blob: 35cc4b1d60f77600e0c4ac3a2b3d42c41ecb8e52 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001ToDo/Notes:
2 - Find and fix bugs.
Anton Altaparmakov98b27032005-10-11 15:40:40 +01003 - The only places in the kernel where a file is resized are
Anton Altaparmakov3672b632006-02-24 09:55:07 +00004 ntfs_file_write*() and ntfs_truncate() for both of which i_mutex is
Anton Altaparmakov98b27032005-10-11 15:40:40 +01005 held. Just have to be careful in read-/writepage and other helpers
Anton Altaparmakov3672b632006-02-24 09:55:07 +00006 not running under i_mutex that we play nice. Also need to be careful
Anton Altaparmakov98b27032005-10-11 15:40:40 +01007 with initialized_size extension in ntfs_file_write*() and writepage.
8 UPDATE: The only things that need to be checked are the compressed
9 write and the other attribute resize/write cases like index
10 attributes, etc. For now none of these are implemented so are safe.
11 - Implement filling in of holes in aops.c::ntfs_writepage() and its
12 helpers.
Linus Torvalds1da177e2005-04-16 15:20:36 -070013 - Implement mft.c::sync_mft_mirror_umount(). We currently will just
14 leave the volume dirty on umount if the final iput(vol->mft_ino)
15 causes a write of any mirrored mft records due to the mft mirror
16 inode having been discarded already. Whether this can actually ever
17 happen is unclear however so it is worth waiting until someone hits
18 the problem.
Linus Torvalds1da177e2005-04-16 15:20:36 -070019
Anton Altaparmakov78264bd2006-03-23 15:06:18 +0000202.1.27 - Various bug fixes and cleanups.
Anton Altaparmakov67b1dfe2006-03-23 14:57:43 +000021
22 - Fix two compiler warnings on Alpha. Thanks to Andrew Morton for
23 reporting them.
24 - Fix an (innocent) off-by-one error in the runlist code.
Anton Altaparmakov3ccc7382006-03-23 15:03:11 +000025 - Fix a buggette in an "should be impossible" case handling where we
26 continued the attribute lookup loop instead of aborting it.
Anton Altaparmakov78264bd2006-03-23 15:06:18 +000027 - Use buffer_migrate_page() for the ->migratepage function of all ntfs
28 address space operations.
Anton Altaparmakov949763b2006-03-23 15:34:13 +000029 - Fix comparison of $MFT and $MFTMirr to not bail out when there are
30 unused, invalid mft records which are the same in both $MFT and
31 $MFTMirr.
Anton Altaparmakova0646a12006-03-23 15:53:03 +000032 - Add support for sparse files which have a compression unit of 0.
Anton Altaparmakovf95c4012006-03-23 15:59:32 +000033 - Remove all the make_bad_inode() calls. This should only be called
34 from read inode and new inode code paths.
Anton Altaparmakovd4faf632006-03-23 16:05:11 +000035 - Limit name length in fs/ntfs/unistr.c::ntfs_nlstoucs() to maximum
36 allowed by NTFS, i.e. 255 Unicode characters, not including the
37 terminating NULL (which is not stored on disk).
Anton Altaparmakov2c2c8c12006-03-23 16:09:40 +000038 - Improve comments on file attribute flags in fs/ntfs/layout.h.
Anton Altaparmakova778f212006-03-23 16:18:23 +000039 - Fix a bug in fs/ntfs/inode.c::ntfs_read_locked_index_inode() where we
40 forgot to update a temporary variable so loading index inodes which
41 have an index allocation attribute failed.
Anton Altaparmakov20fdcf12006-03-23 16:21:02 +000042 - Add a missing call to flush_dcache_mft_record_page() in
43 fs/ntfs/inode.c::ntfs_write_inode().
Anton Altaparmakov834ba602006-03-23 16:25:23 +000044 - Handle the recently introduced -ENAMETOOLONG return value from
45 fs/ntfs/unistr.c::ntfs_nlstoucs() in fs/ntfs/namei.c::ntfs_lookup().
Ingo Molnar4e5e5292006-03-23 16:57:48 +000046 - Semaphore to mutex conversion. (Ingo Molnar)
Anton Altaparmakov67b1dfe2006-03-23 14:57:43 +000047
Anton Altaparmakov3672b632006-02-24 09:55:07 +0000482.1.26 - Minor bug fixes and updates.
49
Anton Altaparmakov78af34f2006-02-24 10:32:33 +000050 - Fix a potential overflow in file.c where a cast to s64 was missing in
51 a left shift of a page index.
52 - The struct inode has had its i_sem semaphore changed to a mutex named
53 i_mutex.
Anton Altaparmakov3672b632006-02-24 09:55:07 +000054 - We have struct kmem_cache now so use it instead of the typedef
55 kmem_cache_t. (Pekka Enberg)
Anton Altaparmakov78af34f2006-02-24 10:32:33 +000056 - Implement support for sector sizes above 512 bytes (up to the maximum
57 supported by NTFS which is 4096 bytes).
Anton Altaparmakov1cf3109f2006-02-24 10:48:14 +000058 - Do more detailed reporting of why we cannot mount read-write by
59 special casing the VOLUME_MODIFIED_BY_CHKDSK flag.
Anton Altaparmakov3672b632006-02-24 09:55:07 +000060 - Miscellaneous updates to layout.h.
61 - Cope with attribute list attribute having invalid flags. Windows
62 copes with this and even chkdsk does not detect or fix this so we
63 have to cope with it, too. Thanks to Pawel Kot for reporting the
64 problem.
65
Anton Altaparmakov98b27032005-10-11 15:40:40 +0100662.1.25 - (Almost) fully implement write(2) and truncate(2).
Anton Altaparmakovfd9d6362005-10-04 13:44:48 +010067
Anton Altaparmakov511bea52005-10-04 14:24:21 +010068 - Change ntfs_map_runlist_nolock(), ntfs_attr_find_vcn_nolock() and
69 {__,}ntfs_cluster_free() to also take an optional attribute search
70 context as argument. This allows calling these functions with the
71 mft record mapped. Update all callers.
72 - Fix potential deadlock in ntfs_mft_data_extend_allocation_nolock()
73 error handling by passing in the active search context when calling
74 ntfs_cluster_free().
Anton Altaparmakovfc0fa7d2005-10-04 14:36:56 +010075 - Change ntfs_cluster_alloc() to take an extra boolean parameter
76 specifying whether the cluster are being allocated to extend an
77 attribute or to fill a hole.
78 - Change ntfs_attr_make_non_resident() to call ntfs_cluster_alloc()
79 with @is_extension set to TRUE and remove the runlist terminator
80 fixup code as this is now done by ntfs_cluster_alloc().
Anton Altaparmakov8925d4f2005-10-04 14:48:20 +010081 - Change ntfs_attr_make_non_resident to take the attribute value size
82 as an extra parameter. This is needed since we need to know the size
83 before we can map the mft record and our callers always know it. The
84 reason we cannot simply read the size from the vfs inode i_size is
85 that this is not necessarily uptodate. This happens when
86 ntfs_attr_make_non_resident() is called in the ->truncate call path.
Anton Altaparmakov2a6fc4e12005-10-04 14:57:15 +010087 - Fix ntfs_attr_make_non_resident() to update the vfs inode i_blocks
88 which is zero for a resident attribute but should no longer be zero
89 once the attribute is non-resident as it then has real clusters
90 allocated.
Anton Altaparmakov2d868292005-10-04 15:18:56 +010091 - Add fs/ntfs/attrib.[hc]::ntfs_attr_extend_allocation(), a function to
92 extend the allocation of an attributes. Optionally, the data size,
93 but not the initialized size can be extended, too.
Anton Altaparmakovdd072332005-10-04 15:39:02 +010094 - Implement fs/ntfs/inode.[hc]::ntfs_truncate(). It only supports
Anton Altaparmakov98b27032005-10-11 15:40:40 +010095 uncompressed and unencrypted files and it never creates sparse files
96 at least for the moment (making a file sparse requires us to modify
97 its directory entries and we do not support directory operations at
98 the moment). Also, support for highly fragmented files, i.e. ones
99 whose data attribute is split across multiple extents, is severly
100 limited. When such a case is encountered, EOPNOTSUPP is returned.
Anton Altaparmakove9438252005-10-04 16:01:06 +0100101 - Enable ATTR_SIZE attribute changes in ntfs_setattr(). This completes
102 the initial implementation of file truncation. Now both open(2)ing
103 a file with the O_TRUNC flag and the {,f}truncate(2) system calls
104 will resize a file appropriately. The limitations are that only
105 uncompressed and unencrypted files are supported. Also, there is
106 only very limited support for highly fragmented files (the ones whose
107 $DATA attribute is split into multiple attribute extents).
Anton Altaparmakov29b89902005-10-11 14:54:42 +0100108 - In attrib.c::ntfs_attr_set() call balance_dirty_pages_ratelimited()
109 and cond_resched() in the main loop as we could be dirtying a lot of
110 pages and this ensures we play nice with the VM and the system as a
111 whole.
Anton Altaparmakov98b27032005-10-11 15:40:40 +0100112 - Implement file operations ->write, ->aio_write, ->writev for regular
113 files. This replaces the old use of generic_file_write(), et al and
114 the address space operations ->prepare_write and ->commit_write.
115 This means that both sparse and non-sparse (unencrypted and
116 uncompressed) files can now be extended using the normal write(2)
117 code path. There are two limitations at present and these are that
118 we never create sparse files and that we only have limited support
119 for highly fragmented files, i.e. ones whose data attribute is split
120 across multiple extents. When such a case is encountered,
121 EOPNOTSUPP is returned.
Anton Altaparmakovd04bd1f2005-10-24 08:41:24 +0100122 - $EA attributes can be both resident and non-resident.
123 - Use %z for size_t to fix compilation warnings. (Andrew Morton)
Anton Altaparmakovdda65b92005-10-24 08:57:59 +0100124 - Fix compilation warnings with gcc-4.0.2 on SUSE 10.0.
Anton Altaparmakovc9c20092005-10-24 09:00:51 +0100125 - Document extended attribute ($EA) NEED_EA flag. (Based on libntfs
126 patch by Yura Pakhuchiy.)
Anton Altaparmakovfd9d6362005-10-04 13:44:48 +0100127
Anton Altaparmakov7d333d62005-09-08 23:01:16 +01001282.1.24 - Lots of bug fixes and support more clean journal states.
Anton Altaparmakove7a10332005-09-08 16:12:28 +0100129
130 - Support journals ($LogFile) which have been modified by chkdsk. This
131 means users can boot into Windows after we marked the volume dirty.
