blob: 32a4150ebcb2f3f48e9308829e466edab6906e8d [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001ToDo/Notes:
2 - Find and fix bugs.
Linus Torvalds1da177e2005-04-16 15:20:36 -07003 - In between ntfs_prepare/commit_write, need exclusion between
Anton Altaparmakov66129f82004-11-11 12:34:00 +00004 simultaneous file extensions. This is given to us by holding i_sem
5 on the inode. The only places in the kernel when a file is resized
6 are prepare/commit write and truncate for both of which i_sem is
7 held. Just have to be careful in readpage/writepage and all other
8 helpers not running under i_sem that we play nice...
9 Also need to be careful with initialized_size extention in
Linus Torvalds1da177e2005-04-16 15:20:36 -070010 ntfs_prepare_write. Basically, just be _very_ careful in this code...
Anton Altaparmakov66129f82004-11-11 12:34:00 +000011 UPDATE: The only things that need to be checked are read/writepage
12 which do not hold i_sem. Note writepage cannot change i_size but it
13 needs to cope with a concurrent i_size change, just like readpage.
14 Also both need to cope with concurrent changes to the other sizes,
15 i.e. initialized/allocated/compressed size, as well.
Linus Torvalds1da177e2005-04-16 15:20:36 -070016 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
17 leave the volume dirty on umount if the final iput(vol->mft_ino)
18 causes a write of any mirrored mft records due to the mft mirror
19 inode having been discarded already. Whether this can actually ever
20 happen is unclear however so it is worth waiting until someone hits
21 the problem.
22 - Enable the code for setting the NT4 compatibility flag when we start
23 making NTFS 1.2 specific modifications.
24
Anton Altaparmakove7a10332005-09-08 16:12:28 +0100252.1.24-WIP
26
27 - Support journals ($LogFile) which have been modified by chkdsk. This
28 means users can boot into Windows after we marked the volume dirty.
29 The Windows boot will run chkdsk and then reboot. The user can then
30 immediately boot into Linux rather than having to do a full Windows
31 boot first before rebooting into Linux and we will recognize such a
32 journal and empty it as it is clean by definition.
33 - Support journals ($LogFile) with only one restart page as well as
34 journals with two different restart pages. We sanity check both and
35 either use the only sane one or the more recent one of the two in the
36 case that both are valid.
Anton Altaparmakov06d0e3c2005-09-08 16:28:25 +010037 - Modify fs/ntfs/malloc.h::ntfs_malloc_nofs() to do the kmalloc() based
38 allocations with __GFP_HIGHMEM, analogous to how the vmalloc() based
39 allocations are done.
40 - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
41 ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
42 hence cannot fail.
Anton Altaparmakov9529d462005-09-08 16:33:12 +010043 - Use ntfs_malloc_nofs_nofail() in the two critical regions in
44 fs/ntfs/runlist.c::ntfs_runlists_merge(). This means we no longer
45 need to panic() if the allocation fails as it now cannot fail.
Anton Altaparmakov84d6ebe2005-09-08 16:46:55 +010046 - Fix two nasty runlist merging bugs that had gone unnoticed so far.
47 Thanks to Stefano Picerno for the bug report.
Anton Altaparmakov8bb73522005-09-08 16:48:28 +010048 - Remove two bogus BUG_ON()s from fs/ntfs/mft.c.
Anton Altaparmakov2b0ada22005-09-08 16:52:31 +010049 - Fix handling of valid but empty mapping pairs array in
50 fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress().
Anton Altaparmakovf94ad382005-09-08 17:04:11 +010051 - Report unrepresentable inodes during ntfs_readdir() as KERN_WARNING
52 messages and include the inode number. Thanks to Yura Pakhuchiy for
53 pointing this out.
Anton Altaparmakov3ffc5a42005-09-08 20:23:06 +010054 - Change ntfs_rl_truncate_nolock() to throw away the runlist if the new
55 length is zero.
Anton Altaparmakov6e483212005-09-08 20:26:34 +010056 - Add runlist.[hc]::ntfs_rl_punch_nolock() which punches a caller
57 specified hole into a runlist.
Anton Altaparmakov8e08cea2005-09-08 20:29:50 +010058 - Fix a bug in fs/ntfs/index.c::ntfs_index_lookup(). When the returned
59 index entry is in the index root, we forgot to set the @ir pointer in
60 the index context. Thanks to Yura Pakhuchiy for finding this bug.
Anton Altaparmakovf25dfb52005-09-08 20:35:33 +010061 - Remove bogus setting of PageError in ntfs_read_compressed_block().
Anton Altaparmakov0aaccea2005-09-08 20:40:32 +010062 - Add fs/ntfs/attrib.[hc]::ntfs_resident_attr_value_resize().
Anton Altaparmakov2983d1b2005-09-08 20:56:09 +010063 - Fix a bug in ntfs_map_runlist_nolock() where we forgot to protect
64 access to the allocated size in the ntfs inode with the size lock.
65 - Fix ntfs_attr_vcn_to_lcn_nolock() and ntfs_attr_find_vcn_nolock() to
66 return LCN_ENOENT when there is no runlist and the allocated size is
67 zero.
68 - Fix load_attribute_list() to handle the case of a NULL runlist.
Anton Altaparmakov807c4532005-09-08 21:01:17 +010069 - Fix handling of sparse attributes in ntfs_attr_make_non_resident().
70 - Add BUG() checks to ntfs_attr_make_non_resident() and ntfs_attr_set()
71 to ensure that these functions are never called for compressed or
72 encrypted attributes.
Anton Altaparmakovbbf18132005-09-08 21:09:06 +010073 - Fix cluster (de)allocators to work when the runlist is NULL and more
74 importantly to take a locked runlist rather than them locking it
75 which leads to lock reversal.
Anton Altaparmakov1c7d4692005-09-08 21:15:09 +010076 - Truncate {a,c,m}time to the ntfs supported time granularity when
77 updating the times in the inode in ntfs_setattr().
Anton Altaparmakov67bb1032005-09-08 21:19:45 +010078 - Fixup handling of sparse, compressed, and encrypted attributes in
Anton Altaparmakov311120e2005-09-08 22:04:20 +010079 fs/ntfs/inode.c::ntfs_read_locked_{,attr_,index_}inode(),
80 fs/ntfs/aops.c::ntfs_{read,write}page().
Anton Altaparmakov8dcdeba2005-09-08 21:25:48 +010081 - Make ntfs_write_block() not instantiate sparse blocks if they contain
82 only zeroes.
Anton Altaparmakov54b02eb2005-09-08 21:43:47 +010083 - Optimize fs/ntfs/aops.c::ntfs_write_block() by extending the page
84 lock protection over the buffer submission for i/o which allows the
85 removal of the get_bh()/put_bh() pairs for each buffer.
Anton Altaparmakov8273d5d2005-09-08 22:00:33 +010086 - Fix fs/ntfs/aops.c::ntfs_{read,write}_block() to handle the case
87 where a concurrent truncate has truncated the runlist under our feet.
Anton Altaparmakova01ac532005-09-08 22:08:11 +010088 - Fix page_has_buffers()/page_buffers() handling in fs/ntfs/aops.c.
Anton Altaparmakove6046352005-09-08 22:13:02 +010089 - In fs/ntfs/aops.c::ntfs_end_buffer_async_read(), use a bit spin lock
90 in the first buffer head instead of a driver global spin lock to
91 improve scalability.
Anton Altaparmakove7a10332005-09-08 16:12:28 +010092
Anton Altaparmakovaf859a42005-06-25 21:07:27 +0100932.1.23 - Implement extension of resident files and make writing safe as well as
94 many bug fixes, cleanups, and enhancements...
Linus Torvalds1da177e2005-04-16 15:20:36 -070095
96 - Add printk rate limiting for ntfs_warning() and ntfs_error() when
97 compiled without debug. This avoids a possible denial of service
98 attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
99 out.
Randy Dunlap89075472005-03-03 11:19:53 +0000100 - Fix compilation warnings on ia64. (Randy Dunlap)
Anton Altaparmakovaf859a42005-06-25 21:07:27 +0100101 - Use i_size_{read,write}() instead of reading i_size by hand and cache
102 the value where apropriate.
Anton Altaparmakov36763672004-11-18 13:46:45 +0000103 - Add size_lock to the ntfs_inode structure. This is an rw spinlock
104 and it locks against access to the inode sizes. Note, ->size_lock
105 is also accessed from irq context so you must use the _irqsave and
Anton Altaparmakovaf859a42005-06-25 21:07:27 +0100106 _irqrestore lock and unlock functions, respectively. Protect all
107 accesses to allocated_size, initialized_size, and compressed_size.
Anton Altaparmakov218357f2004-11-18 20:34:59 +0000108 - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
Anton Altaparmakov946929d2005-01-13 15:26:29 +0000109 - Implement extension of resident files in the regular file write code
110 paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()). At present
111 this only works until the data attribute becomes too big for the mft
112 record after which we abort the write returning -EOPNOTSUPP from
113 ntfs_prepare_write().
