Re: [hatari-devel] Reeking rumber analysis

[ Thread Index | Date Index | More Archives ]

OK, I think I know where the problem is now.

In fact, ioMemTabFalcon.c contains the following lines :

{ 0xff8240, 32 , IoMem_ReadWithoutInterception, VIDEL_ColorRegsWrite }, /* ST color regs */


{ 0xff9800, 0x400, IoMem_ReadWithoutInterception, VIDEL_ColorRegsWrite }, /* Falcon Videl palette */

Both lines are calling VIDEL_ColorRegsWrite for the write mode which is :

 * Monitor write access to ST/E color palette registers
void VIDEL_ColorRegsWrite(void)
    videl.hostColorsSync = false;

So, in the both calls, videl.hostColorsSync = false (for both St palette write and true color palette write) and the value of the writing is not saved into the register (or is it ?).

There's something to fix here.



Le 09/09/2013 00:28, Laurent Sallafranque a écrit :
George, I think you're intuition is good :

I've done the following :

> b $ffff9800!$ffff9800
> b $ff9800!$ff9800

and launched reeking rumber.
The program never breaks.

Eero, any idea why the registers are not updated ?


Mail converted by MHonArc 2.6.19+