[FEATURE REQUEST] Distributed boot

[jleidel]

Rev Feature Request
Enable distributed boot of Rev procs. Proc[0] would execute the loader and quiesce the memory subsystem before triggering the remaining Procs (HARTs) to begin executing. This ensures:

1. we don't execute multiple copies of the loader
2. the instruction text, stacks, etc are all written before we begin executing
3. memory is in known good state before we begin executing

Version of SST
13.0.0; 13.1.0

Additional context
@vcave

[rkabrick]

Definitely worth confirming the functionality with @vcave because I think they wanted a distributed boot of RevCPUs. Only 1 Hart on 1 Proc begins execution and the Loader is only executed once. The problem is having multiple RevCPUs being able to 'share' the execution of a single binary instead of each executing the same binary, loading the same data, etc.