Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!harpo!seismo!hao!hplabs!sri-unix!GILLMANN@USC-ISIB
From: GILLMANN@USC-ISIB@sri-unix.UUCP
Newsgroups: net.micro.pc
Subject: Re: Quadram Serial Ports Problem
Message-ID: <4724@sri-arpa.UUCP>
Date: Tue, 30-Aug-83 19:57:00 EDT
Article-I.D.: sri-arpa.4724
Posted: Tue Aug 30 19:57:00 1983
Date-Received: Sat, 3-Sep-83 08:05:02 EDT
Lines: 22

From:  Dick Gillmann <GILLMANN@USC-ISIB>

The problem is that the two ports must be on different interrupt
levels.  When the PC was released there wasn't a definition of COM2:
included.  AST took the route of leaving jumpers that could put COM2:
on any of several interrupt levels.  Quadram (at least my (early)
board) puts all serial ports on level 4, which is COM1:.  The latest
serial ports from IBM can be switched from COM1: to COM2: by flipping
a little jumper plug.  In COM2: mode they are on level 3 (level 5 is
used by the XT's hard disc).  So I suppose this is official: COM1: is
on IRQ4 and COM2: is on IRQ3.

If you have two ports on the same level, they fight, i.e. one tries to
raise an interrupt while the other tries not to.  This leads to some
strange behavior.  Generally you'll get one character through and then
hang.

This problem only happens with interrupt driven I/O on the serial
ports.  The BIOS, which does busy waits, is not affected.

/Dick Gillmann
-------