Re: [hatari-devel] MEMWATCH freezes Hatari

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 22/10/2018 à 22:34, Uwe Seimet a écrit :
> With these changes and CE and PF I still get the segfault in the inlined
> code:
>
> #0  0x00005555578a4f00 in regs ()
> #1  0x0000000000000003 in ?? ()
> #2  0x0000555555ec641f in get_iword_mmu030c_state (o=4)
>      at /home/us/hatari/hatari/src/cpu/cpummu030.h:818
> #3  op_41e8_35_ff (opcode=<optimized out>)
>      at /home/us/hatari/hatari/src/cpu/cpuemu_35.c:16676
> #4  0x000055555593852b in m68k_run_mmu030 ()
>      at /home/us/hatari/hatari/src/cpu/newcpu.c:6623
> #5  0x000055555593349a in m68k_go (may_quit=<optimized out>)
>      at /home/us/hatari/hatari/src/cpu/newcpu.c:7719
> #6  0x00005555558701ab in main (argc=1, argv=<optimized out>)
>      at /home/us/hatari/hatari/src/main.c:949
>
> Just before the segfault 11 bombs are displayed.

op_41e8_35 is "LEA.L (d16,An),An" which needs 1 extra word for the d16 
offset, but nothing really specific, apart from reading the word from 
the i-cache.

In cpu/newcpu.c, can you uncomment line 11103 with 
WINUAE_FOR_HATARI_DEBUG_CACHE to enable some cache consistency tests ?
It should not prevent the crash but it might report some cache issues 
before the crash.

Nicolas