Re: [hatari-devel] Linux boot freeze with Hatari git on 040/060 + MMU + CPU prefetch/cache

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 16.3.2024 17.03, Nicolas Pomarède wrote:
> Le 15/03/2024 à 23:57, Eero Tamminen a écrit :
> > On 15.3.2024 12.38, Nicolas Pomarède wrote:
....
> > > does it work if cpu cache is not enabled but prefetch or cycle exact 
> > > are enabled ?  ie, is it cpu cache taht cause the problem ?
> >
> > There's panic in same place regardless of whether "--compatible on" or 
> > "--cycle-exact on" is used.
> >
> > How do I enable cycle-exact, but not cache emulation, I thought 
> > "--cpu-exact" option controls both?
>
> my bad, there's no option for that :(

I'll add separate option for that.


> but in src/m68000.c you can add at line 396
>
> changed_prefs.cpu_data_cache = false
>
> to force data cache off in all cases ; then if you can test again, this 
> could tell if problem is in cache or not (if it doesn't works with cache 
> off, it doesn't mean there's no problem in cache, but at least it 
> narrows the parts to look at first)

CPU exact & compatible options have no effect on this bug, only whether 
CPU cache is enabled (panic/freeze, due to memory not being at expected 
address) or not (everything works, although there are lot of DEBUG log 
messages [1]).


	- Eero

[1] on user-space / kernel-space transition (= MMU mapping changes?) 
there are messages like this:
DEBUG: Your Atari program just did something terribly stupid: 
dummy_xlate($800998ce)