The Design and Implementation of the FreeBSD Operating System, Second Edition
Now available: The Design and Implementation of the FreeBSD Operating System (Second Edition)


[ source navigation ] [ diff markup ] [ identifier search ] [ freetext search ] [ file search ] [ list types ] [ track identifier ]

FreeBSD/Linux Kernel Cross Reference
sys/fs/ntfs/ChangeLog

Version: -  FREEBSD  -  FREEBSD-12-STABLE  -  FREEBSD-12-0  -  FREEBSD-11-STABLE  -  FREEBSD-11-2  -  FREEBSD-11-1  -  FREEBSD-11-0  -  FREEBSD-10-STABLE  -  FREEBSD-10-4  -  FREEBSD-10-3  -  FREEBSD-10-2  -  FREEBSD-10-1  -  FREEBSD-10-0  -  FREEBSD-9-STABLE  -  FREEBSD-9-3  -  FREEBSD-9-2  -  FREEBSD-9-1  -  FREEBSD-9-0  -  FREEBSD-8-STABLE  -  FREEBSD-8-4  -  FREEBSD-8-3  -  FREEBSD-8-2  -  FREEBSD-8-1  -  FREEBSD-8-0  -  FREEBSD-7-STABLE  -  FREEBSD-7-4  -  FREEBSD-7-3  -  FREEBSD-7-2  -  FREEBSD-7-1  -  FREEBSD-7-0  -  FREEBSD-6-STABLE  -  FREEBSD-6-4  -  FREEBSD-6-3  -  FREEBSD-6-2  -  FREEBSD-6-1  -  FREEBSD-6-0  -  FREEBSD-5-STABLE  -  FREEBSD-5-5  -  FREEBSD-5-4  -  FREEBSD-5-3  -  FREEBSD-5-2  -  FREEBSD-5-1  -  FREEBSD-5-0  -  FREEBSD-4-STABLE  -  FREEBSD-3-STABLE  -  FREEBSD22  -  linux-2.6  -  linux-2.4.22  -  MK83  -  MK84  -  PLAN9  -  DFBSD  -  NETBSD  -  NETBSD5  -  NETBSD4  -  NETBSD3  -  NETBSD20  -  OPENBSD  -  xnu-517  -  xnu-792  -  xnu-792.6.70  -  xnu-1228  -  xnu-1456.1.26  -  xnu-1699.24.8  -  xnu-2050.18.24  -  OPENSOLARIS  -  minix-3-1-1 
SearchContext: -  none  -  3  -  10 

    1 ToDo/Notes:
    2         - Find and fix bugs.
    3         - The only places in the kernel where a file is resized are
    4           ntfs_file_write*() and ntfs_truncate() for both of which i_mutex is
    5           held.  Just have to be careful in read-/writepage and other helpers
    6           not running under i_mutex that we play nice.  Also need to be careful
    7           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.
   13         - 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.
   19 
   20 2.1.29 - Fix a deadlock at mount time.
   21 
   22         - During mount the VFS holds s_umount lock on the superblock.  So when
   23           we try to empty the journal $LogFile contents by calling
   24           ntfs_attr_set() when the machine does not have much memory and the
   25           journal is large ntfs_attr_set() results in the VM trying to balance
   26           dirty pages which in turn tries to that the s_umount lock and thus we
   27           get a deadlock.  The solution is to not use ntfs_attr_set() and
   28           instead do the zeroing by hand at the block level rather than page
   29           cache level.
   30         - Fix sparse warnings.
   31 
   32 2.1.28 - Fix a deadlock.
   33 
   34         - Fix deadlock in fs/ntfs/inode.c::ntfs_put_inode().  Thanks to Sergey
   35           Vlasov for the report and detailed analysis of the deadlock.  The fix
   36           involved getting rid of ntfs_put_inode() altogether and hence NTFS no
   37           longer has a ->put_inode super operation.
   38 
   39 2.1.27 - Various bug fixes and cleanups.
   40 
   41         - Fix two compiler warnings on Alpha.  Thanks to Andrew Morton for
   42           reporting them.
   43         - Fix an (innocent) off-by-one error in the runlist code.
   44         - Fix a buggette in an "should be impossible" case handling where we
   45           continued the attribute lookup loop instead of aborting it.
   46         - Use buffer_migrate_page() for the ->migratepage function of all ntfs
   47           address space operations.
   48         - Fix comparison of $MFT and $MFTMirr to not bail out when there are
   49           unused, invalid mft records which are the same in both $MFT and
   50           $MFTMirr.
   51         - Add support for sparse files which have a compression unit of 0.
   52         - Remove all the make_bad_inode() calls.  This should only be called
   53           from read inode and new inode code paths.
   54         - Limit name length in fs/ntfs/unistr.c::ntfs_nlstoucs() to maximum
   55           allowed by NTFS, i.e. 255 Unicode characters, not including the
   56           terminating NULL (which is not stored on disk).
   57         - Improve comments on file attribute flags in fs/ntfs/layout.h.
   58         - Fix a bug in fs/ntfs/inode.c::ntfs_read_locked_index_inode() where we
   59           forgot to update a temporary variable so loading index inodes which
   60           have an index allocation attribute failed.
   61         - Add a missing call to flush_dcache_mft_record_page() in
   62           fs/ntfs/inode.c::ntfs_write_inode().
   63         - Handle the recently introduced -ENAMETOOLONG return value from
   64           fs/ntfs/unistr.c::ntfs_nlstoucs() in fs/ntfs/namei.c::ntfs_lookup().
   65         - Semaphore to mutex conversion.  (Ingo Molnar)
   66 
   67 2.1.26 - Minor bug fixes and updates.
   68 
   69         - Fix a potential overflow in file.c where a cast to s64 was missing in
   70           a left shift of a page index.
   71         - The struct inode has had its i_sem semaphore changed to a mutex named
   72           i_mutex.
   73         - We have struct kmem_cache now so use it instead of the typedef
   74           kmem_cache_t.  (Pekka Enberg)
   75         - Implement support for sector sizes above 512 bytes (up to the maximum
   76           supported by NTFS which is 4096 bytes).
   77         - Do more detailed reporting of why we cannot mount read-write by
   78           special casing the VOLUME_MODIFIED_BY_CHKDSK flag.
   79         - Miscellaneous updates to layout.h.
   80         - Cope with attribute list attribute having invalid flags.  Windows
   81           copes with this and even chkdsk does not detect or fix this so we
   82           have to cope with it, too.  Thanks to Pawel Kot for reporting the
   83           problem.
   84 
   85 2.1.25 - (Almost) fully implement write(2) and truncate(2).
   86 
   87         - Change ntfs_map_runlist_nolock(), ntfs_attr_find_vcn_nolock() and
   88           {__,}ntfs_cluster_free() to also take an optional attribute search
   89           context as argument.  This allows calling these functions with the
   90           mft record mapped.  Update all callers.
   91         - Fix potential deadlock in ntfs_mft_data_extend_allocation_nolock()
   92           error handling by passing in the active search context when calling
   93           ntfs_cluster_free().
   94         - Change ntfs_cluster_alloc() to take an extra boolean parameter
   95           specifying whether the cluster are being allocated to extend an
   96           attribute or to fill a hole.
   97         - Change ntfs_attr_make_non_resident() to call ntfs_cluster_alloc()
   98           with @is_extension set to TRUE and remove the runlist terminator
   99           fixup code as this is now done by ntfs_cluster_alloc().
  100         - Change ntfs_attr_make_non_resident to take the attribute value size
  101           as an extra parameter.  This is needed since we need to know the size
  102           before we can map the mft record and our callers always know it.  The
  103           reason we cannot simply read the size from the vfs inode i_size is
  104           that this is not necessarily uptodate.  This happens when
  105           ntfs_attr_make_non_resident() is called in the ->truncate call path.
  106         - Fix ntfs_attr_make_non_resident() to update the vfs inode i_blocks
  107           which is zero for a resident attribute but should no longer be zero
  108           once the attribute is non-resident as it then has real clusters
  109           allocated.
  110         - Add fs/ntfs/attrib.[hc]::ntfs_attr_extend_allocation(), a function to
  111           extend the allocation of an attributes.  Optionally, the data size,
  112           but not the initialized size can be extended, too.
  113         - Implement fs/ntfs/inode.[hc]::ntfs_truncate().  It only supports
  114           uncompressed and unencrypted files and it never creates sparse files
  115           at least for the moment (making a file sparse requires us to modify
  116           its directory entries and we do not support directory operations at
  117           the moment).  Also, support for highly fragmented files, i.e. ones
  118           whose data attribute is split across multiple extents, is severly
  119           limited.  When such a case is encountered, EOPNOTSUPP is returned.
  120         - Enable ATTR_SIZE attribute changes in ntfs_setattr().  This completes
  121           the initial implementation of file truncation.  Now both open(2)ing
  122           a file with the O_TRUNC flag and the {,f}truncate(2) system calls
  123           will resize a file appropriately.  The limitations are that only
  124           uncompressed and unencrypted files are supported.  Also, there is
  125           only very limited support for highly fragmented files (the ones whose
  126           $DATA attribute is split into multiple attribute extents).
  127         - In attrib.c::ntfs_attr_set() call balance_dirty_pages_ratelimited()
  128           and cond_resched() in the main loop as we could be dirtying a lot of
  129           pages and this ensures we play nice with the VM and the system as a
  130           whole.
  131         - Implement file operations ->write, ->aio_write, ->writev for regular
  132           files.  This replaces the old use of generic_file_write(), et al and
  133           the address space operations ->prepare_write and ->commit_write.
  134           This means that both sparse and non-sparse (unencrypted and
  135           uncompressed) files can now be extended using the normal write(2)
  136           code path.  There are two limitations at present and these are that
  137           we never create sparse files and that we only have limited support
  138           for highly fragmented files, i.e. ones whose data attribute is split
  139           across multiple extents.   When such a case is encountered,
  140           EOPNOTSUPP is returned.
  141         - $EA attributes can be both resident and non-resident.
  142         - Use %z for size_t to fix compilation warnings.  (Andrew Morton)
  143         - Fix compilation warnings with gcc-4.0.2 on SUSE 10.0.
  144         - Document extended attribute ($EA) NEED_EA flag.  (Based on libntfs
  145           patch by Yura Pakhuchiy.)
  146 
  147 2.1.24 - Lots of bug fixes and support more clean journal states.
  148 
  149         - Support journals ($LogFile) which have been modified by chkdsk.  This
  150           means users can boot into Windows after we marked the volume dirty.
  151           The Windows boot will run chkdsk and then reboot.  The user can then
  152           immediately boot into Linux rather than having to do a full Windows
  153           boot first before rebooting into Linux and we will recognize such a
  154           journal and empty it as it is clean by definition.  Note, this only
  155           works if chkdsk left the journal in an obviously clean state.
  156         - Support journals ($LogFile) with only one restart page as well as
  157           journals with two different restart pages.  We sanity check both and
  158           either use the only sane one or the more recent one of the two in the
  159           case that both are valid.
  160         - Add fs/ntfs/malloc.h::ntfs_malloc_nofs_nofail() which is analogous to
  161           ntfs_malloc_nofs() but it performs allocations with __GFP_NOFAIL and
  162           hence cannot fail.
  163         - Use ntfs_malloc_nofs_nofail() in the two critical regions in
  164           fs/ntfs/runlist.c::ntfs_runlists_merge().  This means we no longer
  165           need to panic() if the allocation fails as it now cannot fail.
  166         - Fix two nasty runlist merging bugs that had gone unnoticed so far.
  167           Thanks to Stefano Picerno for the bug report.
  168         - Remove two bogus BUG_ON()s from fs/ntfs/mft.c.
  169         - Fix handling of valid but empty mapping pairs array in
  170           fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress().
  171         - Report unrepresentable inodes during ntfs_readdir() as KERN_WARNING
  172           messages and include the inode number.  Thanks to Yura Pakhuchiy for
  173           pointing this out.
  174         - Change ntfs_rl_truncate_nolock() to throw away the runlist if the new
  175           length is zero.
  176         - Add runlist.[hc]::ntfs_rl_punch_nolock() which punches a caller
  177           specified hole into a runlist.
  178         - Fix a bug in fs/ntfs/index.c::ntfs_index_lookup().  When the returned
  179           index entry is in the index root, we forgot to set the @ir pointer in
  180           the index context.  Thanks to Yura Pakhuchiy for finding this bug.
  181         - Remove bogus setting of PageError in ntfs_read_compressed_block().
  182         - Add fs/ntfs/attrib.[hc]::ntfs_resident_attr_value_resize().
  183         - Fix a bug in ntfs_map_runlist_nolock() where we forgot to protect
  184           access to the allocated size in the ntfs inode with the size lock.
  185         - Fix ntfs_attr_vcn_to_lcn_nolock() and ntfs_attr_find_vcn_nolock() to
  186           return LCN_ENOENT when there is no runlist and the allocated size is
  187           zero.
  188         - Fix load_attribute_list() to handle the case of a NULL runlist.
  189         - Fix handling of sparse attributes in ntfs_attr_make_non_resident().
