Whamcloud - gitweb
tools/e2fsprogs.git
16 years agopo: update nl.po (from translationproject.org)
Benno Schulenberg [Sun, 6 Apr 2008 20:11:56 +0000 (16:11 -0400)]
po: update nl.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agopo: update fr.po (from translationproject.org)
Samuel Thibault [Wed, 2 Apr 2008 00:34:24 +0000 (20:34 -0400)]
po: update fr.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agoRemove bashism from configure script
Theodore Ts'o [Wed, 2 Apr 2008 00:32:55 +0000 (20:32 -0400)]
Remove bashism from configure script

Thanks to Mike Frysinger for pointing this out.

Addresses-Sourceforge-Bug: 1921969

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUse -E instead of the deprecated -R option in the mke2fs man page
Andreas Dilger [Sun, 30 Mar 2008 18:02:00 +0000 (14:02 -0400)]
Use -E instead of the deprecated -R option in the mke2fs man page

Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Avoid core dump when using the -N option
Theodore Ts'o [Sat, 29 Mar 2008 18:46:48 +0000 (14:46 -0400)]
e2fsck: Avoid core dump when using the -N option

This bug was accidentally introduced by commit 1dc506cb.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebian: Use UID/GID ranges from adduser.conf, if present
Michael Spang [Sat, 29 Mar 2008 11:32:47 +0000 (07:32 -0400)]
debian: Use UID/GID ranges from adduser.conf, if present

Addresses-Debian-Bug: #473179

Signed-off-by: Michael Spang <mspang@uwaterloo.ca>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agologsave: Pass any SIGTERM or SIGINT signals to child process
Theodore Ts'o [Wed, 26 Mar 2008 13:42:00 +0000 (09:42 -0400)]
logsave: Pass any SIGTERM or SIGINT signals to child process

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Include the device name in the progress information
Theodore Ts'o [Wed, 26 Mar 2008 13:09:09 +0000 (09:09 -0400)]
e2fsck: Include the device name in the progress information

Also make sure the device name has no spaces in it, to avoid confusing
displays, and make ctx->filesystem_name and ctx->device_name allocated
memory to avoid potential problems in the future.

Addresses-Launchpad-Bug: #203323
Addresses-Sourceforge-Bug: #1926023

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agofsck: Treat "ext4" and "ext4dev" as ext* filesystems
Theodore Ts'o [Wed, 26 Mar 2008 12:58:25 +0000 (08:58 -0400)]
fsck: Treat "ext4" and "ext4dev" as ext* filesystems

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agofsck: Make -Cn work correctly where n != 0
Theodore Ts'o [Wed, 26 Mar 2008 12:53:13 +0000 (08:53 -0400)]
fsck: Make -Cn work correctly where n != 0

Previously, fsck was only passing in -Cn to the first e2fsck process
to start up, and enabling the progress information by sending a
SIGUSR1 signal.  This didn't work if the progress information was
intended to go to file descriptor, since the information was never
passed to e2fsck.

So we now pass the progress fd in as a negative number if the progress
information is intended to be initially suppressed.

Addresses-Launchpad-Bug: #203323
Addresses-Sourceforge-Bug: #1926023

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Support a negative argument to -C to suppress progress information
Theodore Ts'o [Wed, 26 Mar 2008 12:26:01 +0000 (08:26 -0400)]
e2fsck: Support a negative argument to -C to suppress progress information

If a negative progress argument is given to -C, initially suppress the
progress information.  It can be enabled later by sending the e2fsck
process a SIGUSR1 signal.

Addresses-Launchpad-Bug: #203323
Addresses-Sourceforge-Bug: #1926023

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agofsck: Fix -C handling so that subsequent progress bars are enabled
Theodore Ts'o [Wed, 26 Mar 2008 12:11:11 +0000 (08:11 -0400)]
fsck: Fix -C handling so that subsequent progress bars are enabled

The exiting fsck instance wasn't marked as DONE, so the safety checks
thought a progress bar was still in progress, and so we didn't enable
another filesystem's checking.

Addresses-Debian-Bug: #432865
Addresses-Launchpad-Bug: #203323
Addresses-Sourceforge-Bug: #1926023

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agomke2fs: Clarify man page that -T is interpreted using the current timezone
Theodore Ts'o [Fri, 21 Mar 2008 13:25:23 +0000 (09:25 -0400)]
mke2fs: Clarify man page that -T is interpreted using the current timezone

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebugfs, tune2fs: Handle daylight savings time when parsing a time string
Theodore Ts'o [Fri, 21 Mar 2008 13:10:09 +0000 (09:10 -0400)]
debugfs, tune2fs: Handle daylight savings time when parsing a time string

We need to set tm_isdst to -1 so that mktime will automatically
determine whether or not daylight savings time (DST) is in effect.
Previously tm_isdst was set to 0, which caused mktime to interpret the
time as if it was always not using DST.

