Does using the AccessTimingParameters messages to extend the P2 or P3 timeouts help? I would assume for this to work, the instrument cluster would have to eavesdrop on these messages to the ECU.
Hm, from my experience with this, the cluster doesn't evaluate the contents of the messages passing by.
It doesn't interpret the ATP message in the communication with the ECU, just scans for the slowinit pattern to 0x01
and then is happy and shuts up until communication stops/next slowinit comes in.