Re: Aw: Re: [hatari-devel] MegaSTE cache emulation issues

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 03/07/2025 à 00:07, Eero Tamminen a écrit :
> Hi,
>
> On 3.7.2025 0.09, Nicolas Pomarède wrote:
> > Updated code was pushed, Hatari now matches the result of your 
> > mscatest.tos test program, both for DTA structure and for bus error 
> > access.
>  >
> > Accesses causing address error will also prevent data from being 
> > cached, which was not the case with initial cache emulation
Hi

> Couple of comments on the commit:
>
> * None of the updated MegaSTE_Cache_*() functions are referred outside 
> of m68000.c, so they could all be static

I changed that


> * While all accesses seem to be byte or word sized, 
> MegaSTE_Cache_Addr_Cacheable() check could be for "Size > 2" instead of 
> "Size == 2", just in case...

Not needed, MegaSTE cache is only for 68000 and 68000 doesn't do long 
word memory accesses. As it's low level functions, I'd rather limit them 
to only the needed cases than adding catch-all codes where some years 
later you wonder why "is there a long word case here when it doesn't 
exist in HW" :)