  190         - Add BUG() checks to ntfs_attr_make_non_resident() and ntfs_attr_set()
  191           to ensure that these functions are never called for compressed or
  192           encrypted attributes.
  193         - Fix cluster (de)allocators to work when the runlist is NULL and more
  194           importantly to take a locked runlist rather than them locking it
  195           which leads to lock reversal.
  196         - Truncate {a,c,m}time to the ntfs supported time granularity when
  197           updating the times in the inode in ntfs_setattr().
  198         - Fixup handling of sparse, compressed, and encrypted attributes in
  199           fs/ntfs/inode.c::ntfs_read_locked_{,attr_,index_}inode(),
  200           fs/ntfs/aops.c::ntfs_{read,write}page().
  201         - Make ntfs_write_block() not instantiate sparse blocks if they contain
  202           only zeroes.
  203         - Optimize fs/ntfs/aops.c::ntfs_write_block() by extending the page
  204           lock protection over the buffer submission for i/o which allows the
  205           removal of the get_bh()/put_bh() pairs for each buffer.
  206         - Fix fs/ntfs/aops.c::ntfs_{read,write}_block() to handle the case
  207           where a concurrent truncate has truncated the runlist under our feet.
  208         - Fix page_has_buffers()/page_buffers() handling in fs/ntfs/aops.c.
  209         - In fs/ntfs/aops.c::ntfs_end_buffer_async_read(), use a bit spin lock
  210           in the first buffer head instead of a driver global spin lock to
  211           improve scalability.
  212         - Minor fix to error handling and error message display in
  213           fs/ntfs/aops.c::ntfs_prepare_nonresident_write().
  214         - Change the mount options {u,f,d}mask to always parse the number as
  215           an octal number to conform to how chmod(1) works, too.  Thanks to
  216           Giuseppe Bilotta and Horst von Brand for pointing out the errors of
  217           my ways.
  218         - Fix various bugs in the runlist merging code.  (Based on libntfs
  219           changes by Richard Russon.)
  220         - Fix sparse warnings that have crept in over time.
  221         - Change ntfs_cluster_free() to require a write locked runlist on entry
  222           since we otherwise get into a lock reversal deadlock if a read locked
  223           runlist is passed in. In the process also change it to take an ntfs
  224           inode instead of a vfs inode as parameter.
  225         - Fix the definition of the CHKD ntfs record magic.  It had an off by
  226           two error causing it to be CHKB instead of CHKD.
  227         - Fix a stupid bug in __ntfs_bitmap_set_bits_in_run() which caused the
  228           count to become negative and hence we had a wild memset() scribbling
  229           all over the system's ram.
  230 
  231 2.1.23 - Implement extension of resident files and make writing safe as well as
  232          many bug fixes, cleanups, and enhancements...
  233 
  234         - Add printk rate limiting for ntfs_warning() and ntfs_error() when
  235           compiled without debug.  This avoids a possible denial of service
  236           attack.  Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
  237           out.
  238         - Fix compilation warnings on ia64.  (Randy Dunlap)
  239         - Use i_size_{read,write}() instead of reading i_size by hand and cache
  240           the value where apropriate.
  241         - Add size_lock to the ntfs_inode structure.  This is an rw spinlock
  242           and it locks against access to the inode sizes.  Note, ->size_lock
  243           is also accessed from irq context so you must use the _irqsave and
  244           _irqrestore lock and unlock functions, respectively.  Protect all
  245           accesses to allocated_size, initialized_size, and compressed_size.
  246         - Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
  247         - Implement extension of resident files in the regular file write code
  248           paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()).  At present
  249           this only works until the data attribute becomes too big for the mft
  250           record after which we abort the write returning -EOPNOTSUPP from
  251           ntfs_prepare_write().
  252         - Add disable_sparse mount option together with a per volume sparse
  253           enable bit which is set appropriately and a per inode sparse disable
  254           bit which is preset on some system file inodes as appropriate.
  255         - Enforce that sparse support is disabled on NTFS volumes pre 3.0.
  256         - Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
  257           the creation of the unmapped runlist element for the base attribute
  258           extent.
  259         - Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
  260           helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
  261           This allows us to map runlist fragments with the runlist lock already
  262           held without having to drop and reacquire it around the call.  Adapt
  263           all callers.
  264         - Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
  265           runlist.  This allows us to find runlist elements with the runlist
  266           lock already held without having to drop and reacquire it around the
  267           call.  Adapt all callers.
  268         - Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
  269           warning in the do_div() call on sparc32.  Thanks to Meelis Roos for
  270           the report and analysis of the warning.
  271         - Fix a nasty runlist merge bug when merging two holes.
  272         - Set the ntfs_inode->allocated_size to the real allocated size in the
  273           mft record for resident attributes (fs/ntfs/inode.c).
  274         - Small readability cleanup to use "a" instead of "ctx->attr"
  275           everywhere (fs/ntfs/inode.c).
  276         - Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
  277           definition of ntfs_export_ops from fs/ntfs/super.c to namei.c.  Also,
  278           declare ntfs_export_ops in fs/ntfs/ntfs.h.
  279         - Correct sparse file handling.  The compressed values need to be
  280           checked and set in the ntfs inode as done for compressed files and
  281           the compressed size needs to be used for vfs inode->i_blocks instead
  282           of the allocated size, again, as done for compressed files.
  283         - Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
  284           non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
  285         - Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
  286           write code.
  287         - Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
  288           dropping the read lock and taking the write lock we were not checking
  289           whether someone else did not already do the work we wanted to do.
  290         - Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
  291           ntfs_attr_find_vcn_nolock() and update all callers.
  292         - Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
  293         - Fix sign of various error return values to be negative in
  294           fs/ntfs/lcnalloc.c.
  295         - Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
  296           handle the case where an attribute is converted from resident to
  297           non-resident by a concurrent file write.
  298         - Remove checks for NULL before calling kfree() since kfree() does the
  299           checking itself.  (Jesper Juhl)
  300         - Some utilities modify the boot sector but do not update the checksum.
  301           Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
  302           only emit a warning when the checksum is incorrect rather than
  303           refusing the mount.  Thanks to Bernd Casimir for pointing this
  304           problem out.
  305         - Update attribute definition handling.
  306         - Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
  307         - Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
  308         - Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
  309           better code generation and one less sparse warning in fs/ntfs/aops.c.
  310         - Remove spurious void pointer casts from fs/ntfs/.  (Pekka Enberg)
  311         - Use C99 style structure initialization after memory allocation where
  312           possible (fs/ntfs/{attrib.c,index.c,super.c}).  Thanks to Al Viro and
  313           Pekka Enberg.
  314         - Stamp the transaction log ($UsnJrnl), aka user space journal, if it
  315           is active on the volume and we are mounting read-write or remounting
  316           from read-only to read-write.
  317         - Fix a bug in address space operations error recovery code paths where
  318           if the runlist was not mapped at all and a mapping error occured we
  319           would leave the runlist locked on exit to the function so that the
  320           next access to the same file would try to take the lock and deadlock.
  321         - Detect the case when Windows has been suspended to disk on the volume
  322           to be mounted and if this is the case do not allow (re)mounting
  323           read-write.  This is done by parsing hiberfil.sys if present.
  324         - Fix several occurences of a bug where we would perform 'var & ~const'
  325           with a 64-bit variable and a int, i.e. 32-bit, constant.  This causes
  326           the higher order 32-bits of the 64-bit variable to be zeroed.  To fix
  327           this cast the 'const' to the same 64-bit type as 'var'.
  328         - Change the runlist terminator of the newly allocated cluster(s) to
  329           LCN_ENOENT in ntfs_attr_make_non_resident().  Otherwise the runlist
  330           code gets confused.
  331         - Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
  332           and ntfs_mapping_pairs_build() to allow the runlist encoding to be
  333           partial which is desirable when filling holes in sparse attributes.
  334           Update all callers.
  335         - Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
  336           if the requested vcn is inside it.  Otherwise we get into problems
  337           when we try to map an out of bounds vcn because we then try to map
  338           the already mapped runlist fragment which causes
  339           ntfs_mapping_pairs_decompress() to fail and return error.  Update
  340           ntfs_attr_find_vcn_nolock() accordingly.
  341         - Fix a nasty deadlock that appeared in recent kernels.
  342           The situation: VFS inode X on a mounted ntfs volume is dirty.  For
  343           same inode X, the ntfs_inode is dirty and thus corresponding on-disk
  344           inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
  345           to the table of inodes, i.e. $MFT, inode 0.
  346           What happens:
  347           Process 1: sys_sync()/umount()/whatever...  calls
  348           __sync_single_inode() for $MFT -> do_writepages() -> write_page for
  349           the dirty page containing the on-disk inode X, the page is now locked
  350           -> ntfs_write_mst_block() which clears PageUptodate() on the page to
  351           prevent anyone else getting hold of it whilst it does the write out.
  352           This is necessary as the on-disk inode needs "fixups" applied before
  353           the write to disk which are removed again after the write and
  354           PageUptodate is then set again.  It then analyses the page looking
  355           for dirty on-disk inodes and when it finds one it calls
  356           ntfs_may_write_mft_record() to see if it is safe to write this
  357           on-disk inode.  This then calls ilookup5() to check if the
  358           corresponding VFS inode is in icache().  This in turn calls ifind()
  359           which waits on the inode lock via wait_on_inode whilst holding the
  360           global inode_lock.
  361           Process 2: pdflush results in a call to __sync_single_inode for the
  362           same VFS inode X on the ntfs volume.  This locks the inode (I_LOCK)
  363           then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
  364           read_cache_page() for the page (in page cache of table of inodes
  365           $MFT, inode 0) containing the on-disk inode.  This page has
  366           PageUptodate() clear because of Process 1 (see above) so
  367           read_cache_page() blocks when it tries to take the page lock for the
  368           page so it can call ntfs_read_page().
  369           Thus Process 1 is holding the page lock on the page containing the
  370           on-disk inode X and it is waiting on the inode X to be unlocked in
  371           ifind() so it can write the page out and then unlock the page.
  372           And Process 2 is holding the inode lock on inode X and is waiting for
  373           the page to be unlocked so it can call ntfs_readpage() or discover
  374           that Process 1 set PageUptodate() again and use the page.
  375           Thus we have a deadlock due to ifind() waiting on the inode lock.
  376           The solution: The fix is to use the newly introduced
  377           ilookup5_nowait() which does not wait on the inode's lock and hence
  378           avoids the deadlock.  This is safe as we do not care about the VFS
  379           inode and only use the fact that it is in the VFS inode cache and the
  380           fact that the vfs and ntfs inodes are one struct in memory to find
  381           the ntfs inode in memory if present.  Also, the ntfs inode has its
  382           own locking so it does not matter if the vfs inode is locked.
  383         - Fix bug in mft record writing where we forgot to set the device in
  384           the buffers when mapping them after the VM had discarded them.
  385           Thanks to Martin MOKREJÅ for the bug report.
  386 
  387 2.1.22 - Many bug and race fixes and error handling improvements.
  388 
  389         - Improve error handling in fs/ntfs/inode.c::ntfs_truncate().
  390         - Change fs/ntfs/inode.c::ntfs_truncate() to return an error code
  391           instead of void and provide a helper ntfs_truncate_vfs() for the
  392           vfs ->truncate method.
  393         - Add a new ntfs inode flag NInoTruncateFailed() and modify
  394           fs/ntfs/inode.c::ntfs_truncate() to set and clear it appropriately.
  395         - Fix min_size and max_size definitions in ATTR_DEF structure in
  396           fs/ntfs/layout.h to be signed.
  397         - Add attribute definition handling helpers to fs/ntfs/attrib.[hc]:
  398           ntfs_attr_size_bounds_check(), ntfs_attr_can_be_non_resident(), and
  399           ntfs_attr_can_be_resident(), which in turn use the new private helper
  400           ntfs_attr_find_in_attrdef().
  401         - In fs/ntfs/aops.c::mark_ntfs_record_dirty(), take the
  402           mapping->private_lock around the dirtying of the buffer heads
  403           analagous to the way it is done in __set_page_dirty_buffers().
  404         - Ensure the mft record size does not exceed the PAGE_CACHE_SIZE at
  405           mount time as this cannot work with the current implementation.
  406         - Check for location of attribute name and improve error handling in
  407           general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
  408         - In fs/ntfs/aops.c::ntfs_writepage(), if the page is fully outside
  409           i_size, i.e. race with truncate, invalidate the buffers on the page
  410           so that they become freeable and hence the page does not leak.
  411         - Remove unused function fs/ntfs/runlist.c::ntfs_rl_merge().  (Adrian
  412           Bunk)
  413         - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_find() that resulted in
  414           a NULL pointer dereference in the error code path when a corrupt
  415           attribute was found.  (Thanks to Domen Puncer for the bug report.)
  416         - Add MODULE_VERSION() to fs/ntfs/super.c.
  417         - Make several functions and variables static.  (Adrian Bunk)
  418         - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() so it allocates
  419           buffers for the page if they are not present and then marks the
  420           buffers belonging to the ntfs record dirty.  This causes the buffers
  421           to become busy and hence they are safe from removal until the page
  422           has been written out.
