Btrfs: fix clone vs chattr NODATASUM race
authorOmar Sandoval <osandov@fb.com>
Tue, 22 May 2018 22:02:12 +0000 (15:02 -0700)
committerDavid Sterba <dsterba@suse.com>
Tue, 29 May 2018 16:12:59 +0000 (18:12 +0200)
commitb5c40d598f5408bd0ca22dfffa82f03cd9433f23
tree0d5ddcb124f17f0f2b52f880a005429d647aee86
parentb89311efe60d47f8fc23b6f027c6b3f30067c93e
Btrfs: fix clone vs chattr NODATASUM race

In btrfs_clone_files(), we must check the NODATASUM flag while the
inodes are locked. Otherwise, it's possible that btrfs_ioctl_setflags()
will change the flags after we check and we can end up with a party
checksummed file.

The race window is only a few instructions in size, between the if and
the locks which is:

3834         if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode))
3835                 return -EISDIR;

where the setflags must be run and toggle the NODATASUM flag (provided
the file size is 0).  The clone will block on the inode lock, segflags
takes the inode lock, changes flags, releases log and clone continues.

Not impossible but still needs a lot of bad luck to hit unintentionally.

Fixes: 0e7b824c4ef9 ("Btrfs: don't make a file partly checksummed through file clone")
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ update changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/ioctl.c