xref: /external/e2fsprogs/doc/RelNotes/v1.42.txt

E2fsprogs 1.42.13 (May 17, 2015)
================================

Fixed a potential buffer overflow while closing a file system in
libext2fs.  (CVE-2015-1572, Addresses Debian Bug: #778948)

Fixed a bug which could cause e2fsck to corrupt Hurd file systems.
(Addresses Debian Bug: #760275)

Fixed a deadlock which occurs when using systemd and e2fsck.conf's
logging feature.  (Addresses Debian Bug: #775234)

Fixed a bug which could cause programs using libext2's inode table
scanning functions from crashing on certain corrupted file systems.
(Addresses Debian Bug: #773795)

Fixed dumpe2fs so it won't crash if the user doesn't specify a block
device (file system).  (Addresses Debian Bug: #764293)

Fixed e2fsck so if it notices unexpected HTREE blocks in pass 2, it
will report the correct directory inode containing the inconsistency.

If e2fsck fails to grow the dir_info structure due realloc(3) not
having enough memory, it will now fail with explanatory message
instead of staggering on failing with a confusing internal error
messages.

The tune2fs program will zero out the superblock journal backup
information when removing a journal from a file system.

The mke2fs program now enables the large_file feature by default.

Fixed a bug which could cause badblocks to crash if there are millions
and millions of bad blocks.

Fixed some use-after-free bug in resize2fs and e2fsck.

Fixed a memory leak in tune2fs.

Fixed some bigendian bugs that had crept into both indirect and extent
handling inside libext2fs.

Updated/fixed various man pages.

Update Esperanto, German, and Spanish translations.  Added Danish
translation.

Programmer's Notes
------------------

Fixed coverity, sparse, gcc -Wall, and clang warnings/nits.

Clean up some build system problems (build failures with various
configure options, fix Posix portability issues, etc.)

The functions ext2fs_inode_alloc_stats[2]() and
ext2fs_block_alloc_stats[2]() now check the inode and block numbers
passed to them, to avoid crashes caused by buggy callers.

The libext2fs directory iterator will now check for an invalid
directory entry length instead of possibly walking off the end of the
directory buffer.


E2fsprogs 1.42.12 (August 25, 2014)
===================================

Fix various e2fsck bugs when trying to repair bigalloc file systems.

E2fsck can now repair a file system with an overly large
s_first_meta_bg field, which had previously caused all e2fsprogs
programs to crash when trying to open such a file system.

Fix e2fsck so that it can correctly fix a number of rare file system
corruptions that were discovered when using a file system fuzzer.

Fix e2fsck so it does not try to write back block group descriptors if
they have not been modified.

Mke2fs program will now place metadata blocks in the last flex_bg so
they are contiguous.  This reduces free space fragmentation in a
freshly created file system, as well as allowing mke2fs commands which
request extremely large flex_bg size to succeed.

Mke2fs now creates hugefiles more efficiently (with fewer extent tree
blocks).

Fix a 32/64-bit overflow bug that could cause resize2fs to loop
forever. (Addresses-Launchpad-Bug: #1321958)

The resize2fs program will now use much less memory when resizing very
large file systems.

Fix a bug which could cause resize2fs to get confused and produce a
corrupted file system when shrinking a file system that had been
previously expanded and converted to use the meta_bg file system
format, but which no longer needs to use the meta_bg file system
format any longer.  This typically happened with large (> 16TB) file
systems.  (Addresses Debian Bug: #756922)

The e4defrag program will now defrag backwards-allocated files.

Fix tune2fs updating UUID's when manipulating file systems with
external journals (both the file system and journal UUID).

Fix tune2fs so it can remove an external journal for file systems with
a 1k block size.

Add a new debugfs command, "inode_dump", which prints the inode in hex
and ASCII format.

The debugfs's "set_inode_field" will now automatically allocate blocks
for indirect blocks and extent tree blocks.

Fix debugfs's "set_inode_field" so can properly handle
"block[IND|DIND|TIND]".

The debugfs "rdump" command will now take multiple source arguments.

Fixed a double close(2) bug in "rdump" and "rdump -p".

Fix debugfs's argument parsing for the freefrag command.

Fix filefrag to properly handle using the FIBMAP ioctl (with -B).
(Addresses Launchpad Bug: #1356496)

Clarified messages that were confusing users in debugfs, e2fsck,
mke2fs, and resize2fs (Addresses Debian Bugs: #758029, #757543,
#757544)

Dumpe2fs will now complain if extra arguments are given to it.
(Addresses Debian Bug: #758074)

Updated/fixed various man pages.  (Addresses-Debian-Bug: #726760)

Update Czech, Dutch, French, Polish, Spanish, Swedish, Ukrainian, and
Vietnamese translations.

Programmer's Notes
------------------

The Unix I/O layer will now use pread[64]/pwrite[64] to reduce the
number of system calls issued.

Fixed profiled build.

The libext2fs library will now more efficiently split an extent tree
index block when appending to an end of a file.

Fixed free block accounting for 64-bit file systems.

Add a new function ext2s_inode_size_set() which takes care of all of
the required feature flag modifications.

The regression tests no longer require the presence of the md5sum
program.

Fix build failure on MIPS platforms on Debian (Addresses Debian Bug:
#754605)

Fix various Coverity warnings (#1225003, #1229243, #1252003)


E2fsprogs 1.42.11 (July 9, 2014)
================================

Add support so that mke2fs can create hugefiles so that they are
aligned relative to the beginning of the disk, instead of relative to
the beginning of the partition, using the mke2fs.conf configuration
parameter "hugefiles_align_disk".

Fix a bug which causes e2fsck to abort a journal replay on a file
system with bigalloc enabled.  (Addresses Debian Bug: #744953)

Add sanity checks so that mke2fs will refuse insanely large flex_bg
counts specified by the -G option.  Insanely large flex_bg sizes can
result in a file system which can't be unmounted, or with extremely
pessimal metadata layouts.

The mke2fs program will provide a better metadata layout for
moderately large (but believable) flex_bg counts, such as 133,072.

The mke2fs program will also check the kernel version number to
determine whether the lazy_itable_init option is supported, so that
lazy inode table initialization can be used even if the ext4 is
compiled as a module which isn't loaded at the time that mke2fs is
run.

Add description of ext4's mount options to the ext4 section 5 man
page.

Improve the chattr man page and clean up the chattr usage message.

E2fsck will now automatically fix a last mount time or last write time
which is in the future in preen mode, instead of aborting the fsck
check.

Mke2fs will now check the file system revision number requested by the
command line, and reject it if it is too large.  A file system with an
unsupported revision number will not be accepted by any of the other
file system tools, and this command line option is rarely used, so
let's try to help the user not shoot their own foot off.

The debugfs program can now set the error count and associated
first_error and last_error fields in the superblock using the
set_super_value command.

Resize2fs will not try to calculate the minimum size of a file system
if it contains errors.  Otherwise a very badly corrupted file system
image can cause the minimum size calculation to loop forever.  Since
resize2fs will not allow a file system marked as containing errors to
be resized.  It's pointless to calculate the minimum size, so just
skip it.

Fixed a typo in one of mke2fs's error message.

If filefrag runs into an error, it will continue processing the rest
of the files specified on the command line, but then exit with an
error code.

Filefrag now prints some additional new flags (such as
FIEMAP_EXTENT_ENCODED), and print unknown flags as hex values.

Fixed support in filefrag for files with > 2**32 blocks on 32-bit
platforms.

Fixed a file descriptor leak in debugfs when copying files.

