Re: [hatari-devel] Sokoban VGA (Falcon)

[Laurent Sallafranque <laurent.sallafranque@xxxxxxx>]

Hi Thomas,

> Is that game freely available for download somewhere?

Yes, you can get it from 
http://christophe.bray.free.fr/informatique/falcon/falcon_jeux.htm

It's the entry 80



> Could you do a "info blitter" in the debugger when you reach the TAS
instruction?


I've put a breakpoint at address   pc = $e12bcc, then, when the code 
reach this address, I've done a disasm to see the next few instructions,
then info blitter
then C1 (and I get the exception error again.

It seems reproductible.


CPU=$e0258e, VBL=3166, FrameCycles=68, HBL=0, LineCycles=68, DSP=$51
> b pc=$e12bcc
CPU condition breakpoint 1 with 1 condition(s) added:
    pc = $e12bcc

> c
Returning to emulation...

1. CPU breakpoint condition(s) matched 1 times.
    pc = $e12bcc

CPU=$e12bcc, VBL=3274, FrameCycles=2064, HBL=5, LineCycles=48, DSP=$a7

> d

$e12bcc : 4ad5                                 tas       (a5)
$e12bce : 4e71                                 nop
$e12bd0 : 6bfa                                 bmi.s     $e12bcc
$e12bd2 : 2c0c                                 move.l    a4,d6
$e12bd4 : 671c                                 beq.s     $e12bf2
$e12bd6 : 4bed ffc4                            lea $ffc4(a5),a5
$e12bda : 2adc                                 move.l (a4)+,(a5)+
$e12bdc : 2adc                                 move.l (a4)+,(a5)+

> info blitter

src addr:  0x2de9fe
dst addr:  0xff9800
words:     33
lines:     1
src X-inc: -2
src Y-inc: -962
dst X-inc: -2
dst Y-inc: -962
end mask1: 0xffff
end mask2: 0xffff
end mask3: 0xffff
HOP:       0x02
LOP:       0x03
control:   0x80
skew:      0xc0
Note: internally changed register values aren't visible to breakpoints 
or in memdump output until emulated code reads or writes them!)

> c 1
Returning to emulation for 1 CPU instructions...
Exception 2 (e12bcc) at e12bce -> e00fb6!

CPU=$e00fb6, VBL=3274, FrameCycles=2232, HBL=5, LineCycles=216, DSP=$ad
>



Regards

Laurent




Le 07/10/2013 22:58, Thomas Huth a écrit :
> Am Mon, 07 Oct 2013 00:35:32 +0200
> schrieb Laurent Sallafranque <laurent.sallafranque@xxxxxxx>:
>
> > Hi,
> >
> > I've given a quick look at Sokoban (Falcon), and I get the following
> > error:
> Is that game freely available for download somewhere?
>
> > Exception 2 (e12bcc) at e12bce -> e00fb6!
> That's an bus error... can you determine where it comes from (e.g. by
> putting a breakpoint with gdb and do a backtrace when it triggers?).
>
> > The executed code before the error is a linea #7 (Bit block transfer)
> > And the program stops on the TAS instruction.
> >
> > I suspect the TAS instruction not to return to the correct PC
> > address, but I haven't digged more for now (maybe there's a patch of
> > this instruction in winuae, or in the old cpu).
> I hardly believe that TAS itself is wrong (although the code for the
> cycle-exact 68000 mode seems to be dodgy, but that does not matter here
> since you run in 68030 mode instead).
> The code you've added below seems to try to start the blitter. So I
> think the problem could be related to the blitter instead.
> Could you do a "info blitter" in the debugger when you reach the TAS
> instruction?
>
> > The last instructions before the crash
> >
> >
> > $06aac6 : a007                                 linea
> > #7                        ; Bit block transfer
> [...]
> > $e12bcc : 4ad5                                 tas       (a5)
> > Debugger: *CPU breakpoint*
> >   >
> >   > r
> >     D0 0000007E   D1 00000004   D2 00000000   D3 FFFF0000
> >     D4 00000001   D5 00000000   D6 00000000   D7 000080C0
> >     A0 002DEA41   A1 00FF9841   A2 00000002   A3 00000002
> >     A4 00000000   A5 00FF8A3C   A6 00083E02   A7 00087768
> A5 = 0xff8a3c is the blitter control register ... TAS is obviously used
> to test whether it is already running and to start it if it was not
> running yet.
>
>   Thomas