Re: [hatari-devel] MEMWATCH freezes Hatari

[ Thread Index | Date Index | More lists.tuxfamily.org/hatari-devel Archives ]


Hi,

> Well, at least this is a problem solved :)
> I don't remember if I downloaded the version from your website, but 
> maybe you could check it's the correct one too just in case.

Yes, that one is correct.

> cpu emulation completety handles changes of VBR, several programs are 
> using it, so I'm not sure it's related.
> This and the stack reading you do to call th eoriginal vector is not 
> that unusual, you would not imagine all the tricks that were used in 
> game's protection to obfuscate the code, much worse than that and 
> they're supported ;)

:-)

The VBR change must have worked fine, indeed, otherwise the MEMWATCH
exception vector (the one the VRB points to) would not have even been
called. But using the 68030 exception vector offset in order to call the
original vector ($04, $08, ...) is what appears to be not working.
Instead it seems to call the replaced (VBR) vector again, as if the
emulation did not realize that THIS TIME the value of $04, for instance,
and not the one from the VBR+4 must be called. In other words, it's as
if the emulation always uses VBR+4 whenever $04 is used as address. But
VBR+4 must only be used if there is actually an exception/trap, but not
whenever the content of address $04 is loaded.

> Could you upload the whole log file too so I could have a look ; 
> sometimes the context is altered several instructions before the 
> faulting one.

Uploaded again to https://www.hddriver.net/downloads/hatari.log.gz

Best regards

Uwe



Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/