> On Mon, 24 Nov 1997, Peter T. Waltenberg wrote:
>
> > I suspect the existing code is correct.
> >
> > On most systems, you can't actually share com's IRQ's in the sense that they
> > can't be used by both ports at the same time. You can use one or the other,
> > but not both.
> >
> > The com's ports have logic to tri-state their IRQ's so that one can safely sit
> > unused while you use another. But trying to drive one port's interupts while the
> > other is also enabled (but not interupting) results in a bus contention.
> >
> > With the correct hardware, or a few diodes and a resistor, you can make it work
> > properly, but the "standard" PC hardware won't work reliably.
> >
> I won't recommend using two ports on the same irq on standard hardware.
> I've tried it on my system (mouse on /dev/ttyS0, irq 4 and modem on
> /dev/cua2, irq 4) and it slows both ports down to crawling speed. This
> means that if you have two serial ports plus an internal modem, one port
> will be almost unusable if the modem shares an irq with one of the other
> ports. I made the patch to make it possible to set up serial ports 3 and
> 4 with different irqs in a clean way to avoid this problem.
Are you sure of this? On my system I have COM1 as mouse, COM2
occasionally used for InterLNK/Direct CC, and an internal modem on COM4
with no sign of a slowdown.
I have heard of the slowdown occuring if an external modem is used,
however.
>
> --
> Med venlig hilsen / Regards
>
> Niels Kristian Bech Jensen nkbj@image.dk
> http://www.image.dk/~nkbj/
>
>
----------------------------------------------------------------------
Phil Brutsche
"Be of stout heart, Number One. We've handled the Borg. We can
certainly handle Admiral Jellico." - Jean-Luc Picard
----------------------------------------------------------------------