ext4: ignore journal checksum on remount; don't fail
authorEric Sandeen <sandeen@redhat.com>
Fri, 13 Feb 2015 04:07:37 +0000 (23:07 -0500)
committerTheodore Ts'o <tytso@mit.edu>
Fri, 13 Feb 2015 04:07:37 +0000 (23:07 -0500)
As of v3.18, ext4 started rejecting a remount which changes the
journal_checksum option.

Prior to that, it was simply ignored; the problem here is that
if someone has this in their fstab for the root fs, now the box
fails to boot properly, because remount of root with the new options
will fail, and the box proceeds with a readonly root.

I think it is a little nicer behavior to accept the option, but
warn that it's being ignored, rather than failing the mount,
but that might be a subjective matter...

Reported-by: Cónräd <conradsand.arma@gmail.com>
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
fs/ext4/super.c

index 2ecce86..bff3427 100644 (file)
@@ -4858,9 +4858,8 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
        if ((old_opts.s_mount_opt & EXT4_MOUNT_JOURNAL_CHECKSUM) ^
            test_opt(sb, JOURNAL_CHECKSUM)) {
                ext4_msg(sb, KERN_ERR, "changing journal_checksum "
-                        "during remount not supported");
-               err = -EINVAL;
-               goto restore_opts;
+                        "during remount not supported; ignoring");
+               sbi->s_mount_opt ^= EXT4_MOUNT_JOURNAL_CHECKSUM;
        }
 
        if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) {