Fixed a regression introduced in 1.42.10 which caused mke2fs to fail
to create file systems larger than 2GB on 32-bit platforms.
(Addresses Red Hat Bugzilla: #1099892, Debian Bug: #752107)

Add Esperanto and Ukrainian translations, and update Czech, Dutch,
French, German, Italian, Polish, Swedish, and Vietnamese
translations.

Fixed various compiler warnings.


Programmer's Notes
------------------

The gcc "-checker" option has been long deprecated, so remove support
from e2fsprogs.  It was causing a few people to be confused when they
tried enabling the configure option.

Update to the latest version of the config.status and config.rpath
files, as well as newer versions of the autoconf files in aclocal.m4.
Also, move the e2fsprogs-specific macros to acinclude.m4, which allows
people who want to use autoreconf to do so.

Make the use of strptime() function in debugfs to be more portable to
fix regression test failures for FreeBSD and MacOS.

A few miscellaneous changes designed to make life easier for
translators.

Fix the f_quota test and some Makefile dependencies if quota support
is not enabled.

Fix cross-compilation support, which got broken in 1.42.10.
(Addresses Debian Bug: #753375)

Delete lib/ext2_types.h on a make clean, and build it as necessary.

Add a debug program used for libext2fs unit tests.

Clean up the tests so that it avoids using GNU-specific behaviour in
mktemp.  Make sure the temporary files are cleaned up if the tests are
interrupted.

Add a new ext2fs_close_free() helper function which automatically
takes care of freeing the ext2_filsys structure even if ext2fs_close()
returns an error.


E2fsprogs 1.42.10 (May 18, 2014)
================================

Mke2fs now creates file systems in regular files (which is very often
used when maintaining virtual machine images) without requiring the
user to use the force option.  In addition, the mke2fs output has been
made much less verbose and only displays information that users will
be more likely to find useful by default.

Mke2fs now will ask the user to confirm that they want to continue
before wiping out a pre-existing file system, partition table, or LVM
physical volume.

Mke2fs now has the ability to create file systems where all the
metadata is located at the very beginning of the device.  This can be
useful for flash devices which have SLC flash at the beginning of the
disk, for FAT compatibility, for example.  As part of this, mke2fs and
tune2fs can control the location of the data blocks used by the
journal inode.  In addition, the new sparse_super2 feature allows for
even fewer (anywhere from zero to two) backup superblocks.

Mke2fs now uses much less CPU when allocating the blocks used for very
large file systems.

Mke2fs can now support creating a file system at an offset.  This can
be useful when creating a disk image for virtual machines (Addresses
Debian Bug: #497984)

Previously, e2fsck had a number of very serious bugs when checking a
file system which used the new the quota file system option (where the
quota inodes were stored in hidden inodes) and the quota inode was
inconsistent with the actual usage data.  This problem was documented
in https://ext4.wiki.kernel.org/index.php/Quota.  These problems have
been fixed in 1.42.10, so the quota should be safe to use with
e2fsprogs 1.42.10.

Fixed in a bug in resize2fs which could cause shrink operation fail in
the unlikely situation when the inode table needs to be moved to a
location before the current location.

Resize2fs now has a much more accurate (and less conservative)
estimation of how far the file system can be shrunk.  This allows a
mostly empty filesystem which is a few terabytes, to be shrunk to a
few megabytes in a single resize2fs -M operation.

E2fsck will now force a full file system check if there are any file
system inconsistencies detected in the super block.

The filefrag program will now display the shared extent flag, which is
used by file systems such as btrfs.

If the number of inodes in the file systems is larger than the time
that file system was created (as measured by the number of seconds
since January 1, 1970) e2fsck would print a scary (but otherwise
harmless) warning of file system corruption for each inode in the
orphan list.  This false positive has been fixed.

The e4defrag program has been fixed so it will no longer refuse with
filesystem with the 64-bit or the bigalloc feature enabled.

The logsave program will print a much less scary message which could
lead users to believe something has gone very wrong with e2fsck exits
with a non-zero exit status (since this is normal after e2fsck has
automatically fixed a file system corruption during a preen
operation).   (Addresses Debian Bug: #468821)

When creating a file system which is larger than 16TB, the
resize_inode option must be disabled -- since the resize_inode simply
doesn't support reserving metadata block numbers which is larger than
32 bits.  The mke2fs program does this automatically if the file
system size is determined automatically.  It will now also do this if
the file size is specified explicitly.

Fixed bugs associated with resize2fs and shrinking bigalloc file
systems.

The e2fsck program will no longer try to add a UUID on a mounted file
system with checksums enabled, since this could leave the file system
checksums broken.

Tune2fs will allow the removal of an external journal from file system
which is marked as needing the journal replayed when the force ("-f")
option is given twice.  (Addresses Debian Bug: #559301)

Tune2fs will no longer support enabling sparse_super if the meta_bg
file system feature is enabled, since it could result in data loss.
In practice, all modern file systems have sparse_super is enabled, so
it's not worth trying to change how tune2fs handles enabling the
sparse_super feature.

Fixed support for 1k block file systems with the meta_bg feature.

When the superblock is corrupt, e2fsck can't figure out the location
of the alternate superblock.  Unfortunately, the routine that
calculates the location of the alternate superblock uses 8193 if it
can't figure this out, so the message printed by e2fsck always
suggests using "e2fsck -b 8193".  This message has been fixed to
suggest both the superblock location of 8193 and and 32768.
(Addresses Debian Bug: #719185)

The lookback mount detection code that was introduced in 1.42.9 wasn't
actually compiled in due to an autoconf oops, so it's fixed now in
1.42.10.  (Addresses Debian Bug: #497984)

A bug introduced in 1.42.9 would cause debugfs to print two error
messages if it found an error while parsing a user-supplied block
number.  This has been fixed in 1.42.10.

Update Czech, Dutch, French, German, Polish, Spanish, Swedish, and
Vietnamese translations.  (Addresses Debian Bug: #703048)

Updated/fixed various man pages.  (Addresses Debian Bugs: #719189,
#719184)

Fixed various Debian Packaging Issues.  (Addresses Debian Bug: #718725)


Programmer's Notes
------------------

Fixed a lot of coverity, sparse, gcc -Wall, and clang warnings/nits.

Allow the location of pkg-config files to be specified independent of
the libdir location via a makefile variable.

Fixed parse-types.sh not to complain when cross-compiling and the
sizes of types are different between the target architecture and the
architecture of the build system

Allow the regression test suite to work correctly on systems which do
not have the "truncate" or "mksawp" programs.

Allow e2fsck to build correctly on systems (such as Android) that do
not have the signal.h file.

E2fsprogs now has code coverage testing which can be enabled using
"configure --enable-gcov".

The libe2p.h header file can now be used included by C++ programs.

The profile/config file used by e2fsck and mke2fs will interpret
numbers with a leading 0 character to mean that they should be
interpreted as an octal integer.

The extent handling functions in libext2fs have been improved so they
have proper rollback when there is an error splitting an extent.  Also
fix a number of bugs when punching holes in files, and fix an
off-by-one bug when inserting an extent into an empty inode.

The libext2fs library now handles the support of BLOCK_UNINIT by
clearing the portion of block bitmap when it is loaded, instead of
when it is used.  This reduces the chances of bugs, and and simplifies
the code.  It also means that debugfs will properly show that blocks
in uninitialized block groups as being unused when using the testb
command.

The e4defrag program will try to use fadvise64 or posix_fadvise64() if
it is present, which allows 64-bit offsets on 32-bit systems.


