Whamcloud - gitweb
tools/e2fsprogs.git
15 years agoe2fsprogs: Fix tst_extents output on bigendian machine.
Aneesh Kumar K.V [Fri, 18 Jul 2008 12:05:40 +0000 (17:35 +0530)]
e2fsprogs: Fix tst_extents output on bigendian machine.

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agotune2fs: Delete a stray word from a comment.
Benno Schulenberg [Thu, 17 Jul 2008 21:58:35 +0000 (23:58 +0200)]
tune2fs: Delete a stray word from a comment.

Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopartinfo: Print clearer error messages.
Benno Schulenberg [Thu, 17 Jul 2008 21:46:49 +0000 (23:46 +0200)]
partinfo: Print clearer error messages.

Also use complete sentences, instead of separate words filled
into a phrase.  And gettextize the main output message.

Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agobadblocks: Insert a missing "\n" in the usage string.
Benno Schulenberg [Thu, 17 Jul 2008 21:32:02 +0000 (23:32 +0200)]
badblocks: Insert a missing "\n" in the usage string.

Take the opportunity to wrap the string to be more readable,
and sort the options somewhat alphabetically.

Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agoe2undo, tune2fs: Report the correct name for the undo program.
Benno Schulenberg [Thu, 17 Jul 2008 21:24:09 +0000 (23:24 +0200)]
e2undo, tune2fs: Report the correct name for the undo program.

Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agomke2fs: Fix enforcement of no features for revision 0 filesystems
Theodore Ts'o [Sat, 19 Jul 2008 01:08:56 +0000 (21:08 -0400)]
mke2fs: Fix enforcement of no features for revision 0 filesystems

Fix typo; instead of checking for s_feature_incompat twice, add check
for s_feature_ro_compat.

Thanks to Benno Schulenberg for noticing this problem.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoFix cross compilation issues in e2fsck/gen_crc32table
Theodore Ts'o [Sat, 19 Jul 2008 01:05:30 +0000 (21:05 -0400)]
Fix cross compilation issues in e2fsck/gen_crc32table

The Makefile should use BUILD_CFLAGS instead of ALL_CFLAGS since it
will be built for the host, not the target.

Addresses-Sourceforge-Bug: #2019287

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agopo: update de.po (from translationproject.org)
Philipp Thomas [Thu, 17 Jul 2008 15:47:00 +0000 (11:47 -0400)]
po: update de.po (from translationproject.org)

Addresses-Debian-Bugs: #313697, #401092

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agoFix inode table allocation with flexbg
Frederic Bohe [Thu, 17 Jul 2008 11:34:41 +0000 (13:34 +0200)]
Fix inode table allocation with flexbg

Disordered inode tables may appear when inode_blocks_per_group is lesser
or equal to the number of groups in a flex group.

This bug can be reproduced with:
mkfs.ext4 -t ext4dev -G512 70G

In that case, you can see with dump2fs that inode tables for groups 510
and 511 are placed just after group 51's inode table instead of being
placed after group 509's inode table.

Signed-off-by: Frederic Bohe <frederic.bohe@bull.net>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update vi.po (from translationproject.org)
Clytie Siddall [Thu, 17 Jul 2008 15:08:59 +0000 (11:08 -0400)]
po: update vi.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update cs.po (from translationproject.org)
Petr Pisar [Thu, 17 Jul 2008 15:08:58 +0000 (11:08 -0400)]
po: update cs.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agoFix missing space typo's in partinfo and resize2fs
Theodore Ts'o [Thu, 17 Jul 2008 14:50:26 +0000 (10:50 -0400)]
Fix missing space typo's in partinfo and resize2fs

These were caused by multi-line strings missing a space at the line
break.  Thanks to translator Phillipp Thomas for noticing these typo's.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agopo: update fr.po (from translationproject.org)
Theodore Ts'o [Thu, 17 Jul 2008 13:04:05 +0000 (09:04 -0400)]
po: update fr.po (from translationproject.org)

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agopo: update pl.po (from translationproject.org)
Theodore Ts'o [Thu, 17 Jul 2008 13:03:50 +0000 (09:03 -0400)]
po: update pl.po (from translationproject.org)

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agopo: update nl.po (from translationproject.org)
Theodore Ts'o [Thu, 17 Jul 2008 13:03:30 +0000 (09:03 -0400)]
po: update nl.po (from translationproject.org)

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoAdd more intelligent handling of the mke2fs.conf installation process
Theodore Ts'o [Wed, 16 Jul 2008 17:39:32 +0000 (13:39 -0400)]
Add more intelligent handling of the mke2fs.conf installation process

For people who are compiling mke2fs for their own use outside of a
package manager, we need to make sure the system /etc/mke2fs.conf is
sufficiently up-to-date that it won't cause problems, but at the same
time we don't want to blow away any user-specific customizations.