Addresses-Debian-Bug: #471882

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebugfs: Don't use a pager if stdout is not a tty
Theodore Ts'o [Thu, 20 Mar 2008 14:57:55 +0000 (10:57 -0400)]
debugfs: Don't use a pager if stdout is not a tty

If we are redirecting the output of debugfs to a file or to another
process via a pipe, there's no point sending the output to a pager.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agopo: update de.po (from translationproject.org)
Philipp Thomas [Tue, 18 Mar 2008 18:36:12 +0000 (14:36 -0400)]
po: update de.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agoChange valgrind options in test_script to conform with valgrind 3.2.3
Theodore Ts'o [Sat, 15 Mar 2008 15:16:56 +0000 (11:16 -0400)]
Change valgrind options in test_script to conform with valgrind 3.2.3

Newer versions of valgrind seem to have renamed --wierd-hacks to
--sim-hints.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoext2_fs.h: Rename EXT4_ORPHAN_FS to be EXT3_ORPHAN_FS
Theodore Ts'o [Sat, 15 Mar 2008 05:25:51 +0000 (01:25 -0400)]
ext2_fs.h: Rename EXT4_ORPHAN_FS to be EXT3_ORPHAN_FS

No application will ever use the ORPHAN_FS flag, since it only shows
up in kernel memory, but it's been pointed out it was first used in
ext3, and so it should be renamed for accuracy.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate release notes and debian changelog for 1.40.8-2 release
Theodore Ts'o [Fri, 14 Mar 2008 18:29:51 +0000 (14:29 -0400)]
Update release notes and debian changelog for 1.40.8-2 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebian: Remove preinstall script which is no longer necessary
Theodore Ts'o [Fri, 14 Mar 2008 19:13:45 +0000 (15:13 -0400)]
debian: Remove preinstall script which is no longer necessary

The preinstall script checked for dpkg --asert-support-predepends,
which has been true since 1996.  Also it removed configuration files
which haven't been around since well before Debian sarge, more than
two stable releases ago.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agouuidd: Fix issues identified by SuSE's security team
Matthias Koenig [Fri, 14 Mar 2008 17:29:35 +0000 (13:29 -0400)]
uuidd: Fix issues identified by SuSE's security team

SuSE's security team audited uuidd and came up with these issues.
None of them are serious given that uuidd runs setuid as a
unprivileged user which has no special access other than libuuid
directory, but it's good to get them fixed.

Signed-off-by: Matthias Koenig <mkoenig@suse.de>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix ext2fs_swap_inode_full() for in-inode xattrs on big-endian machines
Eric Sandeen [Fri, 29 Feb 2008 22:10:21 +0000 (16:10 -0600)]
Fix ext2fs_swap_inode_full() for in-inode xattrs on big-endian machines

After the fix for resize2fs's inode mover losing in-inode
extended attributes, the regression test I wrote caught
that the attrs were still getting lost on powerpc.

Looks like the problem is that ext2fs_swap_inode_full()
isn't paying attention to whether or not the EA magic is
in hostorder, so it's not recognized (and not swapped)
on BE machines.  Patch below seems to fix it.

Yay for regression tests.  ;)

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agoUpdate release notes, version files for 1.40.8 release v1.40.8
Theodore Ts'o [Fri, 14 Mar 2008 00:25:51 +0000 (20:25 -0400)]
Update release notes, version files for 1.40.8 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoDon't fail the "make install" if the info files have not been built
Theodore Ts'o [Thu, 13 Mar 2008 22:52:18 +0000 (18:52 -0400)]
Don't fail the "make install" if the info files have not been built

We don't fail the "make all" if the "makeinfo" command isn't there, so
we shouldn't fail the "make install", either.

