3 - Enable NFS exporting of NTFS.
4 - Implement aops->set_page_dirty() in order to take control of buffer
5 dirtying. Not having it means if page_has_buffers(), all buffers
6 will be dirtied with the page. And if not they won't be. That is
7 fine for the moment but will break once we enable metadata updates.
8 - Implement sops->dirty_inode() to implement {a,m,c} time updates and
10 - Implement sops->write_inode().
11 - In between ntfs_prepare/commit_write, need exclusion between
12 simultaneous file extensions. Need perhaps an NInoResizeUnderway()
13 flag which we can set in ntfs_prepare_write() and clear again in
14 ntfs_commit_write(). Just have to be careful in readpage/writepage,
15 as well as in truncate, that we play nice... We might need to have
16 a data_size field in the ntfs_inode to store the real attribute
17 length. Also need to be careful with initialized_size extention in
18 ntfs_prepare_write. Basically, just be _very_ careful in this code...
19 OTOH, perhaps i_sem, which is held accross generic_file_write is
20 sufficient for synchronisation here. We then just need to make sure
21 ntfs_readpage/writepage/truncate interoperate properly with us.
23 2.1.5 - Fix minor bug in attribute list attribute handling.
25 - Fix bug in attribute list handling. Actually it is not as much a bug
26 as too much protection in that we were not allowing attribute lists
27 which waste space on disk while Windows XP clearly allows it and in
28 fact creates such attribute lists so our driver was failing.
29 - Update NTFS documentation ready for 2.6 kernel release.
31 2.1.4 - Reduce compiler requirements.
33 - Remove all uses of unnamed structs and unions in the driver to make
34 old and newer gcc versions happy. Makes it a bit uglier IMO but at
35 least people will stop hassling me about it.
37 2.1.3 - Important bug fixes in corner cases.
39 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
40 clusters. (Philipp Thomas)
41 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
42 multiple of the block_size but not the cluster size. (Szabolcs
43 Szakacsits <szaka@sienet.hu>)
45 2.1.2 - Important bug fixes aleviating the hangs in statfs.
47 - Fix buggy free cluster and free inode determination logic.
49 2.1.1 - Minor updates.
51 - Add handling for initialized_size != data_size in compressed files.
52 - Reduce function local stack usage from 0x3d4 bytes to just noise in
53 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
54 - Remove compiler warnings for newer gcc.
56 2.1.0 - First steps towards write support: implement file overwrite.
58 - Add configuration option for developmental write support with an
59 appropriately scary configuration help text.
60 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
61 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
62 overwriting of existing files on ntfs. Note: Resident files are
63 only written into memory, and not written out to disk at present, so
64 avoid writing to files smaller than about 1kiB.
65 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
66 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
67 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
68 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
69 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
70 This enables write(2) based overwriting of existing files on ntfs.
71 Note: As with mmap(2) based overwriting, resident files are only
72 written into memory, and not written out to disk at present, so avoid
73 writing to files smaller than about 1kiB.
74 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
75 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
76 files with the purpose of intercepting and aborting all i_size
77 changes which we do not support yet. ntfs_truncate() actually only
78 emits a warning message but AFAICS our interception of i_size changes
79 elsewhere means ntfs_truncate() never gets called for i_size changes.
80 It is only called from generic_file_write() when we fail in
81 ntfs_prepare_{,nonresident_}write() in order to discard any
82 instantiated buffers beyond i_size. Thus i_size is not actually
83 changed so our warning message is enough. Unfortunately it is not
84 possible to easily determine if i_size is being changed or not hence
85 we just emit an appropriately worded error message.
87 2.0.25 - Small bug fixes and cleanups.
89 - Unlock the page in an out of memory error code path in
90 fs/ntfs/aops.c::ntfs_read_block().
91 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
92 just unlock the page and return. (This can happen due to ->writepage
93 clearing PageUptodate() during write out of MstProtected()
95 - Remove leaked write code again.
