Use ext2fs_group_first_block2() instead of ext2fs_group_first_block()
to avoid dumpe2fs from printing crazy block offsets when we have block
numbers which are larger than 32 bits.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The resize inode only works on 32-bit block numbers, so use blk_t
instead of blk64_t. This avoids some -Wconversion noise, and slims
the compiled code slightly, especially on 32-bit platforms.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
We need to store some error codes using an int to keep recovery.c as
close as possible to the recovery.c source file in the kernel.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The use of ext2fs_write_dir_block() meant that attempts to fix
deleted/unused inodes in a directory would not be fixed for file
systems with 64-bit block numbers. (And some random block with the
high 32-bits cleared would get corrupted.)
Fix a similar problem when expanding directories and when creating the
lost+found dirctory.
Signed-off-by: Kit Westneat <kwestneat@ddn.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
These memory leaks were discovered by using "valgrind
--leak-check=full" while running "e2image -I bar.img foo.e2i"
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
e2image manually opens a new IO channel, and then sets the file system
to use this new IO channel via ext2fs_rewrite+to_io(). We need to
make sure the IO channel is set to the file system's block size to
avoid some nasty buffer overruns.
[ Modified by tytso to use io_channel_set_blksize() ]
Signed-off-by: Kit Westneat <kwestneat@ddn.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Commit bf0449b1a6, which added the ability to write qcow2 files,
generalized the write_header() file to take the size of the header
structure which it writes out. Unfortunately, it changed the call
which supported original e2image format to pass in fs->blocksize,
instead of the actual size of the e2image header structure (which is
substantially smaller than fs->blocksize). This meant that we copied
in stack garbage into the e2image file, and it made valgrind quite
unhappy.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Commit 832cb612: "e2fsprogs: add (optional) sparse checking to the
build" breaks systems that are not using GNU make. In addition, it
breaks if the developer tries to build in a subdirectory (i.e., if he
or she tries running "make" in the misc or e2fsck or lib/ext2fs
directory), since CHECK_CMD is not set.
Fix this by moving the sparse setup to MCONFIG.
Cc: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Don't accept block numbers larger than 2^32 for the badblocks list,
and don't run badblocks on them either.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Passing the "-E resize=NNN" option to mke2fs sets the resize_inode
feature. However, resize_inode and meta_bg are mutually exclusive;
unfortunately, we check this constraint before we parse the extended
options. Fix this by moving this check after the calls
parse_extended_opts().
Reported-by: "Darrick J. Wong" <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When we've succesfully linked an inode into a directory, we can stop
iterating the directory.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
We need to build libquota even if the quota code is disabled. This
fixes a build regression introduced by commit 43075b42bd: 'quota: fix
disabling quota, add quota tests".
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The man page still said it was not possible to change the number of
inodes on a filesystem after creating it. You actually can by
resizing the fs, so clarify this language a bit.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Update all superblock copies when disabling the quota feature.
Added basic tests for the quota feature.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Commit 44a2cca3 disabled tune2fs -O quota when the build
didn't have --enable-quota at configure time, but that
wasn't quite enough.
We need to exclude the "-Q" option as well from tune2fs
when --enable-quota isn't specified at configure time.
Otherwise, tune2fs -Q can set the quota feature, but no other
utility will touch the filesystem due to the unknown flag,
if buitl w/o --enable-quota.
So put everything related to "-Q" under #ifdef CONFIG_QUOTA;
usage output (was missing before) and option parsing.
Addresses-Red-Hat-Bugzilla: #1010709
Reported-by: Bert DeKnuydt <Bert.Deknuydt@esat.kuleuven.be>
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If quota isn't turned on with --enable-quota, then comment
quota documentation out of the mke2fs manpage.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This script uses debugfs command to populate the ext2/3/4 filesystem
from a given directory, it is a little similar to genext2fs, but this
one fully supports ext3 and ext4.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Let debugfs do sparse copy when src is a sparse file, just like
"cp --sparse=auto"
* For the:
#define IO_BUFSIZE 64*1024
this is a suggested value from gnu coreutils:
http://git.savannah.gnu.org/gitweb/?p=coreutils.git;a=blob;f=src/ioblksize.h;h=1ae93255e7d0ccf0855208c7ae5888209997bf16;hb=HEAD
* Use malloc() to allocate memory for the buffer since put 64K (or
more) on the stack seems not a good idea.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The max length of debugfs argument is 256 which is too short, the
arguments are two paths, the PATH_MAX is 4096 according to
/usr/include/linux/limits.h, so use BUFSIZ (which is 8192 on Linux
systems), that's also what the ss library uses.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
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.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Several users have used e2image on a mounted RW filesystem, resulting in
inconsistent, useless e2images for debugging purposes.
This commit will forbid this and print an error message, although the
user can override this using a new force option.