132 The Windows boot will run chkdsk and then reboot. The user can then
133 immediately boot into Linux rather than having to do a full Windows
134 boot first before rebooting into Linux and we will recognize such a
Anton Altaparmakov5a8c0cc2005-09-26 10:48:54 +0100135 journal and empty it as it is clean by definition. Note, this only
136 works if chkdsk left the journal in an obviously clean state.
Anton Altaparmakove7a10332005-09-08 16:12:28 +0100137 - Support journals ($LogFile) with only one restart page as well as
138 journals with two different restart pages. We sanity check both and
139 either use the only sane one or the more recent one of the two in the
140 case that both are valid.
Anton Altaparmakov06d0e3c2005-09-08 16:28:25 +0100141 - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
142 ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
143 hence cannot fail.
Anton Altaparmakov9529d462005-09-08 16:33:12 +0100144 - Use ntfs_malloc_nofs_nofail() in the two critical regions in
145 fs/ntfs/runlist.c::ntfs_runlists_merge(). This means we no longer
146 need to panic() if the allocation fails as it now cannot fail.
Anton Altaparmakov84d6ebe2005-09-08 16:46:55 +0100147 - Fix two nasty runlist merging bugs that had gone unnoticed so far.
148 Thanks to Stefano Picerno for the bug report.
Anton Altaparmakov8bb73522005-09-08 16:48:28 +0100149 - Remove two bogus BUG_ON()s from fs/ntfs/mft.c.
Anton Altaparmakov2b0ada22005-09-08 16:52:31 +0100150 - Fix handling of valid but empty mapping pairs array in
151 fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress().
Anton Altaparmakovf94ad382005-09-08 17:04:11 +0100152 - Report unrepresentable inodes during ntfs_readdir() as KERN_WARNING
153 messages and include the inode number. Thanks to Yura Pakhuchiy for
154 pointing this out.
Anton Altaparmakov3ffc5a42005-09-08 20:23:06 +0100155 - Change ntfs_rl_truncate_nolock() to throw away the runlist if the new
156 length is zero.
Anton Altaparmakov6e483212005-09-08 20:26:34 +0100157 - Add runlist.[hc]::ntfs_rl_punch_nolock() which punches a caller
158 specified hole into a runlist.
Anton Altaparmakov8e08cea2005-09-08 20:29:50 +0100159 - Fix a bug in fs/ntfs/index.c::ntfs_index_lookup(). When the returned
160 index entry is in the index root, we forgot to set the @ir pointer in
161 the index context. Thanks to Yura Pakhuchiy for finding this bug.
Anton Altaparmakovf25dfb52005-09-08 20:35:33 +0100162 - Remove bogus setting of PageError in ntfs_read_compressed_block().
Anton Altaparmakov0aaccea2005-09-08 20:40:32 +0100163 - Add fs/ntfs/attrib.[hc]::ntfs_resident_attr_value_resize().
Anton Altaparmakov2983d1b2005-09-08 20:56:09 +0100164 - Fix a bug in ntfs_map_runlist_nolock() where we forgot to protect
165 access to the allocated size in the ntfs inode with the size lock.
166 - Fix ntfs_attr_vcn_to_lcn_nolock() and ntfs_attr_find_vcn_nolock() to
167 return LCN_ENOENT when there is no runlist and the allocated size is
168 zero.
169 - Fix load_attribute_list() to handle the case of a NULL runlist.
Anton Altaparmakov807c4532005-09-08 21:01:17 +0100170 - Fix handling of sparse attributes in ntfs_attr_make_non_resident().
171 - Add BUG() checks to ntfs_attr_make_non_resident() and ntfs_attr_set()
172 to ensure that these functions are never called for compressed or
173 encrypted attributes.
Anton Altaparmakovbbf18132005-09-08 21:09:06 +0100174 - Fix cluster (de)allocators to work when the runlist is NULL and more
175 importantly to take a locked runlist rather than them locking it
176 which leads to lock reversal.
Anton Altaparmakov1c7d4692005-09-08 21:15:09 +0100177 - Truncate {a,c,m}time to the ntfs supported time granularity when
178 updating the times in the inode in ntfs_setattr().
Anton Altaparmakov67bb1032005-09-08 21:19:45 +0100179 - Fixup handling of sparse, compressed, and encrypted attributes in
Anton Altaparmakov311120e2005-09-08 22:04:20 +0100180 fs/ntfs/inode.c::ntfs_read_locked_{,attr_,index_}inode(),
181 fs/ntfs/aops.c::ntfs_{read,write}page().
Anton Altaparmakov8dcdeba2005-09-08 21:25:48 +0100182 - Make ntfs_write_block() not instantiate sparse blocks if they contain
183 only zeroes.
Anton Altaparmakov54b02eb2005-09-08 21:43:47 +0100184 - Optimize fs/ntfs/aops.c::ntfs_write_block() by extending the page
185 lock protection over the buffer submission for i/o which allows the
186 removal of the get_bh()/put_bh() pairs for each buffer.
Anton Altaparmakov8273d5d2005-09-08 22:00:33 +0100187 - Fix fs/ntfs/aops.c::ntfs_{read,write}_block() to handle the case
188 where a concurrent truncate has truncated the runlist under our feet.
Anton Altaparmakova01ac532005-09-08 22:08:11 +0100189 - Fix page_has_buffers()/page_buffers() handling in fs/ntfs/aops.c.
Anton Altaparmakove6046352005-09-08 22:13:02 +0100190 - In fs/ntfs/aops.c::ntfs_end_buffer_async_read(), use a bit spin lock
191 in the first buffer head instead of a driver global spin lock to
192 improve scalability.
Anton Altaparmakov7d333d62005-09-08 23:01:16 +0100193 - Minor fix to error handling and error message display in
Anton Altaparmakov5d467702005-09-12 14:33:47 +0100194 fs/ntfs/aops.c::ntfs_prepare_nonresident_write().
195 - Change the mount options {u,f,d}mask to always parse the number as
196 an octal number to conform to how chmod(1) works, too. Thanks to
197 Giuseppe Bilotta and Horst von Brand for pointing out the errors of
198 my ways.
Anton Altaparmakov5c9f6de2005-09-19 09:33:40 +0100199 - Fix various bugs in the runlist merging code. (Based on libntfs
200 changes by Richard Russon.)
Anton Altaparmakov5a8c0cc2005-09-26 10:48:54 +0100201 - Fix sparse warnings that have crept in over time.
202 - Change ntfs_cluster_free() to require a write locked runlist on entry
203 since we otherwise get into a lock reversal deadlock if a read locked
204 runlist is passed in. In the process also change it to take an ntfs
205 inode instead of a vfs inode as parameter.
206 - Fix the definition of the CHKD ntfs record magic. It had an off by
207 two error causing it to be CHKB instead of CHKD.
Anton Altaparmakov18efefa2005-10-04 13:06:00 +0100208 - Fix a stupid bug in __ntfs_bitmap_set_bits_in_run() which caused the
209 count to become negative and hence we had a wild memset() scribbling
210 all over the system's ram.
Anton Altaparmakove7a10332005-09-08 16:12:28 +0100211
Anton Altaparmakovaf859a42005-06-25 21:07:27 +01002122.1.23 - Implement extension of resident files and make writing safe as well as
213 many bug fixes, cleanups, and enhancements...
Linus Torvalds1da177e2005-04-16 15:20:36 -0700214
215 - Add printk rate limiting for ntfs_warning() and ntfs_error() when
216 compiled without debug. This avoids a possible denial of service
217 attack. Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
218 out.
Randy Dunlap89075472005-03-03 11:19:53 +0000219 - Fix compilation warnings on ia64. (Randy Dunlap)
Anton Altaparmakovaf859a42005-06-25 21:07:27 +0100220 - Use i_size_{read,write}() instead of reading i_size by hand and cache
221 the value where apropriate.
Anton Altaparmakov36763672004-11-18 13:46:45 +0000222 - Add size_lock to the ntfs_inode structure. This is an rw spinlock
223 and it locks against access to the inode sizes. Note, ->size_lock
224 is also accessed from irq context so you must use the _irqsave and
Anton Altaparmakovaf859a42005-06-25 21:07:27 +0100225 _irqrestore lock and unlock functions, respectively. Protect all
226 accesses to allocated_size, initialized_size, and compressed_size.
Anton Altaparmakov218357f2004-11-18 20:34:59 +0000227 - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
Anton Altaparmakov946929d2005-01-13 15:26:29 +0000228 - Implement extension of resident files in the regular file write code
229 paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()). At present
230 this only works until the data attribute becomes too big for the mft
231 record after which we abort the write returning -EOPNOTSUPP from
232 ntfs_prepare_write().
Anton Altaparmakovc002f422005-02-03 12:02:56 +0000233 - Add disable_sparse mount option together with a per volume sparse
234 enable bit which is set appropriately and a per inode sparse disable
235 bit which is preset on some system file inodes as appropriate.
236 - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
Anton Altaparmakov1a0df152005-02-03 12:04:36 +0000237 - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
238 the creation of the unmapped runlist element for the base attribute
239 extent.
Anton Altaparmakovb6ad6c52005-02-15 10:08:43 +0000240 - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
241 helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
242 This allows us to map runlist fragments with the runlist lock already
243 held without having to drop and reacquire it around the call. Adapt
244 all callers.
245 - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
246 runlist. This allows us to find runlist elements with the runlist
247 lock already held without having to drop and reacquire it around the
248 call. Adapt all callers.
Anton Altaparmakov37e4c132005-02-18 10:03:13 +0000249 - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
Anton Altaparmakovd8ec7852005-02-18 09:23:39 +0000250 warning in the do_div() call on sparc32. Thanks to Meelis Roos for
251 the report and analysis of the warning.
Anton Altaparmakov37e4c132005-02-18 10:03:13 +0000252 - Fix a nasty runlist merge bug when merging two holes.
Anton Altaparmakov5ae9fcf2005-03-02 17:03:24 +0000253 - Set the ntfs_inode->allocated_size to the real allocated size in the
254 mft record for resident attributes (fs/ntfs/inode.c).
255 - Small readability cleanup to use "a" instead of "ctx->attr"
256 everywhere (fs/ntfs/inode.c).
Anton Altaparmakov41382682005-03-03 13:44:15 +0000257 - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
258 definition of ntfs_export_ops from fs/ntfs/super.c to namei.c. Also,
259 declare ntfs_export_ops in fs/ntfs/ntfs.h.
Anton Altaparmakov9451f852005-03-03 14:43:43 +0000260 - Correct sparse file handling. The compressed values need to be
261 checked and set in the ntfs inode as done for compressed files and
262 the compressed size needs to be used for vfs inode->i_blocks instead
263 of the allocated size, again, as done for compressed files.
Anton Altaparmakov271849a2005-03-07 21:36:18 +0000264 - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
265 non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
266 - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
267 write code.
Anton Altaparmakovc0c1cc02005-03-07 21:43:38 +0000268 - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
269 dropping the read lock and taking the write lock we were not checking
270 whether someone else did not already do the work we wanted to do.