99 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
100 inside BUG_ON(). (Adam J. Richter)
101 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
102 calls for improved debugging. (Adam J. Richter)
103 - Add errors flag to the ntfs volume state, accessed via
104 NVol{,Set,Clear}Errors(vol).
105 - Do not allow read-write remounts of read-only volumes with errors.
106 - Clarify comment for ntfs file operation sendfile which was added by
107 Christoph Hellwig a while ago (just using generic_file_sendfile())
108 to say that ntfs ->sendfile is only used for the case where the
109 source data is on the ntfs partition and the destination is
110 somewhere else, i.e. nothing we need to concern ourselves with.
111 - Add generic_file_write() as our ntfs file write operation.
113 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
115 - Massive internal locking changes to mft record locking. Fixes lock
116 recursion and replaces the mrec_lock read/write semaphore with a
117 mutex. Also removes the now superfluous mft_count. This fixes several
118 race conditions and deadlocks, especially in the future write code.
119 - Fix ntfs over loopback for compressed files by adding an
120 optimization barrier. (gcc was screwing up otherwise ?)
121 - Miscellaneous cleanups all over the code and a fix or two in error
123 Thanks go to Christoph Hellwig for pointing out the following two:
124 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
125 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
127 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
129 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
130 at entry/exit respectively.
131 - Use C99 initializers for structures.
132 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
134 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
136 - Limit volume size at mount time to 2TiB on architectures where
137 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
138 This is the most we can do without overflowing the 32-bit limit of
139 the block device size imposed on us by sb_bread() and sb_getblk()
141 - Limit file/directory size at open() time to 16TiB on architectures
142 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
143 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
144 overflowing the page cache page index.
146 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
148 - Move the directory index bitmap to use an attribute inode instead of
149 having special fields for it inside the ntfs inode structure. This
150 means that the index bitmaps now use the page cache for i/o, too,
151 and also as a side effect we get support for non-resident index
153 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
154 fix a page leak that manifested itself in some cases.
155 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
156 index bitmap inode on the final iput().
158 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
160 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
161 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
162 - Drop the "file" from ntfs_file_read_compressed_block().
163 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
164 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
165 - Update ntfs_end_buffer_async_read() with the improved logic from
166 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
167 further logic improvements to better determine when we set PageError.
168 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
169 check for the buffers being uptodate first in line with the updated
170 fs/buffer.c::block_read_full_page(). This plugs a small race
173 2.0.18 - Fix race condition in reading of compressed files.
175 - There was a narrow window between checking a buffer head for being
176 uptodate and locking it in ntfs_file_read_compressed_block(). We now
177 lock the buffer and then check whether it is uptodate or not.
179 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
181 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
182 code and update callers, i.e. ntfs_iget(), to pass that error code
183 up instead of just using -EIO.
184 - Modifications to super.c to ensure that both mount and remount
185 cannot set any write related options when the driver is compiled
187 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
188 cache the current run list element. This should improve performance
189 when reading very large and/or very fragmented data.
191 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
193 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
194 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
195 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
196 - Convert $MFT/$BITMAP access to attribute inode API and remove all
197 remnants of the ugly mftbmp address space and operations hack. This
198 means we finally have only one readpage function as well as only one
199 async io completion handler. Yey! The mft bitmap is now just an
200 attribute inode and is accessed from vol->mftbmp_ino just as if it
201 were a normal file. Fake inodes rule. (-:
203 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
205 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
206 remounts to fail when the partition had an entry in /etc/fstab and
207 the entry specified the nls= option.
208 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
209 expand all the helper functions NVolFoo(), NVolSetFoo(), and
211 - Move copyright statement from driver initialisation message to
212 module description (fs/super.c). This makes the initialisation
213 message fit on one line and fits in better with rest of kernel.
214 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
215 attribute inodes, and both for files and directories.
216 - Implement fake attribute inodes allowing all attribute i/o to go via
217 the page cache and to use all the normal vfs/mm functionality:
218 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
220 - Add needed cleanup code to ntfs_clear_big_inode().
