Commit Graph

3146 Commits (48c336fd6aece5d56e246eeb49f734a1b17546cd)
 

Author SHA1 Message Date
Yura Pakhuchiy 48c336fd6a O(1) mapping pairs update for normal files 2007-06-08 13:45:28 +03:00
Yura Pakhuchiy b99d331f78 We do not need to map whole runlist in this case, only until @pos. Introduce ntfs_attr_map_runlist_range to implement this. 2007-06-08 13:41:10 +03:00
Yura Pakhuchiy e404ce8203 Fix 64bit warnings (Yuval) 2007-06-08 13:36:37 +03:00
Yura Pakhuchiy e6620be16d Do not create holes in ntfs_attr_pwrite that will be instantiated right after creation
* Add new API: __ntfs_attr_truncate that allow user to select whether holes should be created or clusters allocated
* Update ntfs_attr_pwrite and ntfscp to use it.
2007-06-08 13:35:47 +03:00
Yura Pakhuchiy 52d034e5f3 ntfsmount: fix errno in truncate() operation 2007-06-08 13:34:16 +03:00
aia21 62d49cb175 Minor updates for OSX compatibility, i.e. glibtoolize not libtoolize, and
automake 1.10 fix, etc.
2007-03-19 14:39:44 +00:00
yura 2c7fd02349 ntfsmount: use utimens() instead of deprecated utime(). 2007-03-13 18:52:02 +00:00
yura 48399b98b3 cleanup 2007-03-13 18:45:51 +00:00
yura 01dad951c2 minor cleanups 2007-03-13 18:39:47 +00:00
aia21 3575e2089e - mkntfs: Use Vista $Boot rather than XP one. (Anton) 2007-03-06 15:44:30 +00:00
yura 857573a2e4 minor style fix and changelog description 2007-01-07 19:19:53 +00:00
uvman dcc13dca8b ni->allocated_size for resident files is rounded up to multiples of 8. A '>> 9' on such value may round up or down. Fix it to always round up. Samba 3 use st_blocks to detect if a file is sparse. ActiveSync 4 fails on sparse files. 2007-01-07 19:08:22 +00:00
yura 197a4cc756 - upgrade to new shiny fuse option parsing API instead of my old hackish one (more readable and -4kb code)
- remove show_sys_files and add case_insensitive options
2007-01-03 18:27:10 +00:00
aia21 f6adc3fc78 - mkntfs: Create more Vista like volumes. We now match the security
descriptor attributes (but not yet the security descriptors stored in
  $Secure).  (Anton)
- libntfs: Rewrite ntfs_upcase_table_build() to generate a Vista
  compatible upcase table ($UpCase).  (Anton)
- mkntfs: Remove own generation of upcase table, i.e. delete
  ntfsprogs/upcase.[ch] and use ntfs_upcase_table_build() supplied by
  libntfs.  (Anton)
2006-12-19 16:31:52 +00:00
yura 452b99f01b oops, committed temporary thing by accident
hope this will work for everyone
2006-12-18 13:39:23 +00:00
yura bedd6f4a13 add some debug output useful on freebsd for aligned access testing 2006-12-18 13:32:58 +00:00
yura 3c668b7d03 - O_EXCL already handled by MNT_NOT_EXCLUSIVE at more early stage
- fix comments
2006-12-15 17:00:02 +00:00
aia21 1fa522a0a1 - libntfs: Add support for FreeBSD 5.0+ sector aligned access
requirements.  (Max Khon)

With some modifications from me...
2006-12-15 15:21:27 +00:00
aia21 0911eb8b0d Make volumes created with mkntfs more compliant with Vista (still not finished). 2006-12-15 14:52:43 +00:00
aia21 5f3079c2bb Fix ntfstruncate build breakage. 2006-12-13 10:51:32 +00:00
aia21 19257df7f0 Update version to 2.0.0, update NEWS, update mkntfs man page and mkntfs itself.
- mkntfs: As announced, remove the deprecated support for creation of
  NTFS 1.2/3.0 volumes.  We now create NTFS 3.1 volumes only.  (Anton)
- mkntfs: Remove lots of unused/unneeded debugging code.  (Anton)
2006-12-13 10:46:49 +00:00
aia21 78c46fd3c3 - Do not create object id attribute on NTFS 1.2 volumes.
- Create object id attribute after volume flags attribute.
2006-12-13 09:17:14 +00:00
aia21 1a4fb3431c - mkntfs: Generate a random DCE compliant UUID for the created volume
and include --with-uuid[=PFX] and --without-uuid options.  (Anton)
- configure.ac: Set language to C.  (Anton)
- mkntfs: Always set default cluster size to 4096 bytes regardless of
  volume size.  This is what Windows Vista does and it makes perfect
  sense from a performance point of view.  (Anton)
2006-12-13 00:11:50 +00:00
yura 914cf8f2be Workaround for this bogus s64 warnings. Thanks, Linus! 2006-12-12 20:27:32 +00:00
yura 10b22550cd more sparse fixes
left only endianness and this stupid problems with s64
2006-12-12 19:00:00 +00:00
yura 2e44b6ca7c fix some sparse warnings
btw, it looks like sparse loses it head when it see something like:

void foo(s64 *bar)
{
	...
}

...

