Re: [hatari-devel] Hatari 2.0.0 has been released

[Thomas Huth <th.huth@xxxxxx>]

Am Sun, 13 Nov 2016 16:19:42 +0100
schrieb Thorsten Otto <admin@xxxxxxxxxxx>:

> On Sunday 13 November 2016 15:57:33 Christian Zietz wrote:
> >  It seems that TOS
> > also takes into account the memory contents at $4ba (number of 200
> > Hz interrupts) to decide whether it is a cold boot and it should
> > run a memory test:  
> 
> Yes, thats true:
> 
> /* 306de: 00e009a2 */
> /* 206de: 00e0097a */
> dmaboot:
>         move.l    #80*200,d7
>         cmp.l     _hz_200,d7                /* is the system running
> for > 80s? */
>         bcs       dmadone                   /* (no memory test) */
>         movea.w   #$0000,a5                 /* done status = false */
> dmaloop:
>         cmpa.w    #$0000,a5                 /* memtest done? */
>         bne       skipsp                    /* (yes) */
>         bsr       memtest

Very good hint, thanks! I've added a corresponding piece of code to
Hatari to patch $4ba and now the memory test is skipped with TOS 3.06
and 4.04, too!

> > I wonder why fast booting TOS 3.06 worked at all in some
> >previous version, as you mention.  
> 
> I'm quite sure it worked. All you have to do is, beside setting the
> memvalid flags, to set _hz_200 to a value >= 16000.

I doubt that this was ever working, we never had code in Hatari that
patched the _hz200 variable. I've also checked with Hatari v1.8 and I
can also see the memory test there. Maybe you confused this with TOS
2.06 ?

 Thomas