271 - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
272 ntfs_attr_find_vcn_nolock() and update all callers.
Anton Altaparmakov2bfb4ff2005-03-09 15:15:06 +0000273 - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
Anton Altaparmakov43b01fd2005-03-09 15:18:43 +0000274 - Fix sign of various error return values to be negative in
275 fs/ntfs/lcnalloc.c.
Anton Altaparmakov905685f2005-03-10 11:06:19 +0000276 - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
277 handle the case where an attribute is converted from resident to
278 non-resident by a concurrent file write.
Jesper Juhl251c8422005-04-04 14:59:56 +0100279 - Remove checks for NULL before calling kfree() since kfree() does the
280 checking itself. (Jesper Juhl)
Anton Altaparmakovb0d23742005-04-04 16:20:14 +0100281 - Some utilities modify the boot sector but do not update the checksum.
282 Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
283 only emit a warning when the checksum is incorrect rather than
284 refusing the mount. Thanks to Bernd Casimir for pointing this
285 problem out.
Anton Altaparmakovbb3cf332005-04-06 13:34:31 +0100286 - Update attribute definition handling.
Anton Altaparmakov7fafb8b2005-04-06 16:09:21 +0100287 - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
288 - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
Anton Altaparmakovd53ee322005-04-06 16:11:20 +0100289 - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
290 better code generation and one less sparse warning in fs/ntfs/aops.c.
Pekka Enberg2fb21db2005-05-25 21:15:34 +0300291 - Remove spurious void pointer casts from fs/ntfs/. (Pekka Enberg)
Anton Altaparmakov442d2072005-05-27 16:42:56 +0100292 - Use C99 style structure initialization after memory allocation where
293 possible (fs/ntfs/{attrib.c,index.c,super.c}). Thanks to Al Viro and
294 Pekka Enberg.
Anton Altaparmakov3f2faef2005-06-25 15:28:56 +0100295 - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
296 is active on the volume and we are mounting read-write or remounting
297 from read-only to read-write.
Anton Altaparmakov9f993fe2005-06-25 16:15:36 +0100298 - Fix a bug in address space operations error recovery code paths where
299 if the runlist was not mapped at all and a mapping error occured we
300 would leave the runlist locked on exit to the function so that the
301 next access to the same file would try to take the lock and deadlock.
Anton Altaparmakovca8fd7a2005-06-25 16:31:27 +0100302 - Detect the case when Windows has been suspended to disk on the volume
303 to be mounted and if this is the case do not allow (re)mounting
304 read-write. This is done by parsing hiberfil.sys if present.
Anton Altaparmakov3bd1f4a2005-06-25 16:51:58 +0100305 - Fix several occurences of a bug where we would perform 'var & ~const'
306 with a 64-bit variable and a int, i.e. 32-bit, constant. This causes
307 the higher order 32-bits of the 64-bit variable to be zeroed. To fix
308 this cast the 'const' to the same 64-bit type as 'var'.
Anton Altaparmakov1d58b272005-06-25 17:04:55 +0100309 - Change the runlist terminator of the newly allocated cluster(s) to
310 LCN_ENOENT in ntfs_attr_make_non_resident(). Otherwise the runlist
311 code gets confused.
Anton Altaparmakovfa3be922005-06-25 17:15:36 +0100312 - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
313 and ntfs_mapping_pairs_build() to allow the runlist encoding to be
314 partial which is desirable when filling holes in sparse attributes.
315 Update all callers.
Anton Altaparmakov4757d7df2005-06-25 17:24:08 +0100316 - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
317 if the requested vcn is inside it. Otherwise we get into problems
318 when we try to map an out of bounds vcn because we then try to map
319 the already mapped runlist fragment which causes
320 ntfs_mapping_pairs_decompress() to fail and return error. Update
321 ntfs_attr_find_vcn_nolock() accordingly.
Anton Altaparmakovba6d2372005-06-26 22:12:02 +0100322 - Fix a nasty deadlock that appeared in recent kernels.
323 The situation: VFS inode X on a mounted ntfs volume is dirty. For
324 same inode X, the ntfs_inode is dirty and thus corresponding on-disk
325 inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
326 to the table of inodes, i.e. $MFT, inode 0.
327 What happens:
328 Process 1: sys_sync()/umount()/whatever... calls
329 __sync_single_inode() for $MFT -> do_writepages() -> write_page for
330 the dirty page containing the on-disk inode X, the page is now locked
331 -> ntfs_write_mst_block() which clears PageUptodate() on the page to
332 prevent anyone else getting hold of it whilst it does the write out.
333 This is necessary as the on-disk inode needs "fixups" applied before
334 the write to disk which are removed again after the write and
335 PageUptodate is then set again. It then analyses the page looking
336 for dirty on-disk inodes and when it finds one it calls
337 ntfs_may_write_mft_record() to see if it is safe to write this
338 on-disk inode. This then calls ilookup5() to check if the
339 corresponding VFS inode is in icache(). This in turn calls ifind()
340 which waits on the inode lock via wait_on_inode whilst holding the
341 global inode_lock.
342 Process 2: pdflush results in a call to __sync_single_inode for the
343 same VFS inode X on the ntfs volume. This locks the inode (I_LOCK)
344 then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
345 read_cache_page() for the page (in page cache of table of inodes
346 $MFT, inode 0) containing the on-disk inode. This page has
347 PageUptodate() clear because of Process 1 (see above) so
348 read_cache_page() blocks when it tries to take the page lock for the
349 page so it can call ntfs_read_page().
350 Thus Process 1 is holding the page lock on the page containing the
351 on-disk inode X and it is waiting on the inode X to be unlocked in
352 ifind() so it can write the page out and then unlock the page.
353 And Process 2 is holding the inode lock on inode X and is waiting for
354 the page to be unlocked so it can call ntfs_readpage() or discover
355 that Process 1 set PageUptodate() again and use the page.
356 Thus we have a deadlock due to ifind() waiting on the inode lock.
357 The solution: The fix is to use the newly introduced
358 ilookup5_nowait() which does not wait on the inode's lock and hence
359 avoids the deadlock. This is safe as we do not care about the VFS
360 inode and only use the fact that it is in the VFS inode cache and the
361 fact that the vfs and ntfs inodes are one struct in memory to find
362 the ntfs inode in memory if present. Also, the ntfs inode has its
363 own locking so it does not matter if the vfs inode is locked.
Anton Altaparmakove74589a2005-08-16 16:38:28 +0100364 - Fix bug in mft record writing where we forgot to set the device in
Anton Altaparmakov481d0372005-08-16 19:42:56 +0100365 the buffers when mapping them after the VM had discarded them.
Anton Altaparmakove74589a2005-08-16 16:38:28 +0100366 Thanks to Martin MOKREJÅ  for the bug report.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700367
3682.1.22 - Many bug and race fixes and error handling improvements.
369
370 - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
371 - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
372 instead of void and provide a helper ntfs_truncate_vfs() for the
373 vfs ->truncate method.
374 - Add a new ntfs inode flag NInoTruncateFailed() and modify
375 fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
376 - Fix min_size and max_size definitions in ATTR_DEF structure in
377 fs/ntfs/layout.h to be signed.
378 - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
379 ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
380 ntfs_attr_can_be_resident(), which in turn use the new private helper
381 ntfs_attr_find_in_attrdef().
382 - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
383 mapping->private_lock around the dirtying of the buffer heads
384 analagous to the way it is done in __set_page_dirty_buffers().
385 - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
386 mount time as this cannot work with the current implementation.
387 - Check for location of attribute name and improve error handling in
388 general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
389 - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
390 i_size, i.e. race with truncate, invalidate the buffers on the page
391 so that they become freeable and hence the page does not leak.
392 - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge(). (Adrian
393 Bunk)
394 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
395 a NULL pointer dereference in the error code path when a corrupt
396 attribute was found. (Thanks to Domen Puncer for the bug report.)
397 - Add MODULE_VERSION() to fs/ntfs/super.c.
398 - Make several functions and variables static. (Adrian Bunk)
399 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
400 buffers for the page if they are not present and then marks the
401 buffers belonging to the ntfs record dirty. This causes the buffers
402 to become busy and hence they are safe from removal until the page
403 has been written out.
404 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
405 error handling code path that resulted in a BUG() due to trying to
406 unmap an extent mft record when the mapping of it had failed and it
407 thus was not mapped. (Thanks to Ken MacFerrin for the bug report.)
408 - Drop the runlist lock after the vcn has been read in
409 fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
410 - Rewrite handling of multi sector transfer errors. We now do not set
411 PageError() when such errors are detected in the async i/o handler
412 fs/ntfs/aops.c::ntfs_end_buffer_async_read(). All users of mst
413 protected attributes now check the magic of each ntfs record as they
414 use it and act appropriately. This has the effect of making errors
415 granular per ntfs record rather than per page which solves the case
416 where we cannot access any of the ntfs records in a page when a
417 single one of them had an mst error. (Thanks to Ken MacFerrin for
418 the bug report.)
419 - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
Anton Altaparmakov3672b632006-02-24 09:55:07 +0000420 where we failed to release i_mutex on the $Quota/$Q attribute inode.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700421 - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
422 - Add mapping of unmapped buffers to all remaining code paths, i.e.
423 fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
424 and write_mft_record_nolock(). From now on we require that the
425 complete runlist for the mft mirror is always mapped into memory.
426 - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
427 - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
428 - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
429 resident attribute will be smaller than a page which makes the code
430 simpler. Also make the code more tolerant to concurrent ->truncate.
431
4322.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
433
434 - Implement extent mft record deallocation
435 fs/ntfs/mft.c::ntfs_extent_mft_record_free().
436 - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
437 - Add vol->mft_data_pos and initialize it at mount time.
438 - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
439 ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
440 ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
441 ntfs_runlists_merge() and adapt all callers.
442 - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
443 ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
444 and ntfs_mapping_pairs_build(), adapted from libntfs.
445 - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
446 static and add a declaration for it to lcnalloc.h.
447 - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
448 inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
449 cluster bitmap lock for the duration of the call.
450 - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
451 - Implement the equivalent of memset() for an ntfs attribute in
452 fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
453 fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
454 - Remove unnecessary casts from LCN_* constants.
455 - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
456 - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
457 change MFT_RECORD to contain the NTFS 3.1+ specific fields.
458 - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
459 marks all buffers belonging to an ntfs record dirty, followed by
460 marking the page the ntfs record is in dirty and also marking the vfs
461 inode containing the ntfs record dirty (I_DIRTY_PAGES).
462 - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
463 new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
464 longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
465 - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
466 include errors.
467 - Move the typedefs for runlist_element and runlist from types.h to
468 runlist.h and fix resulting include errors.