  423         - Fix stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() in the
  424           error handling code path that resulted in a BUG() due to trying to
  425           unmap an extent mft record when the mapping of it had failed and it
  426           thus was not mapped.  (Thanks to Ken MacFerrin for the bug report.)
  427         - Drop the runlist lock after the vcn has been read in
  428           fs/ntfs/lcnalloc.c::__ntfs_cluster_free().
  429         - Rewrite handling of multi sector transfer errors.  We now do not set
  430           PageError() when such errors are detected in the async i/o handler
  431           fs/ntfs/aops.c::ntfs_end_buffer_async_read().  All users of mst
  432           protected attributes now check the magic of each ntfs record as they
  433           use it and act appropriately.  This has the effect of making errors
  434           granular per ntfs record rather than per page which solves the case
  435           where we cannot access any of the ntfs records in a page when a
  436           single one of them had an mst error.  (Thanks to Ken MacFerrin for
  437           the bug report.)
  438         - Fix error handling in fs/ntfs/quota.c::ntfs_mark_quotas_out_of_date()
  439           where we failed to release i_mutex on the $Quota/$Q attribute inode.
  440         - Fix bug in handling of bad inodes in fs/ntfs/namei.c::ntfs_lookup().
  441         - Add mapping of unmapped buffers to all remaining code paths, i.e.
  442           fs/ntfs/aops.c::ntfs_write_mst_block(), mft.c::ntfs_sync_mft_mirror(),
  443           and write_mft_record_nolock().  From now on we require that the
  444           complete runlist for the mft mirror is always mapped into memory.
  445         - Add creation of buffers to fs/ntfs/mft.c::ntfs_sync_mft_mirror().
  446         - Improve error handling in fs/ntfs/aops.c::ntfs_{read,write}_block().
  447         - Cleanup fs/ntfs/aops.c::ntfs_{read,write}page() since we know that a
  448           resident attribute will be smaller than a page which makes the code
  449           simpler.  Also make the code more tolerant to concurrent ->truncate.
  450 
  451 2.1.21 - Fix some races and bugs, rewrite mft write code, add mft allocator.
  452 
  453         - Implement extent mft record deallocation
  454           fs/ntfs/mft.c::ntfs_extent_mft_record_free().
  455         - Splitt runlist related functions off from attrib.[hc] to runlist.[hc].
  456         - Add vol->mft_data_pos and initialize it at mount time.
  457         - Rename init_runlist() to ntfs_init_runlist(), ntfs_vcn_to_lcn() to
  458           ntfs_rl_vcn_to_lcn(), decompress_mapping_pairs() to
  459           ntfs_mapping_pairs_decompress(), ntfs_merge_runlists() to
  460           ntfs_runlists_merge() and adapt all callers.
  461         - Add fs/ntfs/runlist.[hc]::ntfs_get_nr_significant_bytes(),
  462           ntfs_get_size_for_mapping_pairs(), ntfs_write_significant_bytes(),
  463           and ntfs_mapping_pairs_build(), adapted from libntfs.
  464         - Make fs/ntfs/lcnalloc.c::ntfs_cluster_free_from_rl_nolock() not
  465           static and add a declaration for it to lcnalloc.h.
  466         - Add fs/ntfs/lcnalloc.h::ntfs_cluster_free_from_rl() which is a static
  467           inline wrapper for ntfs_cluster_free_from_rl_nolock() which takes the
  468           cluster bitmap lock for the duration of the call.
  469         - Add fs/ntfs/attrib.[hc]::ntfs_attr_record_resize().
  470         - Implement the equivalent of memset() for an ntfs attribute in
  471           fs/ntfs/attrib.[hc]::ntfs_attr_set() and switch
  472           fs/ntfs/logfile.c::ntfs_empty_logfile() to using it.
  473         - Remove unnecessary casts from LCN_* constants.
  474         - Implement fs/ntfs/runlist.c::ntfs_rl_truncate_nolock().
  475         - Add MFT_RECORD_OLD as a copy of MFT_RECORD in fs/ntfs/layout.h and
  476           change MFT_RECORD to contain the NTFS 3.1+ specific fields.
  477         - Add a helper function fs/ntfs/aops.c::mark_ntfs_record_dirty() which
  478           marks all buffers belonging to an ntfs record dirty, followed by
  479           marking the page the ntfs record is in dirty and also marking the vfs
  480           inode containing the ntfs record dirty (I_DIRTY_PAGES).
  481         - Switch fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to using the
  482           new helper fs/ntfs/aops.c::mark_ntfs_record_dirty() and remove the no
  483           longer needed fs/ntfs/index.[hc]::__ntfs_index_entry_mark_dirty().
  484         - Move ntfs_{un,}map_page() from ntfs.h to aops.h and fix resulting
  485           include errors.
  486         - Move the typedefs for runlist_element and runlist from types.h to
  487           runlist.h and fix resulting include errors.
  488         - Remove unused {__,}format_mft_record() from fs/ntfs/mft.c.
  489         - Modify fs/ntfs/mft.c::__mark_mft_record_dirty() to use the helper
  490           mark_ntfs_record_dirty() which also changes the behaviour in that we
  491           now set the buffers belonging to the mft record dirty as well as the
  492           page itself.
  493         - Update fs/ntfs/mft.c::write_mft_record_nolock() and sync_mft_mirror()
  494           to cope with the fact that there now are dirty buffers in mft pages.
  495         - Update fs/ntfs/inode.c::ntfs_write_inode() to also use the helper
  496           mark_ntfs_record_dirty() and thus to set the buffers belonging to the
  497           mft record dirty as well as the page itself.
  498         - Fix compiler warnings on x86-64 in fs/ntfs/dir.c.  (Randy Dunlap,
  499           slightly modified by me)
  500         - Add fs/ntfs/mft.c::try_map_mft_record() which fails with -EALREADY if
  501           the mft record is already locked and otherwise behaves the same way
  502           as fs/ntfs/mft.c::map_mft_record().
  503         - Modify fs/ntfs/mft.c::write_mft_record_nolock() so that it only
  504           writes the mft record if the buffers belonging to it are dirty.
  505           Otherwise we assume that it was written out by other means already.
  506         - Attempting to write outside initialized size is _not_ a bug so remove
  507           the bug check from fs/ntfs/aops.c::ntfs_write_mst_block().  It is in
  508           fact required to write outside initialized size when preparing to
  509           extend the initialized size.
  510         - Map the page instead of using page_address() before writing to it in
  511           fs/ntfs/aops.c::ntfs_mft_writepage().
  512         - Provide exclusion between opening an inode / mapping an mft record
  513           and accessing the mft record in fs/ntfs/mft.c::ntfs_mft_writepage()
  514           by setting the page not uptodate throughout ntfs_mft_writepage().
  515         - Clear the page uptodate flag in fs/ntfs/aops.c::ntfs_write_mst_block()
  516           to ensure noone can see the page whilst the mst fixups are applied.
  517         - Add the helper fs/ntfs/mft.c::ntfs_may_write_mft_record() which
  518           checks if an mft record may be written out safely obtaining any
  519           necessary locks in the process.  This is used by
  520           fs/ntfs/aops.c::ntfs_write_mst_block().
  521         - Modify fs/ntfs/aops.c::ntfs_write_mst_block() to also work for
  522           writing mft records and improve its error handling in the process.
  523           Now if any of the records in the page fail to be written out, all
  524           other records will be written out instead of aborting completely.
  525         - Remove ntfs_mft_aops and update all users to use ntfs_mst_aops.
  526         - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to set the
  527           ntfs_mst_aops for all inodes which are NInoMstProtected() and
  528           ntfs_aops for all other inodes.
  529         - Rename fs/ntfs/mft.c::sync_mft_mirror{,_umount}() to
  530           ntfs_sync_mft_mirror{,_umount}() and change their parameters so they
  531           no longer require an ntfs inode to be present.  Update all callers.
  532         - Cleanup the error handling in fs/ntfs/mft.c::ntfs_sync_mft_mirror().
  533         - Clear the page uptodate flag in fs/ntfs/mft.c::ntfs_sync_mft_mirror()
  534           to ensure noone can see the page whilst the mst fixups are applied.
  535         - Remove the no longer needed fs/ntfs/mft.c::ntfs_mft_writepage() and
  536           fs/ntfs/mft.c::try_map_mft_record().
  537         - Fix callers of fs/ntfs/aops.c::mark_ntfs_record_dirty() to call it
  538           with the ntfs inode which contains the page rather than the ntfs
  539           inode the mft record of which is in the page.
  540         - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by moving the
  541           index inode bitmap inode release code from there to
  542           fs/ntfs/inode.c::ntfs_clear_big_inode().  (Thanks to Christoph
  543           Hellwig for spotting this.)
  544         - Fix race condition in fs/ntfs/inode.c::ntfs_put_inode() by taking the
  545           inode semaphore around the code that sets ni->itype.index.bmp_ino to
  546           NULL and reorganize the code to optimize it a bit.  (Thanks to
  547           Christoph Hellwig for spotting this.)
  548         - Modify fs/ntfs/aops.c::mark_ntfs_record_dirty() to no longer take the
  549           ntfs inode as a parameter as this is confusing and misleading and the
  550           needed ntfs inode is available via NTFS_I(page->mapping->host).
  551           Adapt all callers to this change.
  552         - Modify fs/ntfs/mft.c::write_mft_record_nolock() and
  553           fs/ntfs/aops.c::ntfs_write_mst_block() to only check the dirty state
  554           of the first buffer in a record and to take this as the ntfs record
  555           dirty state.  We cannot look at the dirty state for subsequent
  556           buffers because we might be racing with
  557           fs/ntfs/aops.c::mark_ntfs_record_dirty().
  558         - Move the static inline ntfs_init_big_inode() from fs/ntfs/inode.c to
  559           inode.h and make fs/ntfs/inode.c::__ntfs_init_inode() non-static and
  560           add a declaration for it to inode.h.  Fix some compilation issues
  561           that resulted due to #includes and header file interdependencies.
  562         - Simplify setup of i_mode in fs/ntfs/inode.c::ntfs_read_locked_inode().
  563         - Add helpers fs/ntfs/layout.h::MK_MREF() and MK_LE_MREF().
  564         - Modify fs/ntfs/mft.c::map_extent_mft_record() to only verify the mft
  565           record sequence number if it is specified (i.e. not zero).
  566         - Add fs/ntfs/mft.[hc]::ntfs_mft_record_alloc() and various helper
  567           functions used by it.
  568         - Update Documentation/filesystems/ntfs.txt with instructions on how to
  569           use the Device-Mapper driver with NTFS ftdisk/LDM raid.  This removes
  570           the linear raid problem with the Software RAID / MD driver when one
  571           or more of the devices has an odd number of sectors.
  572 
  573 2.1.20 - Fix two stupid bugs introduced in 2.1.18 release.
  574 
  575         - Fix stupid bug in fs/ntfs/attrib.c::ntfs_attr_reinit_search_ctx()
  576           where we did not clear ctx->al_entry but it was still set due to
  577           changes in ntfs_attr_lookup() and ntfs_external_attr_find() in
  578           particular.
  579         - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find()
  580           where we forgot to unmap the extent mft record when we had finished
  581           enumerating an attribute which caused a bug check to trigger when the
  582           VFS calls ->clear_inode.
  583 
  584 2.1.19 - Many cleanups, improvements, and a minor bug fix.
  585 
  586         - Update ->setattr (fs/ntfs/inode.c::ntfs_setattr()) to refuse to
  587           change the uid, gid, and mode of an inode as we do not support NTFS
  588           ACLs yet.
  589         - Remove BKL use from ntfs_setattr() syncing up with the rest of the
  590           kernel.
  591         - Get rid of the ugly transparent union in fs/ntfs/dir.c::ntfs_readdir()
  592           and ntfs_filldir() as per suggestion from Al Viro.
  593         - Change '\0' and L'\0' to simply 0 as per advice from Linus Torvalds.
  594         - Update ->truncate (fs/ntfs/inode.c::ntfs_truncate()) to check if the
  595           inode size has changed and to only output an error if so.
  596         - Rename fs/ntfs/attrib.h::attribute_value_length() to ntfs_attr_size().
  597         - Add le{16,32,64} as well as sle{16,32,64} data types to
  598           fs/ntfs/types.h.
  599         - Change ntfschar to be le16 instead of u16 in fs/ntfs/types.h.
  600         - Add le versions of VCN, LCN, and LSN called leVCN, leLCN, and leLSN,
  601           respectively, to fs/ntfs/types.h.
  602         - Update endianness conversion macros in fs/ntfs/endian.h to use the
  603           new types as appropriate.
  604         - Do proper type casting when using sle64_to_cpup() in fs/ntfs/dir.c
  605           and index.c.
  606         - Add leMFT_REF data type to fs/ntfs/layout.h.
  607         - Update all NTFS header files with the new little endian data types.
  608           Affected files are fs/ntfs/layout.h, logfile.h, and time.h.
  609         - Do proper type casting when using ntfs_is_*_recordp() in
  610           fs/ntfs/logfile.c, mft.c, and super.c. 
