Current status of new m68k pmap

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

Hi,

With the following fix for zero-ing bug, the new pmap kernels work on
several real machines:
 https://mail-index.netbsd.org/source-changes/2025/11/13/msg158934.html

luna68k:
 - Ready to switch to new pmap:
  https://mail-index.netbsd.org/source-changes/2025/11/12/msg158910.html
  https://mail-index.netbsd.org/source-changes/2025/11/12/msg158923.html
 - GENERIC boots on LUNA (68030) and reaches multi-user (yay),
   but occasionally some userland processes die.
 - GENERIC boots on LUNA-II (68040) but the MPU reports a fault
   (via its LED indicator) just before init(8) starts.
 - GENERIC boots on nono (emulating LUNA) but init(8) gets a bus error
   (possibly a nono issue?)

news68k:
 - Ready to switch to new pmap:
  https://mail-index.netbsd.org/source-changes/2025/11/12/msg158927.html
  https://mail-index.netbsd.org/source-changes/2025/11/13/msg158935.html
 - GENERIC boots on NWS-1750 (68030) and almost reaches multi-user,
   but userland programs stop working around "Updating motd."
   (L2 cache issue?)
 - GENERIC boots on nono (partial NWS-1750 emulation) but init(8) also
   gets a bus error.

Note that both luna68k and news68k can use TT0/TT1 registers to map I/O
spaces so static mappings in pmap_bootstrap is not required.
This makes them relatively easy to switch to the new common pmap,
while other ports may require more work.

virt68k:
 - GENERIC boots on qemu-system-m68k (68040) and reaches multi-user.
 - GENERIC boots on nono (68030), but init(8) also gets a bus error

Thanks,
---
Izumi Tsutsui