btrfs: use proper endianness accessors for super_copy
authorAnand Jain <anand.jain@oracle.com>
Thu, 22 Feb 2018 13:58:42 +0000 (21:58 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 9 Mar 2018 06:41:05 +0000 (22:41 -0800)
commiteae6179f55396f7ee708c7da1f9d84688412b62e
treeae69733c59c1365092f1ede06f296f7f249fe5e0
parentdabf89052e8e1bea5f4cba2f4533e8718acdbdc9
btrfs: use proper endianness accessors for super_copy

commit 3c181c12c431fe33b669410d663beb9cceefcd1b upstream.

The fs_info::super_copy is a byte copy of the on-disk structure and all
members must use the accessor macros/functions to obtain the right
value.  This was missing in update_super_roots and in sysfs readers.

Moving between opposite endianness hosts will report bogus numbers in
sysfs, and mount may fail as the root will not be restored correctly. If
the filesystem is always used on a same endian host, this will not be a
problem.

Fix this by using the btrfs_set_super...() functions to set
fs_info::super_copy values, and for the sysfs, use the cached
fs_info::nodesize/sectorsize values.

CC: stable@vger.kernel.org
Fixes: df93589a17378 ("btrfs: export more from FS_INFO to sysfs")
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ update changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/btrfs/sysfs.c
fs/btrfs/transaction.c