Hi.
>
> On 15 Apr 98 at 9:13, [log in to unmask] wrote:
>
> > >   Is there any chance you have a COM4 port in this machine?  Many S3
> > > chips include emulation of the IBM 8514/A adapter, and this conflicts
> > > with the "industry standard" COM4 address range.
> >
> >         I'm imagining a method to get rid of this inconvenience, which I
> > myself have to test: map COM4/IRQ3 into one of the on mobo com ports, and
> > your modem or whatever to COM2/IRQ5. Then probably the port addresses are
> > not aliased.
> >         To check:
> >         -With DEBUG test aliasing on COM4 on modem or card:
> >                 -i2EA. You will probably read 1.
> >                 -i6EA. You will probably read also 1, which shows aliasing.
> >         -Now do the change above, and repeat:
> >                 -i2EA. You will probably read 1.
> >                 -i6EA. If you get a different reading, aliasing is gone.
> >
> >         I'll appreciate any feedback.
>
>   I don't see where 6EA comes into it.  Maybe you're talking about
> something else.

        First PC designs only considered 10 bits for decoding ports
on the ISA bus, and this limit was on the mobo design, and perhaps
on the ISA cards. So reading or writing on port xx or xx+400H was
irrelevant. Above 6EAH=2EAH+400H, but this is only one of thousands
examples of the aliasing (any or many bits 11 up).

>
>   To quote from one of the many hits when I search the web for "8514
> COM4":
> > The 8514-compatible mode of the ... video chip uses I/O port
> > addresses from 2E0 to 2E8. This will conflict with systems using
> > COM4.
>   This particular quote comes from a Dell technical advisory, but
> there are also diagnostics to detect the conflict, products which
> avoid it, and FAQ answers that identify it as a cause of strange
> behaviour.  [These do not, however, always correctly explain the
> details of the problem.  One I looked at seemed to imply that a
> less-featured video driver could fix the problem, and that's simply
> not true.]
>
> David G
>

        I have seen very few explanations of what happens, and indeed
I have never experienced the problem. Only I have seen that S3 chips
use a port at something like 22E8H (not sure of exact number, but
surely something aliased on the 2E8H-2EFH). To my observation, on modern
mobos, its com port addresses are not aliased. It is very easy to
design on mobo decoding so as to avoid aliasing. So test above method
and tell me.

************************************
Javier Vizcaino. Ability Electronics. [log in to unmask]
  Starting point:        (-1)^(-1) = -1
  Applying logarithms: (-1)*ln(-1) = ln(-1)
  Since ln(-1) <> 0, dividing:  -1 = 1     (ln(-1) is complex, but exists)