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

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 15/03/2024 à 11:30, Eero Tamminen a écrit :
> Hi,
>
> On 14.3.2024 23.51, Nicolas Pomarède wrote:
> > stupid bug spotted and fixed, please try the attached patch on current 
> > devel sources, it should use the correct MFP vector in case of interrupt
> > (iack_cycle() result was incorrectly saved in 'nr' instead of 
> > 'vector_nr', as in Exception_mmu030() )
>
> This fixes the EmuTOS 040 / 060 + MMU freeze on boot.
>
> When CPU caches & prefetch are disabled [1], it also gets Linux booting 
> both on 040 & 060, which hasn't worked for a long time!
>
> (I can understand how your patch fixes spurious interrupts for 040 case, 
> but not how the bug caused Linux to use non-implemented opcodes for 060 
> case...)

hi

maybe that due to not calling the correct exception handler (before the 
patch with vector_nr) it called an unrelated handler that has some 
opcodes in its code path. Now that correct handler is used, other path 
is not called anymore

>      - Eero
>
> [1] If CPU caches & prefetch are enabled, very early on boot Linux 
> double bus errors on 060, and freezes on 040. Same happens also with 
> v2.4.1 though, so it's not a regression.

could be useful to get traces about this, even if it's not a regression.
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 ?