This bug affected us on 9.3(5) on pair of 93180YC-EX TOR switches in vPC. 9.3(5) is not presently on the versions affected list.
The interfaces stay up, so connected equipment that is watching link-state will not fail over to alternate paths. In vPC mode, the secondary switch continues to believe the primary switch experiencing this bug is online until the reload occurs. In our case it took 22 minutes from traffic flow stopping to the kernel panic/reload. This killed our HyperConverged server stack serviced by these switches for the entirety of that 22 minutes, plus a lot of recovery time afterwards. If you are running this in a mission critical application, treat as severe risk!!