E2fsprogs 1.42.9 (December 28, 2013)
====================================

Mke2fs will detect an attempt to create a file system on a loop
mounted file and complain without the -FF option.  Also fixed mke2fs
so it allows the creation of a file system on a mounted device with
two -F options, as documented in the man page, instead of three -F
options.

Fixed a large number of bugs in resize2fs, e2fsck, debugfs, and
libext2fs to correctly handle bigalloc and 64-bit file systems.  There
were many corner cases that had not been noticed in previous uses of
these file systems, since they are not as common.  Some of the bugs
could cause file system corruption or data loss, so users of 64-bit or
bigalloc file systems are strongly urged to upgrade to e2fsprogs
1.42.9.

The tune2fs program will now not allow changing the uuid on a mounted
file system which has the uninit_bg feature enabled.  This avoids a
nasty race where the kernel and tune2fs are both retrying rewrite the
group descriptors at the same time, with different ideas about what
the UUID should be.

When e2fsck is rehashing an extent-mapped directory, it's possible
(although very rare) that an extent block will need to be allocated;
fix e2fsck to make sure that the block gets marked as used.

Mke2fs will now properly set the LARGE_FILE feature when creating a
journal >= 2GB --- which can happen when using 64k block size.

Fixed debugfs so that its freei command will correctly handle a request
to free a range of consecutive inodes.

Fixed 2fsck so it will not erroneously complain that an external journal
is invalid if it is exactly 2**32 blocks.

Fixed e2fsck so it won't try checking for, and adding, a missing
lost+found directory when running in read-only mode.

Fixed e2image so that progress information won't get leaked to stdout.
(Addresses Red Hat Bugzilla: #1327329)

Fixed e2image to avoid some buffer overruns which would cause it to when
creating a "standard (non-raw, non-qcow2) image file.  Standard
e2image files are actually very rarely used, so we didn't notice when
the changes to enable the qcow2 format broke this e2image mode.

Fixed mke2fs so that the extended option "-E resize=NNN" will not turn
on the resize_inode feature when the meta_bg feature is set, since
these two features can not be set at the same time.

Fixed tune2fs so that when it disables the quota feature, it updates all
of the backup superblocks.

Fixed dumpe2fs that would cause it to abort when run using an image file
when trying to print the journal information (which is not present in
an e2image created image file).

Fixed a potential integer overflow in e2reefrag.
(Addresses-Debian-Bug: #718205)

Enhance debugfs so that when copying a sparse file from a native file
system into the file system image using the "write" command, it will
create a sparse file into the destination file system.

Enhanced debugfs so it can support a command line which is up to 8k
long.

E2image will refuse (unless the -f option is specified to force the
issue) to create a raw or qcow image using a mounted file system,
unless the -f option is specified to force the issue.

E2image has been optimized for using it to efficiently copy a file
system by only copying the allocated blocks, by using the options -ra.
New options (-o and -O) have been added so that a source and
destination offset can be given.  The -p option will print progress
information so the user will know how far along the copy is going.
And finally, the new option -c is useful for updating a file system on
an SSD, by avoiding unnecessary writes.  E2image can also shift a file
system image by doing an in place move.

Fix a regression introduced in 1.42.8 which would cause e2fsck to
erroneously report uninitialized extents past the EOF (as determined by
i_size) to be invalid.

Fixed resize2fs so under a corner case when an inode has a complex
extent tree, it will not corrupt an interior node in the extent tree.

Fixed resize2fs which would sometimes corrupt a file system when
shrinking a file system to a minimum size using resize2fs -M.
(Addresses Debian Bug: #660793)

Fixed resize2fs so that it will relocate inode table blocks if this
becomes necessary when shrinking the file system.

Fixed resize2fs and e2fsck so they will not crash when hit a failure
while operating on a file system with the MMP feature enabled.

Fixed a bug in debugfs which caused it create an invalid inode when
trying to write a zero-length file.

E2fsck will no longer crash if it tries to delete an invalid
extent-mapped symlink.

E2fsck will no longer crash if it comes across an directory which is
larger than 2GB (which is not allowed in valid file systems).

Fixed debugfs's help texts to fully document all options and otherwise
be more helpful.

Updated/fixed various man pages.  (Addresses Debian Bugs: #586218,
#669730, #698076)

Fixed various Debian Packaging Issues (#698879, #721365)


Programmer's Notes
------------------

Fix sparse, gcc -Wall and clang nits.

Update config.{guess,sub} to the latest version

Fixed various memory and file descriptor leaks on various error paths,
as well as some missing error return checks, which were found using
Coverity.

Run sparse against source files when building e2fsprogs with 'make
C=1'.  If instead C=2, it configures basic ext2 types for bitwise
checking with sparse, which can help find the (many many) spots where
conversion errors are (possibly) happening.

Allow the regression test to be run in chrooted environments where
/etc/mtab might be missing.

The ext2fs_punch() function, which was introduced in 1.42, was broken
in many ways, but this was never noticed since it wasn't used for
anything significant.  Some of the bugs include failing when trying to
punch a completely sparse file, failing when punching an extent-mapped
inode when the starting block was at the beginning of the inode, and
not being able to punch a single block (where start_blk == end_block).
It also didn't handle being passed an invalid, too-large ending block
number, and didn't handle properly terminate at the right place when
operating on an indirect-mapped inode.

Fixed some minor typo's in the error catalog for libext2fs.

Fixed ext2fs_file_set_size2() so that if it truncates an inode by
setting the file size, to zero the rest of the block to the end of the
file, so that if an ext4 FUSE driver tries to extended the file, that
we don't avoid stale data from being returned.

Fixed ext2fs_bmap() to disallow clients from trying to map or set
logical blocks which are larger than what an extent-mapped or indirect
block-mapped inode can allow.

If debugfs (or some userspace program using libext2fs) creates a file
which is larger than 2GB, make sure the large_file feature flag gets
set.

Fix a bug in ext2fs_link() where if there is multiple empty slots in the
directory which are large enough, the directory entry could get
inserted more than once in the directory.

If quota support is disabled (which is the default), make sure that
all traces of the quota support is removed from usage messages, man
pages, and tune2fs must not be able to enable the quota file system
feature.  (Addresses Red Hat Bugzilla: #1010709)

The ext2fs_file_write() now updates i_size on a successful write,
instead of only updating i_size wen the file is closed via
ext2fs_file_close().

Added a shell script, populate-extfs.sh which uses debugfs to populate
an ext2/3/4 file system image from a given directory.  It is similar
to the genext2fs program, but it supports ext3 and ext4 file system.

Add changes to the libext2fs library to support block group
descriptors larger than 64 bytes (for future compatibility).

Fixed an off-by-one bug in ext2fs_file_set_size2() so that it will not
leave an extra block in the file when truncating the file down to
size.

The html version info pages are now built using makeinfo --html
instead of the unmaintained and now-obsolete texi2html program.


E2fsprogs 1.42.8 (June 20, 2013)
================================

As a part of mke2fs's option parsing cleanup, the use of the -R option
will give a warning that it is deprecated (it has been so documented
since 2005) and -E should be used instead.

Mke2fs will not give warnings about the bigalloc and quota options in
quiet mode.

If an invalid journal size is given to mke2fs, it will now complain
and exit sooner.

Debugfs was erroneously giving spurious error messages for certain
extent_inode subcommands which take arguments (split_node,
replace_node, and insert_node).  This has been fixed.

Fix the parsing of the 's' (sectors) in parse_num_blocks2, which among
other programs is used by mke2fs.