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently, only the new 64-bit bitmap implementation supports the
block<->cluster conversions that bigalloc requires. Therefore, if we
have a bigalloc filesystem, require EXT2_FLAGS_64BITS be passed in to
ext2fs_open(). This does not mean that bigalloc file systems have to
be 64-bits; just that the userspace utilities have to be able to use
the new 64-bit capable library functions.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
Since it's impossible to address all blocks of a 64bit filesystem
without extents, have e2fsck turn on the feature if it finds (64bit &&
!extents).
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
A 64bit filesystem without extents is not terribly useful, because the
old block map format does not support pointing to high block numbers.
Warn the user who tries to create such an animal.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
implied_cluster_alloc() is written such that if the the user passes in
a logical block that is the zeroth block in a logical cluster (lblk %
cluster_ratio == 0), then it will assume that there is no physical
cluster mapped to any other part of the logical cluster.
This is not true if we happen to be allocating logical blocks in
reverse order. Therefore, search the whole cluster, except for the
lblk that we passed in.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When told to truncate a file, ext2fs_file_set_size2() should start with
the first block past the end of the file. The current calculation
jumps one more block ahead, with the result that it fails to hack off
the last block. Adding blocksize-1 and dividing is sufficient to find
the last block.
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext2fs_file_write() needs to update i_size on successful write,
otherwise, ext2fs_file_read() in same open/close cycle will not
be able to read the just written data.
This fixes a bug which results in the the problem of quotacheck
triggered on 'tune2fs -O quota' failed to write back multiple
users/groups accounting information.
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The inode and block relocation functions aren't currently compiled in
(so we don't need to worry about breaking ABI compatibility). They
were originally intended for use by resize2fs, but we never ended up
using them, so (wisely) they weren't ever included in libext2fs as an
exported interface (they're not even compiled by the Makefile).
Fix them so that in case we ever use them, so that in places where raw
data types (int, long, etc.) stood in for blk_t and blk64_t. Also fix
some sites where we should probably be using blk64_t.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Fix all the places where we should be using a blk64_t instead of a
blk_t. These fixes are more severe because 64bit values could be
truncated silently.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When we're iterating the main loop in ind_punch(), "offset" tracks how
far we've progressed into the block map, "start" tells us where to
start punching, and "count" tells us how many blocks we are to punch
after "start". Therefore, we would like to break out of the loop once
the "offset" that we're looking at has progressed past the end of the
punch range. Unfortunately, if start !=0, the if-break clause in the
loop causes us to break out of the loop early.
Therefore, change the breakout test to terminate the loop at the
correct time.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The range of blocks to punch is treated as an inclusive range on both
ends, i.e. if start=1 and end=2, both blocks 1 and 2 are punched out.
Thus, start == end means that the caller wishes to punch a single
block. Remove the check that prevents us from punching a single
block.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
During a punch operation, if we decide to delete an extent out of the
extent tree, the subsequent extents are moved on top of the current
extent (that is to say, they're memmmove'd down one slot). Therefore
it is not correct to advance to the next leaf because that means we
miss half the extents in the range! Rereading the current pointer
should be fine.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If someone tries to write a file that is larger than 2GB, we need to
set the large_file feature flag to affirm that i_size_hi can hold
meaningful contents.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The ext2fs_link helper function link_proc does not check the value of
ls->done, which means that if the function finds multiple empty spaces
that will fit the new directory entry, it will create a directory
entry in each of the spaces. Instead of doing that, check the done
value and don't do anything more if we've already added the directory
entry.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Eric Sandeen reported that Fedora's mke2fs when compiled for ppc was
creating a file system which caused problems with resize2fs -M.
Closer examination showed that the problem was file system which
looked like this:
Filesystem features: ext_attr dir_index filetype sparse_super
Inode count: 512
Block count: 1247
...
Group 0: (Blocks 1-1024)
Primary superblock at 1, Group descriptors at 2-2
Block bitmap at 66 (+65), Inode bitmap at 67 (+66)
Inode table at 68-99 (+67)
Group 1: (Blocks 1025-1246)
Backup superblock at 1025, Group descriptors at 1026-1026
Block bitmap at 1090 (+65), Inode bitmap at 1091 (+66)
Inode table at 1092-1123 (+67)
It's not obvious to me why Fedora's ppc mke2fs is creating file
systems like this (I can't reproduce this on debian ppc systems), but
resize2fs -M should be able to deal with such file systems, which is
what this test is designed to check.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If the file system is being shrunk, and a block group's inode table
falls beyond the end of the inode table, we need to try to relocate
the inode table blocks.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If the file system's inode table blocks in the last block group are
located in the middle or the end of the block group, it's possible for
resize2fs -M to use a size which will require relocating the inode
table blocks in the last block group. This can lead to all sorts of
problems, so solve it by simply guaranteeing that we will never do
that.