Addresses-Sourceforge-Bug: #1586610

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agopo: update de.po (from translationproject.org)
Philipp Thomas [Thu, 13 Mar 2008 22:21:09 +0000 (18:21 -0400)]
po: update de.po (from translationproject.org)

Addresses-Debian-Bug: #302512
Addresses-Debian-Bug: #370247
Addresses-Debian-Bug: #401092
Addresses-Debian-Bug: #412882

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agoe2fsck: Use "Clear Htree" instead of "Clear" when the Htree is corrupt
Theodore Ts'o [Thu, 13 Mar 2008 14:44:54 +0000 (10:44 -0400)]
e2fsck: Use "Clear Htree" instead of "Clear" when the Htree is corrupt

Change the prompt so it is clear to the user that e2fsck will be
clearing the htree information, not the directory inode itself, when
the htree information has proven to be corrupt.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2image: Use open64() so that "e2image -I" works on image files > 2GB
Theodore Ts'o [Thu, 13 Mar 2008 05:03:28 +0000 (01:03 -0400)]
e2image: Use open64() so that "e2image -I" works on image files > 2GB

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Fix directory i_size handling
Theodore Ts'o [Thu, 13 Mar 2008 04:58:54 +0000 (00:58 -0400)]
e2fsck: Fix directory i_size handling

If a directory's i_size is bigger than the number of blocks, don't try
to allocate extra empty blocks to the end of the directory; there's no
real point to do that.  Also, if a directory's i_size is not a
multiple of the blocksize, flag that as a mistake so it can be fixed.

This more elegantly addresses the problem which was found on Bas van
Schaik's filesystem.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agolibext2fs: Add ext2fs_dblist_get_last() and ext2fs_dblist_drop_last()
Theodore Ts'o [Thu, 13 Mar 2008 04:34:16 +0000 (00:34 -0400)]
libext2fs: Add ext2fs_dblist_get_last() and ext2fs_dblist_drop_last()

Add two new functions which allows the caller to examine the last
directory block entry added to the list, and to drop if it necessary.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Handle a pass 2 "should never happen" error gracefully
Theodore Ts'o [Wed, 12 Mar 2008 20:10:48 +0000 (16:10 -0400)]
e2fsck: Handle a pass 2 "should never happen" error gracefully

Turns out a "should never happen" error can indeed happen very easily
if a directory with an htree index has an incorrect, and too-large,
i_size field.  This patch fixes this so that we handle this situation
gracefully, allowing filesystems with this error to be fixed.

In another patch I will clean up the specific problem which caused the
internal "should never happen" error from happening at all, but patch
will prevent e2fsck from crashing, and prompt the user to remove the
htree index, so it can be rebuilt again after pass 3.

Thanks to Bas van Schaik at Tetra for giving me access to his system
so this problem could be debugged.

Addresses-Launchpad-Bug: #129395

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Fix obvious typo in an "internal error" message
Theodore Ts'o [Wed, 12 Mar 2008 16:12:50 +0000 (12:12 -0400)]
e2fsck: Fix obvious typo in an "internal error" message

Thanks to Philipp Thomas for pointing this out.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agopo: update cs.po (from translationproject.org)
Petr Pisar [Tue, 11 Mar 2008 16:08:52 +0000 (12:08 -0400)]
po: update cs.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agopo: update vi.po (from translationproject.org)
Clytie Siddall [Mon, 10 Mar 2008 19:08:29 +0000 (15:08 -0400)]
po: update vi.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agopo: update sv.po (from translationproject.org)
Göran Uddeborg [Mon, 10 Mar 2008 19:08:29 +0000 (15:08 -0400)]
po: update sv.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agopo: update pl.po (from translationproject.org)
Jakub Bogusz [Mon, 10 Mar 2008 19:08:29 +0000 (15:08 -0400)]
po: update pl.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agopo: update nl.po (from translationproject.org)
Benno Schulenberg [Mon, 10 Mar 2008 19:08:29 +0000 (15:08 -0400)]
po: update nl.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agoresize2fs: Fix resizing filesystems with large inodes
Theodore Ts'o [Sun, 9 Mar 2008 01:20:40 +0000 (20:20 -0500)]
resize2fs: Fix resizing filesystems with large inodes

Use ext2fs_get_next_inode_full() in resize2fs and clean up large inode
handling; previous attempt was not properly handling all cases, and
was incorrectly setting i_extra_isize.  This caused some extended
attributes to get removed or randomly assigned to other inodes as a
result of the resize, which can be unfortunate on systems using
SELinux.

The previous commit didn't fix things completely on big-endian systems
like PowerPC.