Change mke2fs so that it does not set the root directory to the real
uid/gid of the mke2fs process.  Add the extended option root_owner to
override this behavior.

Fix resize2fs when shrinking file systems to make sure that bitmap
blocks aren't left outside the bounds of the shrunken file system.
This could happen with flex_bg file systems that were grown using
the old online resizing algorithm.

E2fsck will now detect and repair corrupted extent trees which contain
invalid extents at the end of the extent tree leaf block.

E2fsck will now longer complain about zero length extended attribute
values.

Fix a regression introduced in e2fsprogs v1.42 which caused e2image -s
to crash.

Add safety check so tune2fs will not attempt to set the inode size to
be larger than the block size.

Fix e2fsck so it can check a read-only root file system with an
external journal.  (Addresses Debian Bug: #707030

Fix off-line resizing of file systems with flex_bg && !resize_inode
(Addresses Debian Bug: #696746)

Fix e2image with large (> 32-bit) file systems (Addresses Debian Bug:
#703067)

Enhance chattr to allow clearing the extent flag if the kernel allows
migrating extent based files to use indirect blocks.

Update German translation.

Updated/fixed various man pages.  (Addresses Debian Bugs: #712429,
#712430, #707609)

Fixed various Debian Packaging Issues (Addresses Debian Bug #708307)

Programmer's Notes
------------------

Use secure_getenv() in preference to __secure_getenv().

Optimize CPU utilization of ext2fs_bg_has_super().

Fix ext2fs_llseek() on 32-bit i386 systems to work correctly when
SEEK_CUR is used with large files.

The ext2fs_read_inode_full() function will no longer use
fs->read_inode() if the caller has requested more than the base 128
byte inode structure and the inode size is greater than 128 bytes.

Fix build failure with --enable-jbd-debug.

Clean up filtering of outputs for the regression tests by using a
common sed script.

Fix gcc -Wall and clang nits.


E2fsprogs 1.42.7 (January 21, 2013)
===================================

Add warnings to mke2fs, resize2fs, and tune2fs that the bigalloc and
quota features are still under development.  For more information
please see:
    * https://ext4.wiki.kernel.org/index.php/Bigalloc
    * https://ext4.wiki.kernel.org/index.php/Quota

Add some new options to filefrag from Lustre's patches to e2fsprogs:
    * add -k option to print extents in kB-sized units (like df -k)
    * add -b {blocksize} to print extents in blocksize units
    * add -e option to print extent format, even when FIBMAP is used
    * add -X option to print extents in hexadecimal format

Fix resize2fs so that it can handle off-line resizes of file systems
with the flex_bg feature but without a resize_inode (or if we run out
of reserved gdt blocks).  This also fixes a problem where if the user
creates a filesystem with a restricted number of reserved gdt blocks,
an off-line resize which grows the file system could potentially
result in file system corruption.

Fix a resize2fs bug which could cause it to corrupt bigalloc file
systems when doing an off-line resize.

Further optimize resize2fs so it doesn't use quite as much CPU when
resizing very large file systems.

Fixed 32-bit overflow bugs which could cause resize2fs to fail and
possibly corrupt the file system while resizing 64-bit file systems.

Fix a bug in mke2fs where parsing "-E resize=NNN" will result in a
mke2fs crash due to a divide-by-zero if the 64bit file system feature
is enabled.

Add better error checking to mke2fs to check for invalid parameters
when creating bigalloc file system.

When creating bigalloc filesystems, the -g option to mke2fs will now
specify the number of clusters per block group.

Add to debugfs the functionality to corrupt a specific file system
block via the "zap_block" command.

Add to debugfs the functionality to print out a hex dump of a block in
the file system via the "block_dump" command.

Add to debugfs the functionality to manipulate the extent tree
directly via the "extent_open" command.

Fixed debugfs's mknod command so that it updates the block group
statistics.

Fix e2fsck so it can detect and fix inconsistencies in the interior
nodes of an inode's extent tree.

Fix a potential memory corruption failure in e2fsck's error path if
the call to ext2fs_open2() fails.

Fix e2fsck if its logging function is enabled in e2fsck.conf, and the
resulting file name for the log file is longer than 100 bytes, that it
properly handles this situation instead of crashing.

E2fsck will now report the amount of memory that it attempted to
allocate when a memory allocation request fails, to make it easier to
track down the problem.

Fix mke2fs's handling of the mmp_update_interval option.  (Addresses
Lustre Bug: LU-1888)

E2image can now include all data blocks in the e2image output file
when the user specifies the -a option.

If e2fsprogs is compiled without --enable-quota, make sure that
tune2fs can not turn on the feature for new-style quota support, since
afterwards, none of the e2fsprogs tools will be willing to touch that
file system.

Optimize e2fsck's so that it uses much less CPU for large file
systems.  This can result in significant speedups, especially on
CPU-constrained systems.  This was primarily done by optimizing
libext2fs's bitmap functions.  (Addresses Google Bug: #7534813)

Fix debugfs's htree command so that all its messages are sent through
the pager.

Fixed debugfs's "dump_file" and "cat" functions work correctly
on file systems whose block size is greater than 8k.

Fix e2freefrag so it works on 64-bit file systems, and so it uses much
less memory.  (Addresses Google Bug: 7269948)

Update the spd_readdir.c file in the contrib directory to include some
additions which were made in 2008 that didn't get folded into the
version which we checked into the e2fsprogs source tree.  These
enhancements include thread safety, support for readdir64_r(), and
safe_getenv() support in case spd_readdir.so is used for setuid binaries.

Update Czech, Dutch, French, German, Polish, Swedish, and Vietnamese
translations

Add a command to debugfs to create symlinks.

Document the bigalloc feature in the mke2fs man page.  (Addresses
Debian Bug: #669730)


Programmer's Notes
------------------

Fix gcc -Wall nits.

Fix a spelling typo in the libext2fs texinfo documentation.

Change the output from "make check" so that tools such as emacs's "M-x
compile" does not mistake the output as containing a compiler error.

Export two new functions from libext2fs: ext2fs_extent_node_split()
and ext2fs_extent_goto2().

The ext2fs_extents_fix_parents() was fixed so it does not modify the
location of the extent handle if its current location is at an
interior node, rather than a leaf node of an inode's extent tree.

Add a regression test to assure that e2fsck can correctly fix an
inconsistent interior node in an inode's extent tree.

The ext2fs_{mark,unmark,test}_block_bitmap_range2() functions now
correctly support bigalloc file systems which use store block usage
information in units of clusters.

Fixed the help text in the configure script for --enable-quota.

The m68k-specific bitops code has been removed since they were
incorrectly treating bit numbers with the high bit set as signed
integers.  Furthermore, modern compilers do a good enough job
optimizing the generic code there is no point in having the m68k
specific asm statements.

Fixed how we link the test programs so they always use the static
libraries, so that we test using the libraries which we have just
built.

Update config.guess and config.sub to the latest versions from the GNU
project.

Fixed the com_err.texinfo file so that it can produce a valid
postscript/pdf printed output.

Add a regression test which checks debugfs's ability to create
symlinks, named FIFO's, and device nodes.

Add a function ext2fs_symlink(), which creates symlinks to the
libext2fs library.

Add debugging code so we can test old kernel interfaces for online
resize to resize2fs.  This backwards compatibility checking is keyed
off of the RESIZE2FS_KERNEL_VERSION, which is designed to allow us to
test the functionality of the kernel's older resize ioctls without
needing to install an old version of resize2fs, and to also test a
modern resize2fs's ability to work with older kernels without having
to install an older version of the kernel.


