I mount a directory from cephfs to local filesystem.

mount -t ceph -o rw,relatime,seclabel,name=admin,secret=<hidden>,acl,mds_namespace=myfs,wsize=16777216 10.99.226.165:6790,10.99.238.119:6790,10.99.227.59:6790:/users/arch-admin /mnt/data-storage

And I can change directory to /mnt/data-storage, as follow:

[root@33e9b5337f0f data-storage]# cd /mnt/data-storage
[root@33e9b5337f0f data-storage]# ll -alh
drwxr-xr-x.  5 root root  13G Jun 21 12:50 .
drwxr-xr-x.  1 root root 4.0K Jun 21 12:47 ..
drwxr-xr-x.  2 root root  13G Jun 19 08:42 test-data

All is well before a long interrupted time between ceph client and ceph mon, mds, osd.
Since the failed network, I try to resume it in ceph client, mon, mds and osd, the mount point of /mnt/data-storage has been failed, and I can not cd to there, as follow:

[root@33e9b5337f0f data-storage]# ll -alh
drwxr-xr-x.  5 root root 4.0K Jun 21 14:43 .
drwxr-xr-x.  1 root root 4.0K Jun 21 12:08 ..
d?????????? ? ?    ?       ?            ? test-data

There exist a trouble shooting in http://docs.ceph.com/docs/master/cephfs/troubleshooting/#disconnected-remounted-fs, as you see the it ascribes mds reject a old session to mount point losted.
I want to remount the directory, like this:

mount -t ceph -o remount,rw

But nothings happen after remount, remount interface is not implemented by libceph, quote from libceph source code of linux kernel:

static int ceph_remount(struct super_block *sb, int *flags, char *data)
{
    sync_filesystem(sb);
    return 0;
}

But the user mode's ceph-fuse can be automatically resume permission denied point.

MY QUESTION IS:
why the kernel mode's libceph not unabridged implement the function of remount? Just like ceph-fuse, flushing all dirty data after the collapsed network.¶