469 - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
470 - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
471 mark_ntfs_record_dirty() which also changes the behaviour in that we
472 now set the buffers belonging to the mft record dirty as well as the
473 page itself.
474 - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
475 to cope with the fact that there now are dirty buffers in mft pages.
476 - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
477 mark_ntfs_record_dirty() and thus to set the buffers belonging to the
478 mft record dirty as well as the page itself.
479 - Fix compiler warnings on x86-64 in fs/ntfs/dir.c. (Randy Dunlap,
480 slightly modified by me)
481 - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
482 the mft record is already locked and otherwise behaves the same way
483 as fs/ntfs/mft.c::map_mft_record().
484 - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
485 writes the mft record if the buffers belonging to it are dirty.
486 Otherwise we assume that it was written out by other means already.
487 - Attempting to write outside initialized size is _not_ a bug so remove
488 the bug check from fs/ntfs/aops.c::ntfs_write_mst_block(). It is in
489 fact required to write outside initialized size when preparing to
490 extend the initialized size.
491 - Map the page instead of using page_address() before writing to it in
492 fs/ntfs/aops.c::ntfs_mft_writepage().
493 - Provide exclusion between opening an inode / mapping an mft record
494 and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
495 by setting the page not uptodate throughout ntfs_mft_writepage().
496 - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
497 to ensure noone can see the page whilst the mst fixups are applied.
498 - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
499 checks if an mft record may be written out safely obtaining any
500 necessary locks in the process. This is used by
501 fs/ntfs/aops.c::ntfs_write_mst_block().
502 - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
503 writing mft records and improve its error handling in the process.
504 Now if any of the records in the page fail to be written out, all
505 other records will be written out instead of aborting completely.
506 - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
507 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
508 ntfs_mst_aops for all inodes which are NInoMstProtected() and
509 ntfs_aops for all other inodes.
510 - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
511 ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
512 no longer require an ntfs inode to be present. Update all callers.
513 - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
514 - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
515 to ensure noone can see the page whilst the mst fixups are applied.
516 - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
517 fs/ntfs/mft.c::try_map_mft_record().
518 - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
519 with the ntfs inode which contains the page rather than the ntfs
520 inode the mft record of which is in the page.
521 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
522 index inode bitmap inode release code from there to
523 fs/ntfs/inode.c::ntfs_clear_big_inode(). (Thanks to Christoph
524 Hellwig for spotting this.)
525 - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
526 inode semaphore around the code that sets ni->itype.index.bmp_ino to
527 NULL and reorganize the code to optimize it a bit. (Thanks to
528 Christoph Hellwig for spotting this.)
529 - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
530 ntfs inode as a parameter as this is confusing and misleading and the
531 needed ntfs inode is available via NTFS_I(page->mapping->host).
532 Adapt all callers to this change.
533 - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
534 fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
535 of the first buffer in a record and to take this as the ntfs record
536 dirty state. We cannot look at the dirty state for subsequent
537 buffers because we might be racing with
538 fs/ntfs/aops.c::mark_ntfs_record_dirty().
539 - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
540 inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
541 add a declaration for it to inode.h. Fix some compilation issues
542 that resulted due to #includes and header file interdependencies.
543 - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
544 - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
545 - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
546 record sequence number if it is specified (i.e. not zero).
547 - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
548 functions used by it.
549 - Update Documentation/filesystems/ntfs.txt with instructions on how to
550 use the Device-Mapper driver with NTFS ftdisk/LDM raid. This removes
551 the linear raid problem with the Software RAID / MD driver when one
552 or more of the devices has an odd number of sectors.
553
5542.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
555
556 - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
557 where we did not clear ctx->al_entry but it was still set due to
558 changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
559 particular.
560 - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
561 where we forgot to unmap the extent mft record when we had finished
562 enumerating an attribute which caused a bug check to trigger when the
563 VFS calls ->clear_inode.
564
5652.1.19 - Many cleanups, improvements, and a minor bug fix.
566
567 - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
568 change the uid, gid, and mode of an inode as we do not support NTFS
569 ACLs yet.
570 - Remove BKL use from ntfs_setattr() syncing up with the rest of the
571 kernel.
572 - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
573 and ntfs_filldir() as per suggestion from Al Viro.
574 - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
575 - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
576 inode size has changed and to only output an error if so.
577 - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
578 - Add le{16,32,64} as well as sle{16,32,64} data types to
579 fs/ntfs/types.h.
580 - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
581 - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
582 respectively, to fs/ntfs/types.h.
583 - Update endianness conversion macros in fs/ntfs/endian.h to use the
584 new types as appropriate.
585 - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
586 and index.c.
587 - Add leMFT_REF data type to fs/ntfs/layout.h.
588 - Update all NTFS header files with the new little endian data types.
589 Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
590 - Do proper type casting when using ntfs_is_*_recordp() in
591 fs/ntfs/logfile.c, mft.c, and super.c.
592 - Fix all the sparse bitwise warnings. Had to change all the typedef
593 enums storing little endian values to simple enums plus a typedef for
594 the datatype to make sparse happy.
595 - Fix a bug found by the new sparse bitwise warnings where the default
596 upcase table was defined as a pointer to wchar_t rather than ntfschar
597 in fs/ntfs/ntfs.h and super.c.
598 - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
599
6002.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
601
602 - Remove vol->nr_mft_records as it was pretty meaningless and optimize
603 the calculation of total/free inodes as used by statfs().
604 - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
605 because the code itself is using the ntfs_lock semaphore which
606 provides safe locking. (Ingo Molnar)
607 - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
608 could occur in the future for when we start closing/freeing extent
609 inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
610 we free it.
611 - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
612 find_external_attr() to ntfs_external_attr_find() to cleanup the
613 namespace a bit and to be more consistent with libntfs.
614 - Rename {{re,}init,get,put}_attr_search_ctx() to
615 ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
616 attr_search_context to ntfs_attr_search_ctx.
617 - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
618 for the attribute list attribute itself.
619 - Fix endianness bug in ntfs_external_attr_find().
620 - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
621 if the attribute is not found, and -EIO on real error. In the case
622 of -ENOENT, the search context is updated to describe the attribute
623 before which the attribute being searched for would need to be
624 inserted if such an action were to be desired and in the case of
625 ntfs_external_attr_find() the search context is also updated to
626 indicate the attribute list entry before which the attribute list
627 entry of the attribute being searched for would need to be inserted
628 if such an action were to be desired. Also make ntfs_find_attr()
629 static and remove its prototype from attrib.h as it is not used
630 anywhere other than attrib.c. Update ntfs_attr_lookup() and all
631 callers of ntfs_{external,}attr_{find,lookup}() for the new return
632 values.
633 - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
634
6352.1.17 - Fix bugs in mount time error code paths and other updates.
636
637 - Implement bitmap modification code (fs/ntfs/bitmap.[hc]). This
638 includes functions to set/clear a single bit or a run of bits.
639 - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
640 runlist element containing a particular vcn. It also takes care of
641 mapping any needed runlist fragments.
642 - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
643 - Load attribute definition table from $AttrDef at mount time.
644 - Fix bugs in mount time error code paths involving (de)allocation of
645 the default and volume upcase tables.
646 - Remove ntfs_nr_mounts as it is no longer used.
647
6482.1.16 - Implement access time updates, file sync, async io, and read/writev.
649
650 - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
651 This is done by setting the appropriate file operations pointers to
652 the generic helper functions provided by mm/filemap.c.
653 - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
654 and directories (fs/ntfs/dir.c).
655 - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
656 Note, except for the root directory and any other system files opened
657 by the user, the system files will not have their access times
658 updated as they are only accessed at the inode level an hence the
659 file level functions which cause the times to be updated are never
660 invoked.
661
6622.1.15 - Invalidate quotas when (re)mounting read-write.
663
664 - Add new element itype.index.collation_rule to the ntfs inode
665 structure and set it appropriately in ntfs_read_locked_inode().
666 - Implement a new inode type "index" to allow efficient access to the
667 indices found in various system files and adapt inode handling
668 accordingly (fs/ntfs/inode.[hc]). An index inode is essentially an
669 attribute inode (NInoAttr() is true) with an attribute type of
670 AT_INDEX_ALLOCATION. As such, it is no longer allowed to call
671 ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
672 there would be no way to distinguish between normal attribute inodes
673 and index inodes. The function to obtain an index inode is
674 ntfs_index_iget() and it uses the helper function
675 ntfs_read_locked_index_inode(). Note, we do not overload
676 ntfs_attr_iget() as indices consist of multiple attributes so using
677 ntfs_attr_iget() to obtain an index inode would be confusing.
678 - Ensure that there is no overflow when doing page->index <<
679 PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
680 - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
681 and ntfs_read_block().
682 - Use case sensitive attribute lookups instead of case insensitive ones.
683 - Lock all page cache pages belonging to mst protected attributes while
684 accessing them to ensure we never see corrupt data while the page is
685 under writeout.
686 - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
687 We have ntfs_is_collation_rule_supported() to check if the collation
688 rule you want to use is supported and ntfs_collation() which actually
689 collates two data items. We currently only support COLLATION_BINARY
690 and COLLATION_NTOFS_ULONG but support for other collation rules will
691 be added as the need arises.
692 - Add a new type, ntfs_index_context, to allow retrieval of an index
693 entry using the corresponding index key. To get an index context,
694 use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
695 This also adds a new slab cache for the index contexts. To lookup a
696 key in an index inode, use ntfs_index_lookup(). After modifying an
697 index entry, call ntfs_index_entry_flush_dcache_page() followed by
698 ntfs_index_entry_mark_dirty() to ensure the changes are written out
699 to disk. For details see fs/ntfs/index.[hc]. Note, at present, if
700 an index entry is in the index allocation attribute rather than the
701 index root attribute it will not be written out (you will get a
702 warning message about discarded changes instead).
703 - Load the quota file ($Quota) and check if quota tracking is enabled
704 and if so, mark the quotas out of date. This causes windows to
705 rescan the volume on boot and update all quota entries.
706 - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
707 It is simply set to __set_page_dirty_nobuffers() to make sure that
708 running set_page_dirty() on a page containing mft/ntfs records will
709 not affect the dirty state of the page buffers.
710 - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
711 buffers that are inside the ntfs record in the page dirty after which
712 it sets the page dirty. This allows ->writepage to only write the
713 dirty index records rather than having to write all the records in
714 the page. Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
715 use this rather than __set_page_dirty_nobuffers().
716 - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
717 writing of page cache pages belonging to mst protected attributes
718 like the index allocation attribute in directory indices and other
719 indices like $Quota/$Q, etc. This means that the quota is now marked
720 out of date on all volumes rather than only on ones where the quota
721 defaults entry is in the index root attribute of the $Quota/$Q index.
722
7232.1.14 - Fix an NFSd caused deadlock reported by several users.
