Oh, strange. I suppose there isn’t a way to prevent it from being advertised on neighbor A if it’s a bug, then?
Does it have any actively harmful effects? Loopbacks obviously shouldn’t be in a routing table but I’m trying to figure out what would actually break if you did. For routers that even implement IPv4 loopback addresses automatically, nothing should believe an OSPF route over a connected route to itself … you would think?
As said - after reboot of the misbehaving router that destination cannot be found among other routers' imported routes, so it's a bug and no, the misbehaving router is a minor part of an appliance and has no route filters configurable by user.
As for actively harmful effects - yes, nothing should beat a connected route, but knowing how SW development often goes, it is hard to predict what it might actually cause at some recipient.
What may or may not be related - I have a router in the same area that now and then stopped forwarding packets from a particular source address to
some destinations within a single remote subnet imported to its routing table via OSPF. To make it forward them again, I had to disable that subnet at its home router (so it disappeared from the routing table of the affected router) and then re-enable it. Then it worked for a while, and then it broke down again. To my cautious relief, it has not happened since the route to 127.0.0.1 stopped being advertised. But it's too early to conclude it was related. The home router of the affected subnet is not the same device like the one advertising the 127.0.0.1, it's its neighbor.