systemd complaining about left-over dnsmasq process when restarting NetworkManager.service

[Jens Schmidt]

I'm on Debian testing with NM 1.50.1-1, using dnsmasq as DNS
provider:

  [main]
  dns=dnsmasq

When I restart NM with

  systemctl restart NetworkManager.service

NM obviously does not kill its child dnsmasq process, resulting
in the following messages ("W": warning, rest info priority):

  [...]
  Jan 17 23:01:50 host01 NetworkManager[3311]: <info>  [1737151310.2544] device (p2p-dev-wlp170s0): state change: disconnected -> unmanaged (reason 'unmanaged-link-not-init', managed-type: 'removed')
  Jan 17 23:01:50 host01 NetworkManager[3311]: <info>  [1737151310.2642] exiting (success)
  Jan 17 23:01:50 host01 systemd[1]: NetworkManager.service: Deactivated successfully.
  Jan 17 23:01:50 host01 systemd[1]: NetworkManager.service: Unit process 3541 (dnsmasq) remains running after unit stopped.
  Jan 17 23:01:50 host01 systemd[1]: Stopped NetworkManager.service - Network Manager.
W Jan 17 23:01:50 host01 systemd[1]: NetworkManager.service: Found left-over process 3541 (dnsmasq) in control group while starting unit. Ignoring.
W Jan 17 23:01:50 host01 systemd[1]: NetworkManager.service: This usually indicates unclean termination of a previous run, or service implementation deficiencies.
  Jan 17 23:01:50 host01 systemd[1]: Starting NetworkManager.service - Network Manager...
  [...]

Is it intentional that NM does not kill the dnsmasq process when
it exits orderly?  If yes, why?

Thanks!