[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: A multiport serial board would be nice...
dempson@actrix.gen.nz (David Empson) writes:
>In article <365lak$2f6@zircon.cs.uow.edu.au>,
>David E A Wilson <david@zircon.cs.uow.edu.au> wrote:
>> It occurred to me a few hours after posting this that I could write
>> the first character device SmartPort implementation. This would allow
>> all N ports to be used and would make the card identifiable in a slot
>> scan.
>Wow! Unfortunately, no existing software would recognise it. I
>suppose ProTerm can have a custom driver written. How about BBS
>software drivers? For that matter, is there any multi-user BBS
>software for the Apple II?
The mail I have received so far indicates that people would like to
be able to use existing telecom programs (at least on the first port).
Unfortunately this means that I would be unable to use any of the newer serial
chips (I am getting info on the Signetics SC26C94 Quad UART and SCC2698B Octal
UART chips) as I believe that all comms programs talk directly to the hardware
rather than just using the Pascal 1.1 interface (as it has no support for
interrupts [or at least no cards have implemented it] - correct?).
Even worse would be if I go with the suggestion that the card have an onboard
CPU to lessen the interrupt load on the //. There is no way that this interface
would look like any serial chip known to man...
I suppose I could add a 65c51 chip as port 0 for backwards compatibility.
It does not look as simple as it did a day ago...
>One other point to consider: the SmartPort and Pascal 1.1 firmware ID
>bytes conflict, so you couldn't have the card support both SmartPort
>and Pascal (unless you did two ROMs).
This is indeed a problem (aside: my Cirtech InSyder SCSI card has an optional
printer box that plugs into the SCSI bus and provides a simple Centronics port,
the Basic entry points are provided but Pascal & CP/M cannot use it as it has
the wrong ID bytes - GS/OS gets a loaded driver for printing). The Pascal 1.1
protocol does not provide any guaranteed way of me choosing an ID and having
no one else choose the same one. This could make automatic identification of
the card difficult.
>I don't know how well GS/OS supports generated drivers for SmartPort
>character devices. The device driver reference implies that they are
>supported.
I guess until someone writes one it will be difficult to test this implication.
Imagine trying to get Apple to fix a possible bug in it!
>I could help with writing a loaded GS/OS drive for any card (also GNO
>drivers).
Assuming this project gets off the ground that will be very useful.