Anton Altaparmakovc002f422005-02-03 12:02:56 +0000114 - Add disable_sparse mount option together with a per volume sparse
115 enable bit which is set appropriately and a per inode sparse disable
116 bit which is preset on some system file inodes as appropriate.
117 - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
Anton Altaparmakov1a0df152005-02-03 12:04:36 +0000118 - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
119 the creation of the unmapped runlist element for the base attribute
120 extent.
Anton Altaparmakovb6ad6c52005-02-15 10:08:43 +0000121 - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
122 helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
123 This allows us to map runlist fragments with the runlist lock already
124 held without having to drop and reacquire it around the call. Adapt
125 all callers.
126 - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
127 runlist. This allows us to find runlist elements with the runlist
128 lock already held without having to drop and reacquire it around the
129 call. Adapt all callers.
Anton Altaparmakov37e4c132005-02-18 10:03:13 +0000130 - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
Anton Altaparmakovd8ec7852005-02-18 09:23:39 +0000131 warning in the do_div() call on sparc32. Thanks to Meelis Roos for
132 the report and analysis of the warning.
Anton Altaparmakov37e4c132005-02-18 10:03:13 +0000133 - Fix a nasty runlist merge bug when merging two holes.
Anton Altaparmakov5ae9fcf2005-03-02 17:03:24 +0000134 - Set the ntfs_inode->allocated_size to the real allocated size in the
135 mft record for resident attributes (fs/ntfs/inode.c).
136 - Small readability cleanup to use "a" instead of "ctx->attr"
137 everywhere (fs/ntfs/inode.c).
Anton Altaparmakov41382682005-03-03 13:44:15 +0000138 - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
139 definition of ntfs_export_ops from fs/ntfs/super.c to namei.c. Also,
140 declare ntfs_export_ops in fs/ntfs/ntfs.h.
Anton Altaparmakov9451f852005-03-03 14:43:43 +0000141 - Correct sparse file handling. The compressed values need to be
142 checked and set in the ntfs inode as done for compressed files and
143 the compressed size needs to be used for vfs inode->i_blocks instead
144 of the allocated size, again, as done for compressed files.
Anton Altaparmakov271849a2005-03-07 21:36:18 +0000145 - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
146 non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
147 - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
148 write code.
Anton Altaparmakovc0c1cc02005-03-07 21:43:38 +0000149 - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
150 dropping the read lock and taking the write lock we were not checking
151 whether someone else did not already do the work we wanted to do.
152 - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
153 ntfs_attr_find_vcn_nolock() and update all callers.
Anton Altaparmakov2bfb4ff2005-03-09 15:15:06 +0000154 - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
Anton Altaparmakov43b01fd2005-03-09 15:18:43 +0000155 - Fix sign of various error return values to be negative in
156 fs/ntfs/lcnalloc.c.
Anton Altaparmakov905685f2005-03-10 11:06:19 +0000157 - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
158 handle the case where an attribute is converted from resident to
159 non-resident by a concurrent file write.
Jesper Juhl251c8422005-04-04 14:59:56 +0100160 - Remove checks for NULL before calling kfree() since kfree() does the
161 checking itself. (Jesper Juhl)
Anton Altaparmakovb0d23742005-04-04 16:20:14 +0100162 - Some utilities modify the boot sector but do not update the checksum.
163 Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
164 only emit a warning when the checksum is incorrect rather than
165 refusing the mount. Thanks to Bernd Casimir for pointing this
166 problem out.
Anton Altaparmakovbb3cf332005-04-06 13:34:31 +0100167 - Update attribute definition handling.
Anton Altaparmakov7fafb8b2005-04-06 16:09:21 +0100168 - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
169 - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
Anton Altaparmakovd53ee322005-04-06 16:11:20 +0100170 - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
171 better code generation and one less sparse warning in fs/ntfs/aops.c.
Pekka Enberg2fb21db2005-05-25 21:15:34 +0300172 - Remove spurious void pointer casts from fs/ntfs/. (Pekka Enberg)
Anton Altaparmakov442d2072005-05-27 16:42:56 +0100173 - Use C99 style structure initialization after memory allocation where
174 possible (fs/ntfs/{attrib.c,index.c,super.c}). Thanks to Al Viro and
175 Pekka Enberg.
Anton Altaparmakov3f2faef2005-06-25 15:28:56 +0100176 - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
177 is active on the volume and we are mounting read-write or remounting
178 from read-only to read-write.
Anton Altaparmakov9f993fe2005-06-25 16:15:36 +0100179 - Fix a bug in address space operations error recovery code paths where
180 if the runlist was not mapped at all and a mapping error occured we
181 would leave the runlist locked on exit to the function so that the
182 next access to the same file would try to take the lock and deadlock.
Anton Altaparmakovca8fd7a2005-06-25 16:31:27 +0100183 - Detect the case when Windows has been suspended to disk on the volume
184 to be mounted and if this is the case do not allow (re)mounting
185 read-write. This is done by parsing hiberfil.sys if present.
Anton Altaparmakov3bd1f4a2005-06-25 16:51:58 +0100186 - Fix several occurences of a bug where we would perform 'var & ~const'
187 with a 64-bit variable and a int, i.e. 32-bit, constant. This causes
188 the higher order 32-bits of the 64-bit variable to be zeroed. To fix
189 this cast the 'const' to the same 64-bit type as 'var'.
Anton Altaparmakov1d58b272005-06-25 17:04:55 +0100190 - Change the runlist terminator of the newly allocated cluster(s) to
191 LCN_ENOENT in ntfs_attr_make_non_resident(). Otherwise the runlist
192 code gets confused.
Anton Altaparmakovfa3be922005-06-25 17:15:36 +0100193 - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
194 and ntfs_mapping_pairs_build() to allow the runlist encoding to be
195 partial which is desirable when filling holes in sparse attributes.
196 Update all callers.
Anton Altaparmakov4757d7df2005-06-25 17:24:08 +0100197 - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
198 if the requested vcn is inside it. Otherwise we get into problems
199 when we try to map an out of bounds vcn because we then try to map
200 the already mapped runlist fragment which causes
201 ntfs_mapping_pairs_decompress() to fail and return error. Update
202 ntfs_attr_find_vcn_nolock() accordingly.
Anton Altaparmakovba6d2372005-06-26 22:12:02 +0100203 - Fix a nasty deadlock that appeared in recent kernels.
204 The situation: VFS inode X on a mounted ntfs volume is dirty. For
205 same inode X, the ntfs_inode is dirty and thus corresponding on-disk
206 inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
207 to the table of inodes, i.e. $MFT, inode 0.
208 What happens:
209 Process 1: sys_sync()/umount()/whatever... calls
210 __sync_single_inode() for $MFT -> do_writepages() -> write_page for
211 the dirty page containing the on-disk inode X, the page is now locked
212 -> ntfs_write_mst_block() which clears PageUptodate() on the page to
213 prevent anyone else getting hold of it whilst it does the write out.
214 This is necessary as the on-disk inode needs "fixups" applied before
215 the write to disk which are removed again after the write and
216 PageUptodate is then set again. It then analyses the page looking
217 for dirty on-disk inodes and when it finds one it calls
218 ntfs_may_write_mft_record() to see if it is safe to write this
219 on-disk inode. This then calls ilookup5() to check if the
220 corresponding VFS inode is in icache(). This in turn calls ifind()
221 which waits on the inode lock via wait_on_inode whilst holding the
222 global inode_lock.
223 Process 2: pdflush results in a call to __sync_single_inode for the
224 same VFS inode X on the ntfs volume. This locks the inode (I_LOCK)
225 then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
226 read_cache_page() for the page (in page cache of table of inodes
227 $MFT, inode 0) containing the on-disk inode. This page has
228 PageUptodate() clear because of Process 1 (see above) so
229 read_cache_page() blocks when it tries to take the page lock for the
230 page so it can call ntfs_read_page().
231 Thus Process 1 is holding the page lock on the page containing the
232 on-disk inode X and it is waiting on the inode X to be unlocked in
233 ifind() so it can write the page out and then unlock the page.
234 And Process 2 is holding the inode lock on inode X and is waiting for
235 the page to be unlocked so it can call ntfs_readpage() or discover
236 that Process 1 set PageUptodate() again and use the page.
237 Thus we have a deadlock due to ifind() waiting on the inode lock.
238 The solution: The fix is to use the newly introduced
239 ilookup5_nowait() which does not wait on the inode's lock and hence
240 avoids the deadlock. This is safe as we do not care about the VFS
241 inode and only use the fact that it is in the VFS inode cache and the
242 fact that the vfs and ntfs inodes are one struct in memory to find
243 the ntfs inode in memory if present. Also, the ntfs inode has its
244 own locking so it does not matter if the vfs inode is locked.
Anton Altaparmakove74589a2005-08-16 16:38:28 +0100245 - Fix bug in mft record writing where we forgot to set the device in
Anton Altaparmakov481d0372005-08-16 19:42:56 +0100246 the buffers when mapping them after the VM had discarded them.