E2fsprogs 1.42.6 (September 21, 2012)
=====================================

When mke2fs creates file systems with lazy itable initialization, the
progress updates for writing the inode table happens so quickly that
on a serial console, the time to write the progress updates can be the
bottleneck.  So mke2fs will now only update its progress indicators
once a second.

Resize2fs will skip initializing the inode tables if the kernel
supports lazy_itable_init, which speeds up growing off-line growth of
uninit_bg file systems significantly.  Resize2fs will now also
correctly set the itable_unused field in the block group descriptor to
speed up the first e2fsck after the file system is grown.

Resize2fs has been fixed so that on-line resizing of meta_bg file
systems work correctly.  This is needed to grow file system which are
larger than 16T.

Resize2fs will now correctly handle resizing file systems to 16TB on
32-bit file systems when "16TB" is specified on the command line.

Fix mke2fs so that it will be careful to set the reserved blocks ratio
larger than 50%; this can happen when creating small file systems and
when the last block group is dropped because there are not enough
blocks to support the metadata blocks in the last block group.

Fixed spelling mistake in debugfs's help message.

Fixed a potential seg fault in e2fsck when there is an I/O error while
reading the superblock.

Fixed various Debian Packaging Issues (Addresses Debian Bug #677497)

Updated/fixed various man pages.  (Addresses Sourceforge Bug:
#3559210)

Programmer's Notes
------------------

The configure option --enable-relative-symlink was broken so that it
needed to be --enable-symlink-relative-symlinks.  We will support both
for at least two years, but then the wrong configure option will be
removed.

Fixed a regression introduced in 1.42.5 so the link order for
e2fsprogs' libraries will be correct for both static and shared
linking.  (Addresses Sourceforge Bug: #3554345)

Add support for e2fsprogs to be compiled using clang/LLVM.

Fix portability problems on non-Linux systems: avoid compile failures
on systems that don't have malloc.h

Fix f_mmp regression test suite so that debugfs gets killed if the
test is interrupted.



E2fsprogs 1.42.5 (July 29, 2012)
================================

Fixed a bug with mke2fs where if there is only 8 inodes per block
group, the calculation of the number of uninitialized inodes in the
first block group would go negative.  This resulted in "mke2fs -N 256
-t ext4 /tmp/foo.img 256m" trying to write so many blocks that /tmp
would run out of space.  (Addresses Sourceforge Bug: #3528892)

Fixed a bug in how e2fsck would uniquify directory entry names.
(AddressesSourceforge Bug: #3540545)

Previously, e2fsck would only allow a mounted file system to be
checked if it was the root file system and it was mounted read-only.
Now it will allow any file system mounted read-only to be checked if
the -f option is specified.  This makes it easier to test how e2fsck
handles checking file systems which are mounted without having to test
on the root file system.

Fixed a problem if e2fsck where if the root file system is mounted
read-only, e2fsck would not clear an error indication in the journal
superblock.  Combined with a kernel bug, this would cause the e2fsck
to check the file system after every single boot.

The e4defrag program can now handle device symlinks, such as
/dev/mapper/testvg-testlv, instead of insisting on a less
human-friendly name such as /dev/dm-2.  (Addresses Red Hat Bugzilla:
#707209)

Fixed filefrag so it will not crash with a segfault on files from a
virtual file system such as /proc.  (e.g., "filefrag
/proc/partitions")

Fixed filefrag so that it correctly reports the number of extents.
(Addresses Red Hat Bugzilla: #840848)

Fixed a file descriptor leak in logsave which could cause it to hang.
(Addresses Debian Bug: #682592)

Fixed e2fsck so that the file system is marked as containing an error
if the user chooses not to fix the quota usage information.

Fixed tune2fs so that it correctly removes the quota feature when the
last quota inode is removed.

Fix tune2fs so that after removing a quota inode, the block bitmap is
updated; otherwise, e2fsck would complain after running 'tune2fs -O
^quota <dev>'.

Fix tune2fs so that when converting a file system from using legacy
quota files to the new quota file system feature with hidden quota
files, the accounting for these files is handled correctly so that
e2fsck doesn't complain.

Improved e2fsck's verbose reporting statistics, and allow the more
verbose reporting to be enabled via /etc/e2fsck.conf.

Fixed various Debian Packaging Issues (Addresses Debian Bug #678395)

Updated/fixed various man pages.  (Addresses Debian Bugs: #680114)


Programmer's Notes
------------------

Fixed portability problems on other operating systems (e.g., Hurd and
FreeBsd) caused by the attempted inclusion of <sys/quota.h>.

Make sure that shared libraries link with the shared libraries built
in the build tree, instead of the system provided libraries.
Previously, libraries and executables were linked with the system
libraries if present, and possibly using static archives instead of
shared libraries.  This was also problematic since if libext2fs.so is
linked with a static libcom_err.a from system, the build system would
attempt to link without -lpthread.  (Addresses Sourceforge Bug:
#3542572)


E2fsprogs 1.42.4 (June 12, 2012)
================================

Fixed more 64-bit block number bugs (which could end up corrupting
file systems!) in e2fsck, debugfs, and libext2fs.

Fixed e2fsck's handling of the journal's s_errno field.  E2fsck was
not properly propagating the journal's s_errno field to the superblock
field; it was not checking this field if the journal had already been
replayed, and if the journal *was* being replayed, the "error bit"
wasn't getting flushed out to disk.

Fixed a false positive complaint by e2fsck if all of the extents in
the last extent tree block are uninitialized and located after the end
of the file as defined by i_size.

The dumpe2fs will now display the journal's s_errno field if it is
non-zero, and it will also display the journal's 64-bit feature flag
if present.

Fix e2fsck so that it always opens the device file in exclusive mode
when it might need to modify the file system, and never if the -n
option is specified.  (Previously, there were a few corner cases where
it might get things wrong in either direction.)

E2fsck now correctly truncates or deallocates extent-mapped inodes on
the orphan list.  The root cause was a bug in libext2fs's block
iterator which could end up skipping an extent when the last block in
an extent is removed, causing the current extent to be removed from
the extent tree.

E2fsck now correctly sets the global free block and inode counts when
truncating or removing inodes on the orphan list in preen mode.
Previously, it would leave these values would be set incorrectly,
which is largely a cosmetic issue since the kernel no longer pays
attention to those fields, but it can cause spurious complaints in
subsequent e2fsck runs.

Fix i_blocks accounting when the libext2fs library needs to add or
remove an extent tree block  on bigalloc file systems.

The lsattr and chattr programs now support the No_COW flag for the
benefit of btrfs.

Debugfs now interprets date strings of the form @ddd as ddd seconds
after the beginning of the epoch.  This is handy when setting an inode
number into the d_time field when debugging orphan list handling.

Fix a precedence bug with built-in quota support which might result in
e2fsck paying attention to the quota inode field even if the built-in
quota feature flag is not set.  Fortunately, in practice that
superblock field should be zero for non-built-in quota file systems,
so it's unlikely this bug would have caused problems.

Updated/fixed various man pages.  (Addresses Debian Bugs: #674453,
#674694)

Programmer's Notes
------------------

The regression test suite can now run the integration tests in the
tests directory in parallel, via "make -jN check".

Add new test, f_zero_extent_length which tests e2fsck's handling of
the case where all of the extents in the last extent tree block are
uninitialized extents after i_size.

Add a new test, f_jnl_errno, which checks handling of an error
indication set in the journal superblock.

