Question about the project's scope

[wiktor-k]

Hi,

I've just learned about this project and I'm happy to see that.

I wonder if developing a software simulator which runs the TPM-rs firmware is also in scope? I've seen similar approach in the smartcard space where some applets can be run in software and I'm using that for integration tests, which allow me to be sure that my apps will work the same with real-world smartcards.

(Just for the record I'm aware of software simulator / tpm2-tools but still, I think running tpm-rs in software would make the user testing validation easier).

Thanks for your time! 👋

[trent-reed]

Hello!

Sorry for the huge delay here, I guess I don't have my email notifications set up properly and somehow missed this message.

Anyways, I would love to support a simulator case! I think there exists a world where we run the firmware implementation as a simulation. However, I think we're quite a ways away from that world currently (~years). My safe response will be that I think it's a nice idea, but that we're not currently committed to it, so I wouldn't take any hard dependencies on the idea for now.

We first we need to get the fundamental common code in place, and then we can probably branch off into client and server (read: firmware) parts, and then when we're dealing with the server part we could consider simulation a use-case for that part of the code. At least that's my ideal version of this.

That's my feelings on it at least, not sure how others feel.

I will be transparent though and say that the priority order of this is:

Fundamentals > Client > Server > Simulator

I'll try to come back to this when we can say more though (after the fundamentals are figured out).

[wiktor-k]

Sorry for the huge delay here, I guess I don't have my email notifications set up properly and somehow missed this message.

No worries, completely understandable 👍

I will be transparent though and say that the priority order of this is:
Fundamentals > Client > Server > Simulator
I'll try to come back to this when we can say more though (after the fundamentals are figured out).

My point was having a simulator helps implement some kind of end-to-end/integration tests where you'd deploy the final firmware and exercise it by a sample client but I understand this is an additional effort.

Thanks for keeping me in the loop and good luck! 👋

[bradlitterell]

Agree with Trent's comments that a simulator (much like the existing C TPM code simulator) is reasonable, but still future. Anticipating no further discussion, closing this discussion for now.