Addresses-Red-Hat-Bugzilla: #434893

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoblkid: Fix portability problem caused by using uint instead of unsigned int
Theodore Ts'o [Sun, 9 Mar 2008 01:01:05 +0000 (20:01 -0500)]
blkid: Fix portability problem caused by using uint instead of unsigned int

This is needed to fix compilation with DJGPP; thanks to Cristophe
Grenier for pointing this out.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix mke2fs man page regarding 128-byte default inode size
Theodore Ts'o [Sun, 9 Mar 2008 00:48:56 +0000 (19:48 -0500)]
Fix mke2fs man page regarding 128-byte default inode size

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate release notes, version files for 1.40.7 release v1.40.7
Theodore Ts'o [Fri, 29 Feb 2008 03:51:49 +0000 (22:51 -0500)]
Update release notes, version files for 1.40.7 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebian: Avoid adding an empty directory in the uuid-dev package
Theodore Ts'o [Fri, 29 Feb 2008 04:43:34 +0000 (23:43 -0500)]
debian: Avoid adding an empty directory in the uuid-dev package

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebian: Add required ${shlibs:Depends} to libss2's depend header
Theodore Ts'o [Fri, 29 Feb 2008 04:43:03 +0000 (23:43 -0500)]
debian: Add required ${shlibs:Depends} to libss2's depend header

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix up debian copyright files to avoid Lintian errors
Theodore Ts'o [Fri, 29 Feb 2008 04:42:26 +0000 (23:42 -0500)]
Fix up debian copyright files to avoid Lintian errors

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate e2fsprogs translation template and po/gmo files
Theodore Ts'o [Fri, 29 Feb 2008 02:47:05 +0000 (21:47 -0500)]
Update e2fsprogs translation template and po/gmo files

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agotune2fs: Remove support for clearing the SPARSE_SUPER feature
Theodore Ts'o [Fri, 29 Feb 2008 02:26:01 +0000 (21:26 -0500)]
tune2fs: Remove support for clearing the SPARSE_SUPER feature

Clearing SPARSE_SUPER is dangerous; it can result in a filesystem
which e2fsck can't fix easily.  Since there is very few good reasons
for wanting to turn this feature off, disable tune2fs's abiity to do
this.  Users who really want this can use debugfs.

Also, deprecate the tune2fs -s option.  Remove it from the man page
and usage message.

Addresses-Sourceforge-Bug: #1840286

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate mke2fs usage message
Theodore Ts'o [Fri, 29 Feb 2008 02:12:31 +0000 (21:12 -0500)]
Update mke2fs usage message

Add options supported by mke2fs to its usage message.

Addresses-Sourceforge-Bug: #1751393

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoAdd manpage references to debugfs/tune2fs commands
Andreas Mohr [Fri, 29 Feb 2008 01:50:05 +0000 (20:50 -0500)]
Add manpage references to debugfs/tune2fs commands

Make debugfs and tune2fs reference each other in the "SEE ALSO"
section.

Addresses-Sourceforge-Patches: #1399325

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix makefile dependency issues for various install targets
Mike Frysinger [Fri, 29 Feb 2008 01:41:17 +0000 (20:41 -0500)]
Fix makefile dependency issues for various install targets

Addresses-Sourceforge-Patches: #1903484
Addresses-Sourceforge-Patches: #1903466
Addresses-Sourceforge-Patches: #1903456

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoImprove descriptions for the r_move_itable and r_resize_inode tests
Theodore Ts'o [Fri, 29 Feb 2008 00:54:03 +0000 (19:54 -0500)]
Improve descriptions for the r_move_itable and r_resize_inode tests

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoAdd regression test for resize2fs moving inodes with inline xattrs
Eric Sandeen [Thu, 28 Feb 2008 19:22:12 +0000 (13:22 -0600)]
Add regression test for resize2fs moving inodes with inline xattrs

The image contains an inode past the resize point with an inline xattr
value of "propervalue".  It also contains deleted inodes which had the
xattr value "oldvalue".

When resize2fs moves the inode, it should copy the xattr as well, so
that post-resize we see "propervalue" not "oldvalue".

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoresize2fs: Fix movement of large (greater than 128 byte) inodes
Eric Sandeen [Thu, 28 Feb 2008 19:24:17 +0000 (13:24 -0600)]
resize2fs: Fix movement of large (greater than 128 byte) inodes

inode_scan_and_fix() in resize2fs needs to do read/write of the full
inode to be sure it gets all data from larger (>128 byte) inodes.

Addresses-Red-Hat-Bugzilla: #434893

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Print an explicit message when the bad block inode is updated
Theodore Ts'o [Wed, 27 Feb 2008 20:10:20 +0000 (15:10 -0500)]
e2fsck: Print an explicit message when the bad block inode is updated

Some users were confused about why the filesystem was modified.

Addresses-Sourceforge-Bug: #756460

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoAdd support for manipulating large_file feature in mke2fs and tune2fs
Theodore Ts'o [Wed, 27 Feb 2008 20:01:19 +0000 (15:01 -0500)]
Add support for manipulating large_file feature in mke2fs and tune2fs

Previously we just let the kernel and e2fsck do this automatically,
but e2fsck will no longer automatically clear the large_file feature.
It still isn't really necessary to worry about this feature flag
explicitly, but some users seem to care.

