Re: [hatari-devel] EmuTOS boot freeze with Hatari git on 040/060 + MMU

[ Thread Index | Date Index | More lists.tuxfamily.org/hatari-devel Archives ]


Le 13/03/2024 à 12:09, Eero Tamminen a écrit :
Hi,

(I cross-posted both to Hatari and EmuTOS lists, remove the other one depending on which one your answer concerns.)


With the Git Hatari version, EmuTOS freezes early in the boot (before boot info screen), when >030 CPU with MMU is enabled:
--machine tt/falcon --cpulevel 4/6 --mmu on

When it's frozen, Hatari "--log-level debug" output shows constant flow of these:
------------------------------------
DEBUG: SCC: get status for CTS 1
DEBUG: SCC: get status for DCD 1
DEBUG: SCC: get status for CTS 0
DEBUG: SCC: get status for DCD 0
------------------------------------

And EmuTOS CPU symbol tracing shows only VBL interrupt running after the freeze:
------------------------------------
...
_jdisint
_rsconf1
_setup_timer
_dsp_init
_memcpy
_dsp_flushsubroutines
_bzero_nobuiltin
_dsp_execboot
_int_vbl
8 repeats of: _int_vbl
16 repeats of: _int_vbl
...
------------------------------------


There's no freeze with June version of Hatari Git, so it could be something related to SCC support added after that, and SCC handling either in Hatari or EmuTOS.

This happens both with EmuTOS 1.2.1 and git version, so it's not a regression on EmuTOS side, only potential latent bug in it.

Hi

I don't see anything related to the cpu model/mmu in recent scc changes, as it's cpu agnostic and works on 68000

maybe it's in Emutos handling routine, I can't tell (or a bug in MMU for 68040/60 core ?)

would need more info / logs on the emutos side to see what condition can cause this loop and see how it could be related to MMU

Nicolas

PS : I'm not subscribed to emutos ML, so this mail from me will not reach them.




Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/