[hatari-devel] 68030 Debug mode and new CPU

[Laurent Sallafranque]

It seems to work for me, no regression detected.

Laurent

Le 20/02/2011 23:07, Eero Tamminen a écrit :
> Hi,
>
> On sunnuntai 20 helmikuu 2011, Laurent Sallafranque wrote:
>> I don't understand everything here (sorry).
>> If you think this patch is needed, please, fell free to upload it and
>> I'll test it.
> I commited it with largish explanation, but didn't test it more than
> to see that it compiles.  Now WinAUE disassembly should work also
> if you redirect it to file.
>
> To test it I guess you first need set that you want to use the WinAUE
> disassembly instead of Markus' code in in debug/68kDisass.h, currently it's
> pointing to Markus version:
> #define        DISASM_ENGINE_UAE       0               /* Use UAE's internal
> disassembler */
> #define        DISASM_ENGINE_EXT       1               /* Use external
> disassembler from 68kdisass.c */
>
>
> Btw. While testing the compilation, I noticed this in cpu/cpummu.h:
> 	#define likely(x) x
> 	#define unlikely(x) x
>
> Why the optimization macros from main.h (intended to mark code branches
> as cold/hot) are nullified?
>
>
> 	- Eero
>
>> Best regards
>>
>> Laurent
>>
>> Le 17/02/2011 22:25, Eero Tamminen a écrit :
>>> Hi,
>>>
>>> On keskiviikko 16 helmikuu 2011, Laurent Sallafranque wrote:
>>>>> You could add:
>>>>> 	#define f_out fprintf
>>>>>
>>>>> to one of the headers, remove the f_out stuff from cpu/custom.c and
>>>>> replace all "void *" args for functions calling f_out() with "FILE
>>>>> *".
>>>> Done, it works well now.
>>>> And it seems to have removed some compiling warnings.
>>> Actually, as f_out() seems to be used only in newcpu.c, the define
>>> could be there too.  Unless you want all this kind of stuff to be in
>>> compat.h so that it's easier to find.
>>>
>>>
>>> While looking into newcpu.c more, I noticed that the "buf"&   "bufsize"
>>> arguments to function:
>>> void m68k_disasm_2 (TCHAR *buf, int bufsize, uaecptr addr, uaecptr
>>> *nextpc, int cnt, uae_u32 *seaddr, uae_u32 *deaddr, int safemode)
>>>
>>> Are redundant with your changes (instead of using buf_out(), you're
>>> calling printf() directly and not buffering the string for later
>>> printing).
>>>
>>> Attached (untested) patch would clean that.
>>>
>>> 	- Eero
>>>
>>> _______________________________________________
>>> hatari-devel mailing list
>>> hatari-devel at lists.berlios.de
>>> https://lists.berlios.de/mailman/listinfo/hatari-devel
> _______________________________________________
> hatari-devel mailing list
> hatari-devel at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/hatari-devel