  611         - Fix all the sparse bitwise warnings.  Had to change all the typedef
  612           enums storing little endian values to simple enums plus a typedef for
  613           the datatype to make sparse happy.
  614         - Fix a bug found by the new sparse bitwise warnings where the default
  615           upcase table was defined as a pointer to wchar_t rather than ntfschar
  616           in fs/ntfs/ntfs.h and super.c.
  617         - Change {const_,}cpu_to_le{16,32}(0) to just 0 as suggested by Al Viro.
  618 
  619 2.1.18 - Fix scheduling latencies at mount time as well as an endianness bug.
  620 
  621         - Remove vol->nr_mft_records as it was pretty meaningless and optimize
  622           the calculation of total/free inodes as used by statfs().
  623         - Fix scheduling latencies in ntfs_fill_super() by dropping the BKL
  624           because the code itself is using the ntfs_lock semaphore which
  625           provides safe locking.  (Ingo Molnar)
  626         - Fix a potential bug in fs/ntfs/mft.c::map_extent_mft_record() that
  627           could occur in the future for when we start closing/freeing extent
  628           inodes if we don't set base_ni->ext.extent_ntfs_inos to NULL after
  629           we free it.
  630         - Rename {find,lookup}_attr() to ntfs_attr_{find,lookup}() as well as
  631           find_external_attr() to ntfs_external_attr_find() to cleanup the
  632           namespace a bit and to be more consistent with libntfs.
  633         - Rename {{re,}init,get,put}_attr_search_ctx() to
  634           ntfs_attr_{{re,}init,get,put}_search_ctx() as well as the type
  635           attr_search_context to ntfs_attr_search_ctx.
  636         - Force use of ntfs_attr_find() in ntfs_attr_lookup() when searching
  637           for the attribute list attribute itself.
  638         - Fix endianness bug in ntfs_external_attr_find().
  639         - Change ntfs_{external_,}attr_find() to return 0 on success, -ENOENT
  640           if the attribute is not found, and -EIO on real error.  In the case
  641           of -ENOENT, the search context is updated to describe the attribute
  642           before which the attribute being searched for would need to be
  643           inserted if such an action were to be desired and in the case of
  644           ntfs_external_attr_find() the search context is also updated to
  645           indicate the attribute list entry before which the attribute list
  646           entry of the attribute being searched for would need to be inserted
  647           if such an action were to be desired.  Also make ntfs_find_attr()
  648           static and remove its prototype from attrib.h as it is not used
  649           anywhere other than attrib.c.  Update ntfs_attr_lookup() and all
  650           callers of ntfs_{external,}attr_{find,lookup}() for the new return
  651           values.
  652         - Minor cleanup of fs/ntfs/inode.c::ntfs_init_locked_inode().
  653 
  654 2.1.17 - Fix bugs in mount time error code paths and other updates.
  655 
  656         - Implement bitmap modification code (fs/ntfs/bitmap.[hc]).  This
  657           includes functions to set/clear a single bit or a run of bits.
  658         - Add fs/ntfs/attrib.[hc]::ntfs_find_vcn() which returns the locked
  659           runlist element containing a particular vcn.  It also takes care of
  660           mapping any needed runlist fragments.
  661         - Implement cluster (de-)allocation code (fs/ntfs/lcnalloc.[hc]).
  662         - Load attribute definition table from $AttrDef at mount time.
  663         - Fix bugs in mount time error code paths involving (de)allocation of
  664           the default and volume upcase tables.
  665         - Remove ntfs_nr_mounts as it is no longer used.
  666 
  667 2.1.16 - Implement access time updates, file sync, async io, and read/writev.
  668 
  669         - Add support for readv/writev and aio_read/aio_write (fs/ntfs/file.c).
  670           This is done by setting the appropriate file operations pointers to
  671           the generic helper functions provided by mm/filemap.c.
  672         - Implement fsync, fdatasync, and msync both for files (fs/ntfs/file.c)
  673           and directories (fs/ntfs/dir.c).
  674         - Add support for {a,m,c}time updates to inode.c::ntfs_write_inode().
  675           Note, except for the root directory and any other system files opened
  676           by the user, the system files will not have their access times
  677           updated as they are only accessed at the inode level an hence the
  678           file level functions which cause the times to be updated are never
  679           invoked.
  680 
  681 2.1.15 - Invalidate quotas when (re)mounting read-write.
  682 
  683         - Add new element itype.index.collation_rule to the ntfs inode
  684           structure and set it appropriately in ntfs_read_locked_inode().
  685         - Implement a new inode type "index" to allow efficient access to the
  686           indices found in various system files and adapt inode handling
  687           accordingly (fs/ntfs/inode.[hc]).  An index inode is essentially an
  688           attribute inode (NInoAttr() is true) with an attribute type of
  689           AT_INDEX_ALLOCATION.  As such, it is no longer allowed to call
  690           ntfs_attr_iget() with an attribute type of AT_INDEX_ALLOCATION as
  691           there would be no way to distinguish between normal attribute inodes
  692           and index inodes.  The function to obtain an index inode is
  693           ntfs_index_iget() and it uses the helper function
  694           ntfs_read_locked_index_inode().  Note, we do not overload
  695           ntfs_attr_iget() as indices consist of multiple attributes so using
  696           ntfs_attr_iget() to obtain an index inode would be confusing.
  697         - Ensure that there is no overflow when doing page->index <<
  698           PAGE_CACHE_SHIFT by casting page->index to s64 in fs/ntfs/aops.c.
  699         - Use atomic kmap instead of kmap() in fs/ntfs/aops.c::ntfs_read_page()
  700           and ntfs_read_block().
  701         - Use case sensitive attribute lookups instead of case insensitive ones.
  702         - Lock all page cache pages belonging to mst protected attributes while
  703           accessing them to ensure we never see corrupt data while the page is
  704           under writeout.
  705         - Add framework for generic ntfs collation (fs/ntfs/collation.[hc]).
  706           We have ntfs_is_collation_rule_supported() to check if the collation
  707           rule you want to use is supported and ntfs_collation() which actually
  708           collates two data items.  We currently only support COLLATION_BINARY
  709           and COLLATION_NTOFS_ULONG but support for other collation rules will
  710           be added as the need arises.
  711         - Add a new type, ntfs_index_context, to allow retrieval of an index
  712           entry using the corresponding index key.  To get an index context,
  713           use ntfs_index_ctx_get() and to release it, use ntfs_index_ctx_put().
  714           This also adds a new slab cache for the index contexts.  To lookup a
  715           key in an index inode, use ntfs_index_lookup().  After modifying an
  716           index entry, call ntfs_index_entry_flush_dcache_page() followed by
  717           ntfs_index_entry_mark_dirty() to ensure the changes are written out
  718           to disk.  For details see fs/ntfs/index.[hc].  Note, at present, if
  719           an index entry is in the index allocation attribute rather than the
  720           index root attribute it will not be written out (you will get a
  721           warning message about discarded changes instead).
  722         - Load the quota file ($Quota) and check if quota tracking is enabled
  723           and if so, mark the quotas out of date.  This causes windows to
  724           rescan the volume on boot and update all quota entries.
  725         - Add a set_page_dirty address space operation for ntfs_m[fs]t_aops.
  726           It is simply set to __set_page_dirty_nobuffers() to make sure that
  727           running set_page_dirty() on a page containing mft/ntfs records will
  728           not affect the dirty state of the page buffers.
  729         - Add fs/ntfs/index.c::__ntfs_index_entry_mark_dirty() which sets all
  730           buffers that are inside the ntfs record in the page dirty after which
  731           it sets the page dirty.  This allows ->writepage to only write the
  732           dirty index records rather than having to write all the records in
  733           the page.  Modify fs/ntfs/index.h::ntfs_index_entry_mark_dirty() to
  734           use this rather than __set_page_dirty_nobuffers().
  735         - Implement fs/ntfs/aops.c::ntfs_write_mst_block() which enables the
  736           writing of page cache pages belonging to mst protected attributes
  737           like the index allocation attribute in directory indices and other
  738           indices like $Quota/$Q, etc.  This means that the quota is now marked
  739           out of date on all volumes rather than only on ones where the quota
  740           defaults entry is in the index root attribute of the $Quota/$Q index.
  741 
  742 2.1.14 - Fix an NFSd caused deadlock reported by several users.
  743 
  744         - Modify fs/ntfs/ntfs_readdir() to copy the index root attribute value
  745           to a buffer so that we can put the search context and unmap the mft
  746           record before calling the filldir() callback.  We need to do this
  747           because of NFSd which calls ->lookup() from its filldir callback()
  748           and this causes NTFS to deadlock as ntfs_lookup() maps the mft record
  749           of the directory and since ntfs_readdir() has got it mapped already
  750           ntfs_lookup() deadlocks.
  751 
  752 2.1.13 - Enable overwriting of resident files and housekeeping of system files.
  753 
  754         - Implement writing of mft records (fs/ntfs/mft.[hc]), which includes
  755           keeping the mft mirror in sync with the mft when mirrored mft records
  756           are written.  The functions are write_mft_record{,_nolock}().  The
  757           implementation is quite rudimentary for now with lots of things not
  758           implemented yet but I am not sure any of them can actually occur so
  759           I will wait for people to hit each one and only then implement it.
  760         - Commit open system inodes at umount time.  This should make it
  761           virtually impossible for sync_mft_mirror_umount() to ever be needed.
  762         - Implement ->write_inode (fs/ntfs/inode.c::ntfs_write_inode()) for the
  763           ntfs super operations.  This gives us inode writing via the VFS inode
  764           dirty code paths.  Note:  Access time updates are not implemented yet.
  765         - Implement fs/ntfs/mft.[hc]::{,__}mark_mft_record_dirty() and make
  766           fs/ntfs/aops.c::ntfs_writepage() and ntfs_commit_write() use it, thus
  767           finally enabling resident file overwrite!  (-8  This also includes a
  768           placeholder for ->writepage (ntfs_mft_writepage()), which for now
  769           just redirties the page and returns.  Also, at umount time, we for
  770           now throw away all mft data page cache pages after the last call to
  771           ntfs_commit_inode() in the hope that all inodes will have been
  772           written out by then and hence no dirty (meta)data will be lost.  We
  773           also check for this case and emit an error message telling the user
  774           to run chkdsk.
  775         - Use set_page_writeback() and end_page_writeback() in the resident
  776           attribute code path of fs/ntfs/aops.c::ntfs_writepage() otherwise
  777           the radix-tree tag PAGECACHE_TAG_DIRTY remains set even though the
  778           page is clean.
  779         - Implement ntfs_mft_writepage() so it now checks if any of the mft
  780           records in the page are dirty and if so redirties the page and
  781           returns.  Otherwise it just returns (after doing set_page_writeback(),
  782           unlock_page(), end_page_writeback() or the radix-tree tag
  783           PAGECACHE_TAG_DIRTY remains set even though the page is clean), thus
  784           alowing the VM to do with the page as it pleases.  Also, at umount
  785           time, now only throw away dirty mft (meta)data pages if dirty inodes
  786           are present and ask the user to email us if they see this happening.
  787         - Add functions ntfs_{clear,set}_volume_flags(), to modify the volume
  788           information flags (fs/ntfs/super.c).
  789         - Mark the volume dirty when (re)mounting read-write and mark it clean
  790           when unmounting or remounting read-only.  If any volume errors are
  791           found, the volume is left marked dirty to force chkdsk to run.
  792         - Add code to set the NT4 compatibility flag when (re)mounting
  793           read-write for newer NTFS versions but leave it commented out for now
  794           since we do not make any modifications that are NTFS 1.2 specific yet
  795           and since setting this flag breaks Captive-NTFS which is not nice.
  796           This code must be enabled once we start writing NTFS 1.2 specific
  797           changes otherwise Windows NTFS driver might crash / cause corruption.
  798 
  799 2.1.12 - Fix the second fix to the decompression engine and some cleanups.
  800 
  801         - Add a new address space operations struct, ntfs_mst_aops, for mst
  802           protected attributes.  This is because the default ntfs_aops do not
  803           make sense with mst protected data and were they to write anything to
  804           such an attribute they would cause data corruption so we provide
  805           ntfs_mst_aops which does not have any write related operations set.
  806         - Cleanup dirty ntfs inode handling (fs/ntfs/inode.[hc]) which also
  807           includes an adapted ntfs_commit_inode() and an implementation of
  808           ntfs_write_inode() which for now just cleans dirty inodes without
  809           writing them (it does emit a warning that this is happening).
  810         - Undo the second decompression engine fix (see 2.1.9 release ChangeLog
  811           entry) as it was only fixing a theoretical bug but at the same time
  812           it badly broke the handling of sparse and uncompressed compression
  813           blocks.
  814 
  815 2.1.11 - Driver internal cleanups.
  816 
  817         - Only build logfile.o if building the driver with read-write support.
  818         - Really final white space cleanups.
  819         - Use generic_ffs() instead of ffs() in logfile.c which allows the
  820           log_page_size variable to be optimized by gcc into a constant.