So if /etc/mk2fs.conf exists, but does not mention ext4dev, we will
move it aside to /etc/mke2fs.conf.e2fsprogs-old and then install the
new mke2fs.conf.  If the /etc/mke2fs.conf file exists but does mention
ext4dev, we install the new mke2fs.conf file as
/etc/mke2fs.conf.e2fsprogs-new.  In both cases we print a warning
mesage to the user so they can manually update /etc/mke2fs.conf with
any changes if they so desire.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agolibext2fs : Fix typo in function names in libext2fs.texinfo
Manish Katiyar [Tue, 15 Jul 2008 19:08:40 +0000 (00:38 +0530)]
libext2fs : Fix typo in function names in libext2fs.texinfo

Signed-off-by: "Manish Katiyar" <mkatiyar@gmail.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agodebian: wrap copyright files to avoid "line too long" lintian warnings
Theodore Ts'o [Tue, 15 Jul 2008 16:38:55 +0000 (12:38 -0400)]
debian: wrap copyright files to avoid "line too long" lintian warnings

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: Fix minor typo in uuid-runtime's package description
Theodore Ts'o [Tue, 15 Jul 2008 16:06:06 +0000 (12:06 -0400)]
debian: Fix minor typo in uuid-runtime's package description

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: Update changelog for 1.40.0-3 release
Theodore Ts'o [Tue, 15 Jul 2008 15:40:45 +0000 (11:40 -0400)]
debian: Update changelog for 1.40.0-3 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: Fix FTBFS problem on mips
Theodore Ts'o [Tue, 15 Jul 2008 15:38:33 +0000 (11:38 -0400)]
debian: Fix FTBFS problem on mips

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoext2fs_read_inode: Check the validity of the inode number earlier
Manish Katiyar [Tue, 15 Jul 2008 15:10:19 +0000 (11:10 -0400)]
ext2fs_read_inode: Check the validity of the inode number earlier

It looks like the right place to check for ino=0 in
ext2fs_read_inode_full() is before creating the inode cache, otherwise
since we set icache[i].ino = 0 in create_icache(), it will match the
loop below and thus we return a wrong value.

Signed-off-by: "Manish Katiyar" <mkatiyar@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoAdd an tune2fs.static target in misc/Makefile.in
Theodore Ts'o [Tue, 15 Jul 2008 04:03:16 +0000 (00:03 -0400)]
Add an tune2fs.static target in misc/Makefile.in

This is added as a convenience for distro's that need a static tune2fs
for whatever reason.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoDisable %.sh -> % GNU make rule in the top-level Makefile
Theodore Ts'o [Tue, 15 Jul 2008 02:38:01 +0000 (22:38 -0400)]
Disable %.sh -> % GNU make rule in the top-level Makefile

It's not needed for e2fsprogs, but one niche distribution likes to use
configure.sh files which can cause GNU make to inappropriately fire an
implicit rule to overwrite the configure file resulting in a fork bomb
leading to the OOM crash.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoblkid: Fix portability issue if TIOCGSIZE and TIOCGWINSZ are present
Theodore Ts'o [Mon, 14 Jul 2008 21:49:38 +0000 (17:49 -0400)]
blkid: Fix portability issue if TIOCGSIZE and TIOCGWINSZ are present

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoDon't assume that /bin/true is always in /bin
Theodore Ts'o [Mon, 14 Jul 2008 21:47:52 +0000 (17:47 -0400)]
Don't assume that /bin/true is always in /bin

... because it isn't on MacOS X

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: Update section and priority of -dbg packages to match override file
Theodore Ts'o [Mon, 14 Jul 2008 00:51:33 +0000 (20:51 -0400)]
debian: Update section and priority of -dbg packages to match override file