Fix the test f_jnl_64bit so that it properly checks e2fsck's handling
of a 64-bit journal.

Add two tests, f_orphan_indirect_inode and f_orphan_extent_inode which
tests e2fsck's handling of orphan inodes in preen mode, and truncation
of extent inodes on the orphan list.

Fixed more OS X portability issues.


E2fsprogs 1.42.3 (May 14, 2012)
===============================

Fix a bug in the Unix I/O manager which could cause corruption of file
systems with more than 16TB when e2fsprogs is compiled in 32-bit mode
(i.e., when unsigned long is 32-bits).  Also fix a bug which caused
dumpe2fs to incorrectly display block numbers > 32-bits.

Improve the support for integrated quota files (where quota is a first
class supported feature using hidden files in the ext4 file system).
Previously the quota file was getting rewritten even when it was not
necessary, and e2fsck would erroneously try to hide quota files which
were already hidden.

Quiet complaints in e2fsck when the total free blocks or inodes are
incorrect in the superblock after an system crash, since we don't
update nor depend on the superblock summaries at each commit boundary.

Fixed a regression introduced in 1.42.2 which would cause applications
outside of e2fsprogs which did not pass the EXT2_FLAG_64BIT (and so
would were still using the legacy 32-bit bitmaps) to crash.  This was
due to missing 32-bit compat code in side the function
ext2fs_find_first_zero_generic_bmap().  (Addresses Red Hat Bugzilla:
#808421)

Fix a bug which would cause mke2fs to fail creating the journal if
/etc/mtab and /proc/mounts are missing.  (Addresses Sourceforge Bug:
#3509398)

Updated/fixed various man pages.

Update Czech, Dutch, French, German, Polish, Swedish, and Vietnamese
translations

Fixed various Debian Packaging issues.


Programmer's Notes
------------------

Change the nonsensically wrong types in the function signature of the
inline function ext2fs_find_first_zero_block_bitmap2().  This was
caused by a cut and paste error; fortunately no code in e2fsprogs used
this inline function, and there are any users of this functions
outside of e2fsprogs.

Add support for systems which have valloc(), but which do have
posix_memalign() nor memalign() (such as MacOS 10.5).

Refactor and clean up the allocation of aligned buffers for Direct I/O
support.  Previously some allocations were requesting a greater
alignment factor that what was strictly necessary.  Also optimize
reading and writing bitmaps using Direct I/O when the size of the
bitmap did not fully cover the file system blocksize.

Reserve the codepoints for the INCOMPAT features LARGEDATA and
INLINEDATA.

Improved the regression test suite by adding some new integration
tests (f_jnl_32bit, f_jnl_64bit) which detect breakage of the on-disk
jbd2 format, as well as f_eofblocks which tests the new handling of
uninitialized and initialized blocks beyond i_size.  Also add a new
unit test which verifies 32-bit bitmaps support and the new
find_first_zero primitives.

Add a few dependencies to fix parallel (make -j) builds.

Removed bash'isms which were breaking the regression test suite on
systems where /bin/sh is not bash.

The config.guess and config.sub have been updated to the 2012-02-10
version.

Fix a portability problem caused by assuming the present of mntent.h
means that setmntent() exists.  Instead, explicitly test for this in
the configure script.

If the sys/signal.h header file does not exist, don't try to include
it, since it's not available on all systems.

Add support for systems that do not support getpwuid_r()

The configure script now supports a new option,
--enable-relative-symlinks, which will install relative symlinks for
the ELF shared library files.  (Addresses Sourceforge Bug: #3520767).

When building BSD shared libraries make sure the LDFLAGS variable is
passed to the linker.  Fixing this allows, for example, e2fsprogs to
be built in 32-bit mode on Mac OS X Lion (Addresses Sourceforge Bug:
#3517272)

Fix gcc -Wall nitpicks


E2fsprogs 1.42.2 (March 27, 2012)
=================================

The resize2fs program uses much less CPU and is much faster for very
large file systems.  (Addresses Debian Bug: #663237)

The seti and freei commands in debugfs can now take an optional length
argument to set and clear a contiguous range of inodes.

E2fsck will now make explicit checks for the EOFBLOCKS_FL, since we
plan to remove support for it from the kernel file system driver.  It
really wasn't very useful and was causing more problems than it
solves.  Since e2fsck will complain if inodes that should have
EOFBLOCKS_FL do not have the flag set, we are going to remove this
check from e2fsprogs first, and then only remove the flag from the
kernel much later.

The mke2fs program can now use direct I/O via "mke2fs -D".  This will
slow down the mke2fs, but it makes it more polite on a loaded server
by limiting the amount of memory that gets dirtied by mke2fs when it
is using buffered I/O.

E2fsck was needlessly closing and re-opening the file system as a side
effect of adding Multiple Mount Protection (MMP).  This isn't
necessary for non-MMP file systems, so drop it.

Print errors returned by ext2fs_open2() and ext2fs_check_desc() so we
can more easily diagnose memory allocation failures caused by
insufficient memory.  E2fsck will now abort if there are memory
allocation failures when the file system is initially opened and
during the block group descriptor checks.  (Addresses Google Bug:
#6208183)

If there are incorrect block group checks, e2fsck will now report the
incorrect and corrected checksum values.

The e2fsck program can now write log files containing the details of
the problems that were found and fixed directly, via configuration
parameters in /etc/e2fsck.conf.

Added the ability to limit the number of messages reported by e2fsck
for a given problem type.  This avoids a potential bottleneck if there
is a serial console which can cause a boot sequence to take a long
time if e2fsck needs to report many, many file system errors.

The dumpe2fs, debugfs, and tune2fs now use rbtree bitmaps, which cause
them to use much less memory for large file systems.

The dumpe2fs program will now print the expected block group checksum
if it is incorrect.  This helps to diagnose problems caused by
incorrect block group checksums.

E2fsck now checks for extents with a zero length, since the kernel
will oops if it comes across such a corrupted data structure.  (See
https://bugzilla.kernel.org/show_bug.cgi?id=42859)

E2fsck has a number of bugs relating to discard that have been fixed.
(1) Fixed a bug which could cause e2fsck to discard portions of the
inode table which were actually in use.  (2) E2fsck will now avoid
using discard if the block device doesn't zero data on discard, since
otherwise this could cause problems if the file system gets corrupted
in the future.  (3) E2fsck will now avoid using discard when it is run
in read-only mode.  (4) Fixed a bug which caused e2fsck to not issue
discards in the last block group.

E2fsck's CPU utilization in pass 5 has been optimized, which will
speed up e2fsck slightly.

E2image will now skip copying uninitialized bitmap and inode table
blocks.

Fixed mke2fs -S so it does not corrupt the first block group's
information.

E2fsck will now check the new sysfs interface to determine if we are
using the battery or AC mains.  (Addresses SourceForge Bug: #3439277)

Updated/fixed various man pages.  (Addresses Debian Bug: #665427)

Fixed various Debian Packaging issues.  (Addresses Debian Bug: #665885)

Programmer's Notes
------------------

Fixed various portability issues for non-Linux systems, particularly
MacOS X, as well as Linux systems running with the just-released glibc
2.15.

Fix file descriptor leak in ext2fs_close() if the file system with
uninit_bg is opened read/only with a backup superblock.  (Addresses
SourceForge Bug: #3444351)

Fixed an invalid return in a non-void function in the quota code.
(Addresses SourceForge Bug: #3468423)

Fixed the debian rules file so that the calls to dpkg-buildflags works
when the shell is dash.