  821         - Rename uchar_t to ntfschar everywhere as uchar_t is unsigned 1-byte
  822           char as defined by POSIX and as found on some systems.
  823 
  824 2.1.10 - Force read-only (re)mounting of volumes with unsupported volume flags.
  825 
  826         - Finish off the white space cleanups (remove trailing spaces, etc).
  827         - Clean up ntfs_fill_super() and ntfs_read_inode_mount() by removing
  828           the kludges around the first iget().  Instead of (re)setting ->s_op
  829           we have the $MFT inode set up by explicit new_inode() / set ->i_ino /
  830           insert_inode_hash() / call ntfs_read_inode_mount() directly.  This
  831           kills the need for second super_operations and allows to return error
  832           from ntfs_read_inode_mount() without resorting to ugly "poisoning"
  833           tricks.  (Al Viro)
  834         - Force read-only (re)mounting if any of the following bits are set in
  835           the volume information flags:
  836                 VOLUME_IS_DIRTY, VOLUME_RESIZE_LOG_FILE,
  837                 VOLUME_UPGRADE_ON_MOUNT, VOLUME_DELETE_USN_UNDERWAY,
  838                 VOLUME_REPAIR_OBJECT_ID, VOLUME_MODIFIED_BY_CHKDSK
  839           To make this easier we define VOLUME_MUST_MOUNT_RO_MASK with all the
  840           above bits set so the test is made easy.
  841 
  842 2.1.9 - Fix two bugs in decompression engine.
  843 
  844         - Fix a bug where we would not always detect that we have reached the
  845           end of a compression block because we were ending at minus one byte
  846           which is effectively the same as being at the end.  The fix is to
  847           check whether the uncompressed buffer has been fully filled and if so
  848           we assume we have reached the end of the compression block.  A big
  849           thank you to Marcin Gibuła for the bug report, the assistance in
  850           tracking down the bug and testing the fix.
  851         - Fix a possible bug where when a compressed read is truncated to the
  852           end of the file, the offset inside the last page was not truncated.
  853 
  854 2.1.8 - Handle $MFT mirror and $LogFile, improve time handling, and cleanups.
  855 
  856         - Use get_bh() instead of manual atomic_inc() in fs/ntfs/compress.c.
  857         - Modify fs/ntfs/time.c::ntfs2utc(), get_current_ntfs_time(), and
  858           utc2ntfs() to work with struct timespec instead of time_t on the
  859           Linux UTC time side thus preserving the full precision of the NTFS
  860           time and only loosing up to 99 nano-seconds in the Linux UTC time.
  861         - Move fs/ntfs/time.c to fs/ntfs/time.h and make the time functions
  862           static inline.
  863         - Remove unused ntfs_dirty_inode().
  864         - Cleanup super operations declaration in fs/ntfs/super.c.
  865         - Wrap flush_dcache_mft_record_page() in #ifdef NTFS_RW.
  866         - Add NInoTestSetFoo() and NInoTestClearFoo() macro magic to
  867           fs/ntfs/inode.h and use it to declare NInoTest{Set,Clear}Dirty.
  868         - Move typedefs for ntfs_attr and test_t from fs/ntfs/inode.c to
  869           fs/ntfs/inode.h so they can be used elsewhere.
  870         - Determine the mft mirror size as the number of mirrored mft records
  871           and store it in ntfs_volume->mftmirr_size (fs/ntfs/super.c).
  872         - Load the mft mirror at mount time and compare the mft records stored
  873           in it to the ones in the mft.  Force a read-only mount if the two do
  874           not match (fs/ntfs/super.c).
  875         - Fix type casting related warnings on 64-bit architectures.  Thanks
  876           to Meelis Roos for reporting them.
  877         - Move %L to %ll as %L is floating point and %ll is integer which is
  878           what we want.
  879         - Read the journal ($LogFile) and determine if the volume has been
  880           shutdown cleanly and force a read-only mount if not (fs/ntfs/super.c
  881           and fs/ntfs/logfile.c).  This is a little bit of a crude check in
  882           that we only look at the restart areas and not at the actual log
  883           records so that there will be a very small number of cases where we
  884           think that a volume is dirty when in fact it is clean.  This should
  885           only affect volumes that have not been shutdown cleanly and did not
  886           have any pending, non-check-pointed i/o.
  887         - If the $LogFile indicates a clean shutdown and a read-write (re)mount
  888           is requested, empty $LogFile by overwriting it with 0xff bytes to
  889           ensure that Windows cannot cause data corruption by replaying a stale
  890           journal after Linux has written to the volume.
  891 
  892 2.1.7 - Enable NFS exporting of mounted NTFS volumes.
  893 
  894         - Set i_generation in the VFS inode from the seq_no of the NTFS inode.
  895         - Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc.
  896         - Implement ->get_dentry() in fs/ntfs/namei.c::ntfs_get_dentry() as the
  897           default doesn't allow inode number 0 which is a valid inode on NTFS
  898           and even if it did allow that it uses iget() instead of ntfs_iget()
  899           which makes it useless for us.
  900         - Implement ->get_parent() in fs/ntfs/namei.c::ntfs_get_parent() as the
  901           default just returns -EACCES which is not very useful.
  902         - Define export operations (->s_export_op) for NTFS (ntfs_export_ops)
  903           and set them up in the super block at mount time (super.c) this
  904           allows mounted NTFS volumes to be exported via NFS.
  905         - Add missing return -EOPNOTSUPP; in
  906           fs/ntfs/aops.c::ntfs_commit_nonresident_write().
  907         - Enforce no atime and no dir atime updates at mount/remount time as
  908           they are not implemented yet anyway.
  909         - Move a few assignments in fs/ntfs/attrib.c::load_attribute_list() to
  910           after a NULL check.  Thanks to Dave Jones for pointing this out.
  911 
  912 2.1.6 - Fix minor bug in handling of compressed directories.
  913 
  914         - Fix bug in handling of compressed directories.  A compressed
  915           directory is not really compressed so when we set the ->i_blocks
  916           field of a compressed directory inode we were setting it from the
  917           non-existing field ni->itype.compressed.size which gave random
  918           results...  For directories we now always use ni->allocated_size.
  919 
  920 2.1.5 - Fix minor bug in attribute list attribute handling.
  921 
  922         - Fix bug in attribute list handling.  Actually it is not as much a bug
  923           as too much protection in that we were not allowing attribute lists
  924           which waste space on disk while Windows XP clearly allows it and in
  925           fact creates such attribute lists so our driver was failing.
  926         - Update NTFS documentation ready for 2.6 kernel release.
  927 
  928 2.1.4 - Reduce compiler requirements.
  929 
  930         - Remove all uses of unnamed structs and unions in the driver to make
  931           old and newer gcc versions happy. Makes it a bit uglier IMO but at
  932           least people will stop hassling me about it.
  933 
  934 2.1.3 - Important bug fixes in corner cases.
  935 
  936         - super.c::parse_ntfs_boot_sector(): Correct the check for 64-bit
  937           clusters. (Philipp Thomas)
  938         - attrib.c::load_attribute_list(): Fix bug when initialized_size is a
  939           multiple of the block_size but not the cluster size. (Szabolcs
  940           Szakacsits)
  941 
  942 2.1.2 - Important bug fixes aleviating the hangs in statfs.
  943 
  944         - Fix buggy free cluster and free inode determination logic.
  945 
  946 2.1.1 - Minor updates.
  947 
  948         - Add handling for initialized_size != data_size in compressed files.
  949         - Reduce function local stack usage from 0x3d4 bytes to just noise in
  950           fs/ntfs/upcase.c. (Randy Dunlap)
  951         - Remove compiler warnings for newer gcc.
  952         - Pages are no longer kmapped by mm/filemap.c::generic_file_write()
  953           around calls to ->{prepare,commit}_write.  Adapt NTFS appropriately
  954           in fs/ntfs/aops.c::ntfs_prepare_nonresident_write() by using
  955           kmap_atomic(KM_USER0).
  956 
  957 2.1.0 - First steps towards write support: implement file overwrite.
  958 
  959         - Add configuration option for developmental write support with an
  960           appropriately scary configuration help text.
  961         - Initial implementation of fs/ntfs/aops.c::ntfs_writepage() and its
  962           helper fs/ntfs/aops.c::ntfs_write_block(). This enables mmap(2) based
  963           overwriting of existing files on ntfs. Note: Resident files are
  964           only written into memory, and not written out to disk at present, so
  965           avoid writing to files smaller than about 1kiB.
  966         - Initial implementation of fs/ntfs/aops.c::ntfs_prepare_write(), its
  967           helper fs/ntfs/aops.c::ntfs_prepare_nonresident_write() and their
  968           counterparts, fs/ntfs/aops.c::ntfs_commit_write(), and
  969           fs/ntfs/aops.c::ntfs_commit_nonresident_write(), respectively. Also,
  970           add generic_file_write() to the ntfs file operations (fs/ntfs/file.c).
  971           This enables write(2) based overwriting of existing files on ntfs.
  972           Note: As with mmap(2) based overwriting, resident files are only
  973           written into memory, and not written out to disk at present, so avoid
  974           writing to files smaller than about 1kiB.
  975         - Implement ->truncate (fs/ntfs/inode.c::ntfs_truncate()) and
  976           ->setattr() (fs/ntfs/inode.c::ntfs_setattr()) inode operations for
  977           files with the purpose of intercepting and aborting all i_size
  978           changes which we do not support yet. ntfs_truncate() actually only
  979           emits a warning message but AFAICS our interception of i_size changes
  980           elsewhere means ntfs_truncate() never gets called for i_size changes.
  981           It is only called from generic_file_write() when we fail in
  982           ntfs_prepare_{,nonresident_}write() in order to discard any
  983           instantiated buffers beyond i_size. Thus i_size is not actually
  984           changed so our warning message is enough. Unfortunately it is not
  985           possible to easily determine if i_size is being changed or not hence
  986           we just emit an appropriately worded error message.
  987 
  988 2.0.25 - Small bug fixes and cleanups.
  989 
  990         - Unlock the page in an out of memory error code path in
  991           fs/ntfs/aops.c::ntfs_read_block().
  992         - If fs/ntfs/aops.c::ntfs_read_page() is called on an uptodate page,
  993           just unlock the page and return. (This can happen due to ->writepage
  994           clearing PageUptodate() during write out of MstProtected()
  995           attributes.
  996         - Remove leaked write code again.
  997 
  998 2.0.24 - Cleanups.
  999 
 1000         - Treat BUG_ON() as ASSERT() not VERIFY(), i.e. do not use side effects
 1001           inside BUG_ON(). (Adam J. Richter)
 1002         - Split logical OR expressions inside BUG_ON() into individual BUG_ON()
 1003           calls for improved debugging. (Adam J. Richter)
 1004         - Add errors flag to the ntfs volume state, accessed via
 1005           NVol{,Set,Clear}Errors(vol).
 1006         - Do not allow read-write remounts of read-only volumes with errors.
 1007         - Clarify comment for ntfs file operation sendfile which was added by
 1008           Christoph Hellwig a while ago (just using generic_file_sendfile())
 1009           to say that ntfs ->sendfile is only used for the case where the
 1010           source data is on the ntfs partition and the destination is
 1011           somewhere else, i.e. nothing we need to concern ourselves with.
 1012         - Add generic_file_write() as our ntfs file write operation.
 1013 
 1014 2.0.23 - Major bug fixes (races, deadlocks, non-i386 architectures).
 1015 
 1016         - Massive internal locking changes to mft record locking. Fixes lock
 1017           recursion and replaces the mrec_lock read/write semaphore with a
 1018           mutex. Also removes the now superfluous mft_count. This fixes several
 1019           race conditions and deadlocks, especially in the future write code.
 1020         - Fix ntfs over loopback for compressed files by adding an
 1021           optimization barrier. (gcc was screwing up otherwise ?)
 1022         - Miscellaneous cleanups all over the code and a fix or two in error
 1023           handling code paths.
 1024         Thanks go to Christoph Hellwig for pointing out the following two:
 1025         - Remove now unused function fs/ntfs/malloc.h::vmalloc_nofs().
 1026         - Fix ntfs_free() for ia64 and parisc by checking for VMALLOC_END, too.
 1027 
 1028 2.0.22 - Cleanups, mainly to ntfs_readdir(), and use C99 initializers.
 1029 
 1030         - Change fs/ntfs/dir.c::ntfs_reddir() to only read/write ->f_pos once
 1031           at entry/exit respectively.
 1032         - Use C99 initializers for structures.
 1033         - Remove unused variable blocks from fs/ntfs/aops.c::ntfs_read_block().
 1034 
 1035 2.0.21 - Check for, and refuse to work with too large files/directories/volumes.
 1036 
 1037         - Limit volume size at mount time to 2TiB on architectures where
 1038           unsigned long is 32-bits (fs/ntfs/super.c::parse_ntfs_boot_sector()).
 1039           This is the most we can do without overflowing the 32-bit limit of
 1040           the block device size imposed on us by sb_bread() and sb_getblk()
 1041           for the time being.