221 - Merge address space operations for files and directories (aops.c),
222 now just have ntfs_aops:
224 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
225 ntfs_attr_read_block() -> ntfs_read_block(),
226 ntfs_file_read_page() -> ntfs_readpage().
227 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
228 attribute inodes, and both for files and directories.
229 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
231 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
233 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
234 the locking out of super.c::get_nr_free_mft_records() and taking and
235 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
236 - Bring attribute run list merging code (fs/ntfs/attrib.c) in sync with
237 current userspace ntfs library code. This means that if a merge
238 fails the original run lists are always left unmodified instead of
239 being silently corrupted.
242 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
244 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
245 from ntfs_volume structure.
246 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
247 from ntfs_volume structure.
248 - Use iget5_locked() and friends instead of conventional iget(). Wrap
249 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
250 to use ntfs_iget(). Leave only one iget() call at mount time so we
251 don't need an ntfs_iget_mount().
252 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
255 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
257 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
258 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
259 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
260 to determine whether to apply mst fixups or not.
261 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
262 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
263 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
264 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
265 the VFS readpage function prototype to the ntfs_attr_read_block()
268 2.0.11 - Initial preparations for fake inode based attribute i/o.
270 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
271 do some macro magic (adapted from include/linux/buffer_head.h) to
272 expand all the helper functions NInoFoo(), NInoSetFoo(), and
274 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
275 - Add new fields to ntfs_inode structure to allow use of fake inodes
276 for attribute i/o: type, name, name_len. Also add new state bits:
277 NI_Attr, which, if set, indicates the inode is a fake inode, and
278 NI_MstProtected, which, if set, indicates the attribute uses multi
279 sector transfer protection, i.e. fixups need to be applied after
280 reads and before/after writes.
281 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
282 ntfs_{new,clear,destroy}_extent_inode() and update callers.
283 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
284 instead of ntfs_destroy_extent_inode().
285 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
286 - Make all operations on ntfs inode state bits use the NIno* functions.
287 - Set up the new ntfs inode fields and state bits in
288 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
289 allocated memory to __ntfs_clear_inode().
290 - Cleanup ntfs_inode structure a bit for better ordering of elements
291 w.r.t. their size to allow better packing of the structure in memory.
293 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
295 - Add check at mount time to verify that the number of inodes on the
296 volume does not exceed 2^32 - 1, which is the maximum allowed for
297 NTFS according to Microsoft.
298 - Change mft_no member of ntfs_inode structure to be unsigned long.
299 Update all users. This makes ntfs_inode->mft_no just a copy of struct
300 inode->i_ino. But we can't just always use struct inode->i_ino and
301 remove mft_no because extent inodes do not have an attached struct
304 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
306 - Change decompression engine to use a single buffer protected by a
307 spin lock instead of per-CPU buffers. (Rusty Russell)
308 - Do not update cb_pos when handling a partial final page during
309 decompression of a sparse compression block, as the value is later
310 reset without being read/used. (Rusty Russell)
311 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
313 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
314 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
315 it also makes everything safer so it is a good thing.
316 - Miscellaneous minor cleanups to comments.
318 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
320 Big thanks go to Al Viro and other inhabitants of #kernel for investing
321 their time to discuss the case sensitivity and dcache aliasing issues.
323 - Remove unused source file fs/ntfs/attraops.c.
324 - Remove show_inodes mount option(s), thus dropping support for
325 displaying of short file names.
326 - Remove deprecated mount option posix.
327 - Restore show_sys_files mount option.
328 - Add new mount option case_sensitive, to determine if the driver
329 treats file names as case sensitive or not. If case sensitive, create
330 file names in the POSIX namespace. Otherwise create file names in the
331 LONG/WIN32 namespace. Note, files remain accessible via their short
332 file name, if it exists.
333 - Remove really dumb logic bug in boot sector recovery code.
