fma: FMA for MT/64-bit Cannon

[pauldowman]

This is the failure modes analysis for multi-threaded & 64-bit Cannon.

[pauldowman]

@Inphi @mbaxter I did rebase and force push this but I won't do that again so that we can share the branch if you want to make edits in here.

[Inphi]

@pauldowman could you fill in the "Initial reviewers" and "Need approval from" fields in the table.

[Ethnical]

@Inphi For the failures that require monitoring like op-dispute-mon for the detection.
I would like to know if a scenario that exhaust or DoS op-program can also DoS the op-dispute-mon.
I am looking for case where we can crash the CHALLENGER and also the monitoring, thus making the game resolving incorrectly but impossible to detect and thus allowing an malicious actor to exploit the bridge.
In other words is there is part of the op-program code shared with op-dispute-mon.
I am thinking about case like Insufficient memory in the program or the Unimplemented syscalls or opcodes needed by op-program.

[pauldowman]

Hooray, all resolved, and the vm-runner health check alert has been added. There's still the one open action item on here to add a dedicated Validation file for ensuring the absolute prestate is valid, but we agreed it's not blocking so I set the status to Final. Can we close it?

[Ethnical]

Thanks a lot for the work there!
Yes, sound good to me to switch to FINAL here.
I have approved ✅

I will ask also @mds1 how long can we stay with the status implementing Action with the current process.

[Ethnical]

Thanks for the work!
The status of the FMA is set to FINAL.