SEABIOS and DOS compatibility

[Gerhard Wiesinger]

Hello,

I bisected down 2 DOS 6.22 compatibility issues:

1.) COMMAND.COM can not be loaded high, 
9c98517c938d20c38f537d516c71b30bb60c3ea0 is the first bad commit
Looks like and UMB generated which is not recognizeable ...
Before: 9E80-9F7F: 4k EBDA, 9F80-9FFF: 2k UNUSED, really unused or used 
for option ROMs?
After: 9F00-9FFF: 4k EBDA, now the 2k area is somewhere above and will 
be used by memory managers???

2.) 59d6ca52a7eba5b1f4f2becf70fd446dccaf0a2e is the first bad commit
Opening a file and closing a file raises a 386 exception

I'm using LSI 53C895A adapter with LSI BIOS and disabled LSI BIOS .

Any ideas?

Release are in the state:
git checkout rel-1.7.1 => NOT OK
git checkout rel-1.7.0 => OK

Thank you.

Ciao,
Gerhard

[Kevin O'Connor]

On Sun, Nov 11, 2012 at 06:44:53PM +0100, Gerhard Wiesinger wrote:
> Hello,
>
> I bisected down 2 DOS 6.22 compatibility issues:
>
> 1.) COMMAND.COM can not be loaded high,
> 9c98517c938d20c38f537d516c71b30bb60c3ea0 is the first bad commit
> Looks like and UMB generated which is not recognizeable ...
> Before: 9E80-9F7F: 4k EBDA, 9F80-9FFF: 2k UNUSED, really unused or
> used for option ROMs?
> After: 9F00-9FFF: 4k EBDA, now the 2k area is somewhere above and
> will be used by memory managers???

Thanks.  Can you post your qemu command line and the SeaBIOS output
(-chardev stdio,id=seabios -device
isa-debugcon,iobase=0x402,chardev=seabios) from both a working run and
the non-working run?  Also, can you provide the exact text reported by
msdos on success and failure.
> 2.) 59d6ca52a7eba5b1f4f2becf70fd446dccaf0a2e is the first bad commit
> Opening a file and closing a file raises a 386 exception

That's odd - nothing in this commit impacts the disk code.  Maybe a
slightly different memory layout is triggering some other bug.

I have an MSDOS 6.22 floppy image - but I don't see either of these
issues.  Can you provide the steps needed to reproduce them?

-Kevin

[Fred .]

Anyone wishing to test this, can get a bootdisk from bootdisk.com.

http://bootdisk.com/bootdisk.htm

On Sun, Nov 11, 2012 at 8:32 PM, Kevin O'Connor  wrote:
> On Sun, Nov 11, 2012 at 06:44:53PM +0100, Gerhard Wiesinger wrote:
> > Hello,
> >
> > I bisected down 2 DOS 6.22 compatibility issues:
> >
> > 1.) COMMAND.COM can not be loaded high,
> > 9c98517c938d20c38f537d516c71b30bb60c3ea0 is the first bad commit
> > Looks like and UMB generated which is not recognizeable ...
> > Before: 9E80-9F7F: 4k EBDA, 9F80-9FFF: 2k UNUSED, really unused or
> > used for option ROMs?
> > After: 9F00-9FFF: 4k EBDA, now the 2k area is somewhere above and
> > will be used by memory managers???
>
> Thanks.  Can you post your qemu command line and the SeaBIOS output
> (-chardev stdio,id=seabios -device
> isa-debugcon,iobase=0x402,chardev=seabios) from both a working run and
> the non-working run?  Also, can you provide the exact text reported by
> msdos on success and failure.
> > 2.) 59d6ca52a7eba5b1f4f2becf70fd446dccaf0a2e is the first bad commit
> > Opening a file and closing a file raises a 386 exception
>
> That's odd - nothing in this commit impacts the disk code.  Maybe a
> slightly different memory layout is triggering some other bug.
>
> I have an MSDOS 6.22 floppy image - but I don't see either of these
> issues.  Can you provide the steps needed to reproduce them?
>
> -Kevin
>
> SeaBIOS mailing list
> [e-mail]
> http://www.seabios.org/mailman/listinfo/seabios