build: portability fixes ...

- Invoke Python scripts with `python3` if `python` is not available.

  Despite SciresM preferring Python 2, the scripts used in the build all
  worked in Python 3 already, so this is just a Makefile change.

- Export `MAKE` to work around a GNU bug.

  See the comment for details.

- `rmdir --ignore-fail-on-non-empty` -> `rmdir 2>/dev/null`

  macOS `rmdir` doesn't support `--ignore-fail-on-non-empty`.

  This is a slight downgrade in functionality, because now we ignore all
  errors from `rmdir` rather than just 'Directory not empty'.  I could
  have avoided this with a more complicated fix, but the benefit strikes
  me as not worth the complexity.  Let me know if you disagree.

- Append `$(DEVKITPRO)/tools/bin/` to `$PATH` when invoking
  `build_romfs`.

  This avoids the need to add devkitPro directories to `$PATH` when
  building Atmosphere.

  If you already have `build_romfs` in your PATH, then that will still
  take precedence.  (This seemed like the nicer option, though I don't
  have strong opinions.)

kern: move spendsgir static assert

kern: save/restore spendsgir in KInterruptController::Save/RestoreCoreLocal

kern: wait 100us after synchronizing cores before saving interrupt state on sleep

kern: use KLightLock for KCoreBarrierInterruptHandler

fs.mitm: add and use memlet module to temporarily steal applet memory while building romfs images. ...

Starting in 20.0.0, the browser needs more applet memory to function, so we can't steal as much any more.
Thus, we now steal 14 MB on 20.0.0+ instead of 40MB.

However, since this reduces memory available for custom system modules, we are adjusting to compensate.
ams.mitm's heap size has been reduced from 32MB to 12MB (recovering 20MB).
In addition, fs.mitm now uses a new mechanism for stealing memory from the applet pool while romfs is being built.

On net, we are compromising:
    * Custom sysmodules lose memory available to them.
        On 19.0.0/AMS 1.8.0, there was 30 MB available for custom sysmodules.
        Stealing 14 MB instead of 40 MB, we lose 26 MB of that. Reducing ams.mitm's usage will gain us back 20.
        Nintendo also appears to...use 4 extra MB, in 20.0.0, from my test homebrew.
        So on 20.0.0/AMS 1.9.0, there should be 20 MB available for custom sysmodules.
        On the bright side, on <20.0.0/AMS 1.9.0, I guess there will be 50 MB available for custom sysmodules now?
    * totk mods will lose the ability to...put every file in the romfs on sd card. There will be some unknown maximum filecount for totk mods.
        On the bright side, implementing the transient memory stealing should improve compatibility for some mods which strictly add files?

kern: plutoo is an intellectual, I am a fool, fix chicanery

kern: use callback to note pte updates in KPageTableImpl

kern: refactor FindFreeArea region search logic per 20.0.0 changes

kern: update instruction cache invalidation logic in KPageTableBase to reflect 20.0.0 changes

kern: reorder resource manager implementation

kern: support null resource limit in KSecureSystemResource

kern: adjust system registers during exception handling on MTE-violation or kernel address fault

kern: update UserspaceAccess functions for 20.0.0 changes

kern: update KAddressSpaceInfo to reflect 20.0.0 changes

kern: invert meaning of KTargetSystem/KSystemControl bools

ams: various fixes to support booting debug configuration

kern: update KernelLdr for 19.0.0 (new checks, dummy function call). ...

Also, fix a few very embarassing mistakes in kernel ldr:
* We have been mapping the page table region RWX for a few years now, accidentally.
* My attempt at making initial page tables not use bit 58 was broken in multiple ways.

kern: support reboot to fatal error on mariko

kern: Perform page table validity pass during KPageTableImpl::InitializeForKernel

kern: fix more page table refactor bugs

kern: first round of page table refactor bug fixes

kern: update KPageTable::Map for new refactor

kern: update ChangePermissions to use new iteration logic

kern: update KPageTable::Finalize for the refactor

kern: use new merge pages api

kern: implement KPageTableImpl merge

kern: continue page table refactor, implement separate/unmap

kern: start KPageTable(Impl) refactor, use array-with-levels for KPageTableImpl

kern: invoke supervisor mode thread functions from C++ context with valid stack frame