334 - Fix dcache aliasing issues wrt short/long file names via changes
335 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
336 fs/ntfs/namei.c::ntfs_lookup():
337 - Add additional argument to ntfs_lookup_inode_by_name() in which we
338 return information about the matching file name if the case is not
339 matching or the match is a short file name. See comments above the
340 function definition for details.
341 - Change ntfs_lookup() to only create dcache entries for the correctly
342 cased file name and only for the WIN32 namespace counterpart of DOS
343 namespace file names. This ensures we have only one dentry per
344 directory and also removes all dcache aliasing issues between short
345 and long file names once we add write support. See comments above
346 function for details.
347 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
349 2.0.7 - Minor cleanups and updates for changes in core kernel code.
351 - Remove much of the NULL struct element initializers.
352 - Various updates to make compatible with recent kernels.
353 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
354 in fs/ntfs/ntfs.h instead.
355 - Remove no longer needed KERNEL_VERSION checks. We are now in the
356 kernel proper so they are no longer needed.
358 2.0.6 - Major bugfix to make compatible with other kernel changes.
360 - Initialize the mftbmp address space properly now that there are more
361 fields in the struct address_space. This was leading to hangs and
362 oopses on umount since 2.5.12 because of changes to other parts of
363 the kernel. We probably want a kernel generic init_address_space()
365 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
366 only caller of ->readdir() is vfs_readdir() which holds i_sem during
367 the call, and i_sem is sufficient protection against changes in the
368 directory inode (including ->i_size).
369 - Use generic_file_llseek() for directories (as opposed to
370 default_llseek()) as this downs i_sem instead of the BKL which is
371 what we now need for exclusion against ->f_pos changes considering we
372 no longer take the BKL in ntfs_readdir().
374 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
376 - No need to set old blocksize in super.c::ntfs_fill_super() as the
377 VFS does so via invocation of deactivate_super() calling
378 fs->fill_super() calling block_kill_super() which does it.
379 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
380 -> Do we really need it? I don't think so as we have exclusion on
381 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
382 move the ->f_pos accesses under the mrec_lock though. Check this...
383 - Fix really, really, really stupid buffer overflow in extent inode
384 handling in mft.c::map_extent_mft_record().
386 2.0.4 - Cleanups and updates for kernel 2.5.11.
388 - Add documentation on how to use the MD driver to be able to use NTFS
389 stripe and volume sets in Linux and generally cleanup documentation
391 Remove all uses of kdev_t in favour of struct block_device *:
392 - Change compress.c::ntfs_file_read_compressed_block() to use
393 sb_getblk() instead of getblk().
394 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
395 of get_hardsect_size().
396 - No need to get old blocksize in super.c::ntfs_fill_super() as
397 fs/super.c::get_sb_bdev() already does this.
398 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
400 2.0.3 - Small bug fixes, cleanups, and performance improvements.
402 - Remove some dead code from mft.c.
403 - Optimize readpage and read_block functions throughout aops.c so that
404 only initialized blocks are read. Non-initialized ones have their
405 buffer head mapped, zeroed, and set up to date, without scheduling
406 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
407 Thanks go to Andrew Morton for spotting the below:
408 - Fix buglet in allocate_compression_buffers() error code path.
409 - Call flush_dcache_page() after modifying page cache page contents in
410 ntfs_file_readpage().
411 - Check for existence of page buffers throughout aops.c before calling
412 create_empty_buffers(). This happens when an I/O error occurs and the
413 read is retried. (It also happens once writing is implemented so that
414 needed doing anyway but I had left it for later...)
415 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
416 readpage and read_block functions. Reasoning same as above (i.e. I/O
417 error retries and future write code paths.)
419 2.0.2 - Minor updates and cleanups.
421 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
422 and cleanup the code a bit, removing the unused size parameter.
423 - Change default fmask to 0177 and update documentation.
424 - Change attrib.c::get_attr_search_ctx() to return the search context
425 directly instead of taking the address of a pointer. A return value
426 of NULL means the allocation failed. Updated all callers
428 - Update to 2.5.9 kernel (preserving backwards compatibility) by
429 replacing all occurences of page->buffers with page_buffers(page).
