When we are interrupting threads for stepdown, we only check if the context has ever acquired the global lock in a write conflicting mode, we do not directly check if a thread is holding the RSTL.

When we are acquiring the global lock we first acquire the RSTL in MODE_IX, then attempt to acquire a write ticket, then finally acquire the global lock

If a system is starved of write tickets that means we can have a thread blocked indefinitely holding the RSTL, which can crash stepdown. Is this dependency easily resolvable?

Note that this is not a concern for 9.0+ releases since the Intent Registry system handles this case already, but for RSTL based timeouts (8.3 and below) this can occur