The debian package build now uses V=1 so that there is more
information about potential build failures in debian build logs.

If the uninit flags get cleared by functions such as
ext2fs_new_inode() or ext2fs_new_block2(), we now make sure the
superblock is marked dirty and the block group descriptor checksum is
updated if necessary.

The debian rules file will now try to load debian/rules.custom of it
exists.  This file can skip various builds for speed reasons if there
is no need for the e2fsck-static or udeb packages.  Available
customizations in the rules file includes SKIP_STATIC=yes,
SKIP_BF=yes, and SKIP_DIETLIBC=yes.

In addition, if the file misc/mke2fs.conf.custom.in exists in the
source tree, it will be used instead of the standard misc.conf file in
the upstream sources.  This makes it easier for Debian-derived systems
to distribute a custom mke2fs.conf file without having to worry about
merge issues if future versions of e2fsprogs makes changes in the
upstream default version of mke2fs.conf.


E2fsprogs 1.42.1 (February 17, 2012)
===================================

The mke2fs and e2fsck now use significantly less memory when creating
or checking very large file systems.  This was enabled by adding
extent-based bitmaps which are stored using a red-block tree, since
block and inode allocations tend to be contiguous.

The command mke2fs -S is used as a last ditch recovery command to
write new superblock and block group descriptors, but _not_ to destroy
the inode table in hopes of recovering from a badly corrupted file
system.  So if the uninit_bg feature is enabled, mke2fs -S will now
set the unused inodes count field to zero.  Otherwise, e2fsck -fy
after using mke2fs -S would leave the file system completely empty.

Since mke2fs recognizes mke3fs in argv[0] to mean "mkfs.ext3", also
honor "mke4fs" to work the same as "mke2fs.ext4", since RHEL5 has
installed an mke2fs binary using that name.

The usage and help messages for the -G, -t and -T options in mke2fs
have been fixed.

If e2fsck needs to use the backup group descriptors, the
ext2fs_open2() function clears the UNINIT bits to ensure all of the
inodes in the file systems get scanned.  However, the code which reset
the UNINIT flags did not also recalculate the checksum, which produced
many spurious (and scary) e2fsck messages.  This has been fixed by
resetting checksums when the UNINIT bits are cleared.

Relax a check in e2fsck which required that the block bitmap to be
initialized when the inode bitmap is in use.  This will allow us to
eventually eliminate code from the kernel which forcibly initialized
the block bitmap when the inode bitmap is first used, requiring an
extra journal credit and disk write.  (Addresses Google Bug: #5944440)

Make sure rdebugfs (which may be installed setuid or setgid disk) does
not honor environment variables if euid != uid or egid != gid.

Debugfs's ncheck command has been optimized and now is much more
robust with faced with corrupted file systems.  The ncheck command
also now has a -c option which will verify the file type information
in the directory entry to see if matches the inode's mode bits.  This
is extremely useful when trying to use debugsfs to determine which
parts of the file system metadata can be trusted.

E2image will try to use ftruncate64() to set the i_size for raw
images, instead of writing a single null byte.  This avoid allocating
an extra block to the raw image, for those file systems and/or
operating systems that support this.  (Linux does.)  In addition, fix
a logic bug that caused the file to not be properly extended if the
size of the last hole was exactly an multiple of a megabyte.

Fixed a bug in resize2fs where for 1k and 2k file systems, where
s_first_data_block is non-zero, this wasn't taken into account when
calculate the minimum file system size for use with the -M option.

Fixed the badblocks program to honor the -s flag when in read-only -t
mode.  (Addresses Debian Bug #646629)

Update Czech, Dutch, French, Polish, and Swedish translations from the
Translation Project.

Fixed various Debian Packaging issues so that dpkg-buildflags is used
if present, which allows e2fsprogs to be built with security hardening
flags.  (Addresses Debian Bugs: #654457)

Programmer's Notes
------------------

Fix a bug in ext2fs_clear_generic_bmap() when used for 32-bit bitmaps.
This was only an issue for programs compiled against e2fsprogs 1.41
that manipulate bitmaps directly.  (Addresses Sourceforge Bugs:
#3451486)

The libext2fs library now uses sysconf() to fetch the page size, instead
of the deprecated getpagesize().

The ext2fs_get_pathname() function will return a partial path if an a
directory in the path is not a directory, displaying it as an inode
number in angle brackets instead of giving up and displaying an error.
This is much more helpful when a user is trying to debug a corrupted
file system.

Codepoints for the RO_COMPAT_REPLICA feature has been reserved.

Added a new library function, ext2fs_file_get_inode_num(), for use by
fuse2fs.

Fixed a bug in ext2fs_file_set_size2() so that when it is truncating a
file, it actually works.

The block iterator now properly honors the BLOCK_ABORT flag for
extent-based flags.  Previously, it didn't, which generally made code
be less efficient, but it could cause bugs in ext2fs_link(), for
example, by causing it to insert multiple directory entries.

Fixed an (harmless other than causing a compiler warning) use of an
uninitialized variable in e2fsck's MMP code.


E2fsprogs 1.42 (November 29, 2011)
==================================

This release of e2fsprogs has support for file systems > 16TB.  Online
resize requires kernel support which will hopefully be in Linux
version 3.2.  Offline support is not yet available for > 16TB file
systems, but will be coming.

This release of e2fsprogs has support for clustered allocation.  This
reduces the number of block (now cluster) bitmaps by allocating and
deallocating space in contiguous power-of-2 collections of blocks,
which are called clustered.  This is a file system level feature,
called 'bigalloc', which must be enabled when the file system is
initially formatted.  It is not backwards compatible with older
kernels.

Added support for the Multi-Mount Protection (MMP) feature.

E2fsck more efficiently uses scratch files for really big file
systems.  (This is a feature that has to be turned on explicitly; see
[scratch_files] in the e2fsck.conf man page.)

Fix a bug in e2fsck where if the free blocks and inodes counts are
incorrect, e2fsck would fix them without printing an error message.
This would cause a "*** FILE SYSTEM WAS MODIFIED ***" message without
any explanation of what was fixed.

E2fsck will no longer attempt to clone an extended attribute block in
pass1b handling if the file system does not support extended
attributes.

E2fsck will be more careful accidentally asking the user to continue
if the file system is mounted, so that an escape sequence won't cause
a false positive.  (Addresses Debian Bug: #619859)

E2fsck now uses less cpu time in pass 5 when large portions of the
bitmaps are uninitialized.

E2fsck will no longer segfault when a corrupted file system has a bad
extent, and removing it leads to a block needing to be deallocated.
(Addresses SourceForge Bug: #2971800)

E2fsck will catch termination signals (segfaults, bus errors, sigfpe)
and print debugging information to make it easier to find potential
problems.

E2fsck will check to see if the bad block inode looks insane, and will
skip trying to use if it certain fields which should be zero look
non-zero.  This is to avoid a corrupted bad block inode causing e2fsck
to get confused and do more harm than good.

If e2fsck modifies certain superblock fields which the kernel doesn't
look at, it will now mark the superblock as dirty without marking the
file system as changed.  This avoids signaling the init scripts that
a reboot is necessary, since the kernel isn't going to look at those
fields, so it won't care if they have been changed.

Fixed a bug in the libext2fs library (in the binary search routine of
the icount abstraction) that could (very, very rarely) cause e2fsck to
die in the middle of pass 1 or pass 2 processing.

E2fsck will not try to do a discard operation if the -n option was
specified on the command line.