Addresses-Red-Hat-Bugzilla: #258381

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agofsck: Don't complain about an old-style fstab if the fstab is empty
Theodore Ts'o [Wed, 27 Feb 2008 19:19:18 +0000 (14:19 -0500)]
fsck: Don't complain about an old-style fstab if the fstab is empty

Addresses-Debian-Bug: 468176

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebugfs: Fix find_free_block to avoid reporting the block more than once
Theodore Ts'o [Wed, 27 Feb 2008 18:38:56 +0000 (13:38 -0500)]
debugfs: Fix find_free_block to avoid reporting the block more than once

Addresses-Sourceforge-Bug: #1096315

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix gcc -Wall warnings for lib/blkid/probe.c
Theodore Ts'o [Wed, 27 Feb 2008 07:06:08 +0000 (02:06 -0500)]
Fix gcc -Wall warnings for lib/blkid/probe.c

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Don't clear the LARGE_FILES feature flag
Theodore Ts'o [Wed, 27 Feb 2008 05:00:30 +0000 (00:00 -0500)]
e2fsck: Don't clear the LARGE_FILES feature flag

Stop clearing the EXT2_FEATURE_RO_COMPAT_LARGE_FILE flag automatically
if there are no large files in the filesystem.  It's been almost a
decade since there have been kernels that don't support this flag, and
e2fsck clears it quietly without telling the user why the filesystem
has been changed.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agolibblkid: Add error checking to devicemapper code to avoid segfaults
Theodore Ts'o [Wed, 27 Feb 2008 03:24:42 +0000 (22:24 -0500)]
libblkid: Add error checking to devicemapper code to avoid segfaults

If a device mapper volume disappears while libblkid code is running,
it is possible for the devicemapper code to return errors, and since
libblkid wasn't checking for error returns, it would dereference a
null pointer and crash.  Add error checking to prevent this.

Addresses-RedHat-Bugzilla: #433857

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoEnhance e2fsck's reporting for reporting unsupported filesystem features
Theodore Ts'o [Wed, 27 Feb 2008 01:59:56 +0000 (20:59 -0500)]
Enhance e2fsck's reporting for reporting unsupported filesystem features

Addresses-Sourceforge-Feature-Request: #1175808

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agolibext2fs: Add EXT2_FLAG_NONFREE_ON_ERROR to ext2fs_open2()
Theodore Ts'o [Wed, 27 Feb 2008 01:45:36 +0000 (20:45 -0500)]
libext2fs: Add EXT2_FLAG_NONFREE_ON_ERROR to ext2fs_open2()

Add a flag which returns the partially completed filesystem object so
e2fsck can print more intelligent error messages.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix dumpe2fs parsing of explicit superblock/blocksize parameters
Theodore Ts'o [Wed, 27 Feb 2008 00:05:33 +0000 (19:05 -0500)]
Fix dumpe2fs parsing of explicit superblock/blocksize parameters

The dumpe2fs syntax documented in the man page has been broken for
some time due to getopt() changes.  Change the option syntax in
dumpe2fs to be one which is more extensible and consistent with the
format for extended options in mke2fs and tune2fs.

Addresses-Sourceforge-Bug: #1830994

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoEliminate really obsolete references to Linux 2.0/2.1/2.2 in tune2fs man page
Theodore Ts'o [Tue, 26 Feb 2008 22:38:48 +0000 (17:38 -0500)]
Eliminate really obsolete references to Linux 2.0/2.1/2.2 in tune2fs man page

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agotune2fs: Add support to clear the resize_inode feature
Theodore Ts'o [Tue, 26 Feb 2008 22:31:06 +0000 (17:31 -0500)]
tune2fs: Add support to clear the resize_inode feature

This requires an fsck aftwards.  We don't allow setting the
resize_inode feature because extensive work to tune2fs or e2fsck to
safely relocate blocks is necessary in order to reserve the blocks
needed by the resize inode.