Reported-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Commit d3f32c2db8 was intended to detect extents found outside their
proper location in the extent tree, including invalid extents at the
end of an extent block. However, it incorrectly reported legal
uninitialized extents created by fallocate() at the end of file with
the FALLOC_FL_KEEP_SIZE flag as false positives. xfstests
generic/263 (among others) caught this problem, while the e2fsprogs
test f_uninit_ext_past_eof did not. The latter test failed to
detect the problem in part because it uses a test file whose i_size
is 0.
Add a test derived from the fsx-based test case in xfstests
generic/263 consisting of a file with non-zero length, more than
four extents total, and two uninitialized extents past EOF to
reliably reproduce commit d3f32c2db8's false positive behavior.
Signed-off-by: Eric Whitney <enwlinux@gmail.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Commit d3f32c2db8 introduced a regression that caused e2fsck failures
in xfstests generic 013, 070, 083, 091, and 263. Uninitialized
extents created by fallocate() at the end of file with the
FALLOC_FL_KEEP_SIZE flag were identified as invalid. However,
because the file size is not increased when FALLOC_FL_KEEP_SIZE is
used, uninitialized extents can correctly contain blocks located past
the end of file.
Fix this by filtering out possible invalid extents if they are
uninitialized and extend past the block containing the end of file.
Signed-off-by: Eric Whitney <enwlinux@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When we define an error in lib/ext2fs/ext2_err.et.in, we will always use
EXT2_ET_* prefix for a new error. But EXT2_NO_MTAB_FILE doesn't obey
this rule. So fix it.
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If we have an extent tree like this (from debuge2fs's "ex" command):
Level Entries Logical Physical Length Flags
...
2/ 2 60/ 63 13096 - 13117 650024 - 650045 22
2/ 2 61/ 63 13134 - 13142 650062 - 650070 9
2/ 2 62/ 63 13193 - 13194 650121 - 650122 2
2/ 2 63/ 63 13227 - 13227 650155 - 650155 1 A)
1/ 2 4/ 14 13228 - 17108 655367 3881 B)
2/ 2 1/117 13228 - 13251 650156 - 650179 24 C)
2/ 2 2/117 13275 - 13287 650203 - 650215 13
2/ 2 3/117 13348 - 13353 650276 - 650281 6
...
and we resize the fs in such a way that all of those blocks must
be moved down, we do them one at a time. Eventually we move 1-block
extent A) to a lower block, and then follow it with the other
blocks in the next logical offsets from extent C) in the next
interior node B).
The userspace extent code tries to merge, so when it finds that
logical 13228 can be merged with logical 13227 into a single extent,
it does. And so on, all through extent C), up to block 13250 (why
not 13251? [1]), and eventually move the node block as well.
So we end up with this when all the blocks are moved post-resize:
Level Entries Logical Physical Length Flags
...
2/ 2 120/122 13193 - 13193 33220 - 33220 1
2/ 2 121/122 13194 - 13194 33221 - 33221 1
2/ 2 122/122 13227 - 13250 33222 - 33245 24 D)
1/ 2 5/ 19 13228 - 17108 34676 3881 E) ***
2/ 2 1/222 13251 - 13251 33246 - 33246 1 F)
2/ 2 2/222 13275 - 13286 33247 - 33258 12
...
All those adjacent blocks got moved into extent D), which is nice -
but the next interior node E) was never updated to reflect its new
starting point - it says the leaf extents beneath it start at 13228,
when in fact they start at 13251.
So as we move blocks one by one out of original extent C) above, we
need to keep updating C)'s parent node B) for a proper starting point.
fix_parents() does this.
Once the tree is corrupted like this, more corruption can
ensue post-resize, because we traverse the tree by interior nodes,
relying on their start block to know where we are in the tree.
If it gets off, we'll end up inserting blocks into the wrong part
of the tree, etc.
I have a testcase using fsx to create a complex extent tree which
is then moved during resize; it hit this corruption quite easily,
and with this fix, it succeeds.
Note the first hunk in the commit is for going the other way,
moving the last block of an extent to the extent after it; this
needs the same sort of fix-up, although I haven't seen it in
practice.
[1] We leave the last block because a single-block extent is its
own case, and there is no merging code in that case. \o/
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
free_gdp_blocks needs to be taught to use 64-bit fields and the appropriate
getters, otherwise it'll truncate high block numbers (when, say, resizing a
>16T fs) and mark the low numbered group descriptor blocks as free. Yikes.
Reported-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
It turns out that resize2fs uses ext2fs_dup_handle to duplicate fs handles. If
MMP is enabled, this causes both handles to share MMP buffers, which is bad
news when it comes time to free both handles. Change the code to (we hope) fix
this. This prevents resize2fs from failing with a double-free error when
handed a MMP filesystem.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Under some failure cases, we can get to fatal_error()
without even having a superblock set up. In that case,
ext2fs_mmp_stop() will segfault when it tries to dereference
fs->super.
Check for the existence of a superblock before we go
down the ext2fs_mmp_stop() path to avoid this problem.
Reported-by: Hubert Kario <hkario@redhat.com>
Addresses-Red-Hat-Bugzilla: #997972
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>