Re: [hatari-devel] Re: IDE byte swapping options (was: Problems with GEMDOS drive support)

[Thomas Huth <th.huth@xxxxxxxxx>]

Am Sun, 23 Sep 2018 20:17:56 +0300
schrieb Eero Tamminen <oak@xxxxxxxxxxxxxx>:

> Hi,
> 
> On 9/21/18 7:17 PM, Roger Burrows wrote:
> > On 21 Sep 2018 at 10:33, Vincent Rivière wrote:  
> >> Hatari should have an option to disable its automatic byteswapping
> >> for emulated IDE drives, then everything will be fine.
> >>  
> > There still has to be a fix in IDE_Init()/HDC_Partition_Count(),
> > because Hatari now reads the disk in two different ways: when
> > passing the data to the user, it byteswaps, but when analyzing the
> > boot sector for its own purposes, it doesn't. It needs to do the
> > same thing for both.  An extra option not to byteswap might be
> > nice, but it would have to apply to both situations (and it runs
> > the risk of being confusing to users).  
> 
> I looked at Aranym code quickly, and it would seem to do 16-bit
> swaps for the whole buffer.
> 
> So, is that what "byteswap" means?

Yes.

> If yes, attached patch "should" implement such option for IDE
> partition counting.  If it works fine, same options can be used
> to toggle swapping for rest of IDE handling too.

Thanks for the patch, that's a first step... but actually, I think this
area (especially the HD dialog in the GUI) needs a major clean-up, e.g.
I also need to add SCSI setup to the GUI, add support for multiple ACSI
images, fix the nomenclature for IDE ("master" and "slave" is
nonstandard, the official terms are IDE 0 and IDE 1), etc.

Using "--byteswap master" to only enable byteswapping is also a bad
idea - what if a user already enabled byteswapping in the config file
and wants to disable it on the CLI again?

So I'd appreciate if you could hold back your patch, I'll take care of
this, I just need to find some spare time to work on this...

 Thomas