 1042         - Limit file/directory size at open() time to 16TiB on architectures
 1043           where unsigned long is 32-bits (fs/ntfs/file.c::ntfs_file_open() and
 1044           fs/ntfs/dir.c::ntfs_dir_open()). This is the most we can do without
 1045           overflowing the page cache page index.
 1046 
 1047 2.0.20 - Support non-resident directory index bitmaps, fix page leak in readdir.
 1048 
 1049         - Move the directory index bitmap to use an attribute inode instead of
 1050           having special fields for it inside the ntfs inode structure. This
 1051           means that the index bitmaps now use the page cache for i/o, too,
 1052           and also as a side effect we get support for non-resident index
 1053           bitmaps for free.
 1054         - Simplify/cleanup error handling in fs/ntfs/dir.c::ntfs_readdir() and
 1055           fix a page leak that manifested itself in some cases.
 1056         - Add fs/ntfs/inode.c::ntfs_put_inode(), which we need to release the
 1057           index bitmap inode on the final iput().
 1058 
 1059 2.0.19 - Fix race condition, improvements, and optimizations in i/o interface.
 1060 
 1061         - Apply block optimization added to fs/ntfs/aops.c::ntfs_read_block()
 1062           to fs/ntfs/compress.c::ntfs_file_read_compressed_block() as well.
 1063         - Drop the "file" from ntfs_file_read_compressed_block().
 1064         - Rename fs/ntfs/aops.c::ntfs_enb_buffer_read_async() to
 1065           ntfs_end_buffer_async_read() (more like the fs/buffer.c counterpart).
 1066         - Update ntfs_end_buffer_async_read() with the improved logic from
 1067           its updated counterpart fs/buffer.c::end_buffer_async_read(). Apply
 1068           further logic improvements to better determine when we set PageError.
 1069         - Update submission of buffers in fs/ntfs/aops.c::ntfs_read_block() to
 1070           check for the buffers being uptodate first in line with the updated
 1071           fs/buffer.c::block_read_full_page(). This plugs a small race
 1072           condition.
 1073 
 1074 2.0.18 - Fix race condition in reading of compressed files.
 1075 
 1076         - There was a narrow window between checking a buffer head for being
 1077           uptodate and locking it in ntfs_file_read_compressed_block(). We now
 1078           lock the buffer and then check whether it is uptodate or not.
 1079 
 1080 2.0.17 - Cleanups and optimizations - shrinking the ToDo list.
 1081 
 1082         - Modify fs/ntfs/inode.c::ntfs_read_locked_inode() to return an error
 1083           code and update callers, i.e. ntfs_iget(), to pass that error code
 1084           up instead of just using -EIO.
 1085         - Modifications to super.c to ensure that both mount and remount
 1086           cannot set any write related options when the driver is compiled
 1087           read-only.
 1088         - Optimize block resolution in fs/ntfs/aops.c::ntfs_read_block() to
 1089           cache the current runlist element. This should improve performance
 1090           when reading very large and/or very fragmented data.
 1091 
 1092 2.0.16 - Convert access to $MFT/$BITMAP to attribute inode API.
 1093 
 1094         - Fix a stupid bug introduced in 2.0.15 where we were unmapping the
 1095           wrong inode in fs/ntfs/inode.c::ntfs_attr_iget().
 1096         - Fix debugging check in fs/ntfs/aops.c::ntfs_read_block().
 1097         - Convert $MFT/$BITMAP access to attribute inode API and remove all
 1098           remnants of the ugly mftbmp address space and operations hack. This
 1099           means we finally have only one readpage function as well as only one
 1100           async io completion handler. Yey! The mft bitmap is now just an
 1101           attribute inode and is accessed from vol->mftbmp_ino just as if it
 1102           were a normal file. Fake inodes rule. (-:
 1103 
 1104 2.0.15 - Fake inodes based attribute i/o via the pagecache, fixes and cleanups.
 1105 
 1106         - Fix silly bug in fs/ntfs/super.c::parse_options() which was causing
 1107           remounts to fail when the partition had an entry in /etc/fstab and
 1108           the entry specified the nls= option.
 1109         - Apply same macro magic used in fs/ntfs/inode.h to fs/ntfs/volume.h to
 1110           expand all the helper functions NVolFoo(), NVolSetFoo(), and
 1111           NVolClearFoo().
 1112         - Move copyright statement from driver initialisation message to
 1113           module description (fs/super.c). This makes the initialisation
 1114           message fit on one line and fits in better with rest of kernel.
 1115         - Update fs/ntfs/attrib.c::map_run_list() to work on both real and
 1116           attribute inodes, and both for files and directories.
 1117         - Implement fake attribute inodes allowing all attribute i/o to go via
 1118           the page cache and to use all the normal vfs/mm functionality:
 1119           - Add ntfs_attr_iget() and its helper ntfs_read_locked_attr_inode()
 1120             to fs/ntfs/inode.c.
 1121           - Add needed cleanup code to ntfs_clear_big_inode().
 1122         - Merge address space operations for files and directories (aops.c),
 1123           now just have ntfs_aops:
 1124           - Rename:
 1125                 end_buffer_read_attr_async() -> ntfs_end_buffer_read_async(),
 1126                 ntfs_attr_read_block()       -> ntfs_read_block(),
 1127                 ntfs_file_read_page()        -> ntfs_readpage().
 1128           - Rewrite fs/ntfs/aops.c::ntfs_readpage() to work on both real and
 1129             attribute inodes, and both for files and directories.
 1130           - Remove obsolete fs/ntfs/aops.c::ntfs_mst_readpage().
 1131 
 1132 2.0.14 - Run list merging code cleanup, minor locking changes, typo fixes.
 1133 
 1134         - Change fs/ntfs/super.c::ntfs_statfs() to not rely on BKL by moving
 1135           the locking out of super.c::get_nr_free_mft_records() and taking and
 1136           dropping the mftbmp_lock rw_semaphore in ntfs_statfs() itself.
 1137         - Bring attribute runlist merging code (fs/ntfs/attrib.c) in sync with
 1138           current userspace ntfs library code. This means that if a merge
 1139           fails the original runlists are always left unmodified instead of
 1140           being silently corrupted.
 1141         - Misc typo fixes.
 1142 
 1143 2.0.13 - Use iget5_locked() in preparation for fake inodes and small cleanups.
 1144 
 1145         - Remove nr_mft_bits and the now superfluous union with nr_mft_records
 1146           from ntfs_volume structure.
 1147         - Remove nr_lcn_bits and the now superfluous union with nr_clusters
 1148           from ntfs_volume structure.
 1149         - Use iget5_locked() and friends instead of conventional iget(). Wrap
 1150           the call in fs/ntfs/inode.c::ntfs_iget() and update callers of iget()
 1151           to use ntfs_iget(). Leave only one iget() call at mount time so we
 1152           don't need an ntfs_iget_mount().
 1153         - Change fs/ntfs/inode.c::ntfs_new_extent_inode() to take mft_no as an
 1154           additional argument.
 1155 
 1156 2.0.12 - Initial cleanup of address space operations following 2.0.11 changes.
 1157 
 1158         - Merge fs/ntfs/aops.c::end_buffer_read_mst_async() and
 1159           fs/ntfs/aops.c::end_buffer_read_file_async() into one function
 1160           fs/ntfs/aops.c::end_buffer_read_attr_async() using NInoMstProtected()
 1161           to determine whether to apply mst fixups or not.
 1162         - Above change allows merging fs/ntfs/aops.c::ntfs_file_read_block()
 1163           and fs/ntfs/aops.c::ntfs_mst_readpage() into one function
 1164           fs/ntfs/aops.c::ntfs_attr_read_block(). Also, create a tiny wrapper
 1165           fs/ntfs/aops.c::ntfs_mst_readpage() to transform the parameters from
 1166           the VFS readpage function prototype to the ntfs_attr_read_block()
 1167           function prototype.
 1168 
 1169 2.0.11 - Initial preparations for fake inode based attribute i/o.
 1170 
 1171         - Move definition of ntfs_inode_state_bits to fs/ntfs/inode.h and
 1172           do some macro magic (adapted from include/linux/buffer_head.h) to
 1173           expand all the helper functions NInoFoo(), NInoSetFoo(), and
 1174           NInoClearFoo().
 1175         - Add new flag to ntfs_inode_state_bits: NI_Sparse.
 1176         - Add new fields to ntfs_inode structure to allow use of fake inodes
 1177           for attribute i/o: type, name, name_len. Also add new state bits:
 1178           NI_Attr, which, if set, indicates the inode is a fake inode, and
 1179           NI_MstProtected, which, if set, indicates the attribute uses multi
 1180           sector transfer protection, i.e. fixups need to be applied after
 1181           reads and before/after writes.
 1182         - Rename fs/ntfs/inode.c::ntfs_{new,clear,destroy}_inode() to
 1183           ntfs_{new,clear,destroy}_extent_inode() and update callers.
 1184         - Use ntfs_clear_extent_inode() in fs/ntfs/inode.c::__ntfs_clear_inode()
 1185           instead of ntfs_destroy_extent_inode().
 1186         - Cleanup memory deallocations in {__,}ntfs_clear_{,big_}inode().
 1187         - Make all operations on ntfs inode state bits use the NIno* functions.
 1188         - Set up the new ntfs inode fields and state bits in
 1189           fs/ntfs/inode.c::ntfs_read_inode() and add appropriate cleanup of
 1190           allocated memory to __ntfs_clear_inode().
 1191         - Cleanup ntfs_inode structure a bit for better ordering of elements
 1192           w.r.t. their size to allow better packing of the structure in memory.
 1193 
 1194 2.0.10 - There can only be 2^32 - 1 inodes on an NTFS volume.
 1195 
 1196         - Add check at mount time to verify that the number of inodes on the
 1197           volume does not exceed 2^32 - 1, which is the maximum allowed for
 1198           NTFS according to Microsoft.
 1199         - Change mft_no member of ntfs_inode structure to be unsigned long.
 1200           Update all users. This makes ntfs_inode->mft_no just a copy of struct
 1201           inode->i_ino. But we can't just always use struct inode->i_ino and
 1202           remove mft_no because extent inodes do not have an attached struct
 1203           inode.
 1204 
 1205 2.0.9 - Decompression engine now uses a single buffer and other cleanups.
 1206 
 1207         - Change decompression engine to use a single buffer protected by a
 1208           spin lock instead of per-CPU buffers. (Rusty Russell)
 1209         - Do not update cb_pos when handling a partial final page during
 1210           decompression of a sparse compression block, as the value is later
 1211           reset without being read/used. (Rusty Russell)
 1212         - Switch to using the new KM_BIO_SRC_IRQ for atomic kmap()s. (Andrew
 1213           Morton)
 1214         - Change buffer size in ntfs_readdir()/ntfs_filldir() to use
 1215           NLS_MAX_CHARSET_SIZE which makes the buffers almost 1kiB each but
 1216           it also makes everything safer so it is a good thing.
 1217         - Miscellaneous minor cleanups to comments.
 1218 
 1219 2.0.8 - Major updates for handling of case sensitivity and dcache aliasing.
 1220 
 1221         Big thanks go to Al Viro and other inhabitants of #kernel for investing
 1222         their time to discuss the case sensitivity and dcache aliasing issues.
 1223 
 1224         - Remove unused source file fs/ntfs/attraops.c.
 1225         - Remove show_inodes mount option(s), thus dropping support for
 1226           displaying of short file names.
 1227         - Remove deprecated mount option posix.
 1228         - Restore show_sys_files mount option.
 1229         - Add new mount option case_sensitive, to determine if the driver
 1230           treats file names as case sensitive or not. If case sensitive, create
 1231           file names in the POSIX namespace. Otherwise create file names in the
 1232           LONG/WIN32 namespace. Note, files remain accessible via their short
 1233           file name, if it exists.
 1234         - Remove really dumb logic bug in boot sector recovery code.
 1235         - Fix dcache aliasing issues wrt short/long file names via changes
 1236           to fs/ntfs/dir.c::ntfs_lookup_inode_by_name() and
 1237           fs/ntfs/namei.c::ntfs_lookup():
 1238           - Add additional argument to ntfs_lookup_inode_by_name() in which we
 1239             return information about the matching file name if the case is not
 1240             matching or the match is a short file name. See comments above the
 1241             function definition for details.
 1242           - Change ntfs_lookup() to only create dcache entries for the correctly
 1243             cased file name and only for the WIN32 namespace counterpart of DOS
 1244             namespace file names. This ensures we have only one dentry per
 1245             directory and also removes all dcache aliasing issues between short
 1246             and long file names once we add write support. See comments above
 1247             function for details.
 1248         - Fix potential 1 byte overflow in fs/ntfs/unistr.c::ntfs_ucstonls().
 1249 
 1250 2.0.7 - Minor cleanups and updates for changes in core kernel code.
 1251 
 1252         - Remove much of the NULL struct element initializers.
 1253         - Various updates to make compatible with recent kernels.
 1254         - Remove defines of MAX_BUF_PER_PAGE and include linux/buffer_head.h
 1255           in fs/ntfs/ntfs.h instead.
 1256         - Remove no longer needed KERNEL_VERSION checks. We are now in the
 1257           kernel proper so they are no longer needed.
