elsewhere in the kernel... Otherwise will need to use a
542 - Complete implementation of proper handling of allocated_size vs
543 initialized_size vs data_size (i.e. i_size) in whole driver.
544 Basically aops.c is now completely rewritten.
545 - Change NTFS driver name to just NTFS and set version number to 2.0.0
546 to make a clear distinction from the old driver which is still on
549 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
551 - Replace bdevname(sb->s_dev) with sb->s_id.
552 - Remove now superfluous new-line characters in all callers of
554 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
555 directories. Without this the "find" utility gets very upset which is
556 fair enough as Linux/Unix do not support directory hard links.
557 - Further run list merging work. (Richard Russon)
558 - Backwards compatibility for gcc-2.95. (Richard Russon)
559 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
560 - Convert to new file system declaration using ->ntfs_get_sb() and
561 replacing ntfs_read_super() with ntfs_fill_super().
562 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
563 overflow on 32-bit architectures.
564 - Cleanup upcase loading code to use ntfs_(un)map_page().
565 - Disable/reenable preemtion in critical sections of compession engine.
566 - Replace device size determination in ntfs_fill_super() with
567 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
568 function super.c::get_nr_blocks().
569 - Implement a mount time option (show_inodes) allowing choice of which
570 types of inode names readdir() returns and modify ntfs_filldir()
571 accordingly. There are several parameters to show_inodes:
573 win32: long file names (including POSIX file names) [DEFAULT]
575 dos: short file names only (excluding POSIX file names)
577 posix: same as both win32 and dos
579 Note that the options are additive, i.e. specifying:
580 -o show_inodes=system,show_inodes=win32,show_inodes=dos
581 is the same as specifying:
583 Note that the "posix" and "all" options will show all directory
584 names, BUT the link count on each directory inode entry is set to 1,
585 due to Linux not supporting directory hard links. This may well
586 confuse some userspace applications, since the directory names will
587 have the same inode numbers. Thus it is NOT advisable to use the
588 "posix" or "all" options. We provide them only for completeness sake.
589 - Add copies of allocated_size, initialized_size, and compressed_size to
590 the ntfs inode structure and set them up in
591 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
592 for files and the index allocation attribute for directories.
593 - Add copies of allocated_size and initialized_size to ntfs inode for
594 $BITMAP attribute of large directories and set them up in
595 inode.c::ntfs_read_inode().
596 - Add copies of allocated_size and initialized_size to ntfs volume for
597 $BITMAP attribute of $MFT and set them up in
598 super.c::load_system_files().
599 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
600 posix, and utf8) and tell user what the new options to use are. Note
601 we still do support them but they will be removed with kernel 2.7.x.
602 - Change all occurences of integer long long printf formatting to hex
603 as printk() will not support long long integer format if/when the
604 div64 patch goes into the kernel.
605 - Make slab caches have stable names and change the names to what they
606 were intended to be. These changes are required/made possible by the
607 new slab cache name handling which removes the length limitation by
608 requiring the caller of kmem_cache_create() to supply a stable name
609 which is then referenced but not copied.
610 - Rename run_list structure to run_list_element and create a new
611 run_list structure containing a pointer to a run_list_element
612 structure and a read/write semaphore. Adapt all users of run lists
613 to new scheme and take and release the lock as needed. This fixes a
614 nasty race as the run_list changes even when inodes are locked for
615 reading and even when the inode isn't locked at all, so we really
616 needed the serialization. We use a semaphore rather than a spinlock
617 as memory allocations can sleep and doing everything GFP_ATOMIC
619 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
620 This can never happen due to the nature of lookup_attr() and how we
621 support attribute lists. If it did happen it would imply the inode
623 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
625 - Update to 2.5.6-pre2 changes in struct address_space.
626 - Use parent_ino() when accessing d_parent inode number in dir.c.
627 - Import Sourceforge CVS repository into BitKeeper repository:
628 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
629 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
630 Documentation/filesystems/ntfs.txt for NTFS TNG.
631 - Create kernel configuration option controlling whether debugging
633 - Add the required export of end_buffer_io_sync() from the patches
634 directory to the kernel code.
635 - Update inode.c::ntfs_show_options() with show_inodes mount option.
636 - Update errors mount option.
638 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
640 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
641 bug in mapping of extent inodes. Update all the comments to fit all
642 the recent code changes.
643 - Modify vcn_to_lcn() to cope with entirely unmapped run lists.
644 - Cleanups in compress.c, mostly comments and folding help.
645 - Implement attrib.c::map_run_list() as a generic helper.
646 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
647 thus making code shorter and enabling attribute list support.