Anton Altaparmakove74589a2005-08-16 16:38:28 +0100247 Thanks to Martin MOKREJÅ  for the bug report.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700248
2492.1.22 - Many bug and race fixes and error handling improvements.
250
251 - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
252 - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
253 instead of void and provide a helper ntfs_truncate_vfs() for the
254 vfs ->truncate method.
255 - Add a new ntfs inode flag NInoTruncateFailed() and modify
256 fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
257 - Fix min_size and max_size definitions in ATTR_DEF structure in
258 fs/ntfs/layout.h to be signed.
259 - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
260 ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
261 ntfs_attr_can_be_resident(), which in turn use the new private helper
262 ntfs_attr_find_in_attrdef().
263 - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
264 mapping->private_lock around the dirtying of the buffer heads
265 analagous to the way it is done in __set_page_dirty_buffers().
266 - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
267 mount time as this cannot work with the current implementation.
268 - Check for location of attribute name and improve error handling in
269 general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
270 - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
271 i_size, i.e. race with truncate, invalidate the buffers on the page
272 so that they become freeable and hence the page does not leak.
273 - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
274 Bunk)
275 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
276 a NULL pointer dereference in the error code path when a corrupt
277 attribute was found. (Thanks to Domen Puncer for the bug report.)
278 - Add MODULE_VERSION() to fs/ntfs/super.c.
279 - Make several functions and variables static. (Adrian Bunk)
280 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
281 buffers for the page if they are not present and then marks the
282 buffers belonging to the ntfs record dirty. This causes the buffers
283 to become busy and hence they are safe from removal until the page
284 has been written out.
285 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
286 error handling code path that resulted in a BUG() due to trying to
287 unmap an extent mft record when the mapping of it had failed and it
288 thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
289 - Drop the runlist lock after the vcn has been read in
290 fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
291 - Rewrite handling of multi sector transfer errors. We now do not set
292 PageError() when such errors are detected in the async i/o handler
293 fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
294 protected attributes now check the magic of each ntfs record as they
295 use it and act appropriately. This has the effect of making errors
296 granular per ntfs record rather than per page which solves the case
297 where we cannot access any of the ntfs records in a page when a
298 single one of them had an mst error. (Thanks to Ken MacFerrin for
299 the bug report.)
300 - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
301 where we failed to release i_sem on the $Quota/$Q attribute inode.
302 - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
303 - Add mapping of unmapped buffers to all remaining code paths, i.e.
304 fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
305 and write_mft_record_nolock(). From now on we require that the
306 complete runlist for the mft mirror is always mapped into memory.
307 - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
308 - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
309 - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
310 resident attribute will be smaller than a page which makes the code
311 simpler. Also make the code more tolerant to concurrent ->truncate.
312
3132.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
314
315 - Implement extent mft record deallocation
316 fs/ntfs/mft.c::ntfs_extent_mft_record_free().
317 - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
318 - Add vol->mft_data_pos and initialize it at mount time.
319 - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
320 ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
321 ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
322 ntfs_runlists_merge() and adapt all callers.
323 - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
324 ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
325 and ntfs_mapping_pairs_build(), adapted from libntfs.
326 - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
327 static and add a declaration for it to lcnalloc.h.
328 - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
329 inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
330 cluster bitmap lock for the duration of the call.
331 - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
332 - Implement the equivalent of memset() for an ntfs attribute in
333 fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
334 fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
335 - Remove unnecessary casts from LCN_* constants.
336 - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
337 - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
338 change MFT_RECORD to contain the NTFS 3.1+ specific fields.
339 - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
340 marks all buffers belonging to an ntfs record dirty, followed by
341 marking the page the ntfs record is in dirty and also marking the vfs
342 inode containing the ntfs record dirty (I_DIRTY_PAGES).
343 - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
344 new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
345 longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
346 - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
347 include errors.
348 - Move the typedefs for runlist_element and runlist from types.h to
349 runlist.h and fix resulting include errors.
350 - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
351 - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
352 mark_ntfs_record_dirty() which also changes the behaviour in that we
353 now set the buffers belonging to the mft record dirty as well as the
354 page itself.
355 - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
356 to cope with the fact that there now are dirty buffers in mft pages.
357 - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
358 mark_ntfs_record_dirty() and thus to set the buffers belonging to the
359 mft record dirty as well as the page itself.
360 - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
361 slightly modified by me)
362 - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
363 the mft record is already locked and otherwise behaves the same way
364 as fs/ntfs/mft.c::map_mft_record().
365 - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
366 writes the mft record if the buffers belonging to it are dirty.
367 Otherwise we assume that it was written out by other means already.
368 - Attempting to write outside initialized size is _not_ a bug so remove
369 the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
370 fact required to write outside initialized size when preparing to
371 extend the initialized size.
372 - Map the page instead of using page_address() before writing to it in
373 fs/ntfs/aops.c::ntfs_mft_writepage().
374 - Provide exclusion between opening an inode / mapping an mft record
375 and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
376 by setting the page not uptodate throughout ntfs_mft_writepage().
377 - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
378 to ensure noone can see the page whilst the mst fixups are applied.
379 - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
380 checks if an mft record may be written out safely obtaining any
381 necessary locks in the process. This is used by
382 fs/ntfs/aops.c::ntfs_write_mst_block().
383 - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
384 writing mft records and improve its error handling in the process.
385 Now if any of the records in the page fail to be written out, all
386 other records will be written out instead of aborting completely.
387 - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
388 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
389 ntfs_mst_aops for all inodes which are NInoMstProtected() and
390 ntfs_aops for all other inodes.
391 - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
392 ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
393 no longer require an ntfs inode to be present. Update all callers.
394 - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
395 - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
396 to ensure noone can see the page whilst the mst fixups are applied.
397 - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
398 fs/ntfs/mft.c::try_map_mft_record().
399 - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
400 with the ntfs inode which contains the page rather than the ntfs
401 inode the mft record of which is in the page.
402 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
403 index inode bitmap inode release code from there to
404 fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
405 Hellwig for spotting this.)
406 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
407 inode semaphore around the code that sets ni->itype.index.bmp_ino to
408 NULL and reorganize the code to optimize it a bit. (Thanks to
409 Christoph Hellwig for spotting this.)
410 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
411 ntfs inode as a parameter as this is confusing and misleading and the
412 needed ntfs inode is available via NTFS_I(page->mapping->host).
413 Adapt all callers to this change.
414 - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
415 fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
416 of the first buffer in a record and to take this as the ntfs record
417 dirty state. We cannot look at the dirty state for subsequent
418 buffers because we might be racing with
419 fs/ntfs/aops.c::mark_ntfs_record_dirty().
420 - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
421 inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
422 add a declaration for it to inode.h. Fix some compilation issues
423 that resulted due to #includes and header file interdependencies.
424 - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
425 - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
426 - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
427 record sequence number if it is specified (i.e. not zero).
428 - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
429 functions used by it.
430 - Update Documentation/filesystems/ntfs.txt with instructions on how to
431 use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
432 the linear raid problem with the Software RAID / MD driver when one
433 or more of the devices has an odd number of sectors.
434
4352.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
436
437 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
438 where we did not clear ctx->al_entry but it was still set due to
439 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
440 particular.
441 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
442 where we forgot to unmap the extent mft record when we had finished
443 enumerating an attribute which caused a bug check to trigger when the
444 VFS calls ->clear_inode.
445
4462.1.19 - Many cleanups, improvements, and a minor bug fix.
447
448 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
449 change the uid, gid, and mode of an inode as we do not support NTFS
450 ACLs yet.
451 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
452 kernel.
453 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
454 and ntfs_filldir() as per suggestion from Al Viro.
455 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
456 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
457 inode size has changed and to only output an error if so.
458 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
459 - Add le{16,32,64} as well as sle{16,32,64} data types to
460 fs/ntfs/types.h.
461 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
462 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
463 respectively, to fs/ntfs/types.h.
464 - Update endianness conversion macros in fs/ntfs/endian.h to use the
465 new types as appropriate.
466 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
467 and index.c.
468 - Add leMFT_REF data type to fs/ntfs/layout.h.
469 - Update all NTFS header files with the new little endian data types.
470 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
471 - Do proper type casting when using ntfs_is_*_recordp() in
472 fs/ntfs/logfile.c, mft.c, and super.c.
473 - Fix all the sparse bitwise warnings. Had to change all the typedef
474 enums storing little endian values to simple enums plus a typedef for
475 the datatype to make sparse happy.
476 - Fix a bug found by the new sparse bitwise warnings where the default
477 upcase table was defined as a pointer to wchar_t rather than ntfschar
478 in fs/ntfs/ntfs.h and super.c.
479 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
480
4812.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
482
483 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
484 the calculation of total/free inodes as used by statfs().
485 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
486 because the code itself is using the ntfs_lock semaphore which
487 provides safe locking. (Ingo Molnar)
488 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
489 could occur in the future for when we start closing/freeing extent
490 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
491 we free it.
492 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
493 find_external_attr() to ntfs_external_attr_find() to cleanup the
494 namespace a bit and to be more consistent with libntfs.
