Ml-PPP and QoS

For a ML-PPP configuration requiring a prioritized QoS model, looking for confirmation if “mangle pre-routing” and “global-in” is the recommended approach?

sure. why ML-PPP should be different? no reason =)