Addresses-Red-Hat-Bugzilla: #167816

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agotune2fs: Cleanup feature handling to avoid merge conflicts
Theodore Ts'o [Tue, 26 Feb 2008 20:08:14 +0000 (15:08 -0500)]
tune2fs: Cleanup feature handling to avoid merge conflicts

Use a more abstract set of feature tests to avoid merge conflicts as
we add support for new features in the maint, master, next, and pu git
branches.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoTeach tune2fs to use e2p_edit_feature2() to provide better error handling
Theodore Ts'o [Tue, 26 Feb 2008 19:27:57 +0000 (14:27 -0500)]
Teach tune2fs to use e2p_edit_feature2() to provide better error handling

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agolibe2p: New e2p_edit_feature2 which provides better error handling
Theodore Ts'o [Tue, 26 Feb 2008 19:26:01 +0000 (14:26 -0500)]
libe2p: New e2p_edit_feature2 which provides better error handling

This creates a new enhanced edit_feature function for libe2p which
supports a different set of feature flags that are OK to clear as
opposed to set, and which returns more specific information about why
the user provided an invalid edit feature command.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoe2fsck: Document that the -D option finds duplicate file names
Theodore Ts'o [Tue, 26 Feb 2008 17:50:51 +0000 (12:50 -0500)]
e2fsck: Document that the -D option finds duplicate file names

Document in the e2fsck man page that e2fsck finds duplicate filenames
only when the -D option is passed to e2fsck.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebian: Pass 100 to UID_MIN and GID_MIN, not 1
Theodore Ts'o [Tue, 26 Feb 2008 17:11:47 +0000 (12:11 -0500)]
debian: Pass 100 to UID_MIN and GID_MIN, not 1

According to the policy, UIDs and GIDs in the range 1-100 are reserved
to be globally allocated by the base-passwd package. The libuuid1
postinstall script passes in UID_MIN=1 and GID_MIN=1.  The useradd and
groupadd commands seems to skip UID's between 1 and 100 anyway, but
it's confusing, so we should pass in UID_MIN=100 and GID_MIN=100 for
clarity and in case useradd and groupadd ever changes their behaviour.

Addresses-Debian-Bug: #466929

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agolibe2p: Change Raid to RAID in display option
Theodore Ts'o [Fri, 22 Feb 2008 21:52:27 +0000 (16:52 -0500)]
libe2p: Change Raid to RAID in display option

Update m_raid_opt test so that it reflects the code change.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix memory leak in ext2fs_alloc_block()
Theodore Ts'o [Tue, 19 Feb 2008 13:33:50 +0000 (08:33 -0500)]
Fix memory leak in ext2fs_alloc_block()

If a block buffer was not supplied and ext2fs_alloc_block() returned
with no errors, it would leak a temporary block buffer.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoblkid: detect LVM2 physical volumes
Eric Sandeen [Wed, 30 Jan 2008 23:25:03 +0000 (17:25 -0600)]
blkid: detect LVM2 physical volumes

Bits liberally stolen from lvm2 userspace.

Addresses-Red-Hat-Bugzilla: #409321

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agolibe2p: Make list_super2() print the RAID stride and stripe-width
Theodore Ts'o [Tue, 19 Feb 2008 03:59:42 +0000 (22:59 -0500)]
libe2p: Make list_super2() print the RAID stride and stripe-width

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoAdd support for setting RAID stride and strip-width via mke2fs and tune2fs
Theodore Ts'o [Tue, 19 Feb 2008 03:56:25 +0000 (22:56 -0500)]
Add support for setting RAID stride and strip-width via mke2fs and tune2fs

This is useful for mballoc to align block allocation on the RAID
stripe boundaries.

Signed-off-by: Rupesh Thakare <rupesh@clusterfs.com>
Signed-off-by: Andreas Dilger <adilger@clusterfs.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebugfs: Add new superblock fields to the set_super_value command
Theodore Ts'o [Tue, 19 Feb 2008 03:16:52 +0000 (22:16 -0500)]
debugfs: Add new superblock fields to the set_super_value command

Signed-off-by: Rupesh Thakare <rupesh@clusterfs.com>
Signed-off-by: Andreas Dilger <adilger@clusterfs.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoprint "mostly-printable" xattr strings in debugfs
Eric Sandeen [Wed, 30 Jan 2008 03:30:46 +0000 (21:30 -0600)]
print "mostly-printable" xattr strings in debugfs

Taking a cue from getfattr... if a string is "mostly"
printable characters, go ahead & print as a string,
and escape what's left over.

so we get:

Extended attributes stored in inode body:
  selinux = "system_u:object_r:root_t:s0\000" (28)

instead of:

Extended attributes stored in inode body:
  selinux = "73 79 73 74 65 6d 5f 75 3a 6f 62 6a 65 63 74 5f 72 3a 72 6f 6f 74 5f 74 3a 73 30 00 " (28)

(selinux includes the trailing null in "len" so it
never prints as a string today)

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agoUpdate to latest samba tdb library before LGPLv3 change
Theodore Ts'o [Sun, 17 Feb 2008 11:59:21 +0000 (06:59 -0500)]
Update to latest samba tdb library before LGPLv3 change

The major changes were:

* Fix realloc() leak on failure case from Jim Meyering
* Fixed various problems in transaction lock code
* Made transaction_brlock() static
* Added more fine-grained locking features

Moved from svn revision #22080 to #23590

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoMake fsck ignore mounted filesystems if given the -M option
Theodore Ts'o [Sat, 16 Feb 2008 19:14:24 +0000 (14:14 -0500)]
Make fsck ignore mounted filesystems if given the -M option

Adapted from the SuSE patch, but fixes a number of very serious
problems with the patch in SLES:

1) This changeset uses -M instead of -m; most lowercase options are
reserved for use by the filesystem-specific fsck programs.  All new
fsck options must be upper case.

