The special tests for accepting ftruncate(2) and access(W_OK) applied
to a file being created read-only have been implemented in the fuse
kernel module, so there is no need to duplicate them in the file system.
On OpenIndiana the fuse kernel module does not prevent unlink(2)
from deleting directories, nor prevent rmdir(2) from deleting files,
so this has to be done in the lowntfs-3g driver.
Checking supplementary groups permissions to access a file relies on
a supplementary group list available in /proc/PID. This patch adds a
variant to implement the checks based on the specific format used
by OpenIndiana.
Defined new options --new-serial and --new-half-serial to set a new
serial number (either the given one or a random one) to a file system.
Useful for mounting the original and the cloned file system at the
same time.
Defined new options --new-serial and --new-half-serial to set a new
random serial number when cloning or restoring a file system.
Useful for mounting the original and the cloned file system at the
same time.
Clearing of user data was wrong when extracting the metadata into a
special format image. In this situation the clearing has to be done
without mounting the file system.
When a file is created with no write permissions, ftruncate and
access(W_OK) should be allowed, but this check has to be made by the
file system on OpenIndiana.
This patch is not active on Linux (#ifdef'ed)
When a user queries an extended attribute in the trusted namespace,
the traditional error return was EPERM. This has been recently changed
to ENODATA. See https://lkml.org/lkml/2011/5/27/199
On Windows XP and Vista, the system hibernation is identified by the
sequence "hibr" at the beginning of hiberfil.sys. This had been changed
to "HIBR" in Windows 7, so both sequences have to be accepted as
hibernation criteria.
When extracting the metadata, unused data at the end of MFT records is
supposed to be cleared. This was done for the base record of each file,
but not for the extent records.
Clearing the bad cluster list was done by truncating $BadClus:$Bad,
this turned out not to be reliable because chkdsk does not adjust
the size of $BadClus:$Bad when declaring a cluster bad.
This avoids name collisions with Mac OS X system headers (specifically
/usr/include/sys/queue.h). It's quite possible that other operating
systems also have similarly named macros in their system headers since
the function/macro names are very generic.
'wipe_mft' didn't write any $MFTMirr entries for newly manufactured
entries (i.e. entries that were unused and therefore competely wiped and
replaced with a newly initialized copy).
The dereferencing of 'offset' and 'usa' from the MFT record wasn't an
le16* dereference but a u8* dereference, leading to only the least
significant byte (little-endian systems) or the most significant byte
(big-endian systems) being part of the value. (So while this bug could
go unnoticed on little-endian systems for volumes with small values of
'usa', it caused even more significant problems on big-endian systems.)
Fixed by properly casting the address to le16* before dereferencing.
'wipe_mft' was hardcoded to use the USA size of a 1024 byte MFT record
when creating new blank MFT entries. It now calculates the USA size
based on vol->mft_record_size.
This bug led 'utils_mftrec_in_use' to use 'cached' data before it had
been initialized (i.e. its contents was undefined), producing incorrect
results for the first 3584 MFT entries.
When the cluster size is bigger than the index block size, the index
block size unit is 512 (not the sector size) instead of the cluster size.
The partitions formatted by mkntfs and used by ntfs-3g were not
interoperable with Windows when the cluster size is bigger than 4K
and the sector size is not 512.
The booting process has recently been changed so that partitions
which are requested to be checked at mount time with no checker
available, cause a specific user interaction while booting.
The patch warns users against requesting such a check, as there
is no known open-source ntfs file system checker.
When computing the runlist for the first non-resident write to an
attribute, an inconsistency was created between the attribute image
and the ntfs_attr structure, which could cause an MFT record overflow
when the first write is huge and fragmented (reported by Vito Caputo).
With the default mount options, compression of new files are now done
if the parent directory is marked for compression. The mount option
"compression" is not needed any more, but the option "nocompression"
can be used to disable compression of new files.
The default option also applies to applications using libntfs-3g with
no mount command.
When undeleting a file whose name cannot be recovered, this patch
defines a name to which the inode number is appended to facilitate
a bulk recovery (recovering the unwanted deletion of a subtree).
When a file is deleted, ntfs-3g deletes the name from the MFT record,
so the name is not available when the file has to be undeleted.
This patch tries to recover the name when it has not been overwritten.
This is mostly possible when data is non-resident and not fragmented.