495 - Rename {{re,}init,get,put}_attr_search_ctx() to
496 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
497 attr_search_context to ntfs_attr_search_ctx.
498 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
499 for the attribute list attribute itself.
500 - Fix endianness bug in ntfs_external_attr_find().
501 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
502 if the attribute is not found, and -EIO on real error. In the case
503 of -ENOENT, the search context is updated to describe the attribute
504 before which the attribute being searched for would need to be
505 inserted if such an action were to be desired and in the case of
506 ntfs_external_attr_find() the search context is also updated to
507 indicate the attribute list entry before which the attribute list
508 entry of the attribute being searched for would need to be inserted
509 if such an action were to be desired. Also make ntfs_find_attr()
510 static and remove its prototype from attrib.h as it is not used
511 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
512 callers of ntfs_{external,}attr_{find,lookup}() for the new return
513 values.
514 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
515
5162.1.17 - Fix bugs in mount time error code paths and other updates.
517
518 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
519 includes functions to set/clear a single bit or a run of bits.
520 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
521 runlist element containing a particular vcn. It also takes care of
522 mapping any needed runlist fragments.
523 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
524 - Load attribute definition table from $AttrDef at mount time.
525 - Fix bugs in mount time error code paths involving (de)allocation of
526 the default and volume upcase tables.
527 - Remove ntfs_nr_mounts as it is no longer used.
528
5292.1.16 - Implement access time updates, file sync, async io, and read/writev.
530
531 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
532 This is done by setting the appropriate file operations pointers to
533 the generic helper functions provided by mm/filemap.c.
534 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
535 and directories (fs/ntfs/dir.c).
536 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
537 Note, except for the root directory and any other system files opened
538 by the user, the system files will not have their access times
539 updated as they are only accessed at the inode level an hence the
540 file level functions which cause the times to be updated are never
541 invoked.
542
5432.1.15 - Invalidate quotas when (re)mounting read-write.
544
545 - Add new element itype.index.collation_rule to the ntfs inode
546 structure and set it appropriately in ntfs_read_locked_inode().
547 - Implement a new inode type "index" to allow efficient access to the
548 indices found in various system files and adapt inode handling
549 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
550 attribute inode (NInoAttr() is true) with an attribute type of
551 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
552 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
553 there would be no way to distinguish between normal attribute inodes
554 and index inodes. The function to obtain an index inode is
555 ntfs_index_iget() and it uses the helper function
556 ntfs_read_locked_index_inode(). Note, we do not overload
557 ntfs_attr_iget() as indices consist of multiple attributes so using
558 ntfs_attr_iget() to obtain an index inode would be confusing.
559 - Ensure that there is no overflow when doing page->index <<
560 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
561 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
562 and ntfs_read_block().
563 - Use case sensitive attribute lookups instead of case insensitive ones.
564 - Lock all page cache pages belonging to mst protected attributes while
565 accessing them to ensure we never see corrupt data while the page is
566 under writeout.
567 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
568 We have ntfs_is_collation_rule_supported() to check if the collation
569 rule you want to use is supported and ntfs_collation() which actually
570 collates two data items. We currently only support COLLATION_BINARY
571 and COLLATION_NTOFS_ULONG but support for other collation rules will
572 be added as the need arises.
573 - Add a new type, ntfs_index_context, to allow retrieval of an index
574 entry using the corresponding index key. To get an index context,
575 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
576 This also adds a new slab cache for the index contexts. To lookup a
577 key in an index inode, use ntfs_index_lookup(). After modifying an
578 index entry, call ntfs_index_entry_flush_dcache_page() followed by
579 ntfs_index_entry_mark_dirty() to ensure the changes are written out
580 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
581 an index entry is in the index allocation attribute rather than the
582 index root attribute it will not be written out (you will get a
583 warning message about discarded changes instead).
584 - Load the quota file ($Quota) and check if quota tracking is enabled
585 and if so, mark the quotas out of date. This causes windows to
586 rescan the volume on boot and update all quota entries.
587 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
588 It is simply set to __set_page_dirty_nobuffers() to make sure that
589 running set_page_dirty() on a page containing mft/ntfs records will
590 not affect the dirty state of the page buffers.
591 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
592 buffers that are inside the ntfs record in the page dirty after which
593 it sets the page dirty. This allows ->writepage to only write the
594 dirty index records rather than having to write all the records in
595 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
596 use this rather than __set_page_dirty_nobuffers().
597 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
598 writing of page cache pages belonging to mst protected attributes
599 like the index allocation attribute in directory indices and other
600 indices like $Quota/$Q, etc. This means that the quota is now marked
601 out of date on all volumes rather than only on ones where the quota
602 defaults entry is in the index root attribute of the $Quota/$Q index.
603
6042.1.14 - Fix an NFSd caused deadlock reported by several users.
605
606 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
607 to a buffer so that we can put the search context and unmap the mft
608 record before calling the filldir() callback. We need to do this
609 because of NFSd which calls ->lookup() from its filldir callback()
610 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
611 of the directory and since ntfs_readdir() has got it mapped already
612 ntfs_lookup() deadlocks.
613
6142.1.13 - Enable overwriting of resident files and housekeeping of system files.
615
616 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
617 keeping the mft mirror in sync with the mft when mirrored mft records
618 are written. The functions are write_mft_record{,_nolock}(). The
619 implementation is quite rudimentary for now with lots of things not
620 implemented yet but I am not sure any of them can actually occur so
621 I will wait for people to hit each one and only then implement it.
622 - Commit open system inodes at umount time. This should make it
623 virtually impossible for sync_mft_mirror_umount() to ever be needed.
624 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
625 ntfs super operations. This gives us inode writing via the VFS inode
626 dirty code paths. Note: Access time updates are not implemented yet.
627 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
628 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
629 finally enabling resident file overwrite! (-8 This also includes a
630 placeholder for ->writepage (ntfs_mft_writepage()), which for now
631 just redirties the page and returns. Also, at umount time, we for
632 now throw away all mft data page cache pages after the last call to
633 ntfs_commit_inode() in the hope that all inodes will have been
634 written out by then and hence no dirty (meta)data will be lost. We
635 also check for this case and emit an error message telling the user
636 to run chkdsk.
637 - Use set_page_writeback() and end_page_writeback() in the resident
638 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
639 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
640 page is clean.
641 - Implement ntfs_mft_writepage() so it now checks if any of the mft
642 records in the page are dirty and if so redirties the page and
643 returns. Otherwise it just returns (after doing set_page_writeback(),
644 unlock_page(), end_page_writeback() or the radix-tree tag
645 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
646 alowing the VM to do with the page as it pleases. Also, at umount
647 time, now only throw away dirty mft (meta)data pages if dirty inodes
648 are present and ask the user to email us if they see this happening.
649 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
650 information flags (fs/ntfs/super.c).
651 - Mark the volume dirty when (re)mounting read-write and mark it clean
652 when unmounting or remounting read-only. If any volume errors are
653 found, the volume is left marked dirty to force chkdsk to run.
654 - Add code to set the NT4 compatibility flag when (re)mounting
655 read-write for newer NTFS versions but leave it commented out for now
656 since we do not make any modifications that are NTFS 1.2 specific yet
657 and since setting this flag breaks Captive-NTFS which is not nice.
658 This code must be enabled once we start writing NTFS 1.2 specific
659 changes otherwise Windows NTFS driver might crash / cause corruption.
660
6612.1.12 - Fix the second fix to the decompression engine and some cleanups.
662
663 - Add a new address space operations struct, ntfs_mst_aops, for mst
664 protected attributes. This is because the default ntfs_aops do not
665 make sense with mst protected data and were they to write anything to
666 such an attribute they would cause data corruption so we provide
667 ntfs_mst_aops which does not have any write related operations set.
668 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
669 includes an adapted ntfs_commit_inode() and an implementation of
670 ntfs_write_inode() which for now just cleans dirty inodes without
671 writing them (it does emit a warning that this is happening).
672 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
673 entry) as it was only fixing a theoretical bug but at the same time
674 it badly broke the handling of sparse and uncompressed compression
675 blocks.
676
6772.1.11 - Driver internal cleanups.
678
679 - Only build logfile.o if building the driver with read-write support.
680 - Really final white space cleanups.
681 - Use generic_ffs() instead of ffs() in logfile.c which allows the
682 log_page_size variable to be optimized by gcc into a constant.
683 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
684 char as defined by POSIX and as found on some systems.
685
6862.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
687
688 - Finish off the white space cleanups (remove trailing spaces, etc).
689 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
690 the kludges around the first iget(). Instead of (re)setting ->s_op
691 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
692 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
693 kills the need for second super_operations and allows to return error
694 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
695 tricks. (Al Viro)
696 - Force read-only (re)mounting if any of the following bits are set in
697 the volume information flags:
698 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
699 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
700 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
701 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
702 above bits set so the test is made easy.
703
7042.1.9 - Fix two bugs in decompression engine.
