It still doesn't compile. I'll need to work on it a little bit more
for it to compile.

* Mm: Add a platform function called MiFreePageMapping(HPAGEMAP).

* Mm: Add basic implementation of ARM page tables.
* Rtl: Add ARM ELF relocation types.

* Add libgcc for arm, thanks to https://github.com/osdev0/libgcc-binaries

…tup interrupts.

Now interrupt stacks are per-thread.  This isn't that great since
memory usage is increased but I honestly don't care at this point.

…ss a ctx switch.

* ARM progress, now it actually runs but does tend to break

…mode

* Ke(armv6): Remove leftover debug print at boot.

* Finish the init procedure on armv6

I know there won't be any framebuffer but at least
we'll be on a more stable platform, and also closer
to the platform I actually want to target in 2026.

I'll pick this up later. Right now, I want to work on something else.

However, I won't actually do anything with it yet.

Not sure why these show up in the first place, but we don't need them.

…opySub.

…andlers not restoring stack properly.

wtf was I cooking?

…operly.

…m calls.

I'm too lazy to fix the underlying issue...

I don't know why this fails in the first place.  Normally the thread
has already entered TERMINATED state by the time we unlock the dispatcher.
This bug seems to be similar to that one monkuous found a few weeks ago...

This is temporary.

This reverts commit a5a7444.

…witches.

… to arch-specific code.

This fixes a triple fault on amd64.

Flanterm already converts to native, so this will corrupt the colors.