724
725 - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
726 to a buffer so that we can put the search context and unmap the mft
727 record before calling the filldir() callback. We need to do this
728 because of NFSd which calls ->lookup() from its filldir callback()
729 and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
730 of the directory and since ntfs_readdir() has got it mapped already
731 ntfs_lookup() deadlocks.
732
7332.1.13 - Enable overwriting of resident files and housekeeping of system files.
734
735 - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
736 keeping the mft mirror in sync with the mft when mirrored mft records
737 are written. The functions are write_mft_record{,_nolock}(). The
738 implementation is quite rudimentary for now with lots of things not
739 implemented yet but I am not sure any of them can actually occur so
740 I will wait for people to hit each one and only then implement it.
741 - Commit open system inodes at umount time. This should make it
742 virtually impossible for sync_mft_mirror_umount() to ever be needed.
743 - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
744 ntfs super operations. This gives us inode writing via the VFS inode
745 dirty code paths. Note: Access time updates are not implemented yet.
746 - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
747 fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
748 finally enabling resident file overwrite! (-8 This also includes a
749 placeholder for ->writepage (ntfs_mft_writepage()), which for now
750 just redirties the page and returns. Also, at umount time, we for
751 now throw away all mft data page cache pages after the last call to
752 ntfs_commit_inode() in the hope that all inodes will have been
753 written out by then and hence no dirty (meta)data will be lost. We
754 also check for this case and emit an error message telling the user
755 to run chkdsk.
756 - Use set_page_writeback() and end_page_writeback() in the resident
757 attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
758 the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
759 page is clean.
760 - Implement ntfs_mft_writepage() so it now checks if any of the mft
761 records in the page are dirty and if so redirties the page and
762 returns. Otherwise it just returns (after doing set_page_writeback(),
763 unlock_page(), end_page_writeback() or the radix-tree tag
764 PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
765 alowing the VM to do with the page as it pleases. Also, at umount
766 time, now only throw away dirty mft (meta)data pages if dirty inodes
767 are present and ask the user to email us if they see this happening.
768 - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
769 information flags (fs/ntfs/super.c).
770 - Mark the volume dirty when (re)mounting read-write and mark it clean
771 when unmounting or remounting read-only. If any volume errors are
772 found, the volume is left marked dirty to force chkdsk to run.
773 - Add code to set the NT4 compatibility flag when (re)mounting
774 read-write for newer NTFS versions but leave it commented out for now
775 since we do not make any modifications that are NTFS 1.2 specific yet
776 and since setting this flag breaks Captive-NTFS which is not nice.
777 This code must be enabled once we start writing NTFS 1.2 specific
778 changes otherwise Windows NTFS driver might crash / cause corruption.
779
7802.1.12 - Fix the second fix to the decompression engine and some cleanups.
781
782 - Add a new address space operations struct, ntfs_mst_aops, for mst
783 protected attributes. This is because the default ntfs_aops do not
784 make sense with mst protected data and were they to write anything to
785 such an attribute they would cause data corruption so we provide
786 ntfs_mst_aops which does not have any write related operations set.
787 - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
788 includes an adapted ntfs_commit_inode() and an implementation of
789 ntfs_write_inode() which for now just cleans dirty inodes without
790 writing them (it does emit a warning that this is happening).
791 - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
792 entry) as it was only fixing a theoretical bug but at the same time
793 it badly broke the handling of sparse and uncompressed compression
794 blocks.
795
7962.1.11 - Driver internal cleanups.
797
798 - Only build logfile.o if building the driver with read-write support.
799 - Really final white space cleanups.
800 - Use generic_ffs() instead of ffs() in logfile.c which allows the
801 log_page_size variable to be optimized by gcc into a constant.
802 - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
803 char as defined by POSIX and as found on some systems.
804
8052.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
806
807 - Finish off the white space cleanups (remove trailing spaces, etc).
808 - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
809 the kludges around the first iget(). Instead of (re)setting ->s_op
810 we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
811 insert_inode_hash() / call ntfs_read_inode_mount() directly. This
812 kills the need for second super_operations and allows to return error
813 from ntfs_read_inode_mount() without resorting to ugly "poisoning"
814 tricks. (Al Viro)
815 - Force read-only (re)mounting if any of the following bits are set in
816 the volume information flags:
817 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
818 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
819 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
820 To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
821 above bits set so the test is made easy.
822
8232.1.9 - Fix two bugs in decompression engine.
824
825 - Fix a bug where we would not always detect that we have reached the
826 end of a compression block because we were ending at minus one byte
827 which is effectively the same as being at the end. The fix is to
828 check whether the uncompressed buffer has been fully filled and if so
829 we assume we have reached the end of the compression block. A big
830 thank you to Marcin Gibuła for the bug report, the assistance in
831 tracking down the bug and testing the fix.
832 - Fix a possible bug where when a compressed read is truncated to the
833 end of the file, the offset inside the last page was not truncated.
834
8352.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
836
837 - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
838 - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
839 utc2ntfs() to work with struct timespec instead of time_t on the
840 Linux UTC time side thus preserving the full precision of the NTFS
841 time and only loosing up to 99 nano-seconds in the Linux UTC time.
842 - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
843 static inline.
844 - Remove unused ntfs_dirty_inode().
845 - Cleanup super operations declaration in fs/ntfs/super.c.
846 - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
847 - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
848 fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
849 - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
850 fs/ntfs/inode.h so they can be used elsewhere.
851 - Determine the mft mirror size as the number of mirrored mft records
852 and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
853 - Load the mft mirror at mount time and compare the mft records stored
854 in it to the ones in the mft. Force a read-only mount if the two do
855 not match (fs/ntfs/super.c).
856 - Fix type casting related warnings on 64-bit architectures. Thanks
857 to Meelis Roos for reporting them.
858 - Move %L to %ll as %L is floating point and %ll is integer which is
859 what we want.
860 - Read the journal ($LogFile) and determine if the volume has been
861 shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
862 and fs/ntfs/logfile.c). This is a little bit of a crude check in
863 that we only look at the restart areas and not at the actual log
864 records so that there will be a very small number of cases where we
865 think that a volume is dirty when in fact it is clean. This should
866 only affect volumes that have not been shutdown cleanly and did not
867 have any pending, non-check-pointed i/o.
868 - If the $LogFile indicates a clean shutdown and a read-write (re)mount
869 is requested, empty $LogFile by overwriting it with 0xff bytes to
870 ensure that Windows cannot cause data corruption by replaying a stale
871 journal after Linux has written to the volume.
872
8732.1.7 - Enable NFS exporting of mounted NTFS volumes.
874
875 - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
876 - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
877 - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
878 default doesn't allow inode number 0 which is a valid inode on NTFS
879 and even if it did allow that it uses iget() instead of ntfs_iget()
880 which makes it useless for us.
881 - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
882 default just returns -EACCES which is not very useful.
883 - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
884 and set them up in the super block at mount time (super.c) this
885 allows mounted NTFS volumes to be exported via NFS.
886 - Add missing return -EOPNOTSUPP; in
887 fs/ntfs/aops.c::ntfs_commit_nonresident_write().
888 - Enforce no atime and no dir atime updates at mount/remount time as
889 they are not implemented yet anyway.
890 - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
891 after a NULL check. Thanks to Dave Jones for pointing this out.
892
8932.1.6 - Fix minor bug in handling of compressed directories.
894
895 - Fix bug in handling of compressed directories. A compressed
896 directory is not really compressed so when we set the ->i_blocks
897 field of a compressed directory inode we were setting it from the
898 non-existing field ni->itype.compressed.size which gave random
899 results... For directories we now always use ni->allocated_size.
900
9012.1.5 - Fix minor bug in attribute list attribute handling.
902
903 - Fix bug in attribute list handling. Actually it is not as much a bug
904 as too much protection in that we were not allowing attribute lists
905 which waste space on disk while Windows XP clearly allows it and in
906 fact creates such attribute lists so our driver was failing.
907 - Update NTFS documentation ready for 2.6 kernel release.
908
9092.1.4 - Reduce compiler requirements.
910
911 - Remove all uses of unnamed structs and unions in the driver to make
912 old and newer gcc versions happy. Makes it a bit uglier IMO but at
913 least people will stop hassling me about it.
914
9152.1.3 - Important bug fixes in corner cases.
916
917 - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
918 clusters. (Philipp Thomas)
919 - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
920 multiple of the block_size but not the cluster size. (Szabolcs
Anton Altaparmakov3672b632006-02-24 09:55:07 +0000921 Szakacsits)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700922
9232.1.2 - Important bug fixes aleviating the hangs in statfs.
924
925 - Fix buggy free cluster and free inode determination logic.
926
9272.1.1 - Minor updates.
928
929 - Add handling for initialized_size != data_size in compressed files.
930 - Reduce function local stack usage from 0x3d4 bytes to just noise in
Anton Altaparmakov3672b632006-02-24 09:55:07 +0000931 fs/ntfs/upcase.c. (Randy Dunlap)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700932 - Remove compiler warnings for newer gcc.
933 - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
934 around calls to ->{prepare,commit}_write. Adapt NTFS appropriately
935 in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
936 kmap_atomic(KM_USER0).
937
9382.1.0 - First steps towards write support: implement file overwrite.
939
940 - Add configuration option for developmental write support with an
941 appropriately scary configuration help text.
942 - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
943 helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
944 overwriting of existing files on ntfs. Note: Resident files are
945 only written into memory, and not written out to disk at present, so
946 avoid writing to files smaller than about 1kiB.
947 - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
948 helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
949 counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
950 fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
951 add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
952 This enables write(2) based overwriting of existing files on ntfs.
953 Note: As with mmap(2) based overwriting, resident files are only
954 written into memory, and not written out to disk at present, so avoid
955 writing to files smaller than about 1kiB.
956 - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
957 ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
958 files with the purpose of intercepting and aborting all i_size
959 changes which we do not support yet. ntfs_truncate() actually only
960 emits a warning message but AFAICS our interception of i_size changes
961 elsewhere means ntfs_truncate() never gets called for i_size changes.
962 It is only called from generic_file_write() when we fail in
963 ntfs_prepare_{,nonresident_}write() in order to discard any
964 instantiated buffers beyond i_size. Thus i_size is not actually
965 changed so our warning message is enough. Unfortunately it is not
966 possible to easily determine if i_size is being changed or not hence
967 we just emit an appropriately worded error message.
968
9692.0.25 - Small bug fixes and cleanups.
970
971 - Unlock the page in an out of memory error code path in
972 fs/ntfs/aops.c::ntfs_read_block().
973 - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
974 just unlock the page and return. (This can happen due to ->writepage
975 clearing PageUptodate() during write out of MstProtected()
976 attributes.
977 - Remove leaked write code again.
978
9792.0.24 - Cleanups.
