Based on systemd/systemd#12018. Systemd uses libmount to parse /proc/self/mountinfo.

$ mkdir "$(echo -e foo\\rbar)"
$ sudo mount -t tmpfs tmpfs foo^Mbar/
$ findmnt | grep foo
│ └─/tmp/foo                                tmpfs                          tmpfs       rw,bar,seclabel

Quoting systemd/systemd#12018 (comment):
fscanf() will split fields on whitespace, and what counts as whitespace is both locale-specific and different to the actual file format (even in the C locale). The Linux kernel, the GNU C library, and other utilities are quite specific that only SPC and TAB are field separator characters; whereas fscanf() will also incorrectly split fields on VT and CR.

• GNU C library using __strsep() with TAB and SPC and decoding field contents
• netdata specifying only TAB and SPC as the the separators in the call to procfile_open()
• Linux only encoding some whitespace characters in the first place
• source of fstab-decode

Whether or not CR is a good idea in a mount point name, one should parse the data file correctly.