648 - Cleanup incorrect use of [su]64 with %L printf format specifier in
649 all source files. Type casts to [unsigned] long long added to correct
650 the mismatches (important for architectures which have long long not
652 - Merge async io completion handlers for directory indexes and $MFT
653 data into one by setting the index_block_size{_bits} of the ntfs
654 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
655 - Cleanup aops.c, update comments.
656 - Make ntfs_file_get_block() use map_run_list() so all files now
657 support attribute lists.
658 - Make ntfs_dir_readpage() almost verbatim copy of
659 block_read_full_page() by using ntfs_file_get_block() with only real
660 difference being the use of our own async io completion handler
661 rather than the default one, thus reducing the amount of code and
662 automatically enabling attribute list support for directory indices.
663 - Fix bug in load_attribute_list() - forgot to call brelse in error
665 - Change parameters to find_attr() and lookup_attr(). We no longer
666 pass in the upcase table and its length. These can be gotten from
667 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
668 - Cleanups in attrib.c.
669 - Implement merging of run lists, attrib.c::merge_run_lists() and its
670 helpers. (Richard Russon)
671 - Attribute lists part 2, attribute extents and multi part run lists:
672 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
673 further run list parts via attrib.c::map_run_list().
674 - Tiny endianness bug fix in decompress_mapping_pairs().
676 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
678 - Enable encrypted directories. (Their index root is marked encrypted
679 to indicate that new files in that directory should be created
681 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
682 - Enable $Extend system directory. Most (if not all) extended system
683 files do not have unnamed data attributes so ntfs_read_inode() had to
684 special case them but that is ok, as the special casing recovery
685 happens inside an error code path so there is zero slow down in the
686 normal fast path. The special casing is done by introducing a new
687 function inode.c::ntfs_is_extended_system_file() which checks if any
688 of the hard links in the inode point to $Extend as being their parent
689 directory and if they do we assume this is an extended system file.
690 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
691 when compiled with -DDEBUG. Default is debug messages to be disabled.
692 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
693 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
694 interface is enabled). Inspired by old ntfs driver.
695 - Add debug_msgs insmod/kernel boot parameter to set whether debug
696 messages are {dis,en}abled. This is useful to enable debug messages
697 during ntfs initialization and is the only way to activate debugging
698 when the sysctl interface is not enabled.
699 - Cleanup debug output in various places.
700 - Remove all dollar signs ($) from the source (except comments) to
701 enable compilation on architectures whose gcc compiler does not
702 support dollar signs in the names of variables/constants. Attribute
703 types now start with AT_ instead of $ and $I30 is now just I30.
704 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
705 - Load complete run list for $MFT/$BITMAP during mount and cleanup
706 access functions. This means we now cope with $MFT/$BITMAP being
707 spread accross several mft records.
708 - Disable modification of mft_zone_multiplier on remount. We can always
709 reenable this later on if we really want to, but we will need to make
710 sure we readjust the mft_zone size / layout accordingly.
712 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
714 - Use sb_set_blocksize() instead of set_blocksize() and verify the
716 - Use sb_bread() instead of bread() throughout.
717 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
718 of a directory index block vcn. Apply resulting simplifications in
720 - Fix a small bug somewhere (but forgot what it was).
721 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
722 on the printf-format parameter list and fix bugs reported by gcc
723 as a result. (Richard Russon)
724 - Move inode allocation strategy to Al's new stuff but maintain the
725 divorce of ntfs_inode from struct inode. To achieve this we have two
726 separate slab caches, one for big ntfs inodes containing a struct
727 inode and pure ntfs inodes and at the same time fix some faulty
728 error code paths in ntfs_read_inode().
729 - Show mount options in proc (inode.c::ntfs_show_options()).
731 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
733 - Modified (un)map_mft_record functions to be common for read and write
734 case. To specify which is which, added extra parameter at front of
735 parameter list. Pass either READ or WRITE to this, each has the
737 - General cleanups to allow for easier folding in vi.
738 - attrib.c::decompress_mapping_pairs() now accepts the old run list
739 argument, and invokes attrib.c::merge_run_lists() to merge the old
740 and the new run lists.
741 - Removed attrib.c::find_first_attr().
742 - Implemented loading of attribute list and complete run list for $MFT.
743 This means we now cope with $MFT being spread across several mft
745 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
746 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
747 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
748 cache. There are too little ntfs_volume structures at any one time
749 to justify a private slab cache.
750 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
751 Use KM_BIO_IRQ on advice from IRC/kernel...
752 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
753 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
754 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
755 removed the now superfluous exports from the kernel core patch.
756 - Fix a bug where kfree() was used insted of ntfs_free().
757 - Change map_mft_record() to take ntfs_inode as argument instead of
758 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
759 - Add pointer to ntfs_volume to ntfs_inode.
