Description of problem:

We have a customer mounting SMB3 share hosted on a NetApp appliance (no details on appliance yet).
In order to mount the share, he has to use the "iocharset=iso8859-1" option.
The share contains non-ascii character "U Umlaut" (ü) in its name, which leads to having /proc/mounts displays the mount badly (see "umlaut_xxx�xxx"):
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
//netapp.server/path/to/share/containing/umlaut_xxx�xxx /mount/point cifs rw,relatime,vers=3.0,xxx,iocharset=iso8859-1,xxx 0 0
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------

This is an issue for systemd, which reads the information and sends it through DBus directly, causing non-UTF8 messages to be sent and later DBus to complain/fail.

I understand that something has to be done on systemd side to avoid breakage, but I think this should be handled at /proc/mounts level already.

Version-Release number of selected component (if applicable):

All kernel including kernel-4.18.0-425.3.1.el8.x86_64.
Was also seen on RHEL7 kernel.

How reproducible:

Always on customer system with NetApp appliance.
Cannot reproduce internally (tried Windows & Linux smb server)