 1258 
 1259 2.0.6 - Major bugfix to make compatible with other kernel changes.
 1260 
 1261         - Initialize the mftbmp address space properly now that there are more
 1262           fields in the struct address_space. This was leading to hangs and
 1263           oopses on umount since 2.5.12 because of changes to other parts of
 1264           the kernel. We probably want a kernel generic init_address_space()
 1265           function...
 1266         - Drop BKL from ntfs_readdir() after consultation with Al Viro. The
 1267           only caller of ->readdir() is vfs_readdir() which holds i_mutex
 1268           during the call, and i_mutex is sufficient protection against changes
 1269           in the directory inode (including ->i_size).
 1270         - Use generic_file_llseek() for directories (as opposed to
 1271           default_llseek()) as this downs i_mutex instead of the BKL which is
 1272           what we now need for exclusion against ->f_pos changes considering we
 1273           no longer take the BKL in ntfs_readdir().
 1274 
 1275 2.0.5 - Major bugfix. Buffer overflow in extent inode handling.
 1276 
 1277         - No need to set old blocksize in super.c::ntfs_fill_super() as the
 1278           VFS does so via invocation of deactivate_super() calling
 1279           fs->fill_super() calling block_kill_super() which does it.
 1280         - BKL moved from VFS into dir.c::ntfs_readdir(). (Linus Torvalds)
 1281           -> Do we really need it? I don't think so as we have exclusion on
 1282           the directory ntfs_inode rw_semaphore mrec_lock. We mmight have to
 1283           move the ->f_pos accesses under the mrec_lock though. Check this...
 1284         - Fix really, really, really stupid buffer overflow in extent inode
 1285           handling in mft.c::map_extent_mft_record().
 1286 
 1287 2.0.4 - Cleanups and updates for kernel 2.5.11.
 1288 
 1289         - Add documentation on how to use the MD driver to be able to use NTFS
 1290           stripe and volume sets in Linux and generally cleanup documentation
 1291           a bit.
 1292         Remove all uses of kdev_t in favour of struct block_device *:
 1293         - Change compress.c::ntfs_file_read_compressed_block() to use
 1294           sb_getblk() instead of getblk().
 1295         - Change super.c::ntfs_fill_super() to use bdev_hardsect_size() instead
 1296           of get_hardsect_size().
 1297         - No need to get old blocksize in super.c::ntfs_fill_super() as
 1298           fs/super.c::get_sb_bdev() already does this.
 1299         - Set bh->b_bdev instead of bh->b_dev throughout aops.c.
 1300 
 1301 2.0.3 - Small bug fixes, cleanups, and performance improvements.
 1302 
 1303         - Remove some dead code from mft.c.
 1304         - Optimize readpage and read_block functions throughout aops.c so that
 1305           only initialized blocks are read. Non-initialized ones have their
 1306           buffer head mapped, zeroed, and set up to date, without scheduling
 1307           any i/o. Thanks to Al Viro for advice on how to avoid the device i/o.
 1308         Thanks go to Andrew Morton for spotting the below:
 1309         - Fix buglet in allocate_compression_buffers() error code path.
 1310         - Call flush_dcache_page() after modifying page cache page contents in
 1311           ntfs_file_readpage().
 1312         - Check for existence of page buffers throughout aops.c before calling
 1313           create_empty_buffers(). This happens when an I/O error occurs and the
 1314           read is retried. (It also happens once writing is implemented so that
 1315           needed doing anyway but I had left it for later...)
 1316         - Don't BUG_ON() uptodate and/or mapped buffers throughout aops.c in
 1317           readpage and read_block functions. Reasoning same as above (i.e. I/O
 1318           error retries and future write code paths.)
 1319 
 1320 2.0.2 - Minor updates and cleanups.
 1321 
 1322         - Cleanup: rename mst.c::__post_read_mst_fixup to post_write_mst_fixup
 1323           and cleanup the code a bit, removing the unused size parameter.
 1324         - Change default fmask to 0177 and update documentation.
 1325         - Change attrib.c::get_attr_search_ctx() to return the search context
 1326           directly instead of taking the address of a pointer. A return value
 1327           of NULL means the allocation failed. Updated all callers
 1328           appropriately.
 1329         - Update to 2.5.9 kernel (preserving backwards compatibility) by
 1330           replacing all occurences of page->buffers with page_buffers(page).
 1331         - Fix minor bugs in runlist merging, also minor cleanup.
 1332         - Updates to bootsector layout and mft mirror contents descriptions.
 1333         - Small bug fix in error detection in unistr.c and some cleanups.
 1334         - Grow name buffer allocations in unistr.c in aligned mutlipled of 64
 1335           bytes.
 1336 
 1337 2.0.1 - Minor updates.
 1338 
 1339         - Make default umask correspond to documentation.
 1340         - Improve documentation.
 1341         - Set default mode to include execute bit. The {u,f,d}mask can be used
 1342           to take it away if desired. This allows binaries to be executed from
 1343           a mounted ntfs partition.
 1344 
 1345 2.0.0 - New version number. Remove TNG from the name. Now in the kernel.
 1346 
 1347         - Add kill_super, just keeping up with the vfs changes in the kernel.
 1348         - Repeat some changes from tng-0.0.8 that somehow got lost on the way
 1349           from the CVS import into BitKeeper.
 1350         - Begin to implement proper handling of allocated_size vs
 1351           initialized_size vs data_size (i.e. i_size). Done are
 1352           mft.c::ntfs_mft_readpage(), aops.c::end_buffer_read_index_async(),
 1353           and attrib.c::load_attribute_list().
 1354         - Lock the runlist in attrib.c::load_attribute_list() while using it.
 1355         - Fix memory leak in ntfs_file_read_compressed_block() and generally
 1356           clean up compress.c a little, removing some uncommented/unused debug
 1357           code.
 1358         - Tidy up dir.c a little bit.
 1359         - Don't bother getting the runlist in inode.c::ntfs_read_inode().
 1360         - Merge mft.c::ntfs_mft_readpage() and aops.c::ntfs_index_readpage()
 1361           creating aops.c::ntfs_mst_readpage(), improving the handling of
 1362           holes and overflow in the process and implementing the correct
 1363           equivalent of ntfs_file_get_block() in ntfs_mst_readpage() itself.
 1364           I am aiming for correctness at the moment. Modularisation can come
 1365           later.
 1366         - Rename aops.c::end_buffer_read_index_async() to
 1367           end_buffer_read_mst_async() and optimize the overflow checking and
 1368           handling.
 1369         - Use the host of the mftbmp address space mapping to hold the ntfs
 1370           volume. This is needed so the async i/o completion handler can
 1371           retrieve a pointer to the volume. Hopefully this will not cause
 1372           problems elsewhere in the kernel... Otherwise will need to use a
 1373           fake inode.
 1374         - Complete implementation of proper handling of allocated_size vs
 1375           initialized_size vs data_size (i.e. i_size) in whole driver.
 1376           Basically aops.c is now completely rewritten.
 1377         - Change NTFS driver name to just NTFS and set version number to 2.0.0
 1378           to make a clear distinction from the old driver which is still on
 1379           version 1.1.22.
 1380 
 1381 tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
 1382 
 1383         - Replace bdevname(sb->s_dev) with sb->s_id.
 1384         - Remove now superfluous new-line characters in all callers of
 1385           ntfs_debug().
 1386         - Apply kludge in ntfs_read_inode(), setting i_nlink to 1 for
 1387           directories. Without this the "find" utility gets very upset which is
 1388           fair enough as Linux/Unix do not support directory hard links.
 1389         - Further runlist merging work. (Richard Russon)
 1390         - Backwards compatibility for gcc-2.95. (Richard Russon)
 1391         - Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
 1392         - Convert to new filesystem declaration using ->ntfs_get_sb() and
 1393           replacing ntfs_read_super() with ntfs_fill_super().
 1394         - Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
 1395           overflow on 32-bit architectures.
 1396         - Cleanup upcase loading code to use ntfs_(un)map_page().
 1397         - Disable/reenable preemtion in critical sections of compession engine.
 1398         - Replace device size determination in ntfs_fill_super() with
 1399           sb->s_bdev->bd_inode->i_size (in bytes) and remove now superfluous
 1400           function super.c::get_nr_blocks().
 1401         - Implement a mount time option (show_inodes) allowing choice of which
 1402           types of inode names readdir() returns and modify ntfs_filldir()
 1403           accordingly. There are several parameters to show_inodes:
 1404                 system: system files
 1405                 win32:  long file names (including POSIX file names) [DEFAULT]
 1406                 long:   same as win32
 1407                 dos:    short file names only (excluding POSIX file names)
 1408                 short:  same as dos
 1409                 posix:  same as both win32 and dos
 1410                 all:    all file names
 1411           Note that the options are additive, i.e. specifying:
 1412                 -o show_inodes=system,show_inodes=win32,show_inodes=dos
 1413           is the same as specifying:
 1414                 -o show_inodes=all
 1415           Note that the "posix" and "all" options will show all directory
 1416           names, BUT the link count on each directory inode entry is set to 1,
 1417           due to Linux not supporting directory hard links. This may well
 1418           confuse some userspace applications, since the directory names will
 1419           have the same inode numbers. Thus it is NOT advisable to use the
 1420           "posix" or "all" options. We provide them only for completeness sake.
 1421         - Add copies of allocated_size, initialized_size, and compressed_size to
 1422           the ntfs inode structure and set them up in
 1423           inode.c::ntfs_read_inode(). These reflect the unnamed data attribute
 1424           for files and the index allocation attribute for directories.
 1425         - Add copies of allocated_size and initialized_size to ntfs inode for
 1426           $BITMAP attribute of large directories and set them up in
 1427           inode.c::ntfs_read_inode().
 1428         - Add copies of allocated_size and initialized_size to ntfs volume for
 1429           $BITMAP attribute of $MFT and set them up in
 1430           super.c::load_system_files().
 1431         - Parse deprecated ntfs driver options (iocharset, show_sys_files,
 1432           posix, and utf8) and tell user what the new options to use are. Note
 1433           we still do support them but they will be removed with kernel 2.7.x.
 1434         - Change all occurences of integer long long printf formatting to hex
 1435           as printk() will not support long long integer format if/when the
 1436           div64 patch goes into the kernel.
 1437         - Make slab caches have stable names and change the names to what they
 1438           were intended to be. These changes are required/made possible by the
 1439           new slab cache name handling which removes the length limitation by
 1440           requiring the caller of kmem_cache_create() to supply a stable name
 1441           which is then referenced but not copied.
 1442         - Rename run_list structure to run_list_element and create a new
 1443           run_list structure containing a pointer to a run_list_element
 1444           structure and a read/write semaphore. Adapt all users of runlists
 1445           to new scheme and take and release the lock as needed. This fixes a
 1446           nasty race as the run_list changes even when inodes are locked for
 1447           reading and even when the inode isn't locked at all, so we really
 1448           needed the serialization. We use a semaphore rather than a spinlock
 1449           as memory allocations can sleep and doing everything GFP_ATOMIC
 1450           would be silly.
 1451         - Cleanup read_inode() removing all code checking for lowest_vcn != 0.
 1452           This can never happen due to the nature of lookup_attr() and how we
 1453           support attribute lists. If it did happen it would imply the inode
 1454           being corrupt.
 1455         - Check for lowest_vcn != 0 in ntfs_read_inode() and mark the inode as
 1456           bad if found.
 1457         - Update to 2.5.6-pre2 changes in struct address_space.
 1458         - Use parent_ino() when accessing d_parent inode number in dir.c.
 1459         - Import Sourceforge CVS repository into BitKeeper repository:
 1460                 http://linux-ntfs.bkbits.net/ntfs-tng-2.5
 1461         - Update fs/Makefile, fs/Config.help, fs/Config.in, and
 1462           Documentation/filesystems/ntfs.txt for NTFS TNG.
 1463         - Create kernel configuration option controlling whether debugging
 1464           is enabled or not.
 1465         - Add the required export of end_buffer_io_sync() from the patches
 1466           directory to the kernel code.
 1467         - Update inode.c::ntfs_show_options() with show_inodes mount option.
 1468         - Update errors mount option.
 1469 
 1470 tng-0.0.7 - 13/02/2002 - The driver is now feature complete for read-only!
 1471 
 1472         - Cleanup mft.c and it's debug/error output in particular. Fix a minor
 1473           bug in mapping of extent inodes. Update all the comments to fit all
 1474           the recent code changes.
 1475         - Modify vcn_to_lcn() to cope with entirely unmapped runlists.
 1476         - Cleanups in compress.c, mostly comments and folding help.
 1477         - Implement attrib.c::map_run_list() as a generic helper.
 1478         - Make compress.c::ntfs_file_read_compressed_block() use map_run_list()
 1479           thus making code shorter and enabling attribute list support.
 1480         - Cleanup incorrect use of [su]64 with %L printf format specifier in
 1481           all source files. Type casts to [unsigned] long long added to correct
 1482           the mismatches (important for architectures which have long long not
 1483           being 64 bits).