980
981 - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
982 inside BUG_ON(). (Adam J. Richter)
983 - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
984 calls for improved debugging. (Adam J. Richter)
985 - Add errors flag to the ntfs volume state, accessed via
986 NVol{,Set,Clear}Errors(vol).
987 - Do not allow read-write remounts of read-only volumes with errors.
988 - Clarify comment for ntfs file operation sendfile which was added by
989 Christoph Hellwig a while ago (just using generic_file_sendfile())
990 to say that ntfs ->sendfile is only used for the case where the
991 source data is on the ntfs partition and the destination is
992 somewhere else, i.e. nothing we need to concern ourselves with.
993 - Add generic_file_write() as our ntfs file write operation.
994
9952.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
996
997 - Massive internal locking changes to mft record locking. Fixes lock
998 recursion and replaces the mrec_lock read/write semaphore with a
999 mutex. Also removes the now superfluous mft_count. This fixes several
1000 race conditions and deadlocks, especially in the future write code.
1001 - Fix ntfs over loopback for compressed files by adding an
1002 optimization barrier. (gcc was screwing up otherwise ?)
1003 - Miscellaneous cleanups all over the code and a fix or two in error
1004 handling code paths.
1005 Thanks go to Christoph Hellwig for pointing out the following two:
1006 - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
1007 - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
1008
10092.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
1010
1011 - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
1012 at entry/exit respectively.
1013 - Use C99 initializers for structures.
1014 - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
1015
10162.0.21 - Check for, and refuse to work with too large files/directories/volumes.
1017
1018 - Limit volume size at mount time to 2TiB on architectures where
1019 unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
1020 This is the most we can do without overflowing the 32-bit limit of
1021 the block device size imposed on us by sb_bread() and sb_getblk()
1022 for the time being.
1023 - Limit file/directory size at open() time to 16TiB on architectures
1024 where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
1025 fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
1026 overflowing the page cache page index.
1027
10282.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
1029
1030 - Move the directory index bitmap to use an attribute inode instead of
1031 having special fields for it inside the ntfs inode structure. This
1032 means that the index bitmaps now use the page cache for i/o, too,
1033 and also as a side effect we get support for non-resident index
1034 bitmaps for free.
1035 - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
1036 fix a page leak that manifested itself in some cases.
1037 - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
1038 index bitmap inode on the final iput().
1039
10402.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
1041
1042 - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
1043 to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
1044 - Drop the "file" from ntfs_file_read_compressed_block().
1045 - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
1046 ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
1047 - Update ntfs_end_buffer_async_read() with the improved logic from
1048 its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
1049 further logic improvements to better determine when we set PageError.
1050 - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
1051 check for the buffers being uptodate first in line with the updated
1052 fs/buffer.c::block_read_full_page(). This plugs a small race
1053 condition.
1054
10552.0.18 - Fix race condition in reading of compressed files.
1056
1057 - There was a narrow window between checking a buffer head for being
1058 uptodate and locking it in ntfs_file_read_compressed_block(). We now
1059 lock the buffer and then check whether it is uptodate or not.
1060
10612.0.17 - Cleanups and optimizations - shrinking the ToDo list.
1062
1063 - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
1064 code and update callers, i.e. ntfs_iget(), to pass that error code
1065 up instead of just using -EIO.
1066 - Modifications to super.c to ensure that both mount and remount
1067 cannot set any write related options when the driver is compiled
1068 read-only.
1069 - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
1070 cache the current runlist element. This should improve performance
1071 when reading very large and/or very fragmented data.
1072
10732.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
1074
1075 - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
1076 wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
1077 - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
1078 - Convert $MFT/$BITMAP access to attribute inode API and remove all
1079 remnants of the ugly mftbmp address space and operations hack. This
1080 means we finally have only one readpage function as well as only one
1081 async io completion handler. Yey! The mft bitmap is now just an
1082 attribute inode and is accessed from vol->mftbmp_ino just as if it
1083 were a normal file. Fake inodes rule. (-:
1084
10852.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
1086
1087 - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
1088 remounts to fail when the partition had an entry in /etc/fstab and
1089 the entry specified the nls= option.
1090 - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
1091 expand all the helper functions NVolFoo(), NVolSetFoo(), and
1092 NVolClearFoo().
1093 - Move copyright statement from driver initialisation message to
1094 module description (fs/super.c). This makes the initialisation
1095 message fit on one line and fits in better with rest of kernel.
1096 - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
1097 attribute inodes, and both for files and directories.
1098 - Implement fake attribute inodes allowing all attribute i/o to go via
1099 the page cache and to use all the normal vfs/mm functionality:
1100 - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
1101 to fs/ntfs/inode.c.
1102 - Add needed cleanup code to ntfs_clear_big_inode().
1103 - Merge address space operations for files and directories (aops.c),
1104 now just have ntfs_aops:
1105 - Rename:
1106 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
1107 ntfs_attr_read_block() -> ntfs_read_block(),
1108 ntfs_file_read_page() -> ntfs_readpage().
1109 - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
1110 attribute inodes, and both for files and directories.
1111 - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
1112
11132.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
1114
1115 - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
1116 the locking out of super.c::get_nr_free_mft_records() and taking and
1117 dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
1118 - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
1119 current userspace ntfs library code. This means that if a merge
1120 fails the original runlists are always left unmodified instead of
1121 being silently corrupted.
1122 - Misc typo fixes.
1123
11242.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
1125
1126 - Remove nr_mft_bits and the now superfluous union with nr_mft_records
1127 from ntfs_volume structure.
1128 - Remove nr_lcn_bits and the now superfluous union with nr_clusters
1129 from ntfs_volume structure.
1130 - Use iget5_locked() and friends instead of conventional iget(). Wrap
1131 the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
1132 to use ntfs_iget(). Leave only one iget() call at mount time so we
1133 don't need an ntfs_iget_mount().
1134 - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
1135 additional argument.
1136
11372.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
1138
1139 - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
1140 fs/ntfs/aops.c::end_buffer_read_file_async() into one function
1141 fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
1142 to determine whether to apply mst fixups or not.
1143 - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
1144 and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
1145 fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
1146 fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
1147 the VFS readpage function prototype to the ntfs_attr_read_block()
1148 function prototype.
1149
11502.0.11 - Initial preparations for fake inode based attribute i/o.
1151
1152 - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
1153 do some macro magic (adapted from include/linux/buffer_head.h) to
1154 expand all the helper functions NInoFoo(), NInoSetFoo(), and
1155 NInoClearFoo().
1156 - Add new flag to ntfs_inode_state_bits: NI_Sparse.
1157 - Add new fields to ntfs_inode structure to allow use of fake inodes
1158 for attribute i/o: type, name, name_len. Also add new state bits:
1159 NI_Attr, which, if set, indicates the inode is a fake inode, and
1160 NI_MstProtected, which, if set, indicates the attribute uses multi
1161 sector transfer protection, i.e. fixups need to be applied after
1162 reads and before/after writes.
1163 - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
1164 ntfs_{new,clear,destroy}_extent_inode() and update callers.
1165 - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
1166 instead of ntfs_destroy_extent_inode().
1167 - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
1168 - Make all operations on ntfs inode state bits use the NIno* functions.
1169 - Set up the new ntfs inode fields and state bits in
1170 fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
1171 allocated memory to __ntfs_clear_inode().
1172 - Cleanup ntfs_inode structure a bit for better ordering of elements
1173 w.r.t. their size to allow better packing of the structure in memory.
1174
11752.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
1176
1177 - Add check at mount time to verify that the number of inodes on the
1178 volume does not exceed 2^32 - 1, which is the maximum allowed for
1179 NTFS according to Microsoft.
1180 - Change mft_no member of ntfs_inode structure to be unsigned long.
1181 Update all users. This makes ntfs_inode->mft_no just a copy of struct
1182 inode->i_ino. But we can't just always use struct inode->i_ino and
1183 remove mft_no because extent inodes do not have an attached struct
1184 inode.
1185
11862.0.9 - Decompression engine now uses a single buffer and other cleanups.
1187
1188 - Change decompression engine to use a single buffer protected by a
1189 spin lock instead of per-CPU buffers. (Rusty Russell)
1190 - Do not update cb_pos when handling a partial final page during
1191 decompression of a sparse compression block, as the value is later
1192 reset without being read/used. (Rusty Russell)
1193 - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
1194 Morton)
1195 - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
1196 NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
1197 it also makes everything safer so it is a good thing.
1198 - Miscellaneous minor cleanups to comments.
1199
12002.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
1201
1202 Big thanks go to Al Viro and other inhabitants of #kernel for investing
1203 their time to discuss the case sensitivity and dcache aliasing issues.
1204
1205 - Remove unused source file fs/ntfs/attraops.c.
1206 - Remove show_inodes mount option(s), thus dropping support for
1207 displaying of short file names.
1208 - Remove deprecated mount option posix.
1209 - Restore show_sys_files mount option.
1210 - Add new mount option case_sensitive, to determine if the driver
1211 treats file names as case sensitive or not. If case sensitive, create
1212 file names in the POSIX namespace. Otherwise create file names in the
1213 LONG/WIN32 namespace. Note, files remain accessible via their short
1214 file name, if it exists.
1215 - Remove really dumb logic bug in boot sector recovery code.
1216 - Fix dcache aliasing issues wrt short/long file names via changes
1217 to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
1218 fs/ntfs/namei.c::ntfs_lookup():
1219 - Add additional argument to ntfs_lookup_inode_by_name() in which we
1220 return information about the matching file name if the case is not
1221 matching or the match is a short file name. See comments above the
1222 function definition for details.
1223 - Change ntfs_lookup() to only create dcache entries for the correctly
1224 cased file name and only for the WIN32 namespace counterpart of DOS
1225 namespace file names. This ensures we have only one dentry per
1226 directory and also removes all dcache aliasing issues between short
1227 and long file names once we add write support. See comments above
1228 function for details.
1229 - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
1230
12312.0.7 - Minor cleanups and updates for changes in core kernel code.
1232
1233 - Remove much of the NULL struct element initializers.
1234 - Various updates to make compatible with recent kernels.
1235 - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
1236 in fs/ntfs/ntfs.h instead.
1237 - Remove no longer needed KERNEL_VERSION checks. We are now in the
1238 kernel proper so they are no longer needed.
1239
12402.0.6 - Major bugfix to make compatible with other kernel changes.
1241
1242 - Initialize the mftbmp address space properly now that there are more
1243 fields in the struct address_space. This was leading to hangs and
1244 oopses on umount since 2.5.12 because of changes to other parts of
1245 the kernel. We probably want a kernel generic init_address_space()
1246 function...
1247 - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
Anton Altaparmakov3672b632006-02-24 09:55:07 +00001248 only caller of ->readdir() is vfs_readdir() which holds i_mutex
1249 during the call, and i_mutex is sufficient protection against changes
1250 in the directory inode (including ->i_size).