705
706 - Fix a bug where we would not always detect that we have reached the
707 end of a compression block because we were ending at minus one byte
708 which is effectively the same as being at the end. The fix is to
709 check whether the uncompressed buffer has been fully filled and if so
710 we assume we have reached the end of the compression block. A big
711 thank you to Marcin Gibuła for the bug report, the assistance in
712 tracking down the bug and testing the fix.
713 - Fix a possible bug where when a compressed read is truncated to the
714 end of the file, the offset inside the last page was not truncated.
715
7162.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
717
718 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
719 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
720 utc2ntfs() to work with struct timespec instead of time_t on the
721 Linux UTC time side thus preserving the full precision of the NTFS
722 time and only loosing up to 99 nano-seconds in the Linux UTC time.
723 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
724 static inline.
725 - Remove unused ntfs_dirty_inode().
726 - Cleanup super operations declaration in fs/ntfs/super.c.
727 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
728 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
729 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
730 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
731 fs/ntfs/inode.h so they can be used elsewhere.
732 - Determine the mft mirror size as the number of mirrored mft records
733 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
734 - Load the mft mirror at mount time and compare the mft records stored
735 in it to the ones in the mft. Force a read-only mount if the two do
736 not match (fs/ntfs/super.c).
737 - Fix type casting related warnings on 64-bit architectures. Thanks
738 to Meelis Roos for reporting them.
739 - Move %L to %ll as %L is floating point and %ll is integer which is
740 what we want.
741 - Read the journal ($LogFile) and determine if the volume has been
742 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
743 and fs/ntfs/logfile.c). This is a little bit of a crude check in
744 that we only look at the restart areas and not at the actual log
745 records so that there will be a very small number of cases where we
746 think that a volume is dirty when in fact it is clean. This should
747 only affect volumes that have not been shutdown cleanly and did not
748 have any pending, non-check-pointed i/o.
749 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
750 is requested, empty $LogFile by overwriting it with 0xff bytes to
751 ensure that Windows cannot cause data corruption by replaying a stale
752 journal after Linux has written to the volume.
753
7542.1.7 - Enable NFS exporting of mounted NTFS volumes.
755
756 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
757 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
758 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
759 default doesn't allow inode number 0 which is a valid inode on NTFS
760 and even if it did allow that it uses iget() instead of ntfs_iget()
761 which makes it useless for us.
762 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
763 default just returns -EACCES which is not very useful.
764 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
765 and set them up in the super block at mount time (super.c) this
766 allows mounted NTFS volumes to be exported via NFS.
767 - Add missing return -EOPNOTSUPP; in
768 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
769 - Enforce no atime and no dir atime updates at mount/remount time as
770 they are not implemented yet anyway.
771 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
772 after a NULL check. Thanks to Dave Jones for pointing this out.
773
7742.1.6 - Fix minor bug in handling of compressed directories.
775
776 - Fix bug in handling of compressed directories. A compressed
777 directory is not really compressed so when we set the ->i_blocks
778 field of a compressed directory inode we were setting it from the
779 non-existing field ni->itype.compressed.size which gave random
780 results... For directories we now always use ni->allocated_size.
781
7822.1.5 - Fix minor bug in attribute list attribute handling.
783
784 - Fix bug in attribute list handling. Actually it is not as much a bug
785 as too much protection in that we were not allowing attribute lists
786 which waste space on disk while Windows XP clearly allows it and in
787 fact creates such attribute lists so our driver was failing.
788 - Update NTFS documentation ready for 2.6 kernel release.
789
7902.1.4 - Reduce compiler requirements.
791
792 - Remove all uses of unnamed structs and unions in the driver to make
793 old and newer gcc versions happy. Makes it a bit uglier IMO but at
794 least people will stop hassling me about it.
795
7962.1.3 - Important bug fixes in corner cases.
797
798 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
799 clusters. (Philipp Thomas)
800 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
801 multiple of the block_size but not the cluster size. (Szabolcs
802 Szakacsits <szaka@sienet.hu>)
803
8042.1.2 - Important bug fixes aleviating the hangs in statfs.
805
806 - Fix buggy free cluster and free inode determination logic.
807
8082.1.1 - Minor updates.
809
810 - Add handling for initialized_size != data_size in compressed files.
811 - Reduce function local stack usage from 0x3d4 bytes to just noise in
812 fs/ntfs/upcase.c. (Randy Dunlap <rddunlap@osdl.ord>)
813 - Remove compiler warnings for newer gcc.
814 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
815 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
816 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
817 kmap_atomic(KM_USER0).
818
8192.1.0 - First steps towards write support: implement file overwrite.
820
821 - Add configuration option for developmental write support with an
822 appropriately scary configuration help text.
823 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
824 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
825 overwriting of existing files on ntfs. Note: Resident files are
826 only written into memory, and not written out to disk at present, so
827 avoid writing to files smaller than about 1kiB.
828 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
829 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
830 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
831 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
832 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
833 This enables write(2) based overwriting of existing files on ntfs.
834 Note: As with mmap(2) based overwriting, resident files are only
835 written into memory, and not written out to disk at present, so avoid
836 writing to files smaller than about 1kiB.
837 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
838 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
839 files with the purpose of intercepting and aborting all i_size
840 changes which we do not support yet. ntfs_truncate() actually only
841 emits a warning message but AFAICS our interception of i_size changes
842 elsewhere means ntfs_truncate() never gets called for i_size changes.
843 It is only called from generic_file_write() when we fail in
844 ntfs_prepare_{,nonresident_}write() in order to discard any
845 instantiated buffers beyond i_size. Thus i_size is not actually
846 changed so our warning message is enough. Unfortunately it is not
847 possible to easily determine if i_size is being changed or not hence
848 we just emit an appropriately worded error message.
849
8502.0.25 - Small bug fixes and cleanups.
851
852 - Unlock the page in an out of memory error code path in
853 fs/ntfs/aops.c::ntfs_read_block().
854 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
855 just unlock the page and return. (This can happen due to ->writepage
856 clearing PageUptodate() during write out of MstProtected()
857 attributes.
858 - Remove leaked write code again.
859
8602.0.24 - Cleanups.
861
862 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
863 inside BUG_ON(). (Adam J. Richter)
864 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
865 calls for improved debugging. (Adam J. Richter)
866 - Add errors flag to the ntfs volume state, accessed via
867 NVol{,Set,Clear}Errors(vol).
868 - Do not allow read-write remounts of read-only volumes with errors.
869 - Clarify comment for ntfs file operation sendfile which was added by
870 Christoph Hellwig a while ago (just using generic_file_sendfile())
871 to say that ntfs ->sendfile is only used for the case where the
872 source data is on the ntfs partition and the destination is
873 somewhere else, i.e. nothing we need to concern ourselves with.
874 - Add generic_file_write() as our ntfs file write operation.
875
8762.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
877
878 - Massive internal locking changes to mft record locking. Fixes lock
879 recursion and replaces the mrec_lock read/write semaphore with a
880 mutex. Also removes the now superfluous mft_count. This fixes several
881 race conditions and deadlocks, especially in the future write code.
882 - Fix ntfs over loopback for compressed files by adding an
883 optimization barrier. (gcc was screwing up otherwise ?)
884 - Miscellaneous cleanups all over the code and a fix or two in error
885 handling code paths.
886 Thanks go to Christoph Hellwig for pointing out the following two:
887 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
888 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
889
8902.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
891
892 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
893 at entry/exit respectively.
894 - Use C99 initializers for structures.
895 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
896
8972.0.21 - Check for, and refuse to work with too large files/directories/volumes.
898
899 - Limit volume size at mount time to 2TiB on architectures where
900 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
901 This is the most we can do without overflowing the 32-bit limit of
902 the block device size imposed on us by sb_bread() and sb_getblk()
903 for the time being.
904 - Limit file/directory size at open() time to 16TiB on architectures
905 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
906 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
907 overflowing the page cache page index.
908
9092.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
910
911 - Move the directory index bitmap to use an attribute inode instead of
912 having special fields for it inside the ntfs inode structure. This
913 means that the index bitmaps now use the page cache for i/o, too,
914 and also as a side effect we get support for non-resident index
915 bitmaps for free.
916 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
917 fix a page leak that manifested itself in some cases.
918 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
919 index bitmap inode on the final iput().
920
9212.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
922
923 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
924 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
925 - Drop the "file" from ntfs_file_read_compressed_block().
926 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
927 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
928 - Update ntfs_end_buffer_async_read() with the improved logic from
929 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
930 further logic improvements to better determine when we set PageError.
931 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
932 check for the buffers being uptodate first in line with the updated
933 fs/buffer.c::block_read_full_page(). This plugs a small race
934 condition.
935
9362.0.18 - Fix race condition in reading of compressed files.
937
938 - There was a narrow window between checking a buffer head for being
939 uptodate and locking it in ntfs_file_read_compressed_block(). We now
940 lock the buffer and then check whether it is uptodate or not.
941
9422.0.17 - Cleanups and optimizations - shrinking the ToDo list.
943
944 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
945 code and update callers, i.e. ntfs_iget(), to pass that error code
946 up instead of just using -EIO.
