Looks great already, thanks a lot! The serial printing seems to work already on the CI, for example we see the bootloader output here: https://github.com/rust-osdev/bootloader/actions/runs/3846887378/jobs/6552698281#step:11:223

Just for information: I just restarted the Windows CI job manually in the hope of triggering the timeout issue e.g. seen in #304. I suspect that the cause of that timeout could be a panic in the bootloader, which this PR should make visible.

Makes sense. I'll also block config.txt on this as well, once the kernel config is in place.
It may yet again be wise to wait on #302 to avoid conflicts before doing the kernel config. So, perhaps do what is currently in this PR separately, as it'll be a huge help.

I can't apply the changes proposed by @jasoncouture in the linked issue because it's very difficult to refactor BIOS stages due incompatibility problems of the used crates of the logger. So, I think that this PR is ready to merge.

Thanks!