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoAdd debian changelog for 1.41.0-2 release
Theodore Ts'o [Sun, 13 Jul 2008 23:20:02 +0000 (19:20 -0400)]
Add debian changelog for 1.41.0-2 release

15 years agolibblkid: Make sure we never reference a null dev->bid_type
Theodore Ts'o [Sun, 13 Jul 2008 23:10:10 +0000 (19:10 -0400)]
libblkid: Make sure we never reference a null dev->bid_type

There were a few places where we don't check to make sure
dev->bid_type is non-NULL before dereferencing the pointer, mostly in
debug code.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agolibext2fs: Don't check the group checksum when !GDT_CSUM
Theodore Ts'o [Sun, 13 Jul 2008 23:03:59 +0000 (19:03 -0400)]
libext2fs: Don't check the group checksum when !GDT_CSUM

ext2fs_group_desc_csum_verify() is always checking the bg_checksum (to
make sure it is zero) even when the GDT_CSUM feature is not present.
This is normally OK, but apparently there are filesystems in the wild
where this field has not be initialized to zero.

Addresses-Debian-Bug: #490637

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agobadblocks: If nanosleep() does not exist, try using usleep() instead
Theodore Ts'o [Sun, 13 Jul 2008 20:17:57 +0000 (16:17 -0400)]
badblocks: If nanosleep() does not exist, try using usleep() instead

For portability on systems that don't have nanosleep().

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoe2undo: Don't use __FUNCTION__ for portability reasons
Theodore Ts'o [Sun, 13 Jul 2008 19:44:32 +0000 (15:44 -0400)]
e2undo: Don't use __FUNCTION__ for portability reasons

Solaris C99 apparently doesn't support it.  We should report the
program name, not the internal function name, when printing an error
in any case.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoe2fsck: Set LD_LIBRARY_PATH before running "make check" programs
Theodore Ts'o [Sun, 13 Jul 2008 19:41:35 +0000 (15:41 -0400)]
e2fsck: Set LD_LIBRARY_PATH before running "make check" programs

This is needed so that "make check" works in the e2fsck library even
if the shared libraries are not yet installed, and so that we run
those programs against the version of the libraries built in the build
tree.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoFlush stdio buffers before calling setbuf()
Theodore Ts'o [Sun, 13 Jul 2008 19:40:15 +0000 (15:40 -0400)]
Flush stdio buffers before calling setbuf()

On Solaris setbuf() will discard any pending output to the stream, so
make we call fflush() before calling setbuf().

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoDefine _XOPEN_SOURCE to 600 consistently for Solaris C99 support
Theodore Ts'o [Sun, 13 Jul 2008 19:32:37 +0000 (15:32 -0400)]
Define _XOPEN_SOURCE to 600 consistently for Solaris C99 support

Solaris's header files are very picky about which C compiler can be
used for SUSv3 conformance.  Use of C99 is not compatible with SUSv2
(_XOPEN_SOURCE=500), and C89 is not compatible with SUSv3
(_XOPEN_SOURCE=600).  Since we need some SUSv3 functions, consistently
use SUSv3 so that e2fsprogs will build on Solaris using c99.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoRemove gcc-specific varargs macro in extent.c
Theodore Ts'o [Sun, 13 Jul 2008 18:19:14 +0000 (14:19 -0400)]
Remove gcc-specific varargs macro in extent.c

Use "#ifdef DEBUG // printf... // #endif" instead of dbg_printf().

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoRemove vararg macro which causes portability problems from ext3_extent.h
Theodore Ts'o [Sun, 13 Jul 2008 18:00:53 +0000 (14:00 -0400)]
Remove vararg macro which causes portability problems from ext3_extent.h

Also remove other unneeded kernel-only cruft for ext3_extent.h.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUse K&R style comments instead of C99 comments for portability
Theodore Ts'o [Sun, 13 Jul 2008 16:27:33 +0000 (12:27 -0400)]
Use K&R style comments instead of C99 comments for portability

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoOnly use __builtin_expect if __GNUC__ > 3
Theodore Ts'o [Sun, 13 Jul 2008 12:05:54 +0000 (08:05 -0400)]
Only use __builtin_expect if __GNUC__ > 3

Gcc only supports __builtin_expect for gcc versions 2.96 and up.
Since it's tricky to check for gcc 2.95 vs 2.96 (and either are only
used on really ancient systems anyway), we only use this optimization
on 3.x and newer versions of gcc.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoe2fsck: Change kmem_cache_t to lkmem_cache_t for Solaris
Theodore Ts'o [Sun, 13 Jul 2008 12:04:36 +0000 (08:04 -0400)]
e2fsck: Change kmem_cache_t to lkmem_cache_t for Solaris

Solaris polutes the C namespace with kmem_cache_t when
you include in/netinet.h is included, so rename kmem_cache_t
to lkmem_cache_t.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoFix Solaris shell script portability issues
Theodore Ts'o [Sun, 13 Jul 2008 12:03:20 +0000 (08:03 -0400)]
Fix Solaris shell script portability issues

Solaris ships with a pathetically ancient shell in /bin/sh,
so fix various shell scripts to accomodate.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoblkid: Add new option -L which pretty-prints the device list
Theodore Ts'o [Sun, 13 Jul 2008 02:06:30 +0000 (22:06 -0400)]
blkid: Add new option -L which pretty-prints the device list

Also accessed via -o list, this new output format is much more
user-friendly and lists whether or not a particular device is mounted.

Addresses-Debian-Bug #490527

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agolibext2fs: Add stricter/earlier tests for blocksize in ext2fs_open()
Manish Katiyar [Fri, 11 Jul 2008 21:45:07 +0000 (17:45 -0400)]
libext2fs: Add stricter/earlier tests for blocksize in ext2fs_open()

Signed-off-by: "Manish Katiyar" <mkatiyar@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agolibext2fs: fix potential memory leak in ext2fs_initialize()
Manish Katiyar [Fri, 11 Jul 2008 18:42:57 +0000 (14:42 -0400)]
libext2fs: fix potential memory leak in ext2fs_initialize()

If we fail doing ext2fs_allocate_block_bitmap() or
ext2fs_allocate_inode_bitmap() we directly goto cleanup and don't free
the memory allocated to buf.

Signed-off-by: "Manish Katiyar" <mkatiyar@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate release notes, changelog, etc., for e2fsprogs 1.41 release v1.41.0
Theodore Ts'o [Thu, 10 Jul 2008 21:24:22 +0000 (17:24 -0400)]
Update release notes, changelog, etc., for e2fsprogs 1.41 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate dependencies in the Makefile for lib/ext2fs
Theodore Ts'o [Thu, 10 Jul 2008 21:17:55 +0000 (17:17 -0400)]
Update dependencies in the Makefile for lib/ext2fs

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate e2fsprogs.pot I18N translation template file
Theodore Ts'o [Thu, 10 Jul 2008 21:16:57 +0000 (17:16 -0400)]
Update e2fsprogs.pot I18N translation template file

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoMake ext2fs_check_desc() more stringent to force use of backup superbocks
Theodore Ts'o [Thu, 10 Jul 2008 20:35:05 +0000 (16:35 -0400)]
Make ext2fs_check_desc() more stringent to force use of backup superbocks

E2fsck could to do more damage to a filesystem by trying to relocate
inode tables due to corrupted block group descriptors, and the
relocation could seriously damage the filesystem.

This patch enhances ext2fs_check_desk() so it detects more
self-inconsistent block group descriptors, including the cases where
e2sck might be tempted to relocate the inode table, and reports the
block group descriptors as invalid; this will cause e2fsck to attempt
to use the backup superblocks, which hopefully have not been trashed.

Addresses-Sourceforge-Bug: #1840291

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate release checklist to include reminder to update translation
Theodore Ts'o [Thu, 10 Jul 2008 18:20:18 +0000 (14:20 -0400)]
Update release checklist to include reminder to update translation
template.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoresize2fs: Clean up the resize inode properly if necessary
Theodore Ts'o [Thu, 10 Jul 2008 18:18:41 +0000 (14:18 -0400)]
resize2fs: Clean up the resize inode properly if necessary

If the filesystem is grown to the point where the resize_inode is no
longer needed, clean it up properly so e2fsck doesn't have to.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agomke2fs,tune2fs: Enable huge_file, dir_nlink, and extra_isize features
Theodore Ts'o [Thu, 10 Jul 2008 14:49:59 +0000 (10:49 -0400)]
mke2fs,tune2fs: Enable huge_file, dir_nlink, and extra_isize features

Also change mke2fs.conf to enable huge_file,dir_nlink,extra_isize, and
uninit_bg by default for ext4 filesystems, and enable extra_isize in
the library as well.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoDon't use thread local storage if using diet libc
Theodore Ts'o [Thu, 10 Jul 2008 14:21:42 +0000 (10:21 -0400)]
Don't use thread local storage if using diet libc

...because the diet libc doesn't support TLS.

Addresses-Sourceforge-Bug: #2000654

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agobadblocks: Use standard terminology for first/last block
Theodore Ts'o [Thu, 10 Jul 2008 14:08:40 +0000 (10:08 -0400)]
badblocks: Use standard terminology for first/last block

Use consistent and standard terminology for the starting and ending
blocks for the badblocks test in the man page and in the messages
printed by the program.

Addresses-Debian-Bug: #440983, #440981

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoFix incorrect definition of stripe-width in mke2fs man page
Theodore Ts'o [Thu, 10 Jul 2008 13:40:48 +0000 (09:40 -0400)]
Fix incorrect definition of stripe-width in mke2fs man page

Also clarified the definition of the stride-size extended option as
well.

Addresses-Debian-Bug: #487849

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agomke2fs: Dumb down filesystems for GNU Hurd
Theodore Ts'o [Thu, 10 Jul 2008 13:24:25 +0000 (09:24 -0400)]
mke2fs: Dumb down filesystems for GNU Hurd

The Hurd only supports filesystems with a blocksize of 4096 bytes, and
128 byte inodes.  It also doesn't understand the journal.  So force
the defaults to be something which the Hurd can handle if "-o hurd" is
specified on the command line.

Addresses-Debian-Bug: #471977

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoAdd a comment in INSTALL regarding --disable-tls on older systems
Theodore Ts'o [Thu, 10 Jul 2008 13:08:44 +0000 (09:08 -0400)]
Add a comment in INSTALL regarding --disable-tls on older systems

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoblkid: Make regression test tolerate older versions of mkswap
Andreas Dilger [Tue, 8 Jul 2008 05:37:34 +0000 (23:37 -0600)]
blkid: Make regression test tolerate older versions of mkswap

The blkid/tests contains new tests for swap, but the type-1 swapfile
test depends on mkswap supporting the "-U" option to specify the UUID.
This is not available even on relatively recent versions of mkswap
(2.13.1 16-Jan-2008) so the test needs to be changed to handle this.
If the "-U" option is not supported, don't verify the UUID in the blkid
output during testing.

Signed-off-by: Andreas Dilger <adilger@sun.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agodebian: Add bash-static to the list of shells recommended for e2fsck-static
Theodore Ts'o [Thu, 10 Jul 2008 12:47:15 +0000 (08:47 -0400)]
debian: Add bash-static to the list of shells recommended for e2fsck-static

Addresses-Debian-Bug: #490003

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate release notes for e2fsprogs 1.41-WIP-0707 release. v1.41-WIP-0707
Theodore Ts'o [Mon, 7 Jul 2008 19:10:16 +0000 (15:10 -0400)]
Update release notes for e2fsprogs 1.41-WIP-0707 release.

15 years agoAdd build.static to util/subset.exclude as well.
Theodore Ts'o [Mon, 7 Jul 2008 20:33:53 +0000 (16:33 -0400)]
Add build.static to util/subset.exclude as well.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: Update debian policy compliance to 3.8.0.0
Theodore Ts'o [Mon, 7 Jul 2008 20:28:38 +0000 (16:28 -0400)]
debian: Update debian policy compliance to 3.8.0.0

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agolibext2fs: Fix FLEX_BG offset used by ext2fs_allocate_group_table
Theodore Ts'o [Mon, 7 Jul 2008 20:11:11 +0000 (16:11 -0400)]
libext2fs: Fix FLEX_BG offset used by ext2fs_allocate_group_table

The offset for both inode bitmaps and inode tables is overshot by one
block causing a hole between the group of bitmaps and inode tables
when initializing a filesystem using mke2fs.

Signed-off-by: Jose R. Santos <jrs@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: Create debian release tarball to obey new WIP naming convention
Theodore Ts'o [Mon, 7 Jul 2008 20:07:43 +0000 (16:07 -0400)]
debian: Create debian release tarball to obey new WIP naming convention

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoAdd ZFS test image to blkid test suite
Ricardo M. Correia [Fri, 4 Jul 2008 18:13:40 +0000 (19:13 +0100)]
Add ZFS test image to blkid test suite

Signed-off-by: Ricardo M. Correia <Ricardo.M.Correia@Sun.COM>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoe2fsck: Fix i_size checking for large extent-based files
Eric Sandeen [Mon, 7 Jul 2008 04:02:47 +0000 (23:02 -0500)]
e2fsck: Fix i_size checking for large extent-based files

This:

Truncating bigfile to 14680064000000
Pass 1: Checking inodes, blocks, and sizes
Inode 49154, i_size is 14680064000000, should be 0.  Fix<y>?

is a bit unexpected.  It's because the size is being checked against
the max sizes for bitmap files, not extent-based files.

Nick saw this with his 14TB file.

Patch below applies different size limits to the different file
formats.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: Add dbg packages which contain the stripped out debuging symbols
Theodore Ts'o [Mon, 7 Jul 2008 13:43:26 +0000 (09:43 -0400)]
debian: Add dbg packages which contain the stripped out debuging symbols

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoTreat build.static as an ingored file in the source directory
Theodore Ts'o [Mon, 7 Jul 2008 12:37:40 +0000 (08:37 -0400)]
Treat build.static as an ingored file in the source directory

Don't include it in distribution tarballs, and add it to the
.gitignore file.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agomke2fs: Print a better error msg when ext2fs_get_device_size() returns EFBIG
Theodore Ts'o [Mon, 7 Jul 2008 00:57:17 +0000 (20:57 -0400)]
mke2fs: Print a better error msg when ext2fs_get_device_size() returns EFBIG

Print a message when mke2fs uses a default blocksize from an external
journal device, and print a more self-explanatory message so that if
that blocksize is used and ext2fs_get_device_size() returns EFBIG, the
user has a better chance of understanding why mke2fs issued that error
message.

Addresses-Debian-Bug: #488663

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agomke2fs: Use a fs_type of 'journal' when creating an external journal
Theodore Ts'o [Mon, 7 Jul 2008 00:24:29 +0000 (20:24 -0400)]
mke2fs: Use a fs_type of 'journal' when creating an external journal

If creating a an external journal via "mke2fs -O journal_dev",
override the fs_type list (i.e., "ext2", "small"), and replace it with
an fs_type list of "journal".  This will prevent external journals
smaller than 512MB from being created with a block size of 1k, which
is not very useful and leads to much confusion.

Addresses-Debian-Bug: #488663

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agobadblocks: Fix crash when lots of -t patterns given
Theodore Ts'o [Sun, 6 Jul 2008 22:50:44 +0000 (18:50 -0400)]
badblocks: Fix crash when lots of -t patterns given

With more than 8 -t patterns given, badblocks will overwrite the
t_patts array boundary due to realloc not taking into account the size
of an int.  Oops.

Addresses-Debian-Bug: 487298

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoClean up some trivial build warnings
Eric Sandeen [Sun, 6 Jul 2008 22:36:56 +0000 (18:36 -0400)]
Clean up some trivial build warnings

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoblkid: Eliminate stale entries that duplicate a verified device
Theodore Ts'o [Sun, 29 Jun 2008 02:09:43 +0000 (22:09 -0400)]
blkid: Eliminate stale entries that duplicate a verified device

Addresses-Debian-Bug: #487758, #487783

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoblkid: Make sure a device returned by blkid_find_dev_with_tag exists
Theodore Ts'o [Sun, 29 Jun 2008 01:02:01 +0000 (21:02 -0400)]
blkid: Make sure a device returned by blkid_find_dev_with_tag exists

There could be stale entries in blkid file, so if the device does not
exist, skip it.

Addresses-Debian-Bug: #487758, #487783

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoblkid: use list_for_each_safe in garbage collection
Eric Sandeen [Sat, 21 Jun 2008 03:32:49 +0000 (22:32 -0500)]
blkid: use list_for_each_safe in garbage collection

We need to use list_for_each_safe in case a device gets removed from
the list during garbage collection.

Also make the manpage slightly more informative about
what the -g garbage collection option does.

Addresses-Debian-Bug: #487758, #487783

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update nl.po (from translationproject.org)
Benno Schulenberg [Tue, 24 Jun 2008 01:35:08 +0000 (21:35 -0400)]
po: update nl.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update cs.po (from translationproject.org)
Petr Pisar [Tue, 24 Jun 2008 01:35:07 +0000 (21:35 -0400)]
po: update cs.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agolibext2fs: Don't include ss/ss.h except when debugging
Theodore Ts'o [Mon, 23 Jun 2008 18:10:43 +0000 (14:10 -0400)]
libext2fs: Don't include ss/ss.h except when debugging

extent.c should only try to include ss/ssh.h when it is compiled with
-DDEBUG.  Otherwise it's not necessary and it breaks the Debian MIPS
build (and the Debian MIPS build only) because it tries to build
libext2fs without building libss as part of a MIPS-specific build
rule.

Addresses-Debian-Bug: #487675

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agodebian: parse /etc/adduser.conf file manually
Theodore Ts'o [Mon, 23 Jun 2008 15:57:54 +0000 (11:57 -0400)]
debian: parse /etc/adduser.conf file manually

The syntax for /etc/adduser.conf allows spaces around the equals sign,
i.e.:

GROUPHOMES = no

We were previously asking /bin/sh to source the file, which doesn't
support the above syntax.  So pull out the necessary fields via
sed/grep.

Thanks to Juan A. Diaz for reporting the bug and suggesting the fix.

Addresses-Debian-Bug: #487443

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agobadblocks: fix a bug in the read throttling
Iustin Pop [Wed, 18 Jun 2008 20:26:26 +0000 (22:26 +0200)]
badblocks: fix a bug in the read throttling

The previous patch was missing an #include and thus the compiler didn't
catch the (now obvious) error.

Signed-off-by: Iustin Pop <iustin@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update vi.po (from translationproject.org)
Clytie Siddall [Sat, 21 Jun 2008 18:06:00 +0000 (14:06 -0400)]
po: update vi.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update pl.po (from translationproject.org)
Jakub Bogusz [Sat, 21 Jun 2008 18:06:00 +0000 (14:06 -0400)]
po: update pl.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update fr.po (from translationproject.org)
Samuel Thibault [Sat, 21 Jun 2008 18:06:00 +0000 (14:06 -0400)]
po: update fr.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update es.po (from translationproject.org)
Benno Schulenberg [Sat, 21 Jun 2008 18:06:00 +0000 (14:06 -0400)]
po: update es.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update de.po (from translationproject.org)
Philipp Thomas [Sat, 21 Jun 2008 18:06:00 +0000 (14:06 -0400)]
po: update de.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agopo: update ca.po (from translationproject.org)
David Planella Molas [Sat, 21 Jun 2008 18:05:59 +0000 (14:05 -0400)]
po: update ca.po (from translationproject.org)

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agoblkid: Fix cc portability problem in probe_lvm2()
Theodore Ts'o [Sat, 21 Jun 2008 18:04:36 +0000 (14:04 -0400)]
blkid: Fix cc portability problem in probe_lvm2()

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoblkid: fix swap tests (again)
Eric Sandeen [Wed, 18 Jun 2008 19:58:55 +0000 (14:58 -0500)]
blkid: fix swap tests (again)

Some architectures (ppc ...) need a bigger swapfile than is shipped,
in the test image so the current re-make of swap was failing.

We could either ship a bigger image or just dd a bigger file...

There is one more minor problem with the tests; older mkswap does not
support the -U uuid specification.  I'm not sure offhand what to do
about that problem, or if it really needs fixing.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agoFix "make check" in e2fsck subdirectory to use the in-tree com_err library
Theodore Ts'o [Thu, 19 Jun 2008 14:09:08 +0000 (10:09 -0400)]
Fix "make check" in e2fsck subdirectory to use the in-tree com_err library

Thanks to Jose Santos for pointing this out

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoAdd e2undo files to the RPM specfile template
Eric Sandeen [Wed, 18 Jun 2008 21:16:35 +0000 (16:16 -0500)]
Add e2undo files to the RPM specfile template

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agoAdd a reference to e2fsck.conf(5) to the e2fsck man page
Theodore Ts'o [Wed, 18 Jun 2008 15:27:21 +0000 (11:27 -0400)]
Add a reference to e2fsck.conf(5) to the e2fsck man page

Addresses-Debian-Bug: #393313

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate e2fsprogs.pot and POTFILES.in
Theodore Ts'o [Wed, 18 Jun 2008 10:50:34 +0000 (06:50 -0400)]
Update e2fsprogs.pot and POTFILES.in

