After an IPC channel count we realised we had two spare channels which we could use for the IOP and user model watchdog transmission from SUS ITMY and ISI ITMY. This would allow unattended operation of BSC1 with an active ISI.
IOP: we modifed the iop models (h1iopsusb123 and h1iopseib1) to reinstall the Dolphin IPC (SUS sender, ISI receiver). We modified h1boot to only run the iop models, and shutdown all running models on h1susb123 and h1seib1. We then started the IOP models and verified that the output of the SUS was being received by the SEI. We manually panic'ed the SUS DACKILL and verified the SEI DACKILL was activated (all DACS were disabled).
User Models: Mark and Vincent made the appropriate user model changes to h1susitmy and h1isiitmy. These models were then restarted. Watchdog testing for these systems was handed over to Mark and Vincent.
We verified the safe.snap burt restore of both h1susitmy and h1isiitmy.
Ooops, by reverting to an earlier h1iopsusb123 model we undid my change covered by the original WP 3582 which took the Beam Splitter out of the trip logic. So after about 90 seconds the DACKILL was getting triggered due to the missing BS OSEM signals.
We shutdown h1susitmy, rebuilt and reinstalled h1iopsusb123 sans BS trigger, tested by waiting a couple of minutes with no trip, then restarted h1susitmy (manually safe.snapped to be extra safe).
We think now we have finished this change. We kept the SEI DACKILL triggers (DACs killed) because team ISI were shaking the stack manually out on BSC1.