The scratch_files feature is not really needed except on 32-bit
platforms, since tdb's performance is pretty awful given how we are
using it. Maybe SQLite would be faster, but for 64-bit platforms,
enabling swap works fairly well, especially using the rbtree for the
bitmap abstraction.
We leave tdb for Android since it's unlikely that someone will be
trying to connect petabyte+ sized file systems to a mobile handset.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If the journal superblock is corrupt and the user declines to fix it
(or runs e2fsck -n), make sure the error messages are clear and
explain that e2fsck cannot (safely) proceed.
Addresses-Debian-Bug: #768162
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Enable the following security features: stack protection, fortify,
read-only relocation tables, immediate dynamic symbol binding, and
text segment ASLR by enabling position independent executable
(PIE).
Special handling is provided for shared library and statically linked
executables. For all the gory details please see:
https://lists.debian.org/debian-devel/2016/05/msg00302.html
Distributions who want to do their own special thing can set CFLAGS,
CFLAGS_SHLIB, CLFAGS_STLIB, LDFLAGS, LDFLAGS_SHLIB and LDFLAGS_STATIC
as appropriate.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Use a large_inode so that when e2fsck is fixing a file system with
project quota enabled, the correct project id's quota is adjusted when
a corrupted inode is deleted.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Some operating systems may not define EBADMSG and EUCLEAN, so for
better portability use our own private error code numbers.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The fs_offset entry stores the filesystem offset. This allows for an
easy undo, because one does not have to remember/specify the
filesystem offset manually.
The fs_offset entry is implemented as a compatible feature.
Signed-off-by: Marcus Huewe <suse-tux@gmx.de>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Support key extension if the tdb_data_size is an arbitrary integer
multiple of the channel's block size. Before, key extension was only
possible if the tdb_data_size and the channel's block size were
equal.
Note: a key, whose data is the result of a short read, will be
extended if the tdb_data_size and the channel's block size are equal
(that's what the old code did) (if tdb_data_size is an arbitrary
integer multiple (> 1) of the channel's block size, the key might
be extended as well (depending on the keysize)).
Signed-off-by: Marcus Huewe <suse-tux@gmx.de>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The old code has some issues, for example, when backing up fs block 0
(can be reproduced via "mke2fs -z undo -b 1024 -E offset=1024 out 1024"):
* backing_blk_num is set to ULLONG_MAX instead of 0
* data is read from the beginning of the file instead of offset 1024
* data_ptr is set to read_ptr - 1024 ("invalid" address)
Hence, the wrong fs block is associated with the wrong data.
For details, see also commit 76da764639cbfcc998f13c263a11a4601bcb9961.
Signed-off-by: Marcus Huewe <suse-tux@gmx.de>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Avoid a potential out-of-bounds memory access if the group passed to
ext2fs_clear_block_uninit() or ext2fs_clear_inode_uninit() is greater
than the number of groups in the file system. This prevents a failure
in resize2fs when to allocate a block when growing the file system
significantly.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Hurd and *BSD is not going to have FS_IOC_FIEMAP, at least not at
Linux's codepoint.
Addresses-Debian-Bug: #822576
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Fix resize2fs so that the location of the backup superblocks when the
sparse_super2 feature is enabled is properly set when growing the file
system from a single block group to larger file system sizes.
Also fix a bug where the block group summary statistics in some cases
when exapnding a sparse_super2 file system.
Finally, accurately calculate the file system metadata overhead of the
last block group in sparse_super2 file systems.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Added offset support to the following functions:
- unix_cache_readahead
- unix_discard
- unix_zeroout
Signed-off-by: Marcus Huewe <suse-tux@gmx.de>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
After
http://git.kernel.org/cgit/fs/ext2/e2fsprogs.git/commit/lib/ext2fs/llseek.c?id=274d46e1d35af423d0292d63c4d0ad7a03be82ba
with
__linux__
defined(HAVE_LSEEK64) && defined(HAVE_LSEEK64_PROTOTYPE)
SIZEOF_OFF_T >= SIZEOF_LONG_LONG
it leads to ext2fs_llseek() doing a "return lseek(fd, offset, origin);"
Which fails for offsets > 32bit.
Also, with
__linux__
!(defined(HAVE_LSEEK64) && defined(HAVE_LSEEK64_PROTOTYPE))
defined(HAVE_LLSEEK)
SIZEOF_OFF_T == SIZEOF_LONG_LONG
my_llseek is not defined at all. And there is no need to define
llseek as lseek, as llseek is never used.
