Re: New pre-2.0.31 patches..

David Parsons (orc@pell.chi.il.us)
Wed, 6 Aug 1997 08:08:07 -0700 (PDT)


Rob Hagopian wrote:
[On enhanced memory detection]
>
> How is that possible? I thought that the code only looked for more than
> 64M if the normal routines reported 64M... Thus, even if the probe for
> >64M of RAM failed you'd still have your 64M and you'd have to use an
> append statement or such... If this isn't how it works, maybe it should?

No, that's Larry's memory detection code. My patch calls
int 15h, ax=0e801h, grabs that result if the call succeeded,
then calls int 15h, ah=88h, grabs that result, and uses the
largest of the two values as the memory size. You can't
just check the return from ah=88h against 64mb, because
some bioses take a conservative approach and only return
a maximum of 16mb (Compaq comes immediately to mind here)

____
david parsons \bi/ hardwiring memory sizes is icky
\/