The previous commit exposed bugs in the calculation for flex_bg file
systems. The problem is that since (by default) we keep the metadata
blocks for the flex_bg in the first block group of the flex_bg, and
because we don't want to overwrite metadata blocks used by the
original file system with data blocks make life easier in case the
resize is aborted for some reason, we need to treat all of the
metadata blocks in the existing flex_bg has in use for the purposes of
calculate_minimum_resize_size().
Even though this means we need to reserve more data blocks to avoid
running out of space, the net result of these two commits is a net
savings in how much we can shrink a file system.
Using the following test sequence:
mke2fs -F -t ext4 /tmp/foo.img 2T
resize2fs -M /tmp/foo.img
resize2fs -M /tmp/foo.img
resize2fs -M /tmp/foo.img
Here is the comparison in the resulting file systems between the old
and new resize2fs (units are in 4k blocks):
resize #1 resize #2 resize #3
old resize2fs 1117186 45679 43536
new resize2fs 48784 37413 37392
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
These images contain various forms of corrupted filesystem which
e2fsck will correct. They are used as a regression test for e2fsck.
The test_script program will automatically run e2fsck against the
filesystem images. It will run them two times, and display the exit
status for each run. The meaning of the exit status codes are as
follows:
0 No filesystem errors were detected
1 Filesystem errors detected, but corrected
2 System should be rebooted
4 Filesystem errors left uncorrected
8 Operational error (generally means internal error,
or filesystem error that the e2fsck was not
prepared to deal with)
16 Usage or syntax error
During the regression test, the first exit code should be 1, and the
second exit code should be 0. In other words, all (with one
exception) of the test filesystems in this directory have some sort of
filesystem corruption, which e2fsck should fix on the first pass.
After the first pass, e2fsck should leave a fully consistent
filesystem with no detectable errors found in the second pass. The
exception is the okgroup.img filesystem, which contains no errors, and
so both exit codes should be 0.
NOTE: It appears that at least some versions of the original e2fsck do
not exit with an exit status code of 1 after correcting filesystem
errors. So if you modify the test_script to try running these
filesystems against the original e2fsck, you will have to inspect the
test_script.log file manually.
--------------------------------------------------------------
Here's a one-line descriptons of the various test images in this
directory:
baddir.img Filesystem with a corrupted directory
badbblocks.img Filesystem with illegal blocks in the bad block inode.
badinode.img Filesystem with various different corrupted inode
entries.
badlkcnt.img Filesystem with deleted files with non-zero link count
badroot.img Filesystem with a file for a root directory
badtable.img Filesystem with blocks shared between the bitmaps and
inode table blocks and the bad block inode
bbfile.img Filesystem with files containing bad blocks
bitmaps.img Filesystem with corrupted inode and block bitmaps
dirlink.img Filesystem with a hard link to a directory
dup.img Filesystem with blocks claimed by two different files
dup2.img Filesystem with blocks claimed by three different files
dupfsblks.img Filesystem with blocks claimed by a file and
inode/block bitmaps and inode tables
dupsuper.img Filesystem with blocks claimed by a file and
the superblock / group descriptors
end-bitmap.img Filesystem with corruption at the end of the block
bitmap
expand.img Tests e2fsck's ability to expand lost+found if
necessary
lpf.img Filesystem with disconnected files and no /lost+found
directory
mke2fs2b.img Filesystem with corruption similar to that
created by mke2fs version 0.2b
noroot.img Filesystem with a deleted root directory
okgroup.img Filesystem that's exactly 8193 blocks long
(otherwise OK)
overfsblks.img Filesystem with overlapping inode and block bitmaps
symlinks.img Filesystem with bad symlink sizes