Luckily ext2fs_llseek() then does "return lseek(...);"
It would seem that my_llseek should be used in both places.
Addresses-Google-Bug: #13340735
Change-Id: Ie7330300c9c1ca103eaaef97536dcf10adbbba02
Signed-off-by: JP Abgrall <jpa@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
FreeBSD doesn't have libintl support built into the libc, and
libsupport requires libintl support. So we need to make sure it is
available when we link against the libsupport library.
Also, work around *BSD's interesting interpretation of how
_XOPEN_SOURCE is supposed to work.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The major() cpp macro is defined as requiring sys/types.h to be
included with _BSD_SOURCE defined. However, in older glibc's this
hasn't been strictly required and the stdlib.h header file included
sys/types.h implicitly. Fix this so that more aggressive
distributions run into build errors.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
in with andchanges. Lines starting
Glibc has depcreated _SVID_SOURCE and _BSD_SOURCE in favor of
_DEFAULT_SOURCE. So define _DEFAULT_SOURCE to shut up glibc 2.20
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Most libraries use the "-host" suffix when building for the host. This
patch renames all the libraries to use -host instead of _host.
Addresses-Google-Bug: #24619596
TEST=make dist
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When adding 'libext2fs' as a library dependency from any other binary,
the binary needs to find the library headers. Users of this library
should not hard-code the include path themselves. This patch exports the
library headers on all three version of the libext2fs library and
removes the redundant include path from executables in e2fsprogs.
Addresses-Google-Bug: #23084776
TEST=mma
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Older kernels on 64-bit machines would incorrectly encode pre-1970
ext4 dates as post-2311 dates. Detect and correct this (assuming the
current date is before 2242).
Include tests for this, as well as changes to debugfs to correctly
set crtimes.
Signed-off-by: David Turner <novalis@novalis.org>
Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If resize2fs_get_alloc_block() allocates from a BLOCK_UNINIT group, we
need to make sure that the UNINIT flag is cleared on both file system
structures which are maintained by resize2fs. This causes the
modified bitmaps to not get written out, which leads to post-resize2fs
e2fsck errors; used blocks in UNINIT groups, not marked in the block
bitmap. This was seen on r_ext4_small_bg.
This patch uses clear_block_uninit() to clear the flag,
and my problem goes away.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Replace the EXT2FS_RB_EMPTY_ROOT(), EXT2FS_RB_EMPTY_NODE(),
EXT2FS_RB_CLEAR_NODE(), IS_BLOCK_BM(), IS_INODE_BM(), and
IS_INODE_TB() macros with static inline functions to avoid
suprious compiler warnings with clang:
pass1.c:618:28: warning: equality comparison with extraneous
parentheses [-Wparentheses-equality]
if ((*((__u32 *)(entry)) == 0UL)) {
~~~~~~~~~~~~~~~~~~~~^~~~~~
pass1.c:618:28: note: remove extraneous parentheses
around the comparison to silence this warning
if ((*((__u32 *)(entry)) == 0UL)) {
^~
pass1.c:618:28: note: use '=' to turn this equality
comparison into an assignment
if ((*((__u32 *)(entry)) == 0UL)) {
^~
=
The static inline functions should compile identically, and allow
some extra compile-time checking for the arguments over macros.
Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Richard Purdie reports that libext2fs doesn't sort attribute keys in
the xattr block correctly, causing the kernel to return -ENODATA when
querying attributes that should be there. Therefore, sort attributes
so that whatever ends up in the xattr block is sorted according to
what the kernel expects.
Cc: Darren Hart <dvhart@linux.intel.com>
Reported-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tested-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Allow the filesystem to store the metadata checksum seed in the
superblock and add an incompat feature to say that we're using it.
This enables tune2fs to change the UUID on a mounted metadata_csum
FS without having to (racy!) rewrite all disk metadata.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This patch add EXT4_PROJINHERIT_FL to enable inherit feature for
project ID. If an directory has its inherit flag set, all its
newly created children will inherit its project ID. Conversely,
new inodes will get a default project ID (i.e. zero). Also, no
hard link or rename is permitted if the directory and child has
different project ID.
Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This patch adds project quota support. An new quota type PRJQUOTA(2)
is added. EXT4_PRJ_QUOTA_INO(11) is reserved for project quota inode.
The super block reservers an field s_prj_quota_inum for saving
project quota inode. And each inode adds an internal field i_projid
for saving its project ID.
Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This patch add project feature flag EXT4_FEATURE_RO_COMPAT_PROJECT.
Project feature is a read-only compat feature. Thus, an ext4 file
system with project feature enabled could only be read by ext4
kernel module without project feature support.
Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
FreeBSD 10.2 will blow up compiling its own header files in sys/file.h
if _XOPEN_SOURCE is defined.
In file included from tdb.c:59:
/usr/include/sys/file.h:209:2: error: unknown type name 'u_int'
u_int xf_flag; /* flags (see fcntl.h) */
^
1 error generated.
This is despite the fact that POSIX.1 requires comforming applications
to define _XOPEN_SOURCE (to different numbers depending on the version
of POSIX.1 the program is expecting to work against). See section
2.2.1 in POSIX.1 for chapter and verse.
Work around this by removing the _XOPEN_SOURCE declaration. This will
cause compiler warnings (and will cause builds against some versions
of Solaris to break), so only do this for FreeBSD.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If ext2fs_punch() was called with "end = ~0ULL" to indicate truncate
to the end of file it tried to compute "count" for ext2fs_punch_ind()
based on "start" and "end", but incorrectly passed "count = ~0U" even
when "start" was non-zero, causing an overflow in some cases.
The calling convention for ext2fs_punch_ind() was also gratuitously
different from ext2fs_punch() and ext2fs_punch_extent(), passing
"count" instead of "end" as the last parameter. Fix this by passing
it "end" like the other functions, and handle "count" internally.
Add checks to ext2fs_punch_ind() if "end" is at or beyond the 2^32
indirect block limit so the 32-bit internal variables don't overflow.
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The test2[] array is not 0-terminated and the create_test_list() for
loop does not terminate properly at the end of this array, but
continues until it hits the 0 at the end of test3[].
Reported-by: Hanno Boeck <hanno@hboeck.de>
Addresses: https://bugzilla.kernel.org/show_bug.cgi?id=104311
Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This code is partially derived from patches from David Turner to allow
debugfs to properly support extended timestamps.
Cc: David Turner <novalis@novalis.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext2fs_extent_set_bmap() and ext2fs_punch_extent(), fix the parents
when altering either end of an extent so that the parent nodes reflect
the added mapping.
There's a slight complication to using fix_parents: if there are two
mappings to an lblk in the tree, the value of handle->path->curr can
point to either extent afterwards), which is documented in a comment.
Some additional color commentary from Darrick:
In the _set_bmap() case, I noticed that the "remapping last block in
extent" case would produce symptoms if we are trying to remap a
block from "extent" to "next_extent", and the two extents are
pointed to by different index nodes. _extent_replace(...,
next_extent) updates e_lblk in the leaf extent, but because there's
no _extent_fix_parents() call, the index nodes never get updated.
In the _punch_extent() case, we conclude that we need to split an
extent into two pieces since we're punching out the middle. If the
extent is the last extent in the block, the second extent will be
inserted into a new leaf node block. Without _fix_parents(), the
index node doesn't seem to get updated.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Create separate predicate functions to test/set/clear feature flags,
thereby replacing the wordy old macros. Furthermore, clean out the
places where we open-coded feature tests.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If during formatting we'd lose more than 75% a block group to group
descriptors and other metadata, enable the meta_bg feature. This
enables us to create >500T filesystems with default options.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When meta_bg is enabled at mkfs time, we put at most one group
descriptor block in each blockgroup. Unfortunately, the calculation
of max overhead per bg doesn't know this, so mkfs fails when it isn't
strictly necessary. Fix it, since Dave reported that he couldn't
create a 500TB ext4 filesystem.
Reported-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If there are directory entries with file names which are less than 16
bytes, it turns out that passing less than the crypto block size to
the kernel's crypto layer will cause the kernel to crash.
However, since there never should be encrypted directory entries where
the file name is less than 16 bytes (the AES block size), change
e2fsck to offer to address this corruption by deleting the directory
entry.
(We need to checks for this condition into the kernel as well.)
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Also change ext2fs_symlink() so that the target parameter is a const
char *, thus promising that we will never change the incoming string.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Theodore Ts'o