2)  This changeset will skip the root filesystem in "fsck -AM", which
the SLES patch will not do.

3)  Loading /proc/mounts into the fs_info can cause -t opts matching to
malfuction.  So this changeset uses a simplified version of the
ismounted.c function from the ext2fs library.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoAdd portability checks to support DJGPP
Christophe GRENIER [Sat, 16 Feb 2008 11:10:56 +0000 (12:10 +0100)]
Add portability checks to support DJGPP

DJGPP lacks sys/select.h and sys/un.h; add header checks to be more
portable.

Signed-off-by: Christophe Grenier <grenier@cgsecurity.org>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 years agoconfigure.in: Don't use the dc command unless it is required
Theodore Ts'o [Sat, 16 Feb 2008 02:28:48 +0000 (21:28 -0500)]
configure.in: Don't use the dc command unless it is required

Some systems don't have the 'dc' command installed, and this causes
configure to print a warning message unnecessarily for a standard
(non-WIP and non-pre) release of e2fsprogs.

It's easy enough to avoid this problem, so let's do it.

Addresses-Sourceforge-Bug: #1893024

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agouuidd: Fix pid file so it has the correct pid number
Theodore Ts'o [Sat, 16 Feb 2008 02:24:30 +0000 (21:24 -0500)]
uuidd: Fix pid file so it has the correct pid number

The pid file was getting created before the fork(), so it had the
incorrect pid number.  No one noticed for a while, since "uuidd -k"
will kill the daemon and it has enough automatic convenience functions
that it's usually not necessary to refer to the pid file except as a
convenient place for uuidd to lock against multiple instances of the
daemon starting up.

Addresses-Sourceforge-Bug: #1893244

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agolibuuid: use fcntl locking instead of lockf
Theodore Ts'o [Fri, 15 Feb 2008 22:41:38 +0000 (17:41 -0500)]
libuuid: use fcntl locking instead of lockf

Cygwin doesn't support lockf(), so move to fcntl() locking as more
portable.  Also fix a bug which could cause get_lock() to loop forever
if the attempt to lock the file fails for some reason.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate Czech, Dutch, Polish, Sweedish, and Vietnamese translations
Theodore Ts'o [Fri, 15 Feb 2008 22:12:18 +0000 (17:12 -0500)]
Update Czech, Dutch, Polish, Sweedish, and Vietnamese translations

Install updates from the Translation Project.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoSet the C locale in the tests/test_script driver
Theodore Ts'o [Sun, 10 Feb 2008 12:58:08 +0000 (07:58 -0500)]
Set the C locale in the tests/test_script driver

Since the test_script driver uses [A-Za-z], we need to set the locale
so we don't get bitten on locals where the sort order might be
different.

Addresses-Sourceforge-Bug: #1890526

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix incorrect distribution name in Debian's Changelog
Theodore Ts'o [Sun, 10 Feb 2008 05:24:53 +0000 (00:24 -0500)]
Fix incorrect distribution name in Debian's Changelog

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate release notes, version files for 1.40.6 release v1.40.6
Theodore Ts'o [Sun, 10 Feb 2008 05:16:37 +0000 (00:16 -0500)]
Update release notes, version files for 1.40.6 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agolibcom_err: Use thread local storage to fix reentrancy problems
Theodore Ts'o [Sun, 10 Feb 2008 04:39:27 +0000 (23:39 -0500)]
libcom_err: Use thread local storage to fix reentrancy problems

Address the theoretical problem of two threads trying to format a
different unknown error code by using TLS.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoDocument the BLKID_FILE environment variable in the libblkid man page
Theodore Ts'o [Sun, 10 Feb 2008 04:35:03 +0000 (23:35 -0500)]
Document the BLKID_FILE environment variable in the libblkid man page

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate e2fsprogs translation template and Vietnamese and Czech translations
Theodore Ts'o [Sun, 10 Feb 2008 04:19:35 +0000 (23:19 -0500)]
Update e2fsprogs translation template and Vietnamese and Czech translations

