Subject: Re: A1230 clock and SCSI, hints
To: DE SMET Olivier <de_smet@lurpa13.ens-cachan.fr>
From: Geert Uytterhoeven <Geert.Uytterhoeven@CS.kuleuven.ac.be>
List: amiga
Date: 03/24/1995 14:11:10
e_smet@lurpa13.ens-cachan.fr (DE SMET Olivier) wrote:
> Some news about the clock for an A1230 GVP board. Finally I convert
                                                              ^^^^^^^
> the ASM to the MIT format by hand and make 2 quick C programmes to
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Gcc 2.6.3 for AmigaOS comes with mit2mot and mot2mit.

> use with. The main problem is the fact that the routine is VERY
> UNRELIABLE .. I got a bad date for 50% of the call.

	[...]

> But I think there are some trouble: with my previous 1230 board (a
> 40Mhz one without MMU) the reading of the clock was always right. But
> when I changed to an 50Mhz one (with MMU), now the reading of the
> clock is VERY unstable. I think the change of speed cause some
> 'drift'in the timing, because I always got the right first bits, but
                                             ^^^^^^^^^^^^^^^^^^^^
> the last one are false witch cause a bad date of plus or minus one
  ^^^^^^^^^^^^^^^^^^^^^^
> month.

Use the same `trick' as is recommended for reading the CIA timers:

  - read low byte
  - read high byte
  - read low byte again

If the two reads of the low byte indicate that a low byte overflow (with carry
to the high byte) occurred in between, repeat this procedure.

Greetings,

						Geert

+--------------------------------------------------------------------+
| Geert Uytterhoeven -->> Wavelets, Amiga, Linux/68k, MultiUser,...  |
| Geert.Uytterhoeven@cs.kuleuven.ac.be                               |
| Dept. of Computer Science, Katholieke Universiteit Leuven, Belgium |
+--------------------------------------------------------------------+
  Click!