430 - Fix minor bugs in run list merging, also minor cleanup.
431 - Updates to bootsector layout and mft mirror contents descriptions.
432 - Small bug fix in error detection in unistr.c and some cleanups.
433 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
436 2.0.1 - Minor updates.
438 - Make default umask correspond to documentation.
439 - Improve documentation.
440 - Set default mode to include execute bit. The {u,f,d}mask can be used
441 to take it away if desired. This allows binaries to be executed from
442 a mounted ntfs partition.
444 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
446 - Add kill_super, just keeping up with the vfs changes in the kernel.
447 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
448 from the CVS import into BitKeeper.
449 - Begin to implement proper handling of allocated_size vs
450 initialized_size vs data_size (i.e. i_size). Done are
451 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
452 and attrib.c::load_attribute_list().
453 - Lock the run list in attrib.c::load_attribute_list() while using it.
454 - Fix memory leak in ntfs_file_read_compressed_block() and generally
455 clean up compress.c a little, removing some uncommented/unused debug
457 - Tidy up dir.c a little bit.
458 - Don't bother getting the run list in inode.c::ntfs_read_inode().
459 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
460 creating aops.c::ntfs_mst_readpage(), improving the handling of
461 holes and overflow in the process and implementing the correct
462 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
463 I am aiming for correctness at the moment. Modularisation can come
465 - Rename aops.c::end_buffer_read_index_async() to
466 end_buffer_read_mst_async() and optimize the overflow checking and
468 - Use the host of the mftbmp address space mapping to hold the ntfs
469 volume. This is needed so the async i/o completion handler can
470 retrieve a pointer to the volume. Hopefully this will not cause
471 problems elsewhere in the kernel... Otherwise will need to use a
473 - Complete implementation of proper handling of allocated_size vs
474 initialized_size vs data_size (i.e. i_size) in whole driver.
475 Basically aops.c is now completely rewritten.
476 - Change NTFS driver name to just NTFS and set version number to 2.0.0
477 to make a clear distinction from the old driver which is still on
480 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
482 - Replace bdevname(sb->s_dev) with sb->s_id.
483 - Remove now superfluous new-line characters in all callers of
485 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
486 directories. Without this the "find" utility gets very upset which is
487 fair enough as Linux/Unix do not support directory hard links.
488 - Further run list merging work. (Richard Russon)
489 - Backwards compatibility for gcc-2.95. (Richard Russon)
490 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
491 - Convert to new file system declaration using ->ntfs_get_sb() and
492 replacing ntfs_read_super() with ntfs_fill_super().
493 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
494 overflow on 32-bit architectures.
495 - Cleanup upcase loading code to use ntfs_(un)map_page().
496 - Disable/reenable preemtion in critical sections of compession engine.
497 - Replace device size determination in ntfs_fill_super() with
498 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
499 function super.c::get_nr_blocks().
500 - Implement a mount time option (show_inodes) allowing choice of which
501 types of inode names readdir() returns and modify ntfs_filldir()
502 accordingly. There are several parameters to show_inodes:
504 win32: long file names (including POSIX file names) [DEFAULT]
506 dos: short file names only (excluding POSIX file names)
508 posix: same as both win32 and dos
510 Note that the options are additive, i.e. specifying:
511 -o show_inodes=system,show_inodes=win32,show_inodes=dos
512 is the same as specifying:
514 Note that the "posix" and "all" options will show all directory
515 names, BUT the link count on each directory inode entry is set to 1,
516 due to Linux not supporting directory hard links. This may well
517 confuse some userspace applications, since the directory names will
518 have the same inode numbers. Thus it is NOT advisable to use the
519 "posix" or "all" options. We provide them only for completeness sake.
520 - Add copies of allocated_size, initialized_size, and compressed_size to
521 the ntfs inode structure and set them up in
522 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
523 for files and the index allocation attribute for directories.
