[syslinux] [PATCH][git-pull] core/diskstart.inc

[Gene Cumm]

On Thu, Dec 23, 2010 at 17:00, H. Peter Anvin <hpa at zytor.com> wrote:
> On 12/21/2010 04:54 AM, Gene Cumm wrote:
>>
>> Feri: the push/pop around INT 13h I think did the trick.  My Dell
>> Latitude E6410 with an unpartitioned UFD (USB Flash Drive) works
>> perfectly.  I'm glad that BOCHS also behaved similarly as this would
>> have been much harder to nail down otherwise.  This may help out with
>> some of those Debian bugs.
>>
>> HPA: Could I request a code review and new -pre, at least for the
>> push/pop commit (d8cc533), as it fixes the above bug?
>>
>> Thinking about it, it makes complete sense.  An unpartitioned HDD is
>> essentially a huge superfloppy.  With the device number being an HDD,
>> use of the high two cylinder bits would be expected while some other
>> INT13h functions might act more like a floppy is present.
>>
>
> It isn't, actually, from a BIOS point of view (partitions are, or at
> least *SHOULD BE* irrelevant from the BIOS' standpoint), but it has more

Agreed.  It should be irrelevant.  This particular laptop was
definitely exhibiting behaviors that I normally don't see out of a
POST.

> to do with reusing common implementations, which is understandable and
> probably a good idea.

That's another idea and is probably the more correct idea.  Instead of
not returning anything in ES:DI for HDD, do the same thing regardless
of if it's a floppy or HDD device.

Thanks for the pull and -pre4.

-- 
-Gene