{
	s64 baz;

	foo(&baz);
	...
}
2006-12-12 15:44:52 +00:00
yura 7b60afcc2f fix sparse errors. and first steps towards endianness checking 2006-12-12 15:02:25 +00:00
aia21 a38ead2f20 Szaka prefers to have the lofgile reset and dirty bit setting done even if they
are already done so remove the conditionals I added.
2006-12-11 10:41:34 +00:00
yura 2b611a1461 fix clusters per index block calculation (Szaka) 2006-12-11 03:16:22 +00:00
yura c616e29663 ntfsck by Yuval to extra dist with minor changes from me 2006-12-11 03:06:35 +00:00
yura 19089b08bd one place with mount option description 2006-12-11 03:03:30 +00:00
yura 03fabce6eb o fix magicNTFS defenition
o remove NTFS_SB_MAGIC because it is unusefull
o rename magicNTFS -> NTFS_SB_MAGIC
o make bootsect.c use NTFS_SB_MAGIC instead of hardcoded value
2006-12-11 03:02:23 +00:00
aia21 43b61b82e7 If pread/pwrite are not supported by the OS, set the device pread()/pwrite()
pointers to NULL so we automatically use seek() + read()/write() from now on.
2006-12-10 19:38:02 +00:00
aia21 f73d77d568 - Implement ntfs_pread() and ntfs_pwrite() in terms of device
operations pread() and pwrite() respectively and fall back to using
  seek() + read()/write() if no pread()/pwrite() device operation is
  supplied or the OS does not support the pread()/pwrite() system call.
  Adapt unix_io pread()/pwrite() device operations to use pread()/
  pwrite() system call and adapt win32_io device operations to not
  supply pread()/pwrite().  (Csaba Henk, Anton)
2006-12-09 14:01:12 +00:00
aia21 8640540b53 Warn about VOLUME_MOUNTED_ON_NT4 wrt Vista in layout.h and do not set the flag
in ntfsfix.
2006-12-09 11:27:37 +00:00
yura 82747f3c47 ntfsmount: fix rename if destination already exists. 2006-12-06 19:51:59 +00:00
yura c475778b47 Turn ntfs_pathname_to_inode() into ntfs_pathname_to_inode_num() which
returns ntfs inode number instead of opened inode itself.
Reimplement ntfs_pathname_to_inode() as wrapper to new API.
2006-12-06 18:50:13 +00:00
szaka 3cd5cb9c8a Don't check free space if output file is FIFO (Andree Leidenfrost) 2006-12-03 18:26:58 +00:00
yura 33ffbb2af7 rename --enable-fuse-module to more clear --enable-ntfsmount and cleanup autotools scripts a bit 2006-12-03 14:19:29 +00:00
aia21 0089670e1c Use NTFS_MNT_FORENSIC with ntfsresize to restore old behaviour as it is
documented, i.e. that ntfsresize does not do any modifciations until you say
so.  Also reinstate code to do the journal emptying and setting of dirty flag
on the volume because of it.  However update the code so that those things are
only done if they were not already the case.
2006-12-03 08:59:57 +00:00
yura fb40779709 Force using FUSE 2.6.1 that have proper error message in case of missing fuseblk 2006-12-01 16:04:56 +00:00
aia21 f40bef7246 Disable unmounting at the point of no return when we start doing serious
modifications to the volume.  Hopefully Szaka will be happy now.  (-:
2006-11-28 13:45:42 +00:00
aia21 a32aa26d0b - Fix ntfsresize to unmount the volume when finished/exiting so it does
not leave the volume in an inconsistent state.  Somewhat crude
  solution using atexit() but it works...  (Anton)
2006-11-28 11:00:52 +00:00
aia21 5302d23f7b - Empty the journal at mount time. (Anton)
- Set the volume dirty bit at mount time (if it is not set already and
  clear it again at umount time but only if it was not set to start
  with.  (Anton)
2006-11-28 10:09:56 +00:00
yura fbfbe3b48e force distro developers to fix their broken init scripts 2006-11-25 21:44:35 +00:00
yura f738f39bdf oops, introduced stupiedness during one of recent cleanups 2006-11-25 21:35:39 +00:00
yura f12f4aefb7 ntfs_log_error("Failed.\n") is evil. TODO: Implement ntfs_log_errortrace 2006-11-25 21:10:35 +00:00
yura 3cd59df400 bit more nice error messages 2006-11-25 18:38:47 +00:00
yura 8df298aec2 Introduce MNT_NTFS_NOT_EXCLUSIVE mount option that tells libntfs do
not open volume exclusively. Useful if libntfs user cares about this
himself, eg. FUSE with blkdev option.
2006-11-25 17:37:37 +00:00
yura 4b7868ddc2 ntfsmount: Umount volume in DESTROY. This is guarantees that all data would be s
ynced before umount return for volumes mounted with blkdev option.

Thanks to Miklos for information about blkdev.
2006-11-25 14:57:54 +00:00