🚀 go-pugleaf

In article 3484@doortje.and.nl, berto@and.nl (Bert van Oortmarssen) writes:
>I have a piece of code that reads one bitplane of a VGA-screen from disk,
>( fread(buf, 1, VGASIZE, fp)  where VGASIZE is defines as 38400u). If 
>smartdrive is disabled, the code works fine, but when smartdrv is enabled
>the contents of buf will be corrupted (when loaded a second time).
>
>Is smartdrv wrong ?? Anyone who can give me any information on this ??
>
>
>Bert van Oortmarssen
>berto@and.nl


This sounds like the kind of thing that happens when you are using a disk
controller (like an Adaptec 1542, maybe?) that does bus mastering and deals
only in physical (real-mode) addresses.  Smartdrv 4.0 deals only with virtual
addresses, as does the ASPI4DOS driver, by the way.

If your controller gets passed a virtual address when it is expecting a physical
address your data winds up somewhere in West Hyperspace, or comes from there.
Might that be your problem?

There is a paragraph or two in your Windows 3.1 User's Guide that explains the
installation of Smartdrv 4.0 with the Double Buffering option, and why you may
need it.  Double Buffering in this case means it has one set of buffers for
virtual addresses, and one for physical addresses, and it takes care of the
translation in both directions, as I understand it.

Also, if you really are using the Adaptec stuff, and you put your drives under
control of the ASPI4DOS manager, you can skip the Double Buffering option in
Smartdrv.

-Bob