This review of -12 updates my previous review of -11

I find that the ongoing discussion of a "reflection attack" echoes the point that I raised with the -11 Security Considerations section. On the whole, I think that calling out the attack vector and noting that 5800 already observes the risk of spoofed echo requests. It's notable that in "classic" BFD it is the responder that is attacked, but in this work, it is the "sender" that is attacked.

Thanks to the authors for resolving my previous review.