Re: [hatari-devel] New drive image settings

[Thomas Huth <th.huth@xxxxxxxxx>]

Am Sun, 7 Oct 2018 10:26:56 +0200
schrieb Uwe Seimet <Uwe.Seimet@xxxxxxxxx>:

> Hi,
> 
> > I'm currently re-working the IDE code to get away from the
> > hard-coded "master/slave" config settings to a more generic
> > enumerated approach (IDE0 + IDE1). With that it should finally be
> > easier to support multiple IDE interfaces, too. But I don't have
> > real plans to implement this, there are other more important things
> > on my queue...  
> 
> Like emulating a Riebel ethernet card I assume ;-). By the way
> currently the naming pattern of the emulated drives differs. It says
> "IDE0 disk", but "EmulatedHarddisk" for ACSI. Shouldn't this better
> be ACSI0, in order to be more consistent?

Yes, maybe ... but does it really matter? It's just a string...
(actually I wonder whether the ID should be encoder there at all, since
you would not expect a change of the string just because it is
connected to a different port of the bus)

> > > 3. Thomas, some time ago you mentioned problems with
> > > the Falcon SCSI emulation. Were you able to resolve them?  
> > 
> > No, I was not able to find the issue, so HDDriver currently still
> > locks up in certain cases (when loading bigger blocks in one go) in
> > Falcon mode. Help would be very appreciated ;-)  
> 
> So checking issues related to the DMA address count (see my PM some
> time ago) being 1 byte off did not help, i.e. Hatari emulates this
> behavior correctly?

I've played with the settings there a little bit (you've mentioned 16
bytes, not 1 byte), but it did not help. So currently this is not
emulated yet, since I did not find the right setting yet. If you want
to have a try, have a look at the dma_check() function in ncr5380.c.

 Thomas