 1484         - Merge async io completion handlers for directory indexes and $MFT
 1485           data into one by setting the index_block_size{_bits} of the ntfs
 1486           inode for $MFT to the mft_record_size{_bits} of the ntfs_volume.
 1487         - Cleanup aops.c, update comments.
 1488         - Make ntfs_file_get_block() use map_run_list() so all files now
 1489           support attribute lists.
 1490         - Make ntfs_dir_readpage() almost verbatim copy of
 1491           block_read_full_page() by using ntfs_file_get_block() with only real
 1492           difference being the use of our own async io completion handler
 1493           rather than the default one, thus reducing the amount of code and
 1494           automatically enabling attribute list support for directory indices.
 1495         - Fix bug in load_attribute_list() - forgot to call brelse in error
 1496           code path.
 1497         - Change parameters to find_attr() and lookup_attr(). We no longer
 1498           pass in the upcase table and its length. These can be gotten from
 1499           ctx->ntfs_ino->vol->upcase{_len}. Update all callers.
 1500         - Cleanups in attrib.c.
 1501         - Implement merging of runlists, attrib.c::merge_run_lists() and its
 1502           helpers. (Richard Russon)
 1503         - Attribute lists part 2, attribute extents and multi part runlists:
 1504           enable proper support for LCN_RL_NOT_MAPPED and automatic mapping of
 1505           further runlist parts via attrib.c::map_run_list().
 1506         - Tiny endianness bug fix in decompress_mapping_pairs().
 1507 
 1508 tng-0.0.6 - Encrypted directories, bug fixes, cleanups, debugging enhancements.
 1509 
 1510         - Enable encrypted directories. (Their index root is marked encrypted
 1511           to indicate that new files in that directory should be created
 1512           encrypted.)
 1513         - Fix bug in NInoBmpNonResident() macro. (Cut and paste error.)
 1514         - Enable $Extend system directory. Most (if not all) extended system
 1515           files do not have unnamed data attributes so ntfs_read_inode() had to
 1516           special case them but that is ok, as the special casing recovery
 1517           happens inside an error code path so there is zero slow down in the
 1518           normal fast path. The special casing is done by introducing a new
 1519           function inode.c::ntfs_is_extended_system_file() which checks if any
 1520           of the hard links in the inode point to $Extend as being their parent
 1521           directory and if they do we assume this is an extended system file.
 1522         - Create a sysctl/proc interface to allow {dis,en}abling of debug output
 1523           when compiled with -DDEBUG. Default is debug messages to be disabled.
 1524           To enable them, one writes a non-zero value to /proc/sys/fs/ntfs-debug
 1525           (if /proc is enabled) or uses sysctl(2) to effect the same (if sysctl
 1526           interface is enabled). Inspired by old ntfs driver.
 1527         - Add debug_msgs insmod/kernel boot parameter to set whether debug
 1528           messages are {dis,en}abled. This is useful to enable debug messages
 1529           during ntfs initialization and is the only way to activate debugging
 1530           when the sysctl interface is not enabled.
 1531         - Cleanup debug output in various places.
 1532         - Remove all dollar signs ($) from the source (except comments) to
 1533           enable compilation on architectures whose gcc compiler does not
 1534           support dollar signs in the names of variables/constants. Attribute
 1535           types now start with AT_ instead of $ and $I30 is now just I30.
 1536         - Cleanup ntfs_lookup() and add consistency check of sequence numbers.
 1537         - Load complete runlist for $MFT/$BITMAP during mount and cleanup
 1538           access functions. This means we now cope with $MFT/$BITMAP being
 1539           spread accross several mft records.
 1540         - Disable modification of mft_zone_multiplier on remount. We can always
 1541           reenable this later on if we really want to, but we will need to make
 1542           sure we readjust the mft_zone size / layout accordingly.
 1543 
 1544 tng-0.0.5 - Modernize for 2.5.x and further in line-ing with Al Viro's comments.
 1545 
 1546         - Use sb_set_blocksize() instead of set_blocksize() and verify the
 1547           return value.
 1548         - Use sb_bread() instead of bread() throughout.
 1549         - Add index_vcn_size{_bits} to ntfs_inode structure to store the size
 1550           of a directory index block vcn. Apply resulting simplifications in
 1551           dir.c everywhere.
 1552         - Fix a small bug somewhere (but forgot what it was).
 1553         - Change ntfs_{debug,error,warning} to enable gcc to do type checking
 1554           on the printf-format parameter list and fix bugs reported by gcc
 1555           as a result. (Richard Russon)
 1556         - Move inode allocation strategy to Al's new stuff but maintain the
 1557           divorce of ntfs_inode from struct inode. To achieve this we have two
 1558           separate slab caches, one for big ntfs inodes containing a struct
 1559           inode and pure ntfs inodes and at the same time fix some faulty
 1560           error code paths in ntfs_read_inode().
 1561         - Show mount options in proc (inode.c::ntfs_show_options()).
 1562 
 1563 tng-0.0.4 - Big changes, getting in line with Al Viro's comments.
 1564 
 1565         - Modified (un)map_mft_record functions to be common for read and write
 1566           case. To specify which is which, added extra parameter at front of
 1567           parameter list. Pass either READ or WRITE to this, each has the
 1568           obvious meaning.
 1569         - General cleanups to allow for easier folding in vi.
 1570         - attrib.c::decompress_mapping_pairs() now accepts the old runlist
 1571           argument, and invokes attrib.c::merge_run_lists() to merge the old
 1572           and the new runlists.
 1573         - Removed attrib.c::find_first_attr().
 1574         - Implemented loading of attribute list and complete runlist for $MFT.
 1575           This means we now cope with $MFT being spread across several mft
 1576           records.
 1577         - Adapt to 2.5.2-pre9 and the changed create_empty_buffers() syntax.
 1578         - Adapt major/minor/kdev_t/[bk]devname stuff to new 2.5.x kernels.
 1579         - Make ntfs_volume be allocated via kmalloc() instead of using a slab
 1580           cache. There are too little ntfs_volume structures at any one time
 1581           to justify a private slab cache.
 1582         - Fix bogus kmap() use in async io completion. Now use kmap_atomic().
 1583           Use KM_BIO_IRQ on advice from IRC/kernel...
 1584         - Use ntfs_map_page() in map_mft_record() and create ->readpage method
 1585           for reading $MFT (ntfs_mft_readpage). In the process create dedicated
 1586           address space operations (ntfs_mft_aops) for $MFT inode mapping. Also
 1587           removed the now superfluous exports from the kernel core patch.
 1588         - Fix a bug where kfree() was used instead of ntfs_free().
 1589         - Change map_mft_record() to take ntfs_inode as argument instead of
 1590           vfs inode. Dito for unmap_mft_record(). Adapt all callers.
 1591         - Add pointer to ntfs_volume to ntfs_inode.
 1592         - Add mft record number and sequence number to ntfs_inode. Stop using
 1593           i_ino and i_generation for in-driver purposes.
 1594         - Implement attrib.c::merge_run_lists(). (Richard Russon)
 1595         - Remove use of proper inodes by extent inodes. Move i_ino and
 1596           i_generation to ntfs_inode to do this. Apply simplifications that
 1597           result and remove iget_no_wait(), etc.
 1598         - Pass ntfs_inode everywhere in the driver (used to be struct inode).
 1599         - Add reference counting in ntfs_inode for the ntfs inode itself and
 1600           for the mapped mft record.
 1601         - Extend mft record mapping so we can (un)map extent mft records (new
 1602           functions (un)map_extent_mft_record), and so mappings are reference
 1603           counted and don't have to happen twice if already mapped - just ref
 1604           count increases.
 1605         - Add -o iocharset as alias to -o nls for backwards compatibility.
 1606         - The latest core patch is now tiny. In fact just a single additional
 1607           export is necessary over the base kernel.
 1608 
 1609 tng-0.0.3 - Cleanups, enhancements, bug fixes.
 1610 
 1611         - Work on attrib.c::decompress_mapping_pairs() to detect base extents
 1612           and setup the runlist appropriately using knowledge provided by the
 1613           sizes in the base attribute record.
 1614         - Balance the get_/put_attr_search_ctx() calls so we don't leak memory
 1615           any more.
 1616         - Introduce ntfs_malloc_nofs() and ntfs_free() to allocate/free a single
 1617           page or use vmalloc depending on the amount of memory requested.
 1618         - Cleanup error output. The __FUNCTION__ "(): " is now added
 1619           automatically. Introduced a new header file debug.h to support this
 1620           and also moved ntfs_debug() function into it.
 1621         - Make reading of compressed files more intelligent and especially get
 1622           rid of the vmalloc_nofs() from readpage(). This now uses per CPU
 1623           buffers (allocated at first mount with cluster size <= 4kiB and
 1624           deallocated on last umount with cluster size <= 4kiB), and
 1625           asynchronous io for the compressed data using a list of buffer heads.
 1626           Er, we use synchronous io as async io only works on whole pages
 1627           covered by buffers and not on individual buffer heads...
 1628         - Bug fix for reading compressed files with sparse compression blocks.
 1629 
 1630 tng-0.0.2 - Now handles larger/fragmented/compressed volumes/files/dirs.
 1631 
 1632         - Fixed handling of directories when cluster size exceeds index block
 1633           size.
 1634         - Hide DOS only name space directory entries from readdir() but allow
 1635           them in lookup(). This should fix the problem that Linux doesn't
 1636           support directory hard links, while still allowing access to entries
 1637           via their short file name. This also has the benefit of mimicking
 1638           what Windows users are used to, so it is the ideal solution.
 1639         - Implemented sync_page everywhere so no more hangs in D state when
 1640           waiting for a page.
 1641         - Stop using bforget() in favour of brelse().
 1642         - Stop locking buffers unnecessarily.
 1643         - Implemented compressed files (inode->mapping contains uncompressed
 1644           data, raw compressed data is currently bread() into a vmalloc()ed
 1645           memory buffer).
 1646         - Enable compressed directories. (Their index root is marked compressed
 1647           to indicate that new files in that directory should be created
 1648           compressed.)
 1649         - Use vsnprintf rather than vsprintf in the ntfs_error and ntfs_warning
 1650           functions. (Thanks to Will Dyson for pointing this out.)
 1651         - Moved the ntfs_inode and ntfs_volume (the former ntfs_inode_info and
 1652           ntfs_sb_info) out of the common inode and super_block structures and
 1653           started using the generic_ip and generic_sbp pointers instead. This
 1654           makes ntfs entirely private with respect to the kernel tree.
 1655         - Detect compiler version and abort with error message if gcc less than
 1656           2.96 is used.
 1657         - Fix bug in name comparison function in unistr.c.
 1658         - Implement attribute lists part 1, the infrastructure: search contexts
 1659           and operations, find_external_attr(), lookup_attr()) and make the
 1660           code use the infrastructure.
 1661         - Fix stupid buffer overflow bug that became apparent on larger run
 1662           list containing attributes.
 1663         - Fix bugs in readdir() that became apparent on larger directories.
 1664 
 1665         The driver is now really useful and survives the test
 1666                 find . -type f -exec md5sum "{}" \;
 1667         without any error messages on a over 1GiB sized partition with >16k
 1668         files on it, including compressed files and directories and many files
 1669         and directories with attribute lists.
 1670 
 1671 tng-0.0.1 - The first useful version.
 1672 
 1673         - Added ntfs_lookup().
 1674         - Added default upcase generation and handling.
 1675         - Added compile options to be shown on module init.
 1676         - Many bug fixes that were "hidden" before.
 1677         - Update to latest kernel.
 1678         - Added ntfs_readdir().
 1679         - Added file operations for mmap(), read(), open() and llseek(). We just
 1680           use the generic ones. The whole point of going through implementing
 1681           readpage() methods and where possible get_block() call backs is that
 1682           this allows us to make use of the generic high level methods provided
 1683           by the kernel.
 1684 
 1685         The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
 1686         though and it doesn't implement accesssing compressed files yet. Also,
 1687         accessing files with attribute list attributes is not implemented yet
 1688         either. But for small or simple filesystems it should work and allow
 1689         you to list directories, use stat on directory entries and the file
 1690         system, open, read, mmap and llseek around in files. A big mile stone
 1691         has been reached!
 1692 
 1693 tng-0.0.0 - Initial version tag.
 1694 
 1695         Initial driver implementation. The driver can mount and umount simple
 1696         NTFS filesystems (i.e. ones without attribute lists in the system
 1697         files). If the mount fails there might be problems in the error handling
 1698         code paths, so be warned. Otherwise it seems to be loading the system
 1699         files nicely and the mft record read mapping/unmapping seems to be
 1700         working nicely, too. Proof of inode metadata in the page cache and non-
 1701         resident file unnamed stream data in the page cache concepts is thus
 1702         complete.

Cache object: efb1e58c1f6c7a09c3907edc7321db97


[ source navigation ] [ diff markup ] [ identifier search ] [ freetext search ] [ file search ] [ list types ] [ track identifier ]


This page is part of the FreeBSD/Linux Linux Kernel Cross-Reference, and was automatically generated using a modified version of the LXR engine.