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoblkid: Add support for returning labels for UDF filesystems
Theodore Ts'o [Sun, 10 Feb 2008 04:12:56 +0000 (23:12 -0500)]
blkid: Add support for returning labels for UDF filesystems

Addresses-Sourceforge-Bug: #1886394

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoDon't try to create $DESTDIR/etc/init.d as part of make install
Theodore Ts'o [Sun, 10 Feb 2008 03:34:43 +0000 (22:34 -0500)]
Don't try to create $DESTDIR/etc/init.d as part of make install

This isn't necessary since we don't install the init.d script (and
it's not the recommended way to start uuidd anyway).

Addresses-Sourceforge-Bug: #1885085

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoblkid: Flush cached filesystem information on any error other than EPERM
Theodore Ts'o [Sun, 10 Feb 2008 03:30:20 +0000 (22:30 -0500)]
blkid: Flush cached filesystem information on any error other than EPERM

USB devices can return ENOMEDIUM, and when the filesystem cached
information wasn't flushed, it resulted in the wrong location of a
filesystem to be returned to the caller.  The only justification for
using cached information when the open fails is in the case of a
permission denied error.

Addresses-Debian-Bug: #463787

16 years agoAllow tune2fs to set and clear the test_fs flag on ext4 filesystems
Theodore Ts'o [Sun, 10 Feb 2008 03:22:38 +0000 (22:22 -0500)]
Allow tune2fs to set and clear the test_fs flag on ext4 filesystems

Also allow the label to be set/cleared on ext4 filesystems via e2label
and tune2fs.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoblkid: Automatically chose between ext4 and ext4dev as appropriate
Theodore Ts'o [Fri, 8 Feb 2008 21:04:12 +0000 (16:04 -0500)]
blkid: Automatically chose between ext4 and ext4dev as appropriate

Add logic that on Linux systems will check for the presence of the
ext4dev filesystem; if it isn't present, fall back to ext4 for
filesystems that are marked as being "OK for use on test filesystem
code".  If they are OK for use for in-development filesystem code, it
should also be fine to use stable filesystem code if there is no test
filesystem code (ext4dev) available.

The reverse is not true, of course.  We don't ever want to mount a
production filesystem using test filesystem code unless the user gives
us explicit permission via "tune2fs -E test_fs".

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoFix bug in e2fsck which caused it to core dump if --enable-jbd-debug is used
Theodore Ts'o [Thu, 31 Jan 2008 19:22:24 +0000 (14:22 -0500)]
Fix bug in e2fsck which caused it to core dump if --enable-jbd-debug is used

Missing curly braces from a python programmer; my bad for noticing it!

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebian: Fix packaging problem caused by dpkg 1.14.16
Theodore Ts'o [Wed, 30 Jan 2008 12:20:54 +0000 (07:20 -0500)]
debian: Fix packaging problem caused by dpkg 1.14.16

Addresses-Debian-Bug: #436058

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoUpdate release notes, version files for 1.40.5 release v1.40.5
Theodore Ts'o [Mon, 28 Jan 2008 01:01:25 +0000 (20:01 -0500)]
Update release notes, version files for 1.40.5 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoCreate filesystems with the ext_attr feature by default
Theodore Ts'o [Mon, 28 Jan 2008 00:38:46 +0000 (19:38 -0500)]
Create filesystems with the ext_attr feature by default

Since recent kernels have a tendency to set this feature willy-nilly,
let's just enable by default.  It's only very old kernels that don't
support it any more.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoCreate new filesystems with 256-byte inodes by default
Theodore Ts'o [Mon, 28 Jan 2008 00:30:27 +0000 (19:30 -0500)]
Create new filesystems with 256-byte inodes by default

This makes it easier to upgrade to ext4 in the future, and it speeds
up extended attributes handling --- important on SELinux systems!

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agoDon't build e2fsck statically by default anymore
Theodore Ts'o [Sun, 27 Jan 2008 22:43:10 +0000 (17:43 -0500)]
Don't build e2fsck statically by default anymore

Also removed the --enable-dynamic-static configure option.

Unfortunately the usefulness of building e2fsck statically is gone on
all modern distributions, since everything else on the system is built
dynamically these days.  In fact on some distributions it is almost
impossible to build programs statically any more.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
16 years agodebian: Don't use dietlibc on platforms that don't support it
Theodore Ts'o [Sun, 27 Jan 2008 22:20:44 +0000 (17:20 -0500)]
debian: Don't use dietlibc on platforms that don't support it

Addresses-Debian-Bug: #459475

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>