Description of problem:

A customer running Oracle ASM sees tons of the following messages:
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
systemd[SOMEPID]: Failed to stat() device node file /dev/asm/XXX: Permission denied
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------

These messages are due to "systemd --user" instances not being able to stat the device node, because /dev/asm directory has not the read/exec permissions for "others":
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------

1. ls -ld /dev/asm
   drwxrwx--- 2 0 oraadmin 980 Sep 3 14:46 asm
   • • • • • • • 8< ---------------- 8< ---------------- 8< ---------------- 8< --------

When such device (e.g. /dev/asm/XXX) is a mount point, the "systemd --user" instances complain in the following piece of code:
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
994 static int validate_node(Manager *m, const char *node, struct udev_device **ret) {
:
1010 if (stat(node, &st) < 0) {
1011 if (errno != ENOENT)
1012 return log_error_errno(errno, "Failed to stat() device node file %s: %m", node);
:
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------

The message is printed every time a mount changes in /proc/self/mountinfo, which is not rare at all.

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

systemd-239 and Upstream

How reproducible:

Always

Steps to Reproduce:
1. Change permissions on /dev/mapper

1. chmod 750 /dev/mapper

2. Login as a user

Actual results:

systemd[26327]: Failed to stat() device node file /dev/mapper/rhel-root: Permission denied

Expected results:

No error if issue is "cosmetic" for the "systemd --user" instance.