Subj : MCP?
To : Bob Jones
From : Mike Luther
Date : Mon Aug 20 2007 05:50 am
Bob ..
BJ> I'm catching up on a few items..... temporarily....
BJ> Are you running SIO for the com ports and/or vmodem
BJ> definitions? If so, is this SIO version 1.60d? If
BJ> so, is it a registered version that is registered for
BJ> 8 or more nodes? If not registered, it will only
BJ> allow up to for "ports" to be used, if I remember
BJ> correctly.....
I've never thought about this and haven't researched it in the manual, just
assuming your thought line is correct. Does the SIO2K version have the same
'limitation' of no more than four 'ports' and/or vmodem definitions?
I have both SIO 1.60d as well as SIO2K registered. A number of systems around
me use all four serial ports with four port accessory plugin serial port cards
.. some old style buss type and some PCI. At least in my case each comm port
for those hardware cards is on a separate IRQ. In my case no more than two
instances of MAX are running against telephone lines at one time, but the other
two hard comm ports are being used for other telecommunications purposes. That
varies from serial port use for Rochelle phone line data monitoring for up to
four phone lines simulteously, to my own programs which use serial port
connections for radio transmitter send/receive purposes for telegraph or other
digital data, plus antenna and other discrete station site hardware control
devices. Which in the case of Rose KVM switched sites might even include still
serial mouse port activity as one of the hard comm port uses.
On top of that is the SIO Vmodem issue which can 'expand' the use of 'comm'
ports to what I guess is four ports. In this case the IRQ that is 'required'
in the SIO control parameters is 'listed' for the control setup for the system,
but is an 'upper' IRQ number and if I recall this correctly is really not
'used' as such for Vmodem? Right? That's how we get more than four 'serial'
ports into use for the four hardware comm ports and Vmodem in SIO/SIO2K, no?
And SIO/SIO2K, in theory, doesn't 'care' about that VModem IRQ number, as it is
not one of the sixteen IRQ's that are 'reserved' for such things?
Which can be confused in some cases because PCI slot hardware can share IRQ's
.. provided that the device driver for that PCI slot hardware is written
properly to share IRQ's. Which sometimes it doesn't no matter what .. in the
case of audio and NIC card isses and OS/2, specifically. sigh.
Now I don't have MAX attempting to run on eight different users at one time
like is in question here. But the Integrand Research pig iron relay rack OS/2
system which is running this BBS is using all four serial ports at one time on
SIO2K registered. Two of the phone lines are in fact MAX BBS systems still on
phone line connections, while the other serial ports are in use for other
things. Which obviously, SIO2K is servicing.
But in this case another MAX BBS is, in fact, using Vmodem and TelNet for a
third MAX operation while all this is going on. And .. that does work at the
same time as the other four comm port lines are in use. The other two serial
ports in the box on SIO2K are being used for the mouse and for an RF gateway to
an AEA packet operation between the MAX BBS operation and VHF Packet radion
work to gate Fido to VHF Packet operations for emergency data service if
needed. Thus providing a Vmodem Telnet operation which can gateway Fido to
either land line Fido or other service as well as VHF/HF radio interfaces for
emergency uses.
Which works.
That would be, in this case, five comm port uses, including the Vmodem Telent
instance. But .. that's not five MAX BBS users. Say four on modem connections
and four more on Vmodem Telnet connections, like I think is the question here?
Part of the reason I'm participating here is that there is another quirk I have
noticed in all this over the years. I also use HyperAccess for OS/2 for remote
desktop operations under OS/2 between sites over phone line modems using one of
the phone lines at some sites as well. And curiously, while the remote site
can use SIO2K for the HHost part of the connection, the local site I am using
to connect to the SIO2K remote HHost operation cannot use SIO2K. It has to be
on the older SIO 1.6d version as well. Or I can never complete local
HyperAccess Pro desktop connection to the SIO2K remote HHost site connection.
Which opens up the next issue that may be at work here for the poster with the
questions! You must use a DIFFERENT serial number SIO installation for each
interconnection via Vmodem and Telnet between sites! That means, if you were
using four different serial ports on hard phone lines with SIO or SIO2K, that
part would work fine. But any user with OS/2 and SIO or SIO2K which attempts
to connect with a Vmodem and TelNet connection back to this same box, has to
have an install version of SIO or SIO2K with a different serial number for that
connection, if my memory is correct! Same with HyperAccess Pro and HHost. In
this case when you register that, it comes with two different install diskette
files, so you can connect between two sites that way.
But .. in the case of SIO or SIO2K, is the reason this is failing is that the
test boxes are also running OS/2 for the four Vmodem and Telnet connections and
all from the same registered version of them?
In my case I registered more than one version of SIO and SIO2K. They are
different for each install remote site which I know is expected to connect to
my systems .. site to site. And I know from past experience that without that
I cannot successfully get Vmodem and Telnet service between sites.
So might this be what is wrong?
--> Sleep well; OS/2's still awake! ;)
Mike @ 1:117/3001
--- Maximus/2 3.01
* Origin: Ziplog Public Port (1:117/3001)