Linus Torvalds1da177e2005-04-16 15:20:36 -07001251 - Use generic_file_llseek() for directories (as opposed to
Anton Altaparmakov3672b632006-02-24 09:55:07 +00001252 default_llseek()) as this downs i_mutex instead of the BKL which is
Linus Torvalds1da177e2005-04-16 15:20:36 -07001253 what we now need for exclusion against ->f_pos changes considering we
1254 no longer take the BKL in ntfs_readdir().
1255
12562.0.5 - Major bugfix. Buffer overflow in extent inode handling.
1257
1258 - No need to set old blocksize in super.c::ntfs_fill_super() as the
1259 VFS does so via invocation of deactivate_super() calling
1260 fs->fill_super() calling block_kill_super() which does it.
1261 - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
1262 -> Do we really need it? I don't think so as we have exclusion on
1263 the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
1264 move the ->f_pos accesses under the mrec_lock though. Check this...
1265 - Fix really, really, really stupid buffer overflow in extent inode
1266 handling in mft.c::map_extent_mft_record().
1267
12682.0.4 - Cleanups and updates for kernel 2.5.11.
1269
1270 - Add documentation on how to use the MD driver to be able to use NTFS
1271 stripe and volume sets in Linux and generally cleanup documentation
1272 a bit.
1273 Remove all uses of kdev_t in favour of struct block_device *:
1274 - Change compress.c::ntfs_file_read_compressed_block() to use
1275 sb_getblk() instead of getblk().
1276 - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
1277 of get_hardsect_size().
1278 - No need to get old blocksize in super.c::ntfs_fill_super() as
1279 fs/super.c::get_sb_bdev() already does this.
1280 - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
1281
12822.0.3 - Small bug fixes, cleanups, and performance improvements.
1283
1284 - Remove some dead code from mft.c.
1285 - Optimize readpage and read_block functions throughout aops.c so that
1286 only initialized blocks are read. Non-initialized ones have their
1287 buffer head mapped, zeroed, and set up to date, without scheduling
1288 any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
1289 Thanks go to Andrew Morton for spotting the below:
1290 - Fix buglet in allocate_compression_buffers() error code path.
1291 - Call flush_dcache_page() after modifying page cache page contents in
1292 ntfs_file_readpage().
1293 - Check for existence of page buffers throughout aops.c before calling
1294 create_empty_buffers(). This happens when an I/O error occurs and the
1295 read is retried. (It also happens once writing is implemented so that
1296 needed doing anyway but I had left it for later...)
1297 - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
1298 readpage and read_block functions. Reasoning same as above (i.e. I/O
1299 error retries and future write code paths.)
1300
13012.0.2 - Minor updates and cleanups.
1302
1303 - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
1304 and cleanup the code a bit, removing the unused size parameter.
1305 - Change default fmask to 0177 and update documentation.
1306 - Change attrib.c::get_attr_search_ctx() to return the search context
1307 directly instead of taking the address of a pointer. A return value
1308 of NULL means the allocation failed. Updated all callers
1309 appropriately.
1310 - Update to 2.5.9 kernel (preserving backwards compatibility) by
1311 replacing all occurences of page->buffers with page_buffers(page).
1312 - Fix minor bugs in runlist merging, also minor cleanup.
1313 - Updates to bootsector layout and mft mirror contents descriptions.
1314 - Small bug fix in error detection in unistr.c and some cleanups.
1315 - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
1316 bytes.
1317
13182.0.1 - Minor updates.
1319
1320 - Make default umask correspond to documentation.
1321 - Improve documentation.
1322 - Set default mode to include execute bit. The {u,f,d}mask can be used
1323 to take it away if desired. This allows binaries to be executed from
1324 a mounted ntfs partition.
1325
13262.0.0 - New version number. Remove TNG from the name. Now in the kernel.
1327
1328 - Add kill_super, just keeping up with the vfs changes in the kernel.
1329 - Repeat some changes from tng-0.0.8 that somehow got lost on the way
1330 from the CVS import into BitKeeper.
1331 - Begin to implement proper handling of allocated_size vs
1332 initialized_size vs data_size (i.e. i_size). Done are
1333 mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
1334 and attrib.c::load_attribute_list().
1335 - Lock the runlist in attrib.c::load_attribute_list() while using it.
1336 - Fix memory leak in ntfs_file_read_compressed_block() and generally
1337 clean up compress.c a little, removing some uncommented/unused debug
1338 code.
1339 - Tidy up dir.c a little bit.
1340 - Don't bother getting the runlist in inode.c::ntfs_read_inode().
1341 - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
1342 creating aops.c::ntfs_mst_readpage(), improving the handling of
1343 holes and overflow in the process and implementing the correct
1344 equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
1345 I am aiming for correctness at the moment. Modularisation can come
1346 later.
1347 - Rename aops.c::end_buffer_read_index_async() to
1348 end_buffer_read_mst_async() and optimize the overflow checking and
1349 handling.
1350 - Use the host of the mftbmp address space mapping to hold the ntfs
1351 volume. This is needed so the async i/o completion handler can
1352 retrieve a pointer to the volume. Hopefully this will not cause
1353 problems elsewhere in the kernel... Otherwise will need to use a
1354 fake inode.
1355 - Complete implementation of proper handling of allocated_size vs
1356 initialized_size vs data_size (i.e. i_size) in whole driver.
1357 Basically aops.c is now completely rewritten.
1358 - Change NTFS driver name to just NTFS and set version number to 2.0.0
1359 to make a clear distinction from the old driver which is still on
1360 version 1.1.22.
1361
1362tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
1363
1364 - Replace bdevname(sb->s_dev) with sb->s_id.
1365 - Remove now superfluous new-line characters in all callers of
1366 ntfs_debug().
1367 - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
1368 directories. Without this the "find" utility gets very upset which is
1369 fair enough as Linux/Unix do not support directory hard links.
1370 - Further runlist merging work. (Richard Russon)
1371 - Backwards compatibility for gcc-2.95. (Richard Russon)
1372 - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001373 - Convert to new filesystem declaration using ->ntfs_get_sb() and
Linus Torvalds1da177e2005-04-16 15:20:36 -07001374 replacing ntfs_read_super() with ntfs_fill_super().
1375 - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
1376 overflow on 32-bit architectures.
1377 - Cleanup upcase loading code to use ntfs_(un)map_page().
1378 - Disable/reenable preemtion in critical sections of compession engine.
1379 - Replace device size determination in ntfs_fill_super() with
1380 sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
1381 function super.c::get_nr_blocks().
1382 - Implement a mount time option (show_inodes) allowing choice of which
1383 types of inode names readdir() returns and modify ntfs_filldir()
1384 accordingly. There are several parameters to show_inodes:
1385 system: system files
1386 win32: long file names (including POSIX file names) [DEFAULT]
1387 long: same as win32
1388 dos: short file names only (excluding POSIX file names)
1389 short: same as dos
1390 posix: same as both win32 and dos
1391 all: all file names
1392 Note that the options are additive, i.e. specifying:
1393 -o show_inodes=system,show_inodes=win32,show_inodes=dos
1394 is the same as specifying:
1395 -o show_inodes=all
1396 Note that the "posix" and "all" options will show all directory
1397 names, BUT the link count on each directory inode entry is set to 1,
1398 due to Linux not supporting directory hard links. This may well
1399 confuse some userspace applications, since the directory names will
1400 have the same inode numbers. Thus it is NOT advisable to use the
1401 "posix" or "all" options. We provide them only for completeness sake.
1402 - Add copies of allocated_size, initialized_size, and compressed_size to
1403 the ntfs inode structure and set them up in
1404 inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
1405 for files and the index allocation attribute for directories.
1406 - Add copies of allocated_size and initialized_size to ntfs inode for
1407 $BITMAP attribute of large directories and set them up in
1408 inode.c::ntfs_read_inode().
1409 - Add copies of allocated_size and initialized_size to ntfs volume for
1410 $BITMAP attribute of $MFT and set them up in
1411 super.c::load_system_files().
1412 - Parse deprecated ntfs driver options (iocharset, show_sys_files,
1413 posix, and utf8) and tell user what the new options to use are. Note
1414 we still do support them but they will be removed with kernel 2.7.x.
1415 - Change all occurences of integer long long printf formatting to hex
1416 as printk() will not support long long integer format if/when the
1417 div64 patch goes into the kernel.
1418 - Make slab caches have stable names and change the names to what they
1419 were intended to be. These changes are required/made possible by the
1420 new slab cache name handling which removes the length limitation by
1421 requiring the caller of kmem_cache_create() to supply a stable name
1422 which is then referenced but not copied.
1423 - Rename run_list structure to run_list_element and create a new
1424 run_list structure containing a pointer to a run_list_element
1425 structure and a read/write semaphore. Adapt all users of runlists
1426 to new scheme and take and release the lock as needed. This fixes a
1427 nasty race as the run_list changes even when inodes are locked for
1428 reading and even when the inode isn't locked at all, so we really
1429 needed the serialization. We use a semaphore rather than a spinlock
1430 as memory allocations can sleep and doing everything GFP_ATOMIC
1431 would be silly.
1432 - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
1433 This can never happen due to the nature of lookup_attr() and how we
1434 support attribute lists. If it did happen it would imply the inode
1435 being corrupt.
1436 - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
1437 bad if found.
1438 - Update to 2.5.6-pre2 changes in struct address_space.
1439 - Use parent_ino() when accessing d_parent inode number in dir.c.
1440 - Import Sourceforge CVS repository into BitKeeper repository:
1441 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
1442 - Update fs/Makefile, fs/Config.help, fs/Config.in, and
1443 Documentation/filesystems/ntfs.txt for NTFS TNG.
1444 - Create kernel configuration option controlling whether debugging
1445 is enabled or not.
1446 - Add the required export of end_buffer_io_sync() from the patches
1447 directory to the kernel code.
1448 - Update inode.c::ntfs_show_options() with show_inodes mount option.
1449 - Update errors mount option.
1450
1451tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
1452
1453 - Cleanup mft.c and it's debug/error output in particular. Fix a minor
1454 bug in mapping of extent inodes. Update all the comments to fit all
1455 the recent code changes.
1456 - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
1457 - Cleanups in compress.c, mostly comments and folding help.
1458 - Implement attrib.c::map_run_list() as a generic helper.
1459 - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
1460 thus making code shorter and enabling attribute list support.
1461 - Cleanup incorrect use of [su]64 with %L printf format specifier in
1462 all source files. Type casts to [unsigned] long long added to correct
1463 the mismatches (important for architectures which have long long not
1464 being 64 bits).
1465 - Merge async io completion handlers for directory indexes and $MFT
1466 data into one by setting the index_block_size{_bits} of the ntfs
1467 inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
1468 - Cleanup aops.c, update comments.
