hey @zapper
generally I can recommend reading some of the docs for heads: https://osresearch.net/ as explaining all details of how heads works, which threat model it addresses would be a little out of scope here
But anyways, some words for your questions:
A: Would this function at all without any major hiccups? I get it would not be supported for the future as something you would help with
The problem/challenge is not the fact that there are multiple hard-drives within a machine. The issue arises once you would like to have multiple, bootable operating systems (OS) as HEADS will sign all files within /boot
for a specific OS - so it can promise you that only an untouched OS (the one you signed) will be booted.
So if you have multiple drives within a machine running HEADS there will be no problem, you just have to ensure that you boot just one OS.
Yes, of course it is possible to have two harddisks within one system, it’s just not possible to boot two different OS.
the difference is that HEADS gives you (among others) a guarantee that nobody touched/modified your files in /boot
for the set up main OS. If you boot from a USB-stick you still get the promise that your firmware hasn’t been modified, but if something or somebody changed your usb-stick’s contents (e.g., the kernel) heads will not recognize this.
This simply means, that if there is not a ton of users, which justify this development invest here, then likely never. Another option would be (as it is open source) that the development would be made within the HEADS community, but I cannot see this there at this moment. Despite the fact that your idea doesn’t really match with the overall HEADS architecture, goals and threat model.
Generally your hypothetical idea, without a fully thought through concept is really tough to compare with other already existing approaches - and “how much” implies you would like to get a metric or number from me - I have to admit: this is nothing I can do - simply because there are no absolute nor relative metrics for security. At least none I can give you in the context of a forum thread without a thorough threat analysis - hope you understand. Generally, open-source firmware is good and better as a proprietary blob, by how much is a very complex question to answer.