E2fsck now supports an extended "discard" option which will cause
e2fsck to attempt discard all unused blocks after a full, successful
file system check.

E2fsck will check for the bad block inode to make sure it looks sane
before trusting it, to avoid causing more harm than good to the file
system.

E2fsck now returns additional status bits in its exit code if it
aborts early in the e2fsck run.

E2fsck now correctly calculates the maximum file size in the case of
the huge_file file system feature enabled without extents.

The mke2fs and e2fsck programs now tries to use the punch hole command
as a "discard" when operating on normal files.

The e2image program now supports the qcow2 format, which is a more
efficient way of capturing file system dumps.

Mke2fs now supports the [devices] stanza in mke2fs.conf which allows
per-device defaults to be specified in the configuration file.

Mke2fs now supports the reserved_ratio relation in the [defaults] and
[fs_types] section in mke2fs.conf.

Mke2fs now creates extent-mapped directories for the root and
lost+found directories.

Mke2fs will skip zero'ing the journal if the extended option
"lazy_journal_init" is specified.  This can save a lot of time, but it
does add a small amount of risk if the system crashes before the
journal is overwritten entirely once.  It is especially useful for
testing.

Mke2fs will now create file systems that enable user namespace
extended attributes and with time- and mount count-based file
system checks disabled.

Mke2fs will not set a stride or strip size of one block based on block
device attributes obtained from sysfs.

Mke2fs now displays a progress report during the discard process.

Mke2fs now handles extreme file system parameters correctly which
previously caused the inodes per group to drop below 8, leading to a
segfault.  (The inodes per group must be a multiple of 8, but the code
didn't correctly deal with an inodes per group count less than 8.)

Mke2fs and tune2fs previously would give an error if the user tried
setting the stride and stripe-width parameters to zero; but this is
necessary to disable the stride and stripe-width settings.  So allow
setting these superblock fields to zero.  (Addresses Google Bug:
#4988557)

Mke2fs now gives a warning if the auto-detected block size exceeds the
system's page size.

If the enable_periodic_fsck option is false in /etc/mke2fs.conf (which
is the default), mke2fs will now set the s_max_mnt_count superblock
field to -1, instead of 0.  Kernels older then 3.0 will print a
spurious message on each mount then they see a s_max_mnt_count set to
0, which will annoy users.  (Addresses Debian Bug: #632637)

The default mke2fs.conf now has entries for "big" and "huge", which
are needed for very big file systems.

The resize2fs program now has support for a new online resize ioctl
that can support file systems > 16TB, once it arrives in v3.x kernels.

Fixed bug which caused resize2fs to fail when shrinking an empty file
system down to its minimal size.  (Addresses Sourceforge Bug #3404051)

Fixed tune2fs's mount options parsing.  (Addresses Debian Bug: #641667)

Allow tune2fs to remove external journals if the device can not be
found.

Debugfs's icheck will now correctly find inodes which use the
searched-for block as an extended attribute block.

Debugfs now has a new "punch" command which remove blocks from the
middle of an inode.

Debugfs now has a new "e2freefrag" command which analyzes the free
space fragmentation of the file system, using the same code as the
e2freefrag program.

Debugfs now has a "filefrag" command which displays information about
a file's fragmentation.

Add support to build a metadata-only, read-only, stripped-down version
of debugfs called rdebugfs.

Fixed a potential stack overrun bug in debugfs.

The badblocks program now correctly recovers from I/O errors when
direct I/O is being used.  The badblocks command now also supports a
-B option which forces the use of buffered I/O, and the -v option will
provide a more detailed breakdown of read, write, and failed
comparison errors.

Added e4defrag tool which uses the EXT4_IOC_MOVE_EXT ioctl.

Added support for journals larger than 2GB.

Support using both hard links and symlinks when installing e2fsprogs.

Add overflow checking to tune2fs -i's fsck interval, which must fit in
a 32-bit field.

The debugfs command now has a new 'blocks' command which prints out
data blocks of a particular inode in a format which is useful for
scripting.

Filefrag will report 0 extents correctly in verbose mode.  (Addresses
RedHat Bugzilla: #653234)

Filefrag has been fixed so its -v report prints the correct expected
block number (previously there had been an off-by-one error).  In
addition, it will now display the number of contiguous extents when -v
is not specified.  This makes it consistent with the number of extents
printed when the -v option was specified.  In addition, the number of
contiguous extents is far more interesting/useful than the number of
physical extents for very large files.  (Addresses Debian Bug:
#631498, #644792)

Logsave's usage message has been fixed.  (Addresses Debian Bug:
#619788)

Avoid an infinite loop in ext2fs_find_block_device() if there are
symlink loops in /dev caused by a buggy udev.

Added a useful "fallocate" program to the contrib directory.

Fixed an ABI compatibility problem in libext2fs which broke the dump
program.  Also added back some macros which dump needed so it could
compile against the latest version of ext2_fs.h (Addresses Debian Bug:
#636418)

Fixed parsing of MNTOPT_ options for tune2fs and debugfs (Addresses
Debian Bug: #641667)

Added internationalization support for libcom_err error table strings.

Fixed various spelling mistakes found in various output strings found
by I18N translators.

Update translations: French, Chinese, Germany, Indonesian, Swedish,
Vietnamese, Polish, Dutch, Czech.  (Addresses Debian Bugs: #520985,
#620659)

Fixed various Debian Packaging issues.  (Addresses Debian Bugs:
#614662, #632169, #641838, #627535, #629355)

Updated/clarified man pages.  (Addresses Debian Bugs: #639411,
#642193, #634883)

Programmer's Notes
------------------

Initial support for quota as an integrated feature, where the quota
files are hidden system files that are automatically maintained by
e2fsck is present, although disabled by default.  It must be enabled
by using the configure option --enable-quota.

Reserved the on-disk fields for the metadata checksum and snapshot
features, which are currently in development.

The ext2fs library now has the new functions ext2fs_punch(),
ext2fs_get_memzero() and ext2fs_file_get_inode().

The ext2fs library now has support for calculating the crc32c checksum
(via the new functions ext2fs_crc32c_be and ext2fs_crc32c_le).

The I/O manager now supports the discard operation.

Reserved file system code points for new 1st class quota feature.

Shortened the compilation lines by moving the autoconf defines to
a config.h header file.

Fixed a potential free of an uninitialized pointer in
ext2fs_update_bb_inode().

Fixed miscellaneous gcc -Wall and coverity warnings.

Fixed portability issues for Mac OS X, Hurd, and FreeBSD.

Fixed a build failure when OMIT_COM_ERR is defined.

Improved error checking and fixed memory leaks caused by error return
paths.

Add ext2fs_flush2() and ext2fs_close2() which takes a flag to allow
the fsync() to be skipped.

Added a test for extent-mapped journals by mke2fs and tune2fs.

Added a test for creating a large (over 4GB) journal using mke2fs.

Added a test to make sure the inode size remains constant.

The regression test script now prints the list of which tests failed.

The regression test system now uses /tmp for its scratch files, which
is often a tmpfs mounted file system and hence much faster.

The i_e2image test was fixed so it works with valgrind.

Tests that rely on debugfs are now skipped if debugfs wasn't built.

Fixed the dependencies for "make check" so all required dependencies
are built before running the regression tests.

A link to com_err.h is installed in $(includedir) during a "make install".

The po/*.gmo files are automatically rebuilt if they are missing or
out of date.  This helps out Debian packaging.

Allow ext2fs_get_memalign() to compile on systems that don't have
posix_memalign().

Fixed a namespace leak in libext2fs (tdb_null).