1469 - Make ntfs_file_get_block() use map_run_list() so all files now
1470 support attribute lists.
1471 - Make ntfs_dir_readpage() almost verbatim copy of
1472 block_read_full_page() by using ntfs_file_get_block() with only real
1473 difference being the use of our own async io completion handler
1474 rather than the default one, thus reducing the amount of code and
1475 automatically enabling attribute list support for directory indices.
1476 - Fix bug in load_attribute_list() - forgot to call brelse in error
1477 code path.
1478 - Change parameters to find_attr() and lookup_attr(). We no longer
1479 pass in the upcase table and its length. These can be gotten from
1480 ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
1481 - Cleanups in attrib.c.
1482 - Implement merging of runlists, attrib.c::merge_run_lists() and its
1483 helpers. (Richard Russon)
1484 - Attribute lists part 2, attribute extents and multi part runlists:
1485 enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
1486 further runlist parts via attrib.c::map_run_list().
1487 - Tiny endianness bug fix in decompress_mapping_pairs().
1488
1489tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
1490
1491 - Enable encrypted directories. (Their index root is marked encrypted
1492 to indicate that new files in that directory should be created
1493 encrypted.)
1494 - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
1495 - Enable $Extend system directory. Most (if not all) extended system
1496 files do not have unnamed data attributes so ntfs_read_inode() had to
1497 special case them but that is ok, as the special casing recovery
1498 happens inside an error code path so there is zero slow down in the
1499 normal fast path. The special casing is done by introducing a new
1500 function inode.c::ntfs_is_extended_system_file() which checks if any
1501 of the hard links in the inode point to $Extend as being their parent
1502 directory and if they do we assume this is an extended system file.
1503 - Create a sysctl/proc interface to allow {dis,en}abling of debug output
1504 when compiled with -DDEBUG. Default is debug messages to be disabled.
1505 To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
1506 (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
1507 interface is enabled). Inspired by old ntfs driver.
1508 - Add debug_msgs insmod/kernel boot parameter to set whether debug
1509 messages are {dis,en}abled. This is useful to enable debug messages
1510 during ntfs initialization and is the only way to activate debugging
1511 when the sysctl interface is not enabled.
1512 - Cleanup debug output in various places.
1513 - Remove all dollar signs ($) from the source (except comments) to
1514 enable compilation on architectures whose gcc compiler does not
1515 support dollar signs in the names of variables/constants. Attribute
1516 types now start with AT_ instead of $ and $I30 is now just I30.
1517 - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
1518 - Load complete runlist for $MFT/$BITMAP during mount and cleanup
1519 access functions. This means we now cope with $MFT/$BITMAP being
1520 spread accross several mft records.
1521 - Disable modification of mft_zone_multiplier on remount. We can always
1522 reenable this later on if we really want to, but we will need to make
1523 sure we readjust the mft_zone size / layout accordingly.
1524
1525tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
1526
1527 - Use sb_set_blocksize() instead of set_blocksize() and verify the
1528 return value.
1529 - Use sb_bread() instead of bread() throughout.
1530 - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
1531 of a directory index block vcn. Apply resulting simplifications in
1532 dir.c everywhere.
1533 - Fix a small bug somewhere (but forgot what it was).
1534 - Change ntfs_{debug,error,warning} to enable gcc to do type checking
1535 on the printf-format parameter list and fix bugs reported by gcc
1536 as a result. (Richard Russon)
1537 - Move inode allocation strategy to Al's new stuff but maintain the
1538 divorce of ntfs_inode from struct inode. To achieve this we have two
1539 separate slab caches, one for big ntfs inodes containing a struct
1540 inode and pure ntfs inodes and at the same time fix some faulty
1541 error code paths in ntfs_read_inode().
1542 - Show mount options in proc (inode.c::ntfs_show_options()).
1543
1544tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
1545
1546 - Modified (un)map_mft_record functions to be common for read and write
1547 case. To specify which is which, added extra parameter at front of
1548 parameter list. Pass either READ or WRITE to this, each has the
1549 obvious meaning.
1550 - General cleanups to allow for easier folding in vi.
1551 - attrib.c::decompress_mapping_pairs() now accepts the old runlist
1552 argument, and invokes attrib.c::merge_run_lists() to merge the old
1553 and the new runlists.
1554 - Removed attrib.c::find_first_attr().
1555 - Implemented loading of attribute list and complete runlist for $MFT.
1556 This means we now cope with $MFT being spread across several mft
1557 records.
1558 - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
1559 - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
1560 - Make ntfs_volume be allocated via kmalloc() instead of using a slab
1561 cache. There are too little ntfs_volume structures at any one time
1562 to justify a private slab cache.
1563 - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
1564 Use KM_BIO_IRQ on advice from IRC/kernel...
1565 - Use ntfs_map_page() in map_mft_record() and create ->readpage method
1566 for reading $MFT (ntfs_mft_readpage). In the process create dedicated
1567 address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
1568 removed the now superfluous exports from the kernel core patch.
1569 - Fix a bug where kfree() was used insted of ntfs_free().
1570 - Change map_mft_record() to take ntfs_inode as argument instead of
1571 vfs inode. Dito for unmap_mft_record(). Adapt all callers.
1572 - Add pointer to ntfs_volume to ntfs_inode.
1573 - Add mft record number and sequence number to ntfs_inode. Stop using
1574 i_ino and i_generation for in-driver purposes.
1575 - Implement attrib.c::merge_run_lists(). (Richard Russon)
1576 - Remove use of proper inodes by extent inodes. Move i_ino and
1577 i_generation to ntfs_inode to do this. Apply simplifications that
1578 result and remove iget_no_wait(), etc.
1579 - Pass ntfs_inode everywhere in the driver (used to be struct inode).
1580 - Add reference counting in ntfs_inode for the ntfs inode itself and
1581 for the mapped mft record.
1582 - Extend mft record mapping so we can (un)map extent mft records (new
1583 functions (un)map_extent_mft_record), and so mappings are reference
1584 counted and don't have to happen twice if already mapped - just ref
1585 count increases.
1586 - Add -o iocharset as alias to -o nls for backwards compatibility.
1587 - The latest core patch is now tiny. In fact just a single additional
1588 export is necessary over the base kernel.
1589
1590tng-0.0.3 - Cleanups, enhancements, bug fixes.
1591
1592 - Work on attrib.c::decompress_mapping_pairs() to detect base extents
1593 and setup the runlist appropriately using knowledge provided by the
1594 sizes in the base attribute record.
1595 - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
1596 any more.
1597 - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
1598 page or use vmalloc depending on the amount of memory requested.
1599 - Cleanup error output. The __FUNCTION__ "(): " is now added
1600 automatically. Introduced a new header file debug.h to support this
1601 and also moved ntfs_debug() function into it.
1602 - Make reading of compressed files more intelligent and especially get
1603 rid of the vmalloc_nofs() from readpage(). This now uses per CPU
1604 buffers (allocated at first mount with cluster size <= 4kiB and
1605 deallocated on last umount with cluster size <= 4kiB), and
1606 asynchronous io for the compressed data using a list of buffer heads.
1607 Er, we use synchronous io as async io only works on whole pages
1608 covered by buffers and not on individual buffer heads...
1609 - Bug fix for reading compressed files with sparse compression blocks.
1610
1611tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
1612
1613 - Fixed handling of directories when cluster size exceeds index block
1614 size.
1615 - Hide DOS only name space directory entries from readdir() but allow
1616 them in lookup(). This should fix the problem that Linux doesn't
1617 support directory hard links, while still allowing access to entries
1618 via their short file name. This also has the benefit of mimicking
1619 what Windows users are used to, so it is the ideal solution.
1620 - Implemented sync_page everywhere so no more hangs in D state when
1621 waiting for a page.
1622 - Stop using bforget() in favour of brelse().
1623 - Stop locking buffers unnecessarily.
1624 - Implemented compressed files (inode->mapping contains uncompressed
1625 data, raw compressed data is currently bread() into a vmalloc()ed
1626 memory buffer).
1627 - Enable compressed directories. (Their index root is marked compressed
1628 to indicate that new files in that directory should be created
1629 compressed.)
1630 - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
1631 functions. (Thanks to Will Dyson for pointing this out.)
1632 - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
1633 ntfs_sb_info) out of the common inode and super_block structures and
1634 started using the generic_ip and generic_sbp pointers instead. This
1635 makes ntfs entirely private with respect to the kernel tree.
1636 - Detect compiler version and abort with error message if gcc less than
1637 2.96 is used.
1638 - Fix bug in name comparison function in unistr.c.
1639 - Implement attribute lists part 1, the infrastructure: search contexts
1640 and operations, find_external_attr(), lookup_attr()) and make the
1641 code use the infrastructure.
1642 - Fix stupid buffer overflow bug that became apparent on larger run
1643 list containing attributes.
1644 - Fix bugs in readdir() that became apparent on larger directories.
1645
1646 The driver is now really useful and survives the test
1647 find . -type f -exec md5sum "{}" \;
1648 without any error messages on a over 1GiB sized partition with >16k
1649 files on it, including compressed files and directories and many files
1650 and directories with attribute lists.
1651
1652tng-0.0.1 - The first useful version.
1653
1654 - Added ntfs_lookup().
1655 - Added default upcase generation and handling.
1656 - Added compile options to be shown on module init.
1657 - Many bug fixes that were "hidden" before.
1658 - Update to latest kernel.
1659 - Added ntfs_readdir().
1660 - Added file operations for mmap(), read(), open() and llseek(). We just
1661 use the generic ones. The whole point of going through implementing
1662 readpage() methods and where possible get_block() call backs is that
1663 this allows us to make use of the generic high level methods provided
1664 by the kernel.
1665
1666 The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
1667 though and it doesn't implement accesssing compressed files yet. Also,
1668 accessing files with attribute list attributes is not implemented yet
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001669 either. But for small or simple filesystems it should work and allow
Linus Torvalds1da177e2005-04-16 15:20:36 -07001670 you to list directories, use stat on directory entries and the file
1671 system, open, read, mmap and llseek around in files. A big mile stone
1672 has been reached!
1673
1674tng-0.0.0 - Initial version tag.
1675
1676 Initial driver implementation. The driver can mount and umount simple
Anton Altaparmakovc002f422005-02-03 12:02:56 +00001677 NTFS filesystems (i.e. ones without attribute lists in the system
Linus Torvalds1da177e2005-04-16 15:20:36 -07001678 files). If the mount fails there might be problems in the error handling
1679 code paths, so be warned. Otherwise it seems to be loading the system
1680 files nicely and the mft record read mapping/unmapping seems to be
1681 working nicely, too. Proof of inode metadata in the page cache and non-
1682 resident file unnamed stream data in the page cache concepts is thus
1683 complete.