On 11/10/2024 4:42 PM, Paul Dufresne via Freedos-devel wrote:
https://github.com/86Box/86Box/issues/4818#issuecomment-2350705648
For me, it is kind of a revelation that some BIOS don't implements int 13h.
That was after listening to: https://www.youtube.com/watch?v=4GCvF-chgn0
I frankly don't think it is worth to do direct port and DMA programming of the
disk controller...
Maybe we could detect if int 13h exist... frankly not sure how, and give a
clearer error message sooner.
Also, I think we might want to make the test version visible on:
https://freedos.org/download/ page... although, it would proably not helped
much in this case.
As a general rule I would say: Yes, FreeDOS assumes the presents of
INT13h. It is one of the very basic "staples" of an IBM compatible PC. A
BIOS that doesn't support INT 13h at all just isn't...
What however needs to be treated with care is to use extensions beyond
what was published in the various IBM Technical Reference manuals. And
THAT is what the 86Box maintainer was eluding to. Assumption of a
(random) extension, not that INT13h didn't exist at all.
Ralf
PS: Just to be clear, there was MS-DOS available on non-IBM compatible
PCs, which did not have a BIOS with INT13h (or INT10h for that matter),
like the Sirius 1/Victor 9000, DEC Rainbow or HP-150, just to name a
couple of systems I used to work with in the now distant past. But they
required custom version of MS-DOS, which replaced all those BIOS calls
with the appropriate function calls as they existed (or not) in those
non-IBM systems. Which over time became a major pain in the posterior
and contributed to the rather quick extinction of non-IBM PC compatible
clones...
_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel
