If overhead is not recorded in the super block, it is calculated
during mount in kernel, for bigalloc file systems it takes
O(groups**2) in time. For a 1PB device with 32K cluster size it takes
~12 mins to mount, with most of the time spent on figuring out
overhead.
While we can not improve the overhead algorithm in kernel due to the
nature of bigalloc, we can work out the overhead during mke2fs and set
it in the super block, avoiding calculating it every time when it
mounts.
Overhead is s_first_data_block plus internal journal blocks plus the
block and inode bitmaps, inode table, super block backups and group
descriptor blocks for every group. This patch introduces
ext2fs_count_used_clusters(), which calculates the clusters used in
the block bitmap for the given range.
When bad blocks are involved, it gets tricky because the blocks
counted as overhead and the bad blocks can end up in the same
allocation cluster. In this case we will unmark the bad blocks from
the block bitmap, convert to cluster bitmap and get the overhead, then
mark the bad blocks back in the cluster bitmap.
Reset the overhead to zero when resizing, we can not simply count the
used blocks as overhead like we do when mke2fs. The overhead can be
calculated by kernel side during mount.
Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Don't call sed multiple times on the output, and avoid the use
of temporary files, or if possible. It would be convenient to
use "sed -i" to only update the output file once, but this is
not portable to all platforms.
[ Fixed a few test regression failures --tytso ]
Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Now that the majority of device name filtering is in filter.sed,
it does not need to be specified explicitly for every test.
Fix the error message printed in debugfs when opening the device
to match that used in other tools. This simplifies the filtering,
and will be helpful if debugfs messages are internationalized.
[ Fixed up some test failures in m_hugefile t_uninit_bg_rm --tytso ]
Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Instead of putting the entire test script under an implicit
"if test -x $DEBUGFS_EXE" conditional (sometimes indenting
the code, and sometimes not), rather check for the reverse
and exit the test script early if $DEBUGFS_EXE is missing.
In some places, tests were running $DEBUGFS_EXE directly,
when they should be running $DEBUGFS (which will run with
Fortify, or other options).
[ Fixed up missing exit statement in f_detect_junk. --tytso ]
Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently tune2fs just says without any explanation, "run fsck -f".
Add a short explanation that a freshly checked file system is required
to reduce user confusion. (We could add even more details, but
hopefully this is enough.)
Addresses-Debian-Bug: #857336
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The message requesting that the user run e2fsck doesn't include the -f
option, and this is needed to force a file system check.
Addresses-Debian-Bug: #828022
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Skip tests that create large filesystem on MacOS, since HFS doesn't
have sparse file support and this causes testing to be uninterruptible
for minutes while mke2fs is writing to some large non-zero offset and
filling up the test filesystem. Since most testing is done on Linux
this shouldn't cause a significant gap in testing coverage.
Tests skipped are d_dumpe2fs_group_only, m_bigjournal, m_hugefile,
t_iexpand_full, t_iexpand_mcsum, and t_uninit_bg_rm.
Signed-off-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When we're turning on metadata checksumming /and/ resizing the inode
at the same time, disable checksum verification during the
resize_inode() call because the subroutines it calls will try to
verify the checksums (which have not yet been set), causing the
operation to fail unnecessarily.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
