diff --git a/src/basic/mountpoint-util.c b/src/basic/mountpoint-util.c
index bf67f7e01a..97e0615047 100644
--- a/src/basic/mountpoint-util.c
+++ b/src/basic/mountpoint-util.c
@@ -605,23 +605,9 @@ int mount_nofollow(
                 const char *filesystemtype,
                 unsigned long mountflags,
                 const void *data) {
+        // Disabled due to ChromiumOS kernel already protecting against mount paths with symlinks.
 
-        _cleanup_close_ int fd = -EBADF;
-
-        /* In almost all cases we want to manipulate the mount table without following symlinks, hence
-         * mount_nofollow() is usually the way to go. The only exceptions are environments where /proc/ is
-         * not available yet, since we need /proc/self/fd/ for this logic to work. i.e. during the early
-         * initialization of namespacing/container stuff where /proc is not yet mounted (and maybe even the
-         * fs to mount) we can only use traditional mount() directly.
-         *
-         * Note that this disables following only for the final component of the target, i.e symlinks within
-         * the path of the target are honoured, as are symlinks in the source path everywhere. */
-
-        fd = open(target, O_PATH|O_CLOEXEC|O_NOFOLLOW);
-        if (fd < 0)
-                return -errno;
-
-        return mount_fd(source, fd, filesystemtype, mountflags, data);
+        return RET_NERRNO(mount(source, target, filesystemtype, mountflags, data));
 }
 
 const char *mount_propagation_flag_to_string(unsigned long flags) {