524 - Add copies of allocated_size and initialized_size to ntfs inode for
525 $BITMAP attribute of large directories and set them up in
526 inode.c::ntfs_read_inode().
527 - Add copies of allocated_size and initialized_size to ntfs volume for
528 $BITMAP attribute of $MFT and set them up in
529 super.c::load_system_files().
530 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
531 posix, and utf8) and tell user what the new options to use are. Note
532 we still do support them but they will be removed with kernel 2.7.x.
533 - Change all occurences of integer long long printf formatting to hex
534 as printk() will not support long long integer format if/when the
535 div64 patch goes into the kernel.
536 - Make slab caches have stable names and change the names to what they
537 were intended to be. These changes are required/made possible by the
538 new slab cache name handling which removes the length limitation by
539 requiring the caller of kmem_cache_create() to supply a stable name
540 which is then referenced but not copied.
541 - Rename run_list structure to run_list_element and create a new
542 run_list structure containing a pointer to a run_list_element
543 structure and a read/write semaphore. Adapt all users of run lists
544 to new scheme and take and release the lock as needed. This fixes a
545 nasty race as the run_list changes even when inodes are locked for
546 reading and even when the inode isn't locked at all, so we really
547 needed the serialization. We use a semaphore rather than a spinlock
548 as memory allocations can sleep and doing everything GFP_ATOMIC
550 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
551 This can never happen due to the nature of lookup_attr() and how we
552 support attribute lists. If it did happen it would imply the inode
554 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
556 - Update to 2.5.6-pre2 changes in struct address_space.
557 - Use parent_ino() when accessing d_parent inode number in dir.c.
558 - Import Sourceforge CVS repository into BitKeeper repository:
559 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
560 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
561 Documentation/filesystems/ntfs.txt for NTFS TNG.
562 - Create kernel configuration option controlling whether debugging
564 - Add the required export of end_buffer_io_sync() from the patches
565 directory to the kernel code.
566 - Update inode.c::ntfs_show_options() with show_inodes mount option.
567 - Update errors mount option.
569 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
571 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
572 bug in mapping of extent inodes. Update all the comments to fit all
573 the recent code changes.
574 - Modify vcn_to_lcn() to cope with entirely unmapped run lists.
575 - Cleanups in compress.c, mostly comments and folding help.
576 - Implement attrib.c::map_run_list() as a generic helper.
577 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
578 thus making code shorter and enabling attribute list support.
579 - Cleanup incorrect use of [su]64 with %L printf format specifier in
580 all source files. Type casts to [unsigned] long long added to correct
581 the mismatches (important for architectures which have long long not
583 - Merge async io completion handlers for directory indexes and $MFT
584 data into one by setting the index_block_size{_bits} of the ntfs
585 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
586 - Cleanup aops.c, update comments.
587 - Make ntfs_file_get_block() use map_run_list() so all files now
588 support attribute lists.
589 - Make ntfs_dir_readpage() almost verbatim copy of
590 block_read_full_page() by using ntfs_file_get_block() with only real
591 difference being the use of our own async io completion handler
592 rather than the default one, thus reducing the amount of code and
593 automatically enabling attribute list support for directory indices.
594 - Fix bug in load_attribute_list() - forgot to call brelse in error
596 - Change parameters to find_attr() and lookup_attr(). We no longer
597 pass in the upcase table and its length. These can be gotten from
598 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
599 - Cleanups in attrib.c.
600 - Implement merging of run lists, attrib.c::merge_run_lists() and its
601 helpers. (Richard Russon)
602 - Attribute lists part 2, attribute extents and multi part run lists:
603 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
604 further run list parts via attrib.c::map_run_list().
605 - Tiny endianness bug fix in decompress_mapping_pairs().
