If this flag is passed to open_ctree(), we'll clear the
FREE_SPACE_TREE_VALID compat_ro bit. The kernel will then reconstruct
the free space tree the next time the filesystem is mounted.
Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Run more tests in the travis build environment. The mkfs tests are still
left out due to missing capabilities of losetup.
Other minor travis adjustments.
Signed-off-by: David Sterba <dsterba@suse.com>
Add new keyword to dump the log file after any test fails. Can be useful
for remote analysis of test failures.
Signed-off-by: David Sterba <dsterba@suse.com>
The option --list might not be available on older versions, the
equvalent is --all. Discovered via failed travis build.
Signed-off-by: David Sterba <dsterba@suse.com>
The current user might not be able to peek into the loop files directly,
use the helper. Discovered by running tests in travis.
Signed-off-by: David Sterba <dsterba@suse.com>
The read_buf does not verify that we've read the expected number of
bytes. A corrupted of malformated stream will not be detdcted.
Signed-off-by: David Sterba <dsterba@suse.com>
Fixing the problems by one does not scale now. Add more images despite
the fuzz tests will fail. They have been for some time already.
Reported-by: Lukas Lueg <lukas.lueg@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Definition of NULL is missing if ioctl.h is included standalone or when
no other include pulls the definition.
Reported-by: Sergei Trofimovich <slyich@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The find_most_right_entry() tends to miss on the best entry if it is the
first one on the list and there are only two entries in the list. So,
we assign both prev and best to entry.
To do this, the selection process (rather the rejection) has to be
performed earlier to skip on broken==count.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In get_running_kernel_version() function, we directly pass return
pointer from strtok_r() to string_is_numerical().
Return pointer from strok_r() can be NULL, but string_is_numerical()
can't handle it and will cause NULL pointer derefernces.
Fix it by check if it's a NULL pointer first.
Resolves-Coverity-CID: 1374097
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Separate checksum test from convert_test_post_check and use it to fix
the broken test 005 as reported.
References: https://bugzilla.kernel.org/show_bug.cgi?id=177141
Signed-off-by: David Sterba <dsterba@suse.com>
Enhance the mkfs_features list with the minimum kernel versions that
will allow for turning on compatible and/or safe options.
Signed-off-by: David Sterba <dsterba@suse.com>
Ebs and pointers are allocated, but if any of the allocation failed, we
should free the allocated memory.
Resolves-Coverity-CID: 1374101
Resolves-Coverity-CID: 1374100
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Sending stream size of clone-src(-c) option is checked.
Fixed by "btrfs-progs: send: fix handling of -c option".
Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In some cases the root might not be able to create the output file (and
streaming to stdout is not an option). Make the output file creation two
step and let it work if the file is already created.
Signed-off-by: David Sterba <dsterba@suse.com>
Tsutomu Itoh
Omar Sandoval
David Sterba
Goldwyn Rodrigues
Qu Wenruo