Hi Dhaval,

In general, I didn't not want to add many extra code/design  as part of 
this series to keep it simple  refactor and port from edk2-platforms 
patches. We are the first one to do this type of refactoring work and 
hence not easy to convince people. So, I am still keeping many things 
hardcoded as in edk2-platforms. Once this series gets merged, next set 
of patches will be mostly within RISC-V code which will be much easier 
to upstream.

Irrespective of SATP (>= sv39), the physical address space is 56 bits as 
per the priv spec. We don't plan to support RV32 with edk2 anyway.

Thanks

Sunil

On 30/01/23 10:47, dhaval@rivosinc.com wrote:
> BuildCpuHob (56, 32); Can we find a better programmatic way of doing 
> this? SATP register could be on method, or PCD? Also would help if you 
> can clarify how we arrive at this value 56.