607 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
609 - Enable encrypted directories. (Their index root is marked encrypted
610 to indicate that new files in that directory should be created
612 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
613 - Enable $Extend system directory. Most (if not all) extended system
614 files do not have unnamed data attributes so ntfs_read_inode() had to
615 special case them but that is ok, as the special casing recovery
616 happens inside an error code path so there is zero slow down in the
617 normal fast path. The special casing is done by introducing a new
618 function inode.c::ntfs_is_extended_system_file() which checks if any
619 of the hard links in the inode point to $Extend as being their parent
620 directory and if they do we assume this is an extended system file.
621 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
622 when compiled with -DDEBUG. Default is debug messages to be disabled.
623 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
624 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
625 interface is enabled). Inspired by old ntfs driver.
626 - Add debug_msgs insmod/kernel boot parameter to set whether debug
627 messages are {dis,en}abled. This is useful to enable debug messages
628 during ntfs initialization and is the only way to activate debugging
629 when the sysctl interface is not enabled.
630 - Cleanup debug output in various places.
631 - Remove all dollar signs ($) from the source (except comments) to
632 enable compilation on architectures whose gcc compiler does not
633 support dollar signs in the names of variables/constants. Attribute
634 types now start with AT_ instead of $ and $I30 is now just I30.
635 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
636 - Load complete run list for $MFT/$BITMAP during mount and cleanup
637 access functions. This means we now cope with $MFT/$BITMAP being
638 spread accross several mft records.
639 - Disable modification of mft_zone_multiplier on remount. We can always
640 reenable this later on if we really want to, but we will need to make
641 sure we readjust the mft_zone size / layout accordingly.
643 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
645 - Use sb_set_blocksize() instead of set_blocksize() and verify the
647 - Use sb_bread() instead of bread() throughout.
648 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
649 of a directory index block vcn. Apply resulting simplifications in
651 - Fix a small bug somewhere (but forgot what it was).
652 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
653 on the printf-format parameter list and fix bugs reported by gcc
654 as a result. (Richard Russon)
655 - Move inode allocation strategy to Al's new stuff but maintain the
656 divorce of ntfs_inode from struct inode. To achieve this we have two
657 separate slab caches, one for big ntfs inodes containing a struct
658 inode and pure ntfs inodes and at the same time fix some faulty
659 error code paths in ntfs_read_inode().
660 - Show mount options in proc (inode.c::ntfs_show_options()).
662 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
664 - Modified (un)map_mft_record functions to be common for read and write
665 case. To specify which is which, added extra parameter at front of
666 parameter list. Pass either READ or WRITE to this, each has the
668 - General cleanups to allow for easier folding in vi.
669 - attrib.c::decompress_mapping_pairs() now accepts the old run list
670 argument, and invokes attrib.c::merge_run_lists() to merge the old
671 and the new run lists.
672 - Removed attrib.c::find_first_attr().
673 - Implemented loading of attribute list and complete run list for $MFT.
674 This means we now cope with $MFT being spread across several mft
676 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
677 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
678 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
679 cache. There are too little ntfs_volume structures at any one time
680 to justify a private slab cache.
681 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
682 Use KM_BIO_IRQ on advice from IRC/kernel...
683 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
684 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
685 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
686 removed the now superfluous exports from the kernel core patch.
687 - Fix a bug where kfree() was used insted of ntfs_free().
688 - Change map_mft_record() to take ntfs_inode as argument instead of
689 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
690 - Add pointer to ntfs_volume to ntfs_inode.
691 - Add mft record number and sequence number to ntfs_inode. Stop using
692 i_ino and i_generation for in-driver purposes.
693 - Implement attrib.c::merge_run_lists(). (Richard Russon)
694 - Remove use of proper inodes by extent inodes. Move i_ino and
695 i_generation to ntfs_inode to do this. Apply simplifications that
696 result and remove iget_no_wait(), etc.
697 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
698 - Add reference counting in ntfs_inode for the ntfs inode itself and
699 for the mapped mft record.
700 - Extend mft record mapping so we can (un)map extent mft records (new
701 functions (un)map_extent_mft_record), and so mappings are reference
702 counted and don't have to happen twice if already mapped - just ref
704 - Add -o iocharset as alias to -o nls for backwards compatibility.