947 - Modifications to super.c to ensure that both mount and remount
948 cannot set any write related options when the driver is compiled
949 read-only.
950 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
951 cache the current runlist element. This should improve performance
952 when reading very large and/or very fragmented data.
953
9542.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
955
956 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
957 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
958 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
959 - Convert $MFT/$BITMAP access to attribute inode API and remove all
960 remnants of the ugly mftbmp address space and operations hack. This
961 means we finally have only one readpage function as well as only one
962 async io completion handler. Yey! The mft bitmap is now just an
963 attribute inode and is accessed from vol->mftbmp_ino just as if it
964 were a normal file. Fake inodes rule. (-:
965
9662.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
967
968 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
969 remounts to fail when the partition had an entry in /etc/fstab and
970 the entry specified the nls= option.
971 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
972 expand all the helper functions NVolFoo(), NVolSetFoo(), and
973 NVolClearFoo().
974 - Move copyright statement from driver initialisation message to
975 module description (fs/super.c). This makes the initialisation
976 message fit on one line and fits in better with rest of kernel.
977 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
978 attribute inodes, and both for files and directories.
979 - Implement fake attribute inodes allowing all attribute i/o to go via
980 the page cache and to use all the normal vfs/mm functionality:
981 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
982 to fs/ntfs/inode.c.
983 - Add needed cleanup code to ntfs_clear_big_inode().
984 - Merge address space operations for files and directories (aops.c),
985 now just have ntfs_aops:
986 - Rename:
987 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
988 ntfs_attr_read_block() -> ntfs_read_block(),
989 ntfs_file_read_page() -> ntfs_readpage().
990 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
991 attribute inodes, and both for files and directories.
992 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
993
9942.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
995
996 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
997 the locking out of super.c::get_nr_free_mft_records() and taking and
998 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
999 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
1000 current userspace ntfs library code. This means that if a merge
1001 fails the original runlists are always left unmodified instead of
1002 being silently corrupted.
1003 - Misc typo fixes.
1004
10052.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
1006
1007 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
1008 from ntfs_volume structure.
1009 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
1010 from ntfs_volume structure.
1011 - Use iget5_locked() and friends instead of conventional iget(). Wrap
1012 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
1013 to use ntfs_iget(). Leave only one iget() call at mount time so we
1014 don't need an ntfs_iget_mount().
1015 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
1016 additional argument.
1017
10182.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
1019
1020 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
1021 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
1022 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
1023 to determine whether to apply mst fixups or not.
1024 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
1025 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
1026 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
1027 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
1028 the VFS readpage function prototype to the ntfs_attr_read_block()
1029 function prototype.
1030
10312.0.11 - Initial preparations for fake inode based attribute i/o.
1032
1033 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
1034 do some macro magic (adapted from include/linux/buffer_head.h) to
1035 expand all the helper functions NInoFoo(), NInoSetFoo(), and
1036 NInoClearFoo().
1037 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
1038 - Add new fields to ntfs_inode structure to allow use of fake inodes
1039 for attribute i/o: type, name, name_len. Also add new state bits:
1040 NI_Attr, which, if set, indicates the inode is a fake inode, and
1041 NI_MstProtected, which, if set, indicates the attribute uses multi
1042 sector transfer protection, i.e. fixups need to be applied after
1043 reads and before/after writes.
1044 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
1045 ntfs_{new,clear,destroy}_extent_inode() and update callers.
1046 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
1047 instead of ntfs_destroy_extent_inode().
1048 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
1049 - Make all operations on ntfs inode state bits use the NIno* functions.
1050 - Set up the new ntfs inode fields and state bits in
1051 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
1052 allocated memory to __ntfs_clear_inode().
1053 - Cleanup ntfs_inode structure a bit for better ordering of elements
1054 w.r.t. their size to allow better packing of the structure in memory.
1055
10562.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
1057
1058 - Add check at mount time to verify that the number of inodes on the
1059 volume does not exceed 2^32 - 1, which is the maximum allowed for
1060 NTFS according to Microsoft.
1061 - Change mft_no member of ntfs_inode structure to be unsigned long.
1062 Update all users. This makes ntfs_inode->mft_no just a copy of struct
1063 inode->i_ino. But we can't just always use struct inode->i_ino and
1064 remove mft_no because extent inodes do not have an attached struct
1065 inode.
1066
10672.0.9 - Decompression engine now uses a single buffer and other cleanups.
1068
1069 - Change decompression engine to use a single buffer protected by a
1070 spin lock instead of per-CPU buffers. (Rusty Russell)
1071 - Do not update cb_pos when handling a partial final page during
1072 decompression of a sparse compression block, as the value is later
1073 reset without being read/used. (Rusty Russell)
1074 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
1075 Morton)
1076 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
1077 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
1078 it also makes everything safer so it is a good thing.
1079 - Miscellaneous minor cleanups to comments.
1080
10812.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
1082
1083 Big thanks go to Al Viro and other inhabitants of #kernel for investing
1084 their time to discuss the case sensitivity and dcache aliasing issues.
1085
1086 - Remove unused source file fs/ntfs/attraops.c.
1087 - Remove show_inodes mount option(s), thus dropping support for
1088 displaying of short file names.
1089 - Remove deprecated mount option posix.
1090 - Restore show_sys_files mount option.
1091 - Add new mount option case_sensitive, to determine if the driver
1092 treats file names as case sensitive or not. If case sensitive, create
1093 file names in the POSIX namespace. Otherwise create file names in the
1094 LONG/WIN32 namespace. Note, files remain accessible via their short
1095 file name, if it exists.
1096 - Remove really dumb logic bug in boot sector recovery code.
1097 - Fix dcache aliasing issues wrt short/long file names via changes
1098 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
1099 fs/ntfs/namei.c::ntfs_lookup():
1100 - Add additional argument to ntfs_lookup_inode_by_name() in which we
1101 return information about the matching file name if the case is not
1102 matching or the match is a short file name. See comments above the
1103 function definition for details.
1104 - Change ntfs_lookup() to only create dcache entries for the correctly
1105 cased file name and only for the WIN32 namespace counterpart of DOS
1106 namespace file names. This ensures we have only one dentry per
1107 directory and also removes all dcache aliasing issues between short
1108 and long file names once we add write support. See comments above
1109 function for details.
1110 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
1111
11122.0.7 - Minor cleanups and updates for changes in core kernel code.
1113
1114 - Remove much of the NULL struct element initializers.
1115 - Various updates to make compatible with recent kernels.
1116 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
1117 in fs/ntfs/ntfs.h instead.
1118 - Remove no longer needed KERNEL_VERSION checks. We are now in the
1119 kernel proper so they are no longer needed.
1120
11212.0.6 - Major bugfix to make compatible with other kernel changes.
1122
1123 - Initialize the mftbmp address space properly now that there are more
1124 fields in the struct address_space. This was leading to hangs and
1125 oopses on umount since 2.5.12 because of changes to other parts of
1126 the kernel. We probably want a kernel generic init_address_space()
1127 function...
1128 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
1129 only caller of ->readdir() is vfs_readdir() which holds i_sem during
1130 the call, and i_sem is sufficient protection against changes in the
1131 directory inode (including ->i_size).
1132 - Use generic_file_llseek() for directories (as opposed to
1133 default_llseek()) as this downs i_sem instead of the BKL which is
1134 what we now need for exclusion against ->f_pos changes considering we
1135 no longer take the BKL in ntfs_readdir().
1136
11372.0.5 - Major bugfix. Buffer overflow in extent inode handling.
1138
1139 - No need to set old blocksize in super.c::ntfs_fill_super() as the
1140 VFS does so via invocation of deactivate_super() calling
1141 fs->fill_super() calling block_kill_super() which does it.
1142 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
1143 -> Do we really need it? I don't think so as we have exclusion on
1144 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
1145 move the ->f_pos accesses under the mrec_lock though. Check this...
1146 - Fix really, really, really stupid buffer overflow in extent inode
1147 handling in mft.c::map_extent_mft_record().
1148
11492.0.4 - Cleanups and updates for kernel 2.5.11.
1150
1151 - Add documentation on how to use the MD driver to be able to use NTFS
1152 stripe and volume sets in Linux and generally cleanup documentation
1153 a bit.
1154 Remove all uses of kdev_t in favour of struct block_device *:
1155 - Change compress.c::ntfs_file_read_compressed_block() to use
1156 sb_getblk() instead of getblk().
1157 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
1158 of get_hardsect_size().
1159 - No need to get old blocksize in super.c::ntfs_fill_super() as
1160 fs/super.c::get_sb_bdev() already does this.
1161 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
1162
11632.0.3 - Small bug fixes, cleanups, and performance improvements.
1164
1165 - Remove some dead code from mft.c.
1166 - Optimize readpage and read_block functions throughout aops.c so that
1167 only initialized blocks are read. Non-initialized ones have their
1168 buffer head mapped, zeroed, and set up to date, without scheduling
1169 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
1170 Thanks go to Andrew Morton for spotting the below:
1171 - Fix buglet in allocate_compression_buffers() error code path.
