Re: [hatari-devel] Solved: Pure debugger, illegal opcodes, exceptions

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


Hi,

On torstai 11 lokakuu 2012, Laurent Sallafranque wrote:
> I've just downloaded your patch, compiled and tested moongame.
> The mouse still not move at all and I still get the illegal access
> message in the console.
> 
> Do you still have the following lines when starting hatari (newcore) ?
> 
> Exception 2 (0) at e02ce2 -> e02ce6!
> A-Trap a000 at e010a0 (0x1b25c40)
> A-Trap a00e at e010c0 (0x1b25c60)
> A-Trap a000 at fa022c (0x1cc4dcc)
> Exception 2 (0) at e01854 -> e01836!
> Exception 2 (0) at e01854 -> e01836!
> Exception 2 (0) at 20346 -> 20356!
> A-Trap a000 at 20e02 (0xd459a2)

With TOS v4?  Yes.

Part of this is because WinUAE core is more verbose
than old UAE core.  That outputs the above kind of
debug only when Exception debugging is enabled.


For some reason -D doesn't catch exceptions with normal
TOS versions, but it does them with EmuTOS.  With that
I can compare the cores...


WinUAE core:

$ src/hatari --tos etos512k-0.8.7.img 
....
Exception 2 (0) at e00676 -> e0067a!
Illegal instruction: 7300 at 00E00D02 -> 00E00D0A
Exception 2 (0) at e00676 -> e0067a!
Exception 2 (0) at e00676 -> e0067a!
Exception 2 (0) at e00676 -> e0067a!
Exception 2 (0) at e00676 -> e0067a!
Exception 2 (0) at e00676 -> e0067a!
Exception 2 (0) at e00676 -> e0067a!
Illegal instruction: 42c0 at 00E00718 -> 00E0074A
Exception 2 (0) at e007d4 -> e007d6!


Old UAE core:

$ hatari -D --tos /home/emulation/atari/tos/etos512k-0.8.7.img 
Exception debugging enabled.
....
Bus Error at address $ff8400, PC=$e00676
....
Bus Error at address $ff8282, PC=$e00676
....
Bus Error at address $ff8e09, PC=$e00676
....
Bus Error at address $ff8961, PC=$e00676
....
Bus Error at address $ff8901, PC=$e00676
....
Bus Error at address $ff8a3c, PC=$e00676
....
Bus Error at address $fffa40, PC=$e007d4

The interesting thing is that the addresses
are slightly different between the cores.

I'm not sure whether this is due to when
they're printed, but I suspect they could be
due to exception frame stuff mentioned by
Thomas.



	- Eero



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