760 - Add mft record number and sequence number to ntfs_inode. Stop using
761 i_ino and i_generation for in-driver purposes.
762 - Implement attrib.c::merge_run_lists(). (Richard Russon)
763 - Remove use of proper inodes by extent inodes. Move i_ino and
764 i_generation to ntfs_inode to do this. Apply simplifications that
765 result and remove iget_no_wait(), etc.
766 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
767 - Add reference counting in ntfs_inode for the ntfs inode itself and
768 for the mapped mft record.
769 - Extend mft record mapping so we can (un)map extent mft records (new
770 functions (un)map_extent_mft_record), and so mappings are reference
771 counted and don't have to happen twice if already mapped - just ref
773 - Add -o iocharset as alias to -o nls for backwards compatibility.
774 - The latest core patch is now tiny. In fact just a single additional
775 export is necessary over the base kernel.
777 tng-0.0.3 - Cleanups, enhancements, bug fixes.
779 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
780 and setup the run list appropriately using knowledge provided by the
781 sizes in the base attribute record.
782 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
784 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
785 page or use vmalloc depending on the amount of memory requested.
786 - Cleanup error output. The __FUNCTION__ "(): " is now added
787 automatically. Introduced a new header file debug.h to support this
788 and also moved ntfs_debug() function into it.
789 - Make reading of compressed files more intelligent and especially get
790 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
791 buffers (allocated at first mount with cluster size <= 4kiB and
792 deallocated on last umount with cluster size <= 4kiB), and
793 asynchronous io for the compressed data using a list of buffer heads.
794 Er, we use synchronous io as async io only works on whole pages
795 covered by buffers and not on individual buffer heads...
796 - Bug fix for reading compressed files with sparse compression blocks.
798 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
800 - Fixed handling of directories when cluster size exceeds index block
802 - Hide DOS only name space directory entries from readdir() but allow
803 them in lookup(). This should fix the problem that Linux doesn't
804 support directory hard links, while still allowing access to entries
805 via their short file name. This also has the benefit of mimicking
806 what Windows users are used to, so it is the ideal solution.
807 - Implemented sync_page everywhere so no more hangs in D state when
809 - Stop using bforget() in favour of brelse().
810 - Stop locking buffers unnecessarily.
811 - Implemented compressed files (inode->mapping contains uncompressed
812 data, raw compressed data is currently bread() into a vmalloc()ed
814 - Enable compressed directories. (Their index root is marked compressed
815 to indicate that new files in that directory should be created
817 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
818 functions. (Thanks to Will Dyson for pointing this out.)
819 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
820 ntfs_sb_info) out of the common inode and super_block structures and
821 started using the generic_ip and generic_sbp pointers instead. This
822 makes ntfs entirely private with respect to the kernel tree.
823 - Detect compiler version and abort with error message if gcc less than
825 - Fix bug in name comparison function in unistr.c.
826 - Implement attribute lists part 1, the infrastructure: search contexts
827 and operations, find_external_attr(), lookup_attr()) and make the
828 code use the infrastructure.
829 - Fix stupid buffer overflow bug that became apparent on larger run
830 list containing attributes.
831 - Fix bugs in readdir() that became apparent on larger directories.
833 The driver is now really useful and survives the test
834 find . -type f -exec md5sum "{}" \;
835 without any error messages on a over 1GiB sized partition with >16k
836 files on it, including compressed files and directories and many files
837 and directories with attribute lists.
839 tng-0.0.1 - The first useful version.
841 - Added ntfs_lookup().
842 - Added default upcase generation and handling.
843 - Added compile options to be shown on module init.
844 - Many bug fixes that were "hidden" before.
845 - Update to latest kernel.
846 - Added ntfs_readdir().
847 - Added file operations for mmap(), read(), open() and llseek(). We just
848 use the generic ones. The whole point of going through implementing
849 readpage() methods and where possible get_block() call backs is that
850 this allows us to make use of the generic high level methods provided
853 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
854 though and it doesn't implement accesssing compressed files yet. Also,
855 accessing files with attribute list attributes is not implemented yet
856 either. But for small or simple file systems it should work and allow
857 you to list directories, use stat on directory entries and the file
858 system, open, read, mmap and llseek around in files. A big mile stone
861 tng-0.0.0 - Initial version tag.
863 Initial driver implementation. The driver can mount and umount simple
864 NTFS file systems (i.e. ones without attribute lists in the system
865 files). If the mount fails there might be problems in the error handling
866 code paths, so be warned. Otherwise it seems to be loading the system
867 files nicely and the mft record read mapping/unmapping seems to be
868 working nicely, too. Proof of inode metadata in the page cache and non-
869 resident file unnamed stream data in the page cache concepts is thus