RPF (Reverse Path Forwarding) checks are used to prevent multicast loops. An RPF check uses theunicast routing table(ormulticast static routes, if configured) to determine the correct incoming interface for a multicast packet.
Multicast protocols like PIM use the RPF mechanism, and the unicast routing table is the primary reference.
Multicast dynamic routing protocolsdo not maintain their own separate tablesfor RPF by default.
Reference from Huawei HCIP-Datacom-Core Technology Study Guide:
"RPF checks are performed using the unicast routing table or multicast static routes. The interface through which the multicast source is reachable is considered valid for RPF." (Chapter: Multicast - Section: RPF Principle)

Let's break it down step by step based on Huawei's routing protocol behavior:
R1imports thedirect route 10.1.1.0/24intoOSPF.
This route is advertised byR1 to R2 and R3viaOSPF.
BothR2 and R3are runningIS-IS, andbidirectional route redistributionis configured between OSPF and IS-IS.
R2 and R3 bothre-advertisethe OSPF-learned route into IS-IS and then potentiallyre-learnit back from the IS- IS domain, causingredistribution loopsunless properly controlled.
Now, the question focuses onR3's routing table:
R3 already learns the routedirectly via OSPFfrom R1.
It also learns the same route again viaIS-ISfrom R2 (looped back due to bidirectional redistribution).
Reference from Huawei HCIP-Datacom-Core Technology Study Guide:
"OSPF ASE routes are assigned a default preference of 150. If a route is redistributed from another protocol (like IS-IS) into OSPF, and then re-learned, it is considered an ASE route."(Chapter: OSPF Advanced Features - Section: Route Redistribution Behavior)

When OSPF routers exchange LSDBs using Database Description (DD) packets, they first establish a master- slave relationship. The router with the larger Router ID becomes the master, setting the MS (Master/Slave) bit in the DD packets. This hierarchical relationship ensures an orderly exchange of routing information .