1172 - Call flush_dcache_page() after modifying page cache page contents in
1173 ntfs_file_readpage().
1174 - Check for existence of page buffers throughout aops.c before calling
1175 create_empty_buffers(). This happens when an I/O error occurs and the
1176 read is retried. (It also happens once writing is implemented so that
1177 needed doing anyway but I had left it for later...)
1178 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
1179 readpage and read_block functions. Reasoning same as above (i.e. I/O
1180 error retries and future write code paths.)
1181
11822.0.2 - Minor updates and cleanups.
1183
1184 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
1185 and cleanup the code a bit, removing the unused size parameter.
1186 - Change default fmask to 0177 and update documentation.
1187 - Change attrib.c::get_attr_search_ctx() to return the search context
1188 directly instead of taking the address of a pointer. A return value
1189 of NULL means the allocation failed. Updated all callers
1190 appropriately.
1191 - Update to 2.5.9 kernel (preserving backwards compatibility) by
1192 replacing all occurences of page->buffers with page_buffers(page).
1193 - Fix minor bugs in runlist merging, also minor cleanup.
1194 - Updates to bootsector layout and mft mirror contents descriptions.
1195 - Small bug fix in error detection in unistr.c and some cleanups.
1196 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
1197 bytes.
1198
11992.0.1 - Minor updates.
1200
1201 - Make default umask correspond to documentation.
1202 - Improve documentation.
1203 - Set default mode to include execute bit. The {u,f,d}mask can be used
1204 to take it away if desired. This allows binaries to be executed from
1205 a mounted ntfs partition.
1206
12072.0.0 - New version number. Remove TNG from the name. Now in the kernel.
1208
1209 - Add kill_super, just keeping up with the vfs changes in the kernel.
1210 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
1211 from the CVS import into BitKeeper.
1212 - Begin to implement proper handling of allocated_size vs
1213 initialized_size vs data_size (i.e. i_size). Done are
1214 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
1215 and attrib.c::load_attribute_list().
1216 - Lock the runlist in attrib.c::load_attribute_list() while using it.
1217 - Fix memory leak in ntfs_file_read_compressed_block() and generally
1218 clean up compress.c a little, removing some uncommented/unused debug
1219 code.
1220 - Tidy up dir.c a little bit.
1221 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
1222 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
1223 creating aops.c::ntfs_mst_readpage(), improving the handling of
1224 holes and overflow in the process and implementing the correct
1225 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
1226 I am aiming for correctness at the moment. Modularisation can come
1227 later.
1228 - Rename aops.c::end_buffer_read_index_async() to
1229 end_buffer_read_mst_async() and optimize the overflow checking and
1230 handling.
1231 - Use the host of the mftbmp address space mapping to hold the ntfs
1232 volume. This is needed so the async i/o completion handler can
1233 retrieve a pointer to the volume. Hopefully this will not cause
1234 problems elsewhere in the kernel... Otherwise will need to use a
1235 fake inode.
1236 - Complete implementation of proper handling of allocated_size vs
1237 initialized_size vs data_size (i.e. i_size) in whole driver.
1238 Basically aops.c is now completely rewritten.
1239 - Change NTFS driver name to just NTFS and set version number to 2.0.0
1240 to make a clear distinction from the old driver which is still on
1241 version 1.1.22.
1242
1243tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
1244
1245 - Replace bdevname(sb->s_dev) with sb->s_id.
1246 - Remove now superfluous new-line characters in all callers of
1247 ntfs_debug().
1248 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
1249 directories. Without this the "find" utility gets very upset which is
1250 fair enough as Linux/Unix do not support directory hard links.
1251 - Further runlist merging work. (Richard Russon)
1252 - Backwards compatibility for gcc-2.95. (Richard Russon)
1253 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001254 - Convert to new filesystem declaration using ->ntfs_get_sb() and
Linus Torvalds1da177e2005-04-16 15:20:36 -07001255 replacing ntfs_read_super() with ntfs_fill_super().
1256 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
1257 overflow on 32-bit architectures.
1258 - Cleanup upcase loading code to use ntfs_(un)map_page().
1259 - Disable/reenable preemtion in critical sections of compession engine.
1260 - Replace device size determination in ntfs_fill_super() with
1261 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
1262 function super.c::get_nr_blocks().
1263 - Implement a mount time option (show_inodes) allowing choice of which
1264 types of inode names readdir() returns and modify ntfs_filldir()
1265 accordingly. There are several parameters to show_inodes:
1266 system: system files
1267 win32: long file names (including POSIX file names) [DEFAULT]
1268 long: same as win32
1269 dos: short file names only (excluding POSIX file names)
1270 short: same as dos
1271 posix: same as both win32 and dos
1272 all: all file names
1273 Note that the options are additive, i.e. specifying:
1274 -o show_inodes=system,show_inodes=win32,show_inodes=dos
1275 is the same as specifying:
1276 -o show_inodes=all
1277 Note that the "posix" and "all" options will show all directory
1278 names, BUT the link count on each directory inode entry is set to 1,
1279 due to Linux not supporting directory hard links. This may well
1280 confuse some userspace applications, since the directory names will
1281 have the same inode numbers. Thus it is NOT advisable to use the
1282 "posix" or "all" options. We provide them only for completeness sake.
1283 - Add copies of allocated_size, initialized_size, and compressed_size to
1284 the ntfs inode structure and set them up in
1285 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
1286 for files and the index allocation attribute for directories.
1287 - Add copies of allocated_size and initialized_size to ntfs inode for
1288 $BITMAP attribute of large directories and set them up in
1289 inode.c::ntfs_read_inode().
1290 - Add copies of allocated_size and initialized_size to ntfs volume for
1291 $BITMAP attribute of $MFT and set them up in
1292 super.c::load_system_files().
1293 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
1294 posix, and utf8) and tell user what the new options to use are. Note
1295 we still do support them but they will be removed with kernel 2.7.x.
1296 - Change all occurences of integer long long printf formatting to hex
1297 as printk() will not support long long integer format if/when the
1298 div64 patch goes into the kernel.
1299 - Make slab caches have stable names and change the names to what they
1300 were intended to be. These changes are required/made possible by the
1301 new slab cache name handling which removes the length limitation by
1302 requiring the caller of kmem_cache_create() to supply a stable name
1303 which is then referenced but not copied.
1304 - Rename run_list structure to run_list_element and create a new
1305 run_list structure containing a pointer to a run_list_element
1306 structure and a read/write semaphore. Adapt all users of runlists
1307 to new scheme and take and release the lock as needed. This fixes a
1308 nasty race as the run_list changes even when inodes are locked for
1309 reading and even when the inode isn't locked at all, so we really
1310 needed the serialization. We use a semaphore rather than a spinlock
1311 as memory allocations can sleep and doing everything GFP_ATOMIC
1312 would be silly.
1313 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
1314 This can never happen due to the nature of lookup_attr() and how we
1315 support attribute lists. If it did happen it would imply the inode
1316 being corrupt.
1317 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
1318 bad if found.
1319 - Update to 2.5.6-pre2 changes in struct address_space.
1320 - Use parent_ino() when accessing d_parent inode number in dir.c.
1321 - Import Sourceforge CVS repository into BitKeeper repository:
1322 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
1323 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
1324 Documentation/filesystems/ntfs.txt for NTFS TNG.
1325 - Create kernel configuration option controlling whether debugging
1326 is enabled or not.
1327 - Add the required export of end_buffer_io_sync() from the patches
1328 directory to the kernel code.
1329 - Update inode.c::ntfs_show_options() with show_inodes mount option.
1330 - Update errors mount option.
1331
1332tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
1333
1334 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
1335 bug in mapping of extent inodes. Update all the comments to fit all
1336 the recent code changes.
1337 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
1338 - Cleanups in compress.c, mostly comments and folding help.
1339 - Implement attrib.c::map_run_list() as a generic helper.
1340 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
1341 thus making code shorter and enabling attribute list support.
1342 - Cleanup incorrect use of [su]64 with %L printf format specifier in
1343 all source files. Type casts to [unsigned] long long added to correct
1344 the mismatches (important for architectures which have long long not
1345 being 64 bits).
1346 - Merge async io completion handlers for directory indexes and $MFT
1347 data into one by setting the index_block_size{_bits} of the ntfs
1348 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
1349 - Cleanup aops.c, update comments.
1350 - Make ntfs_file_get_block() use map_run_list() so all files now
1351 support attribute lists.
1352 - Make ntfs_dir_readpage() almost verbatim copy of
1353 block_read_full_page() by using ntfs_file_get_block() with only real
1354 difference being the use of our own async io completion handler
1355 rather than the default one, thus reducing the amount of code and
1356 automatically enabling attribute list support for directory indices.
1357 - Fix bug in load_attribute_list() - forgot to call brelse in error
1358 code path.
1359 - Change parameters to find_attr() and lookup_attr(). We no longer
1360 pass in the upcase table and its length. These can be gotten from
1361 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
1362 - Cleanups in attrib.c.
