Re: [hatari-devel] IKBD regression with EmuTOS

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

On 10/02/2013 19:53, George Nakos wrote:
> I haven't taken a look at the EmuTOS sources in years, so I'm not sure
> how  it's  handling the IKBD reset, but I do know that if you send any
> command  during  the  IKBD  reset (i.e. within 300ms after issuing the
> reset  command)  it  gets ignored. Maybe there's a deliberate pause in
> EmuTOS?
>
> Proof:  http://pouet.net/prod.php?which=60836,  which  I  debugged for
> !Cube.  He  was  doing  what  I  described  above, so his "kill mouse"
> command  was  being  ignored  by  real hardware, while it worked ok on
> STEem (and possibly hatari).

Hi,

The problem is more complex, most commands will be ignored during a 
reset, but some specific commands ($12, $1a, $8, $14) can put the ikbd 
in a different state if you send them during the reset before receiving 
the $f1 byte.

I tried to document this by looking at the ikbd's code, but it's not 
that clear when studying the HD6301 disassembly code and I haven't fully 
found all the cases.

Nicolas