705 - The latest core patch is now tiny. In fact just a single additional
706 export is necessary over the base kernel.
708 tng-0.0.3 - Cleanups, enhancements, bug fixes.
710 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
711 and setup the run list appropriately using knowledge provided by the
712 sizes in the base attribute record.
713 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
715 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
716 page or use vmalloc depending on the amount of memory requested.
717 - Cleanup error output. The __FUNCTION__ "(): " is now added
718 automatically. Introduced a new header file debug.h to support this
719 and also moved ntfs_debug() function into it.
720 - Make reading of compressed files more intelligent and especially get
721 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
722 buffers (allocated at first mount with cluster size <= 4kiB and
723 deallocated on last umount with cluster size <= 4kiB), and
724 asynchronous io for the compressed data using a list of buffer heads.
725 Er, we use synchronous io as async io only works on whole pages
726 covered by buffers and not on individual buffer heads...
727 - Bug fix for reading compressed files with sparse compression blocks.
729 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
731 - Fixed handling of directories when cluster size exceeds index block
733 - Hide DOS only name space directory entries from readdir() but allow
734 them in lookup(). This should fix the problem that Linux doesn't
735 support directory hard links, while still allowing access to entries
736 via their short file name. This also has the benefit of mimicking
737 what Windows users are used to, so it is the ideal solution.
738 - Implemented sync_page everywhere so no more hangs in D state when
740 - Stop using bforget() in favour of brelse().
741 - Stop locking buffers unnecessarily.
742 - Implemented compressed files (inode->mapping contains uncompressed
743 data, raw compressed data is currently bread() into a vmalloc()ed
745 - Enable compressed directories. (Their index root is marked compressed
746 to indicate that new files in that directory should be created
748 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
749 functions. (Thanks to Will Dyson for pointing this out.)
750 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
751 ntfs_sb_info) out of the common inode and super_block structures and
752 started using the generic_ip and generic_sbp pointers instead. This
753 makes ntfs entirely private with respect to the kernel tree.
754 - Detect compiler version and abort with error message if gcc less than
756 - Fix bug in name comparison function in unistr.c.
757 - Implement attribute lists part 1, the infrastructure: search contexts
758 and operations, find_external_attr(), lookup_attr()) and make the
759 code use the infrastructure.
760 - Fix stupid buffer overflow bug that became apparent on larger run
761 list containing attributes.
762 - Fix bugs in readdir() that became apparent on larger directories.
764 The driver is now really useful and survives the test
765 find . -type f -exec md5sum "{}" \;
766 without any error messages on a over 1GiB sized partition with >16k
767 files on it, including compressed files and directories and many files
768 and directories with attribute lists.
770 tng-0.0.1 - The first useful version.
772 - Added ntfs_lookup().
773 - Added default upcase generation and handling.
774 - Added compile options to be shown on module init.
775 - Many bug fixes that were "hidden" before.
776 - Update to latest kernel.
777 - Added ntfs_readdir().
778 - Added file operations for mmap(), read(), open() and llseek(). We just
779 use the generic ones. The whole point of going through implementing
780 readpage() methods and where possible get_block() call backs is that
781 this allows us to make use of the generic high level methods provided
784 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
785 though and it doesn't implement accesssing compressed files yet. Also,
786 accessing files with attribute list attributes is not implemented yet
787 either. But for small or simple file systems it should work and allow
788 you to list directories, use stat on directory entries and the file
789 system, open, read, mmap and llseek around in files. A big mile stone
792 tng-0.0.0 - Initial version tag.
794 Initial driver implementation. The driver can mount and umount simple
795 NTFS file systems (i.e. ones without attribute lists in the system
796 files). If the mount fails there might be problems in the error handling
797 code paths, so be warned. Otherwise it seems to be loading the system
798 files nicely and the mft record read mapping/unmapping seems to be
799 working nicely, too. Proof of inode metadata in the page cache and non-
800 resident file unnamed stream data in the page cache concepts is thus
projects / reactos.git / blob