This is to allow the translation project to get started on translating
messages for e2fsprogs 1.41.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate release notes for debian experimental package
Theodore Ts'o [Wed, 18 Jun 2008 04:50:05 +0000 (00:50 -0400)]
Update release notes for debian experimental package

15 years agoMerge branch 'maint'
Theodore Ts'o [Wed, 18 Jun 2008 03:54:51 +0000 (23:54 -0400)]
Merge branch 'maint'

Conflicts:

README
resize/online.c
version.h

15 years agoUpdate release notes, changelog, and other files for v1.40.11 release v1.40.11
Theodore Ts'o [Wed, 18 Jun 2008 02:17:09 +0000 (22:17 -0400)]
Update release notes, changelog, and other files for v1.40.11 release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoClarify the debugfs man page regarding the default superblock
Theodore Ts'o [Wed, 18 Jun 2008 01:55:38 +0000 (21:55 -0400)]
Clarify the debugfs man page regarding the default superblock

Addresses-Debian-Bug: #486463

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agomke2fs, tune2fs, resize2fs: Use floating point to calculate percentages
Theodore Ts'o [Tue, 17 Jun 2008 23:58:29 +0000 (19:58 -0400)]
mke2fs, tune2fs, resize2fs: Use floating point to calculate percentages

When calculating the number reserved blocks, use floating point for
better accuracy, since for big filesystems it really makes a
difference.  In addition, mke2fs and tune2fs accepts a floating point
number from the user, so they should provide that level of accuracy.

Addresses-Debian-Bug: #452639

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agolibblkid: Fix marginal C code in probe_lvm2()
Theodore Ts'o [Tue, 17 Jun 2008 17:03:47 +0000 (13:03 -0400)]
libblkid: Fix marginal C code in probe_lvm2()

Modern gcc accepted what was there previously, but it's clearly not
correct C code, and this may have been the explanation for why a user
trying to compile a recent version of e2fsprogs failed to do so on Red
Hat 7.3.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agoUpdate version.h for 1.41-WIP-0617 snapshot release v1.41-WIP-0617
Theodore Ts'o [Tue, 17 Jun 2008 11:48:59 +0000 (07:48 -0400)]
Update version.h for 1.41-WIP-0617 snapshot release

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
15 years agobadblocks: implement read throttling
Iustin Pop [Thu, 12 Jun 2008 07:30:04 +0000 (09:30 +0200)]
badblocks: implement read throttling

Currently, badblocks will read as fast as it can from the drive. While
this is what one wants usually, if badblocks is run in read-only mode on
a drive that is in use, it will greatly degrade the other users of this
disk.

This patch adds a throttling mode for reads where each read will be
delayed by a percentage of the time the previous read took; i.e., an
invocation of '-d 100' will cause the sleep to be the same as the read
took, a value of 200 will cause the sleep to be twice as high, and a
value of 50 will cause it to be half.  This will not be done if the
previous read had errors, since then the hardware will possibly have
timeouts and that would decrease the speed too much.

This algorithm helps when the disk is used by other processes as then,
due to the increased load, the time spent doing the reads will be
higher, and correspondingly badblocks will sleep even more and thus it
will use less of the drive's bandwidth. This is different from using
ionice, as it is a voluntary (and partial) throttling.

Signed-off-by: Iustin Pop <iustin@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
15 years agobadblocks: fix a possible bug in parse_uint
Iustin Pop [Wed, 11 Jun 2008 15:55:18 +0000 (17:55 +0200)]
badblocks: fix a possible bug in parse_uint

Currently, the parse_uint() function checks errno after the strtoul()
call. But, according to the man page of strtoul():
  Since strtoul() can legitimately return 0 or LONG_MAX (LLONG_MAX for
  strtoull()) on  both  success  and failure,  the  calling program
  should  set errno to 0 before the call, and then determine if an error
  occurred by checking whether errno has a nonzero value after the call.

When using locales, it can happen that looking for the locale files is
not successful, and therefore errno will have a nonzero value from this.
And since the argument parsing is one of the first things done after
startup, parse_uint() will wrongly report errors.

The fix is to simply reset errno to zero before calling strtoul().

Signed-off-by: Iustin Pop <iustin@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>