Re: [hatari-devel] Hatari startup freeze regression with 030 MMU + EmuTOS + non-VDI screen

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 31.5.2022 11.17, Nicolas Pomarède wrote:
> this is now fixed ; I tested Falcon or TT in CE mode with or without MMU 
> and everything runs correctly now.
>
> There was an error when updating cycles during the STOP state which 
> altered the internal timers in cycInt ; unlike WinUAE Hatari has 2 
> different cases to count cycles depending on whether the cpu runs in 
> cycle exact mode or not and the check for MMU was not correct there 
> (eventually this 2 ways of counting cycles (dating from old UAE core) 
> will go away one day to have more common logic with WinUAE)

Verified.

I ran all the previous TOS-boot test combinations plus some extra, and 
everything seems fine one.

Of these thousands of combos (running through the night), these failed 
on first run:
* tos102us-st-vdi1-floppy-14M-0TTRAM-cpuexactoff
* tos205de-megaste-mono-floppy-14M-0TTRAM-cpuexactoff

But I was not able to reproduce the fails today, so I guess we can 
ignore them.


	- Eero