1363 - Implement merging of runlists, attrib.c::merge_run_lists() and its
1364 helpers. (Richard Russon)
1365 - Attribute lists part 2, attribute extents and multi part runlists:
1366 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
1367 further runlist parts via attrib.c::map_run_list().
1368 - Tiny endianness bug fix in decompress_mapping_pairs().
1369
1370tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
1371
1372 - Enable encrypted directories. (Their index root is marked encrypted
1373 to indicate that new files in that directory should be created
1374 encrypted.)
1375 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
1376 - Enable $Extend system directory. Most (if not all) extended system
1377 files do not have unnamed data attributes so ntfs_read_inode() had to
1378 special case them but that is ok, as the special casing recovery
1379 happens inside an error code path so there is zero slow down in the
1380 normal fast path. The special casing is done by introducing a new
1381 function inode.c::ntfs_is_extended_system_file() which checks if any
1382 of the hard links in the inode point to $Extend as being their parent
1383 directory and if they do we assume this is an extended system file.
1384 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
1385 when compiled with -DDEBUG. Default is debug messages to be disabled.
1386 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
1387 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
1388 interface is enabled). Inspired by old ntfs driver.
1389 - Add debug_msgs insmod/kernel boot parameter to set whether debug
1390 messages are {dis,en}abled. This is useful to enable debug messages
1391 during ntfs initialization and is the only way to activate debugging
1392 when the sysctl interface is not enabled.
1393 - Cleanup debug output in various places.
1394 - Remove all dollar signs ($) from the source (except comments) to
1395 enable compilation on architectures whose gcc compiler does not
1396 support dollar signs in the names of variables/constants. Attribute
1397 types now start with AT_ instead of $ and $I30 is now just I30.
1398 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
1399 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
1400 access functions. This means we now cope with $MFT/$BITMAP being
1401 spread accross several mft records.
1402 - Disable modification of mft_zone_multiplier on remount. We can always
1403 reenable this later on if we really want to, but we will need to make
1404 sure we readjust the mft_zone size / layout accordingly.
1405
1406tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
1407
1408 - Use sb_set_blocksize() instead of set_blocksize() and verify the
1409 return value.
1410 - Use sb_bread() instead of bread() throughout.
1411 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1412 of a directory index block vcn. Apply resulting simplifications in
1413 dir.c everywhere.
1414 - Fix a small bug somewhere (but forgot what it was).
1415 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1416 on the printf-format parameter list and fix bugs reported by gcc
1417 as a result. (Richard Russon)
1418 - Move inode allocation strategy to Al's new stuff but maintain the
1419 divorce of ntfs_inode from struct inode. To achieve this we have two
1420 separate slab caches, one for big ntfs inodes containing a struct
1421 inode and pure ntfs inodes and at the same time fix some faulty
1422 error code paths in ntfs_read_inode().
1423 - Show mount options in proc (inode.c::ntfs_show_options()).
1424
1425tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1426
1427 - Modified (un)map_mft_record functions to be common for read and write
1428 case. To specify which is which, added extra parameter at front of
1429 parameter list. Pass either READ or WRITE to this, each has the
1430 obvious meaning.
1431 - General cleanups to allow for easier folding in vi.
1432 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1433 argument, and invokes attrib.c::merge_run_lists() to merge the old
1434 and the new runlists.
1435 - Removed attrib.c::find_first_attr().
1436 - Implemented loading of attribute list and complete runlist for $MFT.
1437 This means we now cope with $MFT being spread across several mft
1438 records.
1439 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1440 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1441 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1442 cache. There are too little ntfs_volume structures at any one time
1443 to justify a private slab cache.
1444 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1445 Use KM_BIO_IRQ on advice from IRC/kernel...
1446 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1447 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1448 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1449 removed the now superfluous exports from the kernel core patch.
1450 - Fix a bug where kfree() was used insted of ntfs_free().
1451 - Change map_mft_record() to take ntfs_inode as argument instead of
1452 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1453 - Add pointer to ntfs_volume to ntfs_inode.
1454 - Add mft record number and sequence number to ntfs_inode. Stop using
1455 i_ino and i_generation for in-driver purposes.
1456 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1457 - Remove use of proper inodes by extent inodes. Move i_ino and
1458 i_generation to ntfs_inode to do this. Apply simplifications that
1459 result and remove iget_no_wait(), etc.
1460 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1461 - Add reference counting in ntfs_inode for the ntfs inode itself and
1462 for the mapped mft record.
1463 - Extend mft record mapping so we can (un)map extent mft records (new
1464 functions (un)map_extent_mft_record), and so mappings are reference
1465 counted and don't have to happen twice if already mapped - just ref
1466 count increases.
1467 - Add -o iocharset as alias to -o nls for backwards compatibility.
1468 - The latest core patch is now tiny. In fact just a single additional
1469 export is necessary over the base kernel.
1470
1471tng-0.0.3 - Cleanups, enhancements, bug fixes.
1472
1473 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1474 and setup the runlist appropriately using knowledge provided by the
1475 sizes in the base attribute record.
1476 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1477 any more.
1478 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1479 page or use vmalloc depending on the amount of memory requested.
1480 - Cleanup error output. The __FUNCTION__ "(): " is now added
1481 automatically. Introduced a new header file debug.h to support this
1482 and also moved ntfs_debug() function into it.
1483 - Make reading of compressed files more intelligent and especially get
1484 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1485 buffers (allocated at first mount with cluster size <= 4kiB and
1486 deallocated on last umount with cluster size <= 4kiB), and
1487 asynchronous io for the compressed data using a list of buffer heads.
1488 Er, we use synchronous io as async io only works on whole pages
1489 covered by buffers and not on individual buffer heads...
1490 - Bug fix for reading compressed files with sparse compression blocks.
1491
1492tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1493
1494 - Fixed handling of directories when cluster size exceeds index block
1495 size.
1496 - Hide DOS only name space directory entries from readdir() but allow
1497 them in lookup(). This should fix the problem that Linux doesn't
1498 support directory hard links, while still allowing access to entries
1499 via their short file name. This also has the benefit of mimicking
1500 what Windows users are used to, so it is the ideal solution.
1501 - Implemented sync_page everywhere so no more hangs in D state when
1502 waiting for a page.
1503 - Stop using bforget() in favour of brelse().
1504 - Stop locking buffers unnecessarily.
1505 - Implemented compressed files (inode->mapping contains uncompressed
1506 data, raw compressed data is currently bread() into a vmalloc()ed
1507 memory buffer).
1508 - Enable compressed directories. (Their index root is marked compressed
1509 to indicate that new files in that directory should be created
1510 compressed.)
1511 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1512 functions. (Thanks to Will Dyson for pointing this out.)
1513 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1514 ntfs_sb_info) out of the common inode and super_block structures and
1515 started using the generic_ip and generic_sbp pointers instead. This
1516 makes ntfs entirely private with respect to the kernel tree.
1517 - Detect compiler version and abort with error message if gcc less than
1518 2.96 is used.
1519 - Fix bug in name comparison function in unistr.c.
1520 - Implement attribute lists part 1, the infrastructure: search contexts
1521 and operations, find_external_attr(), lookup_attr()) and make the
1522 code use the infrastructure.
1523 - Fix stupid buffer overflow bug that became apparent on larger run
1524 list containing attributes.
1525 - Fix bugs in readdir() that became apparent on larger directories.
1526
1527 The driver is now really useful and survives the test
1528 find . -type f -exec md5sum "{}" \;
1529 without any error messages on a over 1GiB sized partition with >16k
1530 files on it, including compressed files and directories and many files
1531 and directories with attribute lists.
1532
1533tng-0.0.1 - The first useful version.
1534
1535 - Added ntfs_lookup().
1536 - Added default upcase generation and handling.
1537 - Added compile options to be shown on module init.
1538 - Many bug fixes that were "hidden" before.
1539 - Update to latest kernel.
1540 - Added ntfs_readdir().
1541 - Added file operations for mmap(), read(), open() and llseek(). We just
1542 use the generic ones. The whole point of going through implementing
1543 readpage() methods and where possible get_block() call backs is that
1544 this allows us to make use of the generic high level methods provided
1545 by the kernel.
1546
1547 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1548 though and it doesn't implement accesssing compressed files yet. Also,
1549 accessing files with attribute list attributes is not implemented yet
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001550 either. But for small or simple filesystems it should work and allow
Linus Torvalds1da177e2005-04-16 15:20:36 -07001551 you to list directories, use stat on directory entries and the file
1552 system, open, read, mmap and llseek around in files. A big mile stone
1553 has been reached!
1554
1555tng-0.0.0 - Initial version tag.
1556
1557 Initial driver implementation. The driver can mount and umount simple
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001558 NTFS filesystems (i.e. ones without attribute lists in the system
Linus Torvalds1da177e2005-04-16 15:20:36 -07001559 files). If the mount fails there might be problems in the error handling
1560 code paths, so be warned. Otherwise it seems to be loading the system
1561 files nicely and the mft record read mapping/unmapping seems to be
1562 working nicely, too. Proof of inode metadata in the page cache and non-
1563 resident file unnamed stream data in the page cache concepts is thus
1564 complete.